Sunteți pe pagina 1din 140

Journal of the International Association of

Buddhist Studies
Volume 19 Number 2 Winter 1996
On MaJ}.galas
. The Mm:t<;lala at Ellora / Ellora in the Mm:t<;lala
Mandalas on the Move: Reflections from Chinese
Esoteric Buddhism Circa 800 C. E.
Mandala, Mandala on the Wall:
Varhitions of Usage in the Shingon School
Reevaluating the Eighth-Ninth Century Pala Milieu:
leona-Conservatism and the Persistence of Sakyamuni
The Moves Mal,l<;lalas Make
e watermark
The Journal o/the International
Association of Buddhist Studies
(ISSN 0193-600XX) organ of
the International AssocIaoon of
Buddhist Studies, Inc. It welcomes
scholarly contributions pertaining to
all facets of Buddhist Studies.
JIABS is published twice yearly, in
the summer and winter.
Address manuscripts (two copies)
and books for review to The
Editors, JIABS, Section de
langues et civilisations orientales,
Universite de Lausanne, BFSH 2,
CH-1015 Lausanne, Switzerland.
Address subscription orders and
dues, changes of address, and
business correspondence
(including advertising orders) to
Professor Joe B. Wilson, Treasurer
lABS, Department of Philosophy
and Religion, University of North
Carolina at Wilmington,
Wilmington NC 28403 USA
Fax: 910-962-7107
Subscriptions to JIABS are $40
per year for individuals and $70
per year for libraries and other
institutions. For information on
in lABS, see back
1996 by the International
Association of Buddhist Studies,
Inc. The editor gratefully
acknowledges the support of the
Center for Chinese Studies and the
Department of Asian Languages
and Cultures of the University of
Donald S. Lopez, Jr.
Robert Buswell
Steven Collins
Collett Cox
Luis O. Gomez
Oskar von Hinuber
Roger Jackson
. Padmanabh S. J aini
Shoryu Katsura
Alexander Macdonald
D. Seyfort Ruegg
Ernst Steinkellner
Erik Ztircher
Editorial Assistant
Alexander Vesey
The term of the current Editor-in-chief of the Journal of the International
Association of Buddhist Studies will end on July 1, 1997. He is pleased to
announce that at that time, two distinguished scholars will together assume
the editorship of the journal. They are Professor Tom J. F. Tillemans and
Professor Cristina A. Scherrer-Schaub of the University of Lausanne.
Beginning immediately, all submissions to the journal should be sent to:
The Editors
Journal of the International Association of Buddhist Studies
Section de langues et civilisations orientales
Universite de Lausanne, BFSH 2
CH-1015 Lausanne, Switzerland
Fax: (41) 21 692 30 45
The Twelfth Congress of the International Association of Buddhist
will be held in 1999 in Lausanne, Switzerland from August 23rd until
August 28th.
For information please contact:
Professor Tom I.F. Tillemans
Section de langues et civilisations orientales
Universite de Lausanne, BFSH 2
CH-1015 Lausanne, Switzerland
Fax: (41) 21 6923045
Contributors to this issue:
DAVID L. GARDINER received his Ph. D. in 1995 from the Religious
Studies Department of Stanford University. He now teaches at Hawaii
Pacific University in Honolulu.
JACOB N. KINNARD received his Ph. D. in History of Religions
from the University of Chicago in 1996. He is currently a Mellon
Postdoctoral Fellow in the Department of Religion at Northwestern
GERI H. MALANDRA is department head of Professional Develop-
ment and Conference Services at the University of Minnesota's Uni-
versity College (continuing education), and Adjunct Assistant Profes-
sor of History. She is the author of Unfolding a Mandala: The Bud-
dhist Cave Temples at Ellora (SUNY Press: 1993).
CHARLES D. ORZECH is Associate Professor of Religious Studies at
the University of North Carolina Greensboro. His Politics and Tran-
scendent Wisdom: The Scripture for Humane Kings and the Creation
of National Protection Buddhism is forthcoming in the Hermeneutics
Series of Penn State Press.
FRANK REYNOLDS is Professor of the History of Religions and
Buddhist Studies at the University of Chicago where he also serves as
the Program Director of the Institute for Advanced Study of Religion.
JOHN STRQNG is Professor of Religion at Bates College. His most
recent publIcations include The Legend and Cult of Upagupta and The
Experience of Buddhism.
Frank Reynolds
For many years Buddhist Studies has been dominated by research that has
focused on particular "national" traditions on the one hand, and on written
texts and textual traditions on the other, The project which has culminated
in this Special Issue of JIABS represents an attempt to move beyond the
limitations on our understanding that the dominance of these two
components of the received Buddhological orthodoxy has imposed.
Stated in a more positive manner, there are two closely correlated
hypotheses that the conference that spawned this Special Issue was
designed to explore. These two hypotheses can be stated at a very broad
level as follows. The fIrst is that for a period that extends roughly from the
3rd century B. C. E. through at least the 13th century C. E., Buddhists were
the proponents and bearers of a civilization ai, "interregna1" religious
tradition that gradually became more or less fIrmly established throughout
most areas of Asia. 1 The second is that during this period, as in other
periods of their history, Buddhists expressed, communicated, reconfigured
and implemented their various orientations toward the world (including
worldly power), and salvation beyond it, not only through the medium of
written texts, but also-and very importantly-through a variety of sensory
media as well.
For the purposes of this particular conference, the temporal scope and
topical focus were necessarily delimited in a much more precise manner.
The period from the beginning of the 8th century C. E. to the middle of the
9th century C. E. was chosen as an appropriate temporal span because it
constitutes what is arguably the highpoint of the Pan-Asian spread of
Buddhism as a dynamic interregnal tradition. M a I ) . ~ a l a imagery, symbolism
and practice was chosen as an appropriate thematic focus because during
this century and a half various mwujalic expressions, along with associated
forms of ritual practice, were emerging as especially prominent media
1. The strategy of substituting the term "interregnal" for the term "international"
is explained and employed by Jonathan Walters in his superb but as-yet unpub-
lished paper entitled "Finding Buddhists in Global History."
178 JIABS 19.2
through which Buddhist orientations were being experienced, molded, and
adapted to changing situations.
The immediate stimulus for organizing the "Ma:t;l<;lalas on the Move"
conference that was held at the University of Chicago on April 21, and 22,
1995 came from a fascinating book by Geri Malandra entitled Unfolding a
Mandala: The Buddhist Cave Temples at Ellora (State University of New
York: Albany, 1993). Written by an art historian from the University of
Minnesota, Unfolding a Mandala focuses on the central role of ma:t;l<;lala
imagery and patterns at an important Buddhist cave site in the Indian
Deccan; and particularly on the development of this imagery and patterning
during the late 7th and early 8th centuries C. E. Using art historical methods
(directly correlated written texts could not be identified) Malandra traces the
emergence of a series of complex ma1Jq,aiic structures that she associates
with an early form of Esoteric Buddhism.
In addition, she points to
variety of interesting correspondences between the Esoteric ma:t;l<;lala
imagery that she had discerned at Ellora, and the emergence-in the 8th to
9th centuries-of similar kinds of Esoteric expression all across the
Buddhist world.
The original versions of the five essays published in this Special Issue
were presented, along with a superb paper by Julie Gifford (Whitman
College). In addition, important contributions to the discussion were made
by four scholars who served as respondents to specific papers: Jonathan
Walters (Whitman College), Richard Cooler (Northern Illinois University),
Robert Campany (Indiana University at Bloomington) and John Holt
(Bowdoin College).
Among the papers that were presented at the conference and included in
this Special Issue, three highlight the emerging importance of Esoteric
traditions and describe the meanings that Esoteric ma:t;l<;lalas conveyed, and
the uses (both. soteriological and mundane) to which they were put. Geri
Malanda's p ~ p e r sets the stage by placing the Esoteric / ma:t;l<;lala
developments that can be discerned in the Buddhist cave temples at Ellora
within the context of an Asiatic "world system" in which Buddhists were
actively and creatively involved.
2. In the conference, and in the papers included in this volume, the terms
"Tantric Buddhism" and "Esoteric Buddhism" are both used, and no attempt to
formulate a systematic differentiation between them has been made. In this
"Preface" I have chosen "Esoteric Buddhism" because it is the term which I take
to be the more comprehensive and encompassing.
Charles Orzech follows with a paper in which he discusses the slightly
later coming to prominence of Esoteric / maI)qala elements in China,
especially in the higher reaches of the Tang court. He provides a
fascinating description of the maw;ialas that were employed, and analyzes
the rich complexity and variety of the correlated practices generated by
adept practitioners. David Gardiner, in the third essay, emphasizes both the
very early presence of Esoteric elements in Japanese Buddhism, and the
crucial role played by Kl1kai-after his return from China in 806 C.E.-in
establishing a systematic basis for the emergence of the Esoteric / Shingon
tradition. In his discussion of the central role that maI)qalas played in
Kukai's fascinating and highly complex system, Gardiner focuses special
attention on the rich maI)qala symbolism that he integrated into the
construction of the important new monastic complex established on Mt.
The other two papers presented at the conference strike notes of caution.
They warn us not to become too exclusively captivated by the emerging
importance of Esoteric orientations and associated maI)qala imagery during
this period. They do so by pointing to extremely important and highly
innovative examples of contemporary Buddhist imagery that are not
Esoteric in character and not associated with maI)qalas-at least with the
kind of maI)qalas that were so central in specifically Esoteric contexts.
Julie Gifford, in the conference paper that is not included in the present
collection, made a strong case for her contention that Borobodur, the great
Buddhist monument constructed during the late 8th / early 9th centuries in
Java, is an architectural embodiment of an orientation that is not-despite
the interpretations of several previous scholars--ciistinctively Esoteric and
does not involve distinctively Esoteric maI)qala symbolism. Rather, she
argued, this great Buddhist monument displays and makes available a
classically Mahayana soteriology of emanation and reabsorbition in which
specifically Mahayana visualization techniques playa central role. (Those
interested in Gifford's very original and fascinating interpretation will have
to wait until her University of Chicago dissertation on the topic is
Jacob Kinnard, in a paper that follows Gardiner's in the present
collection, provides another cautionary example. Kinnard argues that
previous scholars have greatly exaggerated the role that Esoteric traditions
and Esoteric maI)qalas played during the 8th and early 9th centuries in the
Buddhist "homeland" in northeastern India. Rather, he contends, the
Buddhists who lived in this tradition-rich area adopted a rather conservative
iconographic stance. Following up on an insight of Paul Mus, and basing
1.80 JIABS 19.2
his argument on contemporary archaeological, art historical evidence rather
than the much later textual evidence from Tibet, he maintains that the
historical Sakyamuni was the primary focus of iconographic attention.
What is more, he demonstrates that the most prominent imagery involved
local pilgrimage sites where, according to long established tradition, the
major events in Sakyamuni's historical life had actually taken place.
Kinnard's argument leads him to suggest a major revisionist hypothesis
that received much support in the discussion that followed his presentation.
The hypothesis which he proposes is that Buddhist Esoteric / maI}.<;lala:
traditions, rather than developing at the "center" of the Buddhist world and
spreading outward to the periphery, actually developed primarily on the
periphery where they provided marvelously efficacious means for
constituting a Buddhist world or worlds in contexts in which the traditions
focused on the historical Sakyamuni were less accessible and less
John Strong's essay-which is a revised version of the "summary
response" that he gave at the closing session of the conference--contains a
series of brilliant reflections on the topic of maI}.<;lalas and the "moves
maI}.<;lalas make." In this essay Strong identifies many of the central themes
that were highlighted in the papers and conference discussions; he
introduces a number of fascinating new insights of his own; and he weaves
these various strands together in a way that provides not only a most
appropriate conclusion for this Special Issue, but also a most exciting
stimulus for further research and reflection.
All of those who have participated in the "MaI}.<;lalas on the Move" project
greatly appreciate the financial support that we have received from three
administrative units at the University of Chicago: the Institute for the
Advanced Study of Religion (a research arm of the Divinity School), the
Committee ~ p Southern Asian Studies, and the Buddhism Workshop. We
are also grateful to Donald Lopez who has presided over the publication of
this Special Issue of JIABS.
The MaJ.).qala at Ellara / Ellara in the Mal).qala
This paper originated as the introductory contribution to a conference on the
mru;u;lala as an exemplar of the ways Buddhism moved and evolved through
Asia in the eighth to tenth centuries. I was asked to set the context for the
conference, working from the evidence at Ellora, the case most familiar to
me and one which, I will argue below, suggests that a constellation of
beliefs and practices surrounding the malJ.<;lala was already influencing the
layout and use of certain Buddhist sites in India as early as the seventh
century. During the conference and while editing this paper for publi-
cation, I was challenged to clarify the interrelated methodological issues of
treating images (and sites) as "texts" and extrapolating from the presence of
mal)<;lala-like images that esoteric practices took place at a particular site.
The fIrst issue has to do with the appropriateness and utility of interpret-
ing visual images as texts. Historians of religion, with access to written
texts and observations of ritual behavior, may have differing opinions about
the validity of deriving meaning from what is seen or, more often the case,
from what is seen when it is only loosely connectable to what is known
from a written or observed tradition. For art historians, images are com-
monly treated both as "texts" that can be explicated and "read," and also as
evidence that reflects and / or amplifies evidence from written traditions.
As W. J. T. Mitchell elegantly describes the assumption underlying this
interplay, "the dialectic of word and image seems to be a constant in the
fabric of signs that a culture weaves around itself." I Among art historians
of South Asia, there may be disagreement about which texts to connect to
particular images, but not about the appropriateness of using visual and
written texts "dialectically."2 I would argue that we have to start with what
1. W. J. T. Mitchell, !conology, Image, Text, Ideology (Chicago: University of
Chicago Press, 1986) 43.
2. For instance, in a recent and influential debate about aniconism and emblems
in early Buddhist art, Susan Huntington and Vidya Dehejia interpret visual evi-
dence in strikingly different ways. Yet, while they critique the interpretative
182 JIABS 19.2
is known and, in the case of Ellora, what we can know most about is the
visual evidence. To explicate this material requires interaction with a textual
tradition and whatever strands of the historical record and practice seem
most probably to be connected. This should be an interactive process;
visual evidence is analyzed until a pattern seems to emerge. The pattern can
be checked with known documents or other sites for possible correlations.
Ideally these will come from the same or a similar milieu. And, the more
complicated the pattern of repetition among sites or sites and texts, the
greater the likelihood that a similar pattern of meanings underlies them.3 If
relationships seem to appear, identifications of images and their positions
can be tried out. If the "tests" do not work, the images can be examined
again to see whether a different pattern might emerge and, simultaneously,
additional documents or sites can be sought for alternative identifications. 4
At the same time, there are limitations to this approach. For instance, if
esoteric Buddhism was practiced at Ellora, not all of its meaning would
have resided in the sculptures themselves, or even in a written textual tradi-
tion. Some of the knowledge necessary to worship there would have been
passed on orally and that tradition, invisible as it may be, would also be part
of the larger "text" of the site.
context of this material over the past century; as well as one another's
approaches, they both work from a double assumption that, on the one hand,
early Buddhist texts can be used to identify and explain the subject matter of the
images they discuss and, on the other, that if we adopt the right approach, the
"narrative" content of the images is sufficient for us to decode their meaning.
See Susan L. Huntington, "Early Buddhist Art and the Theory of Aniconism,"
Art loumal49.4 (1990): 401-408; Vidya Dehejia, "Aniconism and the Multiva-
lence of Emblems," Ars Orientalis 21 (1991): 45-66 (see particularly p. 51);
Susan L. Huntington, "Aniconism and the Multivalence of Emblems: Another
Look," Ars Orientalis 22 (1992): 111-156 (see particularly, pp. 124-125).
3. This methodology is not exclusive to a Buddhist or South Asian context; see
Henry Maguire, Earth and Ocean: the Terrestrial World in Early Byzantine Art
(University Park, Penn.: 1987) 2-3. In this case, too, text and image rarely come
from precisely the same time and place.
4. This can go in contrary directions. While further exploration of the Tibetan
Buddhist tradition might reveal a written text that better matches or "explains"
Ellora than the scheme I discuss here, searching the considerably later Tibetan
tradition would force anachronistic connections that may have no historic
5. The question was raised during the conference, whether there was physical
evidence of esoteric rituals at Ellora, for instance, traces of smoke on the ceiling
where ritual fires would have been lit. Inside the shrines, there is considerable
If we accept viewing images as texts in this broad sense, then we can
focus on the second question, whether we can postulate from the visual
evidence of a maJ).Qala (assuming there is agreement that a particular image
or groups of images is a maJ).Qala) at a site, that esoteric or tantric Buddhism
was practiced there. As I will discuss below, Ellora is a particularly
challenging place because of the absence of evidence other than its layout
and sculptures themselves. The presence of maI).Qala-like. groupings of
images provides the possibility of making analogies to better documented
places. However, it became clear during our conference discussions that
we should not expect point-by-point coherence in the application of
maI).Qalas, even when we are certain that maJ).Qalas were used at particular
places. Instead, esoteric iconography and practice seem to have been
applied in a more "modular" fashion, elements chosen to meet the needs of a
specific time and place. Still, the overall patterns-that I have used the con-
cept of the maQ.Qala to represent-seem to have been similar from place to
Although this paper may not resolve these questions, I hope that by
bringing them to the surface, other scholars will be inspired to contribute to
their solutions. Many loose ends may continue to remain untied, and many
analogies may defy logical proof. However, our conference's discourse
across the various cultures in which Buddhist took root generated new
insights that would not have occurred without this kind of intellectual travel
across boundaries. Challenging places like Ellora will become comprehen-
sible if we can continue to compare, contrast, stretch and test our means of
Historical Context
In 806, as the well-know story goes, the Japanese monk Kukai returned
home from China, after two and one-half years of esoteric training in the
Tang capital of Ch'ang-an. For presentation to his emperor, Kukai carried
smoke damage. In the halls outside the shrines, most plaster has fallen away
and with it, traces of smoke. Without archaeological dating, which has not yet
been attempted to my knowledge, it would be difficult to determine whether the
shrine smoke was contemporary with use of the caves, or accumulated later by
the non-Buddhists known to have frequented the site. For instance, the 1278
describes the visit to the caves by the famous Maharashtrian saint
Cakradhara who used them as a place of refuge; see S. G. Tulpule, ed., Mhai
Bhat, vol. 1 (Nagpur and Poona: 1964): 22-26, 44. I am indebted
to the late Professor Tulpule for directing me to these passages and for his assis-
tance in reading them.
184 JIABS 19.2
hundreds of text scrolls, many statues of Buddhist deities, ritual objects,
and numerous mal).galas, including the Vajradhiitumahiima1J4ala and the
Mahiikarur;adhiitumar;cjala. These would form the core of Shingon Bud-
dhist teachings and practice, revolutionizing temple layouts and- linking
them closely to the official activities of the Heian court. 6
At the other end of the Buddhist world, finishing touches on the largest
and last of the great Buddhist cave temples had been completed at Ellora.
seventy-five years earlier. There, at the ritual center of the rising Ra.;;.trakii.ta
empire, a three-tiered temple was filled with Buddhist sculptures, arranged
in unique, mal).<;l.ala-like patterns brought, perhaps by an Indian "Kukai" to
Ellora around 700. By the early 800s, the nearby Kailasa temple was
completed under patronage, even larger than its Buddhist and
Brahmanical predecessors and equally unique in its use of the rock-cut site
to express an iconographic vision. With such wonders, Ellora attracted
visitors and worshippers from India and abroad who, witnessing worship
in the extraordinary Hindu and Buddhist shrines, might have been as aston-
ished as modern visitors at the sagacity and power of the to
support "state-of-the-art" temple building for not one, but two major reli-
gions of their time.?
Unfortunately, apart from the sites themselves and inscriptions recording
the expansion and political and economic life of the little pri-
mary evidence for their activities exists. In the absence of direct evidence,
the legendary-and admittedly rudimentary-image of Kukai, laden with
mal).<;l.alas, provides a seductive analogy for the wayan individual, under
official auspices, could have physically imported new ideas about icono-
graphy and practice to Ellora and initiated the creation of new structures to
house those precious images.
6. Chikyo History of Mantrayana in Japan (New Delhi: 1987): 37.
David Gardner's paper at the conference decoded the mythic aspects of this
7. The Kailasa project was reportedly a source of amazement to its own artisan.
In the Baroda plates of the Ral'trakUta king Karkaraja, it was put this way:
... a temple, the architect-builder of Which, in consequence of the failure of his
energy as regards [the construction of] another such work, was himself sud-
denly struck with astonishment, saying, "Oh, how was it that I built it!"
(R. G. Bhandarkar, "The Rashtrakuta King I and Elapura," Indian
Antiquary 12 [1883]: 229.)
While we have other records of the transmission of Buddhist texts,
Kukai's, because it is an idealized version, offers a much fuller account of
the role that maQQalas played in the movement of esoteric Buddhism
through A s i ~ in the eighth and ninth centuries.
Even when the myth is
stripped away, we know that they were crucially important and precious,
displayed in temples throughout the Buddhist world (some of which were
constructed specifically for worship prescribed in these sutras). But they
were just part of a complex of material items and knowledge that were
requisite to the transmission of esoteric Buddhist teachings. So, when we
look at a very different kind of place, like Ellora, where all we have are the
maQQala-like iconographic programs of the temples, we should use the
"Kukai analogy" in exploring the visual record to determine if it might
reveal traces of individual action and royal support similar to the pattern of
KUkai's accomplishments.
As an official visitor to the Tang capital, Kukai was able to study with
Hui-kuo, a key teacher who was, himself, part of a direct line of transmis-
sion of Vajrayana teachings from India. The three great teachers of this
line-Subhakarasimha (637-735), Vajrabodhi (671-741), and
Amoghavajra (705-774)-all taught both the Mahiivairocanasutra and the
Sarvatathiigatatattvasaritgrahasutra, and were responsible for introducing
them in China.
SubhakarasirQha, who studied at Nalanda, was also an
official emissary, carrying the Mahiivairocanasutra to Ch'ang-an in 716 at
the invitation of the Tang emperor. It was under one of his disciples that
Hui-kuo studied this sutra, and he is traditionally thought to have studied
the Sarvatathiigatatattvasa11J.grahasutra under Amoghavajra, a disciple of
Vajrabodhi, who brought that sUtra to China in 720.!0 We know, of
course, that these teachers were concerned not just about text transmission
but about initiating disciples, properly setting up maQQalas, and ultimately,
with the efficacy of the rituals they conducted to support the ruler and the
8. The widely repeated story of Kiikai's accomplishments should not be
accepted at face value, as Charles Orzech convincingly demonstrated in "Seeing
Chen-yen Buddhism: Traditional Scholarship and the Vajrayana in China,"
History of Religions 29.2 (1989): 87-114 and in his paper at this conference.
9. Orzech, ibid.: 90-93, argues that both siitras, and m8.I}.Qala traditions, were
promulgated together by all three teachers.
10. Minoru Kiyota, Shingon Buddhism: Theory and Practice (Los Angeles:
1978): 17-19.
11. Orzech, op.cit. 91, n. 6.
186 JIABS 19.2
International travel was a significant factor in these transmissions.
Subhakarasill)ha and Vajrabodhi, who may both have been from Central
India originally, and Amoghavajra, were reputed to have traveled exten-
sively, in India and southeast Asia, to gain advanced experience irr esoteric
practices. It seems evident that, in the seventh and eighth centuries, there
was no single seat of esoteric knowledge; world-famous teachers circu-
lated throughout India, and ambitious practitioners traveled widely from
within India and from abroad to seek them out.
Krrkai's story is, then,
emblematic of an idealized pattern for the transmission of esoteric teach-
ings in Asia. An official monk / traveler is eager to collect texts and
maI).<;lalas from abroad; his findings are of great interest to his ruler, who
believes that esoteric Buddhist rituals can help ensure the safety and health
of the state. He makes contact with a top teacher of the tradition, creating a
legitimate line of succession, and transmission of carefully translated texts,
but he also creates something new-an original synthesis- to be practiced
in his homeland. This pattern followed a precedent going back centuries,
for international travel to gather the highest and most current Buddhist
12. Although later than the period of Ellora and this conference, medieval
Tibetan texts are the richest source on the history of tantric Buddhism. It is in
this literature that we find traces of a link, through Saraha, between Vidarbha
(eastern Maharashtra) and Orissa. Saraha was the teacher of Nagarjuna, the
teacher of Nagabodhi, who was Vajrabodhi's teacher; Vajrabodhi was supposed
to have been born in Central India around 670. See Andre Bareau, "Der
Tantrismus," Die Religionen Indiens, II: Buddhismus, linismus, Primitivvolker
(Stuttgart: 1964) 173; Alex Wayman, The Buddhist Tantras (New York: 1973)
13-14, argues for later dates for these individuals. In various accounts, Saraha
is said to have been born in Vidarbha, performed a mahiimudrii ritual there, and
converted the people; H. V. Guenther, The Royal Song of Saraha (Seattle:
1969) 4-12; Lama Chimpa and Alaka Chattopadhyaya, trans., Taranatha's His-
tory of Buddhjsm in India (Simla: 1970) 102-106. Saraha, elsewhere called
Rahula, is also connected to Orissa where a seventh- or eighth-century inscrip-
tion refers to a Rahularuci, a mahiimar;rjaliicarya and paramaguru; A. Ghosh,
"Kha4ipada Image Inscription of the Time of Subhakara," Epigraphia Indica 26
(1942): 247-248; S. C. De, "The Orissan Museum Image Inscription of the
Time of Subhakaradeva," Proceedings of the Indian History Congress, 1949
(Allahabad: 1950) 66-74. Even if the Tibetan texts and inscriptions do not refer
to the "historic" Saraha, they demonstrate that, at least in the later tradition, it was
believed that tantric Buddhism was pursued and taught in the region abutting the
domain and the one from which more parallels to Ellora's iconogra-
phy appear than from any other region. For a more extensive discussion, see
Geri H. Malandra, Unfolding a Mar;rjala, The Buddhist Cave Temples at Ellora
(Albany: 1993) 16-17, and 133-134, notes 92-97.
teachings, linked directly to the original texts and instruction in India itself.
Disseminated both by land and sea, esoteric Buddhism in the eighth to tenth
centuries was an international movement, capable of transfer from one
. country to another within a single gerieration. The pace of the exchange,
and the concern for accurate translation and practice, might suggest on one
hand, that if we found two places where the same teaching was pro-
pounded, we should see similar visual evidence of it. On the other hand,
the ideal also provides for creative adaptation to local circumstances, so we
may find that this "mal).c;lala" of esoteric centers was loose, indeed, capable
of creative adaptations to local contexts.
These travels and transmissions should also be considered within the
broader context of a developing "world system." Although monks had
traveled by land and sea to view sacred places and collect texts for cen-
turies, with the advent of Islam and the rise of the Tang dynasty in the early
600s, the motivations for and pattern of land and ocean voyages shifted and
expanded, linking India, Southeast Asia and China in a network of
increasing economic as well as religious exchanges.!3 By the ninth
century, Arabic geographical accounts show in detail the growing
knowledge of routes between the Persian Gulf to India and on to China. 14
Whether transcontinental or oceanic, these routes drew together
international communities of traders and others. In these centuries, there
c l ~ a r l y was not the kind of economic system that developed in the next
millennium but, as Chaudhuri has suggested, on some levels and in some
ways people in the "Indian Ocean civilisations" (Islam, India, Southeast
Asia, China) considered themselves to be part of an "entire structure." 15
The adaptation of related teachings of a world religion in three of these four
civilization areas (and in Japan, to add a fourth relevant to Buddhist culture)
depended on the connections being made by more secular travelers within
this "structure." Thanks to Arabic geographers and historians, we have
detailed eyewitness accounts of these activities. They give us a glimpse of
a nascent "world system," which not only contained esoteric Buddhism, but
may have been just the right medium in which it could develop and expand.
13. K. N. Chaudhuri, Trade and Civilization in the Indian Ocean: An
Economic History from the Rise of Islam to 1750 (London: 1985) 44.
14. In such works as the Relation of China and India (851) and Ibn
Khurdadhbih's Book of Roads and Provinces (846-85); see Chaudhuri, Trade
and Civilization, 49-50.
15. K. N. Chaudhuri, Asia Before Europe: Economy and Civilisation of the
Indian Ocean from the Rise of Islam to 1750 (Cambridge: 1990) 48.
188 JIABS 19.2
Foreign enclaves were, for example, welcomed on the west coast of
India, it has been suggested, because of the Brahmanical aversion to mar-
itime travel. As "Hindu" empires grew from the ninth century onward,
these enclaves and the trade contacts they brought, became increasingly
From this point of view, there was interaction with a wider
world and developments on the Indian subcontinent cannot be viewed in
isolation even if most internal records seem to be inward-looking. So, for.
instance, in the ninth century the empire was recognized as one
of the four imperial formations of the medieval world, even as this dynasty
described itself only in terms of its Indian conquests and territories. The
king, one of whose titles was Vallabharaja, appears as "Balhara" in an
Arabic list of four great kings in the world, together with rulers of China,
Greece, and the Arabs. His capital was at Mankir, that is or
Malkhed, known to be the capital in the ninth century. 17 The
Balhara was considered by these writers to be the greatest of the kings of
aI-Hind, able to maintain his position in the face of numerous threats from
surrounding, lesser rulers. 18 Arabic geographers noted the
kings' wealth, displayed in the capital, filled with thousands of elephants,
and adorned with an "idolhouse" containing twenty thousand idols made of
a variety of precious materials.
The exotica ofIndia's royal and religious
presentations was "world class," at least from an Arab point of view.
By this time, the dominated India from the Ganges valley
south and west to Gujarat, across the Deccan, and south again to Kanci.
Their wealth may, in fact, be linked to the expansion ofIslamic trade with
India. 20 They likely benefited, at least indirectly, from the resources
accrued from trade along the coasts and, as one source reports, "no king
had more friendship for the Arabs."21 Yet, this system is not evident in
epigraphic records, typically inward-looking and concerned with
the more traditional matters such as royal donations and military campaigns.
So, we tempted to conclude that from the inside looking out, this
"world system" was not apparent or important to the How-
ever, both the Buddhist temples that were created in the early years of their
16. Andre Wink, Al-Hind. The Making of the Indo-Islamic World, vol. 1
(Leiden: 1990) 68-69, 101.
17. Quoted from SUlayman, merchant, A. D. 851, in Ronald lnden, Imagining
India (Cambridge: 1990) 213-215.
18. Wink, op. cit. 304-305.
19. From the Kitab al-Fihrisht, ibid. 305-306.
20. Ibid. 308.
21. Ibid. 306.
rise to power, and their Brahmanical temples, like those described in Mus-
lim geographers' accounts, exhibit through new scale, style, and iconogra-
phy a claim to transregional authority, the Kailasa temple for instance refer-
ring to earlier Kailasas constructed byPallava and CaJukya rulers in the
To the extent that the Pallavas, Ca]ukyas, or had
dominion over the coasts of the Deccan, this authority was ultimately con-
nected-if only through trading diasporas-to the expanding world system
of the Indian Ocean. And, it was in this system that esoteric Buddhism as
an international movement was to thrive, perhaps directly supported by
rulers who sought its special help in securing their power. From this point
of view, developments in the Indian subcontinent cannot be viewed in iso-
lation even if most internal records seem to be inward looking; there was
interaction with a wider world that, if only indirectly, related to what was
happening within India.
This was, then, the broader context in which
ma1J<;lalas were moving and changing.
Meanings a/the MalJq.ala
As a pretext for this conference, the mar:t<;lala served as a tangible symbol of
the depth and complexity of the exchanges and transformations we hoped
to explore. Often considered "diagnostic" of the presence of a certain
"stage" of Buddhism, they have been defined in myriad ways. In the
Tibetan tradition, ma1J<;lalas were grouped into those made of powdered
colors, those painted on textiles, those formed by meditation, and that
. formed by the human body. In a different kind of scheme, they were clas-
sified as five types: receptacle, causal, means, path, or fruit ma1J<;lalas.
22. A inscription in EUora Cave 15 describes Dantidurga as having
made the CaJukya king VaUabharaja (Klrtivarman II) his tributary; James
Burgess, Report on the Elura Cave Temples and the Brahmanical and Jaina
Caves in Western India, Archaeological Survey of Western India, vol. 5 (1882;
rpt., Varanasi: 1971) 88, verse 23. Various elements of style and iconography
in the KaiIasa temple appear to have been derived directly from the
temple at Pattadakal, itself patterned in part on the KaiIasanatha temple at Kanci
(see D. Chatham, "The Stylistic Sources and Relationships of the Kailasa Temple
at EUora," [diss., University of California at Berkeley, 1977] for a detailed dis-
cussion). See Malandra, op. cit., 9-10 and notes 42-50 for a fuller discussion of
the relationship of early history to Ellora.
23. Wink, op. cit. 225.
24. Alex Wayman, "Reflections on the Theory of Barabuc.lur as a Mm:ujala,"
Barabudur, History and Significance of a Buddhist Monument, eds. Luis O.
Gomez and Hiram W. Woodward (Berkeley: 1981) 146-149. The latter classi-
fication is found in Vajravarman's commentary, the Sundariilanikariiniima, on
190 JIABS 19.2
appearance, they are generally symmetrical or even geometrically arranged
groups of images,. organized in order of (and to guide) worship for those
initiated into the proper way of "reading" them. The point of using
maJ.lQalas is to engender enlightenment, through proper initiation and ritual
practice. They are, thus, sacred ground, which an initiate approaches in
carefully orchestrated steps, and into which the gods are invited to descend;
they are "the whole universe in its essential plan, in its process of emanation
and reabsorption."25 They are, in other words, as conceptualized the Shin-
gon tradition, the seat of realization of specific Buddhist insights: "a
maI)Qala is what gives birth to all buddhas ... "26
MaI)Qalas conceived as diagrams could be extended into a visualization of
concrete architectural space and, thus, were transformed into actual temple
architecture and sculpture, as I have argued Ellora was.
The universe-in-
a-maI)Qala may thus be described and represented as a palace and, con-
versely, the maI;lQala as a whole is conceived as being located in a ku.tiigiira,
a three-storied eaved palace resting on top of mount Sumeru.
maI)Qalas would contain layers or galleries, in which reside numerous
manifestations of buddhas, bodhisattvas, and other deities, whose
arrangement and numbers vary from mal).Qala to mal).Qala. These group ings
have been collected in iconographic lists in such texts as the Mafijusrf-
mulakalpa, Siidhanamalii, and the Kriyiisal1}graha, and
might be associated with specific teachers and / or schools.
The maI)Qalabecomes a kind of sacred ground and as such, can confer
advantages even to the uninitiated. Thus, in the Tibetan tradition, the ques-
tion was asked rhetorically:
the Sarvadurgatiparisodhana-tantra. Vajravarman was possibly a contempo-
rary of Atisa, who brought tantric Buddhism to Tibet in 1042 after study in the
tantric school at Snvijaya. Ibid.,140-14l.
25. Guiseppe Tucci, The Theory and Practice of the Mandala (New York:
1970) 23.
26. This according to the Mahiivairocanasiitra and Subhakarasinilia's commen-
tary on it; see Adrian Snodgrass, The Matrix and Diamond World MalJ!j.alas in
Shingon Buddhism (New Delhi: 1988) 120.
27. This was not an exclusively Buddhist phenomenon. Dennis Huston has
argued that a Vai(>I)ava maI)c;Iala was applied to the architectural form and icon-
ographic program of the 770 C. E. Vaikuntha Perumal temple at Kanchipuram;
see his "Vasudeva Kr(>I)a in Theology and Architecture: A Background for
Snvai(>l).avism," Journal ofVaisnava Studies, 2.1 (1993): 139-170. I am grateful
to Charles Orzech for bringing this study to my attention.
28. For instance, the specifically describes the ku.tiigiira on
Sumeru as housing maI)c;Ialas with the main deity in the center.
If someone were made only to enter the maIJ.Q.ala and not be conferred initia-
tion, what would be the advantage? [The answer given was:] If one takes the
refuge vow and beholds the maIJ.Q.ala of faith, there is the advantage that he
becomes purified from sins accumulated for many aeons and plants in his
stream of consciousness ... the disposition ... of becoming in future times a
receptacle fit for entering the profound mantra path. 29
In this, the maJ)Q.ala functions as a tlrtha does; I will return to this point.
But, is it sufficient to observe that Barabudur or Ellora are "like a
maJ)Q.ala?" On what basis might we postulate that a monument is a
maJ)Q.ala? Which one? If we cannot name it, does that weaken the analogy?
Does the presence of a maJ)Q.ala ensure that esoteric Buddhism was
practiced at that site? What more complex ranges of practice does their
presence suggest? Are there alternative analogies or models that would be
more productive in explicating what gave such sites as these their particular
Ellora: A Case Study
Ellora is a fitting case study because at no other Indian site of this period is
evidence for sculpted maJ)Q.alas so well preserved as early as it was there.
Yet, it has fallen into cracks between or overlapped the boundaries that
appear in standard historical and art historical accounts of this period and
has been generally missing from discussions of early tantric Buddhism and
its art. Much of what we see at EBora has its roots in its history at Ajanta,
Aurangabad and other cave temple sites. But, there is also much that was
new to the Deccan, and was connected more to places like Sirpur, Sanci,
Bodhgaya, and Ratnagiri in Orissa. Moreover, iconographic features
among all suggest a transregional diffusion of a teaching or teachings that
shared a core of common belief, across various dynastic and geographical
boundaries. 30
Best known as a major Brahmanical site and tlrtha, Ellora is located near
Aurangabad, in the "cave temple" region of western Maharashtra, about 150
miles northeast of Bombay. As a tlrtha, it was relatively easy to get to, ac-
cessible for centuries by a land route still in use today. Bus, train, and air
29. It is worth remember Wayman's caution, that " ... there is no revelation of
the maIJ.Q.alajust by exhibiting it, or by the disciple's mere seeing it" (The Bud-
dhist Tantras , op. cit. 59).
30. Detailed background, discussion, and illustrations of the points made here
can be found in Malandra, op. cit., passim.
192 JIABS 19.2
travel have made it even more accessIble to the tens of thousands of mod-
em-day tourists and devotees who visit each year.
The Kailasa temple has overwhelmed first impressions and serious study
of Ellora, but there is much more-thirty-four Brahmanical, Jain, and Bud-
dhist caves in all, dating from the late sixth to the early tenth century. It
was significant enough, as a Brahmanical site, that it appears in the litera-
ture of the emerging Islamic world system, in the travelogue of the
century traveler, al-Masudi (contemporary with, if historically unconnected
to KiIkai), who noted:
... the great temple named Aladra [Ellora], where Indians come on pil-
grimage from the farthest regions. The temple has an entire city dedicated
to its support and it is surrounded by thousands of cells where devotees
consecrated to the worship of the idol dwell)!
From the early eighth century, the tfrtha was visited by leaders,
long before Kri$I).a I took credit (in the 812 Baroda plates) for constructing
the Kailasa temple. They used it as a ritual capital even before they
assumed all the titles of empire, and continued to use it as such at least until
they moved their capital south to Malkhed (Mankir of Arab geography). In
plates issued at Ellora in 742, Dantidurga recorded his worship at
guhesvaratzrtha in connection with a gift of a village. The tfrtha also has a
place in the Puranic lists of jyotirlfngas where sriiddha should be per-
formed and, later, it appears in a list of 50 Siikta Local legends
provide a paradigm for Saiva worship at Ellora. In one version, a linga
arises from the "lake" at Elapura, a place where worship will absolve con-
flict and sin. In a medieval Marathi story about queen MaI).ikavatI and the
king of Elapura, even accidental worship-bathing in the tank there-alle-
viates suffering caused by sin. In gratitude for this expiation, which
1.,,"'- .'"
answered MaJJikavati's prayers, she had an entue temple to Siva con-
structed, perhaps the Kailasa excavation itself.
We might expect patrons of rock-cut architecture on this scale in such a
numinous place publicly to claim and bequeath the credit for such an
extraordinary expenditure of time, funds, and human resources. Yet, not
atypical for monuments of this period, what could be read of the only in
situ dedicatory inscription does not refer explicitly to patronage
or practice at the site. As noted above, their patronage was only briefly
31. C. Barbier de Meynard, ed. and trans., Mar;oudi-Les Prairies d'Or:
Muruj-ul-Zahab, vol. 4 (Paris: 1865) 95-96.
acknowledged in the Baroda plates issued in the early ninth century in
The Kailasa and other Ra$trakl1ta caves were not, however, the fIrst
Brahrnanical excavations at Ellora. Cave excavation began in the late sixth
century; one of its earliest caves dedicated to the worship of Siva, and most
similar to the great cave temple at Elephanta.3
So, when Buddhist
teachers, practitioners and patrons moved to Ellora around 600, they were
locating their worship at a Brahmanical site, a tlrtha "on the rise." And,
somewhat later, during the late seventh and early eighth centuries when the
Ra$.tra1..'l1tas were forming their empire, the Buddhist community was also
experiencing change. That this change may have included a considerable
increase in support is suggested by the increased scale and complexity of
Cave 11 and 12, created during in the early 700s. Without external evi-
dence, the changes at the site itself are our best indication of a Ra$trakl1ta
connection to the changes in Buddhism expressed at Ellora. Similarly dra-
matic changes in China and Japan, initiated with the support of royal
houses concerned about expansion and stability, suggest such a connection.
Moreover, this was the time frame in which Islamic merchants and armies
were just beginning to expand the systems in which goods and ideas circu-
lated in Asia, and that included the continuous movement of monks back
and forth to India from China and Southeast Asia to study and teach new
esoteric texts and practices.
The juxtaposition of Buddhist and Saiva (and later Jain) shrines shows
that Ellora's space was considered sacred in more than a Brahmanical con-
text. Just as the Kailasa temple refers explicitly to another sacred place
(suggesting a regional transposition of Maharashtrian for Himalayan
sacrality), so it is possible to consider that as a Buddhist site, it came to
suggest a similarly monumental transposition; Ellora for Bodhgaya. (Later,
the Jain community was to define it as its own tlrtha, making an explicit
analogy between Caranadri and KaiIasa.3
) I will return to this point later.
32. Recent excavations have revealed foundations of what appears to be an even
earlier Hinayana establishment in the vicinity of the caves; this was reported by
the Xinhua News Agency, New Delhi (October 24, 1994), reference via Inter-
net; thanks to Richard Lariviere for sending me this notice. The layering of
Hlnayana and Mahayana temples at the same site is well known in Maharashtra:
Ajanta, Aurangabad, Nasik, Kanheri, and Karle are among the most well-known
examples of this phenomenon. Until now, Ellora has been the exception with its
exclusively, and relatively late, Mahayana / esoteric temples.
33. In the donative inscription dated 1234-35, on an image of Parsvanatha in
Ellora's Jain excavations, the donor is said to have made " ... many huge images
,194 JIABS 19.2
But, despite its later fame and the weighty evidence of the Buddhist caves
themselves, we have no direct evidence to identify patrons or teachers.
Unlike many earlier cave temple sites, where donative inscriptions in situ
help locate them fairly precisely in time and dynastic affiliation, EIlora's"
Buddhist caves are anonymous. It has appeared to be, therefore, lacking
historical, religious, and historiographic importance. It has been treated as
marginal and derivative, as "the end of the line" by most art and religious
historians, who have typically looked to eastern India to explain changes in
style, iconography, and sectarian affiliation that we can observe in other
parts of the subcontinent. This is connected to the even broader tendency
many historians have had to see decline and deterioration in the religion, art,
and politics of late classical and early medieval India. Such work as Wink's,
AI-Hind and Inden's Imagining India are helpful counters to that attitude,
offering fresh explorations of the economic / cultural / political and ritual
context of religious monuments and objects. But even here, the importance
of Brahmanical establishments in the creation of "scales of forms" over-
shadows what are viewed as waning Buddhist activities.
Contrary to these opinions, Ellora embodies considerable, significant
change. It appears to be on a kind of boundary where transitions in iconog-
raphy, and then style, occurred. At Ellora we see the culmination of a mil-
lennium-long tradition of rock-cut Buddhist architecture in India. At the
start of the 12-cave Buddhist sequence around 600 C. E., style, iconogra-
phy, (and, by extension, teachings) derive in part from other nearby sites.
But, by the end of the sequence, around 730 C. E., more is different than
similar. Techniques, stylistic, thematic and iconographic idioms were in
place to be applied there; then new idioms were introduced in a "traditional"
style. And, fmally, in its last Buddhist temples, new style and iconography
appear, spanning different cultural zones. At this point, no later than 700,
we cannot.J,lnderstand it by treating it purely as a regional site. The tradi-
tional rock-cut environment was shaped-in places unevenly, experimen-
tally, incompletely, to house a new kind of sect and practice, with as many
connections outside as inside the region.
Given the limitations in the historical record, I would argue that we
should simultaneously treat E1lora as a text about itself, possessing an
internal logic, but also as part of a larger system. If its former "marginal
of the lordly Jinas ... and converted the Charanadri thereby into a holy t1:rtha
just as Bharata [made] Mount Kailasa [a tfrtha]"; James Burgess and
BhagwanlaI Indraji, "Elura Inscriptions," Inscriptions from the Cave Temples of
Western India (Bombay: 1881) 99-100.
status" is put aside, significant new internal and external connections
appear that do in fact suggest explanations for much about the site. We still
may not be able to name patrons or teachers and the results may not be neat
correlations to other kinds of sites and texts, but the results will ultimately
be more useful than the more common, linear search for forced correlations
between a known text and site that can be equally frustrating. And, looking
outward, Ellora's meaning can be placed in a network (a more fanciful
"maJ).gala") extending from the Buddhist caves within a Brahmanical tfrtha,
to other early esoteric sites in central India, to the wider range of places
connected by monks and traders traveling through the Buddhist world sys-
tem of the eighth century and onward.
Ellora's Buddhist temples followed patterns used for centuries in the
western caves, the typical layout including a caitya (cave with monolithic
stilpa) and several other excavations that served as worship, study, or resi-
dence halls. Here there is only space to bracket the earliest and latest mani-
festations of the malJgala as a simple, repeated, geometric arrangement of
buddha and / or bodhisattva images within their architectural enclosures.
The earliest at Ellora is worth noting briefly, to highlight continuities and
changes even in the early 600s, and to anticipate what was to happen a
century later.
Cave 6, like most of the early shrines, is a single-level temple. Its wide
entrance hall is filled with pillars, and not much else, empty cells lining the
side walls. In this cave, the only sculpture is found in the shrine and its
antechamber. There, four stunning images are carved: Bhrkuti and
A valokitdvara to the left of the shrine door; Maitreya and MahamayUri to
the right. The bodhisattva dvarapalas follow convention in iconography,
style and location. B'ut, these female figures are the earliest to display the
precise iconographic elements that clearly identify them. They are found in
several other of the seventh-century caves, and would be part of a much
more complex group of female figures in Cave 12 (to which I will return),
an indication that they were important, and original, members of Ellora's
earliest maJ)..gala.
Inside the Cave 6 shrine, a small seated Tara image was carved directly to
the left of the door. The left and right walls are each filled by a nine-bud-
dha maIJ.gala carved in shallow relief. The buddha images are undifferen-
tiated (they were painted, but the colors are not apparent today); all seated in
vajraparyaiikiisana, hands held in dharmacakramudrii. Below the left
wall group are three worshippers; two are crowned and, with attendants
and an elephant, it appears that they should be viewed as royal figures.
Below the right wall group are seated images of A valokitesvara, J ambhala,
196 JIABS 19.2
and Mafijusri (the earliest of the latter two at Ellora). The location ofthese
maI).c;lalas (on either side of a central buddha image) is similar to the location
of relief maI).c;lalas carved in a sub-shrine in Cave 12. They may also be
connected to the ten, nearly three-dimensional buddha images carved in the
left and right mm:ujapa galleries of Cave 2, from the mid-600s, and to the
nine buddha images carved in niches along the left and right mafJ,ejapa
walls of the third floor of Cave 12; and to the two groups of nine small
buddha images carved at ceiling level in the antechamber of that cave. But,
as a group, these nine buddhas have not been precisely matched with
gfOUPS known from other sites or written texts.
The central shrine image in Cave 6, on the back wall, is a dharmacakra-
mudrii, pralambapiidiisana buddha, similar to many at earlier Buddhist
sites throughout the western caves, attended by two chowrie-bearing bod-
hisattvas. And numerous "architectural" and stylistic motifs are so similar to
those of the earlier Brahmanical caves that it could be argued that the same
workshop produced this Buddhist cave and that, therefore, it dates from
around 600. This shows how traditional style, and a key image, could be
embedded in a new iconographic framework.
If we "fast-forward" to look at Ellora's latest Buddhist cave, it will be
apparent how much had changed in a century. Cave 12 is a three-level
excavation, everything executed on a larger scale than in Cave 6. One
approaches through a thick screen wall across a large, bright fore court. A
shallow stairway leads up through two entrance pillars into a dark
mafJ,ejapa filled with pillars and lined with small, empty cells. On the left
rear wall of the mar;ejapa is the first of the Cave 12 eight-bodhisattva relief
maI)c;lalas. Two others were carved in the cell that leads to the stairway up
to the second level; and two more are carved on either side of a buddha
image in a sub shrine between the fIrst and second floor; making a total of
five relief lJlaI).c;lalas in Cave 12. The bodhisattvas identities (according to
objects being held) are similar in all, as is the central, dhyiinamudrii buddha
They appear to be:
34. See Malandra, op. cit. 75-79, for a more detailed discussion of these
Maitreya (nagakesara) Samantabhadra
(sword) ( kalpadruma)
A valokitesvara Buddha Vajrapili).i (vajra)
Akasagarbha Mafijusri (book)
(banner) (bud / jewel)
Cave 12 Relief Maw;lala
Moving toward the shrine on the first floor, panels depicting
dhyanamudra buddhas were carved in large niches on both walls of the
six-pillared antechamber. Outside the shrine door, seated images of
Maitreya and MafijusrI were carved to the left and right, respectively.
Inside the shrine, an image of Tara was carved to the left the door (in the
same position as the one in Cave 6); to the right of the door is an image of
Cunda). On the left and right walls of the shrine are carved eight bod-
hisattvas, all seated in lalitasana. Carved on the back wall, the main shrine
image is a dharmacakramudra buddha attended by two nagas. Where
attributes are preserved, we appear to have:
Lokesvara (padma) A.kasagarbha (frothy bud /
Maitreya (nagakesara) [V ajrapaI).i?]
Samantabhadra (sword) Mafijusrl (book)
K$itigarbha (kalpadruma
(banner) branch)
Cave 12.1 Core Shrine program
The entrance to the second floor is through a cell, in which two more of
the relief maI).<;lalas are carved, which leads to a stairway up to a small front
shrine, where the central bhumispadamudra buddha image is attended by
A valokitdvara and Vajrapili).i. To the left and right of this image are the
last two relief maI).<;lalas; above the one to the right are images of Cunda,
Tara, and Bhrkutl. On the right wall of this small shrine is a triad com-
posed of A valokitesvara, accompanied by J ambhala and Tara. It is note-
198 JIABS 19.2
worthy that a large lotus medallion is carved on the ceiling of this shrine, a
decorative motif typical of eighth-century BrahmanicaI architecture.
All sculptures on the second floor are located along the central front-to-
back axis. The passageway to the ma1Jg,apa (another pillar-fIlled,hall lined
with empty cells) is framed, to the left, by an image of Avalokite.5vara
flanked by Tara and BhrkujI, and, to the right, an image of Mafijusri,
flanked by four female deities. The shrine entrance, protected by
A valokitdvara and Vajrapal)i leads, down two steps, to a relatively spa-
cious shrine. Inside, to the immediate left is an image of Tara; J ambhala is
carved to the right.
Four standing bodhisattva images were carved on the left and right walls
of the shrine; attributes (where preserved) suggest that this is the same
group as in the relief mal)<;lalas, and in the fIrst and third floor shrines of the
cave, with slight adjustments in position:
Maitreya (srupa in hair) Mafijusn (book)
Samantabhadra (sword) Lokesvara (bud)
[V ajrapiil)i?]
Akasagarbha (bud / jewel)
Cave 12.2 Core Shrine program
Above the}:Jodhisattvas to the left, at ceiling level, is a row of seven small
bhumisparsamudrii buddha figures; above and to the right, is a similar
group but with hands held in dhyiinamudrii. The central shrine image, a
bhumisparsamudrii buddha, is attended by A valokitesvara and Vajrapiil)i.
In front of the throne are images of BhUdevi and Aparajita, carved as if
rising out of the floor.
The third floor of Cave 12 is an extraordinary creation, filled with light
and major pieces of sculpture. Unlike the lower two levels, its ma1Jg,apa is
lined with nine buddha images instead of empty cells. The two "central"
ones on each side are seated in pralambapiidiisana, hands held in dharma-
cakramudrii. The remaining fIve are all seated in vajraparyankiisana. The
mudriis of these five vary; taking them in clockwise order from the
"entrance" in the front, left, they appear to have been: dhyiinamudrii,
dharmacakramudrii, dharmacakramudrii, dhyiinamudrii, 35 b humis-
Along the rear marp!dpa walls are two groups of seven
buddhas; those to the left hold their hands in dhyiinamudrii; trees above
their heads distinguish them as the six past buddhas and the buddha of this
age. Those to the right hold their hands in dharmacakramudrii.
The shrine antechamber is framed by female figures, as it was in Cave 6.
But here, in the densest expression of the mal)<;iala, there are twelve,
unprecedented in the western caves (or elsewhere for the early eighth cen-
tury). Distinctive iconographic details include a four-armed Cunda (third
on the left); the three-pronged vajra of SarvakarmavaranavisodhanI (seated
immediately to the left of the shrine door); the snake belt worn by Jafigull
(immediately to the right of the shrine door); the peacock of MahamayuIi,
second on the right; and the four arms and twisted daJ'.Lcja of Bhrkup-, fourth
on the right. Such specific attributes help in identifying the group as the
Dharal)Is who appear (in varying configurations, as described in later
iconographic compendia) in mal)<;ialas of Tara, Dharmadhatu Vagisvara,
and Mahavairocana. 37 Above them, to the left, at ceiling height are nine
bhumisparSamudrii buddhas; to the right, nine dhyiinamudrii buddhas.
Inside the shrine, Tara and JambhaIa again protect the front wall on either
side of the door. Four standing bodhisattvas are carved on the left and right
walls, holding objects that identify them as the same group in the lower
levels and in the relief mal)<;iaIas:
35. This is a correction; in Malandra, op. cit. 86, it is erroneously listed as
bhumispadamudril .
36. It is tempting to read these five as representations of the five Dhyilni bud-
dhas: Mahavairocana, Ratnasambhava, Amitayus, and
Amoghasiddhi, but what is preserved of the mudrils doesn't support this differ-
entiation. (We would expect, instead, to find dharmacakramudril for Vairocana;
bhumispadamudril for varadamudril for Ratnasambhava;
samilhitamudril for Amitabha; and abhayamudril for Amoghasiddhi.) Still, this
is a close as Ellora's iconography seems to come to a five-Buddha system;
otherwise, there seems to be a rather strong emphasis on triads of various sorts.
37. It is worth noting that locally, a compositional precedent for visually similar
female groups existed in the sixth- and the eighth-century Brahmanical caves,
where groups of saptamiltrkil images, are commonly found.
;wo JIABS 19.2
Maitreya (smpa in hair) MafijusrI (book)
Samantabhadra (sword) Aka"sagarbha ?

[K$itigarbha ?]
V ajrapru;ri? Lokesvara
Cave 12.3 Core Shrine program
Above them, at ceiling level on the left and right, are groups of seven bud-
dhas' hands of all held in dhyiinamudrii. The central shrine image is, again,
a bhumisparSamudrii buddha, with images of BhUdevi and Aparajitii
carved on the floor in front of the throne.
What does this condensation of Ellora's "text" tell us? Looking in
overview at the relief maI).c;lalas, while their locations suggest that they were
not part of the original programs of the cave, their content-eight dif-
ferentiated bodhisattvas surrounding a dhyiinamudrii buddha image-con-
nects them directly to the bodhisattva programs in the Cave 12 shrines (and
to the slightly earlier Cave 11). The kernel of the concept was there from
the beginning, but the content changed quite dramatically over the century
and a quarter of Buddhist activity. Carved in shallow relief, they were also
"unfolded" into the three-dimensional space of the cave shrines, in which
groupings of eight bodhisattvas frame the central buddha image: the "top"
row becomes the left shrine wall, the "bottom" row the right wall, and the
"center" ro;,w is the rear wall, containing the main shrine image.3
38. The question has been raised, why "unfold" rather than simply rotate the
maI).c;lala from a vertical to a horizontal position? Implicit in this is the broader
question of how literally a maI).c;lala concept or diagram would have had to be
transposed into a sculpted medium to be comprehensible and useful for ritual.
AtEllora, a more literal transfer would have had the effect of placing the Buddha
image in the center of the shrine instead of on the rear wall, making it possible
physically to circumambulate it. There is, of course, evidence from the BhUdevi
and Aparajitii images that sculptors could carve images in three dimensions.
Moreover, behind the Buddha throne in the Cave 12.3 shrine, rough cutting sug-
gests an attempt to prepare a small (although the image would
still have been essentially on the back wall, not in the center of the shrine).
group as a whole, if identifications based on correlating attributes with
those found in later iconographic texts are correct, corresponds generally to
lists such as those found in the Siidhanamiilii, the
Unfortunately, such overlapping similarities prevent us from identifying
Ellora's the eight bodhisattva group with anyone text.
The central shrine images should also contain important information. As
do two of the three key shrine images in Cave 12, shrine images of the sec-
ond floor of Cave 11, and many of the small "intrusive" images in the two
latest caves depict the Buddha holding his right hand in bhumis-
padamudrii. It is commonly viewed as the emblem of Sakyamuni or
(as contrasted with the dhyiinamudrii in the relief mandalas
which, as with certain forms of dharmacakramudrii, signifies an image of
Vairocana). This gesture has layers of meaning, but on the most basic
level, it symbolizes the event of the Buddha's enlightenment, which took
place at Bodhgaya. It appears that Ellora's creators did not want to leave
this interpretation in doubt: many of the main shrine images in Caves 11
and 12 include sculptures of BhUdevi and Aparajita, rising from the earth in
front of the throne; Bhildevi attesting to the Buddha's integrity as he faced
Mara's attack; Aparajita trampling on the back of a male figure, representa-
tive of the "evil beings" she slaps into submission with a hand raised in
These images condense the lesson to be learned about the
power of enlightenment and of the Buddha himself. They are unique and
strikingly early at Ellora. Similar, although later, images have been found
in eastern India,including Bodhgaya itself, and Ratnagiri in Orissa.
The precision of Ellora's compositions strongly suggests that worship in
the shrines could have been viewed as a substitution or transposition of
worship at Bodhgaya-not an unreasonable expectation at a site that was
However, this circumambulatory was not completed and in the other late-seventh
and early eighth-century shrines, convention seems to have dictated that most
images be carved on the walls of the shrine even when, as in Cave 8, a circum-
ambulatory passage was excavated around the shrine.
39. TIlls connection was suggested to me initially by Janice Leoshko. It is doc-
umented in D. C. Bhattacharyya, "The Vajraviilf-niima-ma1JrJalopayika of
Abhayakaragupta," Tantric and Taoist Studies in Honor of R.A. Stein, ed. M.
Strickmann (Brussels: 1981) 74-75. In the Sadhanamala, Aparajita ispor-
trayed trampling GaI).apati; she is the destroyer of all wicked beings. M.-T. de
Mallmann, Introduction d l'Iconographie de Tantrisme Bouddhique (Paris:
1975) 245-246 and figs. 189 and 190. One hand appears to cup a breast; the
texts say that the left hand rests on the heart, holding the sacred thread and mak-
ing the gesture of danger, tarjanlmudra.
202 JIABS 19.2
also to carry Brahmanical prayers to l;teaven as effectively as pilgrimage to
Mount KaiIasa can. This may be linked back to conceptualizations of ~ h e
maI).Qala, which could be viewed as "an ideal Bodhgaya, an 'adamantine
plane,' that is, an incorruptible surface, the representation of the vefY instant
in which is accomplished the revulsion to the other plane, in which one
becomes Buddha."4o
Buddhist Ellora thus exemplifies the attitude expressed in the later
medieval period by a Maharashtrian saint who advised, "stay in Maha-
rashtra because every place worth going to is there."41 The RastrakiItas in
effect brought to Ellora every place worth going to. Transformations and
interaction of geography, politics, and religion combined to create a power-
ful regional tfrtha, part of a universal sacred system to which architecture,
sculpture, and religious practice refer. The patrons and teacher(s) respon-
sible for this extraordinary transformation of the site must have been
thinking in what we might call transregional terms; terms in which most
historians and art historians have not viewed it.
What might have been on the margin was made the center. But, who
brought it, and from what school? A search in traditional written sources
has revealed only partial, or general connections. The relief maI).Qalas might
suggest a connection to the teachings of the Mahavairocanasutra. But the
central shrine images, so clearly Sakyamuni, may be better seen as connect-
ing Cave 12 to teachings of the Sarvadurgatiparisodhanatantra, centered
on Mahavairocana / SakyasifI).ha. But in this, the five-buddha system is
well developed, as it is not at Ellora. Instead, emphasizing Sakyamuni,
attended by A valokitesvara and VajrapllI).i, Cave 12 may simply reflect an
earlier teaching similar to what was classified in the Tibetan tradition as
kriyatantric Buddhism (as reflected in a text like the Mafijusrfmulakalpa),
that foreshadows what would become differentiated in later tantric tradi-
tions. There are no perfect matches from the known literature; comparisons
as above carFsuggest certain parallels. But Ellora, relatively early, predates
these later systemizations. It gives us a glimpse, still difficult to interpret,
of the expression of one such system in a very early form.
Other Early Expressions of the Mm:uJala at South Asian Buddhist Sites
EIlora, unique as it is, is not an isolated case where esoteric Buddhism was
expressed in a cave temple format, although it is the earliest to exhibit such
40. Tucci, op. cit. 86.
41. Quoted in Anne Feldhaus, "Maharashtra as a Holy Land: A Sectarian Tradi-
tion," Bulletin of the School of Oriental and African Studies, vol. 49 (1986): 544.
systemfltic, extensive tantric influence. Traces of tantric iconography have
been identified at several sixth or early seventh-century sites in the western
caves, ranging from Kanheri on the coast just north of modem Bombay, to
Nasik in the ghats west of Ellora, to Aurangabad, just miles away from
Ellora, and likely the site most closely connected to it in time as well as
space. Aurangabad, whose latest caves probably date to the late 500s, may
exemplify an even earlier form of tantric Buddhism than does Ellora, as
John Huntington argued in a 1981 article. He hypothesized that Caves 6
and 7 were expressions, respectively, of the MahiikanaJiigarbhadhiitu-
mm:u!ala and the Vajradhiituma1J4ala. He argued that, if Subhakarasin)ha
and Vajrabodhi were known to have transmitted the teachings of both
siltras which include these rnaI).qalas, then it is likely that they were
expounded together somewhere earlier in India, as they would be later in
China and then, in Shingon Buddhism. He, therefore, looked for a site that
might display both. 42
He noticed that, although the central buddha images in the shrines at
Aurangabad display a generic dharmacakramudrii, like buddha images
throughout the western caves, buddha images in the small sub shrines dis-
play dhyiinamudrii and dharmacakramudrii. These he connected to the
mudriis of buddhas in the GarbhadhiitumalJ-rJala and the Vajradhiitu-
malJ-rJala, respectively (in the latter, the mudrii is actually bodhyangi-
mudrii. This group of three buddha images is, in Huntington's argument,
the "key" to the cave's program. He described the bodhisattva figures,
PadmapaI).i and VajrapaI).i, as corresponding "exactly to the Mahiivairo-
canasiltra ... in that the two halls of the bodhisattvas VajrapaI).i and
PadmapaI).i flank the central eight-petal lotus hall and manifest the Prajiiii
and KarulJ-ii of the buddha Vairocana." Unfortunately, he was not able to
draw out systematic correspondences; instead, he found, for instance, that
the central buddha image seems resolutely to be the traditional dharma-
cakramudrii pralambapiidiisana form. In Cave 7, which he hypothesized
would represent the VajradhiitumalJ-rJala, we might expect the mudrii to be
bodhyangi; that it is not, he said, must be "because this image type is so
widespread in the western caves, they must be taken as generic images, not
42. John C. Huntington, "Cave Six at Aurangabad: A Tantrayana Monument?"
Kaladarsana, American Studies in the Art of India, ed. Joanna G. Williams
(New Delhi: 1981) 47-55. The article was intended as a test of the comparative
methodology; it hypothesizes an answer to, but does not definitively resolve the
question of whether the connection of these two mar;t<;!alas actually took place in
India, or whether it was a compilation formed somewhat later by esoteric teach-
ers in Tang China.
204 JIABS 19.2
specifically different from those of Cave Six."43 He further read the
remarkable, if iconographically general, group of female images in Cave 7
as portrayals of prajfia, pointing out that sixteen of thirty-two deities in the
Vajradhatumar;ala are female. Given Aurangabad's relatively generalized
iconography such close analogies are difficult to confirm. Moreover,
although it is clear that certain iconographic and stylistic idioms were trans-
ferred to Ellora, the overall program of these caves was not. Instead, at
Ellora, just a few years later, iconographic elements become much more
clearly differentiated, and programs are laid out in very different ways.
This is not suggestive of a comprehensive teaching, comparable to Shin-
gon, that would have been readily transferred from site to site. Instead, we
would have to argue that the teachings of two rather different mal).qalas
went on in the same geographical area in the space of one or two
Important sites exhibiting similar teachings also developed in eastern
Orissa and south to Andhra where, according to several strands of tradi-
tion, tantric masters traveled and taught. Located on the periphery of the
traditional Buddhist heartland, each preserves unique variations and even
innovations in Buddhist art during the period when Buddhist missionary
activity also carried these ideas to Southeast and Eastern Asia. Among
these, Ratnagiri provides more iconographic parallels to Ellora than any
other site.
Located on a tributary of the Mahanadi River (which connected coastal
Orissa to interior centers like Sirpur, which also shares certain iconographic
elements with Ellora), Ratnagiri was developed as a major monastic and
temple site from the mid-eighth century. There, thirteen of sixteen large-
scale buddha images are portrayed in bhumisparSamudra, one including an
image of Aparajita defeating Gal).apati. Ellora's eight-bodhisattva mal).qala
is also strikingly similar-in content, not style or format-to the
carved on steles at Ratnagiri (three-dimensional
images of such groups are also found at the nearby sites of Lalitagiri and .
Udayagiri). Moreover, the iconography and location of larnbhala and Tara
as shrine protectors, are more like those at Ellora than anything else we can
find in Maharashtra.
These, and many other similarities (but not identities) strongly suggest
that the eighth-century Orissan sites shared elements of a cornmon teaching
43. Ibid. 50.
44. See Malandra, op.cit. 16,21,70, 76,97-98, 106-107, 111, 115-116, 121,
for a detailed discussion of these parallels.
with Ellora, despite differences in political regime, artistic culture, and his-
tory of Buddhist development in their respective regions. Nancy Hock,
locating Ratnagiri in the Indo-Tibetan tantric tradition, has made the case
that the earlier bhumisparSamudra buddhas were intended to represent the
more or less ):listoric Sakyamuni (as distinguished from the
focus of more advanced tantric teachings, also depicted with hands held in
this position).45 Although the specific text describing the configuration of
images at Ratnagiri has not yet been recovered, she has shown that the
pantheon of deities found there seems most like that described in relatively
early tantric texts such as the MafijuSrzmulakalpa, classified as a
kriyatantra in the Tibetan canon. This Mantrayana system is distinguished
from the later, anuttarayoga system, practiced later at Ratnagiri, represented
by horrific deities quite different from the benign images of the earlier
The new buddha image, certain female deities, and the proliferation
of bodhisattva images that we find in Ellora's latest Buddhist caves 11 and
12 have counterparts in Ratnagiri's earlier stage, although style, emphasis,
and placement differ considerably. Iconographic parallels can also be
drawn to images from sites along what was the even more extended margin
of the Buddhist world, as missionaries carried these teachings to places
such as Sahlihundum in Andhra Pradesh and Candi Mendut in central Java
(where the eight bodhisattvas, Jambhala and Harm, Cunda, and BhrkutJ
images offer striking similarities to Ellora's "core" mar:l(:lala, even while the
central image is a dharmacakramudra buddha).
45. Nancy Hock, "Buddhist Ideology and the Sculpture of Ratnagiri, Seventh
Through Thirteenth Centuries (India)," diss., University of California, Berkeley,
1987, 1-33. David Snellgrove also makes this distinction, in Indo-Tibetan Bud-
dhism, vol. 1 (Boston: 1987) 117-152.
46. Hock 68-69; A.K. Bemet-Kempers, Ancient Indonesian Art (Cambridge,
1959) 40-41 and Plates 58-61; N. J. Krom, "De bodhisattvas van den Mendut,"
Bijdragen tot de tall-, land- en volkenkunde, door Koninklijk Bataviaasch
Gennotschapp van Kunsten en We tenschapp en, vol. 74 (1918): 419-427; 1. L.
Moens, "De Djandi Mendut," Tijdschrift voor Indische Taal-, Land, en
Volkenkunde, door Bataviaasch Genootschap van Kunsten en Wetenschappen
59 (1921): 529-600. Lokesh Chandra, "Borobudur as a Monument of Esoteric
Buddhism," South East Asian Review 5 (1980): 17-21, compares this
iconography with the Garbhadhatumar;.qala, but also sees a connec tion to the
iconography of the Durgatlparisodhaniimar;.qala where, he says (p. 18) the
Vairocana image also displays dharmacakramudrii.
~ 0 6 JIABS 19.2
Among these and many other places, EIlora appears to reflect esoteric Bud-
dhism on the cusp of change, at a relatively early point. If it was geograph-
ically peripheral from the point of view of the great university at N alanda, it
was central in that it more than "kept up with the times." We might even
say, in aspiring to be the "Bodhgaya of the south" it erased, in sense, the
geographic and sectarian boundaries that separated them.
How far do these selective comparisons get us in understanding who was
responsible for what happened at Ellora? Although the maQ.c;lala is a clear
link throughout Ellora's Buddhist development, its content had changed
radically by the end of the seventh century. The teachings represented in
Cave 12 are not the same as in the earlier caves. And, the new ideas were
carved in a new style. So, new teacher / teachings, new artisans. Where
did they corne from? Taking iconographic details as primary evidence, the
answer would appear more likely to be Orissa than eastern India, or a place
that sent teachers to Orissa and the Deccan. Certainly, legends about eso-
teric teachers who reportedly came from, or preached in the Deccan, sug-
gest but cannot prove this scenario. This was, after all, during the time
when teachers like SubhakarasirQ.ha and Vajrabodhi were active in India
and then in China.
One way to imagine how this happened is in the context of the pre-impe-
rial "opening" of the Deccan in which the early Ra*akujas appear to have
been engaged. The regional and national references in the iconography at
Ellora are not, then simply evidence of a dynastic change. They could be
seen as part of the activities the Ra:*akUjas engaged in to forge the charac-
ter of their new empire. We could see them, already worshipping at the
nearby tfrtha, endorsing if not actively supporting the creation of new
Buddhist temples at the most active center in their region. A "cutting edge"
(if, now, a!{$mymous) teacher would have been recruited to Ellora, or a
local mOnKcould have been sent out to study with such a person, linking
this region to the growing international network of esoteric teachers and
sites. If the ambitious Ra$jrakuja leader took a more personal interest in
these developments then Cave 12 and its maQ.c;lala might indeed have been
as central to the official activities of this new empire, as the movements of
monks supported by emperors were to the north and east in Java, China
and Japan. This analogy suggests that, just as Kukai built on already exist-
ing juxtapositions of beliefs, in support of Buddhism and the Heian empire,
so we might imagine Buddhism at EBora as a century-old tradition primed
for the advent of new ideas in support of a newly broad-thinking dynasty.
Ellora's "Kukai" may well have gone out to seek them, returning, as Kukai
himself.did, to create an original, local synthesis of new and old concepts
and practices.
MaI).<;lalas on the Move: Reflections from Chinese
Esoteric Buddhism Circa 800 C. E.
The construction and use of was an essential part of the Esoteric
or tantric Buddhism which spread throughout Asia from the eighth century
Yet the role of malJ.qalas as the vehicle for a complex, conserva-
tive, lineage-based and initiation-controlled ritual system is seldom exam-
ined. Focusing on malJ.qalas in the propagation of one lineage of late
eighth-century Chinese Esoteric Buddhism, I propose that maI).qalas repre-
sent the traces of a tradition at once conservative and designed to be readily
adapted to new missionary fields.
But before I proceed I must meet the most obvious of objections: There
are no malJ.qalas from eighth-century Chinese Esoteric Buddhism. The
Esoteric Buddhism of India, Tibet, Central Asia, and Japan teems with a
variety of maI).qalas, and through these malJ.qalas we can trace the diffusion
of Esoteric Buddhism from India to Japan. In China maI).qalas dating to the
Yuan Dynasty (1280-1368) testify to the influence of Tibetan Vajrayana.
The Mo Kao cave-temples at Tun-huang and the cave temples at Ta-tsu and
An Yueh in Szechuan contain Esoteric Buddhist iconography and malJ.qalas
dating to the tenth century.2 Yet not a single painted, drawn, or SCUlpted
1. Even if we leave aside modem adaptations of mar:tqala in Jungian psychology
and new-age spiritual movements it is obvious that the "idea" of the mar:tqala is a
very portable one. The use of maI).qalas and pseudo-Sanskrit mantras in Taoist
ritual is a good exmnple of inter-tradition borrowing of maI).qalas. This borrow-
ing between Buddhism and Taoism seems to have been a two way street. As I
have argued elsewhere, Pu-k'ung or one of his heirs seems to have borrowed the
nonary configuration of Taoist cosmogrmns for the distinctive East Asian ver-
sion of the Vajradhatu maI.J.qala.
2. Matsumoto Eiichi, Tonka ga no kenkyu (Tokyo: Toho Bunka Gakuin, 1937)
is still the classic work on the art of the Mo Kao caves. For an overview of the
scholarship and a recent contribution to it see Henrik H. Srensen, "Typology
and Iconography in the Esoteric Buddhist Art of Dunhuang," Silk Road Art and
Archeology 2, 1009-92 (Kamakura: Institute of Silk Road Studies): 285-349.
On Ta-tsu see Liu Zhangjiu, Hu Wenho, and Li Yongqiao, eds., Dazu shike
~ 1 O JIABS 19.2
maI).q.ala can be definitively dated to the eighth century. Generations of
Japanese scholars.have labored to prove indisputably the Tang dynasty
provenance of the twin "Genzu" maI).q.alas which are at the heart of Kukai's
Although Esoteric Buddhism has used painted and sculpted
maI).q.alas, altars configured as maI).q.alas, maI).q.alas composed of syllables
(bfja) or symbols (samaya) visualized by the adept, and the body as
maI).q.ala, the only trace of these maI).q.alas dating from eighth-century China
are descriptions of their construction and use preserved in ritual manuals.
The situation is exactly the opposite of that described by Geri Malandra.
For Ellora all that remains is the sculpture. For Tang China all that remains
are the ritual manuals.
This is not so bad a state of affairs, since maI).q.alas are an artifact of prac-
tice, and what we have are the "how to" manuals. In this essay I will exam-
ine ritual manuals from Pu-k'ung's (Pu-k'ung chin-kang, Amoghavajra,
705-774) lineage connected with the Peifect Wisdom Scripture/or Humane
Kings Who Wish to Protect Their States.
The choice of these manuals is
yanjiu: Collected Works of the Researches on Dam Stone Carvings (English
subtitle) (Chengdu: 1985). For a survey of the An Yueh carvings and maI).q.alas
see Henrik H. Srensen, A Survey of the Religious Sculptures of Anyue, East
Asian Institute Occasional Papers 3 (Copenhagen: East Asian Institute, Univer-
sity of Copenhagen, 1989).
3. A continuing source of scholarly and religious controversy is the way in
which the maQ.Qalas described in Chinese texts differ from those dating from the
earliest period of Japanese Esoteric Buddhism. Partisans of Shingon and Tendai
Esoterism have a vested interest in discovering charters for their interpretations
and practices either in the teachings of PU-k'ung (Amoghavajra) or in those of
his disciple Hui-kuo (Kukai's teacher). Unfortunately this has led both to an
anachronistic Shingonization of Tang Esoteric Buddhism and to its virtual dis-
appearance from the study of Tantric Buddhism. I have addressed these issues
in "Seeing r;Jhen-yen Buddhism: Traditional Scholarship and the Vajrayana in
China," History of Religions 29.2 (1989): 87-114. The literature in Japanese on
the relationship between the Genzu maQ.Qalas in use in Shingon and maI).Qalas in
a variety of sources in Tang translations is massive and the English literature on
the topic is growing rapidly. For a convenient summary see Yamamoto Chikyo,
Introduction to the Ma1Jala, (Kyoto: 1980) 64-82.
4. For Ellora as maQ.q.ala see Geri Malandra's essay in this volume and her
Unfolding a Ma1Jala (Albany: State University of New York Press, 1993).
5. There are two versions of this text. The version attributed to Kumarajlva (T.
245) and that of Pu-k'ung (Amoghavajra, T. 246), are available in Takakusu
Junjiro and Watanabe Kaigyoku, eds., Taisho shinsha daizokyo, (hereafter T)
85 vols. (Tokyo: Taish5 Issaikyo Kankokai, 1924-1935). For an introduction to
and description of the text see M. W. De Visser, Ancient Buddhism in Japan:
not arbitrary. Indeed, they comprise a key link between India and Japan
and they are emblematic of a tradition which is at once highly conservative
and very adaptable. As I will demonstrate, these manuals were central to
Pu-k'ung's Esoteric teachings and are representative of a number of other
manuals in tbe lineage, all of which are based upon principles outlined in
the SarvatathiigatatattvasafIJgraha (also known as the Vajrasekhara or
"Diamond Tip," T. 865-66, henceforth SITS). 6 Moreover, these manuals
and the mal)<;lalas drawn from them epitomize the close connection-found
throughout East and Central Asia-between Esoteric Buddhism and the
state. Pu-k'ung's Esoteric Buddhism, which Osabe has aptly termed "State
Protection Buddhism," sought the fulfillment of two goals, rapid enlight-
enment and benefits for the state.
In response to the needs of his imperial
patrons Pu-k'ung skillfully adapted ritual programs developed in South
Asia to the situation in the Tang court. It was these teachings which Pu-
k'ung's spiritual grandson Kukai imported to Japan. Though the continu-
ities between Pu-k'ung's eighth-century manuals and some ninth-century
Japanese manuals is astounding, Kukai and his heirs readily adapted the
Esoteric teachings to the ninth-century Japanese milieu. The key to this
missionary success was the conjunction of a clearly defined ideology and a
modular ritual structure.
Satras and Ceremonies in Use in the Seventh and Eighth Centuries A. D. and
their History in Later Times, vol. 1 (Leiden: E. J. Brill, 1935) 116-142. De
Visser's study includes a summary-translation of approximately one fourth of the
text. For a full study and translation of the scripture see my Politics and Tran-
scendent Wisdom: The Scripture for Humane Kings and the Creation of
National Protection Buddhism (The Pennsylvania State University Press, forth-
6. The best source on the Tang versions of the Sarvatathiigatatattvasanigraha
(T. nos. 865 and 866) is David L. Snellgrove's introduction to Lokesh Chandra
and David L. Snellgrove, Sarva-Tathagata-Tattva-Sa1igraha, Sata-Pitaka Series
vol. 269 (New Delhi: Mrs. Sharada Rani, 1981) 5-67. One should note that
Shingon references to the SITS are usually references to Pu-k'ung's text (T. no.
865) though refrences to Vajrabodhi's text (T. no. 866) and a host of commen-
taries is not uncommon. Both are truncated translations which focus on the first
major sec tion of the STTS text. The first complete Chinese translation of the
SITS was by Shih-hu (Danapala) in 1002 (T. no. 882).
7. Osabe makes this argument in Todai mikkyoshi zakko (Kobe: Kobe Shoka
daigaku gakujutsu kenkyukai, 1973) 90-91.
212 JIABS 19.2
Chinese Esoteric Buddhism has a complex prehistory which I cannot fully
rehearse here. Suffice it to say that between the third century and the
beginning of the eighth century South and Central Asian texts describing
maJ).<;ialas and studded with mantras and began first to trickle and
then to pour into China. 8 This piecemeal transmission continued until
Subhakarasinilia (arrived Ch'ang-an in 717), and then Vajrabodhi, and his
disciple PU-k'ung arrived in the Tang capital (721) to propagate and articu-
late comprehensive systems of Buddhist Esoterism. 9
The reception of these missionaries was quite different than that which
might be accorded in South Asia. Throughout most of the history of Bud-
dhism in China highly educated monks propagating the teachings quickly
became servants of the state. The acaryas of the eighth century were not
free to do as they pleased. On arrival at the court they were placed under
house arrest as "guests" in a government monastery where they could be
watched and interrogated. Once accepted they were put to work in the ser-
vice of the state with teams of translators, rendering texts and performing
rituals to augment state policy, to ensure seasonable rain, to repel invasion
and put down uprisings, and to help promote the well-being of the imperial
8. Chou I-liang's "Tantrism in China" (HJAS 8 [1945]: 241-332) remains the
best source on Chinese Esoteric Buddhism in English. Material on Japanese
Shingon, by comparison, abounds. Yoshito S. Hakeda's Kfikai: Major Works
(New York:: Columbia University Press, 1978) gives access to the writings of
Pu-k'ung's spiritual grandson, and works by Minoru Kiyota (Shingon Buddhism
[Tokyo and Los Angeles: Buddhist Books International, 1978]) and Taik5
Yamasaki, (Shingon: Japanese Esoteric Buddhism [Boston and London:
Shambala, 1988]) give access to Japanese Esoteric thought and practice. Tajima
Ryujun's, Etude sur le Mahavairocana-Sfitra (Dainichikyo) (Paris: Adrien
Maisonneuvre, 1936) is hard to get and comes from within the Shingon tradi-
tion, but it is more comprehensive than the works of Kiyota or Yamasaki. For
further bibliography in French, Chinese, and Japanese see my "Seeing Chen-yen
9. For the current purpose I will not be examining the tradition stemming from
the iiciirya Subhak:arasinilia. For standard overviews of his teachings and their
confluence with those of Vajrabodhi and Pu-k'ung see Mochizuki,
BukkyOdaijiten 3005a; Matsunaga, Mikkyodaijiten 1366c-1368a; Chou I-liang,
"Tantrism in China,' 251-272; and the transmission flow-charts in MikkyOdaijiten
beginning on p. 1. For more substantial bibliographical references see Orzech,
Politics and Transcendent Wisdom, chs. five and six.
family and its ancestors. Effectively to transplant Esoteric Buddhism to
China meant combining religious ideology and political expediency.
The dominant ideological and ritual template of Pu-k'ung's Esoteric Bud-
dhism was the SITS, and, though certain elements of the Mahiivairocana
Scripture (T. !;l0. 848, henceforth MVS) are used in his teaching, he actively
promoted the SITS over the MVS which had been translated and propa-
gated by Subhakarasinilia and I-hsing. The SITS was the centerpiece of
Vajrabodhi's teaching, and procuring a more complete version of the text
was the object of PU-k'ung's pilgrimage to India and Ceylon.10 As we shall
see, in pursuing the twin goal of enlightenment and protection of the state
PU-k'ung was adapting and extending the Indian teachings of the SITS and
its cycle of texts for his Chinese patrons.
The central role of the SITS as the ideological and ritual template of Pu-
k'ung's Esoteric teachings is enshrined in the story of the origin of the SITS
and its maI,lc:ialas in an "iron stUpa" in central India. 11 The tale begins with a
great worthy (ta-te, Sanskrit bhadanta) during the "Latter Age of the
Teaching" (mo-fa) who through the use of Mahavairocana's mantra had
gained a vision of Mahavairocana and of a teaching. which has traditionally
10. For the MVS see Tajima Ryujun, Etude sur le Mahiivairocana-Sutra:
(DainichikyiJ) , and Iyanaga Nobumi's excellent "Liste des abbreviations
(Bibliographie commentee)" which appears in "Recits de la soumission de
Mahesvara par Trilokyavijaya d'apres les sources Chinoises et Japonaises," in
Tantric and Taoist Studies in Honor of R. A. Stein, ed. by Michel Strickmann,
Melanges Chinois et Bouddhiques, vol. 22 (Bruxelles: Institut BeIge des hautes
etudes Chinois, 1985) 649-655. Yamamoto Chikyo has recently published a full
translation of the text, making it available for the first time in English. The
translation, however, lacks critical apparatus and is cast in a stilted, sometimes
incomprehensible English. Not withstanding, it is a useful and welcome contri-
bution by one of Japan's formost scholars of Shingon Buddhism. See Mahi'i-
vairocana Sutra (Delhi: International Academy of Indian Culture, and Aditya
Prakashan, 1990). I-hsing's biography is in Sung kao-seng chuan T. no. 2061,
11. The legend of the Iron StUpa recounts the origins of Pu-k'ung's Esoteric
Buddhism and the "reappearance" of the SITS and its maI).qalas and rites. The
story was related by Pu-k'ung based on the oral teaching of his master
Vajrabodhi and is found in Instructions on the Gate to the Teaching of the Secret
Heart of Mahii-yoga of the Scripture of the Diamond-Tip (T. no. 1789,
39.808aI9-b28, Chin-kang-ting ching fa yu-ch'ieh pi-mi shin tifa-men yi-kuei].
For a translation and introduction to the tale see Charles D. Orzech, "The Legend
of the Iron SWpa," Donald S. Lopez, Jr. ed., Buddhism in Practice (Princeton:
University Press, 1995) 314-317.
214 JIABS 19.2
been identified with the "Essential Rites for Vairocana" [T no. 849 and
chiian seven of the MVS].12 Using these techniques he then opened-the
iron stiIpa (i. e. he entered the maI).<;J.ala). Once inside the stiIpa his educa-
tion consisted of a course in the SITS and we are informed that the text of
the SITS available in China is but a superficial outline of the truly compre-
hensive scrip ture contained in the iron stilpa. A longer "outline" than that
now extant was supposed to have been brought with Vajrabodhi from
India, but this text which is described as "broad and long like a bed, and
four or five feet thick," was tossed overboard during a typhoon. Weare left
with the obvious conclusion that the total teaching is still available, but only
through initiation. 13
The legend simultaneously encodes the basic process of initiation into the
maI).<;J.ala world of the Esoteric school and into a hierarchy of teachings and
maI).<;J.alas. Through the process of homa (immolation) and consecration
(kuan-ting, Sanskrit abhi$eka) every initiate reenacts the burning away of
defIlements (klesa) and the entry into the iron stilpa with his or her own
entry into the ma1Jejala .14 The MVS represents the lower Esoteric teachings
whose mastery provides entry to the "higher" teachings. Once the initiate is
"inside," the STTS is both the avenue to enlightenment and the basis of
apotropaic ritual.
The dominant place of the SITS and the importance of the rituals related
to the Scripture for Humane Kings is evident in three document collections
compiled by Pu-k'ung's disciple Yuan-chao in the last quarter of the eighth
century. The bulk of Pu-k'ung's correspondence was collected in Tai-tsung
ch'ao seng ssu-/(ung ta-pien cheng kuang-chih san-tsang ho-shang piao-
chih-chi[The Collected Documents Relating to the Monk Pu-k'ung of Tai-
12. There is no proper Sanskrit equivalent for the term moja. For a discussion
see Jan Natti:er, Once Upon A Future Time: Studies in A Buddhist Prophecy of
Decline (Berkeley: Asian Humanities Press, 1991) 95-103.
13. Pu-k'ung's Indications of the Rites of the Eighteen Assemblies [Shih-pa hui
chih-kuei, T no. 869, 18.284c-287c] outlines this more comprehensive text. It is
notable that the fifteenth assembly consists of the Guhyasamiija yoga (Pi-mi chi-
hui yii-ch'ieh, 827a28-b7). for more on this see Kenneth Eastman, "The Eighteen
Tantras of the Vajrasekhara / Mayajala," (unpublished paper presented to the
26th International confe"rence of Orientalists in Japan, Tokyo, 1981). A brief
resume of the paper appeared in Transactions of the International Conference of
Orientalists in Japan 36 (1981): 95-96.
14. Indeed, as Kakuban and other esoteric masters make clear, "the iron stilpa is
this very body." See Kogyo-daishi zenshu (Tokyo: Kaji sekkai shisha, 1910)
tsung's Reign, T. no. 2120, circa 781, hereafter Piao-chih chi]. The corre-
spondance includes letters between PU-k'ung and the emperors Su-tsung,
and Tai-tsung as well as assorted other communications connected with the
Esoteric school, including Pu-k'ung's testamentary memorial. 15 These doc-
uments provide us a glimpse of Pu-k'ung's understanding of Esoteric Bud-
dhism and its role in the court. The Piao-chih-chi is supplemented by two
further works of Yuan-chao. The Ta-f ang chen-yuan hsu kai-yuan shih-
chiao lu [Supplement to the Catalogue of Buddhist Teachings of the Kai-
yuan Period Compiled During the Chen-yuan Period of the Great Tang
Dynasty, T. no. 2156, compiled in 796, hereafter Shih-chiao lu], and the
more expansive Chen-yuan shin-ting shih-chiao mu-lu [New Recension of
the Catalogue of the Buddhist Teaching Compiled in the Chen-yiian Period,
T. no. 2157, compiled in 800, hereafter Shin-ting], are invaluable sources
for the teachings of Pu-k'ung's lineage from just after his death in 774 until
just prior to Kukai's arrival in 804.
Time and again Pu-k'ung frames his life and mission in terms of the Five
Wisdoms in the SITS, and its five-fold mandalic structure became the tem-
plate for new rituals for his imperial and aristocratic patrons. PU-k'ung
states this plainly to his successors and to the emperor in his testamentary
Among the teachings I have translated, the Yoga of the Tip of the Vajra (the
SITS) is the path for quickly becoming a Buddha. As for the remaining sec-
tions of the mantra teachings, these I present to help the state avoid disasters,
to keep the stars on their regular courses, and to insure that the wind and rain
are timely. [T. no. 2120, 52.840bl-2]
Reading these comments from the perspective of later Japanese Shingon
tradition with its dual emphasis on the SITS and the MVS we might mistak-
enly assume that Pu-k'ung relied upon the teachings of the STTS for
enlightenment and on the MVS for "worldly" goals. An examination of
Tang ritual texts from Pu-k'ung's lineage shows instead that the SITS and
15. The text is T. 52.826c-860c. Raffaello Orlando ("A Study of Chinese Doc-
uments Concerning the Life of the Tantric Buddhist Patriarch Amoghavajra [A.
D. 705-774], diss., Princeton University, 1981) has translated some of the docu-
ments in this collection and lists the contents of others. Osabe Kazuo has trans-
lated the entire text into Japanese in "Gokyobu," Kokuyaku Issaikyo voL 98,476-
16. I examine the Piao-chih chi and both of Yuan-chao's catalogues in chapter
six of Politics and Transcendent Wisdom.
216 nABS 19.2
its central teachings not only provided the quick path to enlightenment, they
also provided the basic structure or "template" for the key rituals of Pu-
k'ung's state-protection Buddhism. Only the most advanced students were
granted full initiation to this teaching and Pu-k'ung took care to single them
Many are the disciples who have entered the altar to receive the Dharma.
Eight of them have been nurtured and established in the [Yoga of the] Five
Sections [the STTS], and two of these have died, leaving six persons [so
trained]. These are: Han-kuang of the Chin-k'o [monastery], Hui-chao of
Silla, Hui-kuo of the Ch'ing-Iung [monastery], Hui-Iang of the Ch'iung-fu
[monastery], and Yuan-chao and Chueh-chao of the Pao-shou [monastery].17
[T. no. 2120 844a28-b2]
The MVS is clearly secondary in Pu-k'ung's tradition.
While the SITS provided the basic ideological and ritual template for Pu-
k'ung's system, the Scripture for Humane Kings applied the teachings of
the SITS to the onset of the "Latter Age of the Teaching" and to actualize
the era of the Correct Teaching (Cheng-fa, Saddharma). Pu-k'ung and his
disciples Liang-pi, Fei-hsi, and Yuan-chao produced a "new translation" of
17. The Yuan-chao mentioned here is not the Yuan-chao who complied the
Piao-chih chi.
18. Iyanaga Nobumi summarises the evidence for all three acaryas having both
the SITS and the MVS in his "Recits de la soumission de Mahesvara, 706-707,
note 143. The transmission of these texts is quite complicated, particularly in the
case of the SITS. For the MVS see Tajima Ryfijun, Etude sur Ie Mahavairo-
cana-Siitra: (Dainichikyo), and Iyanaga's excellent "Liste des abreviations
(Bibliographie commentee)," 649-655. For the SITS see David L. Snellgrove's
Lokesh Chandra and D. L. Snellgrove, Sarva-Tathagatha-
Tattva-Smigraha 5-67, and Iyanaga, 656-657. Elsewhere, such as in the Tou-pu
fo-lo-ni mu (T. no. 903, 898c-900a) attributed to Pu-k'ung, but probably the
work of a disciple, both traditions are mentioned, and yet other disciples such as
his biographer Chiao Ch'ien make pointed reference to Pu-k'ung's teachings as
comprised the SITS and "the method of setting up the ma1Jq.ala according to the
Mahakarunagarbhama1Jq.ala of the MVS" (see T. no. 2056, 292c5ff, and 283a6-
9). Some Japanese Scholars, such as Ono Gemmyo have argued on the basis of
ma1Jq.ala iconography that Pu-k'ung is the author of the pure "dual ma1Jq.ala"
tradition. See for instance, Chandra, 37. It seems clear that Pu-k'ung used the
teachings of the MVS but not in a "double" sense as in Japan. The double
maJ:.lQala tradition probably arose in the generation after PU-k'ung and it never
came to be the all-encompassing ideological emblem that it did in Japan.
this anonymous fifth-century Chinese scripture in 765-66.
Liang-pi and
PU-k'ung also produced ritual manuals in tandem with the new scripture
and Liang-pi wrote a massive Commentary on the Scripture for Humane
The importance of the new recension of the scripture and its new1yeso-
tericized rituals is evident in the Pioa-chih chi. The Piao-chih chi begins
with Pu-k'ung's involvement in the denoument of the An Lu-shan rebellion
and establishes a parallel between the emperor (in this case Su-tsung) who
is the head of the state and the Buddha (and, by implication; the QCQrya, Pu-
The first section of the Piao-chih-chi presents us with a major
disaster which is rectified by close cooperation between the emperor and
the QCQrya and by new forms of ritual action and honors by the emperor. It
then goes on to portray the institutionalization of the Correct Teaching in
this dual sovereignty under the new emperor Tai-tsung. Through the care-
ful arrangement of a wealth of documents, Yuan-chao shows again and
again that this is the ideal polity for the empire, and that the Scripture for
Humane Kings is the emblem of that polity.
As I have detailed elsewhere, the structure of Yuan-chao's Piao-chih chi
was a response to particular historical circumstances. Te-tsung came to the
throne in June of 779 as an energetic 40-year-old bent on wresting some of
the control of the empire back from provincial warlords. He did not share
his father's regard for the Esoteric masters, and, indeed, he put an end to
esoteric rites in the palace and cut back patronage to the schoo1.
1 During
the years 782-85 Te-tsung was beset by a series of rebellions brought on
by his stubborn pursuit of renewed imperial control over the nominally
loyal warlords Who had entrenched themselves as an outcome of the An
19. According to Fo-tsu t'ung chi (T. no. 2035, 39.377c-378a) Tai-tsung per-
sonally ordered the new translation and compared the two versions. The imme-
diate cause of the new translation was the 765 Tibetan invasion of the Chinese
heartland lead by the Uigur commander P'u-ku Huai-en. P'u-ku Huai-en
dropped dead in camp and his forces fell into disarray. Pu-k'ung was credited
with this fortunate tum of events. According to Fei-hsi, the ritual that felled P'u-
ku Huai-en invoked the vidyaraja Acala who is often the focus of Humane
Kings rituals. For Fei-hsi's comments see Piao-chih chi 52.849al-5.
20. I have argued elsewhere that the unusual structure of the Piao-chih chi cen-
ters around a dual polity of Emperor and acarya and the role of the Scripture for
Humane Kings. See Politics and Transcendent Wisdom, chapter six, "A New
Buddhist-State Polity: PU-k'ung, Yuan-chao, and the Piao-chih chi."
21. For Te-tsung's reign, see C. A. Peterson, "Court and Province in Mid- and
Late T'ang, The Cambridge History of China, vol. 3 pt. 1 (Cambridge:
Cambridge University Press, 1979) 492ff.
218 JIABS 19.2
Lu-shan rebellion. What began in 782 as moderate success for imperial
forces soon turned into a nightmarish 'repeat of the events 37 years earlier
during the An Lu-shan rebellion. In 783-84 Te-tsung was forced to flee his
capital and to retreat, first to Feng-t'ien and then toward Szechuan. Not
until 785 did he reenter the capital. 22 The parallel with the events of the An
Lu-shan rebellion must have been on everyone's mind. Yuan-chao begins
the Piao-chih chi with a memorial by PU-k'ung congratualting Su-tsung for
retaking the capitol during the An Lu-shan rebellion. In the memorial Pu-
k'ung refers to ritual activity he had undertaken to repulse the invaders, rit-
uals which are elsewhere identified as invoking the vidyiiriija Acala (Pu-
tung) associated with esoteric performances based on the Scripture for
Humane Kings. Yuan-chao's careful sequencing of documents presents an
ideal Esoteric polity, a polity supported by rituals of the "humane kings"
established by Pu-k'ung.
The roles of the SITS and the Scripture for Humane Kings in Pu-k'ung's
teachings are also further evident in Yuan-chao's two "catalogues," the
Shih-chiao lu and the Hsin-ting. The large number of Esoteric scriptures in
Yuan-chao's catalogues reflects both the fact that these were the scriptures
that were "hot" in India and Central Asia in the latter part of the eighth cen-
tury and the dominance of Pu-k'ung's Esoteric teachings. The Hsin-ting is
indeed a massive catalogue of Buddhist scriptures, though one which high-
lights scriptures of the Esoteric school and particularly those aligned with
the SITS. But the Shih-chao lu is no comprehensive catalogue at all.
Instead it has three concerns: the new recension of the Scripture for
Humane Kings and its commentaries; the commentary on the Liu-ch'u
ching (another important text in the SITS orbit) and the role of the monk
Prajfia in its propagation; and the great vinaya commentary produced at the
An-kuo monastery.23 Yuan-chao was involved in these projects as
"recorder" a,pd 75% of the Shih-chiao lu consists of narrative accounts of
these projects.
The Shih-chiao lu is broken into three chiian with translations of scrip-
tures in the first, commentaries in the second, and catalogues, memorial
22. For the An Lu-shan rebellion see The Cambridge History, 453-463. For
Te-tsung's predicament see The Cambridge History, esp. 503-510.
23. The Liu-ch'u ching (Japanese Rishukyo) is a short Prajiiiipiiramitii with
decidedly tantric coloring. Pu-k'ung made a translation of the text (T. no. 243,
8.784a-786b) and Ian Astley-Kristensen has produced an excellent study and
translation of the scripture, The Rishukyo: The Sino-Japanese Tantric
Prajiiiipiiramitii in 150 Verses (Amoghavajra's Version Buddhica Britannica,
Series Continua III (Tring, U. K.: The Institute of Buddhist Studies, 1991).
stele, and other documents comprising the third chiian. 24 The first long
narrative to appear in the Shih-chiao lu is the narrative of the history of the
transmission of the Scripture for Humane Kings and the circumstances of
the production of its new recension, of Pu-k'ung's ritual commentaries on
the scripture,and of Liang-pi's great commentary [749c-753a with short
breaks; 758a-758c; 761c]. When we add the prominent role of the Scrip-
ture for Humane Kings in the Shih-chiao lu to its place in the Piao-chih chi
it is clear that Esoteric rituals connected with the Scripture for Humane
Kings were among the most visible signs of the adaptation of South Asian
Esoteric Buddhism to the Chinese milieu. 25
Before we can examine Pu-k'ung's ritual manuals and how maI).qalas con-
nected with the Scripture for Humane Kings were produced from them, a
brief overview of Esoteric ritual and the templates from the SITS is in
order. The purpose of ritual (siidhana) is siddhi (Chinese ch'eng-chiu,
sometimes transliterated as hsi-ti) a term which literally means the attain-
ment of a goal. In Esoteric Buddhism the basis of siddhi is often defmed as
the realization of the identity of the practitioners body (mudrii), speech
(mantra), and mind (samiidhi) with those of the "basic divinity," (Chinese
pen-tiun, Japanese honzon). Some treatments of Esoteric Buddhism tend
to over-intellectualize the tradition by focusing on the mental component.
Esoteric ritual involves all three components, mental, sonic, and somatic.
When siddhi is considered from the perspective of ultimate enlightenment,
anuttarasan;yaksambodhi, then one refers to it simply as siddhi or more
specifically as lokottara siddhi (ch'u-shih ch'eng-chiu, or ch'eng-chiu hsi-
ti). When this attainment is channeled toward action in the conditioned
24. While much of the material in the collection has been reassembled from the
Piao-chih-chi and other sources, occasional new details do appear. There are
mentions of 100 seat Humane Kings convocations as well as a narrative of the
grand convocation outside the south gate of the city in 765. Y iian-chao mentions
100 seat Humane Kings convocations at T. no. 2156, 55.751c9-18, including an
imperial reply, and another at 55.761c24. The great convocation outside the
south gate of the city in 765 is the one connected to P'u-ku Huai-en and the
Tibetan invasion and the new recension of the Scripture for Humane Kings. The
account appears at 55.752b27-753b8.
25. Together these documents provided the exempla for a "national protection"
Esoteric Buddhism which was exported to Korea and Japan, where it once again
underwent complex evolution. The Piao-chih chi was among the key texts taken
to Japan by Kukai.
220 JIABS 19.2
universe through images, maI).<;lalas, and mantras it is referred to as mun-
dane siddhi (Zaukika siddhi, shih-chien ch' eng-chiu) and is manifested
through application of supernormal powers used to aid in the salvation of
beings. Though the purpose of any given ritual might be predominantly
Zokottara or Zaukika, all rites assume both goals.
This "dual" structure is often described in terms of "inner" and "outer"
dimensions of performance. Rituals are are articulated in terms of the inner
versus the outer cosmos, the human body and the divine body, the samsaric
cosmos and the nirvanic cosmos. These relationships are established and
manipulated mentally, sonically, and physically through the use of visual-
ization, mantra, homa, nyasa.
Such correspondences are
repeatedly articulated in the Esoteric texts and commentaries. For instance,
in discussing homa (immolation) one text in the SITS cycle says that the
outer homa is the fire altar, the sapwood, and so on, while in the
"adamantine inner homa ... total enlightenment is the flame and my own
mouth is the hearth."27 Although siddhi is thus of "two types," each attain-
ment implies and requires the other.
The ultimate soteriological element of Esoteric ritual is "identification," or
the generation of the adept in the body of the divinity for the purpose of
insight into emptiness. Nevertheless, most rites, such as those of the
Scripture for Humane Kings, focus on the effect of such identification in
the world. Thus, most rituals are apotropaic, and the adept, acting as the
divinity, secures various sorts of blessings for a community. The
apotropaic dimension of Esoteric ritual has not escaped scholarly attention.
Stephan Beyer's The CuZt of Tara demonstrates this dual goal in Tibetan
rites to Tara, and the articles of William Stablein demonstrate the process in
Newar healing rituals. 29 Indeed, the two kinds of siddhi may be consid-
26. On nyqlSa (the localization of divine powers in the body) see Bharati, 273-
274; Eliade, Yoga, 210-211; and the Mahavairocana Siltra, T. no. 848, 18. 22a-
22b, 38b-38c.
27. From the Chin-kang-feng-lo-ko i-ch'ieh yu-ch'ieh yu-ch'i attributed to
Vajrabodhi,, 18.266aI2-21. Examples abound throughout the tradition.
28. The root texts of the tradition discuss not only the attainment of enlighten-
ment but also the attainment and use of supernormal powers. Rituals used to
obtain these siddhi comprise a sizable part of the latter portions of the SITS. So
too, texts aimed at immediate "worldly" goals almost always point out the
soteriological and transcendent insight gained in such practices.
29. Stephan Beyer, The Cult of Tara: Magic and Ritual in Tibet (Berkeley: Uni-
versity of California Press, 1973) esp. 254-258. William Stablien, "A Descrip-
tive Analysis of the Content of Nepalese Buddhist Piljas as a Medical-Cultural
ered the ritual realization of the two truths, a realization in which the adept
simultaneously becomes "world renouncer" and "world conqueror."
The cultivation of siddhi is the aim of the Esoteric teaching, and Esoteric
rituals, whether of the Indian, Tibetan,Chinese or Japanese variety, exhibit
a highly regular structure based on the metaphor-harkening back to the
Vedas-of inviting a guest for dinner. "At the most fundamental and overt
level, both Vedic and Tantric rituals are banquets in honor of the gods." 30
This is so in Vedic ritual (much else is, of course, going on), in pujii
("offering") which characterizes popular Hindu worship, and in the various
rites of the tantras. Indeed, if we examine the sixteen traditional upaciiras
or "attendances" of household and temple pujii we find remarkable corre-
spondence with homa and other esoteric rites. 31 Not surprisingly, one
mainstream tradition in Japanese Shingon (Chuinryu of Koyasan) divides
most rituals into five modules based on the guest metaphor: purification,
construction, encounter, identification, and dissociation.
However the
stages of ritual are divided, what occurs is the construction of a world-of
a mm;t<;lala-in which the adept and the buddhas, bodhisattvas, or guardian
divinities can meet. This fundamental mandalic structure is a simulacra of
the cosmos with Mahavairocana enthroned in the palace at the summit of
the realm of form (the heaven). The ritual construction of the
maI,l<;lala is the construction of the universe. The process of construction
culminates in the consecration of the adept)3 Realizing the
System with Reference to Tibetan Parallels," A. Bharati, ed., In the Realm of the
Extra-Human: Ideas and Actions (The Hague: Mouton, 1976) 165-173, and his
"Tantric Medicine and Ritual Blessings,"The Tibetan Journal 1 (1976): 55-69.
30. Wade T. Wheelock, "The Mantra in Vedic and Tantric Ritual," Harvey P.
Alper, ed., Mantra (Albany: State University of New York Press, 1989) 111.
For an introduction to and brief bibliography on the vast topic of Vedic ritual see
Jan C. Heesterman, "Vedism and Brahmanism," Encyclopedia of Religion
l5:217b-242a. For the metaphor of the "guest" in Vedic ritual see J. C.
Heesterman, The Broken World of Sacrifice: An Essay in Ancient Indian Ritual
(Chicago: University of Chicago Press, 1993) esp. 36-39, 188-89.
31. For a convenient overview of pujil and relevant bibliography see Nancy E.
Auer Falk, "Puja," in Encyclopedia of Religion 12:83a-85a. The list of the six-
teen upacilras, "invocation," "offering a seat," "greeting," "water," etc. is on page
32. For this analysis see Richard K. Payne, "Feeding the Gods: The Shingon
Fire Ritual," diss., Graduate Theological Union, 1985, 219ff. Payne's thesis has
been published as The Tantric Ritual of Japan: Feeding the Gods: The Shingon
Fire Ritual (New Delhi: Aditya, 1991).
33. The process begins on 237c and runs through 239b.
222 nABS 19.2
complete identity of his body, speech, and mind with that of
Mahavairocana, the fully initiated adept is consecrated Lord of the Triple
World. This "lordship" involves both world transcendence (lokottara sid-
dhi, Chinese ch'u shih ch'eng-chiu) and world conquest (laukika siddhi,
Chinese shih-chien ch' eng-chiu). This element of "world conquesf' was
taken quite seriously in East Asia and royal patrons endowed grand temples
with pennanent altar / maJ.H;ialas for a b h i ~ e k a and homa.
The Esoteric ritual system of Vajrabodhi, PU-k'ung, and their disciples at
first seems labyrinthian. One is confronted with hosts of divinities, bud-
dhas, and bodhisattvas, with seemingly endless ritual texts and ritual varia-
tions. In fact the system is quite straight-forward, even when we take into
account the tendency to ritual accretion and elaboration over time. As in
Taoist rituals scrutinized by Kristofer Schipper, we find here an architec-
tonic ritual paradigm or template out of which specific rituals are con-
structed to meet specific needs. These rituals are themselves composed by
stringing together a series of "rites" in a sort of boilerplate fashion. 34 Ritual
elaboration typically takes place through the embedding or nesting-some-
times recursively-of independent modules or rites into larger ritual struc-
tures, a process which usually involves some modification of the original
ritual's concern.
At the highest level, the architectonic paradigms-the "cosmologicaf'
outlines or templates-are laid down in the root texts of the tradition,
whether the STTS or the MVS.3
Each scripture details a distinctive
34. Kristofer Schipper, The Taoist Body, trans. Karen C. Duval (Berkeley: The
University of California Press, 1993). For Schipper's discussion of rituals and
rites see, pp. 75-80.
35. Fritz Staal has made similar arguments concerning Vedic ritual. See his
"Ritual Syntax," in Sanskrit and Indian Studies, ed. M. Nagatomi, et al.
(Dordrecht: D. Reidel Publishing Co., 1980) 119-142.
36. In addition to these two scriptures the Susiddhikara (T. no. 893) was seen in
certain late Tang Esoteric lineages as a third, integrating principle uniting the
MVS and the SITS. For these developments see R. Misaki, "On the Thought of
Susiddhi in the Esoteric Buddhism of the Late Tang Dynasty," Studies of Eso-
teric Buddhism and Tantrism in Commemoration of the 1, 150th Anniversary of
the Founding of Koyasan (Koyasan: Koyasan University Press, 1965) 255-281,
and Osabe Kazuo, "On the Two Schools of Garbhodbhava Esoteric Buddhism
in the Latter Period of the Tang Dynasty and the Method of the Three Siddhis,"
in the same collection, pp. 237-254. Osabe goes into greater depth in Todai
mikkyoshi zakko, 209-252.
mal)<;laIa structure. 37 Thus, when we examine the STTS, we find a sys-
tematic procedure for constructing a mal)<;lala / altar based upon the rela-
tionships of its constituent divinities. These relationships govern the con -
struction of the mal)<;lala / altar and the progress of the ritual, from the
establishment of boundaries, through the visualization of its central divinity,
to the visualization of its subsidiary divinities or their bfja or samaya. The
scripture gives the name and mantra of each divinity, his or her mudrii, and
an iconographic description for the puropose of visualization.
In the case
of the MVS the mal)<;lala unfolds as a lotus to reveal a pyramid-like hierar-
chy of Mahavairocana surrounded by the buddhas of the cardinal directions
and the interstitial bodhisattvas arranged on the eight petals of a lotus
throne. In the most common graphical arrangements (the so-called Genzu
mal)<;lala in Shingon) eleven other halls emanate from the petals represent-
ing the activities of the bodhisattvas. The MVS also classifies all of its
divinities into three categories: Buddha, Vajra, and Padma.
In the SITS Mahavairocana, seated in the karma assembly, is surrounded
by four buddhas representing four aspects of his wisdom. In contradis-
tinction to the pyramid-like hierarchy of the MVS each of these buddhas is
the center of another five-fold configuration which recursively replicates the
the larger pattern. Altogether the scripture describes the five buddhas, six-
teen prajfiii bodhisattvas and sixteen samiidhi bodhisattvas. These latter
are divided into female puJii and male prajfiii bodhisattvas. Thus, thirty-
seven divinities comprise this mal)<;lala. The SITS classifies its divinities
37. In the case of the SITS we find a five-fold structure composed of four bud-
dhas arranged at the cardinal directions surrounding Mahavairocana. In tum,
each of these buddhas is the center of a similar cardinal and recursive arrange-
ment. The divinities are seated on a lotus throne which rests on a lunar disk. In
the MVS Mahavairocana is surrounded at the cardinal and interstitial directions
by buddhas and bodhisattvas, but these stand alone and are not, as in the SITS
the center of further cardinal deployments. Each buddha or bodhisattva is seated
on a lunar disk which rests upon a lotus (the reverse of the SITS).
38. For example, the key "template" text for the teachings of the SITS in the mid
to late Tang was Vajrabodhi's Scripture Outlining the Meditations and Chants
in the Yoga of the Vajra Summit (Chin-kang-ting yii-ch'ieh chung Ziieh-ch'u nien-
sung ching [T. no. 866, 18.223b-253c]). The text is it truncated version of a
much longer Sanskrit text. This translation renders only the first part of the full
text and focuses on the establishment of the great mar;t<;lala and instructions for
the initiation of disciples. The actual construction of the maI).<;lala begins at 227a.
Each of the five central divinities is named and their mantras, mudriis and
iconography are detailed. Shih-hu (Danapala) provided the first full Chinese
translation of the SITS (T. no. 882) at the end of the tenth century.
224 JIABS 19.2
into Buddha, Vajra, Padma, Ratna, artd Karma, in accord with the type of
wisdom and maIJ.Qala descdbed in the text. While the MVS includes'the
Vajra-beings, they playa much more prominent role in the SITS.
Ritual "application" for specific purposes starts with the fundamental
template which governs the deployment of the maIJ.Qala / altar itself, the
names and iconography of the divinities in it, and their mantras and
mudriis. Each ritual manual (Sanskrit kalpa, "ordinance," Chinese i-kuei, Or
fa, "method"), is structured by the template in the form of the chief divinity
for the ritual. Thus, in the esoteric ritual for Humane Kings Prajfiii-
piiramitii bodhisattva of the SITS is the centnil divinity. Much of the rest
of a given ritual, its subsidiary divinities and sequences of rites, are drawn
from the SITS. Indeed, the construction of the mal).Qala / altar is largely a
matter of using boilerplate sequences appropriate to the STTS. These
sequences include the purification of the adept and the site, the construction
of the maIJ.Qala / altar, the expulsion of hindrances or evil influences, the
invitation of the three "departments" (in this case buddhas, bodhisattvas,
and vajra-beings) of the SITS, offerings (water, thrones, incense, lamps,
etc.), meditation on the chief divinity, and exit rites (usually the reverse of
preparatory sequences). All are structured according to the template of the
SITS using boilerplate recognizable in a variety of ritual texts by identical
sequences of procedures, mantras, mudriis, and divine names.
modular approach makes the system learnable, infmitely expandable, and
easily adapted to whatever needs a new context might require.
One dimension of the template derived from the SITS is a division of the
manifestations, functions, and attributes of divinities into "wheel bodies"
(lun shen, Sanskrit cakrakiiya).40
39. One example of such "boilerplate" is based on the Shih-pa kuei-yin, which
Shingon exegetes consider to be Kiikai's account of his master Hui-kuo's teach-
ing. The teXt is T. no. 900, 18.781c-783c. It sets out a standard sequence of
worship keyed to a series of mudrii. This sequence does in fact reflect
sequencing found in Tang dynasty manuals. For an outline of the sequence and
the mudriis see MikkyiJdaijiten 2:888a-889b. For the Tang manuals and their
sequencing see the tables of correspondence in Hatta Yukio's Shingonjiten
(Tokyo: Hirakawa shuppansha, 1985) 255 and 264-67. Hatta's dictionary is one
of the most important works for the study of East Asian Esoteric Buddhism to
be produced in recent decades. For a review see Ian Astley-Kristensen, "Two
Sino-Japanese Dharru:.u Dictionaries," Temenos 23 (1987): 131-134. "Boilerplate
Sequences in Pu-k'ung's Teachings" compares these sequences in key T'ang
dynasty texts and is found below.
40. Kiyota translates [un shen as as "Wheel-body" (Shingon, 103-104). Unfor-
tunately this is both meaningless and clumsy in English and it misses the the
According to the Sanskrit text of the Yoga of the Summit of the Vajra (the
SITS) in the possession of Tripitaka (Pu-k'ung) ... the five bodhisattvas
manifest bodies differentiated in accord with two kinds of wheel. In the
first-the Wheel of the [Correct] Teaching-bodhisattvas manifest their
bodies of truth (chen-shih shen) because this is the body received as recom-
pense for the practice of vows. In the second-the Wheel which brings about
the Teaching-[they] display their bodies of wrath (wei-nu shen) because it is
the body which, arising from great compassion, manifests as anger.42
Thus, each of the five buddhas of the STTS have three forms: Buddha,
bodhisattva, and wrathful vidyaraja.
Buddhahood-the state of enlight-
enment itself-is represented as the "Wheel body of the Self-nature" (tzu-
hsing lun shen, Sanskrit svabhavacakrakaya). Apotropaic rites focusing
on the beneficent teaching activities of bodhisattvas invoke the "Body
[which turns] the Wheel of the Correct Teaching" (cheng-fa lun shen, San-
skrit sadharmacakrakaya). These beings have the term vajra (chin-kang)
prefixed to their names. The chastizing and wrathful manifestations who
are transformations of the buddhas and bodhisattvas are designated the
"Body [which turns] the Wheel of Command" (chiao-ling fun shen, San-
metaphorical, mythical and cosmological connotations of the term (discus, realm
/ cakravala ruler / cakravartin etc.). I suggest lun shen abbreviates the phrase
chuan-lun shen, "wheel-turning body" or "body which turns the wheel of ... "
Thus, Cheng-fa-lun shen should be read as "body [which turns] the Wheel of the
Correct Teaching," and Chao-Zing-Zun shen is the "body [which turns] the Wheel
which Commands" or "brings about the Teaching." For the sake of fluid English
I call these the "Body of the Correct Teaching" and "Body of Command" respec-
tively. See Ian Astley-Kristensen's analysis of the term in The Rishukyo 136 and
41. "Bodhisattvas" (p'u-sa) here refers to the buddhas Mahavairocana,
Ratnasambhava, etc., in their compassionate activities.
42. The passage occurs in slightly different form in two places. It occurs in
Liang-pi's great Commentary on the Scripture for Humane Kings (T. no. 1709,
33.515c22-25) and in Instructions for the Rites, Chants, and Meditations of the
Prajiiiipiiramitii-dhiirm;l Scripture for Humane Kings Who Wish to Protect
Their States. (T. no. 994, 19.514a24-28) len-wang hu-kuo po-jo-po-lo-mi-to
ching fo-lo-ni nien-sung i-kuei, hereafter Instructions) a24-28. The "wheel-
bodies" are also covered briefly in Bukkyodaijiten 1857c-1858a, 1315b, and
43. These three are similar to the "families" of the MVS and it is tempting to see
these wheel bodies as indicative of the influence of the MVS. Perhaps their
inspiration is in the MVS but, as the quote indicates, in Pu-k'ung's manuals they
are clearly framed in terms of the SITS.
226 nABS 19.2
skrit Adesaniicakrakiiya). 'The termvidyiiriija (ming-wang) is suffixed to
their names and they are activated in rites of subjugation and in situations
where beings forcefully resist the teaching. 4 4 Liang-pi identifies the
Correct Teaching body of each of the five buddhas. For instance, the
Correct Teaching form of Mahlivairocana of the SITS is Vajraparamita (the
. chief divinity of the Humane Kings ritual), and this form represents the
body of Mahavairocana who, having just achieved the state of total
enlightenment, sets in motion the wheel of the Teaching "to transform and
guide beings" to the other shore (T. no. 1709, 33.516b12-16). In
apotropaic ritual one or two "wheels" (the Correct Teaching form and the
Wrathful form) may be activated. Although there are no extant graphic
representations of these wheels from eighth-century China, the Wheel
bodies and their associations were given physical expression in KiIkai's
Ninnokyo mandara which still survives in the lecture hall of T6ji.
When we examine ritual texts produced by Pu-k.'ung and his successors we
find that ritual aid for the state fell into two broad categories. Some rituals
were for the express welfare of the imperial family, both living and dead,
while others were designed to protect and. maintain the state and the cosmic
order more generally.46 Whether it involved an Esoteric revamping of ear-
44. It is tempting to identify the Wheel body of the Self-nature with the dharma-
kiiya, the Body of the Correct Teaching with the samboghakaya, and the Body
of Command with the nirmiivakllya. From this perspective the Body of Com-
mand proceeds from the Body of the Correct Teaching in a fashion reminiscent
of medieval Christian arguments about the Son and the Holy Spirit. Some Shin-
gon exegetes do exactly this. See, for instance, the entry in "Sanrinzin,"
MikkyOdaijien 2:844a-b.Tang Esoteric teachings do not specify whether the
Body of the Correct Teaching is a sa11J.boghakiiya or a nirmiivakiiya. Indeed, in
terms of ritual practice, both the Body of the Correct Teaching and the Body of
Command are forms of the Buddha's compassion (compassion and wrath
aroused by compassion).
45. See Mikkyodaijiten 4:1764c-1767a.
46. The most prominent of these are the "distribution" rites (shih-shih). Since
their appearance in the late Tang, the shih-shih have formed the ritual core of the
Ghost Festival and of rites for the recently dead. Both PU-k'ung and S i k ~ a n a n d a
translated manuals for rituals used to alleviate the suffering of beings in the
lower realms, rituals which found immediate applicatiori in rites for the dead and
in the yearly Ghost Festival. These texts serve both as guides to practice and as
accounts of the origin of the rites. They include The Conditions and Causes
lier Buddhist texts and rituals or the production of new rituals Pu-k'ung and
his successors consistently followed the template set out in the STTS.47
Although I will focus primarily on the ritual manuals connected with the
Scripture for Humane Kings, several other contemporary manuals pro-
duced under Pu-k'ung's tutelage form its cohort. Among these are the
Kuan-tzu-tsai p'u-sa ju-i-lun nien-sung i-kuei (T. no. 1085, hereafter Ju-i-
lun), the Wu-liang-shou ju-lai kuan-hsing kung-yang i-kuei (T. no. 930,
hereafter Wu-liang-shou), and the Chin-kang-ting lien-hua-pu hsin nien-
sung i-kuei (T. no. 873, hereafter Hsin kuei). I will return to these manuals
Three ritual commentaries on Pu-k'ung's new recension of the Scripture
for Humane Kings are attributed to Pu-k'ung (though likely the joint prod-
uct of Pu-k'ung and his close disciples).48 The most important of these
Which Gave Rise to the Teaching to Ananda Concerning the Essentials of the
Yoga [Tradition] on Distribution of Food to Burning Mouths (T. nos. 1318 and
1319), The DhiirmJI Sutrafor Saving the Burning Mouth Hungry Ghost (T. nos.
1313 and 1314), and the ritual text Distributions of Food and Water to Hungry
Ghosts (T. no. 1315). This last text presents a ritual, centered on the use of
mantra, for magically multiplying offerings of food and water to alleviate the
suffering of the countless beings in the lower realms. The new techniques of
offering are given authoritative charter through an account of the ritual's origins,
an account which is associated with these rites into this century. The earlier rites
connected with the Chinese Yu-lan-p' en Scripture have been treated by Stephen
F. Teiser, The Ghost Festival in Medieval China (Princeton: Princeton
University Press, 1988). Teiser has also explored the emergence of the Chinese
vision of the underworld in his study and translation, The Scripture on the Ten
Kings and the Making of Purgatory in Medieval Chinese Buddhism (Honolulu:
University of Hawaii Press, 1994). For an exploration of the the shih-shih rites
see Charles D. Orzech, "Esoteric Buddhism and the Shishi in China," Henrik H.
Sprensen, ed., The Esoteric Buddhist Tradition, SBS Monographs Series 2
(Copenhagen: Seminar for Buddhist Studies, 1994) 51-72. I have translated the
story of the origin of the shih-shih rites in appendix two of that publication, and
have also done an updated translation titled "Saving the Burning-Mouth Hungry
Ghost," in Donald S. Lopez, ed., Religions of China in Practice (Princeton:
Princeton University Press, 1996) 278-283.
47. Again, we have the various "Ghost Festival" texts already cited as well as
the rainmaking scriptures (T. nos. 989-993), some of the "Peacock" scriptures
(T. nos. 982-988), rites for the worship of Amitabha (T. no. 930), etc.
48. Osabe sees these ritual commentaries as indicative of Pu-k'ung's transforma-
tion of Vajrabodhi's Esoteric Buddhism into an Esoteric Buddhism of state-pro-
tection designed to serve Chinese needs. Indeed, Liang-pi (T. no. 1709 33.516b-
.228 JIABS 19.2
commentaries is Instructions for the Rites, Chants, and Meditations of the
Prajfilipliramitli Dhlirm;f Scripture for Humane Kings Who Wish to Pro-
tect Their States (T. no. 994, 19.514a-519b len-wang hu-kuopo-jo-po-lo-
mi-to ching fo-lo-ni nien-sung i-kuei, or Instructions).49 Instructions out-
lines the establishment of the maI.1c;lala / altar, the order of rites in the ritual,
and gives instructions on the contemplation of the great dhlirm;f.
Methodfor Chanting the Humane Kings Prajfili [paramitlij (T no. 995,
len-wang po-jo nien-sung fa, or Method) also outlines the ritual and
focuses on its mantra sequences. The Translation of the Humane Kings
Prajfilipliramitlij Dhlirar;,f (T. no. 996, len-wang po-jo fo-lo-ni shih) is an
exegesis of the key bfja or seed syllables of the major dhiirar;,f of the text.
The Instructions, the Method, the Translation, and Liang-pi's great
Commentary on the Scripture for Humane Kings include only partial
accounts of specific rites such as the inner visualizations of the adept. Like
rainmaking rituals and other such performances there is no exhaustive ritual
commentary dating from later Chinese use of the text. Nevertheless, when
we examine these ritual manuals with knowledge of the template drawn
from the SITS and of specific boilerplate sequences found in other late-
Tang manuals in PU-k'ung's lineage we can can "flesh out" the full perfor-
mance. Our knowledge of these ritual performances is further corroborated
both by Tang sources contemporary with PU-k'ung and by ritual manuals
preserved in the medieval Japanese Shingon and Tendai collections.
These are intimately linked to state protection. The Zuzosho, the Kakuzen-
sho, the Asabasho, the Bessonzakki, and the Byakuhokusho all preserve
medieval Japanese versions of mal)c;lalas and altar layouts for the Scripture
for Humane Kings. Although these manuals must be used with caution,
ff.) details the rites in his commentary and the continuity is striking. See Osabe
Todai mikkyoshi zakko, 89-95.
49. Though'it is attributed to Pu-k'ung the opening passages and the preface
indicate that it was the product of the master and his disciple Liang-pi. The pref-
ace was composed by Hui-ling of Pu-k'ung's Hsing-shan monastery. For the
comments in the preface see T. no. 994 19.514a6-7.
50. For these ritual commentaries see Osabe, op. cit. and De Visser, Ancient
Buddhism, vol, 1, 158-176. Toganoo, Mandara no kenkyu 370-371, treats these
51. For instance, the Piao-chih chi is replete with references to homa and abhi-
$eka performed at the Esoteric altars ofthe inner palace (nei tao-ch' ang ), at Pu-
k'ung's home monastery the Hsing-shan ssu, and at the Golden Pavilion (Chin-
k'o ssu) on Mt. For the homa rite see Michel Strickmann, "Homa in
East Asia," Fritz Staal, ed., Agni, vol. 2 (Berkeley: University of California
Press, 1982) 418-455; and Richard K. Payne, The Tantric Ritual of Japan.
they clearly reflect the overall structure and sequencing of rites known from
Tang dynasty sources.
Finally, J6kei's ninth-century Kaguraoka shidai
which preserves Shingon ritual sequences connected with the Vajradhiitu
(SITS) demonstrates a remarkable continuity with eighth-century Tang rit-
uals in Pu-k'ung's lineage. Almost certainly PU-k'ung's rituals were not as
punctilious as those found in modern day Shingon or in medieval Taimitsu
and Tomitsu manuals, but the template and the sequencing or ritual modules
are nearly identical.
Any ritual and the maI).qala and altar layouts for its performance will fol-
low the "grammar" or "template" of the SITS-the five-fold maI).qala struc-
ture and the Wheel body taxonomy-but will vary in "vocabulary," details,
divinities, and so forth depending upon its purpose. The specific configu-
ration of the mal)qala / altar-the "vocabulary," if you will-depended
upon the purpose at hand. These variations are explained in Instructions
which is divided into five unequal parts. Part one, 'The Five Bodhisattvas
Manifesting Their Awesome Virtue" discusses the relationship of the key
divinities and their Wheel Bodies to the SITS. Part two, "Rites for Con-.
structing the MaI).qala," sets out the procedures to be followed in establish-
ing the maI).qala / altar. Part three details the rites for "Entering the Ritual
Arena" (actually the order or sequence of rites comprising the ritual pro-
gram). Part four presents a 'Translation of the Phrases of the Dhiira7J.I and
the Method for its Contemplation." Part five sets out the "Visualization of
the Dhiira7J.I According to the Wheel of Characters."53
52. The ZuzoshO (Taish5 supplement vol. 3, compiled by Ejo, 1139); the
Bessonzakki (Taish5 supplement vol. 3, compiled by Shinkaku [1117-1180]);
the KakuzenshO (Taish6 supplement vol. 4, compiled by Kakuzen [1143-1218]
and his Kakuzen hitsu NinnokyohO in the same volume); the AsabashO (Taish5
vol. 9, compiled by Sh5ch5 [1205-1282]; and the Byakuhokusho (TaishB vol. 6,
compiled by Ry5zen [1258-1341]) all contain material on the Humane Kings.
These commentaries proport to be the oral traditions passed on from Chinese to
Japanese initiates. While it is obvious this later material must be used judi-
ciously, the conservative nature of the tradition, the clear continuity with Tang
ritual manuals (immediately obvious in examining Hatta Yukio's tables in Shin-
gonjiten 255-267) and the example of the ghost rites (for which we have later
Chinese ritual commentaries) means that we can use this material with some
53. Section two of Instructions details strikingly Indian procedures for setting
up the ritual arena, and with it we enter the ritual process proper, including the
purification of the ground with cow dung and cow urine. "Chakuji," Hobi5girin
3:279-280, has a discussion of these procedures.
230 JIABS 19.2
The purpose of part one is to situate the divinities of the ritual within their
proper orbits of association with the SITS. Beginning in the east -with
VajrapaI).i and proceeding to the south (Vajraratna), west (VajratIksna),
north and center (Vajrapararnita), Instructions first 'quotes the
initial description of each divinity from Pu-k'ung's translation of the Scrip-
ture for Humane Kings. It then identifies each with its "Wheel body" forms
derived from the five chief divinities of the SITS. For example, Vajrapani
is identified with Samantabhadra (P'u-hsien p'u-sa) as the Body of the Cor-
rect Teaching who eliminates subtle defilements (klda), while his Body of
Command is Trilokyavijayavajra (Chiang san-shih chin-kang) who sub-
dues the miiras and Mahesvara.
This list thus describes three possible
maI).Qala / altar deployments with Vajraparamita / Mahavairocana /
Acalavajra seated at the center. Depending on the need at hand (teaching
and purification of subtle defilements, enlightenment, or the pacification of
calamities, the subjugation of enemies, etc.) the iiciirya employs
Vajrapararnita or Acalavajra as the central divinity. 55 The associations out-
lined in part one of Instructions are presented in figure 9.
Correct SITS Function Command Function
East Va jrapfu.ll Samanta- eliminates Trilokya- Mara!
bhadra klesa vijaya Mahesvara
South Vajraratna
fulfills Km;u;lali- subdues
desires vajra asuras
West VajratIksna Mafijusrl severs Yamantaka subdues
obstacles evil
North Vajrarak$a Vajra- eliminates Vajrayak$a subdues
56 sins rak$asas
and yak$as
54. In the SITS Vajrap1iI).i, the "great rak$a" subdues Mahesvara. For a fascinat-
ing analysis of this conversion of a Hindu divinity into a Buddhist protector see
Iyanaga Nobumi, "Recits de la soumission de Mahesvara par Trilokavijaya."
55. Mahavairocana is likely not to be the central divinity of the ritual since other
rituals have total enlightenment as their primary goal, though the imagery of
visionary light is a central part of the visualization of Prajfiapararnita bodhisattva.
See below, and the original at T. no. 994, 517b20ff.
56. MikkyOdaijiten vol. 2, 676a-677b identifies this figure as "Vajra tooth" from
Amoghasiddhi's court in the first maI).qala from the STTS. He is equated with
Vajrayak$a in the Vajradhatu maI).Qala. is the name given the
Center Vajra-
Acalavajra subdues
Part two of Instructions stipulates requirements for siting and construct-
ing a altar as well as procedures for painting the divinities and
arranging objects on the altar. The sequence of rites which are strung
together to form the ritual is invariant, though particular rites may be nested
into other rites in accordance with the goal of the performance. First, the
iiciirya will decide which sort of mal)<;lala / altar to construct. Although
most rituals will involve the visualization of the maI).<;lala in the body, an
external pre-painted maI.1<;lala or a three dimensional altar mayor may not be
available. Moreover, both interior and exterior maI).<;lalas may consist of
full images in either two or three dimensions, bfja (the fundamental sonic
expressions of divinities), or samaya (the "pledge, in this case a symbolic
representations of the meditative state of divinities).57 Part two also stipu-
lates proper times and colors of vestments, and seating position of the
iiciirya which are appropriate to the divinity evoked and the purpose of the
For example, in the case of the rites outlined in Instructions the
central divinity will commonly be PrajiHiparamita bodhisattva or
The former is the Body of the Correct Teaching of Maha-
vairocana. The latter is Mahavairocana's wrathful Body of Command form.
If the rite is being performed for pacification of calamities (Siintika), the
beneficent form of this bodhisattva in the Garbha maI).Qala, where he is placed in
Vajrapani's court. His vow is to devour the causes of suffering. See Snodgrass,
The Matrix and Diamond World Mandalas 1:332. Osabe, Todai mikkyoshi
zakko 92-93, asserts that the ritual program of the Humane Kings predates the
dichotomy of the Garbhadhatu and Vajradhatu maI).<;lalas and methods. While I
agree with the spirit of Osabe's remark, putting the issue in terms of the Shingon
dual maI).Qala system is still anachronistic.
57. The four main types of maI).<;lalas are the Mahii-ma1J.q.ala, which is the
maI).Qala constructed with painted images of the deities, the Samaya-ma1J.q.ala
consisting of the symbolic forms of the deities, the Dharma-ma1J.q.ala which use
the bIja or "seed-syllables" of the deities, and the karma-ma1J.q.ala which repre-
sents the the forms of the deities in unpainted images.
58. Instructions, T. no. 994, 19.515c-516a18 for the colors, times, and direc-
tions. For types of homa see Snodgrass, 82-96. For a full discussion of con-
temporary homa (which closely match medieval Japanese manuals) see Richard
K. Payne, The Tantric Ritual oj Japan.
59. Although Instructions details the iconography ofPrajfiaparamita bodhisattva
contemporary evidence and evidence from Japanese manuals shows that Acala is
often favored.
.232 JIABS 19.2
iiciirya is dressed in white, faces north-the direction of Vajraralq;a /
Vajrayalq;a-and visualizes the focal divinity, the offerings, etc. as white in
color and he chants calmly and silently. If for increase (paus.t*a) he wears
yellow and faces Vajrapill)i / Trailokyavijaya-yellow in color---'in the east.
He chants calmly and under his breath. If the rite is for SUbjugation
(abhiciiraka) he wears black and faces Vajraratna / KUIJ.Q.ali-visualized as
black in color-in the south. He inwardly arouses great compassion and
outwardly assumes an awsome, angry demeanor and shouts out the chants.
If for attraction (vasfkarana) the iiciirya wears red and faces Vajratlksna /
Yamantaka-visualized as red in color-in the west. He chants in ajoyous
and fierce voice. These four types of rites are homa (immolation) and we
know from other sources that the shapes of homa altars also vary in accor-
dance with the purpose of the ritual. 60 If for pacification the altar is
circular. For increase the altar is square. For SUbjugation the altar is
triangular, while for attraction the altar is in the form of a lotus.
Thus, depending upon the purpose of the ritual, one generates a maIJ.Q.ala
which is simultaneously stereotypical and tailored to specific circumstances.
Throughout the performance the text (in the PU-k'ung recension) resides on
the altar and a group of monks chant it as specified in the Scripture for
Humane Kings itself.
The basic sequence of rites which compose the ritual is outlined in sec tion
three of Instructions and begins with the preparation of the iiciirya.
These rites involve the purification of body, speech, and mind, the perfor-
mance of the "pledges (samaya) of the three "departments" of Buddha,
bodhisattva, and vajra beings, and the protection of the iiciirya by donning
"armor." The next sequence of rites involves visualizing, securing and
embellishing the ritual space. Only then can the deities be welcomed to the
ritual arena where they are offered water, jeweled thrones, incense, etc.
Having and worshipped the divinities the iiciirya moves to the
visualization and identification with the chief divinity of the ritual and the
contemplation of the syllables of the great dhiiralJ.l. Finally, having
accomplished the purpose of the ritual the iiciirya performs a series of exit
rites which reverse the entry rites. Among these are a final set of offerings,
the transfer of merit to all beings, the dissolving of the ritual space and
60. On the shapes of homa altars, the first systematic appearance of rites of
"pacification," "subjugation," and "increase," see Bodhiruci's I-tzu-ting lun-wang
ching ("The Scripture of the Cakravartin of the Single-character T. no.
951, 19.261c-263b) and Strickmann's comments in "Roma in East Asia,' 434-
61. The same order is outlined in Method, T. no. 995, 19.520a-521c.
depart!lre. Homa, when it is performed, focuses on Acala and is nested
into the series of offerings after the divinities have "taken their seats," and
before the identification sequence. Some T'ang manuals carefully detail
some sequences while abbreviating others.
The ritual as it appears in sec-
tion three is translated in the following pages. I have supplied the sequence
divisions numbered 1-4.
1. Preparation of the Acarya and the Arena
If the practitioner seeks relief from calamities [santika]:
First: You must bathe and put on freshly cleaned clothes. If you are a
householder, receive the lay precepts. Caring naught for your own life [you]
should arouse the ferverent mind of the Great Vehicle seeking siddhi. Toward
numberless beings arouse the compassionate vows of the mind of salvation.
In this manner you will be able to swiftly achieve siddhi. On entering the
ritual arena do a full prostration in veneration to all the Triple Jewel throughout
the Dhannadhatu. Kneeling on the right knee, repent all transgressions of the
triple karma (body, speech, mind), request that the buddhas of ten directions
turn the wheel of the Correct Teaching, and beseech all Tathagatas to long
abide in the world. [The practitioner then says,] "All the merit that I, (insert
name) cultivate, shall be dedicated to the achievement of unsurpassed enlight-
enment. I vow that, together with all the beings of the Dharmadhatu, the sid-
dhi we seek shall quickly attain fulfillment."
Next: Assume the cross-legged position. In case there are deficiencies
that have not been cleansed, take the hands and rub [them with] incense and,
arousing the ferverent mind, make the purification mudra. With care and
humility clasp the hands like an unopened lotus blossom. Chant the mantra:
62. For instance, Instructions gives the barest indication of the exit rites (T. no.
994, 19.515c) while Method specifies each step (T. no. 995, 19.521c).
63. The specific rites which consist largely of "boilerplate" have been labled by
me and numbered 1-4 in italics. I have found it convienient to use a somewhat
different terminology and division of the sequence from that used by the
ChUinryfi of Koyasan. Instructions is representative of Tang Esoteric manuals
in Pu-k'ung's lineage. It is structured around the mudra and mantra sequences
and includes ritual instructions concerning what to do, how to make the mudras,
and other "stage" directions (in normal typeface), liturgy to be recited by the
acarya (in quotation marks) and the mantras to be chanted (boldface). Bold ital-
ics represents transliterated Sanskrit terms appearing in the Chinese. Pictures of
the mudtas can be found in Hatta's Shingonjiten under their corresponding
mantras and at the front of volume one of Mikky8daijiten. The spareness of
Instructions contrasts with the elaborate ritual laid out in the ByakuhOkushO
(Taisho supplement vol. 6, 198c-217c), though one should note that exactly the
same sequence of ritual modules is present there.
234 JIABS 19.2
Om. svabhava-suddhaI,. sarva-dharmaI,.
Chant this mantra three times. While you chant it move the mind 'to magna-
nimity and [reflect]: "All dharmas are originally pure, therefore my body is
also completely pure." Then with eyes closed visualize all the multitudes of
ritual arenas, the assemblies of buddhas and bodhisattvas that everywhere fill
the void. Hold every sort of supernal incense and with triple karma resolute
and sincere, face them to pay your respects.
One: Make the mudrii of the Buddha department samaya. The two hands
are placed before the heart, making a fist with the fingers crossed and inside,
while the thumbs are upright. Chant the mantra:
Om. jina-jik svaha 65
Silently chant this mantra three times and release the [mudra] above the head.
By making this mudra and chanting this Buddha-department samaya mantra,
all of the buddhas of the Dharmadhatu of the ten directions will assemble like
a cloud and totally fill the void. [They] empower the practitioner [who will
thus] be freed from all obstacles, and the vow cultivating the purification of the
triple karma will be swiftly accomplished.
Two: Make the mudra of the Bodhisattva depa..-tment samaya. As before
[clasp] the hands before the heart and make a fist [this time] with the left
thumb inside. Chant the mantra:
Om. alolik svaha 66
Just as before chant it three times and release [the mudra] above your head.
Because of making this mudra and chanting this bodhisattva-department
samaya mantra Kuan-yin and all the other bodhisattvas of the Dharmadhatu
of the ten directions will assemble like a cloud and totally fill the void. [They]
empowerJhe practioner [whose] triple karma [thus] becomes pure and without
any affliction. This is called bodhisattvas carrying out the vow of great com-
passion and it will cause one who seeks it to attain complete fulfillment.
Three: Make the mudra of the Vajra department samaya as in the previ-
0us mudra [but] extend the left thumb while enclosing the right thumb in the
palm. Chant the mantra:
64. Hatta, Shingon-jiten no. 1808.
65. Hatta, Shingon-jiten no. 242, the heart mantra of the "Buddha-department."
66. Hatta, Shingon-jiten no. 64, the heart mantra of the "Bodhisattva - depart-
ODJ. vajra-dh:rk svaha 67
As before chant it three times and release [the mudra] above your head.
Because of making this mudra and chanting this Vajra-department samaya
mantra, all of the Vajra [beings] of the Dharmadhatu of the ten directions will
manifest their wrathful bodies and assemble like a cloud and fill the void.
[They] empower the practioner [whose] triple karma becomes firm as dia-
mond. This is called the sages earring out the Buddha's awesome spirit.
Using the strength of their vows [they] are able to protect the state and cause it
to be without calamities, and even this insignificant body will be without
Four: (516c) Make the mudra of protecting the body. Again use the
mudras and chant the mantras of the three departments and empower the five
places-that is the two shoulders, heart, throat, top of head-and release [the
mudra] above your head. Forthwith you will be protected by stout Vajra
annor. Because of this empowerment the entire body of the practitioner glows
with an awesome radiance. All the maras who would obstruct and harass
[you] do not dare to look [at you] and they quickly flee.
Five: (516c-517a) Make the exorcism mudra and then the mudra of the
Vajra-quarter jewel-realm. Use the previous Vajra department mudra and
chanting the mantra circle the altar turning to the left. Make three circuits.
Forthwith you will be able to exorcise all the powerful maras and, as a conse-
quence of the goodness of all the buddhas and bodhisattvas, all those who are
hidden will be exposed and they will flee far from [the Buddha's] world.
Make three circuits to the right, as you like, big or small. This will complete
the Vajra-quarter jewel-realm. All the buddhas and bodhisattvas will not dis-
obey you. How much more is it true for those who would harass you, and
you will be able to obtain their expedient devices. Release the [mudra] above
your head.
2. Summoning and Feting the Divinities
Six: Make the mudra of inviting the sages to descend to the altar. Use the
previous mudras of the three departments and chant their mantras. [This time]
move your thumbs toward your body summoning them three times. Immedi-
ately the air before you will fill up with the sages of the three departments,
each going to his proper place without obstructing one another. They wait
Seven: Mudra offering agra perfumed water. As above using two hands
respectfully offer the vessel filled with perfumed water. Hold
it at eyebrow level and chant the mantra:
ODJ. vajrodaka tha hiiDJ.68
67. Hatta, Shingon-jiten no. 1090, the heart mantra of the "Vajra-departrnent."
68. Hatta, Shingon-jiten no. 1498.
236 nABS 19.2
Just as above chant it three times while moving the heart to magnanimity.
Next bathe all the sages and release [the mudrii] above your head. Because of
this agra water, during each and every stage-from the stage o( victorious
understanding and practicing of the Teaching to the stage of the Dharma-
cloud-the buddhas and bodhisattvas of the Dharmadhiitu of the ten direc-
tions all will protect you and you will obtain all their
Eight: Mudra presenting jeweled thrones. As above, with care and
humility, clasp the hands with thumbs and little fingers matching and slightly
bent. The remaining six fingers are spread and a little bent, like a lotus blossom
just opening. Chant the mantra:
Oil) kamala sviiha 69
By making this mudrii and chanting this mantra, you cause the jeweled thrones
which are presented to be received and used by the sages as though they were
real, and it causes the practitioner to reach the state of fruition 70 and to attain
the Vajra-finn jeweled-throne.
Nine: Make the mudrii of universal offering (517a-b). As above, clasp the
two hands. The five fingers are interlaced with the right pressing on the left.
Place it above the heart and chant the mantra:
Namah sarvatha kham udgate sphara himall) gagana-khall)
By making this mudrii and chanting this mantra-moving the mind to magna-
nimity-it rains all [types] of offering vessels in all of the ritual arenas of all
ocean-like assemblies of buddhas and bodhisattvas all about the Dharma-
dhiitu. On the first recitation numberless vessels are filled with incense paste
which is daubed on all the sages. On the second recitation every sort of flower
garland adorns [the sages]. On the third recitation all sorts of incense is
burned as offering. On the fourth recitation it rains superb divine food and
drink whith is properly arranged in in the jeweled vessels and offered every-
where. On the fifth recitation it rains all sorts of bejeweled lamps which are
offered before all the buddhas and bodhisattvas. Because of the strength of the
empowerment [confered by] chanting this mantra, in all the ocean-like
assemblies the offerings of incense and so on all are completely real and are
used by the sages and, as for the practitioner, he is certain to obtain
69. Hatta, Shingon-jiten no. 123.
70. The "state of fruition" or attainment indicates the completion or outcome or
attainment of the goal of practice.
71. Hatta, Shingon-jiten no. 1711.
3. Contemplation of the Chief Divinity of the Ritual
Ten: Make the fundamental mudrli of Prajiilipliramitli. Place the two hands
back to back with the index and little fingers enclosed in the palms with the
thumbs pressing on the index fingers. Place [the mudrli] above your heart and
. chant the dhliraf}.f from the scripture seven times.72 Because of making this
mudrli and chanting this dhliraf}.f the practitioner's own body is immediately
transformed into Prajiilipliramitli bodhisattva and becomes the mother of all
buddhas. The image of the bodhisattva is seated cross-legged upon a white
lotus. His body is golden colored and he has many precious necklaces adorn-
ing his body. On his head is a jeweled crown with two (pieces of) white silk
hanging down the sides. In the left hand is the Sanskrit text of the Prajiia-
pliramitli]' His right hand is held before his breast making the Dharmacakra-
mudrli; thumbs pressing on the tip of the fourth finger. Now, meditate on the
bodhisattva from head to toe. All the pores of his body emit a multi-colored
radiance which fIlls the Dharmadhlitu. Each ray transforms into countless
buddhas who fill up the void, and on behalf of the assembled beings in all
these worlds they expound the profound teaching of thePrajiilipliramitli
which causes the samlidhi of the abode of enlightened comprehension. After
the practitioner completes this contemplation release the mudrli above the head.
Grasp the prayer beads and clasp hands together and with resolute heart chant
this mantra:
0111 Vairocana mala svaha 73
Chant this three times and empower the rosary by touching it to your head. 74
Then bring it before the heart. With the left hand receiving the bead and the
right hand moving the bead, focus on union and abide in the Buddha-mother
samadhi. Contemplate it without interruption, and chant [the mantra either]
108 or 21 times. When you have finished touch the rosary to your head and
put it back in its place. Make the samadhi mudrli. Lay the hands across one
another just below the navel with the right pressing on the left. [Sit] with
upright posture, closed eyes, and head slightly inclined, and concentrate on
your heart. Visualize a bright round mirror which [expands] from one hasta
in breadth gradually [to fIll] the entire Dharmadhlitu. Set out the characters in
line revolving to the right, and contemplate them in sequence. Their effulgent
radiance shines everywhere. Proceed from the outside toward the inside until
reaching the character ti. Then go from the inside toward the outside. Grad-
ually contemplate all the characters. When you have been around once start
over again. When you reach the third repetition your mind will be quiescent
72. Section four of Instructions gives a word by word explication of the
dhlirafJf. T. no. 994, 19.518a-519a Section five gives instructions for interior
visualization of the dhliralJ.f in "wheels" of words.
73. Hatta, Shingon-jiten no. 1541.
74. Literally, "by wearing it on your head."
238 JIABS 19.2
and concentrated, and you will clearly comprehend the meaning of that which
you contemplate:, "no production, no extinction, all is the same throughout the
Dharmadhiltu. Not moving, not quiescent, meditation and wisdom are the
twin conveyance. Forever beyond all signs, this is the contemplation of
Prajftilpilramitil samildhi." Make the Prajfiaparamita mudril and chant the
dhiira1J! seven times and release [the mudriiJ above the head.
4. Exit Sequence
Next: Make the mudrii of universal offering. As previously move the mind
[to magnaminity] and follow the sequence of offerings. Before the sages dedi-
cate the merit produced to the fulfillment of all vows on behalf of the state and
the family, and for the benefit of others. Thereafter transfer [the merit] to
beings so they may tum to the Pure Land, tum to the edge of reality, tum to
seeking unsurpassed bodhi, and vow that all beings will swiftly arrive at the
other shore.
Next: Make the previous [Vajra]-dhiltu mudrii and chant the previous
mantra three times circling to the left, which will complete the dissolusion of
the [Vajra] realm.
Next: As previously, make the mudriis of the three departments and chant
the previously (used) mantras three times, all the while moving the thumbs
toward the outside. This will complete the departure, and the sages will each
return to their original land. The practitioner should make a prostration and
The modular construction of the ritual set out in Instructions is apparent
when we examine other texts of its cohort. The lineaments of this structure
have recently been laid bare in Hatta Yukio's Shingon-jiten. Hatta's com-
prehensive tables and appendices provide a basis for examination of the
underlying ritual structure and the sequencing of individual Esoteric rites in
a variety of historical contexts.
Hatta's table of rites copnected with the
Vajradhiitu (pp. 264-268) are of particular relevance to the procedures con-
cerning the Humane Kings detailed in Instructions and Methods. In this
table Hatta compares the sequencing of rites in eleven manuals connected
with the teachings of the SITS. He provides a master numbering which
represents all the possible rites which might be nested to produce a specific
ritual program. Among the manuals Hatta uses are Japanese manuals of the
ChUinryii ~ e c t of Koyasan, that attributed to Jokei (866-900) of the Kan-
jujiryu, and the Shih-pa-chih yin (T. no. 900) supposedly the oral teachings
of Pu-k'ung's disciple Hui-kuo as transmitted by Kiikai, and T'ang manuals
75. The appendices and tables run from pp. 254-339.
from the Vajrabodhi I lineage.7
This latter group includes
Vajrabodhi's version of the SITS (Chin-kang-ting yii-ch'ieh ch'ieh chung
liieh ch'u nien-sung ching T. no. 866) as well as three manuals attributed to
PU-k'ung. These are the Kuan-tzu-tsai p'u-sa ju-i-Iun nien-sung i-kuei (T.
no. 1085, lu-i-Iun), the Wu-liang-shou ju-Iai kuan-hsing kung-
yang i-kuei (T no. 930, hereafter Wu-liang-shou), and the Chin-kang-ting
lien-hua-pu hsin nien-sung i-kuei (T. no. 873, hereafter Hsin kuei).
Hatta's table clearly demonstrates the direct connection between these
manuals and later Shingon manuals. The table also demonstrates Japanese
codification of the modular structure and further elaboration on the part of
Shingon ritualists. Looking back to the Tang dynasty context, the table
makes it quite obvious that the lu-i-Iun, the Wu-liang-shou, and the Hsin
mei are closely related to Instructions (T. no. 994) and Method (T. no.
995). Indeed, Osabe has cogently argued that all of these texts are the
product of Pu-k'ung and his heirs.7
In the following table I compare the
sequences of rites in Instructions with those in the lu-i-Iun, Wu-liang-shou,
and Kaguraoka. In each case the ritual programs involve the same
sequences of rites, though some manuals abbreviate, elaborate, or even skip
certain details. All use the same mantras or variants of the same mantras.7
Finally, for our purposes, Hatta's work has one drawback. Shingonjiten is
a dictionary of mantra, and in the many cases where a sequence of standard
rites is briefly refered to without mention of the mudrii or mantra, Hatta is
silent. For instance, Hatta's tables show none of the exit rites for lu-i-Iun,
or Wu-liang-shou. Like these manuals Instructions mentions the sequences
of rites without specifying the mantra or simply notes, "use the three
mantras as before." When we examine the Wu-liang-shou and other man-
uals we find the same kind of abbreviation as in Instructions. Indeed,
when we take into account indications of rites both when they include
mantras and when they merely refer to a rite without actually transcribing a
76. J5kei's Kaguraoka shidai represents teachings on the Vajradhatu in the
Kanjujiryu tradition. The Kaguraoka shidai is the most comprehensive of the
manuals surveyed by Hatta. For the manual see Mikkyi5daijiten 230a. For
J5kei's life see Mikkyi5daijiten 1136b-c.
77. Osabe groups these manuals together under the rubric of "the Esoteric
Teachings of Pu-k'ung and his milieu." He argues that these texts (and a number
of other texts) represent the adaptation of the tantras to the Chinese scene. Osabe
also argues that they represent a joint esoterism of Vajradhatu and Garbhadiitu,
an esoterism influenced by the Susiddhikara (pp. 44-48; 89-105).
78. The numbers in parentheses which follow each rite refer to Hatta's sequenc-
ing numbers, pp. 264-268.
240 JIABS 19.2
mantra the high degree of congruence between T'ang rites and Japanese
Shingon rites is astounding. The table below presents the results of my
examination of these manuals. When a mantra or mudra indicating a corre-
spondence is found in Hatta's table I have marked it with an *. When Hatta
is silent but an abbreviated reference to sequences of rites is mentioned in
the manuals I have marked them with a **. In one case a mantra is in a text
but Hatta skips over it. In this case I have marked it with a #. The boldface
headings indicate logical breaks between sequences of rites. My division
does not completely correspond to those put forward by various Shingon
exegetes. Each of the number:s 1-12 running down the left side of the table
represents a discrete sequence of boilerplate rites. I have not included all of
the sub-rites in each. Thus, in the case of the hama sequence I have not
broken out the establishment of the homa altar, the invitation of its deities,
the offerings, and so on.
l. Worship Triple Jewel (4)
2. Purify Triple Karma (9)
3. Buddha Dept. Samaya (1020)
Bodhisattva / Lotus" " (1 I2l)
Vajra Dept. Samaya (1222)
4. Annoring the Body (13)
5. Establish Vajra Realm (35)
Summoning and Feting the Divinities
6. Offer Agra Water (80) *
Offer ThrQpes (81)
Universal Offering (90)
7. Homa Sequence
JU-i-lun Wu-liang- Kaguraoka
* *
* * *
* * *
* * *
* * *
* * *
* * *
* * *
* * *
79. In its full fonn this sequence of rites includes dispatching a chariot to bring
the divinities, welcoming them and feting them with a variety of offerings includ-
ing water, garlands of flowers, various kinds of incense, and so on.
80. For a full account of the many sub-rites see Payne, The Tantric Ritual of
Chief Divinity
8. Contemplation of
Chief Divinity (95)
Exit Sequence
9. Universal Offering (106)
Dedication of Merit (108)
10. Dissolving the Ritual Arena
Three Depts. I Departure
(110, 112)81
11. Taking Off the Armour (117-
12. Prostration and Exit (124)
* * *
** ** *
** **
** *
** ** *
What made-and still makes-Esoteric Buddhism a great missionary reli-
gion was its modular structure. If the modular structure I have detailed in
PU-k'ung's Esoteric Buddhism is representative of other lineages of Esoteric
Buddhism, as I suspect it is, then interpretation of sites like EIlora,
Aurangabad, or KongobUju becomes more difficult. This structure pro-
vided a degree of freedom and adaptability while maintaining strong ideo-
logical continuity. Even with a living tradition and ritual manuals extant
interpretation is far from clear; in the absence of these we must grope
toward an interpretation. We can look to other Esoteric sites and traditions
as Geri Malandra has so ably demonstrated, but when we do so we need to
take account of both the stereotypical and unique dimensions of any
maI).c.iala. If Rastrakuta rulers endowed these sites (a point we are as yet
uncertain of), what sorts of rituals did they wish to be performed there?
Esoteric Buddhism was connected with the periphery of the Indian world,
if it represented Indian Esoteric Buddhism for "export," and if this export
was targeted for and adapted to those in power, then what sorts of rituals
and what sorts of maI).c.ialas are these? Though an altar can be adapted to
immediate needs as they arise and the appropriate painted mar;tc.iala can be
supplied to fit the circumstances, a permanent rock-cut mar;tc.iala provides no
such flexibility. Malandra puts the problem concisely. If Ellora is a
maI).c.iala or maI).c.ialas, which is it? "If we can't name them, does that
81. This sequence in its full form involves sending off the divinities.
242 JIABS 19.2
weaken the analogy?" 82 I would like to suggest we reframe the question.
Given the modular structure I have outlined above, there are two possible
avenues of inquiry open to us. One the one hand we may be looking at a
particular instantiation of a particular ritual procedure which was,the raison
d'etre of the site. On the other hand Ellora's maI).c;lalas may be "templates"
or generic structures. In either case we should try to determine both the
ideological template of the caves as well as the sort of ritual which would fit
such a space. Looking at the ritual and social context of maI).c;lalas in
eighth-century China shows us not only how maI).c;lalas were created but
also one of the reasons why Esoteric Buddhism was a great missionary
Chen-yiian-hsin-tingshih-chiao mu-lu. Yuan-chao (d. 800) T. no. 2157.
Chin-kang-feng-ko i-cliieh yii-ch'ieh yu-chih ching. Trans. by Vajrabodhi
(671-741). T. no. 867.
Chin-kang-ting i-ch'ieh ju-lai chen-shih she ta-ch' eng hsien-ti ta chiao-
wang ching (Sarvatathagatatattvasamgraha). Trans. Pu-k'ung (705-
774). T. no. 865.
Chin-kang-ting lien-hua-pu hsin nien-sung i-kuei. Pu-k'ung (705-774). T.
no. 873.
Chin-kang-ting yii-ch'ieh chung liieh ch'u nien-sung ching
(Sarvatathiigatatattvasanigraha). Trans. by Vajrabodhi (671-741). T.
no. 866.
F 0 shuo i-ch'ieh ju-lai chen-shih she ta-ch' eng hsien ti san-mei ta chiao-
wang ching (Sarvatathagatatattvasanigraha). Trans. Shih-hu (Danapala,
fl. 980s). T. no. 882
len-wang hu-kuo po-jo-po-lo-mi-to ching. Attributed to Kumarajlva (350-
409), ca. 470-490. T. no. 245.
len-wangho-kuo po-jo-po-lo-mi-to ching. Pu-k.'ung (705-774). T. no. 246.
len-wang hu-kuo po-jo-po-lo-mi-to ching shu. Liang-pi (fl. 760's). T. no.
len-wang ho-kuo po-jo-po-lo-mi-to f o-lo-ni nien-sung i-kuei. PU-k.'ung
(705-744). T. no. 994
len-wang po-jo nien-sungfa. PU-k.'ung (705-744). T. no. 995.
len-wang po-jo fo-lo-ni shih. Pu-k.'ung (705-744). T. no. 996.
Liu-ch'u ching. Pu-k.'ung (705-744). T. no. 243.
Shih-pa kuei yin. KUkai (774-835). T. no. 900.
82. ''The Mar;4ala at Ellora / Ellora in the MaT;ejala" 191
Ta p'i-Iu-ch' e-na ch' eng fo shen-pien chia-chih ching (Mahiivairocanii-
bhisambodhi sutra). Trans. SUbhakarasi.n)ha (637-735) with the aid ofI-
hsing (683-727). T. no. 848.
Ta-fang chen-yuan hsu-liai-yuan s h i h ~ c h i a o lu. Yuan-chao (d. 800). T.
no. 2156.
Tai-tsung-ch' ao ssu-liung ta-pien-cheng kuang-chih san-tsang ho-shang
piao-chih-chi. Yuan-chao (d. 800). T. no. 2120
Wu-liang-shouju-Iai kuan-hsing kung-yang i-kuei. PU-k'ung (705-744). T.
no. 930.
cheng-fa-Iun shen
ch'eng chiu pj(;ff"t
ch'eng-chiu hsi-ti
ChUinryil $
ch'u-shih ch'eng-chiu tf:Ifftpj(;:it
Hui-kuo tI-*
I-hsing -fi
J6kei :&./ffi{.
Kaguraoka shidai .
kuan-ting rtf]}
Kukai :!Em
Liang-pi ft.
lun shen
ming-wang alEE
pen-tsun :2fs:#
PU-k'ung (chin-kang) /f:!E
Pu-tung /fib
tzu-hsing-lun shen
wei-nu shen
Y llan -chao Iillffi
Ma1).c;iala, Ma1).c;iala on the Wall:
Variations .of Usage in the Shingon School
Kayasan is the name of the mountain, as well as the small community that
occupies its peak, in Wakayama prefecture in Japan where KUkai established
an important rural monastic center devoted to the practice of esoteric
Buddhism in the early ninth century. The modern town of Kayasan is
perhaps best known, by both Japanese people and foreigners alike, as a
tourist destination. This is especially the case during the summer months
in western Japan, when the company that owns the train line that carries
people to the mountaintop town two hours south of Osaka pitches a high
visibility advertising campaign to attract visitors to the distant peak. A
recurrent theme in the poster ads stresses the temperature differential
between Osaka and Kayasan (the latter is cooler), with the poster colors
of choice invariably being hues of green and blue. The invitation is to a
respite from the sweltering heat of the urban plain into a land of cool and
natural comfort. Another aspect of the invitation is the opportunity to
visit a grand repository of cultural history as well as the burial site of
many of Japan's greatest cultural heroes. A walk through Kayasan's
Oku-no-in, the country's largest cemetary with its breathtaking paths through
canopies of ancient cryptomeria and massive moss covered gravestones,
with all the paths culminating at the mausoleum of Kukai, is frequently
billed as a glorious trek through Japanese history. Whatever their various
motivations, over one million people flock to Kayasan each year, most
during summer. But people visit during all seasons, and in the climatically
still-tolerable portions of spring and autumn (the mountain has a long,
cold winter) many will also attend one of the numerous annual festivals
there, some of which have been celebrated for over one thousand years.
While these festivals also get billed in full-size posters placed in the midst
of busy urban centers, and the modern quality of the activities surrounding
some of them-the stage sets, brightly clad kimono dancers, amplified
music and colored lights--carry the sheen of a truly polished production,
many of them maintain a core structure that appears to preserve key
elements of the religious observations practiced by the founders of the
246 JIABS 19.2
monastic complex nearly twelve huridred years ago. The town is a curious,
and I think still attractive, blend of the past and present.
It would be as egregious a misrepresentation to portray Koyasan's present
as if bereft of the glorious religious traditions of the past as it would be to
depict the town's earlier history as if it had lacked anything resembling an
element of tourism. Ever since it became a widespread belief in the tenth
century that the founder, Kukai, had never died but was in an eternal
samadhi at Oku-no-in awaiting the arrival of the future Buddha Maitreya,
the mountaintop was an active recipient of the pilgrimages of innumerable
sovereigns, aristocrats and common people. There were also vigorous
advertising campaigns in the medieval period run by the famous Koya-hijiri,
the wandering "holy men" who promised bereaving family members splen-
did rewards for their deceased loved ones in the afterlife if at least some
of the remains were buried near the "resting place" of KobO Daishi (KUkai's
posthumous title and the name by which he is still best known). Monies
collected for the service of transporting remains to the distant mountain
and caring for the burial seem at times to have been used to help maintain
the struggling community of monastics in their remote locale. Thus the
tradition developed of making the visit to Koyasan for the dual purpose of
making a pilgrimage to the site of KobO Daishi's tomb and to pay respects
to one's own ancestors or loved ones who were buried there. This tradition
is maintained strongly up to the present (the cemetary is still a major
source of income for the main temple of the Shingon school, Kongobuji,
that is located at Koyasan), and it manifests a remarkable exception to the
common exclusivity of Japanese sectarian Buddhism by virtue of the
multi-sectarian affiliation of the pilgrims. Both these two common purposes,
in addition of course to the third "purely" touristic motivation mentioned
in my opening, seem to animate visits even of people whose families are
closely affjliated with other Buddhist sects beside Shingon. Koyasan is a
complex community fed by various streams of energy flowing along mul-
tiple currents.
The modem town of Koyasan maintains an old pattern of struggling to
maintain its livelihood on a mountain top far removed from the urban
center. As part of this struggle, the images employed to ensure the tourist
trade, which represents a significant influx of the wealth that supports the
four thousand residents, rely not only upon the natural features of the
locale but on its historical and cultural virtues as well. Borrowing from
both geographical and cultural implications, one theme that has frequently
been employed to characterize the community has been that of a
The circular rim of hills that surrounds the mountaintop basin has often
been likened to the perimeter of one, as has the complex plurality of
religious practices and devotees that have comprised the community's
long history. Indeed, the term maI)qala in Japan has been popularized to
such an extent, in particular since the "mikkya (esoteric Buddhist) boom"
of the 1980sbrought on in part by the 1150th anniversary of Kukai's
passing, that it can often connote little more than an image of a melting-pot
suggesting a sense of the connectedness and equality of allparticipants.
A recent lecture series open to the public at Kayasan utilized just this
theme in its colorful brochures. But a maI)qala has had other more specific
meanings in its history at Kayasan and this paper will address the question
of how the term and its concrete expressions have been employed there.
As as well-known, the key mal).qalas in the Shingon tradition focus on
the Buddha Mahavairocana and the system of Five Buddhas (paiiGa
tathiigata) of which he is a central figure.
Since it is known that
mal).qalas focusing on Vairocana were also associated with the early found-
ing of Tantric Buddhist monasteries in Tibet at around the same time,
some of the features of the development of the monastic complex at
Kayasan might be of general interest to students of the history of Buddhist
Tantric traditions. In the case of Kayasan, and more broadly speaking the
early years of Shingon school in Japan, we have a remarkable wealth of
documents from which to learn about how a system of esoteric Buddhist
practice took institutional form. And, although there have certainly been
changes over the centuries, the contemporary religious community at
Kayasan preserves many symbolic architectural and ritual structures with
very ancient roots.
This paper will address several related issues centering on the theme of
the mal).qala in relation to Kayasan. I will tie these themes together by
referring to several of Kukai's writings on the subject. Some of the
themes addressed include the conception of the esoteric Buddhist teachings
themselves (or, perhaps more precisely, the world as envisioned in these
teachings) as being a maI)qala, the utilization of smpas for housing maI)qalas
and for representing the center of a mal).qala, and the employment of
painted and sculpted mal).qalas in various ritual contexts. Along the way I
will forward a hypothesis regarding the design of the monastic complex at
Kayasan, and furthermore will argue for the importance of additional
study of the ritual use of mal).qala representations for coming to better
1. For a recent and exhaustive study of the variety of Buddha systems in
Tantric Buddhism, utilizing materials from India, Tibet and China, see Yoritomi
Motohiro's Mikkyo butsu no kenkyu (Kyoto: H6z6kan, 1990), which contains a
lengthy English summary on pp. 691-716.
248 JIABS 19.2
understand the development of the esoteric Buddhist cults in Heian Japan.
The paper by Geri Malandra, prior to mine both in this volume and in
presentation at the conference, opens up our topic in a grand manner. It
treats us to images of an international esoteric Buddhist tradition spreading
easily throughout Asia via a variety of means. She portrays it as being a
remarkably adaptable tradition and as thus being eminently portable. Its
diffusion throughout Asia is likened to the unfolding of a maI).Qala from
center to periphery, like the petals of a flower emerging from a bud and
pushing outward. My task is to direct attention to the easternmost edge
of this unfolding in Japan where, in 806 when Kukai returned from two
years of study in the Tang, we can find a reasonably well-documented
case of the consciously directed establishment of an esoteric Buddhist
system of practice. In what appears to have been his first public pro-
nouncement of the unique qualities of what he was transmitting from
Ch'ang-an, Kukai borrowed the words of his master Hui-kuo, which indeed
depict his lineage in floral (or at least botanical) imagery: "In the personal
transmission of this teaching, from the Buddha's Body of Truth down to
my master Pu-k'ung, there are six leaves.,,2 The Shingon tradition counts
Hui-kuo as the seventh and Kukai as the eighth patriarch. Were we to
add one more, we could imagine the tradition representing its lineage as
something akin to the eight red petals surrounding the center of the Garb-
hakosadhatu (Womb World) maI).Qala, wherein resides the Buddha Mahavai-
Within the thirty years since Kukai's return from China (until his death
in 835), he was instrumental in establishing four major centers of esoteric
practice in Japan. These were at the T6ji temple and the Imperial Palace
in Kyoto, ,:"at the T6daiji temple in Nara and at K6yasan. He was also
instrumental in setting up numerous other centers for esoteric practice in
Nara, Kyoto and probably other localities as well. In the span of one
generation Japan acquired a unified network of such establishments, many
of them managed by Kukai's students, which represented a new unfolding
at the periphery of the Asian Buddhist world. As is well-known, this
process only accelerated after Kukai's death such that even in the "new"
Buddhism of the Kamakura period (12th to 14th c.) we find not only
2. KobOdaishi chosaku zenshil, ed. by Katsumata Shunky6, vol. 3 (Tokyo:
Sankib6 1973) 390. This standard collection of Kl1kai's writings will henceforth
be referred to as KCZ, followed by a colon and then the volume and page
numbers, so that the present citation is KCZ 3: 390.
widespread esoteric practice but also the incorporation of various elements
of its tradition, such as the malJ.<;iala, manifesting in the Pure Land and
even Nichiren traditions in the visual forms of malJ.<;ialas depicting the
Buddha Amida or the sacred phrase in praise of the title of the Lotus
Satra, the gohonzon (sacred image, literally, "honorable main deity") of
"Namu myohO renge kyo." Historians of Japanese religion speak of the
permeation of esoteric patterns of thought as the "esotericization" (mikkyo
ka) of religious practice and theory at various levels.
Early on in this development (in the 9th and 10th centuries) the Tendai
school was more aggressive and successful in this regard than was the
Shingon school. This was probably because its founder SaichO was not
well versed in esoteric Buddhism, left very few writings on it, and so his
disciples had to build a workable program on their own. Incidentally,
although the Tendai sect afer Kukai's death (Saicho died several years
earlier) is often portrayed in Shingon scholarship as competitively trying
to "catch up" during the late 9th century due to the fairly complete system
that Kukai had apparently transmitted to his students, there were surely
other more internally driven mechanisms that propelled the propitious
growth of Tendai esoteric Buddhist practice. Despite the Shingon rhetoric
that hails KUkai's accomplishments as the only real core of esoteric Bud-
dhism in Japan, the subsequent journeys to China by Tendai and Shingon
monks alike and their introduction of numerous new texts and practices
speaks clearly of the multifaceted character of esoteric Buddhism in Japan
as well as of its vitality in China at the time (after Hui-kuo's death).
There can be little doubt, however, that Kukai was instrumental in the
founding of the Shingon sect, and that his presence in Kyoto also greatly
stimulated the development of Tendai esoteric practice. It is no secret
that Saicho was eager to learn all that he could from Kukai and that he
sent many of his students to study under him. And for a while, at least,
both men even appeared interested in collaborating in the building of new
forms of esoteric practice in Japan.
Some types of esoteric-style practice
existed already in Nara, but the great interest shown in Kukai's new
synthesis, which was part of a general vogue for anything from the Tang
capital, clearly spurred the efforts of both men's creative energies. Although
Kukai brought to Japan texts (especially translations by Amoghavajra),
3. See for example Furuta Shokin, Nihon Bukkyo shisoshi (Tokyo: Kadokawa
shoten, 1961) ch. 5, 139-65.
4. For a reevaluation of the traditional view that their relationship was
primarily a bitter one, see my doctoral dissertation, "KUkai and the Beginnings
of Shingon Buddhism in Japan," Stanford University, 1994, 194-224.
250 JIABS 19.2
liturgical paraphernalia (such as vajras and mat;lQalas) and forms of ritual
(such as advanced kanjo,Skt. consecration rites) that Were
entirely new to Japan, some forms of Buddhist practice transmitted from
the continent that were already referred to as esoteric were c1early not
The extent to which such practices existed is not well documented
in English, nor have they always been given due attention in Japanese
studies of KUkai. Yoshito Hakeda's oversight in this regard resulted in his
statement that the ftrst esoteric Buddhist ceremony in Japan was conducted
in 805 by Saich6, just after returning from. China, at the behest of the
ailing Emperor Kammu.
Both Kukai and Saich6 were familiar with
5. On the elements of Kfikai's practice that were new to Japan, including
some Sanskrit texts and the five-pointed vajra, see Kushida Ry5k5, Shingon
mikkyo seiritsu katei no kenkyu (Tokyo: Sankib5, 1981) 36-42. As for esoteric
practices during the Nara period, examples of leading research in this area are
Kushida Ry5k5, Shingon mikkyo seiritsu katei no kenkyu (Tokyo: Sankib5,
1981); Ishida M5saku, Shakyo yori mitaru NarachO bukkyo no kenkyu (Tokyo:
Toyo Bunko, 1930) (especially 146-59); Horiike Shunp5, "Nara jidai bukky5
no mikky5teki seikaku," Kukai, Nihon meis5 ronshU 3, ed. Wada ShUj5 and
Takagi Shingen (Tokyo: Yoshikawa K5bunkan, 1982) 22-39 (originally pub-
lished in Nishida sensei koju kin en nihon kodai shi ronso); Hayami Tasuku,
Heian kizoku shakai to bukkyo (Tokyo: Yoshikawa K5bunkan, 1975) 1-33,
and also his Jujutsu shukyo no sekai (Tokyo: K5shob5, 1987) 25-40; Misaki
Ry5shU, "Narajidai no mikky5 ni okeru shomondai," Nanto bukkyo 22 (1968):
55-73; Katsumata Shunky5, Mikkyo no Nihonteki tenkai (Tokyo: Shunjfisha,
1989) 6-13, and Sawa RyUken, Nihon mikkyo -sono tenkai to bijitsu (Tokyo:
Nihon H5s5 Ky5kai, 1966) 30-87. For a "state of the field" account that
includes a review of several of these articles and arguments of its own, see
Miyagi Y5ichir5, "Nara jidai no mikky5 ni kansuru ichi k5satsu," Mikkyogaku
kenkyu 18 (1986): 75-91. This isa growing area of research as evidenced in
the listing under a recently published bibliography of ftfty-eight articles and
books (aUq5ublished between 1985-89) under the heading of "Nara period
esoteric Buddhism." See Takeuchi K5zen, "Mikky5 kankei bunken moku-
roku-ChUgoku, Nihon-hen," Koyasan daigaku mikkyo bunka kenkyusho kiyo
(1991) 136-38.
6. Hakeda, Yoshito, Kukai: Major Works (New York: Columbia University
Press, 1972) 37. While on the one hand such a statement reflects a traditional
bias against recognizing the precursors of Kfikai's Shingon in Japan, on the
other it can be seen as preserving part of Kfikai's emphasis on the disparity
between the existing practices and his own. Stanley Weinstein noted that
Hakeda may have had in mind the earliest instance of a kanjo ceremony, of
which indeed the case he mentions of Saich5's performance would probably
have been the ftrst. See Weinstein's review article of Hakeda's book, "The
Beginnings of Esoteric Buddhism in Japan: The Neglected Tendai Tradition,"
The Journal of Asian Studies, 34.1 (1974): 185. It also appears that although
esoteric practices and texts before they went to China and their importation
of new texts and practices, as well as the tremendous influence these
came to have, is only conceivable in light of the knowledge of and interest
in esoteric Buddhism possessed by each man prior to their travels.?
Nevertheless, it is clear that Kukai brought something new even though
it is hard to say for certain how much of his systematization relied upon
Chinese precedents and how much was due to his own creative synthesis.
Testament to this is the immediate interest shown in his system by SaichO
and the rapid growth of new centers of practice during his lifetime. Still,
it is clear that he felt a need to distinguish his new texts and practices
there are references to kanjo in texts in Japan prior to this time, the term may
have had an altogether different meaning denoting certain flags used in rituals.
On sources designating Saicho's as the first kanjo in Japan as well as on the
meaning of the term in Nara texts, see Paul Groner, Saicho The Establishment
of the Tendai School (Berkeley: 1984) 66, fn. 7.
7. From catalogues recording the copying of scriptures we can know that of
the nearly six-hundred texts classified as related to esoteric Buddhism in the
modem Taisho canon, about one quarter (at least one hundred thirty) existed in
Japan during the Nara period. This percentage is more significant than it sounds
because so many of these six-hundred texts post-date the Nara period. Of texts
translated (or composed) prior to the new tradition marked by Shan wu-wei
(born in India as Subhakarasimha, 637-735), who translated the Mahiivairocana-
sutra (Ta-jih-ching ) and the Susiddhikara-sutra (Su-hsi-ti-chieh-lo-ching),
almost all were on hand in Nara. In fact, a testament to the effectiveness with
which the Japanese were obtaining materials from the continent is the fact that
the Ta-jih-ching was copied in Japan as early in 737, just twelve years after its
translation in China. As for texts translated by Subhakarasimha, Vajrabodhi
(Chin-kang-chih, 670-741) and Amoghavajra (Pu-k'ung, 705-74), which repre-
sented a new type of esoteric Buddhism (characterized within the Shingon
tradition as being more concerned with the goal of enlightenment than earlier
forms), there were only a very small portion available in Nara. It is noteworthy,
however, that the three texts central to the systematized esoteric Buddhism of
the Heian period, the Ta-jih-ching, Chin-kang-ting-ching (Vajrabodhi's transla-
tion) and the Su-hsi-ti-chieh-lo-ching were all available. In fact the first two of
these three, which were the key texts in the Shingon school and are related to
the two main maI;u:j.ala, were even frequently copied and lectured on together
during the Nara period, apparently as a pair (see Kushida, Shingon mikkyo 4,
21-22). Thus in terms of texts, pre-Subhakarasimha materials had been very
well transmitted and there was a smattering of newer texts also available, some
of which were apparently in good use.
8. For a critical perspective on some issues related to our understanding of
Chinese esoteric Buddhism, see Charles Orzech, "Seeing Chen-yen Buddhism:
Traditional Scholarship and the Vajrayana in China," History of Religions 29.2
(November 1989): 87-114.
252 JIABS 19.2
from ones already existing in Japan. For example, in his first formal
declaration of what he brought back from China, the Shorai mokuroku
(Catalogue of Imported Items, an extant copy of which is thought to be
SaichO's hand), Kukai writes that "within those teachings considered es-
oteric, there is [the difference between] source and tributary. The former
transmitters of the Dharma [merely 1 tugged at leaves and swam in tributaries,
but what I now transmit unearths the root itself and fully exposes the very
wellspring.,,9 In an effort to carve a niche for his system of esoteric
Buddhism in the world of early Heian Japan, it needed to be differentiated
on the levels of both doctrine and practice. It appears that the mal)<;lala
was a dc;vice eminently suited to just this purpose.
Flexibility (Multivalence) of the Concept of MalJcjala
In order to understand what was new in the forms of esoteric Buddhism
Kukai brought from China, it is important to look both at his own s e l f ~
congratulatory rhetoric regarding the superiority of his transmissions (his
theory), as well as at what new developments he accomplished within
actual institutional settings (his practice). In the present study the focus is
on the latter but it is worth reiterating the degree to which all the energies
he expended in the former stand as clear indication of the imperative need
he must have felt to differentiate Shingon from other forms of Buddhist
theory and practice in Japan. And, it is most fruitful to consider the two
cornerstones of his doctrinal edifice-the theory that Shingon scriptures
and practices are based upon the preaching of the Dharmakaya Buddha
(hosshin seppo) and the claim that these same practices can enable one to
attain Buddhahood in this very lifetime (sokushin Jobutsu )-in light of the
classificatory wedge he was trying to insert between Shingon practice and
not only the "exoteric" Buddhist schools but also the existing esoteric
practices in Japan, to which he refers in his writings as "shallow" esoter-
.. 10
At the levels of both theory and practice, the mal)<;lala is one item that
served to distinguish the Shingon school from all others. Whether consid-
9. KCZ 2: 14.
10. Later Shingon exegetes of course label these earlier forms of practice
"mixed esoterism" (zomitsu), but there is no evidence that Kukai or anyone at
his time was using such a phrase. Kukai clarifies at different places in his
works that by "shallow" esoteric Buddhism he refers to the practice of dharii1J!
recitation that derives from teachings by Sakyamuni Buddha.
ering maI).Q.ala as a philosophical/religious concept or as a material
support for (or object of worship in) ritual practice, the appearance in
Japan of the new maI).Q.alas centered on Dainichi Nyorai (Mahavairocana)
was a novel presence indeed. Saicha had brought back some maI).Q.alas
associated with esoteric Buddhist practice when he returned from China
nearly two years earlier than Kiikai, but he did not bring the Vajra and
Womb Realm representations nor a doctrinal system describing their sig-
nificance. These mat).Q.alas played a central role in ilie abhigekha ceremonies
Kiikai performed (which were well-attended by SaicM and his disciples
as well as scores of monks from Nara) and in the doctrinal system he
presented as the foundation for Shingon practice. They also played an
important conceptual role in the layout of the Kongabuji ("Vajra Peak")
temple complex atop Kayasan.
The significance of the maI).Q.ala is attested in a document written by
Kiikai in 818 on the occasion of the consecration of the ritual space on top
of Mount Koya. His votive account of the establishment of this center,
whereby "the secret maI).Q.ala was transmitted to J ambudvlpa," contains
the following passage:
Mahavairocana Buddha, the great compassionate one, enjoying for himself
the taste of the equality that is enlightenment, was saddened by the plight of
the beings in the six realms of rebirth. And so it was that the thunder of his
wisdom that is one with reality trembled throughout His dharma-realm palace
and the secret maI).Qala was thereby transmitted to Jambudvlpa [our world].
It was passed on to Vajrasattva, to Nagarjuna and down to the present without
a break in continuity. Vajrabodhi and Amoghavajra came east, bearing their
staffs from India to China, and transmitted the esoteric teaching in order to
liberate all beings. And yet, here across the broad ocean in Japan, worthy
vessels of this teaching had yet to appear and so the teaching remained
hidden in the secret palace of Mahavairocana without being transmitted to
our land.
Fortunately, due to the power of the grace of the Buddha and other forces,
hidden as well as visible, that mature beings for spiritual work, I was able to
travel to T'ang China in 804, whence I safely returned with the two maJ.).Qalas
11. See Kushida RyBkB as cited in fn. 5 on what was new among Kukai's
importations. On what SaichO brought back with him, see the list of esoteric
Buddhist paintings imported by each of the celebrated "eight monks who went
to T'ang" (all from the Tendai and Shingon schools) provided by Yoritomi
Motohiro in Mandara no uchii (Tokyo: ShUeisha, 1988) 37. It is important to
note here that while the material maI).Qalas that Kukai brought from China were
clearly not of his own design, the theoretical system that tied them together
into a structured practice may well have contained elements of his creation.
254 JIABS 19.2
of the Womb Realm of Great Compassion and Vajra Realm and over one
hundred scrolls. of Vajrayana scriptures. Still, people in Japan were not
ready and the time was not right. Month after month passed in rapid succession
and it is now more than twelve years since I returned. At last, since our
devout Sovereign has taken it upon himself to help promote this teaching, we
are in need simply of a place for the practice. I have searched far and wide
and through divination have decided on Mount Koya. The Sovereign, deter-
mined to spread this teaching, has granted the mountain for this purpose.
Thus will a monastic complex be constructed on this land bestowed by the
So it is that now I will promote these esoteric teachings in order to return
favor to the Buddhas above and to liberate beings here below, as well as to
augment the dignity of the various beneficient spirits. Thus, in accordance
with the esoteric Buddhist teachings of the Vajrayana, I will establish in this
space the two maQ.<;I.alas of the Vajra and Womb Realms. Mayall the Buddhas
rejoice in, and all the heavenly beings protect, my efforts here and may all
virtuous spirits please vow to help realize our wishes.
The term maQ.<;lala is used here in several ways. First, there is mention
of the my tho-historical beginnings of the Shingon lineage in the initial
transmission of the esoteric teaching from Mahavairocana to Vajrasattva.
Here the maf.\<;lala indicates the esoteric teaching as a whole. Next, there
is reference to the scroll paintings of the two Realms that Kukai transported
from China. Lastly, he states that he will establish the maf.\<;lalas of the
two Realms at Kongobuji. It is of course possible that by "establishing
the two maf.\<;lalas" Kukai refers to setting up a ritual space within a
temple where painted scrolls depicting the maf.\<;lala will be hung. But a
more suitable interpretation is to understand him as saying that he is
establishing Kongobuji as a center of two greater, all-encompassing
maf.\<;lalas that represent the esoteric Buddhist concept of the entire world
as a diviQe assembly emanating from the Buddha. This is a common
usage of the term maf.\<;lala in Kukai's writings and its significance is
revealed in the title of his famous essay on the stages of religious develop-
ment, the Himitsu mandara Jiijiishinron (Treatise on the Ten Levels of
Mind of the Secret Ma1Jl!.ala). The title of this work is frequently abbreviated
12. KCZ 3: 392-94. See also Allan Grapard's partial translation of the same
text in his "Flying Mountains and Walkers of Emptiness: Toward a Definition
of Sacred Space in Japanese Religions," History of Religions 20.3 (1981): 203.
Grapard translates a section regarding the sacralization of the monastery grounds
that includes the ordering away of maleficient powers from, and the invitation
of all beneficient ones into, the newly established boundaries of the sacred
territory. He also comments on how similar this text is to Shinto texts that are
also used for demarcating sacred space.
resulting in the omission of the reference to the secret maI).<;iala. The full
title expresses the view that all religious teachings, not only esoteric
Buddhist ones, are part of the "secret maI).<;iala" of the Buddha and as such
serve to liberate living beings. The esoteric maI).<;iala established at Koyasan
is one part of this broader one (and clearly the most soteriologically
effective part from Kukai's perspective). Regarding this last reference,
there is another text that Kukai wrote five years later, at a time when he
was seeking patrons to contribute to the construction of smpas at Kongobuji,
in which the following explanation is given:
Thus it is that in the near future, in order to liberate beings and out of
gratitude to our four great benefactors [parents, king, sentient beings, the
Three jewels], I will establish at Kong5buji two smpas that represent Vairocana
as the Essential Nature of the Dharmadhatu, in addition to two mandalas of
13 ..
the Womb and Vajra Realms.
In other words, he is not only consecrating the grounds of the monastic
complex; he is also designating these grounds as the center of a sacred
cosmos, as a place of religious practice where this world becomes trans-
formed, through the power of the Buddhas and the practitioner's mystical
identification with them, into a realm of perfection. In Shingon parlance,
this perfection is depicted as partaking of the interlocking spheres of
compassion and wisdom (and other pairs of qualities) that are visually
represented in the painted maI).<;ialas. This is a traditional understanding
of what it means to "establish" the maI).<;iala.
In the last passage cited, Kukai notes that he intends to establish two
smpas at Kongobuji.
These smpas are traditionally recognized as centers
13. KCZ 3: 366.
14. The maI).<;ialas of the two realms are said to depict the complementarity,
or non-duality, of a variety of concepts or realities. The most common expression
for describing the pair in this regard is to see the Womb Realm as representing
"principle" (n) and the Vajra Realm "wisdom" (chi). Hence the oft-cited phrase
"the non-duality of principle and wisdom" (richi fun i) to describe the interpene-
tration of these realms. For a treatment of the various pairs of terms represented
by the two maI).<;ialas, see Adrian Snodgrass, The Matrix and Diamond World
Mandalas in Shingon Buddhism (New Delhi: Aditya Prakashan, 1988) 124-4l.
15. I use the term smpa rather than the more common term pagoda (to refer
to the Japanese architectural form corresponding to the South Asian stUpa)
because the Sino-Japanese term makes no distinction between these two and is
in fact derived from a transliteration of the word stiIpa. Furthermore, the form
of the present main stUpa at Kong5buji preserves in part the hemispheric shape
so characteristic of the Indian stiIpa but often missing in its East Asian represen-
256 JIABS 19.2
of the Womb and Vajra Realm mal)<;lalas. As the center of a mal)<;lala,
each also symbolizes the Dharmakaya Buddha Mahavairocana, understood
in this particular context as being the "essential nature of the dhannadhiitU'
(hokkai taisho), the very essence of that which pervades all existence. In
fact in the paintings of the symbolic samiiya mar:ujala, where the various
deities are represented by sacred objects that express their vows to liberate
beings, the Dharmakaya Buddha is represented by a srupa. I will have
more to say below on the relation between the mal)<;lala and stupa, the
concept of mal)<;lala as a sacred cosmos and on K6yasan as an axis or
center for the practice that is conducive to the realization of such a world.
For now it will suffice to conclude this section by recalling that the
introduction of various levels of theory and practice related to esoteric
Buddhist mal)<;lalas was one of the most conspicuous contributions that
Kl1kai made to the religious landscape of early Heian Japan.
The Layout of the Kongobuji Complex
For the modern visitor to K6yasan, the center of the monastic complex
(or garan, as it is known in Japanese) would seem to be a single srupa
rather than two.
This large structure is called the "fundamental great
stl1pa" (konpon daito, hereafter Great Srupa) and, as older drawings of
K6yasan indicate, it seems to have stood out for centuries as a central
piece of the landscape.
It has become such a key symbol for the
community of K6yasan itself that photographs of its ringed spire, more
often than not shown piercing a cover of low-lying mist in the midst of
the surrounding rim of green peaks, frequently grace the covers of tourbooks
and pamphlets. It stands on the eastern side of the center of the Garan
and is mirrored even today by another stl1pa on the western side [See
Diagram 1]. This second srupa in the Garan is known simply as the
Western SJupa (saito), and is notably smaller than the Great Stl1pa and
slightly r;moved from the center of ritual activities (monastic as well as
lay) in the Garan that revolve around the Great Srupa. Although there are
no architectural plans dating from the founding of Kong6bl1ji to confirm
the design, most scholars agree that the original conception likely contained
both stl1pas as equally prominent focii of the Garan. This is suggested by
such evidence as the relatively symmetrical placement of the two stl1pas
tatives. The shape is known as the Tah6-tO (Prabhiitaratna-stiipa)
16. The term garan is derived from the Sanskrit word for the residence for
members of the Buddhist order, san:zghiiriima.
17. The present structure is a reconstruction dating from 1937. It stands over
48 meters in height with a square base of over 23 meters on each side.
with regard to what was once the entrance to the Garan, Kukai's writing
testifying to the plan to erect two smpas and, lastly, the convention of
such a double smpa layout in the major monastic complexes in Nara. It is
uncertain, however, whether both smpas were actually ever constructed
on an equal scale. It has been hypothesized by some scholars that the
difficulties inherent in erecting two large stupas at the rural mountain site
prevented this from ever taking place. Thus the reason why one smpa has
corne to play such a central role may have been an expediency due to
economic and geographic conditions.
A more popular interpretation,
18. Indications of the difficulties involved in constructing a stupa can be
found in KUkai's letters of request for aid in the process. One such letter on the
occasion of planning the smpas for Kong6buji was cited above (fn. 40). The
letter also contains the following plea:
We now have many workers engaged in this effort but food supplies are in
shortage. What 1 wish for is that persons of whatever status, whether rich
or poor, lay or cleric, would unite [with us in our purpose]. When it is said
that a huge mountain can be made by innumerable contributions of a single
grain of dust, or that the great ocean is created by the joining of many
drops of water, it is only made possible by the sharing of a common
purpose and the collective union of energies.
And so do I humbly request, that all the patrons contribute anything, even
as little as a single cent or a grain of rice, in order to assist us in this
virtuous task. If you will do so, our work will surely be accomplished in
no time at all and the merit thus produced will last for thousands of kalpas.
Another example is in a similar letter drafted when he was seeking help erecting
a smpa at the T6ji temple in Kyoto. This temple's construction had been
languishing for many years when it was placed under Kukai's supervision in
824. An excerpt from the letter reads:
By some mistake the Sovereign's wish [to complete construction of this
temple] has fallen upon me instead of upon skilled artisans. I run about
day and night, in the east and in the west, supervising the work. The
lumber for the smpa has now been secured in the nearby eastern hills.
Together, monks and layman have been hauling timber ... but since the
trees are large and our strength is insufficient it is an extremely difficult
task. It reminds me of [the story in Chuang-tzu about] the praying mantis
who had to push a cart or the mosquito who tried to shoulder a mountain.
The letter continues by requesting that the Sovereign assign court officers,
even of the highest rank (!), to assist in the labor. See KCZ 3: 374-76. If help
was needed even in the capital, the project at K6yasan must have been all the
more burdensome.
258 JIABS 19.2
based primarily on the appearance of a central srnpa similar to the Great
Srnpa at other Shingon monastic complexes in Japan, is that the sing1e-srupa
design was the original one. I will comment more on this theory later and
for now say only that it ignores the evidence to the contrary listed above.
In reliance on this evidence, the view presented here as most plausible is
that the original plan was for two srnpas.
The present-day Garan at Koyasan appears at a first look to be somewhat
of a sprawling assemblage of buildings without symmetry or center. This
is due to the gradual accretion, over the centuries, of structures adjacent
to, and primarily in the eastern direction of, the center of the Garan,
which is still today separated by stone steps on all four sides leading up to
the elevated plateau where the Great StUpa and the Western StUpa are
located. The "Garan proper," by which I mean the section that became
the center, includes three main buildings: the two stUpas and a Lecture
Hall (kodo).19 There was once a large covered gate from which one
could enter the Garan from the south side (known as the nanmon, or
southern gate) but only the stone foundations remain today. Upon entering
from where the southern gate used to be, one comes first upon the stately
Lecture Hall, behind which are placed more or less symmetrically the
Great StUpa to the east (and right) and the Western StiIpa (smaller and to
the left). Before focusing attention on the stiIpas themselves, it will be
helpful to comment on the basic layout of these three main buildings.
First of all, based on the model of ancient Chinese imperial palaces,
Buddhist temple complexes in China and Japan commonly placed the
entrance gate on the south side, with the central structure, and thus its
main Buddha images (like the Emperor), facing south toward those ap-
proaching from the entrance. This layout was very common in Nara and
seems to hflve been based on T'ang models. Kongobuji was no exception.
It is unclear, however, if KiIkai-or his disciples since the complex was
not completed during his lifetime-had a particular temple in mind as a
model, whether from a Chinese or Japanese precedent, upon which he
based its layout. Nevertheless, when compared with other Buddhist com-
plexes in Nara, the Garan layout at Koyasan has both significant similarities
19. Today there are actually five main structures in all, but I exclude from the
present discussion the Mied6, which enshrines a famous portrait of Kukai, and
the large shrine to the local deity that stands west of the Lecture Hall. In spite
of the significance of these buildings for ritual purposes, the focus is limited
here to those buildings with the most explicit characteristics of esoteric Buddhist
doctrine. It is thought that the Mied6 may have originally housed the practitioners
of the community.
as well as differences.
While there are certainly grounds for thinking that Kl1kai made unique
contributions to Japanese Garan design via his plan for Kongobuji, it is
nevertheless. important to recognize that there were also key continuities
with local precedents. Secondary literature occasionally refers to his
contribution as a revolutionary remaking of Garan layout, but the limitations
of his contributions must be clarified. Beginning with thesmpas, there
was clearly nothing revolutionary about placing two stl1pas on opposite
sides of a central north-south axis. Archeological evidence evinces this
as a common pattern throughout Nara in the eighth century. The correlation
of these smpas with the mru.H;lalas of the Womb and Vajra Realms is, of
course, a separate matter. This could not possibly have occured in Japan
prior to Kl1kai's efforts because the maI:1cjalas did not exist until he brought
them from China. Thus this correlation may be treated as a doctrinal
innovation superimposed on an already existent architectural convention.
Regarding the placement of smpas within Garan complexes in Nara, it
appears that there was an evolution from one-stl1pa to two-smpa layouts.
Many smpas were regarded as reliquaries, and documents from the Nara
period reveal that relics were indeed often buried underneath, or on occasion
stored near the top of, the central pillar. Some scholars believe that the
appearance of two smpas in the complexes was likely due to aesthetic
reasons, with two structures offering a more symmetrical and decorative
layout, and that this development reflects a diminution in the overall
significance of the smpa within the sacred space of the Garan. Whatever
the reason, several major Nara temple complexes in Kl1kai's day-including
TOdaiji, Yakushiji and Daianji-contained two smpas.
The temple complexes also frequently contained two other halls: a
Golden Hall and a Lecture Hall, and this holds for both the one- and
two-smpa Garan layouts. Furthermore, there was a pattern common to
many Garan, which was that upon entering the entrance gate (or on occasion
just before entering) one would first encounter the smpa or stl1pas, then
20. For more on this, including helpful diagrams and photographs of three
dimensional reconstructions based on archeological findings, see Nara no jiin
to Tempyo chokoku, vol. 3 of the series Genshoku Nihon no bujitsu, 1st ed.
(Tokyo: Shogakkan, 1966) 142-53. The author of this portion of the book,
Asano Kiyoshi, notes that during the Nara period stupas gradually lost their
central position, which they had in the earlier Asuka period when their location
suggested that their "ranking" was at least equivalent to that of the Kondo. He
states that there seems to have been a tendency for stllpas to shift away from a
central position and for the Kondo to become more of a focus.
.260 nABS 19.2
21 .
the Golden (or Buddha) Hall and finally the Lecture Hall. The Golden
Hall is usually seen as the central structure of any complex and it traditionally
housed cast or sculpted images of Buddhas and other deities. These were
all placed on a raised platform known as a Shumi-dan or "Sumeru altar,"
which was commonly very large and took up much of the floor space
inside the building. This hall was riot designed for large gatherings, since
there was often very little open space, but rather for worship. The Lecture
Hall was frequently larger and provided space for religious talks as well
as other ceremonies such as the recitation of sutras. These same functions
were available within the Kongobuji complex but the "division of labor"
among the buildings there is rendered different because the Golden Hall
was no longer present.
The Kongobuji complex presents three major features that appear to
some degree to be unique. First of ail, although not mentioned above,
there is the marked absence of an enclosed walkway surrounding the
buildings. It can be supposed that financial considerations as well as
early estimates of the expected number of residents could easily have
rendered this structure unnecessary atop of Mount Koya. Secondly, there
is the equally conspicuous absence of the Golden Hall. And third, there
is the placement of the pair of stl1pas at the back (north end) of the
complex. As a pair, smpas were most commonly located closer to the
entrance before the Golden Hall. In some instances, such as at Horyuji in
Nara, the smpa and the Golden Hall were adjacent (that is, on the same
east-west axis). But the arrangement at Kongobuji would seem to be
Leaving aside the absence of an enclosure, the placement of structures
at Kongobuji poses interpretive challenges. I have yet to encounter any
attempts in scholarship to put these pieces together. It would appear that
there has ~ e e n some kind of reversal in arrangement, such that the Lecture
Hall comes first and the smpas last, but I can forward no explanation for
this. What does seem clear, however, is that although there is no Golden
Hall, the smpas seem clearly to have appropriated the role of this hall by
housing the key images (representing the two maI).qalas fundamental to
Shingon practice) in a hall dedicated to worship. For one of the most
remarkable characteristics of the smpas at Kongobuji is that the interiors
were fully furnished with images, on a raised platform, in a manner most
commonly reserved for a Golden Hall. Thus both the position and the
interior of the Kongobuji smpas suggests that their relative status in
21. The most common name for this hall is Golden Hall (kondo) but it is also
known as the Buddha Hall (butsuden or butsudo).
terms of sacred space seems to have been elevated over Nara counterparts.
I cited earlier a passage from one of Kl1kai's writings describing the
establishment of the Kongobuji complex that states, "I will establish at
Kongobuji two stl1pas that represent Vairocana as the Essential Nature of
the Dharmadhatu, in addition to two maw;lalas of the Womb and Vajra
Realms." There has been some scholarly debate over the centuries as to
how to interpret this statement. At issue is whether the statement means
that the stl1pas and the maI).qalas are "established" separately so that the
stl1pas, on the one hand, represent "Vairocana as the Essential Nature of
the Dharmadhatu," while the maI).qalas alone represent the two Realms.
One historical factor that has led to understanding the stl1pas themselves
as not each representing one of the maI).qala realms is that for several
centuries the Great Stl1pa at Kongobuji has not been a representation of
one of the maI).qalas but of both of the maI).qalas. This is also the case
with the central stl1pa at some other Shingon complexes, such as at Daigoji
in Kyoto. In both Daigoji's five-storied stl1pa and the Great Stl1pa at
Kongobuji, the iconognwhy clearly represents both the Vajra and the
Womb Realms at once.
At Kongobuji, on the central platform within
the Great Stl1pa are five Buddha images. At the center is Mahavairocana
of the Womb Realm (it is the same deity as Mahavairocana of the Vajra
Realm but is differentiated by mudrii), and he is surrounded by four
Buddhas of the Vajra Realm.24 The interior layout of the stl1pa at
Daigoji is somewhat different, but the combining of both realms is never-
theless an unmistakable feature there. Although at Daigoji there is just
22. The mere presence of Buddha statues inside a stl1pa may not have been
an innovation at Kongobuji. Two other temples from the Nara period contain
statues, Saidaiji (wooden) and Horyiiji (plaster). Some scholars have expressed
surprise that, considering the importance of stiipas in the garan ofNara temples,
it is surprising that we have only two extant cases where Buddha statues are
found. See Asano Kiyoshi's comments in Nara no jiin to Tempyi5 chi5koku 96,
section on the four statues at Saidaiji.
23. The interior layout of the stiipa at Daigoji, which is very different from
that of the Great Stiipa, is represented by a diagram in Sawa Takaaki, Art in
Japanese Esoteric Buddhism (New York: Weatherhill, 1972) 134.
24. These four Buddhas are from the group of five sometimes referred to as
'Dhyarn" Buddhas. They are (clockwise from the east, which is at the 'bottom"):
Ratnasambhava, Amoghasiddhi and Amitabha. The five Buddhas
of the Womb Realm are Mahavairocana at the center surrounded by (clockwise
from the west, which is at the "bottom"): Amitayus, Divyadundubhi-megha-
Ratnaketu and Samkusumita-raja. For more on the deities of the two
maI).qalas, see Snodgrass, The Matrix World and Diamond World Mandalas in
Shingon Buddhism.
262 JIABS 19.2
one main stUpa while at Kong6buji there are two stUpas, the degree to
which the Great StUpa has usurped a central ritual position at the latter
makes it virtually a one-stUpa complex today in spite of what was likely
its original plan. The combination of iconography from both malJq.ala
realms within one stUpa makes for a certain economy of architecture
since the need for a second (hard to build) stupa can be made obsolete.
This reason, in addition to the convenient expression that such a combination
makes of the doctrine that the two realms are ultimately indivisible, would
offer ample grounds for understanding how a one-stUpa layout may have
become popular. According to some Japanese scholars, the result of this
economics of construction as well as of doctrinal expression has been that
the original plan has been ignored. Evidence for the existence of an
original plan based on two stupas representing, respectively, the two
malJq.alas can be found in two places: Kukai's statement that he would
construct two stupas and in one additional place: the second or Western
StUpa at Kong6buji contains five Buddha images that are all derived from
one maJ)q.ala, that of the Vajra Realm. And, although the building itself is
not so old, the central image has been dated from the early Heian period.
Thus it is supposed that the Western Stupa has always represented the
Vajra Realm and the Great Stupa the Womb Realm, except that in the
latter a transformation took place whereby the central Buddha image
remained of the Womb Realm while the surrounding four Buddhas came
to represent the Vajra Realm such that the StUpa became symbolic of the
union of the two. If such a tranformation took place, it must be assumed
that it corresponded also with a donning of a new name, "Great Fundamental
StUpa" (perhaps to replace an original name such as "Eastern Srupa,,).25
Another reason why it makes sense to understand the original plan as
having been for two stUpas representing each of the two malJq.ala Realms
25. I h a v e y ~ t to encounter any supposition as to the possible former name of
the Great Srupa, but the suggestions by scholars that both the composition of
its images and. its role changed such that it became a sort of single axis, as it
were, of the community, would be consistent with a change of name. Sawa
Ryuken's Mikkyo jiten (under "konpon daitB," p. 249) explains how the Great
Srupa's combining of the images of both Realms derives from the doctrinal
basis of this integration expressed in the Chin-kangfeng lo-ko i-ch'ieh yu-ch'ieh
yu-chi ching (T. 18, no. 867), an important scripture for the Shingon tradition
that is understood as integrating the teachings of the two key scriptures from
which the two Realms are derived. Sawa adds, however, that the original
Buddha images in the two srupas seem to have been derived from the two
different malJ4a1as. A careful argument in favor of this position appears in
Nakagawa Zenkyo, Koyasan garan kaiso no koso to rinen (Koyasan: Koyasan
Daigaku, 1983) 41-51.
is in the wording of the passage from Kukai stating that he will establish
"two smpas that represent Vairocana as the Essential Nature of the Dhar-
madhatu, in addition to two mru:t<;lalas of the Womb and Vajra Realms."
A factor contributing to the "separate" interpretation must have been that
the smpas are given in this passage a name that is different from that of
the maJ).<;lalas; they are not called "smpas of the two Realms." But the
name of the smpas as representing "Vairocana as the Essential Nature of
the Dharmadhatu" need not be understood as indicating that they are
different from the maJ).<;lalas of the two Realms. Vairocana seen in this
aspect is recognized as the Vairocana who is depicted at the center of
both Realms. So the smpas could be given this name and still each
represent one of the mru:t<;lalas.
Nevertheless, since there are no extant documents clarifying in detail
precisely what the original plan was, the problem as to the exact relationship
between the maJ).<;lalas and the smpas will likely remain unsolved. But
whether each smpa was intended to represent a different maJ).<;lala or
whether, as some think, at least the Great Smpa was envisioned from the
start as an "integrated" model as it stands today, the transformation of a
stupa into a symbolic center of a maJ).<;lala (or of both mal!<;lalas at once),
with the interior of the stupa adorned with images in a manner most
consistent with those of a Golden Hall, was one of the most striking
features of the Garan at Koyasan not to be found in any previous Japanese
temple complex.
I will return to the relationship between stupa and maJ).<;lala after a brief
discussion of the other building central to the Garan at Koyasan, which is
the Lecture Hall. In spite of the central position that the Golden HaIl
held in many temple complexes in Nara, the Lecture Hall played a very
significant role as well. For example, the three most important annual
Buddhist rituals in early Japan took place at Lecture Halls and were
formal ceremonies where large numbers of monks recited, lectured on
and debated about the doctrines of key Mahayana sutras. The ceremonies
were known as the Shoman-e (named after a translation of the Surafigama
sutra, which was recited; held at Hory11ji), the Yuima-e (reciting a translation
of theVimalaklrti-nirdesa sutra; held at KOfuku-ji) and the Saish6-e (re-
citing a translation of the SuvanJaprabhiisa sutra; held at Yakushiji). To
accomodate these gatherings, the Lecture Hall was usually more spacious
than the Golden Hal1.
The Lecture Hall at Kongobuji seems to have
played a role that combined that of both the Lecture Halls and Golden
26. Regarding the relative sizes of Golden Halls and Lecture Halls in Nara,
see the chart in Nara no jiin to Tempyo chokoku 146.
764 JIABS 19.2
Halls of Nara temples since it was at once the central gathering place for
important rituals and a repository of sacred images. In both these rdles,
however, it differed from predecessors in Nara because many of the rituals
and images were new to Japan and reflected Kukai's new systematization
of esoteric practice and doctrine.
The central rectangular section of the Lecture Hall at Kongobuji is
demarcated by a ceiling-to-floor wooden latticed wall that permits the
inside to be seen by all, but only entered by a few [see Diagram 3J.
Surrounding this central section is floor space in which monks commonly
gather for ceremonies. In the central section is an array of statues of
deities in the back (to the north) and a wide space for ritual worship
centered on the two maI}.c;lalas. At the eastern and western ends of this
wide space are hung large (approximately ten feet square) painted scrolls
of the MaI}.c;lalas of the Two Realms (Womb Realm to the east and Vajra
Realm to the west)?7 In front of these painted maI}.c;lalas are square,
horizontal altars (known as daidan and modeled after earlier Indian
maI}.c;lalas) on top of which are placed a variety of ritual imQlements and
before which the worshipper sits facing the hanging maI}.c;lala?8
Si"itras may have been read before these altars, as they were at the
Lecture Hall in other temples, but whether they were exoteric or esoteric
ones, it is clear that Kukai saw that there were certain esoteric rites
associated with the reading of sutras that represented an altogether unique
order of worship. He laid out his views on this matter quite clearly in a
request written to the court in 834 that the annual Saisho-e ceremony held
at the Sovereign's palace during the fIrst week of the New Year be replaced
. 29
Y an esotenc ceremony.
I have heard that there are two kinds of preaching of the Buddha. One is
27. The p ~ i n t e d maI).c;lalas face the center and toward one another, perhaps
ten meters apart, and are hung so as to preserve the directional orientation of
the much earlier Indian convention of horizontal maI).c;lalas. Thus the top of the
Womb Realm maI).c;lala (hung to the east facing west) represents the east, while
the top of the Diamond Realm maI).c;lala (hung to the west) represents the west
so that if one were to lay the paintings flat atop of a horizontal altar their
orientation would accord with the cardinal directions.
28. For more on these altars and drawings of their layouts, see Sawa Ryuken,
Mikkyo daijiten (first one-volume edition of the original six-volume dictionary)
(Kyoto: H6z6kan, 1983) 1493-96.
29. KCZ 3: 370-71. The same text is partially translated and briefly discussed
in De Visser, Ancient Buddhism in Japan, vol. 2 (Leiden: E. J. Brill, 1935)
shallow and incomplete while the other is esoteric. The shallow teaching is
comprised of the scriptures with long passages and verses, whereas the esoteric
teaching is the d h a r a ~ z l found in the scriptures. The shallow teaching is, as
one text says, like the diagnoses of an illness and the prescription of a
medicine. The esoteric method of reciting dhara/'}z is like prescribing ap-
propriate medicine, ingesting it and curing the ailment. If a person is ill,
opening a medical text and reciting its contents will be of no avail in treating
the illness. It is necessary to adapt the medicine to the disease and to ingest it
in accordance with proper methods. Only then will the illness be eliminated
and life preserved. .
However, the present custom of chanting the Satra of Golden Light at the
Imperial palace is simply the reading of sentences and the empty recital of
doctrine. There is no drawing of Buddha images in accordance with proper
technique nor the practice of setting up an altar for offerings and for the
ceremonies of empowerment. Although the reading of the Satra may appear
to be an opportunity to listen to the preaching of the nectar-like teachings of
the Buddha, in actuality it lacks the precious taste of the finest essence [ghee]
of Buddhist truth.
I humbly request that from this year on, fourteen monks skilled in esoteric
ritual and fourteen novices be selected, who while properly reading the Satra
will for seven days arrange the sacred images, perform the requisite offerings
and recite mantra in a specially adorned room. If this is done, both the
exoteric and esoteric teachings, which express the Buddha's true intent, will
cause great happiness in the world and thereby fulfill the compassionate
vows of the holy ones.
We do know that a chapel designed expressly for the purpose of such
esoteric rituals was indeed constructed within the Sovereign's residence,
and what is thought to be the oldest remaining representation of the
painted Womb Realm MaQ<;I.ala is reported to have been used there. It is
not known how the ceremonies performed there might have compared
with those at the Lecture Hall in Kong6buji (surely they were more elab-
orate), but there can be no doubt that representations of the maQ<;I.alas
were employed.
Stu pas and Mar;qalas
The Lecture Hall at Kong6buji, which like the stupas maintains apparent
continuities with Nara temples, nevertheless differs from these structurally
as well as ritually. An interesting feature of the stUpas and the Lecture
Hall is that each of them contains within their structure representations of
both maQ<;I.alas as well as stUpas. In the Lecture Hall, in the middle of the
wide area used for worship there is another flat, square shaped altar. This
is separate from the two such altars that face each of the hanging maQ<;I.alas
to the east and west. It is a third altar between these other two, but with a
266 JIABS 19.2
seat that faces the sacred images to the north side of the central section.
Each of these three altars represents it horizontal maI).c;lala and is covered
with a variety of ritual implements and offerings such as a vajra and
vajra-bell, small bowls for water and incense, candles, flowers and so on,
and is surrounded by a single cord that connects the four corners atop a
two-foot high metal stick at each corner. 30 In this third altar there also
stands at the center a smail bronze st11pa, approximately two feet in height.
A very similar altar is found in the Great St11pa. It also has a seat on the
south side of the altar facing north, toward the five main Buddha images
at center, and is adorned with the same implements and offerings. At its
center again stands a miniature bronze st11pa. It is worth noting that there
is a succession of representations here: the small bronze stilpa. sits within
a maI).c;laIa altar, which is located inside the Great St11pa that enshrines the
central Buddhas of the two maI).c;lalas, which is itself understood as the
center of a larger maI).c;lala that encompasses at least the sacred precincts
of K6yasan if not the entire world beyond. Put simply, there is a motif
represented here on more than one level: the stiIpa as the center of the
maI).c;lala and the maI).c;lala as the center of the stilpa.
Beyond this, however, is another relationship that pertains between the
maI).c;lala and stilpa. The stilpa can itself be a representation of the Dhar-
makaya, and in the Shingon painted maI).c;lalas it is frequently found as a
symbolic representation (samiiya) of the central deity Dainichi (Mahavai-
rocana). There are also drawings of the precincts of K6yasan that show
the Great Stilpa at the center surrounded by the eight mini-mountain
peaks that encompass the mountaintop valley in which the temple complex
sits. These eight peaks do not naturally form a perfect circle but this fact
has not prevented them from being depicted in such drawings as if they
surround the Great Stilpa just as do the eight petals of the red lotus
blossom at the heart of the Womb Realm maI).c;lala. Thus the natural
features of K6yasan have come to be interpreted as an actual embodiment
of the maI).c;lala with the stilpa at center as a samiiya form of Dainichi.
The painted form of maI).c;lala in which the deities are represented by
symbols such as stilpas, wish-fulfilling gems, vajras, bells and so on, is
known as the sammaya- (Sanskrit, samiiya) maI).c;lala. In his writings,
30. A photograph of similar altars from the Kanshin-ji temple in Osaka can
be found in Sawa, Art in Esoteric Japanese Buddhism 18-19.
31. These eight deities are four Buddhas in the cardinal directions and four
bodhisattvas. The red lotus is also literally the "heart" of the maI).c;lala since at
one level of symbolism it is considered to represent the physical organ of the
human heart.
Kukai -described this as one of four types of mal.).qala. The paintings on
which are found anthropomorphic representations of the deities (the most
commonly seen mal.J.qala type) are known as dai- (mahii) mal.J.qala. Then
there are maI).qalas of the same shape where the deities are represented by
their mantric- seed-syllables (bfja) written in Siddham script, and these
depictions are known as ho- (dharma) mal.J.qala. These three types are
often found as painted hanging scrolls (bordered in cloth -in a manner
resembling Tibetan tangkha paintings). Lastly, there are three-dimensional
representations where statues of deities are arranged on a horizontal plane,
whether in miniature on an altar or on a large scale such as that found
inside the Great Stupa. This is known as the katsuma- (karma) mal.J.qala.
These four are all important components of the Shingon symbolic and
ritual systems. An indication of how highly Kukai regarded such repre-
sentations is one of his comments about the power of seeing a mal.J.qala:
'The esoteric essence is profound and mysterious and is not easily captured
with brush and ink [i. e., through writing]. And so it is revealed to the
unenlightened by means of drawings and paintings. The many postures
and mudrii [represented in the mal.J.qalaJ derive from the great compassion
[of the Buddha]. With just one glance [at the mal.).qalaJ, one becomes a
Thus these visual representations are held to be vivid and effective
upiiya, or skillful means. As such, they are invitations to another world
of experience. An additional meaning of these four mal.J.qalas is that they
represent dimensions of enlightened perception, that is, the experience of
one who has truly learned to see the world as a sacred maI).qala. According
to this view, the mahii mar;c/ala is the physical world around us: all the
objects of our senses as integral parts of the body of the deity. The
samiiya ma/p/aZa can be understood as a representation of the deity's
intention, which in general suggests his compassionate wish to liberate
beings and in particular to the myriad forms this compassion assumes.
The dharma malJq,ala represents all sound and speech seen as the words
of the deity and, lastly, karma malJq,ala can be understood as taking all
the movements in the world to be the actions of the deity. Thus a
32. A small representation of a karma-mal).qala is shown in a photograph in
Sawa, Art in Japanese Esoteric Buddhism 135.
33. KCZ 2: 25. See also Hakeda's translation in Kakai 145.
34. The term samiiya has many meanings. Key to understanding its application
to this mal).Q.ala is the meaning of "vow," which can include the meaning of
"promise" or "guarantee." The various symbolic depictions represent aspects of
the deity's vow to awaken sentient beings.
268 JIABS 19.2
maJJ<;lala may be an artistic representation or it may be a demarcated ritual
sacred space, but iUs also the entire world of sentient experience transformed
by religious practice into a sacred realm that is the presence of the deity.
As I noted earlier, it can also refer to the esoteric teachings themselves,
but particularly with respect to their power to transform our experience of
the world. It is in this context thatKukai wrote about the origins of the
esoteric teachings: "Mahavairocana Buddha, the great compassionate one,
enjoying for himself the taste of the equality that is enlightenment, was
saddened by the plight of the beings in the six realms of rebirth. And so
it was that the thunder of his wisdom that is one with reality trembled
throughout His dharma-realm palace and the secret mandala was thereby
35 ..
transmitted to Jambudvlpa."
But the malJ<;lala is not alone in having cosmic connotations, for the
same can be said of the stl1pa. It is not merely an architectural product
nor just a bronze item adorning a ritual altar. If, as in the iconography of
the interior of the stl1pas at Kong6buji, the stupa is represented on top of
the horizontal altar as part of a karma ma1Jrjala, it can be likened to the
"dharma-realm palace" of the deity. Such a combination of stl1pa and
malJ<;lala evokes complex meanings: while the stl1pa is commonly taken
as a symbol of the Buddha's absence, the malJ<;lala is often interpreted as a
symbol of his active presence in our world. Curiously, both of these
meanings of absence and presence can be seen to coincide in the Shingon
school's interpretation of the Buddha's dharmakiiya insofar as this incon-
ceivable, transcendent and ultimate body of the Buddha is understood to
be the "preacher" of its main scriptures. Even outside the Shingon tradition,
however, the stl1pa is often said to represent the Buddha's dharmakiiya?6
As noted above, this is the case in the Shingon samiiya ma1Jrjala when
painted as a hanging scroll where the symbol stands for Mahavairocana,
but it is alsoctrue for the three-dimensional structure. There is as well an
altogether different significance that recalls a legend of an ancient "iron .
stl1pa" in India out of which Nagarjuna is said to have retrieved the
esoteric Vajrasekhara sutra. This legendary metallic "cosmic egg" is thus
said to have given birth to the esoteric Buddhist traditions. The Great
Stl1pa at K6yasan is sometimes said to be a representation of this ancient
one. Kukai wrote about this legend as part of the history of his lineage
and the later Shingon master Kakuban even said of it that "the iron stl1pa
is this very body," referring to the body in which a practitioner is said to
35. KCZ: 3: 392-94.
36. For example, see David Snellgrove, Indo-Tibetan Buddhism, vol. 1 (Boston:
Shambala) 37.
be able to attain enlightenment according to Shingon (sokushinjobutsu).37
"This very body" commonly denotes "this lifetime" and as such is under-
stood to refer to the capacity that esoteric Buddhist practice has to bestow
Buddhahood rapidly. The Shingon tradition refers to itself as "the sudden
among the sudden teachings." But in Kukai's work entitled The Realization
of Buddhahood in This Very Body (Sokushinjobutsugi) "this body" also
has a cosmic dimension and signifies the entire dharmadhiitu. In KUkai's
exposition, all of existence whether mundane or transmundane is comprised
of the same six elements, five material ones and the mind. It is with this
grand "body" as our ground that we engage in the practices that engender
aWakening. 38 And, insofar as this universal "body" can be considered
that of the Buddha, it can also be conceived as being a grand smpa. As
Kukai writes
The mind-palace of the many treasures is high and w ~ d e without limit, and
the sun-residence of bright light extends everywhere.
The "great self' [as
taught] in Shingon originally resides in the lotus of the mind while its myriad
attendant mental-aspects naturally dwell in the moon of aWakening. The
Truth of the Three Equalities [of body, speech and mind, or of beings, the
Buddha and one's own mind] resides in the effulgence of the Buddha-sun
and is continually being taught. The Buddha's mystical embrace [kaji, Skt.
adhi$.thiina] ceaselessly responds to the faculties of beings. How marvellous
is the stupa of the Dharma-nature body; it is truly grand. [emphasis mine]
37. Charles Orzech has translated this legend from the Chinese in Donald
Lopez, ed., Buddhism in Practice, (Princeton: Princeton University Press,
1995) 314-17. Orzech also evinces a further connection between smpa and
ma.I).qala by noting that through "the process of consecration (kuan-ting, Sanskrit
abhi$ekha) every initiate reenacts the entry into the Iron Smpa with his or her
own entry into the ma.I).qala." Taik5 Yamasaki also mentions the legend in his
Shingon: Japanese Esoteric Buddhism, (Boston: Shambala, 1988) 8. Kakuban's
quote is from Kogyo daishi zenshu (Tokyo: Kaji sekkai shisha, 1910) 510 (as
cited in Orzech above, p. 315).
38. See Hakeda's partial translation of Sokushin jobutsu gi in his Kukai:
Major Works, as well as my Master's thesis, "Kukai's 'The Realization of Bud-
dhahood in This Very Body,'" University of Virginia, 1986.
39. KCZ 3: 278, Shoryoshu no. 54. On the same occasion he also announces
that he will make ten statues of deities, too. He dedicates the merit from these
works to all people rich and poor in order to protect the country, improve the
living conditions of all beings and to cause everyone to enter the path of
40. A standard interpretation in commentaries is to take "the mind-palace of
the many treasures" and "the sun-residence of bright light" to refer to the
Womb and Vajra MaIJ.qala Realms, respectively.
JIABS 19.2
Here the entire universe is likened to a palace. The all-pervading Dharma-
nature body, co-extensive with the vast cosmos, is itself a stupa.
yet this same "body" is also a palace and also a mal)<;lala.
This last passage comes from a text entitled "Votive Document on the
Occasion of Producing the Two MaJ;l<;lalas for the Sake of Repaying Our
Benefactors," which Kukai wrote dedicating the production of painted
mahii mar}4ala representations of the Womb and Vajra Realms in 82l.
As the text notes, in the fifteen years since Kukai's return from China
with these maJ;l<;lalas, their continued use in ritual settings had worn them
considerably and it had become necessary to reproduce them. Records
indicate that this was the first of several times that Kukai had them
Drawing attention to this votive document (ganmon) that
Kukai drafted on the occasion of reproducing a pair of large painted
maJ;l<;lalas in 821 provides an entry into one final consideration of the
roles that maJ;l<;lalas played in the early Shingon tradition. The collection
of Kukai's writings known as the Shoryoshu contains at least seven votive
documents in which it is recorded that maJ;l<;lalas were produced. Many of
these documents were drafted on the occasion of a memorial service
commemorating the death of a court noble. It is not clear precisely to
what use these maJ;l<;lalas were put once they were produced. Most
would surely have been used solely for formal initiations into esoteric
practice, but such initiations were not intended only for the monastic
practitioner. Large numbers of laymen also received consecrations from
Kukai in ceremonies for both the Vajra and the Womb Realm maJ;l<;lalas.
It seems that there was widespread interest in Shingon teachings and
practices a!]long clerics and court nobles alike in both Nara and in Kyoto.
41. One last reference to another usage of the term stl1pa takes us to I-hsing's
commentary on the Mahiivairocana sutra. There he refers to "the mind as a
Buddha-stiipa" (shin i hutto). As cited in Sawa Ryuken, Mikkyo Daijiten 1706,
under nanten tetto.
42. The phrase "for the sake of our benefactors" refers specifically to four
objects of gratitude: parents, the sovereign, sentient beings and the Three
Jewels (Buddha, Dharma and Sangha). This term appears frequently in Kilkai's
works as an expression of a sense of indebtedness as well as dedication.
43. Although the patrons of these memorial services must have provided for
the costs of materials and labor required by such work, it is doubtful at this
early date that the painted-scroll malJQalas themselves would have graced the
walls of an aristocrat's living room as they might today.
The first large formal ceremony for which we have a record took place in
812 at the Takaosanji temple in the mountains outside Kyoto. Extant in
Kukai's hand is a list of those who received initiation. The names include
Saicho and several of his disciples, many important monks from the main
Nara temples and some laymen, numbering over one hundred forty in all.
While a few of these initiants would later study further with Kukai, it is
doubtful that for the majority their initiation into the world ofthe mat).<jalas
was the actual beginning of dedicated cultivation of esoteric practice.
These were probably lower level initiations that served to "establish a
bond" (kechi-en) with the deities of the mat).qala, and this bond was con-
sidered to be beneficial in and of itself regardless of the extent of subsequent
Such ceremonies became popular later in the Heian period
among the aristocratic class and we can be sure that for many people,
partaking in a consecration was like receiving a blessing, if not making a
statement of fashion or status, more than it represented any committment
to a spiritual discipline. But participation in such ceremonies was something
desired by both the aciirya and the recipients. That is, while there was
among the aristocracy in general (lay and cleric alike, as they often came
from the same sectors of society) a strong interest, perhaps even fascination,
with the new esoteric rites and ritual paraphernalia that had just recently
been introduced from China, there was Clearly also a keen desire from
Kukai's side to promulgate these practices and related texts widely at a
variety oflevels. As he wrote around 816:
What I wish for now is that people with some affinity for this teaching
might read the texts, lecture on them and make them known to the world,
thereby repaying the kindness of the Buddha. Since the circulation of the
teaching has been inhibited because the texts themselves are so few in number,
I have sent my disciples Koshu and Angyo to distant regions bearing these
If there be persons with predilections for the vehicle of supernatural pow-
er-virtuous men or women, whether cleric or lay-who share the same
wish as I, may they establish a bond with this teaching; copy, read and recite
the texts; practice in accordance with the teaching and harmonize their thoughts
with the principle taught therein. By so doing, they will ascend beyond the
ten stages of the bodhisattva path without having to pass through the incalcu-
lable aeons traditionally required by the path, and in the very body born oK
their parents will quickly realize the Buddhahood inherent within their mind.
44. The exact status of these initiations into the two Realms given by Kukai
in 812 is unclear. Paul Groner treats this topic in his Saicho 81-83.
45. KCZ 3: 386-91 (ShiJryoshfi no. 98).
272 nABS 19.2
This passage is taken from a letter that Kukai sent to numerous people
throughout Japan along with copies of various esoteric scriptures. The
main emphasis of the letter is on the importance of copying and thereby
disseminating these scriptures. Whether or not the texts were actually
ever recited or studied by their recipients, implicit in the enterprise of
copying scripture was, of course, the power that is to be derived from the
mere presence of the texts themselves as well as the merit to be obtained
by those who copy (or sponsor the copying of) them. The letter indicates
that Kukai envisioned a broad field of practice developing in connection
with these texts. Surely it was not long before some of the communities
to which these texts were sent received also the liturgical parapherna-
lia-va]' ras, bells, offering vessels, and so on-and some of the mandalas
46 ..
that were so central to the promulgation of the esoteric cult.
In fact, it is very important to keep in mind the role of financial /
material support provided by court aristocrats if we are to understand
fully the manner in which new religious practices such as those promoted
by Kukai actually took root in Japan. As was noted earlier, support for
the construction of smpas and temple buildings, especially in the remote
region of K6yasan, was not easy to come by. Perhaps in the case of
smaller projects like the production of painted maI;lQ.alas or of a few
carved Buddhist deities, patronage was easier to procure. One can imagine
how this might be particularly true when the aims of a specific ceremony
in which such material is to be employed are clearly spelled out in terms
of the kinds of soteriological benefits that can accrue from memorial
services offered on the behalf of a loved one. The following is a passage
from one of the votive documents written by Kukai on the occasion of a
memorial ceremony:
To come:together and then to drift apart is the law of things in this world.
Alas, the late fifth daughter of the Fujiwara family married in full possession
of virtue as well as beauty; she went to her husband with all the requisite
skills and proper speech. Her numerous children and grandchildren filled
homes and gardens alike. All had hoped that like a crane or a tortoise she
46. I have translated this text, commonly known as the Kan' ensho netter of
appeal to those with interest"), in my dissertation and it will also appear in the
forthcoming volume of the Princeton Readings in Religion series called, Reli-
gions of Japan in Practice, ed. by Robert Morrell, Princeton University Press.
My dissertation contains a study of the contexts in which this letter was sent
(pp. 14-52) as well as an annotated translation, (pp. 244-57). See my "Kukai
and the Beginnings of Shingon Buddhism in Japan," diss., Stanford University,
would live a long life closely united with her husband. .. How pitiful and
desolate we are. Dew and frost vanish quickly and lightning flashes only
briefly. And now many days have passed, the stars have coursed their way
and we face the one year anniversary [of her decease].
On the 22i:td day of the 5th month of the fourth year of Tench5 [827], in
order to deliver her spirit we drew the "one mudra" mandala of Dainichi on
one canvas made of five strips.47 We also copied in' ~ e v e n fascicles the
entire dharma mar:t<;iala [text?] of the broad eye[d one, that is the Dainichi
kyo or Mahiivairocana satra]. In addition, at JingBji temple [Takaosanji]
monks assembled to recite the Dainichi kyo. The sound of bells resounded
throughout the broad valley ...
My wish is that by means of this virtuous cause we might bring succour to
her lost spirit. May the thunder of the Dharma awaken the long slumber of
Buddha nature, and i j ~ sweet ambrosia nourish the roots and leaves of the
tree of bodhimind ...
The text is unclear as to whether the Buddha nature to be awakened by
the thunder of the Dharma is that of the deceased woman alone or of all
beings in general. But since later in the text it is hoped of the latter that
they might enter into the "palace of awakening," it would be natural to
conclude that the ceremony is intended to transfer merit to both, near and
far. Yet the document also refers, on the near side, to the bereaved's
family, chief among which is the husband Nakamori Kasa, officer in the
court Department of Ceremony (shiki bu). Another votive document,
dated the 15th day of 10th month of the sixth year of KBnin (815), also
records a different memorial ceremony that Kukai performed on behalf of
this same Nakamori's deceased parents.
Kukai commonly wrote multiple
votive documents for a given individual, many of them for a member of
the influential court family Fujiwara. It would appear that through these
ceremonies he was offering them something of value. Further studies are
needed in the history of funerary and memorial rites in early Japan before
proffering any judgement as to how novel were these ceremonies that
Kukai performed, but one cannot deny that some of the esoteric elements
of these Shingon rites-the maQ.l;lalas painted, the scriptures copied, the
new esoteric deities carved in statue form, in addition to the overall
mystifying atmosphere of the complex liturgy--contributed greatly to the
attractiveness and subsequent popularity of such rites. In other words,
47. This is the ichi in'e, which forms one of the nine square panels of the
mar:t<;iala of the Vajra Realm. It is the panel at the top center, which depicts a
large image of Dainichi nyorai in the "wisdom fist mudra."
48. KCZ 3: 312-14 (Shoryosha no. 66).
49. KCZ 3: 270-72 (Shoryosha no. 51).
.274 JIABS 19.2
the ceremonies offered more than just the accumulation of merit for both
the deceased and the bereaved. These ceremonies were part of an economy
of exchange between Buddhist priests and the laity (especially members
of the court) that provided, on the one hand, the resources essential for
the propagation of a new cult and, on the other, the solace and perhaps
status to be gained from the participation in what must have been seen as
a grand system of religious meaning and even legitimation.
Although not all esoteric Buddhist practice was devoted to such "worldly"
matters, it was nonetheless in contact with this economy of exchange that
the earliest communities of systematized esoteric Buddhist practice thrived
in Japan. The Kong6buji monastery complex at K6yasan was the first
attempt to establish a center devoted exclusively to this practice though
many centers were being developed at the same time within the already
established "exoteric" monasteries in Nara. The founding of these centers
for esoteric Buddhist practice in the early Heian period is a relatively
well-documented process but one that has yet to receive adequate attention
outside of sectarian studies focused on Kl1kai's life. Nevertheless, the
writings of Kl1kai-especially his various correspondences and votive
documents -preserve a vast amount of information that represents a
valuable record of the emergence of one of the earliest known coherent
systems of esoteric Buddhist theory and practice. By "emergence" Iof
course refer to causes and conditions both intrinsic and extrinsic to the
work of the monastic specialists themselves. Esoteric Buddhism in
Japan did not develope in a vacuum but in the rich atmosphere of an
active social and political world. It is within the dynamics of this particular
world that we must learn to discern the processes whereby "the secret
mal).<;lala was transmitted to Jambudvlpa."
betta J3U 3
Chin-kangfeng lo-ko i-ch'ieh yu-ch'ieh yu-chi ching
15: jiijJU M - -IjJ] f& 11Jo f&*.e;;f,I
chingo kokka
Dainichi *8
hi5 mandara
Dosh6 3!!fIB
ganmon JJlJt
Genb6 .
Gonzo ibtlk
Gyohyo fJ*
hokkai taishO
hosshin seppo
ichi ine
jinenchi shu ?'&ff *
Kai-yuan shih-chiao-Iu
kaji JJu3<#
katsuma mandara
KobO daishi gyoka ki
kokuzo gumonji hO
Konkomyo saishOo-kyo Iljj l&Jlj ::Eff,I
konpon daito tH**m
dai mandara
mikkyo ka
nenbun dosha
Nihon ryoiki
sammaya mandara
Shan wu-wei
shiki bu
shin i butto
shinzui *$71<
'}.76 nABS 19.2
sokushinjobutsu RPJrfi)GMJ
Ta-jih-ching -};:. 8
Tao-hsiian iJ(
Toji chiJja
uchi dojo
Kusuko no ran
Yao-shih liu-li-kuang ju-Iai pen-yuan kung-te ching


0) 0)
f- I
1 Kondo (Golden Hall)
2 Kodo (Lecture Hall)
3 Stilpa
Western Great
Stilpa Stllpa
4 Main (Southern) Gate (Foundation only remaining at Kongobuji)
Diagram 1: The Layout of the Kongobuji Garan Compared with
Three Others from N ara .
.278 JIABS 19.2
0 0
0 0
0 0
0 0
'-------------11 Entrance It------------'
1 Statue of Mahavairocana Buddha of the Womb Realm surrounded by
the Four auddhas who accompany Mahavairocana in the Diamond
2 The sixteen small circles represent round pillars on which are painted
representations of the Sixteen Great Bodhisattvas of the Diamond Realm
3 Altar upon which are various bronze implements laid out in maI}.<;lala
form with asmall stiipa at the center
Diagram 2: The Interior of the Great Stiipa at Kongobuji
[ 5 I
~ @ ~ c 0 ~
1 Hanging Diamond Realm Painted MaJ)c).ala
2 Hanging Womb Realm Mal).c).ala
3 Altars placed before each Mal).c).a1a-accompanying seat faces
hanging mal).c).ala; surface decorated with offering implements and
small stilpa at center
4 Central Altar Facing Statues (with accompanying seat)
5 Statues of Various Esoteric Deities
Diagram 3: Interior of the Lecture Hall at Kongobuji
Reevaluating the Eighth-Ninth Century Milieu:
Icono-Conservatism and the Persistence of Sakyamuni
The typical way to begin an essay such as this-that is, a "reevaluation"-is
to say that the period in question has been neglected, understudied, or
simply passed over. One could hardly say this about the Pala period.
Indeed, since Banerji's 1915 monograph, The Piilas of Bengal, this has
been one of the more intensely examined periods in Indian Buddhist his-
tory. One of the effects of all of this attention, however, is that certain very
assumptions have simply been handed down, paral1Jparii, without
sufficient scrutiny. What I wish to examine here is what is perhaps the
most glaring of all such assumptions: namely, that Pala-period Buddhism
is Vajrayana Buddhism. I shall call in to question this simple equation, and
argue that rather than the hotbed of innovation that is typically seen in the
Pilla period, this is in fact a strikingly conservative period. I argue here that
in the early Pilla milieu what we see appears to be a concerted effort to pre-
serve and conserve the sense of Magadha as the locative center of the Bud-
dhist world, and to assert and reassert Sakyamuni's place at the center of
this center. 1 Paul Mus, of course, made this very point in his Borobutfur,2
I. I should say at the start that I am using the phrase "PaJa period" here as a con-
venient rubric, and that I am thus bracketing the decidedly messy issues involved
in such easy, if not facile, periodization. It is, for instance, virtually impossible
to determine just what constituted the PaJa polity, or to determine the geo-politi-
cal extent of that polity, let alone to determine the extent to which the Palas as
kings influenced the production and use of Buddhist sculpture: see my 1996 Ph.
D. Dissertation, "Wisdom Divine: The Visual Representation of Prajiiii in PaJa-
Period Buddhism," particularly ch. two. One important issue that is indirectly
raised by the present paper is the degree to which the artistic remains from the
early PaJa milieu are reflective of a larger political discourse aimed at, or at least
in tension with, the to the South. The fact that Buddhists in the early
PaJa milieu did not develop the mal)galic programs prevalent in the Deccan dur-
ing the same period is certainly saturated with political significance and issues of
282 JIABS 19.2
but he never fully pursued his own assertion. It is my hope to flesh out
Mus' suggestion here, and I shall argue that the early Pilla period is charac-
terized as much, if not more, by continualion as it is by innovation.
On its face, this is indeed a tall order. What I shall do here., then, is
approach this issue from a very specific angle. I want to focus on the rep-
resentation of the Buddha-or let me say for now of the Buddhas--during
the early Pala period, and in particular the prominence of images of the
Buddha in the bhqmisparsamudrii. What these representations tell us, or
more properly show us, is that, in contrast to what we might see during the
same period in the Deccan or in Orissa, it is not maIJ.Qalas that are being
propagated, but what I shall tentatively call the extended biography of
I shall proceed here by first discussing some of the most commonly per-
petuated assumptions about the sort of Buddhism prevalent in the early PaIa
period; I will then tum to an analysis of the sources of these assumptions,
and offer some alternative interpretations of these sources; and finally, and I
hope most substantially, I shall discuss the iconographic representation of
the Buddha during this period. It is this last discussion that I think pro-
vides the most insight into the nature of Buddhist practice during this
period in the PaIa milieu. For although this period is notable for its innova-
tive and even radical textual practices, the overwhelming amount of hard
evidence provided by the sculptural remains of the period indicates a strik-
dynastic "legitimacy": i. e. the PaJa period Buddhists' development of a kind of
locative maI).<;lala of Sakyamuni's life, the center of which is the PaJa realm, can
be seen in a tense sort of juxtaposition with the period Buddhists'
development of a complex of maI).Qalas that create a different sort of "center," a
"cosmic Bodhgaya" away from Bodhgaya. See Geri Malandra's in Malandra
1993, for the very different view from the Deccan.
2. Mus suggests that it is Sakyamuni who remains in the foreground of Bud-
dhist practice-particularly art-until the "irruption" of Islam into the Indian
Subcontinent: see Borobut/-ur, 1 Off.
3. At Ratnagiri, as well as elsewhere in Orissa, there was a tremendous amount
of iconographic innovation taking place during this period, and, in particular, a
marked emphasis on female deities and a move away from Sakyamuni as the
central figure in Buddhist art see Hock, 1987. However, it is important to note
that Sakyamuni does not simply drop out of the picture, in either Orissa or the
Deccan. As Geri Malandra has argued, at Ellora there is at once an innovative
thrust in the direction of increasingly complex maI).<;laIas, while at the same time a
concerted effort to retain the importance of Sakyamuni and also to equate Ellora
with Bodhgaya and the enlightenment episode: see Malandra 1993, particularly
29,70-71, and 114-15.
ingly conventional and conservative modus operandi. It is also my hope
that this specific discussion will open up some possibilities for understand-
ing the contemporary developments in such "periphery" locales-periphery,
at least, from the Pala point of view-':"as the Deccan and Orissa, not to
mention the cqntemporary developments in Southeast and East Asia.
What We (Think We) Know
One of the first modem scholars to describe Buddhist practice during the
Pilla period was Rajendra Lal Mitra, who, in his 1882 book, Sanskrit Bud-
dhist Literature of Nepal, set the standard for at least the next fifty years by
saying that the texts he believed dated to this period-texts that had been
collected by Hodgson in Nepal in the 1820s and 1830s-were "reeking of
pestilent dogmas and practices" (Mitra 1882, 24). The texts in question
were, of course, tantric in nature: among them, the GuhyasamCljatClntra,
SCldhanamClICl, Mafijusrlmulakalpa. La Vallee Poussin,
in his article on Tantrism for Hastings' Encyclopedia of Religious Ethics,
echoed and even exaggerated Mitra's opinion, saying that these Vajrayana
texts contained and promoted "disgusting practices both obscene and crimi-
nal" (La Vallee Poussin 1908-26, 195). Two other prominent early schol-
ars of Buddhism, Maurice Wintemitz and Benoytosh Bhattacharyya, cer-
tainly did not agree on much, but they agreed that Pala-period Buddhism
was of a decidedly low character: Winternitz said that the Pala-period texts
presented "an unsavory mixture of mysticism, occult pseudo-science, magic
and erotics" that was expressed in "strange and often filthy language"
(Winternitz 1933a, 3-4, and 1933b 389-89), while Bhattacharyya wrote that
the texts of the Vajrayana are "specimens of the worst immorality and sin"
(Bhattacharyya 1929, II, xxi).
These comments, by some of the leading scholars of Buddhism in the
fITst half of this century, were not only seldom challenged, but were perpet-
uated and elaborated on. Thus what started out in 1882 as a reaction to the
sexual imagery and seemingly-lascivious practices of a handful of texts
becomes the standard way to describe the nearly half-millennium of Bud-
dhist practices that is encompassed by the phrase "Pala Period," In a work
that remains one of the standard sources for the period, The History of
Ancient Bengal, R. C. Majumdar is representative of this tendency when he
writes: "Buddhism under the Palas differed essentially from what it was
even in the time of Hiuen Tsang in the seventh century A. D. There was no
trace, not only of the ancient schools of the Hlnayana system, but even of
the pure form of Mahayana. What we find instead were forms of mysti-
284 JIABS 19.2
cism that had developed out of the Mahayana," namely, Vajrayana and
Mantrayana (Majumdar 1971,378). 'In another widely read and quoted
work on the period, Sukumar Dutt's Monks and Monasteries of India, we
read that "during the 'Pala Period' of its history Buddhism itself was already
in a phase wherein it was heavily adulterated by the Tantric cult and its
magic spells and practices. . . . The effect of this on its old cultural stan-
dards," writes Dutt, "was to stunt their catholicity of intellectual interests-
in fact to reduce culture to a cult" (Dutt 1962, 345). In short, according to
Dutt, under the Palas, "The religion had enterea on a phase in which the
Mahayana philosophy, of which Nalanda had hitherto been the intellectual
stronghold, had slanted off to an esoteric cult know as Vajrayiina (Tantric
Buddhism)" (Dutt, 1962 349). We find the same conclusion in Joshi's
Studies in the Buddhistic Culture of India; he too emphasizes the develop-
ment of the Vajrayana, and remarks that "except for some exceptional
examples, few and far between, the figures of Buddha, the historic
Sakyamuni, became rare during this period. He being relegated to the posi-
tion Buddhas, the Dhyaru Buddhas became famous and popular
... " (Joshi 1967, 78).
I shall return to this last point in particular, since one of the assumptions
that goes along with the Pala Buddhism equals Vajrayana Buddhism equa-
tion is that Sakyamuni simply drops out of the picture; in the early part of
the Pala period, as we shall see, this assumption could hardly be less true.
The question I wish to address fIrst, however, is from where these scholars
have gleaned these characterizations of early Pala-period Buddhism as
being almost exclusively Vajrayana.
Perhaps the single most commonly cited source for this early Pala period
has been 'ljranatha, who wrote his unambiguously Vajrayana-centric His-
tory of Buddhism in India (rGya-gar-chos- 'byun or dGos-'dod-kun-'byun)
in Tibet in 1608. Taranatha begins his description of this period with an
account of the ascent of Gopala, the first of the Pala kings, who seems to
have taken over the rule of the region in 750, following the so-called
matsyayana, the "reign of fishes" mentioned in the Khalimpur copper-plate
of Dharmapala (Kielhom 1896-97). Taranatha makes the Vajrayana sym-
pathies of this fIrst Pala king quite clear. Before ascending the throne,
Taranatha tells us, Gopala "received from an iiciirya with instruc-
tions to propitiate the goddess Cunda," and, having "attained siddhi of god-
dess Cunda. . . . he became the king on the next day" (Chimpa and
Chattopadhyaya 1970,257-58). In the remainder of Taranatha's account of
this first Pala king there are several other allusions to the Vajrayana; he
mentions a variety of siddhis and siidhana practices, implying, although not
actually stating, that Vajrayana was practiced at the highest level of this fIrst
Pilla king's realm.
The prominence of Vajrayana in Taranatha's account of the next Pala
king, Dharmapala-here I am correcting Taranatha's somewhat confused
chronology-whose reign was from 775 to 812, is even more explicit.
During this period, several Vajrayana iiciiryas are said to have been active,
including, notably a figure named Kn;IJ.acarya who Taranatha says was a
great palJ-cjita of Cakrasambara, Hevajra, and Yamari (Chimpa and
Chattopadhyaya 1970, 268), some of the very deities to whom Taranatha
himself is know to have been devoted.
One of the most commonly cited passages in Taranatha's History is this
account of Dharmapala: "He accepted as his preceptors Haribhadra and
Jfianapada and filled all directions with the Prajfia-paramita and the Sri
Guhya-samaja. The pa1;cjitas versed in the Guhya-samaja and the Prajfia-
paramita were offered the highest seats of honour etc.," (Chimpa and
Chattopadhyaya 1970, 274). This purported dissemination of the
Guhyasamiijatantra, what might be called the Urtext of the Vajrayana, has
often been taken as proof-positive not only of the Palas' official support of
Vajrayana, but also of the prominence of this text in the region. As we
shall see, however, it is precisely this kind of freewheeling extrapolation
that leads to what I think are misguided assumptions about the sort of Bud-
dhism prevalent between 750 and 850 in Northeast India; the date of this
text is a matter of much debate, and the Buddhism that we see contained
within it does not jibe with the predominance of sculptural remains from the
time. I am not suggesting that the Guhyasamiija had no significance in the
Pala milieu, but rather that its prominence has been, at the very least, over-
stated, and that this overstated emphasis has skewed our understanding of
this period in Buddhist history.
Throughout his account of the Pilla period, Taranatha is most interested in
the goings on at VikramaSIla. At Vikramasila there was, for instance, "a
temple of Vajrasana [where] there were then a large silver-image of Heruka
and many treatises on Tantra." According to Taranatha, however, these
were destroyed by Sravakas from Sri Lanka because they said that these
images and texts were made by Mara; "So they burnt these and smashed the
image into pieces and used the pieces as ordinary money" (Chimpa and
Chattopadhyaya 1970,279). Again, we see quite clearly where Taranatha's
sympathies lie, for in one of his few references to non-tantric monks, he
portrays these Sravakas as heretics. He goes on, however, to tell us that
286 JIABS 19.2
Haribhadra, who was then the preceptor of VikramasIla, saved the people
from these misguided and evil. Sravakas: he explained to them. the
kriyayogas, and "he preached most extensively the five Tantras of the
insiders, namely the Samaja, Mayajala, Buddha-samayoga, Candra-guhya-
tilaka, and Mafijusrl-krodha. Special emphasis," Taranatha tells us, "was
put on the teachings of the Guhya-samaja and so it was very widely spread"
(Chimp a and Chattopadhyaya 1970, 279).
I could go on and on about Taranatha here; his account is brimming with
names of authors and texts and descriptions of monastic complexes. My
point, however, should be clear by now: Taranatha is hardly engaged in
what we would call "objective historiography." On the face of it, certainly,
Taranatha's account gives an unambiguous description of the early Pala
period as being dominated by the three classes of tantra and, in particular,
by the Vajrayana teachings and practices as described in the Guhyasamaja
and other anuttarayoga texts.
Consider, however, the problems with taking Taranatha at face value:
First, Taranatha is writing from Tibet for a Tibetan audience; his history, in
fact, is more or less a chronicle of the significant Tibetan figures and those
Indians who either came to Tibet themselves, or whose work became cen-
tral to Tibetan Buddhism. Second, Taranatha himself seems to have been
an adherent to the Kiilacakratantra, and he wrote several texts within that
vein. Third, he finished his history in 1608; although Goeffrey Samuel
may be correct in asserting that Taranatha objectively used the sources
available to him, (Samuel 1993, 420), we should be skeptical, to say the
least, in taking as "objective" a history composed eight hundred years after
the fact. Clearly Taranatha had a sectarian axe to grind, and part of his his-
toriographical exercise was to legitimize his own preferred brand of Bud-
dhist practice. As David Templeman nicely puts it, "Taranatha's purpose ...
was not tQ.,paint a completely accurate portrait of the Dharma and its
adherents but to glorify them, to make them serve as inspirations to the
Buddhists of Tibet ... ," (Templeman 1981, 45). Thus the mainstream
Mahayana, as exemplified by the Prajfiaparamitii genre, is present in his
History only as a kind of lesser partner to the tantras.
And the Sravakas
who were present in the Pilla realm at this time are mentioned only as icon-
4. This is a particularly significant absence, since the majority of extant texts
from the Pala period are Prajfiiipiiramitii texts, and the evidence provided by
roughly contemporary authors, such as Santideva, indicates that it was the
Prajfiiipiiramitii that was at the center of the early Pala-period Buddhist
oclasts who go so far as to melt down consecrated images and turn them
into money. Furthermore, the locative emphasis in TaranlUha's account is
on Vikramaslla, which does seem eventually, that is post-tenth century, to
have become the center of Vajrayana practice in Pala India; he makes only
passing r e f e r ~ n c e to the obviously major, and decidedly mainstream,
monastic centers of NaIanda and Bodhgaya.
The Chinese Pilgrims
In contrast to Taranatha's Vajrayana-centric account, Xuanzang and
Yijing--both of whom travelled in India in the seventh century, and are
thus at least chronologically much closer to the early Pala period-present
views of a much more mainstream Buddhism in what is roughly the Pala
milieu. Yijing, for instance, describes a Pala region in Northeast India that
is dominated by the Sarvastivadins and other so-called Hlnayana schools:
"In Northern India and the islands of the Southern Sea," he reports, "they
generally belong to the Hlnayana ... " (Takakusu 1896, 14). The environ-
ment Yijing describes conforms very closely, in fact, to Dutt's "catholicity
of intellectual interests," the loss of which Dutt argues distinguishes the
Pala period. Yijing sees, for instance, no real differences between the vari-
ous schools extant in the seventh-century milieu, and writes: 'We can rea-
sonably practise both the Maha(yana) and the HIna(yana) doctrines in obe-
dience to the instruction of the Merciful Honoured One, preventing small
offences, imd meditating upon the great Doctrine of Nothingness"
(Takakusu 1896,51). In short, one comes away from Yijing's report with
the impression that the milieu he saw in Northeast India was decidedly
mainstream; except for a passing mention of the Mahavairocanasutra,
there is no mention of Tantric practices.
Xuanzang, in contrast to Yijing, certainly sees plenty of discord in
Northeastern India: "The tenets of the schools keep these isolated, and
controversy runs high ... Each of the Eighteen schools claims to have
intellectual superiority; and the tenets (or practices) of the Great and Small
Systems (lit. vehicles) differ widely ... and many are the noisy discus-
sions" (Bea11884, 162). Nonetheless, as does Yijing, Xuanzang describes
a mostly mainstream milieu, with a notable absence of tantric and Vajrayana
practices. There is no doubt that in their accounts of their travels in India,
these Chinese pilgrims also have their own sectarian agendas, and I am not
suggesting somehow that they are prima jacie more objective or more reli-
able than Taranatha. Rather, I am positing that they present a more diverse
and more, for lack of a better term, balanced view of the early Pala milieu.
Furthermore, when we examine the sculptural remains of this period, we
288 JIABS 19.2
are left with a portrait of Buddhist practices that bears a much greater
resemblance to the mainstream Buddhism that we see in the Chinese. pil-
grims' accounts than to what we see in Taranatha's History.
Searching for
Scholars have generally characterized the Buddhist sculpture of the early
Pala period in much the same way that they have characterized the overall
religious milieu. They emphasize the expansion of the Buddhist pantheon
to include various Vajrayana deities, and, most significantly here, they cite
the rise of the so-called dhyiini Buddhas and the concomitant decline of
Sakyamuni. Gouriswar Bhattacharya, echoing Joshi's comment that I have
already quoted about the rarity of figures ofSakyamuni, remarks: "In the
medieval period a great change took place in the Buddhist theological con-
ception in Bihar-Bengal when Buddha Sakyamuni, the great monk,
Mahiisramar;.a, lost his primary importance and became a family member
of like a Bodhisattva" (G. Bhattacharya 1989, 353). What I will
argue in the remainder of this essay is that this assertion-which I believe
betrays an over-reliance on Taranatha and other Tibetan historians, such as
Bu-ston, as well as on those scholars who have uncritically accepted the
Tibetan portrayal of the period-is not borne out by the sculptural evidence
from the early Pala period. Indeed, almost the opposite is the case for the
eighth through tenth centuries: although, certainly, a wide variety of bodhi-
sattvas and related deities are represented in the Buddhist sculpture of the
early Pala period-some clearly belonging to the Vajrayana-there is dur-
ing this period a continued and consistent emphasis on Sakyamuni, and
particularly on, from the tenth century onward, the group of eight signifi-
cant events in his life and the places associated with these events, the
As we'have seen, Taranatha emphasizes the prevalence of the
Guhyasamiijatantra during the reigns of Gopala, Dharmapala, and
DevapaIa. Taranatha tells us that Dharmapala "filled all directions" with the
Guhyasamiijatantra (Chimpa and Chattopadhyaya 1970, 274), and at
places such as Vikramaslla "Special emphasis was put on the teachings of
the GUhya-samaja and so it was very widely spread" (Chimp a and
Chattopadhyaya 1970,279). The Guhyasamiija is one of the fundamental
texts of the anuttarayoga class of tantra-Benoytosh Bhattacharyya calls it
"the Bible of the Tantric Buddhists" (B. Bhattacharyya 1931, 24 )-and, as
such, it is something of the root text of the Vajrayana. It is in the
Guhyasamiija that the pafica-tathiigatas-the five Buddhas who are
typically referred to, incorrectly, as dhyaniBuddhas-are first articulated
and placed in a maI).<;iala, with at the center. These five Buddhas
are described in this text as emanations or manifestations of Sakyamuni in
the form of the five skandhas: paficabuddhasvabhavatvat paficaskandha
jinalJ, smrtalJ, (B. Bhattacharyya 1931,41). Given the purported ubiquity of
this text in the eighth and ninth centuries, one could reasonably expect to
find evidence of the practices described there throughout the sculptural
remains of the period. We could expect, for instance, to find evidence of
the five Tathagatas and the various members of their families. In short, we
could expect to find evidence of maJ)<;ialas in the art of the early Pilla period.
There are, however, first some chronological difficulties here, since the
dating of the Guhyasamaja is by no means certain. One of the first to
attempt to assign a date to this text was Benoytosh Bhattacharyya, who in
his introduction to the 1931 Gaekwad's edition of the text argues that the
Guhyasamajatantra dates to the third century, C. E. (B. Bhattacharyya
1931, xxxiv). Bhattacharyya attributes the text to the great AsaIiga.
Although he recognized that there was a potential problem with this early
dating, since the text does not appear to be mentioned anywhere until the
seventh century, Bhattacharyya simply sweeps this problem away. "The
reason why we do not find any mention of the Guhyasamlija before
Nagfujuna [the seventh-century Nagarjuna], is because the Tantra was kept
secret among the professors and the doctrines inculcated therein were con-
fined to a few adepts for three hundred years until Buddhist Tantras of the
Yoga and Yogatantra classes obtained publicity during the time of the Sid-
dhacaryas mainly through their mystic songs, preachings and works," (B.
Bhattacharyya, xxxii). The shakiness of Bhattacharyya's argument is not
limited, however, to this rather absurdly literal thesis about the "secrecy" of
the text (the guhya of the title does mean secret or hidden). He erroneously
connects the AsaIiga to whom a single sadhana in the Sadhanamiila is
attributed with the great AsaIiga of the third century, to whom Taranatha
traces the origin of the tantras, and then he uses this to argue for the early
dating of the Guhyasamaja on the basis of the existence of the full-blown
pafica-tathagata system, as mentioned in the so-labelled Asanga sadhana.
The Sadhanamala, however, is clearly a late text (post-tenth century, at
least), and the attribution of one of the verses by the redactors of the text
and subsequent Tibetan tradition to Asanga seems purely a legitimizing
5. See, for instance, Joshi 1967, 330-33.
290. JIABS 19.2
Alex Wayman has also discussed the dating of the Guhyasamiijatantra at
length, and based on, among other factors, a commentary written by
Lilavajra in the eighth century on a text that mayor may not be the
Guhyasamiijatantra as we know it, he asserts that the text must date to at
least before the eighth century; he concludes, however, on what he himself
calls "a purely tentative basis," that the text dates to the fourth century.6
My purpose here is not to engage further in this dating debate-the
details of which would fill several volumes-although I favor a date of no
earlier than the eighth century for the text, and even then I question the
prevalence and the popularity of the practices described therein; 7 rather, I
wish to point out that these attempts to make the Guhyasamiijatantra a very
early text both fuel and are fueled by the very assumptions concerning the
development and prominence of the Vajrayana that I have been calling into
question here. The argument, in short, proceeds along the logic that if the
Guhyasamiijatantra were written in the fourth century, it would then make
sense, as Taranatha claims, that the text and the practices contained in it
would have been prominent by the eighth century. The question that must
be asked, then, is do we see evidence of this text in the sculpture of the
period? Do we see Sakyamuni nudged aside by the pafica-tathiigatas, par-
ticularly by Do we see, in fact, what Gouriswar Bhattacharya
describes when he writes: "In Magadha or South Bihar, of all
the Transcendent Buddhas, was the most important deity of worship during
the Pala period. Buddha Sakyamuni in bMimispar.samudrii was identified
with and this Transcendent deity attained more importance than
the mortal Sakyamuni?"S The short answer to this question is a simple no.
lt is, however, at least worth looking for since any number of
scholars in any number of studies have misidentified images of Sakyamuni
The over:1Vhelming number of early Pala-period Buddha images depict a
seated Buddha displaying the bl1L7mispar.samudrii. Bhumispar.samudrii
images, also sometimes referred to as miiravijaya images, on the most
obvious level serve what Vidya Dehejia has called a "monoscenic" narrative
and symbolic function, presenting a single episode in the life of
6. See Wayman 1977,84-104, and also Wayman 1973, 17-19.
7. See Joshi, 1967 330-32, for a particularly concise, and sensible, hypothesis:
and also see Yukei Matsunaga 1977, 179, for a view that takes in to account the
Chinese translations of the text.
8. G. Bhattacharya 1989, 352 Bhattacharya makes this assertion in spite of the
fact that he is very clear about the iconographic distinctions between Sakyamuni
and Alq;obhya.
Slikyamuni;9 such images represent the point at which Sakyamuni achieves
enlightenment, the point at which he becomes the Buddha. John
Huntington has called this event "the summum bonum of Buddhist soteri-
ology" (J. Huntington 1987b, 58). "It is this moment," he writes, "the
moment of gaining the right to enlightenment ... that is the validation of the
soteriological methodology taught by the Buddhists" (J. Huntington 1986a,
61). In other words, this is the culmination of the dharma.
Let us briefly recall the event: Just at the point at which he is about to
attain bodhi, Sakyamuni is confronted by Mara, who realizes that he is
about to be defeated by this man who has discovered the means with which
to cut through all artifice and to conquer death. Mara, however, the embod-
iment of subterfuge, creates all manner of illusion and temptation to distract
and defeat the Buddha-to-be. He unleashes his various armies-appropri-
ately named desire, discontent, hunger and thlrst, craving-but Sakyamuni
is unmoved. Mara then uses his own daughters to tempt Sakyamuni, to stir
in him lust and desire, but again to no avail. So finally Mara assaults him
verbally, and challenges his very right to be beneath the bodhi tree, his right
to achieve enlightenment. Sakyamuni responds that all of the millions of
offerings that he has made in the past have given him the right to enlight-
enment. Mara, however, persists; he says there is no witness to support
Sakyamuni's claims. Sakyamuni's response is the exact moment depicted in
bhumisparsamudra. images: he reaches out his right hand and touches the
earth. The bhudevl, the goddess of the earth (who is also sometimes
depicted in the images), is impartial and free from malice, and thus serves
as the ideal character witness.
Scholars have frequently identified bhumisparsa Buddha images from
the early Pala-period as for example, in the volume The Image
of the Buddha, edited by David Snellgrove, several bhumisparSamudra.
Buddha images are identified as although there is virtually no
iconographic rational for doing so other than the fact that in the
Guhyasama.jatantra and other anuttarayoga tantra texts, char-
acteristic hand gesture is the bhumisparsamudra..l
Again, it seems that
the assumption about the prevalence of tantric practices in the early Pala
period has simply clouded the vision of such scholars, transforming
Sakyamuni into Take, for instance, the image from Bodhgaya
9. See Dehejia 1990, for the full context of her discussion of this term, as well
Dehejia 1992. See also S. Huntington 1990, and S. Huntington 1993, for a dif-
ferent perspective on the issues Dehejia discusses.
10. See Snellgrove 1978 ..
292 JIABS 19.2
that Susan Huntington properly identifies as Sakyamuni and that dates
roughly to the late ninth century;l! this is a fairly typical bhumisparsa
image. The Buddha (to-be) is seated cross-legged, his left hand rests in his
lap, and his right hand is draped over his right knee, touching the earth.
The Buddha is flanked by two bodhisattvas-Avalokitesvara and
Maitreya-and above his head, at the top of the stele, are two vidyadharas
and a bit of the asvattha bodhi tree, the latter a fairly standard motif on such
sculptures, serving to visually represent the locale of the scene. There is
also a small vajra directly below the ankles of the Buddha, a common motif
on Pala-period bhumisparSa images, and a motif that might lead one to
identify the image as representing the vajra is, after all, the
symbol par excellence of the Vajrayana. In this case, however, the vajra is
iconographically tied to Sakyamuni's defeat of Mara and, in particular, to
Bodhgaya itself: as Mara tempts him, he sits atop the vajrasana, the
"adamantine," or "diamond-like seat" of enlightenment, pure and
immovable. 12
Although it is uncertain when Buddhist began to sculpturally represent
individual jina-Buddhas, it appears that this practice did not begin until
some time after the tenth century. Of the tathagatas, is in fact
the most common in Northeast India (as opposed to Vairocana or Amitabha
elsewhere in the Buddhist world); this later iconographic development
does, perhaps, indicate the eventual prominence of the anuttarayoga prac-
tices such as those found in the Guhyasamiijatantra. The distinguishing
iconographic detail in such images is not, as has frequently been assumed,
the crown and jewelry that adorn the Buddha in such images (these ele-
ments are typically used with images of Sakyamuni from the later Pilla
period). Rather, the distinct iconographic element that does not occur on
images of Sakyamuni displaying the bhumisparSamudrii is the single, cen-
tral elephant at the base of the stele. 13 According to Abhayakaragupta's
again a late, post-tenth century text that gives the
iconography of many of the deities in the Vajrayana pantheon, s
vehicle is the elephant. It is important to note, however, that this icono-
graphic detail is a late development, and even in the latest periods of the
Pala era sculptures of never even approach the popUlarity of
images of Sakyamuni.
11. S. Huntington 1984, 103.
12. See Janice Leoshko 1988, for an extended discussion of vajriisana images.
13. O. Bhattacharya 1989, was, I believe, the fIrst to point this out.
In sum, then, reports of Sakyamuni's death during this period have been
greatly exaggerated. Indeed, although we do see evidence of iconographic
innovations during this period-the increased popularity of Tara, say, or of
Mafijusn, or the emergence of Prajfiaparamita as a deity in her own right-
Sakyamuni does not get replaced by or by any of the other so-
called 'Transcendent Buddhas," but on the contrary continues to be the
focus of early Pal a-period silpins. Furthermore, even in the -later period,
when the panca-tathagatas are iconographically represented, they remain,
in comparison to Sakyamuni, iconographically marginal figures, frequently
relegated to the periphery, to the tops and bottoms of images of
Searching for a Mmpjala
It is possible, I must admit at this point, that I have been overly rigid in my
discussion of bhumisparSamudra Buddha images in the Pala milieu, and
that in attempting to correct the "Pala-period Buddhism equals Vajrayana"
equation, I have in the process blurred my own vision. Let me, then, step
back from my own thesis for a moment. Both Geri Malandra and Nancy
Hock have suggested that bhumisparsamudra Buddha images, as well as
other Buddha images with what we might call "historical" referents, have a
polyvalent potential, and that seeing them as representing only Sakyamuni
is an overly narrow interpretation.l
Hock, for instance, sees an
"intentional ambiguity" and a "dual nature" in several sculptures from
Ratnagiri that depict a bhumisparsamudra Buddha, with a blurring of the
distinction between Sakyamuni and She argues that the SCUlp-
tural evidence from Ratnagiri reflects the transitional nature of that site,
transitional in the sense that the Buddhism practiced there in the eighth and
ninth centuries was midway between the Mahayana and the Vajrayana,
what she describes as Mantrayana. Thus, suggests Hock, images of
bhumisparSamudra Buddhas are best seen as what she calls "Sakyamuni in
a tantric form,"-that is, not the Sakyamuni of the Pilii and Mahayana texts,
what we frequently call "the historical Buddha," but Sakyamuni as he
appears in several early carya and kriya tantras, such as the Manjusrf-
Hock's "transitional stage" thesis is, I think, largely convincing, for she
presents a great deal of evidence, both sculptural and textual, that indicates
the practice of a variety of early maI).qalas in Orissa. As Hock herself
notes, however, the evidence from eighth / ninth-century Ratnagiri stands
14. See Malandra 1993, 29, 70-7l, 114-15 and Hock 1987, 55-56.
294 nABS 19.2
in contrast to the contemporary. evidence from the Pal a milieu, from
NaIanda and Bodhgaya; as she says about Nalanda, it is a site "that lacks
the iconographically more complex images and programs found at
Mantrayana sites," such as Ratnagiri (Hock 1987, 33). Indeed, as I have
already sug gested here, in the early Pala milieu we see almost an opposite
sort of iconographic thrust;
Thus whereas Hock argues for an intentional ambiguity in the represen-
tation of the Buddha in the Ratnagiri context, I am arguing that there is no
such blurring in the early Pala milieu, but instead a consistent emphasis on
Sakyamuni. At Ratnagiri such images seem to be best seen in the context
of an iconographic expansion-what Hock calls an "explosion" of the
pantheon-that includes increasingly more complex maIJ.q.alas.
Bhiimisparsamudrii Buddha images from the Pala period, however, must
rather be seen in the larger context of the images that
first appear in the early Pala period and which continue to be popular into
the 12th century; for these images, and the pilgrimage centers associated
with them, present the full biography of Sakyamuni in a condensed form,
and thus they recreate the presence of the historical Buddha in the Pala
realm. The viewer of, or we might more accurately say the participant in,
such images is visually transported into the past, into the presence of
Sakyamuni, to the time when-and also to the place where-the Buddha
was alive, defeating Mara, preaching the dharma, and so on.
Another common Buddha image from the early Pala milieu that also
forms part of the conventionalized set of eight scenes is the Buddha dis-
playing the dharmacakrapravartanamudrii (or simply the dharmacakra-
mudrii). As with bhiimisparsamudrii images, these images have a
graphical and also a locative significance, since the fIrst sermon was deliv-
ered by Sakyamuni at Samath, located at the heart of the Pala realm. And
as with bhiimisparsamudrii images, such images have frequently been
as representing one of the panca-tathiigatas, in this case
Vairocana, who is described in later texts as displaying this same hand
gesture. Again, however, such images from the PaIa milieu unambiguously
represent Sakyamuni, as evinced by the common iconographic details at the
bottom of such stelae, such the fIve disciples who are gathered to hear the
frrst sermon, and the two deer who represent both the locale of this first
sermon in the r$ipatana mrgadiiva at Samath and also the power of the
dharma to "tame" and give refuge.
15. It is not insignificant, also, that the Palas adopt this motif for their royal
seals: see, for instance, L. D. Barnett 1925-26.
These two sorts of images-bhiimispada and dharmacakrapravar-
tanamudra-when taken individually are iconographically quite conserva-
tive, repetitions of iconographic conventions that extend to the earliest peri-
ods of Buddhist sculpture. When they are placed in the larger context of
the stelae, however, this conservatism takes on a more
innovative dimension. Four of the scenes that make up the set of eight-
the birth, the enlightenment, the first sermon, and the parinirvaIJa-had
been sculpturally depicted together from a very early period, and the four
places associated with these scenes had been predicted to become pilgrim-
age spots by the Buddha himself, as recorded in the Mahaparinibbana-
sutta. Although various collections of scenes from the life of the Buddha
had been depicted in a variety of groupings since as early as the
period,16 the standardization of the is new to the Pala
period.!7 Furthermore, the four additional scenes-the miracle at SravastI,
the descent from the trayastrin;.sa heaven at Sakasya, the taming of the wild
elephant Nalagiri at Rajagrha, and the gift of honey to the monkey at
Vaisali-all take place within the basic confines of the Pilla realm.
So what, finally, are we to make of this particular collection of eight
scenes? John Huntington has remarked: "The sequence is a kind of epit-
ome of the life of Sakyamuni .... the epitomizes the
whole life of the Buddha, his attainments, his teachings and the benefits of
faith in his life to his followers. In short, the set of eight scenes epitomizes
the whole of Buddhism" (1. Huntington 1987a, 55, and 1987b, 67-68).
How do these stelae signify such a totality? The original set of four scenes
in a sense mark the outer parameter of Sakyamuni's life: his birth, his
attainment of enlightenment, his first articulation of the dharma, and his
death. The other four scenes, the ones that are added in the Pala period, all
have to do with the Buddha's propagation of the dharma.1
The miracle at
SravastI, for instance, although it is perhaps most obviously concerned with
the Buddha's superior rddhi, also represents the triumph of Sakyamuni's
dharma over all other teachings; the descent from the TrayastriIl).sa heaven,
where the Buddha had gone to preach the dharma to his mother, likewise
can be seen to demonstrate the superiority of the Buddha's dharma, even
16. See Joanna Williams 1975.
17. See Janice Leoshko 1993/94.
18. Here I am indebted to J. Huntington's interpretation, particularly as articu-
lated in 1. Huntington 1986a and 1986b.
296 JIABS 19.2
over the potential teachings of the gods; the taming of the wild elephant
NaHigiri represents the dharma's ability to control even the most
lable forces; and the gift of honey by the monkey at Vaisali represents the
importance of dana and the potential the dharma holds for even the most
humble of adherents to it.
A great deal of attention in Buddhist Studies has been devoted to the
whole issue of the "absence" and the "presence" of the Buddha in such
physical objects as stiipas, relics, and images, 19 and we could, I think, eas-
ily use such language to discuss images. It may,
however, be more appropriate in this context to think not in terms of pres-
ence, but in terms of an intentionally emphasized "pastness." Each of the
events that makes up the set refers to Sakyamuni's
specific activities in the past in Magadha: in other words, Buddhists in the
early Pilla milieu did not emphasize the future, the coming of Maitreya; and
they did not emphasize the transcendent, cosmic present of the pure lands
occupied by and the other tathiigatas. Rather, in the Pala milieu
the focus falls squarely on the past, on Sakyamuni, and images such as the
A$.tamahiipratiharya make this past available-allow the Buddhists of the
present to participate in this past-in a condensed, visual sort of pilgrimage.
The temporal reality of the Pala present (marked by the absence of the
Buddha) was thus replaced, via such sculptural images; by the spacial real-
ity of the Sakyamuni-inhabited past (marked by the presence of the
Buddha). To use a distinction somewhere articulated by Paul Ricoeur, the
worldview reflected, if not also created, by such images is archeological,
and not teleological: the present is significant because it resonates with the
past, not because it anticipates the future.
Let me pose a final question, then: Does this set of the eight great events
in Sakyamuni's life constitute a mal)c;lala? The answer to this question
depends 'L(fry much on what we mean by a mal)c;lala If we mean what
GuisepptTucci means when he describes the mal)c;lala as "a map of the
cosmos. , , the whole universe in its essential plan" (Tucci 1961, 23), the
answer is "no." If we mean what Reginald Ray means when he calls the
mal)c;lala "the central and all-integrating symbol in Tantric Buddhism as a
whole" (Ray 1973, 58), the answer again is "no,"
images in the Pilla milieu do, however, represent something similar to such
conceptions of the maI).c;lala, in that they create a kind of totality of the
Buddhist world, but it is a totality as it was specifically conceived by Bud-
dhists in the Pala milieu. This totality encompasses the entire life of
19. See in particular Schopen 1987 and 1988.
Slikyamuni, from birth to Nirvai)a, including the essential point at which the
Buddha discovers the triumphant dharma-as represented by the
miiravijaya-as well as the crucial points at which he makes the dharma,
and its superiority and power, known to the world. Furthermore, each of
the events that constitute this maI}.Qala is connected to a specific place and
the set creates a pilgrimage circle that could have been completed within the
basic confmes of the Pilla realm. This, then, is a distinctly worldly totality.
It makes the idealized past of Sakyamuni available in the Buddha-less
Why did early Pala-period Buddhists might have put such an emphasis on
Sakyamuni, and why they did not develop the more complex and cosmi-
cally resonant iconographic programs and maI}.qalas that we see at contem-
porary sites in the Deccan and Orissa, as well as in locales outside ofIndia?
We can probably never know the answer to this question. But it may be
that they simply did not need to look any further than the locative present.
There was no need to look to the alternative visions of the tantras and the
Vajrayana to find a cosmic center, since early Pala-period Buddhists were
already at the center of the this-worldly totality constituted by the life of
Sakyamuni. To use an image employed by Asvaghosa to describe
Bodhgaya, they already were at the very navel of the earth.
Bagchi, P. C. 1941. "The Eight Great Caityas and Their Cult." Indian Historical
Quarterly 17: 223-36.
Banerji, Rakhal Das. 1917. The Palas of Bengal. Memoirs of the Asiatic Soci-
ety of Bengal. Vol. 5. No.8. Calcutta: The Asiatic Society.
Barnett, L. D. 1925-26. "The Mungir Plate of Devapruadeva: Samvat 33." Epi-
graphica Indica 18: 304-07.
Beal, Samuel, trans, 1884. Si-Yu-Ki: Buddhist Records of the Western World
Translatedfrom the Chinese of Hiuen Tsiang (AD 629). London: Trubner.
B6nisti, Mireille. 1982. "Une Sculpture insolite de l'Inde Museum de Calcutta."
Artibus Asiae 18: 209-16.
Bhattacharya, Gouriswar. 1989. "Buddha Sakyamuni and Panca-Tathagatas:
Dilemma in Bihar-Bengal." South Asian Archaeology 1985. Eds. Karen
Freifelt and Per Sorenson. Scandinavian Institute of Asian Studies Occasional
.Papers 4. London: Curs on Press. 350-71.
Bhattacharyya, Benoytosh, ed. 1929. Sadhanamiilii. 2 vols. Gaekwad's Oriental
Series 26 and 41. Baroda: Oriental Institute.
298 JIABS 19.2
----, ed. 1931. GuhyasamZijatantra. Gaekwad's Oriental Series 63.
Baroda: Oriental. Institute.
----, ed. 1949. of Mahiipalpjita Abhayiikaragupta.
Gaekwad's Oriental Series 109. Baroda: Oriental Institute.
Chimpa, Lama, andAlaka Chattopadhyaya, trans. 1970. Taranathds History of
Buddhism in India. Simla: Indian Institute of Advanced Studies.
Dehejia, Vidya, 1990. "On Modes of Visual Narration in Early Buddhist Art."
Art Bulletin 72.3: 374-92.
----. 1992. "Aniconism and the Multivalence of Emblems." Ars Orien-
taUs 21: 45-66.
Dutt, Sukumar. 1962. Buddhist Monks and Monasteries of India: Their His-
tory and Their Contribution to Indian Culture. London: George Allan and
Hock, Nancy, 1987. Buddhist Ideology and the Sculpture of Ratnagiri, Seventh
through Thirteenth Centuries. Diss. University of California, Berkeley.
Huntington, John, 1985. "Sowing the Seeds of the Lotus: A Journey to the Great
Pilgrimage Sites of Buddhism, Part 1." Orientations 16.12: 46-61.
----. 1986a. "Sowing the Seeds of the Lotus: A Journey to the Great Pil-
grimage Sites of Buddhism, Part II." Orientations 17.2: 28-43.
----. 1986b. "Sowing the Seeds of the Lotus: A Journey to the Great Pil-
grimage Sites of Buddhism, PartIII."_JOrientations 17.3: 32-46.
----.1986c. "Sowing the Seeds of the Lotus: A Journey to the Great Pil-
grimage Sites of Buddhism, Part IV." Orientations 17.7: 28-40.
----. 1986d. "Sowing the Seeds of the Lotus: A Journey to the Great Pil-
grimage Sites of Buddhism, Part V." Orientations 17.8: 46-58.
----. 1987a: "Pilgrimage as Image: The Cult of the A$.tamahiipriitihiirya,
Part 1." Orientations 18.7: 55-63.
----1987b. "Pilgrimage as Image: The Cult of the
Part II." Qrientations 18.8: 56-68.
Huntington, Susan L., and John C. Huntington. 1990. Leaves from the Boddhi
Tree. Dayton: Dayton Museum of Art.
Huntington, Susan L. 1984. The Piila-Sena Schools of Sculpture. South Asian
Culture, v. 10. Ed. J. E. van Lohuizen-de Leeuw. Leiden: E. J. Brill.
----. 1990. "Early Buddhist Art and the Theory of Aniconism." Art
10uma149: 401-408.
----. 1993. "Aniconism and the Multivalence of Emblems: Another
Look." Ars Orientalis 22: 111-56.
Joshi, Lalmani. 1967. Studies in the Buddhistic Culture of India Delhi: Motilal
---.-,. 1991. "Religious Changes in Late Indian Buddhist History, Part l,"
Buddhist Studies Review 8.1-2: 45-64.
----. 1992. "Religious Changes in Late Indian Buddhist History, Part II."
Buddhist Studies Review 9.2: 153-68.
Karetsky, Patricia. 1987. "The Act of Pilgrimage and Guptan Steles with Scenes
from the Life of the Buddha." Oriental Art 33.3: 268-74.
Kielhom, F. 1892. "The Mungir Copper-plate Grant of Devapaladeva." Indian
Antiquary 21: 253-58.
----. 1896-97. "Khalimpur Plate of Dharmapaladeva." Epigraphica
Indica 4: 243-54
Kinnard, Jacob N. 1996. "Wisdom Divine: The Visual Representation of
Prajiia in Pilla-Period Buddhism." Diss. The University of Chicago.
Leoshko,Janice. 1987. "The Iconography of the Buddhist Sculptures of the Pilla
and Sena Periods from Bodhgaya." Diss. The Ohio State University.
----. 1988. "The Vajrasana Buddha." Bodhgaya, the Site of Enlighten-
ment. Ed. Janice Leoshko. Bombay: Marg Publications. 45-60.
----. 1993 / 94. "Scenes of the Buddha's Life in Pala-Period Art." Silk
Road Art and Archaeology 3:251-76.
Majumdar, Ramesh Chandra. 1971. History of Ancient Bengal. Calcutta: G.
Bharadwaj and Co.
Malandra, Geri H. 1993. Unfolding a Mm:u/.ala: The Buddhist Cave Temples at
Ellora. Albany: SUNY.
Matsunaga, Yukei. 1977. "A History of Tantric Buddhism in India with Refer-
ence to Chinese Translations." Buddhist Thought and Asian Civilization:
Essays in Honor of Herbert V. Guenther on His Sixtieth Birthday. Eds. Leslie
S. Kawamura and Keith Scott. Emeryville,CA: Dharma Publishing. 167-81.
Mitra, Rajendra Law. 1982. Sanskrit Buddhist Literature of Nepal. Calcutta.
Mukherji, R. R., and S. K. Maity. 1967. Corpus of Bengal Inscriptions Bearing
on History and Civilization of Bengal. Calcutta: K. L. Mukhopadhyay.
Mus, Paul. 1935. Borobuc/.ur: esquisse dune histoire du bouddhisme fondse
sur la critique archaeologie des textes. 2 vols. Hanoi: Imprimerie d'Extreme
Nakamura, Hajime, 1980. "The and the Chinese
and Tibetan Versions of a Text Similar to It." Indianisme et Bouddhisme,
Melanges offerts a mgr. Etienne Lamotte. Louvain: Institut Orientaliste de
l'Universite Catholique de Louvain. 259-65.
Parimoo, Ratan. 1982. Life of Buddha in Indian Sculpture (Ashta-maha-prati-
harya: An Iconological Analysis. New Delhi: Kanak.
Poussin, L. de la Vallee. 1908-26. "Tantrism." Hastings' Encyclopedia of Reli-
gious Ethics. Vol. 12. Edinborough.
300 nABS 19.2
Ray, Reginald. 1973. "Ma:Q.Qala Symbolism in Tantric Buddhism." Diss. The
University of Chicago.
Roth, Gustav, 1987. "The Physical Presence of the Buddha and its Representa-
tion in Buddhist Literature." Investigating Indian Art. Berlin: Museum fur
Indische Kunst. 291-312.
Sastri, Hirananda. 1923-24. "The Nalanda Copper-plate of Devapaladeva."
EpigraphicaIndica 17: 310-27.
----. 1942. Nalanda and its Epigraphic Material. Memoirs of the
Archaeological Survey of India 66. Delhi: Manager of Publications.
Samuel, Goeffrey. 1993. Civilized Shamans: Buddhism in Tibetan Societies.
Washington: Smithsonian Institution Press.
Schopen, Gregory. 1987. "Burial 'Ad Sanctos' and the Physical Presence of the
Buddha in Early Indian Buddhism: A Study in the Archaeology of Religions."
Religion 17: 193-225.
---. 1988. "On the Buddha and His Bones: The Conception of a Relic in
the Inscriptions of Nagarjunikoa." Journal of the American Oriental Society
108: 527-37.
---. 1991. "Archaeology and the Protestant Pressuppositions in the Study
of Indian Buddhism." History of Re ligions 31: 1-23
Snellgrove, David, ed. 1978. The Image of the Buddha. Paris: UNESCO.
Takakusu, J., trans. 1896. A Record of the Buddhist Religion as Practised in
India and the Malay Archipelago (A.D. 671-695); by I-tsin. Oxford: The
Clarendon Press.
Templeman, David. 1981. "Taranatha the Historian." Tibet Journal 6.2: 41-46.
Tucci, Guisippe. 1961. Theory and Practice of the Mandala. London: Rider
and Company.
Wayman, Alex. 1959. "Studies in Yama and Mara." Indo-Iranian Journal 3:
44-73 and 112-3l.
----,. 1,973. The Buddhist Tantras. New York: Samuel Weiser.
----. 1977 The Yoga of the Guhyasamajatantra. Delhi: Motilal
Williams, Joanna G. 1975. "Samath Gupta Steles of the Buddha's Life." Ars
Orientalis 10: 171-92.
Wintemitz, Maurice. 1933a. History of Indian Literature. Vol. 2. Calcutta:
University of Calcutta.
----,. 1933b. "Notes on the Guhyasamaja-Tantra and the Age of the Tan-
tras." The Indian Historical Quarterly 9.1: 1-10.
Woodward, Hiram, 1990. "The Life of the Buddha in the Pala Monastic Envi-
ronment." Journal of the Walters Art Gallery 48: 13-27.
The Moves MalJ.q.alas Make
In 1895, L. Austine Waddell listed "magic-circles" (mar.lC.;lalas) as being
among the "silly secrets," of Lamaism.! Clearly, the study of mal).<;lalas has
progressed a lot over the last one hundred years, and, judging from the
papers published here,2 it is continuing to evolve. Geri Malandra's
"unfolding" mal).<;lalas at Ellora, Charles Orzech's ritually "unrolling"
mal).<;lalas in Tang China, Jacob Kinnard's search for an "alternative man-
. <;lala" in Pal a-period iconography, and David Gardiner's exploration of the
overlap and "slippage" of mal).<;lala symbolism at Kayasan in Japan-all can
stimulate us to new definitional reflections. One of the thrusts of these
papers is to change the ways, or at least the contexts, in which we may
view mal).<;lalas by presenting them not just as well-ordered, clearly-delim-
ited, pre-defined "cosmograms," 3 but as configurations that are "on the
move" in more ways than one, and whose definitions must try to reflect
those motions.
This paper is a reworking of the "summary-response" I gave to the papers
presented at the symposium entitled "Mal).<;ialas on the Move: Caves, Monuments
and Icons Across the Buddhist World, ca. 750-850 C. E.," held April 21 and 22,
1995 at the Divinity School of the University of Chicago. I would like to thank
Frank Reynolds for organizing the conference and for inviting me to speak.
1. L. Austine Waddell, The Buddhism of Tibet or Lamaism (London: W. H.
Allen, 1895) 145.
2. All of these papers were presented to the symposium in Chicago. In addition,
there was a paper (not included here) by Julie Gifford on Borobudur. Each
paper was the focus of oral comments made by designated respondents, Robert
Campany, Richard Cooler, John Holt, Geri Malandra, and Jonathan Walters. In
what follows, I have tried to limit myself to a consideration of the four papers
here published, although sometimes my citations of them will reflect their origi-
nal versions presented to the symposium. Occasionally, also, I will make use of
or refer to insights expressed at the symposium by some of the other partici-
pants, for which I thank them. .
3. For a presentation of a number of such definitions of mal).<;ialas, see E. Dale
Saunders, "Mal).<;ialas: Buddhist Mal).<;ialas," The Encyclopedia of Religion, ed.
Mircea Eliade, vol. 9 (New York: Macmillan Co., 1987) 155-58.
302 JIABS 19.2
Of course, other scholars have been aware of the "movement" of
mar:t4alas, but, for the most part they have approached this topic within
meditation aI, psychological or doctrinal contexts. For them, mar;t4alas con-
tain polyvalent richness and multidimensional symbolism, but they are
more or less timeless, cultureless, and ahistorical "worlds in themselves." 4
The papers presented here, however; seek to place mar:t4alas themselves "in
the world." They take seriously a number of contexts that are often
neglected by other scholars. These include the historical, cultural, political,
geopolitical, art historical, architectural, physical, ritual, and textual milieux
which mar:t4alas reflect and effect. In them, the nature of mar:t4alas cannot
be separated from time, space, politics, world views.
My task, as I see it, is not so much to attempt to summarize these papers
as to respond to their stimulus. I want to do this by taking seriously the
title of the symposium and talking about the various ways in which the
mar:t4alas described in these papers are "on the move." I will center my
remarks around four topics each reflecting a different sort of motion. Sim-
ply put: (1) mar;t4alas themselves are moved; (2) mar:t4alas make moves by
rearranging relationships; (3) mar:t4alas move other things over distance
and time; and, fmally (4), there is movement within mar:t4alas.
(1) First of all, mar:t4alas can themselves be physically moved. The
paradigmatic example here is that of Kukai literally bringing the two Shin-
gon mar:t4alas with him upon his return from China in 806. The signifi-
cance of this act by this "first mover" is nicely put into its Japanese context
in David Gardiner's paper. Yet as Geri Malandra suggests, there were
probably other "Kukais" in other parts of the world, and the more local use
4. For a variety of examples of this genre, see Mircea Eliade, "Mar;t4ala," Yoga,
ImmortalitifInd Freedom (Princeton: Princeton University Press, 1969) 219-27;
Carl Gustav Jung, Mandala Symbolism, trans. R. F. C. Hull (Princeton:
Bollingen Foundation, 1972); Minoru Kiyota, Shingon Buddhism: Theory and
Practice (Los Angeles: Buddhist Books International, 1978), ch. 4; Michael
Saso, Tantric Art and Meditation (Honolulu: Tendai Educational Foundation,
1990), chs. 2-3; David L. Snellgrove, Indo-Tibetan Buddhism, vol. 1 (Boston:
Sharnbhala, 1987) 198-212; Adrian Snodgrass, The Matrix and Diamond World
Mandalas in Shingon Buddhism, vol. 1 (New Delhi: Aditya Prakashan, 1988)
121-23; Ryiljun Tajima, Les deux grands ma1Jejalas et la doctrine de resoter-
isme shingon (Tokyo: Maison Franco-Japonaise, 1959); Giuseppe Tucci, The
Theory and Practice of the Mandala., trans. Alan Brodrick (London: Rider,
1961); Taik5 Yamasaki, Shingon: Japanese Esoteric Buddhism (Boston:
Shambhala, 1988) ch. 6.
of movable maI).galas or mal)gaIa-like depictions by itinerant story-tellers
has a long history throughout Asia. 5 Mal)galas were thus part and parcel of
the spread of Buddhism. They could be transported across continents or
within communities, but in both cases their purpose seems to have been a
pedagogic one: they attracted attention and communicated meaning whether
at the popular or the esoteric level.
MaI).galas, of course, were not the only objects so moved and so used.
Kilkai himself brought back with him, along with the mal)galas and por-
traits of Chinese masters, various texts, ritual implements, Buddha relics,
and images. 6 Among these, the texts-Sanskrit manuscripts, Chinese
siitras and commentaries-comprised by far the largest portion. It is inter-
esting, therefore, to see what Kiikai has to say about the pedagogic power
of the mal)gaIas in comparison to texts. The passage from his "Memorial,"
quoted by David Gardiner, may be paraphrased here: Esoteric teachings,
according to Kiikai, cannot easily be expressed in writing, but they can be
completely captured and revealed in mal)galas, the simple sight of which
can bring one to Buddhahood.
Thus mal)galas, unlike texts, are efficient
tools that can more readily transmit at least esoteric teaching in toto. Truly
they are pictures worth a thousand words.
In this context, it may be interesting to compare, phenomenologically,
Buddhist stories about the transportation of texts with stories about the
movement of maI).gaIas. Texts seem eminently more likely to be subject to
loss or diminution in the course of transmission. Especially in the move-
ment from India to China (which, it might be pointed out, also necessitated
translation-a further erosion not necessarily afflicting mal)galas), texts
sometimes seem to have been incapable of remaining intact, whole. The
myth recounted by Charles Orzech is a case in point: the original Sar-
vatathagatatattvasal1Jgraha found in the Iron Stiipa was not transmitted to
China, but instead a summary was sent, an "outline" that itself was "broad
and long like a bed." But even this went overboard during a typhoon, and
5. For many examples of this oral genre, see Victor H. Mair, Painting and Per-
formance: Chinese Picture Recitation and its Indian Genesis (Honolulu: Uni-
versity of Hawaii Press, 1988); on the identification of Chinese pien-wen as
"maJ).c;lala-texts," however, see idem, Tang Transformation Texts (Cambridge,
MA: Harvard University Press, 1989) 59-60.
6. This according to his "Memorial Presenting a List of Newly Imported Sutras
and Other Items" (ShOrai mokuroku), trans. Yoshito S. Hakeda, Kakai: Major
Works (New York: Columbia University Press, 1972) 140-50.
7. See ibid., 145-46.
304 JIABS 19.2
what ended up in China was a summai:y of a summary. 8 In the process, as
Orzech points out; there occurred an "esoterization" of the text, the total
teaching of which could now only be attained through initiation. In light of
Kllkai's comments, it might be possible to think of this as a "maI).Q:alization"
of the text, a stripping of it down to its esoteric essence.
Because this
essence, according to KUkai, did not need elaboration and exegesis so much
as viewing, i. e., realization, it may, in fact, have been more easily movable
from one culture to another than exoteric traditions. This fact can help shed
further light on Geri Malandra's claim that maI).Q.alas were an important part
of making tantric Buddhism into an "international movement."
(2) Secondly, maI).Q.alas make moves by systematizing formerly disjunct
elements of the culture or the religion. They give unity to multiplicity, and
this in several different arenas. First of all, they synthesize in a new orga-
nized format structures that were nascent or not yet formed within the tradi-
tion. A clear example of this may be seen in David Gardiner's portrayal of
the role played by Kllkai and his newly imported maI).Q.alas in the process
of creating a new "systematized" esoterism in Heian times out of an older
"fragmentary" Nara esoterism. This was occasionally referred to at the
symposium as "putting old wine into new bottles," but it might more appro-
priately be conceived of as "bottling what had been loose in barrels."
A slightly more complex instance of this process may be found in
Charles Orzech's analysis of the systematization, under SubhakarasifI).ha,
Vajrabodhi and Pu-k'ung, of "piecemeal" mantra and dharalJ-l texts that
had been flooding into China. Here, in his view, the new maI).Q.alaic struc-
ture provided by the Sarvatathagatatattvasa11Jgraha was not single and
static but modular and adaptable to the different demands of ritual situa-
tions-a notion that some discussants at the symposium dubbed as
maI).Q.alas providing the basis for "ritual as jazz." 10 Something similar, per-
8. There are many such Buddhist legends. For another famous example,
Hsiian-tsang's loss of scriptures in a boat while crossing the Indus river on his
way back home, see Li Rongxi, trans., A Biography of the Tripi.taka Master of
the Great Cz"en Monastery of the Great Tang Dynasty, BDK English Tripijaka
77 (Berkeley: Numata Center for Buddhist Translation and Research, 1995) 156
(= T. 50: 249b, no. 2053).
9. It is interesting to recall here that one of the etymologies proposed for the
word "maI).Q.ala" is "seizing the essence." See Ferdinand D. Lessing and Alex
Wayman, Introduction to the Buddhist Tantrica Systems (Delhi: Motilal
Banarsidass, 1978) 270, n.l.
10. This point was frrst raised at the symposium by Robert Campany.
haps, may be found in the iconographic context at Ellora, in the different
"unfoldings" of the "shrine programs" described by Geri Malandra. And
yet another sort of synthesizing movement may be seen in Jacob Kinnard's
analysis of the various ways in which the eight scenes of the a ~ t a m a h i i
priitihiirya "mar;cjala," formalized during the Pala period, create a new
totality that "epitomizes" the whole of Buddhism.
But mar:u;lalas do not just synthesize previously disjunct elements witmn
Buddhism, they also serve to syncretize Buddhist and "non-Buddhist" reli-
gious traditions, as well as Buddhist and state ideologies. Clearly, Bud-
dhist esoterism, in its different cultural venues, interacted with a variety of
other religious traditions: Brahmanism, Taoism, and Shinto, as well as
"indigenous spirit cults." There was, of course, nothing new about this in
the Buddhist world. It was, rather, the continuation of a trend, but it did
affect the development of maI)Qalas. Thus we find, for example, that elabo-
rate compositions such as the Vajradhatu and the MahakaruI)adbatu
maI)Qalas have incorporated into their halls gods, goddesses, and godlings
that in a different context might be thought of as Hindu. More specifically,
at Ellara (where Hindu and Buddhist patterns replicated and reinforced each
other), Geri Malandra notes the presence of Bhudevi , the Earth goddess
who testifies to Sakyamuni's merits when he is challenged by Mara. This
is taken to be a reference to Bodhgaya and to the Buddha's conquest of
Mara there, and as such, it may be compared to the bhumisparsamudrii
images featured in Jacob Kinnard's discussion of the Pala period. But the
conquest of Mara and the witness of the Earth goddess may also be seen as
symbolic of the syncretism between Buddhism and indigenous divinities
that everywhere accompanied the establishment of Buddhism in new
10cales.1l This theme is perhaps less pursued in the papers dealing with
maI)Qalas in East Asia, but it did not go unnoticed by the discussants at the
symposium, several of whom emphasized the fact that the systematizations
taking place in Tang China and Heian Japan need to be considered not only
in the pre-existent Buddhist context, but in light of Buddhist-Taoist paral-
lels on the one hand, and various cults of Shinto kami on the other,12 The
question was posed, for example, of the relationship between maI)Qalas and
the nine-squared Ming Tang (Hall of Light), supposedly used by early
11. BhUdevi (Nang Thorani ) was to become especially important in Southeast
12. These points were particularly raised by Robert Campany and by Gary
.306 JlABS 19.2
Chinese rulers. 13 Elsewhere, Charles Orzech has suggested that Chinese
maI;lc,ialas should be seen in light of the "Taoist" Ho-t'u and Lo-shu cosmo-
grams as well as the Chinese philosophical categories of Ii and chih.14
A second type of syncretic movement going on in maI}.c,ialas is that
between Tantric Buddhism and state ideologies. Thus, as has often been
recognized, the rituals and structures of mal).c,ialas-in India, T'ang China,
and Heian Japan-were in part stimulated by and helped reinforce political
concerns. As Charles Orzech points out, two goals dominated Esoteric
Buddhism in T'ang times: the rapid attainment of enlightenment and the
protection of the state. Similarly, at Ellora, Geri Malandra suggests that the
depiction of royal figures might reflect "an early record of royal participa-
tion in amaI}.c,iala ... ritual."
At the same time, however, she goes further to propose that Ellora as a
whole, and the esoteric Buddhism it nurtured, be viewed in the geopolitical
context of a nascent "world system" that saw Indian Ocean civilisations as
part of an "entire structure" defined in part by Islam. This is an important
point. On the one hand, it may help correct the generally dominant percep-
tion of pan-Asian Buddhism as a continental culture with a view of it
being also a maritime tradition. On the other hand, it may alter the way in
which we see maI}.c,ialas themselves. From this perspective, maI}.c,ialas are
not only cosmograms; they also paradoxically embody both a totality that is
the whole and a totality that is "less than the whole." Because of their polit-
ical connections, they are graphs of palaces, capital cities, kingdoms,
empires, from a ruler's perspective. They are thus privileged places, and are
often marked off by gateways, boundaries, guardians. Socially, they might
be called "aristograms." In terms of state ideology, these very demarcations
set them off and mean that they have to be seen not just as "the whole cos-
mos" but as units operating within a greater world system. In this regard,
it may be '%9rth remembering that when the Lotus Sutra lists various types
of kings, It uses the word "maI}.c,ialin" not to denote a cosmic king, but to
identify a more minor ruler, one who governs only a single region. The
maI}.c,ialin, who controls a "maI}.c,iala," is thereby distinguished from a
cakravartin who rules all four continents (i. e., everything) and a bala-
13. For an interpretation of the Ming T'ang, see the still useful Marcel Granet,
The Religion of the Chinese People, trans. Maurice Freedman (New York:
Harper and Row, 1975) 67.
14. See Charles D. Orzech, "Chen-yen," The Encyclopedia of Religion, ed.
Mircea Eliade, vol. 3 (New York: MacMillan Publishing Co., 1987) 236.
cakravartin who rules over only one.1
A mm;tc;lala may constitute a world,
but it can also be a world within the world.
(3) A third set of moves made by mm;tc;lalas lies in the fact that they make
present the absent Buddha. By this I do not mean only that, as the Mahii-
vairocana sutra puts it, a mm;tc;lala is "what gives birth to all Buddhas," i. e.,
that it provides a matrix for enlightenment, for the development and matur-
ing ofbodhicitta.1
Mal).c;lalas do not only make meditators into Buddhas,
they also are able "magically" to transpose the Buddha through time and
space to the present here and now. I am using the word "magic" here in the
sense employed by Paul Mus in his Barabuij.ur. Mus's focus was on the
stilpa rather than the mal).c;lala, but he saw both as "mesocosms," a term he
created to mean a "place of passage," 17 "a magical structural milieu" 18 that
can "overcome the absence of objects or persons ... wherever they might
be or no longer be." 19 The primary "absent" person, of course, was the
Buddha in Nirval).a, a concept "so ungraspable that one can only propose
for it a formula bristling with contradictory negations."2o According to
Mus, it was in order to cut through this "excessively subtle logic" that Bud-
dhists turned to magical and cultic practices focused on mesocosms "so as
to assure themselves of satisfactions and certainties of an affective order,
without which there is no religion."21
The "mal).c;lala as mesocosm," however, would appear to "make the
Buddha present" in several particular ways. First of all, it can overcome the
barrier of distance by transposing to itself the site (and hence the event) of
the Buddha's enlightenment at Bodhgaya. This is most explicitly stated by
Geri Malandra who convincingly connects Ellora with Bodhgaya even
15. Hendrik Kern and Bunyu Nanjio, ed., Saddharmapur;.4arlka sutra (SC-
Petersburg, 1912) 6, 20, 363. Kern (The Lotus a/the True Law [Oxford: The
Clarendon Press, 1884] 7, 20, 343) translates "mm;tc;lalin" as "governor of a
region." Eugene Burnouf (Le Lotus de fa bonne loi [Paris: Imprimerie
Nationale, 1852], p. 4, n.1 has "rai d'un pays."
16. See on this, Snodgrass, 120, 122. The original citation may be found in T.
848,18: 5.
17. Paul Mus, Barabu4ur, vol. 1 (Hanoi: Imprimerie d'Extreme-Orient, 1935)
18. Ibid., 94.
19. Ibid., 74.
20. Paul Mus, "La mythologie primitive et la pensee de I'Inde," Bulletin de la
Societe Fran(aise de Philosophie 37 (1937): 91.
21. Ibid.
308 JIABS 19.2
calling it the "Bodhgaya of the South." But more generally, of course,
every maI).<;lala could be seen as the Vajrasana.
The transposition of Bodhgaya and other "central" Buddhist sites where
the Buddha had lived and taught to localities on the periphery was not an
uncommon phenomenon in the Buddhist world. Thus, according to certain
local traditions, the various sacred places of Magadha could be found
reduplicated in Gandhara,22 in Yiinnan,23 and in Southeast Asia.2
Even as
far away as Japan, the tradition developed that it was no longer necessary to
travel to India because the sites most intimately associated with the Buddha
(and thus the Buddha himself) could be found right there in Japan, at the
Sennyu-ji temple in Kyoto, for example,25 or at the Kasuga shrine in
These views recall the words of the Maharashtrian saint quoted by
Geri Malandra to the effect that there is no need to go anywhere, because
every place worth visiting-in this context, those places where the Buddha
can be found-has been made present right at home. MaI).<;lalas thus can
make pilgrimages unnecessary because they move the center to the periph-
ery, thereby transforming it into a place of the Buddha.
It may be, however, that that "center" is not always located on this earth.
Thus the Buddha's dwelling place could be thought of not as Bodhgaya but
as a Pure Land or as one of the heavens. Here too, however, the maI).<;lala
can serve to make present that distant site. Alex Wayman, for instance,
describes a text in which "the maI).<;lala can be understood to represent the
palace of the heaven, where . . . Gautama was initiated as a
Complete Buddha."27 By entering into the maI).<;lala, bodhisattvas could
thus find themselves in the heaven and receive instruction from
the Buddha himself (in his sarr;tbhogakaya).28
22. See Edouard Chavannes, "Voyage de Song Yun dans l'Udyana et Ie
Gandhara,";Bulletin de fEcole Franr;,:aise dExtreme-Orient 3 (1903): 381-87
23. See Paul Pelliot, "Deux itineraires de Chine en Inde a la fin du VIle siec1e,"
Bulletin de fEcole Franr;,:aise dExtreme-Orient 4 (1904): 161-62
24. See Michael Aung-Thwin, "JambudIpa: Classical Burma's Camelot," Con-
tributions to Asian Studies 16 (1981): 38-6l.
25. See John S. Strong and Sarah M. Strong, "A Tooth Relic of the Buddha in
Japan: An Essay on the Sennyu-ji Tradition and a Translation of Zeami's No
Play 'Shari'," Japanese Religions 20 (1995): 15
26. See Robert E. Morrell, "Passage to India Denied," Monumenta Nipponica
37 (1982): 192-93.
27. Alex Wayman, "Symbolism of the MaI).<;lala Palace," The Buddhist Tantras
(New York: Samuel Weiser, 1973) 9l.
28. Ibid.
Secondly, the "maJ.lqala as mesocosm" can overcome the barrier of time,
and here the directions it takes can be several. On the one hand, the orien-
tation may be to the past, and the maJ.lqala may act as the milieu for a trans-
position to the life and time of the Buddha Sakyamuni. Such a move lies at
the heart of the icono-conservatism described by Jacob Kinnard, with its
freezing in time the sequence of eight scenes from the biography of the
founder. This is akin to a "move to Bodhgaya" but the metaphor is tempo-
ral rather than spatial. On the other hand, the orientation may be to the
future, and the maJ.lqala may enact a projection to the presence of the
Buddha Maitreya who is yet to come. Such a move does not figure in the
papers published here but was much emphasized in Julie Gifford's presen-
tation to the symposium on "The Place of Maitreya in the Borobodur
MaJ.lqala." The same theme, perhaps, may be found at present day K6ya-
san where Kukai and many others lie waiting for Maitreya in the midst of
the maJ.lqala-like cemetery at the Okuno-in, or more explicitly at
Kasagidera, not far from N ara, where the original presence of the future
Maitreya was preserved in what is now called the "Kasagi mandara."29 At
the same time, of course, the maJ.lqala as mesocosm can eschew both the
past and the future and seek the Buddha in the "eternal," finding him in fig-
ures such as Aksobhya and / or Vairocana.
(4) Finally, we come to what might be called "movements within the
maJ.lqala." I want to address this by speaking in terms of several different
"oscillations," which are not unconnected to the movements we have dis-
cussed so far, but present them in a different light. These oscillations are
the result of a maJ.lqala's inherent multivalence: as one pole of significance
within the maJ.lqala is emphasized, there is a tendency for it be counterbal-
anced and replaced by the opposite pole.
The first oscillation concerns the question of the identity of divinities
within the maJ.lqala. Here I would like to return to the ambiguity pointed
out in Jacob Kinnard's paper between Sakyamuni and Aksobhya, both of
whom are represented as figures in bhumisparsamudra.
Kinnard cites
Nancy Hock's view that this is an "intentional ambiguity," a "dual nature"
reflective of a transitional stage epitomized by Sakyamuni in a tantric
29. See Karen L. Brock, "Awaiting Maitreya at Kasagi," Maitreya, the Future
Buddha, ed. Alan Sponberg and Helen Hardacre (Cambridge: Cambridge Uni-
versity Press, 1988) 214-47.
30. The same ambiguity between Sakyamuni and Aksobhya can be found in
Southeast Asia, as was pointed out by Richard Cooler at the symposium.
.310 JIABS 19.2
There is thus some oscillation here between Tantra and non-Tantra
within the malJ.Q.ala. This could also be viewed as an oscillation between
different bodies of the Buddha or different levels of understanding: in a
certain sense, Sakyamuni is in another he is not. Ill' a slightly
different context, it is similarly possible to state that all the divinities in the
two great Shingon maI).Q.alas-the Taizokai and the Kongokai-are finally
none other than Vairocana (Dainichi Nyorai) who occupies the center of
both of them. Here too, there may be constant opportunities for oscillation.
For example, in the first level rite mentioned by David Gardiner,
the kechien kanjo meant to "establish a bond" between the initiate and a par-
ticular deity of the malJ.Q.ala, the blindfolded candidate must toss a flower
onto the surface of the malJ.Q.ala; the particular divinity on whom the flower
lands thereby takes on a special importance and is established in a tutelary
relationship (embodied in an esoteric name) with the initiate. 32 In the midst
of the ritual, however, at least in the Tendai tradition, as soon as the flower
is thrown, the name "Dainichi Nyorai" (Vairocana) is shouted out by the
attendant priests, regardless of what particular divinity or bodhisattva the
flower falls on. This is because all deities in the malJ.Q.ala are Vairocana,33
A second oscillation may be found in the structure of the Kongobuji at
Koyasan as described by David Gardiner: was the plan for one stUpa or
two stUpas? Was the intention to map out on the landscape of the mountain
two malJ.Q.alas in distinction or overlapping as one?34 The ambiguity, of
course, may be due to such mundane matters as the lack of funds or mate-
rials for buildings, but it nicely captures an oscillation that is perhaps also
summed up in the Japanese expression "ryobufuni," meaning the two
paired maI).Q.alas are "not two"-i. e., not different from each other. This
31. See Nancy Hock, "Buddhist Ideology and the Sculpture of Ratnagiri," diss.,
UniversityBf California, Berkeley, 1987.
32. For brief descriptions of the ceremony, see Tajima, 271; Snodgrass, 732.
33. lowe this detail of "oral tradition" to the late Michel Strickmann, with whom
I visited the hall set up for the abhi$eka in Yokawa, on Mount Hiei, in 1977. It
may also be due to the fact that these initiations can be seen as repetitions of
Kl1kai's abhi$eka. When Kl1kai was initiated, his flower landed on the image of
Dainichi Nyorai both times-in the Taizokai and the Kong6kai maI).Q.alas. See
Snodgrass, 732.
34. A similar oscillation may be found in the mountains of the Kii peninsula
where, for Japanese yamabushi, Mount Kimpu (near Yoshino) represents the
Kongokai, the Kumano peaks represent the Taiz6kai, and Mount Ornine repre-
sents the superimposition of both maI).c;ialas. See Carmen Blacker, The Catalpa
Bow (London: Unwin, 1986) 211-12.
paradoxical assertion, of course, reflects an oscillation of major importance
to the whole Tantric tradition: the twin emphasis on Wisdom and Compas-
sion. Thus the Taizokai mm;u:;lala is traditionally taken to show the unfold-
ing womb of the Buddha's compassion and the Kongokai maI).<;iala the
penetrating pqwer of his wisdom. But even within the Kongokai, there can
be a double motion: a centrifugal one spinning out compassion from the
center and a centripetal one seeking wisdom by spiralling inward.3
where, as in the case of the monument of Borobudur, this double motion
may be thought of as ascending (to enlightenment) and descending (to this
world of suffering). 36 In each instance, there is a realization within the
motion of the maI).<;iala of the duality-yet-unity of NirvaI).a and safI).sara, that
forms such a crucial part of Mahayana Buddhism and of Tantra in particu-
1ar. One might recall here David Snellgrove's distinction between the
"vertical core" of the maI).<;iala in which stages of realization and reality are
distinguished, and the "horizontal maI).<;iala" which is "an idealized represen-
tation of the identity of n i r v ~ a and safI).sara."37
Finally, these papers bring to light a third kind of movement within the
maI).<;iala that might be described as an oscillation between the inside and the
outside. Simply put, in this context, what one finds when one penetrates
into the center of the structure is the structure. Some of this may be found
in Geri Malandra's paper in which "the illaI).<;iala is at Ellora," and "Ellora is
in the maQ.<;iala," but the full dimensions of this oscillation are perhaps best
seen in David Gardiner's discussion of the relationship of stupa and
maI).<;iala. At Koyasan, what one finds inside the maQ.<;iala is the srnpa and
what one finds inside the srnpa is the maQ.<;iala. In so far as the stupa may
represent the Buddha's absence and the maI).<;iala his "active presence in our
world," this also represents an oscillation between absence and presence.
But other elements-texts and bodies-may enter into this game of tantric
leapfrogging. The Great Srnpa at Koyasan is sometimes identified with the
Iron Srnpa in India which, as was pointed out, was the source of texts that
were the source of maI).<;ialas, but which here in Shingon is also "this very
body in which one realizes Buddhahood" (sokushin jobutsu), which
35. See Daigan and Alicia Matsunaga, Foundation of Japanese Buddhism, vol. 1
(Los Angeles: Buddhist Books Internationa.l, 1974) 189.
36. This was one of the insights from Julie Gifford's paper not published here.
37. D. L. Snellgrove, The Hevajratantra, vol. 1 (London: Oxford University
Press, 1959) 29.
312 JIABS 19.2
Buddha body is a stlipa, which is a malJ.<;lala. Truly the word "malJ.<;lala"
means a "circle" in more than the usual sense of the term. 38
One could probably keep going round and round on the meaning of
maIJ.<;lalas which themselves are not static entities, and which are in constant
interaction with the contexts that inform them. But enough has been said, I
think, to show that there is much to be learned from the approaches taken in
these four papers. Though dealing with very different situations in very
different parts of Asia,39 they nonetheless make it possible to think of 8th-
9th centuries in pan-Buddhist terms.
38. The model for all this may well be an Indian one rooted in Brahmanical tra-
ditions of sacrifice, in which the sacrifice itself creates the gods, creates the
priests, creates the texts, creates the structure that creates the sacrifice . .. The
necessity of thinking about maIJ.<;lalas in the context of srauta rites was empha-
sized by John Holt at the symposium.
39. It should be noted that no paper was presented to the symposium on
malJ.<;lalas on the move in Tibet or in Mainland Southeast Asia or in Nepal, a fact
that was noted and regretted.
Treasurer's Report 1994
Beginning Balance
Dues $24,127.67
Bank Credit
I.B.S. 3825.00
D. Lopez Advance 3357.00
L. Lancaster Adv. 1044.44
Wisconsin Secretary of State
Bank Debits
Bank Account Fee
Electronic Transfers
Returned Checks & fees
Final Balance
Signed: Lewis Lancaster, Treasurer
US $3933.35
US $8717.03
Treasurer's Report 1995
Beginning Balance US $8717.03
Dues and Subscriptions $ 9880.00
Bank Interest 47.67
1994 Royalties 113.41
Back Issues 1044.50
TOTAL income 11085.58
Postage - billing, shipping 435.00
Photocopying; office costs 105.93
Typesetting, Postage [16-17] 4578.14
Printing [vol. 17.2] 3043.61
Postage [vol. 17.2] 774.09
Typesetting [voL 18.1] 2190.00
Printing [vol. 18.1] 3012.38
Postage [voL 18.1] 1465.35
[Subtotal 15604.50]
Bank Fees
~ a n k F e e s 35.00
Check Printing 27.50
Returned Checks & Fees 40.00
[Subtotal 102.50]
TOTAL Expenses 15707.00
Final Balance US $4095.61
Signed: Joe Wilson, Treasurer