Anda di halaman 1dari 31



time of death he asked to be bathed and dressed in clean clothing, following which he sat up facing west, entered into meditative concentration, and
died. Purple clouds hovered overhead, and people came running to witness
his demise. These are all familiar events in the description of the death of a
Buddhist saint in East Asia.
The portrait sculpture of Eison was transported to the cremation grounds
during his funerary rites and, following the cremation, was returned to his
living quarrers at Saidaiji; there, a nun assigned to the task made daily offerings to it. T. Griffith Foulk and Robert H. Sharf describe similar funerary and


Robert H. Sharf

memorial rites used for Ch'an abbots in China, in which "the portrait served
not only to represent the spirit of the deceased abbot, but also, in some sense,
to embody it." 108 Eison's image, likewise, embodying the spirit of the deceased
master, would serve his spiritual descendants as an object of veneration, and
a source of inspiration, from this time onward.

Traditional biographies of Eison situate him among the monks who strictly
observed the precepts, a characterization that would suggest a rather con-

One of the truisms in the study of East Asian Buddhist Tantra is that the

servative figure who wished to return to traditional forms of monastic life.

depictions of deities associated with Tantric practice-notably the often com-

My investigation into Eison's use of images reveals a side of Eison's person-

plex geometric arrays of divinities known as mandalas (fig. 4.1 and pl. ?)-

ality and religious practice absent in many accounts. I have noted the influence

function as aids for visualization practices. Such practices, which are purport-

of Esoteric Buddhism in Eison's practice, as is evident from his worship of

edly the mainstay of Tantric Buddhist meditation, are understood as exercises

Aizen, and have argued that his devotion to Sakyamuni was based on his view

in which the practitioner attempts to construct an image of the "principal

of Sakyamuni as a still-vital presence in the world, rather than as the long-

deity" (]: honzon ::;$: Jl!lf.) associated with a given rite in the "mind's eye." ~c
complishment at visualization is regarded as an essential step in the realiza-

departed founder of the precepts.

Most important for the theme of this volume, I have found that figures

tion of the ultimate identity of the practitioner and the principal deity. Since

such as Aizen, Sakyamuni, and Maiijusri were not mere transcendent deities,

the principal deity is invariably declared to be a manifestation or emanation

of Mahavairocana Buddha, and since Mahavairocana is none other than the
dharmadhatu (J: hokkai '$. !ff- ), or absolute truth itself, identification with the

much less abstract ideals, for Eison. Rather, these buddhas and bodhisattvas
were physically present in the world in sacred images that were enlivened
through the incorporation of relics, scriptures, and dharm:zl and through rites
of consecration and periodic invocations and offerings. Finally, the treatment

In another rubric, it is the affirmation of inherent "buddha nature." The goal

of Eison's own image, brimming with relics, scriptures, and personal docu-

of Shingon mikkyo it ~ ("esoterism"), thus understood, is consonant with the

ments, suggests that it was also more than a mere likeness; it was a living force
that was expected, after the monk's demise, to safeguard the order he founded

aims of Mahayana soteriology in general.

The claim that mandalas function as aids for visualization can be found in

and to remain the locus of Eison's enduring presence.

a wide variety of works on Shingon and mikkyo mandalas. In his study of the

principal deity is tantamount to the realization of the absolute within oneself.

Shingon fire sacrifice (goma 8&~), for example, Richard Payne writes that one

---~--------.~--------------------------------------Visualization and Mandala


characteristic feature of Buddhist Tantra is the use of "images, both paintings

and sculptures, as a part of ritual and as the objective base for visualization." 2
Ishida Hisatoyo, the preeminent Japanese scholar of East Asian Tantric mandalas, is even more explicit: "It is ... extremely difficult to perceive the Buddha in one's head by concentrating one's thoughts. For this reason the image
mandala [gyiizii mandara] , in which objects or statues are placed or painted
on the altar, was created to help the devotee experience the depths of contemplation and perceive the Buddha." 3
Statements to this effect are ubiquitous in the scholarly literature.4 Yet
rarely, if ever, do scholars bother to substantiate the claim with historical or
ethnographic evidence. The notion that mandalas function as aids for visualization seems to be one of those truisms so widely and unquestioningly held
that corroboration of any kind is deemed unnecessary. Indeed, the complex
epistemological problems entailed in the use of the term "visualization" are
rarely, if ever, acknowledged.
As I began my own study of Shingon ritual systems I was, accordingly, surprised to find that Shingon rituals themselves offer little support for this view.
To begin with, neither the manuals used for the performance of major Shingon
rituals (shidai !.-\~, Sk: vidhi), nor the available "oral commentaries" (kuden

Taizokai mandala (Sk: garbhadhiitu marzrfala, matrix-realm

mandala ). r693. One of a pair of hanging scrolls (Ryogai mandara, mandala
of ~~e two worlds!,mk and colors on silk, 410.9 x 37.9.0 em. each. Kanjoin,
Kyoogokokup (Top), Kyoto Prefecture. Photo: Ky66gokokuji.

D j.~ ) associated with various lineages or "streams" (ryii 7ftU of Shingon mikkyo, instruct the practitioner to use mandalas in such a manner.5 This might
not be significant were it not for the fact that explicit instructions are provided for the use of virtually every other piece of ritual paraphernalia arrayed
on and about the altar.
Even more striking is the fact that there is little obvious correlation between
the elaborate graphic detail of the major Shingon mandal;1s, on the one hand,
and the content of the specific rites with which they are associated, on the
other. Finally, the commonly accepted understanding of "visualization" -the
notion that Shingon rites involve fixing a technicolor image of one or more
deities in the mind's eye-is borne out neither by an examination of the ritual
manuals, nor by ethnographic evidence pertaining to the utilization of such
Accordingly, my goal in this chapter is to raise some problems concerning the claim that (I) Shingon meditative practices center on the mental construction or inner visualization of mandala-like images, and (2) Shingon mandalas are used as aids in visualization exercises. I will hereafter refer to both



Visualization and Mandala

claims under the rubric of the "phenomenological model," because they are
enmeshed in an approach to the subject that privileges the "inner experience"
of the practitioner over the performative and sacerdotal dimensions of the rite.
I intend to cast doubt on the veracity of the phenomenological model by focusing on a sequence of Shingon initiations known as the Shidokegy6 1Z9 Ji 1JD 1T,
or "four emancipatory practices." More specifically, my analysis will concentrate on the first two of these rites, namely the J uhachido -t- f\ i]~t( eighteenmethods practice) and the Kongokai 1iZ:: ~lj W (vajra-rea!m practice). My selection of these two complex rituals out of the dozens commonly performed in
Shingon monasteries is not arbitrary: the Juhachido not only is the first major
ritual to be mastered by Shingon initiates but also is paradigmatic for virtually
all others; all subsequent ritual initiations are structured as variants or expansions of the "eighteen methods." The Kongokai is the second major practice
to be undertaken. Because it is associated with the elaborate Kongokai mandala, this practice is particularly well suited to serve as a test case for the relationship between rite and icon. Despite the need to restrict the scope of this
study to this small corner of the Buddhist Tantric universe, I expect the methodological issues raised here will be of some interest to scholars working on
other Buddhist ritual traditions as well.

Aids to Visualization
I certainly do not want to imply that the use of images or physical objects as
foci for meditation is foreign to Buddhism. Indeed, a host of Theravadin concentration exercises, as systematized in the Visuddhimagga for example, use
natural or fabricated objects as the subject of meditative visualizations, and
there is no reason to doubt the antiquity of such practices.
Typical of the Theravadin exercises are the kasir;a meditations, which use
the four "physical elements" as a means of concentrating the mind and attaining absorption (jhima). The meditation on the earth kasir;a, for example,
involves the use of a smooth clay disk "the color of the dawn." Such a disk
can be fabricated on a portable canvas made by "tying rags or leather or matting onto four sticks," or alternatively constructed on a fixed spot "made by
knocking stakes into the ground in the form of a lotus calyx, lacing them over
with creepers." Having made ready the kasir;a, the practitioner cleans the surrounding area, takes a bath, seats himself on a raised chair a short distance
from the disk, and proceeds to develop a mental image of the earth disk. He


does this by alternately gazing at the disk, then trying to visualize it with eyes
closed. In order to keep the mind focused he may repeat a word to himself
that characterizes the earth element, such as "earth, earth." Eventually the disk
will appear with eyes closed exactly as it appears with eyes open, .at which
point he is to withdraw to his own quarters and continue to practi~e th~re.
Should his concentration flag, however, he must return to the place m which
the kasina is installed and begin again? Exercises on the other kasi1:za (water,
fire, air: blue, yellow, red, white, light, and "limited space") are developed in
much the same manner: in each instance the practitioner begins by using a
natural or fabricated object in order to develop a mental image.8
Kasir;a practices are merely one group of concentration techniques that use
physical objects as supports; one might also mention the meditations on foulness (asubhabhavanii) in which the monk meditates on a corpse 111 one of ten
specified stages of decomposition. 9 Once the adept fixes the mental image, he
uses the image to attain meditative absorption in much the same manner as a
kasina is used.
Neither the kasir;a practices nor the meditations on foulness ever became
popular in China. However, from the dawn of Chinese Bud~hism devo~t Buddhists meditated upon Amitabha and his Pure Land, occasionally makmg use
of iconographic depictions of Amitabha to engender faith and to inspirevisions.w Such practices, generally subsumed under the heading "recollection
of the Buddha" (C: nien-fo ~f~g, J: nenbutsu, Sk: buddhanusmrti), may well
have evolved from earlier exercises in which devotees would meditate on the
qualities of the Buddha, a form of which is still practiced in Theravadi.n countries today. 11 In any case, scholars of East Asian Pure Land often depict such
practices in terms reminiscent of the "phenomenological model":
Amongst meditative types of buddha-reflection we may distinguis~ a form
called buddha-contemplation [kuan-fofkanbutsu]. This is the practice of
gazing upon an image or painting of Amitiibha or his land until a mental .
image of this can be retained in the mind's eye when the eyes are closed. Th1s
vision is then developed-or develops-into a presence of the actual Buddha such that we may describe it as a "buddhophany," a manifestation or
appearance of the actual Buddha. 12
It would move this discussion too far afield to chronicle the development and

significance of Buddhist visualization techniques in East Asia. Suffice itt~ say

that references to such practices are found in Buddhist literature associated
with a variety of traditions, sects, and cultural spheres. All of this would ap-



Visualization and Mandala

pear to render the phenomenological model-the notion that Tantric mandalas are used as supports for visualization exercises-unproblematic.
Indeed, there are unambiguous references to the use of painted and sculpted
images as props for meditation in a number of Chinese Tantric scriptures and
commentaries. Typical is the following passage from fascicle u of the Ta p'i-

lu-che-na ch'eng-fo ching shu :*!Ji~f&#U~iF!iE, I-hsing's - ff (683-727)

commentary to the Ta-jih ching :};: 8 *~ : B
When the practitioner first cultivates the skillful means of the contemplation .m! of the super-mundane he begins by contemplating the principal deity.
Relymg on a painted image fi'< il: 13! he contemplates thoroughly. At first
he attains illumination Bfj ) ' with eyes closed; later, gradually opening his
eyes, he perceives [the deity and the deity becomes) fully manifest and illumined without any obscurity. [However] this is not yet absorption.14 Should
he attain this [illumination] the mind will give rise to both faith and joy,
and from this faith the mind is purified. Gradually he attains the [stage of]
nonduality wherein the mind is free of all craving and attachment. But nonduality must also be relinquished in order to attain the apprehension of the
non-distinction between the middle and the extremes. When all mental conditions have been relinquished the myriad dharmas are equal and the same;
thts moment is called the mark of absolute equality '{ff : ;f. But [prior to
this], when the various marks are still manifest, although there is not yet consummation, one is still able to gradually attain consummation. Therefore this
is called the stage wherein there is not yet absorption 5f;: '{ff iJ J '(QY


Shingon Ritual
The Shingon ritual tradition comprises a systematic if staggeringly complex system of invocation rites centered upon particular deities or families of
deities. In order to better understand the nature of Shingon "visualization"
we must turn directly to this ritual system as delineated in ritual manuals and
liturgies, rather than rely on the theoretical and ideological formulations of
the scholastic literature. Before turning to specific rites, however, a word of
introduction is necessary for those unfamiliar with the structure of Shingon
When Kiikai '1:iffl (774 - 835) returned to Japan in 8o6 after spending
slightly less than two years in China, he carried with him a substantial body of
ritual manuals and implements associated with a multitude of invocation and
empowerment (kaji IJOt#) rituals. 18 He and his disciples are credited with systematizing this vast body of procedures and instituting a series of four initiations, the Shidokegyo or "four emancipatory practices," as a common course
of training for all Shingon monks. 19 These rituals, mastered during the course
of an ascetic retreat often lasting one hundred days or more,20 constitute what
is in effect a primer of ritual grammar, wherein the novice acquires proficiency
in the underlying syntax and basic lexicon of the Shingon system.
The essential structure of the rites is mastered in the first practice, the Juhachido or "eighteen methods." 21 A tenth-century ritual manual for the eighteen

This passage goes on to describe the process of contemplation, wherein the

image of the deity is manifest before the adept in fine detail, and the body of
the practitioner becomes one with the body of the principal deity.
Such passages from the hands of authoritative Chinese masters, insofar as
they explicitly enjoin the practitioner to "rely on a painted image," would
seem to offer unequivocal support for the phenomenological model. However, explicit directives countenancing the use of an image to assist in Tantric
"visualizations" are not as common as one might suppose in the East Asian

methods articulates a sequence of seventy-one procedures; these component

rites form the underlying structure for all subsequent initiations.22 An additional fifty to eighty rites are added with each of the three major rituals that
follow the eighteen methods, namely the Kongokai (vajra-realm practice), Taizokai nii ~:W (matrix-realm practice), and Goma (fire sacrifice). By the completion of the Shidokegyo and the anointment that follows (J: kanio 7il/H~ , Sk:
abhi~eka), the adherent has learned literally hundreds of ritual segments, comprising a lexicon that is ideally at his or her beck and call.

Buddhist corpus. One must also be cautious lest one ascribe undue authority
to 1-hsing's remarks; although there are scriptural precedents for such injunctions/6 1-hsing's commentary is nonetheless a scholastic compendium whose
function vis-a-vis Tantric practice is prescriptive rather than descriptiveY The

All Shingon invocation rituals are structured around a narrative concerning a visit by an honored guest (the principal deity) who is entertained and

evidence provided by Shingon ritual manuals, as well as the ethnographic data

and treats a distinguished visitor. 23 The narrative functions as an overarching schema, lending coherence and narrative unity to the various isolated and
often fragmentary units that comprise the rites. The schema is clearly manifest

bearing on the performance of Shingon rites, suggests that the major Shingon
mandalas were not necessarily employed in the manner described by I-hsing.

feted by the host (the practitioner). According to tradition, this narrative has
its roots in ancient Indian customs governing the manner in which one receives

in the sequence of rites comprising the Jiihachido.

Visualization and Mandala




The ritual day is broken into three periods, each marked by a single performance or "sitting" (ichiza gyobO - Rl~ f) !1;) of the Juhachido. 24 Prior to each
performance of the rite, the practitioner prepares the altar (mitsudan

\1 ),

setting out in carefully prescribed fashion fresh flowers, food (in the form of
uncooked rice), candles, and censers prepared with pressed incense powder.
The practitioner further prepares six small cups (rokki 7\'i* ) containing arrangements of cut shikimi ;jW; leaves and water that will be offered to the visiting
gods. (O ne pair of cups represents holy water [aka ~A ihO ] for bathing the feet of
the gods, one pair represents unguents [zuko w~ ], and one pair a garland of
flowers [keman :ijt~ [.)The practitioner then purifies him- or herself by washing the hands, rinsing the mouth, and putting on clean robes prior to entering
the practice hall.
What follows is an overview of the Juhachido ritual following a traditional
method of parsing the ritual into nine sections. 25 This outline takes as its
model the Juhachido used in the Sanboinryu .=::::. "f,)'f ll"fc

m{ initiatory line; the dif-

ferences between the rites as performed by one Shingon lineage and another
are, however, minor. 26


" Procedure for Adorning the Practitioner" (shogon gyoja ho iff~

ff*ii};). The ritual begins with a series of procedures that purify and
empower the practitioner. One enters the hall chasing away demons
with mudrii and mantra and then imagines it overflowing with tathagatas. One ano ints one's body with incense and uses a variety of rites,
each consisting of a mudrii, mantra, and contemplation, to purify mind
and body.
"Samantabhadra's Vows'' (Fugen gyogan hO 1-Wl\I fl"!li.fii~) . This section
opens with a sequence of rites that purify and empower the ritual implements themselves. Thereupon various and sundry divine spirits (kami
tEP) are invoked in two rites: the "declaration of intent" (hyohyaku
:Nl::.l) and the "supplication to the spirits" (jinbun kigan t$5tNrm!) .
In the former, done only during the first performance of the rite, the
practitioner declares his or her mission to the kami and solicits their
assistance. The latter consists of specific requests, covering everything
from appeals for the salvation of all beings to entreaties for the health
of the emperor and seasonal rains. Each request is followed by offerings
consisting of the recitation of siitras (done in most cases by intoning
the title alone), and the recitation of the names of buddhas and bodhisattvas. This section closes with vows of repentance and refuge and
further purifications.


"Procedure for Binding the [Sacred] Realm" (kekkai ho k:i5 W7:1; ). This I
.sts of a set of three rites that prepare the sanctuary for the arnva
templat1ons the
of the deity. Using ritual gestures, mantras, an con
practitioner drives a pillar from his or her seat to the center of t~e cart
. .t "l.mlnobile, and an indestrucnble vafra wallis erected
ren d enng '
h. h the
around the four sides of the sanctuary. The roof, through w IC

principal deity will later enter, is not yet sealed.

- - - . - . , .. .
" Procedure for Adorning the Sanctuary" (shOgon dofO ho }lt/l~QrFl~ 1~).
h d --k ">!;rail (contemplatiOn of the
This consists pnmanly of t e OfO an m ?JJ '

sanctuary), an elaborate meditation on the principal detty m his pure
land. l will examine this segment in some detail below.

"Procedure for Inviting the Deities [into the Sanctuary]" (ka~jo ho

d d contemplations to
Wl~i~). This segment uses mu ra, mantras, ai1
dispatch a carriage to pick up the principal detty and his ret:u~, ennce
them into the carriage, and convey them to the sanctuary. T e emes

are greeted with applause.

6 " Procedure for Binding and Protecting [the Sanctuary]" (ketsugo ho
. ~,3fii -~) Here one undertakes a series of procedures that complete
~~;' p~ep~ration of the sanctuary by ritually sealing it off from the outd
ld A va,ra net covers the roof and th.e entire sanctuary IS surSI e wor
_.. . :a
h d d K nrounded by a ring of fire. Bato Kannon .\\;,pfHil:i (Horse- ea e . a
non), a wrathful incarnation of Avalokitc5vara, guards the precmcts.
7 "Procedure for Offerings" (kuyo hO {Jt i'-Hi). Various offerings are then
made to the guests. Following what is taken to be Indian custom, one
. b offering water to wash the deities' feet. The demes are then
b egms Y
d h
1c Th1s
offered lotus thrones to sit upon and are entertame w~t mus .
is followed by individual offerings of mcense, unguents, flowers, foo '

drink, light, eulogies, and prayers .

. ,
170- ~ ~,.11r,) This is the centerp1ece
8. " Procedure for RecitatiOn (nenftt
,;,:, ,;h ' .
of all Esoteric rituals proper, in which one reahzes the ldennty of
the practitioner and the principal deityP It consists of three proce- .
dures, punctuated by the invocation of the prmclpal deity through hts
mantra( s). Some exegetes hold that this section of the nte effects a
merger of the principal deity and the practitioner through the successive identification of the body, speech, and mind of the pracnnoner
with the body, speech, and mind of the deity. H owever, It can also be
as the ritual enactment of the fundamental unity or nonduahty of
" mewhat misleadseen
deity and practitioner, rendering the term merger so
.. .
ing. The " procedure for recitation" is broken down mto the followmg



a. "[The Deity! Enters Me and I Enter [the Deity]" (nyuga-ga 'nytl

Jdltf.ltA.) . This enacts the merger or essential identity of the body
of the principal deity and that of the practitioner. I will examine it
in some detail below.
b. "Empowerment of the Principal Deity" (honzon kaji ::$: ~1JDt1' ), also
known as the "three mudrii and mantra of the principal deity" (honzon sanshuin shingon '* lliL:::=. ;flltC/1[\\f! 'i'). In the Jiihachido of the Sanboinrytl, in which Nyoirin Kannon ~[J;\it!JiAAIIDHf(Cinramanicakra
Avalokitesvara, fig. 4.2) is the principal deity, this segmen~ consists of multiple recitations of Nyoirin's three mantras: the "great
mantra" (daiju )( )"[ ), the "middle" (or " heart") mantra (chtiju i:j:J )"[ ,
shinju {.' '}1:) , and the "heart within the heart mantra" (shinchushinju
{ .' rj:J{,- !t ), along with a specified mudrii for each.
c. "Formal Recitation " (shonenju TL~:~~j) . This contemplation focuses
on the second of the "three mysteries," that of speech. It involves
the manipulation of an ornate rosary reserved exclusively for the
performance of this rite. This rosary is first carefully removed from
its lacquer box and put throu gh a series of purifications and empowerments. The practi tioner then makes 108 slow repetitions of
the " heart mantra " of the principal deity, manipulating the rosary
in such a fashion that his hands form the mudrii known as " teaching of the law" (seppo no in f.5t it- .L E/1 ). The mantra is accompanied
by an elaborate contemplation wherein the syllables of the mantra
are imagined to circulate through the body of the principal deity,
emerge from his mouth, enter the head of the practitioner, circulate
through his body, emerge from his mouth, and enter the belly of the
principal deity, where the process begins all over again.

Visualization and Mandala


tion of the rite can take up to one hour to complete, that is, one-half
to one-third of the total duration of the ritual.
9 "Final Offerings " (go kuyo j~{J:li1lt) . The ritual winds down with a
repetition of many of the offerings and purifications found in the first
seven sections of the rite, although they are subject to considerable abbreviation. Near the close of the performance the ritual seal around
the sanctuary is broken, and the barrier of fire, the net over the roof,
and the vajra wall are removed in the reverse order in which they were
erected. The deities are sent on their way and the goshinbo ~ !4- i:l;
(protection of the body) purifications found in the opening section
are repeated.
The outline above is sufficient to give a general idea of the structure of
the Juhachido. At the completion of the Juhachido retreat, which typically
lasts four weeks, one moves on directly to the Kongokai practice. The underlying structure of the Kongokai is precisely that of the Jiihachid(>, although the
overall sequence has been considerably expanded by the addition of dozens of
o ther ritual segments (see below). Furthermore, the liturgical content of many
segments common to the Jiihachido is altered to reflect the fact that the principal deity is now Kongokai Mahii.vairocana. The result is a more complex
and longer procedure.
After the Kongokai the practitioner moves on to the Taizokai, also structured as an expansion of the Jiihachido. After four weeks spent on the Kongokai and another four on the Taizokai, one comes to the final and most arduous
rite of the Shidokegyo, namely the Goma or fire ritual. The Shidokegyo Goma
consists of a five-tiered fire sacrifice (it involves five rounds of offerings) de-

d. "Empowerment of the Principal Deity" (as above).

voted to Fudo Myoo ;p wh DH :, which is inserted whole into the Juhachido

e. "Contemplation of the Syllable Wheel" (jirinkan :f:ili!iijl!). This

represents the third of the " three mysteries," that of the identity of
the mind of the practitioner and the mind of the principal deity. Due
to the importance of this segment, which is arguably the climax of
the entire ritual, I will look at it roo in detail below.

sequence in the midst of the sannenju, or "dispersed recitations."

f. "Empowerment of the Principal Deity" (as above).

g. "Dispersed Recitations" (sannenju ~~;{t.\r!i). This is variously interpreted as "miscellaneous recitations," "scattering recitations," and
so on. In the Jiihachido manuals of both the Sanboin and Chiiin lineages it involves the repetition of twelve mantras of varying lengths,
each repeated anywhere from seven to one thousand times. This sec-

Upon completing the four Shidokegyo initiations the practitioner is eligible to receive denbo kanjo lWititiH~ (consecration of dharma transmission )
making him (or, more rarely, her) a "master," or ajari


(Sk: iiciirya).

Although this chapter focu ses on Shingon mikkyo, that is, the tomitsu


tradition of Japanese Esoterism, it is worth mentioning that the taimitsu iJZfi

Shidokegyo initiations used in the Tendai school are essentially identical, the
main difference being that in Tendai the Taiz<Jkai practice precedes the Kongokai.28 In both the Shingon and Tendai schools it is highly unusual to perform the Jiihachido, Kongokai, or Taizokai after receiving denbO. An abbreviated form of the Goma, by contrast, is performed regularly in Shingon and



Visualization and Mandala

Tendai monasteries. Should a practitioner wish to continue the ascetic training, he or she may seek formal initiation into one of the many supplementary mikkyo practices such as the rishukyo bo l'.fU!Ull!: $ or the komyoshingon
ho JtB~~~ $.

Contemplation of the Three Mysteries

As should now be apparent, a Shingon ritual invocation is comprised of dozens, sometimes hundreds, of small ritual segments. Some of these are primarily
liturgical recitations such as the jinbun kigan (supplication to the spirits), the
go daigan E.:km'i (five great bodhisattva vows), or the eko ~lol (transference
of merit). Such liturgies are composed in classical Japanese or Chinese and
are typically performed with palms joined together in gassho 15 ~ or with
hands ceremonially clasping the egoro M~ 1P (a hand-held censer) and rosary.
Other ritual segments consist of combinations of a mudrii and mantra known
as inmyo ED B~ (literally "mudra-mantra"). The segments that are of primary
concern to this study, however, consist of three elements performed more or
less in unison: (1) a mudra; (2) an utterance, usually a mantra or dharat;l but
sometimes a verse, vow, or prayer; and (3) a "contemplation." These "tripartite rites" constitute the core of the Shingon ritual system: they incorporate
the body, speech, and mind of the practitioner and thus ritually instantiate
the "three mysteries" (sanmitsu =:.;t) of Shingon doctrine. When one talks of
Shingon "visualizations" one is thus properly referring to the third component of these tripartite rites-the ritual procedure linked to the "mystery of
mind" (imitsu :W:\t ). In order to understand the nature of Shingon visualizations we must focus our attention on this specific component of the tripartite
I will begin with a relatively straightforward rite, the sanmikkan =:. ;til?,, or
"contemplation of the three mysteries," which appears in virtually all major
Tantric rituals in the Sanb6inryu tradition. This segment is performed near
the beginning of each sitting, after the practitioner has assumed his seat in
the sanctuary, made final adjustments to the altar, and anointed his body with
unguents. The instructions read as follows:
Place the palms together to form the lotus blossom mudra. Then imagine ;@
that in between the palms and on top of the tongue and the heart there is a
moon disk. On the disk is an eight-petaled lotus blossom, on top of which is
the syllable un. The syllable changes and becomes a five-pronged vajra emit-


ting rays of light that destroy the defilements and impurities of body, speech,
and mind. Intone the syllable un three times.29
Although relatively simple, the sanmikkan "contemplation" or "visualization"
contains many characteristic elements found in more elaborate tripartite rites.
First, note that the contemplation is introduced by the term so~' "to think,"
"contemplate," "imagine," and so on. In fact, a wide variety of terms are used
in Shingon materials to refer to the "mental component," or imitsu, of the rite,
including kan 1!?,, kanso il.:t!\t, kansatsu il.~, teikan ~IlL kannen il.~, nenso
~~ .and shii ,~, '\t. There is little to be gained by trying to distinguish the
meaning of these terms in English, as Sino-Japanese Buddhist lexicons state
that they are used more or less interchangeably-a position quickly borne out
by a survey of Shingon ritual texts (see below).
Secondly, one finds that the contemplation of the sanmikkan is not static,
but rather consists of a series of changing or mutating images: the final image
of a vajra emitting light evolves from a "seed-syllable" (shuji ~ r, Sk: bija) sitting on a lotus blossom set upon a lunar disk. Such seed syllables figure prominently in many of the contemplations, especially in cardinal segments such
as the dojokan and the jirinkan. Finally, note that the contemplation includes
a "discursive gloss," specifically the statement that the rays of light shining
from the vajra "destroy the defilements and impurities of body, speech, and
mind." I characterize this phrase as "discursive" because it does not suggest
any obvious visual or pictorial correlate.
Indeed, as will become evident below, "visualization" is a dubious choice
for an English equivalent of terms such as kanso and kannen. These technical
Sino-Japanese terms refer to procedures whose elements are often more discursive, literary, or tropical than they are visual or graphic. Accordingly, in
the discussion below, in which I examine considerably more elaborate segments drawn from the Shidokegy6, I will avoid using the terms "visualize"
and "visualization" in favor of "think," "imagine," "contemplate," "discern,"
and so on.

Contemplation of the Sanctuary

Certain segments of the Shidokegyo practices entail rather elaborate contemplations, including the dojokan, or "contemplation of the sanctuary," found
in virtually all Shingon invocation rituals. The dojokan occurs in the first half
of the ritual sequence, after the practitioner's body, the ritual implements,



and the sanctuary have been individually purified and empowered, but prior
to summoning the principal deiry to the sanctuary. The dojokan consists of a
meditation on the principal deiry and his retinue emerging from a series of
seed syllables. Since the contemplative element of this rite includes a graphic
description of the principal deity in anthropomorphic form, this is a good
place to begin to examine the relationship between a ritual procedure and a
painted mandala.
Before the practitioner begins the Juhachido retreat he or she must prepare
(more technically, "adorn," shogon H:t frW ) the sanctuary and its altar. As part
of these preparations three icons (typically paintings) are installed in front of
the altar: an image of the principal deity flanked by a portrait of Kukai on the
right, and a lineage patriarch on the left. The identiry of the principal deity
and lineage patriarch will differ depending on the lineage; the Chuinryu i:f:l


popular on Mount Koya, for example, uses Dainichi Nyorai

1c 13 t.!D*

(Mahavairocana) as the principal deity, and the Sanboinryu stemming from

Daigoji ~W!ctr in Kyoto uses Nyoirin Kannon (fig. 4.2). 30 Sanboinryu monks
will thus place an image of Nyoirin directly in front of the altar, with Kukai
to the right, and Shobo ~}!\' (Rigen Daishi f.\I!j}jj\:f;:~iji , 832-909) on the left.3 1

The pictorial depiction of Nyoirin used in the Sanboinryu will ideally conform to the description of the deity found in the dojokan of the Sanboinryu
Juhachido manual, which reads as follows:
Assume the "tathagata fist mudrii" [nyorai kenin !ZD*$E[J] .... Contemplate ll!\I:i'2: as follows: In front [of me] is the syllable ah (]: aku). The syllable
changes into a palatial hall of jewels. Inside is an altar with stepped walkways on all four sides. Arrayed in rows are jeweled trees with embroidered
silk pennants suspended from each. On the altar is the syllable hrlh (kiriku)
which changes and becomes a crimson lotus blossom terrace. On top is the
syllable a (a) which changes and becomes a full moon disk. On top is the
syllable hrlh (kiriku), and to the left and right there are two triih (taraku) syllables. The three syllables change and become a vajra jewel lotus. The jewel
lotus changes into the principal deity, with six arms and a body the color of
gold. The top of his head is adorned with a jeweled crown. He sits in the posture of the Freedom King (jizai o tti:), assuming the attribute of preaching the dharma. From his body flow a thousand rays of light, and his upper
torso is encircled by a radiant halo . His upper right arm is in the posture of
contemplation. His second right arm holds the wish-fulfilling gem. His third
right arm holds a rosary. His upper left arm touches the mountain [beneath
him]. His second left arm holds a lotus blossom. His third left arm holds a

F 1 G u R E . 2 Nyoirin Kannon (Sk: Cintamanica kra Avalokitesvara). Twelfth

century. Panel, ink and color on silk, 98.5 x 445 em. Courtesy, Museum of
Fine Arts, Boston (Fenollosa-Weld Collection :rr.4032). Reproduced w1th
permission. 2000 Museum of Fine Arts, Boston. All rights reserved.



wheel. His magnificent body of six arms is able to roam the six realms, employing the skillful means of great compassion to end the suffering of all sentient beings. The eight great Kannons and the innumerable members of the
Lotus-realm assembly surround him on all sides.
When finished with this contemplation take the mudra and empower the
following seven spots [that is, touch each place on the body with the mudra]:
left knee, abdomen, right knee, heart, forehead, throat, and crown. [Recite)
the mantra ombokken (seven times [along with each empowerment)). (This
mudra, mantra, and empowerment transforms this world into a pure realm.) 32
This is one point in the Ji.ihachido where there is an undeniable correspondence between image and kanso or "contemplation," where one seems to end
up with the practitioner imagining, albeit after a complex series of morphing
images, a golden, bejeweled, six-armed deity whose form precisely matches
the iconography of the mandala (fig. 4.2). But it would be a mistake to make
too much of this correspondence. Here, too, the contemplation culminates in
a discursive gloss, namely the statement, "His magnificent body of six arms is
able to roam the six realms, employing the skillful means of great compassion
to end the suffering of all sentient beings." The painting on the altar makes
no more reference to this gloss than it does to the sequence of images leading up to the anthropomorphic appearance of the deity. Moreover, none of
the many premodern and modern ritual manuals and commentaries at my disposal direct the practitioner to cast his gaze on the icon of Nyoirin Kannon
during this (or any other) sequence. Even if the practitioner did attempt to use
the painting to help him in his contemplation, it would ultimately be of limited value in his efforts to visualize this complex montage of rapidly mutating
1mages and discursive contemplations.
Ethnographic evidence points to yet another problem with the phenomenological model: the model presupposes that practitioners approach the text of
the dojokan kanso as a "guided contemplation." When one looks at the manner
in which contemporary Shingon monks actually perform this rite, however,
one finds that the kanso is treated liturgically-it is intoned quietly or vocalIzed mwardly. This is a crucial point: the execution of the kanso consists not
in "visualization" or even in "meditation" so much as in recitation. And even
if the practitioner did want to linger over or meditate upon the content of the
liturgy he would find himself severely constrained by the need to finish the rite
within the time allotted. I will return to this issue below.

Visualization and Mandala


Moving Through the Vajradhiitu Mandala

Although Shingon materials prescribe the use of an image of the principal
deity during the Ji.ihachido, they remain silent on the relationship between
image and rite. The situation with the Kongokai or "vajra-realm practice,"
however, is considerably more complex, because there is in fact a tradition
according to which the Kongokai ritual moves the practitioner through the
nine assemblies of the Kongokai mandala that is used for the rite (pl. 7).33 At
first glance this might seem to constitute evidence for the phenomenological
model, but once again a close examination suggests otherwise.
The locus classicus for this tradition is a tenth-century commentary by
Gengo :lC* (914-995), the Kongokai kuemikki :11z:~IJWfL\tit ~c.34 According to Gengo's analysis, the movement through the mandala takes place in a
counterclockwise spiral toward the center (fig. 4.3), beginning with the gozanze sanmaya e ~~=ill: =~!jj) \t (Trailokyavijaya-samaya assembly, fig. 4.4)
in the lower right corner of the mandala, and ending with the jojin-ne I& 5l \t
(consummate body assembly, fig. 4.5) in the center. (There is an alternative
but less common tradition wherein the practitioner begins with the central assembly and moves around the mandala in a clockwise movement ending in the
bottom right corner-a movement known as the "descending rotation" [geden
rilii] in contrast to the "ascending rotation" [iaden ..tilii]. I will return to this
briefly below.) 35 Each assembly of the mandala is associated with one rite or a
short sequence of rites, beginning with gokki sanmaya f.ti ~ = ~!jj) (samaya of
ultimate bliss)-a "tripartite rite" that occurs well into the ritual sequence (see
below). Thus, according to Gengo, the performance of the gokki sanmaya corresponds to the gozanze sanmaya e (Trailokyavijaya-samaya assembly, fig. 4.4);
and the next rite-the gozanze ~~=ill: (Trailokyavijaya)-corresponds to the
next assembly in the mandala, the gozanze e ~~ _::: ill:lt (Trailokyavijaya assembly, fig. 4.6). The remaining correlations run as follows: the dairaku fukushin
::*:~1'~ 5l moves one to the rishu e i!~\t (rishu assembly, fig. 4.7); thegoso
jojinkan li fEll& 5l il. (contemplation of the attainment of the [buddha] body
through the five marks) corresponds to the ichiin-ne
fl (assembly of the
single seal, pl. 8); the shibutsu kaji 031?#1JD:fi!f (empowerment of the four Buddhas) moves one to the shiin-ne 03Enl't (assembly of the four seals, fig. 4.8);
the gobutsu kanjo li {?#itrn (anointment of the five Buddhas), shibutsu keiman
031?#Jit. (offering garlands to the four Buddhas), katchu Ffl ~ (armor and
helmet), and ketchu *5 ~ (fastening armor and helmet) correspond to the kuyo

- En


R 0 BERT H . S H A R F


Four Seal Assembly


tfl 1:.?
Single Seal Assembly

Rishu e

Rishu Assembly

f----.----+---------+---- t - - - END
Kuy(i e



fj~ $1-

Offering Assembly

Consummate Body



f--__ ---+---- ----+---- t _____,

Misai c

Subtle Assembly


Sanmaya e

Gozanze sanmaya e

Samaya Assembly

TrailokyavijayaSamaya Assembly


4. 3

"Ascending rotation" through the Kongokai mandala.

tf$T (offering assembly, fig. 4.9); the genchishin

fest knowledge) begins the misai e

half of the di5j6kan
the sanmaya e

B~dlT5 tf$T

')W !'} (body of mani-

(subtle assembly, fig.; the first

(contemplation of the sanctuary) corresponds to

(samaya assembly, fig. 4-II); and the latter half of

the same rite corresponds to the central j6jin-ne (consummate body assembly,
fig. 45).36
However, neither the directions found in the ritual manuals nor the content
of the contemplations or the recitations of the Kongokai ritual give the least
hint to the practitioner that the rites he or she is performing bear any relationship to the assemblies of the mandala. The correlations enumerated above
can be ascertained only from written or oral commentaries to which the student may or may not have access; there would be simply no way to determine

Trailokvavijaya-samaya a_ssembly (gozanze samnayaf e), from
44 mandala.
' KanJom,
._. Kyoogo
--, ko kuj.i(Ti'lji) ' KyotoPreecture.
Photo: Kyc>Dgokokuji. Detail of plate 7

FIGuRE 4 . .5 Consummate body assembly ( j6jin-ne), from the Kongokai

mandala. Kanj<)in, Kyo<)gokokuji (T6ji), Kyoto Prefecture. Photo:
Ky66gokokuji. Detail of plate 7

F 1 c; u RE 4. 6 Trailokyavijaya assembly (gozanze e), from the Kong6kai

mandala. Kanj6in, Ky66gokokuji (T6ji), Kyoto Prefecture. Photo:
Ky66gokokuji. Detail of plate 7

F r G u R E 4. 7
Rishu assembly (rishu e), from the Kongokai mandala.
Kanj6in, Ky()6gokokuji (T6ji), Kyoto Prefecture. Photo: Ky66gokokuji.
Detail of plate 7

FIGURE 4.8 Assembly of the four seals (shiin-ne), from the Kong6kai
mandala. Kanj6in, Ky66gokokuji (T6ji), Kyoto Prefecture. Photo:
Ky66gokokuji. Detail of plate 7

FIG u RE 4. 9
Offering assembly (kuyo e), fro m the Kongokai mandala.
KanjcJin, Ky6ogokokuji (Toji), Kyoto Prefecture. Photo: Kyoogoko kuji .
Derail of plate 7

F 1 G u R E 4 . I o Subtle assembly (misai e), from the Kong<>kai mandala.

Kanj oin, Kyoogokokuji (Toji ), Kyoto Prefecture. Photo: Ky66gokokuji.
Detail of plate 7

Visualization and Mandala 177

these correlations through a comparison of the iconographic content of the
mandala with the liturgical content of the ritual.
To illustrate the incongruity between rite and mandala I will examine the
first of the above correspondences in detail, that is, the relation between the

gozanze sanmaya assembly of the mandala (fig. 4.4) and the rite known as
sanmaya or "samaya of ultimate bliss." The gazanze sanmaya assembly, situated in the bottom right corner of the mandala, consists of the samaya form
of the Trailokyavijaya assembly that is placed immediately above it. In other
words, where the Trailokyavijaya assembly consists of seventy-three deities
represented in anthropomorphic form, the samaya version of the same assembly represents these deities by symbols such as a lotus, vajra, pennant,
or various weapons, placed on a lotus within a lunar disk. The principal figures comprising both the Trailokyavijaya and the Trailokyavijaya-samaya assemblies are the five Buddhas (Mahavairocana, Ak~obhya, Ratnasambhava,
Amitabha, and Amoghasiddhi), the four piiramitii bodhisattvas, and the sixteen great bodhisattvas-that is, the same figures that appear in the central

jojin-ne group (fig. 4.5). There are, however, a number of differences. For example, Vajrasattva, who normally appears in Ak~obhya's group, here manifests in the wrathful form of Trailokyavijaya Vidyaraja. In addition, the four
"Mantra Queens" replace the tridents of wrath ((unnu-sanko ;tt~ -=:. ~) in
the corners, and the four gods of the elements are replaced by lotuses .37 The
buddhas and bodhisattvas of the gazanze sanmaya e are understood as manifest in their wrathful forms, represented in the gazanze e by the use of the

funnuken in ft~~ ED (wrathful fist mudrii) in which the deities' hands are
clenched into fists crossed at the wrists. These two assemblies are the only
ones in the Kongokai mandala that depict buddhas, bodhisattvas, and other
deities in their wrathful forms-forms assumed by the deities in order to conquer greed, hatred, and delusion in the three worlds. However, as the deities
of the gazanze sanmaya e assembly are represented in samaya form, this assembly ends up looking rather similar to the samaya asse mbly situated directly to
its left (fig. 4-n) .
FIG u_ R E 4 ~ r_ Sa maya ~:sembly (sanmaya e), from the Kongokai mandala.
KanJom,. Kyoogokokup (I C>JI), Kyoto Prefecture. Photo: Kyoogokoku 'i.
Detad of plare 7

Doctrinally, the gozanze sanmaya e represents, in general terms, the "mind

aspect" of the wrathful emanation of Mahavairocana, capable of vanquishing
evil and bringing particularly recalcitrant beings to enlightenment. Snodgrass
summarizes the technical literature as follows:
The Trailokyavijaya Assembly represents the actions of the [Buddha's] Body
that vanquish the beings who are difficult to subdue and the Trailokyavijaya


Visualization and Mandala


Samaya Assembly represents the original vows that embody the actions of the
Buddha's Mind. The former is the mandala of the Tathagata's Body Mystery
expressed as the wrathful forms of the Doctrine Command Cakra Body, and
the latter is the mandala of the Tathagata's Mind Mystery expressed in the
same way. 38
The gokki sanmaya rite- the rite said to correspond to the gozanze sanmaya e
and its seventy-three major divinities-is short. As is usual with the tripartite
rites, the manual begins with a description of the mudra: "The two hands are
clasped with fingers on the outside, and the tips of the little fingers and thumbs
touching and extended out straight. The middle fingers are crossed and folded
into the fist, and the mudra is held in front of the face. The two middle fingers pierce the heart, as the arrow of great compassion." Having formed the
mudra, one then repeats the mantra sanmayakoku sorata satoban and contemplates as follows: "Not wearying of sarrzsara I abide peacefully in the mind of
awakening. I shoot the arrow of great compassion into the mind that wearies
and seeks escape [from sarrzsiira]." The mantra is repeated three times, and
with some experience the entire rite can be performed in no more than ten or
fifteen seconds. 39
Note that there is simply no due in the rite itself as to its relationship with
the gozanze sanmaya e of the mandala, and the contemplation of the "arrow
of compassion piercing the mind that wearies of sarrzsiira" bears no obvious
connection to the iconography of the Kongokai mandala. Indeed, this is really
not a "visualization" at all-the "arrow of great compassion" is again more
of a literary trope than a graphic image, and neither the ritual manuals nor
the commentaries offer any clue as to how the practitioner is to visualize the
"mind that wearies and seeks escape [from sa'!lsiira]."
Another instructive example is the simplest of the nine assemblies, the
ichiin-ne (pl. 8), found at the top center of the mandala. This assembly consists of a single anthropomorphic image of Mahavairocana sitting on a lotus,
wearing a crown of five buddhas, hands clasped in the "knowledge fist mudrii"
(chi ken in ~~ED), surrounded by water pots and lotuses. 40 One might expect
this assembly to correspond to the dojokan examined above. But such is not the
case; according to Geng6's commentary, the rite corresponding to the ichiin-ne
assembly is the goso jojinkan (contemplation of the attainment of the [buddha]
body through the five marks), a complex meditation involving, among other
things, a dialogue between the practitioner and all the tathagatas of the world,
and various contemplations of an expanding and contracting stupa. 41 There is


no obvious allusion to Mahavairocana in anthropomorphic form at all in this

particular ritual sequence. Indeed, later commentators admit confusion over
the textual antecedents, iconography, and ritual procedures associated with
this particular assembly. 42
It would take too much space to examine in detail all nine assemblies of
the Kongokai mandala and the ritual segments to which they supposedly correspond. Suffice it to say that in each and every case the practitioner would be
hard pressed to match the liturgical content of the rite to the iconography of
the mandala. I have already mentioned an alternative geden tradition, in which
the practitioner moves through the mandala in reverse (clockwise) direction
beginning with the jojin-ne assembly in the center, and ending at the gozanze
sanmaya e in the bottom right-hand corner. The existence of two alternative
paths through the mandala underscores the post hoc nature of the correlations
between rite and image. Indeed, as Todaro has argued, Gengo's attempts to
relate the Kongokai rite with the nine-assembly mandala was likely a response
to the recognition of a more general problem: the manuals then in circulation
did allude to the deities of the Kongokai, but only to those associated with six
of the nine assemblies. Gengo's strategy involved viewing the preeminent assembly-the jojin-ne-as encompassing the remaining eight, and he developed
a creative if ad hoc series of correspondences to make his point.43 Gengo's
opaque if not outright arbitrary scholastic correlations are ultimately of little
consequence to this analysis, however, because there is no hint in either the
ritual manuals or the ethnographic evidence that they play any effective role
in the performance of the rite itself.
This is not to deny a connection between the Kongokai rite and its mandala. There is one section of the ritual that does bear a clear and comprehensive relationship to the iconography of the mandala. This is the sanjushichison
ingon == -t !fED (mudrii-mantra for the thirty-seven deities), a sequence
consisting of the invocation of the thirty-seven figures featured in the central
jojin-ne (fig. 4 .5).44 Here each of the thirty-seven main deities is individually
invoked in rapid succession by means of the single repetition of a short mantra
and mudrii. The sanjushichison ingon sequence is immediately followed by the
invocation of the "sixteen worthies of the [present! auspicious kalpa" (gengo
juroku son WM+7\!f),45 and the twenty deities of the "outer vajra realm"
(gai kongobu niju ten 7Hi~l5iJijili) _-:::. +7(),46 which complete the invocation of

the central assembly.

The sanjushichison ingon is thus the one segment of the Kongokai rite where


Visualization and Mandala


one finds a more or less precise correspondence to one of the assemblies of

the mandala. However, somewhat surprisingly, this is one of the few ingon or
inmyo segments that does not include any kanso or "contemplative" component whatsoever.47 Moreover, the invocations in the two segments devoted to
the sixteen worthies and twenty deities are perfunctory to say the least; none
of the individual figures are specifically identified. Rather, the mantra un is
repeated sixteen times for the worthies, then twenty times for the deities, to
the accompaniment of a single up-and-down motion of the mudra with each
repetition. The contemplation, or kanso, for both rites reads in its entirety as
follows: "Think of encircling the four sides of the altar."
It should now be clear that the relationship between the complex contemplations found in the Juhachid6 and Kongokai practices, and the iconography
of their respective mandalas, is tenuous at best. The same turns out to be true
of the latter two rituals of the Shidokegy6: the Taiz6kai and the Goma.48 In
each case the mandala or icon represents or "embodies" the principal deity
to whom the rites are directed, but this fact alone does not render the icon of
much help to a practitioner bent upon "visualizing" the dozens of elaborate

kanso scattered throughout these rites.

Visualization or Contemplation?
In the end, the most compelling evidence against the phenomenological model
is provided by the content of the kanso themselves. As seen above, the discursive character of many Shingon contemplations renders them poor subjects for
"visualization" as the term is commonly understood. This becomes increasingly obvious in looking at the more complex and soteriologically cardinal
sections of these invocation practices. Accordingly, I have translated below
the kanso from three such segments: the goso jojinkan found exclusively in the
Kong6kai practice, and the nyuga-ga 'nyu and jirinkan found in virtually all
Shingon invocation rituals.


blossom situated on a lunar disk in the heart, as well as more unusual passages in which the practitioner converses directly with the tathagatas of the
dharmadhatu. Because this is one of the more fascinating sections of the Kong6kai from both a narrative and doctrinal perspective, I have translated the
Sanboinryu version in full:

First: The Wisdom of Marvelous Contemplation IJ'J; il]ll'*'IJ!i':

Form the Amitabha samadhi mudra and say the mantra on sanmaji handomei kiriku.
Contemplate [as follows] ~~liM: The nature of all things gl it.~tf. arises
from one's own mind. The afflictions, defilements, aggregates, realms, all the
sense fields, and so on are like a magical illusion or a flickering flame, like
the castle of the Gandharvas, like a whirling wheel of fire, like an echo in an
empty valley. [Thereupon] all the Buddhas who abide in ultimate and true
knowledge of quiescence, extinction, and sameness, and who pervade and fill
the realm of space startle [you] awake by snapping their fingers and saying:
" Good son, that which you have realized is the purity of the single path, but
you are not yet able to realize the knowledge of the vajra-like samadhi and
full omniscience ~~;6 , and thus you must not rest content with this. You
must still perfect the way of Samantabhadra and attain supreme awakening."
Next: Penetration of the Bodhi Mind iifii~Ft!fi~{} (meditation mudra):
Hearing this the practitioner is startled out of his meditation and pays
homage to all the Buddhas reciting the mantra of universal homage as before
[on saraba tatagyata hannamannano kyaromi]. [The practitioner then addresses
the Buddhas:] "I only wish that all the Tathagatas would appear here in my
place of practice." All the Buddhas respond in unison: "You must contemplate II your own mind. In accordance with the teachings when discerning
your own mind do not view \!. its distinguishing marks." [The practitioner]
recites the mantra of universal homage once again and says to the Buddhas:
"I am unable to see my own mind. What are the distinguishing marks of this
mind?" All the Buddhas respond : "The distinguishing marks of the mind
are difficult to fathom. The mind is like a moon disk in a fine mist." Say the
mantra: on a sowaka; on shitta harachibeito kyaromi.



Nowhere is the discursive nature of Shingon kanso more apparent than in

the "contemplation of the attainment of the [buddha] body through the five
marks," or goso jojinkan found in the Kongokai practice. 49 This sequence involves some standard contemplations of an expanding and contracting lotus

Next: Cultivation of the Mind of Bodhi ~~-fftitL' (meditation mudrii):

The storehouse consciousness is essentially unstained, pure, and without
blemish. Because it is endowed with merit and knowledge, one's own mind
is like a full moon. Seeing the pure moon disk one is able to realize the bodhi
mind . Say the mantra: on bochishittahadayami.



Next: Consummation of the Vajra Mind JJ.)Z{Jzjljlj{,- (meditation mudra):

Again all the Buddhas say: "Now that hodhi is stable, receive this heart
mantra and imagine 1m a vajra lotus blossom; that is, imagine an eightpetaled lotus blossom on the moon disk at the heart." Say the mantra: on
chishuta hazara handoma.
Next: Expanding the Vajra }ii![{!z:l&jlj (meditation mudra):
Think ~ of an eight-petaled lotus blossom on the heart moon disc.
Gradually it opens and expands to fill the three-thousand realms all the way
to the dharma realm bringing material abundance and joy to all living beings.
Say the mantra: on sahara hazara.
Next: Contracting the Vajra ~{!z:l&jlj (meditation mudra):
Think of this lotus blossom at the heart gradually contracting and becoming smaller, returning to its previous size. Say the mantra: on sokara
Next: Realizing the Vajra Body ~{Jz:~lj !it (meditation mudra):
Know that one's own body is the realm of the vajra lotus blossom. All the
Buddhas pervading the dharma realm enter the lotus blossom of one's own
body, just as myriad images appear in a [single] mirror. Say the mantra: on
hazara handoma tamakukan.
Next: Perfection of the Buddha Body m!it lim TJM] (meditation mudra):
All the Buddhas further say: "Contemplate your body as the principal
deity and receive this mantra." Say: on yata saraha tatagyata satatakan.
Next: The Empowerment of All the Buddhas ~mt.m~ (meditation mudra):
Having become the body of the principal deity I receive the empowerment
of all the Tathagatas; all the worthies of the vajra realm envelop me. Say the
mantra: on saraha tatagyata hisanhoji jiricha hazara
The textual antecedents and doctrinal significance of the "contemplation of
the attainment of the [buddha] body through the five marks" has been examined by a variety of Japanese Buddhist scholars, but these complex issues need
not be of concern here. 51 For the purposes of the present argument I would
note only the elaborate narrative structure and discursive content of the rite
that renders much of it inappropriate for "visualizing in the mind's eye." It is
particularly striking that the Tathagatas featured in this contemplation do not
appear in physical form; according to the narrative they make their presence

Visualization and Mandala


known only through their spoken discourse. Indeed, the gist of their teaching is that one "sees" the Tathagata only when one is able to "see" the nature
of one's own mind-a venerable Buddhist tenet found in a host of Mahayana

The climax of virtually all Shingon invocation practices is found in the "procedure for recitation" (nenju ho) section comprising three main rites: the nyugaga'nyu ([the deity] enters me and I enter [the deity]), the shonenju (formal recitation), and the jirinkan (contemplation of the syllable wheel). As mentioned
above, these three rites correspond to the three mysteries of body, speech, and
mind respectively; that is to say, these rites enact the identity of the body,
speech, and mind of the practitioner with the body, speech, and mind of the
principal deity.
Considering the significance of the nyuga-ga'nyu-it is here that the practitioner affirms his or her identity with the Buddha or absolute truth itself-the
kanso of the rite is relatively simple. Since the rite is short and differs slightly
from one practice to the next, I have translated the nyuga-ga'nyu from both
the Jiihachido and Kongokai practices taken from the same set of Sanboinryii
ritual manuals. I begin with the Jiihachido version:
Assume the Amida meditation mudra ...
Contemplate [as follows] !1~: The principal deity sits on a mandala.
I sit on a mandala. The principal deity enters my body and my body
enters the body of the principal deity. It is like many luminous mirrors
facing each other, their images interpenetrating each other !zO ~ BR. tEl
f.t.B:~:mi~ A th. 52

Commentators note that the body of the principal deity does not actually
"enter" the practitioner, as there is ultimately no original duality between the
two. This is seen by some commentators as the significance of the mirrors: one
must look upon the principal deity as if one is looking at one's own reflection
in a clear mirror.53
The contemplation from the Kongokai nyuga-ga'nyu adds an interpretative
gloss reflecting further on the relation between practitioner and deity:
Assume the meditation mudra.
Contemplate [as follows) !1~: Facing the principal deity I have now become the body of Tathagata Mahavairocana. The principal deity enters my



body empowering me. I enter the body of the principal deity raking refuge
m h1m. We are of one body, not two. Because [this rite] manifests the meaning of both the root and the traces :+ liE, it constitutes a contemplation of
empowerment and refuge.54

Visualization and Mandala


rained. Because the speech of the syllable ha cannot be obtained, the ceaseless
flow of the syllable on cannot be obtained. (This is called the counterclockwise contemplation :i!/!W?.. Contemplate the sequence clockwise and counterclockwise.) 56
According to Shingon doctrine, the mantra of a particular deity is consubstan-

With the jirinkan, or "contemplation of the syllable wheel," one arrives at the

tial with the deity itself (that is, it bears the same relationship to the deity that,

climax of the rite-the ritual identification of the mind of the practitioner and

according to certain Mahayana scriptures, the title of a sutra bears to the sutra). Thus to grasp rhe mantra is to grasp the deity, just as to intone the title of
a siitra is tantamount to reciting the entire sutra. The "heart within the heart
mantra" is the essence of all the mantras associated with a particular deity. In
the jirinkan procedure this core mantra is made the subject of a Madhyamikalike critique that renders it empty of "own being." This is accomplished by
disassembling the mantra into its component parts or syllables and then contemplating the individual "attributes" of each syllable. Since the syllables are
meaningful only in dependence upon one another, once the mantra is disassembled the meaning of each syllable, like the mantra itself, proves "impossible to obtain." In short, the realization of the mind of the principal deity
consists in contemplating the constructed, contingent, or "empty" nature of
the deity's mantra, which is to realize the emptiness of the deity itself.
The Juhachido thus culminates in a ritualized deconstruction of the principal deity. And yet, according to the narrative logic of the rite, to deconstruct
the deity in this manner is precisely to become one with the mind of the deity.
The deity's mind is the pure contemplation of emptiness. The performance
of the syllable wheel discernment is thus the ritual instantiation of the identity of the practitioner's mind and the mind of the principal deity. And this is
precisely the mind of Mahiivairocana - the dharmadhiitu itself.

the mind of the deity. Here, too, the structure of the rite is more or less the
same from one practice to the next, although the specific "seed syllables" featured in the kanso will differ depending on the identity of the principal deity. I
have translated the jirinkan from the Sanboinryu Juhachido, in which Nyoirin
Kannon serves as principal deity.
Form the Amida samiidhi mudrii and contemplate ID! as follows:
. In my heart there is an eight-petaled white lotus blossom, on top of which
IS a full moon d1sk. On top of the disk there are the syllables on ha ra da han
domei un, along with each syllable's meaning. 55 Contemplate W?, [the disk]
rotating once clockwise and once counterclockwise.
The ceaseless flow (rushu iiTUt) of the syllable on cannot be obtained. Because the c~~seless flow of the syllable on cannot be obtained, the speech
(gonzetsu 1'~-~11. ) of the syllable ha cannot be obtained. Because the speech of
the syllable ha cannot be obtained, the defilements (jinku ~!5 ) of the syllable ra cannot be obtained. Because the defilements of the syllable ra cannot
be obtained, the charity (seyo fit!j Si~) of the syllable da cannot be obtained. Becau:~e the .:harity of the syllable da cannot be obtained, the supreme principle
(shogz ll#;fi) of the syllable han cannot be obtained. Because the supreme
p~mCiple of the syllable han cannot be obtained, the self-attachment (gashu
:fx$")\) of the syllable domei cannot be obtained. Because the self-attachment
of the syllable domei cannot be obtained, the conditions and karma (ingo
of the syllable un cannot be obtained. (This is called the clockwise
contemplation Ill~ .)


Because the conditions and karma of the syllable un cannot be obtained,

the self-attachment of the syllable domei cannot be obtained. Because the
self-attachment of the syllable domei cannot be obtained, the supreme prinCiple of the syllable han cannot be obtained. Because the supreme principle
of the syllable han cannot be obtained, the charity of the syllable da cannot
be obtained. Because the charity of the syllable da cannot be obtained, the
defilements of the syllable ra cannot be obtained. Because the defilements of
the syllable ra cannot be obtained, the speech of the syllable ha cannot be ob-

Having examined in detail some key sections from Shingon invocation rituals
it should be clear that they offer little support for the widespread assumption that such rites involve the mental construction of mandala-like images.
What one finds instead is that the contemplative elements are often discursive
meditations with little or no "visual" component. Moreover, they are typically treated as liturgies to be verbalized softly or silently. Indeed, nowhere
is the liturgical nature of Shingon kanso more apparent than in the repetitive
phrasing and parallelism of the syllable-wheel contemplation.
The notion that Shingon contemplations are essentially discursive is con-



Visualization and Mandala

firmed by a casual glance through some of the standard East Asian Buddhist
encyclopedias and Shingon lexicons. The entry for kanso in the Mikkyo daijiten
W~:::k;$W , for example, reads:

Kanso: Kannen, teikan, kansatsu, kan, so, shii, and so on all have roughly the
same meamng. They mean to bring to mind and reflect upon; that is to say:
to contemplate the meaning of the mutual interpenetration of the practitioner
and the principal deity or saint, or to contemplate the real characteristics of a
syllable wheel, or to contemplate the true significance of forming the mudra
and intoning the mantra, and so on. Among the three mysteries [kanso corresponds to] the operation of the mystery of mind. If one were to speak generally, any contemplation of the phenomena and principle of all dharmas is
called a kanso ~~ ~ ' ~ ~~' ~~ ~ ' mJl' ~, WZIJ:,I[l, tt~*.Elfiil ~'d: I) {.

C: ;fJJ,Ui!J!-~'9 ~~L: L l ' ' f'T ~ IJ; :;t;: ~ ~ i:::fEl ff! iJPA t" ~ ~~
II t ' :s.x ii ~ilii1liO)'fffEl ~ ll. t , WZtJ:En a~ ~ *68iii L -c ~ ~ ~ ~ ~il
'9 ~ ~ ~ ii: -\, .=: ;t O) !fJ O) ~ W O)f'f ~ 'd: r) ;5 L fl~t:.IJ:M!it 0) $J'!

~il'f ~ ~ ~~~~

1:: fil}t'.57

A similar set of definitions can be found in the Bukkyogo daijiten f9!l~~!t:::k

I$W, where the relevant terms are glossed with phrases such as "fully concentrating one's thoughts" ~
f; l ~ 1:- c:=. c:, -g c:=. 1::, "bringing to mind"
{ : (,:, ,l[l, l >?!J! i.l' ,.)~ , "thinking precisely and clearly" *HI f.i' (,:, B~ c:, f.i' (,:, ~ ;t ~
C:. 1::, "careful deliberation" and "careful consideration" eJ...~ 1"{.'~
11th I~\
~ ..,j2 L.. .
~ '0)





(_ '

<1.~ ~ T ~ C:. 1::, and so on.58 And the detailed treatment of kanso in Mo-

chizuki's Bukkyo daijiten is explicit concerning the continuity between Shingon kanso practices and earlier Buddhist meditation exercises in which the
practitioner concentrates his mind on various "correct thoughts" in order to
avert or expunge mental defilements such as lust and desire.59 Thus, according to the lexicons virtually all of the terms that might qualify as Japanese
eqUivalents of the English "visualization" are perhaps better understood as
"thinking," "contemplation," "discernment," and so on.
There are additional factors in the Shingon treatment of mandalas that
mitigate against their use as "aids to visualization." To begin with, many
temples use shuji, or "seed-syllable," versions of the Kongokai and Taizokai
mandalas in which each deity is represented by a single siddham (Sanskrit)
character. (The popularity of such shuji mandara fll r ~ ~*-1, also known
as ho mandara it~~ *-1, may be due in part to the fact that they are easier
and cheaper to produce than full-blown polychrome daimandara :::k~~*-f.)
Such seed-syllable mandalas would be of little help in visualizing the complex


iconographic detail of the deities. Besides, many Shingon trainees are not sufficiently proficient at siddham to be able to decipher the seed-syllable mandalas
in the first place.
In addition, although it is traditional for a practitioner to undergo Shidokegyo training in cloistered isolation, nowadays many Shingon novices undergo training in groups at headquarter monasteries (honzan :.$: UJ) such as
those at Koyasan or Daigoji. Such monasteries offer instruction to as many
as several dozen monks or nuns at a time who perform the practices together
in a single large hall. Each trainee has his own ritual altar, but only a single
set of mandalas is installed at the front of the sanctuary-a situation that renders the mandalas of limited value as visualization aids. Group training also
obliges the monks to move through each section of the rite more or less in
unison, because all must wait for the slowest to finish before leaving the hall.
Peer pressure thus tends to ensure the expeditious performance of the rites
and to discourage lingering over the contemplations.
But even before the group training of priests became commonplace, monks
would have been under considerable pressure to hasten through the rites; the
practices are so involved and time-consuming that a conscientious monk practicing in isolation who tarried over sections would soon find himself seriously
deprived of time for meals and rest. Indeed, no matter how expeditiously the
rites are performed, priests still find themselves forced to survive on little or
no sleep during the final week of the Shidokegyo devoted to the Goma-an
onerous practice that can take as many as four or five hours per sitting. All of
these factors would mitigate against the development of visualization skills, if
the development of such skills were in fact the goal of the training.

Mandala as Divine Presence

In questioning the supposition that Shingon mandalas are used as aids to
visualization, I do not mean to detract from their significance in the Shingon
tradition. The following oft-cited passage from Kukai's Goshorai mokuroku
iiffl~31<: ilk, an annotated catalog of the texts, icons, and ritual implements
that Kukai brought back from China, readily attests to the cardinal role accorded such images:
The Esoteric teachings are so profound and mysterious that they are difficult
to record with quill and ink [that is, in writing]. Thus we resort to the expedient of diagrams and paintings il to reveal them to the unenlightened.


Visualization and Mandala


The various postures and mudrii [depicted in the mandalas] emerge from [the
Buddha's! great compassion; with a single glance [at them] one becomes a
buddha. The secrets of the siitras and commentaries are recorded in a general
way in diagrams and images lffil! i~ , and the essentials of the Esoteric teachings
are actuall y set forth therein. Should these be discarded there will be difficulty in transmitting and receiving the dharma, for they are none other than
the foundation f!Hffi of the ocean-like assembly [of enlightened ones]. 60
Similar sentiments are reiterated later in the same text: "The master [Hui-kuo]
said that the scriptures and commentaries of the Mantrayana Esoteric teachings are so recondite that without the expedient of diagrams and paintings
they could not be transmitted." 6 !
These well-known passages do not explicitly countenance the use of images
for "visualizing" deities. They do, however, suggest alternative ways of understanding Shingon mandalas. The first is that of hensozu ~ f i! , or "transformation images," a term used for visual representations of regions of the Buddhist cosmos derived from written sittras and commentaries-that is, what


dhist sacred images. As I argued in the Introduction to this volume, Buddhists

have, throughout history, regarded icons of buddhas and saints as animate
entities possessed of considerable apotropaic and redemptive powers. This
attitude is well attested in textual, archaeological, and ethnographic sources;
the completion of a painted or sculpted Buddhist icon, for example, typically
involved an elaborate "eye-opening ceremony" in which the pupils of the icon
were "dotted" to the accompaniment of invocation rites and offerings. Such
ritual consecrations, still widely performed today, were intended to transform
an inanimate image into a living presence, and icons thus empowered were
worshipped with regular offerings of incense, flowers, food, light, and sittras.
Chinese Buddhist biographies and temple records are replete with tales of
miraculous occurrences associated with such images; images were known to
fly through the air, to sweat, to communicate in dreams, to prophecy, and
so on.65
The treatment of sacred images as animate beings applied not only to
images of budd has, bodhisattvas, and other deities but also to portraits of emi-

As scholars of Shingon know all too

nent Buddhist masters. Funerary practices and memorial rites for Chinese and

well, Tantric scriptures typically contain extended lists of buddhas, bodhisattvas, and other deities, along with descriptions of their physical attributes
and symbolic accouterments. Indeed, it is easy to find oneself overwhelmed

Japanese Buddhist patriarchs reveal that their spirits were believed to cohere
to their portraits long after death, rendering such portraits sacred icons to be
worshipped in the same manner as a buddha image or relic.66 This is true of
Shingon as much as of any other Buddhist school: Kukai returned with five

one might simply call "illustrations."


in an exasperating accumulation of detail. One obvious but nonetheless noteworthy function of mandalas is that of visual commentary: they are capable
of representing a vast number of polychrome deities on a single, albeit often
rather large, expanse of silk or paper. Looking at a mandala, one is able to
grasp at a glance not only the color, posture, demeanor, and ritual implements
associated with each of hundreds of depicted deities but also the hierarchical
and spatial relationships that exist between them (fig. 4.1 and pl. 7).
But clearly Kukai believes that Shingon images are more than mere "illustrations "; note his striking statement that "with a single glance one becomes
a buddha" -- ~ fi\({iJll . Here Kukai points to a more "magical" dimension of
Tantric art that modern commentators are seemingly unable or unwilling to
confront directly. Hakeda Yoshito, for example, modifies the passage in his
translation by the insertion of a potential verb: "The sight of [the images] may
well enable one to attain Buddhahood." 63 Such a translation mutes the force
of Kukai's original statement, which unequivocally attributes transformative
or salvific powers to the sacred images he carried back from China.
There is nothing particularly revolutionary in Kukai's attitude toward Bud-

such patriarchal portraits and grouped them together with mandalas in his

Goshorai mokuroku (fig. 4.12.).67 His comments cited above on the power of
mikkyo images thus refer not only to the Kongokai and Taizokai mandalas but
also to these patriarchal portraits. Furthermore, portraits of Shingon patriarchs are hung on either side of the central mandala during the Shidokegyo
practices. Each portrait then functions as the "principal deity" of separate selfcontained r ites-the Daishi horaku :kBflii~~ (or Daishi mimae 7dli!i~M)
and the Sonshi horaku ~ p,iji i~ ~ -performed daily during the Shidokegyo
retreat. 68
Like all Buddhist icons, a Shingon mandala is not so much a representation of the divine as it is the locus of the divine-the ground upon which
the principal deity is made manifest. Scholars believe that painted mandalas
originally evolved from mandalas set upon the earth- geometrically arranged
altars that functioned as the earthly abodes of divine beings. The power or
"charisma" believed to cohere to mandalas is mentioned explicitly in a variety
of Tantric sources, including the commentary to the Mahiivairocana-sittra by

Visualization and Mandala


Subhakarasil!lha and I-hsing, which explains that a mandala is called an "assembly"~ . ~ because "the actual meritorious power of the Tathiigatas is now
gathered together [there) in a single place" ~ t),~O* ~ Jp:)Jf,~~:t:f - ~.69
To come into the presence of a mandala is to enter the presence of the Tathiigata-one literally "sees the Buddha" and partakes of his lwji, or empowerment. According to Shingon doctrine, such empowerment is no more and no
less than the realization of one's essential identity with the Tathiigata, a realization ritually reaffirmed or instantiated every time one takes a seat before
the honzon. This, then, is the doctrinal basis behind Kiikai's bold claim that
one becomes a buddha through a single glance at an Esoteric icon.
Shingon ritual manuals say virtually nothing with respect to the use of man-

dalas during the course of the initiatory practices. In fact, the only section of
the manuals that specifically mentions the mandalas is the set of preliminary
instructions for "adorning the sanctuary" (dojo shogon )j ijjjllfJlit\().7 Here detailed directions are given for the construction and preparation of the altar,
the arrangement of various ritual implements, and the installation of the icons,
all of which must be completed prior to beginning the rites proper.
In Buddhism the expression shOgon, or "adornment," functions as a technical term, often used as an equivalent to the Sanskrit terms ala'!fkiira or vyuha. 71
In sutra literature these terms are associated with the magnificence, splendor,
and supernal adornments of a "pure land," which include a resplendent array of exquisite pennants and banners, rare jewels, precious metals, exotic
flowers, and so on. Shingon practice halls, like their architectural counterparts in other Buddhist traditions, are accordingly decorated with silk brocade hangings, ornate altar pieces and implements of fine lacquer and gold,
fresh flowers, and incense, all of which are intended to reflect the glory of a
pure buddha realm. The "adornment of the sanctuary" is thus a formal procedure wherein the place of practice is ritually transformed into the world of enlightenment-the instantiation of the Mahayana tenet that nirviit;a is sa'!fsiira
correctly perceived. This transformation of the sanctuary into a pure land is
effected in large part through the agency of the image that constitutes the
sacred presence of the principal deity.72 If anything, the presence of the manF 1 G u R E 4. 12 (Opposite)
Amoghavajra (C: Pu-k'ung), one of five portraits
of patriarchs of Shingon Buddhism painted by Li Chen in the early ninth
century and imported from China to Japan by Kiikai in 8o6. Hanging scroll,
ink and colors on silk, 212.1 x 150 . 6 em. Ky66gokokuji (T6ji), Kyoto
Prefecture. Photo: Ky66gokokuji.



Visualization and Mandala

dala-an eminently visible supernatural being positioned directly in front of
the practitioner-does not so much serve as an aid for visualizing the deity as
it abrogates the need for visualization at all.

Shingon Apologetics and the Hermeneutic of Experience

The "supernatural" or "magical" properties of Buddhist icons should be evident to all those familiar with the treatment of such images in Japan. Yet all
too often this aspect of Buddhist images is overlooked by contemporary apologists, art historians, and buddhologists alike. Rather than treat icons in the
context of the miraculous powers attributed to them, contemporary writers
tend to emphasize their didactic function, as if Buddhist images were intended
merely to symbolize the virtues of buddhahood, or to nurture a sense of reverence toward the Buddha's teachings. Alternatively, belief in the power of icons
is often dismissed as a popular accretion antithetical to the tenets of "true
The tendency to disregard or dismiss the Buddhist veneration of images
is yet another manifestation of what has become known as the Protestantizarion of Buddhism: the widespread penchant to present "real Buddhism"
as a rational and humanistic creed that rejects sacerdotalism, magic, idolatry, and empty ritual.7 3 Thus Theravadin apologists will insist on the Buddha's "humanity" and the purely pedagogical role of images, despite extensive
ethnographic and textual evidence that suggests otherwise.74 Some advocates
of Zen have gone even further, propagating the fiction that Zen eschews the
use of images altogether. In the case of Shingon, however, where the invocation and veneration of deities constitutes the heart of clerical practice, modern
commentators are unable to avail themselves of these defensive strategies. One
finds instead the reconfiguration of image worship in psychologistic terms,
such that the elaborate rituals revolving around the invocation of deities and
the worship of icons are construed as meditative exercises intended to inculcate a mystical or transformative "religious experience."
I have discussed the roots of the Japanese Buddhist "hermeneutic of experience" elsewhere and thus will limit myself to the briefest overview here. 75 In
the late nineteenth and early twentieth centuries, Buddhism was the subject of
a sustained critique by Japanese intellectuals, many of whom were schooled
in Western thought. Japanese opponents of Buddhism adopted many of the
arguments used by European Enlightenment critics of Christianity; Buddhism


is denounced as a "primitive," "illogical," and "unscientific" creed propagated

by a self-serving, morally degenerate priesthood. Buddhism-construed as a
foreign import at odds with the genuine spiritual and eth~cal values. of ~he
Japanese-was even held responsible for Japan's technological and scientific
"backwardness." 76
In response, Buddhist leaders, many of whom were versed in Western philosophy, invoked the traditional rhetoric of upiiya, or "skillful means," in defense of their faith. They argued that Buddhist ritualism must not be seen as
an end in itself, nor as a primitive method of manipulating natural,
but rather as a means of engendering a profoundly liberating, nonsectanan
spiritual experience. In short, Buddhism was not so much "bad science" as it
was "enlightened mysticism" or "transformative psychology." Ap~log1s~: for
the Zen and Pure Land schools would thus emphasize the centrality of personal experience" (keiken *~~'or taiken fit~) in their respective traditions,
a strategy that had the felicitous result of rendering Buddhism intellectually
respectable and immune to external critique at one and the same time.
Shingon was particularly vulnerable to Western-influenced cnuques of religion due to the emphasis it placed on sacred icons and the sacerdotal powers
of the priesthood. Following the lead of their Zen and Pure Land coun_terparts
Shingon exegetes-including such eminent authorities as Toganoo Shoun, Yamasaki Taiko, and Matsunaga Yiikei-accentuate the "meditative," "expenential," and "psychological" dimensions of Shingon practice. Matsunaga ~s
unequivocal in this regard; in a short talk on the meaning of th~ word mzkkyo,
he says, "Esoteric Buddhism is not theory but religious expenence, the mtuition of the essential oneness of the macrocosm and microcosm. Through religious experience one gains direct intuition of the fact that the macroco~m~
the universe of Mother Nature-and oneself are essentially one." n And m his
recent general introduction to Buddhist Esoterism Matsunaga writes:
In Buddhism the realm of religious experience ff;f:,( f*Ml is called the "inner
witness" [jinaishO 13 pg ru ], and as is understood from the fact that the common Japanese word "private" [or "secret," naisho pg *?% ] is derived from It, It
lies within the domain of personal experience that IS Impossible to communicate to others. While we may grasp it with our ordinary sense faculties, It
is not something that can be transmitted by usual modes of communication
such as speech or writing .. .. In order to acquire it we can.not rely on others,
but must experience it personally by mystical intuition which remams the
only means of comprehending it. 78


R 0 BE R T H S H A R F

Visualization and Mandala

According to Matsunaga, the "privacy" of the world of inner experience

explains the ubiquitous Buddhist rhetoric concerning the "inexpressibility"
of absolute truth. Matsunaga goes on to say, however, that this is one area in
which Esoteric Buddhism parts ways with exoteric Buddhism, for although
the latter insists that the realm of absolute truth is utterly beyond expression,
Esoteric Buddhism holds that the absolute can indeed be expressed in sacred
signs (hyoji H\'l$/W ), sacred speech (mantra), and above all sacred art. Citing the
same passage from Kukai 's Goshorai mokuroku that I discussed above, Matsunaga argues that Shingon art must thus be understood in the context of religious experience- Esoteric images both express and elicit the experience of
the absolute .79
The same claim is made by Yamasaki Taiko, who says, "The source of
the Mikkyo mandala is Shakyamuni's experience at Buddhagaya, where he
realized enlightenment through meditation under a bodhi tree. Mikkyo seeks
to convey this experience of the source of the self by means of painted and
sculpted forms, and by meditation and ritual employing these forms. This is
the inner meaning of the mandala." 80 Ishida Hisatoyo concurs, appealing to
the hermeneutic of experience to explain certain stylistic and iconographic
features of the mandalas such as the frontal presentation of Ishida
As long as contemplative experience is a living, meaningful one for Esoteric
master and painter, it generates the creative impulse. As a result Esoteric
paintings possess a vital power that is not present in those of other types of
Buddhism. Indeed, if Esoteric images are executed mechanically according
to the manuals, their essential life is lost, and they turn into facile formalizations . . .. Ultimately, the mandala is a symbolic and concrete representation,
through images, of the process of human enlightenment.82


as mediated through the writings of Nishida Kitar6, D. T. Suzuki, and their

intellectual heirs. 83
Japanese Shingon exegetes were not the only ones to tout the experiential or
psychological dimensions of Tantric art. The preeminent and still influential
early Western scholar of Tantric art, Giuseppe Tucci, taking Carl Jung and his
theory of archetypes as his point of departure, refers to mandalas as "psychocosmogrammata" and as a "way to the reintegration of consciousness." 84 For
Tucci the mandala is an external representation of the psychospiritual forces
that lie in the depths of human consciousness, and each feature of the mandala
corresponds to some specific psychological aspect of the self and the process
of spiritual evolution. Tucci was also instrumental in popularizing the notion
that mandalas serve as aids to visualization: "The mandala born, thus, of an
interior impulse became, in its turn, a support for meditation, an external
instrument to provoke and procure such visions in quiet concentration and
meditation." 85 Here, too, in promoting the mandala as a sophisticated means
of occasioning psychological and spiritual transformation, Tucci successfully
masks the patently "idolatrous" nature of Buddhist icon worship; nary a mention is made of the soteriological, apotropaic, and redemptive powers that traditional Buddhist sources categorically attribute to Tantric rites and images.

I have examined in considerable detail some of the problems with the notion
that Shingon mandalas are used as aids to visualization. For one thing, there is
surprisingly little correlation between the iconographic content of the major
Shingon mandalas-the Kongokai and Taizokai-and the liturgical content of
the invocation rituals with which they are associated. For another, the kanso,

The questionable claim that Shingon mandalas function as aids for medi~
ration is then simply one facet of an all-encompassing approach to mikkyo, an

or "contemplative," material in the rites is often more discursive than visual.

approach that privileges inner experience over public performance. The prob-

rather as liturgical recitations that constitute the ritual enactment of the "mys-

lem with this exegetical strategy is that traditional Buddhist sources, whether

tery of mind."

These contemplations are treated not so much as guided meditations, but

esoteric or exoteric, simply fail to provide support for it. The roots of the

In emphasizing the performative dimension of Shingon rites, my "eric"

Japanese Buddhist "hermeneutic of experience" lie rather in the writings of

analysis is largely commensurate with orthodox "ernie" doctrinal formula-

twentieth-century Buddhist exegetes who found themselves forced to respond

tions. As is evident from central tenets such as "becoming buddha in this very

to powerful rationalist or empiricist critiques of their faith. These exegetes

body" and "the nonduality of practitioner and deity," the prescribed sequence

adopted in part the Western rubric of "religious experience" in formulating

their response, drawing on sources such as Rudolph Otto and William James

of ritual acts that comprises a Shingon invocation ritual was not traditionally
understood as leading to some private mystical experience. Rather, they were


Visualization and Mandala 197


viewed as the enactment of buddhahood- the practitioner literally mimics the

body, speech, and mind of the Tathagata, thereby attaining his kaji. (Shingoo kaji might then profitably be analyzed under the rubric of "sympathetic
Be that as it may, there is no denying the many patently visual or pictorial
elements in the ritual contemplations. The kanso found in Shidokegy6 invocations contain numerous references to lunar disks, lotus blossoms, mutating seed-syllables, stupas, and, of course, a wide variety of anthropomorphic
deities. It is certainly possible, if not probable, that this store of visual imagery
would contribute to the construction of an elaborate imaginative world in
which the sanctuary is construed as a pure buddha field populated by a host
of benevolent deities. In entering the sanctuary and undertaking the rites a
priest learns to behave as if he were dwelling in a sacred realm, as if he were in
the presence of the principal deity, as if he had merged with Mahavairocana.
No doubt, sustained training in these arduous practices would effectively alter
one's affective response to the liturgies, the implements, and the images used
in the rites. But at the same time, this imaginative "as if" aspect of Shingon
performance demands that the practitioner, like any accomplished stage actor,
remain fully cognizant of his immediate physical environs. As such my analysis
stands in contrast to the phenomenological model, which holds that Shingon
meditations involve the psychological projection or "inner visualization" of
an alternative universe.
This chapter raises many issues that I am unable to pursue here, perhaps the
most conspicuous being the disparity between the iconography of the mandalas and the content of their respective rites. It is almost as if the conjunction of rite and image were the result of a not-altogether successful attempt
to synthesize two independent traditions-one liturgical and the other iconographic.86 Speculation on this issue would be premature, however, as the historical development of Shingon rites and images is still poorly understood.
The orthodox Shingon position has always been that the Kong6kai and Taiz6kai lineages, including their scriptures, images, and rituals, were independently transmitted in unadulterated form from India to China. In China, the
theory goes, they were finally brought together into a single unified system of
doctrine and practice, which was then transmitted directly to Japan. Scholars are now beginning to appreciate, however, the extent of the East Asian
contribution to the tradition now known as Shingon. Indeed, the fundamental categories used to delineate and define the Shingon school- that is, the

opposition between "esoteric teachings" (mikkyo) and "exoter.ic teachings"

(kengyo ~~),and that between "pure" versus "mixed" esotensm (Junm1tsu
~W and zomitsu *ltW)-are not attested in India and may well have been
Japanese innovations. Kiikai appears to have been responsible for the notion
of a lineal genealogy of mikkyo masters originating with Mahavatrocana, and
his disciples were the first to systematize the major invocation rites int~ the
Shidokegy6 sequence. Scholars are also beginning to scrutinize the pedtgree
of the Taizokai and Kong6kai mandalas; it now appears likely that the configuration of the two mandalas, as well as some of the iconographic detail,
evolved in China rather than IndiaP In short, scholars are starting to grasp
the significant part played by Amoghavajra, Hui-kuo, Kukai, and their followers in shaping East Asian mikkyo, but much work remains to be done on
the specific role played by each of these figures and the sources on which they
drew. 88
These issues lie well beyond the modest scope of this study. It should now
be clear, however, that speculation as to the conception and function of Buddhist mandalas-and indeed all Buddhist religious icons-is of limited value
unless predicated on the critical analysis of their ritual and institutional


Notes to Pages 1 42- 47

Notes to Pages 148-5I

Shigaku zasshi R9, _no. 2 .( 191\o): 66-9r. Many scholars have argued that hinin
were co'ntrolled by Kofukup;,see, for example, Nagahara, "Eta-Hinin,"
_ .
393 94
However,.Matsuo notes that E1son and Nmsho brought many hinin under the control of_ Sa1dmp. Hosokawa Ryoichi has criticized Matsuo on this point, noting
that K.ofukup monks contmued to attend many events connected with Eison and
Sa!da1p (Chusei no risshii jiin to minshu, 56-57). He also argues that the Kamakura
bakufu (military government) may have actively supported Eison's proselytization
and orgamzatwn of the hinin.

89. Sasaki Kaoru f i /.;

seigai bukkya no sokoku r[1

Chtisei kokka no shukyo kozo: taisei hukkyo to tai-

: ft,:ffiij
t (4;:if,lj
(Tokyo: Yoshikawa h>bunkan, r9RR), 210-20.
90. Taikan, pls. 12-q, so-53; Saidaijiten, 24-29.
91. Kobayashi Takeshi 'J
"Saidaiji eisonzo ni tsuite" [/Ej
(; l > C:' BukkyiJ geijutsu 2R (1956): 37 However, the tradition of painting portrai~
of deceased ~~ment. monks in Vinaya lineages was well established. Two portrait
P_amtmgs of Ia(~-hsuan . and Yuan-chao
had been brought back from
Chma by ShunJo m r2u Subsequently, many such paintings were commissioned
or earned back from China.

92. It IS not certain that the fly whisk, together with certain other objects currently handed down With the image such as a copper-handled censer placed to
one s1de, were later replacements of original objects belonging to Eison, or not
see Tatkan, 43
93 Sec ibid., 41-43.
9_4 Shiisei, .J36. The image also contained a formulaic dedicatory statement
concernmg such ISsues as the Three Jewels; unfortunately the text includes several
undecipherable characters making it difficult to read (ibid., 6_ ).
33 37
95 KanJm gakushoki, Shusei, 4 s.
96. For a complete list of the items in the image, see Taikan, _ 6; counts of
45 4
the numb~r of documents mcluded in the image vary because some items comprise
several texts. For the texts of documents concerning Eison and the col h

. . d
op ons o
scnptures cop1e by those who Wished to be associated with the icon, see Shusei
97 Kobayashi Takeshi, "Saidaiji eisonzo," 6.
. 98. Kurata Bunsaku fJ! ){
"Tainai nonyuhin"
f1,1 A. r'ib, Nihon 110 biJtttsu R6 (1973): 17.

~akao Takashi rp I

"Eison ni okeru sh<ljinbutsu no shinko"

jrp, in Kamakura jidai bunka denpa no kenkyi ;ijt f!f fJi {-t


jl}J :tllf
ed. Osun11 Kazuo
(Tokyo: Yoshikawa kobunka
~ d '"hJ:L,
. .
n,1993 ,131.
roo. en a1s u A. 1.1
Koyasan hterzan meihiJten
!lilt fiX Ui :g
(Tokyo: Sanke1 shmbun, 1997), 73, 195-96; Asada Masahiro i~EBIE
shmp1tsuhon no hakken: shogoin shoz6 'sanbu manda' ni tsuite"
.~~ r,c~.: I~.
~- 111111.
L:: ') l '"(, in ChisM daishi kenkyu henshu
unka1, Clnsho datshr kenkyzi
(Kyoto Japan: Dohosh
8 ) 8
a, 19 9 ' 99
101 F
. or a rsc~;:swn o~,the materials contained in the Seiryoji image of Sakyamum, see Kurata, fama1, 19-27.


102. The colophons for these texts are found in Taikan, 44-45, nos. 5, 12. For
a thorough study of Soji, see Hosokawa Ryoichi *ln) II iJ)( "., "Sairinji s6ji to ama:
chi.isei risshu to nyonin kyusai"
t 1[2 : rjJ
Sukui to oshie :f:iz l ' t ~,: ;t, ed. Osumi Kazuo
[ffi D
(Tokyo: Heibonsha, 1989 ), 121-64.
103. Kyoe was one of Eison's closest disciples and is mentioned in a number
of sources. He compiled the Saidaiii den'en mokuroku ril:i J\. <11' IJ:l !iill l:l
an Important source for research into Saidaiji landholdings. This text lists all of the
landholdings either given to or bought by Saidaiji during a sixty-five-year penod
ranging from 12 33 to 1298.
104. The colophon on the dhiiral}l mentions protection of the state (Tatkan,
44). The Suvar~wprabhiisa-sutra is one of a set of three scriptures traditionally used
to protect the state.
ros. See note 30 to the Introduction of this volume.
ro6. Kurata, "Tainai," 29.
107. For the text of the Saidaiii k6sh6 bosatsu gonyiimetsu no ki rZEi k.
see Shusei, 287-300; the provenance of the text is discussed by
Narita Teikan EB ~
"Saidaiji koshcl bosatsu gonyumetsuki ni tsuite" r!Ei 7-.:.
C:. -) ll"(, Indogaku bukkyiJgaku kenkyu 5, no. 2 (1957): 23134. 108. T. Griffith Foulk and Robert H. Sharf, "On the Ritua I Use o f C'h' an Portraiture in Medieval China," Cahiers d'Extreme-Asie 7 (1993-1994): 192; the function of portraits in Ch'an funerary rituals as described in this article apply to
Eison's image, but the data in Shokai's record (Shusei, 294) is insufficient to determine how close the similarities might have been. Funds for the daily offerings
who had contributed to other
were contributed by the nun Myoh6b6 ~1; i};
services at Eison's quarters (Saidaji den'en mokuroku: uragaki, Shiisei, 426, 440).

Chapter 4
Research for this paper was supported in part by a grant from the Social Sciences
and Humanities Research Council of Canada. I would like to extend my thanks
to Abbot Tagawa Shunei for the opportunity to study mikkyiJ ritual at KMukuji
in Nara, and to the other monks and support staff of Ki'ifukuji who went out of
their way to accommodate me during several extended stays at the temple.
r. The term honzon (C: pen-tsun) is thought to be derived from Tantric scriptures modern lexical works cite the Chinese translation of the Mahiivairocanasutra,as the locus classicus (see esp. the section entitled Pen-tsun san-mei /jl:
T 848: r8.44a-b). The term apparently lost its explicitly Tantric overtones rather
early and came to be used by all sects in Japan; see Mikky<> Jiten Hensankai
kr~f.ijl! (1931, 1970; reprint, 6 vols.
ed., Mikkyo daijiten
in r, Kyoto: H6z6kan, r9R3), 2o68b-c; MZ 5: 4697b-9Ra; and Roger Goepper,
"Some Thoughts on the Icon in Esoteric Buddhism of East Asia," in Studia SinoMongolica, Festschrift fur Herbert Franke, Mi.inchener Ostasiatische Studien, bd.
25 (Wiesbaden, Germany: Steiner, 1979 ), 245-54.


Notes to Pages 153-55

Notes to Pages 155-57

2. Richard Karl Payne, The Tantric Ritual of japan: Feeding the Gads- The
Shingon Fire Ritual, Sata-Pitaka Series, vol. 365 (New Delhi, India: Aditya Prakashan, 1991), 21.
3 Ishida Hisatoyo, Esoteric Buddhist Painting, Japanese Arts Library, no. 15,
trans. E. Dale Saunders (Tokyo: Kodansha International, 1987; first published in
Japanese in 1969), 30.
4 See also Ulrich H. R. Mammitzsch, Evolution of the Garbhadhatu Mandala,
Sata-Pitaka Series, vol. 363 (New Delhi, India: International Academy of Indian
Culture and Aditya Prakashan, 1991), r5; Dale E. Saunders, "Some Tantric Techniques," in Mikkyogaku mikkyoshi ronbunshii
Studies of
Esoteric Buddhism and Tantrism in Commemoration of the r, I)oth Anniuersary of the
Founding of Koyasan (Koyasan, Japan: Koyasan University, r965), r69; Guiseppe
Tucci, The Theory and Practice of the Mandala, trans. A. H. Brodrick (New York:
Samuel Weiser, 1970), 37; and Denise Patry Leidy and Robert A. F. Thurman,
Mandala: The Architecture of Enlightenment (New York: Asia Society Galleries and
Tibet House, r997). Michael Saso's studies of Tendai mikkyo practices are predicated on the notion that Tantric art functions as the basis for meditative practice;
see his Tantric Art and Meditation: The Tendai Tradition (Honolulu, Hawaii: Tendai Educational Foundation, 1990); and Homa Rites and Mandala Meditation in
Tendai Buddhism, Sata-Pitaka Series, vol. 362 (New Delhi, India: Aditya Prakashan, 1991).
5 In recent years a number of scholarly publications have appeared that make
accessible the kuden associated with various Shingon lineages; see esp. Oyama
Kojun k. rU i~
Chiiinryit no kenkyt"i 1t1
(Osaka, Japan: Toho shuppan, 1987; originally published as Himitsu hukkyo koyasan clnlinryit no kenkyu
Koyasan, Japan: Oyama kojun hoin sh<)shin ki}c .]0
nen shuppankai, 1962); Takai Kankai J:fll.iM, Mikkyo jiso taikei
(Kyoto, Japan: Takai zenkeshu chosaku kankokai, 1953); Tanaka Kaio l:ll r:p !!!,
Mikkyo jiso no kaisetsu
(Tokyo: Rokuyaon, 1962); Toganoo
Himitsu jiso no kenkyii
!ilfJ'E, vol. 2 of Toganoo
shoun zenshtl
(1935; reprint, Koyasan, Japan: Mikkyo bunka kenkyujo, 1982); and Ueda Reijo .L fJH~#5\:, Shingon mikkyo jiso gaisetsu-shidobushin'anryit o chitshin toshite
- [7g flit if[)- llfr~ilfL ~ rj:q::_, (: L
T (Tokyo: D6hosha, 1986). Thus, unqualified claims by scholars such as Michael
Saso to the effect that the mikkyo oral traditions (Saso calls it the "oral hermeneutic") are available only to initiates working directly with masters (ajari
overstated (Saso, Homa Rites, 13). This is not to say that firsthand experience of the
rituals is of no value; such experience often proves indispensable in deciphering
the cryptic language of mikkyo ritual manuals. Moreover, the publications of the
oral commentaries mentioned above are intended primarily for initiated Shingon
priests and thus they presume familiarity with the rites.
6. Buddhaghosa, The Path of Purification (Visuddhimagga), trans. Bhikkhu
Nyiinamoli, 2 vols. (Sri Lanka, 1956; reprint, Berkeley, Calif.: Shambhala Publications, r976), r: 128-29.
7 Ibid., T}O.



8. The replica image first created in the mind's eye, known ~s t,?e "acqu~red
sign" (uggaha-nimitta), is later replaced by the "counterpart stgn (pattbhaganimitta). This second image is "pure," lacking any faults that .may be prese~t m the
kasir;a, and is likened to a round mirror, a mother~of-pearl drsh, a .~oon drsk, ~nd
so on (ibid., 13o; see also the discussion in Rodenck S. Buckndl, Remterpretmg
the jhanas," journal of the International Assoctattan of Buddhtst Studtes r6, no. ~
[1993 ]: 3s7 ). The notion that the counterpart stgn of the kasn;a resembles a lum.t
nous circle such as a mirror or moon disk suggests a connectiOn With later Tantnc
practices and iconography, in which the moon disk comes to play a central role.
9 Buddhaghosa, Path of Purification, I: r85-203.
ro. One of the earliest Buddhist scriptures translated mto Chmese, the Panchou san-mei ching ~Jiflr :=:B;;f:*! (T 418) translated in 179 A.D. by Lokak~ema
(Chih-lou-chia-ch'an :liill!Jili) describes practices intended to mvoke a VISIOn
of Amitiibha.
II. For a standard Theraviidin account of this practice, see Buddhaghosa, Path
of Purification, r:206- 3o. Among the many discussions of the development of
East Asian nien-fo-style practices from antecedents m the Agama htera_ture, see
Pau1 H arnson,

. in the Pratyutpanna-buddha-sammukhauasthtta.
samadhi-sutra," journal of Indian Philosophy 6 (1978): _35-57; Shth Heng-c mg,
The Syncretism of Ch'an and Pure Land Buddhism, Asian Thought and Culture
Series, vol. 9 (New York: Peter Lang, 1992), 26-29; Ujitani,~ii~en
"Agon ni okeru shomyogyo ni tsuite" ~ilJiH: .})(to _ ~ L "J G)~' Ntppo~
bukkyo gakkai nenpo 30 (1954): 51-70; and Yoshioka Ryoon_YJ IMJ
, Ago~ky
ni okeru nenbutsu no kigen" jlilJ t3"*-~ 1:::. :j"j (to ;fr;{9f,O)j~J!tr,, Indogaku bukkyogaku
kenkyu 9, no. 2 (1961): 130-31.
12. Allan A. Andrews, "Lay and Monastic Forms of Pure Land Devotionahsm:
Typology and History," Numen 40, no. I (1993): 18-19.
13 . Ta-jih ching is short for Ta p'i-lu-che-na ch'eng-fo shen-pten chta-ch th _cht~g
J::. ~ J:l ~ ;lj~ JiX {9f,.f$ ~ /JO tH~~ (Sk: Mahavazrocana-abhtsa~bodht-utkurvttadhts
tiina-vaipulya-sittra), translated by Subhakarasif)1ha and 1-hsmg I~ 725 (T 84~):
14 . The term I am translating as "absorption" is. teng-yin $iff r] I, Sk: samahtta,
which refers to a stage in which practitioner and deity are one.
15 . T 179 6: 39 .695 br9 -26; see also 783b19-c6, for a similar discussiOn of the
use of a fabricated image to help "see" the deity.
" .
. . , .
16. Reference to the use of an image as a support for visuahzano~ m a sutra (as opposed to a commentary) can be found in Chih-ye~'s translatiO~ ~f _the
Anantamukhasadhakadharm;i-sittra (Ch'u-sheng wu-pten-men t o-1~-m chmg W1: ~
JIF~~EfllliE*!), T 1018: 19.705c23-6ar; see also the discussiOn m Goepper,
"Some Thoughts on the Icon," 245-46.
17. The same is largely true of the works of Buddhagh?sa; see Ro?,ert H. Sharf,
"Buddhist Modernism and the Rhetoric of Meditative Expenence, Numen 42,
no. 3 (1995): 239
_ .,
18. For an analysis of the ideological agenda behmd Kuka1 s activities, see
Ryiiichi Abe, The Weauing of Mantra: Kitkai and the Constructton of Esotertc Buddhist Textuality (New York: Columbia Umvemty Press, 1999).


Notes to Pages r58-64

Notes to Pages I 57-58

19. It is unclear just who was responsible for the system that emerged by the
end of the Heian period (794-II85). The authenticity of many of the ritual texts,
or shidai ;)\ J41, traditionally attributed to KC1kai is in doubt. It is clear, however,
that the Shidokegy6 had become the standard course of ritual training for all Shingon monks by the end of the Heian . It continues today to be a prerequisite for
denbo fit} i~ consecration; see Toganoo, Himitsu jiso, 26 - 27; Dale Allen Todaro,
"An Annotated Translation of the Tattvasa'?tgraha (Part r ) with an Explanation of
the Role of the Tattvasmrrgraha Lineage in the Teachings of Kukai" (Ph.D. diss.,
Columbia University, 1985), 72-74; and idem, "A Study of the Earliest Garbha
Vidhi of the Shingon Sect," journal of the International Association of Buddhist
Studies 9, no. 2 (1986): 109-10.
20. The overall length of the retreat differs depending upon one's lineage and
the tradition of one's local temple. See the discussions in Toganoo, Himitsu jiso,
29 - 31; Takai, Mikkyo jiso, 90-91; and Oyama, Chtlinryii, 61.
21. Initiation into the Juhachido is restricted to those who have been ordained
as priests (tokudo fli}. r'fi ). In addition, prior to the Juhachido proper the practitioner
must undergo a preliminary purification practice known as raihai kegyo ffr~tftm fj
(prostration practice), which usually takes one to four weeks to complete.
22. The manual, Shonyoirin-kanjizai-bosatsu nenju shidai :'! ~D!l, !MiiWl. El f:E:g
~;{t ilrn ;);: ~ 'was edited by Geng6 it: 4f: (914-995) and remains highly influential
today. Gengo's manual is in turn based on the }tlhachi geiin
f\ 4Zi Ell, the ]iihachido nenju shidai -t J\ ill ;{t ~r'li ;:);:
and the Jiihachido kubi shidai f \ ill J~ ;:);:
all of which are attributed to Kukai. Shingon scholars believe that these manuals
were based on both the oral transmission from Hui-kuo TM* (746-Sos), as well
as on Amoghavajra's Wu-liang-shou ju-lai kuan-hsing kung-yang i-kuei J!!.!i f!l: ~ ~0.:31?:
fi 1r 1JH~ f;fHJL (T 930: 19.67b-72b), and the Kuan tzu-tsai p'u-sa ju-i-ilm niensung i-kuei liDH::J ti:TI=~i,ziH\H~ ~": ;]flififU)L (T1085: 20.203c-6c). Kukai explicitly
cites T 930 as the basis for his ]tihachido in the postscript to that work. In his Sangakuroku he mentions T 930 and T ro86, a variation ofT roSs. See Takai, Mikkyo
jiso, n 6 -17; and Todaro, "Annotated Translation," 103-4
23. See Takai, Mikkyo jiso, 109; Michel Strickmann, "Canons of Giant Art:
Ritual of Land and Water" (paper presented at the conference "Art and the Emperor: A Public Interdisciplinary Symposium on the Arts of China," Ohio State
University, 14 Aprilr989 ), 16 - 17; Yamasaki Taiko, Shingon: Japanese Esoteric Buddhism, trans. Richard and Cynthia Peterson (Boston: Shambhala, 1988), r62; and
Payne, Tantric Ritual, 88 - 89.
24. The daily schedule for the Juhachido used at KOfukuji today, following
the format favored at Saidaiji (Nara), is more or less representative of that used
at mikkyo training monasteries:




2:00 P.M. "Commencement of the night"

{1i . This is the first performance of the Juhachido.
Ritual for feeding the hungry ghosts 1i{!j ~Jt)l.
Bath A. i{i .
I: 30 A.M. Rite for drawing holy water [from the well] iJ.H~ ilJU {'fj;t;; . This
water is used for the preparation of altar offerings.


"Late night" ~~ iJZ , the second performance of the JCthachid6.

"Midday" El rp , the third performance of the Juhachido. This performance follows the "Late Night" performance, allowing 20 minutes or so to
reset the altar. It begins at about s: oo A.M.
Breakfast ft 1'f i!.
Cleaning ill t~j til [;'f<.
Rounds of the temple buildings 1?~ ~ -This is a short daily "pilgrimage" of
the monastic grounds, during which offerings of siitra recitations and mantras
are made to the principal deity of each building and shrine.
II: 00 A.M. Lunch ft 1t rt~. This completes the ritual day.
25. The analysis of the Juhachido into nine major units is only one of several
traditional ways of approaching the structure of the rite; see, for example, Miyata
Taisen, A Study of the Ritual Mudriis in the Shingon Tradition: A Phenomenologt-

cal Study of the Eighteen Ways of Esoteric Recitation (jiihachido nenju kubi shidai:
chiiin) in the Koyasan Tradition-Including Translation of Kiikai's "] iihachido Nenju
Kubi Shidai" (published privately, 1984), rn; and Takai, Mikkyo jiso, ru-16.
26. There are, according to tradition, twelve major Shingon initiatory lineages,
six associated with the Ono line ,J , ~r iift and six associated with the Hirosawa
line f~~iftt . There are, however, dozens of sublincages, a situation that developed
in part because each ajari is empowered to alter the manuals according to his or
her own understanding of the rites. The only major feature distinguishing the Juhachid6 of one line from the jllhachid(J of another concerns the identity of the
principal deity: the Chuinryu rl:~ l~feiilE (the tradition currently dominant on Mount
Koya), for example, uses Dainichi Nyorai f.:. Fl /,![1
(Mahavairocana) as pnnCIpal deity, while the Sanboinryu stemming from Daigoji reaN! 41' (Kyoto) favors
Nyoirin Kannon /,ZQ ~ !MlillM '&f (Cintamaryicakra Avalokitesvara). Even then, the
differences between the rites as actually performed are relatively minor. For a full
discussion of the complex relationship between the various lineages, see Toganoo
Sh6un, Himitsu bukkyoshi fiif.'f91:~ if_, vol. r of Toganoo shoun zenshii (1933; reprint, Koyasan, Japan: Mikky6 bunka kenkyi:1jo, r982), 239-66; and Takai, Mikkyo jiso, 25 - 58.
27. Tanaka, Mikkyo jiso no kaisetsu, 95 - 96.
28. For an English-language treatment of contemporary Tendai Shidokegyo
rites, see Saso, Tantric Art; and idem, Homa Rites.
29. Miyano Yuchi '/i~IDf ~iW and Mizuhara Gyoei 7]( 81{ :% ~ , eds., Shidokegyo
shidai [LqfJit!JD fi' :xm , 6 vols. (Koyasan, Japan: Matsumoto nisshindo honten,
1933), "Nyoirin," 2. (This text is based on a set of manuals dated Tensho X IE 4
[1576] .) See also Ozawa Shoki '] ' 7:~ [l.tlffii!! , Shido shidai: koshin ho lm Tf!f. ;'x_ ~ ' {,. 1i
(Kyoto, Japan: Sohonzan chishakuin, 1962), 22.
30. In Shingon history a variety of figures have been favored at one time or
another, in one lineage or another, as honzon for the jl1hachido, including Amitabha, Sakyamuni, and Fud6-myo6; see Toganoo, Himitsu jiso, 34
31. Rigen Daishi is celebrated as the founder of the Onoryu ,j,~!'f'rfri tradition,
from which sprang the Sanb6inryi:1; see Nihon Bukkyo Jinmei Jiten Hensan linkai
El i>:19U~ A. :t:Jf*- !,!4 k*hlti =~ ffil. tfgr , ed., Nihon bukkyo jinmei jiten E1 4:f?U5( A. :t'J ii$


Notes to Pages r66-78

~~!- (Kyoto, Japan: Hozokan, 1992), 384. Kiikai and Shobo are the subjects of
separate commemorative services held immediately following the "midday sitting" (nicchu El $) during each day of the Shidokegyo retreat-the Daishi horaku :*:liliflr:t~ (or Daishi mimae :J;:Iiliflf,l!Jrlfj), and the Sonshi horaku ~liliflrt~
(or Rigen Daishi mimae l'!f!ijjf,:J;:ijjf,l!J
On the positioning of the portraits, see
Takai, Mikkyo jiso, 94-96.
32. Miyano and Mizuhara, Shidokegyo shidai, "Nyoirin," 13-14; cf. Ozawa,
Shido shidai, "Jiihachido," 51-53. The Chiiinryii dojokan for the Jiihachido is basically identical except for the identity of the principal deity.
33 On the historical and textual issues surrounding the Kongokai mandala
itself, see esp. Ishida Hisatoyo EEBMl
Mandara no kenkyu ~~il0lilf3'1:,
2 vols. (Tokyo: Tokyo bijutsu, 1975); idem, Ryogai mandara no chie ii!ii 3'f. ~ ~*10
t\'!1: (Tokyo: Tokyo bijutsu, 1979); Toganoo Shoun, Mandara no kenkyu ~~*!
0lilf3'1:, vol. 4 of Toganoo shoun zenshu (1927; reprint, Koyasan, Japan: Mikkyo bunka kenkyiijo, 1982), r89-353; Horiuchi Tomohito :fiffi pq
kue mandara no meisho ni tsuite" ~ ~lj !f)!
7 / )1" 7 0) ~fill(:_-::> C; T, Mikkyo bunka
{[:: 88 (1969): r2o-ro2 (reverse numbering); Ueda, Shingon mikkyo, 261-65; and David L. Snellgrove, "Introduction," in Lokesh Chandra and
David L. Snellgrove, Sarva- Tathiigata- Tattva-Saiigraha, Sata-Pitaka Series, vol. 269
(New Delhi, India: Sharada Rani, I98I), 5-14. David L. Snellgrove discusses a
Tibetan version of the mandala in his Buddhist Himalaya: Travels and Studies in
Quest of the Origins and Nature of Tibetan Religion (New York: Philosophical
Library, 1957), 227-42.
34 T 2471: 78.72arff. Gengo's analysis of the relationship between the nine
assemblies of the Kongokai mandala and the ritual sequence of the Kongokai
rite was treated as authoritative by virtually all later Shingon commentators. Although Shingon scholars have claimed that Gengo's work represents an oral tradition running through Kiikai, to my knowledge there is no compelling evidence
for this. On the relationship between the Kongokai rite and the mandala, see
Todaro, "Annotated Translation," 62-roo; Hatta Yukio /\EB:;~, "Kongokaiho no mondai" ~~IJW 7;t(l)ffi]/l!H, Mikkyobunka124 (1978): 104-78 (reverse numbering); Ueda, Shingon mikkyo, 265-67; as well as the comments interspersed
in Oyama, Chuinryu, r69ff. (for the Chiiinryii); Takai, Mikkyo jiso, 246ff. (Sanboinryii); and Saso, Homa Rites, r85-235 (Tendai).
35 See Takai, Mikkyo jiso, 276-79; Hatta Yukio, Gobushinkan no kenkyu 1i '1m
{-'II 0) lilf3'1: (Kyoto, Japan: Hozokan, r98r), 278-8o; idem, "Kongokai-ho," 103;
Ueda, Shingon mikkyo, 265; and Todaro, "Annotated Translation," 98.
36. The jojin-ne is also referred to as the kongokai daimandara ~ ~lj 3'f. :*: ~ ~
the kamma e
and the konpon e tN:;$:.; see Takai, Mikkyo jiso, 274-79;
and Adrian Snodgrass, The Matrix and Diamond World Mandalas in Shingon Buddhism, 2 vols., Sata-Pitaka Series, vol. 354 (New Delhi, India: Aditya Prakashan,
1988), 2:555-58.
37 A description of the iconography can be found in Ishida, Ryogai mandara,
141; and Snodgrass, Matrix and Diamond World Mandalas, 2:716-27.
38. Ibid., 2:725.

Notes to Pages 178-79


39. Miyano and Mizuhara, Shidokegyo shidai, "Kongokai," 17; Ozawa, Shido
shidai, "Kongokai," 46. The Chiiinryii version is identical to that of the Sanboinryii given above, except for the omission of the "discernment"; see Nakagawa
Zenkyo rp Jll
ed., Shidokegyo shidai: chuin pg !tim ff
Japan: Shinnoin, 1986).
4 o. For a description of this assembly, see Ishida, Ryogai mandara, r36-37; and
Snodgrass, Matrix and Diamond World Mandalas, 2:698-702.
4L T 2471: 7872b22-26.
42. One problem is that the scriptural references for this assembly mention
thirteen and seventeen deities, yet the iconography of the ichiin-ne of the nineassembly mandala contains a single figure; see Snodgrass, Matrix and Diamond
World Mandalas, 2:7or-2; and Mikkyo daijiten, 68c-69b.
43. Early Kongokai ritual manuals contain references to the mudriis and mantras of the jojin-ne, kamma e (that is, misai e), sanmaya e, kuyo e, shiin-ne, and ichiinne. Indeed, much of the available evidence suggests that a six-assembly Kongokai mandala preceded the nine-assembly model. Japanese scholars generally credit
Hui-kuo with the development of the nine-assembly mandala, but there is little evidence that would rule out a Japanese origin. In fact, according to Todaro, detailed
exegesis of the final two assemblies, the gozanze e and gozanze sanmaya e, are not
(1306-1362; see
found in any Kongokai ritual manual until the time of Goho
the Sanjukkan kyoo gyomon shidai
IX~, T 2226: 6r.37rb-87b).
Todaro argues that the evolution of what he calls "abbreviated manuals," a
category that includes virtually every Kongokai manual in use today, was made
possible by Gengo's explanation of the central jojin-ne as containing all nme assemblies. This allowed the last three assemblies (the gozanze e, gozanze sanmaya e,
and the rishu e) to be incorporated into the Kongokai cycle, thus ensconcing them
within the Tattvasarpgraha lineage. See the full discussion in Todaro, "Annotated
Translation," 84-99.
4 4 . For the rite itself, see Ozawa, Shido shidai, "Kongokai," 92-107; and
Takai, Mikkyo jiso, 279-89. Accounts of the thirty-seven deities of the jojin-ne
can be found in Ishida, Ryogai mandara, 121-28; Tajima Ryiijun, Les deux grands
mandalas et Ia doctrine de l'Esoterisme Shingon, Bulletin de Ia Maison FrancoJap~naise, n.s., 6 (Tokyo: Maison Franco-Japonaise; Paris: Presses Universitaires
de France, 1959 ), r66 -90; Shashibala, Comparative Iconography of the VajradhatuMandala and the Tattva-Sangraha, Sata-Pitaka Series, vol. 344 (New Delhi, India:
Sharada Rani, 1986), 103-285; and Snodgrass, Matrix and Diamond World Mandalas, 2:576-633. On the relationship between this specific rite and the mandala,
see Takai, Mikkyo jiso, 275-79; and Ishida, Ryogai mandara, u8-19.
45. Ozawa, Shido shidai, "Kongokai," 107. These sixteen properly represent
the thousand Buddhas of the bhadrakalpa. Although the jojin-ne represents the
thousand Buddhas, the sixteen are graphically represented in the sanmaya e, the
misai e, and the kuyo e. See Takai, Mikkyo jiso, 289-90; Ishida, Ryogai mandara,
128-29; Tajima, Les deux grands mandalas, 196-97; and Snodgrass, Matrix and
Diamond World Mandalas, 2:634-36.
46. The standard rite can be found in Ozawa, Shido shidai, "Kongokai," w8.


Notes to Page 180

Notes to Pages r8o-85

On the twenty deities, see Takai, Mikkyo jisi5, 291-92; Ishida, Ryogai mandara,
129-30; Tajima, Les deux grands mandalas, 191-96; Shashibala, Comparative Iconography, 286-309; and Snodgrass, Matrix and Diamond World Mandalas, 2:
47 Interestingly, one monk at KOfukuji saw the absence of a kansi5 in this segment of the nte as an advantage because it allowed a particularly tedious section
of the Kongokai ritual to be performed relatively quickly.
48. Whereas the manuals associated with the Kongokai ritual make no explicit
reference to the nine assemblies of the corresponding mandala, the same is not true
of the Taizokai. The Taizokai rite includes explicit allusions to the construction
of the altar/mandala described in the Mahavairocana-sittra. It is thus not surprismg to find references within the rite to the thirteen assemblies of the Taizokai
mandala and to the individual deities associated with some of the assemblies. For
example, the major assemblies of the Taizokai mandala are individually invoked
m the shoe shoshi5 bun ~l'l
segment of the rite (Toganoo, Himitsu jisi5,
So), ar~d the eight deities of the "Central Platform Eight-Petal Assembly" (chittai
hachzyo m rf1
arrayed around the central Vairocana are the subject of
the segment known as the "secret eight mudriis." (This rite does not appear in the
manuals associated with the Sanboinryii.) In addition, the culminating section of
the "contemplation of the sanctuary" (di5jokan) describes the central section of the
mandala in detail:
This stupa changes and becomes the Tathagata Mahii.vairocana. He is wearing a jeweled crown of five buddhas. He sits erect, dressed in fine silk robes
and wearing a necklace of many precious jewels. His body emits an exquisite light that illumines the dharmadhii.tu. In the eight directions above
the leaves of the lotus there are eight syllables: rm!l, val!l, Satfl, ha~, al!l, a,
vu, and yu. These transform into Ratnaketu, Sa111kusurnitarii.ja, Amitii.bha,
Divyadundubhi-meghanirgho~a, Samantabhadra, Manjusrl, Avalokite5vara,
Maitreya, and so on. They are surrounded front and behind, left and right,
~y all_the~eitres of the three-tiered mandala. (Nakagawa, Shidokegyo shidai,
Ta1zokm ; cf. M1yata Tmsen and Dale Todaro, eds., Handbook on the Four
Stages of Prayoga, Chitin Branch of Shingon Tradition fKoyasan, Japan: Department of Koyasan Shingon Foreign Mission, 1988), "Garbhako~adhatu,"
In short, there is greater consonance between rite and mandala in the case of the
Taizokai than in the Kongokai.
Be that as it may, the relationship between the Taizokai rite and its mandala is
subject to the same critical analysis given above for the Kongokai. Liturgical allusions to the assemblies and deities of the mandala within the rite, while explicit,
are perfunctory: With the sole exception of the central "eight-petal" assembly, the
mvocanon of the assemblies is performed en masse with a quick mudrajmantra
combmauon and no visualization component whatsoever (much like the sanjitshichtson mgon sequence of the Kongokai). And although the deities of the central


eight-petallotus of the mandala are invoked individually, here too it is in the midst
of a series of morphing images. Like the Kongokai, the presence of the sacred Taizokai mandala may well contribute to the imaginative or affective efficacy of rite
(see below), but it is not used as an actual prop or support for visualization in the
course of the rite itself.
Finally, the Goma is not associated with a pictorial or graphic mandala at all,
although an image of Fudo is usually enshrined as the principal deity of the Goma
Hall. If there is a mandala present, it is the elaborate Goma altar itself. (The fire pit
in the center of the altar, which receives the offerings, is Fudo's gaping mouth.) In
other words, Fudo's presence in the hall is not constructed in the "mind's eye"in an interior eidetic visualization. Rather, the blazing altar is itself the living body
of Fudo. I will return to this point in the conclusion to this chapter.
49 This rite is also known as the goho jojin
J;t, goso nikkan 1i 1fl A 1illl,
go ten jojin 1i ~i hx !1f, and so on.
so. Miyano and Mizuhara, Shidokegyo shidai, "Kongokai," 23-28; cf. Ozawa,
Shido shidai, "Kongokai," 54-61, and Nakagawa, Shidokegyo shidai. For Englishlanguage treatments of the Chiiinryii version, see Miyata and Todaro, Handbook
on the Four Stages, 30-33; and Payne, Tantric Ritual, 237-38.
sr. Canonical sources on the goso jiijinkan include the Chin-kang-ting ching
yii-ch 'ieh shih-pa-hui chih-kuei {JZ:
translated by Amoghavajra, where the five elements of this discernment are enumerated as (r) penetration of the original mind iiH
(2) cultivation of the bodhi mind ~~
(3) attainment of the vajra mind
(4) realization of the vajra body
~,and (5) perfection of the buddha body {?~ .f!J ll'ID iilfil (T 869: r8.284c2223). See also the Chin-kang-ting yii-ch'ieh chung fa a-nou-to-lo-san-miao-san-p'u-t'ihsin lun 'if: l!ijlj Nlfffi< {IJO $
, translated by Amoghavajra
(Tr665: 32574br7ff.), and the discussions in MZ 2:1236a-c; Oyama, Chuinryit,
174-78; Takai, Mikkyo jiso, 253-58; and Ueda, Shingon mikkyo, 285-91.
52. Miyano and Mizuhara, Shidokegyo shidai, "Nyoirin," 27-28; cf. Ozawa,
Shido shidai, "Jiihachido," 78-79.
53 See, for example, the analysis in Takai, Mikkyo jiso, I94
54 Miyano and Mizuhara, Shidokegyo shidai, "Kongflkai," 59; cf. Ozawa,
Shido shidai, "Kongokai," rr6-r7.
55. This is the shinchitshin shingon {,> r:p
("heart within the heart mantra") of Nyoirin Kannon, derived from the Sanskrit om varada padme hum. Each
syllable of the Sanskrit mantra has, in the Tantric system, one or more fundamental meanings. Thus the Japanese mantric syllable ha is from the Sanskrit va, which
stands for vac or "speech"; the syllable ra is from the Sanskrit ra, which stands for
rajas or "defilement"; dais from Sanskrit da, which stands for dana or "charity,"
and so on.
56. Miyano and Mizuhara, Shidokegyo shidai, "Nyoirin," 32-34; cf. Ozawa,
Shido shidai, "Jiihachidfl," 90-91; Takai, Mikkyo jiso, 203-5. The jirinkan is not
included in the Jiihachido of the Chiiinryii, although it is included in each of the
other three rituals of the Shidokegyo. Although the syllables that appear on the


Notes to Pages r86-89

Notes to Pages r89-92

moon disk differ depending on the identity of the principal deity, there are few
differences m the structure of the contemplation itself.
57 Mikkyo daijiten, 407c; see also s.v. "kannen," 4r8a. Another standard Shingon dictionary defines kanso as follows: "In Buddhism kan, like kansatsu, does not
merely mean to see
but to discern and illuminate with wisdom ~ ~ t "J l"5t
~lj !f~ .~ -9 <S
.. In mikkyo it refers to the operation of the mystery of mind
through 'bearing in mind' following mikkyo doctrines, as found in practices such
as the contemplation of the syllable A, the contemplation of the moon disk, the
contemplation of the attainment of the (buddha) body through the five marks and
the contemplation of the syllable wheel" (Sawa Ryuken {1:fD~iliJf, ed., Mikkyo
Jtten W~~\!if!!- [Kyoto, Japan: Hozokan, 1975], noa).
. 58. Nakamura Hajime r.p
Bukkyogo daijiten {:!ll~~,j;:~:if!!- (1975 ; repnnt 3 vols. m r, Tokyo: Tokyo shoseki, 198r), 197c-d, r96a (s.v. "kansatsu"),
197c-d (s.v. "kanso"); see also s.v. "kannen," 198a.
59: MZ 1:8o9b; see also s.v. kansatsu, 767a. Such continuity should not be
surpnsmg: all too often scholars of East Asian Tantric Buddhism are the unwitting
heirs ~f Shmgon polemics that exalt the uniqueness and superiority of Japanese
mtkkyo .. As such, scholars are prone to exaggerate the gap between kengyolJj~
(exoteric teachmgs) and mrkkyo (esoteric teachings), as well as the differences be(pure esoterism) and zomitsu 'flEW (mixed esoterism). On
tween junmitsu
these d1stmcnons see esp. Misaki Ryoshu :.:: ~)ili], "'Junmitsu to zomitsu' ni
. (::_ -) ~ >l", Indogaku bukkyogaku kenkyu 15, no. 2 (r96 7 ):
535-40; and Abe, Weavmg of Mantra, 152-54.
6o. T 216r: ss.ro64b26-29; and Kukai
Kobodaishi kukai zenshu "lJE.);:
8 vols., ed. Miyasaka Yusho
Onozuka Kicho tHffJ:*~
eta!. (Tokyo: Chikuma shobo, 1983), 2: 553; cf. the translation in Yoshito S.
Hakeda, Kukai: Major Works; Translated, with an Account of His Life and a Study
of Hrs Thought (New York: Columbia University Press, 1972), r 4 5- 4 6.
6r. T 2I6I: ss.ro65bio-rr; cf. Hakeda, Kukai, 148.
62. Hensozu, or simply henso
is most commonly used for depictions of
P:rr~ lands and hell realms. Nakamura Hajime notes, "In Esoteric Buddhism [hensozu] are known as mandalas" (Bukkyogo daijiten, ursd). See also Elizabeth ten
Grotenhuis, japanese Mandalas: Representations of Sacred Geography (Honolulu:
Umvers1ty of Hawall Press, 1999), 3, 15, 124.
63. Hakeda, Kukai, 145-46 (emphasis added).
64- On image consecration rituals, see, for example, Yael Bentor, Consecration
of Images and Stupas Indo- Tibetan Tantric Buddhism (Leiden: E. J. Brill, 199 6);
Richard Gombnch, The Consecration of a Buddhist Image," journal of Asian
Studres 26 (1966): 23-36; and idem, "Kosala-Bimba-Vannana," in Buddhism in
Ceylon and Studies on Religious Syncretism in Buddhist Countries ed. Heinz Bechert
(Gottingen, Germany: Vandenhoeck & Ruprecht, 1978), 28r-~o .
65. On miracle stories associated with Chinese Buddhist images, see, for
example, K01ch1 Shmohara, "The Maitreya Image in Shicheng and Guanding's
BIOgraphy of Zhiyi," in From Benares to Beijing: Essays on Buddhism and Chinese



Religions in Honour of Professor jan Yiin-hua, ed. Gregory Schopen and Koichi Shinohara (Oakville, Ont.: Mosaic Press, 1991), 203-28; idem, "Changing Roles for
Miraculous Images in Medieval Chinese Buddhism: A Study of the Miracle Image
Section in Daoxuan's 'Collected Records,'" and "Dynastic Politics and Miraculous Images: The Example of Zhuli of the Changlesi Temple in Yangzhou," in
Images, Miracles, and Authority in Asian Religious Traditions, ed. Richard H. Davis
(Boulder, Col.: Westview Press, 1998), qr-88 and 189-206, respectively; and Wu
Hung, "Inventing a Living Icon and a Theory of Divine Images in Medieval China"
(paper presented at the symposium "Living Icons in Five Traditions: Theories and
Practices," University of Chicago, 3r January 1998).
66. See esp. Bernard Faure, The Rhetoric of Immediacy: A Cultural Critique of
ChanjZen Buddhism (Princeton, N.J.: Princeton University Press, 1991); T. Griffith
Foulk and Robert H. Sharf, "On the Ritual Use of Ch'an Portraiture in Medieval
China," Cahiers d'Extreme-Asie 7 (1993-1994): 149-219; and Robert H. Sharf,
"The Idolization of Enlightenment: On the Mummification of Ch'an Masters in
Medieval China," History of Religions _32, no. r (1992): 1-31.
67. The list of images consists of three versions of the Taizokai mandala, two
of the Kongokai mandala, and one portrait each of Subhakarasif!lha (Shan-wu-wei
637-735), Vajrabodhi (Chin-kang-chih
671-741), I-hsing
705-774), and Hui-kuo
(746(683-727), Amoghavajra (Pu-k'ung
805). All are listed under the general heading "Buddhist images etcetera" {:!llf~~
(T 2r6r: 55.I064bii-21).
68. See note 31.
69. Ta-jih-ching shu k. E3
T1796: 39.626a9-rr.
70. This is also known as the "adornment of the altar place" (danjo shogon
li'ii~:Ul~); see Takai, Mikkyo jiso, 94-96; and Ueda, Shingon mikkyo, 40-45.
71. MZ 3: 2607a-9b; and Franklin Edgerton, Buddhist Hybrid Sanskrit Grammar and Dictionary, 2 vols., William Dwight Whitney Linguistic Series (New
Haven, Conn.: Yale University Press, 1953), 2:520.
7 2. On this issue, see the comments in Goepper, "Some Thoughts on the Icon,"
73 Gananath Obeyesekere coined the term "Protestant Buddhism" to refer to
the late-nineteenth-century Ceylonese revival of Theravada Buddhism (Gananath
Obeyesekere, "Religious Symbolism and Political Change in Ceylon," Modern
Ceylon Studies r, no. 1 [1970]: 43-63). The term has since been used by a variety
of scholars to characterize Western scholarly conceptions of the Buddhist tradition; see esp. Gregory Schopen, "Archaeology and Protestant Presuppositions in
the Study of Indian Buddhism," History of Religions 31, no. r (1991): 1-23.
74 The writings of David Kalupahana, Walpola Rahula, and Nyanaponika
Thera might be cited as examples. For an alternative reading of the Pali tradition,
see Peter Masefield, Divine Revelation in Pali Buddhism (Colombo: The Sri Lanka
Institute of Traditional Studies; London: George Allen and Unwin, 1986), as well
as the anthropologically oriented studies by Melford Spiro, Stanley Tambiah, and
Gananath Obeyesekere.


Notes to Pages 192-97

75 See Sharf, "Buddhist Modernism"; idem, "The Zen of Japanese Nationalism," in Curators of the Buddha: The Study of Buddhism Under Colonialism, ed.
Donald S. Lopez Jr. (Chicago: University of Chicago Press, 1995), 107-60; and
idem, "Experience," in Critical Terms in Religious Studies, ed. Mark C. Taylor
(Chicago: University of Chicago Press, 1998), 94-n6.
76. On the Meiji critique and persecution of Buddhism, see esp. James Edward
Ketelaar, Of Heretics and Martyrs in Meiji Japan: Buddhism and Its Persecution
(Princeton, N.J.: Princeton University Press, 1990).
77 Matsunaga Yukei
"Esoteric Buddhism: A Definition," in Mikkyo: Kobo Daishi Kukai and Shingon Buddhism, Special Issue of the Bulletin of the
Research Institute of Esoteric Buddhist Culture (K6yasan, Japan: Research Institute of Esoteric Buddhist Culture, Koyasan University, 1990), 27.
78. Matsunaga Yukei, Mikkyo: indo kara nihon e no denshO "(Ef'l. : 1 / r: iJ' C;,
~] ~4>="(Tokyo: Chii6k6ron, 1989), 25.
79 Ibid., 26-27.
So. Yamasaki, Shingon, 123.
81. Ishida, Esoteric Buddhist Painting, 27.
82. Ibid., 27.
83. Note that the key Japanese terms for "experience" -keiken and taiken-are
not found in premodern Japanese or Buddhist literature, but rather are neologisms
coined in the early Meiji as translation equivalents for the English "experience"
and the German erleben (or Erlebnis) respectively; see Sharf, "The Zen of Japanese
Nationalism," r24-25.
84. Tucci, Theory and Practice, vii, passim.
85. Ibid., 37
86. An analogous disparity between image and rite is found in some of the
forms of Hindu Saivism studied by Helene Brunner. Brunner argues that the
rites she investigated may well have developed independently of the images with
which they came to be associated; see Helene Brunner, "Mal}qala et yantra dans
le sivai'sme agamigue," in Mantras et Diagrammes Rituels dans l'Hindouisme, ed.
Centre National de Ia Recherche Scientifigue (Paris: Centre National de Ia Recherche Scientifigue, 1986), II-35; and idem, "L'image divine dans le culte agamigue
de Siva: Rapport entre !'image mentale et le support concret du culte," in L'Image
Divine: Culte et Meditation dans l'Hindouisme, ed. Andre Padoux (Paris: Centre
National de la Recherche Scientifigue, 1990), 9-29.
87. Charles D. Orzech, "Seeing Chen-yen Buddhism: Traditional Scholarship
and the Vajrayana in China," History ol Religions 29, no. I (1989): 109-13; and
ten Grotenhuis, Japanese Mandalas, 53-57 Many Shingon scholars continue to insist that the 1:1izclkai mandala is derived primarily from the Mahavairocana-siitra,
while the Kong6kai is based on the Sarvatathagatatattvasaf!lgraha, but all concede
problems in tracing certain features of the mandalas to these (or any other) Indian
scriptures. The Japanese literature bearing on the origins and evolution of the mandalas is extensive; see esp. Toganoo, Mandara no kenkyu; and Ishida, Mandara no
kenkyu. In addition to Orzech and ten Grotenhuis, works in English include Lo-

Note to Page I97


kesh Chandra, The Esoteric Iconography oflapanese Mandalas, Sata-Pitaka Series,

vol. 92 (New Delhi, India: International Academy of Indian Culture, I97I), 3349; Mammitzsch, Evolution ol the Garbhadhatu Mandala; Shashtbala, ComparatiVe
Iconography; and Snodgrass, Matrix and Diamond World Mandalas, 2:558-64 ..
88. The most sophisticated revisionist analysis to date of early Shmgon lS
found in Abe, Weaving of Mantra.