Sunteți pe pagina 1din 227

Journal of the International

Association of Buddhist Studies

Volume 28 Number 1 2005
David Seyfor RUEGG
The Kalawan Copper-plate Inscription: Early Evidence for Maha-
yana-type Thinking?......... .......................................... ...................... 3
Jinhua CHEN
Fazang (643-712): The Holy Man ................................................... 11
Richard D. McBRIDE
DMral}l and Spells in Medieval Sinitic Buddhism .... ........ .............. 85
A Crisis of Doxography: How Tibetans Organized Tantra during the
_12th Centuries............................................. .................................. 115
Studies in Indo-Tibetan Buddhist Hermeneutics (5). The mKhas pa
mams 'jug pa'i sgo by Sa skya Pal}cjita Kun dga' rgyal mkhan...... 183
Index to JIABS vol. 22-26................................................................. 221
e watermark
The Journal of the International
Association of Buddhist Studies
(ISSN 0193-600XX) is the organ
of the International Association of
Buddhist Studies, Inc. It welcomes
scholarly contributions pertaining to
all facets of Buddhist Studies.
JIABS is published twice yearly, in
the summer and winter.
Address manuscripts (two copies)
and books for review to:
The Editors, JIABS, Section de
langues et civilisations orientales,
Universite de Lausanne,
BFSH 2, CH-1015 Lausanne,
Address subscription orders and
dues, changes of address, and
business correspondence
(including advertising orders) to:
Dr. Jerome Ducor,
Treasurer IABS,
Section de langues et civilisations
orientales, Faculte des lettres
Universite de Lausanne, BFSH 2
1015 Lausanne-Dorigny
email: iabs. treasurer@orient.uniLch
Fax: +41 21 6923045
Subscriptions to JIABS are USD 40
per year for individuals and USD 70
per year for libraries and other
institutions. For informations on
membership in IABS, see back cover.
Cover: Cristina Scherrer-Schaub
Copyright 2005 by the International
Association of Buddhist Studies, Inc.
Printed in Belgium
Cox Collet
G6MEZ Luis O.
JAINl Padmanabh S.
LOPEZ, Jr. Donald S.
MAcDONALD Alexander
SHARF Robert
The copper-plate inscription found in the north-western part of the South
Asian subcontinent, at Kalawan (Taxila or raises questions of
importance for the historian of Buddhism and early Mahayana.
Recording the installation (Skt. of a corporeal relic (Skt.
sarira), and placed in the base of a stUpa (gahathuba: g[hastilpa) at Cha-
qasila, the inscription is dated in the 134th year of Aja (= c. 77nS CE?)l.
Its installer-dedicator was a certain upiisikii named CaqIdrabhi (or bha),
the daughter of the householder (Skt. grhapati) (Skt. Dharma)
and the wife of Bhadravala. In the inscription it is specified that she was
acting together with her householder brother (Skt. Nan-
divardhana), her two sons and daughter, her two daughters-in-law, and
JivaI}.aqIdi (Skt. Jivanandin), the teacher (Skt. iiciirya). For the historian of
Mahayana this inscription is of considerable potential significance, but in
its final portion it is difficult to interpret with complete certainty.
The Prakrit (GandhlirI) text in script as edited by R. Salomon
reads as follows in its final portion (lines 4-5):
... sa[rvajsti[line parigrahe puyafta
puyae pratiae hotu

I On the era of the Saka ruler Aja/Aya (= Azes), equated with the Vikrama sarrzvat of
58/57 BCE, see recently R. Salomon, Indian epigraphy (New York, 1998), p. 182, follow-
ing on A.D.H. Bivar, 'The Azes era and the Indravarma casket', South Asia archaeology
1979 (ed. H. Hartel, Berlin, 1981), pp. 369-76, and G. Fussman, BEFEO 1980, p. 43. The
identification of 'Aja', here without title or biruda, has been much debated (compare below,
n. 4), as has been the starting date of the Azes era.
As for the enigmatic expression gaha-thuba, its significance has been briefly remarked
upon by G. Fussman, JIABS 27 (2004), p. 242 n. 8.
2 R. Salomon, op. cit., pp. 269-70. Salomon has translated the end of the inscription as
follows: 'for the acceptance of the Sarviistiviidins. The country and the town are [hereby]
honored; [and] it is for the honor of all beings. May it lead to [their] attainment of Nirviil}.a.,
Journal of the International Association of Buddhist Studies
Volume 28 Number 1 2005
The precise significance for the history of Mahayana of this part of the
inscription not being entirely clear, it was not considered in the present
writer's recent article 'Aspects of the study of the (earlier) Indian Maha-
yana'3. The epigraph states explicitly that the installation was pJaced
by the donor in the parigraha - the 'acceptance' or perhaps rather 'keep-
ing' - of persons described (by a Prakrit equivalent) as Sarvastivadins,
one of the main schools (nikiiya) of the Sravakayana. At the same time
it hopes for the attainment of nirvii1)a by the dedicator herself and per-
haps by her entourage - possibly even by all sentient beings, an idea
characteristic of various stages of Mahayanist thought. At this point the
syntax is unfortunately ambiguous. The wish for nirvii1)a with no refer-
ence to all sentient beings appears frequently in inscriptions and texts
In developed Mahayanist thought the hope for the attainment of nirvii1)a
- nirvii1)iiviipti or its equivalent - by all sentient beings (sarvasattva)
The Kalawan inscription was first published by S. Konow, EI 21 (1931-32), pp. 251-9
(cf. id., 'Kalawan copper-plate inscription of the year 134', fRAS 1932, pp. 949-65). See also
1. Marshall, Taxi/a, i, p. 327; D.C. Sircar, Select inscriptions 2 (Calcutta, 1965), pp. 131-
2; and K. Tsukamoto, Indo-bukkyo himei no kenkyt1- A comprehensive study of the Indian
Buddhist inscriptions, i (Kyoto, 1996), p. 971. The inscription has recently been discussed
by G. Fussman, fIABS 27 (2004), pp. 241-2, who translates the end as follows: 'Given
in trust (parigrahe) to the Sarvastivadins. The kingdom and its corporations are honoured.
All beings are honoured. May it be <for their and our> attainment of nirviil}a. '
The use here of the word puyae 'in honour, for respect of' in connexion with 'all sen-
tient beings' finds numerous parallels in the use of this word, or of etymologically related
words, in connexion not only with, e.g., 'all the Buddhas' but also with 'all beings' and
with the dedicator's parents in Mathura inscriptions (see H. Liiders, Mathurii inscriptions
[Gtittingen, 1961], p. 124 with p. 80), as well as in connexion with Buddhas, with the ded ..
icator's parents, and with all sentient beings in inscriptions (where in relation
to 'all sentient beings' hita-sukha also appears; see S. Konow, Kharoshthf inscriptions
[Cn ii/I, 1929], pp. 77, 114 and 155, with pp. 5, 62, 65, 100).
3 fIABS 27 (2004), pp. 13-18.
4 See e.g. Konow's Kharoshthf inscriptions, p. 77: the Taxila silver scroll inscription
dated in the year 136 of Aya [/Aja] = 79 CE (?), where the dedicator himself may be the
only beneficiary of the wish for nirviil}a; the inscription of Ajitasena, father of Senavarma,
published by G. Fussman, BEFEO 75 (1986), p. 2; and the inscription dated in the year 98
of Azes published by A. Sadakata, fA 1996, pp. 308-09. On Senavarma's inscription dated
to no later than the middle of the first century CE, and also on the inscription from
Hidda dated to the year 28, see below. Compare also G. Schopen, 'Mahayana in Indian
inscriptions', IIl21 (1979), pp. 1-19. Concerning the wish that all sentient beings might
attain nirviil}a, see our 'Aspects of the study of the (earlier) Indian Mahayana' (as in n. 3),
p. 13 J.
stands beside another characteristic wish: that all sentient beings may
attall the Gnosis (ofthe buddha), anuttara-(buddha)jfziina
Now, concerning the possible reference in the Kalawan inscription to the
attainment of nirviil}a by all sattvas and at the same time the mention of
Sarvastivada, the following considerations may be noted:
1) It is possible that the reference to the attainment of nirviil}a is not
to. be connected directly with the preceding sarva[sva]tval}a, from which
5 The Govindnagar (Mathura) Brahmi inscription from the time of HuveOka/HuviOka
dated to the year 26 reads: ... sar(va)(sat)[v]a anut(t)ara(-) bud(dh)ajnana- pra(pnva)-(tu)
... See 'Aspects of the study ... ', p. 13. A further variant of the formula appears in another
inscription from the same reign recently published by H. Falk, 'Two new inscriptions from
the time of BIS 12-14 (2000), p. 32 f.: *imena kualamiilena sarvasattvanut-
tarasya nirantarasya jiianavaptaye*, where nirantara jiiiina 'unimpeded, or inunediate,
Gnosis' does not appear to be the precise equivalent of terms such as anantaratattva-
jiiiina, iinantaryasamiidhi, iinantyamiirga, iinantaryamurdhaprayoga found notably in
Prajiiaparamitii literature ( on which see E. Obermiller, 'The doctrine of Prajiiaparamitii,
AO 11 [1932]).
A variation on this very frequent formula appears in an inscription of Queen MahiidevI
of Gupta lineage on a bronze of the Buddha with his hands in dharmacakra position, now
in the British Museum, and dated to the fIfth century. The recent editor of this inscription,
O. von Hiniiber, reads (Die Palola [Mainz, 2004], p. 127): ... yad atra pUlJyaTfl
tad bhavatu sarvasattviinii(TfI) miitiipit[purvaTflgamana anuttarapadajiiiiniiviiptaye. (In
v. Hiniiber's book, the more usual formula anuttara-jiiiinaO is found in colophons of Qilgit
Mss. published on pp. 18,77 and 79; it is of course frequent elsewhere.) Whether the
form of words on the bronze in the British Museum - also known as an Icchawar inscrip-
tion (H. Liiders, List no. 11; K. Tsukamoto, op. cit., p. 612) - represents a conflation or
contamination with the expression am[ta(pada/dhiitu) (cf. ibid., p. 179 n.) - equivalent
to nirviilJadhiitu (7) - is not perfectly clear. - For amuda dhatu, a Prakrit equivalent of
am[ta-dhiitu (: nirviilJadhiitu), see line 12b of Senavarma's inscription in O. von Hiniiber,
. Beitriige zur Erkliirung der Senavarma-Inschrift (AWL Mainz, Stuttgart, 2003), p. 37.
Cf. lJivalJadhatu [Le. nirviilJadhiitu] in line 7c of the same inscription, ibid., p. 23. Sena-
varma's inscription has again been discussed by G. Fussman, BEFEO 90-91 (2003-04),
It is interesting to note that the inscriptions on the British Museum bronze and on the
Kalawan copper-plate both originate with women, the fIrst a queen and the second an
upiisikii. For a further instance see n. 12 below. It is to be recalled that, in one of the main
canonical texts of the tathiigatagarbha teaching according to which all sattvas carry in them-
selves the potentiality of becoming tathiigatas or buddhas, the SrfmiiliidevfsiTfl-
haniidasutra, the Buddha's interlocutor and speaker of this teaching is Queen SIimruadevI.
The inscriptional materials gathered in this article are to be added to those noticed in
our Theorie du tathiigatagarbha et du gotra (paris, 1969), p. 31 n. 2.
it is separated by the word puyae. In other words, the syntax allows the
rendering: 'May [this installation] be for respect towards all sentient
beings, for attainment of nirviilJa [viz. by the upiisikii and perhaps her cir-
cle, namely her relatives andthe A.carya, but not by all sentient beings]'.
This interpretation would be in conformity with the text of numerous
other inscriptions. The syntax is ambiguous, however, and the words 'for
respect towards all beings' and 'for attainment of nirviilJa' are juxtaposed
asyndetically, with the only verb hotu at the very end. Hence, it does not
seem syntactically impossible, or altogether unnatural, to regard the attain-
ment of nirviilJa as here relating to all sentient beings6.
2) There perhaps existed no problem, at least in the view of the
installer-dedicator. For Sravakayanists may hold that all sentient beings
are able to attain nirviilJa (of the kind classified scholastically as that of the
Sravaka, perhaps along with an anuttara-jftiina distinct from anuttara-bud-
6 This is the way some previous translators of the Kalawan inscription have understood
its final portion (see n. 2 above), but not Konow who translated 'may it be for the obtain-
ment pf Nirvfu:1a' without connecting this phrase with 'all sattvas'.
It may be noted that a question concerning exactly to whom the hoped for attainment
of the amuda dhatu (Skt. amrta-dhiitu: nirviil)a) is relatable arises also in interpreting line
12b of the inscription of Senavarma, ruler of O<;li (dated to no later than the middle of the
1st century CE; see below). See O. von Hiniiber, op. cit., p. 37, who has observed (op. cit.,
pp. 47-48) that it is not certain whether, when speaking in his inscription of the amuda
dhatu, Senavarma was aiming for nirviil)a for himself alone or for all beings; v. Hiniiber
adds that the (unclear) context suggests the latter interpretation. In the inscription referred
to above (n. 4) of Ajitasena, the father of Senavarma, the reference to nirviil)a apparently con-
cerns the dedicator alone.
7 What the view of the iiciirya mentioned in the inscription might have been we do not
know. .
Compare Pali sambodhiparayana (Skt. sambodhiparayal)a). Several references are found
in PTSD, s. u. sambodhipariiyana, and in Nanatiloka, Buddhist dictionary, s. u. bodhi.
Concerning the identity or difference of the (vimukti) of the Sravaka and the bud-
dha, as well as of their path (miirga) and yiina and also of their jiiiinas, see the theses no.
43 of the Sarvastivadins, no. 22 of the Malllsasakas, and no. 2 of the Dharmaguptakas in:
A. Bareau, Les sectes bouddhiques du Petit vehicule (paris, 1955). See also P. Jaini (ed.),
Abhidharmadipa with Vibhiiijiiprabhavrtti (patna, 1959), p. 205 f. (on nirviil)a of the Sravaka,
Pratyekabuddha, and Bodhisattva or Buddha), and p. 358 f. (on the three bodhis); and
W. Rahula, Zen and the taming of the bull (London, 1978), p. 71 ff.; L. de La Vallee Poussin,
L'Abhidharmakosa de Vasubandhu, vol. v (Louvain, 1925), pp. 267-8 (on the soteriological
implications of the avyiikrtavastu concerning whether the world has an end); and D. Sey-
fort Ruegg, Theorie du tathiigatagarbha et du gotra, Part ii (on universal Awakening, the
ekayiina, and whether sarrzsiira has an end). - The precise interrelationship in the history
3) The installer-dedicator of the relic being a lay-follower (upiisikii),
she might possibly have been quite unaware of the doctrinal issue raised
by a reference to nirviiIJ-a as a universal goal for all sattvas.
4) There is in fact no problem here because, within Vinaya schools
there were to be found not only Sravakaylinists but also Mahaylinists.
As is known, the word nikiiya may denote either a dogmatic school of the
Sravakayana or a Vinaya school. (The residence of both Sravakaylinists
and Mahaylinists in the same place is known from the seventh-century
account by HSiian-tsang
.) Here in the Kalawan inscription, the reference
to Sarvastivada may very well reflect the second usage of the word
(despite the dogmatic content ofthe name 'Sarvastivada'). In the history
of Buddhism, the Vinaya of a Sravakaylinist Nikaya has in fact been used
by Mahaylinists (those of Tibet for example have adopted the Vinaya of
the MfiIa-Sarvastivadin school).
On the basis of what is written in the Kalawan inscription, it is scarcely
possible to establish conclusively and without a shadow of doubt which
of the aforementioned considerations are the most pertinent, and which
may be preferable in the present context. In general, what is known from
the history of Buddhism would incline one to attach importance to the
fourth point. But there is no absolute certainty here.
In summary, the view that all sentient beings are to attain nirviiIJ-a as a
universal goal being characteristic of Mahaylinist thinking, if the Kalawan
inscription, dated as it has been to 77n8 CE (?), is to be considered
one of the very oldest known inscriptional attestations of the idea, it does
not seem that, per se, the reference in it to Sarvastivada need inevitably
constitute an insuperable obstacle in the way of regarding the inscription
as providing evidence for (proto-)Mahayana
And the fmd-spot of the
of Buddhist thought between sriivaka-nirviir;za, sambodhi. and anuttarasamyaksambodhi is
perhaps not as clear as could be wished. On sambodhi compare also n. 12 below. As for
the anuttara-jiiana, it has on occasion been specified as being .the anuttara-buddhajiiiina
(in the above-mentioned Govindnagar inscription and then later).
8 Cf. our 'Aspects of the study .. .', p. 31 and p. 50 n. 81.
9 Together perhaps with line 12b of Senavarma's inscription. See n. 6 above, and below.
Beside the ekayiina or One-Vehicle theory of universal Awakening ([samJbodhi)
according to which all sattvas are sooner or later to attain buddhahood, there has existed
in Mahayanist thought a form of the triyiina or Three-Vehicle and triple gotra doctrine
Kalawan inscription would then confirm that the north-western part of the
subcontinent was at this quite early time a hearth of Mahayana-type think-
ing. The problem of the syntactic construction of the [mal portion of the
inscription does, however, leave room for uncertainty as to just what.stage
of Buddhist thinking may be reflected in it. A similar problem arises also
concerning the mention of the attainment of the am[ta-dhatu in Sena-
varma's inscription, also in Gandhfui, dated to no later than the middle of
the fIrst century CEo
Depending on just how it is to be understood, then, the Kalawan inscrip-
tion might be regarded as possibly bearing witness to a line of Mahayiinis-
tic thinking (perhaps even one on the way to developing the universalist
soteriology of tathagatagarbha teaching according to which all sentient
beings without exception are to attain buddhahood). On this depends in
tum whether this inscription - together perhaps with Semivarma's -
according to which sattvas are divided into three detenninate 'lineages' (niyatagotra), each
with its own distinct, and 'expressed', spiritual 'gene', these thIee being the gotra of the
buddha or bodhisattva, that of the pratyekabuddha and that of the sriivaka (whose goal
in principle is arhatship rather than buddhahood); these three gotras then conduce to dif-
ferent goals. In addition, there was recognized an undetennined (aniyata), 'unexpressed',
gotra, which is capable of developing into one or the other of the three gotras and of finally
attaining one or the other of the three distinct' forms of liberation just mentioned. It would
be idle to speculate about which of these main theories was statistically and demograph-
ically predominant in earlier Mahayana; it may indeed be the case that this question is
unanswerable on the basis of the available documentation. At all events, among Maha-
yiinists, the ekaytina theory of universal Awakening became very widely spread among
Madhyamikas, the triytina theory being held by some Y ogacara-Vijfianavadins. But at
the time of the earlier inscriptions discussed in this paper - i.e. just before Nagarjuna, the
source of the Madhyamaka school, and previous to, the source of the Y ogacara-
Vijfianavada - these two directions within Mahayana had presumably not yet crystallized
into 'fwo distinct schools of thought. The tathtigatagarbha theory is a particular form of
the ekaytina theory teaching the universal Awakening and fmal buddhahood of all sattvas.
But it has to be clearly recognized that the characteristic images and metaphors of the con-
stituted tathtigatagarbha doctrine as we now know it have not been employed in the early
inscriptions under discussion. Still, several inscriptions on the one side and on the other
the tathtigatagarbha and prakrtisthagotra doctrines do have in common the notion of a
dhtitu as a precious relic-deposit of the tathtigata in a stapa, which it sanctifies and enlivens,
and as the precious spiritual element or matrix of the tathtigata which is present in all sen-
tient beings and enables them ultimately to attain buddhahood. Concerning some of the many
images and metaphors used in the gotra and tathtigatagarbha teachings, see D. Seyfort
Ruegg, 'The meaning of the term gotra and the textual history of the Ratnagotravibhtiga' ,
BSOAS 39 (1976), pp. 341-63; and on the idea of dhtitu and the values of this word, see
'Aspects of the study .. .', p. 27 n. 36.
constitutes the earliest known inscriptional evidence of what can be
described as Manayana-type thinking, or whether the somewhat later
Hidda inscription from near Jalalabad dated to the year 28 of the K u ~ f u ) . a
era - which refers to the requisites for the nirviil}a of all sentient beings
(sarvasattva)10 - is still to be regarded as providing the earliest known
inscriptional evidence of the kind for Mahayana 11.
n seems in any case advisable to retain S. Konow's rendering (see
n. 6) where nothing has been added in brackets by the translator, reserv-
ing interpretation for the annotation and commentary. The ambiguity and
uncertainties noted above confirm once more how difficult it may be to
cite an inscription as conclusive evidence for doctrinal or religious devel-
opment in Buddhism
It is also necessary to keep in mind that in many
a case there is no neat, clean and abrupt break between (proto-)Mahayana
and what preceded it: often we have to do with continuing development
rather than with total discontinuity.
10 See S. Konow, Kharoshthf inscriptions, p. 158, and EI 23 (1935-36), pp. 35-42; cf.
'Aspects of the study .. .', pp. 14-15.
11 If the Kalawan inscription is to be regarded as attesting Mahayana-type thinking, it
would further contribute to reducing any possible 'non-alignment' between our Indian-language
documentation and the Buddhist sources in Chinese concerning the earlier history of Indian
Mahayana. On such non-alignment see G. Schopen, 'The Mahayana and the Middle Period
in Indian Buddhism: Through a Chinese looking-glass', EB 32 (2000), pp. 1-25.
12 Concerning the problem posed by the mention in Asoka's Rock Edict VIII of his
departure for sambodhi, see our 'Aspects of the study .. .', p. 14 n. (cf. n. 7 above).
Many centuries later, in the colophon of a Gilgit manuscript of the SaJpghl'itasutra ded-
icated by Queen Devasirika, we read: anuttara1?J vimalavirajanirmmalavuddhabodhi(1?J)
sprsatu: (see O. von Hiniiber, 'Die Kolophone der Gilgit-Handschriften', StII 5/6 [1980],
p. 69; id., Die Palola fjahis, p. 25). In this case (unlike the cases cited in n. 5 above where
the beneficiaries are all sentient beings), the attainment of the supreme buddhabodhi is wished
for by the dedicator for herself (and perhaps for her entourage).
By and large, the importance of the Tang Buddhist monk Fazang t ; ; ~
(643-712) has still been so far appraised and appreciated in terms of
his contributions to Buddhist philosophy, and especially his status
as the de facto founder of the East Asian Avata'!lsaka tradition,
which has been well known for its sophisticated and often difficult
philosophical system. The choice of modem scholars to focus on
Fazang's philosophical contributions is certainly justifiable. Most of
his extant writings are indeed philosophical texts. This "Avatarp.saka-
only" vision of Fazang might well give the impression that he was an
armchair philosopher, who was almost exclusively preoccupied with
metaphysical speculations, with little or no interest in other forms of
Fazang's historical and hagio/biographical sources present to us three
different types of images, all quite contrary to the sober, if not stem,
impression that his reputation as a great philosopher might have cast
upon us: first as a politician who deliberately and shrewdly added his
significant weight to the balance of power when it reached a critical
point of exploding into major and fundamental sociopolitical changes
("revolutions" is perhaps not too strong a word); second, as a warrior
who fought the enemy of the empire, not by sword, but by charms; and
eventually, as a mediator between humanity and the heavens when dishar-
mony started to develop between them and threatened the very structure
of the human world. It is probably in terms of such a status as a go-
between of humanity and heaven, or - more in line with Chinese tra-
ditional ideas - an adjustor if not manipulator of yin and yang, that we
ought to discuss Fazang's function and image as a holy man in medieval
Journal of the International Association of Buddhist Studies
Volume 28 Number I 2005
I) Fazang the Court Politician
A twenty years junior of Wu Zhao j \ ' ; ~ (623-705), as Empress Wu
was personally known, Fazang outlived her by seven years. When he
started to distinguish himself as a young Buddhist scholar towards the end
of the 660s, the empress had already managed to place herself at the cen-
ter of power stage. It seems therefore reasonable to say that Fazang spent
the majority of his career under the shadow of Empress Wu, who had
been the actual ruler of China in the half century spanning from 655,
when she became the new empress of Gaozong (r. 649-683), to the begin-
ning of 705, when she was forced into abdication. During this period,
she first (655-683) shared supreme power with her husband, then after a
short interval, during which her first emperor-son Zhongzong (r. 694,
705-71 0) maintained his nominal rule for less than two months, she
wielded the state power as the Regent of her second emperor-son Ruizong
(r. 684-690,710-712), a puppet manipulated by her, until 16 October 690,
when she replaced the Great Tang with her own dynasty the Great Zhou.
This fact alone accounts for the irreplaceable importance of the empress's
influence on Fazang, which, in tum, justifies the amount of attention that
we are going to pay to their relationship.
Ll) Fazang and Empress Wu: 671-690
The earliest dated association between Fazang and Empress Wu started
from Xianheng 1 (27 March 670-14 February 671), when the empress, at
the recommendation of several prestigious monks, assigned Fazang, who
was then still a novice, to the Taiyuansi :;tJ]i: ~ . Built on the foundations
of the old residence of Empress Wu's mother Madam Rongguo !iikJillll
(579-670), who died on 22 August of that year, this monastery was ded-
icated to her posthumous welfare!.
1 See Fazang's funeral epitaph by Yan Chaoyin IiMJlIlJ (?-713?) shortly after his death
in 712, the "Da Tang Da Iianfusi gu Dade Kangzang Fashi zhi bei" (hereafter "Kang Zang
bei"), T 50: 280b15-17; a more detailed account can be found in Tang Tae Ch'onboksa
kosaju pon'gyong taedok Popjang hwasang chOn (hereafter Popjang chOn), T 50: 281b15-
20. For the epigraphic evidence establishing Madame Rongguo's dates, see Forte 1996:
It seems thatfrom the very beginning, Fazang succeeded in capturing
the attention o{the empress, as Ch'oe Ch'iwon tells us that shortly after
he entered the Taiyuansi, in the duanwu Jiffiflf festival (later to be known
as Dragon Boat festival) of an unspecified year, which was either-during
the time when Fazang entered the Taiyuansi or several years after,
Empress Wu showed a significant favor to him by sending him a set of
five monastic robes, as a match to the symbolism implied in the duanwu
festival, which was annually celebrated on the fifth day of the fifth month.
This gift was accompanied with a short but highly laudatory message:
As now the season turns to the fifth month, it is time for enjoying the zongzi
1*'f dumpling (jiaosu Now the weather is gradually getting hot, does
the Master's "spiritual body" (daoti J!!t'lD still feel light and comfortable?
It happens to be the good season for wearing the longevity-thread (changsi
and the excellent time of receiving the "ribbon of life" (minglu iPi:t)2.
Now We have sent the five kinds of monastic dress
to match the number
implied in the festival of duanwu (5.5). It is Our hope that following this
season of collecting Artemisia-leaves\ you, 0 Master, will grow evergreen
like the aging of a pine! The lamp of dharma-transmission will be alight for-
ever and you will always be the guiding head. This brief letter was written
[merely] to convey Our regards and We will not linger on now.
*$, zlX. 1mBiPjlztB1Utzf, ikillitii, ';lit
Given that he was then no more than a Buddhist novice, the amount
of attention that Empress Wu paid to him is remarkable. In addition to his
reputation as an excellent Buddhist scholar, there must have been some
more profound factors contributing to this extraordinary success. They
might include Fazang's prestigious family background. Some of Fazang's
ancestors were state ministers in their original home, the kingdom of
2 Here both the changsi and minglii refer to the changmingsi ft1f!;-i* (or changminglii
ft1f!;-*,), a bunch of five-colored threads, which it was customary to wear during the
duanwu festival in hope of extending one's life, hence the name of changmingsi/chang-
minglii - "longevity thread."
3 The five sets of monastic dress included sanghiiti, uttariisangha, antarviisa, sal'{lkalqikii
and kusula (kusulika).
4 It was also a custom during the duanwu festival to collect the Artemisia-leaves, which,
put on the doors, were allegedly capable of warding off evil spirits.
5 Popjang chOn, T 50: 281b24-28.
Kangju J*Ji5 (Samarqand). Fazang's father received the posthumous func-
tion of Commandant of the Left Guard (zuowei zhonglang jiang ft1f.p
which was "rank four, second class" (4b) in the bureaucratic hier-
archy (Hucker 1985: 191,526). This suggests that Fazang's father might
have been active in contemporary aristocratic circles
Another likely reason for Fazang's access to the royal family was the
close relationship between his teacher Zhiyan ,&-Vi (602-668) and Li
Xian *it (653-684) (posthumously known as Crown Prince ZhaI).ghuai
:!J'tt), a son of Gaozong and Empress Wu, who became the Heir Appar-
ent on 3 July 675, a position he held for five years until he was deposed
on 20 September 680 on a charge of treason
. We do not know for cer-
tain how long Zhiyan associated himself with Li Xian. Since it was in
the capacity of Prince Pei, a princely title he achieved on 18 October 661
that Li Xian started to associate with Zhiyan, who died on 8 December
668, we can assume that the association lasted for a period of time falling
between these two dates. Given the close relationship between Zhiyan
and Li Xian on the one hand and the extent to which Fazang was favored
by Zhiyan on the other, it seems likely that Fazang would have had reg-
ular opportunities to visit the imperial court and attract attention from
Empress Wu.
The importance of the Avatal?1saka siitra in Fazang' s relationship with
Empress Wu becomes more evident when we tum to examine another
event in which the sutra was the subject of a series of religious activities
which were held on the eve of Empress Wu's "usurpation" in 690. On
the night of 2 February 689, the emperor (nominally Ruizong, but actu-
ally Empress Wu, who was then "supervising the court" as the Regent)
ordered Fazang and others to build a "high Avatatpsaka-seat" (Huayan
gaozuo and a bodhimafla of "Eight Assemblies" (bahui
at the Northern Gate ofXuanwu 1rlt\;. The assembly was nominally con-
vened for the purpose of elucidating and promoting the wondrous
A vatal?1saka sutra.
6 P6pjang chOn, T 50: 281a19-21.
7 Zizhi tongjian 202.6377, 6397. The association between Zbiyan and Li Xian is reported
in Fazang's Huayan jing zhuanji, T 51 :3.163c20-22, in which Li Xian is referred to as
Prince of Pei 1rP.
8 Zizhi tongjian 200.6325.
Empress Wu honored the occasion with a poem. In the short preface,
She tells us that in the intervals between conducting national affairs she
attended the A vatmpsaka lectures, which provided her an opportunity to
"watch the depth and breadth of the wisdom and eloquence, and to
observe the performance of the 'dragon and elephants' (that is, 'eminent
monks ')."9 She also congratulates herself that by virtue of her previous
cultivation she was able to understand instantly the parts where she had
deep-rooted doubts
Empress Wu here suggests that her interest in the
AvataY[lsaka siUra was not extemporaneous, but long-lasting (and had
possibly even begun in an earlier life as she suggested). TIlls does not seem
a perfunctory remark given her familiarity with Buddhism and the sutra
in particular, which can also be seen in the poem properll.
TIlls apparently purely religious event turns out to have other dimen-
sions as soon as we scrutinize it against the current social and political
context. The two years from 689 and 690 were crucial for Empress Wu's
political ambition. She was then keenly plotting for her formal usurpation
of supreme power. The histories record a series of important measures that
Empress Wu and her ideologues adopted to justify her aspiration for a new
dynasty in her own right. Since I have delineated elsewhere (Chen 2003:
327-329) these main measures, let me here confine myself to an overall
conclusion on the significances of this AvataY[lsaka assembly had for
Empress Wu and her ideologues:
It was only a couple of days after a series of events related to the com-
pletion and celebration of a huge complex generally known as the ming-
tang aA -g (the "Luminous Hall"), the most important architectural expres-
sion of Empress Wu's sacral-political institution, that Empress Wu ordered
Fazang and other monks to convene the A vatmpsaka assembly and the
subsequent vegetarian feast. What made such a dharma-assembly notice-
able was not only its size, but also Empress Wu's deep involvement with
it - in addition to her own personal participation in the assembly, she
honored it with an elegantly composed poem. It warrants our particular
attention that the assembly was held in the vicinity of the north gate of
9 Huayanjing zhuanji, T 51: 3.l64bl-2.
10 Huayan jing zhuanji, T 51: 3.164b2.
II Huayanjing zhuanji, T 51: 3.164b3-7, translated in Chen 2003: 327.
Xuanwu, close to the place where stood the mingtang complex, which was
successfully brought to completion only ten days earlier. Given the spa-
tial and temporal proximity between the mingtang and the A vatrupsaka
Dharma-assembly, I suspect that the two events (or more accurately, the
two series of events) associated with them, were executed for the same
or similar purposes, among which was the politico-religious propaganda
leading to the formal replacement of the Great Tang with the Great Zhou
Dynasty on 16 October 690. Only by referring to this historical context
can we do full justice to this religious events organized and guided by
Fazang (Chen 2003: 329).
The efforts that Fazang, like other Buddhist leaders at the time, made
to legitimate Empress Wu's unconventional (indeed, anti-conventional)
and unprecedented rule as a female monarch and the enthusiasm with
which he embraced the current dharma-prosperity rendered possible by
Empress Wu and further anticipated even the grander vision of a truly
world empire of Buddhism in China are best expressed in the following
passage that Fazang wrote in featuring this Avatarrzsaka dharma-assem-
bly and introducing the empress's poem dedicated to it:
The August Emperor of Divine Spirit (Shengshen huangdi *$ . 1if) of the
Great Zhou, having planted the seeds of Way in previous kaZpas, has been
widely supported by myriads of people
. As prophesized by the Buddha
in the Dayun jing, Her Majesty has been able to turn and manipUlate the
Golden Wheel. In accordance with the predictions in the Graphs from the
River LUG, Her Majesty has come to rule the country by beating the jade-
drum (yugu Being divine and marvelous, Her Majesty has performed
the "Six Kinds of Supernatural Powers,"13 which know no limit. Being of
supreme goodness and perfect beauty, Her Majesty has expanded to the
;;boundless spheres the transformation in terms of "Ten Good Acts."14 Her
12 This refers to these lines in the Daode jing:
JR (Zhu 1984: 268), which Lau (1963: 73) translates as, "Therefore the sage
takes his place over the people yet is no burden; take his place ahead of the people yet causes
no obstruction. That is why the empire supports him joyfully and never tires of doing so."
13 The liu shentong (Skt. abhijfitil}) refer to the six kinds of supernatural power
attributed to the Buddha.
14 The "Ten Good Acts" (shishan are those of avoiding (1) "killing" (shasheng
(2) "stealing" (toudao 1flI:i), (3) "committing adultery" (xieyin (4) "lying"
(wanyu jHft), (5) "speaking harshly" (ekou jffi\ Q), (6) "speaking divisively" (liangshe
(7) "speaking idly" (qiyu (8) "being greedy" (tanyu Jtl'iX), (9) "being angry"
(chenhui and (10) "having wrong views" (xiejian
Majesty exceeded the rulers of the Xia arid Yin (Le. Shang) in [her com-
. passion to animals by] "opening up the nets"15 (jiewan and [show-
ing sympathy i:o the people by] "wailing over the criminals" (qigu N*)16.
Thus, a jade-citadel (guicheng :;l1;:lJilG) is surrounded by River Fen 717j(, the
sun of wisdom equally spread its light into every tiny being. Therefore,
"wearing herself out from head to foot," 17 Her Majesty has exerted her
energy in helping people with her "ten powers."18 Stopping with only a
mouthful in the middle of eating and binding up her hair in the midst of a
bath [in order to grant audience to those useful to the state] just like Duke
Zhou fi(J0, Her Majesty has kept having "Four Necessities" 19 delivered [to
the s<Up.gha]. With the finest metal cast and the sandalwood carved for [metal
and wooden] statues, the roseate clouds are mirrored as deeply as one thou-
sand gates [of the monasteries]. [Sailing through the oceans by] floating on
wooden cups and [climbing mountains by] shaking their staffs, [eminent
monks] have been coming to gather within the nine-layered walls of the
imperial palaces. Compared to these, how could the extraordinary propi-
tious signs that happened during the Han and Wei dynasties and the profound
faiths in Buddhism displayed in the Liang and Qi dynasties be worth men-
tion? The [government's] efforts to open up the treasure-stores within
the dragon palace and greet the magnificence and beauty of the jade-gates
[of new monasteries] have keep going on just as the sun and moon move
[across the sky], without stopping even for a moment. The compositions of
chanting and eulogizing the virtue of the Buddha and the music singing praises
of the dharma-words have been spread on musical instruments, both strings
and wind, and piling up in paper and ink.

15 This refers to the story that the Shang ifiI King Tang in hunting, ordered to leave
three sides of the four-sided net open so that only the animals without intention to live on
got caught. This enlIanced the feudal princes' (zhuhou admiration for Tang's com-
passion, which they believed extended from the human to aninIals. See Shiji 3.95.
16 According to Liu Xiang JtJ (77BC-6BC), in seeing indicted criminals on the road,
Yu iJ" the King of the Xia, came down from his chariot and became so overwhelmed by
his sympathy that he could not help but wail over their misfortune and his own dereliction
of duty, which caused them to fall into crimes. See Shuoyuan l.4b.
17 This refers to a saying in the Mencius: *mJltJ1l, (Yang Bojun
1960: 43), which Lau (1970: 187-188) translates as, "Mo Tzu advocates love without dis-
crimination. If by shaving his head and showing his heels he could benefit the Empire, he
would do it."
18 The shili +f.7 (dasabalani) indicates ten kinds of powers of awareness specially pos-
sessed by the Buddha.
19 Siyi Il!ItI< indicate the four kinds of necessities required by the monastic life: those
of food, clothing, shelter and medicine.

13 Jj tt.:;,
Fazang here has woven Chinese traditional ideas (mainly Confucian)
and Buddhist ideologies into a coherent discourse with impressive skill.
These two paralkl sentences - "With the finest metal cast and the san-
dalwood carved for [metal and wooden] statues, the roseate clouds are mir-
rored as deeply as one thousand gates [of the monasteries). [Sailing
through the oceans by] floating on wooden cups and [climbing moun-
tains by] shaking their staffs, [eminent monks] have been coming to gather
within the nine-layered walls of the imperial palaces"
- are of particular interest to scholars
interested in Empress Wu's court Buddhism. It seems to me that two
expressions (zhushen SiJt and diaotan in the first sentence refer
to the astronomic device called dayi :;k:1i ("Great Regulator") and the
immense lacquer statue of the Buddha that were installed within, respec-
tively, the observatory lingtai (lit. "Numinous Terrace") and the
Heavenly Hall (tiangong - two essential parts of the mingtang
The second sentence, on the other hand, features the regular
congregations of Buddhist monks within Empress Wu's palace chapels,
which were characteristic of the monastic institution under her rule

It seems that as the empress was approaching the unprecedented step
of taking supreme power, not only in fact but also in name, her reliance
on pazang increased daily. A couple of years before the Avatarp.saka
assembly, she had just asked for Fazang's help in abating the damage
20 Since here is involved a pair of parallel sentences, one character must be redundant
in the sentence (paralleled by 1Jf!i Given
can find their parallels in 1Jf!i (1J, S, respectively), either or 11 is
redundant. While I believe that it is 11, should be emended to ll. The whole sentence
should be reconstructed as
21 Huayanjing zhuanji, T 51: 3.164a12-22.
22 See Forte, 1988 (passim), for dayi and liantang.
23 For the latest study of Empress Wu's palace chapels, see Chen Jinhua 2004: 113-
caused by a drought. We are told that Fazang did a great service
to the state by constructing a platfonn at Xirning si Iffi Ij, to pray for rain.
That monastery had been built by her and her husband in 658 to celebrate
the successful recovery from illness of their Heir Apparent, the four year
old Li Hong (652-675).
According to the Korean monk Kyunyo :l$]:I!1l (923-973), Fazang's
career suffered a severe setback in 694 or early 695, sometime before the
arrival of (652-710) in China. Exasperated by Fazang's inter-
pretation of a Buddha as a "provisionally-named bodhisattva" (jiaming
pusa Fuli 1ltft (fl. 680-705), a Buddhist monk who was also
very influential under the reigns of Gaozong and Empress Wu, impeached
Fazang for advocating such a heterodox theory and urged that Fazang be
punished in accordance with the law. As a result, Empress Wu decreed
Fazang's exile to the Jiangnan rrli area, whence he was not called back
to the capital until and Fuli encountered insunnountable dif-
ficulties in translating the chapter on "Puxian" of the new version
of the A vatarrzsaka siitra that brought to China. In the course
of cooperating with Fazang in the translation project, Fuli even once went
so far as to coerce him to alter some passages in the original text in order
to fit his own theories
This record is not found in any other sources.
However, Fazang's banishment from the capital (though only a brief one)
seems likely given his absence from two extremely important religio-
political projects that were carried out in 693 and 695 respectively - the
re-translation of the Ratnamegha sutra, which resulted in the ten fascicle
Chinese text titled "Baoyu jing" (Skt. Ratnamegha sutra; Satra
of the Precious Rain), and the compilation of an officially sanctioned
Buddhist catalogue which included (and thereby canonized) those texts
(some of dubious origins) that had been newly translated under the aegis
of the empress.
The Baoyu jing is believed to have contained passages interpolated by
the translators for the purposes of providing further ideological support
for Empress Wu's female rule. The translation project, led by Bodhiruci,
involved almost all the major Buddhist monks in Chang'an and Luoyang
24 Kyunyo, S6k hwa6m kyobun w6nt'ong ch'o, HPC 4: 256c19-257a11. For a detailed
discussion of this critical turning point in Fazang's life, see Chen forthcoming: Chapter 2.
at the time
. The other project, the compilation of the Da Zhou kanding
zhongjing mulu :kfiil fIJ J;E};IU in which at least seventy major monk-
scholars were involved according to a list that was attached to the cata-
Fazang's name was - conspicuously - absent from the above
two lists, a fact which strongly suggests his absence in the two capitals
at that time given that his eminence as a Buddhist leader and his extra-
ordinary capacity as a Buddhist translator should have made him a very
likely candidate to be included in either of the two enterprises, (;m which
so much was staked by Empress Wu's government and the Buddhist church
at the time.
1.2) Fazang and Empress Wu: 695-705
Empress Wu's interest in the Avatarrzsaka siitra remained unabated
after her accession to the throne. It was under her auspices that a new Chi-
nese translation of the A vatarrzsaka sutra, which turned out to be the most
complete of its kind, was fInished. Under the supervision of
and joined by over twenty first-rate Chinese and non-Chinese Buddhist
scholars, the translation project was started on 1 May 695
The empress
attended the initiating ceremony, and personally acted (although no more
than symbolically) as a scribe (bishou for the translation, as is
described by Fazang
When the huge translation, in total of eighty fas-
cicles (thirty-eight parivarta [chapters]), was successfully brought to com-
pletion on 5 November 699
, Empress Wu honoured it with a preface.
A few weeks later, when Fazang was delivering a lecture on the new
version of the Avatarrzsaka sutra at the Foshoujisi as a celebration of this
sigriificant achievement, an earthquake occurred, allegedly as a response
25 The names of its thirty-two translators (both Buddhist monks and court officials,
Chinese and non-Chinese) appear in a Dunhuang manuscript, S. 2278; the full list is trans-
lated in Forte 1976: 171-176.
26 Da Zhou kanding zhongjing mulu, T 55: 15.475a-476a.
27 If Kyunyo's account about the rivalry between Fazang and FuIi is credible, we should
concede that Fazang was not actually involved in the project when it was started on 1 May
695, although he defInitely took part in it afterwards.
28 Huayanjing zhuanji, Tno51: 1.155a14-19; the quotation is from 155a16-17.
29 for this date, see Empress Wu's "Da Zhou xinyi Da fangguangfo Huayanjing xu,"
T 10: Ibll-12 (QTW 97.7a6-7).
to Fazang's lecture on a sentence "Huazang shijie-hai zhendong" i(t!t
%T-#IfflUb ("the Seas of the Avataf!1saka-realm started to shake"). The
strong tremors were felt around the area of the monastery. The report of
this episode greatly pleased Empress Wu, who issued an edict to praise
this auspicious sign and ordered it to be recorded in the historical texts

Empress Wu's enthusiasm for the Avatarrzsaka sutra caused a "boom"
of Avataf!1saka worship throughout the empire. The Da Fangguangfo
huayanjing ganying zhuan which was originally
compiled by one of Fazang's chief disciples, records two such AvataJ:!l-
saka-related miracle stories featuring the popularity of the sutra among
the lay and religious communities, and the empress's efforts to promote
people's enthusiasm for the sutra
!. It is also against the same historical
background that we must understand a series of legends and stories fea-
turing Empress Wu's admiration for Fazang's expertise on the Avatarrzsaka
sidra. Of these legends/stories, the following three are perhaps the most
1) The Ordination Episode: Sometime around 696 a white ray of light was
emitted from his mouth while he was delivering an A vataqlsaka lecture,
shooting into the sky where it turned into a canopy and remained there for
a long while. Hearing of this, Empress Wu immediately ordered ten of the
most prestigious preceptors in the capital to confer full ordination on Fazang,
who was then still a novice. She also bestowed on him the title of "Xian-
shou" tftr (Saintliness and Eminence) and then summoned him to the
palace chapel the Great Biankongsi to participate in the A vatamsaka trans-
lation office headed by .
2) The Golden-lion Lecture: Sometime between 26 November 701 - 1 Feb-
ruary 702, or as another source has it, sometime between 29 October-
26 November 699, when Fazang explained to Empress Wu the Avataqlsaka
teaching on interpenetration, the interdependence between all the dharmas
of any space and time; the teaching was so abstruse that it confounded a brain
even as brilliant as Empress Wu's. Recognizing this, Fazang resorted to a
golden-lion in the palace as a metaphor. He finally awakened Empress Wu
30 The earliest known source for this H uayan jing episode is a commentary on the Ava-
ta'!1saka siltra by Huiyuan (673 ?-743 ?), a chief disciple of Fazang. See Xu lileshu
kandingji, XZJ 5: 25b-c.
31 Da Fangguangfo huayanjing ganying zhuan, T 51: 177a, 1 77a-b.
32 The fIrst known source for a fully-fledged version of this episode is a Southern Song
dynasty, non-AvataIpsaka source. See Longxing fojiao biannian tonglun, XZJ 130: 280a2-6.
to the Avatarp.saka teaching. This was the alleged provenance for Fazang's
short but extremely popular essay called" Jin shizi zhang" (Essay
on the Golden Lion)33.
3) The Mirror-hall Device: For the same purpose of explaining to Empress
Wu the complicated tenet of universal interconnectedness, Fazang, on some
unspecified date, created for her a "hall of mirrors" in which all images
replicated themselves infInitely as in the legendary Indra's net woven with
numerous jewels
. .
As I have argued elsewhere, the Ordination Story was probably con-
cocted by later Huayan followers to dispel people's doubts concerning
Fazang's possible lack of full ordination (juzujie J:l.JEt!lt) as a fully qual-
ified Buddhist monk (Chen forthcoming: Chapter 3). The "Jin shizi
zhang" was, on the other hand, actually written much earlier and might
have had nothing to do with the empress. As for the mirror-hall, we
certainly cannot exclude the possibility that Fazang did construct such
a device for some pedagogical purposes, but he might have done so for
his disciples, rather than for the empress, and what is more interesting is
that he here seems to have only reproduced a scheme that had been envi-
sioned by one of his senior contemporaries (the learned monk scholar
Daoxuan) several decades earlier

Although it is naive to accept all these legends/stories uncritically,
it is not too far from the truth to assume that Express Wu's esteem for
Fazang was largely derived from her respect for his superior expertise in
the A vatrupsaka teaching. That said, we should also realize that Fazang
served the empress and her government not merely through his advanced
philosophical and philological skills, but also by his capacity as a per-
forn;ter of esoteric rituals aimed at some worldly benefits (like bringing
down rain, snow and so on) or simply as a magician. During Empress
Wu's regency and reign, some local officials around the Chang'an area,
33 The earliest source promoting this idea is Zongmi's *It: (780-841) Huayan jing
xingyuan pin shuchao, XV 7: 487a7-8.
34 While Zanning just ambiguously observes that it was for those who failed to under-
stand his teachings, rather than specifically Empress Wu, that this ingenious device was
designed (Song gaoseng zhuan, T 50: 5.732a28-b2), the Longxing fojiao biannian tonglun
(XV 130: 28Ic13-16) is the first known source which attempted to correlate this story with
Empress Wu.
35 Shimen guijing yi, T 45: 2.865b4ff, discussed in Chen 2003: 335-336.
who suffered from the ravages of a drought, had already repeatedly enrolled
the kind of supernatural power that Fazang allegedly possessed. They were
one of Empress Wu's first cousins once removed, and a local official two
of whose nephews were to become her favorites in the last decade of her
life (see [ill. I]).
What might appear more startling to modem scholars who are accus-
tomed to Fazang's reputation as a sophisticated religious theoretician is
the fact that Fazang was also believed to have wrought some magic in
the battles that the Chinese army fought - some time from 16 June 696
to 23 June 697 - against the rebellious Khitan and thus played a crucial
role in overcoming a military and political crisis that was then severely
threatening the national security of the Great Zhou dynasty. We will dis-
cuss this unexpected exploit of Fazang in the next part.
This feat of Fazang must have earned more respect from Empress Wu,
although we have no more documentation on their relationship in the suc-
ceeding several years except for the following example of their cooper-
ation. In the summer of 700, Fuli, and other monks who might
or might not include Fazang, were in the empress's company in one of
her summer palaces at the Songshan area when they were preparing for
a new translation of the LaiLkiivatiira siUra
The translation project was
continued at the Qingchansi in Chang'an after followed
the empress to there on 26 November 701. was only able to
fmish a draft of the translation before going back to Khotan. The draft was
then polished by the Tokharian monk Mituoshan smWt; w (a.k.a. Mituoxian
(Mitrasena or Mitrasanta?, ?-704t), who arrived in China prob-
ably in 702, with the assistance of Fuli, Fazang and other monks
Its completion was officially announced on 24 February 704 (Chang'an
4.zheng.15?8. The empress was then starting to show increasing interest
in Chan Buddhism, which at least partly accounted for her determination
36 In her preface to the new Chinese translation of the Lankavatara sutra, Empress Wu
mentions Silqanada and Fuli, but not Fazang. See "Xinyi Dasheng ru Lengqie jing xu,"
QTW 97.lOa8-9.
37 The history of this important translation is surveyed in Fazang's Ru Lengqiexin
xuanyi, T 39: 430b16-23.
38 This date is provided by Empress Wu; see "Xinyi Dasheng ru Lengqie jing xu," QTW
to sponsor a new translation of the sutra since it was recegnized as the
primary theoretic basis of that tradition. In spite of his commitment to the
A vatarp.saka tradition; Fazang still decided to cooperate with the empress
in fostering this type of Buddhism separate from his tradition. His effort
in this respect is fully shown in the commentary that he wrote on the new
Lalikiivatiira translation, the Ru Lengqiexin xuanyi (Ishii 2002).
Starting from the very beginning of the eighth century, taking advan-
tage of Empress Wu's age and poor health, those court officials loyal to
the Li royal house conspired to re-enthrone one of the disposed Tang
emperors. They found an easy target: the empress's two favorites, Zhang
Yizhi ~ ~ Z (676?-705) and Zhang Changzong ~ ~ * (676?-705)39.
It was in this delicate political environment that Empress Wu launched a
major politico-religious campaign which she entrusted Fazang to steer.
It so happened that this campaign developed into a watershed not only in
the life of the empress but also in that of the monk.
At the start of the year 705, at the instigation of Fazang, Empress Wu
decided to bring the Famensi relic to her palace in Luoyang. In view of
Empress Wu's rapidly deteriorating health at the time, the Famensi relic
was then also consulted for its putative therapeutic power, not unlike the
situation forty-five years earlier when she and her husband had turned to
the same "sacred bone" for the personal welfare of the emperor. How-
ever, in view of the political situation at the time, one might assume that
Empress Wu also sponsored this relic veneration with an eye to rallying
her declining political support.
Contrary to her expectation, this grand religious ceremony did not
perpetuate her fortune. Only one week later, a court coup broke out, result-
ingm the killing of the two Zhang brothers, and Empress Wu's abdica-
tion of the throne to her son Zhongzong, who was then ranked as "Heir
Apparent." Empress Wu was subsequently transferred to the Shangyang
palace . L ~ g , where she died less subsequently ten months later, on 16
December 705
There is evidence that Fazang was actually an "accomplice" of
some Pro-Tang royalists, with whom he worked to facilitate the end of
39 Zizhi tongjian 207.6563-67.
40 Jiu Tang shu 6.132.
Empress Wu's rule (Chen 2003: 341-352). Fazang seems to have leaked
to Zhongzong some damaging privileged information about the Zhang
brothers, who, according to the two Tang dynastic histories, were then
intensively plotting with their group in order to pre-empt any possible
offensive on the part of their rivals after the death of Empress Wu, whose
health was then rapidly worsening
Fazang's information was deemed
crucial in helping him and his supporters to suppress the Zhang broth-
ers. His access to secrets about the Zhang brothers was probably made
possible by his special status as a court priest at that time. We know from
Ch'oe Ch'iwon's ~ 3 & m (857-9Q4t) report that Fazang was then a chief
directorof the relic veneration in the court, especially the enshrinement
ceremony in the Luminous Hall complex. A 708 inscription confirms
Fazang's role as the superintendent of the Famensi relic while it was
stored in the imperial palace (Wu and Han 1998: 70; Barrett 2001: 16).
We can imagine that after he brought the relic to Luoyang on 9 Febru-
ary 705, he must have stayed close to Empress Wu (and therefore close
to her favorites, the two Zhang brothers) in the course of orchestrating
this important ceremony. This provided him some opportunities to keep
abreast of what the two Zhangs and their clique were then planning.
He thus cunningly turned his close relationship with his patroness into
a valuable political asset that he used to ingratiate himself with Zhong-
zong, who was then waiting beside his mother's sickbed for the chance
to rule again. This reveals Fazang as a politically sophisticated and shrewd
monk, who was ready to abandon his most important secular supporter
when he sensed that the political situation had started to spin out of her
control, making his continued association with her increasingly to his
own disadvantage (or as he might have thought of it, to the disadvan-
tage of his religion). Fazang thus ended up being a "betrayer," rather
than a supporter and sympathizer, of Empress Wu. This switch of loyalty
also partly explains the glory and success that he continued to enjoy
under the reigns of the three successors of Empress Wu, Zhongzong,
Ruizong and Xuanzong (r. 712-756), an issue we are going to discuss in
the next section.
41 Jiu Tang shu 6.132; 7.135; Xin Tang shu 4.103, 102.4015.
1.3) Fazang under the Reigns oiZhongzong and Ruizong (705-712)
Several months after his re.-enthronement on 24 February 705, Zhong-
zong ordered Fazang to be rewarded for his role in the'court coup remov-
ing the two Zhang brothers. Fazang was awarded a fifth-rank title, which
he resolutely declined. Since the govemment insisted in rewarding Fazang,
he proposed a compromise that this award be transferred to his younger
brother, Kang Baozang .1l. (?-706
), who was then serving as a Gen-
tleman for Court Discussion (chaoyilang and the Vice Director
(jujian in the Tongwan City42. The government approved the
proposal and in the following year Zhongzong issued an order to the
effect that Kang Fabao be appointed as Mobile Corps Commander (youji
jiangjun hfim'.!l[) and the Left Commandant of the Courageous Garri-
son (zuo guoyiffu] duowei [belonging to] the Awesome
Guard (Weiwei based in the Commandery (fu Iff) of Longping
The emperor further specified that in order for Baozang to take
care of his mother at home, he should not be given any actual responsi-
This must have been the same occasion recorded in other sources,
in which Fazang, along with other eight monks, were awarded a fifth-rank
42 Tongwan probably referred to the city of Tongwan, the capital of Helian .Bobo
(a.k.a. Helian Qugai r. 407-425), who ordered it to be built in 413 and
had it completed five years later. The city was named in this way allegedly because Helian
wanted it to embody his ambition of "unifying [the land] under heavens and looking down
on the ten thousand states like a sovereign" (Tongyi tianxia, junlin wanbang
See lin shu 130.3205, Zizhi tongjian 116.3658.
43 Hucker 1985: 565: "2 prefIXed left and right, included among the sixteen Guards
wei) at the dynastic capital, generally responsible for defense of the eastern sec-
tor 6f'the capital city; created in 622 to replace the Left and Right Encampment Guards
(t'un-wei) inherited with the Sui dynasty's Twelve Guards (Shih-erh wei) organization; in
684 renamed Guards of the Leopard Strategy (pao-t'ao wei); in 705 briefly named Awe-
some Guards; from late 705 to 711 again called Encampment Guards; from 711 once agin
called Awesome Guards." Thus, although Ch'oe dates this edict to Shenlong 2 (706), the
appearance of the title Weiwei WGl$J therein reveals that it had actually been drafted by late
705 - when the title was still in use, although it was probably announced in early 706 -
so shortly after the official reversion of the title back to Dunwei (Tun-wei) Jt!l$J (Encamp-
ment Guards) that there was no time to make the necessary correction in the edict to reflect
this change.
44 Popjang chOn, T 50: 283b18-cl. Although according to Ch'oe, Fazang received a
third-rank title at the time, I have argued elsewhere (Chen in preparation: Chapter 1) that
the title was actually fifth-ranked.
title, although some of these sources tell us that he and his colleagues were
rewarded for their merits in reconstructing a major monastery dedicated
to the posthumous benefits of Empress Wu Shengshansi
The high degree of esteem that the emperor held for Fazang also
clearly shown by four verses that he dedicated to Fazang's portrait.
They are still preserved in the Popjang chon, another Korean source and
the Quan Tang wen as well
j;5] With the luminous causes planted from the past [lives],
_>J<:iEJ; [He] has single-mindedly searched for the right and true.
'lii;Il!lfij1fiJ} Although the [Buddha's] traces turned into obscurity in the "park
of amra,"
his body appeared in the Realm of Lotus.
Expounding the teaching of the Sakya,
saving and delivering people stuck in the swamp of illusion.
Always pouring out the rain of oneness,
'tJ"I*/\. For the constant purification of the "six dusts."
When the garden of eloquence opens,
" the spring of words gushes out widely.
Protect and maintain the dharma in the sprit of enduring humili-
diligently cultivating the way of vigor.
m4,*x. His lectures caused the gathering of heavenly flowers,
45 See, for example, Da Song sengshi liie, T 54: 3.250b3-11. Cf. Jiu Tang shu 7.141
and Zizhi tongjian 208.6598, which tells us that three more Daoist priests including Shi
Chongxuan ..t*tf (?-713) and Ye Jingneng (?-710) were also among those who
were rewarded for their merits in building this monastery of exceptional importance for
Zhongzong. All these relevant sources and their implications are discussed in Chen in
preparation: Chapter l.
46 POpjang chOn, T 50: 284aI8-29; Uichon (1055-1101), Wonjong mullyu, HPC 4:
22.631b-c; QTW 17.21b-22a. Ku Cheng-mei (pinyin: Gu Zhengmei) understands
the phrase as "Zhongzong ordered Fazang to draw a portrait of Zangzhen
that is, Qiujiuque frff,lti)p (Kujiila Kadphises, 5 BC - 78 AD)," who was the found-
ing emperor of the Kushan dynasty, whom some scholars - including Professor Ku - believe
to be the prototype for the Buddhist king Asoka. On the basis of this understanding, Ku
has read the following verses written by Zhong zong as dedicated to Qiujiuque, rather than
to Fazang. See Ku 1996: 175-76. This reading seems questionable given that the character
zang l in the phrase obviously refers to Fazang and therefore that zangzhen cannot be
understood as a separate term. In other words, I read as xie Zang zhenyi ("to draw
[xie the portrait [zhenyi Jtfil of Fazang"), rather thanxie zangzhen yi ("to draw a por-
trait of Zangzhen") as is suggested by Ku. Consequently, the four verses should be regarded
as dedicated to Fazang, rather than to Qiujiuque.
r.&1B'iE.11t An unusual sign emerged in response to [the sentence of] earth-

Exerting this dharma-power,
he got the evil camps removed.
The ten contemplations are raised,
to accord with the "Four Dhyanas."
Universally cutting off the afflictions,
ridding himself of the secular ties from afar.
With the source of mind mirrored and penetrated,
the dharma-mirror brilliantly suspended .
The boat of wisdom steered perfectly,
the lamp of compassion to be transmitted forever.
;g f!!'i I1t His names echoing in the imperial palace,
J.t.iiLtttJiJt\ his reputation circulating among the monastic world.
The guiding principle for the Brahmanic Congregation (i.e. sarp.-
the standard and example for Buddhist followers.
[9 ':t In protecting those born in four ways48,
'l!!' he never feels fatigue.
Spreading the beautiful [name] in the three thousand worlds,
ilJ5'fE'f1! transmitting the fragrant [reputation] to the ten billions of gene-
According to Ch'oe Ch'iwon, these verses were written in the winter
of the first year of the Shenlong reign-era (30 January 705-18 January 706).
However, it seems more likely that Zhongzong wrote them some time
In my article on the palace chapels under the Tang dynasty, I have col-
lectyd some materials on this important Buddhist institution within the
Tarig imperial palace in Chang'an
Some time between 7 December 706
and 23 March 709, Zhongzong summoned to the Linguang chapel twenty
47 Chengguan, Da Fangguangfo huayanjing suishu yanyi chao, T 36: 3.17a21-23:
Jil!ii1'1iWo 'ilti"::iJ)iJo M1!lJW!i!l. This suggests that
Chengguan had access to Zhongzong's verses dedicated to Fazang.
48 Four fonns of birth (sisheng 11Y:i:): taisheng Bl:i: (jarayu-ja) birth from the womb
(humans, animals); luansheng (alJrja-ja) birth from the egg (birds), shisheng ;!:i:
(sarrzsveda-ja) birth from moisture (insects), and huasheng 1-t:i: (upapiidu-ja) birth by trans-
fonnation (dwellers in the heavens and hells).
49 Chen Jinhua 2004: 124-128.
or so eminent Buddhist monks from all over the country. Seven of these
palace chaplains are still identifiable. They are (1) Hongjing (var.
Hengjing ,t1iij1;, 643-712), (2) Xuanzang (d. after 709) (not to be
confused with the homonymous great Buddhist translator and pilgrim),
(3) Daojun (?-709
), (4) Daoan i!!J$ (654-717), (5) Wengang
(636-727), (6) Sengqie f'!!l'11Jn (Sarp.gha?, 628-710), and (7) Siheng .11'H1ii
(653-726)50. These monks were requested to perform Buddhist rituals for
the welfare of the state, while ten of them served on the ten-member com-
mittee known as shidade +::kAt (Ten Buddhist monks of "Great Virtue"),
which was in charge of national monastic affairs. While some of these
monks left the palace chapel shortly after 23 March 709, when a parting
banquet was held for their behalf (during this banquet Zhongzong wrote
verses for the departing monks and his verses were responded to by the
academicians who participated in the banquet), other monks remained
there. One year later, in Jinglong 4 (4 February-4 July 710), Zhongzong
invited Bodhiruci (a.k.a. Dharmaruci, 572?-727) and his colleagues for a
vegetarian banquet at the Linguang Palace, where the emperor watched
the monks discussing Buddhist teachings. He then ordered the painter
Zhang Shun (otherwise unknown) to draw on the wall of the palace
the portraits of all the bhadanta-translators and the academicians who
participated in the translation. On these portraits, Zhongzong himself
wrote eulogies in verse
From the above discussion, we get the impression that during the Shen-
long and Jinglong eras, Zhongzong invited some Buddhist monks to his
palace chapel Linguang[siJ for at least two banquets, during both of which
he penned laudatory verses for these monks. It seems quite likely that
the verses that Zhongzong wrote for Fazang might have been related to
one (or several) of these similar occasions. In other words, we have rea-
son to believe that Fazang might have been among the twenty or so monks
50 Song gaoseng zhuan, T50: 5.732b20-27, 24.863c19-20, 8.758a5ff, 14.793b14,
14.792a22-23, 18.822a19-23; for Siheng, see "Da Tang gu dade Siheng liishi muzhiwen,"
in Zhou 1992: 1321-22.
51 Fozu tongji, T 49: 38.372c21ff. Zhipan dates this event to Shenlong 4, which was
apparently an error for Jinglong 4 (4 February-4 July 71 0) given that the Shenlong reign-
era only lasted from 30 January 705 (Shenlong 1.1.1 [renwu]) to 4 October 707 (Shenlong
3.9.4 [yihai]).
who were invited to reside at the Linguang palace and that like
Siheng, who was his acquaintance if not friend
, Fazang might have been
a member of the shidade committee considering his eminence at the time.
Fazang's crucial role in Ii series of events that centered on the contin-
uous veneration of the Famensi relic, to which Empress Wu turned in the
last phase of her life, also reveals Zhongzong's extraordinary trust of
and reliance on him. In the spring of708, Zhongzong entrusted him with
the task of escorting the relic, which was brought to the imperial palace
at the end of 704 at Empress Wu's request, back to its home temple
(see [ID.3]).
Fazang's reputation as a great Buddhist expounder and translator, and
especially his important role in the 705 court coup were certainly chief
factors contributing to the preeminent position that he had managed to
achieve (or maintain) in this period. However, evidence shows that, not
unlike his relationship with Empress Wu and his status under her regency
and rule, Fazang's continuing success as a Buddhist leader also depended
to a large extent on the service that he rendered to the Tang rulers through
his mastery of some esoteric (or even shamanic) skills, which made him
a top candidate whenever the capital area was threatened by some natu-
ral calamities like drought. As we have an opportunity to talk in detail
about the stories and legends on this type of supernatural ability attribu-
ted to Fazang, suffice it here to a brief mention of these feats recorded in
the sources.
In the mid-summer (i.e. the fifth month) of Jinglong 2 (24 May-22
June 708), Fazang successfully performed a rain-praying ritual at Jianfu
si which was the monastery that Zhongzong dedicated to the
welfare of his father Gaozong. The next year, when the drought
recurred, Fazang rose to alleviate the ravages of the drought as he did
before and was once again praised by Zhongzong. Ch'oe Ch'iwon continues
by saying that from then on, Zhongzong and Ruizhong relied on Fazang
as their bodhisattva-preceptor. This might refer to the possibility that
some. time during the Shenlong or Jinglong era Fazang was invited to the
Linguang Palace Chapel, as was suggested above.
S2 Siheng's relationship with Fazang is discussed in Chen forthcoming: Chapter 3.
In the winter of Jingyun 2 (24 January-22 April 711), one year before
his own death, Fazang performed an esoteric ritual at a temple on Mount
Zhongnan and aliegedly brought down some snows, thus significantly
alleviating the drought that was threatening the capital area. Fazang was
highly praised by Ruizong because of this.
The high esteem that Ruizong maintained towards Fazang can be seen
by the fact that on Fazang's sixty-ninth birthday (4 December 712 - Xian-
tian 1.11.2 [dingmao D, which turned out to be his last as he died a mere
twelve days later, Ruizong, who had by then abdicated in favor of his son
Xuanzong but who still maintained a part of supreme power in the capacity
of Taishanghuang :*-.t (Emperor Emeritus), sent him some gifts (a set
of monastic robes and some noodles of longevity [changming suobing
along with a congratulatory letter quite respectfully
Ruizong's letter amply expresses his respect for and fond-
ness of Fazang. Far more than a perfunctionary greeting from a secular
monarch toward a prestigious religious leader, the letter conveyed a taste
of the very genuine and personal sense of friendship that was usually
only cherished between two close friends
Ch'oe Ch'iwon continues by
telling us that in order to show his appreciation of Fazang's unflagging
effort to serve the Tang royal family and his constant respect for Fazang
as a teacher, Ruizong presented him two thousand bolts of silk to cover
the expenses caused by the religious rituals that Fazang was to carry out
for people's benefits.
Most notably, according to D6chil, it is by following Fazang's advice
that Ruziong decided to relinquish the throne to Xuanzong
nately, D6chil did not tell us the source for this claim, which, if true, would
testify Fazang's crucial role in the power-transition at the highest level in
712 that ushered in one of the most prosperous eras in imperial China

The amiable personal relationship between Fazang and Ruizong is also
reflected in the good terms that he maintained with a couople of family
53 Popjang chon, T 50: 284c2-7.
54 Popjang chon, T 50: 284c2-7.
55 "Shinkan Genju hiden shOgo," T 50: 288c1: 'i'tl\:*fI\\z.i!Jl;lJtE.o
56 It is interesting to note that Fazang died on 16 December 712 (Xiantian 1.11.14), only
four months after Ruizong officially handed over supreme power to Xuanzong on 8 August
712 (Yanhe 1.8.3 Uiazi]).
members of the emperor, including one of daughters and one of his
sons-in-law who married another of his daughters. It is at the request of
Zheng Wanjun (7-734t), who married Ruizong's fourth daughter
Li Hua * (style-name Huawan (687-734)57, that Fazangwrote
a commentary on the Heart Satra in Chang'an 2 (February 2, 702 - Jan-
uary 21, 703) at the Qingchansi, while he engaged in preparing some
Buddhist translations. To this commentary, Zhang Yue (667-730),
a prestigious statesman and author, wrote a preface, "Bore xinjing zanxu"
(A Preface to the Comments on the Bore xinjing)'58.
Fazang also befriended another daughter of Ruizong, Princess Jinxian
(689-732), who, along with her blood sister Yuzhen.:EJ.: (6927-
7627), was famous for her devotion to Daoism. Her friendship with
Fazang was made likely not only because of the monk's good relation-
ship with both her father Ruizong and her brother-in-law Zheng Wanjun,
but also by the fact that her Daoist teacher Shi Chongxuan (7-713)
was obviously a friend of Fazang
Shi Chongxuan's friendship with
Fazang can be deduced from their shared efforts in building the impor-
tant monastery Shengshansi, as was noted in the preceding section (I.3).
It is quite unusual that four Daoist priests should have become involved
in such a project. I speculate that their function might have mainly con-
sisted in raising funds, not unlike the role Shi Chongxuan played in
the course of constructing the two convents for Jinxian and her sister.
No matter what Shi Chongxuan's real role was in the Shengshansi proj-
ect, his friendship with Fazang seems of little doubt. As I have suggested
elsewhere, Princess Jinxian was probably such a close friend of Fazang
that she, though already an ordained Daoist priest at the time, was will-
ingto honor Fazang's fond memory ofYunjusi by requesting in 730 (two
years before her death and eighteen years after Fazang's) her brother-
emperor to send to the temple a copy of the Kaiyuan canon, which must
57 For Zhang Wanjun's marriage with Li Hua, see Xin Tang shu 83.3656.
58 This preface is now preserved in QTW 225. lObI la, and attached to the Taish6 edi-
tion of the Bore boluomiduo xinjing lueshu (T 33: 555a24-b9).
59 For Shi Chongxuan's status as a teacher of Jinxian and her sister, see Chaoye qian-
zai 5.114. Jinxian and her sister's ordination ceremony was superintended by Shi Chongxuan.
This important ceremony is the subject of Charles D. Benn's excellent monograph (Benn:
1991). Shi Chongxuan was believed to have raised a huge amount of money for building
two Daoist convents for her two royal disciples. See Xin Tang shu 83.3656-3657.
have constituted a precious gift that the marginal local temple would have
been very hard to secure but for the forceful intervention from a figure
with Jinxian's infiuences

Fazang died on 16 December 712 (Xiantian 1.11.14) at the Great
Jianfusi. Five days later, Ruizong issued an edict to praise his outstanding
performance as a Buddhist leader and his valuable service to the state as
According to Ch'oe Ch'iwon, it was the Tang government policy
that on the death of an official, no matter whether military or civil, the
government would make a donation in proportion to his rank - from a
donation worth two hundred duan ftffij62 and two hundred shuo mi (i.e. shi
;:P-) of millet for a first-rank office, down to only ten duan of silk for a
ninth-ranked office. Ch'oe Ch'iwon observes that the value of the dona-
tion the government made on Fazang's death revealed the extent of the
respect that the imperial house held for him. The government also offered
to pay the labor needed to build Fazang's tomb. The additional donations
made by princes dukes and commoners were innumerable. His funeral was
conducted with the ceremony reserved for a third-rank official

ll) The Magician as a Warrior?: Fazang and the Suppression of the
Khitan Rebellion (696-697)
We have already highlighted above several essential elements that
contributed to Fazang's success as a court priest: his accomplishment as
a Buddhist philosopher, his political skills, his reputation as a miracle
worker, and - rather unusually for a Buddhist priest - his battle skills
in a series of campaigns that the Great Zhou army launched against the
Khitan rebels. This last feat was made particularly noteworthy by the fact
that it was allegedly achieved by virtue of his prowess with black magic.
Due to its significance for revealing a hidden aspect of his intellectual
life, Fazang's role in the suppression of the Khitan rebellion warrants an
60 Chen forthcoming (a).
61 Popjang chOn, T 50: 285b.
62 Duan !;jIij was a unit of measurement for cloth (bu ;(p), while pi Il1: (bolt) a unit for
silk (juan #,/l). Under the Tang, four and six zhangs equaled one pi and duan respectively.
63 Popjang chOn, T 50: 285b18-21.
in-depth investigation. In this part, I will therefore examine this issue
from three perspectives. After a survey of how his biography by Ch'oe
Ch'iwon (by far the most important biographical source on Fazang) pres-
ents this side of Fazang's life, I will contextualize this important aCGount
against the larger political, military and religious background - the 696-
697 Khitan rebellion and its suppression described in secular sources on
the one hand and on the other, the Avalokitesvara cult developed under
the rule of Empress Wu.
ILl) Description in Fazang's Biography
In the first year of the Shengong reign-era (29 September-19 Decem-
ber 697), the tribe known as Khitan (Ch. Qidan ~ f J - ) , then a vassal state
based in the northeastern part of the empire, refused to pledge loyalty
any longer. Empress Wu dispatched an army to suppress the "rebellious"
tribe. At the same time, the empress sought advice from Fazang, con-
sulting him on the possibility of employing the assistance of the Buddha
to help the imperial army defeat the Khitans. Fazang told the empress,
"In order to destroy and subdue the ferocious enemies, please allow me
to resort to the 'left-hand (that is, Buddhistically unorthodox) path' (zuo-
dao ,ft:)!!)." Imperial permission was swiftly granted. Fazang took a bath
and changed his dress before building a bodhimal;uJa (daochang i!!.l1iJ; i.e.
"ritual-precinct") of the Eleven-faced Avalokitesvara, in which he placed
images of that bodhisattva and started to carry out the observance. The
effect of this esoteric procedure was rapid and astonishing.' Within several
days, the barbarians saw to their panic that they were faced not only by
countless warriors of the Great Zhou army, but also that the troops were
backed by a congregation of deities. Some of the enemies saw images of
Avalokitesvara floating in the sky and then slowly descending to the bat-
tlefield. In addition, flocks of goats and packs of dogs started to harass
the Khitan soldiers. Within a month, Empress Wu received the news of
victory. In her great joy, she rewarded the monk's merits with a nicely-
worded decree, which says,
Outside the city of Kuai, the warriors heard the sound of heavenly drums;
within the district of Uangxiang, the enemy crowd saw images of Avaloki-
tesvara. Pure wine spread its sweetness in the battalions, while the chariots
of the Transcendent led the flags in front of the army. This [victory] was
accomplished-by the divine army sweeping away [the enemy], and that must
have been aided by the [Buddha's] compassionate power!

Ch'oe Ch'iwon plainly states that Fazang perfonned these rituals one
month before the victory over the Khitans was declared, which happened
on July 27, 697 according to the secular sources
This implies that
Fazang was invited to resolve the military conflict in June 697. Contrast
to the clear way Ch'oe Ch'iwon provides a timeframe for this event, his
locating of these rituals is problematic and requires further clarification.
The two locations in which the miraculous effects of Fazang's rituals
were allegedly carried out, Kuaicheng and Liangxiang, were in present-
day Baoji JUi in Shaanxi and Fangshan mill in Beijing respectively.
Given that the uprising Khitan army had never been able to reach its
spearhead to the Kuaicheng area bur rather that it had throughout engaged
in close combat with the Great Zhou army in some areas of Hebei Cir-
cuit including Tanzhou 1113+1, Pingzhou f-1+1, Dingzhou 51+1,
Yizhou Zhaozhou Ml.HI, and particularly Youzhou which had
decisive importance for the defense system of the Sui and Tang empires
I suspect that in the current edition of the Popjang chon the character ji
{ij, which indicated a place - in present-day Daxing Beijing -
very close to Liangxiang, was miswritten as kuai ,iliitl due to their similar-
ity in fonn. I am therefore inclined to believe that Ch'oe Chiwon believed
that the miraculous effect of Fazang's rituals took place in two battle-
fields close to each other, both falling in present-day Beijing.
Further, it is important to note that Liangxiang happened to be in the
proximity of Fangshan, where is located the Yunjusi, the monastery which
has over the past several decades earned a world-wise reputation for
the immense repository of Buddhist scriptures carved on the stone slabs
(the so-called "Fangshan shijing" m that it has enshrined. It is
64 This episode is recorded in the P6pjang chOn, T 50: 283cl6-25. The passage quoted
here is located at 283c22-25.
65 Li 2003: 100-10 1.
66 See below for a survey of the military conflicts between the Khitan and Zhou armies
in this one-year period.
possible that Fazang might have carried out the ritual at the Yunjusi or
at a neighboring location.
The secular sources make no mention whatsoever of Fazang's role
in this year-long military endeavor. Here we must note that a fust Gousin
of Empress Wu, Wu Youyi, who played a significant role in suppressing
the Khitan, was a friend of Fazang, who had just (a mere one year before)
helped the prince to end a drought afflicting the area under his jurisdic-
tion, by praying for rain (see [III.2]). This relationship suggests that
Fazang's role in the suppression of the Khitan army was not rinlikely.
However, it is Empress Wu's edict quoted above and a poem that Zhong-
zong wrote for Fazang, which confirms and commends Fazang's role in
"destroying these devils' camps" (dan zi mozhen ~ ~ f l ~ ) (very likely
:referring to the Khitan rebels)67 that force us to consider this role of Fazang
more seriously. It seems undeniable that Fazang did contribute to the
overcoming of this severe socio-political crisis, or at least was perceived
to have done so.
The lack of historical evidence has left us no alternative but to specu-
late that Fazang or some of his followers might have performed some
forms of black magic (the so-called "Left Path" he was reported to have
recommended to Empress Wu) on the battlefield, bringing up the illu-
sion of some images of Avalokitesvara floating in the sky, which scared
away some Khitan soldiers
. Although the effect of this feat might not
have been as decisive and far-reaching as it was depicted in the Buddhist
sources, Fazang's intervention in this crisis and Empress Wu's appreci-
ation of it seem beyond doubt. It is not hard to imagine that both Fazang
(and his group) and the empress were more than happy to play up the
effett of this feat, although they may have done so with different purposes
in mind: for Fazang and his group, they must have interpreted this episode
as a telling demonstration of the divine power of both the Bodhisattva and
67 POpjang chOn, T 50: 284a23-24; QTW 17.22a2.
68 Eugene Wang (2005: 259) suggests that in helping the Zhou army battle the Khitan
rebels, Fazang brought up sO,me frightening reflections with a device composed of eleven
faces of mirror. Although this interpretation is not supported by Ch'oe Ch'iwon's biography,
in which shiyimian guanyin -t-OOIl.{f just means Eleven-faced Avalokitesvara, Wang
does raise a possible stratagem that Fazang might have employed to defeat the Khitan aimy,
especially this seems to have been related to his skill in magic.
Fazang, while Empress Wu and her ideologues must have treasured it as
a potent sign from the heavens that justified and protected her rule.
Ch'oe Ch'iwon here has given the reader the impression that with the
help of Fazang Empress Wu and her government smashed the Khitan
rebels without the slightest effort. TIris impression is by no means supported
by the secular historical sources, which depict the two major campaigns
that the Great Zhou launched against the Khitans as two of the bloodiest
in the history of the Great Tang and ZhOU

ll.2) Historical Background
The historical sources date the outbreak of this rebellion to 16 June 696
and identify the two rebel leaders as Li Jinzhong *",$ (?-696), the
Commander-in-chief (dudu of Songmo and Sun Wanrong
(?-697), the governor of the Guicheng Prefecture (in present-
day Hebei) and whose younger sister was married to Li Jinzhong. They
also inform us that this rebellion was triggered by the haughty and humi-
liating attitude that the Commander-in-chief of Yingzhou it1+1, Zhao
Wenhui (?-696), showed to the Khitan chieftains and the cal-
lousness that he displayed towards the Khitans during a famine. The angry
Khitans killed Zhao Wenhui and occupied Yingzhou. Judging by the fact
that thousands of them joined the rebellious army within a mere ten days,
the Khitans' animosity towards their Chinese rulers must have run rather
deep and wide. Apparently shocked by this largely unexpected rebellion,
Empress Wu dispatched an army to suppress it thirteen days after the
Khitans rose. Although its strength is not specified, the imperial army
must have had an impressive size given that it was under the joint leader-
ship of twenty-eight generals, including the prestigious Cao Renshi
lf1=SiP (?-696
) (General of the Left Soaring Hawk Guard [Yingyang
wei Nlm,i;]), Zhang Xuanyu (?-696?) (the Great General of
the Right Imperial Insignia Guard [Jinwuwei Li Duozuo *$11=
69 The following account of the Khitan rebellion and the quelling thereof is mainly
based on Zizhi (ongjian 205.6505-6523. See also Guisso 1978: 138-143, 1979: 314-316,
Li 2003. For a survey of the history of warfare under the Tang and Great Zhou dynasties,
see Graff 2002 (esp. pp. 8-11).
(?-707) (Great General of the Left Awesome Guard [Weiwei ,WI;wr]), Ma
Renjie )$f(1=llP (Vice Chamberlain for the National Treasury [Sinong Shao-
qing ijJl&9Ilm. Still uncertain about the strength of this army, Empress
Wu, on 15 August of the same year, appointed her first cousin. once
removed, the Prince of Liang Wu Sansi ftt'=:'Ji!1, (?-707), who was then
her Minister of Rites (Chunguan shangshu as the Pacification
Corfunander-in-chief (Anwu dashi of the Yuguan Circuit
(Dao i!!), with Yao Shu (? -705) as his associate.
Li Jinzhong and Sun Wanrong turned out to be two exceptionally
shrewd warriors. The historical sources portray the military success they
achieved in the early phase of the rebellion in this way: "wherever their
spearheads pointed, those places fell into their hands." Their military tal-
ents were amply displayed in the first major battle they fought against the
imperial army on 29 October 696 at the Valley of Xiashi (probably
in present-day Mengjin :;fuW, Henan). With some brilliant tactics they
easily defeated the Zhou army, almost entirely wiping them off the sur-
face of the earth. They were able to expand their victory by luring the
Zhou relief force into an ambush with forged orders, which they coun-
terfeited with the seals that they captured from the Zhou army.
The extent to which this traumatic defeat was felt in the Zhou court is
dramatically shown by the unprecedented offer the empress made in the
ninth month to reward any criminals and private slaves willing to serve
in the army. For the first time, the prefectures to the east of the Taihang
7.\.iT Ranges (the so-called Shandong LlJ* areas) set up cavalry units
(wuqi bingtuan who were expected to fight the horsemen of
the Khitans. She appointed another of her nephews, Prince Jian'an
Wu Youyi ftti!lt1[ (d. between 705 and 710), the Tongzhou 1'iJ1+1 Gover-
nor, as the Grand General of Right Militant and Awesome Guard (Wuwei-
wei ftt,Wl;wr) , the Adjunct (Xingjun iT!!!) Commander-in-chief of the
Qingbian Circuit, obviously in preparation for another major attack
on the Khitan.
Fortunately for the empress, at this crucial moment the new qaghan of
the Northern Turks, Qapaghan (Mochuo offered to help, on the
condition that he be accepted as her son and that an imperial marriage be
arranged for his daughter, both of which were apparently envisioned out
of his ambition for the Chinese throne. Although not at all blind to the
hidd,en agenda of the Turks, Empress Wu still welcomed this offer and
rewarded Mochuo with titles of distinction. The Turkish support was
compounded by an unexpected turn of events which was very favorable
to Empress Wu: Li Jinzhong died on 22 November 696, of an unspeci-
fied cause. Although Sun Wanrong rapidly took over the Khitan leader-
ship, Mochuo wasted no time in taking advantage of the chaos inevitably
created among the Khitans by this power transition. He raided the Khi-
tan base in Songmo, capturing Li Jinzhong and Sun Wanrong's wives
and sons. However, it did not take Sun Wanrong long to recover from this
setback, rapidly managing as he did to rally the scattered Khitan soldiers.
Using two of his valiant subordinates Luowuzheng and He Axiao
fiiJJliiJ/J\ as vanguards, he seized Jizhou A7+1 (present-day Jixian
Hebei) and massacred the inhabitants of the city, killing several thousand
officials and commoners, including the governor Lu Baoji (?-696).
He also moved on to attack the whole Yingzhou area, making the Chinese-
inhabited areas to the north of the River shiver in anticipation of further
military aggression.
On 8 April 697, another major battle was fought between the Khitan
and Zhou annies, the latter of which, comprising one hundred and seventy
thousand troops, was commanded by Wang Xiaojie, the Commander-in-
chief of the Qingbian Circuit. History repeated itself: conducted in the
Eastern Xiashi Valley, this campaign resulted in the complete annihilation
of the 170,000 Chinese soldiers, including even Wang Xiaojie himself,
who was driven off an overhanging cliff.
On 13 May 697, Empress Wu appointed one of her [ust cousins twice
removed Wu Yizong (641-706), who was then the Grand General
of the Right Imperial Insignia Guard, as the Adjunct Commander-in-chief
of the Shenbing Circuit ordering him and He Jiami fiiJ:lb!l1.', the
General of the Right Guard of Leopard Strategy (Baotaowei
to prepare for another round of battle with the Khitans. On 2 June 697,
Empress Wu appointed Lou Shide (631-700) the Vice Comman-
der-in-chief of the Qingbian Circuit, Satuo Zhongyi the Gen-
eral of the Right Militant Awesome Guard, as the Commander of the
Army of the Front (qianjun iW!I!). They led two hundred thousand sol-
diers to attack the Khitans. Obviously, the Empress had staked virtually
the whole of her empire on this single strike. Once again, she turned out
to be extraordinarily fortunate, largely thanks to a tussle betWeen the Turks
and Khitans.
The decisive victory overW ang Xiaojie in April of 697 turned Sun
Wanrong's head. He pondered on one more overwhelming raid on another
major Chinese city Youzhou @j:J+1 (close to present-day Beijing). In hope
of freeing himself of any possible threat to the rear while dealing with the
Chinese, Sun Wamong tried for a provisional alliance with the Turks,
intending to turn against them as soon as he got his way in Y Quzhou.
The Turks saw through his trickery by chance and turned it against him.
They attacked the Khitan base in Liucheng (present-day Chaoyang
Liaoning), seizing all the booty that Sun Wanrong had storedthere.
When news of this reached the Khitan army, which was then battling the
Chinese army, they panicked. One Khitan tribe, the Xi mutinied and
this eventually led to the dispersion of the whole army. Sun Wanrong fled,
followed only by some remnants of his routed army. He did not run too
far before he was beheaded by a servant on 23 June 697.
On 27 July 697, Wu Youyi returned in triumph to the capital from
Y ouzhou. This marked the successful suppression of the Khitan rebellion,
which was not achieved without an enormous loss of life and property on
the side of the Great Zhou government. From 16 June 696, when the
rebellion broke out, to 23 June 697, when Sun Wanrong died, it took
the Chinese army a whole year to suppress the Khitan rebels. In order to
celebrate this hard-fought victory, and probably also for the casting of the
jiuzhou-ding fL:J+IWII (Tripods of the Nine Prefectures)1o, the empress
ordered on 29 September 697 a change of the reign-name from Tiance-
wansui to Shengong (The Divine Feat), apparently attributing the
overcoming of the Khitans to divine intervention.
II.3) Impact
It does not seem a mere coincidence that Fazang availed himself of
the rituals of A valokitesvara in this endeavor. This bodhisattva was then
widely worshipped within China. A Buddhist monk from Uq.q.hyana,
70 Jiu Tang shu 22.867-68; cf. Zizhi tongjian 205.6499, 206.6512.
Damozhantuo once drew on a fme carpet a portrait of the One-
thousand-anned Ayalokitesvara, which he presented to Empress Wu along
with the Sanskrit original of Qianyan qianbi Guanshiyin pusa tuoluoni
shenzhou jing (Sutra of Divine dhii- [spells] [spoken by] the One-thousand-eyed and One thousand-
armed Avalokitesvara). Empress Wu ordered her palace maidens to
embroider the portrait. She also requested a craftsman to draw portraits
of the bodhisattva. The portraits were then distributed throughout the
empire in the hope of perpetrating his "numinous shape" (lingzi
This episode attests to the exceptional degree of esteem that Empress Wu
and her Buddhist supporters rendered to the bodhisattva. Damozhantuo
must have been the monk who is elsewhere simply known-as Zhantuo
a major translator very active under the reign of Empress WU72.
More remarkably, in Changshou 2 (6 December 690-25 November
691), shortly after the empress declared herself as the founding emperor
of the Great Zhou dynasty, another Indian monk closely associated with
Empress Wu, Huizhi (fl. 676-703), composed in Sanskrit a set of odes
in praise of the bodhisattva and then translated it into Chinese. To the end
of the translation Huizhi makes it clear that these odes were dedicated to
Empress Wu, implying that he regarded her as one of the reincarnations
of the bodhisattva

However, it should be noted that it is the Eleven-faced Avalokitesvara,
rather than the One-thousand-anned A valokitesvara, who was invoked
by Fazang in the course of the service he rendered to the Great Zhou
government in 696 or 697. Therefore, it must have been an esoteric text
other than the Qianyan qianbi Guanshiyin pusa tuoluoni shenzhou jing
71 Qianyan qianbi Guanshiyin pusa tuoluoni shenzhou jing, T 20: 83c7 -II.
72 In the capacity of "translator of Sanskrit words" (yiyu Zhantuo was active in
the translation bureaus supervised by Divakara (Xu Gujin yijing tuji 368cl4, Kaiyuan Shi-
jiao lu 9.564aI8, Song gaoseng zhuan 2.719a27), Devendraprajiia (Xu Gujin yijing tuji
369b14, Katyuan Shijiao lu 9.565b22), and Bodhiruci (}{u Gujin yijing tuji 371bI5, Kaiyuan
Shijiao lu 9.570aI8, Zhenyuan xinding shijiao mulul4. 873a9). A Dunhuang manuscript
identifies him, as of October 7,693, as a monk of Jifasi lJ!f7i:;'i in Chang'an and a trans-
lator of Bodhiruci's Ratnamegha translation office. See S 2278; Forte 1976: 172.
73 This text is now extant as Zan Guanshiyin pusa song (T20: 67a-68a).
Forte (1985: 118-122) has convincingly argued that Huizhi was not only its translator, but
also its author.
that was used as the scriptural support for Fazang's Eleven-'faced Avalo-
kitesvara bodhima/:uja.
The shiyimian +-00, ekadasamukha in Sanskrit, means" eleven of the
utmost," or "eleven heads." The Eleven-faced Avalokitesvara was some-
time called Daguang puzhao Guanyin *:Yt-'WJtUll.if (Avalokitesvara with
Great Light and Universallllumination). One of the most popular esoteric
siitras dedicated to this type of Avalokitesvara is Avalokitdvaraekada-
samukha dharaf}l. It appears in four Chinese versions: (1) the Foshuo
Shiyimian Guanshiyin shenzhou jing fas-
cicle, T no. 1070), translated by Yeshejueduo (Yasogupta?) of
the Northern Zhou dynasty74; (2) Shiyimian Guanshiyin shenzhou jing
+-ooftt!tiHI\l5'-E#& by Adiquduo (Wujigao skt. Ati-
kiita, fl. 650s), which was completed sometime between 16 April 653
and 6 May 654, and was included as a part (fascicle 4) in the Tuoluoni
ji jing (3) Shiyimian shenzhou xinjing +- by
Xuanzang 1:"* (602-664) on 27 April 656, only two to three years after
the appearance of AtikUta's version
; and (4) Shiyimian Guanzizai Pusa
xin miyan niansong yigui jing +-ooft (in
three fascicles) by Bukong /G.j:: (Amoghavajra, 705-774) (Tno. 1069)77.
Although in principle Fazang could have used any of the three former ver-
sions, in all likelihood he may have used Xuanzang's given his prestige
as a great translator and that his version was made so shortly after Ati-
kiita's, a fact which attested to the importance given to it by Xuanzang's
patrons, Gaozong and very likely also Empress Wu, who had then suc-
cessfully achieved the hard-fought status as Gaozong's empress.
Regarding the image of the Eleven-faced Avalokitesvara, the earliest
of these Chinese versions gives us the following description. Its height
measures one chi R. and three cun --t, with eleven heads. The three front
faces are those of Bodhisattvas; the three left faces are angry faces; the
three right faces look like those of bodhisattvas, with dog-teeth protrud-
ing [from the mouths]; the rear face is one with wild laughter; the face
at the top is one of the Buddha. All the faces are looking forward, with
74 Lidai sanbao ji (T 49: lUOOe); Kaiyuan shijiao lu, T 55: 7.545a.
75 T 18: 4.812b-825e; see Kaiyuan shijiao lu, T 55: 8.562e.
76 T no. 1071; Kaiyuan shijiao lu, T 55: 8.556a.
77 Zhenyuan xinding Shijiao mulu, T 55: 15.879b, 20.92ge.
lights attached to. the rear. Further, all the eleven faces have flo.wer-crests
(headdresses?), each co.ntaining an image o.f the Amitabha Buddha. The
Eleven-faced Avalo.kitesvara a water jar (kur;u!ikii) in his left hand,
with a Io.tus flo.wer spro.uting o.ut from the mo.uth [of the jar]. Stretching
o.ut his right hand surrounded by jade bracelets, he forms the mUdra o.f
fearlessness 78.
This siUra promises that a ritual devoted to the Eleven-faced Avaloki-
tesvara is able to draw away any enemies pillaging on the bo.rder79. No.t only
was he celebrated for his military pro.wess, he was also believed to be effi-
cacious in dispelling natural disasters like epidemics, as is sho.wn by a
sto.ry recorded in a Chinese collection of Buddhism-related miracles
Although the sutra had already appeared in Chinese translation as early
as the Northern Zhou dynasty, it seems that it did not start to. gain wide-
spread popularity until the Tang, especially after it was successively trans-
lated by both Atikiita and Xuanzang within a two. or three year perio.d.
A telling example of its popularity is that in the third month of Longshuo 1
(5 April 661 - 3 May 661), five years after the appearance of Xuanzang's
version, a Daoist priest of the Xihua Abbey il!f., Guo Xingzhen
(? -663), who then the official title of Grand Master for Clos-
ing Co.urt (chaosan daifu and who was aneo.phyte o.fBuddhism,
made two sandalwood (tan ;tI) statues of the Eleven-faced Avalokites-
vara in additio.n to five gold o.r copper statues o.f the Buddha
. Significantly,
Guo Xingzhen was a confidant of Empress Wu. He started to associate
with her probably in or shortly after 655, when she became's
new empress and when she, out of her sense of insecurity over her posi-
tion in the court, regularly invited Guo Xingzhen to the inner palace to
perform black magic (yasheng aiming at dispelling malicious
spirits and inflicting disaster on her enemies. Ho.wever, it turned out that
his newly aro.used piety towards the Buddha did not bring him good for-
tune. After his black magic was eXPo.sed in 663, which would have
destroyed the empress herself but for her shrewdness and reso.luteness, he
78 Foshuo Shiyimian Guanshiyin shenzhoujing, T 20: 150c. See also Yii 2001 (esp.
79 T 20: 151b25-28.
80 Sanbao ganying yaolUe lu, T 51: 3.852c.
81 Ji gujinfodao lunheng, T 52: 4.395cb-397a.
was banished to Aizhou (in present-day Qinghua mit in Guangxi)
where he died
The unusual closeness of Guo Xingzhen's relationship
with Empress Wu means that both his decision to switch his religious
faith and his efforts to cast the statue of Eleven-faced Avalokitesvara
must have been tacitly approved if not instigated by her. Thus, it seems
that Empress Wu might have been exposed to the worship of the Eleven-
faced Avalokitesvara as early as the beginning of the 660s, three and half
decades before Fazang invoked the power of the bodhisattva to serv-
ice. It is also noteworthy that one of her confidants became a devotee of
the Eleven-faced Avalokitesvara when his patroness was deliberately
working through to the political summit from where she was able to rule
as a co-emperor8
On either 24 October 696 or 22 December 696 - almost simultane-
ous with Fazang's availing of the dhiirar.zz of the Eleven-faced Aval-
okitesvara - a Buddhist thaumaturge Qingxu m (active 696-712), who
was to become a friend of Fazang in ten years or so, allegedly succeeded
in saving his own hermitage from a fire, which devastated other neigh-
boring buildings, on Mount Sanzhong =!\!!, to the north of Lingyansi
(probably an error of Lingyansi in Qizhou (Shan-
dong), not very far from the battlefield of the Great Zhou and Khitan
Some time in 702, the same Qingxu resorted to the dhiirar.z;: of
the Eleven-faced Avalokitesvara and that (or those) in the Jin'gan bore
jing to pray for rain at the request of Fuli, whom we have
already identified as a foe of Fazang. It worked
However, the same
Avalokitesvara dhiirar.zz did not prove efficacious two years later (around
8 May 704), when Qingxu was requested to pacify a malicious spirit
hatiIiting a Buddha-hall at Shaolinsi on Mount Song i'i!li LlJ86.
82 See Xin Tang shu 76.3474, Zizhi tongjian 201.6342, for Guo Xingzhen's involve-
ment in Empress Wu's court strife. Gaozong's edict, dated 17 January 664, condemning
Guo Xingzhen is fully quoted in Fayuan zhulin (T 53: 55.705b5-18), which Daoshi iiilt
(5967-683) precedes with a summary (705a27-b5).
83 Sima Guang (Zizhi tongjian 200.6322) remarks that after overcoming this crisis, Empress
Wu became the de facto ruler of the empire. See Twitchett and Wechsler 1979: 255.
84 lin' gang bore jing jiyan ji, XZl149: 2.47c6-15; Qingxu-Fazang relationship discus-
sed in Cill.2).
85 lin'gang bore jing jiyanji, XZl149: 3.53bI6-cI4.
86 lin'gang bore jing jiyanji, XZll49: 2.48b7-d4.
At any rate, Fazang's feat seems to have further contributed to the pop-
ularity of the Eleyen-faced Avalokitesvara, as is shown by the fact that
it was exactly such an image of A valokitesvara, rather than that of the
one-thousand-armed and -eyed A valokitesvara, that was materialized in
a statue within the Qibaotai -I::;.1i (Tower of Seven Jewels). This tower
(actually very likely a pagoda enshrining some relics of the Buddha) was
completed around 703 under the supervision of Degan (ca. 640-
after 703)87 - another major Buddhist ideologue of Empress Wu - at
the Guangzhaisi a monastery in Chang'an of essential impor-
tance to Empress Wu's pursuit and wielding of supreme power.
The cult of the Eleven-faced Avalokitesvara eventually infiltrated so
deeply into society that he became embodied in Sengqie (Satpgha?,
628-710), a Central Asian Buddhist thaumaturge, who arrived in China
in the early Longshuo era (661-663) (Yii 2001: 211-222). This embodi-
ment, in turn, catalyzed the cult of Sengqie and its integration with the
cult of the Eleven-faced Avalokitesvara. After spending some time in
Xiliang Prefecture ifflrffii-M', he settled in the Longxingsi of Shan-
yang W where he wrought various miracles. Then, he moved to Lin-
huai Wo$, where he impressed a local householder so much that he sur-
rendered a plot of land for building a temple. From there was unearthed
an old epitaph, which revealed that the place was the old site of the
Xiangjisi of the Northern Qi dynasty, and an image of a Buddha,
who was called "Puzhao wang" 'W]ffiE (The King of the Universal illu-
mination). There was a legend that when he was staying at the home of
one of his patrons (called Reba .tt), his body suddenly grew so much
that it exceeded the whole bed by three chi. Re subsequently turned him-
self into the Eleven-faced Avalokitesvara. Some time after being called
to Zhongzong's court in Jinglong 2 (January 28, 708-February 14, 709),
Sengqie proposed to the emperor that his temple at Sizhou 71!!11+1 be
renamed Puzhaowangsi - the temple of the King of Universal
illumination. Given that the character zhao ]ffi was then tabooed because
of Empress Wu' s personal name Zhao Zhongzong modified the tem-
ple's name as Puguangsi The name of puzhao 'W]ffi or Puguang
87 For this monk, see Forte 1976 (esp. pp. 100-108); for Qibaotai's importance under
the late years of Empress Wu, see Yen 1986, Chen Jinhua 2002: 92-97.
tf1t obviously echoes the Dazhao puguang Wang, a name of the Eleven-
faced Avalokitesvara. That Sengqie was an avatar of Avalokitesvara
was verified by Wanhui ~ j ) g L (632-711), another Buddhist thaumaturge
in Zhongzong's favor. When Zhongzong, who was amazed by a series
of miracles that arose following Sengqie's death, asked him to reveal the
real identity of Sengqie, Wanhui confIrmed this to the emperor (Yii 2001:
ill) Fazang the Wonderworker
As a politician, Fazang adroitly interacted with his secular patroness and
patrons, and with other leaders of political groups at the time, exerting a
subtle influence on contemporary court politics that was hard to ignore.
As a warrior, Fazang combated the enemies of the Great Zhou empire with
a special weapon - black magic. His roles as a skillful politician and an
awe-inspiring warrior are the topics which the preceding two parts of this
article have addressed. Let us now turn to another of his multiple roles,
which was even grander compared with the former two; i.e. that of an
intermediary between the heavenly and human realms. According to his
followers, he undertook this role not only with his supernatural abilities
to conjure up miracles, but also through his passion for some special forms
of religious practices, including relic veneration and self-immolation.
ill.l) Miracle Stories Related to Fazang's Mastery of the AvatafJlsaka
Centering around the theme of Fazang's extraordinary capacity as an
Avatarpsaka preacher, a series of stories and legendswere created and pro-
moted both within and without the Chinese A vatarpsaka tradition. These
stories and legends can be roughly divided into two categories, one in
which the legendary elements are so clear and overwhelming that they can
be taken as no more than faithful accounts, or simply legends without
historical veracity. The second consists in those which are manufactured
in such a way that legendary and semi-legendary elements are mixed with
some accounts which mayor may not be verified historically. Compared
with the first category, the second appears more complicated and deserves
more attention. :rn. this section, let us try to study one example of the first
and two more belonging to the second.
Of the fIrst category, the following legend is quite characteristic and
telling. During the Yonglong era (21 September 680-24 January 681),
a native of Yongzhou Guo Shenliang $Jl:fljl:;t, who had continued
to cultivate pure practices, died suddenly. The deities led him to
Heaven to pay homage to Maitreya. A bodhisattva there asked him, "Why
didn't you receive and uphold the Avatarrzsaka sutra?" Guo Shenliang
replied, "nobody preaches on that sutra." The bodhisattva said, "There
is indeed someone who preaches [on the sutra], why do you say that there
is not?" Later Guo Shenliang returned to life and recounted in detail this
experience to Dhanna Master Baochen (?-688
), one of Fazang's
mentors, who discussed it with him in detail. The author (editor) of this
story then remarks, "Looking closely into this, we fInd that Xianshou's
[Fazang's] expositions and turning the dhanna-wheel were such that their
powers was known even in the most ethereal [realms]! "88
We cannot exclude the existence of a layman named Guo Shenliang,
who was obviously an acquaintance (or even a friend) of Fazang. How-
ever, his experiences as a traveler to Heaven, where he is allegedly
instructed on the superiority of the Avata1'Jlsaka sutra and the avail-
ability of an A vatatp.saka preacher in the area, can only be accepted as
a piece of religious narrative concocted and promoted by a believer of
the Avata1'Jlsaka sutra - very likely Fazang himself given that the leg-
end made its first known appearance in a collection originally compiled
by him.
Although it is possible that the main body of the story was written by
Fazang himself, we can certainly reject the idea that the last couple of sen-
tences in praise of the supernatural power of his lectures were by his own
hand; they must have been, rather, added by his disciples. A comparison
of the two versions of the same story in the Huayan jing zhuanji and
Ch'oe Ch'iwon's biography shows that Ch'oe Ch'iwon seems to have
slightly recast it to the extent that it further features Fazang's lectures
88 Popjang chOn (T 50: 281cll-16) contains a largely identical version of the same story.
having attracted attentions from both worldly and celestial beings. If this
analysis of the formation and development of the legend can be accepted,
then we can see an interesting process through which a legend which
originally focused on the superiority of the A vatarp.saka teachings in gen-
eral was recast into a new one in which Fazang's brilliance as an Bud-
dhist preacher (especially his skill in lecturing on the Avataf!lsaka sidra)
became the central theme.
As for the second category, we have the two most famous ane!- repre-
sentative stories/legends, the Ordination Episode and the "Earthquake
Story." Not only do these two stories feature Fazang's exceptional capac-
ity as an expounder of the A vataf!lsaka siltra, but they also portray the high
esteem that Fazang evinced from Empress Wu by some miracles associ-
ated with or directly brought out by his A vatarp.saka lectures. They have
been deeply embedded in Buddhist historiography, to the extent that they
have been taken for granted and few scholars have ever given a second
thought to their historical credibility. However, as was noted above, the
Ordination Story does not have any historical support and cannot be taken
as more than a legend.
The earthquake episode proves more complicated. Given that this record
directly quotes from the reply from Empress Wu and that it was found in
a text compiled shortly after the death of Fazang, a time so close to
Empress Wu's reign that it would be virtually impossible for anyone to
fabricate such an edict in the name of Empress Wu, I believe that it should
have some historical basis, although the event might not have happened
exactly the way as is described here. The following scenario appears close
to the truth. In the course of lecturing on the new A vatarp.saka translation,
probably on 7 January 700, a small-scale earthquake broke out in the
region close to the Foshoujisi, not necessarily at the moment when Fazang
lectured on the sentence regarding the quake in the A vatarp.saka-sea. Very
likely, the Foshoujisi monks correlated the earthquake with the sentence
in the siitra in an attempt to recast it as a propitious portent related to the
sidra. Given that an earthquake was generally understood as a punitive omen
from the heavens, this reaction of the Foshoujisi monks can also be read
as a deliberate act of turning a sign, which would be taken as unfavorable
in traditional Chinese thought, into a favorable one that accorded with Bud-
dhistideology (Chen 2003: 329-336).
JII.2) Fazang's Supernatural Ability to Bring down Snow and Rain
Fazang had begun to enjoy a high reputation as an efficacious invoker
of rain long before Zhongzong began to rule again in 705. During Empress
Wu's regency and reign, some local officials around the Chang'an area,
who suffered from ravages of a drought, had already repeatedly emolled
this kind of supernatural power that Fazang allegedly possessed. As early
as Chuigong 3 (19 January 687-6 February 688), when Empress Wu did
not formally rule in the right of an emperor but as the regent of her
emperor-son Ruizong, a serious drought struck the capital area. Empress
Wu ordered Fazang to construct a platform at Ximing si to pray for rain.
The Magistrate of the Chang'an District Zhang Luke ~ ~ ~ (?-687t),
an uncle of Zhang Yizhi and Zhang Changzong (Fujiyoshi 1997: 320),
acted as the "host of the prayers" (qingzhu @). After strictly observing
both fast and precepts for less than seven nights, the rain started to pelt
down. During the Tiancewansui reign-era (22 October 695-20 January
696), while the Senior Subaltern (zhangli f t ~ ) ofYongzhou ~ 1 + 1 , Prince
of Jian' an 9t1< (i.e. Wu Y ouyi, a first cousin once removed of Empress
Wu who, as noted above, played an important role in the suppression of
the Khitan rebellion)89, performed his duties in Yongzhou, a drought
attacked the area. Like Zhang Luke, Wu Youyi turned to Fazang for help.
It was reported that the rain poured down while Fazang prayed, as swiftly
as echoes responded to a sound.
Now, let us see how Fazang was sought out by Zhongzong and Ruizong
for his expertise in invoking rain during the seasons of drought. In the
mid-summer (i.e. the fifth month) of Jinglong 2 (24 May-22 June 708),
a drought started to threaten the capital area once again. The emperor
ordered Fazang to gather one hundred dharma-masters at Jianfu si to pray
for rain with proper religious rituals and procedures
. Approaching the
dawn of the seventh day, a heavy downpour fell from the sky. It lasted
89 Wu Youyi became a prince on 20 October 690 (Jiu Tang shu 183.4729, Zizhi tongjian
204.6467-68), a mere four days after Empress Wu proclaimed the foundation of her own
90 At least two records are left of Zhongzong's Jianfusi visits, once on 28 May 707
(Shenlong 3.4.23 [gengyin]), and the other 3 May 709 (Jiu Tang shu 7.144,147); discussed
in Sun 2003: 135-137.
for ten nights, until everyone was satisfied with the rainfall. Zhongzong
made manifest his satisfaction and excitement over the performance of
Fazang and his colleagues in his reply to the memorial submitted to report
the result of this rain-praying ritual
. The next year, when the drought
recurred, Fazang came to people's rescue again. Zhongzong issued another
edict to extol his merits92. Ch'oe Ch'iwon continues by saying that from
then on, Zhongzong and Ruizhong relied on Fazang as their bodhisattva-
preceptor. This might refer to the possibility that some time during the
Shenlong or Jinglong era Fazang was invited to the Linguang Palace
Chapel, as was suggested at the beginning of (I.3).
In the spring of Jingyun 2 (24 January-22 April 711), it did not rain
enough, causing a shortage of water. To make things worse, it did not snow
in the winter. The Chang'an area was on the verge of another severe drought.
Ruizong summoned Fazang to the inner palace, eagerly seeking from him
the method to counter the damage to crops threatened by the imminent
drought. Fazang recommended to the emperor an esoteric Buddhist scrip-
ture called Suiqiu zede Da zizai tuoluoni 1EWt:*IfB. He also
proposed that a platform be set up so that Buddhist priests, with peace and
purity in their minds, could copy and recite the dhiiralJ-1 in the sidra before
throwing the dhiiralJ-z-scripts into a dragon-pond. He anticipated that this
would cause some snow to fall. Ruizong was convinced and ordered that
the proposed procedures be carried out under Fazang's guidance beside
the dragon-pool at or beside the Wuzhensi in Lantian ii: EEl valley
on Mount Zhongnan, where Fazang had started his search for Avatrup.saka
teachings and also some Daoist practices in his youth.
The Suiqiu zede Dazizai tuoluoni [jingJ was probably the same text that
descended to us by the title "Foshuo suiqiu jide Da zizai tuoluoni
shenzhou jing" 1EWt:*IfB1$ (The Satra preached
by the Buddha on the DhiiralJ-f-riddhimantra of Great Self-existence to
be obtained as one wishes) (in one fascicle) (T no. 1154) translated by
Baosiwei (Manicintana?, ?_721)93 in 693. This text contains an
extensive dhiiralJ-f. The whole text is partly identical with the Pubian
91 Popjang chOn, T 50: 284b3-5.
92 Popjang chon, T 50: 284b5-7.
93 For this Kashmiri monk, see Forte 1984.
guangming [yanmangJ qingjing chisheng ruyi baoyin xin wunengsheng
da mingwang da suiqiu tuoluoni jing 1f:@Jft Ij)j En
(in two fascicles) (T no. 1153), trans-
lated by Amoghavajra (Bukong::f:5:: [705-774]), although the latter rep-
resents a more developed version.
Fazang's choice of the pool at or beside the Wuzhensi to perform this
ritual for snow should not be easily passed over as a coincidence. Since
the Sui dynasty (581-617) the temple had attracted a number of eminent
monks, including (1) Jingye (564-616), the founder of the monastery,
who was a joint disciple of two Sui Buddhist leaders Jingying Huiyuan
(523-592) and Tanqian (542-607) (Chen 2002a: 41 [no 85]);
(2) Huichao VJm (546-622), a major disciple of the Tiantai patriarch
Huisi V,I[!!, (515-568) (Chen 2002a: 200-201); (3) Facheng (563-
640), who was an admirer (or even a disciple) of Huichao and who
contributed immensely to the renovation and expansion of the temple
(4) Huiyuan VJ! (597-647), who, as perforce the most important disci-
ple of the Sanlun master Jizang (549-623), spent his last decade or
so at the Wuzhensi
; (5) Baogong (542-621) and (6) Huiyin VI!!
(539-627) (Chen 2002a: 41 (n. 85), 170 [no 56]); and last but definitely
not the least, (7) the Pure-land master Shandao (613-683), who
achieved his sobriquet "Great Master Zhongnan" allegedly
thanks to his lengthy stay at the Wuzhensi

Furthermore, Fazang himself had developed a close tie with this tem-
ple starting from an early phase of his career. Although we are not clear
as to exactly when he started to be affiliated with the Wuzhensi, this
must have happened no later than Yifeng 2 (8 February 677-27 January
678), assisting as he did on 29 August 677 (Yifeng 2.7.26) a Pure-land
aspirant called Xuanji (a.k.a. Jingwu (640-706) andten more
monks to erect at the monastery a pagoda celebrating the various mira-
cles that emerged in the process of copying the Diamond Sutra
evidence even suggests that he had been a leader of the monastery by
94 See Facheng's Xu gaoseng zhuan biography at T 50: 28.688c-689b.
95 See Xu gaoseng zhuan, T 50: 11.514c13-17, 515a5-8; Hongzan Fahua zhuan, T 51:
3.19b-c, 17c21.
96 Shinshil ojo den chitkan itzubun, Zokujodo-shit zensho 16: 92a17-b3.
97 See HongzanJahua zhuan, T 51: 1O.47b7-13.
cases. In contrast to Daoism, in which dragon, as noted above, was explic-
itly employed as a messenger, Fazang and his colleagues seem to have
contented themselves by implicitly calling a dragon into service. Actually,
I am willing to believe that Fazang's proposal of throwing the Esoteric
scriptural scripts into the "dragon pool" at the Wuzhensi was not merely
based on the usual association between dragon and water (and by exten-
sion, rain or snow), it might have also been - although apparently not
exclusively - inspired by the crucial role that the legendary s:reature
played in the Daoist practice of toujian.
Thus, the snow-prayer ritual that Fazang performed at the Wuzhensi in
711 must be taken as a very significant example of his interest in some
Daoist practices and his capacity to perform them. Fazang's interest in
Daoism probably derived from his long seclusion at Mount Zhongnan in
his youth. His friend Yan Chaoyin, who wrote the funeral epitaph for
him, characterizes Fazang's experiences at Mount Zhongnan by a general
expression, ya yi chongxuan ("often investigating into the dou-
ble mysteries"), which might at least partly refer to some general Daoist
theories, if not specifically the Daoist trend known as chongxuan
assumption is corroborated by Ch'oe Ch'iwon, who unambiguously tells
us that the year following his burning off a finger in front of the Famensi
pagoda when he was only sixteen sui (i.e. the year 658), he "entered the
mountain (i.e. Mount Zhongnan) to learn the Way (xuedao an
expression which in Classical Chinese usually indicated one's effort to pur-
sue Daoist ideas and practices. On another occasion, Ch'oe Ch'iwon
also notes that during his seclusion at Mount Zhongnan in search of the
dharma, Fazang ate "numinous fungi" (zhu JIt) for several years!05. This
suggests that Fazang practiced both Buddhism and Daoism on the moun-
tain. Thus, as is implicitly insinuated by Yan Chaoyin and explicitly indi-
cated by Ch'oe Ch'iwon, who are respectively the earliest and the best
biographers of Fazang, when the young Fazang climbed Mount Zhongnan
in 659, he was primarily (if not exclusively) in search of arts of longevity
and immortality, rather than the Buddhist wisdom of how to achieve enlight-
enment and nirvar;a - this despite his previous startling self-immolation
104 On this significant trend in Tang thought, see, most recently, Sharf 2001: 52-7l.
105 Popjang chOn, T 50: 281a29-bl.
act at the Farnensi, which unmistakably conveyed his devotion to Bud-
It seems after entering Mount Zhongnan, Fazang continued
to learn and practice Buddhism along with Daoism. It nQt surprise
us that a religious environment like Mount Zhongnan, where Buddhism
and Daoism reached an exceptionally intense degree of convergence
and interaction, allowed and encouraged Fazang to maintain and develop
- simultaneously (if not even-handedly) - his enthusiasm and
edge in the two religions. The mountain was at the time frequented by a
number of world-renouncers who were both A experts and
Daoist adepts, some of them also self-immolators
- for whom Fazang
must have felt a great deal of sympathy given his own experiences in this
After this excursion on Fazang's association with the Wuzhensi and his
previous interests as a Daoist adept, let us now return to the result of his
711 snow-prayer rituals. As reported in his biography, the ritual was quite
successfuL Before ten days passed, it started to snow heavily. Fazang
or the monks at the Wuzhensi sent a memorial to the throne, to which the
emperor responded with high appreciation, urging Fazang and his col-
leagues to remain on the mountain to continue their efforts for more
After it snowed six times and in all four directions, a decree
was issued again, to inquire after Fazang's health. Ruizong attributed all
this plentiful snowfall to the compassion of the Tathagatha and Fazang's
sincere prayers as well

ill.3) Other Miracle Stories Associated with Fazang
In view of the importance of Daoism in the earliest phase of Fazang's
career, one might get the impression that he must be rather friendly to
Daoist priests. This impression would not be borne out insofar as we can
accept the historicity of a story told by one of his direct disciples, which
106 Huayanjing zhuanji (T 51: 4.165C11-166aI5), for example, records an AvataTflSaka
adept acquiring immortality after making acquaintance with some local deities, who resem-
bled, according to Fazang's description, Daoist immortals. For three outstanding examples
of Avatarpsaka masters who were self-immolators at Mount Zhongnan, see (ill.3).
107 PiJpjang chiJn, T 50: 284b22-26.
108 POpjang chiJn, T 50: 284b26-29.
reflects at least some of his real attitude toward that religion after he
became an A vataIp.saka master
Reaching the second year [of the Tianshou era] (6 December 69025 Novem-
ber 691), [people in Cengzhoulentreated [Fazang] to lecture on the Huayan
jingo After preaching on the dharma, the discussion [between him and his
audience] carried them to issues of what was orthodox and what heterodox.
A young Daoist priest, who was then present, returned to report to the Head
of Hongdao Abbey saying, "The preacher in the temple to the north
has disparaged the Daoist Worthies." This exasperated the Head. Next morn-
ing, he led over thirty Daoist priests to go to the lecture center. With a face
contorted with anger, he uttered coarse words, asking Master [Fa]zang, "It
would be all right if you just focused on your lectures. [But] why did you
[rashly] comment on things related to Daoism?" Master [Fa]zang replied,
"A poor monk [like me] has only lectured on the Huayan [teachings], with
no intention whatsoever to comment on or disparage other [teachings]." The
Chief of the abbey asked, "Are all the dharmas equal?" Master [Fa]zang
replied, "All the dharmas are both equal and unequal." The Head asked
agairi, "Which dharmas are equal, and which not?" [Fazang] replied, "None
of the dharmasgoes beyond the sphere of two categories, one Absolute
Truth and the other Provisional. In view of the Absolute Truth, there is nei-
ther this-ness nor that-ness, neither self nor others, neither purities nor impu-
rities, [since they] are all detached from [any characteristics]. Therefore, [all
the dharmas] are equal. However, when judging from the view of Provi-
sional Truth, there are distinctions between the good and evil, the honorable
and humble, the orthodox and heterodox - how could they be e@al?"
Although the priest found himself unable to respond to the argument, he
couldn't constrain his anger, raising ppisonous and harmful words in the
"Place of Tathligata" (i.e. the Buddhist temple).

U.3::.rJ:lB, .
U.3::.Xr"" "-tJ]
-tJJ'&M, :ff#:ff., :ff$:ffiE,

He then returned to the abbey. The night passed [without anything abnor-
mal happening]. But in the morning, when he washed his face and hands,
109 Da fanguangfo Huayan jing ganyin ji, T 51: 176a-b.
110 Obviously an error for 1:. '
III ;J:I; is probably an error for if.
I1is eyebrows and hair all started to fall out suddenly, and boils erupted all
over his body. Not until then did he start to repent, and take refuge in the
"Three Treasures." He pitifully begged [pardon] from Master [Fa]zang,
vowing to recite and uphold the Huayanjing one hundred times. After chant-
ing the sutra for about two years, there were still ten times [short of the one
hundred times], he felt [to his delight] that his eyebrows and hair start to grow
out again and the sores in his body start to heal. This was seen and heard
by both the religious and lay people in Cengzhou.

It is interesting to note that this miracle tale attributes Pazang's suc-
cess in defeating a Daoist challenger not so much to his skill in the expli-
cation and application of the Madhyamika theory oQthe "Two Truths,"
as to his prowess in exerting a kind of black magic, which is, as sug-
gested by the plot of the tale, responsible for the symptoms of leprosy that
befell his unlucky rival. It is quite ironic for a Buddhist theoretician of
Fazang's fame that his magic prowess is here depicted as more decisive
than his eloquence and theoretical sophistication in proselytizing a dis-
In addition to the light that such a miracle story throws on Pazang's
image among his admirers of later generations, it also contains some his-
torical elements valuable for our efforts in constructing some aspects of
Fazang's life and his intellectual background as well. Although it is hard
for a modern scholar to believe that a Daoist leader was indeed defeated
and converted by a Buddhist master in the way described here, it was
probably true that there was indeed such a Daoist priest, even one as
important as the head of a Daoist abbey at the local level, who was con-
verted by Fazang. We should take this possibility more seriously in view
of the fact that at the time some Buddhist monks, probably encouraged
by Empress Wu's pro-Buddhist policies, were rather aggressive in
approaching Daoists and their religion, sometimes even going to the
extreme of converting them by violence, as is shown by the case of Huaiyi
112 Qiuai *:a is probably an error for aiqiu :a*.
113 Da Fangguangfo huayanjing ganyinji, T 51: 176a-b. Afar briefer account of this
standoff is found in P6pjang chOn, T 50: 283cll-16.
(? -695) - this extraordinary monk was accused of having physi-
cally tortured those Daoist priests who refused to convert (Forte 1998,
1999). What is more intereSting in this regard is that in 696 the head of
the Hongdao Abbey in Luoyang, Du Yi 11)(, gave up his faith in Dao-
ism and had himself ordained as a Buddhist monk, taking the dharma-
name Xuanyi To the irinnense dismay and exasperation of his for-
mer religious brothers, Du Yi/Xuanyi wrote a three-fascicle work, titled
"Zhenzheng 1un" to criticize Daoism and defend Bud<;lhism
We note with interest that the author of our story identifies his Daoist
priest also as the head of the Hongdao Abbey, although he locates this
abbey in Cengzhou ;1+1 (actually an error for Xiazhou present-day
Baichengzi in Shaanxi) (For this correction, see Chen Forthcom-
ing: Chapt. 2), rather than in Luoyang. Is the author recasting the Du Yij
Xuanyi conversion in such a way that Fazang is represented as his tamer,
or he is here simply partly reproducing the famous case by contextualizing
his tale in a marginal area with a homonymous Daoist abbey? Whatever
the real situation, this tale is definitely worth serious note for those who
are interested in the Buddho-Daoist and state-saIp.gha relationships during
this period, when Buddhism was reaching its heyday under the patron-
age of Empress Wu.
Fazang is also associated with some other miracles that are related to
the Famensi relic, which seems to have played a very crucial role in his
life, starting from the very beginning of his career. As a matter of fact,
his biographers are unanimous in telling us that the finger-bone enshrined
in the Famensi pagoda which was believed to have been Slikyamuni's, was
a major catalyst that triggered his enthusiasm for Buddhism. As noted
above, he burned a finger in front of the Famensi pagoda when he was
only sixteen sui. We have also already noted that the last example of
cooperation between Fazang and Empress Wu was also mediated through
the same relic and ironically it was also through a process of venerating
this very relic that Fazang eventually ended up as a betrayer of empress
Wu. This is a very important aspect of the Famensi relic veneration at the
beginning of 705 that we attempted to investigate in the first part of this
article. The time is now ripe for us to examine another of its dimensions.
114 For Du Yi and his authorship of the Zhenzheng lun, see Palumbo 1997.
At the end of Chang'an 4 (10 February 704-29 January 705), Empress
Wu, whose healfu .was then deteriorating, had an audience with Fazang
in her palace chapel within the Longevity Basilica (Changsheng-dian
During this audience, Fazang raised the Famensi relic, .with
which she was by no means unfamiliar. Forty-five years earlier, in April
or May of 660, when Gaozong started to have some severe health prob-
lems he and his empress ordered the transferal of the relic to the inner
palace, where they worshipped it for almost two years before sending it
back. As scholars generally believe, in this relic veneration it was mainly
the relic's alleged therapeutic power that was invoked for the sake of the
ailing emperor (e.g. Sen 2002: 69). Given this former tie with the Famensi
relic and her own deteriorating health at the time, it is quite understand-
able that when Fazang mentioned the relic Empress Wu responded very
enthusiastically. She immediately ordered one of her Vice Directors of the
Secretariat Cui Xuanwei (638-705), Fazang and nine more bha-
danta-monks to go to the Famensi to fetch the relic to Luoyang.
Before opening the Famensi pagoda, the imperial emissaries and their
entourages performed a seven-day observance, probably in front of the
pagoda. When it was brought out, the relic emitted dazzling rays of light.
Fazang was emotionally overwhelmed. He held his votive text in hands,
reading it aloud to the people present there. The relic shone on the palm
of his hand, lighting up places both near and far away. In accordance
with the power of the merit that they accumulated over their past lives,
people on the spot saw different divine phenomena. Driven by their flam-
ing religious passion, they competed with each other in performing acts of
self-immolation. They also feared lagging behind in offering donations

The imperial team returned to the Chongfusi in Chang'an (that
is, the Western Taiyuansi as it was then known at the time) with the relic
on the very last day of that year (29 January 705). On this day, the Regent
(liushou ii'if) of Chang 'an, who was very likely no other than Wu Youyi
(Chen 2002: 99nI65), led all the officials and five congregations of
Buddhist believers in Chang'an to prostrate themselves at the left side of
the road, greeting the relic with extravagant offerings including fragrant
flowers and various types of music. The relic allegedly brought sight and
lIS Popjang chOn, T 50; 283c-284a.
hearing back to the deaf and blind, enabling them to see the relic and hear
the music honoring it.
On the eleventh day of the fITst month of the new year (9 February 705),
the relic entered Luoyang. The empress ordered the officials below the
ranks of Prince and Duke, along with commoners in Luoyang and its
adjacent areas, to carefully prepare banners, flowers and canopies. She also
ordered the Chamberlain for Ceremoillals (taichang J,\:1!t) to perform
music and to greet the relic as it was placed in the Luminous Hall. As the
third storey of the Luminous Hall was actually a pagoda (Forte 1988:
161-163), it should not come as a surprise at all that Empress Wu chose
this building as the location for the ceremony of honoring the Famensi relic.
Then, on the day of "Lantern Watching [Eve]" (guandeng-ri B; i.e.
the fifteenth day of the first month [13 February 705]), Empress Wu, with
her mind and body properly maintained and purified and with an expres-
sion of supreme piety on her face, asked Fazang to hold up the relic as
she herself prayed for universal good.
This relic-veneration allegedly brought up a number of miracles, as is
described by Ch'oe Ch'iwon. First, on the first day when the relic was
unloaded from its reliquary, it emitted some light. Second, on the border
of the Wugong sub-prefecture (in present-day Baoji Shaanxi),
the light from the relic shot back to the Famensi and came to encircle it.
Third, on the night spent at the Chongfusi, where the relic was kept in
the Grand Hall (huangtang probably a hall reservred for the spirit
of Gaozong), some lights, as bright as flames and shooting stars, issued
from the relic. Fourth, when the relic arrived at the gate to the Chongren
,*{= quarter, an aura appeared around the sun. Fifth, on the night spent
atthe Xingfasi in the Weinan mm sub-prefecture, the light from
the relic made the night as bright as if it were daytime. Sixth, when arriv-
ing on the border of the Shouan sub-prefecture, the light from the
relic shot into the sky, bringing forth, once again, an aura around the sun.
Seventh, the relic produced light again when Empress Wu and the Crown
Prince (Zhongzong) carried the relic, wrapped with the tUla silk, on the
crowns of their heads116.
116 Popjang chOn, T 50: 284a14-19.
Since it was brought to Luoyang at the very end of 704 at the order of
Empress Wu, the relic had not been returned to its home tem-
ple until almost forty months later, when the Great Zhou dynasty was
over and the Great Tang had been restored for about three years.
and Wengang (636-727), two of the ten monks who accompanied
the relic to the capital from Famensi, were among the monks escorting
the "sacred bone" back to Famensi on 11 March 708. Fazang, in partic-
ular, made for the relic a "spirit canopy" (lingzharig which was
excavated in 1987117. A stone stde unearthed in 1978 from near the
Famensi pagoda reveals an extraordinary practice on the part of the royal
family-Zhongzong, Empress Wei and their prince, two princesses and
the empress's two sisters cut off their hair to be buried with the newly
re-enshrined relic at the Famensi on 11 March 708 (Han and Luo 1983).
We do not know whether the relic was sent back to Famensi from Luo-
yang or Chang'an, to which Zhongzong switched his imperial court on
7 December 706. It could be that Zhongzong brought the relic with him
when he left Luoyang or that he just left it there 118
Started at the very end of Empress Wu's reign, this series of "Fazang-
directed" relic-veneration activities was carried on toward that of Zhong-
zong's. It is noteworthy not only because of the various miracles that
adorned its repeated climaxes, but also those acts of self-immolation
that were inspired by and emphatically punctuated the whole process.
On the one hand, either out of some deliberate pre-planning or largely
acting in accordance with the volatile sociopolitical conditions, Fazang
had aptly turned this series of seemingly pious acts to the best service of
his religious tradition; on the other, this complicated religious drama
vividly reflects some long-obscured aspects of his intellectual and religious
background, most notably his inextricable involvement in the practice of
117 This role of Wengang is recorded in his Song gaoseng zhuan biography at T 50:
14.792a21-22. Fazang's role on this important occasion, however, is recorded in none of
his biographical sources, including the most thorough one, that by Ch'oe. We fortunately
know this from an inscription on the "spirit canopy" unearthed from the Famensi under-
ground chamber containing the fmger-bone relic. See Wu and Han 1998: 70.
118 Two years later, Zhongzong decided to honor the Famensi relic once again by
bestowing the title, "Dasheng zhensheng baota" ("Treasure-pagoda for the
True Body of the Great Sage"), on the pagoda; Chen Jinhua 2002: 102-103.
self-immolation, which most of modem scholars are still not so prepared
to associate with such a learnt and elite Buddhist priest like Fazang.
It will take us too far afield to discuss how orthodox Buddhism in
medieval China treated the issue of self-immolation. Suffice here. to say
that Buddhist doctors in the tradition were rather divided on this topic.
Although some of them enthusiastically endorsed and promoted it, a
majority of them were quite reluctant to do so - some of them were
simply harsh critics. For example, Yijing (635-713), a collaborator
of Fazang, devotes one section exclusively to self-immolation in the forty-
section report that he sent from South Asia to his Chinese colleagues.
He rejected it as an inappropriate practice (shaoshen buhe
(Wang 1995: 222-223, discussed in Benn 1998). It is therefore of par-
ticular interest to see how Fazang, a self-immolator himself, wrote about
this issue.
One of the strongest "scriptural" source for self-immolation in medieval
China was the apocryphal Fanwangjing James Benn (1998) has
recently convincingly shown that a chief motive of the Chinese author of
this text was perhaps to legitimate self-immolation. In view of this, it is
nothing but natural that it is in his commentary on the F anwang jing (of
whose apocryphal nature Fazang might or might not have been aware) that
Fazang expresses himself most explicitly on this highly controversial
issue. According to Fazang, the Buddhist stories about various heroic acts
of self-immolations not only can and should be understood literally,
but they were also to be seriously emulated, the more closely the bet-
terll9. The unreservedness with which Fazang sanctioned those self-immo-
lation acts, even those as radical as burning one's forearms or feeding a
hUJlgry tigress with one's own body, is really striking in the light of their
controversial nature and the relatively lukewarm attitudes held by other
Buddhist exegetes, either his contemporaries or predecessors. They either
understood these stories metaphorically (jukuang zhi qi which
could not be interpreted -let alone emulated -literally (e.g. Zhiyi
[538-597]); or were of the opinion that these exceptionally arduous deeds
were highlighted just in order to test the steadfastness of people's faith -
in other words, these stories were only used for pedagogical, and educational
119 Fanwang jing pusajieben shu, T 40: 5.641clO-15.
purposes (and should not be put into practice) (e.g. [d.u.]);
. or conceded that can only be committed by lay believers,
and not by monks for its damaging effects on their awe-inspiring man-
ners, which had irreplaceable importance for promoting religious causes
(e.g. Siingjang Mj5lf [d.u.])12o. .
As James Benn (2001) has eloquently shown with ample examples,
self-immolation was a widespread practice among medieval Chinese Bud-
dhist believers. However, it is noteworthy that Avata1?lsaka followers who
also happened to be mown as self-immolators seem particularly numerous.
In addition to Fazang's case, another similarly famous example involved
the Avata1?lSaka master Zongmi, who was once entangled in a lawsuit His
lectures in Luoyang excited the attendants so much that one of them cut
off his forearm to express his devotion to Buddhism. The subject of this
occident was called Taigong (? -811 t), whose self-mutilation Zongmi
has graphically depicted and enthusiastically praised in his letter (dated
4 October 811) to Chengguan (738-839), who was then staying in
Chang'an as the "State Master" (guoshi iIDllBrP) of Xianzong (r. 805-820)121.
Zongmi has given such a firm approval to Taigong's self-mutilation
that one might even suspect whether Taigong was privately encouraged
by him. Chengguan's reply does not, however, echo the same degree of
appreciation. Probably apprehensive of any further legal troubles that
Zongmi was to incur, Chengguan asked him not to encourage this kind
of radical act, although he admitted that it does not lack in scriptural sup-
I am of little doubt that at the midst of this emotional episode, the fin-
ger that Fazang set to ftre in front of the Famensi smpa one hundred and
ftfty-two years ago must have been aflame before the eyes of Zongmi,
Taigong and other followers of the A vatrup.ska teachings in the Luoyang
area, and Chengguan as well, who warily watched the development of the
situation afar from Chang'an. They must have also thought of others of
120 These three authors express their opinions on self-immolation in their respective com-
mentaries on the Fanwang jingo See Zhiyi, Pusajie yishu, T 40: 2.576b6-14; UichOk,
Pusa jieben shu, T 40: 2A.675c5-676a14; Sungjang, F anwang jing pusa jieben shuji, X7J
60: 3.135d9-13.
121 "Guifeng Dinghui Chanshi yaobing Qingliang Guoshi shu," T 39: 577b26-28.
122 "Qingliang guoshi huida," T 39: 577c17-21.
their respectable predecessors who did not hesitate to demonstrate their
religious passion with the sacrifice of a certain part of their bodies. Dao-
xuan, for example, records three Avata7f1Saka experts, one master (Puyuan
'iii! [?-560
]) and two his disciples (Puji 'i7!Pt and
[530-609]), who shared their zeal for self-sacrifice
Their friend Iing'ai
(534-578), who was not only close to them personally, but also in
appreciation of the Avatarp,saka teachings, was perhaps one of the most
renowned self-immolator in medieval China. After spreading on a stone-
slab slices of flesh that he cut off from his own body, Iing'ai scooped out
his heart with a knife and, which is more astonishing, died sitting at the
posture of meditation and with his hands holding his heart! 124 In addition
to Puyuan and two of his disciples, Fazang in his collection of Avata7f1.-
saka-related accounts (i.e. Huayan jing zhuanji) also mentions at least
three more Buddhist practitioners, both monks and laymen, who were
Avata7f1.saka admirers and self-immolators as well. First, a eunuch called
Liu Qianzhi (d.u.), an author of a six-hundred fascicle commen-
tary on the Avata7f1.saka sutra, who, though himself not a self-immolator,
was inspired by a very special self-immolator - a Northern Qi dynasty
(550-577) prince who burned himself to death at Mount Wutai out of
desperation derived from his failure to encounter Mafijusn there
ond, (477-522), another Avata7f1.saka commentator (with a
commentary less voluminous, only [I] one hundred fascicles), driven by
his desire to see Mafijusn, had been crawling on the road, wearing a copy
of the Avata7f1.saka sutra on the crown of his head, for a whole year, until
his feet was broken, the blood flowing from his body and the flesh on his
123 See the three monks' biographies at Xu gaoseng zhuan (T 50: 27.680b-c, 680c-
681a, 681a-682b; discussed in Benn 2001: 81-90.
124 See his biography at Xu gaoseng zhuan, T 50: 23.625c-628a; other biographical sources
mentioned in Benn 2001: 223n33. His life, esp. his self-immolation, is extensively stud-
ied in Teiser 1988: 437-439; see also Jan 1965: 252-253. His association with Puyuan's
group is briefly mentioned in Chen Jinhua 2002a: 203n72.
125 Fazang does not tell us whether Liu Qianzhi became a eunuch before or after he wit-
nessed the prince's religious suicide, although the Gu Qingliang zhuan (T 51: 1.1094c14-
21), on which Fazang might have been based, suggests that he had already been castrated
when the prince committed suicide. The possibility exists that LiiI Qianzhi might have
castrated himself as a consequence to the impact that the prince's religious zeal1eft on him.
In the case, he can also be taken as an AvataIp.saka self-immolator.
feet all gone, even completely exposing his kneecaps. The third is Seng-
fan 1i1ftfB (476-555), who burnt a finger (or fingers) as an offering to the
Buddha, when he turned his mind to Buddhism at youth126.
This long list of Avata.rp.saka self-immolators does not necessarily mean
that the Avata.rp.saka tradition produced more self-immolators than other
non-AvataJ1lsaka tradition did (since the documentation of the Buddhist
self-immolators in medieval China was far from being exhaustive). It does,
however, suggest that compared with other traditions the A vataJ1lsaka tra-
dition seems to have been more willing to promote this practice and that
Fazang's attitude to and personal involvement in self-immolation defInitely
played a signifIcant role in affecting how his followers of later generations
approached this practice.
Fazang's attitude toward and involvement in self-immolation contin-
ued and reinforced the self-immolation practiced in the Chinese Avataf[!-
saka tradition. They were, of course, primarily derived from his own
understanding of Buddhism in general and in particular, his fascination
with those paradigms for self-immolators - especially the Buddha
and several Buddhist princes understood to be Sakyamuni
in his former lives - extolled in these Buddhist classics like the Lotus
sutra and the Jiitaka literature. However, we should also consider the
possibility that they might have had something to do with his Sogdian
Following the lead of Egami N arnio tI:.l: is x, scholars have come to
recognize some acts of bodily devotion, such as severing one's ear(s) ,
cutting the face, or even piercing through one's heart and cutting open
one's belly, were part of mourning ceremonies that were executed among
some medieval nomadic tribes living in the Euro-Asian prairies, includ-
ing Fazang's mother country Sogdiana
. They sometime extended this
custom beyond their own cultural spheres. When Taizong died in 649, for
example, people from the "four barbarian regions" (siyi 1ZY:5&) who served
in the Tang court and those barbarian envoys who came to pay tributes
126 These three examples are recorded in Huayan jing zhuanji, T 51: 1.156c18-27,
157b6-16; 2.158b16-19. The cases of Liu Qianzhi and Lingbian are mentioned in Guang
Qingliang zhuan, T 51: 1.1094c.
127 Egarni 1948, Mitani 1984, eai 1998: 24-25.
to the Tang, numbering several hundred, are described as wailing, cutting
off their hair, incising their faces, chopping off their ears and shedding
blood to the ground
Under some particular circumstances, such self-mutilation acts could
also take on different (political or legal) purposes, including those of
protesting, appealing or claiming for innocence to the secular authority.
At the beginning of Ruziong's reign, when Guo Yuanzhen n5G* (?-722),
who was then commanding the Anxi Protectorate (Duhu was
summoned to serve in the court, the chiefs of the tribes under the gover-
nance of the Anxi Protectorate, were said to have cut off their ears and
cut their faces before filing a memorial to the court appealing for Guo
Yuanzhen's being retained as their govemor129. A well-known example
of slicing the abdomen as a radical legal means is provided by a Sogdian
immigrant in Chilla, An Jinzang (before 664-732?), a case which
has been studied for the technique of abdominal suturing in medieval
East Asia
o. An Jirlzang was a son of An Pu (601-664), whose
ancestors were chiefs of the city-state Anguo (Boukhara). An Pu or
his father submitted to the Tang by leaving a Turk tribe and entering
Chang'an during the Zhenguan era (626-649)131. Sometime after January
9, 693 (Changshou 2.1.23 Uiayin])132, An Jinzang served as an attendant
128 Zizhi tongjian 206.6537.
]29 See the biography that Zhang Yue wrote for Guo Yuanzhen, "Bingbu shangshu
Daiguo gong zeng Shaobao Guo gong xingzhuan," 5a-5b.
130 See, for example, Okano 2000. Although in China the application of the technique
of abdominal suturing has been associated with the semi-legendary Hua Tuo (? -208),
this technique, like many other things (including his name!) about this mysterious physi-
cian, was probably of Indian origin. Hua Tuo' s biography in the Sanguo zhi is translated
in DeWoskin 1983: 140-53. For the Indian (Buddhist) origins of some legends about Hua
Tuo (and his name), see Chen Yinque 1992: 36-40; Mair 1993: 331-341. Egarni Namio,
on the other hand, raises the possibility that some medical techniques attributed to Hua Tuo
might have been derived from some magicians (huanren JA) from Central Asia. See
Egami 1965-67: 135-152. These Central Asian techniques could still, however, have been
derived from India. An early Chinese Buddhist self-immolator related with the practice of
abdominal suturing has already been recorded by Huijiao (497-554). See Gaoseng
zhuan 12.404b-c. The case is now capably studied by James Benn (Benn 2001: 45).
13l "Tang gu Luhu zhou Da Anjun muzhi," 1104-1105. Rong 1999: 51; Lei 2003.
132 The Jiu Tang shu here seems to have placed this event to the Zaichu era (Decem-
ber 18, 689-0ctober 15,690). However, according to the Zizhi tongjian (205.6490), which
was based on Xin Tang shu (4.93), this happened sometime after January 9, 693, when sev-
eral of Ruizong's confidants were executed on the ground of visiting him secretly.
of Ruziong in the capacity of taichang gongren (an artisan in
the Court of hnperial Sacrifice). When Ruizong was accused of treason,
Empress Wu ordered Lai Junchen (651-697) to interrogate his
attendants including An Jinzang. Broken by torture, other attendants .were
about to succumb to the false charge, when Jinzang
shouted loudly to [Lai] Iunchen, "If you, master, do not believe my words,
let me cut out my heart in order to show that the heir apparent has no inten-
tion to rebel." He then pulled out the knife that he carried and opened up
his breast [and belly]. As the five internal organs spilled out and his blood
gushed onto the ground, his breath stopped and he fell down. Hearing of this,
[Wu] Zetian ordered him brought into the palace by cart, asking the [impe-
rial] physicians to put his internal organs back into his body. After sewing
close stitches on the wounds with threads manufactured by the root bark of
white mulberry (sangbaipi the physicians applied medicinal oint-
ments to the wounds. [An] Iinzang regained his consciousness in one night.
Zetian visited him in person, sighing, "My own son, who is unable to vin-
dicate himself, is incomparable with this person in loyalty." She thus ordered
[Lai] Iunchen to terminate the prosecution, and Ruizong was thus able to
avoid being hurt because of this. r 1f,
113, l!\E1: .. tB. J
. 133
Not all of these belly-slitting acts were perceived as real acts of self-
immolation. Some Central Asians were believed to perform them as
magic, as described with remarkable vividness by Fazang's contemporary
Zhang (660-733): ..
There are Zoroastrian shrines of the barbarians in Lide Ward
and the
western ward to the west of South Market
Every year, on the occasion
of praying for the [divine] blessings, the barbarian merchants cooked pigs
and goats, played the pipa instruments, drums and flutes, sang to the full and
danced in intoxication. After making offering to the deities, they recruited
one barbarian as the xianzhu *JC.:l:: (Zoroastrian Head?). The onlookers
133 Jiu Tang shu 187A.4885; cf. Xin Tang shu 191.5506.
134 Luoyang had no ward named Lide. There existed two neighboring wards named Lix-
ing:fLrr and Demao Here Zhang Zhuo might refer to these two wards by Lide :fLit!.
135 There were two wards, Fushan m}11ll and Sishun }1!l.}IIll, to the west of South Market
in Luoyang. It is not clear as to which ward Zhang Zhuo is meaning here.
donated their monies, which were to be given to him. The Zoroastrian Head
pulled out a knife, which was as sharp as frost and snow (xiangxue
and which was able to severe the hair that was blew against it - he inserted
such a sharp knife into his belly until the blade pierced through his back.
He further crazily shook the knife inside his body, making the blood shed-
ding out of his bowels and belly. For the space of a single meal, after spray-
ing water on the wound and empowering it with spells, his body was restored
to its original form. This is the magic from the Western Regions. ilJlffff.:lz:1.lii
:ttJ liZ.lfrlJlffi:ttJ, tfjl Jtij 0 '& 1l'li PJHJT:ffli, 2: R ll1t i1&l!$

Il)z-'B/f:@, 0 1JJOOl;ff, Jl1Jli'tt.IfIl.o llJf7.K
0 136
As is revealed by Zhang Zhuo, this belly-slitting show was a magic that
was derived from the Western Regions, which here refers to Central Asia,
including Fazang's original place (Sogdiana). Such magic was performed
not merely for a religious assembly, but also for some secular occasions
like a carnival sponsored by the government:
On February 21,656 (Xianqing l.zheng.20 [bingxu])137, Gaozong ascended
the tower of Gate Anfu to watch the government-sponsored drinking
feast (dapu :;kllm)138. A barbarian proposed that he perform a magic to enter-
tain the people by slitting his belly with a knife. The emperor did not approve
it. A decree was then issued declaring, "It is heard that outside [the palaces]
there are some Brahmin-barbarians who on the occasions of entertainment
often pierce their bellies with swords and cut their tongues with knives,
cheating on people with magic. This very much contravenes the way and
principles [of true government]. It is proper that these people be repatriated
and not be allowed to stay long." Subsequently, the prefectures on the bor-
ders were required not to send this kind of barbarian to the court. ilili *!.m.
:ffiiFiEJjW*, U:;kllmo
Zo 7YHB B, W-JJ%'tl%,
0 139
136 Chaoye qianzai 3.64-65.
137 The original here has the day as bingchen. However, there was no bingchen day in
this month. According to the Xin Tang shu (3.57), the edict prohibiting the magic was issued
on the bingxu (the twentieth) day of this month. I have therefore emended bingchen in the
Cefu yuangui (159.lOb) to bingxu.
138 The expenses of the dapu ceremony were borne by the government, see Schafer 1965.
139 Cefu yuangui 159.10b. Cf. Tang huiyao 34.628, Taiping yulang 737.9a.
Public performance associated with acts' of apparent self-mutilation
was by no means a new thing in Tang China. It could be traced back to
the East Han dynasty (25-220), although it seems that it was not indige-
nous, but imported from the "Western barbarians." These publicly staged
acts of self-mutilation included the performers' (or their assistants') cut-
ting off their tongues, piercing through their ears, slicing their abdomen
and so on. Without any exception, all the mutilated organs are said to
have mysteriously healed shortly afterwards
. Fazang, as a Sogdian immi-
grant, was certainly quite familiar with all these unusual acts, no matter
either attempted as genuine religious self-immolation or simply staged as
a hoax. It is quite likely that his devotion to the self-mutilation was at
partly stimulated by the passion that his compatriots showed to this prac-
tice. A scrutiny of Ch'oe Ch'iwon's account of Fazang's involvement in
the 705 relic veneration might even suggest Fazang's mastery of the belly-
slitting magic, which he employed to manipulate people's emotion:
Before opening the pagoda, a seven-day observance was perfonned. [The
relic emitted] divine rays of light that were shining. Fazang, who once burnt
off a finger here in the past, further destroyed his liver at that time. Hold-
ing a votive text in his hand, he showed it to the religious and lay people
around. Radiating on his palm, the relic projected its illumination from the
near to the far. In accordance with the power of their blissful retribution,
people witnessed different miracles - some seeing the radiant image of the
Buddha made of the most brilliant gold and silver, some watching the
extraordinary vision of the [Buddha-statues embellished with] fringes
The relic, with its jade-like shape and quality, sometime appeared big
and sometime turned small. It measured several chi when it became big and
only several cun when turning small. Therefore, people competed to set
fire to the crown of their heads (dinggang nUl), or bum their fingers (zhiju
They also feared lagging behind in offering donations. 1Tm-t:a:w:,

!Pje!P/J\o jeWl:ticR, IJ\J!X;tJ(i" 0
140 A wide range of Chinese sources on these public performances of self-mutilation
aiming at entertaining the audience can be found in Wu Yugui 2001: 783-786, where he
also covers the instance reported in the Chaoye qianzai. My thanks to Ian Chapman for
referring me to these fascinating materials.
141 This might refer to Genben Shuoyiqieyoubu binaiye zashi, T 24: 10.246c17.
142 Popjang chOn, T 50: 284al-6.
A comparison of this account with that provided by Zhang Zhuo reveals
some remarkable similarities. First, the two occasions were both religious,
one Buddhist and the other Zoroastrian. Second, each consisted in a
grand assembly that seems to have been open to the public, a kind of wuzhe
fahui (paiicavilr:jika) as it was called in Buddhism. Third, both
involved fund-raising: in the case of Fazang, "people feared lagging
behind in offering donations," while in the Zoroastrian assembly described
by Zhang Zhuo, "the onlookers donated their monies, which were to be
given to the Zoroastrian Head." Fourth, both seem to have culminated in
the belly-slitting, which, in the case of Fazang, was depicted by the expres-
sion huigan ("destroying one's liver"). Finally, it is most interest-
ing that in both cases self-mutilation seems to have been employed as
a means to raise money. If the belly-slitting that was performed in the
Zoroastrian assembly was, according to Zhang Zhuo, no more than a
magic trick, then can the same be spoken of Fazang's self-mutilation on
this occasion? This seems highly likely when we consider that Fazang
lived for eight more years after he allegedly "destroyed" his liver. Since
Fazang was believed to have taken out his liver, he was certainly perceived
to have cut off his belly at the time. However, I cannot imagine how
one, under the medical condition in Fazang's time, could have continued
to live for several years after having his liver removed. The only logical
conclusion could be that he here merely performed a magic and that his
self-mutilation was, at least in part, a staged show. In other words, like
some of his Sogdian compatriots, Fazang was also an adroit magician.
Such a newly revealed capacity of Fazang is compatible with his role in
another crucial point of his career, when he availed himself of his magic
skill in helping Empress Wu to overcome the severe crisis posed by the
rebellious Khitans.
In discussing Fazang's political career from the 670s to 71Os, we have
focused on his complicated and oft-misunderstood relationship with Empress
Wu, who projected on his career an influence that can never be exagger-
ated. For their dramatic effects, the legends and stories that featuring
Fazang's brilliant success as a Buddhist expounder and the exceptional
esteem these skills had helped inspire from Empress Wu are very much
celebrated, both m.historical sources and among modem Buddhist schol-
ars, so much so that they have overshadowed the 689 AvataI!lsaka
Dharma-assembly, which has so far remained largely unnoticed by schol-
ars. It turns out, however, that most of those stories/legends are of little
if any historical veracity and that the 689 AvataI!lsaka Dharma-assembly
appears to have been a significant link in a series of deliberate and com-
plicated operations working for the political revolution in the secular
world in the tum of 690. This does not necessarily imply that the empress
lacks historical knowledge and personal fondness of the Avata1!lSaka sutra
- on the contrary, her preface to the new translation of the A vatarJ1saka
sutra demonstrates her impressive knowledge of Buddhist teachings in
general and the sutra in particular.
However, we still have to recognize some political considerations as
the more profound factors driving her to the sutra and its most compe-
tent expounder at the time, Fazang. In addition to what has been pointed
out by Stanley Weinstein (1973: 302), I have elsewhere (Chen 2003)
highlighted several of these factors, including the complicated ideologi-
cal program of turning Mount Wutai into a Buddhist "sacred site" by
identifying it as the abode of Maiijusrl, and the empress's effort to fos-
ter diplomatic ties with the kingdom of Khotan, which was, in tum, an
important link in her policies toward other Central Asian states. In view
of this, Fazang's "international" roles need also to be evaluated in this
highly political and diplomatic regard, and not merely in terms of his s t a ~
tus as the chief founder of the AvatarJ1saka tradition in East Asia.
Fazang's relationship with Empress Wu also proves far more complex,
volatile and even devious than traditional Buddhist historiography has
led us to believe or has been generally understood in modem scholar-
ship. According to an intriguing episode told in a Korean source, Fazang
once fell afoul of another powerful Buddhist monk at the time, Fuli, and
through him, of Empress Wu herself, who was then apparently more under
the influence of that monk. The conflicts were so intense and irreconcil-
able that Fazang was said to have been exiled to the south, although prob-
ably only briefly. We do not know if there were any other more profound
reasons behind this political setback that Fazang suffered sometime
between 690 and 695, or how much his eminence as a Buddhist scholar
contributed to this reverse of his fate. However, it seems certain that
Fazang definitely succeeded in regaining the empress's trust and having
it reinforced through his much desired service in the 696-697 suppression
of the Khitan rebellion, which has been widely recognized as one of the
most crucial points in the empress's eventful life. This brings us to another
little-known way in which Fazang served his patroness and her state.
It so happened that Fazang did not limit himself to serving the empress
in the enterprise of diplomatic pacifism by working towards a new
Chinese translation of the Avataytlsaka sutra and promoting its teachings.
He also actively engaged in undermining and suppressing the "barbarian"
enemies of the empire, not through his philosophical and philological
expertise, but by resorting to his talents in spells and conjuring. Opaque
as it may be, the account of Fazang's role in the 696-697 crisis found in
a Buddhist source (which is, interestingly enough, Korean once again)
not only attests to Fazang's participation in the military endeavors
undertaken by the Great Zhou army, but also suggests that Fazang's role
was highly appreciated by the court. As a matter of fact, his role was per-
ceived to be so decisive that Empress Wu issued an edict to praise him
and about a decade later Zhongzong also fondly recalled and eulogized
his merits in one of his poems dedicated to him. Fazang's role in this
crucial episode in the history of the Great Zhou is also noteworthy for one
particular reason - the deepening and diversification of the A valokites-
vara cult in the years that followed. Fazang's role in this crucial point in
Zhou history is also noteworthy for one particular reason - the expan-
sion of the Avalokitesvara cult in the years that followed. Moreover, it
seems that Fazang's effort to serve the Zhou government in 697 yielded
a result unexpected by anyone (including himself) - that is, it constituted
a decisive factor in the religious and political machinery that was even-
tually to accelerate the pace of an enormous religio-cultural project - the
Yunjusi stone canon.
Not only does it seem ill-founded to assume that Fazang enjoyed sus-
tained favor and support from the empress throughout the whole period
of their association, which lasted for at least three and half decades, but
the long-standing belief among Buddhist scholars that Fazang was a per-
sistently staunch supporter of the empress also seems likewise in doubt.
Evidence shows that he actually worked with some pro-Tang activists in
neutralizing Empress Wu by removing her tWo favorites. Although Fazang
might have "betrayed" his chief patron in this sense, this political move
saved Buddhism from being associated too closely with the Zhou Dynasty
and it also succeeded in stabilizing the current politico-social structure,
which was then jeopardized by Empress Wu' s deteriorating health and her
increasing reliance on her two favorites of questionable personality and
political capacity.
In addition to the political shrewdness he demonstrated in the 705 court
struggles, Fazang's reputation as an effective "trouble-shooter" also
greatly contributed to his continuing success as a religious and political
leader in the last eight years of his life. The two Tang emperors Zhong-
zong and RUizong repeatedly resorted to his esoteric (or shamanic) expert-
ise and his reputed skill in praying for rain and snow whenever. their
country was plagued by drought and other natural disasters. This history
provides us with yet more chances to scrutinize Fazang's image as a won-
derworker. Through three quite typical examples - a large-scale cere-
mony that he supervised in 708 in order to pray for rain, an extraordinary
ritual for snow that he performed at Mount Zhongnan in 711, and his
leadership of a series of relic veneration that lasted from the reign of
Empress Wu to that of Zhongzong -, we are able to recover several
more deeply-hidden layers in Fazang's intellectual and religious life that
have been so far largely lain untouched. They include - but are not lim-
ited to - his promotion of relic veneration, his ideas of and personal
engagement in self-immolation, and at last, quite unexpectedly, his deep
involvement in some Daoist practices, which could be traced back to his
early years as a religious seeker on Mount Zhongnan, a mountain with
time-honored relationship with both Daoism and Buddhism. We are par-
ticularly interested in the ingenious way that he brought religious ele-
ments of different traditions into a highly creative and dynamic combi-
nation, as is most tellingly exemplified in the Esoteric-Daoist ritual that
he performed at the banks of a pond either within or beside the Wuzhensi,
a truly prestigious Buddhist monastery with strong Daoist ties.
In parallel to the Wuzhensi rain-prayer ritual, we should pay particu-
lar attention to the presence and dynamic interactions of various religious
and political concerns in the relic veneration that seems to have persisted
throughout most stages of Fazang's career. One cannot help but feel
amazed at the exceptional skill and subtlety with which sO many diverse
(at times quite incompatible) fibers were woven into the texture of this
series of apparently pious acts. First and foremost, one's attention is drawn
irresistibly towards Fazang's and his followers' passion for self-immola-
tion. Fazang was so enthralled by the Famensi relic that the "sacred bone"
brought out a sense of abhorrence for his own fmger, the burning of which
was subsequently intended as an offering to the former; in the mean-
while, this was also attempted as a catalyst for transforming his physical,
destructible body into a diamond-like one - a personal and direct par-
taking of the dharmaktiya. In this sense, a general remark John Kieschnick
(1997: 44) makes on self-mutilation before relics of the Buddha seems
also applicable to Fazang: it "was not only a sacrifice; it was an appro-
priation. By burning himself, the adept drew on the power of the Bud-
dha's body, purifying his own body and transforming himself into a holy,
living relic."
Fazang's involvement in relic veneration turns out to be far more multi-
dimensional than just his body offerings. Before turning to the multiple
sociopolitical and religious layers so deeply embedded in this series of
relic veneration activities, let me stress one more long-hidden aspect of
Fazang's intellectual life that would never have been exposed to us but
for a brief note that Ch'oe Ch'iwon makes concerning Fazang's per-
formance during the 705 Famensi relic worship. Although it seems his-
torically true that Fazang did burn off a finger in front of the Famensi
pagoda at the age of sixteen, as is affirmed by both Yan Chaoyin and
Ch'oe Ch'iwon, his earliest and best biographer respectively, the same
carmot be said of another far more startling act of self-mutilation that he
was alleged to have committed before the same pagoda almost half a cen-
tury later. According to Ch'oe Ch'iwon, Fazang greeted the Buddha's
fmger-bone, which was then newly exhumed from underneath the Famensi
pagoda, by "destroying his liver." On the basis of the fact that he contin-
ued to live for eight years and that slicing of the abdomen was a common
component of the magical tradition from Central Asia, I have broached
the possibility that on the occasion Fazang simply enacted such a sleight
of hand, without really cutting open his belly and destroying his liver.
Be that as it may, in addition to being an enthusiastic and skillful manipu-
lator of esoteric and shamanic rituals, Fazang was also an adroit showman,
a capacity which fits quite well with his experiences of conjuring up illu-
sionary scenes frightening enough to drive away the Khitan army.
Contrary to the apparently damaging effects that the sacred bone caused
to Fazang and other participants of the Famensi relic veneration -cere-
mony, the same sacred object was sought as the source of therapeutic
power that Fazang attempted to invoke on behalf of his patroness when
she was struggling with her health, and of blessings for personal welfare
(above all, health and longevity) that Fazang's new patron Zhongzong
and other chief members of the imperial family were eager for when the
court politics could make an immediate turn against them. This was, nev-
ertheless, only the beginning of the story.
No matter whether of his own accord or against his will, in the course
of the protracted relic saga, Fazang gradually found himself sitting in
the hot seat of a "triple middleman" functioning at several different
levels. First, on the grandest level envied by all religionists, he was
expected to mediate between the sacred and secular spheres - while
the former exerted its transforming impact on the latter through the
medium of the spirituality of a religious leader like him, the latter, usu-
ally considered too ignorant and worldly to behold or directly get into
contact with the former, had to look up to a religious paragon as its
embodiment. Second, at a lesser level, as this series of relic veneration
was turned into a special form of in which people from
all walks of the society, from the most powerful to the most helpless,
were all invited to participate, Fazang - as its heart and mind - acted
as the mediator between all these members coming from so diverse
social and cultural backgrounds. Eventually, at the most isolated - and
by far the most powerful - level, we come to the innermost part of
Empress Wu's palaces, the mingtang complex (to be specific, its third
story, which was built as and also functioned as a pagoda), at which this
relic series reached its climax. In addition to its therapeutic effects,
Empress Wu also pursued the Famensi relic as a new politico-sacral
symbol around which she wished to rally various interest groups, sev-
eral of which were then - to her intense alarm - starting to spin out
of her control. To re-enshrine the relic in the mingtang was a key step
towards recasting her imperial palace as the center of both the divine
and human realms. Fazang was not only the chief escort of the relic in
the process of this reliquary relocation, but he was also supposed to
sanctify a paradoxical transformation that was inevitably brought to the
sacred bone by this relocation: it was an exaltation in that the relic was
moved from the margin of a local monastery to the very centre ofpolit-
ical power, and simultaneously it was also a fall- from the sacred to
the mundane. Following this successful reliquary transferal, Fazang fur-
ther acted as the guardian of the sacred bone; and more importantly, the
orchestrator of the series of political and religious ceremonies aimed at
mobilizing the broadest possible support for the aging and politically
weakened empress. Thus, as far as this series of relic venerations was
concerned, the primary role that Empress Wu had assigned to Fazang
- at least at the level of court politics - was that of a chief coordi-
nator between different religious and political forces, the representa-
tives of which the empress wished Fazang to attract to this series of
relic-centered ideological maneuverings. Completely unexpected to her
(and probably to Fazang too), he was gradually drawn so close to the
top of the power pyramid that he must have felt almost crushed by the
pressure from two rival political forces, which were rapidly racing
towards a head-on crash. As a result, he had no choice but to side with
one of them and fortunately for him (and unfortunately for his patroness),
he ended up in the right side of the vicious political infighting that led
to the closure of Empress Wu's political life and with it the end of an
unparalleled chapter in the history of imperial China.
Fazang's role as a middleman reminds us of Peter Brown's saint, who,
as an outsider to a social group, is perceived by that group as distant,
unknown and thus mysterious, and able to resolve disputes within the
group and act as a mediator between the group and external entities
(Brown 1971). Fazang appears to have been a typical "saint" in that he
served as a "middle" man in several senses. He was half Han Chinese
(in culture) and half Sogdian (in ethnicity), and both lay and monastic if
we consider the possibility that he was never fully ordained. In the eyes
of his followers and later hagiographers he was even half human and
divine due to his alleged ability to communicate with deities on behalf of
human beings. All these characteristics qualified him as a mediator and
arbiter within a society that was in the grip of intensifying conflicts
between diverse forces of distinct origins.
No matter at which of the above three levels, Fazang's status as an
intermediary seems to have been primarily derived from a kind of unique
power that he possessed - or was perceived to possess. It was a com-
plex combination of spirituality, personal charisma and - let me add
without any intention of doubting his religious sincerity - political
shrewdness in acting between rival forces and a superior capability in
manipulating mobs. Such a mysterious aura of power that surrounded
Fazang contributed to the massive production of various miracle tales
about him, which, in turn, reinforced his image as a wonderworker.
(I) Abbreviations:
HPC Han'guk Pulgyo chOnso (Tongguk taehakkyo pu1chOn kan-
haeng wiwonhoe nae Han' guk Pu1gyo chOnso pyonch' an wiwon [compiled],
1979-1984; see Bibliography Ill).
P Chinese manuscripts from Dunhuang in the Pe1liot Collection, Bibliothe-
que Nationa1e, Paris.
QTS Quan Tang shi (see Bibliography II).
QTW Quan Tang wen (see Bibliography II).
S Chinese manuscripts from Dunhuang in the Stein Collection, British
Library, London.
SKQS Yingyin Wenyuan-ge siku quanshu (see Bibliography
T TaishO shinshii daizokyo (Takakusu, et a1 [comp.], 1924-
1932; see Bibliography Ill).
TMH Tangdai muzhi huibian (Zhou and Zhao [comp.], 1992 [see
Bibliography Ill])
XZJ (Wan) Xuzang jing (rt!) (see Bibliography Ill, Nakano Tatsue OP
et al [compiled], 1905-1912).
(II) Primary Sources
"Bingbu shangshu Daiguo gong zeng Shaobao Guo gong xingzhuan"
by Zhang Yue (667-731) in 722. QTW233.1a-7b.
Bore boluomiduo xinjing liieshu 1 juan, by Fazang !j(
(643-712) in 702. T vol. 33, no. 1712.
Cefu yuangui -IffiM:7Gml, 1,000 juan, compiled by Wang Qinruo (?-1013
between 1005-1013. Beijing: Zhonghua shuju op*efi'5, 1960.
Chaoye qianzai 6 juan, by Zhang Zhuo (660-733); references
made to Cheng and Zhao (colla. & annot.), 1979.
Da Fangguangfo huayan jing ganying zhuan 1 juan,
edited by Hu Youzhen (?-783
) some time after 783 (Jianzhong 4
[February 6, 783-January 26, 784]) on the basis of a two-juan text prepared
by Huiying (?-712t) sometime after 701 (and more likely, after 712).
T vol. 51, no. 2074.
Da Fangguangfo huayanjing suishu yanyi chao 90
juan, by Chengguan (738-839). T vol. 36, no. 1736. .
Da Song Sengshi We 3 juan, by Zanning _* (919?-1001?) in 977.
T no. 2126, vol. 54.
"Da Tang Da lianfusi gu dade Kang Zang fashi zhi bei"
by Yan Chaoyin (?-713?) either in the very end of 712
or at the very beginning of 713. T no. 2054, vol. 50, 280b-c.
"Da Tang gu dade Siheng Hishi muzhiwen" by
Chang Dongming 1t%;g (?-726
) in 726. TMH, 1321-1322.
Da Zhou kanding zhongjing mulu by MJngquan (?-712t)
and others in 695. T no. 2154, vol. 55.
"Da Zhou xinyi Da Fangguangfo Huayan jing xu"
by Wu Zhao (i.e. Empress Wu, 623/625-705) in 699. T no. 297, voL 10,
la24-b2 (QTW 97.5b-7a).
Fanwang jing pusa jieben shuji 4 juan, by Si"ingjang Mfllf
). XZJ no. 565, vol. 60.
Fanwang jing pusajie ben shu 6 juan, by Fazang !. (643-
712). T no. 1813, vol. 40.
F ayuan zhulin 100 juan, by Daoshi @t!t (ca. 596 - after 668) in 668.
T no. 2122, vol. 53.
Fozu tongji 54 juan, compiled by (?-1269
) between 1258
and 1269. T no. 2035, vol. 49.
Gaoseng zhuan i@lif:i111J, 14 juan, initially completed by Huijiao (497-554)
sometime between 519 and 522 (final version probably completed ca. 530).
T no. 2059, vol. 50.
Genben shuoyiqieyoubu binaiye zashi 40 juan, trans-
lated by Yijing (635-713) in 710. Tno. 1451, vol. 24.
"Guifeng Dinghui chanshi yoabing Guoshi shu" iEw!J"Eif*iBjjJmi*:ml'ffi!IDllBjjJ-:- .
Zongmi's two letters (first one written shortly before November 12, 810, and
the second dated November 12, 811). In Zongmi (780-841), Da Fang-
guang yuanjue xiuduoluo liaoyijing liieshu zhu
iOff.a:, T no. 1795, voL 39, 4.57 6c2-577 c8, 577 c25-578a6.
Huayanjing zhuanji 5 juan, by Fazang (643-712) (largely com-
pleted by 690). T no. 2073, vol. 5l.
Ii gujinJodao lunheng 4 juan, compiled by Daoxuan @1[ (596-
667) in 661. T no. 2104, vol. 52.
lin shu Hif, 130 juan, completed under the supervision of Fang Xuanling J%
(578-648) in 648. Beijing: Zhonghua shuju <j:l.if fijJ, 1975.
lin' gang bore jing jiyan ji 3 juan, by Meng Xianzhong\t\
) in 718. XZJ no. 1606, vol. 149.
liu Tang shu if Jlif, 200 juan, completed in 945 under the direction of Liu Xu
(887-946), Beijing; Zhonghua shuju, 1975.
Kaiyuan shijiao lu 20 juan, by Zhisheng (fl. 700-740) in 730 .
. T no. 2154, vol. 55.
Lidai sanbao ji 15 juan, submitted by Fei Zhangfang Jltftl% (?-
) to the court at the very beginning of 598. T no. 2034, vol. 49.
Longxing [fojiao] biannian tonglun 30 juan, by Zuxiu
) in 1164. XZJ no. 1489, vol. 130.
PiJpjang hwasang chOn Tang Tae Ch'iJnboksa kosaju piJn'gyiJng
taediJk PiJpjang hwasang chOn
Pusa jie yishu 2 juan, by Zhiyi (538-597). T no. 1811, vol. 40.
"Qingliang guoshi hui da" dated October 4,811. Chengguan's
reply to Zongmi's letter addressed to him shortly before November 12, 810.
In Zongmi, Da F angguang yuanjue xiuduoluo liaoyi jing lileshu zhu, T no.
1795, vol. 39, 4.577c9-24.
Quan Tang shi 900 juan, compiled by Peng Dingqiu (1645-1719)
and others ca. 1707. Beijing: Zhonghua shuju, 1960.
Quan Tang wen :i:r.!fX (Le. Qinding Quan Tang wen 1,000 juan,
completed in 1814 by Dong Hao :i:Mr (1740-1818) and others. Taibei:
Hualian chubanshe 1965.
Ru Lengqiexin xuanyi 1juan, by Fazang (643-712) some-
time between 704 and 705. Tvol. 39, no. 1790.
Sanbao ganying yaolile lu .::=:. 3 juan, compiled by Feizhuo (?-
1063). T no. 2084, vol. 51.
Sanguo zhi '::=:JP]jit;, 65 juan, by Chen Shou 11*$ (233-97). Beijing: Zhonghua
shuji, 1982 (rpt. of 1959).
Shiji 130juan, by Sima Qian (145-86 BC) ca 90BC. Beijing: Zhong-
hua shuju, 1959.
Shimen guijing yi 2 juan, by Daoxuan (596-667). Tno. 1896,
vol. 45.
"Shinkan Genju hiden shOgo" by Mujaku DochU
(1653-1744) sometime between 1699 and 1744. Tno. 2054, vol. 50. 286c-289c.
Shiyimian Gu(;mshiyin shenzhou jing 1 juan, translated by
Yeshejueduo Jm*JIjIlli$ (Y!iSogupta?) sometime between 557-572. Tno. 1070,
vol. 20.
Shuoyuan 20 juan, by Liu Xiang ItlJ1;] (77BC-6BC). SKQS vol. 696.
SiJk hwaiJm kyobun ki wont'ong ch'o 10 kwiJn, by Kyunyo
tI:tL1l (923-973) on the basis of his delivered on Fazang's Huayan wujiao zhang
lectures from 958 to 962. HPC 4: 239-510.
Song gaoseng zhuan *jiiljiifl{$, 30 juan, by Zanning tt$ (919?-1001?) and others
in 988. T no. 2061, vol. 50.
Taiping yulang 1,000 juan, compiled by Li Fang *B.iJ (925-996) and
others. SKQS vols. 893-901.
Tang Tae Ch'iJnboksa kosaju piJn' gyiJng taediJk PiJpjang hwasang chOn m:k:lIm
1 kwiJn, by Choe Jiweon (857-904t)
in 904. T no. 2054, vol. 50.
"Tang gu Luhu zhou Da Anjun muzhi" by an anonymous
author in 709. TMZ, 1104-1105.
Tang Huiyao 100 juan,. completed by Wang Pu (922-982). Shang-
hai: Shangwu yinshu guan 1935 (3 vols.).
Wanjong mullyu originally 23 lewan, only two lewon (14 and 22) extant.
Compiled by UichOn (1055-1101) in late 11th century. HPC 4: 597-647
(DZ vol. 103).
Xin Tang shu 225 juan, compiled by Ouyang Xiu (1007-1072),
Song Qi (998-1061) and others between 1043 and 1060. Beijing:
Zhonghua shuju 9=t.ii=}lU, 1975.
"Xinyi Dasheng ru Lengqie jing xu" by Wu Zhao
(i.e. Empress Wu, 623/625-705) in 704. QTW97.9b-l0b.
Xu Gujin yijing tuji 1 juan, by Zhisheng (fl. 700-740) in
730. T no. 2152, vol. 55.
Xu Huayan jing liieshu kanding ji 15 juan, by Huiyuan
(673?-743?). XZl no. 194, vol. 5.
"You Wuzhensi.shi" by Bai Juyi (772-846) in 814. QTS
429.4734 (Zhu Jincheng [colla. & annot.], Bai Juyi ji jianjiao, 339-344).
Zan Guanshiyin pusa song 1 juan, translated (and probably also
composed) by Huizhi (fl. 676-703) in Changshou 2 (6 December 690-
25 November 691). Tno. 1052, vol. 20.
Zhenyuan xinding shijiao mulu 30 juan, by Yuanzhao illm
(727-809) between 799 and 800. T no. 2157, vol. 55.
Zizhi tongjian 294 juan, compiled by Sima Guang ifj1.lJ'(; (1019-1086)
and others and presented to the court in 1084. Beijing: Zhonghua shuju,
(Ill) Secondary Sources
Barrett, T.H., 2001. The Rise and Spread of Printing: A New Account of Reli-
gious Factors (SOAS Working Papers in the Study of Religions), London:
,:>London University.
-, 2001a. "Stiipa, siltra and sama in China, c. 656-706 CE." Buddhist Studies
Review 18.1: 1-64.
Benn, Charles D., 1991. The Cavern-mystery Transmission: A Taoist Ordination
Rite of A.D. 711. Honolulu: University of Hawaii Press.
Benn, James, 1998. "Where Texts meet Flesh: burning the body as an apocry-
phal practice in Chinese Buddhism", History of Religions 37 (4):
-, 2001. Burning for the Buddha: Self-immolation in Chinese Buddhism.
lished Ph.D Dissertation, University of California, Los Angeles, Los Angeles.
Brown, Peter Robert Lamont, 1971. "The and Function of the Holy Man in
Late Antiquity." Journal of Roman Studies 61: 81-101.
Cai Hongsheng 1998. Tangdai Jiuxinghu yu tujue wenhua
Beijing: Zhonghua shuju.
Chavannes, Edouard, 1919. "Le jet des Dragons," Memoires concernant /'Asie
Orientale 3: 55-220.
Chen, Jinhua, 2002.' "Sarira and Scepter: Empress Wu's Political Use of Bud-
dhist Relics," 1 ournal of the International Association of Buddhist Studies
25.2: 33-150.
-, 2002a. Monks and Monarchs, Kinship and Kingship: Tanqian in Sui Buddhism
and Politics. Kyoto: Italian School of East Asian Studies.
-,2003. "More Than a Philosopher: Fazang (643-712) as a Politician and Mir-
acle-worker." History of Religion 42.4: 320-358.
-,2004. "Tang Buddhist Palace Chapels," lournal of Chinese Religions 32: 101-173.
-,2004a. "Another Look at Tang Zhongzong's (r. 684, 705-710) Preface to
Yijing's (635-713) Translations: With a Special Reference to Its Date,"
Indo Tetsugaku bukkyogaku kenkyu -1 ;.,- (Studies in Indian
Philosophy and Buddhism, Tokyo University) 11: 3-27.
-, forthcoming. History and His Stories: New Perspectives on the Intellectual
and Religious Life of the Avatal"{lsaka Master Fazang.
-, forthcoming(a). "A Daoist Princess and a Buddhist Temple: A New Theory
on the Causes of the Canon-delivering Mission Commissioned by Princess
Jinxian (689-732) in 740," Bulletin of the School of Oriental and African
Studies 69.1 (February 2006) (forthcoming)
-, In Preparation. Collusion and Collision: Buddhism and Taoism's Politico-eco-
nomic Roles in the Tang Restoration (704-713).
Chen Yinque 1992. "Sanguozhi Cao Chong Hua Tuo zhuan yu fojiao
gushi" In Chen Yinque shixue lunwen xuanji
(Shanghai: Shanghai guji chubanshe
1992), pp. 36-40.
Cheng Yizhong fU9:<):l, Zhao Shouyan Mi'if{l/& (collated and annotated), 1979. Sui
Tang jiahua Chaoye qianzai Beijing: Zhonghua shuju.
DeWoskin, Kenneth J. 1983. Doctors, Diviners, and Magicians of Ancient China:
Biographies of Fang-shih. New York: Columbia University Press.
Egami Namio rrJ:$Jt7c, 1948. "Yiirashia hoppo minzoku no sorei ni okeru limen
setsui senbatsu ni tsuite" ::J.. '7 '7 .,.. 7 0) (.: :b"
0' -c. In Yurashia kodai hoppo bunka: Kyodo bunka ronko::J.. '7 '7
.,.. 7 (Tokyo: Yamakawa shuppansha w)11 tJjjl&
*, 1948), pp. 144-157.
-,1965-67. Ajia bunka-shi kenkyu 7 :l 7 jc1t5:.liJf3i. Tokyo: Yamakawa shup-
pansha w)1/ tJjjl&H (2 vols; vol. 1, Yosetsu hen vol. 2, Ronko hen

Forte, Antonino, 1976. Political Propaganda and Ideology in China at the End
of the Seventh Century: Inquiry into the Nature, Author; and Function of the
Tunhuang Document S. 6502. Followed by an Annotated Translation. Napoli:
Istituto Universitario Orientale, Seminario di Studi Asiatici.
-, 1984. "The Activities in China of the Tantric Master Manicintana (Pao-ssu-
wei" }i!l,'!1E: ?-721 A.D.) from Kashmir and of his Northern Indian Collab-
orators," East and West 34/1-3: 301-347.
-, 1985. "Hui-chih (fl. 676-703 A.D.), a Brahmin Born in China," Annali del-
l'Istituto Orientale di Napoli 45: 105-134.
-, 1988. Mingtang and BudtJhist Utopias in the History of the Astronomical
Clock: The Tower, Statue 'and Armillary Sphere Constructed by Empress Wu.
Roma: Istituto Italiano per il Medio ed Estremo Oriente [Series Orientale
Roma, vol. LIX], and Paris: Ecole d'Extreme-Orient [Publications
de l'Ecole d'Extreme-Orient, vol. CXL V.
-, 1996. "The in Chang'an: Foundation and Name Changes,"
L'inscription nestorienne de Si-ngan-fou: A Posthumous Work by Paul Pel-
liot (ed. Antonino Forte, Kyoto and Paris: Scuola di Studi sull'Asia Orien-
tale and College de France, Institut des Hautes Etudes Chinoises), pp. 429-
472. .
-, 1998. "The Maitreyist Huaiyi (d. 695) and Taoism," Tang yanjiu
(Journal of Tang Studies) 4: 15-29.
-, 1999. "The Maitreyist Huaiyi (d. 695) and Taoism. Additions and Correc-
tions," Tang yanjiu 5: 35-38.
Fujiyoshi Masumi 1997. "Kegonkyo denki no kanta. Hozo to Taigenji"
Q) tJJi1J - c. -, in Kegongaku ronshU .M*ii6U-.
(ed. KamataShigeo hakushikokikinenkai Tokyo:
Daizo shuppan, 1997), pp. 311-333.
Graff, David A., Medieval Chinese Warfare, 300-900. London: Routledge.
Guisso, R[ichard]. W.L., 1978. Wu Tse-t'ien and the Politics of Legitimation in
T' ang China. Western Washington University, Occasional Papers Volume 1 L
-, 1979. "The Reigns of the Empress Wu, Chung-tsung and Jui-tsung (684-712),"
in The Cambridge History of China (vol. 3, Sui and Tang China [589-906],
Part 1) (ed. Denis Twitchett, Cambridge,London, New Yok, Melbourne:
Cambridge University Press, 1979), pp. 290-332.
Han Wei Luo Xizhang Bi2:!i., 1983. "Famensi chutu Tang Zhongzong
xiafa rota ming" Wenwu Jt!lm 6: 14-16.
Hucker, Charles O. (compiled), 1985. A Dictionary of Official Titles in Imperial
China. Stanford: Stanford University Press.
Ishii Kosei ;fi*0JJX;, 2002. "Sokuten Buko Daijo nyii Ryoga kyo jo to Hozo Nyii
Ryoga shin gengi - Zenshii to no kankei ni ryui shite" j{IJ:7i:lEtJ
J c. r - * c. Q) 1HIf* [., -c -, Koma-
zawa daigaku zen kenkyujo nenbO 13.14: 25-44.
Jan, Yun-hua, 1965. "Buddhist Self-immolation in Medieval China," History of
Religions 4: 243-265.
Kamitsuka Yoshiko "Sokuten buko ki no dokyo" in
T6dai no shuky6 (ed. Yoshikawa Tadao E JI L\!J.7c , Kyoto: Hoyii
shoten, 2000), pp. 247-268.
Kieschnick, John, 1997. The Eminent Monk: Buddhist Ideals in Medieval Chinese
Hagiography. Kuroda Institute Studies in East Asian Buddhism 10, Hono-
lulu: University of Hawai'i Press.
Ku Cheng-mei (pinyin: Gu Zhengmei) 1996. "Longmen 1eigu-tai san-
dong de kaizao, xingzhi yu dingnian"
In Longmen shiku yiqianwubai zhounian guoji xueshu taolunhui lunwenji
(ed. Longmen shiku yanjiu-
suo llHHiJilIJiJf'nWf. Beijing: Wenwu chubanshe 1996), pp.
Lau, D.C. (trans.), 1963. Tao-te ching. Harmondsworth: Penguin Books.
- (trans.), 1970. Mencius. Harmondsworth: Penguin Books.
Lei Wen 'i'Pli, 2003. "Geer limian yu cixin poufu - Cong Dunhuang 158 ku
beibi niepan bian wangzijuai tu shuoqi" - .1):$5:1 158
Zhongguo dianji yu wenhua <P
2003 (4): 95-104.
Li Songtao 2003. "Lun Qidan Li Jinzhong Sun Wanrong zhi luan"
Shengtang shidai yu Dongbeiya zhengju :illl!)l!flf(f
(ed. Wang Xiaofu XI} it, Shanghai: Shanghai cishu chu-
banshe -ti4UH!Ft:!:lJl&f), pp. 94-115.
Mair, Victor, 1993. "The Biography of Hua-t'o from the History of the Three
Kingdoms." In The Columbia Anthology of Traditional Chinese Literature
(ed. Victor Mair, New York: Columbia University Press), pp. 331-341.
Mitani Gen 1984. "Nainiku ajia no sh5shin ni kansuru ichi shiron" r*J Ili 7
,; 7 I: T 1.> Shigaku zasshi 93.6: 144-157.
Nakano Tatsue et al (compiled), 1905-1912. Dai Nihon zokuzokyo
120 cases. Kyoto: Z5ky6 shoin
Okano Makoto 2000. "T5 no An Kinz5 no kappuku" )l!f (7) (7) tuJtl[.
Hoshigaku kenkyakai kaibt5 !.'t*litf'nWrWrR 5 (2000): 33-37.
Ono Katsutoshi 'J'lfmf,*", 1989. ChLlgoku Zui To Cht5anjiin shiryo shasei <P1mil
2 vols., Kyoto: H5z5kan.
Palumbo, Antonello, 1997. "On the Author and Date of the Zhenzheng lun
An Obscure Page in the Struggle between Buddhists and Taoists in Medieval
China." Annali dell'Istituto Orientale di Napoli 57.3-4: 305-322.
Robson, James, 2002. Imagining Nanyue: A Religious History of The Southern
Marchmount through the Tang [618-907J. Ph.D. Dissertation, Stanford
Rong Xinjiang 1999. "Beichao Sui Tang Sute ren zhi qianxi jiqi juluo"
Guoxue yanjiu Iill1J*JiJf'n 6: 27-85.
Schafer, Edward H., 1965. "Notes on T'ang Culture II." Monumenta Serica 24:
Sen, Tansen, 2002. Buddhism, Trade, and Diplomacy in the Realignment of Sino-
Indian Relations, 600-1400. (Honolulu: University of Hawaii Press.
Sharf, Robert H., 2001. Coming to Terms with Chinese Buddhism: a Reading of
the Treasure Store Treatise. Honolulu: University of Hawai'i Press.
Shilsho Hozonkai *'I'f*f?:Wr (compiled), 1915-1928. Zoku Jodo-shLl zensho
Tokyo: Shasho hozonkai *'I'f*f?:Wr.
Sun Yinggang 2003. "Chang'an yu Jingzhou zhijian: Tang Zhongzong
yu fojiao" in Tangdai Zongjiao xinyang yu
shehui (ed. Rong Xinjiang Shanghai: Shang-
hai cishu chubanshe, 2003), pp. 125-150.
Takakusu Junjiro Watanabe Kaigyoku et al (compiled),
1924-1932. TaishO shinshu daizokyo 100 vols. Tokyo:
Taisho issaikyo kankokai
Teiser, Stephen, 1988. "Having once Died and Returned to Life: Representation of
Hell in Medieval China," Harvard Journal of Asiatic Studies 48: 433-464.
Tongguk taehakkyo pulchon kanhaeng wiwonhoe nae Han'guk Pulgyo chOnso
pyonch'an wiwon *Wll::k:* &
(compiled), 1979-1984. Han'guk Pulgyo chOnso 6 vols. Seoul:
Tongguk taehakkyo ch'ulp'anbu *Wll::k:*!'i&'i!!l.
Twitchett, Denis C. and Howard J. Wechsler, 1979. "Kao-tsung and the Empress
Wu: The Inheritor and the Usurper," in Cambridge History of China, Vol. 3.1,
(ed. Denis C. Twitchett; Cambridge: Cambridge University Press, 1979),
pp. 242-289.
Wang Bangwei .:E;I'ilKiI (edited and annotated), 1995. Nanhai jigui neifa zhuan
jiaozhu t.ltff. Beijing: Zhonghua shuju.
Wang, Eugene, 2005. Shaping the Lotus Sutra: Buddhist Visual Culture in
Medieval China. Seattle: University of Washington Press.
Weinstein, Stanley, 1973. "Imperial Patronage in T'ang Buddhism," in Perspec-
tives on the T'ang (eds. A. Wright and D. Twitchett; New Heaven: Yale Uni-
versity Press, 1973), pp. 265-306.
Wu Limin !RiLlO\.':; and Han Jinke 1998. Famen digong Tang mi mantuo-
luo zhi yanjiu Hong Kong: Zhonguo fojiao
you-xian gongsi
Wu Yugui !R3sJt, 2001. Zhongguo fengshu tongshi (Sui Tang wudai juan)
Shanghai: Shanghai wenyi chubanshe
:l !'i&f.
Yang Bojun (annotated and translated), 1960. Mengzi yizhu
" Beijing: Zhonghua shuju.
Yen Chuan-ying, 1986. "The Sculpture from the Tower of Seven Je'Yels: The
Style, Patronage Iconography of the Monument." Unpublished Ph.D Dis-
sertation, Harvard University.
Yingyin Wenyuan-ge siku quanshu I1!IJ!I[3::i', 1,500 vols. Tabei: Tai-
wan shangwu yinshuguan 1983-1986.
Yii, Chiin-fang, 2001. Kuan-yin: the Chinese Transformation of Avalokitesvara.
New York: Columbia University Press.
Zhou Shaoliang [;!iJ!.\B et al., compiled, 1992. Tangdai muzhi huibian
Shanghai: Shanghai guji chubanshe J:;4!j:llflli:l!'i&f.
Zhu Qianzhi (collated and annotated), 1984. Laozijiaoshi Bei-
jing: Zhonghua shuju.
It has become common for scholars to interpret the ubiquitous presence
of dharaI)I (tuoluoni ~ t ~ m ) and spells (zhou 1l7L) in medieval Sinitic Bud-
dhism! as evidence of proto-Tantrism in China
For this reason, infor-
mation associated with monk-theurgists and thaumaturges has been organ-
ized in a teleological manner that presupposes the characteristics of a
mature Tantric system and projects them backward over time onto an
earlier period. Recently, however, scholars such as Robert H. Sharf have
begun to point out the limitations of this approach to understanding the
nature of Chinese Buddhism and religion
. This essay will address two
inter-related questions: (1) How did eminent monks in medieval China
conceptualize and spells? And (2) did they conceive of them as
belonging exclusively to some defined tradition (proto-Tantric, Tantric, or
something else)?
In this essay I will present a more nuanced view of the mainstream
Sinitic Buddhist understanding of dharaI)I and spells by providing back-
ground on the role of spell techniques and spell masters in Buddhism
and medieval Chinese religion and by focusing on the way three select
The author of this article wishes to express gratitude to Gregory Schopen, Robert
Buswell, George Keyworth, James Benn, Chen Jinhua, and the anonymous reviewer for
their comments and suggestions on how to improve the article.
1 In this essay I deploy word "dhar<n:lI" following traditional Buddhist convention in
both the singular and plural senses. I also use the word "medieval" rather loosely to refer
to the period extending from the Northern and Southern Dynasties period through the end
of the Tang, roughly 317-907 C.E.
2 In this essay I use the words "proto-Tantric" and "Tantric" instead of the commonly-
deployed but problematic term "Esoteric Buddhism" (mijiao W ~ ) . For problems with the
word mijiao see my essay "Is There Really 'Esoteric' Buddhism?" Journal of the Inter-
national Association of Buddhist Studies 27, no. 2 (2005): 329-356.
3 See, for instance, Robert H. Sharf's essay "On Esoteric Buddhism in China," which
comprises Appendix 1 to his Coming to Terms with Chinese Buddhism: A Reading of the
Treasure Store Treatise (Honolulu: University of Hawai'i Press, 2002), 263-278.
Journal of the International Association of Buddhist Studies
Volume 28 Number 1 2005
intellectuals conceptualized them: Iingying Huiyuan (523-592),
an influential sixth-century scholiast and dharaI)I practitioner; Daoshi
31!ii!: (ca. 596-683), the seventh-century compiler of an important Bud-
dhist encyclopedia; and Amoghavajra (Bukong 705-774), the third
of the three "Tantric" masters of the eighth century. I selected these three
individuals because each one composed an essay on dhararM following dif-
ferent approaches. Huiyuan represents the emerging Chinese Buddhist
intellectual community that mastered Sino-Indian literature,. Daoshi
embodies the mature cominunity in the mid-seventh century that seeks to
demonstrate how Buddhism is Chinese, and Amoghavajra serves as a
spokesperson of the putative "Tantric" perspective. In this essay I will
not attempt to define the terms "dharaI)I" and "spell" but will let the lit-
erature speak: for itself. The literary evidence will demonstrate that dhararM
were not conceptualized as "proto-Tantric" in medieval Sinitic Buddhism.
In fact, to the contrary, defined as "spell techniques" (zhoushu
they were a common component of mainstream Chinese religion.
For much of the twentieth century scholars have debated the nature
and defmition of dhararM and their problematic association with Tantric
Buddhism. There are essentially two ways that researchers have approached
this topic: theoretically and historically. Most scholarship on dharaI)I has
followed the theoretical approach, but this also falls roughly into two
camps: (1) scholars following the work of Etienne Lamotte, who hold that
dharaI)I are actually mnemonic devices or codes for storing or maintain-
ing information
; and (2) those following the writings ofL. Austine Wad-
dell and Guiseppe Tucci, who hold the teleological position that "dhararM
the kernel from which the first Tantras developed."5 Much of
4 See Lamotte, trans., Le traite de la grande vertu de sagesse de Niigiirjuna (Mahiipra-
jfiiipiiramitiisastra), 5 vols. (Louvain: Institut orientaliste, Universite de Louvain, 1944-
1981),4:1854-1869; Jens Braarvig, and Pratibhana: Memory and Eloquence of
the Bodhisattvas," Journal of the International Association of Buddhist Studies 8, no. 1
(1985): 17-29.
5 See L.A. Waddell, "The 'DhiiraI}I' Cult in Buddhism, Its Origin, Deified Literature
and Images," Ostasiatische Zeitschrift 1 (1912-1913): 160-165, 169-178; for some early
translations of dhiiraI}I from Tibetan sources see L. Austine Waddell, "The Dharani or
Indian Buddhist Protective Spell," Indian Antiquary 43 (1914): 37-42,49-54,92-95; and,
for the quote, see Guiseppe Tucci, Tibetan Painted Scrolls, An artistic and symbolic illus-
tration of 172 Tibetan paintings preceded by a survey of the historical, artistic, literary and
religious development of Tibetan culture with an article of P. Pelliot on a Mongol Edict,
the sc:,:holarship dealing with dharaI)I is sectanan in nature. Japanese sec-
tarian scholars of the Shingon school i3 *, for the most part, understand
dharat;U as precursors to their own Tantric system
Although some per-
niciously false sectarian views are now being discarded, many scholars
still hold to the position that the "true" understanding and usage of dhBraIfi
is in the Tantric or "Esoteric" context?
There are a few scholars who, viewing the literary materials and archeo-
logical remains historically, suggest a contrary reading of the evidence.
Gregory Schopen, who deploys a strict definition of Tantric Buddhism, has
demonstrated that some dharat;U actually used in the Indian cultural sphere
should not be classified as "Tantric" because there is nothing Tantric about
. Also, Arthur Waley suggested that dhiiraI)I did not become associated
with Tantric Buddhism until the eighth century and coined the term "DharaI)I
Buddhism" to describe the Buddhism of Dunhuang from the fifth
to the eighth centuries
These scholars, however, represent the minority.
the translation qf historical documents and an appendix on pre-Buddhistic ideas of Tibet,
vol. 1 & vol. 2 (Roma: La Libreria Dello Stato, 1949), 1 :224.
6 See Sharf, "On Esoteric Buddhism in China," 263-278, which contains an overview
of early and important Japanese scholarship; see also, for instance, Takubo Shuyo
fflI'ii', Shingon Daraniz6 no kaisetsu (An Explanation of the Shingon
DhliraI)I Storehouse) (Tokyo: Kanoen 1967); and Ujilce Kakusho Darani
no sekai (The World of Dhlir3l)l) (Osaka: Tohii Shuppan :*:OntBJl&,
7 See, for instance, Abe Ryiiichi, The Weaving of Mantra: Kakai and the Construction
of Esoteric Buddhist Discourse (New York: Columbia University Press, 1999), 152-157,
165-177, 182.
8 Schopen suggests that most dhliraI)I are not Tantric "if by 'Tantric' we mean that phase
of Buddhist doctrinal development which is characterized by an emphasis on the central func-
tion of the guru as religious preceptor; by sets - usually graded - of specific initiations;
by esotericism of doctrine, language and organization; and by a strong emphasis on the real-
ization of the goal through highly structured ritual and meditative techniques. If 'Tantric' is
to be used to refer to something other than this, then the term must be clearly defined and its
boundaries must be clearly drawn. Otherwise the term is meaningless and quite certainly mis-
leading." See Schopen, "The Text of the 'DhliraI)I Stones from Abhayagiriya': A Minor Con-
tribution to the Study of Mahayana Literature in Ceylon," Journal of the International Asso-
ciation of Buddhist Studies 5, no. 1 (1982): 105; see also Schopen,
and DhliraI)Is in Indian Inscriptions: Two Sources for the Practice of Bud-
dhism in Medieval India," Wiener ZeitschriftjUr die Kunde Siidasians 39 (1985): 147.
9 See Waley, A Catalogue of Paintings recovered from Tun-huang by Sir Aurel Stein
(London: Printed by the Order of the Trustees of the British Museum and of the Govern-
ment of India, 1931), xiii-xiv.
For the case of China, mainstream scholarship has alSO' tended toward
the teleological view that dharm:u, spells, and their associated rituals are
proto-Tantric. Based in part on Japanese sectarian scholarship, scholars
have suggested that a Tantric Buddhist "school" was established in China
in the first half of the eighth century through the ministrations of the
"three Tantric masters" - SubhakarasiIpha (Shanwuwei 635-735),
Vajrabodhi (Jin'gangzhi 671-741), and Amoghavajra (Bukong
705-774). However, Tantric Buddhism apparently disappeared as
a distinct "school" in China a little more than a century later. This view
was established in western scholarship by Chou Yi-liang in his ground-
breaking article "Tantrism in China." 10 Michel Strickmann, in some of
his writings, fleshed out this view by emphasizing connections to Daoism,
which he suggests assimilated and preserved Tantric Buddhist elements
and practices
. Other recent studies attempt to account for the supposed
disappearance of Tantric Buddhism in China by demonstrating how
Tantric ideas diffused throughout Chinese Buddhism12.
While these and other works provide much stimulating detail they
tend to ignore the views that prominent Buddhist intellectuals espoused
and promoted concerning dharaI).l and spells in their exegetical works
and in the hagiographical literature written about them. Only a few
works of scholarship have touched on this type of material from this per-
10 See Chou Yi-liang, "Tantrism in China," Harvard Journal of Asiatic Studies 8
(1945): 241-332; Kenneth Ch' en, Buddhism in China: A Historical Survey (Princeton: Prince-
ton.University Press, 1964),325-337.
II See Strickmann, Mantras et mandarins: Ie bouddhisme tantrique en Chine (paris:
Editions Gallimard, 1996),52-53,428 n. 70, 73-74, 120-124; and Chinese Magical Medicine,
ed. Bernard Faure (Stanford: Stanford University Press, 2002); see also Strickmann, "The
Consecration Siltra: A Buddhist Book of Spells," in Chinese Buddhist Apocrypha, ed.
Robert E. Buswell, Jr. (Honolulu: University of Hawai'i Press, 1990), 80-81.
12 See the important and comprehensive work of Lii Jianfu Zhongguo Mijiaoshi
(History of Chinese Tantric Buddhism) (Beijing: Zhongguo Shehui Kexue
Chubanshe, 1995).
13 See Ujike Kakusho, Darani shiso no kenkyil (Research on DharaI)I
Thought) (Osaka: TohO Shuppan 1987); Naomi Gentetsu ii1!#!i:j;;T!T, "Koso-
den no ju" (Spells in the Gaoseng zhuan), Toyo shien!il33 (1989): 32-48;
and John Kieschnick, The Eminent Monk: Buddhist Ideals in Medieval Chinese Hagiography
(Honolulu: A Kuroda Institute Book, University of Hawai'i Press, 1997),82-92.
Spells and Spell Masters in Buddhism and Medieval Chinese Religion
Spells and thaumaturgy were already integral aspects of Chinese reli-
gion long before the introduction of Buddhism to China
. This aspect of
the complex structure of practices, beliefs, and rituals comprising Chinese
religion in Han times (ca. 206 B.C.E.-220 C.E.) and before has been char-
acterized as "the search for personal welfare."l5 Many male and female
shamans, spirit mediums, diviners, and thaumaturges, as well as hermits
and recluses, enjoyed local cult followings due to their skills in using
spells and talismans to control ghosts and illnesses, and in elixirs, medi-
cines, and gymnastic practices for inducing longevity and, so they claimed,
"immortality," from the third century B.C.E. to the third century C.El6.
Many of these thaumaturges were believed to be transcendent beings,
immortals, or sylphs (xian fill, shenxian :flflfill). They were often patronized
by local elites who desired to learn their techniques and some enjoyed
great followingsl7. Both Daoist masters and Buddhist monks competed with
these figures and presented their own spells and practices to prove the effi-
cacy of their respective religious paths; hence, adept monks and bodhisattvas
were popularly conceived of as both miracle workers and sylphs
14 Sawada Mizuho ChUgoku no juht5 (Chinese Spells), rev. ed.
(Tokyo: Heika Shuppansha :lJZfliJtf:lJIlHI::, 1984); Donald Hmper, "Spellbinding," in Religions
of China in Practice, ed. Donald S. Lopez, Jr. (Princeton: Princeton University Press, 1996),
15 Mu-chou Poo, In Search of Personal Welfare: A View of Ancient Chinese Religion
(Albany: SUNY Press, 1998).
16 Rolf A. Stein, "Religious Taoism and Popular Religion from the Second to Seventh
Centuries," in Facets of Taoism: Essays in Chinese Religion, ed. Anna K. Seidel and Holmes
H. Welch (New Haven: Yale University Press, 1979),53-81.
17 Robert Ford Campany, To Live As Long As Heaven and Earth: A Translation and
Study of Ge Hong's Traditions of Divine Transcendents (Berkeley and Los Angeles: Uni-
versity of California Press, 2002), 85-97, and passim.
18 Tsukamoto Zenryii Shina Bukkyoshi kenkyu: Hokugi-hen :
(Studies in Chinese Buddhist History: Northern Wei) (Tokyo: Kobundo, 1942),564-
571,571-581,581-594,605-609; see also Hattori Machihiko "Hokugi Rakuyo
jidai ni miru shinsen shiso" in Dokyo kenkyu ronshU: Dokyo
no shiso to bunka; Yoshioka Hakushi kanreki kinen
(English title: Collected Essays on Taoist Thought and Culture), compo
Yoshioka Yoshitoyo Hakushi Kanreki Kinen Ronshil Kankokai
(Tokyo: Kokusho Kankokai 1977), 193-212; Mu-chou Poo, "The Images
of Immortals and Eminent Monks: Religious Mentality in Early Medieval China," Numen
42 (1995): 172-196.
The supernormal powers traditionally attributed to ordained monks
advanced in meditative cultivation, and more especially associated with
bodhisattvas, placed these figures in both comparison to and competition
with their Chinese counterparts. These powers or "spiritual penetrations"
(shentong :f$ili[), as they became known in China, come in lists of five
or six, and include: the ability to work miracles, supernormal hearing, the
ability to read minds, recollection of one's past lives, the ability to dis-
cern the previous lives of others, and comprehension that one's spiritual
state is no longer plagued by any form of defIlement
One of the earliest
references to, if not the locus classicus of, this term is a short Hfuayana
sutra translated by An Shigao (fl. 148) titled Siitra on the Brah-
mans' Avoiding Death (Poluomen bisi jing which tells
how four brahman monk-sylphs (xianren {lirA, a common translation for
or Indian thaumaturges), cultivated various wholesome dharmas and
the five spiritual penetrations and were able to allay death; thus demon-
strating to the Chinese audience of this sutra that physical immortality is
. Even though the Siitra on the Brahmans' Avoiding Death is a
19 The five spiritual penetrations (Ch. wu shentong litifilm, wutong lilHi, Skt. pafzca-
abhijfziif}) are the 1) divine eye (divyacakeus, tianyan tong 5'(:llNlm), 2) divine ear (divya-
srotra, tianer tong 5'(::EI=lm), 3) knowledge of the thoughts of others (para-citta-jiiiina, taxin
tong ft!!,{,'lm), 4) recollection of former incarnations (pilrvanirviisiinusmrti, suzhu tong
1l1HlHi), 5) "deeds leading to magical power and release" or "direct
experience of magical power (rddhisiikeakriyii, shenjing tong t$:tll:lm). See Apidamo da
piposha lun ([Abhidharma-JMahiivibhiieii) 411, T 1545, 27.728b12-
24; 727b22-24. The six spiritual penetrations (Ch. liu shentong 7\t$lm; Skt.
are 1) psychic power (rddhi-vidhi-jfziina, shenzu tong tifiJElm), magical power; 2) heavenly
ear (divya-srotra-jiiiina, tianer tong 5'(::EI=lm), supernormal hearing; 3) cognition of others'
thoughts (para-citta-jiiiina, taxin tong ft!!,1L.,lm), the ability to read minds; 4) recollection
of past lives (pilrva-nirviisiinusmrti-jfziina, suming tong :@rii'i'lm), 5) heavenly eye (divya-
cakeus-jfziina, tianyan tong the ability to discern the previous lives of others; and
6) cognition of the extinction of outflows (iisrava-keaya-jfziina, loujin tong i,mjl;lm), a state
in which one is no longer plagued by any form of defilement. See Apidamo da piposha
lun 102, T 1545, 27.530a18-blO; and Dazhidu lun 28, T 1509, 25.264a-266b; see also
Etienne Lamotte, trans., Le traite de la grande vertu de sagesse de Niigiirjuna, 4: 1809-1838.
By means of the spiritual penetrations a bodhisattva purifies his buddhakeetra; see Mohe
zhiguan 2a, T 1911, 46.14a-b.
20 Poluomen bisijing, T 131, 2.854b. For more discussion on early Chinese Buddhist
scriptures that demonstrate Daoist and Chinese religious interests see Henri Maspero,
Le taoisme et les religions chinoises, preface de Max Kaltenmark (Paris: Gallimard, 1971),
446; in English see Maspero, Taoism and Chinese Religion, trans. Frank A. Kierman, Jr.
(Amherst: University of Massachusetts Press, 1981), 411.
"Hlnayana" scripture, the powers attributed to monk-adepts became
important characteristics ascribed to Mahayana monks in the Sinitic cul-
tural sphere.
Scholars have long emphasized the role that Buddhist monks such as
the Central Asian thaumaturge Fotudeng (or Fotucheng, d. 348)
played in the conversion of the Chinese to Buddhism. Fotudeng arrived
in North China around 317 when a confederation of Huns, led by the
hegemons Shi Le EJJ (d. 333) and Shi Hu Ere (d. 349) of the Later
Zhao i'&m (319-352), thrust the Jin ff (265-317) out of the Central Plain,
the ancient Chinese heartland. Fotudeng became famous for his ability to
foretell the future and to know the particulars of events taking place hun-
dreds of miles away. He used spell techniques to win Shi Le's support of
Buddhism: he took his begging bowl, fIlled it with water, burned incense,
and chanted a spell over it. In a moment blue lotus flowers sprang up, the
brightness of which dazzled the eyes. Later, Shi Hu had a son named
Bin YJIt, whom Shi Le treated as a foster son. Le loved Bin very dearly,
but Bin was taken ill unexpectedly and died. After two days had passed,
Le called for Fotudeng and charged him with bringing the boy back to life.
The monk enchanted a toothpick by means of a spell. Bin was able to get
up almost immediately and recovered fully after a short time
. Accounts
of marvels performed by monks circulated by word of mouth and even-
tually were amassed in collections of miracle tales. Along with laudatory
information gleaned from stele and stiipa inscriptions, these anecdotes
became the basic source material for the hagiographies contained in the
Lives of Eminent Monks' collections (gaoseng zhuan r,jM1$)22.
After the time of Fotudeng Chinese people became infatuated with
India and Indian Mahayana Buddhism. The Sanskrit spells of Mahayana
21 Gaoseng zhuan fiij{tw 9, T 2059, 50.383b21, c9; 384b24; 385a4, a6, alO, b19; Tang
Yongtong mlflm, Han Wei liang lin Nanbeichao fojiao shi (His-
tory of Buddhism during the Han, Wei, Two Jin, and Southern and Northern Dynasties)
(Shanghai, 1938; rpt. Shanghai: ShanghaiShudian J:lilH!H", 1991), 121-186; ArthurF. Wright,
"Fo-t'u-teng: A Biography," Harvard lournal of Asiatic Studies 11, nos. 3-4 (December,
1948): 321-371; see also Tsukamoto Zenryii, A History of Early Chinese Buddhism, trans.
by Leon Hurvitz, 2 vols. (Tokyo, New York, San Francisco: Kodansha, 1985) 1:257.
22 Koichi Shinohara, "Two Sources of Chinese Buddhist Biographies: Stiipa Inscriptions
and Miracle Stories" in Monks and Magicians: Religious Biographies in Asia, ed. Phyllis
Granoff and Koichi Shinohara (Oakville, Ontario: Mosaic Press, 1988), 119-228.
Buddhist thaumaturges of the fourth and fifth centuries became so pop-
ular that the Daoist Lingbao 1lBf (Numinous Treasure, Spiritual Treasure)
tradition, which flourished in the fifth and sixth centuries, produced a
series of revelations containing incantations in the "Hidden Language of
the Great Brahma." Mimicking the Sanskrit sounds of Buddhist dharaI)I,
these Daoist spells claimed to be celestial language, the secret names of
the gods, by which adepts were able to draw upon the powers of the
Heavens. So attractive was the potent language of the exotic western
lands that fierce competition between Buddhists and Daoists in 'the field
of efficacious spells continued throughout China's great cosmopolitan
age of the Tang (618-907)23. However, this is not the only view pre-
sented in Buddhist literature. One anecdote suggests that Buddhists first
began to use spells in response to harassment by Daoists. The hagiogra-
phy of Tanxian (fl. 504-550), a mysterious monk remembered for
his prowess as a miracle worker, says that Buddhists did not at first learn
thaumaturgy (jangshu 1rvf!j'), but only did so since Daoists (daoshi
chanted spells to pester Buddhist monks - causing their begging bowls
to be thrown into the air and to fall tumbling to the ground and causing
the bridges in a given region to fall to the ground and to stand on end.
Hence, Buddhists were forced to defend themselves by cultivating the
powers of spiritual penetrations

Monks from India and Central Asia were held in high regard and were
esteemed greatly for their knowledge of real Buddhism. Chinese Buddhist
pilgrims, such as Faxian (d. after 423), spent years traveling around
the Indian cultural sphere and recorded many facets of Buddhist belief,
doctrine, and practice so that his fellow monks could institute "real"
Mahayana Buddhism in China
. While these writings are certainly important
23 Stephen R. Bokenkamp, Early Taoist Scriptures (Berkeley, Los Angeles, London:
University of California Press, 1997), 385-389.
24 Xu gaoseng zhuan 23, T 2060, 50.625b5-6, 18.
25 Faxian traveled throughout the Indian cultural sphere from 399-414 C.E. For the bio-
graphy ofFaxian see Gaoseng zhuan 3, T 2059, 50.337b-338b; see also, James Legge, trans.
A Record of Buddhistic Kingdoms (Oxford: Clarendon, 1886; rpt. New York: Dover,
1965), 1-8; and Ch'en, Buddhism in China, 89-91. See also "Dharmasucher" - Reliquien-
Legenden. Der iilteste Bericht eines chinesishen buddhistischen Pilgerm6nchs aber seine
Reise nach Indien: Das Gaoseng-Faxian-zhuan als religionsgeschitliche QueUe (Unter-
suchungen zum Text und Dbersetzung des Textes) (Wiirzburg, 1997; unveri:iffentlichte
documents in any attempt to understand medieval India we must remem-
ber that they were written to be read by an audience fluent in literary
Chinese! As these' books were written by Chinese Buddhists for con-
sumption in the Sinitic cultural sphere they may indeed tell us more about
Chinese interests and concerns than what was really going on in India.
We should also remember that the evidence for Buddhism in India proper
suggests that it was never dominated by the Mahayana; however, the
Mahayana was the Buddhism of choice in many Central Asian oasis towns
and city-states along the Silk Route and in Kashmir. Many of the impor-
tant early Buddhist translators and exegetes in China were from these
areas and, as has been demonstrated by several scholars, crafted their
presentation of Buddhism to Chinese tastes

One such work crafted for a Chinese audience is perhaps the single
most important document for understanding Buddhism in medieval China:
The Treatise on the Great Perfection o/Wisdom (Dazhidu lun * ~ . t ! t ~ ) .
There is nothing in Indian Mahayana literature that remotely approaches
the authority this work enjoyed in medieval Chinese Buddhism. It is a large
compendium of Mahayana views and practices attributed to the monk-
scholar Nagarjuna (Longshu nw, ca. 150-200)27. It was translated into
Chinese between 402 and 406 by Kumarajiva (Iiumoluoshi M!;.*IHt,
344-413), the famous Central Asian translator and explicator of Buddhism
Habilitationschaft; Publikation der aktualisierten Fassung vorgesehen fur das Jahr 2001).
I would like to thank Chen Jinhua for the reference to this recent German scholarship.
26 See Henri Maspero, Le taofsme et les religions chinoises, 277-291, 436-448; see
also Maspero, Taoism and Chinese Religion, 249-262, 400-412; see also Eric Ziircher,
The Buddhist Conquest of China: The Spread and Adoption of Buddhism in Medieval
China, 2nd ed., 2 vols. (Leiden: E.J. Brill, 1972); Tsukamoto, A History of Early Chinese
27 There is a great debate as to whether Nagarjuna actually existed or whether he is a
literary creation concocted by Mahayana writers. This is unimportant to our discussion
because he existed to the Chinese. In India Nagarjuna is referred to variously as the author
of one or another particular essay. However, in China, when a Buddhist exegete says
"Nagarjuna" he is alluding almost invariably to the Dazhidu lun. For the problem of
Nagarjuna's existence and dating in Indian literature see Joseph Walser, "Nagarjuna and
the Ratnavalf: New Ways to Date an Old Philosopher," Journal of the International Asso-
ciation of Buddhist Studies 25, nos. 1-2 (2002): 209-262. On the image of Nagarjuna in
China, see Stuart Young Hawley, "The Dragon Tree, The Middle Way, and the Middle
Kingdom: Images of the Indian Patriarch Nagfujuna in Chinese Buddhism" (M.A. thesis,
School of Oriental and African Studies, University of London, 2000).
to the Chinese and founder of Madhyamaka philosophy in China
recent dissertation of Po-kan Chou presents a strong case for a "partly Chi-
nese" authorship of the work, since the hand of Kumarajlva's editor and
amanuensis Sengrui {till: (352-436) can be seen in the translation and
because some subjects treated by Kumarajlva appear to be responses to
questions by Sengrui and the project's sponsor Yao Xing M l ~ (365-416),
sovereign of the Later Qin 1 & ~ dynasty (384-417)29. It was one of the
most widely read and oft-quoted Buddhist exegetical works from the fifth
through the eighth centuries.
In this text, the writer describes the skills that should be cultivated by
ordained monks. Beyond meditating and strictly observing monastic rules,
a monk develops skills in such varied fields as mixing herbs and medi-
cines, planting cereals and trees, and being accomplished in observing
the stars, the sun and the moon, as well as the movements of clouds and
thunder and lightning. Not only does he fathom the impurities of mun-
dane existence, but he understands portents, such as the speech of animals
and signs of the four cardinal directions. Finally, he is also a student of
all spell techniques (zhoushu), divination practices, charms, and talis-
Furthermore, the writer emphasizes the acquisition of all manner
28 See Etienne Lamotte, trans., Le traire de la grande vertu de sagesse de Niigiirjuna.
On the many different names by which this text was known in medieval China and on the
attribution of the text to Niigarjuna see Paul Demieville's review of the second volume of
Lamotte's translation (originally published in 1950), in Choix d'etudes bouddiques (1929-
1970) (Leiden: E.l. Brill, 1973),470, n. 1,475-476.
29 Some of the most notable evidence provided by Chou is that the Dazhidu fun's com-
mentary on the Mahiiprajfiiipiiramitii-sutra follows Chinese word order rather than Indian
ang. that the whole of the commentary is in the form of a dialogue. Dialogue was not only
commonly employed in Sarviistiviidin commentarial literature, with which KumiiraJlva
was familiar, but also in contemporary xuanxue lr* ("dark learning" or "learning of the
mysterious"). Questions appear to be written into the text and answered as the text pro-
ceeds. Furthermore, Sengrui appears to have written down everything that KumiiraJlva
said and perhaps, due to other concerns, did not edit out old translations of technical terms;
hence, both old and new Buddhist terms remain in the Dazhidu fun. Thus, the Dazhidu lun
seems to reflect the work-in-progress nature of this translation. See Chou Po-kan, "The
Translation of the Dazhidulun: Buddhist Evolution in China in the Early Fifth Century"
(ph.D. dissertation, University of Chicago, 2000), 62, 68, 74-77, 78,80,81-84. I would like
to thank James Benn for referring me to this recent dissertation.
30 Dazhidu lun 3, T 1509, 25.79c-80a; see Lamotte, Le traite de la grande vertu de
sagesse de Niigiirjuna, 1: 199-202.
of dharaI.J.I (listing ten different types) as one of the most prominent qual"
ities of the bodhisattva

Not surprisingly, 'this definition of the traits of a Buddhist monk is very
similar to the traditional qualities associated with brahmans, the chief
competition of Buddhist monks in medieval India; but these traits are
also shared with a lot of other religious specialists, particularly the shamans,
diviners, thaumaturges, and Daoist mages in China. Virtuosity in chanting
spells and working miracles, particularly those associated with healing,
protection, and other aspects of personal welfare, was an important char-
acteristic or quality for a monk to develop. Kumarajlva's hagiography
contains an account that in the year 413, days before his death, Kumara-
jlva chanted a spirit-spell (shenzhou three times in hopes that he
would be healed of his illness. He had his foreign (Indian) disciples chant
on his behalf as well, but it was to no avail. He died a few days later

Yet even this failed attempt at healing by means of spells foreshadows
their popularity and important role in personal welfare in medieval China.
Modem scholars typically classify monks who specialized in such things
as divination, astrology, spells or talismans, as "proto-Tantric" or "Tantric"
- but here, in a mainstream and widely influential
medieval Buddhist text these qualities are presented as quite ordinary and
31 Dazhidu lun 5, T 1509, 25.95c-96c; see Lamotte, Le traite de la grande vertu de
sagesse de Nagarjuna, 1 :316-321.
32 Gaoseng zhuan 2, T 2059, 50.332c25.
33 The idea that modern scholars have labeled monastic practitioners of divination,
astrology, alchemy, spells or talismans as "proto-Tantric" or "Tantric" has a long tradition
in Buddhist scholarship and is indelibly connected to the Japanese Shingon sectarian con-
ceptualizations of "impure" or "miscellaneous" esotericism (zomitsu '\\fEW) and "pure" eso-
tericism (junmitsu Practitioners of Buddhist thaumaturgy from the third through the
early eighth centuries were labeled practitioners of "miscellaneous esotericism" (proto-
Tantric) because they putatively did not really understand the true purpose of "tantric" Bud-
dhism. "Pure esotericism" or real tantrism was then said to have been instituted in China
through the ministration of the "three Tantric masters:" SubhakarasiIpha, Vajrabodhi, and
Arnoghavajra. See Omura Seigai :fc;ftgm, Mikkyo hattaatsushi (Tokyo:
Bukkyo Kankokai Zuzobu, 1918), 1:4, 1:19, 1:21-23, and especially 1 :41-42, and passim
invols. 1 and 2; Benoytosh Bhattacharyya, An Introduction to Buddhist Esoterism (Bom-
bay, 1932; rpt. Benares, 1964),32-42; and Matsunaga YUkei, Mikkyo no rekishi 'iI:.
(History of Esoteric Buddhism) (Tokyo: Heirakuji Shoten, 1969), 13,22-29,29-38,38-
53, 131-154; and also Matsunaga, "Tantric Buddhism and Shingon Buddhism," The East-
ern Buddhist n.s. 2, no. 2 (Nov., 1969); for a brief discussion of the problem see Robert
approved for all monks. It also seems that many medieval Indian and
Chinese monks felt the same way.
in Jingying Mahayana Compendium
The eminent sixth-century Buddhist scholiast and dharaI)I practitioner
Jingying Huiyuan (523-592) analyzed dharaI)I in detail in his
collection of doctrinal exegesis called the Mahayana Compendium
(Dasheng yizhang The treatment of dharaI)I by this exegete
learned in all the major sfitras and treatises of the late sixth century relied
heavily on two mainstream sources of Buddhist doctrine: DharmaIq;ema's
(Tanwuchen 385-433) Chinese translation of the Bodhisattva-
bhUmi, The Stages of the Bodhisattva (Pusa dichi jing trans-
lated ca. 414-421) and the Treatise on the Great Perfection of Wisdom
mentioned previously.
Sharf, Coming to Terms with Chinese Buddhism, 265-267. Although Abe Rytiichi recently
has suggested that scholars need to abandon this sectarian idea (see note 7 above), it has
long since left its mark on and continues to mar and mislead scholarship. Most writings
in English are derivative of sectarian Japanese scholarship, particularly Omura and Mat-
sunaga, mentioned above. See, for instance, Kiyota Minoru, who presents the idea that the
Lotus Satra and the Perfection of Wisdom literature are "miscellaneous tantra" because
they include "incantations" and that Niigiirjuna's deployment of mantras, not to mention
Fotudeng's use of incantations, makes these men tantric practitioners. Furthermore, he
says that this early tantrism incorporated "astronomy, astrology, phrenology, music, art,
and folklore;" see Shingon Buddhism: Theory and Practice (Los Angeles and Tokyo:
Buddhist Books International, 1978),6-7, 13-18. Yamasaki Taikii makes a similar argu-
ment in his Shingon: Japanese Esoteric Buddhism, trans. Richard and Cynthia Peterson, ed.
Ya,suyohi Morimoto and David Kidd (Boston: Shambhala, 1988), 15-17. Though austen-
sibly not dependent upon Japanese scholarship, Chou Yi-liang suggests a similar connec-
tion between tantrism and monks who practice dhiiraIfi and spells; see "Tantrism in China,"
241-248. Instead of scuttling this misleading and completely a-historical discrimination,
Michel Strickmann merely renames "miscellaneous esotericism" (esoterisme eclectique)
as "proto-Tantrism" (prototantrisme) and "pure esotericism" as "Tantric;" see Mantras
et mandarins, 48, 53, 72-79, and passim. This incredibly loose conceptualization of Tantric
practitioners provides the basis for and yet complicates the issue of conceptualizing "Tantri-
cism" and "siddhas" in such recent compilations as Tantra in Practice, ed. David Gor-
don White (Princeton: Princeton University Press, 2000); David Gordon White, The
Alchemical Body: Siddha Traditions in Medieval India (Chicago: University of Chicago
Press, 1996),66-77; and Ronald M. Davidson, Indian Esoteric Buddhism: A Social His-
tory of the Tantric Movement (New York: Columbia University Press, 2002).
As a translator, was instrumental in spreading Buddhism
in Northwestern China during the short-lived Northing Liang regional
regime (397-439) ill the early fIfth-century before it was conquered by the
Northern Wei (386-534) in 439. He was so famous for his knowl-
edge of spells that he acquired the title "Great Spell Master" (dazhou shi
jeq7lgffi). He emphasized the importance of spirit-spells in expelling the
ghosts and goblins responsible for illness and calamity to an unconvinced,
evidence-seeking Juqu Mengxun (368-433), the most impor-
tant hegemon of the Northern Liang. As soon as chanted
. a spell a ghost appeared right in their company, startling the king, but the
monk explained how these ghosts could be expelled by means of spells
chanted by wholesome monks who kept the precepts34. Prior to his death
in 592, Jingying Huiyuan commanded his disciples that twice a day,
before the morning and afternoon lecture assemblies, the entire congre-
gation of monks and others participating in the lecture should chant
the Prajiiaparamita or Perfection of Wisdom spell (bore poluomi zhou
fifty times
. This probably refers to the spell found at the
end of Kumarajlva's translation of the Heart Sutra

Following translation of the Bodhisattva-bhUmi, Hui-
yuan classifIed dharaJ)I in four groups: dharma (fa tuoluoni
meaning (yi tuoluoni spell technique (zhoushu tuoluoni
and acquiescence (ren tuoluoni This taxonomy
of dharaJ)I was employed by several translators and exegetes in medieval
China, such as the eminent translators Bodhiruci (Putiliuzhi
fl. 508-527) and Xuanzang :K* (ca. 600-664)38. Dharma and meaning
dharaJ)I are associated with hearing, completely maintaining, and not for-
getting the Buddhadharma, the Buddhist teaching. These two types of
dharaJ)I may best be thought of as "codes." Spell-technique dharalfI, he
explains, rely on spiritual efficacy (shenyan and are something bod-
hisattvas produce to dispel all adversity. Huiyuan explains that they are
34 Gaoseng zhuan 2, T 2059, 50.335c20, 336a5-7, bl1.
35 Xu gaoseng zhuan 8, T 2060, 50.492c18.
36 Mohe bore poluomi damingzhou jing T 250, 8.847c.
37 Dasheng yizhang 11, T 1851, 44.685a27-28; cf. Pusa dichijing 8, T 1581, 30.934a-b.
38 Shidijing lun 11, T 1522, 26.l91c-192c, translated by Bodhiruci ca. 506; and
Yuga shidi lun ljjtr{J]U+ti!JiiRli 45, T 1579, 30.542c-543b, translated by Xuanzang ca. 646-648.
an important by-product of dhyana or meditative trance, and that their
spiritual efficacy is due to various types of instructions and that there
are many applications. Huiyuan's conceptualization of the role of spiritual
efficacy was a common view held by other Buddhist writers of the sixth
century39. Acquiescence dharaI)l, the final type, allow a bodhisattva to main-
tain his place on the bodhisattva path and abide peacefully or acquiesce
with the true reality of dharmas (i.e. that they are utterly empty of self
nature and are neither produced nor destroyed) - not disappear into the
quiescence of nirvfu:1a - so that they may stay in the world to benefi.t other
beings40. In the Sanskrit version of the Bodhisattva-bhumi, the term that
translated as "spell-technique dharaJ)I" is the compound
It is clear from the context that the writer of the
Bodhisattva-bhumi conceived of "mantra" as a type of dharaI).l. Sinitic
Buddhist intellectuals inherited and maintained this expansive view of
Huiyuan explains three reasons why monks and bodhisattvas are able
to obtain spell-technique dharal)I: (1) they rely on the power of cultiva-
tion and habitual practice in the present, (2) they rely on the efficacy of
dhyana-meditation, and (3) they depend on real knowledge deeply pene-
trating into the approach of the spell-technique dharmas; in other words,
they understand the emptiness and interconnection of all things and the
efficacy of the words of dharaI)l42. Thus, to Huiyuan, the acquisition of
dharaI).l, including magic spells and incantations, is a natural outgrowth
of a bodhisattva's religious cultivation, particularly samadhi or medita-
tive absorption, and is a mark, ornament, or adornment of a bodhisattva's
39 See, for instance, Huijiao's definition of meditation in his "Critical Essay" (lun) on
meditation in the Gaoseng zhaun (T 2059, 50.400b-c), which is discussed in Jan Yun-hua
"Zhonguo zaoqi chanfa de liuchuan he tedian - Huijiao, Daoxuan suozhu 'xichan
pian' yanjiu" cpm!fl-Wlifrll'iMtlmt{lt"m%'lUi - (The transformation
and characteristics of the early Chinese meditation traditions: A study of two treatises on
meditation practices by Huijiao 1iilX [497-554] and Daoxuan ilU[ [596-667]), in his Zhong-
guo chanxue yanjiu lunji (Taibei: Dongchu Chubanshe JlffJ]lIJl&JtL
1990),3-10. I would like to thank Chen Jinhua for this reference.
40 Dasheng yizhang 11, T 1851, 44.685a-b; cf. Pusa dichijing 8, T 1581, 30.934a-b.
41 Bodhisattavabhilmi, ed. U. Wogihara (Tokyo, 1930), 272, 12ft; see also, Franklin
Edgerton, Buddhist Hybrid Sanskrit Grammar and Dictionary (New Haven: Yale Univer-
sity Press, 1953; Rpt. Dehli: Motilal, 1998),2:284 (s.v.
42 Dasheng yizhang 11, T 1851, 44.685b-c.
spiritual attainment. The ability to work wonders is presented as a sign of
a true bodhisattva.
In his study of the tenn "spell" (Jap.julshu 1l5L) as it appears in Huijiao's
(497-554) Lives of Eminent Monks (Gaoseng zhuan com-
pleted ca. 519-554) collection, Naomi Gentetsu isolates the idea that "spir-
itual efficacy" (Jap. reigen/reiken; Ch. lingyan resides in the spells
themselves, but he stresses that monks need "charisma" to access the power
in the spe1l
While his emphasis on the connection between the spiritual
power of monks and of the spells themselves is fundamental, I am hesi-
tant to apply this Weberian tenn. Weber stresses that charisma is an inborn
trait, "a highly individual quality" "that rejects as undignified all method-
ological rational acquisition.'>44 However, Buddhist intellectuals, such as
Jingying Huiyuan, continually stress that dharaI.ll and spells may be
learned, cultivated, and developed as a by-product of meditation and that
their power may be unlocked through a variety of means.
Anecdotal evidence from Chinese Buddhist literature, particularly Daox-
uan's ii '' (596-667) Further Lives of Eminent Monks (Xu gaoseng zhuan
*ir%fiiiff;t, completed in 649 and further revised and edited later), sup-
ports the assertion that the acquisition of dharar:n and proficiency in spells
were important characteristics of eminent monks, particularly monks
remembered as adepts in meditation. In the first quarter of the sixth cen-
tury, for instance, the Northern Indian monk Bodhiruci, mentioned above,
was renowned for his linguistic skills and translation abilities in the serv-
ice of the Northern Wei. He is perhaps most famous in Sinitic Buddhism
as the monk who converted Tanluan (ca. 448-554) to the worship
of Amitabha; but he was also gifted in Buddhist spell techniques. A monk
who happened to see him causing well-water to boil by means of a spell
wanted to offer special reverence to him. Bodhiruci forbade him saying
that all Indian monks cultivate these skills

Chinese monks followed their Indian exemplars quickly in mastering
spell techniques. For instance, the Chinese exegetical monk Sengfan fiiifia
43 Naomi Gentetsu "Kosoden no ju" (Spells in the Gaoseng
zhuan), Toyo shien JltJ"F!il:1li!J33 (1989): 36-37.
44 Max Weber, Economy and Society, trans. and ed. by Guenther Roth and Claus Wittich,
2 vols. (Berkeley: University of California Press, 1968; rpt, 1978),2:1113.
45 Xu gaoseng zhuan 1, T 2060, 50.428c-429a; see also Kieschnick, The Eminent Monk, 87.
(476-555) is remembered as a polymath with a penetratingjntellect. In the
late fifth century, at the yoring age of twenty-three, he is said to have
mastered astronomy, mathematics, and Indian spell-techniques (Tianzhu
zhoushu In the mid-sixth century, Sengchou (480-560),
who is famous for his advanced meditation techniques and as a favorite
of Emperor Xiaowu of the Northern Wei (r. 532-534) and Emperor
Wenxuan :;<:'" of the Northern Qi (r. 550-559), chanted a spell to
reveal to Emperor Wenxuan his former incarnation as a king of evil
demons (luocha *i;fU, Skt. In the early seventh century, a
number of monks renowned for their spiritual abilities used spells to serve
the Sui Emperor Wen I5fi3tW (r. 581-604) and the Sui imperial house. The
Tiantai monk Zbiyue (543-616) was a resident of Guoqing Monastery
on Mount Tiantai :::Rt:rIl.! and is remembered as a specialist in
meditative trance. The first Sui emperor ordered him to the palace in
Chang'an to chant spells and to supervise a vegetarian feast the day of the
death of his beloved empress, Wenxian (nee Dugu 1l1:ml, 553-602)48.
Of course, spell masters were not limited to monks adept in meditation
and academics. Later, the monk-theurgist Faqi (ca. 615), for instance,
healed an illness in the Sui imperial palace by enchanting water by means
of a spell and having everyone in the complex drink it. Faqi was a favorite
of Sui Emperor Wen and for whom the re-unifier of China reportedly
built Xiangtai Monastery in Chang'an

The monk Huibin (574-645) of Hongfu Monastery in
Chang'an was also a monk remembered for his skill in meditation. From
the time that he entered the Buddhist religion, his hagiography remarks, he
was constantly engaged in practice. For the most part he made "spell skills"
(zhouye the heart of his religious cultivation, he chanted the names
"Sakyamuni" (Shijia and "Avalokitesvara" (Guanyin fitf), and he
also practiced a "Mafijusn repentance ritual" (Wenshu
46 Xu gaoseng zhuan 8, T 2060, 50.483b21-22.
47 Xu gaoseng zhuan 16, T 2060, 50.555b7; see also Chen Jinhua, "An Alternative View
of the Meditation Tradition in China: Meditation in the Life and Works of Daoxuan (596-
667)," Toung Pao 88, nos. 4-5 (Dec. 2002): 346-349.
48 Xu gaoseng zhuan 17, T 2060, 50.570c16.
49 Xu gaoseng zhuan 25, T 2060, 50.646a29.
so Xu gaoseng zhuan 20, T 2060, 50.591c8. The compound zhouye appears only twice
in the TaishO canon; and in both cases it refer to sundry spell skills; see Dafaju tuoluoni
In 688, in the preface to his comprehensive biography of Xuanzang, Yan-
cong (d. after 688), a disciple of the famous monk-pilgrim, empha-
sized the practice of meditation, the observance of monastic discipline,
and the employment of dharaI]I or spell techniques (zhoushu) as but three
of the myriads of ways leading to the one goal of dispelling illusion and
benefiting sentient beingsSl. Xuanzang's chanting the Heart Satra and its
spell for protection throughout his famed journey to the Indian kingdoms
is well known
However, what is not well known is that in his biogra-
phy, recorded by his colleague Daoxuan in Further Lives of Eminent
Monks, his translation of the Satra on the Six Approach Spirit Spell (Liu-
men shenzhou jing 1\F51$1I7l*ll!:) is listed among his important works and
translations. This refers to a short dharaI]I text called the Satra on the Six
Approach DhiirmJJ (Liumen tuoluoni jing SalJmukhldhii-
ralJZ) preserved in the Buddhist canon
If there was not a wide interest
in spells and dharaI]I in this period in Sinitic Buddhism and if they were
not important to Xuanzang, why would Daoxuan have bothered to record
it among all the possible choices? Not counting his translation of the
Heart Satra, Xuanzang translated at least eight texts on dharaI]I and spells
that are preserved in the "Esoteric Section" of the Taisho
Among these,
he also translated a spell text on Amoghapasa, the lasso-wielding version
of Avalokitesvara titled The Spirit-Spell Satra of Amoghapasa (Bukong
jing 13, T 1340, 21.718c; Bukongjuansuo tuoluoni zizai wangzhou jing
2, T 1097, 20.424b. With respect to the "Maiijusri repentance
ritual, it may have been derived from the now-lost Wenshu huiguo jing (Sutra
on Repentance of Excesses [taught by] MaiijusrI), also called the Wenshu chanhui jing
(Sutra on Repentance [taught by] Maiijusnj, which was translated by Kumara-
jlva; see Kaiyuan shijiao lu 14, T 2154, 55.636c18-19; Zhenyuan xinding shi-
jiao mulu 24, T 2157, 55.971b17-18; and Yiqie jing yinyi 46,
T 2128, 54.609a-b.
51 Da Tang Daciensi sanzang fashi zhuan T 2053, 50.220c.
52 Aurther Waley, The Real Tripitaka and Other Pieces (London: George Allen and
Unwin, Ltd., 1952), 17, 19, and 98; Ch'en, Buddhism in China, 235; Sally Hovey Wriggins,
Xuanzang: A Buddhist Pilgrim on the Silk Road (Boulder, Colo.: Westview Press, 1996),
119-120; Jan Nattier, "The Heart Sutra: A Chinese Apocryphal Text?" Journal of the Inter-
national Association of Buddhist Studies 15, no. 2 (1992): 153-223.
53 Xu gaoseng zhuan 4, T 2060, 50.455a24; it refers to the Liumen tuoluoni jing, T 1360,
54 See T nos. 1034, 1071, 1094, and 1162 in vol. 20; and T nos. 1360, 1363, 1365, and
1395 in vol. 21.
juansuo shenzhou xin jing - which should not be sur-
prising granted his long interest in the Bodhisattva of Compassion

DharaJ.11 in DaoshPs Buddhist Encyclopedia
In the mid-seventh century, the Chinese Buddhist monk Daoshi
(ca. 596-683), produced a Buddhist "encyclopedia" called A Grove of
Pearls in the Garden of the Dharma (Fayuan zhulin $Jiff*#; in 668)56.
He included two fascicles in this great work of one hundred fascicles that
dealt with the topic of dharru.u. The first of these two rolls begins with an
overview of the topic of dharru.u- he uses the term "spell techniques"
(zhoushu) - that was probably widely held among educated Buddhists
of the day, because the encyclopedia was compiled under imperial patron-
Daoshi viewed dharru.u as an ordinary and important aspect of main-
stream Sinitic Buddhism. The overview of the topic presents the subtle
manner in which the Chinese assimilated dharru.u and understood them in
a non-technical sense.
In summary, Daoshi says that dharaI.ll promote a believing mind in
place of ignorance, they further develop truth into wisdom, they rend
55 Bukong juansuo shenzhouxinjing, T 1094, 20.402b-405c. Xuanzang's work is found
in the middle of the other translations of Amoghapiisa dhiiraJ.llliterature: T nos. 1092-
1098 in vol. 20. Xuanzang's translation was the second; the fIrst was by Jiiiinagupta during
the Sui period, T no. 1093. For an English translation of a Tibetan recension see R.O. Mei-
sezahl, "The Amoghaptisa-hrdaya-dhiiralJI," Monumenta Nipponica 17 (1962): 267-328,
esp. 289-300.
56 For discussion on Daoshi's dates see Chen Jinhua, Monks and Monarchs, Kinship
and Kingship: Tanqian in Sui Buddhism and Politics (Kyoto: Italian School of East Asian
Studies, 2002), 24-25, n. 39. For more on the Fayuan zhulin see Stephen F. Teiser, "T'ang
Buddhist Encyclopedias: An Introduction to Fa-yuan chu-lin and Chu-ching yao-chi,"
T'ang Studies 3 (1985): 109-128; Chen Yuzhen "Daoshi yu Fayuan zhulin"
(Daoshi and the Fayuan zhulin), ZhonghuaJoxue xuebao 5
(July 1992): 233-261; see also Kawaguchi GishO "Hoen shurin ni mirareru
isson bessonkyo ni tsuite" (On the Lost and Vari-
ant Versions of Siltras preserved in the Fayuan zhulin), Nanto Bukkyo 37 (Nov. 1976): 82-
100; and "Kyoroku kenkyil yori mita Hoen shurin": Doshi ni tsuite J:: t) l7-tc.
: i!i:t!tt;::-::>v'-C (Research on Siltras from the Viewpoint of the Fayuan zhulin:
Centering on Daoshi), IBK 24, no. 2 (1976): 974-977. I would like to thank James Benn
for introducing me to the F ayuan zhulin and these secondary sources on this important work
many years ago as we attended graduate school together.
57 Fayuan zhulin 60, T 2122, 53.734c13.
massive obstacles into nothingness, and they extenninate bad karma that
has piled up oflifetimes. Daoshi demonstrates his understand-
ing of dhlirm;li in a technical sense, as well as how it was understood by
Buddhist exegetes, by saying that "dhlirm;li" is, of course, an Indic word,
which, if translated literally into Chinese would be called chi :Rf (to hold,
to support, to maintain). Thus, he says that dhlirm;li "are chanted to hold
on to what is wholesome and not lose it and to hold on to what is unwhole-
some so it will not be produced. "58
With these explanatory doctrinal underpinnings he moves to what is
important to him: that the unseen forces of the natural world may be con-
trolled by means of speaking or chanting particular spells. If the spells are
administered and performed in accordance with their prescribed meth-
ods, one will recognize immediate merit and effects. People will be able
to work miracles, or, in other words as he says "to smash rocks or pluck
out trees, remove illness and eradicate disease." Also, people may control
the spirits, take a ride on dangerous and destructive flood dragons, and
rouse the clouds to open and fertile rains to fall

Following his overview of dhlirm;li, Daoshi catalogs some Of the most
important, well-known, and efficacious spells of the mid-seventh cen-
tury. (1) He begins with a section on dhlirm;li used in repentance rituals.
(2) He then turns to spells used to invoke the power of the Buddha Amitabha
and the bodhisattvas Maitreya and AvalokiteSvara (in that order). (3) He con-
cludes with dhliraI}I chanted to eradicate sins. For instance, he says that
in order to repent you must first set up a ritual space called an
ment site" (Ch. daochang Skt. bodhimal;uJa), cover it with a silk
cloth and place a parasol on top of it. Secure various types of incense and
perfume, close your doors, clean your house, keep visitors and intruders
away, bathe, hold blended perfume in your mouth constantly, and pray
to the buddhas of the ten directions for repentance. If you are not lazy,
he promises, you will have an experience and will cease to doubt6.
DhliraI}I were chanted during repentance rituals long before the seventh
century and were produced by the participants to demonstrate the efficacy
58 Fayuan zhulin 60, T 2122, 53.734cI7-23.
59 Fayuan zhulin 60, T 2122, 53.734c23-28.
60 Fayuan zhulin 60, T 2122, 53.735a6-11.
of their repentance. For instance, Buddhist spells deriving originally from
dharal).l siitras were an integral aspect of repentance rituals developed by
Tiantai Zhiyi (523-597) and he is also reported to have "man-
ifested" dharal).I as a consequence of perfonning a repentance ritual based
on the Lotus Sutra

The final section presents dharal).I chanted for the eradication of sins
(miezui bu It is by far the longest section, comprising more than
six pages in the standard edition of the Buddhist canon, and includes
thirty-six dharal).!-62. All manner of dharal).I and the ritual proceciures for
their efficacious use are included in this section. All of the spells address
the ordinary needs and concerns of the common people and the elite in
medieval Chinese society. For instance, Daoshi includes a spell for women
that causes them to stop menstrual bleeding and other awkward and embar-
rassing bodily functions
; a spell that protects against all manner of
calamities, perversities, defilements, and poisons
; a spell that causes
one to remember what he has heard for a long tirne
; a spell invoking
Avalokitesvara in order to fulfill one's wishes or designs (yuan while
traveling on the road (apparently in order to get there safely)66; a spell
for curing a toothache
; and a spell summoning Avalokitesvara for pro-
teCtion against poisonous snakes
. There is even a spell for protection
against all manner of leprous diseases and exposed wounds, which was
also taught by A valokitesvara. The instruction, included by Daoshi, for
people who would use this spell technique is to chant the associated
61 For spells in Zhiyi's repentance rituals see Fangdeng sanmei xingfa :1:N!'=:IJ;f<i''T1*
(The Method of Vaipulya Sam1idhi), T 1940, 46, 943c-944a. The spell procedures out-
lined in the foregoing text are based on the Dafangdeng tuoluoni jing (The
Great Vaipulya 4, T 1339, 46.656a-661a, a sutra translated by the sramaJ).a
Fazhong 1*'* of the Northern Liang regime. I would like to thank the reviewer for remind-
ing me .of this connection. For his manifestation of see Sui Tiantai Zhizhe. dashi
biezhuan T 2050, 50.192a6-7; see also Linda Penkower, "In the
Beginning ... Guanding rim (561-632) and the Creation of Early Tiantai," Journal of the
International Association of Buddhist Studies 23, no. 2 (2000): 261.
62 Fayuan zhulin 60, T 2122, 53.737cll-743c25.
63 Fayuan zhulin 60, T 2122, 53.741b8-17.
64 Fayuan zhulin 60, T 2122, 53.741b29-c9.
65 Fayuan zhulin 60, T 2122, 53.741c19-24.
66 Fayuan zhulin 60, T 2122, 53.741c28-742a5.
67 Fayuan zhulin 60, T 2122, 53.742a22-bl.
68 Fayuan zhulin 60, T 2122, 53.742b23-27.
dhara and repent with all their heart and they will be healed immedi-
ately. If there are 9pen wounds, they are instructed to enchant some dirt
or mud with this spell and to place the mud on the wound and it will be
healed immediately69. The ritual prescriptions for these dhara suggest
that their efficacy is due to two types of factors: first, the inherent magic
power of the words themselves, and, second, the power or merits of gods
and bodhisattvas; yet in both cases the desired-for result is obtained
through faith, sincerity, and correct ritual application. From this standpoint
we can understand why dhara were popular among the Buddhist faith-
ful of both the commoner and cultured elite social classes of medieval
China. More importantly, due to Daoshi's insights, we can contextualize
accurately curious anecdotes about dhara and spells in medieval Chinese
society. For instance, Liang Emperor Yuan (r. 552-554), son of the
fabled Buddhist Emperor Wu (r. 502-549), wrote that he had
memorized several Buddhist spells during his childhood
The entire second roll of Daoshi's work is devoted to retelling stories
about thaumaturges and ordinary people who perfonned miracles by chant-
ing spells. In essence, it emphasizes that dhara are just like Chinese
spells (zhou). Daoshi culled these proofs from a panoply of materials
ranging from collections of miracle tales, to Buddhist hagiography, to
official dynastic histories. But what is most important is that he makes
no distinction between Buddhist and non-Buddhist spells, which suggests
that the Chinese (at least Daoshi) did not perceive of dhara as a com-
pletely foreign commodity. There is nothing inherently unique about his
deploying anecdotes to support his views because Daoshi' s modus
operandi in the encyclopedia is to present doctrinal passages first and
then present examples (ganying "stimulus and response" or "reso-
nance ") that demonstrate the validity of his doctrinal explanation. Of the
twenty-two stories presented in this section only six are Buddhist from
Buddhist sources.
69 Fayuan zhulin 60, T 2122, 53.742c25-743a6.
70 For emperor Liang Yuandi's (504-554) statement on his childhood interest in dharaI)1
see linlilzi :lil:*I'T (var. linlouzi :lil:1l'T) 6:24a5-b2; in Yingyin Wenyuange siku quanshu
(Photofacsimile reprint of the Wenyuan Pavilion Copy of the Siku
quanshu), 1,500 vols. (Taipei: Shangwu, 1983-1986), 880:800; cf. Chou Yi-liang, "Tan-
trism in China," 244.
For instance, Daoshi selects tales from the Baopuzi :fEj;f+=r and the
Liezi JUT, demonstrating his willingness to utilize materials that in his
time had become associated with Daoism, which was supported and pro-
moted as the imperial cult of the Tang dynasty7!. The anecdote he includes
from the Baopuzi is about General He of the ancient state of Wu
!!i2:, who is sent to deal with mountain marauders. However, one of the
thieves is adept in protective measures (spells) against weaponry and the
rules and procedures of their efficacy. General He knows that if he and
his men are armed sharp swords the bad guys will be able' to work
spells against them. So they get rid of all their weapons and outwit the
marauders 72. The account from the Liezi is about King Mu of Zhou
(r. 1001-946 B.C.E.), who is visited by a magician (huanren xJJ-J from
the extreme western countries (xiji E"ti). The thaumaturge is able to live
underwater, penetrate metal and rock, overturn mountains, move cities,
ride in the sky, strike hard and sharp objects and not be hurt, and so forth.
King Mu treats him as a spiritual being worthy of veneration
When we compare the flavor of the doctrinal account of Buddhist spells
presented in the first fascicle with the examples presented in the second
fascicle, Daoshi seems to suggest that the efficacious use of dharaIJI can
become an ordinary aspect of one's religious practice. Buddhist spells
are not seen as foreign so much as extremely beneficial to one's personal
welfare. The exotic Sanskrit-like sounds must have also been a factor in
their popUlarity. Perhaps most importantly, Daoshi is able to demonstrate
a long history of spell-chanting and miracle-working in China that
71 At the time of their composition, though, the Baopuzi and Liezi were probably not
connected with the then-existing religious Daoist tradition of the Celestial Masters (Tianshi
;Rgifj). The author of the Baopuzi Ge Hong's (283-343) knowledge of this tradition
is sketchy and the next religious Daoist tradition, the Supreme Purity (Shangqing ...trw) tra-
dition, the revelations of which were first recorded in 364-370, had not yet been "revealed."
The southern literatus Ge Hong represents a tradition of Chinese religious practitioners of
various techniques, alchemy in particular, who sought to become divine transcendents
(xian fill, shenxian j:$fill). His writings had an impact on the development of the Supreme
Purity Daoist tradition. See Isabelle Robinet, Histoire du taoisme: des origins au XIV' siele
(paris: Les Editions du Cerf, 1991), 85-117; in English, Isabelle Robinet, Taoism: The Growth
of a Religion (Stanford: Stanford University Press, 1997),78-113; Campany, To Live As
Long As Heaven And Earth, 18-97; Stephen R. Bokenkamp, Early Daoist Scriptures
(Berkeley, Los Angeles, London: University of California Press, 1997),7.
72 Fayuan zhulin 61, T 2122, 53.748c24-749a3.
73 Fayuan zhulin 61, T 2122, 53.749c15-18.
embraces and subsumes, while still paralleling, the Buddhist cult of dha-
raIfi. Perhaps better than any other source, this shows the assimilation of
pre-Buddhist Chinese practices into Buddhism and the integration of Bud-
dhism into mainstream Chinese religion.
Amoghavajra's Imperially-Decreed Definition of DharaI;li
Zanning's W$ (919-1001) Lives of Eminent Monks compiled in the
Song (Song gaoseng zhuan *r'aJiw1$, compiled 982-988 and further edited
and revised) records that during the eighth century, the Indian monks
SubhakarasiIpha, Vajrabodhi, and Amoghavajra enjoyed renown for their
ability to make rain and to defeat other magicians and astrologers in dis-
plays of magic power by means of their Sanskrit spells
Since the effi-
cacy of these monks' spells seems to play a prominent role in their
hagiographies scholars have presumed that spells were somehow indeli-
bly connected to Tantrism. However, as we have seen above, Buddhist
monks and lay people conceived of dharaIfi and spells as a common part
of their practice and encouraged lay people to learn and use them as well.
In the second half of the eighth century, the third of the three Indian
Tantric masters in China, Amoghavajra, was ordered by the Tang emperor,
probably Daizong {-\:';* (r. 762-779), to provide an official definition of
dhara!fi, in conjunction with other terms associated with the magical pow-
ers attributed to advanced monks and all bodhisattvas. When completed
it was called the Encomia on a General Interpretation of the Meaning of
DhiiralJl (Zongshi tuoluoni yizan This short essay not
only attempts to define dhara!fi, but relates and equates dhara!fi to three
other classes of terms: "true words" (zhenyan and "esoteric words"
(miyan 2&;) - which are both translations of "mantra"- and "clari-
ties" (ming Amoghavajra's definition of dhara!fi shares
much with that of his predecessors and yet, most surprisingly, shows a
conscious desire to de-emphasize the role of spells. He follows intellectual
convention by placing dhara!fi in four classes: dharma maintenance (jachi
$), meaning maintenance (yichi samadhi maintenance (sanmodi
74 Song gaoseng zhuan 1, T 2061, 50.711c, 713c; and fasc. 2, T 20601.50.715c. See
also Chou Yi-liang, "Tantrism in China," 268, 274, 291-292.
chi and word or pattern maintenance (wenchi Like
Jingying Huiyuan, Amoghavajra emphasizes the role of samadhi in the
development of the five spiritual penetrations, the supernatural powers
of the bodhisattva
. Most importantly, however, the word "spell," zhou,
or any term with that character, is entirely missing from his discussion.
This exegesis on dhlirru;li was probably composed in the later years of
Amoghavajra's life, some time in between 762 and 774. Since it provides
an explanation of many terms associated with dhlirru;li from a
"Tantric" standpoint it should be a crucial piece of evidence for under-
standing how the early Tantric Buddhist masters differentiated their doc-
trines and practices, if at all, from the preexisting Mahayana tradition.
Twice in this short essay Amoghavajra says that dhlirru;li and all related
terms - including zhenyan mantras - are explained in the "exoteric
teachings" (xianjiao MfJt). In the first case he says that they have been
"explained in the exoteric teachings in the Mahayana teachings." In the
second case, found in the final paragraph of the encomia; he says that they
have been "explained also in the exoteric teachings. "77 What does Amo-
ghavajra mean by "exoteric teachings" here? My research on the meaning
and usage of the concepts of "esoteric" and "exoteric" in medieval Sinitic
Buddhist exegesis suggests that both terms are deployed polemically by
scholiasts. "Esoteric" refers to what the writer holds to be a superior
teaching; it is often interchangeable with "the Mahayana," and in par-
ticular is linked to the concept of "acquiescence to the non-production of
dharmas." "Exoteric" refers to ordinary Buddhist teachings, and the teach-
ings of the "HInayana" siitras as well as some Mahayana siitras, such
as the Perfection of Wisdom (Prajfiiipiiramitii) literature
. Thus, what
Anioghavajra means by dhlirru;li having been explained in the "exoteric
teachings" is that many kinds of dhlirru;li are contained in earlier Buddhist
literature, such as Abhidharma literature
, but they are not the most
7S Zongshi tuoluoni yizan, T 902, IS.S9Sall-12.
76 Zongshi tuoluoni yizan, T 902, IS.S9SaI9-20.
77 Zongshi tuoluoni yizan, T 902, IS.S9Sa13, b21-22.
78 See my article "Is There Really 'Esoteric' Buddhism?"
79 See, for instance, Apidamo da piposha lun 25, T 1545, 27.130a, which says that
spells (zhou) able to cure people of illnesses are called clarities (ming, Skt. vidyii), and spells
able to cure people of the sickness of all defilements are also called clarities; and fasc. 102,
T 1545, 27.529b-c, which speaks of all manner of spells (zhongzhong zhou fmfmll5/.): the
imp<;>rtant dharat.JI in his estimation. In other words, he is attempting to dif-
ferentiate between .the dharaI)I found throughout earlier Buddhist literature,
such as the Perfection of Wisdom literature and the dharaI)I rituals described
by Daoshi, and the dharaI)I in the writings he views as superior.
In his concluding remarks, Amoghavajra says that the "esoteric teach-
ing" (mijiao - read as "the Mahayana" - has many types of its
own "true words," referring to mantra, that may be referred to using the
"four designations. "so The context suggests that he means the four terms
he just defined: dhararp:, true words, esoteric words, and clarities as found
in Mahayana literature. Amoghavajra's defInitions of these terms are
unusually uniform to the point that one could consider each one to be
almost identical to the others. That is not surprising for the case of true
words and esoteric words, but less so for dhararp:, which, as we have seen,
was described differently and with greater precision by Jingying Huiyuan
and as virtually interchangeable with "spells" by Daoshi. The signifIcant
thing is that he inverts the relationship between dhararp: and mantra pre-
sumed by the Chinese intellectual tradition. As we have seen above, spell
techniques (zhoushu, viz. mantra) were conceptualized traditionally as a
type of dharaI)I. However, Amoghavajra says that all these terms are just
types of true words (zhenyan). This is amplifted by his stating that they
may be both incredibly short (one syllable) or incredibly long (ten thou-
sand syllables)Sl. This is fundamentally different from the way in which
scholars usually attempt to define terms such as dhararp:, mantra, and
vidya using concise and preCise language. In support of Amoghavajra's
loose defInition of mantra, a Tang-period translation of a ritual text by a
certain Putixian ifmfw d.u.) deploys the terms "true word"
(zhenyan), "esoteric word" (miyan), and "clarity" (ming) interchangeably
for mantras of various lengthss2. Another interesting aspect of this exegesis
peacock clarity (kongque ming dragon and snake clarity (longshe ming
image-hooking clarity (xianggou ming fire clarity (huo ming :kaj]), water clarity
(shui ming asterism claritY (xing ming and the bird clarity (niao ming
80 Zongshi tuoluoni yizan, T 902, 18.898b22-23.
81 Zongshi tuoluoni yizan, T 902, 18.898b23-24.
82 For zhenyan see Da shengmiao jixiang pusa mimi bazi aluoni xiuxin mandaluo cidi
yiguifa T 1184, 20.786b-790b; for
miyan see 788b, 789a-c, and 790b; and for ming: hrdaya-vidya (xinmiming see 786cl,
and "esoteric pearl clarity" (mimi zhuming see 790c.
is that all of Amoghavajra's definitions in one way or another make room
for mantras of the seed-syllable (bfja) variety, the one-syllable true words.
These had been introduced previously as "esoteric speech" (miyu
in. a retranslation of a Perfection of Wisdom sutra by Vajrabodhi83, although
this expression was not included in Amoghavajra's definition of terms.
Amoghavajra's expansive definition of dharrup: as true words resonates
with Subhakarasilpha's inclusive taxonomy of zhenyan in his Commentary
on the Satra on Mahiivairocana's Attaining Buddhahood (Da J>.iluzhena
chengfo jing shu whose narration was "recorded"
by his disciple/colleague Yixing -rr (673-727). Subhakarasilpha explains
that there are five kinds of zhenyan: (1) those explained by Tathagatas
(rulai shuo (2) those explained by bodhisattvas and vajra[sattvas]
(pusa jin 'gang shuo (3) those explained by [adherents of
the] Two Vehicles, (ersheng shuo i.e. arhats and pratyekabuddhas;
(4) those explained by all the deities (zhutian shuo and (5) those
explained by earth-dwelling deities (dijutian shuo t-r:gg7::), such as dragons
(niigas), birds, and asuras (titans)84. The important point I am trying to
make here is that although Amoghavajra's exegesis is hypothetically about
dhararp:, the point he cleverly emphasizes at the end is that they are really
just zhenyan (true words). This is a break with the earlier tradition, and it
appears that it was something not easily accepted or understood in China.
We know that the Chinese did not differentiate between dharaJ:.ll and
mantra in the earlier tradition and this is reflected in the use of dharrup:,
spell, and true word interchangeably in translations as late as the early
eighth century85. Although externally Amoghavajra's exegesis seems to be
an attempt to clarify the confused application of the terms, it is a funda-
mentally polemical document that cleverly privileges the notion of zhenyan
from what we might call in retrospect a "Tantric point of view." Perhaps
this also gives us some insight into a reason why the Tantric masters
retranslated Mahayana sutras including revised or updated dharrup:86. Were
83 lin' gangding yuga liqu bore ling T 241, 8. 779a-781 b.
84 Da Piluzhena chengfo ling shu 7, T 1796, 39.649a.
85 Takubo Shilyo E8;x'{Jjidi!iJ'Ilr, Shingon Daranizo no kaisetsu (An
Explanation of the Shingon Dhararfi Storehouse) (Tokyo: Kanoen Jre!ll.f51I, 1967), 29, 36-37.
86 For example, when Amoghavajra retranslated the Satra of Benevolent Kings (Renwang
ling) in the eighth century he added in a dharllI).l that does not appear in the fifth-century
they attempting to appropriate them, adding new zhenyan components to
them? Since they. did not see their message fundamentally as different to
the Mahayana teaching I do not think this was part of some hidden agenda
to promote an "Esoteric Buddhism" by replacing the preexisting tradition.
But the question remains: what is different about dharal)I in this putatively
"Tantric" point of view?
Amoghavajra provides a crucial clue to understanding what is differ-
ent when he says that all these dhara "mutually resonate with the
approach of the three esoterica" (sanmimen xiangying
What I call the "approach of the three esoterica" (sanmi men) - some-
times called the "three mysterious gates" or "three mysteries" (Jap. san-
mitsu Skt. tri-guhya) - in this case probably refers to the replica-
tion of the body, speech, and mind of the Dharmakaya Buddha
. Although
the exact terminology is not used, Abe Ryfiichi suggests that this idea
may be found in the Sutra on Mahavairocana's Attaining Buddhahood,
in which the Buddha Vairocana displays his enlightenment in language
encouraging ritual replication "in the gestural sequences of mudras, the
chanting of mantras, and the visualization of mru;t<;lala images. "89 Perhaps
more importantly, these ideas were also expressed in Yogaciira literature
translated and introduced into Sinitic Buddhism by Xuanzang during the
seventh century, which may suggest why what academe calls "Tantric
Buddhism" is thoroughly intermixed with what was called the teachings
of the "Yoga" school (yuga fM11JO) in China during the Song * period
text. See Renwang huguo banruo boluomiduo jing 2, T 246,
843cI9-844a9; cf. Charles D. Orzech, Politics and Transcendent Wisdom: The Scripture
for Humane Kings in the Creation of Chinese Buddhism (University Park, Pa.: Pennsyl-
vania State University Press, 1998), 269 n. 76.
87 Zongshi tuoluoni yizan, T 902, l8.898b25.
88 The tenn sanmi appears to have been interpreted differently in different exegetical
traditions, such as between Tendai (Tairnitsu) and Shingon (Tomitsu) esotericism, but a
discussion of this is beyond the scope of the paper. See Kubota Tetsumasa flEBW:iE,
"Nihon Tendai no yuso sanmitsu hOben setsu" 1'1 It Q 'l!f=.JifljJ1j!m (On argu-
ments about the meaning of sanmitsu [Three Mysteries] in the Japanese Tendai Sect),
Nihon Bukkyo gakkai nenpo 1'1 (May, 1992): 145-162; and Okubo Ryo-
shun "Taimitsu no sanmitsu ron" ;tJiflO)='Jiflfnii (The Three Mysteries Theory
of Taimitsu), Tendai gakuhO (Sep., 1992): 109-113; and Otsuka Nobuo
{1jI*" "Sanmitsu shiso nit suite" (On the concept of the tri-guhya), IBK
34, no; 1 (Dec. 1995): 174-176.
89 Da Piluozhena chengfo jing 1, T 848, 18.4a-5a; see Abe, The Weaving of Mantra, 129.
(960-1279)90. This kind of specialized use of mantras by a practitioner
under the guidance of a guru inside a maI).q.ala as part of a ritual meant
to replicate the body, speech, and mind of the Buddha for the purpose of
making him a Buddha immediately is what is actually different about
Tantric Buddhist practice! Although this procedure is described in a
"sutra" and not a "tantra," it resonates with Schopen's more precise and
useful definition of "Tantric Buddhism" alluded to above. Even though
all the components for such a ritual practice (maI).q.ala, mudra, mantra, and
gurus) had existed for a long time before in Mahayana Buddhism they had
not been constructed in such a concise manner for a specific religious
purpose. This seems to be the real break with the preexisting tradition, at
least for the case of China.
Furthermore, since the Tantric masters did not claim that their message
was anything other than that of the Mahayana, it should not be a surprise
that they possessed dharaI).I and were adept in the types of ritual practice
that were common and widespread among Mahayana adherents in the
areas where the Mahayana tradition held sway. In other words, the fact
that Amoghavajra could work miracles or foretell the future by means
of spells does not indicate an inherent connection between spells, thau-
maturgy, divination and Tantric Buddhism. Rather, his success with
dharaI).I and spells provides evidence of the efficacy of his path of prac-
tice in attaining quickly the magic powers associated with enlightenment.
The thaumaturgic skills possessed by the Indian Tantric masters in China,
as we have seen, had long been heralded as the qualities of advanced
monks and bodhisattvas in mainstream Mahayana Buddhism in medieval
China, not to mention religious adepts outside of Buddhism. The Tantric
masters differed in the promotion of special ritual to reproduce the body,
speech, and mind of the Buddha in order to attain Buddhahood quickly.
Hence, I would suggest that Buddhist spells, dharaIji or mantra, used in
that confined context should be seen as "Tantric." Outside of that spe-
cial ritual milieu, dharaI}.l and spells used in Buddhist thaumaturgy,
divination, merit-making, healing practices and repentance rituals,
and the invocation of buddhas and bodhisattvas for the destruction of
unwholesome karma, as well as in astrology and other occult sciences,
90 See Lii Jianfu, Zhongguo Mijiaoshi, 432-513.
were common features of Chinese Buddhism specifically and Chinese
religion in general.
Concluding Remarks
In this essay I have emphasized hitherto ignored evidence from Buddhist
literature that strongly suggests that Buddhist intellectuals and eminent
monks conceptualized dhara and spells as integral components of main-
stream Sinitic Buddhism. Zhoushu, "spell techniques," was one of the most
prevalent translations of "dhara" used by Buddhists in medieval China.
Seminal Buddhist literature describes an ordinary monk as a student of
spell techniques and the acquisition of dhara as a prominent quality of
a bodhisattva. Dhara have a firmly established position in the Maha-
yana doctrine of the bodhisattva path. They were perceived to be part of
an ordinary monk's religious cultivation and a by-product of meditation.
In this respect they are closely associated with an advanced monk or bod-
hisattva's acquisition of the spiritual penetrations, supernormal powers
and the ability to work miracles.
The sixth-century scholiast Jingying Huiyuan promoted the view that
spell-technique dhara were an ordinary by-product of meditation. Fol-
lowing convention Jingying Huiyuan conceptualized spell techniques
(mantra) as a type of dhara. Many Indian and Chinese monks held a sim-
ilar point of view regarding dhara and were renowned for their prowess
in working wonders by means of spells. The seventh-century encyclope-
dist Daoshi endorsed dhara rituals for all people and he used examples
of the efficacious use of spells from Buddhist and, more importantly,
non-Buddhist literature to demonstrate that dharaI}I are just like native
Chinese spells - only better. The eighth-century Tantric master Amogha-
vajra, however, avoided the word "spell" in his imperially-sanctioned
definition of dhara. Instead, this eminent monk attempted to reclassify
dhara as a type of mantra or "true word" (zhenyan).
Dhara were not proto-Tantric in medieval Sinitic Buddhism; in fact,
as spells, they were a common component of mainstream Chinese reli-
gion. Daoshi's demonstration that Buddhist spells correspond well with
native Chinese practices provides nuance and perspective to the ubiquity
of spells and incantations in Chinese religion and to the Daoist borrowing
of Buddhist Brahma language in the Lingbao tradition. DhiiraIfI and spells
were so pervasive that they transcended the confines of strict affiliation
with Buddhism and also caused proponents of what in retrospect we may
call "Tantric" Buddhism in China to relinquish the word "spell" and to
differentiate their practices from mainstream Buddhist and Chinese spells
and dhiiraIjI. DhiiraIjI and spells were functional and fashionable in medieval
Sinitic Buddhism and their role in Chinese religion has continued to the
present. They were understood and used as powerful practices to promote
the Buddhist teaching and to protect the personal and spiritual welfare of
believers. Yet, the very success of Buddhist spells was probably due to
the long-standing value of spells and talismans in Chinese religion.
The four hundred years spanning the eighth to the eleventh centuries wit-
nessed the dramatic rise of a new genre of Indian literature. The tantras.
brought with them a bewildering array of new myths, doctrines, and in
particular new ritual techniques. This was an extraordinarily creative time
for Indian religions, and Buddhism was deeply affected by the new eso-
teric teachings. Thousands of new texts emerged, rewriting Buddhism's
history and reconfiguring its role in Indian society. as well as its cosmo-
logical place in the universe.
Coincidentally, these same years - from the eighth through the eleventh
centuries - also marked the arrival of Buddhism into Tibet. Thus Tibetans
first encountered Buddhism at an exciting time, just as the tantric devel-
opments in India were at their peak. Given the rapidly changing face of
Buddhism, the Tibetans must have experienced considerable difficulties
identifying a stable and authoritative Buddhist religion. Toward this end,
one of the primary strategies they resorted to was doxography, arranging
the tantras into a series of hierarchically ordered classes.
Tibetans were certainly not the first to develop Buddhist doxographi-
cal schemes. Classification systems were popular in China too and had been
for centuries, perhaps even before they were in India 1. Recent scholars have
made much of these Chinese panjiao systems, and how they reflected
cultural interests and anxieties that were uniquely Chinese
The panjiao
often tell us less about the Indian Buddhist teachings that they organize
than about the Chinese concerns that were at stake in the Sino-Indian
lOne of the earliestpanjiao was the five-part system of Huiguan (d.453). This would
predate both Bhar1:fhari and Bhavaviveka.
2 On the early development of panjiao in China, see Gregory (1991) and Ju Mun (2002).
Journal of the International Association of Buddhist Studies
Volume 28 Number 1 2005
encounters. Unlike in India, the early Chinese schemes were usually arranged
around narratives of the Buddha's life, often with the aim of promoting
a particular text over all otl:1.ers. The basic premise of these schemes was
that the siitras taught by the Buddha earlier in life were less defInitive than
those he taught later in life
The early Tibetan classification systems may
have been influenced by the Chinese panjiao systems4, but they did not
tie their classes to periods in the Buddha's life, nor did they adopt the
Chinese promotion of specific texts. The Tibetans' inter-
est in the tantras made the Chinese organizational strategies unworkable.
Where the Mahayana siitras were placed in the mouth of Buddha Sakya-
muni and emphasized a "cult of the book," the tantras were attributed to
cosmic buddhas and spawned all sorts of ritual manuals. We might also
speculate that the Chinese use of historical narrative was a reflection of
their wider interest in historical documentation, an interest not shared by
Tibetans to the same degree, particularly by Tibetans of the ninth and
tenth centuries from which so little historical documentation survived

Given the wealth of scholarship on the Chinese doxographic tradition,
remarkably little work has been done on tantric doxography. This is probably
3 The major exception to this rule is of course Huayen's Avatal?lsaka period. Recog-
nizing that the MahtiparilJirviiJ:za-siltra, known to have been taught from the Buddha's
death-bed, must have been taught last in the Buddha's life, Huayen apologists created the
"Avatal?lsaka period," claiming that the Avatal?lsaka-siltra had been preached by the Bud-
dha under the bodhi tree during the three weeks immediately following his attainment of
enlightenment. These teachings were thus a direct expression of the Buddha's enlighten-
ment experience. However upon fInishing these three weeks of teachings, the Buddha saw
that his audience of pratyekabuddhas and sriivakas were completely unable to understand
what, he had said. Seeing this, the Buddha was moved to teach the siltrapitaka of the
HInayana canon in order to prepare his disciples for the later, more advanced teachings.
4 One possible example is may be the four classes of Gnubs chen sangs rgyas ye shes'
Bsam gtan mig sgron, which are remarkably similar to the "four methods of conversion"
(hua-i) outlined by the tenth century Korean scholar Chegwan (d. 971): gradual (chien),
sudden (tun), secret (Pi mi). and indeterminate (pu-ting). See Chappell (1983), pp. 60-6l.
For more on Gnubs chen's doxographical writings, see below.
5 In his survey of these systems, Iu Mun has observed that later, during the Sui, the
Chinese classification systems did shift somewhat away from the diachronic tendency
toward more doctrinal concerns (see Iu Mun 2002, 146). Iu Mun speculates on a link
between this shift and the rise of sectarianism witnessed during the same historical period,
a link that may warrant further investigation. However clear this tum towards doctrine was,
however, the diachronic element continued to be strong enough to warrant an appearance
in most, if not all, of the panjiao of the Tang.
in large part due to a perceived lack of evidence. Observing this deficit,
Helmut Eimer recently wrote that, "Of other classifications of the Bud-
dhist Tantras [that is, apart from the standard four-fold scheme] only a few
. references in literary sources have survived."6 This is simply not true, even
more so given the recent discovery of two more doxographic systems
among the manuscripts found at the famous "library cave" ofDunhuang.
The only Dunhuang system of this sort to be studied so far has been
the one found in PT849. This important text was translated and published
in 1924 by Joseph Hackin, and since then it has been cited regularly by
many scholars
Two more systems have now come to light. Neither has
been studied to date and both deserve attention. Transcriptions and trans-
lations of both items are appended to the present article. By combining
this new evidence with the classification systems present in the Tibetan
canon, we can now begin to identify some important differences between
the Tibetan and Indian approaches to the Buddhist tantras.
The abundance of early Tibetan doxographical systems presents a chal-
lenge to the singular authority of our received and much-cited scheme -
the famous four classes of Kriya, Carya, Yoga, and * Anuttarayoga. For
over a century when scholars have written on Buddhist tantra from India
. to Japan, they have followed this scheme. These four categories have
gone almost entirely unquestioned; they have been applied with little
mind to their historical context, across space and time, to give a sense of
order to the chaotic mass of esoteric teachings known as tantric Bud-
. dhism
Some scholars have even constructed distinct schools around the
6 Eimer (1993), 224 [my addition].
7 Two of the most useful recent studies on early Tibetan classification schemes have
been Karmay (1988), 137-174, and Kapstein (2000), 10-17. Both rely heavily on PT849.
8 In the appendix to his recent book Coming to Terms with Chinese Buddhism, Robert
Sharf has called attention to scholars' often indiscriminate use of the very term "tantra."
Sharf argues that tantra as a distinct class of teachings never existed in China, and that it
is better understood as a product of Japanese and western imaginations. Sharf's arguments
should at least be considered by all scholars of Buddhist tantra. We must keep in mind,
for example, the ubiquity of ritual practice, from healing rites and divination to oral recita-
tion and visualization techniques, throughout "non-tantric" Buddhism. That said however,
it is clear that in India anyway, by the mid-eighth century at least, Buddhists were distin-
guishing the new tantric literary themes and ritual trends from those of the earlier siltras.
The absence of such distinctions in China may be related to the fact that China, as has been
noted by many other scholars, did not receive the Mahayoga tantras until, well after they
four categories, complete with historically traceable lineages
Others reg-
ularly date the fourfold scheme to the eighth century c.E., apparently only
because by the eighth century certain tantric titles were attested that would
eventually - some four centuries later - come to be classed under the
rubric of * Anuttarayoga
Such an ahistorical conflation of mere titles
with an entire doxographical system inevitably obscures much about the
early history of tantric Buddhism.
This article argues that these four classes that have gained s!lch favor
in modern scholarship are in fact a late (maybe twelfth century) and
uniquely Tibetan innovation. They are part of a long tradition of tantric
doxography that was distinctly Tibetan and therefore they obscure much
about the early development of the tantras in India. In India, the classifi-
cation of tantras was a concern in some (though notably not all) tantric
circles, but the Tibetan treatment of the subject was systematic in unprece-
dented ways. Only in Tibet do we start to see entire texts devoted to the
topic. Only in Tibet do fixed sets of classificatory criteria begin to be
applied. The Tibetan tradition of tantric doxography was a very different
creature from the Indian one, probably with much more at stake, and we
should be careful when we apply these uniquely Tibetan doxographical
categories to the history of tantric Buddhism in India. Fourfold schemes
vaguely resembling the now classic system appear in a couple of Indian
texts (along with a wide variety of alternative schemes), but the system
as we know it was formalized in Tibet and for Tibetan interests.
In order to understand these Tibetan interests more precisely, we should
perhaps first review our use of the term 'doxography.' In the west, 'doxo-
graphy' was originally used to refer to the collected summaries of the
different views asserted by the Greek philosophersll. Doxography was
emerged in India and Tibet, nor did they receive the crucial classificatory terms of Kriya
and Yoga that developed around the same time as the Mahayoga tantras (i.e. the second
half of the eighth century). Thus Chinese Buddhists seem to have experienced a break in
their transmission of Indian tantric Buddhism around the early eighth century, just at the
moment when tantric Buddhism was developing its own distinct identity in India.
9 See for example, Yamasaki (1988), 13-14.
10 For a recent example, see the introduction to the edited volume Tantra in Practice
(White 2000, 22-23).
11 Thus the entry for "doxography" in the O.E.D. reads: "1892 J. BURNET Early Greek
Phi/os. 371 By the term doxographers we understand all those writers who relate the opinions
therefore concerned with categories of philosophical views. As we use the
apply in a Buddhist context, we should be mindful of this origin and hold
open the question of whether Indian thought was ever "philosophical" in
a western sense. That said, many recent Buddhologists have recognized
something useful in the term and have adopted it, labelling certain gen-
res of Buddhist literature" doxographical." Some scholars have even gone
so far as to explore the roots of Indian doxographical thinking, and with
illuminating results

The Indian doxographic tendency had been traced back to the late
fifth century Grammarian scholar, Bhartrhari13. While Bhartrhari's own
writings were not overtly doxographical, they made regular use of the
term darsana ("view") to refer to different philosophical perspectives, and
this concept of darSana came to playa central role in the various doxo-
graphies that emerged over the following centuries. Within the Buddhist
traditions, the doxographic method was soon adopted by Bhavaviveka
(500-570 C.E.), whose writings exerted a strong influence on later Bud-
dhism. At first the doxographic paradigm was resisted by many within
Buddhism, but it was part of a deep and irresistible trend that was sweep-
ing through Indian thought; by the seventh century Candraldrti could
argue against doxography, but only on its own terms

The application of the term "doxography" within a Buddhist context
may be useful, but it can also obscure a crucial difference between west-
ern doxographies and the Indian tantric classification systems: Whereas
the former are generally philosophical works and restrict themselves to
the views held by each school, the tantric classification systems of India
(as we shall see) are largely concerned with differences in ritual practice

of the Greek philosophers. Ibid. 374 The doxography [of the Lucullusl has come through
the hands of Kleitomachus. Ibid. 375 Short doxographical summaries are to be found in
Eusebios [etc.]."
12 For an example of such an exploration, see Halbfass (1988), 263-286 and 349-368.
13 Halbfass (1988), 268.
14 On the history of this dramatic shift within Buddhist thought, see Huntington (2002).
15 A distinction may be in order here. The Indian term that corresponds most closely
with "doxography" is probably the Sanskrit siddhiinta (Tib. grub mtha'). The later Tibetan
tradition often used this term to refer to a genre of literature that concentrates primarily
on the non-tantric philosophical schools. The present study is restricted to the tantric clas-
sification systems; whether these should be considered siddhanta is a question left unan-
swered for now.
This is particularly true of the systems that circulated,in India during
the eighth to tenth centuries, precisely the period when developments in
the tantras were at their most creative
Thus to label these tantric clas-
sificatory systems "doxographical" might obscure the crucial role of rit-
ual in the development of Buddhist tantra.
Once the ritual focus of the Indian classification schemes has been rec-
ognized, a further point of interest emerges. Unlike their Indian contem-
poraries, the Tibetans preferred a more properly "doxographic" approach,
organizing the tantras around differences in doctrine. This is perhaps
the most striking difference between the Indian and the Tibetan schemes

Following the trend established in the earlier Indian Mahayana sl1tras,
early Tibetans divided Buddhism into different approaches, or "vehicles"
(Skt. yana; Tib. theg pa) by which the Buddhist practitioner can travel
the path to enlightenment. In order to evaluate and distinguish between
these vehicles, Tibetans employed a variety of criteria. Within a given
doxographical system the criteria often remained fixed, that they might
function as standards against which each vehicle could be measured and
compared. Among these sets of criteria, the philosophical views (Tib. Ita ba)
were invariably foremost, and when differences of ritual technique were
considered, they were usually framed doctrinally, as "sudden" vs. "grad-
ual" and so on. Thus one of the objectives of the present study is to dis-
tinguish more clearly the criteria used in the early Tibetan classification
systems of tantric Buddhism.
The criteria Tibetans used can tell us much about their concerns as they
worked to assimilate Buddhism into their own cultural milieu. The Tibetan
interest in doctrine should be understood within the wider historical
16 As we shall see in the survey that follows, it was particularly true of the earlier
Indian systems. An increased interest in doctrinal critera begins to be seen in the eleventh
century Indian classification systems. See for example the J)tikinfsarvacitttidvayticintya-
jiitina-vajravtirtihf Tantra (Q.60, 88a.5-6; also discussed below, n.100), which treats the
higher classes according to the mental states they teach.
17 These characterizations of Indian and Tibetan classificatory concerns are of course
a generalization. As we shall see, parts of the Dunhuang manuscript PT656 represent an
exception to the Tibetan concern with doctrine, and certainly Indian treatises on doctrinal
aspects of the different tantric classes can be found, particular after the tenth century. How-
ever, my characterization of these two traditions holds true in a remarkable number of
cases, and it is clear that we are dealing with two distinct sets of classificatory concerns -
one Indian and the other Tibetan.
context of Tibetans trying to comprehend the complexities of Buddhism.
That so many classification schemes flourished in Tibet during the ninth
and tenth centuries in particular reflects the significance of this period in
the Tibetan adaptation of Buddhism. The defInition and classification of
Buddhist vehicles into hierarchical systems was one of the principle strate-
gies used by Tibetans in their assimilation of Buddhism, and the ninth and
tenth centuries produced a bewildering array of classification systems.
Eventually, two principal schemes emerged - one of nine vehicles advo-
cated by the followers of the Rnying rna ("Ancient") school, and one of
four vehicles promoted by all the other Tibetan schools. But behind these
clearly organized systems lay two centuries of contention and confusion,
as Tibetans struggled to make sense of the foreign Buddhist religion and
their own place within it. Today these early classillcation schemes offer a
window onto the history of the Tibetan assimilation process.
What follows is a survey of the classification schemes that circulated
in India and Tibet during these tumultuous years. The survey is arranged
into a roughly chronological order following a hypothetical historical nar-
rative. With research on the early development of tantric Buddhism still
in its infancy, such a narrative is necessarily fraught with shaky conjec-
tures. Nevertheless, I have attempted to arrange the different classifica-
tion schemes so as to reflect the general developmental order of the tantras
from the late eighth to the eleventh centuries. Having completed the sur-
vey, the [mal part of this article turns to the origins of the much-cited four
classes scheme and how it came to dominate our own understanding of
tantric Buddhism in India.
I. Early Indian Classification Systems
I.a. Buddhaguhya
The first system discussed should be the one by Buddhaguhya, the
mid-eighth century Indian commentator on the Mahiivairocana-abhisal'{l-
bodhi Tantra (MVT). The MVT played an influential role in the devel-
opment of tantric Buddhism, and has been the subject of a recent study
and complete translation
Buddhaguhya composed at least two works on
18 Hodge (2003) also includes translations of Buddhaguhya's two commentaries.
the MVT, a summary (pil:u!iirtha) and a more extensive commentary (vrtti
or Both works open with brief introductions to the different
classes of teachings given by the Buddha. Buddhaguhya first distinguished
the teachings of the Mahayana siltras which advocate the cultivation of
the perfections (piiramitii) from those of the tantras which emphasize
mantra recitation. He then divided the latter into two types:
There are two kinds of disciples who engage and practice by means of
mantra: those who principally aspire toward objective supports, and those
who principally aspire toward the profound and vast. For the sake of those
who principally aspire toward objective supports, the Kriya tantras such as
the Arya-Susiddhikiira Tantra, the Vidyiidhara-pitaka and so on, are taught

Similarly, for the sake of those who principally aspire toward the profound
and vast, the Arya-tattvasa/flgraha Tantra and so forth, are taught.
It is not that those who are said to use principally "objective supports" do
not aspire toward and practice the profound and vast, but that they mostly
aspire toward practicing with objective supports. Nor is it the case that those
who principally aspire toward the profound and vast are completely with-
out practices that rely on objective supports, but that they mostly practice
the profound and vast. Clearly in this sense the Arya-tattvasa/flgraha and so
19 The fIrst of these titles, the Susiddhiktira, can probably be identified as Q,43 L The
contents of Vidyiidhara-pitaka, which may not have been a single work but a collection
of early tantric materials, are more difficult to ascertain. Lalou (1955) has identifIed some
of them. Bhavaviveka (c.500-570) cites a passage from the Vidyiidharapitaka of the Sid-
dharthas, whom he classifies in this case under the Mahas1iqlghikas (Tarkajviiiii, Q.5256,
190a6), On this, see Skilling (1992), 114. The late seventh century Chinese scholar Yi jing
purports to have studied a 100,000 verse Vidyiidhara-pitaka (see Hodge 2003, 10, where
he translates the relevant passage from the Record of Eminent Monks who Sought the Dharma
in the West (Xf-yu-qiuja-gao-seng-zhuan, T.2066)). In this regard, the Vidyiidhara-pitaka
was a precursor to the later 100,000 verse tantric collections such as the Vajrasekhara and
the Miiyiijiiia. On the latter two collections and their relationships, see Eastman (1981) and
Giebel (1995); see too n.32 below. There are also a number of short Maiijusn siidhanas
contained in the Peking bstan 'gyur that contain Vidyadhara-pitaka in their title, much as
certain later titles claim to be extractions from, or based upon, the Mayiijaia collection.
Other titles that Buddhaguhya classifIed as Kriya tantras can be culled from his other
works. In his Dhyanottarapataia, 11 b.1-3, Buddhaguhya distinguishes two subclasses of
Kriya tantras: general tantras that are compilations of ritual manuals (spyi'i cho ga bsdus
pa'i rgyud) and specific tantras (bye brag gi rgyud). Under the former type he lists again
the Susiddhiktira, to which he adds the Subiihupariprccha (Q,428) and the Kaipa-iaghu
(Q.319), Under "specific tantras" he lists again the Mahiivairocana-abhisaf(lbodhi and
the Vidyiidhara-pitaka, plus the (Q.130) and the Bodhimal}fja (Q.139).
In his Pil}fjartha, 4a.6, he further adds the Trisamayaraja (Q.134), and the
are principally for the practice of the inward yogas, but this does
. not mean theydo not include some outward practices as well. Similarly,
the Kriya tantras are principally for the outward practices, but this does not
mean they do not also include some inward practices.
One should understarid that the [tantric] Vidyadhara-pi{aka and so forth still
are said to be directed towards the three gates of liberation and the like,
just like those who engage and practice by means of the [sutra-based] per-
fections. Similarly, this Vairocana-abhisarrzbodhi-
is a Yoga tantra, principally for means and wisdom, but in order to help
those disciples who aspire towards activities, it also teaches some practices
which accord with the Kriya tantras. Therefore it can be analysed and is
renowned in both ways, as a Kriya tantra or as a tantra of both (Tib. gnyis ka;
Skt. ubhaya)21.
Thus Buddhaguhya distinguished two main categories, the outward
Kriya tantras and the inward Yoga tantras. This distinction reflected a
major shift that was taking place in Indian ritual technology at the time
of Buddhaguhya's writing. Earlier Buddhist rituals were typically directed
outward toward what Buddhaguhya called an "objective support," an
external focus for one's worship, whether an actual shrine or a
20 In his Pir;lIjiinha, 4a.4, Buddhaguhya adds the Sri-paramiidya (Q.119) as another
example of a Yoga tantra.
21 Mahiivairocana-tantra-vrtti, 2b.1-3a.1. de bzhin du sngags kyi sgo nas 'jug cing
spyod pa'i gdul ba'i 'gro ba rnams la yang rnam pa gnyis te/dmigs pa dang bcas pa la
mos pa gtso bor gyur pa dangl zab cing rgya che ba la mos pa gtso bor gyur pa stel de
la dmigs pa dang bcas pa la mos pa gtso bor gyur pa rnams kyi don tel 'phags pa lags
par grub pa'i rgyud dangl rigs 'dzin gyi sde snod la sogs pa bya ba'i rgyud rnams bstan
tol de bzhin du zab cing rgya che bas 'dul ba'i 'gro ba rnams kyi don du 'phags pa de
kho na nyid bsdus pa'i rgyud la sogs pa bstan tel dmigs pa dang bcas pa la gtso bor gyur
pa zhes pa yangl zab cing rgya che ba la ma mos sing my spyod pa ma yin mod kyil dmigs
padang bcas pa la spyod par mos pa'i shas che ba la bya'ol zab cing rgya che ba la mos
pa gtso bor gyur pa rnams la yang dmigs pa dang bcas pa'i spyod pa med pa ma yin mod
kyil zab cing rgya che ba la spyod pa'i shas che ba stel de Ita bu yin par ni gsal por
'phags pa de nyid bsdus pa la sogs pa nang gi rnal 'byor gtso bor gyur pa yin mod kyil
phyi'i spyod pa rnams kyang med pa ma yin not de bzhin du bya ba'i rgyud rnams kyang
phyi'i spyod pa gtso bor gyur pa yin mod kyil nang gi spyod pa yang med pa ma yin tel
rig 'dzin gyi sde snod la sogs par rnam par thar pa'i sgo rnam pa gsum la gzhol ba la
sogs pa gsungs ba dangl de bzhin du pha rol tu phyin pa'i sgo nas 'jug cing spyod pa rnams
la yang ji ltar rigs par sbyar shes par bya'ol de bzhin rnam par snang mdzad mngon par
rdzogs par byang chub pa rnam par sbrul ba byin gyis rlob pa'i rgyud 'di yang thabs dang
shes rab gtso bor gyur pa rnal 'byor gyi rgyud yin mod kyi/ bya ba la mos pa'i gdul bya'i
'gro ba rnams gzung ba'i phyir bya ba'i rgyud kyi rjes su mthun pa'i spyod pa dang kyang
bstan pas! bya ba'i rgyud daml gnyis ka'i rgyud Ita bur so sor brtags shing grags so.
image. But by the early eighth century ritual worship was beginning to
be directed "inward" toward the practitioner's own body. Thus in the
Yoga tantras one visualized oneself as the central deity and offered obla-
tions towards oneself22.
Following Buddhaguhya, Tibetan exegetes of the ninth and tenth cen-
turies seem to have disagreed on whether a distinct third category should
be recognized. Though Buddhaguhya emphasized the two tantric classes
of Kriya and Yoga, some believed the above-cited passage proposed a
third intermediate class, that of the Ubhaya tantras. Other Tibetans, how-
ever, apparently refused such an interpretation
Buddhaguhya's passage
itself is admittedly unclear on whether a distinct category was intended,
but it is important to recognize that the primary distinction made by Bud-
dhaguhya, as in other early materials, was twofold, while the third inter-
mediate class gained acceptance only gradually24.
lb. Viliisavajra
The confusion surrounding the intermediate class may in part explain
the variety of names applied to it. In addition to Ubhaya, we see Upaya
22 In a recent article (Dalton 2004a) I argue that this inward trend was extended through
the ninth century, as Buddhist ritual technologies focused increasingly on the body's inte-
rior. Thus in the rituals of Mahayoga, many of the same ritual structures at work in the
Kriya and Yoga tantras were mapped onto the practitioner's sexual anatomy. 'This third step
in ritual development may be reflected in Dunhuang references to three kinds of vehicles -
the outward, the inward, and the secret (see PT283 and ITJ576).
23 The question of how Buddhaguhya should be interpreted on this point has been well
addressed by Shinichi Tsuda (1965).
;;24 After the MVT, the next major step in the development of tantra is often said to be
the Yoga tantra, Sarvatathiigata-tattvasa7(lgraha (STTS). It is notable that the early STTS
ritual traditions continued to adhere to the twofold Kriya-Yoga scheme. This is indicated
in the Sarvatathiigata-tattvasa7(lgraha-siidhanopayika, a popular Dunhuang ritual manual
based on the STTS. The STTS' s absence in the 812 C.B. Ldan kar rna catalogue of transc
lated works has led some scholars to conclude that the tantra was not translated during the
early diffusion (snga dar) of Buddhism into Tibet. But the above-mentioned siidhanopayika
manual, which contains a number of passages drawn from the STTS, disproves this theory.
In fact this manual seems to have enjoyed some popularity in Tibet, as at least two ver-
sions appear in the Dunhuang collections (lTJ448/PT270 and ITJ417/PT300), in addition
to a detailed commentary (lTJ447). I am currently preparing a translation and study of these
manuscripts for publication. A passage discussing the differences between Kriya and Yoga
can be found at ITJ447, r19.2-r20.4.
and Upa- being used, as well as Carya
This latter term was eventually
adopted by the later Tibetan Gsar rna schools in their fourfold classifi-
cation system. Perhaps the earliest instance of Carya being used appears
in the writings of another late eighth century Indian tantric scholar,
Vilasavajra. Like Buddhaguhya, the perhaps only slightly later Vilasavajra
often opened his works with discussions of the classes of tantras
In his
influential commentary to the Mafijusriniimasarrzgfti he named three
classes of tantras: Kriya, Carya, and Y oga

Elsewhere, however, Vilasavajra seems to have felt no need for the inter-
mediate category. At the beginning of his other major work, his so-called
Spar khab commentary to the Guhyagarbha Tantra, Vilasavajra set forth
another classification system
In essence there are three [classes of tantras]: the Kriya tantras such as the
the "Conqueror" [i.e. Yoga] tantras such as the [Sar-
vatathiigata-] Tattvasal?'lgraha, and the tantras of the Upaya vehicle. Within
this [latter class] are three further subcategories: the [male] method tantras
such as the Sri Guhyasamiija, the [female] wisdom tantras such as Sri
Sal?'lvara, and the neuter tantras such as the Sri Buddhotpiida. The present
Sri Guhyagarbha completes and joins the aims of all the tantras, their causes
and their effects, and for this reason it is said to be common to all tantras
25 Snellgrove (1988, 1357) suggests that Upayoga was the original term for this class
of tantras which "approximate" (hence the prefix, upa-) the Yoga tantras, and that the other
terms are best understood as later "mistaken corrections." Personally, however, I have seen
no reason for making such a judgement and find Ubhaya a far more convincing original.
26 For a discussion of Vilasavajra (also referred to as Lalitavajra) and his dates, see
Davidson (1981), 6-7.
27 NamasaT{lgftitfka, 31b.2. mal 'byor spyod dang bya ba'i rgyud. See also 33a.5 (mal
'byor dang/ bya ba'i rgyud dang spyod pa'i rgyud).
28 It should be noted that the reliability of this Spar khab attribution has not been estab-
lished. I have seen nothing substantial to contradict the claim that Vilasavajra was the author,
and for this reason I have included it in the present survey. However, this question requires
closer attention than I have given it, and much of what appears in the following section
should for this reason be taken as provisional.
29 Spar khab, 131a.4-6. ngo bo la gsum ste/ phyag na rdo rjes dbang bskur ba la sogs
pa bya ba'i rgyud/ de kho na nyid thub pa la sogs pa thub pa'i rgyud dang/ thabs kyi theg
pa'i rgyud dol de la yang gsum ste/ dpal gsang ba 'dus pa la sogs pa thabs kyi rgyud dang/
dpal bde mchog la sogs pa shes rab kyi rgyud dang/ dpal 'bu ta 'byung ba la sogs pa ma
ning gi rgyud dol de la dpal gsang ba'i snying po 'di ni thams cad kyi don dang rgyu 'bras
tshang zhing 'brei pa'i phyir/ rgyud thams cad kyi spyi yin par gsungs so. As for the titles
mentioned in this passage, the neuter tantra mentioned here, the Sri Buddhotpada, remains
unidentified. The Sri SaT{lvara almost certainly refers to the Sarvabuddhasamayoga, and
In this passage, the class of "Upaya" tantras should pot be identified
with Buddhaguhya's Ubhaya class, nor with Vilasavajra's own Carya
class. Rather, it represented a new third class that was added above the
standard twofold division into Kriya and Yoga tantras
. The Spar khab's
new third class, its so-called "Upaya vehicle" was elsewhere known as
Mahayoga (literally "greateryoga")31.
During the second half of the eighth century, a new class of radically
transgressive tantras was spreading through India. Like Buddhaguhya's
earlier Yoga tantra, the new Mahayoga was deemed an "inward" class
of tantric teachings. The subject of the above-cited Spar khab commen-
tary, the Guhyagarbha Tantra, was itself a well known Mahayoga work,
though the most influential of the new Mahayoga tantras was certainly the
famous Guhyasamaja Tantra, which shared much in common with the
slightly later Guhyagarbha
The early Mahayoga tantras offered a range
not to the CakrasaJ?1vara. The CakrasaJ?1vara system, like the Sarvabuddhasamayoga,
developed gradually with multiple recensions and explanatory tantras, making it a difficult
system to date. Whatever its dates may have been in India, within Tibet the CakrasaJ?1vara
system and the Hevajra did not appear on the scene until the late tenth century; no men-
tion of either system appears in any Dunhuang manuscript.
30 It is interesting to consider that the Spar khab's name for Yoga !antra - "Conqueror"
tantra - may have been a reference to the themes and imagery of sovereignty which per-
vade the rituals of this class. It is perhaps in the same vein that the Dunhuang manuscript
ITJ423, 4v.6 refers to "the four vehicles of royalty" (rgyal theg bzhi). On the connections
between Yoga tantra ritual, Indian coronation rites, and other royal themes, see Snellgrove
(1987),234 and Davidson (2002), 113-168.
31 This reading is confmned by other early sources that identify Mahayoga with the
tantras of Upaya. See for example the Dunhuang manuscript ITJ508, rl (mal sbyor chen
po nang pa thabs kyi rgyud kyi tan tra) and the Man ngag Ita ba'i phreng ba (Karmay 1988,
165: mal 'byor nang pa thabs kyi rgyud) .
. 32 Both tantras appeared in the lists of the eighteen Mayajala tantras. All of the works
included in the Mayajata corpus were understood to have emerged from a single source,
an original Mayajala !antra which in its complete form was purported to be so lengthy that
it probably only existed as a purely mythical kind of ur-text. It may be, however, that
many of the eighteen Mayajala tantras did indeed emerge from a common social and lit-
erary "matrix." They do share a number of traits, such as an emphasis on the five bud-
dha-families, the use of the three samadhis to describe the generation of the visualized
m a J ) . ~ a l a (not be confused with another system of three samadhis found in the Yoga tantra
STTS, on which see Cozort 1986, 51), and the centrality of sacramental sexual rites).
In this sense they may represent a certain period in the development of Buddhist tantra,
when Mahayoga was still closely associated with the Yoga tantras. Future work will hope-
fully tell us more on this point. As discussed by Eastman (1981), the eighteen Mayajala
tantras as a group mirrored the earlier corpus of eighteen Vajrasekhara tantras, which were
of new ritual technologies. Particularly innovative were their sexual prac-
tices. The Yoga tantra ritualist visualized himself as the deity at the
InaI}.Q.ala's center arid made offerings, real or imagined, to himself. Now
the Mahayoga practitioner visualized the maI).Q.ala at the point of sexual
union between himself and a ritual consort. But most crucial to these early
Mahayoga rituals was the culminating moment when the practitioner
would receive a drop of the resulting male and female sexual fluids on
his tongue as a "supreme sacrament. "33
also meant to have been extracted from a massive and probably ultimately mythological
ur-tantra. Both the MiiyiijiiZa and the Vajrasekhara textual groups may also relate to still
other eighteen-fold collections. Of particular interest may be the eighteen mahiipuriiIJas
and upapuriiIJas. According to Hazra, the formation of the latter grouping dates to the mid-
ninth century at the latest and was in formation between 650-800 (Hazra 1958, I, 14-15).
We have already observed (see n.19) the possible existence of still earlier canonical mod-
els such as the Vidyiidhara-pifaka, which was classified by Buddhaguhya under Kriya
33 Elsewhere (Dalton 2004a) I have argued that this ritual moment was a definitive
characteristic of early Mahayoga in India from roughly the mid-eighth century through at
least the mid-ninth century. The importance of this sacramental ritual form was still greater
in Tibet, where the influence of later developments in sexual practice, particularly of the
complex "channels and winds" (rtsa rlung) systems, seems not to have arrived until Bud-
dhism's "later diffusion" (phyi dar) in late tenth century.
In the same article, as I laid out the evidence for my argument, I considered the possi-
ble meanings of the term "padma ban da" that appears in a number of Dunhuang descrip-
tions of the sacramental rite. In my considerations, I neglected to mention the common use
of the term (Skt. biiIJa) to refer to a tantric skull-cup. The term ban da is used in this way
in other Dunhuang manuscripts describing the iconography of wrathful deities (see for
example ITJ306, v13.2-4 or ITJ484, Iv. 1), and perhaps also related are references such as
one seen in the diaries of Yijing, in which the Chinese pilgrim describes his worshipping
a "padma skull" relic of Sakyamuni while visiting Chia-pi-shih (see Lahiri 1986,69).
A conch shell could also be used according to chapter seven of Candrakirti's CaryiimeZiipaka-
pradfpa of the Arya tradition of Guhyasamaja exegesis (see Wedemeyer (forthcoming, and
a link between skulls and conches is well attested. All this said, however, none of it con-
tradicts the idea that the term padma ban da referred to the consort's vagina. Other passages
make it quite clear that the supreme sacrament was gathered from the vagina. Such is cer-
tainJy the case in both chapter eight of the Guhyasamiija Tantra (109a.8) and the (some-
what later) eighteenth chapter (163a.8), two references that are particularly significant, since
ITJ331 which uses the term padma ban da is based on the Guhyasamiija Tantra
(as are the vast majority of Mahayoga ritual manuals from Dunhuang). In fact, the biiIJa
often symbolizes the vagina in tantric literature. For a Dunhuang passage confIrming this,
we may look to ITJ585, which describes the goddess Ghasmari: "With a vajra of means
she stirs the ban da and drinks from it" (Iv.2: thabs kyi rdo rjes dan da dkrug cing gsa).
Here the "vajra of means" alludes to the male penis "stirring" in the female skull-cup/vagina.
This literary allusion may well have resulted in two ritual methods for gathering the supreme
In the passage cited above, the Spar khab distinguis,hed three further
subdivisions within Mahayoga - the male tantras that focused primarily
on method, the female tantras focusing on wisdom, and the neuter tantras.
This strategy of assigning' genders to tantras was widely adopted after the
tenth century, and the Spar khab's use of it may be the earliest instance
we have. Finally, according to the Spar khab's classification system,
the Guhyagarbha Tantra incorporates and thereby transcends all three of
these Mahayoga classes. As already noted, the Guhyagarbha was widely
held to be a Mahayoga tantra, but the Spar khab seems to have been toy-
ing with another idea - that the Guhyagarbha should be placed in a still
higher class.
The Spar khab's hesitation to classify the Guhyagarbha as Mahayoga
is explained later in the same commentary, in its discussion of the Guhya-
garbha's crucial thirteenth chapter. For there we fmd the root tantra's own
internal classification scheme. This is one of the earliest classification
systems to appear within an actual tantra. The vital passage appears at the
beginning of chapter thirteen, where we read:
Using words which rely on letters, linguistic conventions, and nominal des-
ignations, [the teachings] are well represented in terms of no realization and
wrong realization, partial realization, rnisrealization of the genuine, discipline,
the intention, the secret, the natural secret meaning

A number of late-eighth century commentators worked to unpack this
obscure passage, and the Spar khab, if it were indeed written by Vilas ava-
jra, would be one of the earliest
. It begins by explaining that the first
sacrament, that is, from the consort's vagina or from an actual skull-cup. ill either case, the
sacrament remains the culmination of early Mahayoga ritual practice.
34 Guhyagarbha, 192.1-3. rna rtogs pa dang log par rtogs/ phyogs rtogs yang dag nyid
rna rtogs/ 'dul ba dgongs pa gsang ba dang/ rang bzhin gsang ba'i don rnarns nil yi ge
sgra btags rning tshogs la/ brten pa'i tshig gis rab rntshon teo
35 The relevant passage can be found at Spar khab, l86a.8-l86b.6. Apart from the Man
ngag Ita ba'i phreng ba (discussed below), the only other Guhyagarbha commentary attrib-
uted to an illdian author is the Dpal gsang ba'i snying po'i rgya cher bshad pa'i 'grel pa
(Q.4719) by SiiryaprabhasiI1ilia (Nyi ma'i seng ge'i 'od). After an admittedly cursory look,
however, this attribution seems even more suspicious than the Spar khab. Compared to the
Spar khab, it makes far greater use of Rdzogs chen. ill its discussion of the thirteenth chap-
ter's doxographical system (339b.5-340a.6), it includes the class of Anuyoga, a term that
was rarely if ever used before the early-to-mid ninth century. But even more telling, the
work cites a wide range of Mahayoga tantras and commentaries including Vilasavajra's
two tenus, "no realization" and "wrong realiZation," refer to the two kinds
,of worldly views; ~ a t is, the apathetic (phyaZ ba) who are uninterested
in reflection of any kind, and the nihilists and etemalists who hold the non-
Buddhist philosophical views
The next two terms, "partial realization"
and "inisrealiation of the genuine," refer to the exoteric Buddhist paths.
In this way "partial realization" describes the Sravakas, the Pratyekabud-
dhas, and the Cittamatrins, while the more subtle "misrealization' of the
genuine" is the error of the Madhyiimika.
The remaining four divisions refer to the tantric vehicles, and the Spar
khab explained them in these words:
Regarding, "discipline, the intention, the secret, the natural secret mean-
ing," while certainly a correct teaching, the practice of disciplining the three
doors [of body, speech, and mind] is Kriyii. The practitioner who primarily
performs the inward yogas belongs to Yoga. To abide in the unusual views
and practices is the "secret." To abide in the natural result of those two
inward [practices] and of all things is Atiyoga, which is taught as nothing
apart from the obscurations of the various stages in which one craves after

Thus the practice of ritual "discipline" is taught in the Kriya tantras,
while the inward "intention" is Yoga tantra. The "secret" remains unnamed,
. but we can safely assume the Spar khab intended the new class of Maha-
yoga. And finally the "natural secret meaning" refers to the new and even
higher class of Atiyoga. In short, according to the Spar khab, the Guhya-
garbha's classification system can be summarized as follows:
writings (see, for example, 351b.5). This makes unlikely, though perhaps not impossible,
the claim made by the later tradition (Dudjom 1991, 688) that the work was translated by
the late eighth century TibetanVairocana A closer study is required, but it seems unlikely
c that the work dates from before the tenth century. Because of the work's unreliability and
because it adds little to the picture provided by other textS, I have excluded it from the pres-
ent survey.
36 For more indepth discussions of all these views, see Mun pa'i go cha, vol. 51, 406.1ff,
and Mkhas pa Ide'u, 113-114.
37 Spar khab 186b.3-5. 'dul ba dgongs pa gsang ba dang/ rang bzhin gsang ba'i don
mams nil zhes bya ba nil yang dag par bstan mod kyi spyod pas sgo gsum 'dul ba kri ya
dang/ spyod pa bas nang gi mal 'byor gtsor byed pa yo ga dang/ phalla med pa'i Ita spyod
la gnas pas gsang ba stet nang pa gnyis po dang/ dngos po thams cad kyi rang bzhin
'bras bur gnas kyang/ brtags pa la zhen pa'i rim pa sna tshogs kyi bsgrib pa tsam du ston
pa'i a ti yo ga'o.
l. rna rtogs: phyal ba l. no realization: ,apathetic
2. log par rtogs: rtag chad gnyis 2. wrong realization: nihilists &
3. phyogs rtogs: 3. partial realization:
- nyan thos pa - Sravakas
- rang sangs rgyas - Pratyekabuddhas
- marn par rig pd - Vijiianavadins
4. yang dag nyid rna rtogs: dbu rna 4. misrealization of the genuine:
5. 'dul ba: kri ya 5. discipline: Kriya
6. dgongs pa: mal 'byor 6. intention: Yoga
7. gsang ba: [mal 'byor chen po] 7. secret: Mahayoga
8. rang bzhin gsang ba'i don: a ti yo ga 8. natural secret meaning: Atiyoga
The Spar khab describes the final class of Atiyoga as "nothing apart
from the obscurations," in other words, as ordinary uncontrived reality.
The realization of Atiyoga, he explains, is the result of the "two inward"
practices, probably meaning the two preceding classes of Yoga and Maha-
yoga, though this could alternatively be a reference to the two stages of
development and perfection.
We have noted 'above how the Spar khab's commentary placed the
Guhyagarbha above Mahayoga. This move was likely caused by the
Guhyagarbha's strong ties to the still-emerging class of Atiyoga,. oth-
erwise known as Rdzogs chen ("Great Perfection"). The precise rela-
tionship between the categories of Atiyoga and Mahayoga remained
unclear throughout the late eighth and ninth centuries
, which may
explain why the Spar khab seems hesitant about the relationship between
the Guhyagarbha Tantra - the principal canonical source for
AtiyogaJRdzogs chen - and Mahayoga. As seen above, the Spar khab
describes Atiyoga as the culmination of Mahayoga practice, and in this
way it presents Atiyoga as distinct from Mahayoga yet in practical terms
still dependent on it.
38 This lack of clarity is exemplified by a number of Dunhuang manuscripts. Perhaps
best known is the Rdo rje sems dpa'i zhus Ian by the early ninth century Tibetan scholar
Dnyan dpal dbyangs. In its poetic tone and its rejection of any need for ritual practice, this
work resembles the Rdzogs chen texts of its day, as recognized by the interlinear notes to
the Dunhuang inanuscripts (see ITJ470 and PT837). Yet the author himself labels it a work
on Mahayoga.
Given the discrepancies between the two major commentaries ascribed
to Vilasavajra - on the Niimasaf!!glti in which he named only the classes
of Kriya, Carya, and Yoga tantra, and on the Guhyagarbha in which he
excluded Carya and included Mahayoga and Atiyoga - it is tempting to
see the two works as representing distinct chronological periods in
Vilasavajra's thought. One might believe that the Niimasaf!!gfti com-
mentary stems from an earlier period in Vilasavajra's intellectual devel-
opment, before he had encountered the Guhyagarbha tradition and its
higher tantric classes. Such conclusions are dangerous however, and the
discrepancy is probably better explained by assuming that a looser sense
of doxography was at work. Indeed, if both works are by Vilasavajra
(and this remains a serious question), it is quite likely that he understood
the different classification systems as specific to their respective tantric
traditions; in fact Vilasavajra seems to have been well aware of the
category of Mahayoga when he composed his Niimasaf!!giti work, even
as he excluded it from his discussion of the tantric classes
Whatever the
chronological order of his two purported works, Vilasavajra does not
seem to have followed a rigid classification system into which all tantras
must be neatly arranged. Davidson has observed that theNiimasaf!!gfti has
been classified under almost every category possible
, and it is important
to recognize the arbitrary nature of these classifications, even in today's
modem Tibetan tradition.
I do not believe, however, that the general flexibility of these systems
makes the present endeavour to arrange them into a chronological narra-
tive a pointless one. Doing so can not only force us to define more clearly
the gaps in our knowledge, but also help to emphasize precisely the fluc-
tuations in terminologies that are so often portrayed by modem scholars as
clearly defined or unchanging. Certainly the most dramatic discontinuity
in the history of tantric classification literature was the break between the
systems of India and thos.e of Tibet. The brief passages we have seen in
the works of early Indian tantric scholars became entire treatises in Tibet,
and their numbers too continued to proliferate; Tibetans took their clas-
sifications far more seriously than the Indians.
39 See NamasaT(lgiti{zka, 32a.3, where he uses the telID Mahiiyoga in describing the
scriptural source for the NamasaT(lgiti.
40 Davidson (1981), 15.
II. Tibetan Classification Systems
IT.a. Padmasambhava's Man ngag Ita ba'i pbreng ba
The next commentary we have on the same passage from chapter thir-
teen of the Guhyagarbha is the Man ngag Ita ba 'i phreng ba (henceforth
MTP) attributed to Padmasambhava. This text provides a useful transi-
tion from the Indian classification systems to the Tibetan ones, as it was
purportedly composed by an Indian teacher while visiting TiberI. We can
therefore assume that Padmasambhava would have tailored the MTP for
his Tibetan audience. Padmasambhava was central in bririging Mahii-
yoga to Tiber
, and it is significant thathe did so through doxography.
His MTP represents our earliest extant text entirely devoted to setting
forth a tantric classification system. In his transplantation of tantra into
Tibetan soil, Padmasambhava apparently recognized an unprecedented
need for doxography, and his seminal text bore many fruits in the fonn
of later Tibetan doxographical treatises.
It is certainly relevant in this regard that the two best known non-tantric
Buddhist doxographical treatises produced by Indians are
Ornament for the Middle Way (Madhyamakiilar[l1wra; Q. 5284) and
Kamalaslla's Illumination of the Middle Way (Madhyamakiiloka; Q. 5287).
Siintarak1?ita and KamalaSila both visited Tibet for extended periods to
assist the king, Khri srong lde'u brtsan, in establishing Buddhism as Tibet's
state religion. In addition to these crucial works, we must also consider the
Distinctions among the Views (Lta ba'i khyad par; Q. 5847) by another
important figure of the late eighth century Tibetan court - the Tibetan
. ; 41 All indications point to the reliability of the attribution of the Man ngag Ita ba'i
phreng ba to Padmasambhava: The text clearly predates Gnubs chen sangs rgyas ye shes
(late ninth/early tenth century) who quotes it (J1sam gtan mig sgron, 207.3-6), and it reflects
a stage of tantric development we would expect to see in the late eighth century. The eleventh
century Rong zom chos k:yi bzang po wrote a commentary on it attributing it to Padma-
sambhava, and the Sba bzhed claims Padmasambhava wrote it while visiting central Tibet
Moreover, it seems that Padmasambhava composed at least one other Mahayoga commen-
tary while in Tibet. The Dunhuang manuscript ITJ321 is a complete and lengthy commen-
tary to the Thabs kyi zhags pa (Q.416) that is clearly attributed to the master. It also includes
a description of Siintigarbha, another Indian monk in Tibet at the same time, "checking
[the text] and, finding no errors, praising [padma]sambhava" (lTJ321, 84a5. slobs dpon shan
ti gar bas brtags nas ma nor nas /sam ba bha la stod pa'o).
42 On early evidence ofPadmasambhava's Mahayoga involvement, see Germano (2002),
232-237 and Dalton (2004b).
scholar Ye shes sde. Taken together, these three treatises provide an
important background against whichPadmasambhava's own contribution
should be viewed ..
The MTP has been translated elsewhere by Samten Karmay43. In I?hort,
the work sets forth a system of seven distinct vehicles. The first three are
those of the siltras, i.e. those of Sravakas, Pratyekabuddhas, and Bodhi-
sattvas, and the next three follow Buddhaguhya's lead - Kriya, Ubhaya,
and Yoga. Unlike Buddhaguhya, however, Padmasambhava labels Yoga
an "outward" vehicle, thus demoting the Yoga tantras and lumping them
in with the lower classes of Kriya and Ubhaya. Apparently by the end
of the eighth century when Padmasambhava was writing, in the light of
the more recent Mahayoga ritual developments, the Yoga tantras no longer
looked as "inward" as they once had.
The seventh and fmal vehicle in Padmasambhava's system is that of the
inward yogas, referring, one can assume, to Mahayoga. The language of
ViUisavajra's Spar khab may also be seen in the MTP, as the outward Yoga
is termed the "Conqueror vehicle" and the inward Yoga the "Upaya vehi-
cle." Also like the Spar khab, the latter vehicle is further subdivided into
three, but where the Spar khab distinguished the male, female, and neuter
tantras, Padmasambhava has three "techniques" (Tib. tshul) of develop-
ment, perfection, and great perfection (bskyed rdzogs rdzogs chen). Thus
with some interpolation of terms, we have a system that looks like this:
Sutra Vehicles:
1. Sravaka
2. Pratyekabuddha
3. Bodhisattva
Outward Tantra Vehicles:
4. Kriya
5. Ubhaya
6. Yoga/Conqueror
Inward Tantra Vehicle:
7. Mahayoga/Upaya
a. development techniques
b. perfection techniques
c. great perfection techniques
43 Kannay (1988), 152-163.
The last three techniques refer to three stages in the rituals of Mahayoga.
In the development stage, a visualized maI).qala would gradually be con-
structed with oneself at the center, followed by an exchange of offerings
and blessings. In the perfection stage, the maI).qala was generated suddenly
in the context of a ritualized sexual practice. And the great perfection
refers to the culminating moment of the entire ritual sequence, when the
visualization is dissolved back into emptiness through the ingestion of a
sacramental drop.
Taken as a whole, Padmasambhava's MTP can be understood as a sys-
tematization of much that had come before. The system skillfully wove
together the earlier writings of Buddhaguhya and ViIasavajra. By distin-
guishing between the various ritual techniques described in the tantras,
Padmasambhava provided Tibetans with an interpretive framework for
comprehending the mass of materials arriving from India at the end of the
eighth century.
n.b. DpaZ dbyangs' Lamp for the Mind
Some twenty years after Padmasambhava's visit to Tibet, the first
Tibetan classification systems began to appear. Perhaps inspired by the
MTP, the early ninth century Tibetan scholar Dpal dbyangs wrote an even
more extensive work dedicated to the coded scheme found in the Guhya-
garbha Tantra's thirteenth chapter. Dpal dbyangs has received some atten-
tion in recent scholarship for his influential work, the Questions and
Answers oJVajrasattva (Rdo rje sems dpa'i zhus Zan), of which there are
several copies in the Dunhuangcollections
Another work however, the
Lamp Jor the Mind (Thugs kyi sgron ma), is his longest4
Its significance
as a doxographical work has been so far overlooked, probably because it
is extremely difficult to read and corrupt in many places.
The work opens with a justification of classification systems in gen-
eral. The point is made that ultimately such hierarchical distinctions are
irrelevant. "The Mahayoga system," Dpal dbyangs writes, "does not reject
44 Two full copies are found in ITJ470 and PT837, and a partial copy in PT819.
45 In the Peking edition (Q.5918) it fills twenty-one folio sides, while his Zhus Ian (Q.5082)
fills eleven sides.
the twenty-one ways of clinging; nondual, it neither accepts nor rejects
them. It actualizes. them [all] without resorting to methods." That said
however, "this supreme system of thusness does have three stages of
yoga and of tantras. "46 Here Dpal dbyangs is likely referring to the three
Mahayoga stages already distinguished by Padmasambhava
. The correct
path, he continues,
is not like the practices of the unclean clans or like mistaking a mottled rope
for a snake
[In such cases,] an untrue consciousness appears, so they are
false. [In the correct path, such mistakes] may not be rejected, but this does
not mean they are true. The self-arising wisdom is without edges or center.
It is unwavering, self-illuminating, and devoid of grasping. [Yet] the four
inuneasureables and the four dhyanas are reified into distinct stages, all
forms are conceptualized as selves, and for these reasons, the three truths
of the secret should not be seen as the same

Even though such classifications can be misleading, we are told, they
are a fact of life. Enlightenment may be undifferentiated, but our addic-
tion to reification makes classification systems inevitable.
Dpal dbyangs goes on to explain how the different vehicles appear
when, straying from the knowledge of all into a view of non-discrimina-
tion (Tib. 'du shes med; Skt. asal'{lifia), one takes this view to be the
enlightened ground. From this subtle error the concept of time arises, and
"by resting in the abyss of higher and lower paths, the totality becomes
46 Thugs kyi sgron ma, 275a, 5-6. de bas rnal 'byor chen po'i lugs/ bcu gnyis gcig gi
zhen pa dag/ mi spong gnyis med blang dor med/ thabs kyi rang bzhin bral mngon byed/
de bzhin pa yi lugs mchog 'di/ rgyud dang rnal 'byor gsum rim pa.
47 Note that the "three stages of Mahayoga" (ma ha yo ga'i rnam gsum rims) are also
mentioned in ITJ436, 3v.7. There, the passage introduces the ritual that then immediately
turns to the first stage of development. That these are the same three stages as those referred
to by Dpal dbyangs may be supported by the fact that both texts purport to be based on
the Mayajala traditions (see Thugs kyi sgron ma, 274b.8 and ITJ436, 5vA).
48 Note this same image is used in MTP at the beginning of the Sravaka discussion.
49 Thugs kyi sgron ma, 275b, 1-2. de phyir mi gtsang rus spyod dang/ sab rda thag sbrul
mtho mi itar/ mi bden shes snang phyir rdzun pa' ang/ mi spong mod kyang bden pa mini
rang byung ye shes mtha' dbus medl mi g.yo rang gsal 'dzin bralla/ tshad med bzhi dang
bsam gtan bzhi/ rim par so sos bdag cing/ gzugs rnams kun la bdag rtog pas/ gsang ba'i bden
gsum mnyam ma mthong. (My translation corrects several points on the basis of comparison
with other versions of the text found in the Narthang Bstan 'gyur and the 110 volume
Bka' ma shin tu rgyas pa (vo1.86, ff.283-325). Copies of the latter are now held by the
Tibetan Buddhist Resource Center in New York and by the British Library.)
experienced as happiness or suffering. Thus it is that three stages are dis-
tinguished. "50 In this way Dpal dbyangs warns the reader against taking
such classifications too dogmatically.
Dpal dbyangs' philosophical approach reduces all vehicles to mere grades
of delusion. Such an approach allows for the non-Buddhist religions to be
considered also, and when Dpal dbyangs turns to his classification system
proper, he begins with the mundane vehicle of Gods and Humans:
Five distinct differences are taught because of varying beliefs. Their differ-
ences will be explained just briefly: The fIrst vehicle defends the sixteen
[lawsJ52. The second guards the views and practices of the four [truths].
The third teaches the twelve [links] to be definitive; the fourth, the two truths.
The secret fifth includes the outward ones, of which the fIrst is the total purity
of reflexive awareness; the second follows the former and the following [i.e.
follows both Kriya and Yoga]; the third arranges in stages the seven levels
of the clear light of space itself. However, the path for reaching the Secret
Nucleus (guhyagarbha) is for those who, through renouncing the [ other] four
vehicles, abide in the fruition of the single vehicle. This [mal one is the ulti-
mate resting place. This [mal one is the main one explained here

50 Thugs kyi sgron rna, 275b, 4-5. g.yang sa mtho dman lam zhugs pas/ zad par bde
sdug myong bar 'gyur/ 'dod pa'i 'bras bu ga la yin/ de phyir gsum la khyad par mod.
5l Matthew Kapstein has noted the existence of a similar description of the origin of
the non-Buddhist views in the doxographical writings of the second Karmapa, Karma Pak-
shi. Kapstein points out that this description allowed for an eclecticism in Karma Pakshi' s
classification system, and he argues that this eclecticism is best understood within the con-
text of the Karmapa's close ties with the Mongol emperor Mongke Khan. By taking such
a philosophical view of non-Buddhist religions, Kapstein argues, Karma Pakshi made
allowances for the Mongolians' own religious heritage. (See Kapstein 2003 and also his
early study of the same materials in Kapstein 2000, 97-106.) While this all may be true,
it is important to recognize that Karma Pakshi' s theories were not all that unique within
the wider history of Tibetan exegetical writings on the Guhyagarbha's classification system.
Much of what Karma Pakshi wrote was prefigured by Dpal dbyangs as early as the ninth cen-
tury. It is therefore worth considering whether the eclecticism inherent in this philosophical
presentation of the Guhyagarbha classification system might have been motivated by early
Tibetans' wish to place their own pre-Buddhist religious traditions in a more sympathetic
light vis-a-vis Buddhism. Such a reading would seem to be encouraged by Dpal dbyangs'
inclusion of a vehicle of Gods and Humans (lha mi'i theg pa).
52 See Dudjom (1991), 59-60, for an enumeration of the sixteen pure laws of humans
(mi chos gtsang rna bcu drug) which confirm this first vehicle is that of Gods and Humans.
53 Thugs kyi sgron rna, 276a.6-276b.1. Inga yi bye brag mi mthun pal mos pa tha dad
phyir gsungs pal bye brag mdo tsam bshad par byal theg pa dang po bcu drug mgonJ
gnyis pa gzhi yi Ita spyod bsrung/ gsum pa bcu gnyis nges par bstanJ bzhi pa bden pa mam
pa gnyis/ lnga pa gsang ba 'ang phyi pa 'ang/ dang po rang rig mam par dag/ gnyis pa
The remainder of the text goes on to explain each vehicle in greater
. detail. In short, fh:e vehicles are distinguished
1. Gods and Humans
2. Sravakas
3. Pratyekabuddhas
4. Bodhisattvas
5. Tantras:
i. Kriya
ii. Ubhaya
iii. Yoga
iv. Mahayoga
At first glance, this system resembles that of Padinasambhava' s MTP.
Both authors list Kriyii, Ubhaya, Yoga, and Mahayoga and then subdivided
the latter into three. Yet Dpal dbyangs' system differs on two significant
points: First, it adds the vehicle of Gods and Humans (lha mi'i theg pa)
at the beginning. Second, it describes the vehicles in terms far more doc-
trinal than those used in the MTP, or in any of the Indian presentations
we have examined so far.
The first point, that Dpal dbyangs inserts the vehicle of Gods and Humans,
may well betray the influence of Chinese doxographical systems. In his
1983 article, Peter Gregory suggests this worldly non-Buddhist vehicle
originated in China. Gregory traces the earliest reference to the vehicle
to the writings of Liu Ch'iu (438-495), a lay Buddhist recluse from south-
ern China. In this Chinese context, the teaching of Gods and Humans
focused on the workings of karma, with the ultimate goal being a higher
rebirth within the cycle of sarrzsara. This class of teachings, writes Gregory,
"seems to have been invented by Chinese Buddhists during the second
half of the fifth century in an effort to accommodate Buddhism to the needs
of its growing number of lay adherents by adapting it to the more socially
oriented concerns of Confucianism."55 Following this early instance,
snga phyi rjes su 'brang/ gsum pa dbyings nyid 'ad gsal ba'i/ sa bdun rim par bkod pa yang/
gsang ba'i snying par 'gro ba'i lam! theg pa bzhi yis nges 'byung la/ theg pa gcig gi 'bras
bur gnas/ tha rna' ang bsti gnas mthar thug stel tha rna' ang gtsa mchog 'dir bshad do.
S4 A possibly similar five vehicle tantric system is referred to in at least one Dunhuang
manuscript (see ITJ384, r6.1). Unfortunately, it is unclear if the vehicles intended were the
same as those described by Dpal dbyangs.
ss Gregory (1983), 256.
however, the class fell into some disuse; during the seventh-and early eighth
centuries, the teaching of Gods and Humans was missing from both the
influential classification schemes of Fa-tsang (643-712) and Hui-yuan
(ca. 673-743). Then in the early ninth century, the great Chinese.scholar
Tsung-mi (780-841) broke with his recent predecessors and included the
teaching, "as the first and most elementary level of Buddhist teaching"56
in his new Yiian-jen lun system. Gregory concludes that Tsung-mi's inclu-
sion of the teaching helped, "to reconcile the Confucian moral values
that he had learned in his youth with the teachings of the religion that he
had adopted as an adult. "57 It also, "reflected the growing importance of
lay Buddhist societies throughout different strata of Chinese society during
the latter part of the T'ang dynasty."58
Tsung-mi's dates make him roughly contemporary with Dpal dbyangs,
and Dpal dbyangs' addition of the vehicle of Gods and Humans is best
understood in similar terms. Buddhism was spreading and growing in
importance throughout Tibetan society in the early ninth century, and Dpal
dbyailgs' treatise may be seen in part as an attempt to bring the non-Bud-
dhist Tibetan religions into the Buddhist fold.
The vehicle of Gods and Humans continued to be used throughout
the crucial years of the Tibetan assimilation of Buddhism in the ninth
and tenth centuries. The Tibetan manuscripts discovered near Dunhuang
contain a number of references to the term
. Perhaps the most well known
example appears in ITJ370, studied by Hugh Richardson. The passage sup-
ports our reading of Dpal dbyangs:
They [the two Tibetan kings Srong brtsan gam po and Khri srong Ide brtsan]
received that [Buddhist] doctrine and devoted themselves to it and caused
56 Gregory (1983), 253.
57 Gregory (1983), 279.
58 Gregory (1983), 296. .
59 One instance of particular interest comes in the letters of introduction for a travel-
ling monk that are found in ITJ754. The fourth of these letters is written by a military offi-
cial (Dmog 'bu cang), who ends the request for good treatment of the pilgrim with, "On
the part of gods and men also, like consideration is requested" (lha myi phyogs kyang de
bzhin du dgongs par gso!). This line would seem to suggest that the term "gods and men"
had. entered common parlance as a reference to the lay community within which a mili-
tary commander such as our author might wield power. For a transliteration and transla-
tion of the letter in question, see Thomas (1927), 555.
it to spread among all creatures ... The bounds of the dominion increased
and the land of Tibet was happy. Harvests were good, diseases of men and
cattle rare. The sound qualities and right behaviour of the people increased;
and, far from shunning the rites of gods and men, they revered them and,
clinging even more strongly to those principles, they did not fail in proper
respect and affection towards teachers and parents, brothers, sisters and kins-
men, and to those who through age are in a position of honouti.
As in China, the teaching of Gods and Humans is associated with the
religious practices and the cultural values of non-Buddhist Tibet. The
passage indicates that Tibetans worried about the new foreign religion
jeopardizing their native ways of life, and one can imagine that the ritu-
als and the rhetoric of the Buddhist tantras must have posed a particularly
direct threat.
After the tenth century, when Tibetan Buddhism was cleansed of many
of the non-Indian elements that had accumulated during the so-called
"dark period," the popularity of the vehicle of Gods and Humans faded.
In later centuries it was maintained in only a few pockets of the Bon and
Rnying rna traditions
By the end ofthe tenth century, it seems, the vehi-
cle of Gods and Humans had fulfilled its transitional purpose and was no
longer needed.
The second (and more crucial) point of difference between the two
systems of Dpal dbyangs and Padmasambhava is the more doctrinal focus
of Dpal dbyangs'. Padmasambhava's MTP, like the works of Buddha-
guhya and Vilasavajra before him, describes the classes of tantras almost
entirely in ritual terms. Dpal dbyangs' discussion stands in stark contrast
to these earlier Indian systems. Rather than distinguishing Kriya and Yoga
by their outward vs. inward approaches to ritual, Dpal dbyangs portrays
Kriya as concerned with "the total purity of intrinsic awareness" and
Yoga as focused on "the seven levels of the clear light of space itself."
This doctrinal approach to classification fits with what we have already
60 Richardson (1998), 76. My italics.
61 Kannay (1988), 148, claims there are no references to the vehicle of Gods and
Humans in Rnying rna sources. In fact references abound, not only here in the Thugs kyi
sgron rna but in other Rnying rna writings such as those on the Dgongs pa 'dus pa'i rndo
(for a late seventeenth century example, see Dhannasrl's Mdo dbang gi spyi don, 136.3).
The Bon po classifications systems, which also make use of nine vehicles, are beyond the
scope of the present study; on these, see Mimaki (1994), 126-132.
seen of Dpal dbyangs' couching the entire classification enterprise (and
the three ritual stages of Mahayoga in particular) in philosophical tenns.
The various classes are eI)1phasized as reflections of one's mental state
rather than types of ritual practice. This is the first Tibetan classification
system we have examined, and we shall see that this doctrinal emphasis
continued in later Tibetan writings.
II.c. Gnubs chen sangs rgyas ye shes
After Dpal dbyangs in the early ninth century, our next classification
system appears in the influential tantra titled the Compendium of Inten-
tions Sutra (Dgongs pa 'dus pa'i mdo; henceforth GDD). This work was
probably composed in Tibetan around the mid-ninth century62. A sprawling
work of over six hundred folio sides, it represents an early Tibetan attempt
to organize all of Buddhist tantra into a single, comprehensive system.
Its success made it the locus classicus for the nine vehicle classification
scheme used in later centuries by the Rnying rna school
We have seen both Padmasambhava and Dpal dbyangs divide their
highest class of Mahayoga into the three "techniques" (tshu!) or "stages"
(rim) of development, perfection, and great perfection. In the writings of
Gnubs chen, these three stages are enshrined as three separate classes.
The resulting system reads as follows:
1. Sravaka
2. Pratyekabuddha .
3. Bodhisattva
4. Kriya
5. Ubhaya
6. Yoga
62 According to its colophon, the GDD was translated from the obscure central Asian
. language of Brusha (a kingdom near modem-day Gilgit). While this may have been true
of certain sections, the vast majority of the work appears to have been written in Tibetan.
Regarding the work's date, and for more on this tantra, see Dalton (2002).
63 It is interesting to note that the GDD may postdate its own ritual system as repre-
sented in certain early ritual manuals. The classification system used in these early ritual
manuals begins with a vehicle of Gods and Humans, while excluding the vehicle of the
Ubhaya tantras (see, for example, the Mda dbang gi lag len zab rna attributed to the Indian
master Bde ba gsal mdzad).
7. Mahayoga
8. Anuyoga .
9. Atiyoga
Here we may suggest a further possible trend unfolding in these early
classification systems. In Padmasambhava's MTP, the three Mahayoga
techniques of development, perfection, and great perfection grew directly
out of the ritual sphere. Then in Dpal dbyangs' Lamp for the Mind, Dpal
dbyangs made it clear that he saw the same three subdivisions of Maha-
yoga as both "three stages of yoga and of tantras." Some seventy-five
years later, we reach given Gnubs chen, who was deeply involved in the
codification of the tantras around the turn of the tenth century, and we may
detect a still greater concern with bibliographic taxonomy. These two
classificatbry purposes - for distinguishing the phases of a ritual vs. for
categorizing tantric scriptures - may have led to the different terminology
used by Gnubs chen. That is, the three "techniques" or "stages" may
have emerged initially as a result of new ritual practices developing, while
the distinct "vehicles" of Mahayoga, Anuyoga, and Atiyoga may have
come later to facilitate the classification of tantric scripture

The question of whether these three categories should be mere stages
or full-fledged vehicles continued to be debated for centuries in Tibet. After
the tenth century, the "new" (gsar ma) schools followed a more conser-
vative reading, interpreting the three only as stages
, while the Rnying
ma pa maintained the early Tibetan systems we are seeing here, in which
the three constituted entire vehicles. This was a significant difference,
for doxographic recognition carried with it authority. Samten Karmay has
pointed to the polemical writings of the thirteenth century scholar Sa
pru;t kun dga' rgyal mtshan, in which the Rnying ma pa are criticised for
precisely their imprudent naming of vehicles
"The view of Atiyoga is
64 We have seen a similar shift from ritual innovations to bibliographic concerns in the
case of the Yoga vehicle. The earliest known distinction between Kriya and Yoga was
made in the commentarial writings by the Indian Buddhaguhya for the purpose of classi-
fying tantric scriptures; this despite the fact that the ritual innovations seen in the SITS
predated Buddhaguhya's categorization by some seventy-five years.
6S Tsong kha pa was careful to explain that the classes of tantras should be understood
as "doors of entry" and not as vehicles in their own right. Thus, according to him, there
is only one tantric vehiCle, namely the Vajrayana. See Tsong-ka-pa (1977), 15!.
66 Karmay (1988), 147.
wisdom; it is not a vehicle. To make the inexpressible ah object of dis-
cussion was not the intention of the learned ones. "67 Thus many followers
of the new schools, like Sap3lJ, rejected the Atiyoga tantras as spurious
for reasons rooted in part in doxography68.
Early Tibetans' more liberal attitude towards multiple vehicles may
also have been related to their interest in doctrine. Their creation of distinct
vehicles and distinct doctrines for each class of tantras may have served
to justify one another, that is, a distinct doctrine justified anotht<r vehicle
and a separate vehicle required a distinct doctrine.
The GDD presents its nine vehicles system in chapter forty-four. It sets
the system within a doctrinal discussion of three larger Buddhist vehicles
that function together as an entire cosmology. The first of these three
is the vehicle of the "continuous wheel" ('khor {o rgyun), intended for
those beings who are attached to the desire realm ('dod {a zhen pa). It uses
67 Dom gsum rab dbye, as cited in Karmay (1988), 147. a ti yo ga'i Ita ba nil ye shes
yin gyi theg pa mini brjod bral brjod byar byas pa nil mkhas pa'i dgongs pa min zhes bya.
68 It is important to recognize that Tibetans were not alone in making Atiyoga "an
object of discussion." There is some evidence that some Indians of the early ninth cen-
tury were also toying with the idea of making the three stages of tantric ritual into three
distinct vehicles. Curiously, however, the Indian texts that use the terms do so in a different
order. The alternative order observed in these works is: Yoga, Anuyoga, Atiyoga and
Mahayoga. This order can be observed in both the KNl}ayamari Tantra (p.123) and less
clearly in the Sarvabuddhasamayoga (166a.7-8). One might dismiss this as a mere eso-
tericization of the "normal" Maha-Anu-Ati order seen in Tibetan works of the ninth
century, except that the Indian works are in such close agreement. It seems, rather, that a
separate line of development is represented by the Indian works in question.
That the four classes in this system referred to four stages of ritual practice is clear from
both the KNIJayamari and the Buddhasamayoga themselves and their numerous extant
Indian commentaries. All of these sources agree on how the four classes should be under-
stood. In brief, Yoga tantra refers to the development stage; Anuyoga refers to entering
into union with a consort; Atiyoga is the spread and stabilization of the bliss of that union;
Mahayoga is the ingestion of the sacrament and the resulting experience of enlightenment.
One can see that this system reflects a similar period in the development of Buddhist
tantric ritual as that reflected in the roughly contemporary works of Padmasambhava and
Dpal brtsegs, but here the perfection stage is divided into two parts - the initial union with
the consort (Anuyoga) and the subsequent stabilization of sexual bliss (Atiyoga). For some
of the relevant commentarial passages, see: (1) Kumaracandra's Ratnavaifpaiijikii, found
in the KNIJayamari Tantra, 250-266, (2) Santimitra's, Sarvabuddhasamayogapaiijika,
74b.2-76a.l, (3) *Indranala/Brgya byin sdong po's Srfsarvabuddhasamayogatj.akinf-
jalasarrzbaratantrarthatfkii, 303b.6-304a.3. The latter source is particularly clear in its pres-
natural forces (rang bzhin shugs) to tame their desires for momentary pleas-
ure. Thus, the tantra explains, nature provides the disciple with three things
that satisfy his/her needs so that he/she can progress towards enlighten-
ment: birth, sustenance, and support(skye, 'tsho, rten)69.
The second of the three general vehicles is that of the "magical display
arising obviously" (cho 'phrul mngon par 'byung ba), which manifests
for those disciples who are extremely difficult to tame (gdul dka' drag po).
This vehicle is designed to sever the karmic continuum of those intensely
engrossed in the three poisons of desire, ignorance and anger. This is accom-
plished by means of the apocalyptic aeons leading up to the final confla-
gration at the end of the universe. The crescendo of suffering experienced
in these apocalyptic aeons cause many of these benighted beings to reflect
upon, and feel regretfor, their earlier misdeeds. In this sense, these aeons
are the buddhas' final effort to help those who are so stubborn that they
have not been liberated before this late date

The last of the three general vehicles discussed is the "vehicle for
ascertaining the ultimate" (don dam nges pa'i theg pa). This is where we
find our system of nine vehicles. The GDD's presentation arranges them
into three groups of three. Thus the "vehicle that extracts the source [of
suffering]" (kun 'byung 'dren pa'i theg pa) contains the Sravaka, Pratyeka-
buddha, and Bodhisattva vehicles. The "vehicle of awareness through
69 For the discussion of this "natural" vehicle, see Dgongs pa 'dus pa'i mdo, 349.1-350.7.
Mkhan po nus ldan (Dgongs 'dus 'grel pa, vol. 54, 464.6-476.5) explains that this vehicle
functions simultaneously on five levels, listed in order of increasing subtlety. First, because
all things come from the five physical elements, the buddhas are arising all the time as
whatever is wanted. Second, space provides the opening for everything else; earth gives a
firm ground for beings and plants; water is pliant, clear, constantly flowing and quenching;
fire is warm, bright, and rising upwards; wind is unobstructed, unabiding, formless, power-
ful and scattering. Third, each element brings beings to enlightenment: Space is the all-
pervading opening for appearance and emptiness; earth is everywhere in the sphere of
Mahayana; water is pure calm abiding; fire is insight; wind scatters the objects of con-
sciousness. Fourth, these five elements can also be experienced as the five primordial bud-
dhas. And fifth, the discussion ends with the [mal characteristic that is most useful in all
five elements: nothing is really happening, so everything is already enlightened.
70 Mkhan po nus ldan (Dgongs 'dus 'grel pa, voL 54, 497.2-519.2) explains there [lIst
come a series of three aeons: one of famine, one of plague, and one of war (mu ge bskal
pa, nad bskal pa, mtshon cha'i bskal pa). These last respectively for three years, three
months, and three days, as time speeds up to the vanishing point and the closing aeons of
[lIe, water, wind and space.
asceticism" (dka' thub rig byed theg pa) contains the outward yogas of
Kriya, Ubhaya, and Yoga. And the "vehicle with the powerful methods"
(dbang bsgyur thabs kyi theg pa) has the three inward yogas ofMahayoga,
Anuyoga and Atiyoga
In ibis way the GDD sets the nine vehicles. within
a Buddhist cosmology, embedding its entire scheme in doctrinal tenns,
and it is significant that the GDD is a Tibetan composition, for it exhibits
very different concerns from the much briefer ritual-focused presenta-
tions of classification systems seen in India.
In fact, the actual names of the nine classes are not used in the taIitra itself,
nor are each of the nine actually tenned "vehicles." Their descriptions makes
it clear what was intended, but their labels are made explicit only in the early
commentary by the Tibetan scholar Gnubs chen sangs rgyas ye shes
(b. 844 ?)72. It seems clear, however, that the authors of the GDD had in mind
the same system of nine vehicles - Gnubs chen studied directly under
the GDD's "translators," who we suspect were also the work's authors, so
his commentary probably did not introduce many significant innovations.
Gnubs chen is renowned for his support of Buddhism through the so-
called "dark period" of political chaos that stretched from the collapse
of the Tibetan empire around 842 to the late tenth century. He was par-
ticularly focused, as were many Tibetan exegetes during these dark years,
on the codification of the tantric teachings in Tibet. Under the Tibetan
empire, the translation and dissemination of the tantras had been carefully
controlled, but with the empire's collapse, those restrictions were lifted
and Tibetans eagerly adopted and adapted the tantric myths and rituals.
Compared to the court-driven Buddhism of the imperial period, this tantric
conversion of Tibet seems to have taken place at the local level. Gnubs
c h ~ n was the one great exception to this rule, a Tibetan scholar working
on a large scale, translating new tantras and composing not only shorter
works but long, systematic treatises on tantra such as his two-volume
commentary on the GDD, An Armor Against Darkness (Mun pa'i go cha),
and his famous Lamp for the Eye in Meditation (Bsam gtan mig sgron)13.
71 For the relevant passage, see Dgongs pa 'dus pa'i mdo, 351.1-352.6"
72 On Onubs chen's dates, see Vitali (1996),546-7.
73 The standard source for Gnubs chen's life appears in the seventeenth century col-
lection of biographies for the lineage of the ODD (,Dus pa mdo dbang gi bla ma brgyud
pa'i rnam thar, 160-177).
Gnubs chen's other major work, the Lamp for the Eyes in Meditation
was probably written shortly after the turn of the tenth century. In this
work, Gnubs chen makes no mention of the nine vehicle system, despite
citing the GDD more than any other source. The nine vehicle system may
be inferred, however, from his regular use of its terms, including the high-
est three vehicles of Mahayoga, Anuyoga, and Atiyoga, and it is clear that
Gnubs chen did not see his new fourfold system as contradicting the more
well known nine vehicles system. The Lamp for the Eyes sets forth a new
fourfold hierarchical classification system. Gnubs chen refers to a num-
ber of similar fourfold systems that were circulating in Tibet at the time.
His includes the following four divisions (1) gradual sutra-based teach-
ings, (2) immediate sutra-based teachings, (3) Mahayoga, and (4) Atiyoga.
Gnubs chen wrote the Lamp for the Eyes for two main purposes: to
resolve Tibetans' confusion around the Chinese Chan and around the rela-
tionship between Mahayoga and Atiyoga. In both regards then, his work
addressed uniquely Tibetan concerns. In addressing these concerns, Gnubs
chen again focuses overwhelmingly on issues of doctrine and distinguishes
his four classes according to their philosophical take on non-conceptuality74.
The Lamp for the Eyes' doctrinal focus was certainly related to Gnubs
chen's role as the great Tibetan codifier of the dark age. During the early
development of tantric Buddhism, many of the most significant innova-
tions came out of the ritual sphere. Thus, for example, the class of Yoga
tantras grew out of a new emphasis on the practitioner's own body as the
site for the divinity; Mahayoga grew out of the new ritual emphases on
the sexual yogas (sbyor ba) and the violent liberation rite (sgroI ba);
Anuyoga grew out of the increasing interest in the ritual techniques of the
sexual yoga/the perfection stage; and Atiyoga grew out of the taste of the
mind of enlightenment (bodhicitta) obtained at the culmination of the
sexual yoga. As the new ritual techniques and their corresponding textual
categories were codified for a Tibetan audience, new doctrines were devel-
oped to help distinguish them more clearly from one another. Gnubs chen's
Lamp for the Eyes presents a particularly clear example of this trend
towards doctrine.
74 For detailed discussions of Gnubs chen's fourfold system, see Dalton and van Schaik
(2003) and Meinert (2004).
II.d. The Explanation of the Order of Views
By the eleventh century the Tibetan followers of the early Rnying rna
school looked to Gnubs chen as the principle upholder of Buddhism
during the so-called "dark period." The early Zur-s in particuiar made
Gnubs chen a central figure in their early lineage, and as the Zur clan's
influence spread, so did the nine vehicle system that was first seen in the
Compendium of Intentions Sidra and Gnubs chen's commentary. Before
long, the nine vehicles had become the accepted doxographic scheme
within the Rnying rna school.
For a final example of a Tibetan classification system, we turn now to
a Tibetan work of unclear provenance. The Explanation of the Order of
Views (f-,ta ba'j rimpa bshad pa) is traditionally attributed to the early
ninth century Tibetan scholar, Ska ba dpal brtsegs. However, some doubt
was cast on the authenticity of this attribution by the fourteenth century
scholar Bu ston
The work contains an alternative presentation of the
same nine vehicles seen in Gnubs chen's writings, and it may better be
seen as a result of the eleventh or twelfth century spread of the scheme
within early Rnying rna circles

In any case, the work is clearly of Tibetan origin, and the presentation
of its classification system is similar to both Dpal dbyangs' and Gubs
chen's in that it too focuses on doctrine. Thus on the classes of tantras we
Kriya views the ultimate as the dharmatii and views the conventional as
the good qualities of reflexive awareness. They view three families in their
emanated maIfqala. They assert that errors which [lead to] sadness do not
75 On Bu ston's doubts, see Kannay 1988, 149. At this point in our survey, we might
also be tempted to insert a discussion of the nine vehicle classification system found in the
G.yu'i thang rna khras dgu (QA729). Judging from the work's use of certain technical
tenns, I suspect it too dates from after the tenth century. The unlikely colophon attributes
it to Vimalamitra, further evidence of the work's dubious origin. Its classification system
is mentioned by Kannay (1988), 149 and 172. I have excluded the work only because it
would add little to the central argument being offered here.
76 The central discussion of this system appears at Lta ba'i rim pa bshad pa, 424bA-
425a.5. Note that the work refers to the second tantric vehicle as both u pa ya and gnyis
ka (Skt. ubhayii), indicating that when Tibetans wrote "u pa ya," they may have intended
"ubhaya." This would seem to be further evidence that Ubhaya was the original tenn used
in India (see note 25 above).
exist. The as.sertions of the Ubhayii tantras are in agreement with the views
and the practices of the [classes] above and below. According to Yoga, the
ultimate is completely pure. The deities emanate through the blessings of
realizing the wisdom of the dharmadhiitu. Ordinary beings are those who
are erroneous with regards to seeing in that way. According to Mahiiyoga,
ultimately nondual reflexive awareness is thusness. There is no saT[lsiira
within the state in which the deities, which are the six manifestations of
realization and so forth, manifest. According to Anuyoga, ultimately the
dynamism of knowing, which is the realization of the great bliss, emanates
as the m ~ g a l a of conventional deities. For that reason, error comes from
ordinary beings [trying to] analyze that. Atiyoga is devoid of the two truths

Here the six tantric vehicles are distinguished -by their descriptions of
the maI}-q.a1a emanation process in terms of the Mahayana doctrine of the
two truths. Once again, we see a strong Tibetan interest in doctrine, and
more specifically in the process by which the Buddhist teachings emanate
out of emptiness. .
ID. Evidence from Dunhuang
lILa. New Evidence: Pelliot tibetain 656 and IOL Tib J 644
Having gained -some idea of the Tibetan tradition of tantric doxogra-
phy, we can now turn to the evidence that survives from the "library
cave" of Dunhuang. Until now only one classification system from Dun-
huang has been brought to light. PT849 was first presented in the 1924
study by Joseph Hackin and has been cited regularly ever since. In recent
months, however, two new manuscripts have surfaced, each containing
an extensive discussion of a classification system. Transliterations and
translations of both manuscripts are appended to the present article.
77 Lta ba'i rim pa bshad pa, 425a.1. kri ya don dam chos nyid Ital kun rdzob rang rig
yon tan Ita! rigs gsum dkyil 'khor snang bar Ital skyo ba'i 'khrul pa med par 'dodl gnyis
ka rgyud kyi 'dod pa nil Ita spyod gong 'og rjes su 'thun! yo gas don dam mam dag pa'ol
chos kyi dbying kyi ye shes sui rtogs pa'i byin brlabs Iha soong bal skye bos de Itar mthong
par 'khrul! rna hA yo gas don dam du/ rang rig gnyis med de bzhin nyidl rtogs pa'i cho
'phTul drug sogs pa'illhar snang ngang la 'khor ba medl a nu yo gas don dam dul bde
chen rtogs pa'i rig pa'i rtsal! kun rdzob lha yi dkyil 'khor snangl de phyir skye bos rtags
pas 'khrul! a ti yo ga bden gnyis bral.
The first piece, PT656, is a scroll from the Pelliot c6llection held at
the Bibliotheque Nationale. Its date remains uncertain, but it is likely
from the tenth century, as are most of the Tibetan Dunhuang manuscripts
relating to tantra. The system it describes consists of seven classes, called
"the seven great general scriptural systems" (spyi'i lung chen po bdun).
As the name of the system implies, its principal purpose was to classify
scripture. The term "vehicle" is only 'used once in the manuscript, in the
context of a Yoga tantra vow not to drink water from the same. valley as
people of a lower vehicle. The "seven scriptural systems" consist of the
two sUtric vehicles of Sravaka and Bodhisattva, the two outward tantras
of Kriya and Yoga, and the three inward tantras of Mahayoga, Anuyoga,
and Atiyoga. Thus in comparison to Gnubs chen's nine vehicles scheme,
the ever-obscure classes of the Pratyekabuddhas and the Ubhaya tantras
are excluded to make seven. The manuscript's discussions of the tantric
classes are far more extensive than those of the sutric ones, and we find sev-
eral passages appended to the end of the text dealing with the practices of
union and liberation as interpreted by Maha-, Anu-, and Atiyoga. In this
way the content reflects a strong interest in the tantras.
The work evaluates each vehicle using a fixed set of four criteria: views
held, meditations achieved, practices performed, and vows followed. These
criteria reveal a mix of concerns. "Views" are given pride of place as the
first criterion so that doctrine is given a clearly dominant role. However,
the next two criteria of "meditations achieved" and "practices performed"
introduce considerations of ritual that are really quite unique among the
early Tibetan classification systems. The ritual concerns weaken as the text
proceeds to the higher tantric vehicles, so that Anuyoga and Atiyoga are
discussed in entirely doctrinal terms, but for Kriya, Yoga, and Mahayoga,
significant attention is given to ritual practice. In this way, PT656 pro-
vides a rare glimpse of how Tibetans of the tenth century organized the
different kinds of tantric ritual.
Our second new manuscript from Dunhuang is a three-folio pothi found
in the Stein collection held at the British Library. This manuscript more
certainly dates from the tenth century78 and contains two short items. The
78 The manuscript was found by Stein in the bundle he numbered 73.iii., which has been
identified by Tsuguhito Takeuchi (2003) as containing particularly late materials, many of
which date from the late tenth century.
first item is titled, "A Teaching on Classifying the Deity Systems and the
Measures of Accomplishment" (Lha rgyud dang grub tshad nye ring bstan
pa) and is the principal work on the vehicles. The second is titled, "Classi-
fications of the Vidyiidharas" (Rigs 'dzin dbye ba) and fills out the pic-
ture with a discussion of the levels of realization associated with each
tantric vehicle

The classification systems described in these two manuscripts bear a
number of resemblances to one another. At the end of the British manu-
script's second item on the vidyiidharas, we even find a reference to the
same scheme of "seven scriptural systems" (spyi lung bdun) seen in PT656,
even though ITJ644 itself follows a fuller nine vehicle system. It seems
the seven scriptural systems may have enjoyed some popularity during the
tenth century, at least around Dunhuang, even as the nine vehicles scheme
was becoming the generally accepted standard for followers of the later
Rnying rna school. Unfortunately, the source for this sevenfold system
remains to be identified.
Also like PT656, ITJ644 employs a fixed set of criteria for evaluating
the different classes (it does not use the word "vehicle" even once). This
is where the similarities end however, for ITJ644 is entirely concerned
with issues of doctrine over ritual. The set of criteria used includes: (1)
the deity system, (2) the difference between teacher and disciple, and (3)
the measures of accomplishment. As the language of the text makes clear,
all three of these criteria are aspects of the "views" (Ita ba) held by each
class. According to our analysis of the other Tibetan classification sys-
tems, then, ITJ644 is typical of the Tibetan tradition of tantric classifica-
tion and categorizes the tantras according to their respective doctrinal
The use of fixed sets of criteria seems to have been a Tibetan develop-
ment; to my knowledge such a strategy does not appear in the Indian
sources on the classes of Buddhist tantras
The efficacy of this almost
scientific classification technique seems to have been grounded in its
apparent objectivity, ill its application of seemingly impartial criteria to
79 For more on this item and its possible links to the Padmasambhava legends, see Dal-
ton (2004b).
80 The strategy of applying a fixed set of criteria to the various vehicles continued to
be employed by Tibetans in much later works. The Tibetan scholar Kay thog dam pa bde
all vehicles equally, for the purpose of comparison. TIlls kind of technical
concern with classification systems was a new phenomenon in tantric Bud-
dhism, one unique to the early Tibetan exegetes; Indian commentaries
from the same period simply do not exhibit such elaborate and system-
atic analyses of tantric Buddhism.
m.b. Pelliot tibetain 849
Apart from the two manuscripts just discussed, the only other classifi-
cation system from Dunhuang seems to be the one found in PT849. This
is one of the latest dateable manuscripts found in the famous library cave,
dating from around the tum of the eleventh centuryBl. The system pre-
sented has nine vehicles, but it is quite unlike the standard set we have
seen in other Tibetan works. It provides separate worldly vehicles for the
humans and the gods
, and again separate vehicles for the Sutra Adher-
ents (mdo sde pa) and the Bodhisattvas, which are also usually equiva-
lents. It then mixes up the order of the classes of "outward" tantras to read
Yoga, Kriya, and Ubhaya. The inward classes of Mahayoga, Anuyoga, and
Atiyoga are then dropped, though they do appear as subclasses of the
Yoga vehicle, while Kriya and Ubhaya are also subdivided. The result is
a system that looks like this:
gshegs (1122-1192), for example, used a set of seven criteria in his influential Theg pa spyi
beings, 15: (1) door of entry, (2) view, (3) samiidhi, (4) practices, (5) conduct, (6) duration
of the path, (7) result.
8l A list of the Tibetan kings found in the manuscript traces the royal line through the
Stod mgon gsum and then continues with Btsan po bkra shis stsags pa dpal, Dpal Ide, 'Od
Ide, 'Khri Ide, Btsan po bkra shis mgon po, Tsan po a tsa ra, 'Khri Ide mgon, and Lha cig
cag she (Hackin 1924, 18). Btsan po acarya is probably Ye shes'od (see Karrnay 1998,4).
The penultimate name appears in later discussions of the "men of Gtsang" who restored
the monastic vinaya lineage in central Tibet (see Mkhas pa Zde'u, 391-394). Tibetan sources
vary on how to date this event. Many follow the 978 date suggested by 'Brom ston pa (Vitali
1990,62 n.1), while Mkhas pa lde'useems to suggest 988 (Mkhas pa Zde'u, 394). How to
identify the final name in the list, Lha cig cag she is still unclear, but it may be the son of
'Khri Ide mgon, named Lha chen drag pa in Mkhas pa Zde 'u, 388. Given these identifica-
tions, PT849 would have to be dated to the very end of the tenth century.
82 Though another example of them as separate vehicles, here in the context of a fivefold
siitric classification, can be seen in ITJ526, 2r.2-3: theg pa Znga zhes kyang bya ste! myi'i
theg pa dang!Zha'i theg pa dang! gong rna gsum dang Znga.
1. Vehicle of Humans
2. Vehicle of Gods
3. Sravaka-yarra
4. Pratyekabuddha-yarra
5. Vehicle of Sutra Adherents
6. Bodhisattva-yarra
7. Yoga
a. Yoga
b. Mahayoga
c. Anuyoga
d. Atiyoga
8. Kriya
a. Sravaka Kriya
b. Pratyekabuddha Kriya
c. Siitra Adherent Kriya
d. Bodhisattva Kriya
9. Ubhaya
a. Sravaka Ubhaya
b. Pratyekabuddha Ubhaya
c. Sutra Adherent Ubhaya
d. Bodhisattva Ubhaya
Unfortunately, no explanations are given for how these classes are
being distinguished. Rather than try to make sense of this system, it is
probably better to seeit as the result of the doxographic confusion that
had spread through Tibet by the late tenth century. The essential three-
fold hierarchy of Maha-Anu-Ati is clearly represented in the subclasses
of Yoga, but it is probably unwise to attach too much significance to
the shuffled order of the three principal tantric vehicles. The distinction
between the Bodhisattvas and the Siltra Adherents is not seen anywhere
else in the Dunhuang manuscripts, which all agree that the two terms
should be equivalents. Clearly the author of PT849 was either confused,
working out his own idiosyncratic system, scrambling the order for rea-
sons of secrecy, or some combination of ail three.
Iv. Later Indian Systems
Indian writings from around the same period as PT849 exhibit a sim-
ilarly loose approach to tantric classification systems. This is well attested
in the writings of the famous eleventh century Bengali soholar Atisa, who
arrived in Tibet in 1042. Upon arriving at the royal court in western Tibet,
Atisa perceived a need for an orderly discussion of the entire Buddhist
path, and he composed his influential Bodhipathapradfpa with its auto-
commentary. Within this context we find a presentation of the classes of

But perhaps the most important aspect of Atisa's teachings on the
classes of tantras is simply the fact of their existence. When Padmasam-
bhava visited the Tibetan imperial court in the eighth century, he composed
his Man ngag Ita ba'i phreng ba to show the Tibetans how to classify the
Buddhist tantras. Three centuries later Atisa arrived, and once again the
visiting Indian scholar answered the Tibetans' wish for a discussion of the
classes of tantras. The Tibetan need for classification systems had survived
the intervening "dark period" fully intact.
Atisa presented a range of alternative contemporary classification sys-
tems with four or five classes of tantras, but in his own commentary he
chose to follow a sevenfold scheme:
1. Kriya
2. Carya
3. *Kalpa
4. Ubhaya
5. Yoga
6. Mahayoga
7. *Niruttarayoga
This scheme introduced an unusual distinction between Carya and
Ubhaya, which were interchangeable in most of the earlier schemes

83 See Bodhimiirgadfpapafijikii, 332a.3-333a.2
84 For the past century, the standard Sanskrit reconstruction for the Tibetan Rnal 'byor
bla na med pa has been * Anuttarayoga. This is a time-honoured mistake that needs to be
abandoned. An inspection of the available Sanskrit manuscripts reveals that the Tibetan
more often translates Y oganiruttara. (This observation was confinned for me by Harunaga
Isaacson, to whom my thanks are due.) Thus from this point forward, I will be using *Nirut-
tarayoga where one might expect * Anuttarayoga. As will be argued below, this misreading
has been complicit in allowing the Tibetan origin of the famous fourfold classification
system to remain obscured.
85 The one other source that distinguished these two classes is the Caryiimeliiyana-
pradfpa-niima-tfkii (Q.2703, 324a.3), which argued for a fourfold system of Kriya, Carya,
It alsp added the new class of *Kalpatantra (rtog pa'i rgyuti)86, and tops
. Mahayoga with *Niruttarayoga.
Our confusion is made still worse when we turn to the canonical source
Atisa cites for his system. The Jfiiinavajrasamuccaya is a title that appears
in two canonical versions, a shorter one found in all bka' 'gyur editions
and a longer one found only in the Li thang bka' 'gyur and its descendents

Both versions include discussions of tantric classes that are unusually
detailed for Indian Buddhism. The shorter version presents the following
Ubhaya, and *Niruttarayoga. The traditional attribution of this work to the ninth century
Indian Siikyamitra has recently been shown to be false (see Christian Wedemeyer's paper,
"On the authenticity of the Caryameiiiyanapradfpa, commentary attributed to Siikyami-
tra," presented at the fourteenth conference of the International Association of Buddhist
Studies in London, August 2
, 2005). Wedemeyer's analysis suggests the work is rather
. a Tibetan composition probably dating from the eleventh century at least.
The scriptural source Atisa cites for his own system, the Jfidnavajrasamuccaya, pro-
vides the following explanation for why these two classes of Carya and Ubhaya are dis-
tinguished: "One who accomplishes a thorough analysis of the ten aspects of suchness -
that which is arrayed as the m3I).Qala wheel that is coemergent with the knowledge-god-
desses, as well as the pratices, activities, and accomplishments and so forth that arise from
that ~ such a person is engaged in the Ubhaya tantras. The one who accomplishes, together
with that which is accomplished, through a detailed analysis of perceptible characteristics,
such as the mudras from the Kriya tantras whic.h are for extensively performing the various
activities, is engaged in the Carya tantras" (Jfidnavajrasamuccaya, 293b.4-6). This passage
seems to indicate that the Ubhaya tantras are higher than the Carya because they focus more
on the mandala as a manifestation of suchness than on external forms. This would seem
to agree with the order seen in Siikyamitra's sytem. In both cases it seems that Ubhaya was
playing the role normally played by Yoga.
This latter point is confirmed by another tantra under a similar title, the Srfjfidnava-
jrasamuccaya (missing from the Peking Bka' 'gyur, but in the Oerge at 0.450). According
to Mimaki Katsumi, this is a "slightly later version" (Mimaki 1994, 122 n.17), and here we
see a fivefold scheme for which the fourth class appears both ways: (1) *Kalpa, (2) Kriya,
(3) Carya, (4) Ubhayii/Yoga, (5) *Niruttarayoga (see 0.450, lOb.4-5).
86 Eimer's study of this class (Eimer 1993) concludes that *Kalpa tantra teaches mun-
dane rituals for healing and gaining magical powers, and that in this sense, "it does not direct
the performing adept to any spiritual level" (Eimer 1993, 228). The mid-twelfth century
Sa skya scholar, Bsod nams rtse mo, is in general agreement when he writes, "Given that
the *Kalpa tantras principally teach outward activities, the class of *Kalpa tantras essen-
tially are gathered within the Kriya tantras" (Rgyud sde spyi'i rnampar gzhag pa, 33b.5:
rtog pa'i rgyud kyang phyi'i bya ba gtsor ston pa tsam ia dgongs nas rtog pa'i rgyud du
phye ba ste ngo bo bya ba'i rgyud du 'du'o). Why Atisa places the *Kalpa tantras above
Kriya and Carya remains unexplained.
81 See Eimer (1993), 226.
1. *KaIpa
2. Kriya
3. Carya
4. Ubhaya
5. Mahayoga
The system found in the tantra's longer version is identical to the shorter
except that it gives *Niruttarayoga in place of Mahayoga. When these
two canonical systems are compared with the one taught by AtiSa, we see
that Atisa added Yogatantra and included both Mahayoga and *Nirutta-
A number of conclusions can be drawn from all this: First, it seems
clear that tantric classification systems remained highly flexible in India
through the eleventh century. The famous fourfold scheme was by no
means common to all Indian Buddhists. Second, we may suggest that
Atisa's system included *Niruttarayoga in addition to Mahayoga in order
to bring the Jfiiinavajrasamuccaya into line with the latest terminological
developments. In this sense, Atisa's scheme reflects the gradual eleventh
century rise of *Niruttarayoga as a class apart from Mahayoga.
By the eleventh century a need for a new distinct category was start-
ing to be felt. We have observed that during the eighth and ninth centuries,
new tantric classes arose in large part to distinguish new developments
in ritual practice. Now the new class of *Niruttarayoga was surfacing
for similar reasons of ritual technique. In Atisa' s discussion of Miiliayoga
and *Niruttarayoga, he lists which tantras belong to each class, and it is
certainly relevant that many of his *Niruttarayoga tantras did not enter
- Tibet until the later diffusion (phyi dar) period
Generally speaking,
these appear to be later works. What distinguished these later works from
the slightly earlier Mahayoga tantras? Atisa's *Niruttarayoga tantras tend
to incorporate more complex subtle body (rtsa rlung) systems than his
88 Jfuinavajrasamuccaya, 293a.6-293b.S. Note that in the scripture itself the hierarchy
of classes is presented in reverse order, starting with the highest class and ending with the
89 As Mahayoga tantras he names the Guhyasamiija, the Candraguhyatilaka, the Krlfna-
yamiiri, the Paramiidya, the Vairocana Miiyiijiila, and so forth, while under *Niruttarayoga
tantras he includes the Khasama, the Cakrasalflvara, the Vajrarjiika, the Hevajra, and so
forth (Bodhimiirgadlpapanjikii, 332b.4-S).
generally earlier Mahayoga tantras. It seems that the new class of *NinIt-
tarayoga was needed to recognize this recent elaboration of the subtle body
ritual systems
The beginnings of this new class can already be seen in Indian writings
of the late tenth century. The earliest dateable example may be the late tenth
century writings of Sraddhakaravarma, the Indian scholar who worked with
the famous Tibetan translator RiD chen bzang po. In that scheme, how-
ever, the highest division was still Mahayoga:
There are four doors for entering into the Secret Mantra, the fruition that is
the Vajrayana. These are generally known as Kriyatantra, Caryatantra, Yoga-
tantra, and Mahayogatantra

Though Mahayoga still held the highest place, Sraddhakaravarma goes
on to subdivide the class, and here the term *Niruttarayoga appears as the
final and highest subdivision. The purpose of the new term seems to be
to identify the generally later tantras represented elsewhere as Y oginI tan-
tras, that is, those more focused on female deities:
Mahayoga consists of two types, the natural tantras and the tantras for exam-
ination ... The tantras for examination consist of a further two types, the
90 The two classes of Mahayoga and *Anuttarayoga may also reflect another, non-
chronological distinction in the development of Buddhist tantra. In addition to *Nirut-
tarayoga reflecting a generally later stage of ritual development, the two classes may also
represent two lines of development that developed alongside each other. While this the-
ory remains highly impressionistic, it seems to me that from the original matrix of STTS
and other Vajra-sekhara tantras, we may be able to distinguish two lines of development.
The first would have passed through the Guhyasamaja, to Guhyagarbha and Mayajtila,
while the other passed through the Sarvabuddhasamayoga, to Cakrasa1J'lvara and Heva-
jra. The latter were known as yogin! tantras, apparently because their mlU).<j.alas consisted
of female deities surrounding the central figure. The Tantra, though not usu-
ally termed a YoginI !antra, may also fit into the latter line of development, as it has female
deities. This line may also be marked by its early reliance on the Yoga-Anu-Ati-Maha
system discussed above (see n. 68). Certainly there would be significant cross-pollination
between these two lines of development, and further research is required to confirm such
a theory, but it may also provide a starting point for understanding the historical roots of
the distinction between the so-called father and mother tantras.
91 Yoganuttaratantrarthavatarasa1J'lgraha, I 17a.6. gsang sngags 'bras bu rdo rje theg
pa la ni 'jug pa'i sgo mam pa bzhi ste! bya ba'i rgyud dang! spyod pa'i rgyud dang! mal
'byor gyi rgyud dang! mal 'byor chen po'i rgyud ces spyir grags pa yin no. Note this title
should probably be reconstructed as Yoganiruttaratantrarthavatarasa1J'lgraha.
tantras for the yogas of means and the tantras for the ,yo gas of wisdom.
These are also asserted as the tantras of the supreme yoga and the tantras
of *Niruttarayoga92.
Sraddhakaravarma describes the latter *Niruttarayoga tantrasas those
with maI,1qalas populated by female deities. The equivalence of *Nirutta-
rayoga and Y oginI is confrrmed elsewhere by Abhayakaragupta in his own
discussion of a fivefold system, where he explains that, "the *Niruttara-
yoga tantras are the YoginJ tantras."93
In the writings of Sraddhakaravarma, one can see just
beginning to emerge, though still as a subclass of Mahayoga. Soon after,
*Niruttarayoga detached completely to become its own independent cat-
egory above Mahayoga. Thus in the writings of Ratnakarasanti, who lived
around the turn of the eleventh century, we see a fivefold system contai-
ning Kriya, Carya, Yoga, Mahayoga, *Niruttarayoga
. In including both
Mahayoga and *Niruttaraya, this scheme resembles the one Atisa taught
during his mid-eleventh century visit to Tibet, and in fact similar fivefold
systems appear in a variety of sources from around this period. Some of
these schemes replace the fourth element, Mahayoga, with Yogottara (Tib.
mal 'byor bla rna), but these were generally considered equivalents. Thus
in another work by Ratnakarasanti, we see Kriya, Carya, Yoga, Yogottara,
. We can conclude that, taken together, these eleventh
century systems represent a period in Indian tantric development when the
new class of Y oginI/*Nimttarayoga tantras was being added as the top class,
above and distinct from the earlier Mahayoga/Y ogottara tantras such as
v'92 Yoganuttaratantrarthiivatarasarrzgraha, 118a.5-118b.( mal 'byor chen po'i rgyud
ni rnam pa gnyis tel rang bzhin gyi rgyud dang btags pa'i rgyud dol ... btags pa'i rgyud
kyang rnam pa gnyis tel mal 'byor thabs kyi rgyud dangl mal 'byor shes rab kyi rgyud
dol de dag kyang rnam pa gnyis su 'dod del mal 'byor mchog gi rgyud dangl mal 'byor
bla na med pa'i rgyud ces bya ste.
93 Amnayamafijarf, 121a.5. Abhayakaragupta's own classification system in this work
reads as follows: Kriya, Carya, Yoga, Yogottara, *Niruttarayoga. As discussed below, this
system was common among eleventh century Indian H evajra commentators such as Ratna-
karasanti and Kanha.
94 115a.3-4.
95 Muktavalf, 332b.3-5 and 347a.6-7. For the Sanskrit see Tripathi (2001), 169 and 223.
Note that on page 223 the editors mistakenly give "Yogantara" as the fourth vehicle, despite
the fact that the manuscripts clearly read "Yogottara" (my thanks to Harunaga Isaacson
The same state of affairs is observed in the writings of
carin. Kat;tha probably dates to the early eleventh century, making him
roughly contemporary with (or perhaps slightly later than) Ratnakara-
. In Kat;tha's GuhyatattvaprakiiSa, we see what at first looks like the
later Tibetan fourfold system, only with the Y oginI tantras in the fourth
and highest spot
. The trouble is that here Yoga corresponds to the so-
called "father tantras" such as Guhasamiija and not to the Yoga tantras
such as the STTS and so on
. For this reason it is not a very clear example
of the standard fourfold system, which typically classifies Guhyasamiija
and so forth under *Niruttarayoga.
In Klli)ha's other major work, his.Y ogaratnamiilii commentary on the
H evajra, we find another passage in which the fourfold system again seems
to be present. According to the Tibetan translation, the scheme should read:
Kriya, Carya, Yoga, *Niruttarayoga (mal 'byor bla na med pa). "This
Hevajra Tantra," we are told, "is a *Niruttarayoga tantra."99 Yet when
the Sanskrit is consulted, a very different picture emerges. The fourth
class, for which the Tibetan reads "mal 'byor bla na med pa," in the
Sanskrit reads "Yogottara. "100 And this is quite apart from a fifth class
for bringing this to my attention). Another example of Mahayoga in the fourth position is
found in the Srfvajramiiliimaht'iyogatantratfkii, 4b.l, which lists the five classes of Kriya-
96 The person of KaI)ha, let alone his dates, is difficult to pin down. There are at least
two KaI)has, one associated with the Cakrasrunvara system and the other with Hevajra. The
latter Hevajra KaI)ha appears in the early Sa skya lineage lists as the teacher of Gayadhara,
who in turn worked with 'Brog mi 10 tsa ba in the mid-eleventh century (Rgyud sde spy-
i'i rnam par gzhag pa, 72b.5-73a.l). Later sources push KaI)ha further into the past, as is
common for such siddha characters, but for precisely this reason we should probably fol-
low the later date, that is, early eleventh century. This allows for the possibility that KaI)ha
may, after all, have been the same PaI).Qita who "co-operated in the task of trans-
lating the Yogaratnamiilii into Tibetan" (Snellgrove 1959, 13-14 nA).
97 Guhyatattvaprakiisa, 282aA-5. rgyud ni rnam pa bzhir 'gyur tel bgad pa dang ni bltas
pa dang/ de bzhin bzhan ni lag bcangs dang/ gnyis gnyis sbros pa bzhi pa ste/ de dag rgyu
ni rnam bzhi/ 'dus pa 'ang rnam pa gnyis gsungs tel bskyed dang rdzogs byung rim pa'o/
rnal 'byor bskyed par bstan pa ste/ rdzogs pa rnal 'byor ma du brjod.
98 KaI)ha' s Y oga-Y ogiuI distinction here is similar to the one made in chapter three of
the J)iikinfvajrapafijara Tantra (289bA-5), itself an explanatory tantra to the Hevajra.
In fact, this is the most common fourfold classification system in Indian texts, especially
around the eleventh and twelfth centuries.
99 Yogaratnamiilii,49b.8-50a.5.
100 In Snellgrove'S edition of the Yogaratnamiilii (Snellgrove 1959, voL 2, 142 n. 6),
he notes that the actual manuscript reads "kvacid yottariidau." Snellgrove speculates that
of Niruttarayoga. Why the Tibetan translator, Mgos k h ~ g pa lhas btsas,
conflated Y ogottara and Niruttarayoga by translating both with the same
mal 'byor bla na med pa remains a mystery, particularly given that he
translated them differently later in the very same work as mal 'byor bla
na med pa and mal 'byor gong na med pa respectivelylDl. .
V. Sectarian Closure and the Tibetan Formation of the Four Classes
In Tibet, the Hevajra exegetical tradition of which Kar;tha was a crucial
member went on to become the preserve of the early Sa skya writers. We
have seen that a variety of classification systems with four, five, or seven
classes continued to be taught by Indians through the eleventh century, and
it seems the four classes of tantras that are so well known today began
to dominate Tibetan Buddhism only with the early Sa skya pa.
A brief discussion of the fourfold scheme appears, for example, in the twelfth
century introduction to the tantras by Sa chen kun dga' snying po (1092-
1158), the first Sa skya patriarch
Thus the rise of the standard classifi-
cation system seems to have been tied to the rise of the Sa skya school.
The formation of the classic fourfold scheme should therefore be under-
stood against the historical backdrop of twelfth century Tibet. This was
a time of intense competition between Tibetan clans, and the tantras were
central to these contests, able to bestow wealth, power, and prestige upon
this should read" kvacid anuttaradau," but the fivefold system followed in the Yogarat-
namala is clearly Kriya, Carya, Yoga, Yogottara, and Yoganiruttara, as is made explicit
on Snellgrove (1959), vol. 2, 156. The Tibetan for the latterreference is fof once clear (Yoga-
ratnamala, 70a.1-2: sngags kyi theg pa kun zhes bya ba ni bya ba dang! spyod pa dang!
rnal 'byor dang! rnal 'byor bla na med pa dang! rnal 'byor gong na med pa'i dbyed bas
rnam pa lnga'o). For an English translation of the two passages, see Farrow (1992), 183-4
and 274.
101 It is unclear just how extensive the confusion was, but Mgos khug pa's was not an
isolated case. The {Jakinfsarvacittadvayacintyajfiana-vajravarahi Tantra, which was trans-
lated by Gayadhara and Jo zla ba'i 'od zer, provides the same fivefold system, and here
again the fourth term (almost certainly *Uttaratantra) was translated as bla na med pa'i
rgyud (see 88a.5-6: kri ma'i rgyud sde rnam lnga bstan! spyod pa gtso bor byed pa la!
gnyis pa'i rgyud sde rnam lnga bstan! sems nyid gtso bor byed pa la! rnal 'byor rgyud
sde rnam lnga bstan! bde stong gtso bor byed pa la! bla na med pa'i rgyud lnga bstan!
yid la byed pa med rnams la gong na med pa'i rgyud lnga bstan! de'i phyir rgyud sde rnam
pa lnga! gsang sngags rgyud ces bya bar grags).
102 Rgyud sde spyi'i rnam bzhag chung ngu, 8b.5.
whomever controlled their teachings. Under these socio-political pres-
sures, Tibetan Buddhism was forced into ever more defined forms, and
.the array of classification systems seen during the ninth and tenth centuries
became unworkable. Polemics levelled against the Rnying ma pa were
particularly fierce (and the above-mentioned translator Mgos khug pa was
notably renowned in this regard), and the nine vehicles system from
the Compendium of Intentions Sutra quickIy became ubiquitous among
followers of the Rnying ma school.
Meanwhile the followers of the new (gsar rna) schools, such as the
Sa skya pa, required a legitimate system that reflected the latest Indian
developments. From this period, a particularly influential Tibetan dis-
cussion on the classification of tantras appeared in the General Presen-
tation of the Tantras (Rgyud sde spyi'i rnam bzhag pa) by the second Sa
skya hierarch, Bsod nams rtse mo (1142-1182)103. There the author sur-
veys the various Indian schemes circulating at the time and argues that
they all boil down to the same fourfold system: Kriya, Carya, Yoga, and
*Niruttarayoga. He then subdivides his highest class of *Niruttarayoga into
father tantras, mother tantras, and nondual tantras.
This marks the end of the process that had begun in the second half
of the tenth century in works like those of Sraddhakaravarma. There,
Mahayoga was initially subdivided with *Niruttarayoga as the highest
subdivision. Next, *Niruttarayoga split off completely to form its own
class alongside Mahayoga, as seen in the early eleventh century writings
of Ratnakarasanti and Km;lha. Now in the twelfth century, we see that the
rise of *Niruttarayoga was complete, that the earlier category of Mahayoga
had been subsumed under the class it had itself spawned.
It is significant, however, that Bsod nams rtse mo does not cite any
Indian sources for his own system. Rather, he bases his argument on the
existence of other parallel groups of four, specifically the four tantric inti-
ations and the four varlJas of Brahmanical Indian societyl04. To see how
later Tibetans traced their fourfold system to Indian sources, we tum to
perhaps the singlemost influential presentation of Buddhist tantra in
103 See Rgyud sde spyi'i rnam bzhag pa, 30hA-40h.1. The main section on the four
classes appears on 33aA-36aA
104 See Rgyud sde spyi'i rnam bzhag pa, 34h.2-4.
Tibetan history: the Sngags rim chen ma by Tsong khR pa (1357-1419),
the founder of the Dge lugs pa school. Here the standard four classes are
identified in a series of Indian scriptures
Upon closer examination,
however, none of the sources given by Tsong khapa prove defInitive. His
strongest pieces of evidence are in fact those we have already examined
- the Indian commentaries of Sraddhakaravarma, Abhayakaragupta, and
Ratnakarasanti - and as we have seen, none of these authors employed
the system we know so well.
Tsong kha pa repeatedly refers to the passages from the eleventh
century H evajra exegetical writings that align the first four classes with
the four metaphors of laughing, looking, embracing, and sexual union. Yet
he neglects to mention the existence of a fifth higher class that appears
in each of his sources and transcends all of the other four
. Tsong kha
pa's conflation of the fourth and fifth classes of Yogottara and *Nirut-
tarayoga is not entirely surprising, since, as we have seen, the translations
he was working from sometimes made the same mistake
Whether these
misreadings were deliberate or not, the result was the same: later genera-
tions of scholars, Tibetan and western alike, came to believe that the four-
fold system originated in India.
The obscuration of the four classes' Tibetan origin has been allowed
to continue by a further inaccuracy that has been common among west-
ern scholars for the past century. We have noted above (n. 83) that west-
ern scholars have long reconstructed the Tibetan term mal 'byor bla na
med pa as * Anuttarayoga. In fact the corresponding Sanskrit term seen
in the available manuscripts is Niruttarayoga. Now we can see how this
mistake has been complicit in concealing the Tibetan origin of the famous
classification system. Once Niruttarayoga (mal 'byor gong na med pa)
had been mistaken for * Anuttarayoga, just one small step was required
105 For an English translation ofthe relevent section, see Tsong-ka-pa (1977),151-164.
106 For the relevant passages see Abhayakaragupta's .Amnayamafijarf, 253b.8-254a.3
(cited on Tsong-ka-pa 1977, 158), and Ratnakarasanti's Muktavalf, 332b.3-5 (cited on
Tsong-ka-pa 1977, 159).
107 Again, examples would be Kru:ma's Yogaratnamiila, 50a.3-4 (compare the Sanskrit
at Snellgrove 1959, 1421. 35 - 143 1. 1), and probably the l!akinfsarvacittadvayacintya-
jfiana-vajravarahi Tantra, 88a.5-6. In both cases, the Sanskrit yogottara is translated with
rnal 'byor bla na med instead of rnal 'byor bla ma.
for scholars to overlook the comparatively minor, though crucial, mor-
. phological difference between the fourth and fifth classes of mal 'byor
bla ma'j rgyud and mal 'byor bla na med pa'i rgyud, precisely the same
difference obscured by earlier Tibetan writers in their own justifications
of the fourfold system
. In this way the western inaccuracies have com-
pounded the earlier Tibetan ones.
VI. Concluding Remarks
By now it should be clear just how Tibetan the four classes of tantras
really are. We have seen that the fourfold system that is so well known
among modern scholars was by no means the focus of Indian Buddhists
of the eleventh century. Rather the decisive system emerged in Tibet,
probably during the twelfth century, just as the distinction between the
old Rnying rna and the new Gsar rna schools was crystallizing. Of course
in India such "old" vs. "new"distinctions were unknown, and the devel-
opment of tantric Buddhism was more of a continuum, allowing com-
peting classification schemes to coexist in relative peace. In Tibet, how-
ever, the supposed decrepitude of the old was losing out to the allure of
the modern. The Rnying rna pa clung to the nine vehicles system that
was by now well known in Tibet, while the Gsar rna pa constructed a new
fourfold scheme consisting of Kriya, Carya, Yoga, and *Niruttarayoga
tantras. This was a fairly simple system compared to the elaborate ones
seen in Tibet during the ninth and tenth centuries. In adopting *Nirutta-
rayoga, the neW schools gave precedence to the latest tantric ritual systems,
but in excluding the classes of Mahayoga, Anuyoga, and Atiyoga, they
dismissed the last three centuries of ritual development. The gradual
shaping of ritual technologies that had occurred from the eighth to tenth
centuries was thus obscured, as were the tentative Tibetan attempts to
make sense of those shifting sands.
Its Tibetan origin having been identified, the fourfold classification
system must be understood as a reflection of inimitably Tibetan concerns,
108 Thus for example (as noted above), Snellgrove'S apparently benign emendation of
the corrupt Sanskrit "yottaradau" to "anuttaradau" rather than "yogottaradau" (Snellgrove
1959, vol. 2, 1421. 35). This error was then repeated in Farrow (1992), 184.
as a result of the uniquely Tibetan doxographic tradition. Despite the wide
variety of classification schemes that both India and Tibet
between the eighth and twelfth centuries, the fundamental concerns
reflected in those systems remained remarkably stable. In India it was
about ritual. Buddhaguhya's early writings distinguished the outward rit-
uals of the Kriya tantras from the inward rituals of the Yoga tantras. Over
two centuries later, Sraddhakaravarma and KaI.J.ha were busy distin-
guishing the *Niruttarayoga tantras by their focus on the perfection stage
subtle body practices and the use of maI.J.9.alas populated by femme deities.
The ritual technologies had changed over the intervening two hundred
years, but the basic Indian concern with ritual remained largely intact.
This Indian ritual focus stands in marked contrast to the Tibetan
classification systems from the same period that show a far stronger inter-
est in differences of doctrine, in the philosophical views and tantric cos-
mologies. The Tibetan concern with doctrinal classification systems likely
resulted from the greater urgency these schemes had for Tibetans. In India
tantra developed gradually, as an organic part of society as a whole, but
in Tibet the tantric teachings arrived en masse, as a foreign intrusion of
chaotic texts and rituals. Indians, who stood at the origin of the Buddhist
religion, could more easily justify new developments in tantric ritual.
Tibetans, who dwelt beyond the edges of this original universe, were con-
fronted by the entirety of Buddhism all at once. This foreign religion
demanded justifications and explanations in a way that was simply unnec-
essary in India, and classification systems provided them. The four classes
formed a tidy doctrinal package that could be tied to the four initiations,
the four blisses, the four varf}as, the four metaphors of laughing, look-
ing, touching, and sexual union, and so forth. In this way Tibetans [mally
gained doctrinal closure on the chaotic proliferation of Indian tantra.
Through doxography, they organized Buddhism into a single totality that
could be tamed and converted toward new ends.
IOL TIB J 644: Translation
ill the deity system of the Sravakas, only Siikyamuni is seen as a buddha. They
regard Vairocana merely as a mahii-upiisaka, seeing him as a great virtuous
friend. They regard the directional guardians merely as devoted practitioners.
They see the difference [between teacher and disciple] as that between the Bud-
dha and sentient beings. The measure of their accomplishment is as follows:
With regards to the afflictions of the three realms, the Sravakas have ten subtle
[contaminentsp09 to be abandoned through meditation (bhiivaniiheya). In the
desire realm there are eighty-eight [contaminentsl If eighty-seven of those are
abandoned, one attains entry into liberation. Following that, if one abandons all
ninety-eight [contaminents] to be abandoned, one is called an "arhat." After that,
if one lets go of one's own view, one becomes a transformational Sravaka and
. treads the first bodhisattva level of "the absolute enemy." And following that, it
is asserted, buddhahood will be accomplished after three countless aeons. One who
is on this first level of "the absolute enemy" gains foreknowledge of what has
yet to come. At that time, there are magical displays including hearing the dharma
of a hundred buddhas in a single instant, liberating a hundred sentient beings,
absorption in samiidhi, and sending forth a hundred emanations. From that point
onwards, one moves through each of the bodhisattva levels.
The measures of accomplishment for the Pratyekabuddhas are like those of the
The deity system of the [Mahayana] Satra Adherents is regarded as the three bod-
ies. They see the difference [between teacher and disciple] as that between buddhas
and sentient beings. They too accomplish buddhahood in three countless aeons.
[1 v] The deity system of Kriya mantra is regarded as the protectors of the three
families together with their retinues. The difference [between teacher and disci-
ple] is seen in the manner of slave and master. Their measure of accomplishment
asserts accomplishment within one lifetime.
Upaya [tantra] recognizes four families, recognized as the four families of vajra,
ratna, padma, and karma. The difference [between teacher and disciple] is seen
109 This is a reference to the ten anusaya (Tib. bag la nyal ba). Cox (in Willeman,
Dessein, and Cox 1998, 31 n.150), lists the ten and adds that they are, "a V a i b h a ~ i k a view
point. This is contrary to the Sautrantika viewpoint, according to which there are only
eight contarninents." The anusaya were used to refer to the subtle (anu-) seeds from which
the defIlements (klda) reemerge after a period of interruption. On organizing the ninety-
eight contarninents into three realms, see De la Vallee Poussin (1988), vol. 4, 9 n.2, and
Van den Broeck (1977), 62-65.
as that between lord and servant. Their measure of accomplishment asserts accom-
plishment in half a lifetime. '
The deity system of Yoga is regarded as the four bodies: the dharmakaya Vairo-
cana, of the svabhiivikakaya (lit. "body of resting in the enlightened
essence"), Ratnasambhava and Amitabha of the salJ1bhogakaya, and Anioghasid-
dhi of the nirmalJakaya. The difference [between teacher and disciple] is seen
merely as that between a brother and a sister. Their measure of accomplishment
asserts the appearance of certain signs, followed by accomplishment.
The deity system of Mahayoga regards the five families as a single means 110.
There is no difference [between teacher and disciple], that is, they see them as
the same. Regarding the measure of their accomplishment, they assert accomplish-
ment through understanding, through the realization of, or "union with," primor-
dial Being, or "authenticity."
The view and the deity system of Anuyoga are the same. Nor is there any dif-
ference [between teacher and disciple]. Regarding the measure of their accom-
plishment, they assert a spontaneous accomplishment upon emergence.
The view and the deity system of Atiyoga are also the same, and there is no dif-
ference [between teacher and disciple]. Their measure of accomplishment is seen
as spontaneous accomplishment. This teaching on classifying the deity systems
and the measures of accomplishment is complete.
[2r] Classifications of the vidyadharas: There are three vidyadharas of Kriya:
the vidyadharas of accomplishment, the vidyadharas who dwell on the levels, and
the vidyadharas of spontaneous accomplishment.
A vidyadhara of accomplishment meditates on [that which is] endowed with the
three great reflections
. One assembles the necessities, gathering the [requisite]
causes and conditions. Spring is the cause and autumn the result. Those two times
are then subdivided and the days counted. A mar kham dating system is applied,
with fifteen days at the beginning [of each month], fifteen days at the end, the
new moon, the eighth day of the first half, and the eighth day of the second half.
The planets too, Jupiter or "bung rnyil ba," and the king of constellations should
be in position. The necessities and the implements of the deities should be in
accordance with the textual systems. One's food and clothing should also be like-
wise, and having assembled everything, one performs the worship. One is called
a vidyadhara of accomplishment.
110 This description of Mahayoga is common in the Dunhuang collections. See for
example ITJ436, lr.l (de la rna ha yo ga'i [hal rigs la tshul gcig du Ita ba gang zhe na).
See also PT656, r24 (translated below).
III For a discussion of these three, see PT656, r9-12 (translated below).
Then Vajrapfu,ri arrived and granted the siddhis. Then he went to the Asura Cave,
and upon beholding the visage of an emanation of Vajrapfu,ri present there, he
struck the rock with his foot. It seemed as if he had stuck it into dough. From
that footprint the sacrament (samaya) descended, and from within that there came
a spring with eight streams. One flowed to the south face of Mt. Meru, so the
spring was called AsvakarI}.a. Seven of them fell inside of the Asura Cave. In this
[spring] he cleansed himself and gained accomplishment. Thus he was called a
vidyadhara who dwells on the levels.
A vidyadhara of spontaneous accomplishment is equal in status to glorious
Vajrapfu,ri. Such a one is called a "second buddha," noble Vajrapfu,ri.
A vidyadhara of Yoga is called a "beautifully ornamented second buddha."
A vidyadhara of Mahayoga is called a "vajradhara buddha." Here there are four
further kinds: the deity vidyadhara, the medicinal vidyiidhara, the vidyadhara of
maturation, and the mahamudra vidyadhara
Of those, a deity vidyadhara gains
accomplishment through the deity, a medicinal vidyadhara gains accomplish-
ment through extracted nectars and so forth, a vidyadhara of maturation gains
accomplishment from one who is the highest of experts, and a mahamudra vidya-
dhara is endowed with the five kinds of omniscience, the five miraculous powers.
[3r] This one too is called a "second buddha," and should be understood as such
in everything he does. There is no difference between a second [buddha and the
first]; they are equal. One only says "second" because [they used] different ways
to gain accomplishment.
There are no vidyadharas (lit. "knowledge holder") in Anuyoga and Atiyoga;
there is not even any knowledge to hold! Whoever might be called a vidyadhara
gains accomplishment in anything. Such a one should be understood as part of
the ultimate truth, a "vidyadhara on the levels."
Thus there are sixteen vidyadhara levels in all. Anything said to surpass these
contradicts the three kinds of valid cognition and does not appear in the scrip-
tures and the tantras. There are also sixteen between the ten bodhisattva levels
and the six buddha levels. These sixteen levels are not different [from the sixteen
vidyadhara levels], yet it is unsuitable to explain the various pith instructions as
all the same. The levels are distinct and completely perfect. They have been
explained here in terms of the general levels and Vajrasattva and the seven gen-
eral scriptural systems (spyi lung bdun). Their distinction allows the truth to be
ascertained individually. Their complete perfection means they are pervaded or
gathered by that [truth).
112 Note these differ from the four levels seen in the later Rnying rna tradition. Compare
for example Tsogyal (1993), 294-5: (1) maturation, (2) longevity, (3) mahdmudra, (4) spon-
taneous perfection.
[3v] The [Candra-]Guhyatilaka
was translated by JfianaIIJitra and Brang ti
A.ciirya Jayagoca
. The Sarvabuddhasamayoga
from the collection of a
100,000 [verses]1l6 was translated by A.carya Jfiiinagarbha and Vimalarnitra
IOL TIE J 644: Transcription
[lr] de la nyan thos kyi lha rgyud ni shag gya thub pa gci bu sangs rgyas su Ita!
roam par snang mdzad rna ha u pa si ga tsam du Ita dge bsnyen chen por Ita!
phyogs skyong roams mos pa spyod pa tsam du Ita! khyad par ni sangs rgyas dang
sems can du Ital grub tshad nye ring nil nyan tos kyisl khams gsum na nyon
pa bsgom pas spang bar bya ba phra mo bcu yod pa lal 'dod khams pa'i spang
bar bya brgyad cu rtsa brgyad yod pa lasl brgyad cu rtsa bdun spong pa nal thar
pa la 'jug pa thob bol
de nas yang nyon mongs pa dgu bcu tsa brgyadl ril spong pa nal dgra bcom pa
zhes byal de nas yang rang gi Ita ba bor nal 'gyur ba'i nyan tos tel byang cub
sems pa'i sa dang po rab du dgra ba non nas de nas yang bskal pa grangs myed
pa gsum gis sangs rgyas su grub par 'dod dol sa dang po rab du dgra ba zhes bya
sngon cad yong rna myong ba byal de'i tshe cho 'phrul nil skad cig rna gcig Ia!
sangs rgyas brgya pa chos nyanl sems can brgya sgrol/ ting nge 'dzin la snyoms
par 'jugl sprul pa brgya 'gyedl de yan cad kyangl byang cub kyi sa re res gyung
pa yod dol rang byang cub kyi grub tshad ni nyan tos dang mthun/
mdo' sde'i lha rgyud nil sku gsum du Ital khyad par ni sangs rgyas dang sems
can du Ital de yang bskul grangs myed pa gsum gis sangs su grub bol
113 See Q.l11, Sri-candraguhyatilaka-niima-mahiitantrariija.
;>114 Brang ti Jayagoca is a little-known figure. According to the Tshig mdzod chen mo,
he was an expert in Tibetan medicine and served under the Tibetan king Khri Ide gtsug
brtsan, a.k.a. Me ag tshom (704-754/55). The Brang ti clan continued to play an important
role in later Tibetan history.
115 See Q.8, Srf-sarvabuddhasarnayoga. Note that neither of these translation attributions
match those provided in the colophons to the more recent Peking edition of the canon, nor
those found in the Mtshams brag edition of the Rnying rna rgyud 'bum.
116 Here we see yet another example of how scholars of early tantra explained the exis-
tence of multiple recensions of a given tantra by resorting to a mythical ur-text from which
they are extracted. As noted above (notes 15 and 28), other examples include the Vidyii-
dhara-pitaka, the Vajrasekhara, the Miiyiijiila, and the Vajrakflaya.
117 These two titles often appear together in the Dunhuang documents. See for example,
PT332e, Ir, and PT28l, v7.4-r8.1. Apparently they travelled together, at least in Tibet.
[Iv] sngags kri ya'i lha rgyud nil rigs gsum gf mgon po 'khor dang bcas par
lta'oi khyad par dang rjo bo'i tshul su lta'o/ grub tshad ni tshe gcig gi
grub par 'dod dol u pa ya rigs bzhir Ital rdo Ije dang rin po che dang/ pad ma
dang/ las kyi rigs dang bzhir lta'o/ khyad par ni dpon g.yog du Ita'o/
grub tshad nil tshe phyed kyis grub par 'dod dol yo ga'i Iha rgyud ni sku bzhir
ltar/ chos kyi sku mam par snang mdzad/ a sho bya snying po byang cub na
pa'i sku/ rin cen 'byung ldan dang snang ba mtha yas long spyod rdzogs pa'i sku!
don yod par grub sprul pa'i skiI/ khyad par ni spun dang grogs tsam du Ital
grub tshad ni stags byung nas grub par 'dod dol ma ha yo ga'i lha rgyud nil rigs
lnga tshul gcig par Ital khyad par myed de gcig par lta/ grub tshad nil
ye nas de yin ba ni mal ma zhes byal rtogs pa ni 'byor pa zhes bya ste shes pas
grub par 'dod dol a nu yo ga'i Ita ba dang lha rgyud ni gcig ste/
khyad par myed dol grub tshad nil yong nas lhun kyis grub par 'dod dol a ti yo
ga'i Ita ba dang lha rgyud gcig ste khyad par myed/ grub tshad nil
lhun gyis grub par Ita'o/ lha rgyud dang grub tshad nye ring bstan pa rdzogs sol
[2r] rigs 'dzin dbye bal kri ya'i rigs 'dzin gsum ste grub pa'i rigs 'dzin dang/ sa
la gnas pa'i rigs 'dzin dang/ lhun kyis
grub pa'i rigs 'dzin no/ de la grub pa'i rigs 'dzin pa nil gzugs bmyan chen po
gsum dang ldan bar sgom ba yang/ yo byad dang rgyu
skyen tshogs 'dus tel dpyid ni rgyu yin! ston ni 'bras bu yin tel dus de gnyis tel
tshigs ni tshes grangs yar gyi bco lnga dang! mar
bco lnga dang/ gnam stong dang/ yar gi tshes brgyad dang! mar gi tshes brgyad/
mar kham/ tshes grangs de dang sbyar ro/ gza yang bur bu bur myil ba zhes
dang/ rgyu skar kyi rgyal po rgyalla bab pal yo byad dang lha cha gzhung dang
mthun bal zas dang gos kyang de dang 'dra zhing ril tshigs nasi bsgrub
pa byed pa nil sgrub pa'i rigs 'dzin pa zhes bya'o/ de nas phyag na rdo rje gshegs
nasi dngos grub sbyin ba dang/ a su ra'i brag phug du phyin
pa dang! de na phyag na rdo rje'i sprul pa gcig bzhugs pa'i zhal mthong nas brag
la rkang pa gcig brgyab pa dang! zan la brgyab bzhin snang ilgo/
rjes de nas dam babs nasi nang de na chu myig yan lag brgyad dang ldan ba
brgyad yod pa la/ gcig ni ri rab kyi lho ngos su rdol te
[2v] chu myig rta rna zhes bya'o/ bdun a su ra'i nang na 'bab pa la khrus byed
cing bsgrub pa del sa la gnas pa'i rigs 'dzin ces bya'o/
lhun kyis grub pa'i rigs 'dzin nil dpal phyag rdo Ije dang skal pa mnyam ba ste/
de'i mying nil dpal rdo Ije 'dzin kyi sangs rgyas gnyis pa zhes bya'o/
yo ga'i rigs 'dzin nil sdug pos brgyan pa'i sangs rgyas gnyis pa zhes bya'o/ ma
ha yo ga'i rigs 'dzin nil rdo rje 'chang gi
sangs rgyas zhes bya ste de yang bzhi ste/lha'i rigs 'dzin/ sman kyi rigs 'dzin
dang! rnam par smyin pa'i rigs 'dzin dang! phyag rgya chen po'i
rigs 'dzin pa'o/ de la lha'i rigs 'dzin ni lha las grub pa'o/ sman ni ra sa ya na la
stsogs pa las grub pa'o/ rnam par smin pa'i
rigs 'dzin nil mkhas pa'i rab las grub pa ill dpe za myed pa'o/ phyag rgya chen
po'i rigs 'dzin nil mngon bar shes pa lnga dang ldan ba ste/
rdo rje rdzu 'phrul dang lnga'o/
[3r] sangs rgyas gnyis pa zhes bya ba nil ril 'rna bar 'di ltar chud pa la bya'o/
gnyis la khyad par myed de mnyam/ grub pa'i sgo so so
bas/ gnyis zhes bya'o/ a nu yo ga dang a ti yo ga la ni rigs 'dzin yang myed do/
rigs gang du yang myi 'dzin tel rigs 'dzin ces
gang la bsgrub pa'o/ don tan gi phyogs de la chud par bya/ sa la rigs 'dzin ces
bya' 0/ rigs 'dzin kyi sa yang/ bcu drug du zad de/
de las bzla ces nil tshad ma gsum las kyang 'gal/ lung dang rgyud las kyang myi
'phyung ngo/ de yang byang cub sems dpa'i sa bcu dang/
sangs rgyas kyi sa drug te bcu drug go/ sa bcu drug 'di yang tha rni dad! gcig du
yang so so'i man ngag bshad pa las myi rung ngo/
sa ma 'dres pa dang/ yongs su rdzogs pa'o/ spyi'i sa dang/ rdo rje sems pa dang
spyi lung bdun ka'i 'dir bshad do/ de la ma 'dres pa nil
don so sor nges pa'o/ yongs su rdzogs pa nil des khyab pa 'am bsdus pa la bya'o/
[3v] 'gu hya ti la ka dnya na rni tra dang brang ti a tsa rya dzA ya go tsas bsgyur/
sa rba 'bu ta sa ma yo ga 'bum ste las/
a tsa rya nya na ga rba dang bye ma la rni tras bsgyur cing bstrags so/
IOL Tm J 656: Translation
The topics of the teaching on the views and practices specific to each of the
seven great general scriptural systems are grouped into a total of four pith
instructions. What are these? They are (i) view, (ii) meditation, (iii) practices,
(iv) vows.
Regarding the view of the Sravakas: They regard the external body as made of
subtle particles without parts. The internal mind consists of six collections of
consciousnesses, which they see as ultimately existing permanent continuums
of shapeless substantial entities. Their meditation is on the four truths. Their
practices, are the twelve purificatory virtues. [5] Their discipline (vinaya) consists
of 250 rules.
The view of the Siitra Adherents is the two truths. Their practices are the ten
perfections. Their rules are the twenty oaths. Their meditation is to meditate
without fixation. The result is a gradual progression through the ten bodhisattva
The view of Kriya mantra is unification with internal purity by depending on
external purity. "Reliance on external purity" refers to their gathering the neces-
sities, the causes and conditions, and to their maintai.Iring cleanliness. "Unifica-
tion with internal purity" refers to their realization of the essential mind of enlight-
enment (bodhicitta). Therefore theirs is a view of a pure reality. Their meditation
is to meditate while being endowed with the three great reflections. Regarding
those, the reflection of body [10] means that one meditates [gradually,] like piling
up bricks. The reflection of speech means that from between the tongue, the teeth,
and the roof of the mouth, they recite accurately and quietly, so that only one-
self can hear. The image of the mind means that having performed the blessings
by means of the five recitations of mantra and mudra, one meditates on reality.
Their practices are threefold: not transgressing their vows, not transgressing
the ritual forms, and accurate recitations. Their vows are five: (i) not to abandon
the three jewels, (ii) to regard and respect the vajracarya as if he were a buddha,
(iii) not to generate negative thoughts or criticisms about, and to remain harmonious
with, one's vajra brothers and sisters, [15] (iv) to perform the cleansings three
times a day, (v) not to eat or drink meat, garlic, onions, and alcohol.
In the view of Yoga, it is said that one should see everything as arising from the
blessings, and regard whatever arises from the blessings as the deity. Their medi-
tation is to cultivate [themselves] as endowed with the four mahiimudras. To
meditate on oneself as the yi dam deity is called the "deity mudra." To unite the
wisdom being with oneself is called the "samayamudra." While practicing in that
way, to realize without wavering from that which is without birth or cessation is
called the "dharmamudra." Their practice is to take the mind as primary, and trans-
form it by means of a concentration (samadhi) that is not dependent on causes
and conditions; this is called the" karmamudra." [20] Their vows are seven:
(i) to regard and respect the vajracarya as if he were a buddha, (ii) not to gen-
erate negative thoughts or criticisms about, and to remain harmonious with, one's
vajra brothers and sisters, (iii) not to forsake the mind of enlightenment, (iv) until
they have been granted initiation as a mantra-holder vajracarya, to maintain a
vajra tongue, not saying so much as a word, (v) not to drink water from the same
valley as beings of lower vehicles, and (vi) to always hold their bell and vajra.
In the view of the Mahayoga secret mantra, the five familiesare seen as a single
method. Moreover, the five great elements are the mother, and the forms that come
from those [elements] are the father; [25] both abide pervasively in everything.
Thus everything is seen to be without self and other, as nirviif.Ul. Their medita-
tion involves meditating on the gradual development [of the visualization] using
the three kinds [of samiidhis] 118. Their practice is accomplishment, that is, union
and liberation. Their vows number twenty-eight grouped into three kinds: the
vows of view, the vows of practice, and the vows of accomplishment. .
The view of Anuyoga is the view of the father and the mother, together with the
son. Moreover, all the phenomena of existence are the means, the father. Unborn
thusness is the wisdom (prajftii). The nirvtil}a of them both is the bodhicitta, the
son. Their meditation is to cultivate the generation [of the maI).Qala] using the tech-
niques of the perfection [stage]. [30] Their practice is engagement/ehjoyment,
that is, union and liberation. Their vows are four, the four vows of reality, com-
passion, equality, and union with the sense-objects, also called "the general vows
of those who understand the secret mantra, the vows which are difficult to trans-
gress." These are the four vows of reality, compassion, equality (mnyam ba),
and the union of the sense faculties with their objects.
The view of the Atiyoga secret mantra sees the body, speech, and mind as the
inexhaustible adornment wheels. Furthennore, regarding [on the one hand] the
minds of the Bhagavan Buddha endowed with omniscient wisdom and of sen-
tient beings from the hells on up, [35] and [on the other hand] all that appears as
the physical matter of the external world which is an insentient container: regard-
ing these, their appearance as colors and shapes is the body, their unborn thusness
is the speech, and the nirviilJa of them both is the mind. That they are called "inex-
haustible" means they are unchanging. "Adornments" refers to their being pervaded
by that same body, speech, and mind. "Wheels" means they are non-fixating yet
apprehending; "non-fixating" meaning nirviilJa and "apprehending" meaning that
nirviil}a does not liberate them into some other place. The meditation is not to
waver from that view. The practice is the great enjoyment, that is, union and libe-
ration. The vow is singular: it is whatever. Because nothing exists beyond that, it
is called "non-existent." [40] Because there is only one vow in Atiyoga, it is called
"singular." Why "whatever?" Since [the vow] is maintained for the purpose of
an.8,ccomplishment that is spontaneous, one who is not maintaining that accom-
pliShment cannot be said to be not accomplished. Thus we say, "whatever,"
because the infinite supreme blisses of Ihu bu are automatically not transgressed.
"Whatever" means that the question of whether one is maintaining [the vow] or
not is immeasurable and boundless. "Vow" means something not to transgress.
The teachings on the practices of the seven general scriptural systems have been
briefly referred to above, but the practices of union and liberation [may require
118 The three samiidhis are conunonly used in early Mahayoga treatises to describe the
development stage. They are the thusness samiidhi (de bzhin nyid kyi ting nge 'dzin), the
all-illuminating samiidhi (kun tu snang ba'i ting nge 'dzin), and the causal samiidhi (rgyu'i
ting nge 'dzin).
some further explanation]. These practices do not exist in Yoga and Kriya, nor
among the Sutra and the Sravakas.
[45] Union in the Mahayoga secret mantra is the union of the vajra and the lotus,
which can be further divided into three: (i) the union with the sole ornament [i.e.
the central deity], (ii) the union of the five families in the single method, (iii) the
union with whatever. Regarding those, the sole ornament is avowed to remain as
two [i.e. buddha and consort]. The five families in a single method are practiced
as the principle deity, the four female deities, and a single location. Whatever
[arises] is the supreme path of the three realms. If one is performing union in
accordance with the manuals with ill the women one can find, it is said that one
should [maintain] a vajra-speech and not talk about it. Liberation [in Mahayoga]
involves the liberation of oneself and the liberation of others. Of those, there are
two ways to evaluate the liberation of oneself: [50] by one's nearness to the deity
and by the lama of the lineage. Nearness to the deity is verified when the prac-
titioner floats four fingers above the ground. The lama of the lineage will simply
know. [Regarding the liberation of others,] having gathered a large number of
beings who are experts as [verified] in that way, ten fields [i.e. suitable subjects
for liberation] may be subdued. Here their ties to a self-conceptualizing mental
continuum is [the target of] the liberation of others.
Union in Anuyoga mantra is the union of the senses and their objects. The objects
are the mother and the senses the father. Liberation [in Anuyoga] is liberation by
means of the four mahiimudrii.
Union in Atiyoga mantra is the union of space and wisdom. One's own conscious-
ness is the space present in all the objects distinguished by that wisdom. [55]
Liberation [in Atiyoga] is liberation by means of the great equality. Conven-
tionally speaking, this is the equality of the god and the goddess. Ultimately
speaking, this is the equality without birth or cessation. The causes are equal in
the five major elements. The results are equal in body, speech, and mind. Thus there
cannot be even a term for "sentient beings." ill actuality, practicing is alright and
not practicing is alright.
IOL TIB J 656: Transcription
spyi'i lung chen po bdun so so'i Ita ba dang spyod pa bstan pa gsungs pa'i don
man ngag ni bzhi zhig 'du zad del .
de yang gang bzhes nal Ita ba dang bsgom ba dangl spyod pa dang dam tshig gol
de la nyan thos kyi Ita ba nil phyi'i Ius rnams
ni phyogs cha myed pa'i rdul phra bar kyi rjes su Ital nang gyi sems rnam par
shes pa'i tshogs drug sol dngos po gzugs can rna yin
ba'i rgyun gyi rdag pa'i don dam par yod par lta'o/ bsgom ba ni bde ba' bzhi'o/
spyod pa ni sbyangs pa'i yon tan bcu gnyis sol
[5] 'dul ba ni khrims nyis brgya' lnga bcu'o/ mdo sde'i Ita ba ni bden ba gnyis
sol spyod pa ni pha rol 'du phyin pa bcu'o/
khrims ni sdom 'ba nyi shu pa'o/ bsgom ba ni dmyigs su myed par bsgom pa'o/
'bras bu ni byang cub sems pa'i sa bcu rims
kyis dgrod pa'o/ sngags kri ya'i Ita ba nil phyi'i dag pa la brten nas nang dag pa
la sbyor ste/ phyi dag pa la bltos shes bya ba nil
yo byad rgyu rgyu rkyen tshogs shing btsang sbras phyed pa'o/ nang dag pa la
sbyor zhes bya ba ni byang cub kyis sems ngo bOI rtogs pa'o/
de bas na dag pa'i chos nyid du Ita ba' 0/ bsgom ba nil gzugs brnyan chen po gsum
dang ldan bar sgom ba ste/ de yang sku'i
[10] gzugs brnyan nil tshogs pa brtsegs pa ltar bsgom ba'o/ gsung gyi gzugs
brnyan nil Ie dang/ so dang! drkan gsum gyi bar
nasi zur phyin par rang gyi rna ba thos pa tsam 'du zlos pa'o/ thugs gyi gzugs
brnyan nil sngags dang phyag rgya bzlas pa lngas
byin gyis brlabs nas ehos nyid du bsgom ba'o/ spyod pa ni gsum ste/ dam tshig
rna nyams pa dang/ cho ga rna nyams pa dang/
bzlas pa zur phyin pa'o/ dam tshig ni lnga ste/ dkon mchog gsum myi spang ba
dang/ rdo rje slobs pon sangs rgyas dang
'dra bar blta zhing bkur ba dang/ rdo rje spun la ngan sems dang dpyad sems myi
bskyed cing nang mthun bar bya ba dang/
[15] khrus dus gsum du bya ba dang/ sha chang dang/ sgog tsong la sug myi bza
myi mthung ba'o/ ya ga'i Ita ba nil thams cad byin kyi
brlabs las 'byung bar Ita ste/ byin kyi brIab las lhar bIta zhes bya/ bsgom ba ni
phyag rgya chen po bzhi dang ldan bar bsgom ba ste/
bdag yid dam gyi lhar bsgom bas lha'i phyag rgya zhes bya/ ye shes sems dpa'
bdag la beas pas dam tshig gyi phyag
rgya zhes bya/ de Itar mdzad pa nyid kyi dus na/ skye 'gag myed pa las rna yos
par rtogs pa ni chos kyi phyag rgya zhes
bya/ spyod pa ni sems gtsor spyod pas rgyu rkyen la rag rna las par ti nge 'dzin
gyis dbang bsgyur de mdzad pa ni las kyi
[20] phyag rgya zhes bya/ dam tshig ni bdun ste/ rdo rje slobs pon sangs rgyas
dang 'dra bar blta zhing bkur ba dang/
rdo rje spun la ngan sems dang dpyad sems myi bskyed cing nang mthun bar bya
ba dang/ yi dam kyi lha myi spyad ba dang byang cub
kyi s ~ m s rna dang! sngag 'chang pa'i slob pon du dbang rna bskur gi
bar dul rdo rje Ice y<?d kyang tshig 'ag yang myi smras
ba dang theg pa dma' ba'i skyes bo dangl lung pa gcig kyi chu yang myi btung
ba dang! rdo rje dril bu rtag du bcad pa dang bdun nol
gsang sngags rna ha yo ga Ita ba nil rigs lnga tshul gcig du ltal de yang 'byung
ba chen po lnga ni yuml de las gzug .
[25] su gyur pa ni yab/ thams cad 1a khyab par gnas past bdag dang bzhan myed
ngan las 'das par lta ba'ol bsgom ba ni
mams gsum rims kyis bskyed de bsgom ba'ol spyod pa ni bsgoms pa stel sbyor
sgroI 101 dam tshigs ni nyi shu rtsa brgyad dol
de yang lta ba'i dam tshig dang/ spyod pa'i dam tshig dang/ bsgrubs pa'i dam
tshig gsum du 'dus sol a nu yo ga'i ita ba
ni yab yum sras dang bcas par Ita ba'o/ de yang snang srid kyi chos thams cad
ni thabs ste yab/ de nyid rna skyes pa
ni shes rabl gnyis ga mye ngan las 'das pa ni byang cub kyi sems te sras sol
bsgom ba ni bskyed de rdzogs pa'i tshul
[30] du bsgom ba'o/ spyod pa ni spyad pa ste bsbyor sgroI 10/ dam tshig ni bzhi
stel chos nyid snying rje dang nyams ba dang dbang po
yulla sbyor ba'i dam tshig bzhi po dag gsang sngags shes pa mams kyi spyi'i
dam tshig stet 'da bar drka ba mams kyi
dam tshig yin zhes 'byung ba stel chos nyis dang! mying rje dang mnyam ba dang
dbang po dang yullas sbyor ba'i dam tshig
bzhi'ol gsang sngags a ti yo ga'i Ita ba nil sku gsung thugs myi zad rgyan gi 'khor
lor lta ba stel
de yang sangs rgyas bcom ldan 'das thams cad khyen pa'i ye shes can dang!
sems can dmyal ba yan cad sems yod
[35] pa dang! sems myed pa snod/ gyi 'jig rten phyi'i yul bems por snang ba thams
cadi lal kho dog dang sbyibs su snang ba nil
skul de nyid rna skyes pa ni gsung! gnyi ga mye ngan las 'das pa ni thugs sol
myi zad ces bya ba ni 'gyur ba myed la
byal rgyan ni sku gsang thugs nyid kyis khyab pa 1a byal 'khor 10 ni myi gnas
dang 'dzin pal myi gnas pa ni myi ngan
las 'das pa la byal 'dzin pa ni mye ngan las 'das pa des gang yang bkrol pa myed
pa la byal bsgom ba ni lta ba de nyid
las rna yengs pa'ol spyod pa ni spyad pa chen po ste sbyor sgroI 101 dam tshig
ni gcig ste phyal ba'ol de yi gong
[40] na dam tshig gzhan myed pas myed pa zhes bya! a ti yo ga'i dam tshig gcig
su yin basI de'i phyir gcig su zhes byal
ci'i phyir phyal zhes nat lhungis grub pa'i phyir na bsrungs pasl grub la ma
srungs pas myi grubs ces bya ba myed pasl
phyal zhes byas ba de lhu bu'i bde mchog mu 'byam pa lasl rang bzhin kyis myi
' ~ a ba'i phyir ro/ phyal ba ni bsrung ba dang
myi srungs grangs dang mtshams myed pa la bya! dam tshig ces bya ba ni de las
myi 'da ba la bya! phyi lung bdun kyi spyod pa bstan
pa gong ma'i 'dir 'dus mod kyi spyod pa sbyor sgroI 101 yo ga dang kri ya dang
mdo sde dang nyan thos la myed dol de la
[45] gsang sngags rna ha yo ga'i sbyor ba nil rdo rje dang pad mo sbyor ba ste
de yang gsum mol rgyan gcig par sbyor ba dang
rigs lnga tshul gcig par sgyor ba dang/ phyal bar sbyor ba'ol de la rgyan gcig pa
ni gnyis su dam bcas pa'ol
rigs lnga tshul gcig pa nil gtso bo gcig dang! gtso mo bzhi dang/ yul gcig sgrub
pa'ol phyal ba ni khams
gsum dag kyi lam mchogl nat bud myed ci snyed yod pa mamsl thams cad cho
ga bzhin sbyor nat rdo rje gsung kyis
myi smad do zhes 'byung ba'ol sgroI ba ni bdag bsgrol ba dang gzhan bsgrol ba'ol
de la bdag bsgrol ba tshad mams pa nyis
[50] tel lha nye ba dang rgyud bla rna' 01 de la lha nye ba ni bsgrub pa po de. sor
bzhi yan cad 'phags pa'ol rgyud bla rna ni mkha
pa' 01 de Ita bu skyes bu mkhas pa mang zhig 'dus nasI zhing bcu la bstsags te
bdag du/ rtog pa'i rgyud
sbyor ba ni gzhan bsgrol ba'ol sngags a nu yo ga'i sbyor bit nil yul dang dbang
por sbyor ba'o/ yul ni
YUlidbang po ni yab bol bsgrol ba ni phyag rgya chen po bzhis bsgrol bat sngags
a ti yo ga'i sbyor ba nil dbyings
dang ye shes su sbyor ba'ol bdag kyis mam par shes pa nil ye shes des bcad pa'i
yul thams cad na dbyings
[55] sol bsgrol ba ni mnyam ba chen pos bsgrol ba' 01 kun rdzab du lha dang! lha
mor mnyam/ don dam par skye 'gag
myed par mnyam/ rgyud 'byung ba chen po lnga 1a mnyam/ 'bras bu sku gsung
thugs la mnyam basI sems
can zhes bya ba'i mying yang myed pa la bya! drngos su na spyod kyang rung!
rna spyad kyang rung ngo/
Primary Sources
Dunhuang mss. cited: ITJ306, ITJ321, ITJ331, ITJ370, ITJ384, ITJ417/PT300,
ITJ423, ITJ436, ITJ447, ITJ448/PT270, ITJ470, ITJ484, ITJ508, ITJ526,
ITJ576, ITJ644, ITJ754, PT281, PT283, PT332, PT656, PT837, PT849.
A.mniiyamaiijari. (Srisarf/putatantrariija-tikamniiyamaiijari-niima). Ase. Abhaya-
karagupta. Q.2328.
Tantra. Rinpoehe, Samdhong and Dvivedi, Vrajvallabh, eds. 1992.
with Ratniivali Paiijika of Kumiiracandra. Rare Bud-
dhist Text Series-9. Samath: Central Institute of Higher Tibetan Studies.,
Mkhas pa Ide 'us mdzad pa'i rgya bod kyi chos 'byung rgyas pa. Mkhas pa lde'u.
Lhasa: Bod ljong mi dmangs dpa skrun khang, 1987.
Guhyagarbha Tantra. The mTshams-Brag Manuscript of the rNying-ma rgyud
'bum. Thimphu, Bhutan: National Library: vol. 20,jJ. 152.6-218.7.
Guhyatattvaprakasa. Ase. KaI}ha. Q.2167.
Guhyasamiija Tantra. Q.81
Dgongs 'dus 'grel pa. Mkhan-po Nus-ldan Rdo-rje. In Bdud 'joms 'jigs bral ye shes
rdo rje, ed. Rnying rna bka' rna rgyas pa. 56 vols. Kalimpong, W.B.: Dub-
jang Lama, 1982: vols. 53-56.
Dgongs pa 'dus pa'i mdo. The mTshams-Brag Manuscript of the rNying-ma rgyud
'bum. 46 vols. Thimphu, Bhutan: National Library, 1982: vol. 16, ff. 1.1-617.7.
Rgyud sde spyi'i rnam par gzhag pa. Bsod nam rtse mo. In Bsod nams rgya mtsho,
ed., Sa skya bka' 'bum. 15 vols. Tokyo: Toyo Bunko, 1968: vol. 2, ff. lal-74a.6.
Rgyud sde spyi'i mam gzhag chung ngu. Kun dga' snying po. In Bsod nams rgya
mtsho, ed., Sa skya bka' 'bum. 15 vols. Tokyo: Toyo Bunko, 1968: vol. 1,
Caryiimeliiyanapradipa-niima-pkii. Ase. Sakyamitra. Q.2703.
Jiiiinavajrasamuccaya-niima-tantra. Q.84.
Triyiinavyavasthiina. Ratuakarasanti. Q.4535.
Lta ba'i rim pa bshad pa. Ase. Dpal brtsegs. Q.4728.
Theg pa spyi beings rtsa 'gre/. Dam pa bde gshegs. With commentary by Ye shes
rgyal mtshan. Chengdu: Si khron mi rigs dpe skrun khang, 1997.
Thugs kyi sgron rna. Ase. Dpal dbyangs. Q.5918. .
I;Jiikinivajrapaiijara Tantra. Q.l1.
I;Jiikinisarvaeittiidvayiicintyajiiiinavajraviiriihi-adhibhiiva-tantrariija-niima. Q.60.
Mdo dbang gi spyi don. Loehen Dharmasn. In Lo chen gsung 'bum. 18 volumes.
Dehra Dun: D. G. Khoeehen Trulku, 1975: vol. 12, ff. 1-260.
'Dus pa mdo dbang gi bla ma brgyud pa'i mam thar ngo mtshar dad pa'i phreng
ba. By Padma 'Phrin-Ias, Rdo-rje Brag Rigs-'dzin II. In Bka , ma mdo dbang
gi bla rna brgyud pa'i mam thar and Rig 'dzin ngag gi dbang po'i mam thar.
Leh: S. W. Tashigangpa, 1972: 1-425.
Rdo rje sems dpa'i zhus Ian. Asc. Dpal dbyangs. Q.5082.
Dhyiinottarapatala. Asc. Buddhaguhya. Q.3495.
Niimasmpgftitfkii-niimamantriirthiivalokinf-niima. ViHisavajra. Q.3356.
Pi1Jejiirtha. (Vairocaniibhisambodhitantrapi1Jejiirtha). Buddhaguhya. Q.3486.
Spar khab. (Srfguhyagarbhamahiitantrariija!fkii-niima). Asc. Vilasavajra. Q.4718.
Bodhimiirgadfpapaftjikii-niima. DiparpkarasIijiUina (Atisa). Q.5344.
Mahiivairocana-tantra-vrtti. Buddhaguhya. Q.3490.
Muktiivalf. Ratniilcarasanti. Q.2319.
Mun pa'i go chao Gnubs chen sangs rgyas ye shes. In Bdud 'joms 'jigs bral ye shes
rdo rje, ed. Rnying rna bka' ma rgyas pa. 56 vo1s. Kalimpong, W.B.: Dub-
jang Lama, 1982: vols. 50-5l.
Yogiinuttaratantriirthiivatiirasmpgraha. Sraddhiilcaravarma. Q.4536.
Yogaratnamiilii. Kru;tha. Q.2313. For the Sansmt, see Snellgrove (1959), vol. 2.
Srfvajramiiliimahiiyogatantratfkii-gmpbhiriirthadfpakii-nama. Alarpkakalasa.
Srfsarvabuddhasamayogaejiikinfjiilasaf!lbaratantriirthatfkii. Asc. *Indranala.
Sarvabuddhasamayoga-ejiikinfjiilasaf!lvara. Q.8.
Sarvabuddhasamayogapaftjikii. Asc. Santimitra. Q.2535.
Bsam gtan mig sgron. Gnubs chen sangs rgyas ye shes. Leh, Ladakh: S.W. Tashi-
gangpa, 1974.
Secondary Sources
Bu ston. (1931-32). History of Buddhism. 2 vols. Translated by Dr. E. Obermil-
ler. Heidelberg: O. Harrassowitz.
Chappell, D.W. (1983). T'ien-t'ai Buddhism: An Outline of the Fourfold Teach-
ings. Tokyo: Daiichi shobO.
Cozort, Daniel. (1986). Highest Yoga Tantra: An Introduction to the Esoteric Bud-
dhism of Tibet. Ithaca, New York: Snow Lion Publications.
Dalton, Jacob P. (2002). The Uses of the Dgongs pa 'dus pa'i mdo in the Develop-
ment of the Rnying ma School of Tibetan Buddhism. Unpublished disserta-
. tion, University of Michigan.
Dalton, Jacob P. (2004a). "The Development of Perfection: The Interiorization
of Buddhist Ritual in the Eighth and Ninth Centuries," in Journal of Indian
Philosophy 32.1: 1-30.
Dalton, Jacob P. (2004b). "The Early Development of the Padmasambhava Legend
in Tibet: A Study of IOL Tib J 644 and Pelliot tibetain 307," in Journal of
the American Oriental Society 124.3.
Dalton, Jacob and van Schaik, Sam. (2003) "Lighting the Lamp: An Examination
of the Structure of the Bsam gtan mig sgron." In Acta Orientalia 64: 153-
Davidson, Ronald M. (1981). "The Litany of Names of Mafijusn," in Melanges
chinois et bouddhiques, ed. Michael Strickmann, vol. XX: 1-69.
Davidson, Ronald M. (2002). Indian Esoteric Buddhism: A Social History. New
York: Columbia University Press.
De la Vallee Poussm, Louis. (1988-1990). English
translation by Leo M. Pruden. Berkeley: Asian Humanities Press.
Dudjom Rinpoche. (1991). The Nyingma School of Tibetan Buddhism. Boston:
Wisdom Publications.
Eastman, Kenneth. (1981). "The Eighteen Tantras of the Vajrasekhara/Maya-
jiila." Unpublished paper presented to the 26
illtemational Conference of
Orientalists in Tokyo, Japan on May 8, 1981. A summary can be found in
Transactions of the International Conference of Orientalists in Japan XXVI
(1981), 95-96.
Eimer, Helmut. (1993). "The Classification of the Buddhist Tantras," in Wiener.
Zeitschriftfor die Kunde Sudasiens Supplementband 36,221-228.
Farrow, G.W. and Menon, I. (1992) The Concealed Essence of the Hevajra
Tantra. Delhi: Motilal Banarsidass.
Germano, David. (2002). "The Seven Descents and the Early History of Rnying
rna Transmissions," in The Many Canons of Tibetan Buddhism, ed. Helmut
Eimer and David Germano. Leiden: Brill, 225-263.
Giebel, Rolf W. (1995). "The Chin-kang-ting ching yu-ch 'ieh shih-pa-hui chih-
kuei: An Annotated Translation." ill J ourna! of N aritasan Institute for Bud-
dhist Studies 18, 107-201.
Gregory, Peter. (1983). "The Teaching of Men and Gods: The Doctrinal and
Social Basis of Lay Buddhist Practice in the Hua-yen Tradition," in Studies in
Ch'an and Hua'yen, eds. Robert M. Gimello and Peter N. Gregory. Hono-
lulu: University of Hawaii Press.
Gregory, Peter. (1991). Tsung-mi and the Sinification of Buddhism. Princeton:
Princeton University Press.
Hackin, Joseph. (1924). Formulaire Sanskrit-Tibetain duXe siecle. Paris: Librairie
Orientaliste Paul Geuthener.
Halbfass, Wilhelm. (1988). India and Europe. Albany: State University of New
York Press.
Hazra, Rajendra Chandra. (1958) Studies in the Upapuranas. Calcutta: Sanskrit
Hodge, Stephen. (2003). The Mahii-vairocana-abhisal'[lbodhi Tantra. London:
Routledge Curzon.
Huntington, C.W. (2002) "Was CandrakIrti a Prasangika?" ill The Svatantrika-
Prasmigika Distinction: What Difference Does a Difference Make?, eds.
Sara McClintock and Georges Dreyfus. Boston: Wisdom Publications, 2002.
Ju Mun, Chan. (2002). The History of Doctrinal Classification in Chinese Bud-
dhism: A Study of the Panjiao Systems. Unpublished Ph.D. dissertation from
University of Wisconsin.
Kapstein, Matthew. (2000). The Tibetan Assimilation of Buddhism. Oxford: Oxford
University Press.
Kaptstein, Matthew. (2003). "The Doctrine of Eternal Heaven:.A Recently Dis-
covered Treatise by Karma-pa II Karma Pakshi (1206-1283)." Unpublished
paper delivered at the Circle of Tibetan and Himalayan Studies, School of
Oriental and African Studies, April 2003.
Karmay, Samten. (1988). The Great Perfection. Leiden: E.J. Brill.
Karmay, Samten. (1988). The Arrow and the Spindle. Kathmandu: Mandala Book
Lahiri, Latika. (1986). Chinese Monks in India. Delhi: Motilal Banarsidass,
Lalou, Marcelle. (1955). "A la Recherche du Vidyadharapitaka: Ie cycle du Suba-
hupariprcchii-tantra. '! In Studies in Indology and Buddhology: Presented in
Honour of Professor Susumu Yamaguchi on the Occasion of His Sixtieth Birth-
day, eds. Gadjin M. Nagao and Josho Nozawa. Kyogo: Hozokan, 68-72.
Meinert, Carmen. (2004.) "Structural Analysis of the bSam gtan mig sgron. A Com-
parison of the Fourfold Correct Practice in the Arydvikalpapravesandma-
dMrant and the Contents of the Four Main Chapters of the bSam gtan mig
sgron," in the Journal of the International Association of Buddhist Studies
26.1: 175-195. (
Mimaki, Katsumi. (1994). "Doxographie tiMtaine et classifications indiennes,"
in Bouddhisme et cultures locales. Edite par Fumimasa Fukui et Gerard
Fussman. Paris: Ecole franc;:aise d'Extreme-Orient, 115-136.
Richardson, Hugh. (1998). High Peaks, Pure Earth. London: Serindia Publications.
Skilling, Peter. (1992) "The R a k ~ i i Literature of the Sriivakayiina, " Journal of the
Pali Text Society 16, 109-182.
Skorupski, Tadeusz. (1983). The Sarvadurgatiparisodhana Tantra. Delhi: Motilal
Snellgrove, David L. (1959). The Hevajra Tantra. 2 volumes. London: Oxford Uni-
versity Press.
Snellgrove, DavidL. (1987). Indo-Tibetan Buddhism. 2 volumes. Boston: Sham-
bala Publications.
Snellgrove, David L. (1988). "Categories of Buddhist Tantras," in Orientalia Iosephi
Tucci Memoriae Dicata, eds. Gnoli, G. and Lanciotti, L. Rome: Istituto Ital-
Y" iano per il Medio ed Estremo Oriente.
Tiiraniitha, Jonangpa. (1990) Taranatha's History of Buddhism in India, (Lama
Cbimpa and Alaka Chattopadhyaya, transl.) Delhi: Motilal Banarsidass.
Takeuchi, Tsuguhito. (2003). "Sociolinguistic Implications of the use of Tibetan
in East Turkestan from the end of Tibetan Domination through the Tangut
Period (9
_12th c.)" Unpublished paper delivered at the Turfan Revisited
conference, Berlin: Museum flir Indische Kunst, September 8
, 2003.
Thomas, F.W. (1927). "A Chinese Buddhist Pilgrim's Letters of Introduction," in
Journal of the Royal Asiatic Society, 546-558.
Tripathi, Ram Shankar and Negi, Thakur Sain, eds. (2001). H evajratantram with Muk-
tCivalfpafljikii of Mahiipm:uj.itiiciirya Ratniikarasiinti. Bibliotheca Indo-Tibetica
Series, XL VIll. Samath, Varanasi: Central Institute of Higher Tibetan Studies.
Tsogyal, Yeshe. (19.93). The Lotus Born: The Life Story of Padmasambhava.
Translated by Erik Pema Kunsang. Boston: ShambalaPublications.
Tsong-ka-pa and Hopkins, Jeffrey. (1977) Tantra in Tibet. Ithaca: Snow Lion Pub-
Tsuda, Shinichi. (1965). "Classification of Tantras in dPal brtsegs's ITa bait rim
pa Mad pa and Its Problems," in Journal of Indian and Buddhist Studies
13.1: 397-402.
Van den Broeck, Jose. (1977). Le saveur de l'immortel (A-p'i-t'an Kan Lu Wei
Lun). Louvain-Ia-Neuve: Universite Catholique de Louvain, Institut Orien-
Vitali, Roberto. (1990). Early Temples of Central Tibet. London: Serindia Pub-
Vitali, Roberto. (1996). The Kingdoms of Pu.hrang. Dharamsala, India:
Tho.linggtsug.lag.khang lo.gcig.stong ''i rjes.dran.mdzad sgo'i
go.sgrig tshogs.chung.
Wedemeyer, Christian. (forthcoming). Aryadeva's "Lamp that Integrates the Prac-
tices" (Carya-melapaka-prad'ipa). New York: American Institute of Buddhist
Studies at Columbia University.
Weinberger, Stephen. (2003). The Significance of Yoga Tantra and the Com-
pendium of Principles (Tattvasatpgraha Tantra) within Tantric Buddhism in
India and Tibet. Unpublished Ph.D. dissertation, University of Virginia.
White, David Gordon. (2000). Tantra in Practice. Princeton: Princeton University
Press .
. Willeman, Charles and Dessein, Bart and Cox, Collett. (1998). Sarvastivada Bud-
dhist Scholasticism. Leiden: Brill.
Yamasaki, Taiko. (1988). Shingon: Japanese Esoteric Buddhism. Boston: Sham-
bala Publications.
GDD Dgongs pa 'dus pa'i mdo.
D. Derge edition of the Tibetan canon.
ITJ IOL Tib J: British Library shelf mark for Tibetan manuscripts recov-
ered by Sir Aurel Stein from Dunhuang.
MTP Man ngag Ita ba'i phreng ba.
MVT Mahavairocana-abhisa7[!bodhi Tantra.
Q. Peking edition of the Tibetan canon.
PT Pelliot tlbetain: Biblioteque Nationale shelf mark for Tibetan manu-
scripts recovered by Paul Pelliot from Dunhuang.
STTS Sarvatathiigata-tattvasa7[!graha Tantra.
T. TaishO shinshU daizokyo number in the Chinese canon.
(0) Introduction
In the present series this is the second article which is devoted to the
description of a single treatise within the genre of the commentator's
manual. The preceding title in the series, SIBH 4, surveyed the contents
of the Vyiikhyiiyukti by the Indian scholar Vasubandhu (circa fourth/fifth
century). The present article will focus on a work closely related to Vyii-
khyiiyukti which was written by a Tibetan scholar who can be considered
as one of the earliest exponents of Tibetan scholasticism, and which dates
from the early thirteenth century.
(1) The Mkhas-pa-(mams-) 'jug-pa'i-sgo by Sa-skya Pru;tpta Kun-dga'-
The Mkhas-pa-(rnams-)'jug-pa'i-sgo, lit. the 'Introduction for Schol-
ars'2 (henceforth MJ) is a manual on scholastics by Sa-skya PaJ}.q.ita
Kun-dga'-rgyal-mtshan (1182-1251)3. The author of this treatise was the
famous hierarch of the Sa-skya-pa school of Tibetan Buddhism who is
generally known as Sa-skya P3.I).q.ita, 'the scholar from Sa-skya [monastery]'
1 This research was made possible by a subsidy of the Netherlands Organisation
for Scientific Research (Nederlandse Organisatie voor Wetenschappelijk Onderzoek,
2 Jackson (1987) renders the title 'Entrance Gate for the Wise', Van der Kuijp (1996: 395)
paraphrases it as 'An Introduction to Scholarship'.
3 All references for MJ in this article are to the version of this text in the Sde-dge xylo-
graphic edition of the collected works of Sa-p3I.J. contained in the Sa-skya-pa'i-bka'-'bum
volume tha (10), ff. 163r1-224r6, available in the facsimile reprint Bsod-nams-rgya-mtsho
(ed.) (1968.5: 81-111).
Journal of the International Association of Buddhist Studies
Volume 28 Number 1 2005
(henceforth Sa-paJ.l)4. This nomer clearly signalizes his preeminence not
only as a religious expert but also as a scholastic specialist; Sa-paJ.l is in
. fact considered as one of the founding masters of the scholastic traditions
in the classical Tibetan Buddhist cultures.
This important work is known to western academia primarily through
the groundbreaking study by Prof. Jackson (Hamburg) consisting of an
edition and annotated translation of the third chapter, with an elaborate
introduction (1987)6. Jackson decided on a date of composition for the text
of circa 1220-1230
. Recently Kapstein has argued that Sa-pruj., in his
MJ, has formulated a scholarly ideal that he has based directly on the
classical Indian notions of scholastical excellence, of palJ4itya

This type of text, the Mkhas-'jug, the introduction to scholastics, is
- perhaps somewhat unexpectedly - quite rare in Tibetan literature.
The only other work of this type which has gained some popularity was
that by 'Jam-mgon 'Ju Mi-pham-rgya-mtsho (1846-1912) entitled Mkhas-
pa 'i-tshul-Ia- 'jug-pa 'i-sgo-zhes-bya-ba'i-bstan-bcos
Comparable in some
respects is the genre of bshad-mdzod, lit. 'treasure of explanation', a kind
of compendium of the central Buddhist concepts and doctrines which was
aimed primarily at a lay readership (whereas the mkhas- 'jug type was
written for monastic students)lO, examples of which are the Shes-bya-rab-
gsa/by 'Phags-pa Blo-gros-rgyal-mtshan (1235-1280)11 and the Shes-bya-
kun-khyab by Kong-sprul Blo-gros-mtha' -yas (1813-1899)12.
A quite detailed survey of the contents of MJ can be found in Jackson
(1987: 191-206)13, so I will just give a brief general outline here and
4 A brief biographical sketch of Sa-paI). based on the major indigenous sources can be
found in Jackson (1987: 24-31), with a survey of the sources op. cit. p. 15-23; cf. also e.g.
Tucci (1949: 101-102), Bosson (1969: 2-7), Khetsun Sangpo (1973-1980.10: 137ff.).
5 Cf. e.g. Dreyfus (2003: 23, 103, 139), Kapstein (2003: 776-782).
6 Cordial thanks are due to prof. Jackson for kindly providing me with a draft version
of his annotated translation of the second chapter of MI. Its publication, although pro-
jected, has not yet taken place at the moment of writing the present study.
7 Cf. Jackson (1987: 64-66).
8 Kapstein (2003: 776-782).
9 Tachikawa (1983: no. 1579), cf. Smith (2001: 209-210).
10 Cf. Smith (2001: 209-211).
II Cf. Smith (2001: 210).
12 Cf. Smith (2001: 211,235-237,250-258), HSGLT 1: 166-178.
13 And, as mentioned supra, the publication of an annotated translation of chapter two
by Jackson is also forthcoming.
focus in more detail only on these parts of the text that are specifically
relevant to the present topic, in casu the first and especially the second
MJ constitutes an introduction to the theory and practice of the scholas-
tic enterprise, covering the three aspects of composition (Tib. rtsom-pa),
exposition ('chad-pa) and debate (rtsod-pa), which correspond to the three
chapters of the text
. These three topics constitute a generally current
triad in Tibetan scholastics, albeit not necessarily in this order
(1) Composition (rtsom-pa): 163vl-190r1
(2) Exposition ('chad-pa): 190r2-205r1
(3) Debate
(rtsod-pa): 205rl-223v4
Postscript and colophon: 223v4-224r6
(2) Chapter 1: On Composition
Preceded by a general introduction (I.1_6,18 163vl-165r6), the first
chapter, under the heading 'introduction to composition' (rtsom-pa-la-
'jug-pa, 165r5-6, 190r1), is primarily devoted to linguistical topicS
Mter a brief section on the required elements in the introductory parts of
a scholastical treatise (1.7-12, 165r6-167r6), it deals with a wide range of
topics in the fields of grammar (1.13-51, 167r6-173v2) and poetics (1.52-
end, 173v2-189v6).
Among the topics touched on in the section on grammar we find a
number of typological categorizations, the first and most important of
which is a classification of the basic units of language in a model of
three levels: 'phoneme' (yi-ge), 'word' (ming) and 'phrase' (tshig) which
Sa-pal), introduces sub 1.13-14. This schema corresponds of course to the
14 I am not including any detailed information on the third chapter of MJ for two rea-
sons: its contents are not immediately relevant for the present investigation, and it is acces-
sible through the excellent study and annotated translation of Jackson (1987).
15 For earlier general characterizations of MJ, cf. Jackson (1987: 39-42).
16 Cf. Jackson (1987: 192-193).
17 One is tempted to consider the alternative translation of the three topics as: (1) com-
position, (2) exposition and (3) opposition.
18 Following Jackson (1987: 241-242), I have included the six introductory verses and
the six concluding verses in the consecutive numbering of the fIrst and the third chapter
19 For a brief survey of the contents of the fIrst chapter, cf. Jackson (1987: 193-194).
threefold categorization into the levels or, more literally, 'collectives'
(Skt. kaya, Tib. tshogs) of language, sci!.:
(1) vyaiijana-kiiya (Tib. yz-ge'i-tshogs), 'the collective of phonemes',
(2) nama-kiiya (Tib. ming-gi-tshogs), 'the collective of words'
(3) pada-kiiya (Tib. tshig-gi-tshogs), 'the collective of phrases'
These are generally current among the ontological categories in the
various Abhidharma traditions in Buddhism, locus classicus being Abhi-
dharmakosa ll.47

Sa-pal). defines the three levels as follows:
'That which itself does not indicate a content one wishes to express, [but]
which functions as the basis of all expression[s?], is tenned 'phoneme' (yi-ge)'
[ ... ] '[i.e.] the vowels [and] the consonants' [ ... ] 'their subdivisions and
combinations will not be discussed here.' (1.13)21
'That which consists of a combination of phonemes and indicates [lit. a sin-
gularity of meaning, i.e.] one discrete meaning is a 'word' (ming)' (I.14a-c2)22
(for which he gives as examples: ka-ba 'pillar', bum-pa 'vase', 167v4)
'That which indicates a specification of that [scil. a semantic specification]
is tenned a 'phrase' (tshig)' (I.14c3-d)23 (examples: ka-ba-ring-po 'the long
pillar' or perhaps 'the pillar is long', bum-pa-bzang-po 'the excellent vase' or
'the vase is excellent', 167v5)
Higher levels are added also: firstly that of 'sentence' (ngag, I.1S), then
the levels of paragraph, chapter, etc. (U6), which are all, probably, within
the scope of the three-level model subsumed under the third level of
'phrase' (tshig). This three-level Abhidharma model oflanguage does not
correspond to the derivational model which the indigenous Sanskrit gram-
marians used and which involved the verbal roots as primary bases from
which free lexical word forms are derived on the second level, which in
20 Cf. HSGLT 2: 241-245 and Verhagen (2002: 154-155).
21 [I.l3:] dngos-su-brjod- 'dod-mi-ston-cing- / / brjod-pa-kun [?] gyi-gzhir-gyur-pa / / yi-
ge-zhes-bshad, 167r6-167v1 ( ... ) [1.13c cont.:] dbyangs-gsal-byed, 167v1 ( ... ) [I.13d:] /
de-yi-dbye-bsdu-'dir-ma-bshad, 167v3; on the splitting of the verse, cf. Jackson (1987: 241).
Sub l.13d Sa-paJ). refers to Smra-sgo etc. for a more elaborate description of phonology
22 [I.14:] / yi-ge- 'dus-pa'i-bdag-nyid-can / / don-gyi-ngo-bo-brda-sprod-pa / / ming-yin,
23 [I.14 cont.:] / de-'i-khyad-par-dag (I) / ston-pas-tshig-ces-rab-tu-bshad, 167v5.
their turn form the basis for the derivation of the third level of the bound
syntactic word forms
Sa-pal). seems to have been perfectly aware of this
discrepancy, postponing his treatment of aspects of the grammarians'
model until later in this chapter, in particular in his summary discussions
of case grammar (1.38-39), of verbal formation (1.50) and of a definition
of 'sentence' which is more in line with the grammarians' view (1.51).
Further typological classifications are introduced in the distinction
between 'arbitrary designations' (brda, identified with 'dod-rgyal-gyi-sgra)
and 'conventionally established' or 'derivative designations' (tha-snyad
= rjes-sgrub-kyi-sgra) (I.l7-I8) and the distinction between 'class-words'
(rigs-kyi-sgra), i.e. nouns in general, and 'name-words' (ming-gi-sgra),
i.e. proper nouns (1.30-33). Sa-pal). also used the former categorization
in the second chapter of MI, in the autocommentary on II.I0 (cf. infra,
sub 3.3). In his treatment of the latter categorization, Sa-pal). introduces
a quote from Dignaga which has thus far unfortunately defied identifica-
In fact, both of the latter two classifications may be associated with Dig-
naga, in particular his commentary ad Pramii1)asamuccaya 1.3d, where we
find a fivefold typology of words
1. yadrcchii-sabda, 'arbitrary words' i.e. proper nouns
niimnii 'rtha ucyate Ijittha iti)
2. jiiti-sabda, 'genus-words' i.e. nouns jiityii gaur iti)
3. gUlJa-sabda, 'quality-words' i.e. adjectives gUl}ena sukla iti)
4. kriyii-sabda, 'action-words' i.e. verbs (kriyasabdqu kriyayii piicaka iti)
5. dravya-sabda, 'substance-words' i.e. [another type? of] nouns (dravya-
dravyel}a dal}ljl
In this section Sa-pal). summarily discusses some further details on
various forms of metaphorical designations (1.23-26) and unusual types
of words, inter alia onomatopoeiae (1.27-29), which is followed by a
brief section on epistemological aspects of language (1.30-37) including
24 Cf. HSGLT 2: 240-251.
25 Sub I.30: don-brjod-pa 'i-sgra-thams-cad-rigs-kyi-sgra-dang-ming-gi-sgra-yin-n0-
zhes-phyogs-kyi-glang-pos-gsungs-so, 169v5, which amounts in fact to the paraphrase of
Sa-pru;t's verse 1.30: / don-la- 'jug-pa'i-sgra-ji-snyed / / de-kun-rigs-dang-ming-du- 'dus t.
26 Hattori (1968: 25, 83-85), Franco (1984). Jackson (1987: 194).
references to the apoha (Tib. gzhan-sel) theory of the mePIring of words
(1.35-36) and to the twofold typology of negations (Skt. paryudiisa and
Tib. m/n-dgag and med-dgag; 1.36)27. Other topics
within the field of linguistics which pass under review here are: the main
points of Sanskrit case grammar insofar as relevant for his Tibetan read-
ership involving also some comparisons between Sanskrit and Tibetan
case grammar (1.38-46)28; the role of the speaker's intention (Skt. vivabjii,
Tib. brjod- 'dod; 1.47-48, cf. also 1.26) introducing a quotation from Dhar-
makIrti's Pramiif}aviirttika
; some extremely summary statements about
the verb in Sanskrit (1.50) and, fmally, a defmition of a sentence accord-
ing to the Sanskrit grammarians (1.51)30, different from his discussion of
this subject earlier in this chapter in the context of the Abhidharma model
of language (1.15, cf. supra).
The final major section of this first chapter deals with poetics, basing
its treatment primarily on DaI).qin's Kiivyiidada (seventh-early eighth
century CE), in fact for a considerable part consisting of translations of
portions of the first two chapters of Kiivyiidarsa. This classical Sanskrit
work on the theory and practice of poetical composition, focusing in par-
ticular on a great variety of poetical figures, came to occupy a central
position as a manual for poetics in the Tibetan literary world as well.
A later translation, by Shong-ston Rdo-rje-rgyal-mtshan and
was included in the Bstan-'gyur, and throughout the history of Tibetan,
27 Referring to his own epistemological treatise, the Tshad-ma-rigs-pa'i-gter, for a more
detailed treatment of the apoha theory (17Qvl) and the concepts of negation and affrrma-
tion (170v5), and, in connection with the types of negation, referring generally to 'gram-
matical treatises' for further information (170v4). .
28 Cf. SIBH 7, paragraph 2.1,
29 brjod-par-'dod-pa'i-gzhan-dbang-phyir II
172v5, i.e. PramalJil-varttika 2.16ab: vivalqa-paratantratvan na sabdal;z santi kutra va I.
30 In addition to the definition in the verse 1.51 proper, Sa-paJ;l also quotes a definition
from Amarakosa in the auto-commentary: a-ma-ra-ko-sa-las I sup-dang-ti-nga'i-mtha'-can-
ngag I ces-bshad-pa'i-phyir-ro, 173v2. Ultimately this definition can be traced to the basic
treatise of Sanskrit indigenous grammar, Pfu;Iini's where sutra 1,4.14 introduces
and defmes the technical term pada, in this context meaning 'bound, syntactic word form',
as: sUP-tiN-antam padam, '[An element] ending in a nominal case-ending (sUP) or in a
verbal personal ending (tiN) is [technically termed] a bound, syntactic word form (pada)'.
Note that the Tibetan translation ngag for Sanskrit pada is not the standard rende-
ring, which would be tshig.
literature scholars have occupied themselves with this text
. This partial
translation by Sa-pal.J. in MJ appears to be the earliest Tibetan version of
Kiivyiidarsa, as such forming the first introduction of the Indic theories
of kiivya 'poetry' or alafllkiira-siistra, 'the science of the poetical figures',
in the Tibetan cultural sphere
. .
In the section on grammar, Sa-pal.J. deals primarily with Sanskrit gram-
, basing his treatment - as he himself states at the beginning of this
segment - on the models provided by the indigenous Indic traditions of
grammar and related sciences. In the commentary ad 1.3, he stresses that
his MJ does not involve' own fabrications or the products of mental obscu-
ration'34, but that he has based this MJ compendium on his careful 'study
and investigation'35 of the 'most famous and widespread'36 of the relevant
Indian treatises, documenting this by an impressive enumeration of sources
which he used for the composition of this work (164v2-165rl). The sources
which he lists belong to the fields of grammar, poetics, lexicography etc.,
as well as a wide variety of non-language-related technical disciplines
and, of course, the entire range of Buddhist canonical religious literature.
He lists the following titles and genres
(1) 'The grammars KaliJpa [Le. Kiitantra], Ciindra etc.'38
(2) 'The epistemological treatises (Pramii(w-) Samuccaya [by Dignaga]
and the seven treatises [by DharmakIrti] etc. '39
31 Cf e.g. Ruegg (1995: 126), Van der Kuijp (1996).
32 Cf. Jackson (1987: 194), Van der Kuijp (1996: 395).
33 Now, in afterthought, having investigated the contents of MI in more detail than I had
done at the time, it has become clear to me that it would have been proper to include this
text also in my survey of the Tibetan literature on Sanskrit grammar in HSGLT 2, in par-
ticular on account of the considerable intrinsic interest of the passages on Sanskrit grammar
in MI. I will attempt to make up for this omission in part in the present article and SmH
7, although of course the priulary focus here is not on grammar per se. Perhaps there will
be occasion in the future to document this and other addenda to HSGLT 2, a number of
which have already come to my notice.
34 rang-bzo-dang-mun-sprul, 165r1.
35 mthong-zhing-thos-nas-' dris-par-byas-pa-yin, 165r 1.
36 yongs-su-grags-pa-phal-che-ba, 165r1.
37 This passage is also translated and discussed in Kapstein (2003: 778-780).
38 sgra'i-bstan-bcos-ka-lii-pa-dang- / tsandra-pa-la-sogs-pa, 164v2; for this literature,
cf. HSGLT 1.
39 tshad-ma'i-bstan-bcos-kun-las-btus-dang- / rab-tu-byed-pa-sde-bdun-la-sogs-pa,
164v2; the seven works of Dharmaklrti, of course, beiIlg Pram{i!:zaviirttika, PramiilJavini.caya,
(3) 'The poetical works lataka(-malii?), the three major and the three
lesser [works?] etc.'40 '
(4) 'The treatises on metrics (Chando- )Ratnakara and Sdebsbyor-gyi_
tshoms etc. '41
(5) 'The treatise on the poetical figures [by] Dar;J.(;li[n] [i.e, KiivyiidarSa]
and the Sarasvatfkiinthabharana etc. '42
(6) 'The lexicographicai 'works and Visvaprakasa etc. '43
(7) 'The dramaturgical works [lit. treatises] Nagananda and *Rupamaft-
jar! etc,'44
Nyiiyabindu, Hetubindu, Vadanyiiya, Sal[lbandhaparikeii and Sal[ltiiniintarasiddhi; on Bud-
dhist epistemology and its history in Tibet, cf. recently Dreyfus (1997) and Tillemans
40 snyan-ngag-gi-bstan-bcos-skyes-pa'i-rabs-dang- / chen-po-gsum-dang- / chung-ngu-
gsum-Ia-sogs-pa, 164v2-164v3; Kapstein (2003: 780 n. 94) explicates: "As he explains
elsewhere, "three great" refers to three of the major Sanskrit poets, beginning with Bhar-
avi, while "three lesser" refers specifically to the works of Kilidasa, beginning with Kumii-
rasal[lbhava". CL Ruegg (1995: 111-112, 124-125). Or does the phrase 'the three major
and the three lesser' refer to the three major and three minor works of a specific author,
perhaps the most renowned Sanskrit poet Kilidasa, who wrote three plays (Abhijfiiina-
sakuntalii, Miilavikiignimitra and VikramorvaSf) and four poems (three being the most
famous, sciL Kumiirasal[lbhava, Raghuval[lsa and Meghadilta)?
41 sdeb-sbyor-gyi-bstan-bcos-rin-chen- 'byung-gnas-dang- / sdeb-sbyor-gyi-tshoms-Ia-
sogs-pa, 164v3. The former title is readily identifiable as the famous work on metrics by
Ratnakarasanti which was later included in Bstan-' gyur. It is unclear which text is referred
to under the latter title, lit. 'Bundle [or garland?] of metrics' or 'Chapter on metrics': perhaps
Jiianasrimitra's Vrttamiiliistuti 'Praise in the form of a garland of meters', a thirteenth-
century translation of which was included in Bstan- ' gyur as well; or perhaps the classical
Sanskrit works Chandoviciti, 'Investigation of metrics', or Chandomafijari, 'Cluster of
flowers of metrics', which do not seem to have been rendered into Tibetan? Cf. Ruegg
(1995: 127), Kapstein (2003: 780 n. 94).
42 tshig-gi-rgyan-gyi-bstan-bcos-dalJffi-dang- / dbyangs-can-gyi-mgul-rgyan-Ia-sogs-
pa, 164v3. On the former work, fIrst introduced in Tibetan scholastics by Sa-paJ.! in the
present work, cf. supra. The latter is probably the well known treatise on poetics, attrib-
uted to Bhoja(deva) (eleventh century), though a work on Sanskrit granunar with the same
title and by the same author is also known, cf. Renou (1940: 44); I am not aware of a Tibetan
translation of either work by Bhoja.
43 ming-gi-nges-brjod-a-ma-ra-ko-eii-dang- / sna-tshogs-gsal-ba-Ia-sogs-pa, 164v4;
the second lexicon mentioned here can most likely be identified as the Visvaprakiisa by
the twelfth-century author Mahesvara Kavi, cf. Vogel (1979: 329-331). A second, far less
probable identification would be the ViSvalocana lexicon by Srldharasena, cf. Vogel (1976),
(1979: 348-350), Ruegg (1995: 130), SIBH 7 ad MJ 11.23.
44 zlos-gar-gyi-bstan-bcos-glu [emend: klu] -rnams-rab-tu-dga'-ba-dang- / gzugs-kyi-
snye-ma-Ia-sogs-pa, 164v4. For Niigiinanda, 'Joy of the Serpents', which is
also contained in Bstan-'gyur, cf. Halm (1981), Halm, Steiner and (1991), Ruegg
(1995: 128). Can the latter (Gzugs-kyi-snye-ma) be identified as the well-known play
(8) 'The medical [-hrdaya] and [other] medical traditions
(9) '[Treatises] on arts and crafts, [on] the iconographical proportions,
[on] the expertise regarding earth, water, etc.'46
(10) '[Treatises] on prognostication with regard to external [elements] such
as the lunar mansions etc. and on prognostication with regard to inter-
nal [elements] such as wind etc.'47
(11) '[Texts] that have Buddhist and non-Buddhist aspects, [such as] Kala-
cakra (-tantra) and the treatise written by SrIdhara( -sena)'48
(12) 'Within Buddhism, the three Pi!akas of Sutra, Vinaya and Abhidharma,
and the four Tantra classes of Kriya, Carya, Yoga and Yoga-niruttara,
along with the commentaries and subcommentaries on these
, etc.'50
One should probably regard this statement of Sa-paI].'s sources of expert-
ise - and similar passages elsewhere in his reuvre - as reflecting an ideal
Gzugs-kyi-nyi-ma? Cf. Bacot (1957). Concerning the latter identification, Kapstein (2003:
780 n. 94) observes that its "known versions must postdate Sa-skya by several cen-
45 sman-dpyad-kyi-bstan-bcos-yan-Iag-brgyad-pa-dang- / gso-ba-rig-pa-Ia-sogs-pa,
164v4-164v5; the work mentioned is of course the famous medical handbook by Vagbha!a,
also contained in Bstan-'gyur; cf. Vogel (1965), Ruegg (1995: 110-111).
46 bzo-rig-pa-sku-gzugs-kyi-chag-tshad-dang- / sa-dang-chu-Ia-sogs-pa'i-brtag-pa,
164v5; cf. Ruegg (1995: 109-110).
. 47 phyi-rol-gyi-rgyu-skar-la-sogs-pa'i-rtsis-dang- / nang-gi-rlung-Ia-sogs-pa'i-rtsis,
164v5; a possible alternative translation would be 'Non-Buddhist [treatises] on prognos-
tication with regard to lunar mansions etc. and Buddhist [treatises] on prognostication with
regard to the [elements] wind etc.'; cf. Ruegg (1995: 108-109). Kapstein (2003: 779)
translates here: "The calculation of the constellations among external objects,
. and of the inner vital energies [viiyu], and so on, ( ... )", for continuation of his translation,
see the next note.
48 nang-pa-dang-phyi-rol-pa 'i-bye-brag-dus-kyi-' khor-lo-dang- / dpal-' dzin-gyis-byas-
pa'i-bstan-bcos, 164v5-164v6. The Kiilacakratantra is well-known for containing refer-
ences to other religions and their adherents, in particular to the Islam, cf. e.g. Newman
(1998); Stidharasena was the Jain author of the lexicon Vi.fvalocana which found its way
into the Buddhist literature as well, cf. SffiH 7 ad MJ II.23. Kapstein (2003: 779) com-
bines this and the preceding category, translating the latter part: "( ... ), including the Wheel
of Time [Kiilacakra], which is a speciality of both Buddhists and non-Buddhists, and the
treatise by Stidhara", and suggests that "[t]he work of Stidhara here mentioned is probably
the Trisatika".
49 I.e. on the three Pitakas and the four Tantra-c1asses; or, far less probably, on all the
treatises mentioned above?
50 nang-rig-pa-Ia-mdo-sde-dang- / 'dul-ba-dang- / mngon-pa'i-sde-snod-gsum-dang- /
bya-ba-dang-. / spyod-pa-dang- / rnal-'byor-dang- / mal-'byor-bla-na-med-pa'i-rgyud-sde-
bzhi-dang- / de-dag-gi-'grel-pa-dang- / 'grel-bshad-la-sogs-pa, 164v6-165rl.
of scholarship (Skt. piilJ4itya) derived from the classical lndic culture
which he set forth with great self-confidence but also with full apprecia-
tion of the demands it imposes, not, therefore, an expression of mere self-
aggrandizement; but as an ideal which he himself has emulated and which
other Tibetan scholastics should aspire for
This type of testimonium of sources at the outset of a technical trea-
tise in the Indo-Tibetan traditions is by no means unique. We have, for
instance, comparable enumerations in the introductory sections of the
Kiitantra commentary by Sa-bzang Mati Par:t-chen Blo-gros-rgyal-mtshan
(1292-1376)52 and the Ciindra commentary by Si-tu Par:t-chen Chos-kyi-
'byung-gnas (1699?-1774)53.
(3) Chapter 2: On Exposition
The second chapter of MJ deals with the principles of expounding the
doctrine, in particular in the form of explaining and commenting on doc-
trinal scripture, involving the analysis and interpretation of such scripture,
and the specific techniques required for communicating such matters to
a Tibetan audience. This chapter is structured on the five hermeneutical
categories as formulated in Vasubandhu's Vyiikhyiiyukti, which we have
seen in article (4) in this series
(1) 'Intention', 'purpose' (Skt. prayojana, Tib. dgos-pa): sub II.3, f. 191r5-
(2) 'Summarized meaning' (Skt. piJ:ujiirtha, Tib. bsdus-don): II.4-5, f. 191r6-
(3) 'Meaning of the words' (Skt. padiirtha, Tib. tshig-don): II.6-30, f. 192v2-
(4) 'Connection' (Skt. anusaf!1dhi, Tib. mtshams-sbyor): II.31-32, f. 203r3-
(5) 'Objections and rebuttals' (Skt. codya-parihiira, Tib. brgal-lan); II.33-
34, f. 203v2-204v5
51 Cf. Kapstein (2003: 777-780).
52 Cf. HSGLT 2: 93-94.
53 Cf. HSGLT 2: 172-179.
54 Scil. SIBH 4; for a brief survey of the contents of the second chapter of MI, cf.
Jackson (1987; 195-196).
The body of the chapter, i.e. the treatment of the above five categories
is preceded by brief discusssions of the required properties of the teacher
(II.2a) , the student (II.2b) and the doctrine (ll.2c)55 and the interaction
between these three (ll.3).
(3.1) Chapter 2.1: 1ntention
For the first category, 'purpose' or 'intention', Sa-paJ.l merely states that
this point is well known and does not need any further expatiation
He had indeed already spoken of this subject, albeit briefly, sub MI I.12,
on the necessity of stating the purpose at the beginning of a treatise
There he had introduced a quotation from Vasubandhu's Vyiikhytiyukti
'If he [i.e. the student] has heard the greatness [i.e. importance] of the Sutra
and its meaillng [or: ... the greatness of the meaillng of the Sutra], it gene-
rates respect in the student, so that he will study it and take it [to heart]; there-
fore the intention [of the Sutra] must be stated at the outset [by the com-
He had also listed four aspects of 'purpose': (1) the 'subject matter'
(brjod-bya), (2) the 'purpose' (dgos-pa), (3) the 'ulterior purpose' (dgos-
pa'i-dgos-pa) and (4) their 'connection' ('breZ-pa) there (f. 167r5). This
set ofterms, commonly known as dgos-'breZ (prob. '[the set of] purpose,
connection [etc.] ') usually also contains a:fifth element, viz. the 'text' (rjod-
byed), but as is the case here in MI, this is sometimes omitted, presumably
- as Broido (1983: 7) suggested - "since it is taken for granted"59.
55 On the splitting of II.2a, -b and -c, cf. Jackson (1987: 241 & note 39).
56 thog-mar-dgos-pa'i-don-dgos- 'brel-gyi-ngag-yin-la I de-thams-cad-grags-pas-re-
zhig-bzhag-go, f. 191r5-19lr6.
57 Jackson (1987: 195): "The first of these Sa-paI.1 had already touched on in his dis-
cussion of the preliminary parts of the treatise (I 12). This topic in any case seems to have
been already commonly understood by the Tibetans of his time."; Dreyfus (2003: 185):
"(1) A commentary should explain the purpose of the text, whether through an homage
or through an explicit statement of purpose at the beginning of the text.".
58 dang-po-dgos-pa-ni-rnam-bshad-rigs-pa-las I mdo-don-che-ba-nyid-thos-nas II nyan-
dang-len-pa-la-sogs-la II nyan-pa-po-de-gus-'gyur-bas II thog-mar-dgos-pa-brjod-par-bya
II zhes-g sungs-so, f. 167r3-167r4; corresponding to Vyiikhyiiyukti Peking Bstan- 'gyur f. 34r1-
34r2: mdo-don-che-ba-nyid-thos-na II nyan-pa-dang-ni-' dzin-pa-la II nyan-pa-po-ni-gus-
byed-pas II thog-mar-dgos-pa-brjod-par-bya; cf. also SIBH 4 paragraph 3.
59 On dgos- 'brei in general, cf. e.g. Broido (1983).
(3.2) Chapter 2.2: Summary.
In the brief section on the second category, the 'summary' or 'sum-
marized meaning', Sa-pal). distinguishes two types
, sub II.4: the general
summary of a text
and the summary enumerating the individual topics
dealt with in a text or in a portion of a text
Sa-pal). further elaborates
somewhat (sub II.4 and in particular II.5) on the qualities which a proper
summary should have and what defects should be avoided when com-
posing one, stressing such qualities as brevity and clarity of phrasip.g, and
comprehensiveness with regard to its subject matter.
As regards the fIrst type, in the previous chapter (sub I.11), our author
had already discussed the necessity of a summary presentation of, as he calls
it, 'the body of the treatise' (bstan-bcos-lus, 166v4) at the beginning of a
commentary, outlining the general contents of the text commented on

Here, in the second chapter, he adds that:
'When commenting on a basic text which is both difficult and extensive, at
the outset one should make a summary [stating] "This is the topic of this
basic treatise". [Such a summary statement] may actually be present in that
basic treatise. But if [such a summary statement] is necessary, yet not actu-
ally present in that [basic text], one should present [such] a statement sum-
marizing the topic in such a manner that it is brief, easy to understand and
easy to retain [in memory], basing [that summary] on other basic treatises
on scripture and reasoning. '64
As for the second type of summary, Sa-pal). states
60 Jackson (1987: 195): "(2) summaries, of which he discerned lwo main types: (a) con-
cise summaries of the general topic, and (b) more detailed topical outlines (il4-5). He explained
the desired traits and possible defects of each."; Dreyfus (2003: 185): "(2) A commen-
tary should summarize its subject, either concisely or in more detailed topical outlines."
61 ngag-don-bsdus (19lr6), spyi'i-bsdus-don (191v3).
62 gzhung-don-so-so'i-bsdus-don (191r6), gzhung-Iugs-so-so'i-bsdus-don (191v2), bye-
brag-gi-bsdus-don (191 v3).
63 bshad-sla-ba-dang-gzung-bde-zhing- / / bstan-bcos-Ia-yang-rtsod-bral-phyir / / mkhas-
pa-Ia-Ia-bstan-bcos-Ius 1/ bsdus-te-thog-mar-dgod-pa-mdzad, MJ 1.11, 166v4.
64 gzhung-dka' -zhing-rgya-che-ba-bshad-pa-na / thog-mar-gzhung-' di 'i-ngag-don-' di-
yin-zhes-bsdus-te / gzhung-de-Ia-dngos-su-yod-kyang-rung- / de-Ia-nye-bar-mkho-na-dngos-
su-med-kyang-Iung-dang-rigs-pa 'i-gzhung-gzhan-nas-blangs-te / nyung-zhing-' dus-Ia-go-
bde-zhing-gzung-sla-ba' i-ngag-don-bsdus-te-thog-mar-bshad-do, 191r6-191 v 1.
65 gzhung-gi-tho g-mtha' -ma-Ius-pa-blo-yul-du-byas-te / brjod-bya-rigs-mthun-mi -mthun-
blos-phye-nas-spyi 'i-sdom-chen-po-rnams-so-sor-bzhag / nang-gi-dbye-ba-rnams-mi-' gal-
bar-phye, f. 191v2.
'Taking into consideration the entire basic text, from the beginning to the
end, one should establish the main general sections
[in the basic text] each
separately on the basis of an analysis of the various topics discussed [in that
text] that are categorically similar or dissimilar. [Doing this] one should
parse [the text] in such a manner that the internal subdivisions are consis-
tent [with one another].'
It should be mentioned at this point that by the time of Sa-paJ.? the sum-
mary had actually even developed into a separate genre of commentary,
starting from the numerous 'Summary' (bsdus-don) type of commentaries
written by Rngog-Io-tsa-ba Blo-Idan-shes-rab (1059-1109)67.
I have looked at the possibility that the second type of sum-
mary which Sa-paJ.? discusses here can be identified as the well-known sa-
bead or 'topical outline' device, which is widely used throughout the Tibetan
commentary literature. The origin of this sa-bead form of analysis is
unknown. Thus far no unmistakable models for it have come to light in
Indian literature. It may then be a Tibetan innovation. It is however also quite
conceivable that it stems from antecedents in Chinese scholastics. In the lat-
ter scenario this would imply that, at least at his point, Sa-paJ.? is not reflect-
ing merely Indian models and ideals of scholastics, but also Chinese.
(3.3) Chapter 2.3: Meaning of Words.
The third section, on the 'meaning of words', is by far the most elabo-
rate section of the second chapter (II.6-30). Initially the author distinguishes
two aspects of the explanation of words (ngag-gi-don-bshad, 192v2)69:
66 Tentative translation for spyi'i-sdom = Skt. pilJrj.oddiina "abridged summary or state-
ment of contents", Edgerton (1953-2: 345). The usual Tibetan translation is bsdus-pa'i-
sdom; but one also fmds spyi'i-sdom, cf. Eimer (1983-1: 25).
67 Cf. e.g. Dreyfus (2003: 137).
68 SmH 7, paragraph 3.
69 Cf. Jackson (1987: 195): "when explaining (3) how to expound the sense of the words,
he likewise distinguished two methods: (a) the explaining of compound words, and (b) the
method of commenting word-by-word. The fIrst mainly applies to Sanskrit, so he did not
develop it in much detail (II 6-7)" and Dreyfus (2003: 185): "(3) It [Verhagen: a com-
mentary 1 should explain the meaning of the text by glossing each word, explaining rele-
vant grammatical notions, and providing the literary background of the discussion. It should
analyze compound words - a function far more important in the Indian tradition than the
Tibetan, as such words do not exist in the Tibetan language".
(1) Explanation regarding the 'compounding [or, more literally, joining
together] of words' (tshig-gi-sbyor-ba, 192v2)1
(2) 'Explanation of words' proper (perhaps rather 'the individual explanation
of words' (?), tshig-rnam-par-bshad-pa'i-tshul, 192v3)1!
The former is limited to an extremely terse introduction to the topic,
in fact barely more than a mere enumeration of the six types of nominal
compounds in Sanskrit (II.7; 192v3-192v5), referring the reader who
wishes to know more about the subject to 'other grammatical
by himseJf72 and to Smra-sgo-mtshon-cha by SmrtijfianakIrti, etc
The latter, far broader topic is elaborated on in the remainder of this
section (II.8-30; 192v5-203r3). First Sa-pal! addresses elementary sentence
analysis, offering a brief partial treatment of the kiirakas, the system of
syntactic-semantic relations in indigenous Sanskrit grammar
(II.8-9), this
only 'as far as required for Tibetans'75. He then applies this to three sam-
ple passages, viz. from Nagarjuna's Miila-madhyamaka-kiirikii (193r3-
193v3), from Haribhadra's Abhisamayiilarrzkiiriilokii (193v3-193v6), and
from the Vajra-vidiiralJ.a-dhiiralJ.f (193v6-194r4). Note that earlier in the
text, Sa-pal! had already touched on the topic of the kiirakas in connection
with case-grammar (sub 1.39).
In this section, sub II.9, we also find quotation from a 'grammatical trea-
tise' (sgra'i-bstan-bcos) which has thus far - tantalizingly - defied exact
identification. Speaking of 'the methods of expounding the extensive and
difficult scriptural traditions', he offers the following citation
70 Note that a possible translation for tshig-gi-sbyor-ba would be 'formation of words';
taking into consideration that Sa-paJ) only speaks of nominal compound forms here, and
that word-formation in Sanskrit in general would involve many other types of formation
as well, I have not opted for this interpretation here.
71 Cf. Jackson (1987: 195): "the method of commenting word-by-word".
72 kho-bos-byas-pa'i-sgra'i-bstan-bcos-gzhan, 192v4-192v5; Sa-paJ) gives a more elabo-
rate expose on compound formation in Sgra-Ia-'jug-pa (228vl-232v2); cf. also Yi-ge'i-
sbyor-ba 212-215 (249v3-4); on the former text, cf. HSGLT 2: 64-65, on the latter, cf.
Miller (1993: 130-153), HSGLT 2: 70.
73 smra-sgo-Ia-sogs-par-blta-bar-bya'o, 192v5; Smra-sgo deals with samiisa in verses
235-315; on Smra-sgo in general, cf. HSGLT 2: 37-53.
74 On this system, cf. e.g. HSGLT 2: 278-284.
75 'dir-bod-Ia-nye-bar-mkho-ba 'i-bshad-tshul-cung-zad-brjod-par-bya, 193r1.
76 'di-Ia-sgra'i-bstan-bcos-Ias / seng-ge 'i-Ita-stangs-kyis-khyad-par-gyi-gzhi-blang- / sbal-
pa 'i- 'phar-bas-skabs-don-so-sor-dbye / rus-sbal-gyi-' gros-kyis-tshig-don- 'jebs -par-bshad
"One should identify the specific topic with the gaze of the lion.
One should distinguish the subjects of the [various] sections with the leap
of the frog. .
One should explain the meaning of the words in an elegant manner with the
gait of the tortoise. " .
The bearing of the stanza seems to be that the commentator should
pick out the main topic of a text or passage with the far-reaching all-see-
ing gaze of the lion, surveying the entirety of the text; that he should
bring out the topics of the different segments in a text, dexterously jump-
ing from one to the next like a frog; and, finally, that he should go through
the entire text, commenting on each relevant passage or word as if with
the slow, careful and precise gait of the tortoise.
The terms 'gaze of the lion' and 'leap of the frog' may be traceable to
the technical idiom of Sanskrit indigenous grammar
, although the use
of these terms in that context does not correspond precisely to what we
find here. The 'gaze of the lion' (seng-ge'i-lta-stangs) can be likened to
the sirrzhiivalokita- or sirrzhiivalokana-nyiiya, the 'maxim of the lion's
backward glance'78 which is used, e.g. in the Kiisikiivrtti commentary on
3.3.49, to indicate the 'transportation' of a term 'into' a rule
from a later rule in a phenomenon technically called
'drawing back'79. "It is used when one casts a retrospective glance at
what [one] has left behind, while at the same time [one] is proceeding,
just as the lion, while going onward in search of prey, now and then bends
his neck backwards to see if any thing be within the reach"80, the rule
which occurs later in the text as it were 'glancing backwards' to the pre-
ceding rule to which the term in question is 'transported'. This is a very rare
procedure, being a subtype of the generally applied grammarians' device
of anuvrtti; the 'transporting' of a term or terms from a preceding rule to
a later one
/ ces- 'byung-bas-de-dag-gi-tshul-dang-bstun-te / ghzung-lugs-rgya-che-zhing-dka' -ba-
rnams-bshad-par-bya'o, 194r5-194r6.
77 In his draft translation of this chapter Jackson had already identified the two possible
antecedents in Sanskrit vyiikaralJa terminology.
78 Cf. Renou (1942-2: 339), Abhyankar (1977: 428).
79 Cf. Renou (1942-1: 46-47), Abhyankar (1977: 32, s.v. (li.
80 Vasu (1891-1: 503).
81 Cf. Renou (1942-1: 33), Abhyankar (1977: 26), HSGLT 2: 225,227,229-230.
The vyakaraIJ-a parallel to MJ's 'leap of the frog' (sbal-pa'i- 'phar-ba) is
also a subtype of anuvrtti, of far more common occurrence than the former,
termed mm:uj.ukagati, 'gait of the frog' or maIJ-qukapluti, 'leap ofthe frog'82.
This refers to the 'transportation' of a term from a rule, not to the rule(s)
immediately subsequent to it, but to a (group of) rule(s) that occurs some-
what later in the text, skipping the intermediate rules with the 'leap of a frog'.
Regrettably I have not been able to identify any such parallel for the
third metaphor in this quotation in MI, 'the gait of the tortoise' (
gyi-'gros). One might consider Sanskrit antecedents such as *kurmakranti
or *kurmagati
; I did not find these (or comparable) terms in the
vyakaraIJ-a idiom. Although obviously the procedures which Sa-p3.I). seems
to intend here (one might say, three manners of the commentator's 'look-
ing at' the basic text) are not identical to the types of 'transportation' of
terms from one rule to another which I have pointed out as possible par-
allels in vyakaraIJ-a, the similarities are too striking as to be coincidental.
Note also in this connection that Sa-p3.I). atmounces the stanza as a quota-
tion 'from a grammatical treatise' (sgra'i-bstan-bcos-Ias), so there is every
reason to assume that this terminology may in fact stem from a gramma-
tical background.
One might also recognize echos (albeit faint) of two of the character-
istic marks of a Buddha here: the eleventh of the secondary characteris-
tics (Skt. anuvyafijana), namely sifJ1ha-vikranta-gamin = seng-ge'i-stabs-
su-gshegs-pa, 'having a lion's [only Skt.: valiant] gait'84, and the thirtieth
of the primary characteristics, i.e. = zhabs-shin-
tu-gnas-pa, 'having the feet well [and equally] placed'8s, which is often
82 Cf. Renou (1942-2: 249), Abhyankar (1977: 298).
83 Cf. Mahavyutpatti ed. Sakaki (1916-1925: no. 4837): karma = rus-sbal; the other
two animals referred to here in MJ are listed nearby in Mahiivyutpatti as well, no. 4776:
si1!lha = seng-ge and no. 4854: malyjaka = sbal-ba (all three sub dud-' gro'i-skye-gnas-su-
gtogs-pa'i-ming); cf. also no. 9349: kUrmakrti-khara = rus-sbal-gyi-rgyab-'dra-ba-rtsub-pa.
84 Mahiivyutpatti, ed. Sakaki (1916-1925: no. 279). Cf. Bod-rgya-tshig-mdzod-chen-mo:
Seng-ge 'i-stabs-su-gshegs-pa 'i,dpe-byad = sang s-rgyas-kyi-dpe-byad-bzang-po-brgyad-
cu 'i-nang-gses-shig-ste / mi-zil-gyis-gnon-pa-Ia-mkhas-pa-nyid-kyis-seng-ge 'i-stabs-su-
gshegs-pa. Note also Mahiivyutpatti no. 280: naga-vikninta-gamin = glang-po-che'i-stabs-
su-gshegs-pa, no. 281: hal!lSa-vikranta-gamin = nang-pa'i-stabs-su-gshegs-pa and no. 282:
vNabha-vikranta-gamin = khyu-mchog-gi-stabs-su-gshegs-pa.
85 Mahiivyutpatti, ed. Sakaki (1916-1925: no. 265).
compared to the 'firm footing of the tortoise'86. This, however, seems far
less probable than the correlation with the above-mentioned grammatical
Sub IUD a classification into three types of words is introduced (194vl-
(1) 'Words [generally] current in the world' (,jig-rten-la-grags-pa{'i-sgra))
(2) 'Words [specifically] currrent in technical treatises' (bstan-bcos-la-
grags-pa( 'i-sgra))
(3) 'Words [specifically] current in extraordinary [forms of verbal com-
munication]' (thun-mong-ma-yin-pa-Ia-grags-pa'i-sgra)
In the auto-commentary Sa-paIf explains the three categories as follows
'The fIrst [category] are [words] that are commonly current everywhere [lit.:
in all the world / among all men], such as ka-ba 'pillar' and bum-pa 'vase',
The second [category] are [words] that are current among grammarians, such
as rnam-par-dbye-ba 'case-suffIX' (Skt. vibhakti) and byed-pa'i-tshig 'syn-
tactic-semantic relation' (Skt. karaka).
Therefore [this second category of words can] be comprehended [only?] by
The third [category] are [words] that are not current in the world or in tech-
nical treatises.
The basis for [their] occurrence as words [and] the etymologies [for this
third category of words] are difficult to expound.
86 Bod-rgya-tshig-mdzod-chen-mo: Rus-sbal-zhabs-kyi-mtshan-bzang = sangs-rgyas-
kyi-mtshan-bzang-so-gnyis-kyi-nang-gses-shig-ste / sdom-pa-yang-dag-par-blangs-pa-la-
87 dang-po-ni- 'jig-rten-thams-cad-la-thun-mong-du-grags-pa-ka-ba-dang-bum-pa-la-
sogs-pa' 0 II gnyis-pa-ni-sgra-pa-dag-la-grags-pa-rnam-par-dbye-ba-dang-byed-pa 'i-tshig-
la-sogs-pas-bsgrubs-pa-mkhas-pa-rnams-kyis-go-ba'i-brda '0/ I gsum-pa-ni- 'jig-rten-dang-
bstan-bcos-la-ma-grags-pa / sgra- 'jug-pa 'i-rgyu-mtshan-nges-pa 'i-tshig-bshad-dka' -zhing-
I 'phags-pa 'i-gang-zag-la-dgos-pa-shin-tu-che-ba I mdo-sde-dag-las-kyang-cung-zad-
bshad-mod / rgyud-sde-rnams-las-mang-du-bshad-pa I de-bzhin-gshegs-pa'i-brda-zhes-
grags-pa' 0 II 'di-rgyud-sde-bshad-pa-na-dgos-kyi-' dir-skabs-ma-yin-pas-re-zhig-bzhag-
go 1/ 'dir-skabs-su-nye-bar-dgos-pa- 'jig-rten-dang-bstan-bcos-la-grags-pa 'i-sgra /
mkhas-pa-rnams-kyis-shes-par-bya-ba'i-tshul-cung-zad-bshad-do II de-la- 'jig-rten-la-
grags-pa'i-sgra-la I 'dod-rgyal-dang-rjes- 'jug [sic; = (s)grub?] -gi-sgra-gnyis-sngar-
bshad-pa-bzhin-shes-par-bya, 194v2-194v5.
88 I take bsgrubs-pa to be an adjective with mkhas-pa.
89 Sa-paI). mentions only' grammarians' and gramniatical teclmical terms here. It seems
plausible that the teclmical terminology or jargon of other disciplines might be implied as
[These words] are ofthe utmost importance for the noble individuals [i.e.
Buddhas and Bodhisattvas]; they are used on a small scale in the SiUras, but
they occur frequently in the Tamras; they are known as the vocabulary of
the Tathiigata [i.e. Buddha].
Although it is necessary [to explain] these [words] when explaining a Tantra,
it is not appropriate [to elaborate on this topic] here, so I will leave it for
the moment.
At this point I will [only] briefly explain what is necessary in the present
context, namely how scholars should understand the words that are current
in the world and in the technical treatises.
In this connection one should understand the words current in the world in
terms of the two [types of] word, namely the arbitrary designations (' dod-rgyal-
gyi-sgra) and derivative designations (rjes-sgrub-kyi-sgra; occasionally,
probably erroneously, rjes- 'jug-gi-sgra), which I have discussed earlier
[namely sub 1.17-19].'
It is noteworthy that this passage is quite reminiscent of a passage from
a Guhyasamiija commentary, in fact a set of short notes on CandrakIrti's
Pradfpaddyatanii, by Kumara entitled Pradfpa-dfpa-[ippal)f-hrdayiidarsa90.
As for the former two categories, the opposition [aka lit. 'world' i.e.
'common usage in the world' vs. siistra lit. 'treatise' i.e. 'usage in a tech-
nical treatise' was also well-known in indigenous Sanskrit grammar from
M a h i i b h i i ~ y a onwards. In these contexts also siistra is often equated with
vyiikaral)a, the technical discipline of 'grammar'. Another frequent con-
trastation was made between laka and the 'usage in the sacred scripture'
in casu the Veda
The typological classification of 'arbitrary designations' and 'derivative
designations' can be found in several Indo-Tibetan linguistic sources, the
earliest of which was Smra-sga-mtshon-cha, the eleventh-century treatise
by Smp:ijfianakIrti
Sa-pal). used it also in his Sgra-la- 'jug_pa
which is
for the most part based on Smra-sgo, and he spoke of it earlier in the pres-
ent work, sub L17-19, inter alia involving the identifications brda = 'dod-
rgyal-gyi-sgra and tha-snyad = rjes-sgrub-kyi-sgra (sub L18, 168r4-168r5).
90 The passage is translated and studied by Broido (1988: 97).
91 Cf. Renou (1942-2: 266-267), Abhyankar (1977: 336).
92 Smra-sgo-mtshon-cha,11. 177-198, and vrtti ad idem; on these texts, cf. HSGLT 2:
93 Sa-skya-bka'-'bum, tha f. 227r2-228r3; on this text, cf. HSGLT 2: 64-65.
We find it in works by the eighteenth-century Si-tu Par:t-chen Chos-kyi-
'byung-gnas as well
The 'arbitrary desIgnations' or 'random words' Cdod-rgyal-gyi-sgra),
as I have stated earlier
; amount to terms which are not grammatically
analyzable, but which have an ultimately arbitrary form and are purely
conventionally associated with a specific meaning. The second type of the
'derivative designation' (rjes-sgrub-kyi-sgra) corresponds to these terms
which through linguistic analysis can be shown to derive from other lex-
emes or grammatical elements.
As for possible Indic antecedents for this dichotomy, the former cate-
gory of the 'arbitrary designation', might be associated with the Sanskrit
yad-rcchii-sabda also referring to an arbitrary term for which no analysis
or etymology can be provided, usually in the sense of 'proper name' in
Indic linguistics, but also in Buddhist contexts, for instance in Dignaga's
Sa-pm). then introduces three main techniques for word-interpretation
(sub IUO, f. 194v5-196v2)97:
(1) 'Straightforward word-explanation' (sgra-drang-por-bshad-pa, 194v6-
(2) 'Explanation by means of derivation [or: etymology], sgra'i-khams-so-
sor) drangs-nas-bshad-pa, 195r3-195v1)
(3) 'Explanation involving permutation' (phan-tshun-bsgyur-te-bshad-pa,
Two subtypes (195v1):
(3.1) 'permutation by means of synonyms' (rnam-grangs-bsgyur-ba,
(3.2) 'permutation of phonemes' (yi-ge-bsgyur-ba, 195v3-196v2)
Strictly speaking he associates these only with the second category of
words, the terminology current in technical treatises
. However, from his
examples it would appear that they can - at least also - be applied to
commonly current words.
94 E.g. in one of his dris-Ian collections; cf. SIBH 1: 65-67.
95 Cf. SIBH 1: 65-66.
96 Cf. SIBH 1: 65-66 n. 33, and supra in paragraph (2).
97 Cf. Jackson (1987: 195).
98 bstan-bcos-la-grags-pa'i-sgra-la / sgra-drang-por-bshad-pa / drangs-nas-bshad-pa /
phan-tshun-bsgyur-te-bshad-pa' 0, 194v5-194v6.
As one example of (1) 'straightforward explanation', he quotes a(n uniden-
tified) sutra:
'If one summarizes the entire Dharma, it is: If one is connected, one is
bound, and if one is separated, one is wholly liberated.'99
and offers the following - indeed straightforward - explanation of the
'This statement is a reference to the four [Noble] Truths, namely: If one is
connected with the cause [of suffering], one is bound by suffering,' [and] if
one is separated [from the cause of suffering] by the Path, one is wholly lib-
erated on account of the cessation [of suffering].'IOO
Sa-paIJ. offers a number of examples under the heading (2) 'explanation
by means of etymology', one of them for the Sanskrit term kiiya. Its ety-
mology is traced to a verbal root kai, for which S a - p a ~ cites the phrase
kai gai rai sabde, '[The roots] kai, gai and rai [occur] in [the meaning]
"sound".' This can be identified as a so-called dhiitupiitha-entry, i.e. an
entry from a lexicon of verbal roots which forms an integral part of the
indigenous Sanskrit grammatical systems, in this case Ciindra Dhiitupiitha
1.266 or Kiitantra Dhiitupiitha 1.256

Here also the case of the Sanskrit term arhat is briefly referred to,
implicitly distinguishing between a grammatically formally correct etymo-
logy leading to the translation 'worthy of veneration' (mchod-' os-pa) and
what has been termed a hermeneutical etymology, which is the basis for
the rendering 'he who has defeated his enemies' (dgra-bcom-pa)liJ2. This
dichotomy in the interpretation of the term is also expressed in the eighth-
century Indo-Tibetan lexicographical commentary Sgra-sbyor-bam-
po-gnyis-pa in its comments on the term arhat
The Tibetan scholas-
99 mdo-las / chos-thams-cad-bsdu-na / 'brel-na-'ching-zhing-bral-na-rnam-par-grol, 194v6.
100 zhes-gsungs-pa 'i-don-bden-pa-bzhi-ston-pa 'i-tshig-ste / kun- 'byung-gis- 'brel-na /
sdug-bsngal-gyis-'ching- / lam-gyis-bral-na / 'gog-pas-rnam-par-grol-zhes-bya-ba'i-don-to,
101 Note that S a - p ~ does not cite the Pfu.rinian Dhiitupiitha here, which reads kai gai
sabde (1.965), and which, for instance, is cited in Sgra-sbyor-bam-po-gnyis-pa in its com-
ments on the term geya, cf. HSGLT 1: 39, HSGLT 2: 410.
102 arha-ta'i-sgra / dgra-bcom-pa-dang- / mchod-par-' os-pa-gnyis-ka-Ia-bshad-du-rung,
103 Ed. Ishikawa (1990: 7-8), HSGLT 1: 21-22, SIBH 1: 69,75.
tics refer to these two types of translation as sgra-' gyur, 'translation
[according to] the word' and don-'gyur, 'translation [according to] the
meaning', respectively. Elsewhere I have proposed to interpret this typology
of translations as distinguishing 'convention-based translation' or 'sense-
based translation' , as opposed to 'intention-based translation' or 'reference-
based translation' 104,
As for (3), the technique for word-interpretation involving permu-
tation, its first subtype, 'permutation by means of synonyms', is exem-
plified inter alia by a very common glossing of Skt. gata, 'having
gone' (Tib. gshegs-pa) as 'having understood' (Tib. rtogs-pa) , here
specifically applied to the term Sugata, lit. 'he who has gone well', a
famous epithet of the Buddha
Compare, for instance, again Sgra-sbyor-
bam-po-gnyis-pa which glosses gata (in Tathagata, another epithet of
the Buddha) as 'having gone', or 'having come', or 'knowing', or 'having
said' 106.
Under the heading of the second subtype, 'permutation of phonemes',
Sa-pm;t offers the following observation
'Moreover, we find some instances where language-specialists use words in
a particular manner involving the mutual exchanging of phonemes, the sep-
aration of phonemes [from one another] and the hiding [or elision] of
phonemes, when they see a specific purpose [is served by this] such as for
instance the countering of erroneous opinions.'.
Under this subtype Sa-pm;t appears to subsume a wide range of lin-
guistic phenomena and forms of interpretative manipulation, all of which
involve some kind of changing of phonemes or syllables within the
terms at hand. As one example of such manipulation - which in this
case clearly transgresses the bounds of grammatical convention - Sa-pm;t
104 SIBH 7, paragraph 2.2.
105 su-ga-ta'i-sgra I su-ni-legs-pa'am I bde-ba'am I bzang-po'am I shin-tu-la-sogs-pa-
la- 'jug (/) I ga-ta 'i-sgra-gshegs-pa' am I rtogs-pa-la- 'jug (/) I rnam-' grel-las / rgyu-spangs-
yon-tan-gsum-bde-gshegs /1 gshegs-pa-rtogs-pa'i-don-phyir-te /1 des-ni-phyi-rol-pa-dang-
slob // mi-slob-pas-lhag-de-yi-phyir 1/ zhes-gsungs-pa-Ita-bu'o, 195v1-195v3.
106 gata-ni-gshegs-pa'am-byon-pa'am-mkhyen-pa'am-gsungs-pa-la-bya, ed. Ishikawa
(1990: 7).
107 yang-skabs-' gar-Io g-rto g-bzlo g-pa-la-so gs-pa-dgos-pa-khyad-par -can-mthong-ba' i-
tshe / sgra-pa-dag-gis-yi-ge-phan-tshun-brje-ba-dang- I yi-ge-kha-phral-ba-dang- I yi-ge-
mi-mngon-par-bya-ba 'i-sgra 'i-sbyor-ba-yod-de, 196r1-196r2.
refers to a passage in a work
by Ratnakarasanti. Here ,some form of
identification is established between the terms buddha and bhiitartha,
apparently as some form of comment aria 1 device
'According to the master Ratnakarasanti, the term bhiitartha means 'per-
fect meaning' [?] (yang-dag-pa'i-don) and he established that [term bhii-
tartha] as a word for Buddha [or: the word "Buddha"]' By positing bud
instead of bhiita, and by positing dhainstead of artha, he established [the
term bhiitartha] as [a word for] Buddha [or: the word "Buddha"]'
Another example of the 'permutation of phonemes' method, yet of a
different order entirely, remaining clearly within the boundaries of gram-
matical convention, is the reference to the phenomenon of semantical
variation in Sanskrit verbs due to the combination with various verbal
prepositions (Skt. upasarga)llo:
'Moreover, different words are formed when a single verbal root is combined
with various verbal prepositions. For instance, if the single basis mana
combined with the verbal preposition pra-, [the term] pramalJa [meaning]
'means of valid knowledge' is formed; if combined with [the verbal prepo-
sition] anu-, [the term] anumana [meaning] 'inference' is formed; if combined
with [the verbal preposition] upa-, [the term upamana meaning] 'analogy'
is formed; if combined with [the verbal preposition] abhi-, [the term abhi-
mana meaning] 'self-conceit' [is formed], etc.; one should know the appli-
cation [of such formations] in [their] context.
108 Source thus far unidentified.
109 slob-dpon-rin-chen- 'byung-gnas-zhi-bas-bhUta-artha-zhes-bya-ba-yang-dag-pa 'i-
don-yin / de-sangs-rgyas-kyi-sgrar-sgrub-pa-Ia / bhi1ta'i-gnas-su-bud-bzhag / artha'i-gnas-
su-dha-bzhag-nas-sangs-rgyas-su-sgrub-pa, 196r2-196r3.
110 yang -sgra 'i -khams-gcig -la-nyer -bs gyur-gyi-rkyen-tha-dad-dang -p hrad-na-s gra-du-
mar-' gyur-te / mii-na-zhes-bya-ba'i-khams-gcig-la / nye-bar-sgyur-ba'i-yi-ge-pra-dang-
phrad-na / pra-mii-na-tshad-mar-' gyur / a-nu-dang-phrad-na / a-nu-mii-na-rjes-dpag-tu-
'gyur / u-pa-dang-phrad-na-dpe-ru-' gyur / a-bhi-dang-phrad-na-mngon-pa'i-rgyal-la-
sogs-pa-skabs-dang-sbyar-shes-par-bya' a / / sgra'i-bstan-bcos-las / nye-bar-bsgyur-ba'i-
dbang-gis-ni / / skad-byings-don-ni-rab-'gyur-te / / gangii'i-chu-ni-mar-mod-kyi / / rgya-
mtsho'i-chu-yis-' gyur-ba-bzhin / / zhes-bshad-pa-ltar-ro, 196r6-196v2.
111 Note that when Sa-pru;t claims that a 'single basis miina' underlies the four forms
pramiiIJa up to abhimiina this is in fact incorrect, or at least an oversimplification of the
facts. The first three forms (pramiiIJa, anumiina and upamiina) are derived from the ver-
bal root mii, 'to measure', with a verbal preposition (pra-, anu- and upa- respectively) and
a primary nominal suffix -na, whereas abhimiina derives from root man, 'to think', with
the verbal preposition abhi- and a primary nominal suffix -a.
As it is stated in a [Sanskrit] grammatical treatise:
The meaning of the verbal root can be changed by the force of the verbal
[Similarly,] although the water of the Gailga [river] is sweet, [its taste] is
changed by the water of the ocean.'
The verse cited at the end of the above passage is a well-known
mnemotechnical stanza from the indigenous Sanskrit traditions of gram-
mar, frequently quoted, but of unknown ultimate origin:
upasargelJa dhiitvartho baliid anyatra niyate I
gmigiisalilamiidhuryam siigarelJa yathiimbhasii II
This didactic verse was also current in Mahayana scholastics, as shown
e.g. by its citation by CandrakIrti in the beginning of his Prasannapada,
specifically in his explanation of the element pratltya (derived from ver-
bal root i with verbal preposition prati) in the technical term pratftya-
Finally, a third example which Sa-paI.1 offers here is a set of two par-
allel verses with opposite meanings, one attributed to the evil deity Mara
which begins 'The life of men is long' and a counterpoint verse, 'The life
of men is not long', etc., which was spoken by the Buddha (195v4-196r1).
In fact, Sa-paI.1 describes the two verses as different interpretations of one
and the same verse which are apparently based on the presence of a num-
ber of covert negations, i.e. instances where the Sanskrit negative prefix
a- is indiscernible on account of a sandhi-combination with a preceding
phoneme. I have discussed this passage elsewhere, so I will not repeat the
details here!l3.
Elsewhere I have pointed out the partial similarities between the three
types of explanation introduced here and items in the third chapter
(dealing with a set of hermeneutical issues) of the early, probably ninth-
century, grammatical treatise Gnas-brgyad-chen-mo
. Moreover, com-
parable in particular to the third type of interpretation, involving the
112 Cf. Verhagen (1988 [1996]: 42 note 115, 44-45 note 129).
113 SIBH 4 paragraph 3.
114 Cf. HSGLT 2: 10-11.
manipulation of phonemes, is a typology by Padma Dkar-po, (1527-l592)115
who, when discussing the interpretation of non-literal statements (sgra-
ji-bzhin-ma-yin-pa), distinguishes three types of interpretation which involve
(1) 'adding (or combining) phonemes' (yi-ge-bsnan-pa), (2) 'di.viding
phonemes' (yi-ge-phyes-pa) and (3) 'altering [phonemes] into different
[phonemes]' (gzhan-du-bsgyur-ba)116.
Two more or less isolated verses on anaphoric reference of pronouns
(II.l2) and the recognition of the vocative case (II.13) are followed by a
section on general principles for, and various defects and problems which
can occur in the transmission of teachings from teacher to pupil, in par-
ticular, of course, with regard to the interpretation of scriptural sources
At this point, Sa-paI.J. returns to the topic of word-interpretation proper.
After a verse emphasizing the necessity of not merely comprehending
the general sense of a text, but also of taking effort to understand the
individual terms (II.22), the remainder of this elaborate section is for the
most part concerned with the various aspects of such explication which
are specifically relevant for the Tibetan scholars (II.23-30). He touches
on topics such as lexicography (II.23), defects in Tibetan translations of
Sanskrit terms (II.24-27), types of repetition (11.28-29), and standing
expressions and aphorisms (II.30)1l7.
(3.4) Chapter 2.4: Connection
ill the fourth section of this chapter, on 'connection' (II.3l-32), Sa-pal).
distinguishes two types of connection which the commentator should
bring out
115 In his Dbu-ma-gzhung-Iugs-gsum-gsal-bar-byed-pa-nges-don-grub-pa'i-shing-rta
and Brjod-byed-tshig-gi-rgyud-bshad-pa-mkhas-pa'i-kha-rgyan; cf. Broido (1984: 16,29).
116 Cf. Broido (1984: 16).
117 1 discuss some specific items from these sections in a separate article in the pres-
ent series, viz. SIBH 7.
118 I sgra-dang-don-gyi-dbye-ba-yis II mtshams-sbyar-ba-ni-rnam-pa-gnyis I, MJ 11.31,
203r3. Jackson (1987: 196): "The fourth main topic was the method of explanation by link-
ing together previous and subsequent words and topics. He explains how to do this whether
the linking topics or concepts are explicitly mentioned in the basic text or not (II 31-32).";
Dreyfus (2003: 185): "(4) A commentary should also pay attention to the connection
between words and topics as well as that between the different elements of the text.".
As for the [statement of] connection, there are two types, on account of the dis-
tinction between [connection of] words and [connection of] topics. (MI II.31)
In his commentary he identifies the two types of connection as 'the
connection of preceding and subsequent words' (sgra-snga-phyi-mtshams-
sbyar-ba, 203r4) and 'the mutual connection of topics' (don-phan-tshun-
mtshams-sbyar-ba, 203r4). Sa-pat;l's treatment of these two types is quite
Sa-pat;l explains the first type (203r5-203r6) as 'the statement: "The
meanings of the preceding and following words are connected as fol-
lows: ( ... )" when this [connection] cannot be readily elicited from the
force of what was previously stated, namely the meanings of the words
of the basic text, or, when, even if it can [be elicited from that], [the
connection] is not entirely clear, or when there is a particularly difficult
syntactical construction'119. He then quotes Dharmakirti
to the effect that
communication can take place on the basis of the explicit as well as the
The second type of connection, the statement of the connection between
topics, is again subdivided into two SUbtypes (Mf II.32)122. The first,
connection between 'brief expose and elaborate explanation'123 consists
of 'the type of statement "Mter [the author of the basic text] has thus
expounded [this] in a brief form, he
will now explain that same [topic]
elaborately"'125. Sa-pat;l terms the second subtype the statement of 'con-
nection by means of logical incompatibility and relation' 126. This amounts
119 gzhung-gi-tshig-gi-don-sngar-brjod-pa 'i-nus-pa-la-brten-nas-' dren-mi-nus-pa' am I
gal-te-nus-kyang-dngos-su-mi-gsal-ba' am I tshig-sbyar-ba 'i-tshul-dka' -ba- 'ga' -zhig-tshig-
snga-phyi'i-don- 'di-ltar- 'brel-te-zhes-brjod-pa, 203r4-203r5.
120 dper-na I de-sbyar-med-kyang-brjod- 'dod-las I I de-yi-don-ni-rtogs-par- 'gyur I I zhes-
gsungs-pa-ltar, 203r5. The quote could be PramiilJa-varttika 4.191 :vivakeato 'prayoge 'pi
tasyartho 'yarn pratfyate, or PramaIJa-vini.caya 2.10cd.
121 tshig-dngos-su-ma-sbyar-yang-' ga' -zhig-brjod- 'dod-kyi-stobs-kyis-rto gs-pa-yang-
yod I 'ga' - zhig-dngos-shugs-kyi-sgo-nas-rtogs-pa' ang-yod I 'ga' -zhig-dgos-pa 'i-sgo-nas-
rtogs-pa-yod-do, 203r5-203r6.
122 I don-ni-mdor-bstan-rgyas-bshad-dam I I 'gal-'brel-sgo-nas-mtshams-sbyar-byed I,
M] II.32, 203r6.
123 mdor-bstan-rgyas-bshad, in II.32, 203r6.
124 Or I, viz. the commentator?
125 de-ltar-mdor-bstan-nas-de-nyid-rgyas-par-' chad-ces-pa-Ita-bu, 203r6.
126 'gal- 'brel-sgo-nas-mtshams-sbyar, in II.32, 203r6.
to 'statements of the type "After [the author of the basic text] has thus
expounded the negative factor [lit. that which is to be eliminated], he will
now expound the antidote" or "After [the author of the basic text] has
expounded the cause, he will now expound the result", irrespectively
whether or not [these connections] are made explicit in the basic text'127.
The categorical distinction which Sa-par;t makes here between the
two types of connection as pertaining to 'words' (sgra) and 'meanings'
(or 'topics', 'content'; don) is of course by no means so c1ear,-cut

Obviously matters of meaning and content come into play in the first type
also, but equally obviously there is more emphasis on this aspect in the
second type.
Perhaps the two types of 'connection' can roughly be identified as
corresponding to sentence-structure, or syntactic structure, and textual
structure respectively. The former appears to be restricted primarily to
the correlations existing between terms within a sentence or at least within
smaller textual portions (paragraphs etc.), whereas the latter seems to per-
tain to the identification of the topics discussed in larger segments of a
text or even an entire text, and the interrelations that exist between them.
A question which indubitably merits further investigation, but which
can only be briefly hinted at here, is the possible correlation between the
hermeneutical methods set forth in the Mahayana literature, primarily in
Vasubandhu's Vyakhyayukti which is - of course in an adapted form-
continued here in Sa-par;t's MI, and that of the early Buddhist traditions,
specifically the Theravada
which are primarily laid down in the para-
canonical treatises Netti-ppakaral}a and Petakopadesa.
Here, in Sa-par;t's treatment of 'connection', in certain respects the
expbse on 'consecutive connection' in Netti-ppakaral}a comes to mind.
127 dper-na / de-ltar-spang-bya-bshad-nas-gnyen-po-' chad-ces-bya-ba' am /
rgyu-bshad-nas- 'bras-bu-' chad-ces-bya-ba-lta-bu-dngos-su-bkod-pa' am / gzhung-dngos-
su-ma-bkod-kyang-rung, 203vl-203v2.
128 The identification of the sgra / don (Skt. sabda / artha) dichotomy with the 'con-
vention' / 'intention' or 'sense' / 'reference' opposition of modem linguistics, as proposed
by Broido in the eighties, which I have in tum connected with the sgra-'gyur / don-'gyur
categorization in the context of Indo-Tibetan translations in another article on MJ in this
series (SIBH 7), does not appear to apply here.
129 Although some connection with the Mulasarvastivada tradition may be supposed
also, cf. Von Hiniiber (1996: 80 par. 165).
This is discussed under the heading of the 'fouifold array' (pili catubyaha),
the fourth of which is 'consecutive connection' (paJi pubbiiparasandhi).
It is itself subdivided into four types
(1) 'connection of meaning[s]' (pili atthasandhz)l31 .
(2) 'connection of word[ing], (PaIi byaiijanasandhi)132
(3) 'connection of expounding' (pili desanasandhi)133
(4) 'connection of instruction' (pili niddesasandhz)134
This fourfold typology of 'connection' . is reminiscent - up to a point -
of the four types of 'connection' which Sa-pal). discusses here: atthasandhi
corresponding to don-phan-tshun-mtshams-sbyar-ba (sub II.31), byaii-
janasandhi to (sub II.31); less evident,
but nonetheless possibly desaniisandhi corresponding to the first subtype
sub don, viz. mdor-qstan-rgyas-bshad (sub II.32) and niddesasandhi to
the second subtype 'gal-'brel-sgo-nas-mtshams-sbyar (sub II.32).
(3.5) Chapter 2.5: Objections and Rebuttals.
The fifth, final section of the second chapter of MI, still following the
basic structure provided by Vyiikhyiiyukti, is devoted to the method or
perhaps rather the aspect of the commentary consisting in discourse in the
form of a debate, or as both Vasubandhu and Sa-pal). term it, 'the objec-
tions and [their] rebuttals' (Skt. codya-parihiira, Tib. brgal-lan). It is
indeed extremely common in classical Indic commentaries, both Bud-
dhist and non-Buddhist, to find the form of a debate or discussion between
parties, usually a beginner or pupil representing the so-called
130 Hardy (ed.) (1902: 3, 38-39), (1962: xxxviii, 9, 55-64); cf. also Peta-
kopadesa, (1964: 123-124). Cf. also commentary ad Nettippakara1;la, ed. Hardy
(1902: 202): Pubbiiparanusandhi ti pubbena ca aparena ca anusandhi. Pubbiiparena san-
dhi ti pi patho. Suttassa pubbabhiigena aparabhiigaT[l saT[lsandetva kathanan ti attho.
SaT[lgitivasena va pubbiiparabhutehi suttantarehi saT[lva1;l1;liyamanassa suttassa saTflSan-
danaT[l pubbiiparanusandhi. YaT[l pubbapadena parapadassa sambandhanaT[l, ayaT[l pi
13l (1962: xxxviii, 62): "meaning-sequence".
132 (1962: xxxviii, 62): "phrasing-sequence".
133 (1962: xxxviii, 62-63): "teaching-sequence".
134 (1962: xxxviii, 63-64): "demonstration-sequence".
'preliminary position', and the teacher, or author of the com-
mentary, representing the siddhiinta, 'established conclusion'135.
This fifth section is, again, very brief and does not give any compre-
hensive treatment or general survey of the topic it addresses; perhaps
because the subject is treated at length in the next, third chapter of MI,
which is devoted to the theory and practice of debating
. It opens with
a brief introduction on the importance of basing one's arguments both on
'scripture' (Skt. agama, Tib. lung) as well as on 'logical reasoning' (Skt.
yukti, Tib. rigs-pa), regardless whether one-is facing non-Buddhist objec-
tions or critique stemming from fellow Buddhists belonging to the various
early, Mahiiyana, or Tantric traditions
For the most part the section con-
sists of a slightly more elaborate expose of the 'six alternatives' (Skt.
Tib. mtha' -drug, lit. 'six extremes' or 'six limits')138, a set of three alter-
native pairs of properties that playa role in scriptural interpretation 139. In
his commentary ad II.34 Sa-pat;llists the six as follows:
(1) dgongs-pa(-bshad-pa), '(communicated) with [particular] intention' (Mf
135 Jackson (1987: 196): "Then he treated the fifth main topic, the method of com-
menting on a thorny doctrinal question by means of objections and replies, i.e. through a
presentation that mirrors the exchange of views of participants in a discussion."; Dreyfus
(2003: 185): "(5) Finally, it [Verhagen: a commentary] should examine possible objec-
tions and articulate answers in a way that reflects the actual practice of debate.".
136 Text and annotated translation of this third chapter is, as mentioned supra, available
in Jackson (1987)_
137 Sub II.33: I brgal-lan-sgo-nas-gzhung-gi-don IIlung-dang-rigs-pas-gtan-la-dbab I,
translations for the term mtha'-drug: Thurman (1988: 137,147 n. 24): "the
six parameters", Arenes (2002B: passim): "Six extremes (ou possibilires alternatives)".
139 Sub II.34: I mtha' -drug-gang-gis-shes-pa-de II gzhung-bshad-pa-la-shin-tu-mkhas I,
203v5-203v6. In the topical outline (sa-bead) of MJ based on Olo-bo Mkhan-chen's com-
mentary reproduced in Jackson (1987) appendix I, this expose of the 'six alternatives' is
not identified as part of the section on 'objections and rebuttals', but as a second main sec-
tion of the second chapter, dealing with a second major method of scriptural interpretation ((B)
mtha'-drug-gi-bshad-pa) next to the fivefold system based on Vasubandhu's Vyakhyayukti
((A) rnam-bshad-rigs-pa-las-gsungs-pa-ltar-'ehad-thabs-lngas-bshad-pa)_ However, this
passage which/allows the treatment of the mtha' -drug in MJ seems to indicate that it does
form part of the brgal-lan section: mtha' -drug-gi-tshul-rgyas-par-gzhan-du-shes-par-bya' 0
II 'di-lta-bu'i-bshad-pa'i-tshul-legs-par-shes-na I mdo-rgyud-kyi-dgos-pa-legs-par-shes-
shing- I brgal-lan-gyis-gtan-la-'debs-pa-mkhas-par- 'gyur-ro 1/ brgal-lan-gyi-tshul-' di-
legs-par-shes-par- 'dod-na I rigs-pa'i-gter-du-blta-bar-bya' 0, 204v4-204v5.
(2) dgongs-pa-ma-yin-pa(-bshad-pa), '(cominunicated) without [particular]
intention' (Mf 204r2)140
(3) drang-ba'i-don, 'provisional meaning' (Mf 204r3)
(4) nges-pa'i-don, 'definitive meaning' (Mf 204r3-204vl)
(5) sgra-ji-bzhin-pa, 'literal [statement], (Mf 204vl-204v2)
(6) sgra-ji-bzhin-ma-yin-pa, 'non-literal [statement], (Mf 204v2-204v3)
This set of six 'alternatives' is usually found in Tantric contexts. It should
be noted that this ~ a t k o t i in Tantric hermeneutics is often incorporated into
larger complexes of exegetical categories, most notably the system of
the so-called 'seven ornaments' (Skt. saptiilaT(!kara, Tib. rgyan-bdun)141.
It would take us too far afield to go into more detail at this point.
Within the Sanskrit traditions the set of 'six alternatives' appears to be
attested only in Tantric literature
In the Tibetan context it is also pre-
dominantly represented in Tantric exegesis
3, but in Tibetan Buddhism
their application sometimes is extended to exoteric, SiUra, scripture also.
The earliest Tibetan scholastic to do so appears to have been Sa-pal)..
Here, in the second chapter of MJ as well as in the third chapter, sub
verse 23
, he applies this set to non-Tantric Buddha-Word as well
A later
example is the famous 'Brug-pa Bka'-brgyud-pa scholar Padma-dkar-po
(1527-1592), who did the same in his Dbu-ma-gzhung-lugs-gsum-gsal-bar-
byed-pa-nges-don-grub-pa 'i-shing-rta
140 Here in MJ as well as in Padma-dkar-po's Dbu-ma-gzhung-lugs-gsum-gsal-bar-byed-
pa-nges-don-grub-pa'i-shing-rta (cf. infra) there is a slight difference in tenns with Tantric
usage for the fust two: dgongs-bshad and dgongs-min in Tantric context = dgongs-pa-can
and dgons-pa-can-ma-yin-pa here, cf. Ruegg (1985: 322 n. 10).
141 Cf. e.g. Steinkellner (1978: 449-453), Broido (1983: 34-44), Arenes (2002A),
(2002B: 5-12), (2003).
142 Important Indic sources are e.g. CandrakIrti's Pradfpoddyotanii commentary on
Guhyasamiijatantra, cf. Steinkellner (1978: 450-453), and Sraddhiikaravarman's *JiiiinfIVa-
jrasamuccaya-niimatantrodbhfIVa-saptiilal[lkiiravimocana preserved in Tibetan translation
under the title Ye-shes-rdo-rje-kun-las-btus-pa'i-rgyud-las- 'byung-ba'i-rgyan-bdun-rnam-
par-dgrol-ba, cf. Arenes (2002A: 170-171, 181).
143 The set of six 'alternatives' as discussed by various Tibetan Tantric exegetes follow-
ing Pradfpoddyotanii is studied by Broido (1983B: 21-23, 33-40) and (1984: 9-21, 25-26);
for the treatment of this set in Dbal-mang Dkon-mchog-rgyal-mtshan's Rgyud-sde-bzhi'i-
don-rnam-par-bzhag-pa-sngags-pa'i-'jug-pa'i-sgo, cf. Arenes (2003: 22-25, 37, 39-42) and
in Bdud-'joms 'Jigs-'bral-ye-shes-rdo-rje' s Snga-' gyur-rnying-ma-pa 'i-rnam-gzhag-legs-
bshad-snang-ba'i-dga'-ston, cf. Arenes (20D2B: 6-29).
144 Cf. Jackson (1987: 335, 385 n. 56).
145 Cf. Ruegg (1985: 310), Arenes (20D2B: 8).
146 Cf. Broido (1984: 11-21,25-26), Ruegg (1985: 310,322 n. 10).
It is certainly conceivable that (elements in) this set of 'six alternatives'
and. other hermeneutical categories with which they are often associated
may have circulated in Indic non-Tantric milieus contemporaneous with
or prior to their appearance in Tantristic exegetical practice, as Arenes has
speculated recentlyl47. This applies in particular to the three sets of oppo-
sites brought together in the which are known as separate dicho-
tomies (i.e. not integrated as a set of six) in non-Tantric Buddhism, albeit
not per se with the same function or meaning
. Most notably t1s is the
case for the nitiirtha / neyiirtha (Tib. nges-pa'i-don / drang-ba'i-don) pair
which is attested in early Buddhism as well as in (relatively early)
yiina sources as a pivotal set of hermeneutical criteria 149. It is significant,
in any case, that influential Tibetan scholastic authorities such as Sa-paI).
and Padma-dkar-po did not hesitate to use the system of the 'six alterna-
tives' in their interpretation of Siitra scripture.
In his comments Sa-paI). associates the first 'alternative' in his list with
the four types of 'intention' (Skt. abhipriiya, Tib. dgongs-pa) and the four
types of 'allusion' (Skt. abhisarrzdhi, Tib. Idem-dgongs)150 and he refers to
Mahiiyiinasiitriilarrzkiira as a source for further information on this topic
In his treatment of the fourth 'alternative' the author refers to Sarrzdhinir-
mocanasiitra for the same purpose
. Sa-paI). concludes the section on the
'six alternatives' with a statement which again shows that our author indeed
applies these hermeneutical categories to Siitras and Tantras alike, adding a
quotation from the Pradipoddyotanii by (the Tantristic author) Candrakirti
147 Arenes (2002B: 29-37); cf. also Steinkeilner (1978: 449,451-452).
148 Cf. Steinkeilner (1978: 451-452), Arenes (2002B: 8).
Cf. Lamotte (1949: 349-359) = (1988: 16-23), Broido (1983B: 21), (1988: 72),
SIBH 2: 123-130.
ISO On abhipriiya and abhisaTJ1.dhi, cf. e.g. Broido (1984), (1985), Ruegg (1985), (1989).
lSI 'di-dag-rgyas-par-yi-ge-mangs-kyis-dgos-pas-ma-bris-te / mdo-sde-rgyan-Ia-sogs-
par-blta-bar-l7ya' 0, 204r 1-204r2. A locus classicus for these categories is indeed Mahiiyiina-
sutriilaTJ1.kiira which at 12.16-18 introduces the four types of abhisaTJ1.dhi (16-17) and the
four types of abhipraya (18), ed. Bagchi (1970: 80), cf. e.g. Broido (1984: I, 23-24), Ruegg
(1985: 310).
152 'di-dag-rgyas-par-dgongs-' grel-Ia-sogs-par-shes-par-bya' 0, 204r4; the SafJ1.dhinir-
mocana indeed being a classical source for the categorization at hand here, vi2. the nitiirtha /
neyiirtha opposition, which it discusses e.g. in 7.30 and 7.32 (also abhipraya in 10.11 and
[abhHsaTJ1.dhi in 7.29, 8.24 and 10.8).
153 mtha' -drug-mi-shes-na-mdo-rgyud-gang-bshad-kyang-nor-bar-' gyur-te / slob-dpon-
zla-ba-grags-pas / mtha' -drug-bral-bas-nges-par-rtogs-zhes-gang-smra-ba / / zla-ba-lta-
, If one does not know the 'six alternatives', whatever Sutra or Tantra one is
explaining, errors will occur, as is stated by master Candraldrti:
"One who claims to have an unerring understanding [of the scripture]
without [applying] the 'six alternatives',
is like one who desires to look at the moon, [but] looks [only] Oat the
fmgertip [pointing to the moon]."
The simile of looking solely at the pointing finger and not at that which
the fmger points at, especially in the context of the henneneutical discourse
we have here, is of course reminiscent of the well-known passage in the
Laftkiivatarasiitra, where the Buddha speaks about the distinction between
word and meaning, comparing a word to a pointing finger and its mean-
ing to what the finger points out, warning his disciples not to stare only
at the finger (i.e. the word) and thereby fail to see that which it is point-
ing at (i.e. its meaning) and, so to speak, miss the point

Finally, at the very end of the section on 'objections and rebuttals' Sa-
pal). also refers to his own major work on logic, the Tshad-ma-rigs-pa'i-
gter for further reading on this topic
(4) Concluding Observations
Sa-pal).'s MI has proven to be an eminently important source of infor-
mation on the early foundations of Tibetan scholasticism, a feature inex-
tricably linked with the monastically organized fonns of Buddhism which
were to become dominant in Tibet from the twelfth century onwards. Its
three chapters are devoted to a triad later to become classical in Tibetan
scholastics, namely 'composition', 'exposition' and 'debate' respectively.
The ideal of palJrjitya, of scholarly excellence based on the classical Indian
models, which Sa-pal). sets forth in MI and elsewhere in his (uvre, involves
a wide range of scholastical disciplines. This is particularly highlighted
by the truly impressive listing of classical Indian sources for his work, enu-
merated under twelve gemes, which Sa-pal). offers at the outset of MI.
'dod-sor-rno 'i-rtse-la-lta-dang-rntshungs II zhes-gsungs-pa-ltar-ro, 204v3-204v4; cf. Arlmes
(2003: 16).
154 Ed. Nanjio (1956: 196), cf. Lamotte (1949: 347-348) = (1988: 15).
155 brgal-lan-gyi-tshul- 'di-legs-par-shes-par- 'dod-na I rigs-pa 'i-gter-du-blta-bar-bya' 0,
In the present article I have focussed in particular on matters of hermeneu-
tics in the fIrst two chapters. There we find notions on language and scrip-
tural interpretation stemming from Abhidharma and Mahayana literature,
from Buddhist epistemology (Dignaga and DharmakIrti being referred to
explicitly) and from Sanskrit indigenous grammar. The second chapter of
MJ was based on the structural scheme of Vasubandhu's Vyiikhyiiyukti,
which distinguishes five exegetical categories, viz. 'intention', 'summary',
'meaning of the words', 'connection' and 'objections and MJ
does not follow Vyakhyayukti all too closely; Sa-paI). often deals with the
fIve categories more or less in his own way, in part no doubt due to the
fact that Sa-paI). was introducing a Tibetan readership to a non-indige-
nous originally Indian literature, whereas Vasubandhu was addressing an
Indian audience. We have seen that a considerable variety of grammatical
and interpretational notions and devices pass under review here, involving
forms of exegesis which are particular to Slitric as well as Tantric liter-
Zhang Yisun (1985).
Verhagen (1994)
Verhagen (2001A)
Mkhas-pa- 'jug-pa 'i-sgo
Sa-skya PaI.1Qita Kun-dga' -rgyal-mtshan
Verhagen (200lB)
Verhagen (2001C)
Verhagen (2002)
Verhagen (forthcoming A)
Verhagen (forthcoming B)
Abhyankar, K.V. (& Shukla, 1M.) (1977). A Dictionary of Sanskrit Grammar, sec-
ond revised edition (1: 1961), Baroda. (= Gaekwad's Oriental Series 134)
Arenes, P. (2002A). 'Henneneutique des Tantra: Le Ye shes rdo rje kun las btus
pa'i rgyud las 'byung ba'i rgyan bdun rnam par dgrol ba de Sraddha-
karavarman', in: H. Blezer (ed.), Religion and Secular Culture. Tibetan
Studies II. PlATS 2000: Tibetan Studies: Proceedings of the Ninth Seminar
of the International Association for Tibetan Studies, Leiden 2000, Leiden-
Boston-Koln: Brill, 2002, (= Brill's Tibetan Studies Library 2/2), 163-
Arenes, P. (2002B). 'Henneneutique des Tantra: les "Six extremes (ou possibilites
alternatives)" mtha' drug). A propos d'un exemple de pregnance
des modeles exegetiques des satra', Revue d'Etudes Tibetaines 1,4-43.
Arenes, P. (2003). 'De l'utilite de l'henneneutique des Tantra Bouddhiques a
propos d'un expose de l'appareil des "Sept Ornements" par un doxologue
erudit Dge lugs pa Dbal mang Dkon mchog rgyal mtshan (1764-1863)',
Revue d'Etudes Tibetaines 4, 4-60.
Bacot, I. (1957). Zugiiiima, Texte et Traduction, 2 vols., Paris,
Bagchi, S. (ed.) (1970). MaMyiina-satriilaT(lkiira of Asailga, Darbhanga. (= Bud-
dhist Sanskrit Texts 1.3)
Bosson, I.E. (1969). A Treasury of Aphoristic Jewels. The
of Sa Skya Pa1J4ita in Tibetan and Mongolian, Bloomington The Hague.
(= Indiana University Publications, Uralic and Altaic Series, 92)
Broido, M. (1983A). 'A note on dgos-'brel', Journal of the Tibet Society 3, 5-
Broido, M. (1983B). 'Bshad thabs: Some Tibetan Methods of Explaining the
Tantras', in: Steinkellner, E. & Tauscher, H. (eds.) (1983). Contributions on
Tibetan and Buddhist Religion and Philosophy. Proceedings of the Csoma
de Karas Symposium held at Velm-Vienna, Austria, 13-19 September 1981,
vol. 2, Wien (= Wiener Studien zur Tibetologie und Buddhismuskunde 11),
Broido, M. (1984). 'Abhipriiya and Implication in Tibetan Linguistics', Journal
of Indian Philosophy, 12, 1-33.
Broido, M. (1985). 'Intention and Suggestion in the Abhidharmakosa: Sandhii-
Revisited', Journal of Indian Philosophy, 13, 327-381.
Broido, M. (1988). 'Killing, Lying, Stealing and Adultery: A Problem of Inter-
pretation in the Tantras', in: Lopez, D.S., jr. (ed.) (1988). Buddhist Hermeneu-
tics, Honolulu: University of Hawaii Press, 71-118.
Bsod-nams-rgya-mtsho (ed.) (1968). Sa-skya-pa 'i-bka '-'bum. The complete works
of the great masters of the Sa skya sect of the Tibetan Buddhism, 15 vols,
Tokyo (= Bibliotheca Tibetica I, 1-15).
Dreyfus, G.B.I. (1997). Recognizing Reality. Dharmakrrti's Philosophy and Its
Tibetan Interpretations, Albany: State University of New York Press. (= SUNY
Series in Buddhist Studies)
Dreyfus, G.B.I.(2003). The Sound of Two Hands Clapping. The Education of a
Tibetan Buddhist Monk, Bekeley-Los Angeles-London: University of Cali-
fornia Press.
Edgerton, F. (1953). Buddhist hybrid Sanskrit grammar and dictionary. 2: Dic-
tionary, New Haven.
Eimer, H. (1983) [C]. Rab tu 'byun ba'i gzhi. Die tibetische Ubersetzung des
Pravrajyiivastu im Vinaya der Mfllasarviistiviidins, 2 vols., Wiesbaden. (=
Asiatische Forschungen 82)
Franco, E. (1984). 'On the interpretation of Pramfu).asamuccaya(vpti) I, 3d', J our-
nal of Indian Philosophy 12, 389-400.
HahD, M. (1981). 'Zu einer Neuedition von Nagananda', Wiener
ZeitschriJt fiir die Kunde Siidasiens 25, 131-137.
Hahn, M., Steiner, R & M.C. (1991). The Recensions of the Niigiinanda
by Vol. 1: The North Indian Recension, Edited by Madhava
Candra assisted by Kamala Bhattaciirya. With a general intro-
duction by M. Hahn, and a preface and a bibliography of the editions and
translations of the Nagananda by R Steiner, New Delhi.
Hardy, E. (ed.) (1902). The Netti-pakaralJa with Extracts from Dhammapiila's
Commentary, London: Pali Text Society.
Hattori, M. (1968). Digniiga, On Perception, being the of
Digniiga's PramiilJasamuccaya,from the Sanskrit fragments and the Tibetan
versions, translated and annotated, Cambridge, Mass.: Harvard University
Press. (= Harvard Oriental Series 47)
Hiniiber, O. von (1996). A Handbook of Piili Literature, Berlin - New York: De
Gruyter. (= Indian Philology and South Asian Studies 2)
Ishikawa, M. (ed.) (1990). A Critical Edition of the Sgra sbyor bam po gnyis pa.
An Old and Basic Commentary on the Mahiivyutpatti, Tokyo (= Studia Tibe-
tica 18, Materials for Tibetan-Mongolian Dictionaries vol. 2).
Jackson, D.P. (1987). The Entrance Gate for the Wise (Section III). Sa-skya
Pm:z4ita on Indian and Tibetan Traditions for PramiilJa and Philosophical
Debate, 2 vols., Wien (= Wiener Studien zur Tibetologie und Buddhismus-
kunde 17).
Kanno, H. (2003). 'Chinese Buddhist Siitra Commentaries of the Early Period',
Annual Report of The International Research Institute for Advanced Bud-
dhology at Soka University [Academic Year 2002] 6,301-320.
Kapstein, M. (2003). 'The Indian Literary Identity in Tibet', in: S. Pollock (ed.),
Literary Cultures in History. Reconstructions from South Asia, Berkeley-Los
Angeles-London: University of California Press, 747-802.
Khetsun Sangpo [Mkhas-btsun-bzang-po] (1973-1980). Biographical Dictionary
of Tibet and Tibetan Buddhism, 11 vols, Dharamsala.
Kuijp, L.W.J. van der (1996). 'Tibetan Belles-Lettres: The Influence of Dar;t-
qin and in: Cabez6n, J.I. & Jackson, RR (eds.) (1996).
Tibetan Literature. Studies in Genre. Essays in Honor of Geshe Lhundup
Sopa, Ithaca, New York: Snow Lion (Studies in Indo-Tibetan Buddhism),
Lamotte, E. (1949). 'La critique d'interpretation dans Ie bouddhisme', Annuaire
de /'Institut de Philologie et d'Histoire Orientales et Slaves, 9, Melanges
Gregoire, (Bruxelles), 341-361. English translation of Lamotte (1949):
[Boin-Webb, S. (tr.)] 'The Assessment of TextUal Interpretation in Buddhism',
m: Lopez, D.S., jr. (ed.) (1988). Buddhist Hermeneutics, Honolulu: Univer-
sity of Hawaii Press, 11-27.
Miller, R.A. (1993). Prolegomena to the first two Tibetan grammatical treatises,
Wien. (= Wiener Studien zur Tibetologie und Buddhismuskunde 30) ..
Nlll:lamoli (trl.) (1962). The Guide (Netti-ppakaral)al'{l) According to Kacciina
Thera Translatedfrom the Pali by Bhikkhu Nal)amoli, London (= Pali Text
Society, Translation Series 33).
Nlll:lamoli (trl.) (1964). The Pi{aka-Disclosure (Pe{akopadesa) According to Kac-
cana Thera Translatedfrom the Pali by Bhikkhu Nal)amoli, London (= Pali
Text Society, Translation Series 35).
Nanjio, B. (ed.) (1923; repr. 1956). The Lankavatara Sutra, Kyoto
Nattier, J. (2003). 'The Ten Epithets of the Buddha in the Translations of Zhi
Qian', Annual Report of The International Research Institute for Advanced
Buddhology at Soka University [Academic Year 2002] 6,207-250.
Newman, J. (1998). 'Islam in the Kalacakra Tantra' , Journal of the International
Association of Buddhist Studies 21.2.
Renou, L. (1940). La Durgha{avrtti de Saral)adeva, traite grammatical en sanskrit
du xiie siecle, edite et traduit. Volume I,fascicule I: Introduction, Paris.
(Collection Emile Senart 6)
Renou, L. (1942). Terminologie grammaticale du Sanskrit, 3 vols., Paris. (= Bib-
liotheque de l'Ecole des Hautes Etudes, Sciences Historiques et Philologiques
280-282) [Reprint in 1 vol.: Paris, 1957]
Ruegg, D.S. (1985) & (1988). 'Purport, implicature and presupposition: Sanskrit
abhipraya and Tibetan dgons pa/dgons gzhi as hermeneutical concepts',
Journal of Indian Philosophy, 13, 309-325 & Journal of Indian Philosophy,
16, 1-4.
Ruegg, D.S. (1989). 'Allusiveness and Obliqueness in Buddhist Texts: sG/'(ldha,
sal'{ldhi, sal'{ldhyii and abhisarrzdhi', in: C. Caillat (ed.), Dialectes dans les
litteratures indo-aryennes, Paris, 295-328.
Ruegg, D.S. (1995). Ordre spirituel et ordre temporel dans la pensee Bouddhique
de I'Inde et du Tibet. Quatre conferences au College de France, Paris:
Boccard. (= Publications de l'Institut de Civilisation Indienne, Serie in-8o,
Sakaki, R. (ed.) (1916-1925). Mahavyutpatti, part 1 & 2, Kyoto (= Kyoto Imperial
University Series 3).
Smith, E.G. (2001). Among Tibetan Texts. History and Literature of the Hima-
layan Plateau, Boston: Wisdom Publications (Studies in Indian and Tibetan
Steinkellner, E. (1978). 'Remarks on Tantristic Hermeneutics', in: Ligeti, L. (ed.)
(1978), Proceedings of the Csoma de Karas Memorial Symposium held at
Matrafored, Hungary 24-30 September 1976, Budapest: Akaderniai Kiado,
Tachikawa, M. (1983). A Catalogue of the United States Library of Congress Col-
lection of Tibetan Literature in Microfiche, Tokyo. (= Bibliographia Phi1o-
logica Buddhica, Series Maior Ill)
Thurman, R.A.F. (1988). 'VajraOHermeneutics', in: Lopez, D.S., jr. (ed.) (1988).
Buddhist Hermeneutics, Honolulu: University of Hawaii Press, 119-148.
Tillemans, T.J.F. (1999). Scripture, Logic, Language. Essays on Dharmak'irti and
his Tibetan Successors, Boston: Wisdom Publications. (= Studies in Indian
and Tibetan Buddhism)
Tucci, G. (1949). Tibetan Painted Scrolls, 3 vols., Roma.
Vasu, S.C. (ed., trl.) (1891). The A ~ t i i d h y i i y i of PiilJini, edited and translated into
English, 2 vols., Allahabad. (repr. Delhi 1962)
Verhagen, P.e. (1988). 'Tibetan Expertise in Sanskrit Grammar - a Case Study:
Grammatical Analysis of the Term Pratitya-samutpiida', lournal of the Tibet
Society, 8, 1988 [1996],21-48.
Verhagen, P.C. (1994), A History of Sanskrit Grammatical Literature in Tibet.
Volume I: Transmission of the Canonical Literature, Leiden - New York -
Koln: E.J. Brill (= Handbuch der Orientalistik Abt. 2 Bd. 8).
Verhagen, P.C. (2001A). A History of Sanskrit Grammatical Literature in Tibet.
Volume 2: Assimilation into Indigenous Scholarship, Leiden - Boston - Koln:
E.J. Brill (= Handbuch der Orientalistik Abt. 2 Bd. 8.2).
Verhagen, P.C. (200lB). 'Studies in Indo-Tibetan Buddhist Hermeneutics (1):
Issues of Interpretation and Translation in the Minor Works of Si-tu Pru;t-chen
Chos-kyi-'byung-gnas (1699?-1774)', lournal of the International Associa-
tion of Buddhist Studies, 24.1 (2001), 61-88.
Verhagen, P.C. (2001C) 'Studies in Indo-Tibetan Buddhist Hermeneutics (2):
Notes Apropos to the Role of Hermeneutics in Buddhist Sectarianism in
India and Tibet', The Indian International Journal of Buddhist Studies 2
(Special Issue including Proceedings of the Colloquium on "Buddhism:
Schoqls and Sects" held at the Catholic University of Leuven (Belgium) on
May 26-27, 2000), 121-134.
Verhagen, P.C. (2002). 'Studies in Indo-Tibetan Buddhist Hermeneutics (3):
; Grammatical Models in Buddhist Formulas', in: H. Blezer (ed.), Religion
and Secular Culture. Tibetan Studies II. PlATS 2000: Tibetan Studies: Pro-
ceedings of the Ninth Seminar of the International Association for Tibetan
Studies, Leiden 2000, Leiden-Boston-Koln: Brill (= Brill's Tibetan Studies
Library 2/2), 143-161.
Verhagen, P.C. (forthcoming A). 'Studies in Indo-Tibetan Buddhist Hermeneu-
tics (4): The Vyiikhyiiyukti by Vasubandhu', in: lournai Asiatique.
Verhagen, P.C. (forthcoming B). 'Studies in Indo-Tibetan Buddhist Herme-
neutics (7): Sa-skya Pru;tQita's Mkhas-'jug on the Sanskrit-Tibetan Inter-
face: Synthesis, Comparison and Translation', in: Proceedings of the
Seminar of the International Association of Tibetan Studies (Oxford
Vogel, C. (1965). Vtigbhata's The first five chapters of
its Tibetan version. Edited and rendered into English with the original San-
skrit. Accompanied by a literary introduction and a running commentary on
the Tibetan translating-technique, Wiesbaden. (= Abhandlungen fiir die Kunde
des Morgeulandes XXXVII,2) .
Vogel, C. (1976). Srfdharasenas ViSvalocana. Ein laina-Worterbuch des Sanskrit
im lamaistischen Kanon, Giittingen. (= Nachrichten der Akademie der Wis-
senschaften in Giittingen, Philologisch-Historische Klasse, Jhg. 1976, Nr. 8)
Vogel, C. (1979). Indian Lexicography, Wiesbaden. (= A History of Indian Lite-
rature, ed. J. Gonda, vol. V Fasc. 4)
Zhang Yisun (1985). Bod-rgya-tshig-mdzod-chen-mo (Zang Han Da Cidian) ,
Beijing: Renmin Chubanshe.
The present index covers the issues of the Journal of the International
Association of Buddhist Studies from 1999 (nABS 22.1) until 2003
(nABS 26.2). We have by and large taken as our model the index com-
piled by Bruce Cameron HALL in nABS 10.2 (1987), Le., titles are given
as the main entries for all items. Also following HALL, the titles of books
being reviewed have been placed in italics. The reader can consult the
explanations to be found in HALL 1987 (pp. 181-2) for more information
on the organizational principles which we have adopted.
ABEYSEKARA, Ananda see:
Politics of Higher Ordination, Buddhist Monastic Identity, and Leadership
at the Dambulla Temple in Sri Lanka.
AGOSTINI, Giulio see:
On the Nikaya Affiliation of the Snghanacarasailgraha and the Sphutartha
Snghanacarasailgraha!Ika. .
ARNOLD, Dan see:
CandrakIrti on Dignaga on
(The) Bangkok Conference on Buddhist Studies: Introduction. Donald K. SWEA-
RER. 22/2 (1999): 397-399.
Buddhist Studies in Germany and Austria 1971-1996 with a Contribution on
East Asian Buddhism by Michael Friedrich. Eli FRANCO. 22/2 (1999): 401-
BECHERT, Heinz [et al.] see:
Der Buddhismus I: Der Indische Buddhismus und seine Verzweigungen.
Bhadanta Rama: A Sautr1lntika before Vasubandhu. Takurni FUKUDA. 26/2 (2003):
BLACKBURN, Anne M. see:
Looking for the Vinaya: Monastic Disciple in the Practical Canons of the
Blood Writing in Chinese Buddhism. John KIEsCHNICK. 23/2 (2000): 177-194.
Journal of the International Association of Buddhist Studies
Volume 28 Number 1 2005
BOUCHER, Daniel see:
On Hu and Fan Again: the Transmission of "Barbarian" Manuscripts to
BUCKNELL, Roderick S. see:
Conditioned Arising Evolves: Variation and Change in Textual Accounts of
the Paticca-samuppiida Doctrine.
(The) Buddha's Remains: mantra in MaftjuSrimulakalpa. Glenn WALLIS. 24/1
(2001): 89-125.
(Der) Buddhismus I: Der Indische Buddhismus und seine Verzweigungen: Heinz
BECHERT let al.]. Review by Ulrich PAGEL 24/1 (2001): 127-133.
Buddhist Studies in Germany and Austria 1971-1996 with a Contribution on East
Asian Buddhism by Michael Friedrich. Eli FRANCO. 22/2 (1999): 401-456.
CAILLAT, Colette see:
Gleanings from a Comparative Reading of Early Canonical Buddhist and
Jaina Texts.
Can all Beings Potentially Attain Awakening? Gotra-theory in the Mahiiyiina-
sutriilal"(lkiira. Mario D'AMATO. 26/1 (2003): 115-138.
CandrakIrti on Dignaga on Dan ARNOLD. 26/1 (2003): 139-174.
CHEN, Jinhua see:
One Name, Three Monks: Two Northern Chan Masters Emerge from the
Shadow of Their Contemporary, the Tiantai Master Zhanran (711-782).
Sanra and Scepter. Empress Wu's Political Use of Buddhist Relics.
Chinese Nuns and their Ordination in Fifth Century in China. Ann HEIRMAN. 24/2
(2001): 275-304.
Coming of Age: Buddhist Studies in the United States from 1972 to 1997. Frank
E. REYNOLDS. 22/2 (1999): 457-483.
Conditioned Arising Evolves: Variation and Change in Textual Accounts of the
Pa(icca-samuppiida Doctrine. Roderick S. BUCKNELL 22/2 (1999): 311-342.
Report on the XIIth Conference of the IABS. Oskar VON IIINDBER. 23/1
(2000): 155-159.
Accounts of the XIIth IABS Conference .. Oskar VON HINtiBER. 23/1 (2000):
The XIIlth Conference of the International Association of Buddhist Studies,
Bangkok 8-13 December 2002, First Circular. 23/2 (2000): 351-354.
Report on the XIIlth Conference of the IABS. Oskar VON lliNOBER. 26/2 (2003):
INDEX TO JIABS 22-26 223
Constructing Another Perspective for AjaIJ.ta's Fifth-Century Excavations.
Nadine OWEN. 2.111 (2001): 27-59.
D'AMATO, Mario see:
Can all Beings Potentially Attain Awakening? Gotra-theory in the Mahi!ya-
DASSEIN, Bart see:
Sautriintika and the FIfdaya Treatises.
DEEGALLE, Mahinda see:
A Search for Mahayiina in Sri Lanka.
Doctrinal Refonnation of the Hongzhou School of Chan Buddhism. Jinhua JIA.
24/1 (2001): 7-26.
DRAGONETTI; Carmen see:
Marginal Note on the Idealistic Conception of citta-miitra.
DUCOR, Jerome see:
Shandao and Honen. Apropos of Julian F. Pas's book Visions of Sukhavatf
(English Summary).
DUCOR, Jerome see:
Shandao et Honen, a propos du livre de Julian F. Pas: Visions' of Sukhavatf.
Enacting Words. A Diplomatic Analysis of the Imperial Decrees (bkas bcad) and
their Application in the sGra sbyor bam po gills pa Tradition. Cristina A. SCHER-
RER-SCHAUB. 25/1-2 (2002): 263-340.
E ~ i i agrii: Images of Nuns in (Miila-)Sarviistiviidin Literature. Peter SKllliNG. 24/2
(2001): 135-156.
(The) Finances of a Twentieth Century Buddhist Mission: Building Support for
the Theravada Nuns' Order of Nepal. Sarah LEVINE. 24/2 (2001): 217-239.
FRANCO, Eli see:
Buddhist Studies in Gennany and Austria 1971-1996 with a Contribution on
East Asian Buddhism by Michael Friedrich.
PumA, Kotatsu see:
In Memoriam Professor Akira Hirakawa.
FuKUDA, Takumi see:
Bhadanta Rama: A Sautriintika before Vasubandhu.
General Introduction. Robert KRITzER. 2612 (2003): 201-224.
Gleanings from a Comparative Reading of Early Canonical Buddhist and Jaina
Texts. Colette CAILLAT. 26/1 (2003): 25-50.
224 .rrRiHOLBA
GUTS CHOW, Kim see:
What Makes a Nun? Apprenticeship and Ritual Passage in Zanskar, North India
HARA, Minorusee:
In memoriam J.W. de Jong:
HARRISON, Paul M. see:
Relying on the Dhanna and not the Person: Reflections on Authority and
Transmission in Buddhism and Buddhist Studies.
lIEIRMAN, Ann see:
Chinese Nuns and their Ordination in Fifth Century in China.
What Happened to the Nun Maitreyi?
(A) Hermeneutical Problem in SN 42, 12 (SN IV, 333) and AN X, 91 (AN V, 178).
Bhikkhu PASADIKA. 23/1 (2000): 147-154.
HlNOBER, Oskar von see:
Accounts of the XIIth IABS Conference.
HlNOBER, Oskar von see:
Report on the XIIth Conference of the IABS.
HlNOBER, Oskar von see:
Report on the XIIIth Conference of the IABS.
HONJo, Yoshifumi see:
The Word Sautrantika.
In the Beginning ... Guanding ~ r m (561-632) and the Creation of Early Tiantai.
Linda PENKOWER. 23/2 (2000): 245-296.
Introduction to Alexander von Stael-Holstein's Article "On a Peking Edition of
the Tibetan Kanjur Which Seems to be Unknown in the West". Edited for
publication by Jonathan A. SILK. 22/1 (1999): 211-249.
JIA, Jinhua see:
Doctrinal Reformation of the Hongzhou School of Chan Buddhism.
JONES, Charles B. see:
Mentally Constructing What Already Exists: The Pure Land Thought of Chan
Master Jixing Chewu (1741-1810).
INDEX TO JIABS 22-26 225
Blood Writing in Chinese Buddhism.
Knowing All through Knowing One: Mystical Communion or Logical Trick in
the TattvasaJ'!'lgraha and Tattvasa/?1grahapaftjika. Sara MCCLINTOCK. 23/2
(2000): 225-244. .
KRITZER, Robert see:
General Introduction.
Sautrantika in the
LEVINE, Sarah see:
The Finances of a Twentieth Century Buddhist Mission: Building Support
for the Theravada Nuns' Order of Nepal.
(The) Life of dGe slong rna dPal mo: The Experience of a Leper, Founder of a Fast-
ing Ritual, a Transmitter of Buddhist Teachings on Suffering and Renunciation
in Tibetan Religious History. Ivette M. VARGAS-O'BRIAN. 24/2 (2001): 157-185.
Looking for the Vinaya: Monastic Disciple in the Practical Canons of the Thera-
vada. Anne M. BLACKBURN. 22/2 (1999): 281-309. .
Marginal Note on the Idealistic Conception of citta-miitra. Carmen DRAGONETTI.
23/2 (2000): 165-175.
Marginal Notes on a Study of Buddhism, Economy and Society in China. Jonathan
A. SILK. 22/2 (1999): 359-396.
MATHES, Klaus-Dieter see:
Taranatha's Presentation of trisvabhiiva in the gian ston sftin po.
Knowing All through Knowing One: Mystical Communion or Logical Trick
in the TattvasaJ'!'lgraha and TattvasaJ'!'lgrahapaftjika.
McDANIEL, Justin T. see:
Transformative History. Nihon Ry5iki and JinakiilamiilipakaralJam.
MEINERT, Carmen see:
Structural Analysis of the bSam gtan mig sgron. A Comparison of the Four-
fold Correct Practice in the A.ryiivikalpapraveaniimadhiiralJi and the
tents of the Four Main Chapters of the bSam gtan mig sgron.
Mentally Constructing What Already Exists: The Pure Land Thought of Chan Mas-
ter Jixing Chewu (1741-1810). Charles B. JONES. 23/1 (2000):43-70.
Nagrujuna and the RatnavalI. New Ways to Date an Old Philosopher. J oseph WALSER.
25/1-2 (2002): 209-262.
NATTIER, Jan see:
The Realm of A1a;obhya: A Missing Piece in the History of Pure Land Bud-
Nuns, Laywomen, Donors, Goddesses. Female Roles in Early Indian Buddhism.
Peter SKlLLING. 24/2 (2001): 241-274.
OBITUARIES [Chronological]:
In memoriam Prof. Hajime Nakamura. Jikido TAKASAKI. 23/1 (2000): 1-5.
In memoriam J.W. de Jong. Minoru HARA. 24/1 (2001): 1-5.
In Memoriam Professor Akira Hirakawa. Kotatsu FumAI. 26/1 (2003): 3-7.
OHNUMA, Reiko see:
The Story of RupavatI: A Female Past Birth of the Buddha.
On an Alleged Reference to Amitabha in a Inscription on a Gandharian
Relief. Richard SALOMON and Gregory SCHOPEN. 25/1-2 (2002): 3-31.
On Hu and Fan Again: the Transmission of "Barbarian" Manuscripts to China.
Daniel BOUCHER. 23/1 (2000): 7-28.
On the Nikaya Affiliation of the Snghanacarasangraha and the Sphutarilia Sn-
ghanacarasangrahatIka. Giulio AGOSTINI. 26/1 (2003): 97-114.
On the School Affiliation of "Sautrantika" or "Yogacara"? Nobuyoshi
Y AMABE. 26/2 (2003): 225-254.
One Name, Three Monks: Two Northern Chan Masters Emerge from the Shadow
of Their Contemporary, the Tiantai Master Zhanran (711-782). Jinhua
CHEN. 22/1 (1999): 1-91.
OWEN, Nadine see:
Constructing Another Perspective for Ajru:tta's Fifth-Century Excavations.
PAGEL, Ulrich see:
Three Bodhisattvapitaka Fragments from Tabo: Observations on a West Tibetan
Manuscript Tradition.
Der Buddhismus I: Der Indische Buddhismus und seine Verzweigungen.
PASADIKA, Bhikkhu see:
A Hermeneutical Problem in SN 42, 12 (SN IV, 333) and AN X, 91 (AN V,
PENKOWER, Linda see:
In the Beginning ... Guanding jlJY1 (561-632) and the Creation of Early
Politics of Higher Ordination, Buddhist Monastic Identity, and Leadership at the
Dambulla Temple in Sri Lanka. Ananda ABEYSEKARA. 22/2 (1999): 255-280.
1 A correction: the name of this author was incorrectly spelled in the table of context
of JIABS 26.12003 as "Kotabo Fujija".
INDEX TO JIABS 22-26 227
(The) Realm of A Missing Piece in the History of Pure Land Buddhism.
Jan NATIIER. 23/1 (2000): 71-102.
Relying on the Dharnla and not the Person: Reflections on Authority and Trans-
mission in Buddhism and Buddhist Studies. Paul M. HARRISON. 26/1 (2003):
REYNOLDS, Frank E. see:
Coming of Age: Buddhist Studies in the United States from 1972 to 1997.
SALOMON, Richard see:
On an Alleged Reference to Amitabha in a Kharo!jJ:hI Inscription on a Gan-
dhiirian Relief.
Sarfra and Scepter. Empress Wu's Political Use of Buddhist Relics. Jinhua CHEN.
25/1-2 (2002): 33-150.
Sautrantika and the FJ:rdaya Treatises. Bart DASSEIN. 26/2 (2003): 287-319.
Sautrantika in the Robert KR.rrzIlR. 26/2 (2003): 331-384.
SCHERRER-SCHAUB, Cristina A. see:
Enacting Words. A Diplomatic Analysis of the Imperial Decrees (bkas bc:ad)
and their Application in the sGra sbyor bam po gills pa Tradition.
lABS Treasurer Final Financial Report.
SCHOPEN, Gregory see:
On an Alleged Reference to Amitabha in a Kharo!jJ:hI Inscription on a Gan-
dhiirian Relief.
(A) Search for Mahayana in Sri Lanka, Mahinda DEEGALLE. 22/2 (1999): 343-357.
Shandao et Honen, a propos du livre de Julian F. Pas: Visions of Sukhavati.
Jerome DUCOR. 22/1 (1999): 93-163.
Shandao and Honen. Apropos of JulianF. Pas's book Visions ofSukhavati (Eng-
lish Summary). Jerome DUCOR. 22/1 (1999): 251-252.
SHARF, Robert H. see:
Thinking through Shingon Ritual.
SILK, Jonathan A. see:
Introduction to Alexander von Stael-Holstein's Article "On a Peking Edition
of the Tibetan Kanjur Which Seems to be Unknown in the West". Edited for
Marginal Notes on a Study of Buddhism, Economy and Society in China.
SKILLING, Peter see:
agra: Images of Nuns in (Miila-)Sarvastivadin Literature.
Nuns, Laywomen, Donors, Goddesses. Female Roles in Early Indian Buddhism.
Vasubandhu and Vyakhyayukti Literature.
228 .n:Ri HOLBA
(The) Story ofRiipavatI: A Female Past Birth of the Buddha. Reiko,OHNUMA. 23/1
(2000): 103-145.
Structural Analysis of the bSam gtan mig sgron. A Comparison of the Fowfold
Correct Practice in the Aryavikalpapravesaniimadhiirm;zf and the Contents
of the Four Main Chapters of the bSam gtan mig sgron. Carmen MEINERT.
26/1 (2003): 175-195.
Studies in Indo-Tibetan Buddhist Hermeneutics (1). Issues of Interpretation and
Translation in the Minor Works of S i ~ t u Pru;t-chen Chos-kyi-'bywi-gnas
(1699?-1774). Peter VERHAGEN. 24/1 (2001): 61-88.
SWEARER, Donald K. see:
(The) Bangkok Conference on Buddhist Studies: Introduction.
TAKASAKI, Jikido see:
In memoriam Prof. Hajime Nakamura.
Tiiranatha's Presentation of trisvabhiiva in the gZan ston sflin po. Klaus-Dieter
MATHEs. 23/2 (2000): 195-223.
Thinking through Shingon Ritual. Robert SHARF. 26/1 (2003): 51-96.
Three Bodhisattvapitaka Fragments from Tabo: Observations on a West Tibetan
Manuscript Tradition. Ulrich PAGEL. 22/1 (1999): 165-210.
Transformative History. Nihon Ryoiki and linakiilamiilipakaralJam. Justin
T. McDANIEL. 25/1-2 (2002): 151-207.
Treasurer's Report 2000. 24/1 (2001): 135.
IABS Treasurer Final Financial Report (2001-2002). Cristina SCHERRER-SCHAUB.
26/2 (2003): 391-394.
VARGAS-O'BRIAN, Ivette M. see:
The Life of dGe slong rna dPal mo: The Experience of a Leper, Founder of
a Fasting Ritual, a Transmitter of Buddhist Teachings on Suffering and
Renunciation in Tibetan Religious History.
Vasubandhu and Vyiikhyiiyukti Literature. Peter SKILLING. 23/2 (2000): 297-350.
VERHAGEN, Peter see:
Studies in Indo-Tibetan Buddhist Hermeneutics (1). Issues of Interpretation
and Translation in the Minor Works of Si-tu Pru;t-chen Chos-kyi-'byUIi-gnas
WALLIS, Glenn see:
The Buddha's Remains: mantra in MafljusrimiUakalpa.
INDEX TO JIABS 22-26 229
WALSER, Joseph see:
Nagarjuna andthe RatnavalI. New Ways to Date an Old Philosopher.
What Happened to tlie Nun MaitreyI? Ann HEIRMAN. 23/1 (2000): 29-41.
What Makes a Nun? Apprenticeship and Ritual Passage in Zanskar, North Ip.dia.
Kim GUTSCHOW. 24/2 (2001): 187-215.
(The) Word Sautriintika. Yoshifumi HONJO. 26/2 (2003): 321-330.
YAMABE, Nobuyoshi see:
On the School Affiliation of "Sautriintika" or "Yogaciira"?
David SEYFORT-RUEGG, Indologist and Tibetologist, President of the IABS
from 1991-1998. Translator of Bu-ston's anthology of Sutra texts on the Tatha-
gatagarbha theory. Author inter alia of several publications on the Madhyamaka
school of Mahayana philosophy. Research Professor in the Department for the
Study of Religions, School of Oriental and African Studies, London University.
Jinhua CHEN teaches East Asian Buddhism at the University of British Colum-
bia, Vancouver, where he also serves as the Canada Research Chair in East Asian
Buddhism. His research interests cover Nara and Heian Japan, Tiantai and
Esoteric Buddhism in Sui-Tang China, meditation and vinaya traditions in 5-7
century China, relic-workship and state-church relationship in medieval China.
Richard D. McBRIDE, II is a post-doctoral fellow in Korean Studies and Bud-
dhist Studies at Washington University in St. Louis. His first book Domesti-
cating the Dharma: Buddhist Cults and the Hwaom Synthesis in Silla Korea
(University of Hawai'i Press) will be out in 2007.
Jacob DALTON is an Assistant Professor in the Department of Religious Stud-
ies at Yale University. His current research centers on tantric ritual manuals
in the Dunhuang collections and on the later history of the Rnying rna school
of Tibetan Buddhism. He is also presently completing a book on the role of
violence in the early Tibetan assimilation of Buddhism.
Pieter Comelis VERHAGEN is University Lecturer of Buddhology and Tibetan
at the Department of Languages and Cultures of South and Central Asia, Leiden
University, the Netherlands. His main publications include A History of Sanskrit
Grammatical Literature in Tibet, vol. 1, Transmission of the Canonical Literature
(Leiden, 1994) and vol. 2, Assimilation into Indigenous Scholarship (Leiden,
2001), and three series of articles: 'Tibetan Expertise on Sanskrit Grammar', 'Stud-
ies in Tibetan Indigenous Grammar' and 'Studies in Indo-Tibetan Buddhist
Hermeneutics'. His current research centers on the principles and techniques
involved in scriptural interpretation within Indian and Tibetan Buddhism.
Jiri HOLBA is a member of the Oriental Institute of the Academy of Sciences
of the Czech Republic. His doctorate, from Charles University in Prague, was
on the Vajracchedikasutra and its Indian commentaries.
Journal of the International Association of Buddhist Studies
Volume 28 Number I 2005
The International Association of Buddhist Studies
Jikido Takasaki
Ernst Steinkellner
Tom Tillemans
General Secretary
Jerome Ducor
Regional Representatives:
Janet Gyatso (Cambridge, MA, USA)
Kazunobu Matsuda (Kyoto)
David Seyfort Ruegg (London)
Peter Skilling (Bangkok)
Directors at large:
Nalini Balbir (Paris), Colette Caillat (Paris),
Georges Dreyfus (Williamstown, MA), Robert M. Gimello (Cambridge, MA),
Oskar von Hiniiber (Freiburg), ShOryil Katsura (Hiroshima),
Liying Kuo (paris), Richard Salomon (Seattle), Cristina Scherrer-Schaub (Lausanne),
Lambert Schmithausen (Hamburg)
The International Association of Buddhist Studies, founded in 1976, is devoted to
promoting and supporting scholarship in Buddhist Studies in all its aspects, past and
present, around the world. Membership is open to scholars of all academic disciplines.
Membership dues are USD 40 for full members, USD 20 for student members,
USD 1000 for life members. Subscriptions to the nABS for libraries and other institu-
tions are USD 70. Dues and subscriptions may be paid in the following ways:
1. by direct bank to bank transfers in US dollars or in the equivalent amount in
Euros or Swiss francs. The transfers should be to the IABS accounts at the Banque
Cantonale Vaudoise, PI. St-Franorois, Ch-1001 Lausanne, Switzerland (SWIFT CODE:
BCVLCH2L, clearing no. 7677), account numbers, 983.51.04 for US dollars,
5042.09.82 for EUR and 983.51.02 for Swiss francs.
2. by Visa or Mastercard
3. by cheque payable to the "Association Internationale d'Etudes Bouddhiques
(lABS)." There will be no supplementary charges for cheques drawn in Swiss
currency on Swiss banks. Otherwise, please add 10 dollars or 15 Swiss francs to cover
our processing charges.
4. by Eurocheque in Swiss francs.
Prospective members from developing countries may contact the Treasurer concerning
subsidized membership rates. Dues and subscriptions are payable per calendar year
by December 31 of the previous year. Payments other than direct bank transfers should
be sent to Dr. Jerome Ducor, Section de langues et civilisations orientales, Universite
de Lausanne, BFSH 2, CH-1015 Lausanne, Switzerland.