You are on page 1of 158








Contents o f Yol. Y H

Vol. V II: Buddhist and Jaina Literature

F ase. 1 : D . Seyfort R uegg T he L iterature o f th e M adhyam aka School
o f P h ilosop h y in India

CIP-Kurzfcftelaufnahme der Deutschen Bibliothek

A history of Indian literature / ed. by Jan Gonda. - Wiesbaden : Harrassowitz.
N E : Gonda, Jan (Hrsg.]
Vol. 7. Buddhist and Jalna literature
Vol. 7. Fasc. 1. - Ruegg, David Seyfort: The literature o f the Madhyamaka School o f Philosophy In India
Rnegg, David Seyfort:
The literature of the Madhyamaka School of Philosophy in India / David Seyfort Ruegg.
Wiesbaden : Harrassowitz, 1981.
(A history of Indian literature ; Vol. 7, Fasc. 1)
ISBN 3-447-02204-3

Otto Harrassowitz, Wiesbaden 1981. Alle Rechte Vorbehalten. Photographische und photomechanische
Wiedergabe nur mit ausdrcklicher Genehmigung des Verlages. Gesamtherstellung: Allguer ZeltungBverlag
GmbH, Kempten. Printed in Germany. Sigel: HIL

Introduction ..................................................................................................... 1
The early period: the formation of the Madhyamaka sch o o l................... .. 4
Nagrjuna ..................................................................................................... 4
Commentaries on N grjunas w o r k s .............................................................. 47
ryadeva . 50
R hulabhadra ................................................................................................. 64
N aga .................................................................................................................. 56
The middle period: the systematization of the Madhyamaka school .. .. 68
B u d d h ap lita.......................................................................................................... 60
B h av av iv e k a.......................................................................................................... 61
Later S v ta n trik a s ................................................................................................67
rg u p ta .................................................................................................................. 67
J a n a g a rb h a .......................................................................................................... 68
C a n d r a k i r t i .......................................................................................................... 71
a n tid e v a ................................................................................................................ 82
The Y ogcra-Madhyamaka s y n th e s is ................................................................. 87
n ta ra k s ita .......................................................................................................... 88
Kamalaila .......................................................................................................... 93
L ater Y og cra-M adhyam ikas......................................................................... 99
The Madhyamaka-Prajnpramit s y n th e s is .....................................................101
Madhyamaka and V a jra y n a ............................................................................... 104
The last period of the Indian Madhyamaka sch o o l.............................................109
Bodhibhadra, Dharm akrti and D pam kararijana.................................... 109
Jaynanda ........................................................................................................ 113
A b h ay k arag u p ta................................................................................................ 114
Other later M d h y a m ik a s ............................................................................... 115
Appendix I: Works on practice ascribed to Avaghosa, A rya-ra, Dharma-
Subhtifghosa), and Mtrceta ....................................................................... 119
Appendix I I : RatnlcaraJntVs Vijapti-M a d h y a m a k a '..............................122
Modern editions of the Sanskrit texts of the Madhyamaka school .. .. 125
Modern editions of the Tibetan versions of works of the Madhyamaka
school .................................................................................................................129
VI Contents

A b b r e v ia tio n s ........................................................................................................ 132

I. T i t l e s ................................................................................................................ 133
II. Names ........................................................................................................ 139
II I. Sanskrit key-words ............................................................................... 141
IV. Tibetan terms ....................................................................................... 146

This outline of the philosophical literature of the Indian Madhyamaka

school is based on the Sanskrit sources to the limited extent they are now
available to us. The major part of this literature is unfortunately not extant in
the original and m ust be presumed lost. However, this loss is at least partly
compensated for by the fact th a t a very large number of im portant works of
this school were translated into Chinese and T ibetan; and for the purposes of
the present publication use has been made chiefly of the translations contained
in the Peking edition of the Tibetan bsTan-gyur in the facsimile reprint of
the Tibetan Tripitaka Research Institute (Tokyd-Kyoto, 1958). The fact th at
we possess few good modern editions of the surviving Sanskrit texts and the
very nature of our translated sources inevitably raise im portant philological
and exegetical problems which it has not always been possible to pursue in the
space available.
Limitations of space have moreover allowed the inclusion of short and hence
selective summaries of only some of the most fundamental texts from the 1000-
year history of the Indian Madhyamaka. The m ajority of the works from this
vast literature have had to be mentioned simply by title, with brief indications
being given as to the contents of some of the more im portant ones. A com
prehensive historical treatm ent and systematic analysis of the Madhyamaka in
its entirety will thus have to await much larger publications. Here an attem pt
has of course been made to determine the relativeand whenever possible also
the absolutechronology of the Madhyamikas; but the same name has not
infrequently been borne by more than one person, and in a number of cases
the attribution of a work to an author as well as even its precise Sanskrit title
remain doubtful and will have to be the subject of future investigation. Be
cause of the paucity of reliable historical materials and disagreements between
some of our sources many questions may remain open indefinitely; furthermore,
the fact th a t most of our texts are now available only in translations into
Chinese and Tibetan makes certain essential kinds of historical, literary and
sometimes even philosophical analysis exceedingly precarious if not quite im
possible. Since relatively few texts and masters of the Madhyamaka school have
hitherto been the subject of monographic study, the present work will in many
cases be a preliminary exploration in well-nigh uncharted territory.
Further research will also be needed in order to establish the relation of
the Madhyamaka with other schools of Indian thought, non-Buddhist as well
as Buddhist. I f this has not been done extensively in the present book (on
earlier connexions with non-Buddhist schools see nevertheless p. 6 note 11),
this is because of the complex historical and methodological questions posed
V IH David Seyfort Ruegg The Literature of the Madhyamaka School

by such an investigation and the lack of space to pursue such problem s; it is

not because we would subscribe to the view th a t the schools of Indian philosophy
are to be kept ap art according to the particular religion of their representatives:
Buddhist (and Jain) philosophers are surely as much p art of the history of
Indian philosophy as orthodox H indu philosophers. I t can a t least be pointed,
out here, however, th a t in his K handanakhandakhdya the twelfth-century
Vedntin Sriharsawho admits the vitanda procedure (see 168; cf. 1416)
has throughout regularly employed a method of reasoning based on what he terms
khandanayukti, which is very close to the M adhyamikas form of argument in
the prasanga-type of reasoning. At a much earlier period of the Vedanta also
the links between the Gaudapadiya- or M ndkya-Kriks (c. 500?) and
Buddhist thought are well-known. Moreover, the continuous debates carried
on between Brhmanical, Buddhist and Jain thinkers led not only to a sharp
ening of minds and a refinement of argument on all sides but also to commonly
recognized methods in philosophical discussion; bu t the precise history of
these developments still remains to be traced in detail.
In writing a work of this kind there arises the question of the extent to
which the secondary literature in Asiatic languages can be included by a
single writer and in a limited space. Scholars from E ast and Central Asia in
particular have in fact devoted themselves over the centuries to translating,
commenting on and explaining the Madhyamaka literature. (In Tibet for
example there has existed for over a millennium a tradition of study of a very
considerable portion of Indian literature, including even works th at are not
specifically B uddhist; and with a view to both translation and exegesis Tibetan
scholars developed remarkable philological and interpretative methods th a t
could well justify us in regarding them as Indologists avant la lettre.) The
sheer bulk of the secondary literature in Chinese, Japanese, Tibetan, and
Mongolian relating to the M adhyamaka is, however, so great th a t it will have
to be the subject of separate publications.1

1 A lthough th e T ibetan secondary literature to w hich it has been possible to

refer in this book cannot pretend to be representative o f all schools, th e works
used are all o f the highest interest. A m ong historical sources these are m ainly the
rGya^garchos^byuri b y Trantha (born in 1575) ed ited b y A. S c h i e f n e r (St.
P etersburg, 1868), th e w ell-known Chos*byun b y B u sto n (12901364) in th e ol
edition reprinted b y L o k e s h C h a n d r a (N ew D elhi, 1971) w ith page references
also to E . O b e r m i l I iE r s incom plete English translation (2 parts, Heidelberg,
193132), and the D e b ther-snon-po b y Gosgon*nu*dpal (13921481) in th e
Kun*bdeglin edition reprinted b y L o k e s h C h a n d r a (N ew D elhi, 1976) w ith page
references also to G. N . R o e r i c h s E nglish translation (The B lu e Annals, 2 volum es,
C alcutta, 194953); hagiographies o f several M adhyam aka m asters are also to be
found collected together in th e B yarrch u V lam -gyiT in rparbla-m a'brgyudp ai*
rnam*par*tharpa b y Ye*esTgyal*mhan (171393) reprinted b y N g a w a n g G e l e k
D e m o (N ew D elhi, 1970). F rom th e rich m ine o f T ibetan exegetical and doctrinal
literature reference has been m ade principally to th e L egsbSad'sniii'po (in the
lHa*sa edition of th e gSurrbum) and the Lam*rimchen*mo (in the old bKra*is*-
Preface IX

The above-mentioned considerations of a philological, bibliographical and

material nature will explain why the present work differs in certain respects
from m any of the other volumes appearing in the same series. Some attem pt
a t a connected account treating this body of literature, however tentative, is
clearly a desideratum at the present time.

lh u irp o edition, reprinted b y N g a w a n g G e :l e k D e m o , N ew D elhi, 1977) by

(3orrkha*pa (13571419); to th e 8ToAthun*chen*mo (in the lHa*sa edition o f th e
gSun*bum) and th e rGyud*sde*spyi*mam (in the edition b y F . D . L e s s i n g and
A. W a y m a n , Fundam entals o f the B uddhist Tantras, The H ague, 1968) b y
m K has'grubdGedegs-dpal-bzah (13851438); and to th e invaluable doxographical
works (G rub'm tha) b y Janrd b yan s'b zad pavNagdbaivbrdon-grus (16481722)
reprinted b y N g a w a n g G e l e k D e m o (New D elhi, 1973), and b y lCan*skya*Rol*-
p a rrdo'rje (171786) reprinted b y L o k e s h C h a n d r a (N ew D elhi, 1977).
A s for th e exten sive and im portant m odern secondary literature from Japan,
references h ave had to be lim ited m ainly to one o f th e few generally available
Japanese publications, th e Indogaku bukkyogaku ken k yu (Journal o f Indian and
B u d d h ist studies, abbreviated IB K ). U sefu l bibliographical surveys o f this litera
ture are to be found in th e publications listed on p. 4 n ote 9.

The Madhyamaka school of M ahayana Buddhism goes back to Nagarjuna,

the great Indian Buddhist philosopher who is placed early in the first mil
lennium P.O .2
The school derives its name of Madhyamaka, i.e. middlemost (Chinese
chung, Tibetan dbu ma), from the fact th at Nagarjuna and his followers
developed a particular theory which avoids the twin dogmatic extremes of
eternalism (asvatavada) and annihilationism (ucchedavdda) rejected by the
B uddhist tradition and keeps to a philosophical via media in a specific and very
interesting manner. While all Buddhists would of course observe the Middle
W ay (madhyama pratipat) in their theory and practice, the philosophers of
this school especially have therefore come to be known as Madhyamikas or
followers of the middle (Tibetan dbu ma pa).3 Similarly, the advocates of the
other great Mahayanist school, the Yogacara/Vijnanavada, are known as
Yogacarins although other Buddhists are of course also practisers of Yoga;
and Aryadevas CatufiSataka, an early treatise of the Madhyamaka school,
indeed makes specific reference in its chapter-colophons to the follower of the
M adhyamaka as a Yogacara. In the Sino-Japanese Buddhist tradition the
Madhyamaka is frequently referred to as the Three Treatises (San-lun/Sanron)
School after its three basic sourcesthe Chung-lun (Madhyamaka^astra),4

2 On N agarjunas date see below, p. 4.

3 T he use o f th e term s m adhyam aka (-d arfa n a ) and m adhyam ika as designations
for th e school and its followers w as perhaps a som ew hat later developm ent. A t
least th e y do n o t seem to be found in the body o f early workse.g . th e M ulam adhya-
m akakarikas (MMK) o f N agarjuna and th e Catuh^ataka o f Aryadeva. T hey appear
in th e oldest available Sanskrit com m entary on th e MMK, Candrakirtis Prasan-
n apada M adhyam akavrttih. B ud d h ap alitas earlier com m entary (unavailable in
Sanskrit) seem s to have also had th e title M ad h yam ak avrtti; and B h av av ivek a s
great treatise was entitled M adhyam akahrdayakarika. The Chinese translations
p oin t to a quite early origin for th e term s (cf. Chung-lun = M adhyam akaiastra).
A s for th e title M ulam adhyam akakarikas, it w as adopted b y L. de L a V a l l ^ e
P o u s s i n in his classical edition, apparently on th e basis o f th e title given in the
T ibetan bsTan*gyur (Prajnanam am ulam adhyam akakSrika = d B u m a rca b a i
dhig le ur b yas pa ses rab ces b ya b a ) ; this title has been retained b y J . W. d e J o n g
in his new edition (Adyar, 1977). This te x t is actu ally available in Sanskrit only in
th e version em bedded in Candrakirtis P rasannapada M adhyam akavrttih; in th is
com m entary no title is in fact indicated (and Candrakirti occasionally speaks o f th e
m adhyam aka i dstra, p. 40. 7; cf. p. 548. 5). In th e M adhyam aka^astrastuti (see
below , p. 8) this te x t is referred to as astragadita-kdrikas.
4 T he Chung-lun (Madhyamaka^astra, Taisho 1564) consists o f N agarjunas verses
(k a rik d ) together w ith th e com m entary o f Ching-m u (see below , p. 48).
2 David Seyfort Ruegg The Literature of the Madhyamaka School

Nagarjunas Shih-erh-men-lun (*Dvada6amukha6astra) and Aryadevas Pai-

lun (&ataSastra)which were translated into Chinese by K um arajiva at the
very beginning of the fifth century.
Because of his basic doctrine th at all factors of existence (dharma) and all
entities (bhava) are em pty of own being (svabhava-6unya) and non-substantial
(nihsvabhava)i.e. not ultim ately real substantial entitiesthe Madhyamika
is also known as a &unya(ta)vadin an d Nifisvabhavavadin.5 And the Madhya
maka theory is designated as the unya(ta)vada, Sunyatadar^ana and
Nibs vabhavavada.8
Over the past half-century the doctrine of the M adhyamaka school, and in
particular th a t of Nagarjuna, has been variously described as nihilism, monism,
irrationalism, misology, agnosticism, scepticism, criticism, dialectic, mysticism,
acosmism, absolutism, relativism, nominalism, and linguistic analysis with
therapeutic value. W ith the exception of the first five which are hardly ap-
propriate in any context and become quite misleading when taken in then-
usual senses, such descriptions no doubt correspond to some aspect of Madhya
maka thought.7 Also, as the unyavada par excellence, the M adhyamaka has

6 In th e course o f th e discussion and refutation o f it in th e MMK and th e rest

of th e M adhyam aka literature, svabhdva own being, self-nature, aseity* has been
defined as som e thin g unproduced (ak rtrim a ) which is independent o f all other
things (nirapeksah paratra; see MMK x v . 2 and x x iv . 33); th ose who postu lated a
svabhdva have indeed conceived o f it as not produced through causal conditioning
(see i. 3 and x v . 1). N ow , th e onto-logical correlate o f a svabhdva is a parabhdva
other being' (cf. i. 3; x v . 3 s q .) ; and th e pair comprised o f svabhdva and parabhdva
is th en described as necessary for establishing any (postulated) bhava substantial
thing, entity* h aving a svabhdva or self-nature and d istinct from everyth ing else.
B u t according to N&g&rjuna neither really ex ists; and in their absence there can
be no bhava (xv. 4: svabhdvaparabhdvabhydm rte bhdvah kuiah punah\ svabhdve
parabhdve vd sati bhdvo hi sid h ya ti). N&garjuna has furthermore stated paradoxically
and perhaps b y oxym oronth a t w hatever exists in dependence (p r a tity a ) is still
(dnta) b y nature* (svabhdvatah, vii. 16); it is clear from th e doctrinal co n tex t th a t
w h a t is so m u st be precisely w ithou t the svabhdva p ostu lated b y his opponents.
Cf. MMK xviii. 10 (below, pp. 38 sq., 43).
6 The words unyatddarfana, a ttested in Candrakirtis Prasannapada (PP)
xviii. 5 and x x iv . 13, and unyatdvada h ave to be distinguished from th e term
unyatddrsti, w hich denotes a speculative view th a t hyp ostatizes em ptiness (see
MMK xiii. 8; cf. also th e unyatdkoti in P P xxiii. 14). The word unyatdvadin
m aintainer o f th e philosophy o f i u n yatd is found e .g . in N agarjunas Vigrahavya-
vartani 69 and Vaidalyaprakarana 1.
7 In an article in RO 10 (1934), p. 21 note 1, T h. S t c h e r b a t s k y defended his use
o f th e term m onism b y tak ing it to mean not a m onistic rea lity (svabh dva), but a
single Erklarungsprinzip th a t excludes an y real plurality, in th e sense o f m onistic
knowledge. B u t th is usage is hardly usual and em ploym en t o f the term m onism
can therefore easily mislead.
A s for scepticism (or Pyrrhonism ) as a valid description o f th e M adhyam aka, the
term w ould have to refer to th e S ceptics epoche or m eth odically m o tiva ted w ith
holding o f judgem ent and to their aporetic (as opposed to dogm atism and speculative
m etaphysics). See e .g . S. S c h a y e r , A usgew ahlte K ap itel aus der Prasannapada
(Krakow, 1931), pp. x x x x x xiii.
Introduction : The Name Madhyamaka 3

been neologistically term ed a zerology, the reference being to a philosophical

rather than to a mathem atical feature.8 Yet, since none of these descriptions
defines fully and adequately this school of thought, it will probably be pre
ferable simply to retain the established term madhyamaka th a t came to be used
by the school itself; for by emphasizing the theoretical and practical middle*
on the gnoseological and soteriological levels it fits very well the main concerns
of the school.
Considering then all dharmas to be as it were ciphers em pty (nya) of
substantial own being (svabhva self-existence, aseity), the Madhyamikas
refrain from taking up any speculative or dogmatic position definable in terms
of dichotomizing conceptual thinking polarized into binary (e.g. positive/
negative, identity /difference) or quaternary (e.g. positive/negative/positive-
cum-negative/neither-positive-nor-negative, or conditioned production/non
conditioned production/both/neither) sets of constructs, and from postulating
any absolutely (paramdrthatas) real bhdva entity in terms of these positions.
Although they did not therefore propound any thesis in support of such views,
the Mdhyamikas nevertheless regarded themselves as a philosophical school
(vda = darana) devoted to upholding and explicating what they considered
to be the teaching of the Buddha. Accordingly, the Madhyamaka can properly
be said to have a philosophical theory (d a r a n a as distinct from a speculative
view or dogmatic opinion (drsti) albeit not one founded on any conceptually
constructed hypostatizing of some kind of entity, be it positive, negative, both,
or neither.

8 The expression zerology has been em ployed b y L. M at/l, U n e approche

possible du Snyavda, Tel Quel 32 (H iver 1968), p. 64 sq.A lthough th e arith
m etical zero and a know ledge o f place-value were no doubt current in In dia in
Ngrj u n as tim e, there is no evidence to show th a t he w as actu ally em ploying th e
term nya in its m ath em atical sense. In M adhyam aka usage nya refers to the
fact th a t dharmas are em p ty o f own being (svabhva-nya); th e id ea is therefore
th e philosophical one o f th e onto-logical null valu e o f all dharm as as so to speak
ciphers w ith ou t independent and ultim ate existence (see below ). n y a is an
ep ith et o f all dharm as; and nyatva designates em ptiness as th e quality charac
terizing all dharm as. n ya t on the other hand is the fact, or truth, o f th e em ptiness
o f all dharm as. Thus th e words nya and nyatva pertain to th e conventional
surface lev el o f sam vrti; w hile nyat is used to indicate or p oin t to the level o f
u ltim ate reality (param arth a) (to the e x ten t th a t this is a t all possible in term s of
language and d iscu rsivity).
The use o f th e w ord nya to designate the arithm etical zero is alm ost certainly
later th a n N grjuna (at whose tim e th e usual term s seem to h a ve been kha and
bindu) ; in an y case, it is clearly later th an th e early uses o f nya in th e canonical
literature o f B uddhism , where th e reference is to som ething em pty. On this question
see D . S e y f o r t R t t e g g , M athem atical and linguistic m odels in Indian th ou gh t: the
case o f nyat, W Z K S 22 (1978), p. 174 sq.
For bibliographies o f the M adhyam aka see . L a m o t t e , L e trait de la Grande
V ertu de Sagesse (five volum es, L ouvain, 1944, 1949, 1970, 1976 and 1980) ; J . M a y ,
Candrakrti: Prasannapad M adhyam akavrtti (Paris, 1969), pp. 2345; R . H.
R o b i n s o n , E arly M dhyam ika in In dia and China (Madison, 1967), pp. 3218; and


According to a tradition of Indian origin reported by several of the Buddhist

historiographers and doxographers of Tibet, Nagarjuna had as his immediate
predecessor and teacher a certain Rhulabhadra, to whom is ascribed a well-
known hym n to the Prajpram it. But according to another (probably more
accurate) tradition known from both Chinese and Tibetan sources, this R
hulabhadra was a follower of Nagarjuna.10 No m aster is therefore known to
have preceded Nagarjuna at the head of the Madhyamaka school in the strict
sense. At all events it is Nagarjuna who in the present state of our knowledge
has to be considered the source or originator of this school; and this is the way
he has in fact been regarded by modern historians, as well as by the Mdhya-
mikas themselves including even those who make Rhulabhadra his master.
Ngrjuna is generally believed to have been born and to have worked in
South-Central India (South Kosala or Vidarbha?) early in the first millennium
P .C .11
F . S t r e n g , E m p tiness (N ashville, 1967), p. 237 sq. (in which publications m ost o f
th e W estern literature on th e subject has been listed). See also K . P o t t e r , E n c y c lo
pedia of Indian philosophies, i (Bibliography). A survey of S. Y am aguchis im portant
publications is given b y J . W . de J o n g , I I J 19 (1977), 99103. B rief accounts o f
m o d em Japanese work on th e M adhyam aka are given b y H . N a k a m u r a , A cta
asiatica 1 (1960), p. 56 sq., and G. N a g a o , A cta asiatica 29 (1975), p. I l l sq. A
general bibliographical su rvey o f studies on M adhyam aka literature w ith special
reference to Japanese publications is given b y H . N a k a m u r a , Journal o f inter-
cultural studies no. 4 (1977), pp. 7794, 12631.
D eta iled bibliographies relating to both editions o f and studies on M adhyam aka
te x ts are given b y Y . E j i m a in th e prefaces t o : T he sD e dge T ibetan Tripi^aka bsTan
hgyur preserved at th e F a cu lty o f Letters, U niversity o f T okyo (Deruge-ban
Chibetto D aizokyo, R onsho-bu, Tokyo D aigaku Bungaku-bu shozo; Sekai seiten
kako kykai, Tokyo, 197779).
10 On R hulabhadra and th e question w hether he ^seeded or follow ed Ngrjuna
see below , p. 54.
11 N grjuna has been variously placed a t the end o f th e first century P . C.
(S . L v i , J A 1936, pp. 96, 103 s q . ; D . S h a c k l e t o n B a i l e y , atapacatka o f
M trceta [Cambridge, 1951], p. 9); in th e second century (M. W i n t e r n i t z , H isto ry
o f Indian literature, ii [Calcutta, 1933], p. 342; TLa m o t t e , Trait de la Grande
V ertu de Sagesse, i [Louvain, 1944], p. x ; T. R . V. M u r t i , Central philosophy o f
B uddhism [London, 1955], p. 87); and in th e third century (M. W a l l e s e r , L ife of
N grjuna, in A sia Major, H irt A nniversary V olum e [1923], p. 423). E . F r a u -
w a l l n e r has placed him c. 200 (Die Philosophie des B uddhism us [1969], p. 170);
b u t . L a m o t t e opted for the chronology adopted b y K um arajivas school w hich
The Early Period : The Formation of the Madhyamaka School 5

N grjunas writings are the first philosophic treatises (stra) known to us

in which an attem pt has been made to give a systematic scholastic exposition
places N grjuna and his disciple A ryadeva in th e eighth century after th e B u dd h as
N irvana, i.e . in th e third century P. 0 . (Lenseignem ent de V im alakrti [Louvain,
1962], pp. 7477; cf. Trait, iii [Louvain, 1970], p. li sq.).
For discussions o f th e synchronism betw een N grjuna and a Kanigka and also a
S tavh an a m onarch, see S. L v i , loc. cit. ; K . V e n k a t a R a m a n a n , N grjunas
p hilosophy (R utland and T okyo, 1966), pp. 2730 (who identifies th e Stavhana
as Gautam putra takarni) ; and the contributions b y A. K . W a r d e r , F . W i l h e l m
e t al. in A. L. B a s h a m (ed.), Papers on the date of Kanigka (1960) (Leiden, 1968),
pp. 32831, 3346, 3423, 4278, 433 (according to W arder, p. 334, th e Stav-
hana in question w ould be P ujum yi I I jVsigt/hputra] ; see in addition A. K .
W a r d e r , Indian B ud d h ism [D elhi, 1970], p. 375, and Indian K v y a literature, ii
[D elhi, 1974], pp. 1834, 21314, 230). Cf. also . L a m o t t e , Sur la form ation du
M ahyna, in A siatica (Festschrift F . W eller, Leipzig, 1954), pp. 3868, 3912, and
Trait, iii, p. ix s q .; P . S. S a s t r i , IH Q 31 (1955), p. 198; below, p. 26 n ote 59. In
K alh an as R jatarafigin i. 173 th e B od h isa ttva N grjuna is associated w ith the
Sadarhadvana (Harwan) in Kmir at the tim e of th e Turugka kings Hugka, Jugka
and Kanigka. E ven in recent studies th e first regnal year o f K anigka is variously
d ated b y different scholars betw een 78 (e.g. G . F u s s m a n n ) and 238 (R. G o b l ) ; tw o
recent writers have op ted for 200 (H. P i Ia e s c j h k e and Th. D a m s t e e g t ). For a
survey o f the problem see Th. D a m s t e e g t , E pigraphical hybrid Sanskrit (Leiden,
1978), p. 10 sq.On the other hand, in B n a s Hargacarita (Chapter viii) N grjuna
is linked w ith a Stavh an a king. A B hadanta N grjuncrya is furthermore
m entioned in an inscription discovered near the Jagga yy ap eta Stpa ; see J . B u r g e s s ,
N o tes on th e A m arvat St pa (Madras, 1882), p. 57, and The B uddhist Stupas o f
A m aravati and J ag ga yy ap eta (Archaeological Survey o f South India, i, L ondon,
1887), pp. 11112 (who d ates th e w riting o f th e inscription to c. 600; bu t G. T u co i,
Minor B u ddhist tex ts, ii [R om e, 1958], p. 284, suggests th e d atin g 450600) ; T. N .
R a m a c h a n d r a n , N grjunakonda 1938 (Memoirs o f th e A rchaeological Survey
o f In d ia 71, D elhi, 1953), pp. 2829 (who thinks th e inscription relates to th e
Tantrik Siddha N grjuna). Concerning N grjunakonda and its inscriptions see
also J . P h. V o g e l , E l 20 (1929); N . D u t t , IH Q 7 (1931), p. 633 sq.; A. H .
L o n g h u r s t , The B ud d h ist antiquities o f N grjunakonda (Memoirs o f the Archaeo
logical Survey o f In d ia 54, 1938).
On the life and legend o f Ngrjuna see, e .g ., S. B e a l , Si-yu-ki: B uddhist re
cords of th e w estern w orld (London, 1906), ii, p. 210; T. W a t t e r s , On Y uan
Chwangs travels in India, ii (London, 1905), p. 200 sq. ; J . T a k a k u s u , Record o f the
B u ddhist religion b y I-tsin g (Oxford, 1896), p. 158 sq. (on *Jan taka/*Jetak a and
Stavhana); M. W a l l e s e r , The life o f Ngrjuna from T ibetan and Chinese
8orces, A sia Major, In trodu ctory V olum e (Hirth Anniversary V olum e, Leipzig,
1923), pp. 42155; P. D e m i v t l l e , Sur un passage du M ahm eghastra, B E F E O
1924, pp. 2278; G. T u c c i, J P A S B 26 (1930), p. 138 sq. ( = Opra minora, i [R om e,
1971], p. 209 sq.); S. K . P a t h a x , Life o f N grjuna, IH Q 30 (1954), pp. 9395;
P . S. S a s t r i , N grjuna and ryadeva, IH Q 31 (1955), pp. 193202; K . V e n k a t a
R a m a n a n , N grjunas philosophy as presented in th e M ah-Prajnpram it-
Sstra (R utland and T okyo 1966), p. 25 sq. ; R. H . R o b i n s o n , E arly M dhyam ika
in India and China (Madison, 1967), p. 21 sq. ; D . S e y f o r t R u e g g , Le D harm adhtu-
sta v a de N grjuna, in tu d es tibtaines ddies la m m oire de Marcelle Lalou
(Paris, 1971), p. 448 sq.; J. F i l l i o z a t , Y ogasataka (Pondicherry, 1979), pp. iv x ix
(who proposes, p. x v iii, explaining I-chings S h i-yen -te-ka/*Jan tak a as *S(m)-
taka(ni), i.e . Stakani/Stakarni).
On th e relation betw een th e M adhyam aka and the great philosophical Stras o f
6 David Seyfort Ruegg The Literature of the Madhyamaka School

of the theory of emptiness (nyatd) and non-substantiality ( nihsvabhdvata)

not only of the self (dtman) or individual (pudgala) but also of all factors of
existence (dharma), one of the most fundamental ideas of the Mahynastras.
In Buddhist tradition Ngrjuna is linked especially closely with the Prajn-
pram itstras, the M ahynist scriptures th a t devote much space to this
theory; and he is indeed credited with having rescued parts of them from
Ngrjuna has accordingly been regarded by the Buddhist traditions, and
also by most modern scholars, as one of the very first and most original think
ers of the M ahyna.13 A verse of the M lamadhyamakakriks (MMK xiii. 8)
appears clearly to presuppose a section of the R atnakta collection, the
Kyapaparivarta ; and N grjunas doctrine based on the analysis of dichot-

th e B rahm anical tradition, see th e older discussion b y H . J a c o b i , JA O S 31 (1911),

pp. 129. Concerning th e relation betw een N grjuna (and th e early Mdhyamikas)
and other schools o f Indian philosophy see more recently, for th e Sm khya, W .
L i e b e n t h a l , Satkrya in der D arstellung seiner buddhistischen Gegner (Stuttgart,
1934) ; E . H . J o h n s t o n , E arly Sm khya (London, 1937), pp. 6667 ; N . A i y a s w a m i
S a s t r i , Sino-Indian studies 4 (1951), pp. 4750; and E . F r a u w a b b n e r , W ZKSO 2
(1958), p. 131. For th e N y y a see W . B u b e n , D ie N yyas tras (Leipzig, 1928);
S . Y a m a g u c h i , J A 1929, ii, p. 62 sq.; G. T u c c i, Pre-D innga B u d d h ist te x ts on
logic (Baroda, 1929), p. x v sq. ; E . H . J o h n s t o n and A. K u n s t in th e introduction
to their edition o f th e Y igrahavyvartan, MCB 9 (1951), p. 106; G. O b e r h a m m e r ,
W ZKSO 7 (1963), p. 64 s q .; B . K . M a t e l a e , E p istem ology, grammar and logic in
Indian philosophical analysis (The H agu e, 1971); K . B h a t t a c h a r y a , Journal o f
Indo-E uropean studies 5 (177), p. 265 sq., and The dialectical m ethod o f N grjuna
(Delhi, 1978). For th e Vaieika see G. T u c c i, op. cit., pp. x x iv x x v ii; G. O b e r
h a m m e r , loc. cit., p. 70. The relation betw een N grjuna on th e one hand and on
th e other th e Carakasamhit, th e Vaieikastras and th e N yyastras and Bh^ya
h as been touched on b y A. K . W a r d e r in A. L. B a s h a m (ed.), Papers on the d ate
o f K anika (Leiden, 1968), p. 3301. W ith reference to the T a-chih-tu-lun see
E . L a m o t t e , T rait, iii, p. x x v x x v ii; K . V e n k a t a B a m a n a n , op. cit., passim .
See also P . S. S h a s t r i , IH Q 31 (1955), p. 199201; R . H . R o b i n s o n , E arly
M dhyam ika, p. 68 sq.On parallels betw een M dhyam ika th ou gh t and Bhartrhari
(fifth century), see H . N a k a m u r a , Journ. Ganganath J h a K endriya Sanskrit
V idyapeetha 29 (1973), p. 367 sq.
12 See for exam ple Candrakirti, P P i, p. 23. Cf. R . H . R o b i n s o n , E arly M dhya
mika, pp. 6163, 17780.Contra: A. K . W a r d e r , Indian B uddhism , p. 3889.
12 R ecen tly it has been suggested th a t N grjuna was n ot a M ahynist in the
strict sense since his MMK and th e other m ost closely related works nowhere
m ention th e M ahyna and m ake no exp licit reference to M ahynist te x ts; see
A. K . W a r d e r in M. S p r u n g (ed.), The problem o f th e tw o truths in B uddhism and
V edanta (Dordrecht, 1973), p. 78 sq.I t is indeed true th a t no M ahynist te x t has
been m entioned exp licitly in th e MMKth e only te x t nam ed therein (xv. 7) being
th e K tyy an va v d a, a te x t o f the Sam yuktgam a (cf. S am yu ttanik ya, ii, p. 17)
th a t deals w ith th e tw o extrem e view s o f eternalism and annihilationism and th a t
th e MMK allude, in order to criticize th em , chiefly to Abhidharmika doctrines.
H ow ever, given th e im plicit reference to the R atn ak ta (see below) as w ell as th e
elaboration o f doctrines characteristic of th e M ahyna in th e MMK and the
related works, this view seem s extrem ely difficult to m aintain. The R a tn va l as
w ell as several other works ascribed to N grjuna refer at length to th e M ahyna.
The Early Period : The Formation of the Madhyamaka School 7

omously opposed pairs of concepts is characteristic of this work also. In sum,

in view of his place in the history of Buddhist thought and because of his
development of the theory of the non-substantiality and emptiness of all
dharmas, it seems only natural to regard Ngrjuna as one of the first and most
im portant systematizers of Mahynist thought.
Ngrjuna a t the same time explored philosophical topics known to the old
Canon as well as to the schools of the rvakayna.14 In particular, he was well
acquainted with doctrines taught by the Sarvstivdins.15 And even the
doctrine of dharma-nairtmya and emptiness which Ngrjuna contributed so
much to explicating was certainly not unknown to the canonical texts and
treatises of the rvakayna.16
The M ahynist canonical sources of the Madhyamaka school are chiefly (if
not exclusively) the Prajnpram it, R atnakta and Avatam saka literature.
And the Ta-chih-tu-lun (*Mahprajnpramitopadea), which is ascribed to
(a) N grjuna and is formally a commentary on the Pancavimatishasrik
Prajnpram it, cites among other scriptures the Vimalakrtinirdea, the
ramgamasamdhi, the Saddharmapundarlka, the Daabhmika, the Aksa-
yamatinirdea, the Tathgataguhyaka, and the Kyapaparivarta. Most of
these Stras continued to be regularly quoted by the later Madhyamikas. In
two works by Candrakrti (seventh century) for example, the Prasannapad
and the M adhyamakvatra, among Stras cited in addition to the P rajn
pram it (Astashasrik, Dvyardhaatik and Vajracchadik) we find the
Aksayamatinirdea, Anavataptahradpasam kram ana, Uplipariprcch, K-
yapaparivarta, Gaganaganja, Tathgataguhya, Daabhmika, Drdhdhyaya,
Dharanvararja, Pitputrasam gam a, Manjurpariprcch, R atnakta, Rat-
nacdapariprcch, Ratnamegha, R atnkara, Lankvatra, Lalitavistara,
Vimalakrtinirdea, listamba, Satyadvayvatra, Saddharmapundarlka,
Sam dhirja (Candrapradpa), and Hastikaksya.
14 Se L. d e L a V a l l e P o u s s i n , MCB 2 (193233), pp. 1012; T. R . V . M u b t i ,
Central philosophy o f Buddhism , p. 50 sq. ; K . V e n k a t a R a m a n a n , N grju nas
philosophy, p. 46 sq. Cf. also A. B a r e a u , L absolu en philosophie bouddhique:
v olu tio n de la notion d asam skrta (Paris, 1951), pp. 17486.On proto-M dhya-
m ik a elem ents in th e Pali canon see L. G m e z , P E W 26 (1976), p. 137 sq.
16 The relation betw een the author of th e Ta-chih-tu-lun (*Mahprajnpramito-
padesa, attributed to N grjuna) and the Sarvstivda has been studied in som e
detail b y K . V e n k a t a R a m a n a n , N grjunas philosophy, pp. 28 sq. and 57 sq.,
and b y . L a m o i t e , Trait, iii, p. x iv sq. B u t on th e question of th e authorship o f
th is treatise see below, p. 32.
16 T he M adhyam ikas do n ot th em selves claim th a t th e dharm anairtm ya was
their ow n discovery and th a t it w as unknow n to th e rvakas. See for exam ple
Candrakrti, Prasannapad on MMK xviii. 5 and M adhyam akvatra i. 8 (below,
p. 74).For th e canonical form ula sabbe dhamm anaU see M ajjhim anikya i,
p. 2278; Sam yu ttan ik ya iii, p. 1324, and iv, p. 401 ; A nguttaranikya i, p. 286.
Compare Vasubandhu, A bhidharm akoabhya vi. 14d. This formula is one o f th e
characteristics or seals o f th e D harm a (together w ith sabbe sankhr aniccy and
in A guttaranikya i, p. 286 sabbe sankhr dukkhd). See also B od h isattvab h m i,
Chapter x v ii (p. 2767).
8 David Seyfort Ruegg The Literature of the Madhyamaka School

Any attem pt to trace the history of the Madhyamaka school is beset by the
numerous problems th a t result from the obscurity in which the lives, work and
chronology of so m any of its masters remain shrouded. Only if this obscurity
can be dispelled by the extensive historical, philological and philosophical
study th a t is still required will it eventually become possible fully to elucidate
these problems. Of primary importance is the still embroiled question of the
figure of Nagarjuna as a person and the author of certain works, and of the
relationship between the various compositions ascribed to the m asters who
have borne this great nam e; for it is hardly to be doubted th a t there lived in
India more th an one person having this name and belonging to different
periods in the history of the Madhyamaka.17
The Madhyamaka6astrastuti attributed to Candrakirti has referred to eight
works by N agarjuna: the (Madhyamaka)karikas, the Yuktisastika, the Sun-
yatasaptati, the Vigrahavyavartani, the Vidala (i.e. Vaidalyasutra/Vaidalya-
prakarana), the Ratnavali, the Sutrasamuccaya, and Sam stutis.18 This list
covers not only much less than the grand total of works ascribed to Nagarjuna
in the Chinese and Tibetan collections, but it does not even include all such
works th a t Candrakirti has himself cited in his writings. I t is of particular
interest also to note th a t Sam stutis or hymns are included in this list since
some at least of the hymns ascribed to Nagarjuna pose difficult problems of
authorship owing to the fact th a t their doctrines sometimes appear to differ
appreciably from those of the theoretical scholastic treatises; unfortunately,
however, this list does not give the titles of the hymns its author had in mind,
but quotations from a t least some hymns ascribed to Nagarjuna are to be found
in Candrakirtis works (see below, p. 31).
The first six works enum erated above correspond to the six well-known
theoretical scholastic treatises of Nagarjunathe so-called Yukti-corpiis (rigs
hogs) of the Tibetan tradition (except th a t for the R atnavali a certain non-
extant *Vyavaharasiddhi is often substituted). Beside them the Tibetan
sources place a collection of hymnsthe Stava-corpus (bstod hogs)as well
as a body of tracts and epistlesthe Parikatha-corpus (gtam hogs) which in
cludes the Suhrllekha, sometimes the R ajaparikatha-R atnavall, and some
minor works.
In view of the above-mentioned opacity and confusion in the records as well
as the uncertainty concerning the authorship of several works ascribed to
Nagarjuna, it will be convenient for the historian of the Madhyamaka to take
as his point of departure the treatise universally considered as the Madhyamaka-

17 I t is im portant to observe th a t even if th e T ibetan historiographers and

doxographers have ascribed a variety o f doctrinally d istin ct works to N agarjuna,
th ey generally differentiate clearly betw een different phases in his doctrines,
assigning these works to d istin ct periods in a m iraculously long life. See D . S e y f o r t
R u e g g , Le D harm adhtustava, in tu des tibtaines ddies a la mm oire de M.
Lalou (Paris, 1971), pp. 44853.
18 E d ited b y J . W . de J o n g , OE 9 (1962), p. 49 sq.
The Early Period : The Formation of the Madhyamaka School 9

stra par excellencenam ely the MMKtogether with any other texts ascrib-
able to the same author th a t are doctrinally related, and to regard this textual
corpus as a standard of reference when describing Ngrj unas philosophy. In
this way a relatively homogeneous body of literature can be used as the basis
for the discussion of the earliest period of Madhyamaka thought and also as
one criterion for determining what other texts might be attributed to Ngr-
juna, the originator of the Madhyamaka school.

The Mla-Madhyamakakriks
The extant Sanskrit tex t of the MMK is embedded in the Prasannapad
M adhyamakavrttifi, a commentary w ritten by Candrakrti perhaps half a
millennium after the time of Ngrjuna. This work is now available to us in
three manuscriptsall mediocre copies of an original which was apparently
not faultlessused by L. de La Valle Poussin for his edition of 190313 and
in a fourth m anuscript used by J. W. de Jong for his edition of 1977. This text
of the MMK has to be compared with the tex t contained in other commentaries
on the MMK, which are now accessible only in Chinese and Tibetan, and in
quotations in the exegetical literature of the M ahyna.19
Our MMK number 447 or 449 verses according to whether the two intro
ductory stanzas to the Awakened One, the best of philosophers who taught
origination in dependence (prattyasamutpdda) free of eight limiting qualities,
are included or not in the count. These verses are divided into 27 chapters.
The first chapter of the MMK takes up the topic of the dependent origination
of things in the process of conditioning.20 Things (bhava entity), Ngrjuna
states, are never found anywhere produced from themselves, from others, from
both themselves and others, or from no cause a t all (1). Now, in principle, the
production of things would have to take place in dependence (pralitya) on
conditions (pratyaya); and Ngrjuna enumerates the four kinds of condition
known to Abhidharma philosophy, namely the causal condition (hetu)> the
objective one (lambana), the immediately preceding one ([samjanantara),
and the governing one (adhipati) (2). However, it is not to be supposed th a t
in these conditions a specific and independent nature (svabhdva own being,
aseity) of the thing is to be found; and in the absence of such own being the
correlative and complementary category of other being (parabhdva alterity)
cannot exist either (3).21 Moreover, mention has been made of conditions on

19 For a brief com parison o f the Sanskrit version o f th e MMK w ith th e Chinese
see R . H . R o b i n s o n , E arly M adhyam ika, p. 3031. On th e title M u lam adhyam aka-
karika see above, n ote 3.
20 In th e follow ing resum6, the chapter titles o f th e MMK as found in Candraklrtis
P rasannapada have been t ken as a basis. In th e T ibetan translations o f B uddha-
p alita s and B h avavi vek a s com m entaries on the MMK, th e chapter titles differ in
som e places.
21 See also MMK, Chapter xv.
10 David Seyforfc Ruegg The Literature of the Madhyamaka School

the ground th a t we speak of production in dependence on th em ; but as long

as no entity is actually produced, in the last analysis one must allow th a t no
conditioning entities exist either (5). In fact, for either a real or a non-real
thing no condition is possible; th a t is, an existent thing (sal) needs no ad
ditional condition to make it what it already is by definition, whereas no con
dition could effectively relate to a non-existent thing (asal) (6). I f no factor
(dharma) is produced as real, non-real or a combination of both, there can
then be no cause producing it (nirvartako hetuh) (7); and the same holds,
mutalis mutandis, for the other three kinds of pralyaya (810). Finally, when
a product is not present in conditions taken either singly or all together, what
is not present in conditions could not be produced from them (11). Now, were
a non-existent (asal) product to proceed from conditions, why, being non
existent, should it not proceed equally from non-conditions (12) ? A product
is by definition a result of its conditioning factors, whereas conditions are not
made up of themselves; but a product supposed to be made up of such (in
fact non-substantial) conditions could not in fact really be their product (13).
Here N agarjunas critique is intended to reveal the antinomies and paradoxes
in a notion of causation th a t presupposes the existence of a number of sub
stantial and self-existent entities; and it thus covers both the Abhidharmika
concept of origination in dependence (pralityasamutpada) of dharmas as factors
each possessing a specific characteristic property, and also doctrines such as
those of the Samkhya and VauSesika. And the conclusion then is th a t there
exists no substantive product issuing from a real causal nexus with substantive
conditions (or, indeed, non-conditions, 14), each conceived of as an entity
(bhdva) possessing a positive (or even a negative) own being (svabhdva).
N agarjunas analysis thus serves to relativize and deconstruct our artificially
posited entities with their respective conditions, which are thus annulled
(zeroed) both as substantial entities and ulimately valid philosophical cate
Chapter ii is devoted to a critique and deconstruction of the category of
movement over space in time, as situated in either the two aspects of traversed/
past (gala = son ba) and im traversed/future (agala = ma son ba: andgala)
space-in-time (: adhvajata) or in a third aspect apart from (and between)
these two which might be supposed to be in process of being moved over
(gamyamdna = bgom pa) (I).22 Now going (gali) cannot in fact attach to the

22 For a discussion o f th e philosophical problem taken up in Chapter ii o f th e

MMK, see M. S i d e r i t s and J . D . O B r i e n , Zeno and N agarjuna on m otion,
P E W 26 (1976), pp. 28199, who propose a m ath em atical interpretation o f
N agarjunas analyses in addition to th e conceptual interpretation. T he authors
draw a parallel w ith Zenos paradoxes on th e grounds th a t N agarjuna w as arguing
again st a concept o f m otion th a t postulated spatial and tem poral m inim s (a to m s
and in sta n ts) in th e frame o f a theory th a t reified the term s analysed and assum ed
a language-reality isom orphism and a correspondence theory o f truth.
The com m entary on this chapter o f th e MMK in CoiVkha'pas Rigs*pai*rgya*mcho
has been translated into E nglish b y J . H o p k i n s and N g a w a n g L e n g d e n , Chapter
The Early Period: The Formation of the Madhyamaka Sohool 11

first two aspects because there is no movement (cesta) in them (2). Nor can
it be superadded, as an additional action of going, to the third aspect mentioned
above because movement in process devoid of movement ( vigamana) to begin
with does n o t apply (3). Also, when going (gamarui) is postulated for this third
aspect there would ensue (prasajyate) the existence of movement in process
without an agent (gantr goer) (4). Since in the absence of a goer there is in
fact no going, the occurrence (prasanga) of two gamanas would entail the
occurrence of two goers for each movement, something th a t does not apply (6).
I t is moreover pointed out th a t there is neither going of an independent goer
nor going of a non-goer (agantr); but there exists no third possibility (trtiya,
tertium non datur) (8; cf. 15). A proposition (paksa) asserting: the goer goes
and involving gamana and gantr as two distinct entities would entail (: prasaj-
yaie) for him who maintains it a goer without movement (10), something th a t
has been seen not to apply (9). Since no undertaking (drambha) of going can
be found located in any of the three spatio-temporally defined phases of move
ment, the latter are nothing but the product of dichotomizing conceptualiza
tion (vikalpa, 14). Thus, through conjoined linguistic and conceptual analysis,
the category of movement is relativized and annulled (zeroed') with respect
to its conceptually interrelated and imaginarily constructed componentsviz.
the action of going (gatiy gamana) and its agent (gantr) and also with respect
to its opposed correlate (pratipaksa) of stationariness and one who stands
(i.e. desists from movement) (1517). In addition, it is shown th a t (the cate
gories of) the action and agent of going cannot ontologically stand in a relation
of either oneness (ekibhava) or difference (ridndbhdva) (1821); they are
merely the product of dichotomizing conceptualization (20). Agent and action
of movement cannot then be hypostatized as independent entities having the
ontological status of real, unreal or both-real-and-unreal (24). I t is therefore
concluded th a t the posited factors making up the triad of agent, action and
object ( karman, i.e. gantavya) of movement do not exist per se (25cd; cf.
Prasannapada ii. 6). I t is to be noted th a t Nagarjunas analysis here lends
itself to interpretation in term s of grammatical categories; and Candrakirti
in fact introduces the concept of sadhana or kdraka as a akti in his comment
(PP ii. 6).
In its m ethod of argument founded on dialectical cancellation and analytical
deconstruction of the concepts and categories of discursive thinking and
ordinary language through prasanga-type reasoning, Chapter ii of the MMK
serves as a model for the discussion of other concepts and categories in the
sequel. Since such discussion is essentially analytical and seeks to deconstruct
dogmatically posited conceptual entities issuing from dichotomizing construc
tion, N agarjunas procedure can hardly be regarded as mere sophistical argu

tw o o f th e Ocean of R easoning (Dharamsala, 1974); an English translation of MMK

ii w ith th e Prasannapada has also been m ade b y th em : A nalysis o f going and
com ing (Dharamsala, 1976).
12 David Seyfort Ruegg The Literature of the Madhyamaka School

ment employing deceptive devices or fallacious means to construct some,

Chapter iii subjects to a critique the category of the sense faculties (indriya)
viz. the five internal sensory bases of the visual, auditory, olfactory, gusta
tory, and tactile, and the mental base23founded on the idea of a triad of factors
comprising a sensory function (e.g. sight), an agent (e.g. the seer) and an object
(e.g. a visible thing). All are shown by Nagarjuna to be interdependent con
ceptual factors, not substantial and independent entities having own being
(svabhava) as was supposed by other schools of thought.
Subsequent chapters are devoted to similar analyses of other categories (the
succeeding one, according to Candrakirti, often being introduced by an oppo
nent in the belief th at it presupposes and therefore establishes the reality of
the preceding category th at had just been criticized). In every case Nagarjuna
demonstrates th a t each new category is itself an unreal construct having only
a limited conventional and transactional validity on the surface level without,
however, constituting an entity having own being.
The following concepts/categories are thus analysed and the antinomies
involved in their postulation are revealed, (iv) The five groups (skandha, i.e.
the physical and mental factors, which entail the existence of the indriyas
because, if the skandhas in which they are included exist, the indriyas must
exist). In this chapter unyatd is referred to together with some fundamental
principles of Madhyamaka thought in connexion with problems raised in
philosophical argument and explication. Thus: 4When a disputation is con
ducted by means of unyatd, should [an opponent] state a reply (parihdra)
nothing will serve him as a reply: there emerges a sameness with w hat is to be
established (samam sadhyena); when an explication is made by means of
6unyatd, should someone [a quasi disciple (&isyadeAiya) according to Candra
kirti] state an objection (updlambha) nothing will serve him as an objection:
there emerges a sameness with what is to be established (89). T hat is, the
substantial reality of vedand or whatever other thing is adduced as a counter-
instance in a reply or objection intended to establish the reality of e.g. rupa
the sadhyainevitably itself falls within the scope of unyatd; and it cannot
therefore be used to invalidate the emptiness of own being of any dharma
(I'upa, etc.).24(v) The elements (dhdtu)earth, water, fire, air, dka4ay and
vijndriaaxe next introduced (on the supposition th at since they at least have
not been negated they would have the ontological status necessary to establish
the preceding category by analogy). The discussion excludes the idea th a t these
elements could really be characterized things (laksya 'm arked, i. e. bhavas) in
relation to their (supposed) characteristics (laksana 'm ark or specific prop

23 In B u dd h ap alitas and B h a v a v iv ek a s com m entaries and th e *A kutobhaya,

Chapter iii is entitled dyatana.
24 Cf. A ryadevas Catuhataka viii. 16: bhdvasyaikasya yo drastd drastd sarvasya sa
sm rtahj ekasya unyatd ya iva saiva sarvasya unyatd//On sddhyasam a cf. below,
n ote 49.
The Early Period: The Formation of the Madhyamaka School 13

erty).(vi) Passion and the subject of passion (rdga-rakta) (presupposing the

existence of the skandhas, dyatanas and dhatus as the basis of such samklesa).
The discussion involves in particular an examination of the applicability here
of the relational concepts of oneness (ekatva) and difference (prthaklva)the
two conditions under which any real entity might be supposed to exisf^which
constitute fundamental modes of analysis employed by the Madhyamika to
show th a t an entity having own being can exist under neither of these condi
tions.25(vii) Production, insofar as it is conditioned (samskrta), has three
phases, namely origination, duration with change, and decay (i.e. the three
characteristics of the conditionedsamskrtalaksanawhich might be thought
to belong to, and hence entail the existence of, the skandhas, ayatanas and
dhdtus).2* B ut when applied to a conditioned entity, precisely, these three
laksanas are shown to involve the fault of infinite regress [anavasthd, 3).27 If
on the other hand production were unconditioned, it could present no samskrta-
laksana. Here are also discussed problems arising from the notions of the de-
pendently produced [utpadyamanam pratitya, 15) and the production of an
already existing entity, a non-existing one, and one th a t is both (20), as well
as from the notions of duration (22) and destruction (26). In sum, because
production, duration and decay are not themselves to be established as real,
the conditioned (samskrta) cannot exist. And being complementary to the
samskrta how then could an unconditioned (asamskrta) thing be establish
ed'? (33) The three samskiialaksanas are thus like illusions (34).(viii) Action
and its agent (karma-kdraka) (the causes' entailing the existence of vijhana
and the other conditioned factors).(ix) The pre-existent (purva) (i.e. the
appropriating subject, or upaddtr, supposed to be the precondition for per
ceptions and sensations, and assumed under the name of pudgala by the
Sammatiyas amongst the Buddhists).28(x) Fire and fuel, the relation between
25 See below , p. 38 sq.
28 The three sam skrtalaksanasutpauda origination, sth iti duration (or sth ity-
anyathdtva alteration in duration : Candrakirti, P rasannapada vii. 11) and vya ya =
bhahga d ecay are referred to in the title o f Chapter v ii as found in B u dd h ap alitas
and B h a v a v iv ek a s com m entaries, as w ell as in th e *A kutobhaya. For the three
(or four) laksarias o f th e conditioned, see V asubandhus discussion in his A bhi-
dharmakoabha?ya ii. 45.
27 For an exam ple o f th is anavasthd regressus in infinitum , see MMK vii. 1819
on th e production o f production o f . . . production. In other words, if origination
as an en tity is assum ed, th en the origination o f th e en tity origination has also to
be assum ed, and so on ad infinitum . See also N agarjunas V igrahavyavartani 32,
where th e problem o f infinite regress is show n to arise in connexion w ith the
establishm ent o f an object o f correct knowledge (pram eya) b y m eans o f som ething
else known as the m eans o f correct knowledge (p r a m d n a ), which then itself requires
to be established b y still another pram dna, and so on infinitely. The aparyavasdn a-
dosa fault o f infinite regress is noticed b y Candrakirti (Prasannapada vii. 3), as
well as b y V asubandhu (Abhidharmakoabhaya ii. 4546) w ho, following the
Sautrantika view , poin ts out th e problem posed b y th e notion o f jd tija ti, etc.
28 The u paddtr and u padan a are m entioned in th e title of Chapter ix in B u d d h a
p a litas and B h a v a v iv ek a s com m entaries, and in th e *A kutobhaya.
14 David Seyfort Ruegg Tho Literature of the Madhyamaka School

which is analysed in term s of th at between an appropriator (updddtr) and the

appropriated (upddana). This relation raises the question of oneness or dif
ference and of the five forms of relation between fire as appropriator and the
appropriated fuel.29
The next chapters take up the Buddhists own soteriological and gnoseo
logical categories of samsara and nirvana and their putative experiencing
subjects. Thus we have critiques of the following topics, (xi) T he prior and
posterior limits (koti) of samsara30 (the existence of which might b e thought
to entail the existence in it of some transmigrating experiencer [sa msartr] or
appropriating self [dtman]).(xii) Suffering (duhkha, the existence of which
might be thought to entail the existence of an dtman or pudgala to experience
it). I t is shown how an experiencing entity cannot serve as a base for the
experience of suffering.These points are reinforced by a critique of (xiii) the
conditioned (samskdra = samskrta), defined as false and delusive (l).31 Now
an opponent m ight hold th at bhdvas are not without own being because of
their nyatd (3). B ut: Were there some thing not empty (aunya) there might
be some thing em pty; since there is nothing th at is not empty how could there
be some thing em pty? (7, a reasoning founded on the interrelatedness of all
dichotomously opposed but still complementary categories and concepts which
serves to do away with the idea th a t there might exist some thing qualifiable
as em pty).32 Snyaid is in fact release from (or: the expeller of, nihsarana)
all speculative views (drsti); and it is not to be resorted to as a surrogate third
dogma or speculative position replacing the binary set of extremes th a t have
been rejected (8).(xiv) The conceptual category of contact (samsarga, e.g.
between the object seen, the seer and sight, contact between which might be
thought to entail the existence of these factors as entities having own being).
This discussion raises the question of the ontological and logical status of
otherness (anyatva) in the frame of the principle of origination in depend
ence.(xv) Own being (svabhdva, aseity, which would be entailed by the
causality of hetu-pratyayas). This notion is subjected to a critique showing
th a t it is not compatible with the idea of production by causes and conditions
because by definition svabhdva should be independent of any other thing
causing or conditioning it; it is therefore not produced.33 Nor can there be
'other being (parabhdva) since this would involve an own being in relation to
which it is to be established. This being so, how could there be any kind of
entity (bhdva) ? And in the absence of bhdva there can be no non-entity
29 x. 14. This investigation (vicdra) concerns five points, on w hich see below , p. 40.
30 S am sdra is th e title o f Chapter xi in B u d d hap litas and B h a v a v iv e k a s
com m entaries, and in the *A kutobhaya.
81 T attva is th e title o f Chapter xiii in B u d d hap alitas and B h v a v iv ek a s co m
m entaries, and in th e *A kutobhaya.
32 Cf. A ryadeva, Catuhataka x v i. 7.This does n ot seem to be th e case o f the
logical error o f negation o f th e antecedent since w e do n ot in fact h ave a hyp oth etical
syllogism here.
33 On MMK x v . 12 cf. Candraklrti, M adhyam aka vatra vi. 181183a.
The Early Period: The Formation of the Madhyamaka School 15

(abhdva) either.34(xvi) The twin categories of bondage and release (bandhana-

moksa, the existence of bondageor samsaraentailing the existence of some
svabhdva in things and of a transmigrating pudgala, as well as the idea th a t the
correlative category of releaseor nirvanais an opposed entity in a binary
set).(xvii) Action and its result (karma-phalaywhich would entail the existence
of an entity termed isamsdra>).(xviii) The category of a self (dtman) as the
appropriating subject to which actions, their fruits and the associated defile
ments attach as things belonging to it (dtmiya); whereas once appropriation
(upadana) has ceased birth, etc. also come to a stop (4).35(xix) Time present,
past and future (the existence of which might be thought to presuppose the
own being of things th at, in their various states of temporal being, make time
knowable since it is their support).(xx) The complex aggregate (sdmagri) of
causes and conditions (which would together bring about, in a subsequent
moment of time, th e production of a result whereby the existence of time as a
cooperating causesahakdnkdranamight be supposed to be established). The
discussion here raises once more the question of the two conditionsoneness
(ekatva) and difference (anyatva)m which a cause might be supposed to stand
in relation to its putative effect if they are conceived of as entities (19).38
(xxi) Production and destruction (sambhava-vibhava, which might also be
thought to presuppose time as the continuum in which they would successively
take place). As entities they are shown to be in fact inapplicable categories.
(xxii) The tathdgaia (in virtue of whose continuing existence through aeons a
continuity of existencebhavasamtatiin time might be entailed). The dis
cussion includes an investigation of five modes of relation between a tathdgata
and his appropriated skandhas.*1(xxiii) Error (viparydsa) and defilement
(klea, viz. passion, hate and confusion) (which might entail a continuity of
existence in time by being its ultim ate cause). Special consideration is here
given to the kleas (passion and hate proceed from erroneously grasping
something as agreeable or disagreeable, and confusion proceeds from er
roneously taking something to be what it is no te.g. the impermanent as
permanent). In the absence of any dtman it then appears th a t the kleas are
themselves not things which could be regarded as real entities to be eliminat
ed (3). And just as there is no entity called error and hence nothing real to be
eliminated, there is no correlative non-error to be grasped (16), the binary
category of error and non-error as well as ignorance (avidyd) and the sam-
skdras (23) being thus annulled. Indeed, there could be no real elimination of
defilements th a t are either existent (bhta real, 24) or non-existent (abhta, 25).

34 Bhdvabhava is th e title o f Chapter x v in B u dd h ap alitas and B h a va viv ek a s

com m entaries, and in th e *A kutobhaya.
35 In th is chapter th e relation o f an dtm an to th e skandhas is discussed in term s
o f th e investigation (vica ra ) in five points. See below , p. 40.
36 H etuphala is th e title o f Chapter x x in B u d d hap alitas and B h a v a v iv ek a s
com m entaries, and in th e *Akutobhaya.
37 See below , p. 40.
16 David Seyfort Ruegg The Literature of the Madhyamaka School

When an investigation of the above-mentioned categories has revealed the

paradoxes and antinomies th a t attach to their (postulated) existence, and once
they have been relativized and zeroedboth as entities through the principle
of origination in dependence (pratityasamutpada) as understood by the Madhya-
mika, and as logical categories through a process of analysis and deconstruc
tion th a t shows them to have as their foundation nothing but pairs of dichot-
omously opposed complementary conceptsand have been seen to be em pty
of own being, such basic principles of Buddhism as the four noble truths
(aryasatya)namely suffering which is to be known, its origination which is
to be eliminated, its cessation which is to be realized, and the path to cessation
which is to be cultivatedare themselves relativized; for it becomes plain th a t
they belong n ot to ultim ate reality (paramartha) but, semantically, to the
surface level of dichotomizing conceptualization and, pragmatically, to trans
actional usage. Moreover, in the absence of the four noble truths the four
fruits of sainthood would also not exist as real entities, so th a t there could be
no saints (drya), and hence no community (samgha), no dharma and no buddha
existing as real entities. The question then arises: Have these basic principles of
Buddhist thought been rejected or abolished by Nagarjuna (either by in
advertence and a failure to foresee the power of the critical analysis he was
employing, or intentionally out of a bent towards nihilism) ? In his considered
and acute reply in Chapter xxiv, Nagarjuna sets out several of the key prin
ciples of the Madhyamaka. To begin with, it is observed th a t an opponent who
raises the above question merely reveals his own ignorance of the sense of
emptiness (unyata, 7). One m ust first distinguish between the two levels to
which the B uddhas teachings relate: the worldly surface tru th (lokasamvrti-
satya) and tru th as ultim ate reality (satyam ... paramdrthatah, 8); for the
person who cannot distinguish between them simply does not understand the
deep reality (tattva) of the teaching (9). Now, there can be no communication
of ultim ate reality (paramartha) without resorting to transactional usage
(vyavahdra); and without comprehending the paramartha, nirvana is not to
be attained (10).38 Emptiness may then be likened to a dangerous serpent
because it destroys foolish people who do not grasp it correctly; and this
indeed accounts for the Teachers initial hesitation to teach (1112). Nagarjuna
furthermore observes th a t the criticism levelled against iunyata by the op
ponent does not fault it because this criticism has no applicability to the
em pty: the opponent is in fact only attributing his own errors to another
(13,15). In the M adhyamaka the word unyata is used as a term for origination
in dependence (pratityasamutpada), and it is a conditional (pragmatic) designa
tion (prajhaptir upadaya); this is the middle way (pratipat madhyama, which

38 anagam ya is rendered here as w ithout com prehending follow ing Candrakirti,

w ho glosses anadhigam ya, and the T ibetan translation. The word is som etim es also
understood as m eaning w ithout h aving recourse t o , as in the T ibetan translation
(m a brten p a s) o f th is verse as quoted in th e M adhyam akavatara vi. 80. A nd in
MMK x x vi. 3 agam y a certainly means in dependence on (brten n a s ),
The Early Period: The Formation of the Madhyamaka School 17

falls neither into annihilationism by denying what originates in dependence,

nor into eternalism by hypostatizing as real what are constructs and designa
tions originating in dependence and, consequently, empty of own being) (18).39
Indeed, only when unyata is found to be applicable does everything become
applicable (fitting) (14) ;40 were everything on the contrary non-empty nothing
could either arise or cease, and it is this (rather than the theory of unyata)
th a t has as its undesired consequence the non-existence of the four noble
truths (20). For how could there be suffering if a thing were independently
produced, since the impermanent which by definition is painful would not be
present when something has own being in virtue of its own independent exist
ence (21) ? W ith respect to the second noble truth, to deny unyata involves
one in a position where origination could never take place because anything
existing by own being will clearly not require to be produced in order to
exist (22).41 Nor could there be cessation (nirodha) of suffering existing by
own being, so th at it is just by holding to own being th at one rejects the th ird
noble tru th (23). Finally, were the path to possess own being its cultivation
would not be possible (24). And if the first three noble truths did not exist,
what kind of path is it th a t would have the function of leading to attainm ent
in virtue of the fact of the cessation of suffering (25) ? Moreover, even actions
can appertain only to the em ptyi.e. to what is without own beingbecause
nothing is to be effected for any thing th a t is not em pty; own being is not
effected (33). To reject sunyata residing in the principle of origination in inter
dependence42 is therefore to reject all worldly transactional usages (vyavahdra,

39 The pronoun sd in M K K x x iv . 18 c m ight refer either to pratltyasam utpada in

th e first quarter, or to unyata in th e second quarter; and Candrakirti gives an
explan ation th a t in fact takes into account both these possibilities.B h av av ivek a
(za, fol. 287 b 6) speaks o f unyata called pratltyasam utpada w hich is u pdddya
pra jn a ptih and also (fol. 288b 3) m adhyam d pratipat. In B u ddhapalitas com m entary
(tsa, fol. 306b) the pronoun w ould seem to refer only to pratltyasam u tpada (but the
Sanskrit construction naturally does n o t com e through altogether unam biguously
in th e T ibetan translation).
Compare th e expression pratltyasam ictpdda^unyatd in MMK x x iv . 36 c. See also
P P x x iv . 13 (unyatdrtha is pratityasam utpdddrtha), 40 (sarvadharm apratltyasam ut-
pddalaksand svabhdva&unyata) , and x x v i. 1 (p. 542. 6). In N agarjunas com m entary
on his V igrahavyavartanl 70 there is th e verse:
yah unyatam pratltyasam u tpadam m adhyam dm p ratipadam ca\
ekartham n ijagada pran am am i tam apratimabuddham\\
See also V igrahavyavartanl 22.
40 yu jyate = Tib. run ba. Cf. V igrahavyavartanl 70:
prabhavati ca unyateyam ya sya prabhavanti tasya sarvarthah]
prabhavati n a tasya kirnein na prabhavati unyata yasya] |
where prabhavati is translated into T ibetan b y srid p a . For th e idea see also MMK
x x iv . 36; and below , p. 23 w ith n ote 51.
41 See also Chapters i and x v , and below , p. 60.
42 P ra tlty a s a m u tp d d a ^ n y a td , an appositional (karm adhdraya) com pound. See
also above, note 39.
18 David Seyfort Ruegg The Literature of the Madhyamaka School

36). He who recognizes pratityasamutpada according to the Madhyamakath a t

is, in the perspective of unyaia (18)alone really recognizes the four noble
truths (40).
Chapter xxv then turns to the unconditioned (asamskrta) par excellence
nirvanaand examines its status both ontologically and logically. Earlier it
was said that, to the extent th a t an asamskrta is posited as an entity in relation
to its complementary opposite (satnskrta), it camiot be established because
the samskrta is not established (vii. 33). I t is now shown th a t if nirvana is no
thing {bhava existence*) it is equally not no thing {abhava non-existence);
it cannot be posited ontologically or logically in terms of any of the four
imaginable positions of the tetralem m a (catuskoti), viz. bhava, abhava, both
bhava and abhava, and neither bhava nor abhava. I t is no empirical or conceptual
entity, existent or non-existent, and relatable to a person thought of as a t
taining it (1718; cf. xvi. 4 sq.). Given the fact th a t nirvana as well as all the
factors (dharma) pertaining to samsara are accordingly em pty of own being,
it may be said th a t there is no differentiating property (videsana) between
samsara and nirvana (19; cf. 9). And there is nothing whatsoever between the
terminal (koti lim it) of samsara and the term inal of nirvana (20). In the final
analysis, then, no dharma has ever been communicated anywhere to anybody
by a Buddha; such is the stilling of all objectification (saropalambhopaAama) ,
the peaceful stilling of all discursive development (prapancopaAama) (24). As
observed earlier by Nagarjuna, the object of designation (abhidhatavya) comes
to a stop as the domain of thinking (cittagocara) comes to a stop, dharmatd like
nirvana being unproduced and undestroyed (xviii. 7); and the characteristic
of reality (tattva) is to be n o t dependent on another (i.e. not to be communicat
ed), still, not discursively developed through discursive developments (pra-
panca), without dichotomizing conceptualization (vikalpa), and free from
differentiation (xviii. 9).
The last two chapters of the MMK deal with the twelve members (anga)
viz. the conditioning occasions (nidana)of origination independence (xxvi)
and the sixteen speculative views (drsti) th a t stand in the way of the correct
understanding of pratityasamutpada and Sunyaia (xxvii), the good dharma being
taught with a view to eliminating all such views (30). Although these two
chapters add little th a t is philosophically essential to Madhyamaka doctrine as
expounded in the preceding chapters of the MMK, they deal with topics th at
are traditionally of importance in the history of B uddhist thought.
Together the chapters of the MMK explicate the terms and ideas mentioned
in the two introductory verses placed at the head of the treatise: I pay homage
to the best of teachers, the perfectly awakened one (sambuddha) who has
taught origination in dependence (prraiityasamvipada), the stilling of discursive
development (prapancopaAama) and peaceful (iva) , which is without destruc
tion and production, not annihilated and not eternal, neither undifferentiated
nor differentiated, and without both coming and going.
The Early Period: The Formation of the Madhyamaka School 19

N agarjunas Minor Treatises and Commentaries

About five other works, four of which are more or less minor at least with
respect to their extent when compared with the MMK, may be regarded as
belonging to the same category of scholastic texts as N agarjunas great

The Yuktisastika, a verse-text taking its name from the principle of reason-
ing (yukti) applied in the Madhyamaka, deals with the non-apprehension
(anupaiambha) of both positive and negative entities, and with the under
standing of praiityasamvipada free from the tw in extremes of existence and
non-existence which leads to final release.43 Once the idea of nihilistic non
existence, the source of all faults, has been eliminatedfor in the search for
reality (tattva) one has in fact to start with th e principle All exists (sarvam
asti, 31) even though it is certainly not to be made the object of conceptual
attachm ent (41)this yukti is specified as th a t by which existence also is
excluded (3). Release is then to be defined in terms of neither existence nor
non-existence of any entity whatsoever (5); it is rather stoppage of all im
purities (10). All dJiarmas are in fact to be realized as em pty (6unya), non-
substantial (anatman) and isolated (vivikta, 26) (cf. 3132, 57, 60). Thorough
penetration (parijnana) o f existence is indeed term ed nirvana (7); and this
involves comprehending the cessation (nirodha) of w hat appeared, in the way
of an illusion (may a), as production of an entity (8). Ignorance (avidyd) is the
origin of all; but through exact knowledge (samyagjndna) no birth and destruc
tion whatsoever is apprehended (11). This is then w hat is known as nirvdiia
in the present (drstadharma) (12). Whereas in tru th neither birth nor destruc
tion of an entity is found, references in the teaching to birth and destruction
have as their m otivation the fact th a t through knowing birth one comes to
know destruction, through knowing the latter one comes to know imperma
nence, and through knowing the latter one will understand the Law (dharma)
(2223). However, in tru th pratityasamutpdda excludes the real birth or de
struction of any th in g ; and knowing this one crosses the ocean of existences (24).
Ordinary worldlings (prthagjana) assuming the existence of substantial entities
fall into error concerning existence and non-existence; and being under the
influence of kleAa they are deceived by their own m inds (25). Not only the
teaching concerning the results of acts (karmaphala) and states of existence
(gcdi) (33), but also its references to a self (diman) and what belongs to a self
(atmiya) as well as to th e skandhas, dhatus and dyatanas are all well-motivated
(34). Still, nirvana alone is true (36). Entities originating in dependence are like
48 The num bering o f th e verses o f the Y uktiatika here follows th e T ibetan
version in the bsTan*gyur. For a translation o f Shih-hus Chinese version see
P. S c h a e f f e r , Yukti-a$tika, die 60 Satze des N egativism u s (Materialien zur K unde
des B uddhism us 3, Heidelberg, 1924). A very few fragm ents o f th e Sanskrit te x t
survive in quotations. See also S. Y a m a g u c h i , Chugan bukkyo ronko (Tokyo, 1965),
p. 29 sq.
20 David Seyfort Ruegg The Literature of the Madhyamaka School

the moon reflected in water, i.e. neither real nor altogether unreal; knowing
this one is not carried away by dogmatic views (46). The postulating of entities
is in fact the cause of holding to dogmatic views th a t entail attraction, repulsion
and disputes (47); whereas in the absence of dogmatic views no kleia will
occur (48). A ttachm ent is the source of grasping the untrue as true and of
disputing about it, under the influence of false knowledge (mithyajndna) (50).
B ut those who give up disputes m aintain no unilateral thesis (paksa, i.e. one
seeking to establish some kind of thing as ultim ately true); consequently,
being without a thesis of their own, they do not lay themselves open to another
counterthesis (parapaksa) (51). I t is, then, by freeing th e mind from fixation
on any unilateral position with all the accompanying kleia th at release is
attained (52 sq.). And since a great being (mahatman, cf. 5) sees with the eye
of gnosis (jnanacaksus) th a t entities are like mere reflections (pratibimba), he
does not stick in the mire of sense objects (visaya) (55). Beings can be divided
into three classes: the childish (bdla) who are attached to sense objects such
as visible form (rupa), the middling who become free from passion in this
respect, and those of superior intelligence who are released through knowledge
of the nature of rupa (56). The first class remains attached to what it finds
attractive, the second frees itself from this attachm ent, and the third sees th a t
things are vivikta, like a m an projected in a magical show (maydpurv^a) (57).44

The Sunyatasaptati, a treatise in verse accompanied by a prose commentary

ascribed to Nagarjuna himself, establishes emptiness on the ground of pratitya-
samutpdda, the principle of origination in dependence th a t excludes the idea
of any independent substantial entity. In particular, it discusses the non
substantiality of the twelve members of origination in dependence (avidyd,

44 The Yuktiatika contains a verse (35) statin g th at the four elem ents ( m ahd-
bhuta) are contained in v ij d n a : byu ba che la sogs bad p a j /rnam par es su ya
dag 'du\ jde es pas n i bral gyur na\ /log par m a m brtags m a y in nam /I (The Sanskrit
o f th is verse seem s to be equivalent to Jnarmitra, Skrasamgrahastra 3. 27
[p. 545]: mahbhtdi vijdne proktam sam avarudhyate/ tajjdne vigam am y d ti [read :
y a d i] nanu m ithyd vik a lp ita m f/). Later advocates o f a Y ogcra-M adhyam aka
synthesis have cited this verse in support o f their doctrine; see e .g . ntaraksita,
M adhyam aklam kravrtti 92 (fol. 79b); and Ratnkaranti, Prajpram ito-
padea, fol. 161b162a, 169b.
J u st before this verse, and after a quotation from th e L akavatarastra (x.
256 sq.), ntaraksita has quoted a verse th a t seem s to correspond to Yuktiatik
22, ex cep t th a t th e T ibetan translation has es p a (instead o f dgos p a as in the te x t
o f the Y uktiatika): d i la skye ba ci ya medj 'gag pa r 5gyur ba ci ya m ed/ skye ba
da n i gag pa dagI es p a ba ig kho na'oll (M adhyam aklam kravrtti 92, fol. 79b).
And R atnkaranti has then cited this verse together w ith verse 35 in favour o f
his V ijnapti-M adhyam aka theory (see below, A ppendix II). H ow ever, K am alala
ev id en tly considers this verse to be tak en from th e L akvatrasutra (see his
M adhyam aklam krapajik ad loc., fol. 138ab); cf. also Yidykaraprabha,
M adhyam akanayasrasam saprakarana, fol. 5 2 a 1. R atnkarantis interpretation
has been criticized b y Oorik hapa, Legs*bad*si*po, fol. 56a.
The Early Period: The Formation of the Madhyamaka School 21

samskdra, etc.), as well as klea,y karmariy agency, and the four erroneous con
ceptual inversions (viparyasa). The canonical statem ent Everything is im
perm anent (anitya) shows th a t all (conditioned) things are impermanent; but
it is not to be interpreted as adm itting entities having the property of im
permanence (58). Paramartha is then nothing but pratltyasamutpada; still,
having recourse to worldly transactional usage (vyavahdra), the Buddha has
given a true construction of the world of manifoldness (69). Thus the teaching
of a dharma on the worldly level has not been simply destroyed, but in tru th no
dharma a t all is ta u g h t; those who fail to comprehend this teaching, however,
fear the Buddhas faultless declaration (70). The worldly principle This arises
in dependence on th a t is accordingly not simply cancelled (71). A person
endowed with clear receptivity of mind (raddha faith) who searches for
reality (tattva) will investigate through correct analytic reasoning (yukti) the
principle of dependent origination; and having thereby eliminated the idea of
any positive or negative entity he achieves tranquillity (72).45

The Yaidalya-S u tra with its accompanying Prakarana is a prose work

dealing with eristic (vada). I t subjects to a critique the sixteen categories
(padartha) accepted by other philosophers who m aintain the substantive
reality of a self and of entities in the world.46

The Vigrahavyavartanl, a work in seventy verses accompanied by an auto-

commentary, contains a very im portant discussion of the objections raised
against the M adhyamaka theory of the non-substantial (nihsvabhdva) and
hence em pty (Sunya) character of all dharmas together with N agarjunas
replies, a searching critique of the opponents own logical and epistemological
assumptions, and a demonstration of the validity of the M adhyamikas theory

45 The verses o f th e Sun yatasap tati are to be found in th e bsTan*gyur in three

separate, and often diverging, versions, viz. th e karikds alone, th e verses w ith a
com m entary ascribed to N agarjuna, and the verses w ith Candraklrtis V rtti; th e
variations betw een th ese versions pose a num ber o f philological and historical
problem s. On th e T ibetan tradition o f th is te x t see Gosg^ on m rd pal, Deb'ther*
sn on po, cha, fol. 7 b .The version accom panying th e T ibetan translation o f
Candraklrtis com m entary, and hence this com m entary itself, differs from th e
version accom panying th e com m entary ascribed to Nag&rjuna; and the question
arises as to w hether Candrakirti knew th is com m entary or recognized it to be b y
N agarjuna. The Sunyatasaptati is not included in the Chinese canon. On it see
S. Y a m a g u c h i , B ukkyogaku bunshu, i (Tokyo, 1 9 7 2 ) , p. 5 sq.
46 The sixteen paddrthas o f Indian philosophy are: pramdrw,, pra m eya, sam gaya,
prayojan a , drstanta, siddh an ta, avayava, tarka, n irnaya, va da , jalpa, vitaridd,
hetvdbhasa, chain, jd ti, and nigrahasthdna; cf. e.g . N y ayasu tra I. i. 1.
A m ongst th e philosophers referred to in th e Prakarana (section 8) we find K apila,
M athara, U luka, and V yasa.On th e V aidalya see S. Y a m a g u c h i , Chugan bukkyo
ronko (Tokyo, 1 9 6 5 ) , p. I l l s q .; Y . K a j i y a m a , M iscellanea Indologica K iotiensia
6 7 ( 1 9 6 5 ) , p. 1 3 4 sq., and IB K 5 / 1 ( 1 9 5 7 ) , pp. 1 9 2 5 ; P . W i l l i a m s , J I P 6 ( 1 9 7 8 ) ,
p. 2 8 7 sq.
22 David Seyfort Ruegg The Literature of the Madhyamaka School

and m ethod.47 This treatise is especially noteworthy for its critique of the forms
of correct knowledge (pramdna, viz. pratyaksa, anumanay dgama, and upamana)
current a t the time and of their objects (prameya) (3051). Attention is called
in particular to the infinite regress (anavastha) involved in attem pting to
establish prameyas by means of pramdnas, which would then require further
pramdnas to establish them, and so on ad infinitum (32 sq.). Nagarjuna explains
furthermore th a t it is not the Madhyamikas statem ent (vacana) CA11 things
(bhdva) are without own being th at makes things so; the statem ent simply
serves to make it known th at they are em pty (64). Thus the opponents con
tention th at, on the assumption of unyatd, the Madhyamikas own statem ents
will be as em pty as everything else and hence unable to establish the non
substantiality of things merely reveals the opponents failure to understand
the sense of unyatd and his misapprehension concerning what the Madhyamika
is in fact about (21 sq .); and it cannot be used as an argument to invalidate the
communicative (jnapaka) function of the statem ent in philosophical practice
(cf. 64).48The Madhyamaka, based as it is on the unyatd theory and the prasanga
method, is indeed unassailable; for not only does a Madhyamika restrict him
self to a kind of philosophical deconstructionand therapeutic dehabituation
with respect to dichotomizing conceptualization while refraining from pro
pounding any propositional thesis (pratijnd, 29 and 59) of his own, but any
argument adduced to combat and refute the theory of unyatd is devoid of
cogency, and falls into line with and reinforces the Madhyamaka theory, since
all things can be shown to be equally non-substantial (28).4fl Accordingly, just
as the Madhyamika is not adopting an antiphilosophical stance when he states
th a t he has no thesis, so he is not taking up an antirational position when he
observes th a t M adhyamaka theory is immune from refutation. In sum, while the
Madhyamikas statem ent (vacana) is not supposed to establish some entity

47 See K . B h a t t a c h a b y a , The dialectical m eth od o f N agarjuna (Delhi, 1978), for

an English translation o f th e verses and autocom m entary. The Sanskrit te x t has
been edited b y E . H . J o h n s t o n and A. K u n s t , MCB 9 (194951), pp. 10851,
reprinted in K . B h atta ch a rya s book just cited. See also S. Y a m a g u c h i , B ukkyogaku
bunshu (Tokyo, 19723), i, pp. 186,* ii, pp. 533.
48 Elsew here it is explained th a t unyatd does n o t h ave th e function o f m aking
dharmas em p ty since th is is w hat th ey are; a sentence concerning unyatd therefore
serves to m ake this fact know n (see K asyapaparivarta 6364; cf. Candrakirti,
P P xiii. 8 and M adhyam akavatara vi. 34). Candrakirti (P P iv. 8) has also referred
to unyatd as an instrum ent (karana) revealing th a t ru pa is nihsvabhdva. (Cf. the
notion o f the vyanjaka-hetu) . See also Pancavim ^atisahasrika (ed. N . D u t t ), p. 37 38.
49 H ere th e word sadhyasam a apparently does n o t refer to th e logical fault o f
circularity or petitio p rin c ip ii so named. On sadhyasam a see also V igrahavyavartanl-
V rtti 69, and MMK iv. 89 (above, p. 12).Moreover, th e fault th a t an opponent
m ay allege w ith respect to unyatd does n ot touch th e M adhyam ika since, in the
case o f w hat is em p ty, it sim ply does n ot apply (MMK x x iv . 13; above, p. 13); cf.
Y igrahavyavartani 59 (and Aryadeva, Catuh^ataka x vi. 25).On sadhyasam a as a
fallacy in N y a y a and in B u ddhist usage cf. B. K . M a t i l a l , J I P 2 (1974), pp. 21124,
and K . B h a t t a c h a b y a , ibid., pp. 22530.
The Early Period: The Formation of the Madhyamaka School 23

as being non-substantial in the manner of either a syllogism or a propositional

thesis with probative force, it retains its function as a sentence and statem ent
having communicative (jndpaka) force; and the Madhyamika does not hesitate
to uphold it as expressing a doctrine of Buddhism. As for the im m unity of
Madhyamaka theory against refutation, it evidently stems from the fact that,
once the concept of an entity endowed with own being has been criticized and
deconstructed through the understanding of non-substantiality and emptiness,
not only is no position postulating such an entity to be maintained by the
Madhyamika himself, but also no refutation from a counterposition can be
successfully made by an opponent in the frame of any of the current philo
sophical positions based on the postulation of some sort of entity. The Madhya
mika does not therefore tu rn out to be trying to devise a strategy of immuniza
tion to make his theory empirically and logically unassailable; for the immunity
of Madhyamaka theory is the consequence of the annulment (zeroing) of all
hypostatized and dichotomously structured concepts, and it pertains only to
the domain of an entity conceived of in terms of a conceptual dichotom y: the
Madhyamika simply postulates no entity which could become the object of
conceptualisation involving alternative speculative views and hence of debate.
But, as already noted, the Madhyamika still does have a theory and doctrine
a dariana and vddawhich he takes to be th a t of the Buddha (see above,
pp. 23); and his doctrine then requires validation and eventually defence
against attem pts to falsify and refute it, something th a t the successive masters
of the Madhyamaka have sought to supply.50 The W then concludes with the
observation th a t for someone for whom unyatd applies all applies, whereas
nothing can apply for someone for whom unyatd does not apply (70).51
Although basically concerned with the logical and epistemological implications
of the sunyatd theory, this work does not fail to say something also about
soteriology and the religious life (brahmacarya) (5456, 70).52

The Ratnavali, a treatise in some 500 verses divided into five chapters, has
also been customarily regarded as a minor work of Nagarjuna despite the fact
th a t it is actually longer than the MMK; this is probably because it is somewhat
less philosophically incisive and comprehensive than the latter treatise.53 I t
60 A bsence of thesis and im m u n ity from refutation as tw in features o f M ad hya
m aka theory require further historical stu d y and philosophical analysis.
61 Compare MMK x x iv . 14 (above, p. 17), where the correlative pronouns
y a sya . . . tasya could gram m atically be interpreted as referring n ot to a person but
to a thing. H ow ever, th e gloss to th e P P which equates yu jy a te, rocate and ksamate
can app ly only to a person. The purport o f Candraklrtis explanation supports this
interpretation also; and B h a va viv ek a in his Prajnapradipa (za, fol. 285 a286b)
indicates th a t th e pronouns in th e first hem istich refer to persons, nam ely th ose
B u ddhists who do n ot accept unyata.
62 Compare th e first part o f Chapter x x iv o f the MMK.
68 I t has not been included in th e M adhyam aka section o f th e bsTan-gyur but
in th e sPriiVyig section o f epistles etc. On its inclusion in th e Parikatha-corpus see
above, p. 8.
24 David Seyfort Ruegg * The Literature of the Madhyamaka School

takes the form of a partly paraenetic discourse addressed to a king, usually

identified as a Stavhana monarch, whence its subtitle of
The first two chapters are concerned chiefly with the temporal good (abhyudaya)
to be gained through salutary deeds and with the summum bonum ( naihreyasa)
achieved through knowledge; they accordingly deal with ethics (e.g. the ten
salutary Jcarmapatha, i. 824) and discriminative understanding (prajd).
Prajna is supreme and through it one understands reality (tattva); but faith
(raddh) is a necessary preliminary by which one as it were participates
(bhaj-) in dharma (5; cf. ii. 27). Philosophical egoism (ahamkdra self-grasping)
along with the positing of things belonging to the self (mama) are founded on
error and are eliminated by exact knowledge (;yathbhtaparijna, 28); so
long as they have not yet been removed they are the source of acts (karman)
binding one to the round of existences and hence to birth (3537; cf. ii. 24).
The five psycho-somatic groups (skandha) proceed from ahamkdra (29), which
is itself perceived on the basis of the skandhas (32 a b); but in tru th ahamkdra
has no reality and is like a reflection (pratibimba) in a mirror (32 cd33).
Knowing the skandhas to be ultim ately not true (asaXya) one eliminates
ahamkdra, and as a consequence the skandhas do not arise again (30). The round
of existences proceeding from mutual causal conditioning has no more substance
than a wheel of flames formed by a rotating firebrand (aldtamandala, 36). As
for nirvana, while it is absence of all th a t makes up samsdra it is not to be
regarded as either non-existence or existence: it is cessation of the postulation
of both existence and non-existence (bhvbhdvapardmaraksaya) (42). Yet, as
an ethically valid theory, astitddrsti consists in accepting th a t acts bear fru it;
and this leads to favourable states of existence (sugati) (44). By the stilling
(anti) of the ideas of existence and non-existence in gnosis both sin and merit
are transcended (pdpapunyavyatikrama), and there is release from good and
evil states of existence (of which five are enumerated in 23) (45). Knowing
conditioned production one goes beyond nihilism (ndstitd), and understanding
arrest (nirodha) one does not become involved in existence (astitd) (46).
A cause th at is either antecedent or simultaneous in relation to its effect is in
fact no cause at all;55 and production (utpattiy of self-existent entities) is ac
cordingly not understandable in terms of either conventional designations
(prajapti) or ultim ate reality (tattva) (47). The canonical formula for con-
ditionship This being th a t is (asmin satidam bhavati) does not then imply
self-existence (svabhdva); it refers rather to conditioned relation, as between
the long and the short (4849; cf. 92, 95). The same applies to the canonical
formula for a causal relation, e.g. between a lamp and light (48^19). Because
he resorts to neither existence nor non-existence he who is not fixed in duality

84 The identification w ith K in g b D espyod is m ade b y the com m entator A jita-

m itra, whose work is contained in th e b sT am gyur. Cf. below, n ote 59.
86 B ecause an antecedent cause could n ot enter into im m ediate connexion w ith
its supposed effect, and a cause existing sim ultaneously w ith its supposed effect
would be redundant.
The Early Period: The Formation of the Madhyamaka School 25

(advayanirita) is released (51, 57). On the contrary, if one erroneously takes

the world which is a t the same time causally conditioned and mirage-like
(maricipratima)to be either existence or non-existence of some thing (like
water in a mirage), one is not released (5556).58 Moreover, by the denial of
existence non-existence would be implied (aksip-j, and vice versa (5859).
B ut for those taking recourse in bodhi, there is really no propositional thesis
(pratij), no conduct, no citta relating to nihilism ; and one cannot therefore
speak in this case of nihilists (60, i.e. persons bound for a durgati, 57, the
ndstitadrsti consisting besides in denying th a t deeds bear fruit, 43). The fol
lower of the Madhyamaka differs from such people as the Samkhya, Vaiesika( ?)
and Nirgrantha, as well as from a Pudgalavdin, by transcending all positions
based on existence opposed to non-existence (61): the quintessence of the
B uddhas teaching lies precisely in the fact th a t it transcends any such posi
tion (62). The idea of a thing being either perm anent or m omentary is illogical
too (6670); and the idea of an atomic substance is antinomic (71). A thesis
(paksa) implies a counter-thesis (pratipaksa)i but neither obtains in fact
(ii. 4; cf. i. 72); this world is thus in tru th (paramdrthatah) beyond tru th and
falseness (satyanrtdtita), and neither existence nor non-existence really ap
plies (5). Now, since anything th a t is altogether not so (neti)51 cannot be
stated (in some determ inate manner in terms of a logico-ontological position)
by the omniscient Sage, the teacher has refrained from stating th a t the world
(loka) has a (posterior) limit, does not have one, both has and does no t have
one, and neither has nor does not have one (6).68 T hat the world is like a magical
show (mdyd) is indeed the quintessence of the B uddhas teaching (9). An end
of birth there may be, just as the magicians magically projected elephant
comes to an end; but a t no stage is it some thing th a t is real (1011). To m ain
tain th a t some thing is or is not is then nothing but transactional usage
(vyavahara) (14). Hence the four conceivablepositions (within a caluskoti)
relating to the end of the world have been left unexplicated (avydkrta) by the
B uddha (15) (cf. i. 73 on the silence of the Buddha). And it is this very subtlety
and profundity of the dharma realized by him th a t accounts for the B uddhas
initial hesitation to teach (18). Rejection and nihilism are to be avoided (23).
So long as the dharma putting a stop to ahamkdra has not been fathomed, one
m ust observe the dharma th a t consists in liberality, ethics and patience (25).
T ruth (satya) lies not in agreement (or correspondence: avisamvddavat) but in
total altruism (paraikdntahita); and the false (mrsd) is comprised of non-
beneficialness (ahitatva) (35). Truth, liberality, tranquillity, and discriminative
understanding also make up the four virtuous qualities of a king (39). This
chapter also contains a section on the impure nature of the body of a woman
(a theme developed in Chapter iii of ryadevas Catubataka) and by the same
69 Cf. V igrahavyavartanl 6568 for th e mrgatrsnd sim ile, and also MMK xviii.
8 -9 .
57 Cf. MMK xviii. 8.
68 Cf. MMK xxii. 12 and x x v . 2123.
26 David Seyfort Ruegg *The Literature of the Madhyamaka School

token of th a t of a man (4870); and a section on the 32 marks and 80 signs

common to the Cakravartin and the Buddha (77 sq.), although there is really
no commensurability between these two beings (99100). Chapter iii is devoted
to the collections (sambhdra) of merit and gnosis, from which proceed respec
tively the rupakdya and dharmakdya (1012). Chapter iv is entitled An instruc
tion on the conduct of a king, the maintainer of the dharma, on whom rest in
large part the economic and political dimensions of a religious foundation
(dharmdspada). In addition this chapter deals with the senses, their objects
and the mind (5065), the two sambhdras (66 sq.) and the six paramitds (80 sq.).
I t includes also a defence of the Mahayana whichthough not incompatible
with the Vehicle of the Auditors (rdvakayana) (cf. 86 cd and i. 40-^41)differs
from it by expounding the vow (pranidhi) and conduct of the B odhisattva
and his dedicatory transfer of conduct (carydparinamand) (cf. v. 84), so th a t
only in the M ahayana does the Bodhisattva ideal come to realization (9091).
Furthermore, whereas for others unyatd may be destruction of some thing, in
the M ahayana it is rather non-production (anutpada) (86 a b). As between the
teaching of the one Vehicle (ekayana) and of three Vehicles one has to guard
oneself with im partial equanimity (upeksa), for no demerit accrues from this
equanimity while evil results from aversion; indeed, it is not easy to penetrate
the intentional utterances (abhisamdhyoktani) of the T athagata (8889). The
question further arises as to how the fruit of buddhahood is supreme if it
comes from a path th a t is entirely common with the Sravakas (92); but the
difference lies in the M ahayanas proclamation of the B odhisattvas ideal (93).
Finally, because of the harsh injustice (vaidharmya; Tib. m i bsrun) prevailing
in the world (whereas dharma is essentially compassion, 99), it is truly hard for
a king to exercise his rule according to dharma; hence, with a view to dharma
and glory (yaAas), he m ay enter the life of religion (pravrajyd) (100). While
still a householder (grhastha) one remains characterized mainly by liberality,
ethical discipline and patience (99, qualities which permit one to achieve ones
own and others benefit, 81). Chapter v treats of renunciation and the religious
life, and especially the ethical practices and the ten stages (bhumi) of the
Bodhisattva, the Buddha-stage, and the B odhisattvas vow. I t is observed
th a t the precepts taught in the Ratnavali are useful not only to a king but to
all beings (98).

These five works represent, together with the fundamental MMK, Nagar-
junas above-mentioned theoretical scholastic treatises th a t have been referred
to by both Candrakirti and the Tibetan lists of the Yukti-corpus (rigs hogs,
where the R atnavali is, however, often replaced by the now unvailable *Vyava-

The Suhrllekha is also reported to have been addressed by Nagarjuna to a

king, just like the R atnavali.59 A homily dealing largely with ethical m atters,
69 In verse 14 th e k in gs nam e is given as bDe-byed. The Suhrllekha is
The Early Period : The Formation of the Madhyamaka School 27

it touches on some basic philosophical topics too; and it refers to Amitabha

and his pure field. Evidently intended to be a compendium of the ethico-
philosophical ideas common to the Buddhist traditions, it contains little of
w hat may be regarded as N agarjunas particular contribution to the explica
tion of Buddhist philosophy.80

Three collections of stanzas on the virtues of intelligence and moral conduct

ascribed to Nagarjuna are extant in Tibetan translation. Included in the section
of the bsTan*gyur containing works on good conduct and polity (niti), they
are entitled Prajna^atakaprakarana, NitiSastra-Jantuposanabindu and Niti-
6astra-Prajnadanda.81 In Buddhist tradition Nagarjuna has indeed been
reckoned as one in a group of ascetics concerned with affairs of the realm
(yatayo rajyavrttinak); b u t it is uncertain which of these works, if any, are
actually by the author of the MMK.

Another im portant tex t ascribed to Nagarjuna is the Shih-erh-men-lun or

Twelve-topic treatise (*Dvada3anikaya or *Dvadasamukha-&Lstra ?); one of
the three basic treatises of the Chinese Madhyamaka school, it is extant only

connected w ith K in g b D e sp yod (cf. above, p. 24 n ote 54) b y th e co m

m entator M ahm ati, w hose work is included in the bsTan^gyur. Cf. S. L v i ,
J A 1936, pp. 10310; , L a m o t t e , Trait, iii, p. liiiliv (referring to
U dayan a and *Jan tak al), and i, p. x iix iv (referring to Y ajhair). The nam e
bDe*spyod (bzah'po) has b een identified w ith U dayana(bhadra) ^ S tavhana (or
ta) ^ *A ntivhana/m T har,,gro*zon; see e.g . P. C o r d l e r s In d ex to the P eking
bsTan*gyur, mDo*,grel, vol. gi, no. 32 (Suhrllelka), and vol. ne, nos. 27 and 35
(Suhrllekhatk) ; Bu*ston, Chos*byun, fol. 100b (ii, p. 127); Tarantha, rGya*gar*
chos*byun, pp. 5560, and b K a ^ b ab sL d m rgy rrn a n rth a r, fol. 181b sq. Cf. also
Bna, Haracarita, Chapter viii as w ell as K uthala, L lva 1008, and the later
J aina tradition; J. T a k a x u s t t , Record o f the B uddhist religion b y I-tsing (Oxford,
1896), p. 158; A. K . W a r d e r , Indian K v y a literature, ii (N ew D elhi, 1974),
pp. 183, 21314; J. W . de J o n g , I I J 20 (1978), p. 137 ; above, p. 5 note 11.
60 On th e concept o f th e Pure L and in N grja n a s doctrine see S. Y a m a g t j c h i ,
E B 1 (1966), pp. 3447 (w ith special reference to th e Daabhm ikavibhgstra,
on which see below , p. 29). B o th th e Lankavatrastra (x. 166) and the Manjusr-
m lakalpa (liii. 451) m en tion th a t N agarjuna is to be reborn in th e Sukhvat.
(On A m ityu s in th e D harm adh atu stava see below, p. 32). A link betw een the
doctrine o f nyat and A m itabha is to be found in th e P ratyutpannabuddhasam -
m ukhvasthitasam dhistra for exam ple.
English translations o f th e Suhrllekha h ave been m ade from th e T ibetan version
b y H . W e n z e l , JP T S 1886, pp. 132, S. B e y e r , The B u ddhist experience (Encino
and B elm ont, California, 1974), pp. 1018, and L. K a w a m u r a , Golden Zephyr
(E m eryville, 1975); and from the Chinese version b y S. B e a l , The Suhrillekha
(London, 1892).
61 The Prajndanda (Lugs^kyrbstanb cosSesTab'sdorrbu/po) was published and
translated b y W . L. C a m p b e l l , She-rab D ong-bu or P rajnya D and a (Calcutta,
1919). As for th e Prajnatakaprakarana, it is included in the M adhyam aka section
o f the bsTan*gyur also.
28 David Seyfort Ruogg The Literature of the Madhyamaka School

in K um rajvas Chinese translation (Taish 1568).82 I t has a number of verses

in common with the MMK. This treatise deals with twelve topics : origination
in dependence (pratltyasamutpdda, understood as the outer co-operating
hetu-pratyayas in conditionship as well as the twelve inner members or angas
beginning with avidy), the theory of the pre-existence of an effect in its cause
(satkdrya) and the opposite (asatkdrya) theory, conditions (prctiyaya), the
characteristics of conditioned things, the characterized and uncharacterized,
oneness and difference, existence and non-existence, own being (svabhava),
cause and effect, the agent, the three times (viz. past, present and future), and
birth. The commentary on this work, also available in Chinese only, is ascribed
sometimes to Ngrjuna himself and sometimes to Ching-mu.

Also attributed to Ngrjuna are the Pratityasam utpdahrdayakriks

with a Vykhyna, which deal with the twelve members (anga) of origination
in dependence divided into the categories of klea (comprising avidy, trsnd
and updna), karman (comprising samskdra and bhava)y and duhkha (i.e.
vipdka, comprising the remaining seven members). All these factors are shown
to be em pty of own being, so th a t in reality there is no sattva. The final verse (7)
has a parallel in the Ratnagotravibhga (i. 154) and the Abhisamaylamkra
(v. 21), and partially in Avaghosas Saundaranandakvya (xiii. 44).83 The
basic tex t and commentary take up the question of transfer (samkrdnti) from
one existence to another and deny any real transm igration of a being a t the
time of reconnexion (pratisamdhi) of the skandhas.

The Bhavasam krnti, another short treatise ascribed to Ngrjuna, does not
actually deal with this transfer despite its title. I t treats five topics: (i) the
absence of birth of entities and dharmatd; (ii) vikalpa as the source of the loka
and thinking (citta)y which is by nature without vikalpa, and the emptiness
of the five skandhas; (iii) the non-existence of the dharmas given the ultim ate
non-existence of citta, non-duality (advaya) and prajnd; (iv) the pdramitds as
means ; and (v) the twin principles of updya and prajnd, the fact th a t things are
mere designation (ndmamdtray the division between expressions and w hat

61 See N . A i y a s w a m i S a s t r i , V isvabharati Annals 6 (1954), pp. 165231 ; R . H .

R o b in so n ,E arly M dhyam ika, pp. 3233; R . G a r d , IB K 2/2 (1954), p. 717.
A te x t en titled sToiVpa'hidk yrsgo 'b cu g n isp a is listed in th e lDan*dkar*ma
Catalogue (no. 595), bu t it is n ot found in th e existing editions o f th e bsTan*gyur.
B u ston in fa ct n otes th a t th e sTon*pa,n id ,k y r sg o ,b cu ,gn is,pa*r6a*,grel were n ot
available to him (Chos*byun, fol. 159a 7).
63 See L . de L a V a l l e P o u s s i n , MCB 1 (19312), p. 3945; D . S e y f o r t
R u e g g , L a thorie du tathgatagarbha et du gotra (Paris, 1969), pp. 31336.
K rik 5 is quoted and ascribed to N grjuna-pda in th e P rasannapad x xi. 21
and x x v i. 2 (both passages are m issing in the T ibetan translation o f the P P ).See
V . V. G o k h a l e , Studia Indologica (Festschrift W . Kirfel, B onn, 1955), pp. 1016;
C. D r a g o n e t t i , W ZK S 22 (1978), pp. 8793 (who argues against N grjunas
The Early Period: The Formation of the Madhyamaka School 29

they express being merely conventional), and the two truths.64 On this work
there is a commentary by the Pandit Maitreyanatha, who quotes Santidevas

Causation and the twelve 'inner1 angas are also explained briefly in the
Salistambakarikas ascribed to Nagarjuna, who is in addition credited with a
Tika on the &alistamba(ka)sutra.65

Works on Madhyamaka practice and the path of the Bodhisattva are also
ascribed to Nagarjuna. The *Bodhisambhara-Sastra treating of the Bodhi-
sattv as equipment conducive to Awakening (bodhi) was translated into
Chinese by Dharm agupta in the early seventh century (Taisho 1660), this
being the only version of the work now extant. I t is quoted by Candrakirti in
his commentary on the Catub^ataka.66

An im portant commentary on the DaSabhumikasutra dealing with the stages

(bhumi) of the Bodhisattva is attributed to Nagarjuna. I t contains quotations
from the *Bodhisambhara-6astra, on which it appears to be based.67

Nagarjuna is furthermore credited with an extensive anthology of canonical

texts, the Sutrasamuccaya. As already noticed above, it is included in the
M adhyam akaiastrastutis list of the master's works.

Certain other minor works attributed to him by the Chinese or Tibetan

traditions are of uncertain authorship in view of their doctrinal contents. The
MahayanavimSika68 for example not only speaks of cittamdtra (18)something
84 The treatm ent o f these topics (if n ot th e topics them selves) could suggest th a t
th is treatise m ay be b y a later author.On this problem atic w ork cf. N . A i y a s w a m i
S a s t r i , B havaaahkranti Sutra and Nfig&rj u n as B havasankrSnti S&stra (Adyar,
85 Compare MMK, Chapter x x v i.For a M adhyam aka version o f this Sutra see
V . V . G o k h a x e in M ahayanasutrasam graha i ( e d . P . L. V a i d y a , Darbhanga, 1961),
pp. 1 0 7 -1 6 .
88 Catuh^atakatika, y a , fol. 103 a 2 . On th e B odhisam bharaiastra c f . U r y U z u
R yush in , IB K 17/2 (1969), pp. 5139; J . W . de J o n g , AM 17 (1971), p. 109.
67 The authorship o f th is w ork available only in K um arajivas Chinese translation
(Taisho 1521) has been questioned b y A. H i r a k a w a , I B K 5/2 (1957), p. 504 sq. B u t
see R . H i k a t a , Suvikrantavikrami-Pariprcch& Prajnaparam ita-Sutra (Fukuoka,
1958), pp. lii sq., lxx ii sq., where th e follow ing sequence o f works b y N agarjuna is
suggested: MMK, T a-chih-tu-lun, *Bodhisam bharaastra, *D aiabhum ika(sutra)-
vibhaa. See also S. Y a m a g u c h i , E astern B uddhist 1 (1966), pp. 4547.A n earlier
Chinese translation o f th e *D aiabhum ikavibhaa is reported to h ave been m ade by
Dharmarak^a, who worked at Chang-an betw een 265 and 313. A ccording to E .
L a m o t t e , L enseignem ent de Vim alakirti (Louvain, 1962), p. 76, this could m ean
th a t a work b y N agarjuna reached China b y the year 265, long before K um arajlva.
On th e critique of V edic and U pani?adic doctrines in this work see H . N a k a m u r a ,
H JA S 18 (1955), pp. 8992.
88 This work has been assigned to N agarjuna in Dipamkara^rljnanas B odhi-
30 D avid Seyfort Ruegg T he L iteratu re of th e M adhyam aka School

a Mdhyamika m ight well do in the sense the expression has in the Daa-
bhm ikasutra89but it also denies the existence of external objects (19), a
doctrine characteristic of the Vijnnavda th at does not seem to have been a
tenet in the pure Madhyamaka school and which made its appearance only much
later in the Yogcra-Madhyamaka synthesis.70 The Svabhvatrayapravesa-
siddhi,71 the tex t of which is in any case closely related to Vasubandhus
Trisvabhvanirdea from which it differs only in some details, can scarcely be
by Ngrjuna I (i.e. the author of the MMK) because it concerns the three
svabhdvas of the Vijnnavda school.72
The *Ekaslokastra, extant in a Chinese translation by Gautama Praj n
rui (Taish 1573), explains in a single stanza the sense of non-substantiality.73
Ngrjuna is also credited with a refutation of theism (lvara) .13&
The Aksaraataka although sometimes ascribed to Ngrjuna is likely to be
by ryadeva.
m rgadpapajik (fol. 324a).Cf. V. B h a t t a c h a r y a , M ahynavim aka o f
N grjuna, V ivabharati Studies, I (Calcutta, 1931); S. Y a m a g u c h i , E astern
B u dd h ist 4 (1926), p. 56 sq.; (1927), p. 169 s q .; G. T u c c i, Minor B u ddhist te x ts, I
(R om e, 1956), pp. 201 sq. The w ork exists in tw o T ibetan translations, and in a
Chinese one (Taish 1576).
89 Daabhm ikastra, Chapter vi (E ): ciltam tram idam y a d idam iraidhiukam; c f.
Candrakirti, MA vi. 68 (below, p. 73).
70 Cf. ntarak^ita, Tattvasam graha, Chapter x xiii, w ith K am alaslla, P a jik i. 6
(p. 18).
71 See S. Y a m a g u c h i , Shukyo kenkyu, 1931; L. de L a V a i x e P o u s s i n , MCB 1
(1932), p. 404; 2 (1933), pp. 1 4 7 -6 1 .
72 The fa ct th a t three (at least partly) comparable notions are found in som e o f
th e Prajpram it-Stras does not seem to stand against th is estim ation since the
conception in these Sutras is perhaps at th e m ost a forerunner o f th e developed
Y ogcra/V ijnnavda th eory o f th e three natures, w hich is reflected here in the
Trisvabhvanirdea. On th e Prajpram it-Sutra version o f th e doctrine see
E . C o n z e and S. I i d a , Melanges L ouis R enou (Paris, 1968), pp. 22942; D . S e y f o r t
R u e g g , Theorie du tathagatagarbha et du gotra, pp. 1478, 3256. See however
N . H a k a m a y a , IB K 24/1 (1975), pp. 2 0 - 3 0 .
73 Translated b y H . R . R a n g a s w a m i I y e n g a r , Mysore U n iv ersity Journal for
Arts and Science 1 (1927), no. 2.A T ibetan version, apparently n ot included in th e
bsTan*gyur, is in th e Stein collection in L ondon (no. 595 o f L. de L a V a e l e
P o u s s i n , Catalogue o f th e T ibetan m anuscripts from Tun-huang in th e India
Office Library, Oxford, 1962).
73a The *Ivarakartrtvanirkrtih, Vinor Ekakartrtvanirkaranam included in
th e bsTan*gyur, th e Sanskrit te x t o f which has been published b y F. W . T h o m a s ,
J R A S 1903, pp. 3459 and 703, and Th. S t c h e r b a t s k y , A B uddhist philosopher
on m onotheism , in H . C. G u p t a , ed.-transl., The papers o f Th. S t c h e r b a t s k y (Cal
cu tta, 1969), pp. 116 (translated from th e original R ussian published in Zapiski
vostok. otd. Im p. R usk. Arkheol. Ob., St. Petersburg, 1904, v ol. 16 no. i, pp. 058
074). See also N grjunas Shih-erh-men-lun, Chapter x ; . L a m o t t e , Trait de la
Grande V ertu de Sagesse, i, pp. 137, 141.Cf. H . J a c o b i , D ie E n tw icklung der
Gottesidee bei den Indern (Bonn, 1922), pp. 39, 63; H . N a k a m u r a , H JA S 18
(1955), p. 84 s q .; G. C h e m p a r a t h y , W ZKSO 1213 (196869), pp. 85100; H . von
G e a s e n a p p , B uddhism us und Gottesidee (Akademie der W issenschaften und der
L iteratur, Mainz, 1954).
The Early Period : The Formation of the Madhyamaka School 31

The Hymns
Another im portant category of works ascribed to Nagarjuna consists of
hymns (stava) . They are mentioned under the generic name of samstuti in the
M adhyamaka^astrastuti ;74 and in the Tibetan tradition they are known col
lectively as the hymnic corpus (bstod hogs, as opposed to the rigs hogs or
scholastic corpus).75
Prajnakaram ati has referred in particular to a group of four hymns, the
Catubstava, in his Panjika on the Bodhicaryavatara (ix. 76 which quotes the
Niraupam yastava, and ix. 108 which quotes the Lokatitastava). Our sources
do not however make clear which of the hymns ascribed to Nagarjuna actually
make up this quartet; and they have been variously identified as the Niru-
pama or N iraupam ya0, L okatita0, C ittavajra0, and Param artha-stava,78 as
the N irupam a0, L o katita0, Acintya0, and S tu ty atita0,77 and as the L okatita0,
N iraupam ya0, A cintya0, and P aram artha0 (this last identification follows the
commentator Am rtakara).78
In several of the hymns we find a positive conception of absolute reality
(paramartha) and a cataphatic approach to it, and these features distinguish
the hymns philosophically from the apophaticism which characterizes the
theoretical scholastic treatises comprising N agarjunas Yukti-corpus. The
Niraupam yastava speaks of the B uddhas being perceived by the faithful
(bhakta) and those who are devoted (Idlasdh mos m am s) to the idea of
the deeds accomplished by the Bodhisattva and Buddha (23). The Param artha-
stava also alludes to praise bestowed with bhakti by the devotee on the guru,
i.e. the buddha who, though beyond speech, is praised by the devotees use
of expressions belonging to ordinary linguistic usage. B ut in the final analysis
who/what is the object of praise when all dharmas are in reality empty (unya) ?
Philosophically the Niraupam yastava is remarkable for its reference to the
non-differentiation of the dharmadhatu, which justifies the non-differentiation
of the vehicles (yana) and thus establishes the One Vehicle (ekayana); the three
separate ydnas have been taught merely as a means to introduce living beings
to the teaching (21).79 This hymn adds th at the dharmamaya kaya is permanent
(nitya), stable (dhruva) and peaceful (diva) (22). On these two points in
particular the Niraupam yastava is not far removed from the theory of the
absolute expounded in the doctrine of the tathagatagarbha (to be found in

74 In his P P and M adhyam akvatra Candrakirti quotes verses from the N irau
pam ya- and L ok tita-stava.
75 Cf. D . S e y f o r t R u e g g , Le D harm adhtustava, in : tu d es tibtaines ddies
la mm oire de Marcelle Lalou (Paris, 1971), p. 448 sq.
78 L . d e L a V a l l e P o u s s i n , Muson 14 (1913), pp. 118.
77 P . P a t e l , IH Q 8 (1932), pp. 3 1 6 -3 1 , 6 8 9 -7 0 5 ; 10 (1934), pp. 8 2 -8 9 .
78 G. T u c c i, Minor B u d d hist tex ts, i (Rome, 1956), pp. 2357.In his Trait de
la Grande V ertu de Sagesse, iii (Louvain, 1970), p. xliii, . L a m o t t e opted for La
Valle P oussin s identification.
79 Cf. R a tn va ll iv . 88 (referred to above, p 26).
32 David Seyfort Ruegg The Literature of the Madhyamaka School

certain M ahynist Stras and in the Ratnagotravibhga) and the prakrtistha-

gotra (to be found in the Prajnpram it literature).
Still closer to the talhagcUagarbha/tathdgatadhdtu theory, and even more
consistently cataphatic in approach, is the D harm adhtustava80which, though
not included in the Catubstava collection, has been ascribed to Ngrjuna both
by the Buddhist traditions and by a number of modern scholars.81 At first
sight this hymn would perhaps seem to be by a later Ngrjuna (there are
indeed traces in it of ideas associated in particular with Tantrik thought) ; but
in the present state of our knowledge the possibility cannot be altogether
excluded th a t at least the kernel of this work is an early production of the
Madhyamaka school, if not a work by Ngrjuna I. While the Suhrllekha
(like the commentary on the Daabhmikastra) contains a reference to
Amitbha, the D harm adhtustava alludes to Amityus (56 sq.).82

The *Mahprajnpramitopadea
The Ta-chih-tu-lun (*Mahprajnpramitopadea, Taish 1509), a most
extensive treatise of encyclopaedic scope ascribed to Ngrjuna, is formally a
commentary on a large Prajnpram itstra (the Pancavimatishasrik). Its
authorship has been the subject of recent discussion, and it has been argued
th a t it is the work not of the author of the MMK (Ngrjuna I) but of an author
(a Ngrjuna II)or authorswho lived in northwestern India or Kmr not
earlier th an the beginning of the fourth century, i.e. a t least a century later
than N grjuna I who is besides considered to have been a South Indian.88
Remarkably, this treatise is unknown to the Indian and Tibetan traditions;
and it is available only in the Chinese version made a t the beginning of the
fifth century by Kum rajva, a scholar and translator from Ku who had
studied in Kmr, and his Chinese collaborators.84 I t is in fact not certain
exactly how much of the work we now have goes back to an Indian original,
which K um rajva is stated by his disciple Seng-jui to have used in his work
of translation, and how much of it might be the product of Central Asian (or

80 See D . S e y f o r t R u e o o , Le D harm adhtustava de Ngrjuna, p. 448 sq.

81 S. S c h a y e r , OLZ 1935, col. 402, 406; C. R e g a m e y , Three chapters from th e
Sam dhirjastra (Warsaw, 1938), p. 25; K . V e n k a t a R a m a n a n , N garjunas
p hilosophy as presented in the M ah-Prajnpram it-Sstra (R utland and T okyo,
1966), pp. 35, 37, 368 a; R . H . R o b i n s o n , E arly M dhyam ika, p. 27.
82 Cf. Le D harm adhtustava de Ngrjuna, p. 468 n ote 103.
88 . L a m o t t e , Trait de la Grande V ertu de Sagesse, iii (Louvain, 1970), and
iv (1976). See also A. K . W a r d e r , Indian B uddhism (N ew D elhi, 1970), p. 3889.
B u t J. W . d e J o n g , AM 17 (1971), p. 105 sq., w as n o t convinced b y all o f L a m o ttes
argum ents.
84 On K um rajva (344413 or 350409 ?) see J. N o b e l , K um rajva, Sitzungs-
berichte der Preussischen A kadem ie der W issenschaften, Philos.-hist. K l. (Berlin,
1927), pp. 20633; R . H . R o b i n s o n , E arly M dhyam ika in India and China
(Madison, 1967), p. 71 sq. ; and . L a m o t t e s translation o f th e Ta-chih-tu-lun: Le
trait de la Grande V ertu de Sagesse (Louvain, 1944 sq.).
The Early Period: The Formation of the Madhyamaka School 33

Serindian) scholarship in the fourth century, and also of Chinese exegetical

activity at the beginning of the fifth century.85 A t all events there is much in
the treatise th a t speaks in favour of the hypothesis of multiple or collective
authorship. In addition to explaining a P rajpram itstra the *Upadea
refers to a large number of other Mahynist Sutras; and it quotes also the
Prajpram itstotra of Rhulabhadra, evidently a contemporary and follower
of N grjuna I (see below), as well as Aryadeva.86 I t constitutes furthermore
a very valuable source concerning Srvakaynist doctrines, especially those of
the Sarvstivda which it discusses a t length; and it thus makes possible a
comparison of an im portant current of early Mahynist thought and a leading
school of the rvakayana.
Kum rajivas version consists of two parts. The first, which is said to have
been translated in extenso into Chinese, corresponds to pp. 134 of the Paca-
vimatishasrik (in N. D u t t s edition) and contains 52 chapters (translated
by . L a m o t t e , Trait de la Grande Vertu de Sagesse, volumes iv [Louvain,
194480], pp. 12371). The second part, which is presented as an abridgement
of the original, contains 89 chapters (of which only Chapter xx dealing with the
B odhisattvas entry in the Mahyna and the seven common and ten special
bhmis has been translated by L a m o t t e , Trait, v, pp. 23732445).
Especially noteworthy are the references found in the *Upadea to a positive
theory of reality (dharmatd, tathata, dharrnadhtu, bhtakoti).87

Some methodological and theoretical issues in Ngrjunas philosophy

The procedure (mentioned above, p. 9) th a t consists in regarding the MMK
and related texts of the same category attributable to Ngrjuna as providing
the criterion for establishing the genuineness of a text and defining the earliest
stage of Madhyamaka thought, however appropriate and indeed necessary it
is, clearly does not finally and decisively resolve all the problems th a t confront
the historian who attem pts to establish the authenticity of a work traditionally
ascribed to Ngrjuna and to determine the characteristic features of his
philosophy. Not only is it possible th a t Ngrjunas philosophy underwent
development and change in the course of his life, but he might even have
adopted more than one single approach to certain problems. In the works
ascribed to him we in fact find both a negative theory and apophatic treatm ent
and a positive theory and cataphatic treatm ent of the paramdrtha. And we
85 Cf. P. D e m i v t l l e , J A 1950, p. 380 sq .; R . H ttta t a , Suvikrntavikrm i-
Pariprcch, p. lii s q .; L a m o t t e , Traite, iii, p. x lv sq.In these circum stances, th e
fact th a t th e T a-chih-tu-lun quotes R hulabhadra or A ryadeva (see below) cannot
be considered as conclusive proof th a t its author is a second Ngrjuna.
86 W hether it also refers to a chapter o f A ryadevas Catuhataka, as has been
supposed b y L a m o t t e (Trait, iii, pp. xl, 1370 note), has been discussed b y d e J o n g ,
AM 17 (1971), pp. 1078, w hose conclusion is th a t it does not. See now L a m o t t e ,
Trait iv, pp. x iiixv.
87 Cf. K . V e n x a t a R a m a n a n , op. cit., pp. 16, 4445, 251 sq.
34 David Seyfort Ruegg The Literature of the Madhyamaka School

also meet an approach (that might perhaps be compared with the epocM)
according to which only silencea philosophically m otivated refraining from
the conceptualization and verbalization th a t belong to the discursive level of
relativity and transactional usageis considered to correspond in the last
analysis to the pararndrtha, which is as such inconceivable and inexpressible
in terms of discursivity. While no doubt related to apophaticism, this latter
approach goes beyond a negative as well as a positive theory; and in the
Madhyamaka it is distinguishable from apophaticism as well as cataphaticism.
The MMK and the other related theoretical scholastic treatises which do not
employ such terms as dharmadhatu, dharmakdya, tathata, etc. are mainly
apophatic in their approach to reality. On the contrary, several of the hymns
ascribed to Nagarjuna are cataphatic in approach, mention the dharmadhatu
and tathata, and even qualify the dharmamaya-kaya positively. As for the
epochistic approach, it is exemplified in the MMK by the statem ents th a t no
dharma was ever taught by the buddha to anybody anywhere (xxv. 24cd)88
and th at on the level of absolute reality designata and discursive development
have come to a stop (xviii. 7 and 9; xxii. 15; xxv. 24ab), as well as in hymns
such as the Param arthastava (12ab, 910) and the Niraupam yastava (25a).
This theory was fulfilled in the idea of the silence of the Arya (arya-tusm[m]-
bhava) . "
The negative and positive theories of absolute reality together with the apo
phatic and cataphatic approaches to its description have often been represented
as opposed doctrines. For some Tibetan exegetes the difference between the
two was indeed a t the root of the opposition between the doctrine of emptiness
of own being (ran ston) and th a t of the emptiness only of heterogeneous factors
th a t are not constitutive of the absolute (gzan ston). The advocates of the
gzan ston theory of an absolute th a t is not empty of certain constitutive
factorswhich they developed partly an the basis of the tathagatagarbha
doctrinehave then connected what they hold to be the opposed and uni
laterally negative theory with Candrakirtis M adhyamakavatara and Prasanna-
p ad a; and they have gone so far as to describe it as nihilistic emptiness (chad
ston).90 However, the fact remains th a t works connected with both theories
have been ascribed to Nagarjuna even by the Tibetan gzan ston tradition th at
contrasts them by assigning them to two distinct literary genres (the scholastic
rigs 6hogs and the hymnic bstod chogs) as w^ell as to quite different periods in

88 Cf. nyatsaptati 70.In MMK xx vii. 30 it is said th at Gautam a tau gh t the

dharma for th e purpose o f elim inating all dogm atic view s (drsti; cf. xiii. 8).
89 See R atn val i. 73 and Candrakrti, P P i, p . 57. 8. Cf" G. M. N a g a o , The
silence o f th e Buddha, in: Studies in In dology and B uddhology (Fests. S. Y am a-
guchi, K yo to, 1955), p . 137 sq. ; and . L a m o t t e , Trait, i, p . 30 n. 2; iv, p p . 20217 ;
L enseignem ent de V im alakrti (Louvain, 1962), p p . 10910, 31718.
90 Cf. D . S e y f o r t R u e g q , The Jo nan pas, JAOS 83 (1963), pp. 7391; La
thorie du tathagatagarbha et du gotra (Paris, 1969) ; and Le trait du tathagatagarbha
de B u ston R in chen grub (Paris, 1973), indexes s .v . gzan ston.
The Early Period: The Formation of the Madhyamaka School 35

Ngrjunas life.91 Furtherm ore, it has not so far proved possible to demon
strate th a t a hym n such as the Niraupam yastava with its more positive
theory of reality was composed by a writer different from the author of the
MMK. In these circumstances it may be appropriate to consider whether the
two approaches are not coordinate on the literary level (where a hymn m ay be
contrasted with a scholastic treatise) and complementary on the philosophical
and religious levels (where the two theories, though contrastive are not neces
sarily incompatible). This would appear to be consonant with the basic Mah-
ynist theory th a t ultim ately reality is not verbalizable and discursively con-
ceptualizable, and th a t it can be known only directly and immediately, i.e.
beyond vikalpa and prapaea. And it would conform with the idea th a t only
an as it were semioticized silence could adequately correspond to reality. In
sum, on the level of discursive thinking and language, the positive and negative
lines of approach (comparable respectively with the via eminentiae and via
negativa), though asym ptotic, would be regarded as valid complementary
approximations to reality, for which the silence of the Arya is a still more
adequate signifier.92
Consequently, to characterize N grjunas philosophy as exclusively nega-
tivistic and his method as only apophatic on the basis of numerous statem ents
in the MMK and other closely related treatises, to take this view as the single
doctrinal standard by which the authenticity of any work ascribed to Ngr-
juna has to be judged, and to conclude th a t a work th at does not conform to
this standard cannot be by (the same) Ngrjuna is a procedure th a t poses at
least as m any problems as it is supposed to solve. I t is moreover circular in
asmuch as it would attem pt to settle the problem of the authenticity of a tex t
on the basis of a doctrinal criterion which can itself, however, only be established
on the basis of the entire corpus of Nagarj unas writings. In the present state
of our knowledge it is therefore only proper to concede th a t we possess no
91 Cf. D . S e y f o r t R g e g g , Le D harm adhtustava d e N grjuna, p. 448 sq.
92 Silence as refraining from verbalization and as philosophically m otiv ated
aposiopesis is n ot mere absence o f sem iosis on th e pragm atic level. And it can
therefore be regarded as a sem iotic sign in its ow n right, even though it abolishes
th e ordinary processes o f th e sem antic level.I t w ould also seem to be in som e
respects comparable w ith learned ignorance, th e docta ignorantia o f N icholas
Cusanus (N icolaus o f Cues). The com plem entarity referred to above should, however,
probably n ot be regarded as a coincidence o f opposites ( coincidentia oppositorum )
i f opposed qualities are thereby th ou gh t o f as converging in a supreme en tity . For
th e M adhyamika, reality is certainly n ot to be represented in term s of_the third
position o f the catushoti, where an V is conceived o f as both A and *A\* and he
refrains from h yp ostatizin g th e param artha. U nlike Cusanus, therefore, the
M adhyam ika does not develop an idea like complicatio in contrast to explicatio
(though comparable ideas are perhaps laten t in th e M ahyna and were elaborated
in th e Stras and Sstras dealing w ith th e tathgatagarbha, and then especially in
the gzan stoh tradition). (In B uddhist canonical literature, silence (Pli tunhibhdva)
was a sign o f consent or affirmation regularly used b y th e Buddha. For Vedic links
betw een silence, the unexpressed (a n iru kia ) and th e brahm an, see L. R en o tt, La
valeur du silence dans le cuite vdique, JA O S 69 (1949), p. 11 sq.).
36 David Seyfort Ruegg The Literature of the Madhyamaka School

philologically decisive or doctrinally binding criterion by which to determine

whether, in addition to the MMK and the treatises closely related to it, a t least
some of the other works ascribed to Ngrjuna are by the author of the MMK.
Accordingly, it will be preferable to speak here of texts and philosophical
structures while for the time being reserving the question of the authorship of
some of these works, and without seeking to represent one single body of doc
trine as the philosophy of Ngrjuna.

The following sketch of some salient points of philosophical interest in

Ngrj u n as thought is based mainly on the MMK, the principal theoretical
work and the chief scholastic treatise of the Madhyamaka school.
On the methodological side several points deserve special mention. First,
starting with Ngrjuna the Mdhyamikas have employed a form of philo
sophical argument based on pointing out an eventuality or consequence
(prasanga) th a t results from any proposition or thesis operating with the idea
of the real existence of an entity (bhdva) , this prasanga being unacceptable to
the advocate of the proposition himself. This method is used to reject and annul
any and all speculative views (drsti), whatever their source, th a t involve the
hypostatization of some entity possessing positive or even negative own being
(svabhdva). The Mdhyamikas prasanga reasoning is evidently not strictly
speaking an apagogic proof because he does not seek to establish a contrario
something th at is the reverse of what has been rejected. And it is to be regarded
as comparable with a reduction ad absurdumd* procedure ordinarily used to
prove a proposition by deducing a contradiction from the negation of th a t
proposition taken together with other propositions explicitly or implicity
acceptedonly on condition th a t it is clearly understood th a t propositions
postulating the substantial own being of some entity are the opponents
exclusively, and also th a t the negated proposition stated by the Mdhyamika
is not meant by him to express or imply a dogmatic counter-view intended to
supersede the rejected proposition in a dichotomously structured set of alter
natives according to the semantic principle of bivalence. At the same time the
logical principle of the excluded middle is frequently evoked in N grjunas
98 See Candrakrti, Prasarinapad i, p. 24. 3 sq. : nihsvabhvabhvavdin
sasvabhvabhvavdinah prasanga padyam ne kutahprasagaviparitrthaprasagitl
. . . satym aktau vaktur vivalcsdm anuvidhyante / tata ca parapratijpratisedham -
traphalatvt prasagdpddanasya nsti prasagavipartdrthdpattih W hen the a d v o
cate o f th e doctrine th a t entities are w ithout own being adduces a prasanga against
th e advocate o f th e doctrine th a t entities have own being, how could there exist
[for th e former] th e occurring o f som e thing contrary to th e prasanga? . . . W hen
th ey have sem antic cap acity [words] conform to th e in ten tion o f th e speaker.
Therefore, because th e application of the prasanga results exclusively in th e negation
o f the o p ponen ts thesis, there can arise nothing th a t is contrary to the prasanga
Since all view s postu latin g anything o f an y kind o f e n tity (bhdva) are thus
excluded b y the M dhyam ika, even though he is seen to have a philosophical theory
[darana or vda) his prasanga m ethod has therefore to be distinguished from the
The Early Period: The Formation of the Madhyamaka School 37

W hat came to be known later in th e Madhyamaka school as 'prasajyaprati-

sedhai.e. non-presuppositional and non-implicative absolute negation as
opposed to paryvdasapralisedha or presuppositional and implicative relative
negationis accordingly of fundamental importance for the Madhyamika
because it does not commit him to maintaining the contradictory of the pro
position he has negated, which he would indeed consider to be no less faulty
than the one negated. This prasajya-negation will be discussed below in con
nexion with Bhavaviveka, who appears to be th e first Madhyamika to make
use of the term, and Candrakirti. Suffice it to say here th a t by, for example,
denying by prasajya-negation th a t an entity is produced in a certain way (see
e.g. MMK i. 1 and xxi. 13) the Madhyamika is n o t committed to asserting the

indirect proof (avita ~ 4esavat) b y residue (paridesa) in which one thesis about
an en tity is established b y elim inating all others as im possible. The pariesa
has been defined in th e N yayabha^ya I. i. 5 as follow s: prasaktapratisedhe saty
anyatraprasangac chisyam ane sam pratyayah. The M adhyam ikas prasanga m ethod
thus differs also from th e vitaridd ca vildefined in th e N yay asu tra I. ii. 3 as ja lpa
or specious argum ent lacking th e assertion o f a counter-thesis (pratipaksasth dpan d)
in th e place o f th e op p on en ts paksa being rejectedto the ex ten t th at, according
a t least to Vacaspatimi^ras definition in his N yayavarttik atatp aryatik a, th e
vaitan dika is one who seeks to establish his own paksa b y th e m eth od o f residue
(parii& sya). Follow ing th is definition, then, the vaitan dika too has a paksa even if
he does n ot a ttem p t to establish it directly, bu t only indirectly b y elim inating other
paksas (tasm dd asti vaitan dikasya paksah, na tu parapaksapratisedhad anyd sthdpand/
tendsya pakso s ti, n a sti tu palcsasthdpana). In sum, th e prasanga m ethod as used b y
th e M adhyam ika w ould differ from th e pari^esa-anumdna because th e latter
consists in elim inating in a finite num ber o f alternatives all the paksas belonging to
opponents, th e residual pa ksa being th en m aintained as on es own w ith ou t it being
th o u g h t necessary to establish it. B u t for th e M adhyam ika reality (param drthat
tattva, etc.) cannot be one alternative, even a residual one, in a num ber o f alter
n atives since it is not an e n tity about w hich a paksa or p ratijn d can be stated w ith in
th e frame o f com plem entary and dichotom ous opposites.
Term s used in M adhyam aka literature to denote a com plem entarily opposed
concept or category a r e : pratidvan dvin (Tib. 'gran zla , 'gal zla; see e. g. P P ii. 17, x v. 5,
x v i. 4, and x x. 21), pratibandhin (Tib. gal zla; see e .g . P P x xiii. 21), and also
som etim es pratipaksa (see e.g . P P ii. 14, 17, vii. 33, xiii. 7), th ese expressions
being usually found in th e course o f the discussion o f opponents view s. In the
R atn av all (i. 72 and ii. 34) th e word pratipaksa is used virtually in th e sam e sense.
On th e principle o f th e solidarity o f com plem entary opposites cf. J. M a y , Candra
kirti, Prasannapada M adhyam akavrtti, p. 16 and n otes 68 and 80.
I t does n ot seem th a t a prasa n g a-typ e argum ent is to be reduced to a h ypothetical
syllogism in th e tollendo tollens m ode (if p then q; b u t q ; -*.p); and according to
th e M adhyam ika no real en tities actually correspond to the variables o f the formula.
In any even t, N agarjunas procedure in the case of prasanga-typ e reasoning does
n ot appear to in volve the use o f th e hypothetical syllogism . Cf. D . S e y f o r t R u e g g ,
J I P 5 (1977), pp. 5 5 - 5 6 .
Concerning other early uses o f the word, it is to be recalled th at in the term inology
o f th e Indian grammarians prasanga m eant occasion, [provisional] application
[of a rule]. A nd in scholastic usage prasanga = p rd p ti and prasajyate = prdpn oti
occurs, applies ; cf. P P vii. 3 where pra pta is glossed as p r a s a k ta Cf. also D .
S e y f o r t R u e g g , W Z K S 22 (1978), p. 177 sq.
38 David Seyfort Ruegg The Literature of the Madhyamaka School

contradictory proposition th a t an entity is produced in the opposite way, as he

would be if the negation were of the implicative paryudasa k in d ; similarly, by
denying th a t nirvana is a positive entity (bhava, see e.g. MMK xxv) he is not
committed to asserting the contradictory proposition th a t it is a negative
entity (abhdva).9i

N agarjunas regular way of analysing and deconstructing (that is, emptying

or zeroing) any postulated entity is first to show th a t its substantial self-nature
has been constructed and posited in terms of sets of related terms.
In the MMK we meet with binary sets such as origination/destruction
(utpadalnirodha), own being/other being (svctbhdvalparabhava) , existence/non
existence (bhdva/abhdva), conditioned/unconditioned (samskrtalasamskrta),
defining mark/defined thing (laksana/laksya), and identity/difference.95 An
other binary set, th a t of agent and action (e.g. gantr goer and gati going),
may appear expanded into a ternary set, an example being the goer, the action
of going (gati or gamana, i.e. gatikriyd), and th a t which is to be gone over or
traversed (gantavya as karman)) which can in its tu rn be trichotomized as
(already) traversed (gata), (still) untraversed (agata) and in process of being
traversed (gamyamdna) (Chapter ii). A further such set is made up of a seer
(drastr)y seeing (dar&ana, i.e. dr&ikriya action of seeing) and the object to be
seen (drastavya), which can in its turn be trichotomized into the (already) seen
(drsta), the (yet) unseen (adrsta) and what is in process of being seen (drsya-
mana: see Prasannapada iii. 3; cf. MMK, Chapter xiv). Other comparable sets
are made up of the impassioned person (rakta), passion (raga) and the object
of passion (ranjaniya) (Chapter vi; cf. Chapter xiv); the bound (baddha)
person who has to be liberated and the unbound (abaddha) one, together with
the one becoming bound (see Prasannapada xvi. 7); and the person in error

94 Cf. Candraklrti, P rasannapada i, p. 13: nanu ca: n aiva svata u tpannd ity
avadhdryam dne parata utpannd ity anistam prdp n otij na prdpn oti/ prasajyap rati-
sedhasya vivaksitatvdt parato p y utpddasya pratisetsyam dnatvdtj This procedure is
n ot to be understood as ignoring or rejecting th e principles o f non-contradiction
and th e excluded m iddle; see below, pp. 39, 41, 60, 65, 68, 79, 83n ., 109.
The difference betw een these tw o forms o f negation is in som e respects parallel
to th a t betw een w eak and strong negation, although it does n ot seem th at th ey can
be w holly equated w ith these tw o kinds o f negation in modern logic in view o f the
special use o f p ra sa jya negation in M adhyam aka th ou g h t.In th e usage o f th e
Indian gramm arians, prasajyapratisedha is verbally bound predicate or sentence
negation, and paryuddsapratisedha is nom inally bound term negation. The difference
betw een p ra sa jya and paryuddsa negation was also known to the M mms school.
Cf. D. S e y t o r t R x j e q g , J I P 5 (1977), p. 3 sq. (w ith bibliography).
95 The term s used for id en tity and difference are: e k a (tva )a n ya (tva ) (see e .g .
MMK x. 1), a n a n ya jan ya (xiv. 7), eka (tva )p rth a k (tva ) (e.g. vi. 4 and x x . 20),
ekibhdva/ndnabhdva (ii. 21 cf. xviii. 10 and x xi. 10), and tattvajanyatva (xxii. 8, cf.
xviii. 1011). These relata can be th ou gh t o f as being in a relation o f cause and
effect (e.g. in x x . 20).Cf. Candraklrti, P P iv . 8, xx ii. 1 (p. 435), 5; x xv ii. 6;
M adhyam akvatra vi. 143 on tattvdnyalva.
The Early Period: The Formation of the Madhyamaka School 39

(viparlta), th e one not in error (aviparita) and the one in process of entering
into error (viparyasyamdna) (xxiii. 1718).
Nagar juna shows th a t since these sets are made up of interrelated and hence
dependent concepts or categories, no term can be posited as a real entity
possessing independent and substantial svabhdva or aseity ; for the postulated
svabhdva is by its very definition unable ontologically to exist within the
above-mentioned sets of correlates. Thus goer and going are said to be neither
one nor different (ii. 18); th a t is, they lack the postulated svabhdva of an entity.
The theory of emptiness of own being or non-substantiality of all entities
has also been explicated by means of the negation of the positions (koti) in
term s of which an entity m ay conceivably be posited, viz. a positive one (I), a
negative one (II), a conjunction of the positive and negative ones (III), and an
indeterm inate one consisting in the bi-negation of both the positive and negative
positions (IV).98 Analysis by means of the negation of this fourfold set of
positions is a further feature characteristic of Nagar junas philosophical method.
I t is employed in his discussion of the absolute non-occurrence of a bhava pro
duced from itself (I), from another (II), from both itself and another (III), and
from neither (i.e. without a cause; i. 1; xxi. 13), as well as in his discussion of
the taihdgata as neither unya (I), aMnya (II), both (III) and neither (IV) and
of nirvana as neither bhdva (I), abhava (II), both (III) and neither (IV) (Chapters
xxii and xxv). This mode of analysis is founded on the consideration th a t an
entity and its qualifier can be conceptually related only in terms of these four
limiting positions, which together exhaust all conceivable relationships be
tween a subject and its predicate; a t the same time these positions stand in a
relation of complementarity in the frame of dichotomizing conceptualization
(vikalpa) and discursive development (prapanca), the one therefore both
depending oil and evoking its conceptual correlate. Now, the things in question
are in fact found not to exist at all as self-existent entities since it has been
shown in N agarjunas analyses th a t severally or all together the four positions
do not apply. Moreover, the negation of a proposition expressing one position
does not imply the affirmation of the contrary one because the negation here is
non-presuppositional absolute prasajya-negation, so th a t all four positions are
annulled (zeroed).97

96 T he term catuskoti has n ot actually been em ployed in the MMK, A ryadevas

Catuh^ataka and Candrakirtis Prasannapada. In P rajnakaram atis B odhicaryava-
tarapanjika ix . 2 it is found in connexion w ith a q uotation tak en from a later
M adhyam ika m aster (identified as Saraha).On th e uses of th e catuskoti in th e
M adhyam aka and other schools o f th e M ahay ana, see D . S e y f o r t R u e g g , J I P 5
(1977), pp. 171. A nd on th e tw o d istinct uses o f th e neither . . . nor formula, one
o f w hich presupposes an ineffable en tity in position IV whereas th e other does not,
see ibid., pp. 1620.
97 In his analyses N agarjuna has not in every case m ade use o f all four positions
o f the catuskoti, and in m an y passages w e find m ention o f positions I and II only.
The tetralem m a is o f course itself based in th e last analysis on a binary set o f
tw o opposed term s, in conform ity w ith th e principles o f contradiction and excluded
40 D av id S eyfort R uegg T he L iteratu re of th e M adhyam aka School

A further form of analysis in five points investigates the causal relation be

tween two things, e.g. fire as one with its fuel (I), different from its fuel (II),
possessing fuel (III), the locus of fuel (IV), and located in fuel (V) (x. 14). I t is
explained th a t the relation between fire and fuel is one of appropriator (upa-
datr) and appropriated (updddna), which is thus analogous to th a t between a
self (dtman) as appropriator and the five psycho-physical groups th a t are
appropriated (upddana-skandha, x. 15).98 Nagarjuna then demonstrates th a t
none of these five relations can obtain between the two relata so long as the
latter are supposed to be entities possessing independent and substantial own
being. Factors sustaining such relations since they have originated in depend
ence can, therefore, be shown to be without own being and em pty.As sug
gested by the reference to the relation appropriator/appropriated, this form of
analysis in five points evidently derives from a form of analysis used in the
Buddhist canon to investigate a designational entity with respect to the factors
(dharma) on the basis of which i t is imaginarily constructed. Thus the concept
of an dtman or individual (pudgala)which Buddhist thinkers have regarded
as a fictional construct superimposed on the five skandhasw&s analysed with
respect to whether it is identical with the skandhas (I), different from the
skandhas (II), the locus of the skandhas (III), located in the skandhas (IV), and
possessing the skandhas (V); and in each case the relation postulated has been
shown to be untenable, so th a t no dtman or pudgala can find a place among the
dharmas which are accepted as the only valid and real factors for the purpose
of philosophical analysis.99The vicara in five points has also been applied to
the tathagata1s relation to the skandhas (MMK xxii).100
m id d le; th e full quaternary set is th en m eant to cover exh a u stively all conceptually
im aginable positions in which a p u tativ e en tity m ight be p ostulated .In MMK ii
and in the parallel analyses in Chapters iii and x v i as sum m ed up in th e verses in
Prasannapada iii. 3 and x v i. 7th e analysis rests only on positions I, I I and I V ;
cf. ii. 8 and 15, w ith th e Prasannapada, where the principle o f excluded m iddle is
evoked (B uddhapalita ev id en tly differs here and supposes th e third* to correspond
to position III ). In x x v . 1516 w e find positions I, I I and IV . E lsew heree.g . in
i. 7, ii. 2425, vii. 20, viii. 910, xxiii. 20, and x x v . 1113N agarjunas analysis
rests on positions I, I I and I I I only. (It is to be n oted th a t in x x v ii. 17 and 27 th e
subject is distributed betw een each predicatepartly . . . p a rtlyin the case of
p osition III, and th e proposition is negated b y Nagarjuna.)
In MMK xviii. 8 a fourfold set o f unnegated predicatesso* (tath ya, the anton ym
o f m rsd false according to Candrakirti), n ot so , b oth so and n ot so , and neither
so nor n ot so is applied to all (conditioned) things (sarva = sam skrtadharm as
according to Candrakirti) in conform ity w ith the B u d d has graded and successive
teachings (anuSasana) . A nd in MMK xviii. 6 th e question of an dtm an is taken up
under th e unnegated positions I, I I and IV . B u t in neither case is N agarjuna statin g
a view held b y the M adhyam aka sc h o o l; on th e interpretation o f these tw o passages
see D . S e y f o r t R u e g g , J I P 5 (1977), pp. 59, 3739.
98 Cf. Prasannapada x x v ii. 26.
99 See MMK xv i. 2 on th e pudgala in relation to the skandhas (and xxiii. 5 on the
klista in relation to th e kleas). Only id en tity and difference betw een an dtm an and
th e skandhas are considered in xviii. 1.
A difference betw een these m odes o f fivefold analysis lies in th e fact th a t the first
The E arly P e rio d : T he F orm atio n of th e M adhyam aka School 41

A close examination of language and its categories thus reveals the ground
lessness and untenability of the conventional notions of an entity possessing a
svabhdva and existing in a nexus of conceptual construction or causality. All
things are in fact ultim ately non-substantial (nihsvabhdva) and em pty of own
being (svabhdvaAunya).
When it appears from N agarjunas discussion th a t he considers th a t no thing
is to be posited in terms either of the above-mentioned binary, ternary or
quinary sets of relations or of the four positionsin short, th a t no entity
possessing a svabhdva of any kind is to be postulatedthere can be no ground
for supposing th a t he has either neglected or rejected the principle of the
excluded middle in classical two-valued logic by positing an x th a t is neither
*A nor Nor is there then any reason to feel the need to suggest th a t as a
rational thinker he was using some kind of three-valued logic;101 for in the
MMK and his related treatises Nagarjuna has carefully refrained from postulat
ing the existence of an indeterminate value for a real entity.102
Also, in his analysis and criticism of concepts and categories, N agarjunas
reasoning is clearly based on the principles of contradiction and exclusion.103

m entioned m ode concerns tw o causally related en tities (fire and fuel, for exam ple)
which h ave th e sam e ontological and epistem ological statu s, whereas th e follow ing
m ode relates to a designational en tity (atm an or pudgala) im aginarily constructed
on th e basis o f dharmaSy viz. th e five skandhas w hich possess an accepted va lid ity
in th e B u d d hist tradition, even though the M adhyam ika does n o t allow them
u ltim ate reality. I t seem s likely th a t th e latter served as a m odel for th e application
o f the fivefold analysis to th e relationship betw een dharm as on th e sam e level, and
basically it goes back to th e canon; see e .g . S am yu ttan ikaya iii, p. 44, and iv, p. 287
on th e fourfold in vestigation (i.e. rupam [or: vedandm . . . vinndnam ] attato sam anu-
pa8sati, rupavantam vd attanam , attani vd ru p am , ru pasm im vd attdnam ), (Like th e
canonical passage ju st m entioned, th e M ahavyutpatti 209 also enum erates four
rather th a n five p oints).On th e tw en ty forms o f satkdyadrsti elim inated b y this
form o f analysis, see Candrakirti, M adhyam akavatara vi. 144.Concerning th e
to ta l o f five rather th an four, it is reached b y th a t m ode o f counting w hich en u
m erates a covering item in addition to the item s included under it. T h at is, we have
basically th e tw o item s o f id en tity and difference (see MMK xxii. 8 and xviii. 1),
w ith difference th en subdivided into three ite m s ; and th e to tal can be regarded as
consisting in five or four item s according to whether th e covering item (difference)
is counted separately or n ot.
100 The word tathdgala is in this co n tex t interpreted as referring to th e buddha
(MMK xxii. 15) or bhagavant after nirodha (i.e. in nirvana) (xxv. 17). (Elsewhere
how ever, e. g. in the con tex t o f th e avyakrtavastus, tathdgata can refer to an y in d ivid u
al. Compare Candrakirtis P P xxii. 1, where m ention is m ade also o f atm an),
101 Cf. for exam ple F . S t a a l , E xploring m ysticism (Berkeley, 1975), p. 39 sq.
102 For th e principle of excluded m iddle, see e.g. MMK ii. 8, 15 (cf. x v i. 8, x x i. 14
and x x v . 1516); S u n yatasap tati 72. See also Candrakirti, P rasannapada v. 6, x v . 7
and xxiii. 14.
Concerning th e special use o f a form ula corresponding to th e (unnegated) fourth
kotinam ely th e neither . . . nor form ulain connexion w ith u ltim ate reality
w hich is no en tity o f any kind (e.g. in MMK x x v . 10, where n irvan a is said to be
neither en tity nor absence o f en tity ), see our discussion in J IP 5 (1977), pp. 1620.
103 See e .g . MMK v. 6, vii. 30, viii. 7, x x i. 10, and x x v . 14. Cf. MMK x x v ii. 22 and
42 David Seyfort Ruegg The Literature of the Madhyamaka School

In view of his radical critique of conceptual entities and categories as

grounded in the discursivity of thinking and language, and because of his
m ethod of argument based on pra^an^a-reasoning which points out the unten-
ability of a view without implicitly postulating the counter-view, N agarjunas
theory in his scholastic works can no doubt be properly described as largely
negative, as already observed above. B ut this carefully worked out via negativa
is not nihilistic, nihilism being the extreme of the ucchedavda which Ngrjuna
and his followers have avoided as strictly as its opposite, the etem alist extreme
of the vatavada. Nor can his procedure be dismissed as mere sophistry, hair
splitting and fault-finding contentiousness. Its rationale lies on the one hand
in a philosophically elaborated theory of the non-substantiality of any entity
posited in the dichotomous structure of discursive thinking and language, and
on the other hand in the principle th a t ultim ate reality (tattva) cannot be
expressed or conceived in the frame of vikalpa and prapaca (xviii. 9). For the
same reasons N agarjunas procedure cannot be described as anti-philosophical
even if he maintains no propositional thesis (pratijd) intended to establish a
speculative metaphysical system.

Of fundamental importance in N agarjunas philosophy are the two truth-

levels, th a t of worldly surface-convention (lokasamvrti) and th at which is true
ultim ately (satyam paramrthatah, xxiv. 8). The latter is of course the sole
truth , corresponding as it does to reality (tattva) characterized as still (anta) ,
without discursive development and free from dichotomizing conceptualization
and multiplicity (xviii. 9; cf. xxv. 24). Yet the relative level is not simply to
be dispensed with, for it is on the basis of transactional usage (vyavahara) th a t
the paramdrtha is indicated (deyate; xxiv. 10); in order to penetrate the pro
fundity of the teaching, therefore, one m ust understand the distinction between
the two satyas (xxiv. 9). The proper use of vyavahdra indeed makes it possible
to deal with the factors of relative existence and philosophical analysis, such
as the four truths (<ryasatya, xxiv. 13), the three jewels of the buddha, dharma
and samgha, and the eight kinds of persons who cultivate the path (xxiv. 45 ab),
all of which are recognized by the Buddhist tradition. I t would be altogether
m istaken to suppose th a t nyatd destroys (pratibdh-) these factors along
with the existence of karmic fruits, dharma and adharma, and the totality of
worldly usages (xxiv. 5 cd 6).104 And since nyatd could harm an unintelligent
person who has incorrectly grasped it (xxiv. II ) ,105 in order not to be ruined
one m ust understand nyatd according to the principle of dependent origina
25; Candrakirti, P rasannapada i. 7, x v . 7, and xviii. 6; D . S e y f o r t R u e g g , J I P 5
(1977), pp. 18, 4 8 - 4 9 , 5455.
104 The unconditioned (asam skrta) as well as th e conditioned dharmas and factors
belonging to th e surface level o f transactional usage are accordingly n o t w ith o u t a
certain as i f valid ity , although th ey are in reality (param rthatah) all w ith ou t ow n
being (cf. H . V aihingers Als ob th eory o f fictions). This does n o t in volve a rejection
o f sam vrti and vya va h d ra : the phenom ena are indeed saved .
105 Cf. R atn a va li ii. 2022.
The Early Period : The Formation of the Madhyamaka School 43

tion and non-substantiality in its non-nihilistic sense along with the motive
(prayojana) determining the teaching of emptiness (i.e. the stilling of all
discursive development, according to Candrakirti) (xxiv. 7).
N agarjunas philosophy is conceived with the purpose of revealing the con
vergence of unyatd, pratltyamutpdda and the Middle W ay (madhyama prati-
p a t).10'
As set out in the MMK, the idea of origination in conditioned dependence
(pratityasamutpada) is two-faceted since it evidently relates to both truth-
levels, th a t of the surface (samvrii) and transactional usage (vyavahdra) and
th a t of ultim ate reality (paramartha) . Inasmuch as origination in dependence
embraces all the conditioned factors of the world, it clearly pertains to the
transactional level of surface tru th .107 Now, since whatever dharma originates
in dependence on another is neither identical with it nor different from it
(xviii. 10), it m ight be said th a t for Nagarjuna this level is both unamenable
to ontological construction and antinomicand perhaps even th a t it is char
acterized by causal indeterminism.108 Pratityasamutpada does not then imply,
as it did for the Abhidharmikas whom Nagarjuna criticizes, th a t a dharma
defined as a factor bearing its own specific characteristic (svalaksana) is born
from other such dharmas known as its cause(s) and condition(s).109 Rather, it
refers to interdependent dharmas in a field of conditionship which have trans
actional validity on the surface level, but whose nature consists precisely in
being non-substantial and hence em pty of own being in virtue of the fact of
their dependent origination: they are as it were ciphers having no independent
ontological status as positive, negative, both positive and negative, or in
determinate entities.
W hat arises in conditioned dependence is therefore in the last analysis still
by n atu re (6antam svabhdvatah, vii. 16). Now, looked a t in this manner,
pratityasamutpada pertains to the fact or tru th of dependent origination, i.e.
to the reality of all conditioned dharmas. And as the true state of affairs con
cerning these dharmas, which are thus empty of own being, pratityasamutpada

106 MMK x x iv . 18, and V igrahavyvartan 22 w ith V rtti on 70. Cf. MMK x x iv . 36,
and Candrakirti, P rasannapada x x iv . 40.
107 Cf. Candrakirtis expressions sm vrta-pratityasam utpada (P P i, p. 1011) and
laukika-tattvalaksana (P P x v iii. 10).
108 See Candrakirtis remarks in P P i, pp. 9, 26, and M adhyam akvatra vi
(discussed below , pp. 7278). This indeterm inism is referred to b y N agarjuna as
stillness b y nature in MMK vii. 16 quoted below (compare th e expressions
prakrlinta and p ra k rtivivik ta ). Compare also R atn va li i. 47, where it is stated th a t
production (u tpattiy i.e . o f a bhdva) is understood neither in term s o f reality (tat-
tvatas) nor even in term s o f designation (p r a jn a p ti).
100 Cf. Candrakirti, P P x v . 2 (p. 261. 6) on the samvpta-svarUpa o f th e Abhidharma.
For the definition o f a dharm a as bearing a specific characteristic (svalaksaTm-
dh arana), see V asubandhu, Abhidharm akoabhya i. 2 (ed. Pradhan, p. 2. 9). In
th e Abhidharm a, dharm as are regularly determ ined w ith reference to their charac
teristic natures.
44 David Seyfort Ruegg The Literature of the Madhyamaka School

may then be termed emptiness (dunyatd, xxiv. 18).110 In the introductory verses
to the MMK the pratUyasamutpdda is moreover described as not annihilated
and not eternal, undestroyed and unborn (just like nirvana in xxv. 3), and as
w ithout multiplicity and tranquil (just like ultim ate reality, tattva, in xviii. 9).
The complexity of the idea of pratityasamutpada is further brought out when
Nagarjuna states th a t he who recognizes it sees not only the characteristic
feature and the conditioned processes of the worldviz. suffering* (duhkha)111
and origination (samudaya) respectively, i.e. the first two aryasatyasbut also
unconditioned cessation (nirodha), the third truth, and the path (mdrga), the
fourth tru th th a t as it were leads from the first two truths to the third (cf.
xxiv. 40).
I t would seem th a t the above-mentioned ambivalence of origination in
dependence as pertaining either to the dependently produced conditioned fac
tors (pratityasamutpannd dharmah) or to the fact (or truth) of dependent
origination (pratUyasamutpdda) is paralleled in the terminology used to
express the theory of emptiness. Thus, as already noted, the term unya
empty* is an epithet of all dharmas; and unyatva is then the state or property
of emptiness of all dharmas.112 On the other hand, the term sunyatd tends to be
reserved for this fact or true state of affairs; and it is accordingly a term th a t
110 The tw o aspects o f pratU yasam utpdda m ight perhaps he clarified b y saying
th a t as including (extensionally) all conditioned factors originating in dependence
(pratitya sa m u tp an n a ) whose nature is to be e m p ty (: unyatva), pratityasam u tpada
belongs to th e sa m vrti level. B u t as the fact or true state o f affairs o f dependent
origination th a t relates (intensionally) to all conditioned factors, pratityasam u tpada
concerns the u ltim ate reality or truth o f their em ptiness (unyatd) ; and it accord -
in gly pertains to param drtha (cf. P P vii. 15, p. 159. 6). (Because th e asam skrta,
which is param drthika, is also svabhdvaunya, unya and sam skrta are o f course not
coterm inous). Since pratityasam u tpada thus pertains, in the one w ay or the other,
to both th e sam vrta and param drthika levels, it is equivalent in its range to unyatd.
The tw o term s are n o t exact synonym s, how ever, and th ey have distinct spheres o f
use even if essen tially th ey have a com m on (nominalistic) reference (u p ad aya
p ra jh a ptih ).Compare th e old controversy as to w hether pratityasam u tpada is
sam skrta or asam skrta. See also A bhidharm akosabhasya iii. 28. The old canon
already attests th e application of the term dharm atd to th e principle o f pratityasa-
m utpdda com prising th e tw elve m em bers o f origination in dependence; see S am yu t-
tan ik aya ii, pp. 2526, and N id an asam yu k ta (ed. C. T r ip a th i, Berlin, 1962),
pp. 1479, 164 (cf. Candrakirti, P P i, p. 3940), where this description is found
w ith the k ey form ula . . . utpadad vd tathdgatdnam anutpadad vd sthitd eveyam
dharm atd dharm asthitaye dhatuh. Then in th e D asabhum ikasutra (8 G = p. 136) this
dharm atd is described as sarvadharmaunyata (and sarvadharm anupalabdhi). The
m eaning o f th e word dharmatd has o f course developed during the history o f
B ud d h ist th ou gh t, and it is in th e M ahay ana th a t th e link betw een unyatd and
dharm atd has been worked o u t ; b u t th e connexion betw een pratityasam u tpada and
unyatd is found already e.g . in th e N id an asam yu kta, p. 153.
111 Cf. P P x x iv . 1.
112 See e.g . MMK x xv ii. 29; V igrahavyavartani 2122, 57, 59, 61 w ith V rtti.
E m p ty is o f course n ot to be understood here as th e com plem entary opposite o f
n o t e m p ty in th e binary set unya|aunya (see MMK xiii. 7; cf. vii. 33 on sam skrta/
asam skrta as a dichotom ous pair).
The Early Period: The Formation of the Madhyamaka School 45

relates to the paramdrtha (although as a name it belongs, like any designation,

to the transactional surface-level of samvrti).113
The twin principles of pratityasamutpada and unyata thus found a philo
sophical Middle W ay th a t eschews both the extremes of annihilationism
(ucchedavada) and eternalism (aJvatavada) . The Madhyamaka takes account
of phenom enathe manifoldness of dharmas on the samvrti leveland reality
the pararnarthavtYnlQ refraining from presenting them as opposed factors. And
instead of postulating either a pluralism of entities or a single monistic entity
it operates with the principle of non-duality.
W hat has been tentatively referred to above as the causal indeterminism of
the factors within the process to which the tru th of pratityasamutpada applies
is inseparable from unyatd both as emptiness of own being making everything
possible114 and as release from (or: expeller of) all speculative dogmatic views
(xiii. 8). And the stillness in questioYi is characteristic of both paramdrtha
(gnoseologically speaking) and nirvana (soteriologically speaking).
B oth gnoseological and soteriological aspects are combined also in pratitya-
samutpada because, when ignorance (avidya misknowledge, the first of the
twelve members of origination in dependence) has come to a stop, there is no
more possibility for the conditioning factors (samskdra, the second member of
this set) to arise (xxiii. 23). This stoppage of avidya is achieved through gnosis
(jndna)y as a result of meditative realization (of pratityasamutpada) (xxvi. 11).
Cessation of origination or birth may be said to result, according to Nagar-
juna, also from the cessation of the notions of self (aham[kdra]) and every
thing belonging to self (mama) which lead to appropriative clinging (upadana,
xviii. 3-^4). Release thus follows from the cessation of action (karman) and
the defilements (kleAa) issuing from dichotomizing conceptualization (vikalpa),

113 In th e V igrahavyavartanl-V rtti 22, pratityasam utpan n alva is given as th e

reason for a th in g s h avin g no own being (n ih svabh dva), w hich in turn is given as
th e reason for a th in gs being em p ty (unya) in W - V r t t i 21. In P rasannapada
x x v ii. 29, Candrakirti has given pratU yasam idpannatva as th e reason for unyatva;
cf. x x iv . 19 (p. 505.16) on pratityasam u ipann a and unya.Concerning the difference
in use betw een unyata and u n yaftva) , it is true th a t inconsistencies are encountered
in our te x ts, where Skt. unya has on occasion been translated b y Tib. ston p a n id
instead o f ston p a (exam ples are to be found in P P x x iv . 1, p. 475. 6, 8; conversely,
in th e Sanskrit te x t of P P xx iii. 13 [p. 461. 1516] we find svabhava^unyatva where
th e T ibetan has ran bzin g yis ston pa). A nd it is also true th a t th e situ ation is
further com plicated b y th e fact th a t Tib. ston p a n id renders both unyata and
unyatva. Y et, as a rule, th e T ibetan exegetes h ave distinguished betw een unya =
ston p a as an adjective m odifying all dharrtias (ch os)t and ston p a n id ( = u n yata).
(In a case where the M adhyantavibhaga i. 13 refers to unya w hen unyata is th e
topic in question, Sthiram ati explains in his com m entary th a t th e suffix o f state
[bh dvapratyaya] has been elided [lu p ta ] m etri causa). According to th e Prajnapara-
m ita literature and the M adhyam aka, unyata, is itself em p ty o f own being (and
Sthiram ati has pointed ou t in the sam e place th a t unyata is no more different from
w h a t is em p ty th an are duhkhatd from w hat is painful or an ityata from w h at is
im perm anent, for it is th e dharm ata o f dharmas).
114 MMK x x iv . 14 and V igrahavyavartani 70; cf. MMK x x iv . 36.
46 David Seyfort Ruegg * The Literature of the Madhyamaka School

which in its tu rn proceeds from discursive development (prapaca). Prapaca

itself comes to a stop in unyat.115 Release is, however, not to be reified as one
term in a binary set made up of bondage/release; for bandhana/moksa belongs
like sunya/asunya, samskrta/asamskrta or any other set of dichotomously
opposed complementary concepts or categoriesto the realm of vikalpa.
On the level of paramdrtha, there is neither superaddition of nirvna nor
removal of samsdra as opposed entities (xvi. 10). Furthermore, the state of con
tinuous interrelated and dependent coming-and-going (<djavamjambhdva, in the
round of existences) comes, once stoppage of such interrelatedness and depend
ence is achieved, to be designated as nirvna (xxv. 9). In this way there is really
no differentiating property116 between samsdra and nirvana; and there is no
difference, however subtle, between a term inal limit of nirvna and a terminal
limit of samsra (xxv. 1920).117 Similarly, the tathgata and the world of
living beings (jagat) have the same nature (svabhdva), i.e. both are precisely
w ithout own being (nihsvabhdva, xxii. 16). Nirvnawhich as already observed
is not an entity to be posited in terms of the four positions (xxv. 4)is also said
to be free from elimination and acquisition, and to be undestroyed and non-
eternal, neither stopped nor produced (xxv. 3), so th a t it has been defined
in much the same way as th e pratityasamutpada in the introductory verses to
the MMK.
The characteristic of reality (tattva-laksana) is then to be free from dichoto
mizing conceptualization and, accordingly, to be without multiplicity, still and
undeveloped in discursive development (xviii. 9; cf. xxii. 15 and xxv. 24). On
this point too N agarjunas soteriology is closely linked with gnoseology. Reality,
the level of paramdrtha which is simply shownand as it were plotted through
philosophical analysis of the zeroed dharmasis never hypostatized as an
entity of any kind whatsoever.
As noted above, terms to designate reality such as dharmadhdtuy dharmakdya
and tathatd employed elsewhere in the Mahynist literature of the Sutras and
stras, including even some works connected with the Madhyamaka school,
do not occur in the MMK. To indicate reality N grjuna has there confined
himselfapart from the well-established words nirvna and nyatto the
terms paramdrtha (xxiv. 8, 10), tattva (xviii. 9 and xxiv. 9) and dharmatd
116 n ya ty m y MMK xviii. 5. (There is th e variant sto p a id k yis = n yatay
b y em p tin ess). Candrakirti here explains prapaca in term s o f objectification
(upalam bha) o f a vastu (P P xviii. 5, p. 350. 18); cf. P P x x ii. 15 (vastunibandhand hi
prapacdh) , x x v . 24 (where prapaca is glossed as n im itta )y xv iii. 5 (p. 350. 15),
and i, p. 11. G (where prapaca is said to h ave th e characteristic o f a designatum ,
e t c .: abhidheyadilaksana). In MMK x x v . 24 uplam bhopaama and prapacopaam a
figure side b y side. B u t in P P xviii. 9 (p. 373. 9) prapa ca is said to be vdcy w liich
d evelops arthas; and w hat is undeveloped in discursive d evelopm ent is unexpressed
b y words (prapa cair aprapacitam vagbhir avyahrtam ) .
118 viesana (Tib. khyad par) = viesa difference according to Candrakirti.
117 Candrakirti speaks o f th e im aginary construction o f a prior and posterior
lim it (pu rvparakotikalpan) , which is in fa ct im possible (P P x x v . 21 avataranikd;
cf. xi. 1).
The Early Period: The Formation of the Madhyamaka School 47

(xviii. 7).118 To be recalled here also is Nagarjuna's rejection of an asamsbrta

opposed within dichotomizing conceptualization to the samskrta (vii. 33), of a
unya opposed to the aAunya (xiii. 7), and even of absence of defilement opposed
to the state of defilement (cf. xxiii. 25).

The philosophy of the MMK and the related treatises differs from th a t of the
Abhidharmikas in th a t Nagarjuna moved from the taxonomic and constructive
inventorization of a plurality of dharmas conceived of as individually specifiable
and analytically independent entities, with causal laws connecting them, to a
theory of dharmas operating as non-substantial factors arising in the functional
interdependence of conditionship. This theory is described by Nagarjuna in
terms of dependent origination, absence of own being and emptiness. Unlike
th a t of so m any other philosophers, N agarjunas philosophical theory is not
made up basically of a system of propositional judgements with accompanying
categorical and hypothetical syllogisms; and it takes on not only a dialectical
but a structural character, its terms being conceptually, logically and lin
guistically interrelatedbut nevertheless metaphysically non-substantial
dharmas. The name of the basic principle of the theoryunyatais itself non-
referential, for Nagarjuna refrains from positing an absolute in the frame of any
conceptual position th a t would involve its hypostatization as a positive, nega
tive or indeterminate entity.
Nagarjuna was thus engaged in a radical rethinking of the philosophical
endeavour, th a t is, of the very idea of philosophy and the terms in which it is to
be pursued. And by turning away from the construction of a speculative doc
trine involving the postulation of entities having some kind of self-nature he
clearly sought to keep strictly to the Middle W ay indicated by the Buddha in
the only manner he found commensurate with it.
I t was left to N agarjunas Madhyamika successors to elaborate his philo
sophy and pursue in particular the refutation of the speculative views and
dogmas of both non-Buddhist and Buddhist philosophers (his pupil Aryadeva
was especially renowned for excelling in this task), and then to apply their
m aster's theory by undertaking the systematization of Madhyamaka philo
sophy in its three main schools (those of Bhavaviveka, Buddhapalita/Can-
drakirti and Santaraksita).

Commentaries on Nagarjuna's works

Commentaries on N agarjunas theoretical scholastic works, and in particular
on the MMK, are numerous. The *Akutobhaya was traditionally regarded as
118 T he term s dharm adhdtu and tathatd do appear in Candrakirtis Prasannapada,
in quotations and discussions o f M ahayanist doctrine. On th e usage o f the hym ns
ascribed to N agarjuna see above. (On th e absence o f dharm akdya and dharmadhdtu
in early Chinese versions o f the Prajhaparam ita-Sutra see L . L a n c a s t e r , E astern
B u d d h ist 8 [1975], p. 36 sq. A nd on th e usage o f th e MMK see R . H . R o b i n s o n ,
E arly M adhyam ika, p. 63.)
48 David Seyfort Ruegg The Literature of the Madhyamaka School

Nagarjunas own commentary on his MMK;118 but for some time this ascription
has been questioned.120 I t has been observed th at a close connexion exists be
tween this work and Aryadeva.121 I t has also been suggested th a t it is related
to the commentary, available only in Chinese (Taisho 1564), attributed to a
certain Ching-mu122, an Indian master who was known in addition by a name
th a t appears in Chinese transcription as Pin-lo-chieh but whose identity re
mains to be elucidated.123 This Pin-lo-chieh has even been identified with
Aryadeva.124 At all events both the *Akutobhaya and Ching-mus commentary
show familiarity with Aryadevas work. Ching-mu has also sometimes been
credited with the commentary preserved in Chinese on N agarjunas Shih-erh-
men-lun, a work otherwise ascribed to Nagarjuna himself.125

119 A German translation o f this w ork was published b y M. W a l l e s e r , D ie

m ittlere Lehre des N grjuna, nach der tibetanischen Version bertragen (Die
buddhistische Philosophie in ihrer geschichtlichen E ntw icklung, 2. Teil, H eidel-
ber, 1911).
120 B ecause th e *A kutobhay (chapter x x v ii, fol. 113 a 2) quotes a verse (n an p a
po dan m nan by a d an / \sm ra po byun ba Sin tu dkonj /de p h y ir mdor na khor ba n i/
Imtha yod m a y in m tha med m in \\) th a t is found in A ryadevas C atuhataka (vii. 5).
(Cf. m K h a sgrubdG eleg sdpal'bzah, sToh*thun*chen*mo, fol. 37 b 38 a; E . O b e r -
m i l l e r , AO 11 [1933], p. 4 note 9.) The ID an 'd k arm a Catalogue lists th e *Akuto-
bhay after b oth B h v a v iv ek a s and B u dd h ap litas com m entaries on th e MMK.
Moreover, the m ain earlier Indian com m entariesB u d d hap litas V rtti, B h v a v i
v ek a s Prajnpradpa and Candrakirtis Prasannapaddo n o t appear to m ention
th e *A kutobhay. (At th e sam e tim e, there ex ist num erous sim ilitarities betw een
th e T ibetan version o f th e com m entary o f B ud d h aplita on the MMK and
th e *A ku tobh aya). On th e relation o f th e *A kutobhay to th e Chung-lun
and references to it in th e Chinese tradition, see M. W a l l e s e r , D ie m ittlere Lehre
des N grjuna nach der chinesischen Version bertragen, D ie buddhistische P h ilo
sophie in ihrer geschichtlichen E ntw icklung, 3. Teil (H eidelberg, 1912), p. ix sq. On
its a u th en ticity see also C. I k e d a , Anesaki C om m em oration V olum e (Tokyo, 1934),
pp. 2913.
121 See J . W . d e J o n g , AM 17 (1971), p. 109.
122 Cf. . L a m o t t e , T rait, iii, p. 1373; J. W . d e J o n g , loc. cit.A German
translation o f this work was published b y M. W a l l e s e r , D ie m ittlere Lehre des
N grjuna, D ie buddhistische Philosophie (see above, n. 120), 3. Teil (H eidelberg,
1912). Cf, R . A. G a r d , IB K 3/1 (1954), pp. 3 7 6 -3 7 0 .
123 The Chinese transcription P in-lo-chieh has been interpreted as Pingalka
( = Pingalanetra, Pingalacaksus) ; see . L a m o t t e , Trait, iii, p. 1 3 7 3 . Ching-m u
w ould correspond in m eaning to som ething like N ilanetra. Cf. R . H . R o b i n s o n ,
E arly M dhyam ika, p. 2 9 , who gives Vim alka; N . P r i , B E F E O 1 9 1 1 , p. 3 6 6
n o te; M. W a l l e s e r , D ie m ittlere Lehre, 3 . Teil, p. x sq. (It m a y be noted th a t a
Pingala o f th e B rahm anical tradition, th e supposed author o f th e Chandahstra,
was regarded as a ngarja; cf. A. W e b e r , Indische Studien [Berlin, 1 8 6 3 ] ,
p. 1 5 7 sq.).
124 . L a m o t t e , Trait, iii, pp. x x x ix , 1373. Cf. P . D e m i v i l d e in L. R e n o u et
J. F i k l i o z a t , L Inde classique, ii (Paris, 1953), 2137. 1. The id en tity o f Aryadeva
and Ching-m u w as rejected b y G. T u c c i, P re-D innga B u d d h ist tex ts on logic
(Baroda, 1929), p. xvii.
125 Cf. R . A. G a r d , I B K 2/2 (1954), pp. 747742; P. D e m i v i l b e , op. cit.,
2137. 2; N . A i y a s w a m i S a s t r i , V isvabharati A nnals 6 (1954), p. 166.
The Early Period : The Formation of the Madhyamaka School 49

For the developed Madhyamaka school in its two main classical bran
ches the commentaries on the MMK by Buddhapalita and his follower
Candrakirti and the one by Bhvaviveka are of fundamental importance
(see below).
I t is to be noted th a t among the earlier commentaries on Ngrjunas w rit
ings there are some by im portant masters of the Yogcrin/Vijnnavdin
school. A portion of a commentary ascribed to Asanga on the beginning of the
MMK is preserved in Chinese (Taish 1565, translated in 543); it refers to
Rahulabhadra and comments also on the preliminary stanzas to the MMK con
cerned with the eight negative epithets applied to prattyasamutpda.l2t
A commentary by Sthiram ati is also preserved in Chinese (Taish 1567, tran s
lated about 1000); it evidently knows Bhvavivekas commentary on the
MMK.127 In addition, a commentary by Sthiram atfs m aster Gunamati which
is no longer extant is known to tradition; it seems to have been known to
Bhvaviveka. The existence of such commentaries on the MMK by leading
authorities of the Vijnnavda clearly indicates th at Ngrjunas work was
not considered to be the exclusive property of the Mdhyamikas in the narrow
sense of a particular school, and th a t it was regarded as fundamental by
Mahynist thinkers of more than one tendency.
Commentaries on the MMK are also reported to have been composed by
Devaarman,128 Gunarl129 and Rhula(bhadra).130
There are commentaries on the nyatsaptati and the Y uktisastik by
Candrakirti, one by Ajifcamitra on the R atnvali (above, p. 24), one by Mah-
m ati on the Suhrllekha (above, p. 26), and one by a certain M aitreyantha on
the Bhavasam krnti (above, p. 29).
Oh the hymns ascribed to Ngrjuna, although some of them are referred to
and quoted for example by Candrakirti, the only commentary available in
Sanskrit is one by A m rtkara on the L oktta0, the N iraupam ya0, the A cintya0
and the Param artha-stava.131

128 Cf. P. D e m t v l l e e , L T n d e classique, 2138.

127 Cf. H . N a k a m u r a , Journal o f Intercultural Studies no. 4 (1977), p. 7980;
Y. K a j i y a m a , W ZKSO 1213 (1968), p. 198. (The Taisho edition contains only
Chapters ixiii).
128 Deva^arman is referred to b y A valok itavrata in his Tika on B h a v a v iv e k a s
Prajnapradipa, chapter i (wa, fol. 225a 226b), where the title o f his com m entary
is given as dKar*po*char*ba.
128 H e is referred to b y A valokitavrata, op. cit., fol. 8 5 a .For a list o f eight
com m entariesthose o f N agarjuna him self (i.e. the*A kutobhaya), B u d d h ap alita,
Candrakirti, Deva^arman, Gunasri, G unam ati, Sthiram ati, and B h a va viv ek asee
th e colophon to th e *A kutobhaya (fol. 114a). (In his Bodhim argadipapanjika, fol.
324b, Dipamkara^rijnana has su bstitu ted a com m entary b y G unadatta for D ev a-
armans, which he stated is on th e Prajnapradipa o f B h avya).
130 H is com m entary is said to h ave been translated into Chinese b y Param artha
(500569); see . L a m o t t e , Traite, iii, p. 1374.
131 See G. T u c c i, Minor B u ddhist tex ts, i (Rom e, 1956), pp. 23546.
60 David Seyfort Ruegg The Literature of the Madhyamaka School

The m aster usually considered to be next in importance after N grjuna in
the early undivided Madhyamaka school is ryadeva, whom the sources
represent as a direct disciple of Ngrjuna.132 He was especially famed for his
skill in argument, notably against the representatives of schools such as the
Smkhya and Vaiesika concerning whose doctrines a t this time his work
provides im portant information.133
According to Candrakrti, ryadeva was born in the 'island (glin) of Snga
la (Simhaladvpa) as the son of a king ; but he renounced his station as prince,
left the world and came to South India where he attended on Ngrjuna as his
m aster.134 This Simhaladvpa has often been identified with ri L ank135a
location which has been considered confirmed by references in the Dlpa-
vamsa (xxii. 41 and 50) and Mahvamsa (xxxvi. 29) to a certain Deva, whom
these sources place in the second half of the third century when the
Vetullavda was wide-spread in r Lank; but this identification has been put
in question.18
The commentary on the M adhyamakastra ascribed to Pin-lo-chieh has, as
already mentioned, been considered an authentic work by ryadeva since
Pigala appears as an epithet of his.137 B ut this attribution has been ques
ryadeva has traditionally been credited with the authorship of the ata(ka)-
stra (Pai/Po-lun) extant only in K um rajvas Chinese translation of the
early fifth century (Taish 1569).139 The problem of the connexion between
this work and the Catubataka (available in Sanskrit fragments and in a Tibetan
is* A ryadeva is said to h ave been known also under the nam es K nadeva,
N ilanetra, Pigalanetra, etc. (cf. . L a m o t t e , Trait, iii, p. 1373 n o te; above, p. 4 8 ) .
As for his date, it has to be determ ined in relation to N grjunas, w hose direct
disciple he is considered to be. There is a quotation from A ryadevas Catuhataka
(ii. 8) in Chapter cv o f H arivarm ans T attva/Satyasiddhistra (cf. th e Sanskrit
translation by N . A i y a s w a m i S a s t r i , Baroda, 1975, p. 252).
133 See M. H o n d a , IB K 23/1 (1974), pp. 712, and E . F r a t t w a i / l n e r , W ZKSO 2
(1958), p. 131 for th e Sm khya; G. T u c ci, P re-D inga B u ddhist te x ts on logic,
pp. x iv x x x for th e Vaiesika and N y y a.
134 Candrakrti, C atuhataka-tk, fol. 34b. Cf. T. W a t t e r s , On Y uan C hw angs
travels in India (London, 19045), i, p. 3201; ii, p. 200sq.
135 . L a m o t t e , Trait, iii, p. 1373 note. Cf. V idhushekhara B h a t t a c h a r y a , The
Catuhataka o f A ryadeva (Calcutta, 1931), p. x ix .
138 See N . D u t t , IH Q 10 (1934), pp. 13742, w ho places Sim hapura in N orth w est
India. P . S. S a s t r i , IH Q 31 (1955), p. 196 sq., takes Andhra to be A ryadevas
137 . L a m o t t e , Trait, iii, p. 1373.
138 See J. W . d e J o n g , AM 17 (1971), p. 109.
139 For th e Sata(ka)stra see G. Tuccr, P re-D inga B u d d hist te x ts on logic
(Baroda, 1929), p. x iii sq. Cf. G. T u c c i, La versione ciese del Catuhataka di
A ryadeva confrontata col testo snscrito e la traduzione tibetan a, RSO 10 (1923),
pp. 522.This te x t seem s to be listed in th e lD a n d k a rm a Catalogue (ed. M. L alou,
J A 1953, p. 335, no. 677) ?
The E arly Period : The F orm ation of the M adhyam aka School 51

translation as well as in a partial Chinese version)140 is complicated. The two

texts have much in common as to their contents, but they are not identical
either in the total number of their verses or with respect to the treatm ent of
their subject-m atter ; and the authenticity of the shorter work has even been
queried.141 This uncertainty is compounded by the fact th a t the Catubsataka
is itself frequently referred to simply as Sataka, for example in Candraklrtis
Prasannapad. In his commentary on the Catubataka (fol. 34 b) Candrakrti
furthermore observes th a t the word ccUuh had sometimes been (tendentiously)
om itted; and he states (fol. 34a) th a t when Dharmapla, the master of the
Nland school of the Vijnnavda, commented on the work, he in fact broke
it up into two parts, one dealing with an exposition of the Dharm a and the
other with critical discussions of controversial topics, and then commented only
on the latter part. The second half of the Catubataka was translated into
Chinese together with Dharm apalas commentary in 650 by Hsan-tsang (Taish
The ten chapters (in 50 4lokas or 32-syllable units) of K um rajvas Chinese
version of the first half of the ata(ka)stra are : (i) Renunciation of sin and
merit, (ii) R efutation of the dtman, (iii) Refutation of unity, (iv) Refutation of
diversity, (v) Refutation of sense-perceptions, (vi) R efutation of their objects,
(vii) Refutation of the existence of the effect in the cause, (viii) Refutation of
the non-existence of the effect in the cause, (ix) Refutation of permanence, and
(x) Critique of the em pty. These chapters accordingly deal with approximately
the same subjects as th e 200 verses of the last eight chapters of the CS, though
in different words. This tex t is accompanied by a commentary ascribed to Vasu
(certain passages of which follow almost verbatim the Ta-chih-tu-lun).143
Aryadevas major work is the Catubsataka, which is available in fragments
in Sanskrit, in Hsan-tsangs above-mentioned Chinese translation of the
second part only, and in a Tibetan translation.144The work consists of the foliow-
140 See below.
141 See J . W . d e J o n g , AM 17 (1971), p . 110. Cf. R . G a r d , I B K 2 / 2 (1954),
pp. 7 5 1 -7 4 7 .
142 See . L a m o t t e , Trait, iii, p. 137172; G. T u cc i, R SO 10 (1923), p. 523 sq.,
and Pre-D innga B u ddhist te x ts on logic. Cf. R .H . R o b i n s o n , E arly M dhyamika,
pp. 27, 33; P. D e m t v t l l e , L Tnde classique, ii, 2137. 3.See also E . F r a u -
w a l l n e r , Candramati, in Studia Indologica (Festschrift W . K irfel, B onn, 1955),
p. 66 sq., on D harm apalas com m entary and th e Vaiseika.
145 On V asu see N . P r i , B E F E O 1911, p. 361 sq .; G. T u c c i, Pre-D innga
Buddhist te x ts on logic, p. x iv ; E . F r a u w a l l n e r , On the date of th e Buddhist Master
o f th e Law V ausubandhu (R om e, 1951), pp. 3538, 4849 (who identifies this Vasu
w ith his V asubandhu I) ; R . H . R o b i n s o n , E arly M dhyam ika, pp. 33, 211. On this
V asu s knowledge of th e Sm khya see E . F r a u w a l l n e r , W ZKSO 2 (1958), p. 131.
144 See H a r a p r a s a d S a s t r i , Memoirs o f th e A siatic S ociety o f B engal, vol. iii,
no. 8 (1914), pp. 449514; P . L. V a i d y a , tudes sur ryadeva et son Catuhataka
(Paris, 1923) ; G. T u c c i, R SO 10 (1923), pp. 52167 ; V. B h a t t a c h a r y a , Proceedings
and Transactions o f th e Fourth Oriental Conference, A llahabad (1926), pp. 83171,
and The Catuhsataka o f ryadeva (Calcutta, 1931) ; S. Y a m a g u c h i , Chgan bukkyo
ronk (Tokyo, 1965), p. 197 sq.
52 David Seyfort Ruegg The Literature of the Madhyamaka School

ing 16 chapters, (iiv) Elimination of the erroneous positing of things as per

m anent (nitya), pleasant (sukha), pure (ubha or uci),li5 and self (atman)
(according to Candrakirti these four chapters which dispel the four viparydsas
explain the nature of mundane things so th a t they may be abandoned and
buddhahood may be achieved), (v) The B odhisattvas practice (which makes
it practically possible to achieve Buddhahood). (vi) Elimination of the defile
ments (Iclea) which hinder the preceding, (vii) Elimination of attachm ent to
the enjoyment of seemingly desirable sensory objects (\saya), which causes
the defilements to arise and increase. And (viii) the practice of the disciple. The
first eight chapters of the C are thus concerned with the preparation of those
who practise the path. The last eight chapters then explain the non-substan
tiality of the dharmas. They deal in turn with the negation (pratisedha) of (ix)
perm anent entities, (x) self (atman) , (xi) time, (xii) dogmatic opinions (drsti),
(xiii) sense-faculties and their objects, (xiv) the positing of doctrinal extremes
(ianlagrdha, e.g. existence, non-existence, both, and neither) with special
reference to identity and difference, and (xv) the positing of conditioned
(samskrta) things as real. Finally chapter xvi, entitled An exposition of the
cultivation of ascertainment for master and disciple', is devoted to a con
sideration of logical and epistemological problems in the doctrine of nyatd.
In particular, it is pointed out (in conformity with Vigrahavyavartam 2930)
th a t he who does not m aintain a thesis (palcsa) based on the positions of
existence (sat), non-existence (asat), and both cannot be attacked in logic by
an opponent (xvi. 25).
The titles of the chapters of the C differ slightly in the Tibetan translations
of the basic tex t and of Candraklrtis Tika, the latter adding the word bsgom pa
(bhavand) where it is lacking in the former (e. g. Chapter xvi); this emphasis
on m editative realization (and on bhdvanasamadhi in the Tika version of
Chapter ix) accords with the reference to the Bodhisattvas yoga-practice in the
full title of the work, which in the bsTan*gyur version is Bodhisattvayoga-
carystra-Catubataka-Krik (the Sanskrit fragments have yogdcdra). I t is
noteworthy th a t the term yogdcdrayogacarya is here applied to a work dealing
largely with a form of bhavand directed towards philosophical analysis and the
critical treatm ent of controversial points. The use of this term is remarkable
since it came to be usually associated with the school of the Yogcrins/Vija
a vdins established subsequently by Asaga and appears already in the title of
one of their basic sources, the Yogcrabhmi (known in the bsTan*gyur version
as Yogacarybhmi). Some kind of close relation may well have existed between
Aryadeva and early masters of the Yogcrin school; for not only has the
H astavlaprakarana, a work by Dignga, been ascribed to him by an Indo-
Tibetan tradition but, as already mentioned, the Yogcrin Dharm apla wrote
a commentary on the second portion of the C as late as the sixth century.148
145 See e.g. R atn va l ii. 4870.
146 There also exist, as m entioned above, som e im portant com m entaries on
N grjunas MMK b y masters o f the Y ogcrin/V ijnavdin school.The Chinese
The Early Period: The Formation of the Madhyamaka School 53

Yet the reference in the title of the C to yogdcdra/yogacaryd could hardly have
been intended to express any specific connexion between it and the school of
the Yogcrins/Vijnavdins; and it appears th a t the word was then still
being used as a general term to denote practisers of the Buddhist spiritual and
intellectual disciplines without reference to a particular philosophical school.147
In addition to D harm aplas above-mentioned commentary on the second
p art of the C, there is a complete Tika by Candrakirti, now available only in
Tibetan. In it (fol. 35 a) reference is made to examples (dpt = drstanta) relating
to the first eight chapters which had been supplied earlier by Dharm adasa.148
The Aksaraataka and its V rtti have been ascribed to Ngrjuna by Dipam-
kararljna149 and the Tibetan tradition, but to Aryadeva by the Chinese.150
This work, which is mainly directed against Smkhya and Vaiesika type
theses, discusses inter alia the formal objections against establishing the
existence of entities by means of an inference (anumdna) comprising a thesis
and logical reason, as well as the epistemological difficulties involved in postu
lating their production, duration and destruction. In conventional trans
actional usage (vyavahara) entities (bhdva) are said to be comparable to a
dream; th a t is, they lack the characteristics of existence, non-existence and
both existence and non-existence, and they are also not uncaused because they
arise through effort. I t is quite possible to speak conventionally of the existence
of things, but not from the point of view of ultim ate re a lity ; and it is altogether
impossible to deduce the existence of a thing from its nam e (Vrtti, fol. 164 b
As for the H asta vlaprakarana and its Vrtti, although ascribed to Aryadeva
by an Indo-Tibetan tradition, they are probably by Dignga, who is named as
their author in the Chinese tradition.151
canon has preserved the above-m entioned com m entary on th e atak a b y the
B od h isattva V asu translated, together w ith th e basic tex t, by K um rajiva (Taish
1569); but it is uncertain w hether this person is the sam e as th e V ijnavdin
147 The fact th a t th e term M adhyam aka is still n o t used could perhaps be under
stood as indicating th a t this term also was not y e t in (regular) use to designate a
particular school o f thought.
149 A certain D harm adasa, clearly another V ijnavdin, is know n (along w ith
Dignga) as a m aster o f D harm apla; see Trantha, rGya^gar'chos^byun, p. 124,
and S u n v p a Y e ses'dpabbyor, dPag*bsam*ljon*bzan (ed. S. Ch. D as), pp. 99, 102.
Trantha has also m entioned D harm adsa as a direct pupil o f A saga and V a su
bandhu (p. 105), and as a teacher o f Arya V im uktisena (p. 108). (Is this th e D harm a
dsa m entioned as its author in a m anuscript o f th e V rtti on th e Cndra-Vy-
karana ?)
149 Bodhim rgadlpapajik, fol. 324b.
150 See V. V. G o k h a e e , A kara-atakam , th e hundred letters (Heidelberg, 1930),
w ho points ou t th a t th e introductory verse indicates th at th e com m entary was
composed b y a pupil o f (A rya-)D eva; S. Y a m a g u c h i , ta n i gakuh 9 (1930),
p. 191 sq.
151 See F . W . T h o m a s and H . U i, J R A S 1918, pp. 267310; E . F r a u w a l l n e r ,
54 David Seyfort Ruegg The Literature of the Madhyamaka School

The b sT airgyur also contains two very short works attributed to Aryadeva,
the *Skhalitapram athanayuktihetusiddhi152 and the *Madhyamakabhrama-
ghta. In the Chinese canon there is found a treatise attributed to the Bodhi-
sattva Aryadeva on explanations of nirvana given by tw enty sectarian teachers
mentioned in the Lankvatrastra (Taish 1640, translated in the first p art
of the 6th century by Bodhiruci).153 Another work ascribed to him is a refuta
tion of four H inaynist schools mentioned in the Lankavatrastra (Taish
1639, also translated by Bodhiruci).
On the Jnnasrasam uccaya, which m ay be by a later author named Arya-
deva, see below on the Madhyamaka-Vajrayna synthesis.

In the early Madhyamaka, in addition to the Stavas and Stotras ascribed to
Ngrjuna, the hymnic genre is represented by the Prajpram itstotra of
R ahulabhadra.154 A Tibetan tradition has taken him to be the master of Ngr
juna (perhaps owing to a confusion with a later Ngrjuna whose m aster Sa-
raha was otherwise known as Rhulabhadra). B ut according to other Tibetan
and the Chinese traditions, including one recorded by the Mdhyamika m aster
Chi-tsang (549623), he is regarded as a contemporary and follower of Ngr
juna. In other sources again he is regarded as a successor of Aryadeva.155 The

W ZK SO 3 (1959), pp. 12830, 1526; M. H a t t o r i , D ignga on perception

(Cambridge, Mass., 1968), p. 7.T h o m a s and T. R . V. M u r t i , Central philosophy
o f B uddhism (London, 1955), p. 94, ascribe the work to Aryadeva.
162 E nglish version b y R . W . C l a r k and L o z a n g J a m s p a b , T ibet Journal 4/2
(1979), p. 29 sq.
155 On this w ork see G. T u c c i, U n trait A ryadeva sur le N irvn a des
hrtiques, T P 24 (1926), p p . 1631; H . N a k a m u r a , H J A S 18 (1955), p p . 93102.
154 Ed. R . H i k a t a , Suvikxntavikrm i-Pariprcch Prajpram it-Stra, pp.
12. This 8totra is also found a t th e beginning o f the A tashasrik and P acavim -
atishasrik P rajnpram it (ed. N . D u t t ), as w ell as in part in the Ta-chih-tu-lun
(cf. L a m o t t e , Trait, ii, pp. 10615).
155 See B u'ston, C h o s'h y u n , fol. 9 9a ( O b e r m u / l e r , ii, p. 123) ; Gosgon'nu'dpal,
D eb*therson*po, ka, fol. 18 a ( R o e r i c h , i, p. 35) ; Trantha, rGya^garchos^byu
(ed. S c h i e f n e r ), pp. 53 sq., 83, and b K a -babs-bdun'gyi'rnanvthar, fol. 181 sq.
( G r n w e d e l , p. 10 sq.), w ho also makes R ahulabhadra th e m aster o f N grjuna.
B u t elsewhere in his rGya*gar*chos*byuri (pp. 68, 73) Trantha m akes R hulabha-
dra a successor o f N grjuna and a disciple o f Aryadeva. A nd B u'ston also quotes a
lineage in w hich R hua follows Aryadeva (Chos^byu, fol. 94 a 1 [ O b e r m i d l e r , ii,
p. 109, has Bhula, which is presum ably a m isreading or misprint for R hula]; cf.
Sum*pa*mkhan*po, d P a g bsan rljon bzan, (ed. S. Ch. D as), p. 90).Cf. G. T u c c i,
JP A S B 26 (1930), p. 141 ( = Opera minora, i, p. 212 sq.), and . L a m o t t e , Trait,
iii, p. 1374 note (who refers to the remark o f C hi-tsang th a t R hula w as a co n tem
porary o f N grjuna, and to the inform ation given b y Chan-jan [711782] according
to w hich R hulabhadra com posed a com m entary on N grjunas work). J . W. d e
J o n g , AM 17 (1971), p. 1078, seem s to incline to the view th a t R hulabhadra
preceded N grjuna. H . N a k a m u r a , A cta A siatica 1 (1960), p. 63, placed him in
th e third century, after Aryadeva.
The Early Period : The Formation of the Madhyamaka School 55

M adhyamakaastrastuti of Candrakirti speaks of Ngrjunas being accom

panied (sahita) by Rahulabhadra. Writings attributed to Rhulabhadra are
quoted in the Ta-chih-tu-lun (chapter xviii), as well as in Asagas commentary
on the M adhyamakastra (Taisho 1565) and *Sramatis *Mahynvat-
rastra (Taish 1634).156
In the Prajnpram itstotra the prajpdramitd is described as beyond
dichotomizing conceptualization (nirvikalpd) and discursive development
(nisprapacd) . Rahulabhadra writes : He who sees thee as thou really art (or :
devotedly)157 sees the tathdgcUa. The Buddhas are th y beloved sons, and thou
art thus the tender progenetrix of all Heroes devoted to the weal of others.
Though one, thou art praised under various appellations by the tathdgatas, who
thus make thee accessible to their disciples. Also, when thou a rt reached, faults
and disputes raised by polemists vanish. Thus thou terrifiest the foolish whilst
thou comfortest the wise. N ot coming from anywhere and not going anywhere,
thou a rt not made into an object of perception (npcdabhyase) by the wise;
however, those who see thee not but do truly (or: devotedly, bhavatah) take
recourse (prapad-) to thee are released as soon as they have done so : this is a
great marvel ! By the Protectors of the world (lokandtha) who, for the sake of
conventional communication (prajaptyartham) with incarnate beings, adopt
transactional linguistic usage (vyavahara) thou art expressed out of compas
sionwithout being [really thus] expressed. Indeed, who is able to praise thee
who art without phenomenal m ark (nim im ittd) and pure (nirajand), who
transcendest the entire realm of words, and who a rt not fixed anywhere ?
However, we who have praised thee who art not to be praised according to
convention (samvrti) are fulfilled through verbal expressions.
This hymn, besides containing ideas expressed also in some of the hymns
ascribed to Ngrjuna, is remarkable because it employs (though no doubt in a
different and still unsystematic way) terms such as prapad-, bhdvay daya, and
bhakti th a t were to become so usual in the works of the bhakti movements. As
for the conception of the prajpdramitdboth transcendent and immanent,
and the ground of both bondage and release for all beings whether they see
her or not (15)it is in certain respects comparable with the tathdgatagarbha/
tathagatadhatu theory of the Ratnagotravibhga Commentary.158 And the
question arises as to whether Rahulabhadra represented a related trend in
early M adhyamaka thought. He also composed a hymn to the Saddharma
pundarika,159 a Stra devoted in large part to the theory of the One Vehicle
(ekaydna) which is closely connected with the tathdgatagarbha doctrine of the
159 See . L a m o t t e , Trait, ii, p. 1060; iii, p. 1374; H . U i, Z II 6 (1928), p. 2 2 3 - 4 ;
G. T u c c i, JR A S 1930, p. 612 ( = Opera minora, i, p. 240).
157 bhdvena, really, tru ly or resolutely, d evotedly, affectionately.
169 R atnagotravibhga C om m entary i. 14952.
169 E d. W o g i h a r a and T s t j c h i d a in their edition o f th e Saddharmapundarika
(Tokyo, 19345), Introduction, pp. 3739. This hym n is included in th e Kashgar
Ms. o f th e Saddharm apundarika from th e Petrovskij collection (tentatively dated
to betw een the six th and ninth century).
56 David Seyfort Ruegg The Literature of the Madhyamaka School

Ratnagotravibhga (which in fact cites the Saddharmapundarlka in ii. 58).

Moreover, as already observed, he is quoted by *Sramati, the m aster of the
tcUhdgatagarbha doctrine and the Ratnagotravibhga, in his *Mahynvatra-
stra (Taish 1634).180
R hulabhadra thus represents a fairly distinct current in early Madhyamaka
thought th a t was not elaborated in the theoretical scholastic texts of the clas
sical school based on the MMK, but which is reflected in the hymns ascribed to
Ngrjuna. Still a commentary on the MMK by Rhulabhadra is said to have
been translated into Chinese by Paramrfcha (see above, p. 49) ; but this work
is not extant.

A certain N ghvayaliterally the one named Ngais mentioned in the
Lankavatrastra in the context of a prophecy (vydkarana) where he is
evidently identified with Ngrjuna himself.181 Nga is, however, elsewhere
spoken of together with Aryadeva as a pupil of N grjuna.162
Nghvaya has then been taken as a proper name and identified with a
certain T athgatabhadra by Trantha, who adds th a t he was the author of
hymns to the trilcdya163 and on the (tathgata) garbha1*1 and th a t the verses of
the Tathgatagarbhastra were thereupon widely sung in the South. This
master is thus linked with the so-called Vijnapti-M adhyamaka,165 a theory
associated with the tcUhdgatagarbha doctrine and the positive interpretation of
ultim ate reality as em pty only of the heterogeneous.166 The historical and
180 According to Trantha (rGya*gar*chos*byuh, p. 68), R hulabhadra received
th e m eaning o f th e (tathgata)garbha from A ryadeva in th e neighbourhood o f
K hc !
181 Lankvatrastra x. 1656 (where N gh va ya is translated into T ibetan as
K lu 'zesbod'pa) ; cf. Candrakirti, M adhyam akvatra vi. 3 (p. 76). See also Man-
jurmlakalpa liii. 44950 (ed. Ganapati stri, iii, p. 61617; K . P. Jayasw al
and R . Sarnkftyayana, Im perial H istory o f India [Lahore, 1934], p. 3536], where
N gh vaya is said to be an adept o f th e m yr vidy as w ell as nihsvabhdvdrthatat-
tvavit. Cf. G. T u c c i, J P A S B 26 (1930), p. 142 ( = Opera minora, i, p. 212) ; N. D u t t ,
IH Q 7 (1931), pp. 6 3 3 - 9 .
162 Trantha, rG yagarch os^ b yu n , p . 6 8 (where S c h i e f n e r s edition reads
K h rb os).
183 sK u gsum da'bstodp a (*T rikyastotra/stava or *K yatrayastotra/stava ?).
A hym n to the three kdyas has been published b y A. v o n S t a l - H o ls t e in , Bull, de
lA ca d m ie Im priale des Sciences de St. Ptersbourg, no. 11 (1911), pp. 83745;
this work is quoted b y N dapda (Nrop, Sekoddeatk, p. 5758) and is found at
th e beginning o f Gos*g2onn u d p als Deb*ther*snon*po. B u t in the bsTan-gyur a
K yatrayastotra th e first three verses of which correspond to th e first three of the
Sanskrit te x t just m entioned is ascribed to Ngrjuna him self.Compare the
*K yatrayvatra ascribed in th e bsT am gyur to N gam itra (K lu i'besgnen).
184 shin po'i bstod p a (cf. Sum*pa*mkhan*po, d P a gbsanrljon'bzan, ed. S. C. D as,
p. 90).
165 m a m rig gi dbu m a (Trantha, op. cit., p. 68).
188 gzan stoh; see above, p. 34. As already noticed, R hulabhadra is also linked
w ith this current o f th ou gh t (see above, p. 55).
The Early Period : The Formation of the Madhyamaka School 57

doctrinal context in which the information available from our sources is

embedded often evokes the later Madhyamaka-Vajrayana synthesis, rather
than the early period of the Madhyamaka school.167 Yet a t least certain of the
allusions appear to refer to an earlier time, indeed even to a period prior to the
author of the MMK.168 Taken together the available information points to
some (possibly quite early) doctrinal trends of considerable interest with which
certain currents in the Madhyamaka may have on occasion converged, although
in its classical period and going back as far as the MMK the pure Madhyamaka
school was characterized by apophaticism and appears to have had practically
nothing to do with the tcUhdgatagarbha doctrine.
Since the references in our sources to Naga remain obscure and the history
of this period of Buddhist thought is very imperfectly known to us, considerable
further research is required to untangle the skein and reveal the literary,
philosophical and religious movements th a t stand veiled behind these allu

Another figure whose name contains the word Naga and who is associated
with Nagarjuna is Nagabodhi. He is known to some sources as having lived on
the Srlparvata. I t is not certain whether this shadowy and elusive figure is the
same as the T antrika Nagabodhi, the disciple of the Tantrika Arya Nagarjuna-
pada (see below, p. 104).170

187 See Tranthas rGya*garchos*byu, pp. 6668.

188 A Sthavira nam ed N g a is connected w ith a certain cB h ad ra> (:M ahadeva)
and the five points (gzi Ia) at issue a t th e Council o f P^aliputra, and also w ith a
certain Yid*brtan*pa (S thiram ati/*Sthitam ati/*Saram ati ?) ; see B h vaviveka,
T arkajvl, fol. 162b163a ( = N ik yab hedavibhangavykhyna, u, fol. 179a 8).
Cf. . L a m o t t e , H istoire du bouddhism e indien, i (Louvain, 1958), p. 308; D .
S e y f o r t R t j e g g , L a thorie du tathgatagarbha et du gotra (Paris, 1969), p. 47.
188 For a brief discussion o f som e possible interpretations o f th e inform ation
available to u s on these persons and on th e doctrinal connexions see D . S e y f o r t
R u e g g , op. cit., pp. 4650.
170 I t has been suggested th a t N ga(bodhi) was th e author o f th e T a-chih-tu-lun
(see A. K . W a r d e r , In dian B uddhism [Delhi, 1970], p. 388).As for the Tantrika
N agabodhi, he is reported to have been the teacher of Vajrabodhi (c. 671741),
a m aster o f th e Tattvasam graha-T antra, and A m oghavajra (c. 705774). On him see
G .T u c c i, JP A S B 26 (1930), p. 142 ( = Opera minora, i, p. 212) ; Bu*ston, Chos*byim ,
fol. 102a (Obermiller, ii, p. 132); Gos^gommrdpal, D eb 'th ersn o n-p o , ja, fol.
4 b 5a (Roerich, i, p. 3601); Trantha, rGya^ar'chos^byuri, p. 6869.

By the middle of the sixth century the Madhyamika followers of Nagarjuna

had not only formed into a distinct school known as the Madhyamaka, but
they were dividing into two branches with Buddhapalita (c. 470540 ?) a t the
head of one and Bhavaviveka (Bhavya, c. 500570 ?) as the chief exponent of
the other.
Historians have used the term Prasangika to designate B uddhapalitas
school because it restricted itself to the prasanga-typz reasoning already used
by Nagarjuna, and the term Svatantrika to designate Bhavavivekas school
since it introduced independent (svatantra) inferences into its process of
reasoning. The terms Svatantrika and Prasarigika^-which do not seem to be
actually attested as the names of the two branches of the Madhyamaka in the
Sanskrit sources now accessible to us171correspond respectively to the Tibetan
terms Ran*rgyud*pa and Thal*gyur*ba, which have been regularly employed by
Tibetan scholars to designate the two sections of the pure Madhyamaka school.
In his commentary on the Tattvasam graha Kamalailla (eighth century)
speaks simply of undifferentiated Madhyamikas, while for the Vijnanavadins
he adopts the distinction between Sakaravadins and Nirakaravadins.17* In his
commentary on the Jnanasarasam uccaya (fol. 51b) Bodhibhadra (c. 1000)
mentions as the two schools of the Madhyamaka only Bhavavivekas and
Santaraksitas.173 And in the eleventh century Advayavajra divides the
Madhyamikas into Mayopamadvayavadins and Sarvadharm apratisthana-

171 In m ost o f our sources th e protagonists o f th e various branches are referred

to either by nam e or sim p ly b y general expressions such as eke aome'/apare others,
svayu th ya [a master] o f our own school, deary a m aster, etc. (The word prasangika
appears in another sense in Kamala^Uas T attvasam grahapanjika 3308 and Prajna-
k aram atis Bodhicaryavatarapanjika ix. 27 for exam ple, b u t n ot in Candrakirtis
17* See for exam ple Kamala^ilas T attvasam gragapanjika 191617. For a similar
term in ology cf. R atn ak araian ti, T riyan avyavasth an a, fol. 114 a sq. (kun rjob es
p a 'i rnam p a r sm ra ba and es p a m a m p a med p ar sm ra ba) ; see below , p. 122 sq.
173 In its T ibetan version Bodhibhadras Jnanasarasam uccayanibandhana d is
tin gu ish es, w ith regard to the analysis o f sa m vrti, betw een masters like B h a v y a
w ho do not take a cogn itive im age (abhdsa) as th e object o f v alid cognition fsnan
ba la m i ja l b a), and masters like Santarakita w ho hold th a t inner cognition alone
appears as various ( citra, vicitra) , whereas things so appearing are n ot real (snan
b a 'i dnos po j i Ita ba m a y in g y i nan gi Ses p a kho na sna 6hogs su snan bar sm ra ba)
(fol. 51b).
174 T attvaratnavall, pp. 14, 19 sq. (That is, th e school o f &,ntarak$ita and
The Middle Period : The Systematization of the Madhyamaka School 59

Among Tibetan scholars the appellations mDosde-pai-dbu-ma/mDo'sde*

spyod'pai'dbu'm a ( = Sautrntika-M adhyamaka) and rNal-byorspyod-pai*
dbu-ma ( = Yogacara-Madhyamaka) were current by the end of the eighth
century.175 They are found in two of the earliest known Tibetan doxo graphical
treatises, the lTa*baTm*parm airag written by dPal-brcegs (end of the
eighth century) and the lTa*bark h y a d par written by Ye*ses*sde (c. 800), two
renowned scholars and translators who worked during the first diffusion of
Buddhism in Tibet.176 These two appellations distinguish between the Mdhya-
mikas according to whether they m aintain or reject the existence of an external
object (bahyartha) on the surface-level (samvrti or vyavahara). On the other
hand the appellations Thal*gyur*ba Prsagika and Ramrgyud-pa =
Svatantrika seem to appear only somewhat later in Tibet, towards the beginning
of the second diffusion of Buddhism in the eleventh century, when the Prsa
gika school became established there.
In the following pages Prsagika and Svatantrika will be employed as
convenient designations for the two main divisions of the pure Madhyamaka
school going back respectively to Buddhaplita and Bhvaviveka although
these term s refer to a single feature of each of their doctrines, namely their
methods of ascertaining reality through reasoning.177 And the name Yogacara-
Madhyamaka will be used, in accordance with earlier usage, to designate the
synthesizing school of Madhyamikas whose foremost representative was

Haribhadra and th e school o f CandrakTrti respectively according to mKhas'grub*

dG elegs*dpal*bza, sTo*thun*chen*mo, fol. 4 1 a b. See also Janrdbyas-bzad'pa*
ag*dba*bron#,grus, Grub^mtha^chemmo, ii, fol. 131b 6 and 142b 6; ICamskya*
R ol'p ai'rdoTje, Grub'mtha*, kha, fol. l i b . ) F or the apratisthdnavacLa, reference m ay
be m ade also to th e Param rthabodhicittabhvankram a ascribed to Avaghoa/
Sra (gi, fol. 1 5 6 b 23).In his ITa'bar rin r p a T m a n ag, however, dPal'bregs
appears to app ly a similar term inology differently (fol. 140 ab, where he is speaking
o f the S aut rn tik a-M adhyam ikas and Yogcra-M dhyam ikas).Corresponding
T ibetan term s are sgyu m a rigs griob p a and rab tu m i gnas pa. According to (5okha*pa
(Lanrrinrchen'm o, fol. 342 ab), these tw o appellations were (inappropriately)
given b y som e scholars to th e tw o branches of M adhyam ikas w ith reference to their
respective theories o f th e param drtha. o*kha*pa refers to B lo #ldan*ses T abs
stricture against a ttem p ts to m ake such distinctions w ith respect to th e M adhyam i
k a s theories o f th e param drth a.On the term *jig rten grogs sde spyod p a 'i dbu m a
(p a ) for th e Prsagika school see below, note 259.
175 Bodhibhadra has sta ted th a t Sautrntikas too m ay be regarded as M ahy-
nists. A nd he n otes th a t th e expression Sautrantika has been explained as the nam e
given to those w ho accept literally th e te x t o f th e Sanm ukha and B hadracarystra
(Jnasrasam uccayanibandhana, fol. 49 ab).
176 sK a ba'dPal'bregs, lT a b a i*rinrpai*man-ag, fol. 140ab; Ye'es*sde,
ITa'bai'k h yad par, fol. 252b (cf. M anuscrit P elliot tib tain 814, fol. 5 a s q .; see also
Manuscrit P elliot tib tain 116, p. 112 sq.).
177 These term s m ay also refer to th e m anner o f th e generation in th e m ind
(sam tana) of th e theory ascertaining pdram arthika-nyat; see okha*pa, op.
cit., fol. 343 a 4.
178 For further details see below , p. 87 sq.
60 David Seyfort Ruogg The Literature of the Madhyamaka School

Concerning B uddhaplitas life little is reliably known except th a t he
flourished about the year 500.179 His extensive commentary (Vrtti) on the
MMK is now available only in a Tibetan translation;180 but his great importance
for the history of the Madhyamaka is amply vouched for by his follower
Buddhaplita represents a conservative current in Madhyamaka thought
th a t resisted the adoption of the logico-epistemological innovations which
were a t the time being brought into Mahynist philosophy (e. g. by Dignga,
c. 480540). Thus he did not make use of independent inferences to establish
the Madhyamikas statem ents; and he employed th e well-established prasanga
method, which points out the necessary but undesired consequence resulting
from a thesis or proposition intended to prove something concerning an entity.
From the M dhyamikas standpoint this m ethod has the advantage of not
committing the critic who uses the prasaga to taking up a counter-position
and maintaining the contradictory of what he has denied, which as a Mdhya-
mika he would consider to be just as faulty as the position he has negating.181
B uddhaplitas procedure appears accordingly to be in keeping with Ngr-
junas as expressed in the MMK and the Vigrahavyavartani.
W ith respect to the Mdhyamikas statem ent 'Things are not produced from
themselves (cf. MMK i. 1), Buddhaplita points out th a t such production of a
thing having own being from itself would be quite useless because it would
already exist in virtue of the own being it is supposed to have; in addition,
such production would involve the fault of over-extension (atiprasanga), for
a thing already existing by own being would, on this assumption, never cease
being produced. Regarding the second negative statem ent Things are not
produced from an other, Buddhaplita observes th a t if they were so produced
the consequence would necessarily be th at all things could be produced from
all things (sarvatah sarvasambkavaprasangah). W ith respect to the third
negative statem ent Things are not produced from both [themselves and an
other], B uddhaplita points out th a t the faults attaching to the two preceding
alternatives would combine in this third one (vbhayadosaprasaga). Finally,
concerning the fourth negative statem ent Things are n o t produced from no
cause, Buddhaplita also observes th a t were they so produced the consequence
would necessarily be th a t all things could always be produced from all things
(sada ca sarvata ca sarvasambhavaprasagah) .1Ba
179 Cf. Y . K a j i y a m a , W ZKSO 1213 (19689), p. 194, quoting H . U i, who gave
470540; D . S e y f o r t R u e g g , ibid., p. 306; H . N a k a m u r a , Jo u m . o f Intercultural
Studies 4 (1977), p. 126. E . F r a u w a l u n e r , P hilosophie des Buddhism us (Berlin,
1956), p. 221, p u t him a little earlier, in the fifth century.
180 On Chapter ii cf. M. T a c h i k a w a , A stu d y o f B u d d hap alitas M ulam adhyam a-
kavrtti, Journal o f th e F a cu lty o f Literature, N a g oy a U niversity, N o. 63 (1974).
181 See above, p. 3638.
189 Cf. MMK and Prasannapad i. 6 s q .; vii. 18 s q .; u n yatsap tati 45. I t is to
be recalled th a t all these statem en ts contain p ra sajya -negations.
The Middle Period: The Systematization of the Madhyamaka School 61

In m any passages Buddhapalitas V rtti is closely related to or identical with

the *Akutobhaya. The precise relationship of these two works has yet to be
fully clarified, especially in view of the fact th a t the Tibetan texts of the last
five chapters of both works (translated by Jnanagarbha and Cog*ro*Klur
rgyal*m6han) are practically identical.

Bhavaviveka (c. 500570 ?) was born probably in South India. He was also
known as Bhavya and Bhaviveka.183
Bhavaviveka took up a position radically opposed to Buddhapalitas on the
m atter of the logical establishment of the M adhyamikas philosophical position
in general and of the negative statem ents in particular. In his view the neces
sary co-ordination with scripture (dgama) of an adequate logical method of
reasoning (yukti) requires more than prasanga arguments because, to establish
the M adhyamikas position, there is needed in addition an independent
(svatantra) inference (anumana), which can also be embodied in a proper
syllogism (prayogavdkya). And it is from this characteristic use of a svatantrd-
numdna th a t Bhavavivekas school has received its name of Svatantrika. This
school is also known as the Sautrantika-M adhyamaka, in contradistinction to
the synthesizing Yogacara-M adhyamaka.184
In thus seeking to establish the Madhyamikas doctrinal position by means
of logically elaborated inferences and syllogisms, rather than by the prasanga
method only, Bhavaviveka evidently wished to utilize methods developed by
the Buddhist logico-epistemological school headed by his elder contemporary
Dignaga, who has been regarded as belonging to a branchthe Satyakaravada
or Sakaravadaof the Vijnanavada school while a t the same time continuing
Sautrantika tendencies.
Bhavavivekas views on the philosophical status of the two truth-levels
(samvrti0 and paramartha-satya) and on the formulation of logical proofs made
him, however, the target of criticism by masters of the Vijnanavada: Sthira-
m ati (c. 510570?) of the Valabhi school, who commented on N agarjunas
MMK apparently after Bhavaviveka, and Dharm apala (c. 530561 ?) of the

188 The nam e B h a va viv ek a is attested in Candrakirtis Praaannapad. The form

B h a vy a is found in th e M ah vyu tpatti (no. 3495: sK ahldan). A nd the form B hvi-
vek a is presupposed b y som e Chinese transliterations as w ell as b y th e T ibetan
translation sN a h bral (compare th e nam e B h vivik ta, a N aiy y ik a o f th e six th
century ?). In th e M adhyam aka^astrastuti o f Candrakirti (11) w e find also B h vin .
On him see L. de L a V a l u ^e P o u s s i n , MCB 2 (19323), p. 60 sq .; S. I id a ,
R eason and em ptiness (Tokyo, 1980), p. 5 sq.
184 In th e T ibetan doxographical literature th e tw o schools are th en know n
respectively as th e rnDo'sdefspyod'lp ardbu'm alraiV rgyud'pa] = Sautrantika-
[Svatantrika-]M adhyam aka and the rN a h b y o r sp y o d p a V d b u 'm a lrah T g yu d 'p a ] =
Y ogcra-[Svtantrika-]M adhyam aka.
62 David Seyfort Ruegg The Literature of the Madhyamaka School

N alanda school, who commented for example on Aryadevas Catub^ataka.185

As understood by his commentator Avalokitavrata, Bhavaviveka has referred
critically to a commentary on the MMK written by Sthiram atis teacher
Gunam ati.186 On the contrary, Bhavaviveka has quoted DevaSarmans com
m entary on the MMK with approval.187
Bhavavivekas commentary on the MMK entitled Prajnapradipa, which is
now accessible only in Chinese (Taisho 1566) and Tibetan translations, is ac
cordingly one of the most im portant explanations of th at basic work from both
the historical and philosophical points of view. I t is related to the *Akutobhaya
in many passages. And it contains a critique of B uddhapalitas interpretation
of Nagarjuna, as well as critical discussions of the Samkhya, Vaiesika, Nir-
grantha (Jaina), and Abhidharmika doctrines.

No less im portant is Bhavavivekas great original work comprising the

M adhyamakahrdayakarikas, available in the original Sanskrit and in a Tibetan
translation, and his own very extensive commentary entitled Tarkajvala, now
accessible only in Tibetan. This work is one of the earliest and most valuable
sources we possess for the history not only of Buddhist thought but also of
Indian philosophy in general, for it contains a review and critical discussion
of the doctrines of the main schools known to Bhavaviveka. Chapter ix refers
even to some practices of the Persian Magi (maga) . This work has served as a
basis for later histories of Indian philosophy composed by Buddhist authors
in India and then, especially, in Tibet. I t consists of the following eleven
chapters: (i) Maintenance of the bodhicitta, (ii) Following the Munis vow,188
(iii) Quest for knowledge of reality (tattvajnanaJ,189 (iv) Introduction to the
determination of reality according to the &ravaka, (v) Introduction to the
determination of reality according to the Yogacarin,190 (vi) Introduction to
reality according to the Sam khya,191 (vii) Introduction to reality according to
the VarSesika, (viii) Introduction to reality according to the V edanta,192

185 See A valokitavrata, Tika, wa, fol. 103a sq. Cf. Y . K a j i y a m a , W ZKSO 1213
(1968), pp. 193203, on th e controversy betw een B hava viv ek a and th e Vijnanavada.
188 See Prajnapradipa i. 4, and A valokitavrata, Tika, wa, fol. 198 b 199 a. Cf.
Y . K a j i y a m a , W ZKSO 7 (1963), p. 3738; 8 (1964),*p. 106.
187 See e.g . Prajnapradipa i. 7 d .A valok itavrata notes (wa, fol. 2 2 5 a 7 226b)
th a t B h avavivek a considered D eva^arm ans interpretation here as correct (in
contradistinction to B u d d h ap alitas).
188 See Y. V. G o r h a u e , I I J 14 (1972), pp. 4 0 - 4 5 .
189 See V. Y. G o k h a x e , I I J 5 (19612), pp. 2715. P art o f th is chapter has been
studied b y S. I l d a , op. cit., p. 52 s q .; see also th e sam e au th ors article in M.
S p r u n g (ed.), The problem o f tw o truths in B uddhism and Y edanta, pp. 6477.
190 See S. Y a m a g u c h i , B u k kyo ni okeru m u to u no tairon (K yoto, 1941); S. I i d a ,
K anakura F estschrift (Tokyo, 1966), p. 79 sq.
191 See M. H o n d a , IB K 16/1 (1967), pp. 3 3 -3 8 .
192 This chapter o f th e M H K , w hich seem s to contain th e earliest know n reference
to th e term veddnta in M adhyam aka literature, has been published b y V. V. G o k h a l e
and H . N a k a m u r a , II J 2 (1958), pp. 16589, and H . N a k a m u r a , A L B 39 (1975),
The Middle Period : The Systematization of the Madhyamaka School 63

(ix) Introduction to the determ ination of reality according to the Mmms,193

(x) Exposition of the realization of omniscience (sarvajnatdsiddhi)> and
(xi) Exposition concerning praise and specific characteristics (stutilaksana) .
The Tattvajfinm rtvatra, mentioned by Bhvaviveka in his Jewel in
hand treatise, is evidently the title of a work th a t comprised only chapters
iiii of our M adhyamakahrdayakrikas and contained the main points of its
authors philosophy.194 The Nikyabhedavibhagavykhyna, which appears
separately in the bsTan-gyur, corresponds to part of Chapter iv of the

Bhvavivekas Jewel in hand treatise (Chang-chen-lun, available in Hsiian-

tsangs Chinese version, Taish 1578)196 was apparently composed between the
T attvajnm rtvatra and the expanded work in which this text has been
included, for it contains a reference to the former title. I t presents a summary
of Mdhyamika doctrine and takes the form of a discussion, on the levels of
both reasoning and meditation, of the Mdhyamikas statem ent: Tn reality
conditioned things are em pty (nya) because they are produced from condi
tions, like a magical production; the unconditioned (asamskrta) is not real
because it is not produced, like a sky-flower. This statem ent consists of two
inferential propositions each accompanied by its logical reason and homologous
example (no heterologue being available since the propositions are universal),
and including the im portant qualification in reality (paramdrthatah) which is
characteristic of Bhvavivekas system.197 The treatise also discusses and re
futes Smkhya and Vaiesika doctrines, as well as the Vijanavdins theory
of the paratantra or dependent nature. I t contains a defence of the Mdhya
m ikas negative statem ents, and it emphasizes both th a t they contain non-
presuppositional negation and th a t they have the function of showing th a t
w hat is negated is without own being (svabhva, rather than of annulling some
own being possessed by the thing negated).198 The final section of the treatise
is devoted to an im portant exposition of Madhyamaka gnoseology, bhdvandy
praj and the yogdcdrs understanding of reality,199 elimination of the
discursive concept of emptiness (nyat), and realization of the Middle Way.

p p . 30029. See also H . N a k a m u r a , Shoki no V edanta tetsugak u (Tokyo, 1950),

p . 238 sq.; H JA S 18 (1955), p . 103; JO IB 14 (1965), p p . 2 8 7 -9 6 .
193 See S. K a w a s a k i , I B K 22/2 (1974), pp. 1 - 8 .
194 See V. V. G o k h a l e , IIJ 2 (1958), p. 165 n ote; 14 (1972), p. 41.
195 On M adhyam akahrdayakrik iv. 8 (ja, fol. 161a3 sq .).Cf. A. B a r e a u , J A
1956, pp. 16791 (and Les sectes bouddhiques du P etit V hicule [Paris, 1955],
pp. 1726).
196 See L. de L a V a l l e P o u s s i n , MCB 2 (193233), pp. 60138; N . A i y a s w a m i
S a s t r i , K aratalaratna (V isva-Bharati Studies N o. 9, Santiniketan, 1949).
197 See below.
198 Cf. abov, p. 22, 3738.
199 T hat is, th e Y o g in s understanding (not specifically th e Y ogcrin/V ijhna-
v d in s) ; cf. above, p. 53 and below, p. 72.
64 David Seyfort Kuegg The Literature of the Madhyamaka School

The *M adhyamakrthasamgraha200presents a schematic division of the levels

of truth. The paramdrthasatya, which is without discursive development
(prapanca), comprises not only the pure and ultim ate absolute free from all
conceptualized and verbalized forms (rnam grans: parydya) whatsoever, but
also a second aspect associated with conceptualization and verbalization where
reasoning is employed to negate both the four conceivable positions (koti) and
the production of a thing in terms of any of the four positions mentioned in
MMX i. 1. On the other hand, samvrti is either true (tathyasamvrti), in the
case of things th a t are causally efficient (arthakriyasamartha); or it is false
(mithydsamvrli), in which case it may either be accompanied by imaginary
construction (e.g. the rope mistaken for a snake) or be without this imaginary
construction (e.g. the moon which is seen as double because of a physical
defect actually present in the eye and not because of imaginary construction).
While it has been suggested th a t the saparyaya-paramdrtha and the tathya
samvrti are identical,201 it should be noted th a t the second kind of paramdrtha
(the saparydya) consists only in the negation of the positions in terms of which
an entity is posited as having own being (svabhdva)th at is, in a special kind
of propositional knowledge th a t may be considered pdramdrthika because it
pertains to ultim ate reality or unyatdwhereas the tathya-samvrti covers all
things possessing causal efficiency on the surface level of relative transactional
convention. This discussion has a parallel in passages of the Tarkajvl (iii. 26,
fol. 63 a) where it is stated th a t there is a second form of paramdrtha comprising
discursive development (prapanca) and notional construction (abhisamskara)
which is exemplified in the negative propositions (pratisedha) of MMX i. 1;
it is therefore known as uddha-laukikajhdna (as distinct from lokottarajana),202

Bhvaviveka considered th a t what he termed the prasangavdkya of Buddha-

plita remains open to objection (glags yod pa'i hig = sdvakdavacana) from
an opponent because it lacks the logical reason (helu, liga) and example
(drstdnta) necessary for a valid inference; and he argued th a t it therefore
cannot exclude the opposite thesis and thus acquire full probative force.203
Moreover, according to Bhvaviveka, by simply negating production without
any qualification or restriction the user of this prasangavdkya comes into con
flict with the established Buddhist doctrine of origination in dependence

*00 S e e N . A i y a s w a m i S h a s t r i , Journal o f Oriental R esearch 5 ( 1 9 3 1 ) , p p . 4 1 4 9 .

101 Y . K a j i y a m a , B h v av ivek a and th e Prsagika School, N ava-N land-
M ahvihra Research Publication, i (1957), pp. 302, 311, 31415.
S0J This parallel w ould appear to support th e ascription o f th e *M adhyamakartha-
sarpgraha to th e author o f th e Tarkajvl.
*03 Cf. A valokitavrata, Tk, wa, fol. 8 6 a b: gan ya h de j i Itar rigs p a m a y in
e n a\ glags yod p a 'i chig y in p a 'i y a p h yir tej gnas brtan buddha p li tas bad p a de
n i rgol ba gan g y i klan k a'i glags yod p a 'i hig y in p a 'i p h y ir ya rigs p a m a y in
noj ya es bya ba'i sgra n i gtan higs da dpe m a brjod p a 'i p h yir da gan gyis
sm ras p a 'i es p a m a bsal ba'i p h y ir rigs p a m a y in par lba' ig tu m a zad k yij de n i
glags yod p a 'i hig y in p a 'i p h yir ya rigs p a m a y in no es bya bar sbyar ro/l
The Middle Period : The Systematization of the Madhyamaka School 65

(prattyasamutpada). Therefore, in place of it, Bhavaviveka introduced the

svantardrnumna qualified by the restriction paramrthatah (don dam par na)
in reality; and he expressed this independent inference in the form of a
syllogism (prayogavakya).2Qi When the M adhyamikas negative statem ents
are thus expanded by the restriction paramrthatah, the propositions (pratij)
take the form : na paramrthatah + subject of the proposition + svatah (or : para-
tah, etc.) utpannam x is not in reality produced from itself (or : an other, etc.).205
In his Tarkajvl Bhavaviveka has specified th at the qualification para-
mdrthatah relates to the predicate of the inferential proposition;206 and the
paramdha to which reference is made is Bhavavivekas second kind which
involves discursive development (prapahca) and notional construction
(abhisamslcara) . 207
W ith regard to the type of negation involved in the negative statem ents of
the Madhyamaka, Bhavaviveka appears to be the first thinker of the school to
have explicitly stated th a t it is absolute (non-presuppositional) negation
(prasajyapratisedha)i.e. one th a t consists in pure negation without the
opposite of what is negated being even implicitly affirmedrather than relative
(presuppositional) negation (paryndasapratisedha) where negation is subordi
nate to the implicit affirmation of the opposite of what is negated.208 This being
so, according to Bhavaviveka absolute negation of a proposition does not
involve the indirect affirmation of the contradictory proposition.
Since the negative propositions in question have universal validity for all
dharmas, any example (drstnta) adduced cannot be other than a homologous
(sapahsa) one ; and Bhavavivekas prayogavaJcya therefore lacks the heterologue
(vipaksa) in which the logical reason is absent, and which is an essential p art
of a valid inference according to the logicians.209
Bhavaviveka has observed th a t his use of negative propositions does not
represent mere fault-finding contentiousness (vitand); this is because his pro
position does have as subject (paksa) a term th a t he acceptsnamely nyat
and it is therefore w ithout the fault (paksadosa) of being em pty.210
Bhavavivekas arguments against the Vijnnavda are to be found in his
Prajnapradipa, and in Chapter v of the Madhyamakahrdayakarikas and the
Tarkajvl. I t seems th a t in his time the doctrinal disagreements between the
Mdhyamikas and Vijfinavdins became acute. This opposition did not, how
204 Prajnapradipa, passim . Cf. Candrakirti, P rasannapad i, pp. 14, 25, 36, 38.
205 Prajnapradipa, i. 1, p. 11, quoted b y Candrakirti, Prasannapada i. 1, p. 25 sq.
208 T arkajvl iii. 26, fol. 63b 64a.
207 See above, p. 64.
sos Prajpradipa i. 1, p. 10; Tarkajvl iii. 26.I t is to be n oted th a t th e
T ibetan doxographers tak e up th e tw o typ es o f negation w hen describing th e
Sautrntika doctrines also. The Indian grammarians h ave o f course used the term s
p ra sa jya and paryu ddsa from th e tim e o f P ata jali; see above, note 94.
209 Cf. *K aratalaratna in MCB 2 (193233), p. 73. (On th e absence o f a sapaksa ,
see ibid., p. 78.) From th e poin t o f view o f these logicians, then, this form al defi
ciency is a very serious w eakness in B h va vivek as system .
210 Tarkajvl iii. 26, fol. 64b.
66 David Seyfort Ruegg The Literature of the Madhyamaka School

ever, prevent him from adapting the advances made by Dignagas branch of
the Yogacara-Vijnanavada in either logic and epistemology or in soteriology
and gnoseology whenever he found them appropriate, but without of course
abandoning the basic Madhyamaka principle of the emptiness of all dharmas
including consciousness (vijnana) and the highest non-conceptual gnosis
(jnana) achieved in meditative realization. In other words, for the investiga
tion of the surface level, Bhavaviveka was prepared to make use of methods
and insights developed by another school; but in his theory of the paramartha
he remained faithful to the basic theory of his school.211
Thus not only in the Tarkajvala but also in his Jewel in hand treatisein
connexion with the proposition The unconditioned is not real because it is
not producedBhavaviveka presents a critique of the more or less substantialist
theories of absolute reality propounded by other schools. He rejects the
Vaibhasikas notion of pratisamkhyanirodha as an entity existing in reality,
as well as the conception th at takes akada as some thing non-existent which he
ascribes to the Sautrantikas. And he dismisses the Vijnanavada idea of tathatd
as possessing substantive (dravyasat) own being, and as the object (dlambana)
of either supramundane non-dichotomizing gnosis (nirvikalpakalokottarajndna)
or pure post-concentrative mundane knowledge (prstJi^labdha-laukikajndria);
for a theory of the tathatd th a t conceives it as an absolutely real entity, though
beyond the categories of existence and non-existence, merely assimilates it to
the ineffable atman of the heterodox sectarians. Tathatd, then, is simply cessa
tion of all dichotomizing conceptualization.212

The *M adhyamakaratnapradipa is also ascribed to Bhavya. In the Tibetan

translation, the only extant version of this work, the Tarkajvala is even re
ferred to as having been composed by its author.213 However, the work contains
references to Dharm aklrti and Candrakirti; and its doctrine bears the stam p of
later developments, including some inspired by the Vajrayana. Presumably it
has accordingly to be assigned to another (later) Bhavya.214

211 B h av av ivek a seem s thus to h ave taken a step towards a synthesis o f th e

m ethods (if n o t th e doctrines) o f these tw o great schools o f M ahayanist philosophy,
a synthesis th a t w as developed further b y antarakita in th e eighth century.
B h a v a v iv ek a s Sautrantika Svatantrika-M adhyam aka is nevertheless considered
b y doxographers to differ from &antarakitas Y ogacara-Svatantrika-M adhyam aka
b y accepting external objects (bdhyartha) and rejecting th at form o f direct k n ow l
edge additional to th e cognitions derived through th e live sense-faculties and th e
m ind, and known as self-aw areness (svasam vedana, svasam vitti) . See b elow ,p . 87.
212 See MCB 2 (1 9 3 2 -3 3 ), pp. 1 0 5 -2 2 .
213 F ol. 3 3 5 b 3: bdag gis bkod p a rtog ge bar ba.
214 On this w ork cf. lCaivskya*Rol*pai'rdoTje, Grub*m tha\ kha, fol. 6a ; S .
S c h a y e r , HO 11 (1935), pp. 20611; S . Y a m a g u c h i , B ukkyogaku bunshu, i
(Tokyo, 1972), p. 249 sq. C . L i n d t n e r , AO 40 (1979), p. 90, has ascribed this work
to B h a v y a = B h av av ivek a I, disregarding th e reference to D harm aklrti (and dating
Candrakirti as a consequence to 530600).
A certain B h a v y a (sKaWdan-rgyal'po) was, according to B u 'ston (Chos^byuh,
The Middle Period: The Systematization of the Madhyamaka School 07

Later SvatantriJcas
Avalokitavrata composed a very extensive and detailed Tik on Bhva-
vivekas Prajnpradipa.215 This work is of great value for the historian of
Indian philosophy during this im portant and active period in its development.
Among the great Buddhist masters of the seventh century to whom it refers are
Dharmakirti, the continuator of Digngas logico-epistemological school, and
Candrakirti, who opposed Bhvaviveka on several im portant points of Madhya
maka thought.216
An earlier commentary on the Prajnpradipa by G unadatta is referred to by
A valokitavrata.217
Jhnaprabha, a follower of Bhvaviveka, was known to Fa-tsang (643712)
on the authority of Divkara as a contemporary and opponent at Nland of
labhadra, the disciple of Dharm apla (c. 530561 ?) and teacher of Hsiian-
tsang (602-664).218

rigupta (seventh century ?) is the author of a small treatise entitled T att-
vvatravrtti. According to Trantha he was a teacher of Jnnagarbha;21
and he is often classified as a Yogcra-Mdhyamika.220
fol. 138 a 7 [ii, p. 215]), th e teacher o f rog*Bloddan*es*rab (10591109), w ith w hom
he translated P rajkaraguptas Pram navrttiklam kra; he m ust therefore be
distinguished also from th e author o f th e Prajnpradipa, etc. On th e T ntrika
B h a v y a see below , p. 106.
116 Trantha m akes A valokitavrata a contem porary, in th e South, o f K ing
D evap la (whom he places before Dharm apla, rg. c. 770810 or 775812); see
rGya*gar*chos*byuri, p. 162. H e has been ten ta tiv ely d ated in th e seventh century
b y Y . K a j i y a m a , W Z K S 17 (1973), p. 162.
218 I t has been observed th a t A valokitavrata has n o t gone in to Candrakirtis
critique o f B h v a v iv ek a s position; see Y . K a j i y a m a , W ZKSO 7 (1963), p. 39.
217 Tik, wa, fol. 45 a 3. Dipamkararijna has included a com m entary b y
G unadatta in his list o f th e eight great com m entaries on th e MMK (Bodhimrga-
dipapajik, fol. 3 2 4 b ); b ut this indication seem s to be due sim ply to confusion
w ith Devaarman, w hom ho has m entioned as a com m entator on th e Prajnpradipa
(whereas, as n oted above, A valok itavrata sta tes th a t B h v av ivek a refers to
D evaarm an). In the list o f th e eight com m entaries given b y m K h asgrubdG elegs*-
dpal*bza, D evaarm an in fact figures instead o f G unadatta; and Devaarm ans
dKar*po*rnam*parchar*ba is said to be th e com m entary on th e MMK follow ed b y
B h vavivek a (sTo*thun*chen*mo, fol. 37b).
218 See F a-tsan g s treatise on the Awakening o f F a ith 5 referred to b y L. de L a
V a l e e P o u s s i n , MCB 2 (193233), p p . 6364, 139. Cf. N . A i y a s w a m i S a s t r i ,
K aratalaratna, p . viii; N . P r i , B E F E O 1911, p . 367 n o te; A. F o r t e , D ivkara
(613688), Annali di CaFoscari, 13/3 (Serie Orientale 5, Venice, 1974), p p . 1501.
219 rGya*gar'chos*byu , p. 152. See also th e ordination-lineage quoted below,
p. 89 and n ote 285.T rantha m akes rigupta a contem porary o f D harm akirti
and Vim alacandra (a king o f th e Candra d yn asty w hich preceded the Plas), and
gives his teachers nam e as *Sam praduta ( ?); see op. cit., p. 132.
220 Bu*ston counts rigupta as a Y ogcra-M dhyam ika (Chos*byu, fol. 103a
[ii, p. 135]; cf. fol. 158b).
68 David Seyfort Ruegg The Literature of The Madhyamaka School

His treatise is devoted to the logical elaboration and explication of the

proposition th a t all thingsexternal elements (paramdnu) and internal cogni
tionsare in reality (tattvatah) without own being, for they are all without
either simple or multiple self-nature (ekasvabhdva and anekasvabhdva) just like
a reflection (prcUibimbavat). This argument constitutes one of the four (or five)
great reasons adduced to establish the Madhyamaka theory of non-sub
stantiality by the later Indo-Tibetan tradition. A m aterial element, it is ex
plained, is not simple (i.e. partless and hence spatially unextended) because
there could then never be any extended aggregates of such elements in space;
nor is it multiple since multiplicity presupposes a plurality of simple elements
the possibility of which has just been refuted.221 Similarly it is argued th a t the
notion of either simple or multiple cognition is untenable. And any third
possibility (raM) beside the simple and multiple is unavailable owing to the
principles of contradiction and the excluded middle. Moreover, the images
(akara) of the mind are not false (allka). B ut ultim ately mind as well as the
mental factors are without own being. And the teaching of Vijnapti is pro
visional since it serves as an expedient to bring persons who postulate the
existence of real entities to an understanding of non-substantiality. All things
perish in an instant (ksana) and are transient; permanent things could be
efficient (arthakriydsamartha) neither sequentially nor simultaneously. The
Middle W ay lies then in eliminating the twin extremes of eternalistic im puta
tion (samdropa) of a svabhdva and nihilistic denial (apavada) of everything.
While one does not know reality, and so long as samvrti is not brought to a
stop, one is bound to the wheel of acts and their results, viz. the conditions
culminating in old age and death; passion, hate and the other kleias in fact
arise because of grasping at things misconceived as substantially real. B ut
through exact view (samyagdrsti) both ultim ate good (naiMreyasa, i.e. libera
tion) and well-being in the world (abhyudaya,) are attained.

Jnanagarbha, who is stated to have been a disciple of Srigupta,222 is the
author of the Satyadvayavibhanga and its autocom m entary (Vrtti).223 This
work has been commented on by Santaraksita in his Satyadvayavibhanga-

221 For an earlier critique o f the id ea o f a param dnu along similar lines, see
V asubandhu, Vim^atika 11 sq. A nd for th e idea th a t an elem ent is neither im partite
and sim ple (since it has m ultiple sides in space) nor m ultiple (because this could be
nothing but a m u ltip licity o f the sim ple units ju st rejected), see also R atn avali i. 71
(and Dharm aklrti, Pram anavarttika, Pratyak^apariccheda 360).
222 See Taranatha, rGya*gar*chos*byun, p. 152.
223 These tw o works b y Jnanagarbha are available in the sD e'dge edition
(38812), but not in th e P eking edition o f th e bsTan*gyur.Bu*ston counts
Jnanagarbha as a Yogacara-M adhyam ika (Chos-byun, fol. 103a [ii, p. 135]).
See also (5on*khapa, L egsbsad*shih*po, fol. 63a. According to m K has-grubdGe*-
legs*dpal*bzah, Jnanagarbhas S atyadvayavib h ahga, &antarakitas M adhyam a-
The Middle Period: The Systematization of the Madhyamaka School 69

P ajik.224 Jnagarbha is also reported to have been a teacher of ntarak-

sita.225 In all probability he flourished in the first part of the eighth century.22*
In his treatise he has sought to clarify the theory of the two truths, which
had become the object of misunderstanding even among some Buddhists. The
opponents against whom he was arguing were Dharm apala (c. 530561) and
his followers according to ntaraksita (Pajik, fol. 4 a 2), who also notes
discussions by Jnagarbha of the views of Sthiram ati (c. 510570) (fol. 32 b)
and Devendrabuddhi (c. 630690) (fol. 28a). In conformity with the Yogcra-
Madhyamaka position, the theory of the paratantrathe second of the three
svabhavasis alluded to (fol. 7 a, 12 b); while the cittamatra is regarded as only
a step leading finally to the ultim ate understanding of dharmanairatmya accord
ing to the Madhyamaka (fol. 13a).
Jnagarbha explains th a t the first of the two truths, the paramartha, is non-
deceptive. Its nature is in accord not with appearance {snan ba: pratibhdsa,

klam kra and Kamalaila*s M adhyam akloka were know n as th e ra rgyud ar

gum (rG yudsd e'sp yh m am , ed. Lessing-W aym an, p. 90). B u t th is au th ority counts
Jfinagarbha as a Sautrntika-Svtantrika-M dhyam ika since (he says) th e latter
m aintained th a t ru p a , abda> etc. are other than citta and are therefore non-sentient
and m ateria l; while ntaraksita and liis school held th em to be not other th an citta
and considered th e bhyrtha (corresponding to w hat others held to be jada or
material) to be unestablished (which is, however, n ot to say th a t th e y claim ed all
dharm as to be sim ply ones ow n thought) (see op. cit., p. 92, and sTofi*thun*chen*mo,
fol. 3 7 a).For som e Japanese studies on Jnagarbha see H . I n a o a x i , i n : B uddhist
th o u gh t and A sian civilization (Fests. H . V. Guenther, E m eryville, 1977), p. 132 sq.
214 A Jfinagarbha translated ntarakitas Satyadvayavibhaga-P afijik into
T ib etan together w ith Y e Ses'sde in th e early n in th century; th is fact makes it
un lik ely (though n ot entirely im possible) th a t he is th e sam e Jfinagarbha as th e
author o f th e basic te x t and V rtti o f th e Satyadvayavibhafiga. The translator
Jfinagarbha was also responsible, together w ith Cog*ro*Klui'rgyal*mhan, for the
T ibetan translations o f B h a v a v iv ek a s Prajfiapradipa and A valok itavratas Tk
thereon, so th a t he m a y have been a Svtantrika-M dhyam ika; however, he also
translated B u ddhaplitas com m entary on th e MMK (first translation). In these
circum stances it is difficult to determ ine which Jfinagarbha was th e author o f th e
com m entary on the M aitreya-chapter o f th e Sam dhinirm ocanasutra, and also o f
th e Y ogabhvanm rga (or patha) included in the M adhyam aka section o f the
bsTan*gyur (the latter w ork shares a concern w ith problem s o f m editational
p ractice treated b y K am alaila in his Bhvankram a).A nother Jfinagarbha was
a teacher and collaborator o f Mar*pa*Chos*kyrblogros in the eleventh century; see
e .g . Gos*gon-mrdpal, Deb*therson*po, fia, fol. 2 a (ii, p. 400, together w ith G. N .
R o e r i c h s note on p. 417).
I t is to be noted further th a t T rantha has recorded th e existence o f doubts
concerning whether the author o f th e Satyadvayavibhaga-P afijik and th e author
o f th e M adhyam aklam kra are identical (rG ya*garchos'byufi, p. 1623). See also
oirkha^pa, Legs*bad*siirrpo, fol. 64b, and lCafi'skya*Rol*pai*rdoTje, Grubm tha,
kha, fol. 6 a b, who also reject th e id en tity o f authorship o f these tw o works.
225 Cf. Sunrpa*mkhan*po Y e* esdpal*byor, dPag*bsam*ljon*bza (ed. S. Ch.
D as), p. 112.
223 T rantha m akes Jfinagarbha a contem porary o f K ing *Govicandra (op. cit.,
p. 152).
70 David Seyfort Ruegg The Literature of the Madhyamaka School

nirbhdsa, etc.) but with certain knowledge achieved through reasoned principles
(yukti or nydya) and a proper logical reason (trirupa-linga) (fol. 4a). I t is also
free from the movement of discursive thinking, not to speak of syllables (aksara)
(fol. 5a); indeed, as stated in the VimalakirtinirdeSa, only silence is adequate
to the dharmamukha which is without duality (advaya) and is free from dis
cursive development (nisprapanca) (fol. 6 ab). The samvrti on the contrary
is appearance, and it is not ultim ately real or true. B ut what is perceived by
everybody inclusive even of untutored cowherds is still termed samvrti-saXya
(fol. 4 a). I t involves designations (prajnapti) in terms of worldly pragmatic
usage (lokavyavahara), namely what is indicated by syllables, language and
linguistic convention (samketa) (fol. 5 a). W hat originates on the basis of causes
and conditions (hetupratyaya)conceived of as authentic things (vastu) with
out, however, being imagined to be produced in realityis true samvrti. I t is
free from anything imaginarily construed as produced in reality (this latter
being a pure construct issuing merely from certain philosophical imputations)
as well as from the appearance of vijnana or the transformations (parinama)
of primordial m atter (pradhana, as in the Samkhya system) and m aterial ele
ments (bhuta) (fol. 5 b ; cf. 12b l), not to speak of erroneous perceptions such
as a double moon (dvicandra) (fol. 4 a). The samvrti is accordingly either true
or not true according to whether it has causal efficiency (arthakriydsdmarthya)
as in the case e. g. of water, which is described as non-delusive on the relative
levelo r does not have this causal efficiencyas in the case e.g. of a mirage,
which is delusive on the relative level. B ut both forms of samvrti are similar in
th a t they partake of appearance (snan ba can) (fol. 5 b, 6b). The true samvrti
is then comprised of all things in their entire extension (yavat) as they appear
from causes in the concordant cognitions of all people, including even the
simple-minded (bala) (fol. 5 b). There is also no disagreement concerning their
appearance-aspect (snan bayi cha) in the cognitions of a disputant (vddin) and
his opponent (prativddin) (fol. 5 a); and this fact makes possible meaningful
philosophical debate invoking the subject, attribute and example of an in
ference (anumdna) concerning things the status of which is in dispute and has
to be validly inferred (fol. 9b). Yet in reality, with respect to their thusness
(tathatd), paramdrtha and samvrti are not different (fol. 10 a). Jnanagarbha
has furthermore discussed the nature of selfawareness (svasamvedana or sva-
samvitti self-cognition/ fol. 4b) and the non-reality of the images (dkdra) of
knowledge (fol. 4b, 7a; cf. 13a). Special attention is given to the principle th a t
neither a simple nor a multiple thing can produce either a single or a multiple
effect, which is argued a t some length (fol. 7a sq.). The treatise ends with a
treatm ent of buddhahood and the three kayas of the buddha together with
associated gnoseological m atters (fol. 14a15a).
In this treatise Jnanagarbha has mentioned (fol. 11a sq.) some bad dispu
tan ts who held not only th a t entities such as rupa are not produced in reality
but also th a t they are not produced even in samvrti, so th a t they are comparable
for example with the son of a barren woman (vandhyapulra, etc., i.e. a mere
The Middle Period: The Systematization of the Madhyamaka School 71

flatus vocis) ; and he has criticized their opinion as being inter alia incompatible
with worldly pragmatic usage based on valid knowledge such as direct per
ception. The allusion might be to Candrakirtis theory of causal indeterminism
even on the relative level, or it may be to some other opponent ; the name of
the opponent Jnanagarbha had in view is not given in the appropriate passage
of ntaraksitas Panjik (fol. 36a sq,).
A Yogabhvanmrga (or Yogabhvanpatha) included in the M adhyamaka
section of the bsTamgyur is ascribed to (a) Jnanagarbha.227

Candrakirti, about whose life we have little reliable information, lived in the
seventh century (c. 600650).228
He criticized with acumen and penetration the objections against Buddha-
plitas procedure raised by Bhvaviveka and this Svatantrika m asters
adoption of independent inferences and syllogisms. And he sought to establish
once for all the prasanga method of reasoning, so th a t he has been regarded as
the founder in the strict sense of the Prsangika school. C andrakirtis critique is
directed also against doctrines of the Buddhist logico-epistemological school
which he regarded as contrary to Ngrj unas teaching, and against the m eta
physical and gnoseological theories of the Yogcrins/Vijnnavdins.
Candrakirtis two largest and best-known works are the Madhyamakavatra,
which is available in some Sanskrit fragments and in Tibetan, and the Prasanna-
pad M adhyamakavrttib, an extensive commentary on the MMK available in
the original Sanskrit as well as in Tibetan.228
The M adhyamakvatra-Kriks are accompanied by a detailed commentary
(Bhsya). This independent treatise is conceived as a general introduction
to the M adhyamaka-astra and was composed before the Prasannapad which
refers to it. I t is divided into sections according to the ten productions of
thought (of Awakening, cittotpada), each of which is linked with one of the ten
stages (bhmi) of the Bodhisattva and a corresponding perfection (pramitd).
These ten sections are followed by a short one on the qualities (guna) of the
227 On th e existence o f different persons w ith th e nam e Jnanagarbha see above,
note 224.
228 The dates 530600 h ave been proposed b y C. L i n d t n e r , AO 40 (1979), p. 91,
on th e assum ption th a t th e *M adhyam akaratnapradipa was com posed b y B h av ya =
B h va viv ek a I. B u t this te x t refers n ot only to Candrakirti but also to Dharm akirti
(see above, p. 66), w ho is usually placed in th e seventh century.
229 A T ibetan translation o f th e M adhyam akvatra w as edited, together w ith
th e autocom m entary (Bhya), b y L. de L a V a l l e P o u s s i n , B ibliotheca Buddhica
ix (St. Petersburg, 190712) ; it w as translated b y him , up to vi. 165, in Muson 8
(1907), pp. 2 4 9 -3 1 7 ; 11 (1910), pp. 271358; and 12 (1911), pp. 2353 2 8 .-T h e
Sanskrit te x t o f the P rasannapad was edited b y L. de L a V a l l e P o u s s i n ,
B ibliotheca B uddhica iv (St. Petersburg, 190313). See also J. W . d e J o n g , I I J 20
(1978), pp. 2 5 - 5 9 and pp. 21752.
72 David Seyfort Ruegg The Literature of the Madhyamaka School

Bodhisattva, one on the Buddha-stage (buddhabhumi) , and a brief conclusion.

In this work Candrakirti has frequently mentioned the practiser of Yoga, or
yogdcdra, thus conforming with the usage of ryadeva and also Bhavaviveka.
The emphasis placed in the MA on soteriological and connected gnoseological
topics, following the Daabhmikastra which serves as a point of departure
on the subject, both distinguishes this general treatise from works of the
school which relate to the MMK and brings it into line with what Nagarjuna
has himself w ritten in some of his other works,230 and with topics already
treated by Bhavaviveka in his Jewel in hand treatise.
The sixth section of the M adhyamakvatra dealing with the Abhimukhi-
stage and the prajdparamitd is by far the longest, and it explains in great
detail different themes connected with transcending discriminative knowledge
(prajd) as the Bodhisattvas central perfection. Taking as his point of de
parture the negation of production from self etc. (cf. MMK i. 1), Candrakirti first
considers origination in dependence (pratityasamutpdda) and emptiness (nya-
td). In particular, he discusses (12) Bhvavivekas use of independent inference
to establish the Madhyamikas doctrines; and he examines his qualification of the
negative statem ents by the restriction in reality (paramdrthatas), which he
rejects on the ground th a t production is no less unreal (indeterminate) on the
level of samvrti (this being one of the main points of the Prsagika school).231
Candrakirti also explains (23 sq.) the theory of the two truthsthe samvrti
and the paramartha-satyain relation to the distinction between covering
(vr-), which is characteristic of relative knowledge in the world (loka), and the
highest gnosis (juana) of the saint (drya) which comprehends with perfect
accuracy. The samvrlisatya is then shown to be bkasamvrtisatya (27), i.e. what
originates in dependence and is thus well-established in worldly transactional
usage (in contradistinction to what is due to erroneous cognition, mithydjnana,
and is therefore untrue even on the surface-level) (28). Conventional tran s
actional tru th (vyavaharasatya) is to be regarded as means (updya), and the
paramdrthasatya as what is attained by means of it (upeya) (80).232 Candrakirti
notes th a t all things (bhdva) have these two natures (23 and 27); but since the
first is actually delusive from the point of view of ultim ate reality it is not in
fact tru e (satya) (28). This leads to the examination (31) of what is established
in worldly consensus (lokaprasiddha) , and of the question of valid knowledge
(pramdna) on the worldly surface level. Next, in connexion with the negation
of production from an other entity, Candrakirti takes up (34 sq.) the exam ina
tion and refutation of the Vijnavdins doctrine of the three naturesin
particular the dependent nature (paratantrasvabhava)and of the store-
consciousness (dlayavijdna). And pointing to the fact th a t the Lakavatra-

230 Or: in works a t least ascribed to Nagarjuna. On th e R atnavali, th e *Bo-

dhisam bhra-sstra and th e com m entary on the D aabhm ika see above. Compare
also Chapter v o f A ryadevas Catuhsataka.
231 See above, p. 43; below, p. 77.
232 Cf. MMK x x iv . 10.
The Middle Period : The Systematization of the Madhyamaka School 73

stra has identified the dlayavijndna with the taihagatagarbha while at the same
time describing the latter doctrine as provisional and of indirect meaning
(neydrtha) y Candrakirti seeks to establish th a t the dlayavijndna itself m ust by
the same token be of provisional and indirect meaning according to the Stra
(95).233 The denial of an external object (bdhydrtha, bahirartha) of knowledge
and the theory of self-awareness (svasamvedana, svasamvitli) are also criticized
in some detail (71 sq.). In the course of his discussion of the Vijnnavdins
mind-only doctrine Candrakirti alludes to the cittamdtra concept mentioned in
the D aabhm ikastra;234 and he clearly distinguishes it from the Vijnapti-
m tra doctrine he is refuting (83). Returning to the subject of origination in
dependence on the level of worldly transactional usage, Candrakirti notes th a t
the canonical formulae traditionally used to express itviz. asmin sati idam
bhavati this being, th a t is and asyotpdddd idam utpadyate as a result of the
origination of this th a t originatesare pure doctrinal convention (dharmasam-
Iceta) since no real production of an entity in fact takes place (114).235 As for
the M dhyamikas own method of analysis (vicdra) th a t puts a stop to all
mental construction, it is resorted to not out of a predilection for disputes and
a mere desire to refute an opponents thesis. On the contrary, it is employed
with a view to liberation (vimulcti). And if other doctrines are overwhelmed in
the course of this investigation, this is simply because they are dispelled by the
exposition of reality (tattva), just as darkness is by light (11718). The non
substantiality of all factors (dharmanairaimya) having thus been expounded,
the following p art of the sixth section is devoted to the non-substantiality of
the individual (;pvdgalanairdtmya, 120 sq.) and to a refutation of the view
hypostatizing the individual constituents as real (satkdyadrsti, in twenty
canonical and twenty-five commentarial forms). This is done by means of the
fourfold and fivefold vicdra (144).236 To these five schemataidentity, otherness,
container, containedness, and connexion237Candrakirti adds two more, namely
the aggregation of distinct component parts and shape as belonging either to
these individual components or to their totality (151 sq.). Here the classical

233 Laikvatrastra v i, p. 2356.In this Stra (ii, p. 7778), the doctrine o f

th e tathdgatagarbha w as compared w ith th a t o f th e dtman. The m ain line o f the pure
M adhyam aka school appears either to disregard th e tathdgatagarbha teaching, or to
consider it as an intentional (abhiprdyika) teaching o f indirect m eaning (neydrtha)
similar to th a t o f the dtman. This is th e position e .g . o f B h vavivek a (Tarkajvl,
ja, fol. 169 a) and Candrakirti (M adhyam akvatra vi. 95) ; cf. D . S e y f o r t R t j e g g ,
L a thorie du tathgatagarbha et du gotra (Paris, 1969), pp. 35 note, 402 sq.On a
M dhyam ika current o f th ou gh t associated w ith the tathdgatagarbha doctrine see,
however, above, pp. 5556.
For th e distinction betw een neydrtha and nitdrtha see also Candrakirti, P P x v . 11
(p. 276) (cf. i. 1, p. 42).
234 D aabhm ikastra, Chapter v i (E) (quoted above, note 69).
235 Cf. P P i, p. 9 sq.; 55; above, p. 43.
238 On satkdyadrsti see recently A. W a y m a n in Studies in Pali and Buddhism , ed.
A. K . N a r a i n (Deihi, 1979), pp. 37580.
237 See above, p. 40.
74 David Seyfort Ruegg The Literature of the Madhyamaka School

example of chariot as a mere fictional designation based on the assemblage of

individual parts is adduced with a view to analysing the fiction of the satkdya-
drsti or atmadrsti th a t is falsely constructed on the basis of the skandhas. This
analysis allows the Madhyamika to retain worldly transactional usage ( loka-
vyavahara)in terms of conditional designation (upadayaprajnaptih)without,
however, positing any entity established either on the surface level of samvrti
or in ultim ate reality (158). The sevenfold scheme of analysis thus reveals the
principle of pratityasamiUpdda (158) while still saving the appearances of things
recognized consensually in the world (lokaprasiddha) of samvrti, but which
cannot sustain careful analysis (159); and it thus permits the Yogin to fathom
reality (tattvam avagdh-) (160-4). After the fictional construct of an dtman the
resulting, and a fortiori fictional, construct of something related to the self
(atmiya) is considered (165). The last part of the section comprises an explana
tion of the sixteen and tw enty forms of unyata (179 sq.). Candraklrti observes
th a t the pudgalanairatmya was taught to deliver the Sravaka and Pratyeka-
buddha, who cannot yet fully and completely comprehend the dharmanai-
rdtmya in all its modes (179). Yet he concludes th a t they can comprehend it
within certain limits, inasmuch as they too must have the knowledge th a t all
dharmas are without self-existence (nihsvabhdva) ,238
The final section of the M adhyamakavatara on the buddhabhumi is of special
interest for the study of Madhyamaka thought since it is devoted to the
Prasangika schools buddhology, th at is, to the nature and qualities of the
buddha and the gnosis (jnana) corresponding to this supreme level.
On the M adhyam akavatara there exists in Tibetan translation an extensive
commentary by Jayananda (see below, p. 113).

C andrakirtis commentary on the MMK, the Prasannapada Mulamadhyama-

kavrttib, is considerably more extensive than Buddhapalitas commentary on
the same stanzas. Very often in his philosophical discussion Candraklrti goes
beyond, and improves on, B uddhapalitas interpretations; this he does tacitly
without calling attention to his own contributions. Candrakirtis critique of
Bhavaviveka as well as of the Buddhist logico-epistemological school is to be
found explicitly set forth in his elaborate explanations on Chapter i of the
In the Prasannapada (xxiv. 8) Candraklrti has again discussed the two truths.
Samvrti may, he explains, mean (i) ignorance since it completely covers (vr-)
the reality of all things, or (ii) interrelatedness, or (iii) convention (samketa),
i.e. worldly transactional usage (lokavyavahdra) defined in terms of the relation
of a designation to its designatum and of a cognition to the object of cognition.
The samvrti in worldly usage is termed lokasamvrti; and while it can serve no
real purpose to distinguish an alokasamvrti opposed to it (from the point of
view of ultim ate reality both are unreal, though in different degrees from the
ass MA i. 8, w hich quotes Da^abhuinikasutra, Chapter v ii ( J ) ; cf. P P xviii. 5. See
also MA vi. 179, and above, p. 7.
The Middle Period: The Systematization of the Madhyamaka School 75

relative standpoint), one may nevertheless speak of an abkasamvrti as distinct

from it when considering th a t there exist persons who can be described as not
of the world (alokah) since they have experiences which are falsified because
of the fact th a t their sense-faculties are impaired (and which, therefore, do not
belong to the general worldly consensus).230 As for the bkasamvrtisatya, it is
true in virtue of worldly convention (samvrti), and the whole of transactional
usage made up of designations in relation to their designata and cognitions in
relation to their objects is th en termed true by bkasamvrti; but it is not so
really (paramarthatah).2i0 Candrakirti defines the paramartha on the basis of
the MMK (xviii. 7 and 9) as th a t with respect to which there is no functioning
of words and cognitions; for i t is not communicated by another, and it is still,
directly knowable (pratyatmavedya) by the Aryas, and beyond all discursive
development (prapanca) . Grammatically speaking, the word paramartha is
interpreted as an appositional compound (analysable as paramaA cdsdv arthaA
ca, i.e. a karmadhdraya).241Such is then what is true, the paramdrthasatya. For
a fuller discussion of the subject Candrakirti refers back to his Madhyamakava-
tara (vi. 23 sq.).
Although relative and transactional tru th is regarded as the means through
which the paramartha is realized,242 Candrakirti evidently regards the surface-
level processes of transactional usage as causally indeterminateeven as anti
nomic and unamenable to ontological construction. On this subject he quotes
Sutra texts as well as the verse of the MMK (xviii. 10), where Nagarjuna has
said th a t whatever originates in dependence on some thing is not th a t thing
nor is it different from it, and th a t th a t thing is consequently neither destroyed
nor eternal (PP i. 1, p. 2526).243 In this m atter of the samvrti level and its pro
cesses Candrakirti thus differs from Bhavaviveka, who has stated th a t tathya-
samvrti, consisting as it does in avitatha-bkavyavahara opposed to mithya-
samvrti, is tru th (salya) and valid knowledge (pramana).2ii
239 Cf. M A vi. 2428 on. th e question of factors th a t originate in dependence on
th e surface le vele.g . citta , rupa , vedandy etc.and w hat is false even on th a t
levele .g . a reflection (pratibim ba) , an echo (pratiSrutka) , etc. (The former m ay
be regarded as transactionally efficient and true, whereas th e latter are not trans-
actionally efficient and are thu s false.)H ow ever, in P P x x iv . 11 bkasam vrti is
described as reflection-like (pratibimbdkdra) . In fact th e entire level o f samvrti and
w hat originates through pratityasamutpdda is compared w ith a reflection; cf. xxii. 2,
x x iv . 35, x x v i. 11 and x x v ii. 29.
240 Elsew here Candrakirti has used th e expression loka-(sam)vyavahdra; see P P i,
pp. 57. 10 and x v . 2.
241 H ere Candrakirti agrees w ith th e first interpretation in B h a va viv ek as
T arkajvala (fol. 63a), w here th e term paramdrtha is analysed as (i) a com pound o f
artha (that w hich is cognized) and parama (the supreme), (ii) a tatpurusa m eaning
artha o f th e suprem e, viz. non-conceptual gnosis, and (iii) th a t w hich conform s to
the supreme artha , viz. prajnd. B u t Candrakirti differs from B h av av ivek a in
regarding on ly th e paramdrtha as satya.
242 MA v i. 80; cf. MMK x x iv . 911.
243 See above, p. 43.
244 Tarkajvala iii. 12 (fol. 60a).
76 David Seyfort Huegg The Literature pf the Madhyamaka School

Self-causation, other-causation and causelessness as well as generation from

self, an other or a combination of both self and other having been rejected
(nisiddha) in the Madhyamaka, the relative surface-level nature of relative
things has been shown (udbhvita). And it is this precisely th at constitutes
relative surface-level production in dependence (sdmvrta-pratltyasamutpdda).
Now, since there is no production in virtue of own being (svabhdva aseity) with
regard to the gnosis (juana) of the rya, no destruction (nirodha, i.e. as the
complementary opposite of utpdda production) is to be found there either;
and the same applies to uccheda and dsvata, ekartha and nandrtha, dgama and
nirgama. The prattyasamutpdda is accordingly characterized in the introductory
verses to the MMK as free from all these eight features : without destruction
and production, neither annihilated nor eternal, neither single nor multiple,
and without movement to an fro (PP i, p. 1011). Thus the principle of con-
ditionship (idampratyayatd) has been rethought by the Mdhyamika (PP i,
p. 9 -1 0 and MA vi. 114).
W ith respect to the nature of things, in the discussion of the fourth position
of the so-called tetralem ma (catuskoti) expressed in the form of a bi-negation
(neither ... nor) and elsewhere, it is observed th a t no property may be appro
priately predicated of ultimately unreal entities. To ascribe a property to an
em pty thing would indeed be like applying the predicate dark or pale to the
son of a barren woman (vandhyputra, vandhydsuta, etc.), which is simply a
non-referring designation.245 To exemplify a non-existent (avidyamdna: abhdva,
nsti) construct or non-referring designation we find in addition the comparison
with a sky-flower (khapuspa, gaganakusuma) ,246 Furthermore, in the discussion
of the conditioned (samskrta) , the sky-flowerwhich as a mere designation and
construct is not causally conditionedis cited as a counter-example for w hat
ever is causally conditioned, and hence transient and instantaneous (ksa-
nikaJ ;247 but the perishing (vinda) of things is not due to some separate and
self-existent cause, for it is precisely the nature of conditioned things to perish.248
Candrakirtis comment on the first chapter of the MMK provides an impor
ta n t account of the Prsagikas approach to the negative statem ents of the
Madhyamaka (MMK i. 1). Since all things are unproduced from self, etc. not
only on the level of the paramdrtha (on which point Bhvaviveka of course
agrees) but also on the surface level of samvrti (since the concept of the pro
duction of substantial self-existent entities through causes and conditions is
found to be antinomic and onto-logically untenable), Bhvavivekas restriction,
in his independent inferences and syllogisms, of non-production to the para
mdrtha level is unacceptable.249 Candrakirti adds th a t the qualification in

245 p p Xxii. 1 2 , x x v ii. 29 and xviii. 8 (cf. xiii. 3, x x iv . 1).

248 P P i. 1 (p. 64. 2), v. 2, x x iv . 21.
247 p p v jj 3 2 - cf. xviii. 1 (p. 343), x x v . 4.
248 P P vii. 32 w ith i. 1 (p. 29. 45) and x x i. 4.
249 Cf. how ever P P xiii. 2 (p. 453. 5): itattvataK ...param drthatah svabhavato na
vidyante (speaking o f th e kleJas).
The Middle Period: The Systematization of the Madhyamaka School 77

reality (paramdrthatah) could not possibly be taken to refer exclusively to the

doctrines of causality held by non-Buddhist sectarians (tirthika) because these
doctrines are invalid even for the surface level. Yet people do in fact commonly
speak of an effect proceeding from a cause; and so long as he does not attem pt
to transform this conventional usage into a philosophical system postulating
the causation of substantial entities the Madhyamika may do so likewise. (In
deed N grjuna has himself proceeded in this way following the Buddha, who
is quoted as saying250 th at, although people dispute with him, he does not
dispute with them inasmuch as he assents to what is agreed in the world and
does not assent to what is not so agreed (PP xviii. 8).) The qualification para-
marthatah introduced by Bhvaviveka is therefore quite without justification
(P P i. l ,p p . 2627).
Candraklrti points out th a t Bhavavivekas introduction into his independent
inferences and syllogisms of the qualification paramdrthatah in order to avoid
the negation of production on the relative level is also faulty technically. Por
the inferencena paramdrthata ddhydtmikdny ayatandni svata utpannani,
vidyamdnatvat (pp. 2526)in which it is introduced will have a faulty thesis
(paksadosa), the subject of the propositionthe eye and the other internal
basesbeing (in different ways) unreal and unestablished (asiddhadhdra) for
both its proponent and opponent. I t is so for its proponent, the Madhyamika,
since he does not accept such an entity produced from self, etc., and the subject
is thus em pty (null) for him. And Candraklrti observes th a t it would be of no
avail in this m atter to m aintain th a t the subject of the proposition is established
relatively (samvrtya) because the question then arises as to what the qualifica
tion paramdrthatah relates t o ; if it has been used to qualify the negation of pro
duction the formula should have been sdmvrldndm caksurddindm paramarthato
nsty utpattih, b ut this is not what Bhvaviveka has in fact written. In any
case, Bhavavivekas inference remains faulty in its thesis from the point of
view of the opponent against whom he is arguing because, for this opponent,
the subject of the proposition is real (dravyasat) and an inner base unproduced
in reality is therefore meaningless for him (pp. 2728).251In addition, Bhava
vivekas svatantrdnumdna also contains a faulty reason because its ground or
logical Tensnvidyamdriatvdt (p. 26.1) or sattvat (pp. 30.15, 33.4) because it
existsis defective. This is so because, if existence refers here to the relative
surface-level (samvrti) only, the logical reason will be unestablished (asiddha)
from the point of view of its proponent, who does not actually accept the
reality of the thing. And if existence related here to the level of ultim ate reality
(paramdrtha) the logical reason would be unestablished or contradictory

250 S am yu ttan ik aya iii, p. 38.

251 T hat is, to th e opponent in th e debate, who presupposes th a t th e eye,
etc., are real (dravyasat) , non-production o f things is n ot acceptable; and th e
logical fau lt o f unestablished subject is th en usually considered b y th e logicians,
w ith w hom B h av av iv ek a had allied him self, to be a fault in an argum ent w ith such
an opponent.
78 David S e y fo r t Ruegg The Literature of the Madhyamaka School

(viruddha) from the point of view of the proponent, since the Madhyamika
(including Bhavaviveka) does not in fact accept the real production of any thing
(PP i, p. 31, which refers to MMX i. 7).
W ith regard to the status of a propositional thesis (pratijna) in Madhyamaka
thought, it is of course recognized by Candrakirti (PP i, p. 16) th a t Nagarjuna
clearly stated th a t he maintains no pratijna.2*2 Yet, in his interpretation of
MMX xxi. 2,253 Candrakirti describes the first half-verse as a pratijna, the next
quarter-verse as the adducing of the undesired consequence (prasangapadana),
and the last quarter-verse as the conclusion (nigamana).2bi B ut this explana
tion is not intended to reject the usual M adhyamaka standpoint; for not only
does Candrakirti make it clear th a t we have to do here with a prasanga but
there is no suggestion th a t Nagarjuna was trying to establish the existence of
some entity presented as the subject of a proposition. In other words, notw ith
standing the terminological difficulty raised by the use of the word pratijna,
Candrakirti evidently does not repudiate here what he expressly stated in the
first p art of the Prasannapada, following what Nagarjuna had himself said on
the subject.255
As to the logical status of the prasanga method, and in reply to Bhavavivekas
criticism of B uddhapalitas prasanga-type argument as being open to objection,
Candrakirti writes as follows: How could it be, as is suggested [by Bhavavive
ka], th a t the master Buddhapalita who followed the faultless doctrine of the
m aster Nagarjuna, has propounded a statem ent th a t is open to objection
(savakadavacana), so th a t an opponent would be in a position to impugn it ?
When the advocate of the doctrine th a t entities are without own being adduces
a prasanga against the advocate of the doctrine th a t entities have own being,
how could there exist [for him] the occurring of some thing contrary to the
prasanga? For words do not reduce him who utters them to dependence, in the
manner of policemen armed with sticks and fetters! Rather, when they have
semantic capacity they conform to the intention of the speaker. Therefore,
because the application of the prasanga results exclusively in the negation
(pratisedhamdtra) of the opponents thesis, there can arise nothing th a t is
252 See V igrahavyavartani 2930. Candrakirti also quotes in this connexion
A ryedevas Catuhsataka x v i. 25: sadasatsadasac cSti yasya pakso na vidyatej
upalambha cirendpi tasya vaktum na 6akyatef / H o who has no thesis postu latin g
existence, non-existence and b oth existen ce and non-existence cannot h av e an
objection levelled against h im .
253 H ow w ill there be destruction w ith ou t production ? [Were th is the case, there
w ould be] death w ith ou t birth. There is no destruction w ithout production.
264 For Candrakirtis use o f the word pratijna see also P P iv. 2 and viii. 1 and 7;
for nigamana see v. 5 (and v. 6 [p. 134. 5], vi. 10, ix. 14, and x x v ii. 8 for nigamayati).
The prasangapadana has been explained in P P i. 1, p. 24 (see above, p. 36 n ote 93;
b elow ).
255 T he word pratijna, can m ean either proposition or th esis (serving to
assert som ething o f an en tity). I t is th e second use o f the word th a t can find no
legitim ate application in the M adhyam aka according to N agarjuna and Candrakirti ;
cf. also D . S e y f o r t R u e g g , J I P 5 (1977), p. 4950.
The Middle Period: The Systematization of the Madhyamaka School 79

contrary to the prasanga\2b8 Moreover, in contradistinction to relative (i.e.

presuppositional and implicative) negation (paryudsapratisedha), the Mdhya-
m ikas absolute non-presuppositional negation (prasajyapratisedha) does not
commit him to accepting the contradictory proposition concerning the existence
of any entity ; and its force is fully exhausted in the annulment of the opponents
proposition, as already observed.
In his discussion of the employment in philosophical argument of indepen
dent inferences (svatantrnumdna) and syllogisms (prayogavdkya) Candraklrti
has furthermore made a distinction of considerable interest between what we
could perhaps call logical referential presuppositions for the interpretation of a
proposition and pragmatic use of a proposition. He recalls th a t the logicians
accepted as a binding rule th a t in an anumna i t is necessary in any case to
reason with reference to entities the existence of which one accepts oneself
(svaprasiddha) (whether the reality of these entities has to be presupposed also
by ones opponentubhayaprasiddha, p. 35is another question still). And a
proper anumna is not valid so long as it is founded only on the opponents
presuppositions (paraprasiddha, p. 3435). On the contrary, for the philosopher
employing the above-mentioned form of reasoning th a t simply adduces an
undesired consequence in anothers argum entthe prasangpddanathe situa
tion is altogether different according to Candraklrti; for as the subjects of his
sentences, which contain non-presuppositional prasajya-negations, this philo
sopher does not have to take entities which ho himself supposes to be referen
tial. And his prasangdpddana indeed has as its sole end the annulment of the
opponents thesis (parapratijndnisedhaphalatva, p. 34) without its user being
committed to any counter-thesis concerning the existential nature of the
entities in question. A Prasangika is accordingly fully entitled to reason taking
as his point of departure only his opponents existential presuppositions; and
he is then able to annul his opponents statem ents solely on the basis of argu
ments th a t the latter has himself accepted (p. 34).
Now, in the light of this distinction, the referentiality of the subjects of their
sentences and the truth-value of their propositions become altogether prob
lematical for the Svatantrika, who m ust use his svatantrnumna and prayoga-
vdkya in conformity with the established rules, but no t for the Prasangika. For
whereas the terms of the formers propositions should be svaprasiddha for him
in order to fulfil the accepted logical requirements, they are in fact just as em pty
(null) and non-referential for the Svatantrika as they are for any other Madhya
mika.As for the teaching of the Buddhas, since they adopt a transactional and
pragmatic procedure purely m otivated by their desire to help (: anugraha)
others (p. 36), the terms of their statem ents need also not be dependent on
existential presuppositions of their own concerning the referentiality of any
entity being talked about. And they need take into account^-in a so to speak
purely therapeutic manneronly their auditors (mis)conceptions. The method
25 p p ^ p 24 (quoted above, p. 36 note 93); cf. p. 34. 5 and p. 13 (quoted
above, note 94).
80 David Seyfort Ruegg The Literature of the Madhyamaka School

of their teaching is thus ad kominem in the best and most fitting sense (cf. P P
i. 1, p. 5758; xviii. 8). The criterion is then the pragmatic one of the appro
priateness of a sentence in a given pedagogical situation.
W ith respect to the theory of correct knowledge (pramdrw) having a corre
sponding object (prameya), Candraklrti follows N grjunas critique as set
forth in particular in the Vigrahavyavartani (31 sq.).257 Candraklrtis rejection
of the logico-epistemolgica! schools doctrines extends to its whole theory of
the existence of only two forms of pramana together with their corresponding
prameyasnamely pratyaJcsa relating to the particular (svalalcsana) and
anumana relating to the universal (samdny<aksana)ior the relationship
between them is in no way different from th a t between a lahsana and its
laJcsya which has already been criticized by Ngrjuna (PP i. 1, p. 61 ).258 In
his epistemology the Prsagika then proceeds in what at first appears to be
a quite naively realistic manner, taking what is accepted in worldly consensus
as the basis for his logical procedure (nydya) (PP i. 1, p. 35) ;259 yet, based as
it is on a prenetrating critique going back to Ngrjuna and Aryadeva of the
generally accepted logical and epistemological theories, the Prsagikas
approach is in fact anything but naive or simplistic.260
Candraklrti also rejects the form of knowledge additional to the forms
derived from the five physical sense-faculties and the m ind and known as
self-awareness (svasamvedana or svasamvitti) which was accepted by the
Sautrntikas and Vijnavdins, inclusive of Digngas school, and later by
the Yogcra-Mdhyamikas (PP i. 1, p. 61 sq.).281
For all his rejection of the pluralistic realism of the Sarvstivdins dharma-
theory, the Prsagikas criticism led him to adopt in m atters of epistemology
(if not of ontology and gnoseology) a position th a t is no t unrelated to the
Sarvstivdins.262 For people in the world, Candraklrti observes, understand
ing of a thing is established through four means of knowledge. These are
257 PP i, pp. 5575.
258 E .g . in MMK, Chapter v.
259 H ence the nam e M adhyam aka which goes along w ith th e w orldly consensus
(lo k a p ra sid d h iy given b y som e T ibetan scholars to B ud d h aplitas and Candra-
kirtis school jig rten grogs sde spyod p a 'i dbu m a; see Bu*ston, Chos*byun, fol. 103a
[ii, p. 135], and above, pp. 76, 79; this term inology appears also in the Grub*
p a ,i ,m th a ,m am *par,bsad*pai ,mj/od, fol. 9b, b y d B u sp a 'B lo gsal [fourteenth cen
tury]). B u t see m K has'grubdG edegs'dpalbza, sT o'thun'chenm o, fol. 4 1 b 42a,
on the appropriateness o f this term inology.
260 I t m igh t be described rather as nom inalist (and sceptical in th e sense o f n ot
adopting dogm atic speculative view s).For a later defence o f th e existence o f
external objects b y ubhagupta (eighth century, betw een Dharm akirti and
Dharm ottara) see his B hyrthasiddhi (cf. M. H a t t o r i , I B K 8/1 [1960], p. 9 sq.,
and N . A i y a s w a m i S a s t r i , Bull, o f T ibetology 4 [1967], pp. 196). This position
was criticized n ot only b y th e V ijnavdins but also b y th e Yogcra-M dhyam ikas
(cf. ntarakita, T attvasam graha, verse 1971 sq.).
261 Cf. MA vi. 72 sq.
262 This is a com plicated question th a t has been m uch discussed. Cf. MA xiii. 12
(p. 4067); m K h a sgrubdG elegsdpahbza, sT onthm rchen-m o, fol .42b.
The Middle Period: The Systematization of the Madhyamaka School 81

direct perceptual knowledge (pratyaksa); inference (anumdna), i.e. the form

of knowledge having as its object what is not directly perceptible (paroksa)
and produced by an inferential mark th a t does not deviate from what is to be
established (sddhyavyabhicdrilinga); scriptural testim ony (agama), i.e. the
statem ent of reliable persons (apta) who directly (sdksdt) know things in
accessible to the faculties (atindriydrtha); and analogy (upamdna) , i.e. under
standing through similarity (sddrya) as when we say th at the gayal is like
the cow (PP i, pp. 71-75).
At the end of each chapter of the P P are found extracts, often very lengthy
ones, from a number of Mahynastras (see above, p. 7) which help to ex
emplify and corroborate what has been stated in the stanzas of the MMK and
Candrakirtis comment. In several cases the Tibetan translation of the P P
does not contain all the quotations included in La Valle Poussins edition;
and it is therefore possible th a t some of them are additions to Candraklrtis
tex t made by the commentarial tradition (see for example La Valle Poussins
notes on pp. 388, 428 and 539 of his edition; compare also his note 2 on p. 336).

Vrttis by Candrakirti are available in Tibetan translation on N agarjunas

nyatasaptati and Yuktisastik. Candrakirti also wrote an extensive and
im portant Tik, extant in Sanskrit fragments and a Tibetan translation, on
ryadevas Catufiataka; unlike his predecessor Dharmapala, whom he terms
a poet {man dnags mkhan, fol. 34a) and implicitly accuses of having interpreted
Aryadeva in a manner th a t is contrary to the principles of the Madhyamaka
(fol. 34b), he commented not only on the second portion of the work dealing
with dharmanihsvabhdvatd but alsG on the first p art containing an exposition
of the Dharm a which he considered an essential p art of the work (fol. 34 a). In
his commentary on the Catubataka Candrakirti alludes to examples (dpe) set
forth by Dharmadasa, evidently another Vijnavdin who turned his atten
tion to ryadevas work (fol. 35a).263 Candrakirti has referred in this Ttk to
his own M adhyamakavatara vi. 80 (fol. 63a).264

Candrakirtis Pacaskandhaprakarana is an independent treatise dealing

with the factors subsumed under the five groups (skandha) and with the
related topics of the twelve ay ataas and eighteen dhatus,265
The *M adhyamakvatrapraj or *M adhyamakaprajvatara translated
into Tibetan by its author together with Gos*khug*pa*lHa(s)*bcas must be by
another Candrakirti who lived much later, in the eleventh century.
On Candrakirti-pada, the author of im portant Tantrik works, see below,
p. 105.

263 See above, note 148.

264 On th e C atuhataka-tk see S. Y a m a g u c h i , Chgan bukky ronko (Tokyo,
1965), p. 167 s q .; B ukkygaku bunsh, ii (Tokyo, 1973), pp. 259 sq., 349 sq.
265 Cf. S. Y a m a g u o h i , B ukkygaku bunsh, ii, p. 437 s q .; C. L i n d t n e r , AO 40
(1979), p. 87 sq.
82 D a v id Seyfort R uegg The Literature o f th e M adhyam aka School

Santideva, the next great representative of the (Prasangika ?) Madhyamaka
school, flourished probably in the first part of the eighth century. He is reported
to have been bom as the son of a king of Surastra and, like Aryadeva, to have
renounced his royal estate for a life of religion.266
Following on the elaboration and systematization o f Madhyamaka thought
carried out by his great predecessors, &antideva made a significant contribu
tion to the school by taking the perfections (paramitd) of the Bodhisattva as
the focus for his exposition of the Mahayana, which he presents as the union of
praxis (that is, compassionate method) and theoretical intellection, the two
inseparable and coordinate components of the mind of Awakening (bodhicitta) .
Thus, while Candrakirtis Introduction to the M adhyamaka(-Sastra) is ar
ticulated round the ten stages of the Bodhisattva with their corresponding
pdramitds, Santidevas Bodhicaryavatara constitutes an Introduction to the
practice of Awakening founded on the great pdramitds of the Bodhisattva,
the first five of which are auxiliary to and guided by the sixth, namely the
perfection of transcending discriminative understanding (prajnapdramitd).
The Bodhicaryavatara comprises the following chapters: (i) The advantages
of the bodhicitta in its two forms (viz. the B odhisattvas preliminary resolution
pranidhiand its realizationprasthdnathrough the perfections on the
B odhisattvas path proper, 1517); (ii) Offerings ( puja), homage (vandand)
and confession (papadeAand); (iii) Assumption of the bodhicitta; (iv) Absence
of distraction in cultivating the bodhicitta; (v) Maintenance of awareness of
mind; (vi) The Perfection of patient capacity (Jcsanti); (vii) The Perfection of
energy (virya); (viii) The Perfection of meditation (dhyana); (ix) The Prajna-
param ita; and (x) The dedication (parindmana) of m erit.267
In th is work w hat might be called the mystical aspects of Madhyamaka
thought come to the fore in connexion with the exposition of the purification
of the spirit and contemplation, the B uddhas sustaining power (buddhd-
nubhava, i. 5, which works as a predetermining forceadhisthdnar-according
to the commentator Prajnakaram ati), and the theory of ultim ate reality free

388 S eeB u 'sto n , Chos-byun, fol. 113b (ii, p. 161); Taranatha, rGya*gar*chos,byuh,
p. 125 sq. Cf. A. P e z z a e i , Santideva (Florence, 1968) and J. W . d e J o n g , I I J 16
(1975), pp. 163, 17980.S an tid evas BCA is quoted in Santaraksitas T attvasiddhi,
fol. 3 8 b 67 ( = BCA i. 10) and 3 8 b 7 8 ( = BCA vii. 28).
287 A recension o f th e BCA is reported to have lacked Chapters ii and ix (see
Taranatha, rG ya-garchos-byun, p. 127). In th e lD a n d k a r m a Catalogue (no. 659)
the ex ten t o f th e BCA is given as 600 lokas (in tw o bam po), rather than the 1000
verses indicated b y B u-ston, Chos-byuh, fol. 114b 1, 159a). B u-ston discusses th e
discrepancy and states th a t a difference results from om itting th e papadegand
chapter (ii); and he observes th a t som e have ascribed Chapter ix to a certain
Blo*gros*mrzad'pa ( A k?ayam ati). Cf. Taranatha, op. cit., p. 127; J. W . d e J o n g ,
I I J 16 (1975), p. 181.On th e question o f the incom plete Chinese version o f th e
BCA (ascribed to N agarjuna, Taisho 1662) see S. L v i, B E F E O 1902, p. 253 s q .;
L. d e L a V a i x e P o u s s i n , Mus6on 4 (1903), p. 313 sq.
The Middle Period: The Systematization of the Madhyamaka School 83

from the dichotomy of existence and non-existence. W ith respect to the last
point the Bodhicaryavatara contains a verse considered to be a summing up
of the central idea of the Middle W ay (ix. 35): When neither existence nor
non-existence presents itself before the mind, then, being without any support
[to which to attach itself] because of the absence of any other course [for
thinking],288 the [mind] is still. Noteworthy also is the final chapter on dedica
tory transfer (parinarnand) by the Bodhisattva, who assigns all m erit ac
cruing to him to the Awakening of all sentient beings without distinction. The
authenticity of this final chapter has been questioned on the ground th a t not
all the commentators have commented on it,289 but this point does not appear
to be decisive;270 in any case, the concept of dedicatory transfer is found in an
earlier chapter of the work (v. 101) as well as in the same authors Sikssam-
uccaya (xvi, p. 296), and of course also in a number of Mahyna-Stras.
Chapter ix consists essentially in a detailed study of the basic philosophical
ideas of the Madhyamaka school. We thus find in it a critique of the prarndruis;
a discussion of pratltyasamutpdda; the theory of the two tru th s271 together
with a demonstration establishing th a t the practice of the Mahynist path
is not made redundant and unnecessary by the tru th of the pdramdrthika level
of nyat; and a refutation of the Vijnavda together with critiques of
other schools of Buddhist and non-Buddhist philosophy in India, including
theism (varavda, ix. 119 sq.).272
Because it is concerned also with the religious side of Madhyamaka thought
the Bodhicaryavatara has sometimes been described as a primarily religious
and devotional poem rather than a philosophical treatise. Such a description
however curiously neglects the ninth chapterthe longest in the whole w o rk -
dealing with the prajdpramitd, which clearly places the treatise in the main
stream of Madhyamaka thought; and if due consideration is given to this
chapter it becomes abundantly clear th a t the work is hardly more religious in
any sense exclusive of philosophy than certain earlier works of the school
attributed to Ngrjunae.g. the R atnvalas well as im portant portions of
Aryadevas, Bhvavivekas and Candraklrtis treatises.

ntidevas iksasamuccaya is a work in nineteen chapters comprising a

to tal of twenty-seven verses with an extensive explanation made up largely of
quotations from canonical texts of the Mahyna and dealing inter alia with the
first five pdramitas (in Chapters i, iiviii, ix, x, and xii sq.) as constituting the
268 H ere th e principle o f th e excluded m iddle (trtiya-rai) is clearly recognized.
See above, p. 41.
289 See L. d e L a V a t l e P o u s s i n , Introduction la pratique des futura B ouddhas
(Paris, 1907), pp. 1434; cf. A. P e z z a l i , op. cit., p. 4849.
270 See below , p. 84, on Prajkaram atis reference to Chapter x.
271 On th e samvrti level Prajkaram ati distinguishes betw een tathya - and
mithyasamvrti (ix. 2) (as did B h v av ivek a et al.).
272 See above, p. 30, n ote 73a .On S antidevas attitud e tow ards vijdna
th eory see S. K a n a o k a , IB K 10/2 (1962), p. 749 sq.
84 David Seyfort Ruegg The Literature of the Madhyamaka School

B odhisattvas discipline (vinaya). This anthology of canonical texts has

preserved Sanskrit versions of a large number of Sutra passages otherwise
unavailable in an Indian language. Of particular interest from the philosophical
point of view are Chapter xii on the cittaparikarman (preparation of the m ind;
cf. Chapter vi, p. 122 sq.) and Chapter xiv which is largely devoted to a discus
sion of the emptiness of all dharmas. At the end of Chapter xv is found an
explanation of unyaia as endowed with all excellent modes (sarvakdravaro-
peta), viz. seven paramitds, mahdmaitn, mudita and upeksa, satyajndnd-
vatara, bodhicittasattvapeksd, aiayadhyaiayaprayoga, four samgrahavastus (i.e.
ddna, priyavadya, arthakriyd, and samdnarthatd), smrlisamprajanya, smrtyu-
pastMna-samyakpraMna^rddhipada-iridriya-bala-bodhyanga^astdngamdrga, and
amatha and vipafyana,273

On the basis of what Santideva has written in verses v. 1056 of the Bodhi-
cary a vatara Bu*ston274 and T aranatha275 have ascribed to him a work entitled
Sutrasamuccaya. The passage in question is not altogether clear, however, and
Nagarjuna, the author of the well-known Sutrasamuccaya, is also mentioned
in it. At all events, no work entitled Sutrasamuccaya attributable to Santideva
is known to exist; and it has therefore been concluded th a t the above-mentioned
ascription is erroneous.276
Concerning the master iantideva who is identified with the Tantrika Bhusuku
see below.

On the Bodhicaryavatara there exist a number of commentaries, the best

known of whichPrajnakaram atis Panjikais the only one completely p re
served in Sanskrit; as already noted above, it does not comment on Chapter x
although it does refer to it (in i. 33). Vibhuticandras Tatparyapanjika Vie-
sadyotani is available in a Tibetan translation (c. 1200), as are other earlier
and shorter commentaries on the Bodhicaryavatara by dG e-bafi^lha (Kalya-
nadeva?), Krsnapada (eleventh century?) and Vairocanaraksita (eleventh
century). Two abridgements of the Bodhicaryavatara connected with Dharma-
pala of Suvarnadvipa (c. 1000) and transm itted by his disciples Kam alaraksita
and Dipamkara^rijnana (to whom one of them, the Bodhisattvacaryavatara-
sattrim gatpindartha, seems to be attributed by its colophon, whereas Dhar-
278 The expression fsarvdlcarabalopetal] sarvdkdravaropeta ^ sabbdkdravarupeta
w as used earlier as an epithet o f the Sam buddha, e.g. at th e end o f the Mahapari-
nirvanasiitra 44. 10 (Pali 6. 10) (ed. E. W aldschm idt, p. 400).In th e R atna-
gotravibhaga i. 8892 and elsewhere the concept o f sarvdkdravaropeta unyatd
occupies a prom inent place in th e theory of absolute reality.
274 Chos-byun, fol. 114a (ii, p. 163), 115b (ii, p. 166).
275 rG ya-garchos-byun, p. 1278.
276 Cf. J. F i l u o z a t , J A 1964, pp. 4738; A. P e z z a l i , op. cit., pp. 8086; J. W .
d e J o n g , I I J 16 (1975), p. 182 and p. 171 (giving th e te x t o f V ibhuticandras
account o f S an tid evas life, available in both Sanskrit [ed. H a r a p r a s a d S a s t r i ,
IA 42 (1913)] and T ibetan [at th e beginning o f th e translation o f V ibhuticandras
V i^esadyotani]).
The Middle Period: The Systematization of the Madhyamaka School 85

mapala himself is indicated as the author of the shorter Bodhisattvacaryava-

tarapindartha)277 are also available in Tibetan translations by Dipamkara-
Srijnana and Chul* khrimsTgy al-ba. (The Bodhisattvacarya^utrikrtavavaxia
and Bodhisattvacaryavatarabhasya by DipamkaraSrijnana are not commen
taries on Santidevas text).

As a result of the contributions made to the school by Candrakirti and

Santideva, the doctrine of the Prasangika branch of the Madhyamaka was
firmly and systematically established from the points of view of logic, gnoseology
and soteriology. During the last part of the first millennium Candrakirtis and
Santidevas writings are not known to have inspired commentarial attention
in India; but in Tibet Candrakirtis commentaries on the Yuktisastika and
Sunyatasaptati and Santidevas Bodhicaryavatara and Siksasamuccaya as well
as B uddhapalitas commentary on the MMK were translated by the beginning
of the ninth century, when they were included in the lDan*dkar*ma Catalogue.278
In the history of the Madhyamaka school the last quarter of the first millen
nium was rather the period of the great development and glory of the Yoga-
cara-Madhyamaka synthesis, which will be the subject of the next chapter.
As for the Svatantrika school, the treatises of Bhavaviveka were being studied
especially during the third quarter of the first millennium, as is shown by the
activity carried on by his followers in India (see above, p. 67 sq.). Translations of
works of this branch of the Madhyamaka were made first into Chinese (Pra-
bhakaram itras Chinese translation of the Prajnapradipa, Taisho 1566, dates
from 6302) and later into Tibetan (from about the beginning of the ninth

277 T h e sPyod*jug*gi*donbsdus (translated b y Nag*6ho*(5hul*khrims*rgyabba) is

also ascribed to Suvam advipa(-D harm apala) b y Bu*ston (Chos*byun, fol. 1 5 9 a ) .
278 C om m entators on works o f th e Prasangika school seem to appear only
in th e late ten th century in India, first on S an tid evas (by D harm apala and
Prajnakaramati) and th en on Candraklrtis (by Jayananda, a ICaimiri master
who, however, seem s to h ave worked largely in Central A s ia ) B u d d h a p a l it a s
com m entary on th e MMK w as, however, already translated b y CogTO*Klurrgyab-
m chan (with Jnanagarbha) during th e E arly P ropagation of the D harm a (sna dar)
in T ibet; Candrakirtis V rttis on th e Y uk tisastik a and Sunyatasaptati and Santi-
d e v a s &ikasamuccaya were translated also at th a t tim e b y Ye*es*sde (with
Jinam itra, D an aiila and Silendrabodhi). Santid evas B odhicaryavatara was first
translated then b y dPal*brcegs (with Sarvajnadeva), and revised subsequently,
during th e Later P ropagation (p h yi d a r), b y R in'chen'bzafrpo ( 9 5 8 1 0 5 5 ) ,
rN ogBloddan'ses'rab ( 1 0 5 9 1 1 0 9 ) et al. Candrakirtis major w orksth e M adhya-
m akavatara, P rasannapada and Catuh^atakatikawere translated into T ibetan
only in th e eleventh century, b y Nag*6ho*(5hulkhrim sTgyal'ba (born in 1011),
Pa-6hab*iJrma*grags (born in 1 0 5 5 ) , and rNog*Blo'ldansesTab. E ven A ryadevas
Catuh^ataka w as translated only at th a t tim e, b y P a x h a b .I t should be noted th a t
B u t t o n s catalogue contained in his Chos*byun (fol. 1 5 7 b sq.) differs in certain
respects from th e inform ation given above taken from th e colophons o f th e bsTan*
gyur itself; for exam ple, B u ston m entions only rN og as the translator of the
B odhicaryavatara (fol. 1 5 9 a 2).
86 David Seyfort Ruegg The Literature of the Madhyamaka School

century a t the latest, when they are already found listed in the ID an-dkarm a
The study of both the Svatantrika and Prasangika branches of the M adhya
maka received a new impulse from Bodhibhadra, Dipam karairljnana and
their disciples and contemporaries in India and Tibet starting a t the end of the
tenth century (see below).

179 B h a v a v iv ek a s Prajnapradipa w as translated into T ibetan b y Cog*ro K lu i*-

rgyal*mchan (with Jnanagarbha) along w ith A valok itavratas Tika; and th e y are
included in th e lDan*dkar*ma catalogue. H is M adhyam akahrdayakarikas and
Tarkajvala were translated only in th e eleventh century, b y D ipam karasrijnana
and Nag*6ho (5hul*khrims*rgyal-ba; unlike th e Prajnapradipa, this work is n o t
available in Chinese. Sriguptas T attv av ata ra vrtti (included in th e lD an*dkarm a
catalogue) as w ell as Jnanagarbhas S a tyadvayavibh an ga (Karikas and Vrtti)
together w ith th e P ah jik a on it ascribed to Santarak^ita (all included also in the
lDan'dkar*ma catalogue) were already translated during th e E arly Propagation
(sna d a r) o f th e D harm a in T ibet, th e last three works being translated b y Ye*

Approximately half a millennium after Nagarjuna, and following on the

developments sketched above which resulted in the constitution and system ati
zation of the two branches of the pure Madhyamaka school, there came to the
fore a school which elaborated a synthesis of the Madhyamaka and the Yoga-
cara/Vijnanavada on the level of the philosophical analysis of surface-level
process (but not on th a t of ultim ate reality). This school is known as the
Yogacara-Madhyamaka.280 Santaraksita, who lived in the eight century, is
often regarded as both its founder and leading exponent.
This synthesizing movement was, however, not without precedent before
the time of &antaraksita. As noted above, several major Yogacarin/Vijnana-
vadin masters wrote commentaries on works by Nagarjuna and Aryadeva; and
Bhavaviveka was evidently indebted to Dignagas school for certain logico-
epistemological methods he introduced into Madhyamaka thought, although
he was an opponent of the metaphysics of the Vijnanavada and was himself
criticized by its masters. The origin of the Yogacara-[Svatantrika-]Madhya-
m aka has been traced back to Arya Vimuktisena, a master of the Prajna-
param ita and a commentator on the Abhisamayalamkara, by Tibetan doxo-
graphers who have in fact held th at the Abhisamayalamkara largely agrees in
its systemic tenor with the Svatantrika-M adhyamaka position and th a t most
of the Indian commentaries on it may accordingly be connected with th at
school. This Vimuktisena appears to have been a contemporary of Dignaga
and Bhavaviveka, and he is reported to have been a pupil of theirs and even
of Vasubandhu; he is probably to be placed therefore in the first part of the
sixth century, a couple of centuries earlier than Santaraksita.281 Srlgupta and
in particular his disciple Jnanagarbha were also forerunners of Santaraksita.282
As seen above, the early Madhyamaka developed by Nagarjuna and Arya
deva had devoted little attention to establishing a theory of the factors and
processes of the surface-level samvrti per se, the concept of the real production
of entities being indeed regarded as analytically and causally antinomic. For

280 In T ibetan: rN ah b y o r sp y o d p a rdb irin a. The T ibetan doxographical litera

ture refers to it also as th e Y ogacara-Svatantrika-M adhyam aka (rN ahbyor'spyod*
p aV dbuT naT aiiT gyudpa), in contradistinction to B h a va viv ek as Sautrantika-
Svatantrika-M adhyam aka, from w hich it differs n otably b y denying the existence
o f an external object ( bahyartha) and accepting self-awareness1 ( svasamvedana^
svasam vitti) .
281 On this question see D . S e y f o r t R x j e g g , W ZKSO 1213 (1968), pp. 3067.
A nd on the M adhyam aka-Prajnaparam ita synthesis see below, p. 101 sq.
282 On th e doctrinal affiliations o f these tw o masters see above, p. 67 sq.
88 David Seyfort Ruegg The Literature of the Madhyamaka School

their p art the Prasangikas considered these processes to be characterized by a

certain indeterminism ; and they strictly kept to a method of reasoning criti
cizing all speculative opinions by means of prasanga arguments which preclude
any attem pt to propound theses and syllogisms concerned with processes of
the world involving entities (bhava) conceived of as amenable to onto-logical
construction. On the other hand the (Sautrantika-)Svatantrika-M adhyamikas
were prepared to assign an a t least limited and provisional degree of (non-
pdrarndrthika) reality to the dependent origination of things without stressing
so forcefully the indeterminism of the dharmas and processes of the w orld; in
addition, they gave an im portant place in their doctrine both to independent
inference and to a particular kind of propositional knowledge regarded as
conforming to reality (the saparydya-paramartha).
On their side the Yogacara-(Svatantrika-)Madhyamikas appear to have
continued earlier trends of thought when they undertook to complement the
Madhyamaka theory of the pure paramartha both with a system of metaphysics,
epistemology and logic providing an elaborated philosophical construction and
analysis of samvrti and with a system of ethico-spiritual praxis.283 As for the
impact of the Buddhist logico-epistemological school on these Madhyamikas,
Santaraksita is known to have been influenced by Dharm akirti (seventh
century) on whose Vadanyaya he wrote a long commentary, the V ipancitartha;
and he thus shared certain antecedents with Bhavaviveka. B ut quite unlike
Bhavaviveka the Yogacara-Madhyamaka synthesized the pure M adhyamaka
which it regarded as perfectly valid and adequate with respect to the para-
marthar-w ith a form of philosophical analysis derived from the Yogacara/
Vijnanavada, a school which by the eighth century had attained a high degree
of development and whose achievements could not, it was evidently thought,
be ignored by the Madhyamika.

Santaraksita flourished in the eighth century and is reported to have been
bom in a royal line of Vanga (Bengal, at Sahor, Tibetan Za*hor). He worked
a t the great monastic seminaries of his time and became one of the most
prominent and influential thinkers in the later history of Indian Buddhism.
More is known to us about his later life because he was one of the Indian
teachers chiefly responsible for the im plantation of Buddhism in Tibet, and he
is therefore well known to the Tibetan records of the Dharma. Renowned there
as the U padhyaya and Acarya-Bodhisattva, he was the first abbot (upadhyaya)
of bSam-yas, the main Tibetan monastery at th a t time. On the basis of the
Tibetan records it may be inferred th a t he went from Nepal to Tibet for the

283 On th e question o f the bodhicittotpadavidhi according to th e M adhyam ika

and th e V ijnanavadin see e .g . Gos*g2on*mrdpal, D e b thersrion'po, ca, fol. 15a
(i, p. 272).
The Yogacara-Madhyamaka Synthesis 89

first time in about 763, and th a t he again resided there from about 775 to the
time of his death in about 788.284
Santaraksitas ordination-lineage is given as: Nagarjuna, Bhavya (i.e.
Bhavaviveka), Srlgupta, Jnanagarbha, Santaraksita.285 I t was due to him and
his illustrious disciple KamalaMla th a t the Yogacara-Madhyamaka became the
leading school of the Madhyamaka a t th at time, and so influenced very deeply
much of Buddhist thought in Tibet not only during his lifetime but for centuries
afterwards. He and KamalaIla have been counted by doxographers as Satya-
karavadins (Sakaravadins).288

Santaraksitas most extensive work, the Tattvasam graha, is a critical study

of the doctrines of the different schools of Indian philosophy, the value of
which is enhanced by the detailed commentary on it by his disciple Kamalaiila.
Both these works are extant in the original Sanskrit as well as in Tibetan
translation. The treatise begins with (iiii) a discussion of the Samkhya
prakrti and theory of causation,287 and of the concept of God (isvara);288 and
it then goes on to examine the doctrine of a world endowed with own being (iv).
Next Santaraksita takes up the theories of the abdabrahman (v), and purnsa
(vi); and he examines the views of the Nyaya, the Mimamsa, the Samkhya,
the Jainas, and the Aupanisadikas (i.e. the AdvaitadarSana according to
KamalaSila) on the subject of a self (atman), as well as the Buddhist Vatsi-
putriyas pudgala (vii).289 He then considers the doctrines of the existence of
permanent and stable (i.e. non-momentary) entities (viii), of action (karman)
and its fruit (ix), of the six categories (padartha) of substance (dravya) (x),
quality (guna) (xi), action (karman) (xii), the universal (samanya) (xiii), the
particular (viJesa) (xiv), and inherence (samavaya) (xv). The problem of
words and their intentional objects (abddrtha) is also discussed in detail (xvi).
Next Santaraksita takes up direct perception (pratyaksa) and inference
(anum dna)28Qas the means of correct knowledge (pramana) (xviixviii). And
284 See E , F r a u w a l l n e r , W ZKSO 5 (1961), pp. 1413, and G. T u c ci, Minor
B u dd h ist te x ts, ii (Rome, 1958), pp. 2831. Cf. Bu*ston, Chos^byuh, fol. 125b sq.
(ii, p. 187 sq.); sBa-bzed (ed. R . A. S t e i n , Paris, 1961); Gos*g2on*mrdpal,
Deb*thersnon*po, ka, fol. 21b (i, p. 4244); d P a -bog<3ug*lag*phrerrba, m K h as'p ai*
d ga*ston, ja, fol. 7 4b 114a; S um p a mkhan'po, dP ag'bsam 'ljonbzah (ed. S. C h .
D a s ), p. 112 and p. 49.
285 See e.g. Gosg2on*nu*dpal, Debthersnon-po, ka, fol. 17b (i, p. 34); dPa*
bogOug'lag'phreh'ba, mKhas*pai*dga,*ston, ja, fol. 103 a. Cf. G. T u c c i, Minor
Buddhist texts, ii, pp. 25 sq., 122.
289 See e.g . \Jam *dbyahs,bzadp a ,Nag*dbahbr<5on*Tgrus, GrubTntha^chenmo,
fol. 130a; P h a r p h y in #m th a ,dpyod, i, fol. l i b (p. 23). B u t see below, p. 92.
287 See W . L i e b e n t h a x , Satkarya in der D arstellung seiner buddhistischen
Gegner (Stuttgart, 1934).
288 On th e M adhyam ikas critique o f the iJvaravdda see above, p. 30 note 73 a
and p. 83.
289 See S. S c h a y e r , RO 8 (1932), pp. 6893.
290 See A. K tjnst, Problem e der buddhistischen Logik in der D arstellung des
Tattvasam graha (Krakow, 1939).
90 David Seyfort Ruegg The Literature, of the Madhyamaka School

he also discusses verbal knowledge (idbda), analogy (upamana), presumption

(arthapatti), absence (abhdva), and traditional testimony (aitihya) claimed by
some to be further means of correct knowledge (xix). In the next sections Jaina
tropology (syddvdda) (xx) and Lokayata materialism (xxii) are discussed, as
are also the theory of the existence of past, present and future time (traikalya)
(xxi)291 and the problems of the existence of an external object (bahirartha)
(xxiii), valid verbal knowledge (ridi), the eternity of W ord (abda), the
impersonal nature (apauruseyatva) of the Veda (xxiv) and its self-validation
(svatahpramdnya) (xxv), suprasensory cognition, and omniscience (xxvi).
Since these topics have been discussed with reference to the views the various
Indian philosophical schools held on them, Santaraksitas basic treatise and
Kamala6ilas Pahjika are very valuable sources for the study both of the
Sravakayanist and Mahayanist schools of Buddhist thought a t the peak of
their development and of Indian philosophy as a whole.

Santaraksitas Madhyamakalamkara, a treatise in 97 stanzas accompanied by

an auto commentary (Vrtti), is of capital importance as the basic text of the
Yogacara-Madhyamaka school.292 The V rtti refers to the Tattvasam graha, and
the treatise may therefore be regarded as one of Santaraksitas later works
representing his m ature thinking. In it we find not only th at ian tarak sita has
(like Bhavaviveka) formulated inferences to establish the Madhyamikas
principle of the insubstantiality of all factors of existence (dharma), but th a t
(unlike Bhavaviveka) he regarded the doctrine of Mind-only (cittamdtra) of
the Vijhanavada as a philosophical propaedeutic leading up to the M adhya
m ikas goal of establishing the insubstantiality of all factors including the
mind. As Santaraksita puts the m atter following the Lahkavatarasutra, On
the basis of the cittamdtra one is to know the non-existence of external things,
and on the basis of this [Madhyamaka] system one is to know complete non
substantiality; riding the chariot of the two systems [i.e. the Cittam atra and
Madhyamaka] and holding the reins of reasoning (yukti), [the philosopher]
therefore attains the sense as it is, the Mahayanist one itself. 293

291 See S. S c h a y e r , Contributions to the problem o f tim e in Indian philosophy

(Polska A kadem ia U m iejtnoci, Prace K om isji O rientalistycznej, Nr. 31, K rakow,
1938), pp. 2970.
292 On this work see Y . K a j i y a m a , Later M dhyam ikas on epistem ology and
m editation, in M. K i y o t a , d., M ahyna B uddhist m editation (Honolulu, 1978),
pp. 11443. For a brief analysis of it see M .Ich ig , IB K 20/2 (1972), pp. 995989.
293 M adhyam akalam kara 9293 (M adhyam aklam kravrtti, fol. 79ab ).The
L ankvatrastra passage (x. 2567) reads (in V a i d y a s edition) :
cUtamdtram samdruhya bdhyam arlham na lcalpayet\
tathatdlambane sthitvd cittamdtram atikramet\\
cittamdtram atikramya nirdbhdsam atikramet\
nirdbhdsasthito yogi mahdydnam sa payati\\
(N a jj reads na payate).On the follow ing quotation o f th e Yukti$atika see
above, n ote 44.
The Yogcra-Madhyamaka Synthesis 91

The principle dem onstrated in this treatisefirst by means of reasoning

(yukti, verses 262) with respect to the analysis of knowledge and its objects
and then with regard to the two truths following philosophical tradition
(idgama, 6397)is the basic one th at all entities (bhdva) held to be real by
either Buddhists or others are in fact without self-nature (nihsvabhava).***
To prove this by reasoning ntaraksita adduces as the logical reason the
argument th a t entities have neither a unitary nature (ekasvahhdva) nor a
multiple one (anekasvabhdva), the homologous example here being a reflection
(pratibimbavat). Now, taking the axiom25 th a t an entity would indeed by
definition have to be by nature either unitary or multiple as his basic argument
by which to establish th a t there exists no thing having self-nature (svabhava),
he proceeds to apply it to both conditioned (samskrta) elements such as the
paramdnu (1115) and to the unconditioned (nirvana), as well as to constructs
like the V atslputrlyas pudgala (cf. 9), the composite gross body, and all-
pervading space (aka). He observes moreover (8) th a t an asamskrta th at is
permanent and unchanging cannot have a capacity for causal efficiency
(drthakriyd), which had been established by D harm aklrti as the criterion for
existence (Pram navrttika, Svrthnum na chapter, 166).
Turning from the discussion of rupa to th a t of vijndna, ntaraksita shows
th a t it also is devoid of a unitary and a multiple nature (16 sq.). He divides
theories of knowledge into the realist variety positing a duality between
knower (cognition) and external cognized object, and the m entalis variety
m aintaining the non-duality of cognition and the object of cognition which,
accordingly, is not really an external object (bdhyartha). Examples of the realist
variety are the Vaibhsika theory th a t assumes a formless (anakara) knowl
edgei.e. one th a t per se does not have a cognitive image (dbhdsa) of the
object of cognition (visaya)and the Sautrantika one th a t assumes th at
knowledge has such a form (sdkdra)i.e. th at it does possess a cognitive image
of the cognized (and external) object. ntaraksita goes on to discuss the idea
of the reflexivity of cognition (svasamvedana self-awareness). And he argues
against the assumption th a t either a cognition th a t is imageless (anirbhdsa) or
one th at possesses such an image (sanirbhdsa) could apprehend an external
object (1621 and 2234). In addition, he considers a number of theories on
the subject drawn from other non-Buddhist systems (3540). The conclusion
then is th a t vijndna is not unitary (4143). The theory of knowledge th a t
ntaraksita next examines is the m entalis one of the Vijnavadin, who
rejects any real duality between cognition (as grdhaka) and the cognized (as
grdhya) (4460).

294 Madhyamakalamk&ra 1 (quoted in Prajn&karamatis B odhicaryavatara-

panjika ix. 2):
nihsvabhava ami bhdvds tattvatah svaparoditah\
ekdnekasvabhdvena viyogdt pratibimbavat\\
Compare th e beginning o f Sriguptas T attvavataravrtti (above, p. 68).
296 Cf. above, p. 38 sq.
92 David Seyfort Ruegg The Literature of the Madhyamaka School

The Vijnanavadins differ among themselves, however, as to whether the

non-dual vijnana contains an objective image or not; and they accordingly are
divided into the two schools of the Satyakaravada (or Sakaravada), which
accepts th a t it does, and the Alikakaravada (or Anakaravada), which holds
th a t it does not. Santaraksita records arguments against both the former
doctrine (in its three versions, 4651) and the latter one (5260). And he
finally concludes th a t in the vijnana just like any other entity there can be
neither unitariness nor multiplicity of nature (61). B ut since no third possibility
is available in the case of something supposed to have a svabhdva, the conclu
sion must be th a t vijnana is no ultimately real entity (62); on this point, then,
he agrees with the rest of the Madhyamaka school which refuses to assign a
specially privileged status to mind, which it regards as em pty of self-nature
like any other dharma. Santaraksita also subjects the Vijnanavadins theory of
the three natures (svabhdva, i.e. parikalpita, paratantra and parinispanna) to a
critique (Vrtti, fol. 68 a).296 The last part of the M adhyamakalamkara is devoted
to a discussion of the two truths. The paramdrtha is free from the four extreme
positions of existence, etc. (68), and from all discursive development (pra-
panca) (70). On this level there is no coming into existence (production), etc.,
of any kind of entity; and one cannot therefore even postulate the non-produc
tion of an entity, for such a negation could only relate to what has objective
existence, which any entity does not (7172).297 The samvrti has efficiency
(arthakriydsamarthya)> however (64); and it allows the use of inferences
(anumdna, 75) and logical proofs (7678). In Santaraksitas system eternalism
and annihilation are then both equally rejected (82). And the elimination of
affects (samkleia) and the achievement of purification (vyavaddna) as well as
the assembly of pure equipment (sambhdra) thus remain possible (8385). Com
passion (karund) finally is a distinguishing feature of those following the way
of the T athagata (96). And the Protector becomes the object of the devotion
o f those who perceive th a t other systems lack substance (97).
Although the M adhyamakalamkara and its V rtti are not available in the
original Sanskrit, a number of verses from the mvla are to be found in the
296 vantarakita has been counted b y doxographers as belonging to th e S a ty a
karavada /Sakaravada branch o f th e Y ogacara-M adhyam ikas (see above, p. 89
n ote 286)as h ave his disciple Kamala^Ila and his predecessor Arya V im uktisena
because, while rejecting th e view th a t the vijn an a is an u ltim ately real en tity , he
has inclined tow ard th e theory th a t cognition is endowed w ith an im age at least on
th e level o f relativity (see V rtti, fol. 57b 3). Cf. D . S e y f o r t R u e g g , W ZKSO 1213
(1968), p. 307. B u t in the first part o f his treatise antarakita has criticized th e
Sautrantika and his Sakaravada (in its three versions, 2334) as well as th e
VaibhHika and his A nakaravada (1621, especially 20). And th e A likakaravada is
even presented as a higher philosophical sta g e ; cf. Y . K a j i y a m a , loc. cit., pp. 12032,
140. (The Sanskrit o f m an y o f th e relevant stanzas is to be found in &antarak$itas
T attvasam graha, 1999 sq.).
297 See Haribhadra, A bhisam ayalam karaloka v. 89 (p. 838) who quotes verse 72:
na ca n irvisayah sadhu prayogo vidyate nanahj vikalpdpd^rayatve vd sdm vrtah sydn
na tdttvikahU (See also op. cit., i. 27 [p. 45]).
The Yogcra-Madhyamaka Synthesis 93

Pratyaksalaksanapariks and the Bahirarthaparks of the Tattvasam graha,

while others correspond with D harm akrts Pram navrttika on which work
ntaraksita drew in appropriate places in his treatise;28 it has also been
quoted by H aribhadra in his Abhisamaylamkrlok and by Prajnkaram ati
in his Bodhicaryvatrapajik. Both works are available in Tibetan transla
tion. On this treatise there exists in Tibetan translation an extensive Panjik
ascribed to Kamalaila.299

The Satyadvayavibhagapanjik, a commentary on Jnnagarbhas Satya-

dvayavibhanga, is ascribed to ntaraksita.300 I t is available only in a Tibetan
In addition to the works by ntaraksita included in the Madhyamaka
section of the bsTan-gyur, other sections of this collection contain treatises of
his dealing with logice.g. the Vipancitrth on D harm akirts Vdanyya
which is available also in Sanskritand Tantrae.g. the Tattvasiddhi (which
quotes ntidevas Bodhicaryvatra).301
ntaraksitas Samvaravimakavrti, a work linked with Candragomins
Bodhisattvasamvaravimaka, is of special interest as an example of a Yogcra-
Mdhyamika treatise th a t lays particular emphasis on the Bodhisattvas
ethico-spiritual praxis (caryd). I t is to be observed th a t the Samvaravimaka
is related to the teaching of the Bodhisattvabhumi, a basic tex t of the Yog-
cra/Vijnnavda school, and th at (unlike e.g. the ikssamuccaya) it there
fore does not belong to the Madhyamaka tradition properly speaking.302

Kamalaila (c. 740795) was the disciple of ntaraksita. As mentioned
above, he wrote an extensive and valuable commentary (Panjik), now avail
able only in Tibetan, on his teachers Madhyamaklamkra. As for his Panjik
on ntaraksitas Tattvasam graha, it is as already noted an invaluable source
for the history of Indian philosophy, non-Buddhist as well as Buddhist; it
contains the first known reference in a Buddhist treatise to the Advaitadarana,
which Kamalaila mentions in his chapter on the Aupanisadikas doctrine of
the alman (on Tattvasam graha 328).803
m M adhyam aklam kra 8 corresponds to P ram navrttika i. 211, and M adhya-
m aklam kra 26 a b corresponds to P ram navrttika iii. 1 3 8 cd (cf. Tattvasam graha
** This com m entary is said not to be b y K am alaila in m K h asgrubdGedegs*-
dpal'bzan, rG yudsd e'sp yirnam (ed. L essing-W aym an), p. 90. lCaskya*Rol*-
p a rrdoT je, Grubmtha*, kha, fol. 6 b, states th a t D harm am itra has ascribed it to
Kam alala(! ?).
100 See above, p. 68 and n ote 224.
301 See above, note 266.
302 See above, p. 88.
303 Cf. H . N a k a m u e a , H JA S 18 (1955), p. 104.Concerning K am ala iila s
citations o f his predecessors, see E . S t e i n k e l l n e r , W ZKSO 7 (1963), p. 116 sq.
94 David Seyfort Ruegg The Literature of the Madhyamaka School

Kamalaila was invited to Tibet to carry on ntaraksitas work there; and

like his master he played a leading part in the establishment of Buddhism and
the Yogcra-Madhyamaka in Tibet towards the end of the eighth century.
His role a t the Council of Tibet (or council of bSam*yas) in opposing the
sim ultaneis teachings associated with the Chinese Hoshang Mahyna has
often been regarded as decisive by Tibetan w riters;304 but their accounts have
been questioned in some recent research on the subject.305 At all events, and
whatever the exact historical part played by Kamalaila may in fact have
been during this famous debate on the process of attaining Awakening, his
gradualist theory on the subject is amply set out in his treatises on progressive
realizationthe three Bhvankram asin which he deals with the interrelation
between calm (amatha) of the mind and intellective insight {vipayan or
prajd) with respect to the understanding of reality.306 He died in Tibet in
about 795.307

Kamalaila^ most extensive independent work is his Madhyamakloka,

which is extant only in Tibetan translation. Based on both scripture (dgama)
and reasoning (yulcti), this treatise expounds in detail the theory of the non
substantiality of things according to the Yogcra-[Svtantrika-]Madhyamaka
system which m aintains the doctrine of non-production in reality and of pro
duction on the surface-level. The logical and epistemological problems con
nected with this theory are extensively discussed in this work, which is there
fore one of the most im portant of the Indian Madhyamaka treatises from these
points of view. Much attention is given also to soteriology, and in particular to
establishing th a t the teaching of the one vehicle (ekayana) is of certain mean
ing (ntrtha) by contrast with the opposed doctrine of three ultimately

104 F or a stu d y o f th is and connected even ts see P . D e m i b v i l l e , Le concile de

Lhasa (Paris, 1952); G. T u c c i, Minor B uddhist tex ts, ii (Rome, 1958); Y . I m a e d a ,
J A 1975, pp. 12546. See also th e older article b y E . O b e r m i l l e r , JG IS 2 (1935),
pp. 111, and his preface to his publication o f K am alasilas third Bhvankram a in
P am jatniki literatury naradov vostoka, T eksty, Malaja serija x v i (Moscow, 1963).
806 On this question see D . TJe y a m a , Th gakuh 35 (1964); P . D e m t v i i x e ,
T P 56 (1970), p. 40 sq.; Y . I m a e d a , JA 1975, p. 1 4 0 -1 .
808 Compare in th e M adhyam aka section o f th e bsTan*gyur tw o works b y
V im alam itra concerned w ith this problem : th e Cigcar*jug*pa*rnam,par,mTtog*
p a i'bsgonrdon and th e Rim*gyis*jug*paVsgom*don (below, p. 107).The T ibetan
critics o f the H osh an g M ahayna have attributed to him a teaching th a t is more or
less quietistic; b u t this description does n ot altogether ta lly w ith w hat is known
from th e Chinese records (studied b y P. D em iville and his successors).K am ala-
ilas first and third Bhvankram as are preserved in Sanskrit and h ave been
edited b y G. T u coi, Minor B uddhist tex ts, iiiii (R om e, 1958 and 1971) (see also
above, n ote 304). The first exists also in a Chinese translation (Taish 1664).
F or bibliographical particulars see A. Y u y a m a , I I J 17 (1975), p. 265 sq.
807 See E . F r a t j w a l l n e r , W ZKSO 5 (1961), pp. 1434. (According to som e
T ibetan accounts o f th e even ts, Kam alaila m et his death as th e victim o f his
defeated opponents; see e .g . Bu*ston, ChosMbyun, fol. 129b [ii, p. 196]. Cf. P .
D e m t v i l l e , Le concile de Lhasa, p. 11; G. T u c c i, Minor B u d dh ist tex ts, ii, p. 45).
The Yogcra-Madhyamaka Synthesis 95

distinct vehicles of liberation. In this connexion KamalaSila discusses also the

gotra and tathdgcUagarbha theories, thus assuring them a much more prominent
place in later Madhyamaka thought than they had occupied in the works of
the earlier Madhyamikas.300
As is to be expected of a treatise belonging to this school, the Madhyamaka-
loka contains discussions of Vijnanavada doctrine. Regarding the three natures
the parikalpita0, paratantra0 and parinispanna-svabhdvar-accepted by the
Vijnanavada, it is argued th at they have a propaedeutic function in philosophy
because an understanding of the first two may help one to avoid the twin
philosophical pitfalls of im putation (samaropa) and denial (apavada); yet the
last two are not to be taken to be ultim ate realities, and any canonical state
m ent representing them as such in fact constitutes a statem ent of provisional
indirect meaning (neyartha) th a t is simply m eant to introduce philosophically
inexperienced and fearful beginners to an understanding of non-production
and non-substantiality (avatdranahhisamdhi) (fol. 166 b). Similarly, the doctrine
of cittamatra may assist one to understand th at no external object (bdhydrtha)
separate from cognition exists corresponding to the object of cognition, and
th a t the supposed duality of cognition (grahaka) and cognized object (grahya)
is false; but any canonical statem ent representing mind to be an entity th a t
exists in ultim ate reality is merely a statem ent of provisional indirect meaning

808 K am alaila was perhaps the first o f the leading M adhyam aka masters to
incorporate the theory o f th e tathgatagarbha into one o f the m ain schools o f
M adhyam aka th ou ght. (Bub on R hulabhadra and N g a see above, pp. 6467.)
T he tathgatagarbha and gotra theories occupy an im portant place also in the
w ritings o f D harm am itra (fl. c. 800 ?), and th en in those o f A bhaykaragupta and
other later M dhyam ika masters (see th e n ex t chapter).
Since som e o f th e problem s connected w ith the tathgatagarbha, the gotra and the
ekaydna figured prom inently in the history o f B uddhist thou ght in Central A sia (e.g.
K h otan ; cf. the Book of Z am basta) and China where the Y ogcra-M adhyam aka
is n o t known to have been an influential school, th e significance attach ed to these
philosophical topics cannot, however, be assum ed to stem from the Y ogcra-
M adhyam aka exclusively. This com plex o f doctrines is in fact already attested in a
num ber o f Stras as well as in th e R atn agotravibhga, w hich is n ot frequently
quoted b y the Indian scholars o f the M adhyam aka un til a quite late date ( 1 0 t h
11th century). These doctrines have been especially carefully studied in th e T ibetan
exegetical literature ; and the im portance th ey assum ed in T ibet m ay indeed be due
largely to the influence o f th e Yogcra-M adhyam aka school in th a t country (e.g. in
th e domain o f th e Prajnpram it-Madhyam aka syn th esis; see th e n e x t chapter).
The Indo-T ibetan tradition based on the R atnagotravibhaga, the fundam ental
Indian stra on the tathgatagarbha, however appears to have its origins else
w here.As for the ekaydna theory, it is closely linked w ith the doctrine o f the
tathgatagarbha and the prakftistha-gotra. Cf. D . S e y f o r t R u e g g , L a thorie du
tathgatagarbha et du gotra (Paris, 1969), p. 177 sq. ; Le trait du tathgatagarbha de
Bu*ston Itin-chen grub (Paris, 1973), pp. 9, 27 n ote, 69, 142 sq.; T he gotra , ekaydna
and tathgatagarbha theories o f the Prajnpram it according to D harm am itra and
A bhaykaragupta, in: Prajnpram it and related system s (Studies in honor o f
E . Conze, B erkeley, 1977), pp. 283312; A. K u n s t , Som e aspects o f th e E kayna,
ibid., pp. 31326.
96 David Seyfort Ruegg The Literature of the Madhyamaka School

m eant to introduce the novice to the theory of pitdgala-nairdtmya (avatdrand-

bhisamdhi) (fol. 170 b). The cittamdtra doctrine of the Vijnnavda is accord
ingly nothing but a step on the way toward comprehension of the paramdrtha.
This interpretation is supported with quotations from the Lankvatrastra,
as was already the case in Sntaraksitas M adhyamakalamkra-Vrtti.309 I t is
furthermore to be noted th at, for the Vijnnavda, the tathdgatagarbha teaching
refers only to tathatd, thus making possible the theory th at there exist three
ultim ately distinct vehicles of liberation (fol. 159 b);310 but, as already noticed,
Kamalala rejects this last interpretation in favour of the theory of the
ekaydna, which he holds to be of certain meaning and according to which all
sentient beings (sarvascUtva) will attain Buddhahood. Kamalala thus differed
significantly from the Vijnnavda theory th a t understood the triydna as
three altogether separate paths leading to three ultim ately different forms of
bodhiy and th at also m aintained the existence of a category of persons who are
permanently excluded from attaining Awakening.

Kamalalas first Bhvankram a (in the arrangement of the bsTan-gyur)

begins by specifying th at compassion (Jcrpd, karund), the bodhicitta and the
Bodhisattva's practice (pratipatti) are the triad of factors leading to the omni
science (sarvajnatd) of the buddha. Compassion is described as the root (mula)
of all buddhadharmas ( 1 of Tuccis edition of the Sanskrit text) ; and K am ala
la explains how it is to be cultivated ( 2). I t is explained th a t it is a pre
requisite for the bodhicitta (3). This bodhicitta, which is described as consisting
essentially in the B odhisattvas practice, is then stated to be the seed (bija)
of all buddhadharmas (4). Such practice consists in the conjunction of trans
cending discriminative knowledge (prajnd) and means ( updya), there being no
incompatibility whatsoever between these two factors which, when brought
to completion, are known as the perfections (paramitd) ( 57). This practice
leads then to apratisthitanirvdna, i.e. the nirvana in which the Bodhisattva
does not fix himself because of his full complement of compassionate means,
although by prajnd he will have eliminated all errors th a t would fix him in
samsdra (8). The first two stages of prajnd are the one consisting in learning
(rutamayi) whereby the sense of the scriptural tradition (dgama) is retained,
and the stage consisting in reflection (cintdmayi) whereby the final and certain
sense (nitdrtha) of the teaching is distinguished from the provisional and
uncertain sense (neydrtha) and ultim ate reality (bhuta-artha, bhta-vastu-
svarpa) is realized through analysis based on both reasoning (yukti) and
scriptural tradition ( 9). The function of dgama and yukti are investigated
with respect to penetration of the non-origination (anutpdda) and non-de
struction (anirodha) of all dharmas, in keeping with the theory of the Middle
W ay ( 10). In connexion with cintd there follows a critique, according to the
See above, note 293.
See D . S e y f o r t R u e g g , La thorie du tathdgatagarbha et du gotra, p. 277
note 3; Le trait du tathgatagarbha de Bu*ston Rin*chen grub, pp. 31, 52 note, 99.
The Yogacara-Madhyamaka Synthesis 97

principle of the absence of both a unitary and a multiple self-nature (ekane-

Icasvabhdva), of the notions of material atoms (anu) and a non-material con
sciousness (vijndna), and of the duality of m atter (rupa) and mind (citta)
( 11). Next, the third stage of prajha consisting in m editative realization
(bhavandmayi) is described as leading to direct perception (pratyaksikarana)
of reality (bhuta-artha) ( 12 sq.). In the first place, calm (amatha) is to be
achieved for the sake of im perturbability of mind (cittasthirikarana) ( 13);
Kamala&la explains how this is to be brought about with reference to the nine
phases of the settling of mind, the six defects and the eight counteragents in
concentration (samddhi) ( 14) and the four meditations (dhyana) ( 15).
Following the Lankavatarasutra (x. 2568, referred to above, p. 90), the
transcending of both external objects and idealism (cittamdtra) and the topic
of non-conceptuality (nirvikalpasamadhi) and non-dual gnosis (advayajhdna)
are then taken up ( 16); and KamalaSila expounds the paramaiattvadariana
of the Mahayana (17). Through the process of understanding thus described,
the obstacles of the passions (kleidvarana) and the knowable (jheydvarana)
are both eliminated ( 18). A section is devoted to the distinction between
samvrti and paramdrtha ( 19). Kamala^Ila then returns to the subject of the
P ath th a t leads (to the goal of buddhahood) through the conjunction of discri
minative knowledge and means (prajnopayayugaruiddhavdhi margah) ( 20).
And he concludes with further observations on pratipcUti which does not
neglect analysis (pratyaveksand) (21) and with a section on the stages of the
P ath which begins with the adhimukticaryabhumi ( 22), passes through the
ten stages of the Bodhisattva and culminates in the buddhabhumi (23).
The second Bhavanakram a is similarly devoted to a treatm ent of the causes
and conditions required for achieving the omniscience of the buddha. Such
omniscience cannot be attained without the appropriate cause (hetu), for it
could otherwise always arise without any effort ever being needed. He who
wishes to attain omniscience m ust accordingly cultivate the triad of great
compassion, the bodhicitta and means (updya). This involves both completing
the spiritual equipment (sarribhara) of merit (punya) and gnosis (jndna), and
achieving apratisthitanirvana the root of which is great compassion. In order
to attain this goal, calm (amatha) and intellective insight (vipaAyana, i.e.
bhutapratyaveksd) must be realized in conjunction. Under the rubric of vipaAyand
KamalaSIla discusses the doctrine of the lack of difference in their onto-logical
status between the m aterial (rupin) dharmas and mind (citta) and the m ean
ing of the canonical statem ent The three levels of existence (dhdtu) are simply
citta\ And he observes th a t in ultim ate reality (paramdrtha) citta also cannot
be real (bden pa = satya) because it perceives rupa and the other factors which
are false (alika) in nature; moreover, citta is, as already stated, not different
in its onto-logical status from the false rupa. There follows a detailed exposition
of the procedure to be employed in meditation. Attainm ent consists in the
realization of unyata endowed with all excellent modes (sarvakaravaropeta),
i.e. Emptiness inseparable from the six pdramitas as means. Thus prajhd and
98 David Seyfort Ruegg The Literature of the Madhyamaka School

updya m ust operate simultaneously (cig car = yugapat), and the P ath conveys
through their conjunction (yuganaddhavdhin). In sum, with respect to realiza
tion of the Middle Way, the Buddha has declared th a t the gnosis of the Omni
scient has its root in compassion, its cause in the bodhicitta, and its issue in
means (compare the quotation from the Vairocanbhisambodhi found in the
Sanskrit tex t of Bhvankram a I, 7 : sarvajhajhanam karundmlam bodhicitta-
hetukam updyaparyavasdnam).
The third Bhvankram a is concerned more particularly with realization in
the form of concentration (samadhi), with special reference to the P ath th at
conveys through the conjunction of calm and intellective insight (amcUhavi-
payanayuganaddhavdh mdrgah). In fact all samadhis are subsumed under
amatha and vipayand together. Here four noetic objects (dlambanavastu) are
enumerated for meditation : (i) the nirvikalpapratibirribaka for which the image of
all dharmas and, for example, a Buddha-icon are objectified in amaiha without
any conceptualization (vikalpa) pertaining to the understanding of reality, the
dharmas, being those one has learnt and retained ; (ii) the savikalpapratibimbaka
for which there is analysis through vipayand which marks the vikalpa pertaining
to the understanding of reality; (iii) the vastuparyantatd-dlambana where Thus-
ness marked by the final limit of things (vastuparyaratlaksan tathatd) is pene
trated beginning with the first stage of the Bodhisattva, this being done in both
amatha and vipayand; and (iv) the kdryaparinispaiti-dlambana for which, as
a consequence of amatha and vipayand, there is complete accomplishment of
the required, characterized by elimination of the obstacles and associated with
transm utation of the psychic base (drayapardvrtti) on the succeeding stages
culminating in the buddhabhmi. Here amatha is defined as one-pointedness
of mind (cittaikdgratd)y and vipayand as accurate comprehension of reality
(bhtapratyaveksd) (1). Describing next in detail how these two factors are
to be practised conjointly (24), Kamalala observes th a t the form of
bhdvand wherein the mind (manas) which has stopped analysis (vicdra) is
entirely free from discursivity (nirjalpaikarasa) and flows on in virtue of its
nature (svarasavdhin) will arise only after vicdra in vipayand relating to
reality, i.e. the non-substantiality of pudgala and dharmas, ( 2, pp. 5 and 8).
Kamalala then gives a brief statem ent ( 5) and a long refutation ( 6) of the
thesis of non-mentation (amanasikdra), non-direction of attention (asmrti) ,
non-thinking, and ataraxia or quietistic non-practice of the perfections. In
conclusion, the significance of the cultivation of sarvdkdravaropetanyatd is
explained. There is to be no cultivation of mind-only (cittamdtrasevana) with
out noetic object (dlambana). Expertness in means has to be cultivated, as
does accurate comprehension of Teality; for practice (pratipatti) comprises
both prajnd and updya. This practice is necessarily progressive, gradual. Finally,
special reference is made to N grjunas teaching as comprising both scriptural
tradition and exact reasoning (7).
Between the three Bhvankramas there is considerable overlap and repeti
tion. Kamalala clearly composed his three treatises for the purpose of re-
The Yogcra-Madhyamaka Synthesis 99

futing what he considered to be serious misconceptions about the P ath and

establishing the correct philosophical theory and m editational praxis of the
Mahyna by means of concise explanations supported by quotations from
numerous Stras. And whether he in fact personally m et in debate with the
Chinese Hoshang or not (see above), it seems th a t Kamalala composed the
Bhvankram as in order to respond to and document problems under dis
cussion a t the Council of T ibet. I t is to be noted at the same time th at several
a t least of the Stras quotede. g. the Aksayamatinirdea, Gaysrsa and Vimala-
krtinirdea on the relation between prajnd and updya (I, 57, 20 and III, 6)
and the Avikalpapraveadhran on the meaning of amanasikdra (I, 17)show
th a t some of the theses being propounded by the Hoshang which Kamalala
has refuted were not being proposed for the first time in the history of Buddhist
thought; and just as the Hoshang could cite Sutra texts in support of his posi
tion, Kamalala was able to quote other canonical texts (often from the same
Stras) which take up and reject views ascribed to the Hoshang. Since they
are among the relatively few later Madhyamaka texts extant in Sanskrit, and
because they contain im portant quotations from the Sanskrit texts of Stras
otherwise available only in Chinese and Tibetan translations, the first and third
Bhvankramas now constitute our main original documentation for the study
of this school toward the end of the eighth century.

In addition, other works connected with the Madhyamaka and attributed to

Kamalala are found in the bsTan-gyur : the *Bhvanyogvatra (?) (a work
of the same type as Jnnagarbhas Yogabhvanpatha or mrga), the *Sarva-
dharm(sva)bhvasiddhi, and the Tattvloka. Kamalala also composed com
mentaries on the Saptasatik Prajiipram it, the Vajracchedik,311 and the
lis tam ba (ka) s tra.

Later Yogdcdra-Madhyamikas
The M adhyamakanayasrasamsaprakarana by Vidykaraprabha appears
to belong to the Yogcra-Madhyamaka school; following Sntaraksitas
M adhyamaklam kravrtti (92), it discusses Y uktisastik 35 in which it is
stated th a t the four elements (rnahbhta) are contained in vijndna. This work
employs the Vijnnavdins refutation of external objects and his establish
ment of the internal (i.e. subjective) only (nan cam) as a preliminary to re
futing the subjective also in accordance with the Mdhyamikas theory of non
substantiality and nycdd on the level of ultim ate truth. Vidykaraprabha
collaborated with the Tibetan translator dPahbrcegs raksita in producing the
Tibetan version of his treatise, and he may therefore be placed c. 800.

311 W hile closely connected w ith V asubandhus com m entary on the Vajracche-
dika, this work em phasizes certain concepts of th e M adhyamaka. Cf. G. T u cci,
Minor B ud dh ist tex ts, i (Rom e, 1956), pp. 21 sq., 28, 131 sq.
100 David Seyfort Ruegg * The Literature of the Madhyamaka School

Another work belonging to the current of thought th at sought to synthesize

the Vijhnavda and the Madhyamaka is the short tract entitled Nirkrak-
riks by Nandasri, a Nepalese Pandit.
Jitri (whose name is also written Jetri) is counted by doxographers as a
Yogcra-Svtantrika-Mdhyamika (Samala-Alkkra branch). His Sugata-
matavibhanga-Kriks and Bhsya deal with the four main schools of Buddhist
thought. His discussion of the Madhyamaka begins (fol. 330 b) with arguments
in favour of the principle th at things have neither a unitary nor a multiple
nature in reality (a point stressed by Srigupta and ntaraksita for example)
and th a t they are as non-substantial as a sky-lotus. Unlike the Vijnnavdin
the Mdhyamika does not postulate a vijndna existing in reality (paramdrthcUah);
and his theory is quite free of the four extreme positions (anta) th a t posit
existence, non-existence, both, and neither. In the Bhsya (fol. 354 a) Jitri
endeavours in particular to demonstrate th a t Dharm akirti was in agreement
with Ngrjuna and th a t he taught the Madhyamaka. Jitri also wrote tre a
tises on logic.812
Other late Yogcra-Mdhyamikas will be mentioned below in the chapters
on the M adhyamaka-Prajhparam it and Madhyamaka-Vajrayna syn

812 I t w ould appear th a t there were tw o masters nam ed Jitari (or Jetari), one
o f w hom lived c. 800 and the other c. 1000; cf. G. T u cci, Minor B uddhist tex ts, i,
pp. 24952. The later one is said to have been a teacher o f Dlpam karairijnana and
also o f R atnakaraianti.
The Sugatam atavibhangakarikas are virtually identical w ith the verses o f
A ryadevas Jnanasarasam uccaya (see below, p. 106) which deal w ith the Vaibhaika,
Sautrantika, Y ogacara and M adhyam aka schools. The colophon states th a t the
author was born in Bengal. The Sugatam atavibhangabhaya has been assigned to
th e ten th century by Kenjd S h i r a s a k i , IB K 27/1 (1978), p. 493. (Verses correspond
ing to the Karikas are found quoted in Mok^akaraguptas Tarkabhaa, as w ell as
in th e Subhaitasamgraha where the last three are ascribed to Saraha.)

Beside the Yogcra-Madhyamaka synthesis discussed in the last chapter

there exists a synthesis of the Madhyamaka with the tradition of the Abhisa
maylamkra, a fundamental stra of the Prajpram it doctrine which is
ascribed (like the basic stras of the Yogcra/Vijnavda system) to Mai-
treya(ntha).313This synthesis is traced back at least as far as the sixth century to
Arya Vimuktisena, who composed the earliest commentary now available on the
Abhisam aylam kra: the V rtti which relates this treatise to the tex t of the
Pacavimatishasrik Prajpram it, one of the large Prajpram it-
Stras.314 Arya Vimuktisena was followed in this same line of thought by
B hadanta Vimuktisena (sixth or seventh century?), the author of a V rttika
oin the Abhisamaylamkra.315
The connexion between Madhyamaka thought and the Pram it-yna is of
course even older. Ngrjuna himself is generally supposed to have been closely
linked with it.31* And the *Upadea (Ta-chih-tu-lun) attributed to (a) Ngr
juna is in fact formally a commentary on the Pacavimatishasrik.317 The
Prajpram itstotra ascribed to Rhulabhadra has also been mentioned
above.318 Moreover, while the bond between the prajpramit and M adhya
maka philosophy is evident from many of the treatises of the school, the first
five paramitas are intim ately connected with the Bodhisattvas practice, a topic
th a t has been dealt with in a host of works by Mdhyamika authors beginning
with Ngrjuna and Aryadeva.319
A very im portant later representative of this movement was Haribhadra,
who flourished in the later part of the eighth century, in the reign of the Pla
king Dharm apla (rg. c. 770810 or 775812).320 H aribhadra was a disciple of

313 Cf. D . S e y f o r t R u e g o , L a thorie du tathdgatagarbha et du gotra, pp. 3955,

on the problem o f the id en tity o f M aitreya(ntha), w ith a bibliography o f the
314 See above, p. 87.
315 On Arya V im uktisena and B hadan ta V im uktisena see D . S e y f o r t R u e g g ,
W ZKSO 1 2 - 1 3 (1968), pp. 3 0 3 -1 7 .
316 See above, p. 6 .
317 See above, p. 32.
318 See above, p. 54.
313 For exam ple in the R atnval, *Bodhisam bhra-sstra, and the com m entary
on t h e D aabhm ikastra, all attributed to N grjuna; in A ryadevas C atuhataka;
in Can drak Irtis M adhyam akvatra, etc.
320 See th e colophon o f th e Abhisam aylam krlok. On th e dates o f the P la
kings see D . C. S i r c a r , JA S (Bengal) 18 (1976), p. 98, who proposes a slightly
different chronology from R . C. M a j u m d a r s . See also S i r c a r , X I X . D eutscher
Orientalistentag (ZDMG Supplem enta III , 2, 1977), pp. 9649.
102 David Seyfort Ruegg The Literature of the Madhyamaka School

Vairocanabhadra, and also according to Trantha of ntaraksita;321 he

worked a t the T rikatuka Vihara. He is renowned as the author of the very
extensive Abhisamayalamkrlok P rajpram itvy kliya commentary on
the Abhisamaylamkra which relates it to the text of the Astasahasrik Pra-
jpram it, of a shorter stra-Vrtti in verse (the S phutrth) and of an
explanation of the Samcayagths, and also as the editor of a version of the
Pacavimatishasrik. By some doxographers Haribhadra is considered to
belong to the Nirmala-Allkkravdin branch of the Yogcra-Svtantrika-
To this same current of thought belonged also Buddhajna(pda), a disciple
of H aribhadra,323 who thus was a representative of antaraksitas Yogcra-
Madhyamaka school. He seems to be the author of a Pajik on the Samcaya-
gths324 and, perhaps, of the Mahynalaksanasamuccaya. At the same tim e
he was an im portant master of the Vajrayana and the founder of the Jna-
pda lineage of the Guhyasamja tradition.
Another representative of this synthesis was Dharmamitra, the author of the
Abhisamaylamkrakrik-prajpramitopadeastratlk. In the colophon of
the Tibetan translation of this work, the only text of it available, he is referred to
as an Acarya of the Madhyamaka born in Banda (i.e. Bagala ?). Dharm am itra
is placed in the reign of a son of King Dharmapla by T rantha,325 who makes
him a contemporary of D harm ottara (fl. c. 800), Vimalamitra and Dharm-
kara.326 Dharm am itra would thus be an immediate successor of Haribhadra. At
the very beginning of his commentary he mentions in particular Ngrjuna; he
also refers to Mai trey antha as the Yogcra author of the Mahynastrlam-

821 B u'ston, Chos*byuh, fol. 1 1 1 b 1 1 2 a (ii, p. 157); Taranatha, rGya*garchos*

byun, p. 1669.
322 Yot H aribhadras explanations o f the nirdkarajnanavada and sdkdrajndnavada
see his A bhisam ayalam karaloka iv. 1417. Cf. \Jam 'dbyans*bzadpa Nag*dban*-
brcon-grus, P h a rp h y n rm th a ^ d p y o d , i, fol. l i b ( = p. 23); Grub,m th a,chen*mo,
fol. 131a132a.
823 Bu*ston, Chos-byun, fol. 1 1 2 b (ii, p. 15960); Gos*g2on*mrdpal, D e b th e r -
snon'po, ja, fol. 7 b (i, p. 367 s q .); Taranatha, rGya^garchos^byuri, pp. 157, 1669;
Jam *dbyahsbzad*pa*Nag*dbah*br5on*grus, Grub*mtha*chen*mo, fol. 132a. Tara-
natha (p. 195) m akes him a contem porary o f K ing Dharm apala.
324 This work translated by Vidyakarasim ha and dPal'brcegs (and the author o f
which is Buddha^rijhana according to the bsTan*gyur) is ascribed to Buddhajna-
napada b y Bu*ston (Chos*byuh, fol. 1 1 2 b 6 [ii, p. 159]) and the D eb*ther*shoirpo,
ja, fol. 8 a (i, p. 367). Buddhajnana-pada/Buddharijnana is thus to be distinguished
from the B uddhairljnana m entioned below, p. 117.
325 Taranatha, rG ya,gar*chos',byuh, p. 171.
326 TaranStha distinguishes this Dharm am itra from the Vaibhaika m aster o f the
sam e nam e who w rote a Tika on Gunaprabhas V inayasutra (rGya*gar*chosbyuh,
p. 152153). This earlier D harm am itra is in fact reported to have been a pupil of
Gunaprabha (see Bu*ston, Chos*byuh, fol. 113 b 2 [ii, p. 161]), which would make
him approxim ately a contem porary o f Arya V im uktisena. Dharm&kara(datta)
w as th e religious nam e o f Areata (c. 73090 ?) (cf. D urvekamrira, Arcataloka [GOS
ed.], p. 233).
The Madhyamaka-Prajnaparamita Synthesis 103

kara, and to Kamala&la. His work is generally known as the P rasphutapada;

it is a commentary on H aribhadras shorter explanation of the Abhisamayalam-
kara, the Sastravrtti (S phutartha).327 In his treatm ent of the prakrtistha-gotra,
one of the basic topics of the Abhisamayalamkara, in the first chapter of his
work Dharm am itra discusses the tathdgatagarbha (fol. 54 a ); he thus differs from
the two Vimuktisenas and H aribhadra and follows Kamala&la who, as already
noted, introduced the tathdgatagarbha and ekayana theories into his presenta
tion of Madhyamaka thought.328
Abhayakaragupta, the author of a commentary on the Astasahasrika Prajna-
param ita entitled Marmakaumudi, was also a leading later representative of
this current of thought, as well as of the Yogacara-Svatantrika-Madhyamaka
school in general and of the Madhyamaka-Vajrayana synthesis.829

327 Cf. B irston , Chos,byuh, fol. 105a l2 (ii, p. 140); Taranatha, rG yagarchos*-
byufi, p. 153.
828 See above, p. 96, and D . S e y f o r t R u e g g , The gotra, elcayana and tathagata-
garbha theories according to D harm am itra etc., in Prajhaparam ita and related
system s (Studies in honor o f E . Conze, Berkeley, 1977), pp. 283312.
329 On this master see below, p. 11415.

Another religio-philosophical movement to have a very im portant influence

among the Madhyamikas especially during the last centuries of the first millen
nium P. C. was the Vajrayana, also known as M antrayana (as the correlative
of the Paramitayana) or Tantrism. Since, beside the Yogacara, the chief
theoretical source of the Vajrayana is to be found in Madhyamaka thought the
connexion between the two cannot be deemed secondary or external. The syn
thesis th a t came about between the two, which represents a confluence and
coordination of theory and praxis, is now only partially accessible to us through
written documents and the iconological study of the art of the Vajrayana; it
was indeed to a great extent an acroamatic tradition communicated through
the Gurus oral instructions and adapted in principle to the intellectual capacity
and psychological propensities of each disciple individually. Although it is not
possible here to discuss this im portant current of thought, it is indispensable
to say at least a few words concerning the links between the Vajrayana and the
One of the central figures in this movement was Arya Nagarjuna. This Vajra-
yanist master bore the same name as the teacher revered by the Madhyamikas
as the originator of their school. The first and presumably most im portant of
the Vajrayanists to have this illustrious nam e331 seems to have lived in the
seventh (or at the latest in the eighth) century. The Bodhicittavivarana may

330 A lready I-ching (635713) m entions N grjunas having studied th eV idydha-

rapitaka; see E. C h a v a n n e s , Mmoire com pos l poque de la grande d ynastie
T ang sur les religieux m inents qui allrent chercher la loi dans les pays d Occident
(Paris, 1894), p. 102. A bout the year 400 the great Serindian M dhyam ika m aster
K um rajva, who was established in K u and had close links 'with Kmr, is
credited w ith the translation o f a Tntrik te x t, the Mnhmyr (Vidyrjii)
(Taish 988) a work w ith which N g a has indeed been connected by the Ma-
jurimlakalpa (liii. 450). Much later, c. 1000, Shih-hu translated into Chinese works
b y N grjuna and Tntrik tex ts. In India, beside works b y T ntrika M adhyam ikas
m entioned below, a Tk on th e A nantam ukhanirhradhran was com posed b y
Jnnagarbha in the eighth century (on him see above, p. 69). Since this book is
concerned w ith the philosophical literature of the M adhyam aka school in th e strict
sense, works dealing w ith practice and Tantra can here only be referred to briefly.
(The distinction betw een M adhyam aka treatises as such and other kinds o f works
w ritten b y M adhyam ikas was adopted also in the classifications o f the T ibetan
authorities, who o f course did not thereby wish to contest the com plem entarity and
indeed indivisibility o f theory and praxis, and the ultim ate coincidence o f pram i-
tdnaya and m antranaya).
331 There m ay w ell have been more than one Tntrika nam ed N grjuna (see the
literature cited below, note 333).
Madhyamaka and Vajrayna 105

well be his composition;833 and he was most probably the author also of the
commentary on the Guhyasamja (the Tantratik) and of the Pancakrama,
and thus the founder of the rya lineage belonging to the Guhyasamja tra
dition. This Ngrjuna was presumably the disciple of one of the most famous
of the early Vajraynist masters, Saraha, who is also known as R hulabhadra.333
Ngabodhi has been mentioned above (p. 57, together with Nga); he
appears to have been a disciple of Arya Ngrjuna and a teacher of Vajrabodhi
(c. 671-741).
This Vajraynist lineage evidently included other masters bearing the names
of (earlier) Mdhyamika cryas. Candrakrti-pda is the author of a famous
commentary on the Guhyasamja entitled Pradipoddyotana.334 ryadeva-pda

881 This work is apparently quoted b y Haribhadra, A bhisam aylam klok ii. 1.
A T ibetan translation o f it w as made b y Jayn an da and mDo*sde*bar (see below,
p. 114).
888 This teacher-disciple relationship betw een Saraha R ahulabhadra and Arya
Ngrjuna seem s to h ave been projected back b y som e traditions to th e relationship
b etw een Ngrjuna, the author o f th e MMK, and an earlier R ahulabhadra, th e
author o f th e Prajpram itstotra, who appears to have been a successor o f
N grjuna I; see above, p. 54. A Rahulabhadra is listed as a disciple o f
Buddhajna along w ith Dipam karabhadra (Gosg2on*nu*dpal, Deb*therson*po,
ja, fol. 9b [i, p. 371]). The Bodhisattvagocarapariuddhistrrthasam graha in
cluded in the M adhyam aka section o f the bsTan*gyur is ascribed to (a) R ahula
The h istory and lineages o f the Siddhas and Tntrika masters remains very
obscure for us, and our sources som etim es give quite divergent accounts o f them
so th a t it is not y e t possible to fit th em coherently into the history o f Indian and
B udd hist thought. On th e subject cf. G. T u c c i, Anim adversiones indicae, J P A S B
26 (1930), p. 138 sq. ( = Opera minora, i, p. 209 sq.), and T ibetan painted scrolls, i
(Rome, 1949), pp. 214, 231; S. L v i , BSOS 6 (193032), p. 417 sq.; R hula S a m -
k b t y y a n a , J A 1934, p. 209 s q .; H . H o f f m a n n , B eitrage zur indischen Philologie
und A ltertum skunde (Festschrift W . Schubring, H am burg, 1951), p. 142; A. W a y -
m a n , The B uddliist Tantras (New York, 1973), p. 14 sq., and Y oga o f th e Guhya-
sam jatantra (Delhi, 1977); P. K v a e r n e , An an thology o f B uddhist Tantric songs
(Oslo, 1977), p. 4 sq. (w ith bibliography, p. 270).
884 B uddhist tradition has distinguished betw een those writings o f Candrakirti
th a t belong to th e Sutra sy stem (mdo lugs), such as th e M adhyam akvatra and
th e comm entaries on N grjunas and A ryadevas treatises, and those th a t belong
to the Mantra system (snags lugs), such as th e P rad ip o d d yo tan a; b u t it ascribes
both categories to one and th e sam e person. (The case is th u s comparable w ith th a t
o f Ngrjuna him self, betw een different categories o f w hose works B uddhist
tradition has clearly distinguished while at the sam e tim e ascribing th em to the sam e
The *Triarana[gamana]saptati contained in the M adhyam aka section o f the
bsT an#1gyur is presum ably b y th e Tantrika author since it not only picks up the
akliatjdna them e o f th e B odhicittavivarana bu t refers to seven p iiakas including
th e vidydh arapitaka (fol. 2 9 3 b 5 and 2 9 4 a 7). Verse 33 seem s to be quoted b y
Haribhadra, A bhisam aylam krlok i. 3 (p. 89).
On a later Candrakirti w ho lived in th e eleventh century, see above, p. 81.
Concerning th e P radipoddyotana and the related literature, cf. A. W a y m a n ,
Y oga o f the G uhyasam jatantra (Delhi, 1977).
106 D avid Seyfort Ruogg The L iteratu re of th e M adhyam aka School

is the author of the Cittaviuddhiprakarana,335 which is connected in its sub

ject-m atter with Arya Ngrjuna-pdas Pancakram a.336 He is perhaps also
the author of the Jnasrasamuccaya, a not specifically Tntrika work in
cluded in the Madhyamaka section of the bsTan*gyur which describes inter alia
the doctrines of the later schools of Indian philosophy, Brhmanical as well as
Buddhist.337 According to this treatise the Mdhyamika does not accept the
vijnand as a real entity because it has neither a single nor a multiple own being
(ekanekasvabhdva), just like a sky-lotus; for him reality (tattva) is free from
the four positions postulated in dichotomizing conceptualization (viz. existence,
non-existence, both, and neither) (fol. 30 b). To this line of masters may belong
a (later) ntideva known also as Bhusuku,338 as well as a Bhavya.339
The Indo-Tibetan records frequently identify these Vajrayanist masters with
the illustrious teachers of the earlier Madhyamaka school whose names they
bore, and to whom these records accordingly ascribe extraordinarily long life
spans. However, at the same time, these records often differentiate very clearly
between distinct phases in these m asters teachingssuch as the so-called Sutra-
system and Mantra-systemso th a t the identification of the persons of these
masters did not in fact necessarily result in confusion by the doxographers of
distinct doctrines.
In addition to ntaraksita and Kamalaila who were also Vajraynists
several later masters, many of whose works are included in the Madhyamaka
section of the bsTaivgyur, are known as representatives of the Madhyamaka-
Vajrayana synthesis.
Kam balapda wrote a Mandalavidhi and the Prajnpramitopadea (both
of which works are found in the Madhyamaka section of the Peking bsTan* gyur),
as well as the Navaloki340 on the Prajfipram it together with an abridge
ment and a commentary. He is regarded by doxographers as a Yogcra-Sv-
tantrika-M dhyam ika (Nirmala-Allkkra branch). In his Alokaml he makes
very extensive use of Yogcra ideas.341
335 This is the title of the work as quoted in the Subhitasam graha and as edited
b y P . P a t e l (Santiniketan, 1949). The bsTan*gyur has Cittvaranaviodhana-
nma-prakarana (for the te x t it ascribes to Aryadeva) and Cittaratnaviodhana
(for the te x t it ascribes to Indrabhuti-pda). The relation betw een these tw o works,
apparently translated about th e sam e tim e (eleventh cen tu ry), is n ot clear.
338 The Tntrika A ryadeva is also known as Karnari or Kanheri; cf. G. T u c c i,
JP A S B 26 (1930), p. 141 ( = Opera minora, i, p. 211).
337 Cf. S. Y a m a g u c h i , Chugan bukky ronk (K yoto, 1944), pp. 256431;
K . M i m a k i , La rfutation bouddhique de la perm anence des choses (Paris, 1976),
pp. 1869. See also above, note 312.
338 Cf. B u'ston, Chos*byu, fol. 114a; A. P e z z a l i , ntideva (Florence, 1968),
pp. 4 sq., 4445, 8789; H . H o f f m a n n , Asien: Tradition und F ortschritt (Fests.
H . H am m itzsch, W iesbaden, 1971), p. 207 sq.
339 H e is perhaps the author o f the M adhyam akaratnapradipa m entioned above,
p. 66. A Pajik on the Pancakram a was com posed b y a certain B h avya-kirti, as
well as a com m entary on th e Pradipoddyotana.
340 Cf. G. T u c c i, Minor B uddhist texts, i (Home, 1956), pp. 21131.
841 In th e bsT an'gyur there is found a com m entary entitled Alokamlfclk on
Madhyamaka and Vajrayna 107

Advayavajraalso evidently known as Maitri-pada, a master who is re

ported to have rediscovered the Ratnagotravibhagais the author of the
T attvaratnavall, TattvaprakaSa, M adhyamasatka and other treatises in which
short outlines are given of the main doctrines of the Buddhist philosophical
K rsnapada is the author of the M adhyamakapratityasam utpada and the
Bodhisattvacaryavataraduravabodhananirnaya, as well as of other short works
included in the M adhyamaka section of the bsT airgyur.343
Mention should finally be made here of two works by Vimalamitra th at are
included in the Madhyamaka section of the bsT airgyur. The first is a treatise
on non-conceptual (nirvikalpa) m editative realization (bhdvana) in which
entry is simultaneous (cig car = yugapad) entitled Cig-carjug-paT nanrpar-
m rrtog-pai bsgonrdon, and the second is a longer tex t on m editative realization
in which entry is gradual entitled Rim ,gyis,jug*par (b)sgonrdon. According to
the colophons both texts were translated into Tibetan by Ye*ses*sde. And their
author Vimalamitra, who thus evidently flourished by the end of the eighth
century, is apparently to be identified with the master of the same name, a
contemporary of Santaraksita and KamalaSila, who was one of the more pro
minent and influential figures in the early history of Buddhism in Tibet and
who is counted as one of the chief teachers of the rfogs chen tradition of the
rSfin*ma school of Tibetan Buddhism. These two treatises belong to the same
category as the works on bhavana by Jnanagarbha and KamalaSila mentioned
F urther representatives of the last period of the Madhyamaka school in
India who were at the same time Vajrayanistsand some of whom in addition

K am (b)ala(pda),s lokaml (sNan*bai*phren*ba) ascribed to A svabhva, w hom

Trantha appears to m ake an approxim ate contem porary o f Jnanagarbha and
K am bala and describes as an advocate o f th e Vij napti-M adhyam aka (rnam rig gi
dbu m a ) (rG ys'garchos^byun, p. 152). According to M. H a t t o r i , D ignga on per
ception (Cambridge, Mass., 1968), p. 5, A svab hva preceded D harm apla in th e
Y ogcra/V ijfinavda school. Is this then another A svabh va? On K am bala/
K am ala see G. T u c c i, Minor B uddhist tex ts, i, p. 213. See also Trantha, rGya*
gar*chosbyun, pp. 144, 146, 152; b K a *babs*bdun*gyrrnam*thar (transi. A. G r n -
w e d e l , p. 53 sq.).
342 Cf. G osgZ om m rdpal, D eb -th ersn o n po, cha, fol. 9 b 10a (i, pp. 3479); da,
fol. 2 a sq. (ii, p . 842); D . S e y f o r t R u e g g , La thorie du tathgatagarbha e t du
gotra, pp. 3639. H e lived in th e 10th11th century. On the M adhyam asatka o f
Advayavajra/M aitrpda see S. K . P a t h a k , A L B 25 (1961), pp. 53949.
348 A Krgna is included in a list o f Dpam kararjnnas teachers given by
Gos*g2on*mrdpal, D eb ther-shon'po, ca, fol. 2 a (i, p. 2434); he w as a con tem p o
rary o f N rop (Ndapda), Jitri and R atnkarasnti. D pam karairjnna trans
lated a Triskandhasdhana b y a Krna-pda. This K rnapda m ight perhaps be
th e sam e person as th e collaborator o f Nag*6ho*(5hul*khrims*rgyal*ba (born in 1011)
in th e translation o f Candrakrtis M adhyam akvatra-kriks ; or he m ay be a
nam esake.On the m asters bearing this nam e cf. D . L. S n e i j l g r o v e , H e vaj rat a n
tra, i (London, 1959), p. 13 note.
108 David Seyfort Ruegg The Literature of the Madhyamaka School

made extensive use of the Prajnaparam ita and Yogacara/Vijnanavada philo

sophies and thus contributed to the elaboration of a synthesis of several cur
rents of thought in later Indian Buddhismwill be mentioned in the next

Bodhibhadra, Dharmaklrti (Dharmapala) and Dpamkararjnna

Bodhibhadra (fl. o. 1000), a master at Nland and a teacher of Dipamkara-

rjna, wrote a Nibandhana on ryadeva-pdas Jnnasrasam uccaya.344
Like the tex t on which it comments, it is of interest for the later history of the
doctrines of the Indian philosophical schools, Buddhist and Brhmanical. Ac
cording to this commentary (28), for the Mdhyamika absolute reality (tattva)
is free from the four extreme positions postulated in the form of the tetralem-
m a (catuskoti) within the frame of dichotomizing conceptualization (vikalpa),
viz. existence (of the Vijnavadins vijnna), non-existence (according to the
L okayatas doctrine), neither, and both. I t is explained th at reality has been
stated not to be some thing th a t is without both existence and non-existence
(gis dan bral bai bdag id) in order to remove the idea th a t there exists some
kind of entity having a mode of being between existence and non-existence
(trtya-ri), in conformity with the principle of the excluded middle. And
because the Mdhyamika therefore asserts no dogmatically postulated onto
logical position concerning the existence of any kind of entity his theory cannot
be attacked (fol. 51 b).345 Turning to the systematic analysis of the samvrti level,
Bodhibhadra mentions both the doctrine of Bhavya and his school and th at of
antaraksita and his school; according to the first an image (snan ba) is not
made the object of valid cognition ( jai ba), and according to the second inter
nal knowledge appears in various forms, no mental entity (snan bafi dnos po)
being however ultim ately real. W ith respect to praxis, the Mdhyamika is of
course to engage in spiritual training (iks) (fol. 51b52 a).
Bodhibhadra also composed im portant works on the B odhisattvas vows
(samvara)*** and on yoga and meditation.

Dharm aklrti of Suvarnadvlpa (fl. c. 1000), another of Dlpamkararijnas

principal teachers, was born in a royal house of Indonesia. His major work, a

344 Cf. K . M im a x i , L a rfutation bouddhique de la perm anence des choses (Paris,

1976), pp. 1 9 0 -2 0 7 .
345 See Ngrjuna, Y igrahavyvartan 29 and 59; ryadeva, Catuhataka xvi.
25; Santarakita, M adhyam aklam kra 68.
346 One such work is en titled Bodhisattvasam varavim aka-Pajik. Together
w ith ntarak^itas Saipvaravim akavrtti, this work depends on Candragomins
B odhisattvasam varavim aka and thu s belongs to th e Yogcra-M adhyam aka
synthesizing m ovem en t (see also above, pp. 88, 93). B odhibhadra also wrote
a short B od h isattvasam varavidh i included in th e M adhyam aka section o f the
b sT am gyur.
110 David Seyfort Ruegg *The Literature of the Madhyamaka School

Tk on the Abhisamaylamkra entitled Dur[ava]bodhlok, was composed

in the town of Vijayanagara/Vijayapura (rNanrparTgyal*baigro) in the
kingdom of Suvarnadvipa (gSergli)347 at the time of King r Cdmanivar-
man (dP al^C ug-grnorbui-go-cha).848
Dipam kararjnas master known as Dharm apla (Chosskyo) is credited
with a work entitled ikssamuccaybhisamaya. Two other works, the
Bodhisattvacaryvatra-sattrim atpindrtha and the Bodhisattvacaryvat-
rapindrtha, are connected with him in the bsTan*gyur (which states th a t
they were requested by his disciples Kam alaraksita and Dpamkararjna).349
Although these works are very brief, their author has been given an im portant
place in the history of later Buddhist thought as the teacher of Dipamkarar j na.
The epithet of Suvarnadvpa (gSerglirrpa) is given to the above-mentioned
D harm akirti and to Dharmapla, and these are evidently two names of one
and the same person.
Dharm akirti of Suvarnadvpa was, however, not a Mdhyamika but a follower
of the Satykra- or Skra-Vijnavda who held the Madhyamaka to be of
indirect meaning (neydrtha) according to a biography of Dpamkararjna
(the rNam th ar rgyas pa, fol. 12a). According to this same source (fol. 5ab,
12 ab and 24 b25 a), his great disciples were Jnarim itra and R atnakirti
both of whom are described also as Skra-Vijnavdins, R atnkaranti
a follower of the Alkkra- or Nirkra-Vijnavda who held his version of
this doctrine to be the true Madhyamaka (the so-called Vijapti-Madhya-
m aka), and Dpamkararjnaa Mdhyamika who followed Candrakirti
although he had received instruction also in the synthesizing Yogcra-Madhya-
m aka from Avadhtipda and in the Vijapti-M adhyamaka from R atnkara
nti (see below, Appendix II).

Dpamkararjna, who is very frequently known by the appellation Atia,

was bom in a royal lineageaccording to some sources th a t of Sahor/Za*hor

847 Suvarnadvipa is apparently to be identified as Sum atra (and perhaps also

Ja va, and M alaysia; Suvarnabhm i on the other hand is B urm a); cf. G. C o e d s ,
Les tats hindouiss d Indochine et d Indonsie3 (Paris, 1964), pp. 259, 264.
A certain Guru and B hiku P aindaptika o f Y ava dvip a and Suvarnadvipa is cited
in Dipamkararijnnas Bodhim rgadpapanjik (fol. 336b 3).
348 As an im portant artistic creation a ttesting th e significance of th e Prajna-
pram it in Indonesia m ention m ay be m ade o f th e fam ous im age from C(h)andi
Singhasari w hich has been described as the portrait o f a th irteenth century queen
identified w ith Prajnpram it (this im age is in the R ijksm useum voor Volken-
kunde in Leiden). Such an identification o f a queen w ith Prajnpram it is also
known from Champa in th e tw elfth century. See G. C o e d s , op. cit., pp. 31718,
341. Cf. E . C o n z e , T hirty years o f B uddhist studies (Oxford, 1967), p. 246 sq.
349 T he bsTan*gyur no. 5280 seem s to ascribe at least th e first of the tw o works
directly to these tw o disciples, but th e colophons are n o t quite clear. B u'ston has a ttri
buted the Bodhicaryvatrapindrtha to Suvarnadvpa(-Dharm apla) (Chos^byuh,
fol. 159 a). These tw o te x ts were translated into T ibetan b y Dpamkararjnna
and Nag*6ho Lo*6a*ba; th e y are summ aries oi the Bodhicaryvatra.
The Last Period of the Indian Madhyamaka School 111

in Bengal c. 982. After having been the disciple of the best teachers of N orth
eastern Indiasuch as Bodhibhadra, Jitri, Krsnapada, and R atnkaranti
he is said to have travelled to Suvarnadvlpa, with whose kings the Pla rulers
of Northeastern India had long entertained relations,350 to study there with
Dharm aklrti. Following his return to India Dipamkararljna became re
nowned as one of the most eminent teachers of his time. At Vikramaila he was
instated as abbot under Kong Mahipla (rg. c. 9881038 or 9771027) and was
reconfirmed in this office under King Nayapla {rg. c. 10381054 or 1027
1043). Owing to his great fame as a master at one of the most illustrious B ud
dhist monastic seminariesand perhaps also because of his reputed connexion
with the royal house of Za*hor from which ntaraksita, the great Updhyya
and crya-Bodhisattva of the Tibetans at the time of the old kingdom, was
said to have issuedhe was invited, on the advice of Ye'Ses-od of the royal
house of m N a-ris, by King *od*lde and the Prince-Monk (lha bun) B yair
chub-od to the kingdom of western Tibet. He arrived there in 1042 and col
laborated at the monastery of (m)Tho*l(d)m with the great translator Rhrchen*
bzarrpo (9581055), who had himself studied in Kamr. W ith him Dlpam-
kararjnna undertook the Tibetan translation of his teacher D harm akirtis
Dur[ava]bodhlok and in addition revised the Tibetan translation of the
Abhisamaylamkra. A few years later Dipamkararjnna arrived in central
Tibet. First in India and then in Tibet he worked closely with his disciple
Nag *cho* Chul khrim sTgy a lb a (bom in 1011), with whom he translated many
fundam ental texts of the M adhyamaka including Bhavavivekas Madhyama-
kahrdayakriks, Tarkajvl and Madhyamakrthasamgraha, C andrakrts
Pacaskandhaprakarana, and Dharm apalas Bodhisattvacaryvatrapindr-
tha, Sattrim atpindrtha and ikssamuccaybhisamaya, as well as Vasuban-
dhus Mahynasamgrahabhsya and Gunaprabhas Bodhisattvabhum ivrtti. In
central Tibet he worked also very closely with his disciple Bronrston-rGyal*
b ai-byuirgnas (10051064), the founder of the bK a-gdams-pa school. Dl-
pam kararjna died at s$e*than in central Tibet c. 1054.351
Among D lpam kararijna^ best-known works is the Bodhipathapradipa
composed a t (m)Tho*lifi a t the request of Byan*chub-od. A short guide to the
path of Awakening, it is one of the basic works of the bK a-gdams^pa and
dGe-lugs-pa schools in T ib et; and its classification of practisers of the path into
three typesthe lesser, the middling and the superiorwas employed for ex
ample by o k h a p a in his treatises on the graded path (lam rim ).353

850 Cf. R . C. M a j u m d a r , H istory o f ancient B engal (Calcutta, 1971), pp. 116, 625,
582. The historicity o f th is voya ge has however been questioned.
861 Cf. A. C h a t t o p a d h y a y a , Atia and T ibet (Calcutta, 1967); H . E i m e r , B e
richte ber das Leben des Atia (Dlpamkarasrijnna) (Bonn, 1977). Cf. also G.
T u c c i, R in e en bzan po e la rinascita del buddhism o nel T ibet intorno al mille (Indo-
T ibetica ii, R om e, 1933).
862 On this work see H . E i m e r , Bodhipathapradipa (W iesbaden, 1978), w hich
contains an edition and German translation.
112 David Seyfort Ruegg The Literature of the Madhyamaka School

The Bodhipathapradlpa has a commentary attributed to Dipamkararijna

himself the title of which is given in our sources as Bodhimrgadipa-pajik.358
In addition to expanding on the outline of the path given in the basic text, this
commentary sketches the history of the Madhyamaka school and its basic
doctrines. Among the latter special attention is given to four great reasons
(gtan higs: hetu) employed by the Mdhyamika in establishing the under
standing of the non-substantiality of all dharmas : (i) the reason analysing the
(supposed) production of a product;354 (ii) the so-called *Vajrakana-reason
(rdo rje gzegs ma'i gtan higs) analysing origination from any of the causes
postulated in the frame of the catuskoti355 by either non-Buddhists (e.g. vara,
purusa, svabhdva, the gunas, Brahman, Visnu, Mahdeva) or Buddhists (e.g.
the bhidharmikas six hetus and four pratyayas); (iii) the reason analysing
things as devoid of both a single and a multiple nature (eka-aneka-svabhdva) ;35fl
and (iv) the reason constituted by origination in dependence (pratityasamut-
pdda) establishing unyaid,357
This commentary also provides lists of works which were regarded as being
by Ngrjuna (including, in addition to the writings generally considered today
to be certainly by him, the *Akutobhay, the Mahynavimaka and the
Aksaraataka)358 and by Aryadeva (including the *Madhyamakabhramaghata
[dBu-ma*rnam*par*thag*pa*chen*po], the H asta vlaprakarana and the Jna-
srasamuccaya).359 I t also lists eight commentaries on the MMK: N grjunas
own (i.e. the *Akutobhay), Candrakirtis, Bhvavivekas, B uddhaplitas,
Sthiram atis, Gunamatis, Gunaris, and G unadattas, adding th a t the two
great sub-commentaries on Bhvavivekas Prajpradipa were by Avalokita-
vrata and Devaarman (i.e. the dB um a'dkar-po^charba).360
This list appears to link Dpamkararjna above all with the pure M adhya
m aka traditions transm itted from Buddhaplita, Bhvaviveka and Candrakirti,
rather th an with the Yogcra-Madhyamaka synthesis represented by n-
taraksita and Kamalala. Whereas this latter tradition had become firmly
established in Tibet during the second h a lf of the eighth century, a t the time
of the Early Propagation (sa dar) of Buddhism in th at country, Dipamkara-

353 The title o f th is com m entary is so given in th e Peking and sD e dge editions
o f th e bsT am gyur.
854 The com m entary here speaks o f a catuskoti tetralem m a, b u t ju st like th e
basic te x t it enum erates only three alternatives : production of an existen t product,
o f a n on -existen t one (e.g. a sky-flower), and of one th a t is both existen t and n o n
existen t. E ach alternative is show n on analysis to be unacceptable.
355 See Ngrjuna, MMK i.l.
358 This principle has been elaborated especially b y rigupta (above, p. 68) and
867 Pajik, fol. 322 a323 a.
868 Later T ibetan scholarship has n ot accepted th is list uncritically. Cf. above,
p. 49 note 129. (Dpam kararijnas authorship o f this com m entary has indeed
been called into question.)
359 P ajik, fol. 324a.
880 P ajik, fol. 324b.
The Last Period of the Indian Madhyamaka School 113

rjnna and his collaborators and followers were evidently responsible for the
consolidation of the two branches of the pure Madhyamakathose of Bhva-
viveka and Candrakirtitowards the beginning of the Later Propagation (phyi
dar) of Buddhism in th at country (for the role played by Jaynanda and Pa*
chab in the establishment of Candrakirtis Prsangika school see below, p. 114).

Dpam karasrjnna composed the Satyadvayvatra for a king of Suvarna-

dvipa. In this discussion of the two truths he explains th at the samvrti is two
fold (mithy and tathya) b u t the paramrtha is one only, for there can be no
m ultiplicity in dharmat. After having referred to B havyas view th at uncon-
ceptualizable reality m ay be known by savikalpaka as well as by nirvikalpaka
knowledge, the author states th a t nyat or dharmat is to be known in the
way explained by Candrakirti in his M adhyamakvatra.361

Among Dpam kararjnnas numerous other writings his very brief Ma-
dhyamakopadea (on which there is a commentary by Prajnmoksa) and his
longer Ratnakarandodghtanma-M adhyamakopadea require special men
tion together with his Strasamuccayasamcayrtha and Mahstrasamuccaya.
His Dharm adhtudaranagti incorporates a number of verses from the Dhar-
m adhtustava ascribed to Ngrjuna (making up approximately one seventh
of Dpam kararjnnas text).

Jaynanda (flourished in the second half of the eleventh century), a scholar
of Kmri origin, lived at a time when there took place a renewal of interest in
C andrakirtis branch of the Madhyamaka. He contributed very considerably to
this movement especially in Tibet, where he worked extensively.362
Jaynanda is the author of a large Tk on Candrakirtis M adhyamakva
tra, the only commentary by an Indian master now extant on this work.363 I t
661 Vol. ha, fol. 70 b = vol. gi, fol. 7 b.
ae* This Jaynan da w ould seem n o t to be identical w ith th e person o f th e same
nam e who collaborated in th e T ibetan translation o f the Y uddhajaya[arnava]-
nm atantrarja Svarodaya-nm a (the m n ya appended to the T ibetan version
o f this T antra states th a t th e latter w as from Ya*che [Semj], and th at he was
connected w ith the Km ri spiritual lineage of Som nanda and A bhin avagup ta ;
cf. P . C o r d i e r , Catalogue du fonds tib tain de la B ibliothque N ationale, iii, p. 477).
In the colophons o f the M adhyam akvatratk and Tarkamudgara, our J aynanda
is stated to be from Kmr. T he translator of th e Svarodaya w ould moreover
appear to have lived som ew hat later th an our Jayn an d a (cf. E. Gene S m i t h s
introduction to th e edition o f th e dByaAs*char published at Gangtok in 1970).
A certain Kha*che A nanda/A nanta is m entioned am ong th e ancient translators,
including Vairocana and lDan-ma-rCe'mah, at the end o f the eighth century in
T ib et (cf. d P a *bo*g<5ug*lag*phrefl*ba, mKhas*par d g a *ston, ja, fol. 125a); and he
has to be distinguished from our Jaynanda.
,M Cf. N . A i y a s w a m i S a s t r i , Journ. o f Oriental R esearch (Madras) 6 (1932),
pp. 1 7 1 -8 3 .
114 David Seyfort Ruegg The Literature of the Madhyamaka School

is available only in Tibetan translation; and the colophon informs us th a t

Jaynanda composed it far from his native Kmr, in M rnag (near the Huang-
ho and the W u-tai-shan, as stated in the colophon). I t was translated into
Tibetan by its author with the Tibetan translator Kun-dga-grags. The circum
stances of the composition of this work make it one of the most extensive works
of Indian philosophy to have been written outside the Indian subcontinent.384
Jaynanda is also the author of another work included in the Madhyamaka
section of the bsTan-gyur, the Tarkamudgara, a very short tex t in verse on

In Tibet Jaynanda collaborated with the translator P axhab Srm a-grags

(bom in 1055) whotogether with him and other scholars like Kanakavarm an,
Tilakakalaa and Sksmajana365was largely responsible for the establishment
in the land of snows of Candrakrtis school of the Madhyamaka, revised the
Tibetan translation of the MMK, and translated ryadevas Catubataka along
with Candrakrtis commentary on it, the M adhyamakvatra, and the Prasan-
napad. W ith P ax h ab and Khu mDo,sde,bar Jaynanda translated a Stra-
samuccaya ascribed to Dpamkararjnna; and with mDo*sde*bar he trans
lated the Vaidalyaprakarana and the B odhicittavivarana368 ascribed to Ngr-
juna and revised the Tibetan translation of the Yigrahavyvartan-Kriks.
W ith Grags-byorsesxab he translated the Vaidalya-'Stra, the Prattya-
sam utpdahrdayavykhyna and one of the versions of the Mahynavimaka ;
and he revised the Tibetan translation of the Aksaraataka and its V rtti.387

A bhayakaragupta
Abhaykaragupta flourished about 1100 and was a scholar of the Vikramal
monastic seminary. There, in the thirtieth year of the reign of King Rm apla
(rg. c. 10771130 or 10721126), he composed the M unimatlamkra.888 This
very extensive treatise of encyclopaedic character expounds Mahynist
gnoseology and soteriology in connexion with the Prajnpram it doctrine
and includes many references to the basic texts of both the Madhyamaka and
864 On B uddhism and T ibetan culture in M rhag, am ong th e Tanguts, see E. J .
K yanov, Tibetans and T ibetan culture in the T angut state H si hsia, in L . L i g e t i ,
ed., Proceedings o f th e Csoma de Krs Memorial Sym posium (Budapest, 1978),
p. 205 sq.
865 Cf. J . N a t t d o t j , L es bouddhistes kamriens au m oyen-ge (Paris, 1968), pp.
73, 139.
886 The version in th e Tantra section o f the bsTan*gyur (there is another version
in the Sutra section).
387 In view o f th is w ork accom plished b y Jayn an d a and Pa*6hab and b y their
Indian and T ibetan contemporaries, it is n ot possible to ascribe the great im portance
o f the Prsahgika school in Tibet only to C o A k h a p a s influence and to date it from
th e fourteenth century, as has been done b y som e recent scholars.
888 See the author colophon in th e T ibetan translation. According to Sum 'pa
Ye*es-dpal*byors R eh rm ig , A bhaykaragupta died in in sbrid = 1125.
The Last Period of the Indian Madhyamaka School 115

Yogacara/Vijnanavada schools. In Chapter i, beside a treatm ent of the bodhi-

citta as the main subject, there is an explanation of the tathdgatagarbha theory
and the single vehicle (ekayana) (fol. 189 a). Chapter ii deals chiefly with
bhdvana as concentration (samddhi), attainm ent (samdpatti) and meditation
(dhyana), with the objects of meditation, and with ultim ate reality as com
prehended in meditation. Chapter iii is devoted to a study of the eight abhisa-
may as of the Prajnaparam ita doctrine; the treatm ent of the first abhisamaya
the sarvdkdrajhatdincludes an exposition of the theory of th e spiritual germ5
which exists by nature (prakrtisthagotra), and in this connexion the author
discusses further the tathdgatagarbha and ekayana .369 Chapter iv is devoted to
the qualities th a t are constitutive of buddhahood and comprises an exhaustive
treatm ent of M ahayanist gnoseology and buddhology. Abhayakaragupta de
scribes his exposition as based on Nagarjuna and Arya Vimuktisena (fol. 396b).
The M unimatalamkara is one of the last of the major comprehensive treatises
of Indian Buddhism, and it presents a treatm ent of Mahayanist thought based
on the Prajnaparam ita, Madhyamaka and Yogacara traditions. Although as
such it is not in the narrowest sense a work of the Madhyamaka, it bears
testimony to the efforts made by the later Madhyamikas systematically to
elaborate a synthesis of the entire Mahayanist tradition.
Abhayakaraguptas commentary on the Astasahasrika Prajnaparam ita is
held by some doxographers to conform to the Svatantrika position, like the
commentaries on the Prajnaparam ita doctrine by the two Vimuktisenas,
H aribhadra and Buddhajnanapada. And Abhayakaragupta is counted as a
representative of Santaraksitas Yogacara-Svatantrika-M adhyamaka school.370
In addition, Abhayakaragupta was the author of a number of works on
iconology, ritual and several Tantrik cycles.371

Other later Madhyamikas

To this later period of Indian Buddhism, the last in the history of the Ma
dhyamaka in India, belong also the following Madhyamika masters several of
whose works are included in the Madhyamaka section of the Tibetan

m See D . S e y f o r t R u e g g , The gotra, ekaydna and tathdgatagarbha theories o f

th e Prajnaparam ita according to D harm am itra and A bhayakaragupta, in: The
Prajnaparam ita and related system s (Studies in honor o f E . Conze, Berkeley, 1977),
pp. 283312.
370 Cf. m K h a sgrub'rjedG e'legsdpal'bzah, rG yudsdespyrrnam , p. 96; sTon*-
thun*chen*mo, fol. 37 ab. Cf. above, p. 87 sq., 103.
371 A bhayakaragupta is know n to have worked w ith T ibetan translators, and he
w as for exam ple associated w ith sN u r D a r ma'ior: Dharma')grags in th e transla
tion m ade at N alanda o f Candrakirtis Sunyatasap tativrtti. Cf. Gos*g2on*nu*dpal,
Deb-ther-snon-po, cha, fol. 7 b (i, p. 342).
372 I t is unnecessary to underscore again the fact th a t th e historical information
available on th ese m asters is com plex and difficult to interpret w ith certainty,
116 D avid Seyfort R uegg * The L iteratu re of th e M adhyam aka School

Reference has already been made above to a later Jitri or Jetri who evi
dently lived c. 1000 (p. 100), as well as to a later Bhavya (p. 66), a later Can
drakirti (p. 81), and a later Jnagarbha (p. 69), all of whom flourished in the
eleventh century. On Advayavajra/M aitripada and Krsnapda see also above,
p. 107.
A certain dG e'bafi^lha (Kalynadeva, ubhadeva ?) is known as the author
of the *Bodhisattvacaryvatrasamskra.
Prajkaram ati (fl. c. 9501000) is reported to have been a contemporary of
Ratnkaranti, Jnarim itra and Ndapda (Nrop) at Vikramal.373 He
is the author of an extensive Pajik on Sntidevas Bodhicaryvatra and of
a work on the Prajpram it doctrine entitled Abhisamaylamkravrtti-
Parahitabhadra (eleventh century), a Kmri scholar, is the author of a
Vivrti on N grjunas nyatasaptati, which he translated into Tibetan to
gether with g2oirnu*mchog. He also composed a special commentary on two
verses of the M ahynastrlam kra.374
Vairocanaraksita (eleventh century), a contemporary of Dpamkararjna
and a scholar a t Vikramal, composed the ikskusumamajar and a P a
jik on the Bodhicaryvatra.375
Dharm karam ati (eleventh century), a pupil of Dpamkararjna, wrote a
commentary on the latters Satyadvayvatra.376
Another pupil of Dpamkararjna known as dBu*marsen*ge (*Madhya-
maka-Simha) wrote a short treatise on the philosophical systems.877
especially since the sam e nam e has often been borne by more th an one person. The
w ritings o f these m asters as well as the T ibetan and other records concerning th em
still aw ait detailed s t u d y ; often even their Sanskrit titles are uncertain.
878 See Tarantha, rG ya,g a r c h o s,byu, pp. 178, 181; Gos*g2onnu*dpal,
Deb*ther*son*po, a, fol. 2 a (i, p. 206). Cf. J. a u d o u , op. cit., p. 1323.
874 A (n earlier ?) Parahitabhadra is reported to h ave been th e pupil o f
R atnavajra, th e grandfather o f Sajjana and the great grandfather of Sk^majana.
Tarantha m akes him a contem porary o f K ing M ahipala (rg. c. 9881038 or
9771027), w hose death he however places at the tim e o f K ing K hrrrals death (! ?)
(rGya*garchos*byu, p. 172).A Parahitabhadra (evidently the one in question
above) is know n as th e teacher o f r og'B krldam sesT ab (10591109) and as a
translator at m T hodi. Cf. J. N a u d o u , op. cit., pp. 104, 176, 1823.
875 On (this ?) V airocanaraksita see Gosg2on*nu*dpal, D eb*thorson'po, da,
fol. 3 a sq. (ii, p. 844 sq.), who however m akes him also a disciple o f A bhaykara-
gu pta (who flourished c. 1100). The Tibetan translation o f ryaras Pram it-
sam sa included in the M adhyam aka section o f the bsT am gyur is ascribed to a
V airocanarak^ita.
878 This d K agrel (Pajik) is listed by B u'ston as having been translated into
T ibetan b y rGya^brCSomgrussen ge (Chos^byu, fol. 1 5 8 b 3). I t is n ot to be found
in the Peking and sD e dge editions o f the bsTan^gyur.For another Dharm kara
see above, note 326.
377 The IT a'ba'tha'dad'paTnam parp h y eba (*Sam kiptanndrtivibhga), a
te x t evidently w ritten down b y the U psaka Taro(Tra)srmitra, who translated
it into Tibetan w ith Chos*kyrses*rab.^ d B u m ar s e ge, Lion o f the M adhya
m ak a,* seem s to be an epithet o f D harm karam ati in th e Deb*therson*p> ca,
The Last Period of the Indian Madhyamaka School 117

Danaila wrote the Dhyanasaddharm avyasthanavrtti included in the Ma-

dhyam aka section of the b sT airgyur.378
BuddhaSrijnana was invited to Tibet in 1200 by Khro*phu-loca*ba Byams*
p ard p a l.379 He wrote the Jinam argavatara which contains sections on Ma
dhyam aka theory (nairatmya and prcUUyasamutpada) as well as on practice,
and a commentary on the Abhisamayalamkara entitled Prajnapradipavali.380
Vibhuticandra (fl. c. 1200) from Varendra was one of the group of so-called
Junior Pandits who accompanied the great Pandit of Kaim ir, SakyaSrlbhadra,
to Tibet in 1204 ;381 like his teacher he was connected with the Vihara of Jagad-
dala.382 He composed a quite extensive work the Sanskrit title of which is given
in the bsTan*gyur as Bodhicaiyavataratatparyapanjika ViSesadyotani.383

fol. 10 a (i, p. 2612). H ow ever, D harm akaram ati and d B ir m a r se h ge are listed
as different disciples of D ipam karairijnana in m ost o f the biographies relating to
him , including th e rN am thar rgyas pa and Taranatha, rG ya*garchosbyun, p. 188.
878 This t e x t translated b y R urch en 'b zaripo (9581055) relates to a work
ascribed to A vadhuti-pada also translated b y R in 'chenbzaiVpo.
According to Taranatha, a K&^mlri nam ed Dana^ila w as th e contem porary o f
K in g M ahlpala (rg. c. 9881038 or 9771027, w hose death he however places at
th e tim e o f K in g K h rrrals death in rGya*garchos*byun, p. 172); it m ay be this
Da(naslla) who is m entioned as a teacher o f Dlpamkara^rijnana (under the app el
lation m K has pa chen po D a chen po) b y G os'gion m u'd pal, D e b th e r s n o n p o ,
ca, fol. 2 a (i, p. 243) ( ?). Elsew here T aranatha has m entioned a D a n a iila as the
contem porary o f K ing Gopala (rg. c. 750770 or 775) (op. cit., p. 157); he m ay
then be th e D anasila who collaborated w ith dPal*br<5egs*raki-ta and d P al-byor*
snih'po in th e translation o f the H astavala-V rtti, and w ith Jinam itra, Silendrabodhi
and Ye*8es*sde in th e translation o f Candrakirtis Yukti$atik&-Vrtti (? ). Still
another D a n a iila was the contem porary o f Sakyairibhadra (11271225 or 1145
1243?; see below , note 381); see Bu*ston, Chos*byuh, fol. 140a (ii, p. 222).
Cf. J . N a u d o u , op. cit., pp. 21, 8687, 90, 186; and pp. 14, 198. I t is to bo noted
th a t a T ibetan tradition distinguishes betw een three persons bearing th e nam e
D anaiila; see e .g . Nor*chen*dKon*mchog*lhun*grub et al. D a n r p a i'chos'kyrbyuft*
6huWegs*par*bsad*pa, fol. 124b, 136b, 138a.
879 See Gos*g2on*mrdpal, Deb*ther*shon*po, na, fol. 135b (ii, p. 709); ba, fol. 3a
(ii, p. 1065); S u n rp a 'Y esesdpahbyor, R e u'm ig under th e year Icags sprel = 1200.
880 The Prajnapradipavali was translated into T ibetan b y th e author and gN ubs (!)
B yam s-pardpal. In B u 'ston s Chos*byuh (fol. 1 3 9 b 7 [ii, p. 222]), th e con tex t
suggests the identification o f th e author o f th e Jinam argavatara w ith th e Kasm lri
B uddhairijnana. (On the com m entary on th e Sam cayagathas, th e translation of
w hich is attributed to dPal*br6egs, w ho lived c. 800, see above, p. 102).
881 See B u'ston, Chos*byun, fol. 140a 1 (ii, p. 222); G os'g^om m rdpal, Deb*ther*-
snon*po, na, fol. 87a (ii, p. 600). For the dates of Sakyairibhadra (11271225 or
11451243) see D . S e y f o r t R u e g g , The Life o f Bu*6ton Rnrpo*che (Rom e, 1 9 6 6 ) ,
p. 4243 note.
882 On th e Jagaddala m onastery see th e introduction b y D . D . K o s a m b i to his
edition o f the Subhaitaratnakoga (Cambridge, Mass., 1957), p. x x x v ii note 7;
Y . K a j i y a m a , A n introduction to B uddhist philosophy (K yoto, 1966), p. 11.
383 V ibhuticandra appears to have been know n also b y th e epithet rN a k byor*
zlaba; see fol. 229 b 56 o f his work and the colophon. Cf. J . W . d e J o n g , I I J 16
(1975), p. 1 6 6 -7 .
118 David Seyfort Ruegg The Literature of the Madhyamaka School

In the final section of the Tarkabhs by Moksakaragupta, who was also

associated with the Jagaddala Vihra, the Madhyamikas theories of knowledge
and inference are briefly treated.384 This section follows the Jnasrasamuc-
caya, and it quotes Dharm akirti and Prajakaragupta.
Many of the later treatises mentioned above place special emphasis on the
practice of the Bodhisattva beside philosophical theory, and they are connected
with Madhyamaka thought in particular through Santidevas great works,
which were in fact the source of inspiration for so m any of them.

The fact th a t treatises belonging to the later period of the Indian Madhya
maka or at least closely connected with itin particular the Dur[ava]bodhlok
composed in Suvarnadvipa by Dpamkararjnas teacher Dharm akirti and
the Tik on Candrakirtis M adhyamakavatara composed in the Mi*ag region
by Jaynandawere written outside the Indian subcontinent clearly testifies
to the wide geographical diffusion and appeal of this school of thought.385
During the first 900 years of its existence the Madhyamaka thus spread from
India to both Central and E ast Asia, and perhaps also to some degree to I n

884 The Tarkabhaa has been edited b y E . K r i s h n a m a c h a r y a (Gaekwads Oriental

Series, vol. 94) and H . R . R a n g a s w a m i I y e n g a r (Mysore, 1952); it has been trans
lated b y Y . K a j i y a m a , A n introduction to B udd hist philosophy (K yoto, 1966).
Moksakaragupta is placed b y K ajiyam a betw een 1050 and 1292 since he quotes
R atnakirti and is quoted b y Malli?ena in his Syadvadam ahjari.
386 See above, p. 32 on the T a-chih-tu-lun

Works on practice in the Madhyamaka section of the bsTaw'gyur ascribed to

Avaghosa, rya-ra, Dharma-Subhti(ghosa), and Mtrceta

In the Madhyamaka section of the Tibetan bsT airgyur is found a poetical

work dealing with the six perfections (paramitd)liberality, conduct, patience,
energy,, meditation, and discriminative understandingentitled Pram it-
samsa by Arya-ra.388 If, as seems likely, he is the same person as the author
of the Jtakam l, he is perhaps to be placed in the third or fourth century (at
the latest).387 As to why a work on the pdramitds was included in the Madhya
maka section of the b sT airgyur, attention has already been drawn above to
the close connexion of the Mdhyamikas with Pram it philosophy. I t is also
to be noted th a t parallels to verses from another work tentatively ascribed to
rya-ra (the colophons refer in this case to carya ra)the Subhsitarat-
nakarandakakath found in both the Madhyamaka and Epistle sections of the
Peking bsTan-gyur388are to be traced in Sntidevas Bodhicaryvatra and
in the Prajdanda ascribed to Ngrjuna, as well as in Arya-uras own J ta
kaml and M trcetas atapacatka.389
According to an identification recorded by Tarantha (and apparently going
back to Indian sources) Arya-Sra, M trceta and Dharm a/Dhrmika-Subhti
are nothing but various names under which Avaghosa was known.390 Ava-
ghosas works would seem to have been familiar to Ngrjuna and other M-
dhyam ikas;391 and the Prattyasam utpdahrdaya (verse 7), which is tradition
ally ascribed to Ngrjuna, paraphrases a verse found in Avaghosas Saun-

388 Ed. A. F e r r a r i , II Compendio delle perfezioni (Annali Lateranensi, vol. x ,

V atican City, 1946).
887 On th e question o f his d ate, beside the standard histories o f Sanskrit litera
ture, see L i n Li-kouang, L aide-mmoire de la Vraie Loi (Paris, 1 9 4 9 ) , p p . 1 0 2 ,
3 1 0 -1 3 .
888 B u t n ot in th e M adhyam aka section o f th e sD e dge b3Tan*gyur.
389 See H . Z i m m e r m a n n , D ie Subh^ita-ratna-karanlaka-kath (W iesbaden,
1975), pp. 1 8 -2 1 .
890 Trantha, rGya*gar,chos*byun, p. 71. The question o f th e association of
these persons and their works has been m uch m ooted in the past. See S. L v i , J A
1928, ii, pp. 2047; 1929, ii, pp. 25585; and 1936, p. 101 sq. (on Mtrceta) ; L. de
L a V a l l e P o u s s i n , L Ind e aux tem ps des Maurya . . . (Paris, 1930), p. 324;
P. M us, La lum ire sur les six voies (Paris, 1939), pp. 117; L i n Li-kouang, op.
cit., pp. 103 sq., 306 sq. (who notes, p. 31112, th at according to a Chinese tradition
Aryadeva was a co-author o f th e com posite work in the Chinese canon correspond
ing to the Jtakam l).
891 See E . H . J o h n s t o n , Buddhacarita, ii (Calcutta, 1936), pp. x x ix , 232.
120 David Seyfort Ruegg The Literature of the Madhyamaka School

darananda (xiii. 44).392 As for Matrceta, he is reported to have been a disciple

of Aryadeva;393 and this association might also help to account for the fact
th at works attributed to these authors have been included in the Madhyamaka
section of the bsTan-gyur (in particular in the Peking edition).394
M atrceta and Nagarjunaas well apparently as ASvaghosaare in addition
referred to as ascetics concerned with the affairs of the realm (yatayo rdjya-
vrttinah, ManjuSrimulakalpa liii. 867 sq.).
Examples of such works included in the Madhyamaka and Epistle sections
of the bsTan*gyur are the Maharajakani(s)kalekha (classified alongside Nagar-
junas Suhrllekha and Candragomins Sisyalekha), Caturviparyaya[parihara]-
katha and Kaliyugaparikatha ascribed to M atrceta; the Astaksanakatha, Para-
m arthabodhicittabhavanakram a,395 SamvrtibodhicittabhavanopadeSa, and
Da6aku6alakarmapathanirdeSa398 ascribed to Aivaghosa; and the Saddhar-
masm rtyupasthana-Sadgatikarikas,397 Da6akualakarmapathanirde6a and Bo-
dhisattvacarya[sam graha]pradiparatnam ala ascribed to Dharma or Dharm ika
Matrceta is furthermore renowned as the author of lyric devotional hymns
addressed to the Buddha,399 the cataphaticism of which is comparable to th a t of
392 See above, p. 28.H suan-tsang m akes A ivag h o sa and Nagarjuna co n tem
poraries at th e tim e o f Kanika; cf. S. B e a l , Si-yu-ki: B u dd hist records o f the
w estern world (London, 1906), ii, p. 302; and T. W a t t e r s , On Y uan Chwangs
travels in India, i (London, 1904), p. 245.
393 See Bu*ston, Chos^byun, fol. 103a (ii, p. 136). Cf. Taranatha, rG yagarchos*
byun, p. 70 sq.
394 I t is to be n oted th a t m an y o f these works are found a second tim e elsewhere
in th e P ek in g edition, in th e sPriiVyig (Epistle) section of the bsTan*gyur. In th e
sDe*dge edition several o f these works are n o t to be found in the M adhyam aka
section at all, but in th e sPrih'yig and J o b o rc h o schuh sections.
393 This work in particular is quoted in th e T ibetan Siddhanta (Grubmtha*)
literature on th e M adhyam aka, where it is ascribed to Arya-Sura. I t is clearly a
later com position.
396 E d ited in part b y S. L e v i , J A 1929, ii, p. 2689.
397 See S. L vt, J A 1925, i, p. 37 s q .; P. M us, op. cit.; L in Li-kouang, op. cit.,
p. 104 sq. Beside th e Sanskrit te x t of the Sadgatikarikas (edited by Mus) and
their T ibetan translation, we have versions in Chinese and in P ali; see M us, op. cit.
Contrary to w hat has been stated by Mus (pp. 2, 16), the colophon o f the T ibetan
translation (Peking edition) makes no m ention o f the Karikas being extracted from
th e Saddharmasm rtyupasthana-swira.
398 The *Sarvayanalokabhaya (?) also included in the M adhyam aka section o f
the bsTan*gyur and ascribed to Subhutigho?a was translated into Tibetan by its
author w ith th e assistance of T ih h e ,1jin ,bzah*po; this writer m ust therefore be a
later nam esake o f our D harma-Subhuti(ghoa), who lived much earlier (see below,
p. 121 note 404). The work deals w ith the schools o f Indian philosophy, including
som e later developm ents.
399 See D . R . S h a c k l e t o n B a i l e y , The Varnarhavarnastotra o f M atrceta,
BSOAS 13 (194950), pp. 671 sq., 947 sq., and The Satap an caiatka o f M atrceta
(Cambridge, 1951); D . S c h l i n g l o f f , D ie B uddhastotras M atrcetas (Abh. der
deutschen Akad. der W issenschaften, Berlin, 1968); P. P y t h o n , Sugatapahcatrim -
satstotra, in Yinayavini^caya-Upalipariprccha (Paris, 1973).
Appendix I 121

certain hymns ascribed to Nagarjuna.400 These works are included in the Hymn
section of the bsT airgyur.
W hatever historical connexions may lie behind these associations between
persons and texts, all of which are undoubtedly not of the same period, these
works were evidently regarded (at least by some later Madhyamikas) as forming
a cycle of related tracts (parikathd, etc.), epistles and manuals setting out the
ethical, devotional and ritual praxis followed in the Madhyamaka school.401 For
while the Madhyamaka as a school is no doubt concerned essentially with
philosophical theory, the Madhyamikas have regularly coordinated theory with
practice on the pragmatic level of vyavahdra or samvrti embracing 'worldly*
(laukika) conduct, devotion, ritual, and other forms of realization.
Aivaghosa (I) has been usually placed in the first or second century, at the
time of or shortly before Kaniska I ; his doctrinal affiliations are difficult to
determine with precision.402 M atrceta is generally placed in the second century;
his particular doctrinal affiliations are also difficult to determine because of
the nature of his writings.403 Dharma-Subhuti also m ust be placed fairly early;
he evidently belonged to an Abhidharmika line.404

400 See above, p. 31, and D . S e y f o r t R t j e g g , Le D harm adhtustava de N agar

juna in tu des tibtaines ddies la mmoire de Marcelle Lalou (Paris, 1971),
p. 456 sq.
401 P. M us, op. cit., pp. 18992, has described th e D aiakusalakarm apathanir-
dea, the Dakualakarmapathanirdesa, th e Suhrllekha, and th e Salgatikriks
as elem entary school-m anuals ; but such a description can cover only one o f their
402 See E . H . J o h n s t o n , B uddhacarita, ii (Calcutta, 1 9 3 6 ) , p. xiii sq, on the date.
Joh nston suggests connecting Avaghoa w ith th e Bahurutika section o f the
M ahsm ghikas (p. x x v iii sq.). B ut B . C. L a w , A iva gh osa (Calcutta, 1 9 4 6 ) , pp.
1 6 1 7 , proposes linking him w ith the D harm aguptakas ; and J. W. d e J o n g , I I J 2 0
( 1 9 7 8 ) , pp. 1 2 5 6, envisages a link w ith th e Sautrntikas. Avaghoa I has pre
sum ably to be distinguished from the author o f the Gurupancsik, and also of the
Param rthabodhicittabhvankram a (see above, n ote 3 9 5 ) .
403 D . R . S h a c x x e t o n B a i l e y , The Satapancasatka o f Matrceta, p. 7 sq., has
placed M atrceta in the second century, at th e tim e o f Kani($)ka II ( ?), to w hom his
E pistle is apparently addressed. Sim ilarly A. K . W a r d e r , Indian K v y a literature,
ii (Delhi, 1974), p. 242. H ow ever, . L a m o t t e . H istoire du bouddhism e indien, i
(Louvain, 1958), p. 656, supposed Mtrce^as E pistle to h ave been addressed to
K anika at the end o f the third century. T rantha distinguishes betw een an
earlier and later M atrceta (rGyagar'chos*byuri, p. 75; cf. also p. 152).
404 A verse b y D harm a-Subhuti is found in V asubandhus Abhidharmakoa-
bhya (iii. 59) and in th e M ahvibh, as w ell as in Chinese translations o f th e
earliest period b y A n Shih-kao and Sam ghavarm an. See L i n Li-kouang, op. cit.,
p. 102 sq.

Ratndkaradnti1s Vijapti-Madhyamaka1

In connexion with the synthesis of Yogcra and Madhyamaka thought dis

cussed above (p. 87 sq.) mention should be made of another synthesis of
M aitreyanthas and Asagas Vijnavda philosophy with certain elements
of N grjunas Madhyamaka th a t was worked out by R atnkaranti (c. 1000).405
In his M adhyamaklamkravrtti-M adhyamapratipadsiddhi and Madhya-
m(ak)lamkropadea R atnkaranti firmly rejects the idea th a t all is false,
stating th a t it is incompatible with N grjunas as well as with M aitreyanthas
and Asagas teachings. B ut in both these compositions, which are contained
in the C ittam tra (Sems*cam) section of the bsTan-gyur, none of the main lines
of development of the Madhyamaka discussed above prevails ; and R atnkara-
ntis doctrine is in essence a Nirkra-Vijnavda teaching with the Middle
W ay presented following the system of the three svabhdvas (the parikalpita0,
the paratantra0 and the parinispanna0) of the Madhyntavibhga for example.
The traditions of the logico-epistemological schools also occupy a prominent
place in R atnkarantis doctrine; and he takes his opponents to task for
having neglected the pramdnas and logical procedures.408 R atnkrantis
doctrine has been referred to as the Vijapti-M adhyamaka (rnam rig dbu
ma).1 And the above-mentioned Upadea is said to have been m itte n to
oppose the doctrine of B hadanta Candrakirti, who had strayed from the true
purport of Ngrjuna (but who later abandoned nihilism in his Tantrik com
m entary).408
In the M adhyamaklamkravrtti (fol. 136ab) Ratnkaranti makes the
stages of understanding recognized by the Lankavatrastra into four bhmis
of the Yogcra.409 The first yogabhmi consists in taking all existing dharmas
as (noetic) objects (alambana). The second yogabhmi consists in understanding
th at there is no external object of knowledge (grdhya); since everything ap

405 R atnkaranti is stated to have been a contem porary a t Vikramal o f

Jnarimitra, Prajkaram ati and N dapda; Maitripda m et w ith him ; and
Brog*mi was am ong his Tibetan disciples. See Gos*g2on*nu*dpal. Deb*ther*snon*p>
kha, fol. 2 a (i, p. 206); da, fol. 2 a (ii, p. 842).
400 Madhyam(ak)lamkropadea, fol. 264b sq.
407 See for exam ple the colophon to the M adhyam(ak)lam kropadea, fol.
267 a 3. Cf. above, p. 56 on the rnam rig (g i) dbu ma.
408 See the colophon to th e M adhyam(ak)lamkropadea, fol. 2 6 6 b 45. A
pseudo-M adhyam aka (dbu ma Itar m a n : Madhyamakbhsa) holding th a t all is
false is m entioned in the M adhyam aklam kravrtti, fol. 138 a 2.
409 Lakvatrastra x. 256 sq. See above, p p. 90, 96.
Appendix II 123

pears as mind only (cittamatra) no dharma is different from mind. The third
yogabhumi consists in transcending this cittamatra and understanding that, be
cause no grahya exists, no corresponding cognizing subject (grahaka) can exist
either; cittamatra then becomes residence in the alambana of tathata, a non-dual
gnosis which is without the grahaka-laksana. Finally the fourth yogabhumi is
direct comprehension of the mahdyana consisting in residence in gnosis ab
solutely free from appearance (nirdbhasa), and in which naman and laksana
as well as grahya and grahaka have disappeared. In the same authors Upadea
(fol. 266a) these four stages are referred to as the four bhumis having respec
tively the alambana of the existence of things in terms of the extreme of eter-
nalism (as opposed to nihilism), the alambana of cittamatra, the alambana of
tathata, and absence of alambana.
In his Triynavyavasthna included among the Tantra-Commentaries
(rGyud-grel) of the bsTan-gyur, Ratnkaranti has divided Mdhyamikas
into those who m aintain th a t relative knowledge (samvrtijdna) has images or
modes (dkdra) and those who hold th a t traces (vasand) only are there (fol.
114a), a division inspired by the well-known classification of the Vijnavdins
as Skravdins-Satykravdins and Nirkravdins/Alkkravdins.410 Ac
cording to the first group of Mdhyamikas, while citta and jdna (ye es) do
not exist on the paramdrtha-leve 1, all determinations (vyavastha) on the sur
face-level of samvrtithe akdras of citta and manasare thought to exist in the
form of cognitive objects (visaya). The second group who deny th a t jdna has
modes agree with the former about the paramdrtha-level, but they hold th at
the determinations of samvrti are nothing but vasand; and citta is therefore
free from appearance in the nature of dkdras. Both groups are stated to agree in
rejecting any substantial self-nature for pratityasamutpada (fol. 114 b115 a).
R atnkaranti also cites authorities who stated th a t the two groups differ in
their rejection of other doctrines by means of the catuskoti, which the first
formulate in terms of existence/non-existence whereas the latter formulate it in
terms of permanence/impermanence (fol. 115a).
In another major treatise, the Prajnpramitopadea, R atnkaranti also
discusses the connexion between the Vijfinavda/Cittamtra and the Madhya-
maka. Beside the Lakvatrastra he quotes in particular N grjunas Yuk-
tisastik in support of the doctrine of consciousness only (fol. 161b162a; cf.
169b8).411 And he concludes th a tnotwithstanding any other differencesthe

410 On this classification see for exam ple ntarakita, T attvasam graha 1998,
and Kam alaila, P ajik 2081; Mok^akaragupta, Tarkabh^ (ed. H . R . R anga-
8wami Iyengar, Mysore, 1952), pp. 23, 63 (the first passage is missing in Krishnam a-
charyas edition, p. 11, b u t th e second is found on p. 34); Gunaratna, com m entary
on Haribhadrasuri, Saddaranasam uccaya 11 (p. 47); Sarvadaranasamgraha (ed.
Abhyankar), p. 46; A dvayavajra, T attvaratnval, p. 1819. Cf. Y . K a j i y a m a ,
I B K 14/1 (1965), p. 26 sq., and A n introduction to B ud dh ist philosophy (K yoto,
1966), p. 62, 1 5 4 sq.; H . N a j x a m t o a , A B O R I 4849 (1968), p. 119.
411 Yukti?atika 22 ( ?) w ith however th e variant reading es p a (instead o f dgos
pa) also found in Santarak^itas M adhyam aklam kravrtti (92): dos rnam s skye
124 David Seyfort Ruegg The Literature of the Madhyamaka School

Yogcrins and the Mdhyamikas finally agree concerning the four above-men
tioned yogabhmis (fol. 169 b). In addition, R atnkaranti criticizes the doc
trine of those philosophers, Mdhyamikas as well as Yogcrins, who hold th a t
knowledge has an image (sdkara-jnana) (fol. 168a 4, 170a4).
The only work by R atnkaranti actually included in the Madhyamaka
section of the bsTan-gyur is his Strasamuccayabhsya-ratnloklamkra
which contains extracts from canonical Stras on many of the principal topics
of M ahynist doctrine together with explanations derived from Ngrj unas
and Asagas teachings. Thus, like the Strasamuccaya ascribed to Ngrjuna,
this work is not concerned exclusively with doctrines of the Madhyamaka
school in the strict sense.

ba yod m in zin j /gag p a'ah gah na yod m in iih j /es p a 'd i h id kho na n il \shye Hh
'gag p a r 'gyur ba yin U (but see Kam ala& las P anjika, fol. 138ab). A nd Y uk ti-
$atika 35: 'byun ba che la sogs bSad p a \ Irnam pa r es su ya h dag 'duj lye [read: de]
es k yis n i da bral n a / /log par rnam brtags cis m a y in //. Cf. Y . K a j i y a m a , L ater
M adhyam ikas, in: M ahayana B udd hist m editation, ed. M. K iy o ta (H onolulu, 1978),
p. 132, and the synopsis published b y S. K a t s u r a , I B K 25/1 (1976), pp. 3841;
above, n ote 44.
The source o f the first o f th ese tw o verses and R a tn ak araian tis interpretation
have been critically discussed b y Ooh*kha*pa, L egs'bsadsnirrpo, fol. 56a.

A dvayavajra
Apratithnapraka. E d. Haraprasad Shastri, A dvayavajrasam graha, Gaek-
w ad s Oriental Series xl, Baroda, 1927, p. 48.
Amanasikrdhra. Ed. Haraprasad Shastri, op. cit., p. 6062.
T attvadasaka. Ed. Haraprasad Shastri, op. cit., p. 59.
Tattvapraka. E d. Haraprasad Shastri, op. cit., pp. 4647.
T attvaratnvall. Ed. Haraprasad Shastri, op. cit., pp. 1422; H . U i, Shinri no
hkan, N ago ya D aigaku B ungakubu K enkyronsh iii, T etsugaku, i, pp. 131.
Prem apancaka. E d. Haraprasad Shastri, op. cit., p. 58.
M adhyam aatka, E d . Haraprasad Shastri, op. cit., p. 57 ; S. K . P athak, A L B 25
(1961), pp. 54647 (with an English translation).
M ahynavim ik. E d. Haraprasad Shastri, op. cit., p. 5455.
M ynirukti. E d. Haraprasad Shastri, op. cit. p. 44.
Panjik on th e D ohkoa (of Saraha). Ed. P. C. Bagchi, Calcutta Sanskrit Series,
Calcutta, 1938.
C atuhstavasam srtha. E d. G. Tucci, Minor B ud dh ist T exts, i, Serie Orientale
R om a ix. 1, R om e, 1956, pp. 23846 (from an incom plete MS containing a part
o f the com m entary on th e N iraupam yastava and the com m entary on th e A cin ty a0
and Param artha-stava),

(B odhisattvayogcra-)Catuhataka. Sanskrit fragm ents ed. Haraprasad Shastri,
Memoirs o f the A siatic Society o f B engal, vol. iii, no. 8, pp. 449514, Calcutta,
1914. (Cf. P . L. V aidya, tu d es sur Aryadeva et son Catuhataka, Chapitres
v iiixv i, Paris, 1923 [w ith French translation]. V idhushekhara B hattacharya,
The Catuhataka o f A ryadeva w ith extracts from the com m entary o f Candrakir-
tti, reconstructed from th e T ibetan version w ith an E nglish translation, Chapter
vii, Proceedings and transactions of the F ourth Oriental Conference, Allahabad
1928, pp. 83171; The Catuhataka o f A ryadeva: Sanskrit and Tibetan tex ts
w ith copious extracts from th e com m entary o f Candrakirtti, reconstructed and
edited, P art ii, V isva-B harati Series no. 2, Calcutta, 1931 [includes Chapters
viiixvi]. The K riks o f Chapters viix v i h ave been republished, together w ith
a Sanskrit version o f parts o f Candraklrtis V rtti, and rendered into H indi b y
Bhagcandra Jain Bhskar, Catuhatakam, N agpur, 1971. Chapters ix x v i have
been translated into Italian on the basis o f H s an -tsan gs Chinese version b y
G. Tucci, R SO 10 [1923], pp. 5 2 4 -6 7 .)
Cittaviuddhiprakarana (presumably b y a D eutero-A ryadeva), ed. P . P atel,
V isvabharati Studies 8, Santiniketan, 1949. Cf. H araprasad Shastri, JA S B 67
(1898), pp. 17584; B hagcandra Ja in Bhskar, C atuhatakam, Nagpur, 1971,
pp. 1619.
K am alala
Tattvasam grahapanjik. See under Sntarakita.
B hvankram a. F irst Bhvankram a, ed. G. Tucci, Minor B uddhist tex ts, ii,
126 David Seyfort Ruegg The Literature of the Madhyamaka School

Serie Orientale R om a ix. 2, R om e, 1958, pp. 185229. (French translation b y

J . van den Broeck, La progression dans la m ditation, P ublication de l In stitu t
B elge des H au tes tu des Bouddhiques, Srie tud es et te x te s , N o. 6, Brussels,
1977. A brief analysis o f this te x t based on th e Chinese translation was published
b y P . D em iville, L e concile de L hasa, Paris, 1952, pp. 3335.)Third B hvan-
krama, ed. G. Tucci, Minor B uddhist tex ts, iii, Serie Orientale R om a xliii, R om e,
1971. (Italian translation by C. Pensa, RSO 39 [1964], pp. 21142. A partial
French translation from th e T ibetan was published earlier b y . L am otte in
P . D m ieville, Le concile de Lhasa, pp. 33653).

Kam bala(pda) /K am(b)alm bara(pda) /K amala(pda)

rytashasrikyh Prajnpram ityh Pindrthah (N avasloki). Ed. G. Tucci,
Minor B uddhist tex ts, i, Serie Orientale R om a ix. 1, R om e, 1956, pp. 21617
(with T ibetan versions, Chinese version and English translation).

(Bodhisattvayogcra-)Catuhataka^k. F ragm ents ed. Haraprasad Shastri,
Memoirs of the A siatic Society of B engal, iii, no. 8, pp. 449514, Calcutta, 1914.
(See also under ryadeva).
P rasannapad M adhyam akavrttih. E d. L. de L a Valle Poussin, B ibliotheca
B uddhica iv, St. Petersburg, 190313. (See also under N grjuna).
M adhyam akastrastuti. E d. J. W . de Jong, Oriens E xtrem u s 9 (1962), pp. 4951
(with the Tibetan translation and a French translation).

N grjuna
Catuhstava. The N iraup am yastava and Param arthastava have been edited b y
G. Tucci, JR A S 1932, pp. 30925 (with E nglish translation). (The Sanskrit
versions o f th e N irupam a0, L ok atita0, A cin tya0, and S tu tyatita-stavas published
b y Prabhubhai P atel, IH Q 8 [1932], p. 317 sq., were restorations from the Tibetan
translations o f these hym ns). (An Italian translation was published b y R . Gnoli
as an appendix to his N grjuna: M adhyam aka Krik, Turin, 1961).
Pancakram a (presumably b y a D eu tero -Ngrjuna). E d. L. de L a Valle P o u s
sin, tu des et tex tes tantriques: P ancakram a (U niversit de Gand, R ecueil de
travau x publis par la F acult de Philosophie et L ettres, 16mc fascicule), Gand
e t L ouvain, 1896.
P ratityasam utpdahrdayakrik. E d. V. V . Gokhale, Studia indologica (F est
schrift W . Kirfel, B onn, 1955), pp. 1034 (verses 15). See also L. de L a Valle
Poussin, Bouddhism e, tu des et m atriaux: Theorie des douze causes, Gand,
1913, pp. 1234 (French translation from th e Tibetan) ; V. V. Gokhale, P rattya-
sam utpdastra des U llaigha, Bonn, 1930.
M ahynavim ik (apparently b y a D eutero-N grjuna). Ed. G. Tucci, Minor
B u dd hist tex ts, i, Serie Orientale R om a ix. 1, R om e, 1956, pp. 2013 (with
E nglish translation).
M lam adhyam akakriks (as found in the Prasannapad M adhyam akavrttih b y
Candrakirti). E d. together w ith th e Prasannapad by L. de La Valle Poussin,
B ibliotheca B uddhica iv, St. Petersburg, 190313; reprinted w ith alterations b y
P . L. V aidya, M adhyam akastra o f Ngrjuna, B udd hist Sanskrit T ex ts no. 10,
D arbhanga, 1960. N ew edition (using a fourth manuscript o f the K riks only) b y
J. W . de Jong, Adyar, 1977. (The first translations into European languages were
b y Th. Stcherbatsky, A ppendix to : Conception of B uddhist N irvana, Leningrad,
1927 (Chapter i and x x v , in E nglish); S. Schayer, Feuer und Brennstoff, RO 7
(1931), pp. 2652 (Chap. x, in German), and A usgewhlte K ap itel aus der Pra-
Modern Editions of the Sanskrit Texts of the Madhyamaka School 127

sannapada, K rakow , 1931 (Chap. v , x iixvi, in German); I. L am otte, MCB 4

(1936), pp. 26588 (Chap. xvii, in French); J. W . de Jong, Cinq chapitres de la
Prasannapada, Paris, 1949 (Chap. x v iiixxii, in French); J . May, Candrakirti,
Prasannapada M adhyam akavrtti, Paris, 1959 (Chap. iiiv, v iix, xi, xxiii,
x x iv , x x v i, and x x v ii, in French). There are in addition com plete E nglish tran s
lations b y F . J. Streng, E m ptiness, N ashville and N ew York, 1967, pp. 183220;
and b y K . K . Inada, Nagarjuna, a translation o f his M ulam adhyam akakarika
w ith an Introductory E ssay, Tokyd, 1970. An Italian translation was published
b y R . Gnoli, N agarjuna; M adhyam aka Karika, Le stanze del cam m ino di mezzo,
Turin, 1961. An E nglish translation o f Chapters ivii has been published b y
H . N . Chatterjee, M ula-M adhyamaka-Karika, 2 vol., Calcutta, 195762. A nd
another English translation of Chapters ivi, v iiix, xiii, x v , x v iiixix, x x ii
x x v together w ith an abridged version o f Candrakirtis Prasannapada thereon
has recently been published b y M. Sprung in collaboration w ith T. R . Y. Murti
and U . S. Y yas, L ucid exposition o f th e Middle W ay, London, 1979. A German
translation o f Chapters i, x v , xviii, x x iv , and x x v was published b y E . Frauwall-
ner, D ie Philosophic des Buddhism us, Berlin, 1956, p. 178 sq.
R atnavalL P arts o f Chapters i and ii and the whole o f Chapter iv h ave been
edited b y G. Tucci, JR A S 1934, pp. 30725, and 1936, pp. 23752 and pp.
42335 (with an E nglish translation). This edition has been reprinted b y P. L.
V aidya, M adhyamaka^astra o f Nagarjuna, B udd hist Sanskrit T ex ts no. 10,
Darbhanga, 1960, A ppendix Six, pp. 296310; and b y H . Chatterjee Sastri, The
philosophy o f N agarjuna as contained in the R atn avali, P art i, Calcutta, 1977,
pp. 83100. (An E nglish translation based on th e T ibetan translation has been
published b y J. H opkins and L ati R im poche, The precious garland, London,
Y igrahavyavartani w ith th e auto com m entary. E d. E . H . Joh nston and A. K unst,
MCB 9 (194851), pp. 99152; reprinted, together w ith an English translation,
in K . B hattach arya, The dialectical m ethod o f N agarjuna, D elhi, 1978. (This
edition replaces the inadequate one b y K . P. Jayasw al and R ahula S&nkritya-
yana, J B O R S 23 (1937), which was reprinted w ith alterations b y P. L. V aidya,
M adhyam akasastra o f N agarjuna, B uddhist Sanskrit T exts no. 10, D arbhanga,
1960, A ppendix F iv e, pp. 27795). Earlier E nglish versions o f the Sanskrit
Karikas were published b y S. Mookerjee, The ab solu tists standpoint in logic,
N ava-N alanda-M ahavihara R esearch Publication, i, N alanda, 1957, pp. 741;
and b y F . J . Streng, E m ptiness, pp. 2227. The first E nglish translation, based
on the T ibetan and Chinese translations, was published b y G. Tucci, Pre-D innaga
B uddhist tex ts on logic from Chinese sources, Gaekw ads Oriental Series no. xlix,
Baroda, 1929. A French translation based on th e T ibetan and Chinese transla
tions was published b y S. Y am aguchi, JA , Juillet-septem bre 1929, pp. 186. A n
Italian translation o f the Sanskrit Karikas was published b y R . Gnoli as an a p
p endix to his N agarjuna: M adhyam aka Karika, Turin, 1961, p. 139 sq.

Prajnakaram ati
Bodhicaryavatarapanjika. E d ited , together w ith S an tidevas Bodhicaryavatara,
b y L. de L a Vallee Poussin, B ibliotheca Indica, C alcutta 190214; P . L. Vaidya,
B uddhist Sanskrit T exts N o. 12, Darbhanga, 1960.

B h a vavivek a
M adhyam akahrdayakarikas. Chap. ii. 112, ed. V. V. Gokhale, I I J 14 (1972),
pp. 4244. Chap. iii. 1136, ed. S. Iida, R eason and em ptiness, a S tud y in
logic and m ysticism (Tokyo, 1980); iii. 27584, ed. V. V. Gokhale, I I J 5 (1961
62), p. 273. Chap. iv. 7 and 56, ed. V. V. Gokhale, I I J 2 (1958), p. 179. Chap.
128 David Seyfort Ruegg The Literature of the Madhyamaka School

viii. 116, ed. V. V. Gokhale, I I J 2 (1958), pp. 16777; viii. 1896, ed. H . N a k a
mura, A L B 39 (1975), pp. 30129. (On v. 1, 79, see S. Iida, K anakura kinen-
ronbunsh [Kanakura Festschrift], T okyo, 1966, pp. 7996.)

R hulabhadra
Prajnpram itstotra. E d. R . H ik ata, Suvikrntavikrm i-Paripycch Prajn-
pram it-stra, F ukuoka, 1958, pp. 12. (English translation b y E . Conze,
B u dd hist tex ts through th e ages, Oxford, 1954, pp. 1479).

T attvasam graha. E d ited together w ith K am alalas T attvasam grahapanjik b y
E . K rishnam acharya, G aekw ads Oriental Series vol. x x x x x x i, Baroda, 1926;
S vm D vrikds str, B auddha Bharati Series 12, Varanasi, 1968. (English
translation b y G anganath Jha, The T attvasangraha o f Sntarakita w ith th e
com m entary o f Kam alala, Gaekwads Oriental Series vol. lx x x and lx xxiii,
Baroda, 19379.Individu al chapters have been translated b y W . L iebenthal,
Satkrya in der D arstellung seiner buddhistischen Gegner, Stuttgart, 1934 (Ger
m an translation o f chap. i, Prakrtipark^) ; S. Schayer, Kam alailas K ritik des
Pudgalavda, RO 8 (1932), pp. 6893 (German translation o f chapter vii,
tmapark) and Contributions to th e problem o f tim e in Indian philosophy,
Krakow, 1938, pp. 2970 (English translation o f chapter xx i, Traiklyapark) ;
A. K un st, Problem e der buddliistischen Logik, Krakow, 1939 (German tran sla
tion of Chapter xviii, Anum napark, together w ith an edition o f the Sanskrit
and T ibetan tex ts o f the Kriks).)

n tideva
Bodhicaryvatra. E d. L. de L a Valle Poussin together w ith Prajnkaram atis
Bodhicaryvatrapanjik, B ibliotheca Indica, Calcutta, 190214; P. L. V aidya,
B uddhist Sanskrit T exts N o. 12, Darbhanga, 1960. (French translation b y L. de
L a Valle Poussin, Paris, 1907, originally published in R evu e d histoire et de
littrature religieuses 1012 [19057]. Other translations are b y L. F in ot, La
marche la lum ire, Paris, 1920; R . Schm idt, Der E in tritt in den W andel der
E rleuchtung, Paderborn, 1923; and G. Tucci, II cam m ino verso la luce, Torino,
Siksam uccaya. E d. Cecil Bendall, B ibliotheca B uddhica i, St. Petersburg,
1902. (English translation b y C. B endall and W . H . D . R ouse, A com pendium
o f B udd hist doctrine, London, 1922).

Dohkoa. E d. P . Ch. Bagchi, Dohkoa, P art i, Calcutta Sanskrit Series, pp.
1450, Calcutta, 1938 (with th e com m entary o f A dvayavajra). See also Journal
o f the D epartm ent o f L etters, U niversity o f Calcutta, vol. x x v iii (Calcutta, 1935),
op. 62120; R hul Snkrtyyan, Dohko, P atna, 1957 (w ith T ibetan and
H indi translations).
N . B . T exts o f th e Indian M adhyam aka school m entioned in this book w hich
are n o t in the above list o f Sanskrit tex ts are available, either in single or m u l
tip le translations, in Chinese in th e Taish shinsh daizky and/or T ibetan
in the bsT am gyur (the P eking edition o f which has been used in this book).

bi brgya pa (Catuhataka). Chapters viiix v i edited, together w ith parts o f
Candrakrtis com m entary, b y V idhushekhara B hattacharya, The Catuhataka:
Sanskrit and Tibetan te x ts w ith copious extracts from th e com m entary o f Can-
drakirtfci, P art ii, V isva-B harati Series N o. 2, Calcutta, 1931.
Y i ge brgya pa (and Grel pa) (Akaraataka [and Vrtti], ascribed to N grjuna
b ut probably b y Aryadeva). Cf. V. V. Gokhale, Akara-atakam (the H undred
L etters), M aterialien zur K unde des Buddhism us x iv , H eidelberg, 1930.
Y e es si po kun las btu s pa (Jnnasrasam uccaya, probably b y a Deutero-
A ryadeva). Verses 2028 ed. K . Mimaki, La rfutation bouddhique de la p er
manence des choses, Paris, 1976, pp. 1868 (with French translation).

Kam alala
(b)sGom p a i rim pa (da po) (Bhvankram a [I]). Ed. S. Y oshim ura, Tibetan
buddhistology, K yo to , 1953; G. Tucci, Minor B udd hist tex ts, ii, Serie Orientale
R om a ix. 2, R om e, 1958, pp. 22982.
(b)sGom p a i rim pa (bar ma) (B hvankram a [II]). Ed. S. Y oshim ura, op. cit.
(b)sGom p a i rim pa (tha ma) (B hvankram a [III]). Ed. S. Y oshim ura, op. cit.
Translated in part b y E. L am otte in P. D em iville, Le concile de Lhasa, Paris,
1952, A ppendix II, pp. 33653.
D e kho na id bsdus p a i d k a grel las rJes su dpag pa brtag pa. (Chapter x x o f
th e T ibetan version o f the Tattvasam grahapajik, Anumnapark^). Ed.
A. K unst, Kam alailas com m entary on Sntarakitas A num napark o f the
Tattvasam graha, MCB 8 (1947), pp. 166211.

K am bala(pda)/K am (b)alm bara(pda)/K am ala(pda)

bCom ldan das m a es rab k yi pha roi tu phyin p a i don bsdus p a i higs su bead
p a dgu pa (Navaloki). E d. G. Tucci, Minor B ud dh ist tex ts, i, Serie Orientale
R om a ix.i, R om e, 1956, pp. 21822 (both Tibetan translations w ith th e Sanskrit
original, a Chinese version, and an E nglish translation).

P h u po l a i rab tu b yed p a (Paficaskandhaprakarana). E d. C. Lindtner, AO 40
(1979), pp. 87145.
B y a chub sems d p ai rnal byor spyod pa bi brgya p a i rgya cher grel pa
(B odhisattvayogcra-Catuhatakatk). E xtracts from Chapters viiixv i
published b y Vidhushekhara B hattacharya, The Catuhataka o f A ryadeva, V isva-
B harati Series N o. 2, Calcutta, 1931.
dB u m a rca b ai grel pa hig gsal ba (M lam adhyam akavrttih Prasannapad).
Chapters x v iiixxii ed. J. W . de Jong, Cinq chapitres de la Prasannapad, Paris,
1949; chapters iiiv , v iix, x i, xxiii, x x iv , x x v i, x x v ii, ed. J. May, Candrakirti,
P rasannapad M adhyam akavrtti, Paris, 1959. The T ibetan translation o f the
M adhyam akastrastuti found at the end o f the Prasannapad has been edited
b y J . W . de Jong, OE 9 (1962), pp. 5154.
130 David Seyfort Ruegg The Literatue of the Madhyamaka School

dB u m a la jug p a (M adhyamakvatra) and d B u m a la jug p a i bad p a (Ra

grel) (M adhyam akvatrabh^ya). E d. L. de L a Valle Poussin, B ibliotheca
Buddhica 9, St. Petersburg, 190712. (A partial French translation from the
T ibetan, up to vi. 165, w as published b y L. de L a Valle Poussin, Le M uson 8
[1907], pp. 2 4 9 -3 1 7 ; 11 [1910], pp. 2 7 1 -3 5 8 ; 12 [1911], pp. 2 3 5 -3 2 8 ).

Dpamkarar j fina
B y a chub lam gyi sgron m a (Bodhipathapradpa). Ed. J . van den Broeck, Le
flam beau sur le chem in de l veil, P ublications de l In stitu t B elge des H au tes
tu d es Bouddhiques, Srie tud es e t te x te s no. 5, Brussels, 1976 (with a French
translation) ; H . Eim er, Bodhipathapradpa, A siatische Forschungen 59, W ies
baden, 1978 (with a German translation). (English translation b y A. W aym an,
Calming th e m ind and discerning th e real, N ew Y ork, 1978, pp. 914).

N agarjuna
Ga las jigs m ed (*A kutobhay). E d . M. W alleser, D ie tibetische Version von
N grjunas K om m entar A kutobhay zur M adhyamakakrik. N ach der P e
kinger A usgabe des Tanjur herausgegeben, Materialien zur K unde des B ud d h is
m us, H e ft 2, Heidelberg, 1923 (photographic reprint). (German translation b y
M. W alleser, D ie m ittlere Lehre des Nagarjuna, nach der tibetischen V ersion
bertragen, D ie buddhistische Philosophie in ihrer geschichtlichen E ntw icklung,
2. Teil, Heidelberg, 1911).
J ig rten las das par bstod pa (L oktltastava). E d. L. de La Valle Poussin, M u
son 14 (1913), pp. 714 (with French translation); P . P atel, IH Q 8 (1932),
pp. 32631.
rTen ein brel par b yu b ai sni p o i hig le ur b yas pa (Prattyasam utpda-
hrdayakriks). E d . L. de L a Valle Poussin, B ouddhism e, tud es et m atriaux;
Thorie des douze causes, Gand, 1913, pp. 1223.
bsT od pa las das par bstod pa (Stutyattastava). E d. P . P atel, IH Q 8 (1932),
pp. 7035.
T heg pa chen po ni su pa (Mahynavimik). E d. Vidhusekhara B hattacharya,
V isvabharati Studies i, Calcutta, 1931 (edition o f both T ibetan versions w ith a
Chinese version). Cf. S. Y am aguchi, E B 4 (1926), p. 56 sq. ; (1927), p. 169 sq; G.
Tucci, Minor B uddhist texts, i, R om e, 1956, p. 195 sq.
D on dam par bstod p a (Param rthastava). E d. L. de La Valle Poussin, Muson
14 (1913), pp. 1618 (with French translation) ; G. Tucci, JR A S 1932, pp. 3224.
d P e m ed par bstod pa (N iraupam yastava). E d . L. de La Valle Poussin, M uson
14 (1913), pp. 17 (with French translation) ; P . P atel, IH Q 8 (1932), pp. 31923;
G. Tucci, JR A S 1932, pp. 31220.
d B u m a ra b a i hig le ur byas pa es rab ces b y a ba (Prajn nm a M lam adhya-
makakriks). E d. L. de L a Valle Poussin, M lam adhyamakakriks de N aga r
juna, B iblioth eca B uddhica iv, St. Petersburg, 190313 (Tibetan te x t o f th e
karikas printed in th e footnotes).
r(5od p a bzlog p a i higs le ur byas pa and rod pa bzlog p a i grel pa (Vigraha-
vya vartan i and V igrahavyvartanvrtti). E d. G. Tucci, Pre-D innaga B ud dh ist
te x ts on logic from Chinese sources, G aekw ads Oriental Series vol. x lix, Baroda,
1929 (with English translation).
ib mo rnam par th a g p a (Vaidalyaprakarana). E d. Y . K ajiyam a, M iscellanea
Indologica K iotiensia 67 (1965), pp. 13455.
es rab sdo bu (Prajndanla). E d. W . L. Campbell, Calcutta, 1919 (with E nglish
tran slation ).
Sem s kyi rdo rjei bstod pa (C ittavajrastava). E d. La de L a Valle P oussin,
M uson 14 (1913), pp. 1416 (with French translation).
Modem Editions of the Tibetan versions of Indian works 131

Srid pas las das p a i gtam (Bhavasam krntiparikath) and Srid pa pho ba
(Bhavasam kranti) [and Srid pa pho b a i tik o f M aitreyantha]. Ed. N . A iya-
svam i Sastri, B h avasankrnti Stra and N grjunas Bhavasakranti astra,
A dyar Library, 1938, pp. 73103 (with restoration in Sanskrit).
bSam gyis m i khyab par bstod pa (A cin tyastava ), E d. P . P atel, IH Q 8 (1932),
pp. 6 9 4 -7 0 1 .

B uddhaplita
d B u m a ra bai grel p a B uddha p li ta (B uddhaplita-M lam adhyam akavrtti).
Chapters ixii, ed. M. W alleser, B ibliotheca B uddhica x v i, St. Petersburg,
1 9 1 3 -1 4 .

B odhibhadra
Y e es si po kun las btus pa es b ya b a i bad sbyar (Jnnasrasam uccaya-
nibandhana). Portion corresponding to verses 2028 o f ryadevas Jnnasrasa-
m uccaya, ed. K . Mimaki, L a rfutation bouddhique de la perm anence des choses,
Paris, 1976, pp. 190206 (with French translation).

B h vavivek a
d B u m a rca bai grel p a es rab sgron m a (Prajnpradpa-M lam adhyamaka-
v r tti). Chap, iii, ed. M. W alleser, B ibliotheca Indica, Calcutta, 1914. (German
translation o f Chap. i b y Y . K ajiyam a, W ZKSO 7 [1963], pp. 3762; 8 [1964],
pp. 10030. E nglish translation o f Chapter x x iv b y R . TJryzu, K inki D aigaku
k yy gakubu kenky k iy 2/2 [1971], pp. 1556).
dB u m a i si p o i hig le ur byas pa (M adhyam akahrdayakarikas). Chap. v , ed.
S. Y am aguchi, B u kk y ni okeru m u to u no tairon, K y o to , 1941; Chap. viii.
1896, ed. H . N akam ura, A L B 39 (1975), pp. 30129.
d B u m a i si p o i grel pa rtog ge bar ba (M adhyam akahrdayavrtti-Tarkajvl).
Chap. iii. 1136, w ith th e corresponding M adhyam akahrdayakarikas, ed.
S. Iida, R eason and em ptiness, a Stu d y in logic and m ysticism , Tokyo, 1980.
Chap. viii. 116, w ith th e corresponding M adhyam akahrdayakarikas, ed. H . N a
kamura, I I J 2 (1958), pp. 18190.
sD e pa th a dad par ()byed pa da m a m par bad pa (N ikayabhedavibhaga-
vyk h y n a, corresponding to part o f Chapter iv o f the Tarkajvl). E d. E . Tera-
m oto and T. H iram atsu, Kanz-taish-ibushrinron, K y o to , 1935. (Partial E n g
lish translation b y W . W . R ockhill, Life of the Buddha, London, 1884, pp. 18296.
German translation b y M. W alleser, D ie Sekten des alten Buddhism us, H eid el
berg, 1927, pp. 7893. French translation b y A. Bareau, J A 1956, pp. 16791).

D e kho n a nid bsdus pa, rJes su dpag p a brtag pa (T attvasam graha, A num na-
parik^). E d., together w ith the Sanskrit tex t, b y A. K un st, Problem e der b u d
dhistischen Logik in der D arstellung des Tattvasam graha, P olska A kadem ia
Um iejtnoci, Prace kom isji orientalistycznej Nr. 33, K rakow , 1939.

B ya chub sem s d p a i spyod pa la jug pa (B odhi[sattva]caryvatra). Ed.
F. W eller, ber den Quellenbezug eines mongolischen T anjurtextes, A bhandlun
gen der Schsischen A kadem ie der W issenschaften zu Leipzig, Phil.-hist. K lasse
45/2, Berlin, 1950 (based on the editions of sN ar th a and sD e dge) ; Vidhushe-
khara B hattacharya, B odhicaryvatra, B ibliotheca Indica, Calcutta, 1960.

A B OKI Annals o f th e Bhandarkar Oriental Research In stitu te

ALB A dyar Library B ulletin
AM Asia Major
AO A cta Orientalia
BCA Bodhicaryvatra b y ntideva
BEFEO B ulletin de l cole franaise d E xtrm e-O rient
BSO(A)S B ulletin o f th e School of Oriental (and African) Studies
C Catuhataka b y Aryadeva
EB E astern B uddhist
GOS Gaekwads Oriental Series, Baroda
H JA S H arvard Journal o f A siatic Studies
IA Indian A ntiquary
IB K Indogaku B ukkygaku K enky (Journal o f Indian and B uddhist
IH Q Indian H istorical Quarterly
IIJ Indo-Iranian Journal
JA Journal asiatique
JAOS Journal of th e American Oriental Society
JA SB Journal o f the Asiatic Society of Bengal
JG IS Journal o f the Greater India Society
J IP Journal of Indian Philosophy
JO IB Journal o f the Oriental In stitu te, Baroda
JPASB Journal and Proceedings o f the A siatic Society o f B engal
JR A S Journal o f the R oyal Asiatic Society of Great Britain and Ireland
MA M adhyam akvatra by Candrakirti
MCB Melanges chinois et bouddhiques
MHK Madhyamakahj-dayakriks by B h vavivek a
MMK M lam adhyamakakriks b y Ngrjuna
OE Oriens E xtrem us
OLZ Orientalistische Literaturzeitung
PEW P hilosophy E ast and W est
PP Prasannapad Madhyamakavj-ttih b y Candrakirti
RO R ocznik Orientalistyczny
RSO R iv ista degli Studi Orientali
TP T oung Pao
W V igrahavyvartan b y Ngrjuna
WZKS(O) W iener Zeitschrift fr die K unde Sd- (und Ost)asiens
ZDMG Zeitschrift der D eutschen Morgenlndischen Gesellschaft
Z II Zeitschrift fr Indologie und Iranistik

I. Titles

The following index lists the titles of Sanskrit texts mentioned in this
volume. I t is to be noted, however, th a t in a number of cases these are merely
editors titles, or titles (re)constructed by the different editors of the Tibetan
bsTan-'gyur and its catalogues, and th at the original Sanskrit titles of several
such works cannot therefore be said to be known with certainty. Particularly
doubtful forms are preceded by an asterisk. But the titles listed here can in
practice serve to identify the texts in question. After the titles the names of
their (putative) authors are indicated in parentheses. As already observed, the
authorship of a number of works is uncertain, the attributions indicated here
being taken from the colophons of the available versions (Sanskrit, Chinese
and Tibetan) or, when the colophons fail us, from the catalogues.
The abbreviation P refers to the Peking edition of the Tibetan bK a -gyur
and bsTan -'gyur reprinted by the Tibetan Tripitaka Research Institute under
the direction of D. T. Suzuki (Tky-Kyto, 1958), this being the edition to
which reference is usually made in this book. The numbers following P are the
serial numbers of the texts in question, P 52245480 corresponding to the
dBu ma (Madhyamaka) section in the Peking bsTan -gyur (mDo -grel, volumes
xviixxxiii). D refers to the sDe-dge edition of which the d B u -ma section,
with the serial numbers D 38243980, has been reprinted in The sDe dge
Tibetan Tripitaka bsTan bgyur (Deruge-ban Chibetto Daizky, Ronsho-bu,
Tky Daigaku Bugaku-bu shoz) edited by K. Hayashima, J. Takasaki,
Z. Yamaguchi, and Y. Ejima, and published for the Faculty of Letters, Tokyo
University, by the Sekai Seiten Kank Kykai Co. Ltd. (Tokyo, 19771979).
T refers to the Taish Shinsh Daizky with its serial numbers; its index
(2nd edition), compiled by P. Demiville, H. D urt and A. Seidel, is published
as a Fascicule annexe du Hbgirin under the title Rpertoire du canon
bouddhique sino-japonais (ParisTky, 1978).

Abhisam aylam kra (ascribed to M aitreya[ntha] ; cf. P 5184) 28, 87, 101, 110, 111
A bhisam aylam kra-Bhagavat-Prajhpram itopadeastravrtti Prajhpradp-
val (Buddharijnna, P 5198) 117
A bhisam aylam krakrik-Prajhpram itopadeastratik Prasphu^apad
(Dharmamitra, P 5194) 102103
A bhisam aylam krakrikvrttika (B hadanta Vim uktisena, P 5186) 101
A bhisam aylam krlok P rajhpram itvykhy (Haribhadra; cf. P 5189) 93,
1 0 1 - 1 0 2 , 1 0 5n
A bhisam aylam kra-nama-Prajhaparamitopadesasastravj-tti (Sphutrth) (H ari
bhadra, P 5191) 102, 103
134 Indices

Abhisamaylamkra-nma-Prajnpramitopadeastravj*tti Dur(ava)bodhlok-
n m atk (Dharmaklrti of Suvarnadvpa, P 5192) 110, 111, 118
A bhisam aylam kravrtti (Arya V imuktisena, P 5185) 101
A bhisam aylam kravfttipindrtha (Prajnkaramati, P 5193) 116
A cin tyastava (Ngrjuna, P 2019) 31, 49
Akaraataka (ascribed to N grjuna in the b sT an -gyur, P 5234; but perhaps
rather b y Aryadeva, as in T 1572) 30, 53, 112, 114
Akaraatakavrtti (ascribed to Ngrjuna in the bsT an-gyur, P 5235; but perhaps
rather b y Aryadeva) 53, 114
A kayamatinirdeastra (P 842) 7, 99
*A kutobhay. See M lam adhyam akavrtti *A kutobhay
lokaml (K am ala/K am balapda, P 5866) 106
Alokam ltk (A svabhva, P 5869) 106n
A nantam ukhanirhradhrantk (Arya0) (Jnnagarbha, P 3520) 104n
Atkanakath (Avaghoa, P 5423, 5667) 120

Bhyrthasiddhi (ubhagupta, P 5742) 8 0n

Bhvankram a I, I I and I I I (Kamalala, P 53105312) 94, 9699
B hvan yogam rga or *Bhvanyogapatha (Jnnagarbha, P 5305, 5452). See
*Y ogabhvanm rga/ *Yogabhvanpatha
B h van yo gva tra (Kamalala, P 5313, 5451; cf. D 3918, 4537; read *Yoga-
bhvanvatra ?) 99
Bhavasamkrnti(sfcra) (Ngrjuna, P 5240; cf. 5472) 2829, 49
B havasam krntitk (Maitreyantha, P 5241) 29, 49
Bodhicaryvatra (ntideva; cf. P 5272) 8283, 85n, 93, HOsq, 116117, 119
B odhicaryvatrattparyapanjik Y iseadyotan (Vibhticandra, P 5282) 84, 117
B odhicaryvatrapahjik (Prajnkaramati; cf. P 5273) 31, 3 9n, 8 3 n, 84, 91 n, 93,
B odhicittavivarana I (Ngrjuna [II], P 5470; cf. 2665) 104, 105n
B od hicittavivarana I I (Ngrjuna [II], P 2666) 114
Bodhim rgadlpapanjik (Dpamkararjnna ?, P 5344) 29n, 49n , 53, 67n, 112
Bodhipathapradpa (Dpamkararijnna, P 5343, 5378) 111112
Bodhisam bhrastra (Ngrjuna, T 1660) 29, 101 n
B odhisattvabhm i (ascribed to A sanga; cf. P 5538) 93
B odhisattvabhm ivjtti (Gunaprabha, P 5545) 111
Bodhisattvacary[sam graha]pradparatnam l (Dhrmika Subhtighoa, P 5332)
*B odhisattvacarystrkrtvavda (Dpamkarairjfina, P 5342) 85
B odhisattvacaryvatrabhya (Dpamkararjnna, P 5872) 85
Bodhisattvaearyvatraduravabodhanirnaya (Kj-napda, P 5276) 107
B odhisattvacaryvatrapanjik (Vairocanarakita, P 5277) 116
Bodhisattvacaryvatrapindrtha (Dharmapla o f Suvarnadvpa, P 5281) 85,110,
*Bodhisattvacaryvatrasam skra ( *K alynadeva or *Subhadeva ?, P 5275) 116
Bodhisattvacaryvatra*sattrim atpindrtha (Dharmapla o f Suvarnadvpa ?, P
5280) 8 4 ,1 1 0 ,1 1 1
B odhisattvasam varavidhi (Bodhibhadra, P 5362, 5404) 109n
Bodhisattvasam varavim saka (Candragomin, P 5582) 93, 109n
B odhisattvasam varavim sakapahjik (Bodhibhadra, P 5584) 109

Catuhataka (Aryadeva, P 5246) 33n, 39n, 5053, 78n, 85n, 101, 114
Catuhsatakatk (B odhisattvayogacary0 or yogcra ?) (Candrakrti, P 5266)
5 1 -5 2 , 81, 85n, 114
Titles 135

C atuhstava 3132, 120121

C atuhstavasam asartha (Am ftakara) 31
Caturviparyaya[parihara]katha (Matj-ceta, P 5425, 5669) 120
Chang-chen-lun (Ta cheng (K aratalaratna, b y B havavivek a, T 1578) 63, 66,
Cig car 'jug pa m a m par m i rtog pa'i bsgom don (Vimalamitra, P 5306) 94n, 107
C ittavajrastava (N&garjuna, P 2013) 31
Cittavi^uddhiprakarana (A ryadeva [II]; cf. P 5028 ascribed to Indrabhuti and
P 2669 ascribed to Aryadeva) 106

Da^abhum ikasutra (cf. P 761/31) 30, 32, 7274

*Da4abhumikavibhaa(3astra) (Nagarjuna?, T 1521) 27n, 29, 32, 101 n
D ur(ava)bodhaloka Tika (Dharmakirti o f Suvarnadvlpa). See A bhisam ayalam -
karanma-Prajnaparamitopade6aiastravj-tti D ur(ava)bodhaloka-nam a-Tika
Da4aku6alakarmapathanirde3a (Dharmika Subhutigho?a, P 5417, 5676) 120
D aiakuialakarm apathanirde^a (A6vaghoa, P 5416, 5678) 120
Dharmadhatudar^anagiti (Dipam karairljnana, P 5388, 3153) 113
D harm adhatustava (Nagarjuna [I or I I ?], P 2010) 32, 113, 121 n
D hyanagaddharm avyavasthana (Avadhutipada, P 5321, 5446) 117
D hyanaaddharm avyavasthanavj-tti (Dana^ila, P 5322, 5447) 117
D vad asam u k h aiastra or *D vada6anikayaiastra (Shih-erh-men-lun b y Nagarjuna,
T 1568) 2, 2 7 - 2 8

E kailoka^astra (Nagarjuna, T 1573) 30

G uhyasam ajatantra (cf. P 81) 102, 105

G uhyasam aja-Tantratika (Nagarjuna [II], P 2648) 105

Haracarita (Bana) 5n, 27n

H astavalaprakarana (ascribed to A ryadeva in the bsTan**gyur, P 5244, 5248; but
to D ignaga in th e Chinese versions, T 1620/1) 5253, 112
H astavalaprak aran avftti (ascribed to Aryadeva in the bsT an-'gyur, P 5245, 5249;
but perhaps rather b y D ignaga) 53, 117n

I^varakartj-tvanirakpti (ascribed to Nagarjunapada, P 5905) 30n

Jantupo?anabindu (N itisastra0) (N agarjuna?, P 5822) 27

Jatakam ala (Arya-Sura; cf. P 5650) 119
Jinam argavatara (Buddha^rljnana, P 5359) 117
Jnanasarasam uccaya (A ryadeva [II], P 5251) 54, lOOn, 106, 112, 118
Jhanasarasam uccayanibandhana (Bodhibhadra, P 5252) 58, 59n, 109

K aliyugaparikatha (Matj-ceta, P 5426, 5670) 120

K aratalaratna. See Chang-chen-lun
Ka^yapaparivarta (cf. P 760/43) 07
K a tyaya n avav ad a 6n
K a ya tra y a sto tra /sta v a (Nagarjuna, P 2015; cf. 2016) 56
K ayatrayavataram u k h a (Nagamitra, P 5290) 56n

LankavatarELSutra (cf. P 775) 7, 20n, 54, 56, 7273, 90, 96, 97, 122sq
Lilavai (Lilavati, b y K utuhala/K ouhala) 27 n
Lok&titastava (Nagarjuna, P 2012) 31, 49
136 Indices

M adhyam akabhram aghta (Aryadeva, P 5250) 54, 112

M adhyamakahrdayakriks (B hvaviveka, P 5255) 6263, 65, 86n, 111
M adhyam akahrdayavytti Tarkajvl (B hvavivek a/B havya, P 5256) 6 2 sq, 64, 65,
66, 86n, 111*
M adhyam aklam kra (Sntarakita, P 5284) 9092
M adhyamaklamkrapanjik (Kamalala, P 5286) 93
M adhyam aklam kravrtti (ntarakita, P 5285) 20n, 9092, 99
M adhyamaklamkravrtti-M adhyam apratipadsiddhi (R atnkarainti, P 5573)
*M adhyam(ak)lamkropadea (Ratnkaranti, P 5586) 122124
M adhyam akloka (Kamalala, P 5287) 9496
M adhyam akanayasrasam sa (Vidykaraprabha, P 5293) 99
M adhyam akaprajnvatara (Candrakirti [III], P 5264) 81 (in D 3863 the title is
given as M adhyamakvatra-prajn)
M adhyam akaprattyasam utpda (Krnapda, P 5257) 107
*M adhyamakaratnapradpa (B havya, P 5254) 66, 106n
M adhyam akrthasam graha (B hvaviveka, P 5258) 64, 111
M adhyam akasstra 1
M adhyam akastrastuti (Candrakirti) 8, 29, 31, 55
M adhyam akvatra (Candrakirti, P 5261/2) 7, 34, 7174, 75, 81, 82, 85n, 101 n,
113, 114
M adhyam akvatrabhsya (Candrakirti, P 5263) 71sq
M adhyam akvatra-prajn (Candrakirti [III], D 3863). See M adhyam akapra-
M adhyam akvatratk (Jaynanda, P 5271) 74, 113114, 118
M adhyamakopadea (Dipamkararjnna, P 6324, 5326, 5381) 113
M adhyamakopadeavftti (Prajnmok^a, P 5327) 113
Madhyamaatka (Maitrpda; cf. P 3074) 107
M adhyntavibhga (ascribed to M aitreyafntha] ; cf. P 5522) 45n , 122
M ahm yrividyrajn 104n
*Mahrja-Kani()kalekha (M tfceta, P 5411, 5684) 120
M ahstrasam uccaya (Dipamkarasrjnna, P 5358) 113, 114
M ahynalakanasam uccaya (Buddhajhnapda, P 5301) 102
M ahynasam grahabhya (Vasubandhu, P 5551) 111
M ahynastra 67, 81
M ahynavim ik (N grjuna?, P 5233 ^ 5465) 2930, 112, 114
M andalavidhi (K am balapda, P 5443) 106
Manjurmlakalpa 104n, 120
M armakaumud (Arya-Atashasrikprajnpram itvftti Marmakaumudi) (Abha-
ykaragupta, P 5202) 103, 115
M lam adhyam akakriks (Prajnnma0) (Ngrjuna; cf. P 5224) l n , 6, 8sq,
3 4 - 3 6 , 60, 62, 74, 114
M lam adhyam akavrtti *A kutobhay (ascribed to Ngrjuna, P 5229) 12sq, 4748,
49n, 61, 62, 112
M lam adhyam akavrtti *Buddhaplit, (Buddhaplita, P 5242) 12sq, 60sq, 85, 112
M lam adhyam akavrtti Prajnpradipa (B h vaviveka, P 5253) 12sq, 49, 6 2 sq, 65,
8 5 -8 6 , 112
M lam adhyam akavrtti Prasannapad (Candrakirti; cf. P 5260) 7, 9sq, 34, 3 9 n,
49, 51, 71 sq, 7481, 85n, 112, 114
Munim atlam kra (A bhaykaragupta, P 5299) 114115
N avaslok (K am balapda; cf. P 5210, 5906) 106
N ikyabh edavib han gavyk hyna (B h vaviveka/B h avya, P 5640) 63
Nirkrakriks (Nandar, P 5294) 100
Titles 137

N iraupam yastava (Nirupam a0) (Ngrjuna; cf. P 2011) 31 sq, 34, 35, 49
Ntistra. See Jantupoanabindu and Prajndanda

Pacakram a (Ngrjuna [II], P 2667) 105, 106

Pancaskandhaprakarana (Candrakirti, P 5267) 81, 111
Param rthabodhicittabhvankram avarnasam graha (Avaghoa, P 5308, 5431)
59n, 120, 121n
Param rthastava (N grjuna; cf. P 2014) 31 sq, 34, 49
Pram itsam sa (Arya-ra; cf. P 5340) 119
P radpoddyotana (Candrakirti [II], P 2650) 105, 106n
Prajndanda (Ntistra0) (Ngrjuna, P 5821) 27, 119
Prajn-nm a-M lam adhyam akakriks. See M lam adhyamakakriks
Prajftpramit(stras) 67, 30n, 3233, 4 5 n, 4 7 n, 101 sq
Prajhpram itstotra (Rhulabhadra) 33, 54, 101, 105n
Prajpramitopadea (Arya0) (K ambalapda, P 5314, 3466) 106
Prajpramitopadea (Ratnkaranti, P 5579) 123124
Prajpradipa (B hvaviveka). See M lam adhyam akavrtti Prajnpradpa
Prajhpradpatk (A valokitavrata, P 5259) 67, 8 6 n
Prajnataka (Ngrjuna, P 5414, 5820) 27
Pram navrttika (Dharmakrti) 91, 93
Prasannapad M adhyam akavpttih (Candrakirti). See Mlamadhyamakavjrtti Pra-
Prasphu^apad. See A bhisam aylam krakrik-Prajnpram itopadeastratlk
P rattyasam utpdahfdayakriks (Ngrjuna; cf. P 5236, 5467) 28, 119
P rattyasam u tp dahfd ayavykh yna (Ngrjuna, P 5237, 5468) 28, 114

R jatarangin (Kalhana) 5 n
R atnagotravibhga (ascribed to M aitreyafntha] ; cf. P 5525) 28, 32, 56, 84n, 95n,
R atn agotravibhga-vyk hy (Asaga ? ; cf. P 5526) 55
R atnagunasam cayagthpajik (Bhagavad0) (Haribhadra, P 5190) 102 (see also
Sam cay agthpanj ik )
Ratnakarandodghfa-nm a-M adhyam akopadea (Dipamkararijna, P 5325) 113
R atn val (Ngrjuna; cf. P 5658) 6n , 8, 2326, 49, 68n, 83, 101 n
Ratnvalt/k (Ajitamitra, P 5659) 24 n, 49
R im gyis *jug p a i (b)sgom don (Vimalamitra, P 5334) 94n , 107

Saddharmapundarka (cf. P 781) 7, 5556

Saddharm asm rtyupasthna-Sadgatikriks (Dhrmika Subhtighoa ; cf. P 5415,
5679) 120
listam bakakriks (Arya0) (Ngrjuna, P 5466, 5485) 29
listam baka-nm a-m ahynastratiik (Arya) (Ngrjuna, P 5486) 29
listam bakatk (Arya0) (Kamalala, P 5502) 99
Sam cayagthpajik (Buddharjnna [ = Buddhajnna-pda], P 5196) 102,
*Sarnkiptanndrtivibhga ( *Madhyamaka-simha [dBu m ai sen ge]/Trarmi-
tr, P 5295) 116
S am stuti 8, 31
Sam varavim sakavrtti (ntarak^ita, P 5583) 93, 109n (cf. B odhisattvasam vara-
Sam vrtibodhicittabhvanopadeavam asam graha (Avaghoa, P 5307, 5432) 120
Saptaatikprajnpramit^k (Arya0) (Kamalala, P 5215) 99
138 Indices

Sarvadharm abhavasiddhi (Kamala&la, P 5289; read Sarvadharm asvabhavasid-

dhi, or Sarvadharm anihsvabhavasiddhi as in D 3889) 99
*SarvayanalokaviSe$abha$ya (Subhutigho$a, P 5303; the spelling o f th e title
adopted here follow s D 3907) 120
*&ata(ka)3astra (Aryadeva, T 1569: Pai/Po-lun) 2, 50 sq.
Satapahca^atka (M atfceta; cf. P 2038) 119
S atyadvayavatara (Diparjikarairijnana, P 5298, 5380) 113, 116
Satyadvayavatarapanjika (Dharmakaramati) 116
Satyadvayavibhangakarikas (Jhanagarbha, D 3881; not in P) 6871, 86n, 93
Satyadvayavibhangapahjika (Santarakita, P 5283) 6869, 86n, 93
Satyad vayavibh ahgavrtti (Jhanagarbha, D 3882; not in P) 6871
Saundarananda (Aivagho?a) 28, 119120
Shih-erh-men-lun (T 1568: D vada^am ukha- or D vada^an ikayaiastra b y N a gar
juna) 2, 2728
Sik^akusumamanjarl (Vairocanarak^ita, P 5339) 116
Sik^asamuccaya (Santideva; cf. P 5335/6) 8384, 85, 93
Sik^asamuccayabhisamaya (Dharmapala o f Suvarnadvlpa, P 5338, 5464) 110, H I
Skhalitapram athanayuktihetusiddhi (Aryadeva, P 5247; probably Skhalitapra-
mardana0 should be read w ith D 3847) 54
Sphutartha (Haribhadra). See Abhisamayalamkara-nam a-Prajhaparamitopade^a-
S tav a 8, 3132
S tu tyatltastava (Nagarjuna, P 2020) 31
Subhaitaratnakarandakakatha (Arya-Sura, P 5424, 5668) 119
Sugatam atavibhangakarikas (Jitari/Jetari, P 5296, 5461, 5867) 100
Sugatam atavibhahgabhaya (Jitari/Jetari, P 5868) 100
Suhrllekha (Nagarjuna, P 5409, 5682) 2627, 32, 49, 120
Suhrllekhat>Ika (V yaktapada) (Mahamati, P 5690) 27 n, 49
Sunyatasaptati (Nagarjuna, P 5227) 8, 2021, 49
Sunyatasaptativivj-ti (Parahita, P 5269) 116
Sunyatasaptativj-tti (Nagarjuna, P 5231) 20, 21 n
S u n yatasap tativftti (Candrakirti, P 5268) 21 n, 49, 81
Sutralam karadi^lokadvayavyakhyana (Parahitabhadra, P 5532) 116
Sutrasam uccaya (Nagarjuna, P 5330) 29, 84, 124
Sutrasam uccayabhasya-ratnalokalam kara (R atnakaraianti, P 5331) 124
Sutrasam uccayasam cayartha (Dlpamkara^rljnana, P 5333) 113 (see also Maha-
Svabhavatrayapravesasiddhi (Nagarjuna?, P 5243; D 3843 has Svabhavatraya-
prave&asadhana) 30
Ta-cliih-tu-lun ([M aha-]Prajhaparam itopadeia ascribed to N agarjuna, T 1509)
7, 29n, 3 2 -3 3 , 51, 55, 101, 118n
Tarkabhaa (Mok^akaragupta; cf. P 5762) 100n, 118, 123n
Tarkajvala. See M adhyamakahfdayavftfci Tarkajvala
Tarkamudgara (Jayananda, P 5270) 113n, 114
T athagatagarbhasutra (P 924) 56
T atparyapanjika Viseadyotani. See B odhicaryavataratatparyapanjika
T attvajnanam rtavatara (B havavivek a/B havya) 63
T attvalok a (Kamalasila, P 5288) 99
T a ttv a p ra k ^ a (A dvayavajra; cf. P 3086) 107
T attvaratn avali (A dvayavajra; cf. P 3085) 58, 107, 123n
T attvasam graha (&antarakita; cf. P 5764) 8990, 93
Tattvasam grahapanjika (K am alaiila; cf. P 5765) 58, 8990, 93
T attvasiddhi (Santarak?ita, P 4531) 8 2 n, 93
Titles 139

Tattvvatravj*tti (rgupta, P 5292) 67, 86n, 91 n

T rik yastotra/stava (Ngrjuna, P 2015) 56
Triarana[gamana]saptati (Candraklrti [II], P 5366, 5478) 105
Triskandhasdhana (K|*?napda, P 5509) 107 n
Trisvabhvanirdea (Vasubandha, P 5559) 30
T riynavyavasthna (Ratnkaranti, P 4535) 123

V aidalya-prakarana (Ngrjuna, P 5230) 8, 21, 114

V aid alya-stra (Ngrjuna, P 5226) 8, 21, 114
V airocanbhisam bodhi (P 126) 98
Vajracchedikprajpram ittik (Arya0) (Kamalasila, P 5216) 99
Varnrhavarnastotra (Mtfce^a; cf. P 2029) 120
Vigrahavyvartani(kriks) (Ngrjuna; cf. P 5228) 8, 2123, 60, 80, 114
V igrahavyvartanvrtti (Ngrjuna; cf. P 5232) 21 sq
Vimalakrtinirdeastra (P 843) 7, 70, 99
V ipacitrth V dan yyatik (ntarak^ita; cf. P 5725, 5738) 88, 93
V yavahrasid dh i (ascribed to Ngrjuna) 8, 26

Y ogabhvan m rga or *Y ogabh vanp atha (Jnagarbha, P 5305, 5452 [ = D

3909, 4538]) 69n, 71
Y oga bh va n vatra (Kamalala, D 3918; th e title is B hvanyogvatra in
P 5313) 99
Yogcrabhm i (ascribed to A saga; cf. P 5536sq) 52
Yukti?a$tik(kriks) (Ngrjuna, P 5225) 8, 1920, 49, 90n, 99, 123
Yuktiat>ikvrtti (Candrakirti, P 5265) 49, 81, 85, 117n

I I . N a m es

A bhaykaragupta 95n, 103, 114115 79, 83, 8 5 - 8 6 , 8 7 - 8 9 , 90, 109, 112,

A d vayavajra 58, 107, 116, 123n, 125 113, 127, 131 (B hviveka, Bhvin)
(Maitripda, A vadhtipda) B h vav iv ek a /B h a v y a (II) 6667, 106,
Ajitam itra 2 4 n, 49 116
Ak^ayamati 82n B h avyakirti 106n
A inoghavajra 57 B havyarja 66 n
A m ftkara 31, 49, 125 Bhusuku 84, 106
A ryadeva (I) 1 - 2 , 30, 33, 47, 48, 5 0 -5 4 , B odhibhadra 58, 59n, 86, 109, 111, 131
56, 62, 72, 78n, 80, 81, 83, 85n, 87, *Brom *sto n rGyal ba*i -byun gnas 111
101, 112, 119n, 120, 125, 129 Buddhajna-pda 102, 115
A ryadeva (II) lOOn, 105106, 109 B uddhaplita 9n, 47, 48n , 49, 5862,
Arya-ra (ra) 59 n, 119120 71, 74, 78, 80n, 85, 112, 131
A saga 49, 52, 53n, 55, 122, 124 Buddharijna 102 n, 117
A svab h va 106107 n d B u -m a h -s e n -g e . See *Madhyamaka-
Avagho$a 28, 59 n, 119121 simha
A tia. See Dlpamkararijna dB us*pa B lo -g sa l B y a -c h u b -y e - e s
A vadhtipda 100, 117n (cf. Maitri- 80n
pda, A dvayavajra) B u ston R in chen grub V In , 67n, 6 8 n,
A valokitavrata 49 n, 62, 67, 8 6 n, 112 80n, 84, 85n
B y a h -ch u b od (lH a-bun) 111
B n a 5n , 2 7 n
B hartfhari 6n Candragomin 93, 109n, 120
B h v a v iv ek a /B h a v y a (I) 9n , 37, 47, Candrakirti (I) 7 - 8 , 9, 21 n, 26, 34, 37,
48n , 59sq , 6167, 7 1 - 7 2 , 73n, 74r- 47, 4 8n , 49, 5 0 - 5 2 , 55, 59n, 60, 66,
140 Indices

67, 7 1 - 8 1 , 82, 83, 85, 110, 112, 113, H oshang (Hva*sa) M ahyna 94, 99
114, 122, 126, 129 H san-tsang 5n , 51, 67
Candraklrti (II) 81, 105, 115, 122
Candraklrti (III) 81, 116 I-ching 5n, 2 7 n, 104n
ICan skya R ol p a i rdo rje Indrabhti 106n
(Ye ses bstan p a i sgron me) I X n ,
69n, 93n *Jam dbyas bzad pa
Ching-m u 28, 48 a g -d b a -b rco n -grus I X n , 89n,
Chi-tsang 54 102n
h u l-k h rim sT g yal-b a (N ag-ho) 85 J aynand a 74, 85n, 113114, 118
86, 107n, 111 Jitri/Jetri (I) 100
Cog t o K lu i-rg y a l-m ch a n 61, 69n, Jitri/Jetri (II) 100, 107n, 111, 116
8 5 -8 6 Jnagarbha (I) 6771, 8 6 n, 87, 89,
(5o kha pa B io bza grags pa V IIIn , 93, 104n( ?), 116
20n, 59n , 69n, 111, 114n, 124n Jnagarbha (II/III) 61(? ), 69n, 71,
85n, 86n, 10 4 n (? ), 116
Jnapda. See B uddhajna-pda
Dnala 117
Jnaprabha 67
b D e*byed 26 n
Jnarimitra 20n, 110, 116, 122n
dD e spyod 24n, 27n
Devaarm an 49, 62, 67 n, 112
K alhana 5n
D evendrabuddhi 69
K alyn adeva ( ?) 84, 116
D harm adsa 53, 81
Kam alarakita 84, 110
D harm akara(datta) 102
Kam alala 20n, 58, 89, 9399, 103,
D harm karam ati 116117
106, 107, 112, 123n, 125, 129
Dharm akirti 66, 67, 68n, 88, 91, 93,
K am bala-pda (or K am ala0) 106, 126,
100, 1 0 9 -1 1 0 , 118
D harm am itra (com m entator on Vinaya)
K anakavarm an 114
Kanika 5n, 120, 121
D harm am itra (com m entator on Praj-
m K has-gru b d G e-legs*dpal-b za I X n ,
pramit) 95 n, 102
48n, 5 9n, 67n, 68n
D harm apla (of N land) 5153, 61,
Khro phu B yam s pai dpal 117
67, 69, 81
K h u m D o 8de *bar 114
D harm apla (of Suvarnadvipa,
K fna-pada 84, 107, 111, 116
gS er-g li -pa ) 8485, 110, 111, 118
K nm rajiva 2, 5n, 28, 29 n, 32, 5051,
Dharmarak^a 2 9 n
D harm ottara 102
K n n -d g a *grags 114
D ignga 52, 53, 60, 61, 66, 67, 80, 87
K u t hala (K ohala) 27 n
Dipamkararjna(AtiSa) 848 6 ,100n,
105n, 107n, 1 0 9 -1 1 3 , 116, 130
M adhyam aka-sim ha
D ivkara 67
(dB u*m ai *se*ge) 116
M ahm ati 2 7n, 49
F a-tsang 67
Mai trey a (nt ha) 101, 102, 122
M aitreyantha (Pandita) 29, 49
GainJapada V III M aitripada 107, 116, 122n (see also
G os-k h u g -p a lH a(s) -bas 81 A dvayavajra, A vadhtipda)
Grags-*byor-Ses-rab 114 Mtfce^a 119120
G unadatta 4 9 n, 67, 112 Mok^akaragupta 100n, 118, 123n
Gunam ati 49, 62, 112
Gunari 49, 112 Nglapda (Nrop) 56n, 107n, 116,
H aribhadra 59n , 93, 101102, 103, N aga 5657, 104n, 105
105n, 115 N gabodhi 57, 105
H arivarm an 50 n N gh vaya 56
Names 141

N agam itra 56 n Satakarni (Satakani) 5 n

N agarjuna (I) ls q , 4sq , 50, 54sq, 56 S atavahana (Sata0) 6n , 24, 2 7 n
57, 60, 75, 77, 78sq, 81, 82n, 8384, Seng-jui 32
87, 89, 98, 100, 101, 102, 104, 112, gS er-g lih -p a 110
115, 1 1 9 -1 2 1 , 122, 124, 126, 130 Silabhadra 67
N agarjuna (II) (Arya-) 57, 104sq Srlgupta 6768, 86n, 87, 89, 100, 112n
rN og B i o ld an-ges-rab 59n , 85n Srihar^a V III
Sthiram ati 4 5 n, 49, 61, 62, 69, 112
P a -ch a b N i-m a grags 85n , 113, 114 Subhadeva (? ) 116
P aindapatika llO n Subhagupta 80 n
dPal*br6egs 59, 85 n Suk^majana 114
d P al *brcegs raki ta 99 Sura. See Arya-Sura
Parahita(bhadra) 116 tSuvarriadvIpa, (epithet o f Dharma-
Param&rtha 49 n, 56 pala/Dharm aklrti). See g S er-g lin -p a
Pirigala 4 8 n, 50
Pin-lo-chieh 48, 50 T aranatha V I I I n , 56, 5 9 n, 84, 119
Prabhakaramitra 85 Tara^rimitra 116n
Prajnakaragupta 118 Tathagatabhadra 56
Prajnakaram ati 31, 82, 8 3 n, 84, 85 n, Tilakakalasa 114
.93, 116, 122n, 127
Prajnamoka 113 U dayan a/U dayin ( ?) 27 n
P ulu m ayi 5n
Vairocanabhadra 102
Rahulabhadra (I) 4, 33, 49, 5456, 101, Vairocanarakita 84, 116
105n, 128 Vajrabodhi 57 n, 105
Rahulabhadra (II) 54, 105 Vasi^hiputra 5n
Ratnakara^&nti 20n, lOOn, 107n, 110, V asu 51, 53n
111, 116, 1 2 2 -1 2 4 V asubandhu 30, 51 n, 5 3 n, 68n, 87,
R atnakirti 110, 118n 99n, 121n
R in -ch en -b z a n -p o 85n , 111, 117n V ibhuticandra 84, 117
V idyakaraprabha 99
Sakya^ribhadra 117 V im alam itra 94n, 102, 107
Santarakita 20n, 47, 58, 59, 66n, 69, V im uktisena (Arya) 53n, 87, 9 2 n, 101,
71, 86n, 8 7 - 9 4 , 9 9 -1 0 0 , 102, 106, 102n, 103, 115
107, 109, 111, 112, 123n, 128, 131 V im uktisena (Bhadanta) 101, 103, 115
Santideva 8285, 93, 106, 118, 119,
128, 131 Y e Ses ^od (lH a bla ma) 111
Saraha 39n, 54, lOOn, 105, 128 Y e-sesT g y a l* m 6 h a n V III n
*Saramati 55, 56, 57 n Y e -s e s 's d e 59, 69n, 85n , 86n, 107

m . Sanskrit key-words
atiprasanga 60 anupalam bha 19
ad vay a 25, 28, 45, 70 anubhava 82
advayajnana 97 anum ana 22, 37n, 53, 58, 61 sq, 6365,
Advaitadar^ana 89, 93 70, 79, 80, 81, 89, 90, 92
adhi?thana 82 anekasvabhava 68, 9192, 97, 100, 106,
anavasth& 13, 22 112
AnakaravSda 92. See Nirak&ravada = anta 1, 19, 42, 45, 52, 68, 92, 100
A1 ikakara vda a n yatva 15, 38
an itya 21 apavada 68, 95
anirodha 18, 76, 96 apratihana 58, 59 n
142 Indices

apratit>hitanirvna 96, 97 upalam bha 46 n

abhva 3, 15, 18, 24, 38 sq, 76, 90 upaama 18, 46 n
A bhidharm a 9, 43 n updtj* 1314, 40
abh id htavya 18, 75 upadna 1 3 -1 4 , 15, 28, 40, 45
abhidheya 34, 46 n, 74r-75 updya prajnaptih 16, 44, 74
abhisamskra 64, 65 up ya 28, 72, 82, 9 6 sq
abhisam dhi 26, 95, 96 upalam bha 12, 78n
ab hyud aya 24, 68 ubhayaprasiddha 79
amanasikra 98, 99
A m itbha 27, 32 ekatva 13, 15, 38
A m ityu s 32 ekayna 26, 31, 55, 94, 95n, 96, 103,
arthakriy 64, 68, 70, 91, 92 115
arthpatti 90 ekasvabhva 68, 9192, 97, 100, 106,
altamand&l& 24 1 12
Alikkravda, vdin 68, 92, 100, 102, ekbhva 11, 38 n
106, 110, 122123. See also Nirkra0,
A nkravda aitihya 90
aloka 75
alokasamvj-ti 7475 Aupani$adika 89, 93
a v y a k fta 25, 41 n
avidy 15, 19, 20, 28, 45 karun 82, 92, 96, 97, 98
anya 14, 34, 39, 46 sq karman 13, 15, 19, 21, 24, 28, 38, 42,
asamskj-ta 13, 18, 38, 42n , 44n , 47, 63, 45, 68, 89
9 karm apatha 24, 120
asiddha-hetu 77 kya 5 6 n, 70
asiddhdhra 77 kla 52, 90
astit 24 kles 15, 1 9 -2 0 , 21, 28, 40n , 45, 47,
asmj*ti 98 52, 68
aharnkra 24, 25, 45 klevarana 97
kana 68
kara 68, 70, 9 1 - 9 2 , 102n, 1 2 2 -1 2 3 kanika 76
ka 66 knti 82, 119
gam a 22, 61, 81, 91, 94, 96, 98
javam javbhva 46 khapu^pa 63, 76, 100, 106, 112n
atm an 14, 15, 19, 40, 5152, 66, 73, 74,
89, 93
gaganapupa 76 (cf. khapu^pa)
tm iya 15, 19, 74
gati/gam ana 1011, 38
bhidhrmika 6n, 10, 43, 47, 62, 112,
gan tav ya 38
gantj- 1011, 38
yatana 12, 13, 19, 77, 81
guna 89
rya 16, 34, 72, 75, 76
gj-hastha 26
ryasatya 16, 42, 44
gotra 32, 95, 103, 115
lam bana 9, 122123
grhaka 91, 95, 123
lam banavastu 98
grhya 91, 95, 122123
layavijna 7273
Srayaparvftti 98
catu^ko^i 18, 25, 35n, 39, 41, 46, 64,
idam pratyayat 76 76, 92, 100, 106, 109, 112, 123
indriya 12 cakravartin 26
citta 18, 25, 28, 69n, 75, 97, 123
ivara 30, 83, 89, 112 cittam tra 29, 30n, 69, 73, 90, 95, 96,
97, 98, 1 2 2 -1 2 3
udbhvita 76 cittotpda 71, 88
up am n a 22, 81, 90 cint 96
Sanskript key-worcLs 143

J aina 89, 90 (cf. Nirgrantha) Nirkrav&da, vadin 58, 100, 102, 110,
jana 26, 45, 66, 72, 74, 76, 97, 98, 123 122123. See Ankaravda, Alikak-
japaka 2223 ravda
jeyavarana 97 nirodha 17, 19, 24, 38, 41n, 44, 66, 76,
ta ttv a 16, 18, 19, 21, 24, 37n, 42sq, 46, Nirgrantha 25, 62 (cf. Jaina)
68, 7 3 - 7 4 , 76n, 97, 106, 109 nirvana 15, 16, 18, 19, 24, 38, 39, 41 n,
ta ttv a jana 62 45, 46, 54, 91, 96
ta ttv n y a tv a 38 n nitrtha 73, 94, 96
tath at 33, 34, 46, 66, 70, 96, 98, 123 nti 27
ta th gata 15, 3941, 44n , 46, 55, 92 nti 25
tathgatagarbha 3132, 34, 3 5 n, 55 neyrtha 73, 95, 96, 110
57, 73, 9 5 - 9 6 , 103, 115 naihsreyasa 24, 68
tath gatad htu 32, 55 N y y a 6n , 21 n, 37n, 89
ta th y a 4 0 n
ta th y a sa m v fti 64, 75, 83n, 113 paka 20, 25, 37n, 52, 60, 6465, 77,
t$n[m ]bhva 25, 3435, 70 78n
t|*tlya (tj-tya-ri) 11, 36, 38n, 40n, 41, pakadoa 65, 77
68, 83n , 92, 109 padrtha 21, 89
trikya 56, 70
paratantra (svabhva) 63, 69, 72, 92,
triyana 26, 31, 9 4 - 9 5 , 96 95, 122
trirpaliga 70 paraprasiddha 79
trisvabh va 30, 92, 95, 122 parabhva 5n, 9, 14, 38
paramnu 68, 91
darana 23, 23, 36n paramrtha 3, 16, 21, 25, 31, 338q,
duhkha 16, 28, 44 37n, 42sq , 59n, 61, 63sq, 65sq, 69sq,
df^tnta 63, 64, 65, 70 72sq, 74, 76sq, 83, 88, 92, 96sq, 100,
dp?ti 23, 14, 18, 34n, 36, 45, 47, 52 113, 120, 123
dravya 89 param rthatas 6365, 72, 7677
dravyasat 66, 77 parinm a 70
dharma 23, 6, 7n, 10, 16, 19, 21, 22n, parinm an 26, 8283
24, 25, 26, 28, 31, 34, 40, 42sq, 4647, pariea 37 n
6 5 -6 6 , 69n, 80, 90, 96, 98, 112, 122 paryya 64
dharm akya 26, 34, 46 sq paryudsaprati^edha 3738, 65, 79
dharm ata 28, 33, 44 n , 45n, 46, 113 papa 24
dharm adhtu 31, 33, 34, 46 sq ppadean 82
dharm anairtm ya 67, 69, 73, 74, 98 param it 26, 28, 71, 8 2 sq, 96, 97, 101,
dharm am aya-kya 31, 34 119
dharm asam keta 73 pram itnaya 104n
dharmspada 26 P ram ityna 101, 104
dhtu 1213, 19, 4 4n , 81, 97 punya 24, 26, 83, 97
dhyna 83, 97, 115, 119 pugala 13, 15, 25, 39, 40, 89, 91
dhruva 31 pudgalanairtm ya 6, 7374, 96, 98
puja 82
nnbhva 1 1 ,3 8 n p fth a k tv a 13, 3 8 n
nm am tra 28 prakptisnta 43 n
natit 24, 25 prakftisthagotra. See gotra
nihsarana 14 pranidhi 26, 82
nihsvab hva 2, 2122, 41, 4 5 n, 46, 74, prajapti 16, 24, 43n , 44n, 55, 70, 74,
91 76
n ih svab havavd a 2# 3 6n praj 24, 28, 63, 72, 82, 94, 9697
nigam ana 78 prajpram it 55, 72, 82, 96,10 1103,
n itya 21, 25, 31, 52 108, 110, 1 1 4 -1 1 5 , 119
144 Indices

pratij 22, 25, 36n, 37n, 42, 65, 78, 79 bhm i 26, 29, 33, 71 sq, 82, 97, 98, 122
pratidvandvin 37 n m aga 62
pratipaka 25, 37n , 79 m adhyam a 1
pratipatti 9698 m adhyam aka ls q , 53 n
pratibandhin 37 n m adhyam a pratipat 1, 16, 4 3 sq, 45, 47,
pratibim ba 20, 24, 68, 75, 91 63, 68, 83, 96, 98
prativdin 70 m antranaya 104n, 106
prati?edha 3 6 n, 3 7 sq, 52, 64, 78 M antrayna 104 (cf. Vajrayna)
pratiedhamtra 36n, 78 marlci 25, 70
pratisam khynirodha 66 M ahyna 6, 26, 90, 123
pratltyasam utpanna 44, 45 n, 63 m dhyam ika ls q , 58
pratityasam utpda 9 sq, 1618, 19, m aya 19, 20, 25, 63
2021, 28, 42sq, 45, 63sq, 72sq, 75n, M yopam dvayavdin 58
76, 83, 112, 123 marga 1617, 44, 83, 97, 98, 99, 111
pratyaka 22, 80, 81, 89 112
pratyaya 9sq , 28, 70, 112 m ithyjna 20, 72
pratyavekan 97 mithysam vj'ti 64, 75, 83 n, 113
p ratytm aved ya 75 Mmms 3 8 n, 63, 89
pratyekabuddha 74 mf 25, 40n
pradhna 70
prapaca 18, 34, 35, 39, 42, 46, 55, 64sq, yuk ti 19, 21, 61, 70, 9 0 - 9 1 , 94, 96, 98
70, 75, 92 yuganaddhavhin 9798
pra-pad- 55 yugapad 98, 107
pram na 22, 72, 75, 8081, 83, 89, 122 yoga 52, 72, 109
pram eya 22, 80 yogabhm i 122124
prayoga (vkya) 61, 65, 71, 7879 yogcra 1, 20n, 30n, 5253, 63, 72,
prasaga/prasajyate 11, 22, 36sq, 42, 87 sq, lOOn, 104, 106, 108, 115, 122
58, 60, 64sq, 71 sq, 78sq, 88 Y ogcra- (Svtantrika- )M adhyamaka
prasajyapratiedha 37sq, 39, 60n, 63, 2 0 n, 30, 59, 61, 66n, 6 7 - 6 9 , 80, 85,
65, 79 87100, 102, 103, 106, 109n, 110, 112,
prasthna 82 115
Prsagika 58sq, 7186, 88
ri. See tj-tlya/trtlya-ri
bahirartha. See bhyrtha rpa 20, 69n, 70, 75, 91, 97
bhyrtha 30, 59, 6 6 n, 69n, 73, 80, rpakya 26
87n, 90, 91, 95, 97, 99
buddha 16, 26, 31, 41 n, 42, 70, 72, 82, lakana 1213, 26, 38, 80, 123
9 6 - 9 7 , 115 lakya 12, 38, 80
buddhabhm i 72, 74, 97, 98 liga 63, 64, 70, 81 (cf. hetu)
bodhi 25, 29, 96 loka 25, 28, 72, 75
bodhicitta 62, 71, 82, 88, 96sq, 101 sq, lokaprasiddha, lokaprasiddhi 72, 74, 75,
1 1 5,12 0 80n
b od h isattva 26, 29, 31, 52sq, 71 sq, 82sq, lokasamvj-ti 16, 42, 72, 7475
93, 96, 98, 101, 109, 111112, 118 L ok yata 90, 109
brahm acarya 23 lokottarajna 64, 66
brahman 35 n laukikajna 64, 66
laukikatattvalakana 43 n
bhakta 31
bhakti 31, 55 vacana 22
b havasam tati 15 V ajrayna 57, 66, 102, 103, 104108
bhva 2^-3, 9sq, 18, 22, 24, 36sq, 38sq, (cf. M antrayna)
53, 72, 88, 89, 91 vandhyputra, van d h ysu ta 70, 76
bhvan 52, 63, 94, 96 sq, 107, 115 va stu 4 6 n , 70, 96, 98
bh takoti 33 v c 46 n
Sanskript key-words 145

V atsiputriya 89, 91 nyatdj^ti 2

vda 23, 21, 23, 36n nya(t)vda 2
vdin 70 nyatva 3n, 44, 45 n
vsan 123 raddh 21, 24
vik alp a 11, 18, 23, 28, 35, 39, 42, 45sq, Srvaka(yna) 67, 26, 33, 62, 74, 90
55, 66, 92, 98, 106, 107, 109, 110, 113, ruta 96
1228q ruti 90
vicara 14, 40, 73, 98
vijapti 68 sam vara 93, 109
'Vijapti-Madhyamaka* (m am rig [gi] samvj-ti 3, 16, 42sq, 55, 58n , 59, 61, 64,
d b u m a ) 20n, 56, 107n, 110, 122 8, 69sq, 72sq, 74sq, 76sq, 83, 8788,
vijaptim tra 73 92, 97, 109, 113, 120, 121, 123 (cf.
vijna 20n, 66, 68, 70, 83n, 9192, 97, v y a vahara)
99, 100, 106, 109, 123 samsara 1415, 18, 24, 46, 96
V ijnavda, V ijnavdin 1, 30, 49, samskra 14, 15, 21, 28, 45
5153, 58, 61, 63, 65, 71, 7 2 -7 3 , 80, sam sk fta 13sq, 18, 38, 40n, 42n, 44n ,
81, 83, 87 sq, 9 0 - 9 2 , 9 5 - 9 6 , 99, 100, 47, 52, 63, 76, 91
108, 109, 110, 115, 122, 124 sam keta 70, 73, 74
vitan d a V III, 21 n, 3 7 n, 42, 65 sam krnti 28
VidySdharapi^aka 104 n, 105 n samklea 92
viparysa 15, 21, 3839, 52 sam gha 16, 42
vipasyan 84, 94, 9798 satkyadrt'i 41 n, 7374
vipak a 28 satkryavda 28, 51, 89
vim u kti 73 sa ttv a 28
viruddha-hetu 7778 satya 16, 21, 25, 29, 33sq, 42sq, 59,
viv ik ta 19, 20, 4 3 n 61sq, 64, 68sq, 72sq, 74sq, 7677,
viesa 89 83, 8 7 - 8 8 , 92, 9 7 -9 8 , 100, 109, 113,
viaya 20, 52, 91, 123 121
vlrya 82, 119 Satykravda, vdin 61, 89, 92, 110,
V edanta 62 (cf. Advaitadarana) 122123 (cf. Skravda)
vaitandika 3 7 n saparyyaparam rtha 64^-65, 88
Vaibhaika 66, 91, 9 2n , lOOn (cf. Sar- 8am avya 89
vstivdin) sam dhi 97, 98, 115
Vaie^ika 6n , 10, 25, 50, 51 n, 53, 62, 63 sam ropa 68, 95
vyavad n a 92 sam buddha 18
vyavahra 16, 17, 21, 25, 42, 53, 55, 59, sambhra 26, 29, 92, 97
7 0 - 7 1 , 7 0 - 7 1 , 72, 73, 7 4 - 7 5 , 121 (cf. sam yagjna 19
sarnvfti) 8am yagdfsti 68
sarvajat 63, 96, 97
abda 89, 90 Sarvadharm pratithnavdin 58
abdabrahman 89 8arvkravaropet nyat 84, 97, 98
sabdrtha 89 Sarvstivda, vdin 7, 33, 80 (cf. Vai-
amatha 84, 94, 9798 bh?ika)
nta 2n, 4 2 sq, 75 sm vftaprattyasam utpda 43 n, 76
nti 24 Skravda, vdin 58, 61, 89, 9 2 , 102n,
iva 18, 31 110, 122124 (cf. Satykravda)
uddhalaukikajna 64 Sm khya 6n, 10, 25, 50, 51 n, 53, 62, 63,
nya 3, 14, 19, 21, 22, 31, 34, 39, 41, 70
44sq, 63 sdhyasam a 12n, 22 n
unyat 2, 6, 12, 14, 1618, 20, 22sq, sm nya 89
26, 34, 42sq , 45, 52, 59 n , 63, 64, 65, sm nyalakana 80
72sq, 83, 84, 97, 98, 112, 113 svaksavacana 64, 78
nyatkoti 2n Sukhvat 27 n
unyatdarana 2 Sautrntika 59n, 61, 65 n, 66, 80, 91,
146 Indices

92n , lOOn svabh avatraya 30, 69, 92, 9 5 ,1 2 2

Sautrantika- (S vatantrika -)M adhy am aka svabh&va^unya 21, 41, 44 n
69, 66n, 69 n. See also Svatantrika- svarasavahin 98
M adhyam aka svalak$ana 43 n, 80
skandha 12, 13, 15, 19, 24, 28, 40, 74, 81 svasam vitti, svasam vedana 6 6 n, 70,
syadvad a 90 73, 80, 87n, 91
svatantr&numana 68, 61 sq, 65sq, 71, Svatantrika-M adhyam aka 58, 6171,
77 eq, 79, 88 79, 8 5 - 8 6 , 8 7 -8 8 , 115
svaprasiddha 79
svab hava 23, 10, 14, 24, 28, 30, 36, h ita 25
38sq, 41, 4647, 63, 64, 68, 76, 89, h etu 9sq, 15n, 28, 63, 64, 70, 77, 97,
9 1 - 9 2 , 112 112 (cf. lirtga)

IV. Tibetan terms

sgyu m a rigs grub p a 5 9 n m a m rig (gi) dbu m a 56n, 122

cig car jug pa 94n, 107 rnal *byor spyod pai dbu m a 59
chad B to n 34 rjogs chen 107
jig rten grags ede spyod p ai dbu m a gzan stoh 34, 35n, 56n
(pa) 59n, 80n rah rgyud pa 5859
gtam 6hogs 8 rah stoh 34, 45 n
bstod 6hogs 8, 31 rab tu m i gnas pa 59 n
th al *gyur ba 5859 rigs chogs 8, 31
m do sde p a i dbu m a 59 rim gyis jug pa 94n, 107
m do sde sp yod p ai dbu m a 59 h v a ah 94, 99