SANSKRIT FRAGMENTS OF A LOKOTTARAVĀDIN TRADITION¹
Paul HARRISON (Auckland)
INTRODUCTION
Among the dozen Chinese translations of Mahāyāna sūtras attri-
buted to the Indo-Scythian eramana Lóujiachanine
(Lokakṣema), active in the Han capital of Luoyang during the
latter half of the 2nd century A.D., is to be found a short
text entitled F6 shuo nội sàng bãi bão jing là để tỀU G
經 ​, the Sutra of the Hundred Gems of the Inner Treasury
as Expounded by the Buddha. 2 A comparison of the style and
terminology of this work with those of the best-authenticated
translation by Lokakṣema (his version of the Aṣṭasāhasrikā-
prajñāpāramitā, Taishō No. 224, dated A.D. 179) reveals that
the attribution is fully justified; we can therefore be sure
that we have here one of the oldest extant examples of Buddhist
literature translated into Chinese. A Sanskrit title for the
Nèi sàng bài bão jing, viz. (Arya) Lokānuvartana (namama hāyāna)
sūtra, has been preserved for us by the only other known trans-
lation, the Tibetan, done by Jinamitra, Dānasila and Ye ses sde
around the beginning of the 9th century and entitled 'Phags pa
'jig rten gyi rjes su 'thun par 'jug pa kes bya ba theg pa chen
po'i mdo. 3 It is not surprising that, separated as they are
by some six hundred years, the Chinese and Tibetan versions
of the Lokānuvartanasūtra display certain differences. The
fundamental identity of the work, however, is beyond question.
To begin with, the most salient characteristic of the
Tibetan translation is its form: it is entirely in verse
113 four-line stanzas with seven syllables to the line, except
for a hypermetrical first line, the hallowed formula 'di skad
bdag gis thos pa dus geig na yet it has the conventional
-
opening and ending of a sutra: the Buddha is staying on an
unspecified mountain with an unspecified number of bodhisattvas
and érāvakas; Mañjughoṣa (i.e. Mañjuśrī) requests him to expound
a sūtra describing the qualities of the Buddhas; the Buddha does
so, after which the assembled company praises his words with
rejoicing. In the Chinese version, on the other hand, there
is no equivalent for the opening formula evan mayā śrutam (not
unusual in the earlier translations); the Buddha is staying on
the Gṛdhrakūta near Rājagṛha, attended upon by 12,000 bhikṣus
212
and 72,000 (variant: 102,000) bodhisattvas; and Mañjuśrī is
described as a bodhisattva (in the Tib. version he is merely
mkhas pa, 'the wise'). Apart from this addition of further
detail, the most notable difference is that the Chinese version
is entirely in prose. However, it appears that it was not until
the beginning of the 3rd century that the technique of render-
ing Sanskrit gāthās in unrhymed Chinese verse was developed.
Certainly, in those translations which can definitely be con-
sidered as Lokakṣema's, gāthās are always done in prose, and
therefore it is more than probable that the text of the Lokānu-
vartana (hereafter LAn) which he had before him was in verse,
except perhaps for the opening and closing passages.
-
Turning now to the body of the sutra, the actual exposition
delivered by the Buddha, we find a repetitive structure which is
in fact highly suited to expression in verse. The text falls,
very roughly, into two halves. The first half deals with the
Buddha's person, working item by item through the major events
of his life (e.g. the birth of Rahula) as well as its recurrent
day-to-day activities (e.g. cleaning the teeth) to show that
all these are part of a deliberately fostered illusion of
humanity that the true nature of the Buddha's being and his
real attributes are quite different from those which he projects
among men in order to further their salvation. Despite all
appearances to the contrary, the Buddha remains above this
world he is 'supramundane' or 'transcendent' (lokottara).
The second half of the sūtra addresses itself by and large
to the Buddha's teaching, and once again, the truth about
phenomena is contrasted with what the Buddha actually pro-
claimed during his ministry. The Buddha's teaching, like
his person, is far more than it appears to be. It is in this
'ontological' section of the LAn that one finds expressed a
doctrinal standpoint very similar to that of the early Prajñā-
pāramitā as set forth in the Astasahasrikā, a standpoint which
insists upon the emptiness' of all dharmas, their non-arising
(the term anutpattikadharmakṣānti occurs) and lack of own-being
(svabhāva), the merely conventional validity of verbal distinct-
ions as opposed to the true undifferentiated nature of the
dharmadhātu, and so on.
The procedure throughout the work, in both its 'Buddho-
logical' and 'ontological' sections, is the same: each verse
contrasts a particular aspect of the 'show' with the reality
behind it, after which comes the refrain (in the Chinese version)
'it is in conformity with the ways of the world that he makes
such a show', or (in the Tibetan) 'di ni 'jig rten 'thun 'jug
yin, 'jig rten 'thun (par) 'jug (pa) being the lokānuvartana
('conformity with the world') of the title.
213
-
Now, the central theme of the LAN that the Buddha,
while appearing to share our lot, in reality transcends all
earthly limitations is nothing more than the cardinal
doctrine of the Lokottaravādins, that subsect of the Maha-
samghikas whose Buddhology has so often been compared to
Christian docetism.4 We possess very few sources for the
tenets of the Lokottaravādins, as indeed is the case for
many other Buddhist sects. Of these the most significant
is the Mahāvastu, a massive 'biography' of the Buddha which
served as the introduction to their Vinaya5. Yet even the
Mahāvastu (henceforth Mv) has no more than a sprinkling of
passages which make explicit reference to the docetic ideas
which are regarded as the hallmark of the sect; and only one
of these is of any length. When we examine this passage,
which is in effect the locus classicus for Lokottaravādin
Buddhology (Senart's edition, Vol. I, 167.15-170.10), we find
that it displays so many close correspondences with the text
of the LAN that there can be no doubt that the two are some-
how related. This connection was first discovered and demon-
strated by Takahara Shinichi in his article 'Mahāvastu ni
mirareru fukutokuron', in Fukuoka daigaku sanjūgoshunen
kinen ronbunshu, Jinbunhen (1969), pp. 117-141, and subse-
quently commented upon by Shizutani Masao in his Shoki daijō
bukkyō no seiritsukatei (Kyoto, 1974), pp. 282, 315-18.
However, a proper comparison of the two texts has not yet
been possible because of the unsatisfactory nature of Senart's
edition, a magnificent achievement for its time but now very
much in need of replacement, in view both of the advances in
the study of Buddhist Sanskrit which have since been achieved
and of the number of other mss. of the M. which have come to
light.
This paper will therefore present a new 'edition' of that
portion of the M relevant to the LAn, and then illustrate the
correspondences and connections between the two works.6
THE TEXT: PRELIMINARY REMARKS
Thanks to the researches in 'Buddhist Hybrid Sanskrit'
pioneered by Franklin Edgerton, it is now possible to improve
much of Senart's edition of the Mv merely by reference to its
excellent critical apparatus - such improvements being in most
cases simply the reversal of Senart's own efforts to arrive at
a text better - i.e. more in line with classical Sanskrit -
than that which he found in his mss., which were admittedly
of rather indifferent quality. In addition, a number of other
mss. of the Mo have since been discovered, some of which give
214
-
considerably better readings. A new critical edition of the
entire work on the basis of these mss. - an urgent desideratum
for Buddhist and Sanskrit studies naturally presupposes a
thorough investigation of the language of the text in order
to determine its particular features. Until this is done we
have no proper standard according to which any part of the
work can be tackled. In view of this, the edition of the Mv
passage which follows below is in no sense intended to be
critical: I have adopted the text of the best ms. (P) whenever
no demonstrably better reading is found elsewhere, and I have
made emendations only where no ms. yields sense (the few cases
of this are clearly marked), thus allowing to stand many in-
stances of irregular spelling and sandhi, violations of metre,
and so on (see especially vv. 2,3,8a, 14c, 15a, 26c). The esta-
blishment of a 'perfect text', whatever one might consider
that to be, is outside the scope of this paper, for the pur-
poses of which the sense of the passage is of primary importance;
and the sense is generally clear enough. For the sake of brevity
a full apparatus has been omitted.
In addition to Senart's own critical apparatus, the follow-
ing four mss. have been consulted (arranged in descending order
of quality, inasmuch as this can be determined on the basis of
such a small passage):
P
=
X
=
Y:
2
Ꮓ
=
=
Photocopy held by the Staatsbibliothek Preussischer
Kulturbesitz, Berlin, of a palmleaf ms. (427 leaves,
6 lines, 57 x 5.5 cm.) of the Mo in private possession,
Nepal: Hs. or. sim. 1176.7 Its exact date is not known to
me, but according to information received it is many
centuries older than the oldest ms. used by Senart;
it is certainly far superior to any of his in its
readings. Our passage occurs at ff. 46a6-47a2 (a
recto, b = verso).
=
Microfilm made by the Nepal-German Manuscript Preser-
vation Project of a paper ms. (446 leaves, 10 lines,
29.5 x 16 cm.) of the Mv in Nepal: Running No. D919
(Reel No. D43/3). Our passage occurs at f. 45a4-45b6.
Microfilm made by the N.G.M.P.P. of a paper ms. (256
leaves, 12 lines, 43 x 11 cm.) of the Mv in Nepal:
No. 4-127 (Reel No. B 98/14). Passage occurs at
f. 27a12-2768.
Microfilm made by the N.G.M. P. P. of a paper ms. (264
leaves, 14 lines, 47 x 20.5 cm.) of the Mv in Nepal:
No. 3-236 (Reel No. B 99/3). Passage occurs at f.35a3-
35b3.8
The text is followed by a translation.9
215
TEXT
bahudha bahuprakāram cakṣum bhavate visuddhacakṣūņām /
jaramaranamardanānām buddhanām adāntadamakānāṁ // i
lokottara bhagavato carya lokottaram kusalamulam /
gatam sthitam niṣannam sayitan lokottaram munino // 2
yat tat sugataŝartram bhavate... 10bandhanakṣayakaraṇām /
lokottaram tad api bho ity atra na saṇśayaḥ kāryo // 3
ctvaradharanam munino lokottaram atra samŝayo nāsti /
ahārāharaṇam atho lokottaram eva sugatasya // 4
desanā naranāgānām sarvalokottarā matā /
yathātathā pravakṣyāmi māhātmyam varabuddhināṇ // 5
desakālavasam prāpya paripakam ca karmaṇaḥ /
sadyam va abhinirvṛttam dharmam desenti nāyakāḥ // 6
lokānuvarttanā buddhā anuvarttanti laukikīm /
prajnaptim anuvarttanti yathā lokottarām api // 7
iryapatham¹¹ darŝenti catvāraḥ puruṣottamāḥ /
no cet parisramas teṣām jāyate subhakarmiņām // 8
pādām ca nāma dhovanti na caiṣām sajate rajaḥ /
pādāḥ kamalapatrābhā eṣā lokānuvarttanā // 9
snayanti nama sambuddha na caiṣām vidyate malo /
bimbe kanakabimbabhe eṣā lokānuvarttana // 10
dantaponam ca sevanti mukham cotpalagandhikam // 11
nivasanam nivāsenti pravaramti ca civaram /
vairambha pi ca vātāna¹2 vikopenti na dehakam /
vastram purusasimhānām eṣā lokānuvarttanā // 12
chāyāyāṇ ca niṣidanti ātapaś ca na bādhati /
buddhānām subhaniṣyaṇdām eṣā lokānuvarttanā // 13
auṣadham pratisevanti vyādhis caiṣāṇ na vidyate /
dayakānam phalam bhavatu eṣā lokānuvarttanā // 14
prabhus ca karma dhārayitum karmam darśenti ca jină /
aiśvaryam vinigūhanti eṣā lokānuvarttanā // 15
216
karonti nāma āhāram na caiṣām bādhate kṣudhā /
janatāyānukamparthan eṣā lokānuvarttana // 16
pibanti nāma pānīyaṇ pipāsā ca na badhate /
tad adbhutan maharṣiņām eṣā lokānuvarttanā // 17
cīvarāṇi ca bādheti¹³ ṣadā ca prāvṛtā jinā /
yathā rūpyabhave devā¹4 eṣā lokānuvarttanā // 18
keśām ca orupayanti na caiṣām chindate kṣuraḥ /
keśā nīlāmjananibhā eṣā lokānuvarttanā // 19°
jaram ca upadarśenti na caiṣām vidyate jarā /
jina gunaganopetā eṣā lokānuvarttanā // 20
kalpakoṭīm asamkhyeyam punyeṣu pāramimgato /
alabdhi upadarsenti eṣā lokānuvarttanā // 21
na ca maithunasambhūtam sugatasya samuchrayam
mātāpitṛn ca darŝenti eṣā lokānuvarttanā // 22
dipamkaram upādāya vītarāgas tathāgataḥ /
rahulam putra darsenti eṣā lokānuvarttana // 23
kalpakotim asamkhyeyam prajñāyā pāramimgatā /
balabhavam ca darsenti eṣā lokānuvarttanā // 24
aham agrā ti bhāṣitvā (?)15 asmim loke sadevake /
puna paryeṣanti tirthe eṣā lokānuvarttanā // 25
budhyitvā-m-atulām bodhim sarvasatvāna kāraṇā /
alpotsukatvam pradarsenti eṣā lokānuvarttana // 26
1.
2.
TRANSLATION
Manifold in many ways is the vision of those whose vision
is purified,
The Buddhas, destroyers of old age and death, and tamers
of the untamed.
The conduct of the Lord is supramundane, his root of
goodness is supramundane,
The walking, standing, sitting and lying down of the Sage.
are supramundane.
217
3.
4.
5.
6.
7.
8.
9.
As for the Sugata's body, which causes the destruction of
the fetters,
Even that is supramundane, sirs; there ought to be no doubt
about this.
The Sage's wearing of robes is supramundane; there is no
doubt about this.
The Sugata's eating of food is likewise purely supramundane.
The teaching of the Elephants-among-men is held to be
above all worlds;
In accordance with truth I shall describe the majesty of
those of excellent intelligence.
Having acquired control over place and time, and the
fruition of karma,
The Leaders teach the Dharma, as if it had just come into
existence.
The Buddhas conform to a mundane conformity with the world
In such a way that they also conform to the designation of
supramundane (?).
The Supreme Men display the four modes of deportment,
Although fatigue does not arise in them, whose karma is
good.
They do indeed wash their feet, but no dust sticks to them;
Their feet resemble lotus-leaves; this is conformity with
the world.
10. The Fully Awakened Ones do indeed bathe, but no dirt is
seen on them;
Their forms resemble golden images; this is conformity
with the world.
11. They use a toothbrush, yet their mouth is redolent of blue
lotuses.
12. They put on an undergarment and don the robe,
Yet among winds, even the Vairambhas (extremely violent)
do not disturb their body;
As for the clothing of the Lions-among-men, this is con-
formity with the world.
13. They sit down in the shade, yet the sun's heat does not
trouble
The Buddhas, for whom the outcome (of karma) is good;
this is conformity with the world.
218
14. They make use of medicine, yet there is no sickness in
them.
The fruit (of the act of giving the medicine) is to accrue
to the givers. This is conformity with the world.
15. Although able to suppress karma, the Victors make a show
of karma:
They conceal their sovereign power; this is conformity
with the world.
16. They do indeed eat, although hunger does not trouble them,
Because of compassion for mankind; this is conformity with
the world.
17. They do indeed drink water, yet thirst does not trouble
them.
That is a marvel of the Great Seers. This is conformity
with the world.
18. He puts on (?) robes, yet the Victors are always clothed.
Like the gods in the Realm of Form; this is conformity
with the world.
19. They cut off their hair, yet no razor cuts them;
Their hair is like blue-black pigment; this is conformity
with the world.
20. They make a show of old age, but there is no old age for
them;
The Victors are endowed with a host of good qualities;
this is conformity with the world.
21. Having attained perfection in merits for incalculable kotis
of kalpas,
They make a show of not getting things; this is conformity
with the world.
22. Although the body of the Sugatas is not produced by sexual
intercourse,
They make a show of mother and father; this is conformity
with the world.
23. Since the time of Dīpamkara the Tathāgata is free of passion;
He shows Rahula as his son; this is conformity with the world.
24. Having attained perfection in wisdom for incalculable kotis
of kalpas,
They yet make a show of (foolish) childhood; this is con-
formity with the world.
219
25. Having announced 'I am supreme in this world with its
gods' (?),
They yet seek out (heretical) teachers; this is conformity
with the world.
26. Having awakened to peerless awakening for the sake of all
beings,
They show a state of indifference; this is conformity
with the world.
TEXTUAL PARALLELS
Before presenting the parallels between this section of
the Mu and the Chinese and Tibetan versions of the LAn, a few
words concerning the passage as a whole are in order. To begin
with, the first four verses, which are written in Āryā metre,
act as a kind of introduction to the 22 Slokas which follow
them, and anticipate much of their content, mentioning e.g.
in v. 2 the four 'modes of deportment' subsequently taken up
in v. 8, or in v. 4 the wearing of robes, later dealt with
by vv. 12 and 18. It is likely that they were composed for
that purpose when the Ślokas were first set in their present
position in the text of the Mu, between a eulogy of the Buddha
delivered by one Vāgīsa (163.12ff) and a description of the
qualities of the Buddha's voice (170.11ff). The patchwork
nature of the Mv is fairly evident here, but although our
passage has little to do with its immediate surroundings, v. 1
does in fact connect it with certain material introduced
earlier (pp. 158ff) concerning the five 'eyes' or modes of
vision of the Buddha. There, in the course of setting forth
the marvellous powers of his 'physical eye', the text asserts
(159.3-4): 'For there is nothing about the Perfectly Awakened
Ones which is like the world. Now then, everything about the
Great Seers is supramundane. In the same way the attainment
of the Perfectly Awakened Ones is also supramundane. '16 Thus,
despite its present position, our passage may be seen as an
elaboration on this statement which has possibly been displaced
by even later interpolations.
The Arya stanzas are of little interest for our present
purposes, it being the Ślokas which take us back to the LAn,
with the lost Sanskrit text of which they share rather more
than the same refrain and the same metre (see below). Not all
the My verses have their equivalents in the Tibetan and Chinese
220
versions of the LAn, nor are those correspondences which do
exist always verbatim. The following list is not exhaustive:
there are also several instances where a thematic connection
can be discerned even though the wording of the two texts differs.
=
Mv, v.9 LAn (C), §25 (751c28-752a1):
The Buddha's17 feet, like lotus flowers, do not
get dirty. The Buddha washes his feet. It is in
conformity with the ways of the world that he
makes such a show
=
LAN(T), v. 29:
Mv, v.10 =
Although their feet are like lotus-leaves
In that dust does not soil them,
They wash their feet;
This is conformity with the world.
LAN (C), §26 (752a2-3):
The Buddha's body, like gold, does not get dirty.
The Buddha makes a show of taking a bath. It is
in conformity with the ways of the world that he
makes such a show.
= LAN(T), v. 30:
=
Even though they are free of dirt
And their bodies are like the colour of gold,
The Fully Awakened Ones wash;
This is conformity with the world.
Mu, v.11 LAn (C), §27 (752a4-5):
=
=
The Buddha's mouth is fundamentally clean, with
a fragrance like that of kunkuma. Nevertheless
the Buddha cleans his mouth with poplar twigs.
It is in conformity with the ways of the world
that he makes such a show.
LAN(T), v. 31:
Their mouths have the fragrance of a lotus,
And their teeth are spotless like pearls,
Yet they use a toothbrush;
This is conformity with the world.
Mu, v.14 LAn (C), §31 (752a12-13):
The Buddha's body has never suffered illness, yet
221
he manifests illness, summons doctors and takes
medicine. The giver of the medicine obtains im-
measurable blessings. It is in conformity with
the ways of the world that he makes such a show.
= LAN(T), v. 34:
=
Even though they are free of illness,
In order to bring the Sangha together18
They even take medicine;
This is conformity with the world.
Mv, v.16 LAn (C), §28 (752a6-7):
=
The Buddha has never been hungry. Because of com-
passion for the men of the ten directions 19 he
manifests hunger to them. It is in conformity
with the ways of the world that he makes such a
show.
LAN(T), v. 32:
Even though they are free of hunger (?),
They take food,
And for the donor the fruit is great;
This is conformity with the world.
Mv, v.18 = LAN (C), $38 (752a26-27):
=
Mv, v.19 =
The Buddha is like empty space. He makes a show
to others of always wearing clothes without ever
taking them off, just as the men of the Brahma-
heaven (s) always wear clothes. It is in conformity
with the ways of the world that he makes such a
show.
LAN (T), V. 41:
Like Brahma in Brahma's house,
The Jinas always wear clothes,
Yet they accept clothes;
This is conformity with the world. 20
LAN (C), §39 (752a28-29):
The Buddha's head has never had its hair shaved off.
He simply shows it to others, yet nobody sees it
being removed with a razor. It is in conformity
with the ways of the world that he makes such a
show.
222
=
=
LAN (T), v. 42:
Even though there is nothing which a razor can cut,
Because the crown of the head cannot be looked upon,
They make a show of shaving the head;
This is conformity with the world.
Mv, v.20 LAN (C), $30 (752a10-11):
=
The Buddha's body never grows decrepit with age. He
is only endowed with all virtues, yet he manifests
a body which grows decrepit with age. It is in con-
formity with the ways of the world that he makes
such a show.
LAN(T), v. 35:
Even though the Buddhas are endowed with all qualities
And their bodies are free of change,
They manifest old age;
This is conformity with the world.
Mv, v.21 LAn (C), §42 (752b6-7):
=
The blessings of the Buddha's merits are inexhaustible,
and no-one can surpass them. The Buddha enters the city
for pindapata, and leaves with an empty patra. It is
in conformity with the ways of the world that he makes
such a show.
= LAN(T), v. 45:
The Buddhas are exalted above all
By virtue of all their inexhaustible merits,
Yet they make a show of the unobtained;
This is conformity with the world.
Mv, v.22 LAn (C), §12 (751c2-3):
The bodhisattva is not born from the sexual union of
father and mother. His body is magically produced,
like an illusion. He makes a show of father and mother.
It is in conformity with the ways of the world that he
makes such a show.
=
LAN (T), v. 15:
Although the bodies of the Sugatas
Are not produced from sexual intercourse,
They yet manifest father and mother;
This is conformity with the world.
223
=
Mv, v. 23 LAn (C), §16 (751c10-11):
The Buddha illuminated the men of the ten directions
21
in the blindness of lust. It is in conformity
with the ways of the world that he makes a show of
wife and child.
= LAN (T), v. 20:
Ever since Dipankara
The Tathāgata has been free of desire;
Rahula is called 'son';
This is conformity with the world.
Mv, v.24 = LAn (C), §15 (751c8-9):
Since many thousand myriad kotis of asamkhyeyakalpas
ago the Buddha has realised prajñāpāramitā.
It is
in conformity with the ways of the world that he
makes a show of (being) a small boy.
= LAN(T), v. 16:
Since incalculable koṭis of kalpas
He has arisen from the perfection of wisdom,
Yet he manifests childhood;
This is conformity with the world.
Mv, v.25 LAn (C), §17 (751c12-13):
=
=
When the bodhisattva was born and dropped to the
ground he himself uttered the words: 'In the whole
world no-one surpasses me! I shall deliver the men
of the ten directions!' It is in conformity with
the ways of the world that he still asks questions.
LAN (T), v. 21:
Having said "I am foremost
In the world with its gods!'
He afterwards seeks the Dharma;
This is conformity with the world.
Mv, v.26 = LAN(C), §20 (751a18-19):
Because it has become equal to all (?), the Buddha's
wisdom cannot be surpassed. It is in conformity
with the ways of the world that, having obtained
Buddhahood, he sits in peaceful seclusion, and makes
a show of this to the world.
224
= LAN(T),
LAN (T), v. 24:
Even though he, having awakened to incomparable
awakening,
Is the teacher of the world with its gods,
He manifests a state of little mental activity22;
This is conformity with the world.
As stated above, the LAn falls roughly into two parts,
the first (up to v. 55 in the Tib.) dealing with the person
and life of the Buddha, the second (v. 56 onwards) concerning
itself primarily with his teaching. All the 14 verses with
parallels in the Mv just cited are to be found in the first
part, although the order in which they appear is not the same
for both texts. The LAN is rather more logical in dealing
first with specific events in the life of Sākyamuni and then
with recurrent activities. Not only is this procedure reversed
in the Mv, but one verse is definitely out of order: v. 23, in
which the future Buddha displays Rahula as his child, should
come after v. 24, in which he makes a show of being a child
himself. Here and elsewhere the Mv betrays itself as a
pastiche
-
and a clumsy one at that
-
in which verses have
been taken from, or based on the LAn, either in the form in
which we now have it, or, more probably, in an earlier redaction.
OTHER RELATED TEXTS
There we would have to leave the problem, were it not for
the fact that parallels to the LAn appear in other works besides
the Mv. First of all, there are distinct echoes of the text in
Nāgārjuna's Niraupamyastava 23, which indeed addresses itself
to the same theme as the LAN. Vv. 19 and 20 are especially
good instances of this:
nāmayo nāśuciḥ kaye kṣuttṛṣṇāsambhavo na ca /
tvayā lokānuvṛttyartham darsitā laukiki kriyā //
karmavaranadosas ca sarvathanagha nästi te /
tvayā lokānukampārthan karmaplutiḥ24 pradarśitā //
These and other 'echoes' make it a strong possibility that
Nāgārjuna (c. 2nd century A.D.) was familiar with some form
of the LAn tradition; being a competent poet, however, he
225
-
-
has not incorporated chunks of the text wholesale in his own
work as is the case with the Mv in a way which enables
us to establish a certain connection. It is otherwise with
Candrakirti (late 6th or early 7th century), who, in his com-
mentary on Nāgārjuna's Mūlamadhyamakakārikās, cites two Slokas
(ad MKK XXVI.2) which are of particular interest, in that they
provide independent evidence of the original form of our sūtra.
In his edition of the text? 25 Louis de La Vallée Poussin, faced
with the corruptions of his mediocre mss., converted the two
verses into Sanskrit (see p. 548). However, J.W. de Jong's
'Textcritical Notes on the Prasannapadā' (IIJ 20 (1978), pp. 25-
59, 217-252), which presents readings of a ms. of the work
superior to those which La Vallée Poussin utilised, shows
clearly (p. 247) that Candrakīrti has deliberately enshrined
the verses in his commentary in their original Prakrit form.
It is of course impossible to 'edit' such a tiny fragment of
text, but on the basis of de Jong's 'Notes' and La Vallée
Poussin's transcriptions of his mss. (p. 548, nn. 3 and 4),
it appears that the two verses ought to read something like
this:
āgamasūtreṣu/
na vinasta (or: vinaṭṭha?) na uppanna (?)
dharmmadhausamam jagam /
sattadhaum ca damsesi eṣā loaānuvattaṇā //
tisu adhvāsu sattanam pakati nopalambhati /
sattadhāum ca damsesi eṣā loānuvattanety ādi //
Not destroyed (?), not produced, the world is the same
as the dharmadhātu,
Yet he taught the sattvadhātu; this is conformity to the
world.
He does not apprehend the nature of beings in the three
times,
Yet he taught the sattvadhātu; this is conformity with
the world. 26
Unfortunately the Tibetan translation of the Prasannapada has
a lacuna at this point, but, as La Vallée Poussin himself
pointed out, the same two verses also appear at the end of a
series of seven quoted in Candrakirti's Madhyamakāvatāra, of
which only the Tibetan version has survived. There they run:
27
zig pa med cin skye med la/
chos dbyins dan mam par gyur kyan /
sreg pa'i bskal pa brjod mdzad ρα /
'di ni 'jig rten mthun 'jug yin //
226
dus gsum dag tu sems can oyi /
ran bain dmigs pa ma yin ia /
sems can khams kyan ston mdzad pa /
'di ni jig rten mthun 'jug yin //
The main differences from the Sanskrit are in the first verse,
where there is no equivalent for jagam, and the phrase sreg
pa'i bskal pa, 'the kalpa of burning', occurs instead of sems
can khams. On the basis of this La Vallée Poussin (Prasanna-
padā, p. 548, n.3) suggested that the scribes may have inad-
vertently copied the third pada of the following verse, and
proposed the alternative reading tejaḥkalpam ca. Kalpoddāham
ca is perhaps more likely; cf. Ruegg, op. cit., p. 460, nn. 50
and 54.
"
Now these seven verses in the Madhyamakāvatāra, which are
introduced with the words 'As is stated at length in the Verses
according to the Purvasailas... 28. are all to be found in the
LAn, although not in the order in which Candrakīrti has quoted
them. 29
In most cases the two Tibetan translations agree
verbatim, which suggests that one of them has been based on
or revised against the other. As far as concerns the two
verses which also appear in the Prasannapadā, the wording
of the second in the LAn(T) is exactly the same as in the
Madhyamakāva tāra, but that of the first differs slightly:
Eig pa med cin skye (2 eds.: skyes) med la/
chos kyi dbyins dan mñam gyur kyan /
sreg (3 eds.: bsreg) pa'i bskal pa ston mdzad pa /
'di ni 'jig rten 'thun 'jug yin // (LAN(T), v.94)
In the Chinese version of the LAn the two verses are translated
as follows:
§86 (753b18-19)
Fundamentally nothing is produced and nothing is destroyed.
The original realm of the dharmas (jingběnjiè )
is all-pervasive.
The Buddha shows men the time of the destruction of the
world. 30
It is in conformity with the ways of the world
that he makes such a show.
$58 (752c13-14)
Fundamentally there are no men, past, future or present.
The Buddha shows men within the Five Ways (wuddo
pancagati) of death-and-birth. It is in conformity with
the ways of the world that he makes such a show.
227
Before attempting an explanation of the connection between
the LAN and the quotations in the Prasannapada and the Madhya-
makāvatāra, we must first adduce one further piece of evidence.
In the Tibetan translation of Candrakirti's Sūnyatāsaptativṛtti
two more verses are cited. 31 The first of these, introduced
with the words lun las kyan gsuns pa, corresponds to LAn(T),
v. 13; the second, introduced with the words de nid kyi na
'jig rten mthun 'jug mdor gsuns pa, corresponds with certain
discrepancies to LAN(T), v. 87.
-
32
·
If we now put our findings together, we have Candrakirti
(late 6th, early 7th century) citing in various works of his
a total of eight different verses three of which appear more
than once - from a sutra, apparently in Prakrit, which he at
one point names as the 'Jig rten mthun 'jug mdo (= Lokānuvarta-
nasūtra), twice refers to as agama or lun, and which he also
ascribes to the Purvasailas. These same verses are found in
the LAn, a Mahāyāna sūtra translated into Chinese towards the
end of the 2nd century A.D. It seems probable, therefore,
that Candrakīrti had access to a version of this LAn, although
what its exact form or extent was we cannot know. The fact
that, with the exception of (T) v.13, all the verses he cites
fall in the 'ontological' part of the sūtra is, in my view,
a reflection of Candrakirti's own philosophical pre-occupations
rather than an indication that he had before him a LAn sub-
stantially different from the one which has come down to us.
What is intriguing, however, is that he refers to it not as
a Mahāyāna sūtra but as a canonical text (agama) of the Pūrva-
śailas, who, like the Lokottaravādins, were an offshoot of the
Hinayāna Mahāsānghika school.
CONCLUDING REMARKS
To return now to the Mv. It also has a series of verses
which were in all likelihood based on, or taken from some version
or other of the LAn, and which were introduced into the text to
give expression to those ideas about the nature of the Buddha
which we regard as characteristically 'Lokottaravādin'. However,
as Bareau has pointed out, ideas of this kind were by no means
exclusive to that particular subsect of the Mahāsāmghikas known
as the Lokottaravādins (to whom the Mv belongs), but were the
common property of the Mahāsāmghikas in general. Since this
is the case, they were no doubt also held by those southern
branches of the school which, being located in the Andhra region,
33
228
in such centres as Nagarjunikonda, Amaravati and Dhanyakaṭaka,
were known collectively as the Andhakas, and of which the most
important appears in fact to have been the Purvasailas. 34 Such
Buddhological notions, as well as certain ontological theories
of the Mahāsāmghikas and their subsects, closely resemble those
of the Mahāyāna. For this and other reasons it has long been
held that it was amongst the Mahāsāmghikas that the Mahāyāna
originated, a view suggested and supported by several later
Buddhist traditions. Paramartha, for example, ascribes early
splits within the Mahāsāmghikas to differences of opinion over
the authenticity of Mahāyāna sūtras such as the Avatamsaka and
the Prajñāpāramitā; 36 it is impossible to determine the validity
of this assertion, but the affinities between certain theses
attributed to the Mahāsāmghikas and various ideas propounded
in the Prajñāpāramitā literature suggest that it contains at
least a grain of truth. Other late traditions of some relevance
to the discussion are those recorded by Bu ston, who in his
History states that the language of the Mahāsāṇghika canon
was Prakrit, and by Avalokitavrata, whose Prajñāpradîpaṭīkā
ascribes a Prajñāpāramitāsūtra in Prakrit to the Purvasailas
and the Aparaśailas. Finally, we need hardly mention the
very old tradition of the pradakṣiņa of the Prajñāpāramitā
which appears in many Prajñāpāramitāsūtras, the earliest
occurrence being in Lokakṣema's translation of the Aṣṭasahas-
rikā. 39
Whatever the historical value of this piece of lore,
it does emphasise two regions: the south, where the Prajñā-
pāramitā was first practised on a large scale, and the north,
where it subsequently enjoyed considerable successes.
37
38
As far as the LAn is concerned, such geographical con-
siderations may be of some importance. Both Nagarjuna and
Candrakirti were southerners, and it is to be expected that
they would have been familiar with the literature of southern
branches of the Mahasamghikas like the Purvasailas, especially
if that literature represented a doctrinal standpoint close to
or identical with their own. On the other hand, the Lokottara-
vādins, as far as we know, had one of their main centres in the
north-west: the Chinese pilgrim Xuán Zàng records a large con-
centration of them in Bamiyān, in present-day Afghanistan, at
the beginning of the 7th century, while the early presence of
Mahasamghikas in that region is attested by an interesting in-
scription on a reliquary vase found at Wardak (30 miles west
of Kabul), which is dated in the 51st year of Kanişka, i.e.
around the middle of the 2nd century A.D., according to some
scholars.40
It is quite possible that the Kuṣan śramana Loka-
kṣema acquired or memorised his text of the LAn in this
area before setting out on the hazardous journey to China.
-
229
Such problems as the evolution of Buddhist sects and the
origins of the Mahāyāna are too complex to be gone into here.
At this stage all I wish to propose for general consideration
and comment is that the LAn in its several manifestations could
be an important piece in the puzzle of Buddhist history - a
'missing link', as it were - that it could be a relic, albeit
in a fragmentary or reworked form, of a Prakrit proto-Mahāyāna
tradition current among various subsects of the Mahāsāṇghikas
in different parts of India. The purpose of this paper has
been to present the Sanskrit fragments of this tradition and
to demonstrate their relation to each other in the light of
the complete text which has come down to us in Chinese and
Tibetan translation. At a later date, in the context of a
full study of the LAn, I hope to return to the historical
implications of the subject in more detail, and to discuss
the philosophical standpoint of the sūtra in relation to the
Prajñāpāramitā and to such later developments in Buddhological
theory as the trikaya doctrine.
NOTES
1. This paper presents some of the results of a programme of
research carried out during 1980 both in Japan, where I
spent three months as Visiting Research Fellow at the
Reiyukai Library, Tōkyō, and in Germany, where a Deutscher
Akademischer Austauschdienst grant enabled me to work for
eight months in the Seminar für Kultur und Geschichte
Indiens of the University of Hamburg. My first vote of
thanks goes therefore to these institutions and their
personnel for their generous hospitality. Secondly, I
should like to express my gratitude to Dr Akira Yuyama,
Director of the Reiyukai Library, and to Prof. Dr Lambert
Schmithausen of the University of Hamburg for their kind
encouragement and assistance in matters both scholarly and
personal at various stages of the project, and to Dr Gregory
Schopen (concurrently Reiyukai Visiting Research Fellow),
who also gave me the benefit of his friendly advice on
various aspects of the work. Needless to say, the responsi-
bility for any errors which remain rests entirely with me.
Last but not least, I should like to thank my teacher,
Prof. J.W. de Jong, for first directing my attention to
the ratnākara of Lokakṣema's translations. Since this paper
embodies one of my first discoveries therein, I am very
pleased to be able to dedicate it to Prof. de Jong on the
occasion of his sixtieth birthday.
230
2.
3.
4.
Taishō Shinshū Daizōkyō, Vol. XVII, No. 807, 751b-753c.
Note that the pin-yin system for transliterating Chinese
is used in this paper.
Derge, Mdo Tsa 303a6-308a7; Narthang, Mdo Ba 475b2-483a5;
Peking, Mdo Mu 304b2-309a6 (Peking Edition of the Tibetan
Tripitaka, Vol. XXXIV, No. 866, 250-1-2 to 251-5-6); Lhasa,
Mdo Ba 478b3-486a3 (The Xerox-Copy of the Lhasa Kanjur,
Vol. LXI, pp. 238-43). All these editions except Derge
give the Sanskrit title as Lokānusamānāvatāra, which is
clearly a spurious re-Sanskritisation of the Tibetan.
On the Lokottaravādins see esp. A. Bareau, Les sectes
bouddhiques du petit véhicule (Saigon, 1955), pp. 75-7,
and E. Lamotte, Histoire du bouddhisme indien (Louvain,
1958), pp. 690-93.
5. E. Senart, ed. Le Mahāvastu, 3 vols. (Paris, 1882-1890-
1897). See also the complete English translation by J.J.
Jones, The Mahavastu, 3 vols. (Sacred Books of the Buddhists
XVI, XVIII, XIX) (London, 1949-1952-1956), and a useful
survey of the literature in A. Yuyama, 'A Bibliography of
the Mahāvastu-Avadāna', Indo-Iranian Journal, XI,1 (1968),
pp. 11-23. Possibly completed by the 5th century, the
Mahāvastu incorporates much old material, and was no doubt
many centuries in the making.
6.
7.
It is my intention to publish in the near future a complete
study of the LAn, including: a critical edition of the
Tibetan version based on nine editions of the Kanjur (which
raises many problems of general application in the editing
of Kanjur texts); English translations of both Tibetan and
Chinese versions; an examination of parallels to and quota-
tions from the LAn in works by Nāgārjuna, Candrakīrti and
others, as well as in the Mv; and a general evaluation of
the doctrinal and historical significance of the text. The
present paper constitutes a preliminary to this larger work.
The photocopy was made available to me through the kindness
of Prof. B. Kölver of the University of Kiel.
8. For copies of X, Y and Z I am indebted to Dr D. George and
Dr Feistel of the Staatsbibliothek Preussischer Kulturbe-
sitz, Berlin.
9.
There are several translations already in existence, all
of them more or less misleading, based as they are on
Senart's edition: J.J. Jones, The Mahāvastu, Vol. I, pp.
231
132-4; E. Leumann and Shiraishi (Fujita) Shindo, 'Maha-
vastu, ein buddhistischer Sanskrittext', Heft III, Pro-
ceedings of the Faculty of Liberal Arts and Education,
Yamanashi University, No. 3 (1962), pp. 118-21; E.J.
Thomas, The Quest of Enlightenment, A Selection of the
Buddhist Scriptures Translated from the Sanskrit (London,
1950), pp. 85-6 (only vv. 5-26). There is also a partial
translation in E. Lamotte, Histoire, p. 691.
10. Four morae lacking. Senart conjectures bhavasya.
11. Original reading probably iriyāpatha.
12. I construe a partitive genitive; cf. Buddhist Hybrid
Sanskrit Dictionary, s.v. vairambha. All mss. consulted
read vairambha şi ca.
-
13. The text is unsatisfactory. Senart's mss. read badhanti,
as does Z, but P, X and Y all have bädheti, which, if
correct at all, is more likely to be the causative of
bandh- than of badh- (on the confusion of these two roots
in Pali, see PTSD, s.v. bādhati). The sense is not certain
'to bind on', 'to put on oneself', or perhaps even 'to
piece together' (the civara from rags) may be suggested,
but I can find no parallel passages to support these inter-
pretations. Senart emends to nivāsenti, but in his note
(p. 513) suggests the alternative possibility dhārenti,
which, in view of v. 4 (civaradharanam), is much more
likely.
14. Rūpyabhave probably to be taken as an alternative form of
rupabhave. Such an interpretation is supported by the
Tibetan and Chinese versions of the verse (see below),
which both refer to the Brahma Deva(s); this term was
generally applied to all the gods of the Realm of Form
(see BHSD, s.v. deva, and Hōbōgirin, s.v. Bon).
15. Ms. P has ahamantrā vibhāyitvā, X, Y and Z have aham mantrā
vibhāṣitvā, with which Senart's mss. are in general agree-
ment. My very tentative emendation is suggested by the
Chinese and Tibetan versions (see below), but an orthograph-
ical confusion of gra and tra is a definite possibility,
since in v. 3 P has agra where the other mss. read atra.
For nom. sg. masc. ending in -a see BHSG, §8.24.
16. Na hi kimcit samyaksambuddhanam lokena samamatha khalu
sarvam eva maharṣinam lokottaram / tatha hi samyaksam-
buddhānām samudagamah so pi lokottaro.
232
17. The Chinese version is not specific as to number, whereas
in many verses in the Tibetan the Buddhas are explicitly
plural, as they are in the Sanskrit of the Mv. Although
I have translated the Chinese in the singular and the
Tibetan in the plural (except where the context makes
this sound peculiar), not too much weight should be placed
upon this: in BHS distinction of number seems not to have
been of such great moment (see BHSG, §5.1-5), and in any
case what our sūtra is saying applies to Buddhas in general.
18. The pada concerning the rewards accruing to the Buddha's
benefactors has slipped out of place in the Tibetan text,
and appears in its version of M, v. 16.
19. Commonly used by Lokakṣema to render sarvasattva, 'all
sentient beings'.
20. Although the Sanskrit text is uncertain, the antithesis
it seems to be presenting is that although the Buddha is
in reality always decked out in divine splendour, he ap-
pears to don the humble civara. While the wording of the
Tibetan is close, the sense is obscured. Lokakṣema, how-
ever, has read something quite different into the verse:
since the Buddha does not possess a material body, he has
no real need to wear clothes, even though he pretends to
do so.
21. Perhaps this translation is based on a misreading of the
phrase dipamkaram upādāya.
22. Thugs las chun nu nid, a literal rendering of alpotsukatva,
the meaning of which in this context is the 'apathy' or
'disinclination' of the Buddha to teach the Dharma.
23. G. Tucci, 'Two Hymns of the Catuḥ-stava of Nagarjuna',
JRAS (1932), pp. 309-25.
24. Read karmaplotiḥ? Cf. LAn(T), v. 77: sdig pa thams cad
yons zad de / bsod nams kun gyis minon 'phags kyan / las
kyi rgyud ni ston mdzad pa / 'di ni 'jig rten 'thun 'jug
yin //. See also BHSD, s.v. ploti.
25. Mulamadhyamikakārikās (Madhyamikasūtras) de Nāgārjuna
avec le Prasannapadă Commentaire de Candrakirti (Biblio-
theca Buddhica IV) (St. Petersburg, 1903-1913).
26. Sattvadhātu: 'element', 'sphere' or 'mass' of sentient
beings. See also the translations by Jacques May, Candra-
233
kirti Prasannapada Madhyamakavṛtti (Paris, 1959), pp.
256-7, and D. Seyfort Ruegg, 'Le Dharmadhatustava de
Nagarjuna', Etudes Tibétaines dédiées à la mémoire de
Marcelle Lalou (Paris, 1971), pp. 459-60).
27. See La Vallée Poussin, ed. Madhyamakāvatāra par Candrakirti:
Traduction Tibétaine (Bibliotheca Buddhica IX) (St. Peters-
burg, 1912), pp. 134-5.
28. Ji skad du sar gyi ri bo'i sde pa dan mthun pa'i tshigs
su bead pa dag las Les rgya cher gsuns pa lta bu'o.
...
29. They correspond to LAn (T), vv. 13, 62, 57, 73 (with con-
siderable differences), 72, 94 and 63, and to LAn(C),
§§ 10(?), 57, 51, 68(?), 67(?), 86 and 58.
30. This bears out La Vallée Poussin's suggestion of scribal
error (see above): it may well have been a loose trans-
lation of kalpoddaha. Such free renderings are character-
istic of Lokakṣema, who often found it necessary to alter
or amplify his text in order to make it comprehensible to
an audience almost totally unfamiliar with Indian social
or religious life.
31. Ston nid bdun cu pa'i 'grel pa, Peking Edition, Vol. 99
(No. 5268), 15-3-8 to 15-4-1 (309b8-310al) and 36-3-3 to
3-4 (362a3-4). For this information I am indebted to
Mr Felix Erb, Hamburg.
32. The discrepancies between the two Tibetan translations are
such that at first sight they appear unrelated, even though
the correspondence is perfectly clear in the Chinese version
of the LAn (§79). What is interesting, however, is that in
the Derge edition of the Ston hid bdun cu pa'i 'grel pa (and
therefore in Cone as well) the verse is preceded simply by
the words de nid kyis na gsuns pa. Clearly the editors of
Derge, unable themselves to trace the verse in their copy
of the LAN (T), have deleted what they therefore assumed to
be an erroneous reference · a fine example of both editorial
-
meticulousness and its dangers!
33. See Bareau, Les sectes bouddhiques, pp. 57-9, 76, 301-2.
34. Ibid., pp. 89-103.
35. For a particularly cogent statement of the problem of the
origins of the Mahāyāna and its probable solution, see
Bareau's Appendice III, 'Les origines du Mahâyâna', pp.
296-305.
234
36. See P. Demiéville, 'L'origine des sectes bouddhiques
d'après Paramartha', Mélanges chinois et bouddhiques,
t. I (Brussels, 1932), pp. 41, 43.
37. E. Obermiller, transl. History of Buddhism (Chos-hbyung)
by Bu-ston (Heidelberg, 1931), p. 100; also Bareau, p.56.
A good deal of research on the Prakritic features of the
language used by the Lokottaravādins has been done by
Dr G. Roth; see especially his article 'Particular Features
of the Language of the Arya-Mahāsamghika-Lokottaravādins
and their Importance for Early Buddhist Tradition' in
H. Bechert, ed. Die Sprache der ältesten buddhistischen
Überlieferung (Abhandlungen der Akademie der Wissenschaften
in Göttingen, Philologisch-Historische Klasse, Dritte Folge,
Nr. 117) (Göttingen, 1980), pp. 78-135.
38. According to Shizutani, Shoki daijō bukkyō no seiritsukatei,
p. 282.
I have not had the opportunity to verify this my-
self.
39. See E. Lamotte, Le traité de la grande vertu de sagesse,
t. I (Louvain, 1949), p. 25, n. 1.
40. See Shizutani, Shojō bukkyōshi no kenkyū (Kyōto, 1978),
pp. 64-5. On the vexed question of Kuṣāņa chronology, see
A. L. Basham, ed. Papers on the Date of Kaniska (Leiden,
1968).
235
THE DREAM SIMILE IN VIJÑĀNAVĀDA TREATISES
Masaaki HATTORI (Kyoto)
In Mahāyāna sūtras and sastras, the unreality of phenomenal
existences is often explained by the similes of māyā, mirage
(marici), and the like. I shall cite below a few examples
from the sutras composed in the early period of the Mahāyāna
movement.
...
...
ayuṣmān Subhūtis tan devaputrān etad avocat:
māyopamās te devaputrāḥ sattvāḥ svapnopamās te deva-
putrāḥ sattvāḥ sarvadharma api devaputrā māyopamāḥ
svapnopamāḥ nirvānam api devaputrā māyopamam
svapnopamam iti vadāmi kim punar anyam dharmam
(Astasāhasrikā Prajnāpāramita, Chap. II)
...
yaś ca dharman vijānāti māyāsvapnasvabhāvakān
kadali skandhaniḥsārān pratisrutkāsamānakān
...
...
sa vijānāti nirvṛtim. (Saddharmapundarikasūtra,
Chap. V)2
...
yo 'yam ...bodhisattvaḥ saptamim bodhisattvabhumim
akramati, sa daśabhir upāyaprajñājñānābhinirhṛtair
mārgāntarārambhaviseṣair akramati. katamair dasabhiḥ?
yad uta ... māyāmarīcisvapnapratibhāsapratiśrutkodaka-
candrapratitimbanirmāṇabhāvābhāvasvabhāvādvayānugataś
ca bhavati, ... (Dasabhumikasūtra, VIIth Bhumiḥ)3
In the Mulamadhyamakakārikā (abbrev.: MMK), Chapter VII
(Samskrta-parikṣā), Nāgārjuna discusses the unreality of the
three characteristic features (lakṣaṇa) of conditioned
elements (samskṛtadharma), i.e. origination (utpāda),
duration (sthiti), and destruction (bhanga), and he states
in conclusion:
yathā māyā yathā svapno gandharvanagaram yathā
tathotpādas tathā sthānam tathā bhanga udahṛtam (VII.34)
This verse is explained by Candrakirti as follows: "Just as
māyā and the like, although they are unoriginated and non-
existent, are, for ordinary people, denotable by the words.
'māyā', etc., and apprehensible by the cognition of māyā, etc.,
236
even so these 'origination' etc., although essentially non-
existent, are indicated, merely on account of their being
commonly accepted (lokaprasiddha), by the Blessed One who is
keen to confer benefits upon the people to be educated in
that manner.
114 In MMK, these similes are employed also in
XVII (Karma-parīkṣā), k. 33 and XXIII (Viparyaya-p), k. 8.5
Āryadeva also uses the same similes in the following
verse in his Catuḥśataka, and it is cited by Candrakīrti in
his commentary on MMK, VII.32.
alātacakranirmāṇasvapnamāyambucandrakaiḥ
dhumikantaḥpratiśrutkamaricyabhraiḥ samo bhavaḥ (XII.25)6
Ten similes which serve to explain the voidness (sūnyatā)
of phenomenal existences are enumerated in the *Mahāprajñā-
paramitopadesa (Ta-chih-tu-lun), Chapter XI. They are:
māyā, mirage (marici), the moon in the water (udakacandra),
space (ākāśa), echo (pratiśrutkā), fata morgana (gandharva-
nagara), dream (svapna), shadow (chāyā), the reflection in a
mirror (pratibimba) and metamorphosis (nirmāņa). Each of
these is explained there in detail, but there is no essential
difference in the use of them. All of them are equally intended
to explain that the phenomenal elements are not really existent,
that they are essentially void (śunya).
These similes are found employed also in the Vijñānavāda
treatises. Excepting space and fata morgana, the other eight
similes are mentioned in the Mahāyānasūtrālaṇkāra (abbrev.:
MSA), XI (Dharmaparyeṣty-adhikāra), k. 30, and it is shown
8
there for which element each of these similes stands. It is
the six internal bases of cognition (adhyātmikāyatana), viz.,
the sense-organs and the mind, that are compared with māyā,
because they appear as a self (ātman) despite the fact that
there is in reality no such entity apart from them. The mind
(citta) and mental states (caitasika) are likened to a mirage,
since they produce false notions (bhrānti). In this manner
the elements which have a likeness to a shadow, echo, etc.,
are specified one by one.
Among these eight similes, dream and māyā are those most
frequently used by the Vijñanavādins, and they differ from
Nāgārjuna and his followers in their application of these
two similes. I shall examine in this paper the simile of a
dream as used by the Vijñānavādins.
In the above-mentioned passage of MSA, XI, the simile of
a dream is explained as pertaining to the six external bases
237
9
of cognition (bāhyāyatana). viz., the objects grasped by the
sense-organs and the mind for the reason that the enjoyment
of them is not based on reality. The same eight similes
are enumerated also in the Mahayanasangraha (abbrev.: MS),
III (Jñeyalakṣana), 10 and according to the remark made there
by the author, the simile of a dream is intended to show how
the diversity of experience can be explained from the Vijñā-
navāda standpoint. It is generally admitted that man has
agreeable or disagreeable experience in accordance with the
difference of the object experienced, but the Vijñānavādins
who deny the reality of external objects must be asked to
explain how our experience is diversified. The answer to
this question is given by the simile of a dream. The impli-
cation of this simile is elucidated clearly by Vasubandhu in
his commentary on MS: Just as we have agreeable or disagree-
able experience in a dream despite the fact that there is no
real object, even so we have diversified experience although
there is no real object to be experienced.
The simile of a dream is found also in Vasubandhu's
commentary on MSA, XVIII (Bodhipakṣadhikāra), kk. 43-44.12
It is stated there that the sensation (vedanā), which is
caused by the contact of a sense-organ with an external
object, is comparable to a dream, because it is nothing
other than a false experience (mithyanubhava) like the
experience in a dream.
Thus the simile of a dream is, in the Vijñānavāda
treatises, meant for elucidating that there is in reality
no external object and that our daily experience, whether
it is pleasurable or painful, is untrue inasmuch as it is
based on the contact with an unreal object. No such restrict-
ion regarding the usage of this simile is noticed in the works
expounding the doctrine of voidness. As seen above, in the
instances in MMK, not merely the objects perceived by the
sense-organs (XXIII.8), but also man's mental and physical
activities (klesa, karman XVII.33), and the elements which
do not accompany the mind (cittaviprayuktasamskāra), such as
'origination' etc. (VII.34), are likened to a dream. For
Nāgārjuna and those who follow his philosophical tenet, all
the elements, subjective as well as objective, constituting
the empirical world are unreal without exception. The simile
of a dream is used to explain the unreality of these elements,
and in that respect it does not essentially differ in usage
from the other similes.
On the other hand, for the Vijñānavādins the simile of
a dream has an important implication. It is held by them
that what appears as an external object is nothing other than
238
the idea produced by consciousness (vijñāna). This conscious-
ness is exactly the same as the consciousness in a dream which
creates various objects perceived in a dream. These objects
vanish as soon as man awakes from sleep, but the fact that
there was during the dream a consciousness which produced the
objects to be apprehended remains true. The objects seen in
a dream are unreal, and the pleasure and the pain experienced
in a dream are untrue, but the consciousness in a dream is un-
deniably real. Likewise, according to the Vijñānavādins, the
external objects are unreal, but the consciousness which pro-
duces external objects and grasps them really exists.
At the beginning of the Vimsatikā Vijṇaptimātratāsiddhiḥ
the following four questions are directed to the Vijñānavādins,
who deny the existence of external objects and maintain that
a cognition arises by itself from the impressions of past
experiences accumulated in consciousness: (1) Why does the
cognition of an object arise at a certain place (2) and at
a certain time, and not arbitrarily at any place and at any
time ? (3) The illusive cognition of an unreal object, such
as that of a double moon, occurs only to a person whose eyes
are diseased. On the other hand, the cognition of a pot and
the like is commonly shared by all persons who are present
at the same place. How can this fact be explained without
presupposing the reality of the external object? (4) What
is experienced in a dream does not produce a real effect. For
example, a man who is bitten by a snake in a dream does not
suffer from its poison when he awakes. But the object appre-
hended in the waking state produces a real effect. What is
the explanation of this fact from the Vijñānavāda standpoint?
To answer the questions (1), (2) and (4), Vasubandhu takes
recourse to the simile of a dream. In a dream we perceive,
for example, a village, grove, man and woman at a certain
place and at a certain time, but not at any place and at any
time. It is also a fact of experience that an object seen in
a dream does produce a real effect: for example, seminal
emission is caused in a man who has intercourse with a woman
13
appearing in a dream.
The simile of a dream here is not intended to prove the
unreality of phenomenal existences. As a matter of course
the object which is perceived or produces a real effect in a
dream is unreal, but it appears as if it were real. This
object is created by the consciousness in a dream. Likewise,
the object perceived as real in the waking state is in fact
created by the consciousness in which the impressions of the
experiences from beginningless time are accumulated. This is
the point that Vasubandhu attempted to prove by citing the
simile of a dream. Sthiramati closely follows Vasubandhu
239
when he compares, in his Trimŝikāvijħaptibhāṣya, the con-
sciousness which takes the form of an object (arthākāram
vijñānam) with the consciousness in a dream (svapnavijñāna).14
That the above interpretation of the simile of a dream
was regarded as peculiar to the Vijñānavādins is known from
the fact that their use of this simile is repudiated by the
Madhyamikas. Bhāvaviveka, who made a keenly critical exami-
nation of the doctrine of the Vijñānavādins in the fifth
chapter of his Madhyamakahrdaya, disapproved of their view
that the consciousness in a dream functions by itself inde-
pendently of the object. According to him when a man lies
asleep his bodily eyes cease to function, but he perceives
with his mental eyes (janacakṣus) the objects that have
been brought to experience before. Therefore, he maintains,
the consciousness in a dream is conditioned by the external
object.15
On the other hand, Candrakirti asserts that there
is in reality no consciousness in a dream. He forestalls
the Vijñānavāda argument that, since we remember the expe-
rience in a dream after we awake, it is known that there was
a consciousness in a dream, and rejects this argument with
the remark that, if that is the case, the reality of the
object in a dream is also to be admitted, since it is remem-
16
bered afterwards.
The Vijñānavāda doctrine is criticised in some philosophi-
cal works of non-Buddhistic systems, and in the passage in
which the critical discussion is made, it is often found that
mention is made to the dream simile. For instance, in the
Vṛttikāragrant ha quoted by Sabaras vāmin in his commentary on
the Mimāmsāsūtra, 1.1.4-5, the Vijñānavāda view, which is to
be set aside, is summarised as follows: "Every concept (pratyaya)
is devoid of an object, just like a dream. Because the object-
lessness (nirālambanatā) is observed as the nature of a concept
in the case of a dream. Also for a man who is awake, 'post'
or 'wall' is indeed a concept. Therefore, it is also object-
less. Sankara also presents with similar phrases the Vijñā-
navāda arguments in his Brahmasūtrabhāṣya, 2.2.28.18 It is
17
thus known that the simile of a dream was considered to be
quite suited for the explanation of the Vijñānavāda doctrine
of consciousness-only.
240
Notes
1.
2.
3.
Ed. P.L. Vaidya, Buddhist Sanskrit Texts No. 4, Darbhanga
1960, p. 20.
Ed. H. Kern and B. Nanjio, Bibliotheca Buddhica X, St.
Pétersbourg 1909-12, p. 142.
Ed. P.L. Vaidya, Buddhist Sanskrit Texts No. 7, Darbhanga
1967, p. 36.
4. Prasannapadā, ed. L. de La Vallée Poussin (Mulamadhyamaka-
kārikās de Nagarjuna avec la Prasannapadă Commentaire de
Candrakirti), Bibliotheca Buddhica IV, St. Pétersbourg
1913, p. 177.6-8: yathā māyādayaḥ svabhāvenānutpannā
avidya [mānā] māyādi sabdavācyā māyādivijñānagamyāś ca
lokasya, evan ete 'pi lokaprasiddhimātreṇotpād[ādalyaḥ
svabhavenāvidyamānā api bhagavatā tathāvidhavineyajanānu-
grahacikīrṣuṇā nirdiṣṭā iti.
5. kleśāḥ karmani dehas ca kartāras ca phalani ca
gandharvanagarākārā maricisvapnasamnibhāḥ (XVII.33)
rūpasabdarasasparśā gandhā dharmās ca kevalāḥ
gandharvanagarākārā marīci svapna saṇnibhāḥ (XXIII.8)
6. Catuḥsatikā by Arya Deva, ed. by Haraprasad Shastrī
(Memoirs of the Asiatic Society of Bengal, Vol. III,
No. 8), The Asiatic Society, Calcutta 1914, p. 503.4-5.
Cf. Prasannapada, p. 173.3-4.
7.
T. Vol. 25, No. 1509, p. 101cff. Cf. Étienne Lamotte,
Le traité de la grande vertu de sagesse de Nāgārjuna,
Tome 1 (Bibliothèque du Muséon 19), Louvain 1949, pp.357ff.
8. MSA, ed. Sylvain Lévi (Asanga, Mahāyāna-sutralamkara,
exposé de la doctrine du grand véhicule selon le système
Yogācāra. Tome I Texte), Paris 1907, p. 62:
9.
--
māyāsvapnamaricibimbasadṛśāḥ prodbhāsaśrutkopamā
vijneyodakacandrabimbasadṛśa nirmāṇatulyāḥ punaḥ /
sat sat dvau ca punas ca sat dvayamata ekaikasas ca trayaḥ
samskārāḥ khalu tatra tatra kathită buddhair vibuddhottamaiḥ //
Ibid., p. 62.12 (Comm. on k. 30): svapnopamāḥ ṣaṭ bāhyāny
ayatanani tadupabhogasyāvastukatvāt.
10. T. Vol. 31, No. 1594, p. 140b-c. Cf. E. Lamotte, La somme
241
du grand véhicule d'Asanga (Mahāyānasaṇgraha), Louvain
1938. Tome I: Versions Tibétaine et Chinoise (Hiuan-tsang),
p.38; Tome II: Traduction et Commentaire, p. 123.
11. T. Vol. 31, No. 1597, p. 344b.21-23.
12. MSA, p. 141.6.
13. Vimsatikā, ed. Sylvain Lévi (Vijnaptimātratāsiddhi. Deux
traités de Vasubandhu: Vimsatikā et Trimsikā), Paris 1925,
pp. 3-4.
14. Triméikāvijħaptibhāṣya, ed. Sylvain Lévi (VijЯlaptimātratā-
siddhi, Deux traités ...), Paris 1925, p. 17.2-3 (on k. lc):
evam bāhyārthābhāvād vijñānam evārthākāram utpadyate,
svapnavijñānavad ity abhyupeyam.
15. Cf. Tarkajvālā Madhyamakahṛdayavṛttiḥ, Peking ed., Mdo-
hgrel, Dsa, fol. 252.2ff. (Reprint, Vol. 96: Dbu-ma, II,
p. 92-3-2ff.)
16. Cf. Madhyamakāvatāra par Candrakirti, Traduction Tibétaine,
par L. de La Vallée Poussin, St. Pétersbourg 1907-12 (Bib.
Bud. IX), pp. 140ff.
17. Sābarabhāṣya ad Mīmāmsāsātra, 1.1.4-5. Bib. Ind. ed.,
p. 8.21-24 (E. Frauwallner, Materialien zur ältesten
Erkenntnislehre der Karmamīmāṛṇsā, p. 26.22-24): nanu
sarva eva nirālambanaḥ svapnavat pratyayaḥ, pratyayasya
hi niralambanatā svabhāva upalakṣitaḥ svapne. jāgrato
'pi stambha iti vā kuḍyam iti vā pratyaya eva bhavati.
tasmat so 'pi niralambanaḥ.
18. Brahmasūtrasānkarabhāṣya, 2.2.28 (Nirnaya Sagar ed.,
p.544): svapnadivac cedam [sarvam] draṣṭavyam. yatha hi
svapnamāyāmaricyudakagandharvanagarādipratyayā vinaiva
bāhyenārthena grāhyagrāhakākārā bhavanti, evam jāgarita-
gocarā api stambhadipratyayā bhavitum arhantity avagamyate.