Notes on the Nyaaya-pravesa by Sa^nkarasvaamin

·期刊原文
Notes on the Nyaaya-prave'sa by 'Sa^nkarasvaamin

Giuseppe Tucci
The Journal of Royal Asiatic Society of Great Britain and Ireland
April 1931
PP.381-413

--------------------------------------------------------------------------------
.

p.381

The nyaaya-prave'sa by Sa^nkarasvaamin, recently
printed in the Gaekwad's Oriental Series of Baroda
,is a text of great interest for the study of Indian
logic; in fact, in spite of its conciseness it
contains an extremely clear exposition of the
Buddhist logic as it was taught in India, at least
among some particular schools, as those of the
Yogaacaaras and the Sautraantikas(1) in that lapse
of time which separates Di^nnaaga from Dharmakiirti.
That the books was written after Di^nnaaga, but
before Dharmakiirti, is proved by its peculiarities,
which in many a point differentiate the theories
held by the author from those maintained by the
other two great logicians already quoted.(2)
The fact that the fallacies of the
proposition pak.sa-aabhaasas are here nine, and not
five as in the Nyaaya-mukha or in the
Pramaa.na-samuccaya, and that no mention of the
jaatis, mere appearance of refutation, can be found
under the item of the duu.sanaaa-aabhaasas,
erroneous refutations, betrays mata-bheda in the
connection with the system of Di^nnaaga,(3) while,
on the other hand, there is in the book no trace of
the criticism raised

1 The relation between the two schools is very close,
and this explains why Buddhist doctors belonging to
these sects are called by either name. The school
which claims to have been founded by Asa^nga and
after him by Vasubandhu, has very strict similarities
in its dogmatics with the doctrine of the
Sautraantikas. Books like the Yogacaryaabhuumi-
`saastra, the Mahaayaana-samparigraha-`saastra,etc.,
follow very often the ancient scheme of the
Sautraantikas. On the other hand works like the
La^nkaavataara and the `Sraddhotpaada-`saastra
belong to a quite different tendency in which the
vij~naana is elaborated according to mystic lines. I
hope to publish very soon a note establishing these
facts.
2 See my article " Bhamaha and Di^nnaaga ", Indian
Antiquary, 1930, p.142.
3 See my translation of the Nyaayamukha, Heidelberg,
1930. I must add that Uddyotakara also, refuting
Di^nnaaga, knows only five pak.saabhaasas: Nyaaya-
vaarttika (ed. By Lak.sma.na-saastrii,Benares,1915),
p.113.

p.382

by Dharmakiirti against some particular points of
Di^nnaaya's (1) doctrine.but if the NP. Was not
written by Di^nnaaga, as has been wrongly assumed,
it is quite certain that it expounds theories that
must have had, at least for some time, a wide
circulation. This is proved by the fact that Yuang
Chuang translated it into Chinese, while another
great Chinese scholar, to whom we owe some
fundamental works of exegesis upon the most
important `saastras of the Mahaayaana, viz.
K'uci-chi(632-82), commented on it. Moreover, we
have some evidence that the theories expounded in
the NP. were accepted even by the non-Buddhist
philosophical schools. If we take, for instance, the
Maa.thara-v.rtti, we see that in the commentary on
kaarikaa 5 the list of the aabhaasas corresponds to
that given in the Nyaaya-prave'sa (nine
pak.saaabhaasas, fourteen hetv-aabhaasas, ten
d.r.s.taanta-aabhaasas), which is strictly peculiar
to this book, while it does not occur either in
Di^nnaaga or in Dharmakiirti. This is a rather
important argument in settling the question of the
age of the Maa.tharav.rtti, which is certainly quite
different from the original book attributed to
Maa.thara himself.
K'eui-chi was a disciple of Yuan Chuang, and he
acquired from his master his wide knowledge of the
Indian Philosophy. But as Yuan Chuang himself spent
many years in India, taking part in many
philosophical discussions, and had therefore a
direct knowledge of the traditional exegesis of the
`saastras as was current in the Indian monasteries,
it is quite evident that it will be of interest to
have an idea of the exact content of the commentary
written by K'uei-chi on the Nyaaya-prave'sa and
called in Chinese タ瞶阶波 or simply 波.
Some portions of it, it is true, have been
translated or rather used by Sugiura in his study:
Hindu logic as preserved in China and Japan. But, as
I have already had the opportunity to point out
elsewhere, this book

1 The chief points in which Dharmakiirti differs
from Di^nnaaga are the two fallacies,"i.s.ta-vighaata-
krt" and "viruddhaavyabhicaarin" and the theory of
the example.

p.383

cannot be fully relied upon, as the author was very
imperfectly acquainted with the Indian logic in
general, while it is quite evident that, without a
very sound knowledge of the traditional Nyaaya it is
almost impossible to understand the difficult
discussion of K'uei-chi.
The following notes are not a translation.Wherever
the text of the Nyaaya- prave'sa is easily
intelligible, or when the commentary by the Chinese
logician does not seem to add very much to our
knowledge of the subject, I do not even refer to
k'uei-chi's notes. On the other hand, when his
discussions seemed to me of some importance, I gave
an, I hope, intelligible summary of them, trying to
reproduce, in the clearest way, the ideas expounded
by the Chinese author.
Tatra pak.saadivacanaani saadhanam
It is known that according to di^naaga the saadhana
is in fact represented only by the trairuupya
(threefold characteristic) of the hetu "reason", and
the saadharmya and the vaidharmya of the example.
After Vasubandhu (ぱ克 )the pak.sa " proposition" is
only saadhyaabhidhaana, that is to say, the
expression of the conclusion, but it is not the
saadhana. In this way Vasubandhu and Dinnaaga held
an opinion different from that maintained by the
ancient masters, viz. the Yogaacaaras, who believed
that the pak.sa or the pratij~naa is a member of the
syllogism. Now the question may be asked, why did
Sa^nkarasvaamin say that the pak.sa and the other
members are the saadhana? Two replies are possible:
(a) he meant to insist on the fact that the pak.sa(
in its totality) is the saadhya ⌒ that which is to
be proved ⌒ and that the hetu and the d.r.s.taanta
are the saadhanas "proof" of the saadhya. This he
did in order to avoid the danger of being suspected
of being a follower of the ancient masters, who
maintained about the saadhya a particular theory
which we are going to explain; (b) according to the
ancient masters the proposition to be proved is in
fact double, as the saadhya, is represented by the
subject (┦ ) as well as by the predicate (畉 ).
But for Di^nnaaga the subject and the predicate are
only the

p.384

basis, as it were, (ㄌ aa'sraya) of the pak.sa
"proposition", and, being separately no object of
discussion, cannot be taken as saadhya. This,
therefore, is represented by the pak.sa in its
entirety.
Let us take, for instance, the example : "sound is
noneternal."
According to the ancient logicians the saadhya is
twofold: we have to prove not only the subject "
sound", but also the predicate "non-eternal", and
the entire proposition is considered as the first
member of the saadhana. For Di^nnaaga there is only
one saadhya and this is represented by the entire
pak.sa: " sound is non-eter l."

subject
Ancient masters: saadhya is predicate

pak.sa
Saadhana is hetu
d.r.s.taanta

Saadhya is pak.sa
Di^nnaaga :
hetu
Saadhana is d.r.s.taanta

Prasiddho dharmii prasiddhavi`se.se.na

In order to explain what is dharmin and what is
vi'se.sa K'uei-chi beigns by saying that every
notion (dharma 猭 ) can be considered under two
different aspects: essence 砰 bhava, and attribute
竡 artha. If we take, for instance, the five
skandhas, ruupa, etc., represent the bhava, while
the condition of being infected or not (saa'srava,
anaa'srava), which may be predicated of them, is the
artha. In logic ( hetu-vidyaa) bhava and artha
have respectively three names:

a. svabhaava ┦
bhava b. dharmin Τ猭
c. vi'se.sya ┮

a. vi'se.sa.na 畉
artha b. dharma 猭
c. vi'se.sa 

p.385

svabhaava is that quid which is not common to other
dharmas, while the vi'se.sa.na is common to others.
The svabhaava corresponds to the svalak.sa.na, and
it is realized by an immediate perception, like that
which characterizes the aarya-j~nnaana. All that is
beyond this immediate perception is mere parikalpita
(安 ); that nature or aspect which is expressed by
words is in fact parikalpita, but it is not the real
essence of the dharmas. All this from the
ontological point of view (of the Yogaacaaras).
According to logic three explanations are possible:
svabhaava is (a) the mere fact of existence, it has
a narrow extension; the vi'se.sa.na permeates other
notions and has a wider extension; (b) the relation
between svabhaava and vi'se.sa.na is a relation of
before and after; that which is stated first is the
svabhaava; as there is jot yet previously a dharmin
to be specified. The vi'se.sa.na comes afterwards,
as there must be a dharmin to be specified; (c) a
relation between the subject (┮朝 ) and the
predicate which we suppose can be applied to it.
Thus svabhaava (= bhava(a) ) and vi'se.sa.na(=(a))
have been explained.
Now we must pass to dharmin (bhava (b) ) and dharma
(artha(b )). The svabhaava, expressing the notion in
itself is neither right nor wrong unitil it is
specified; the specifying elements is the dharma,,
and therefore, as the subject can receive it, it is
called dharmin; but the dharma itself cannot be
called a dharma until it specifies something; we can
speak of "wrong" and "right" so far as the
attribution of the dharma to the dharmin is
concerned.
Let us consider the third aspect of bhava and artha,
that is to say, of vi'se.sya and vi'se.sa
respectively. A discussion does not depend on the
subject alone, but on that attribute which is
preddicated of it, since this attribute specifies
the subject and not vice versa; therefore the
svabhaava, or subject, is the vi'se.sya and the
predicate the vi'se.sa.
Now three objections are possible:-
(a) That the bhava is the svabhaava and the artha
is the vi'se.sa.na is contradicted by the pratij~naa
of the Saa^nkhyas: " the

p.386

aattman is intelligence ( ), " quoted down below
in the NP., because expressions like " it is the
aatman" or " it is not the aatman" specify the
intelligence and not vice versa. The reply is that
according to Di^nnaaga's logic the subject of a
syllogism is considered in itself without relation
to other things, while the predicate expresses some
quality common to others (the predicate
"non-eternal" is applicable not only to sound but
also to its saadharmya d.r.s.taanta " the pot"). The
second dharma specifies the first, and therefore the
subject of a proposition is called the svabhaava and
the predicate is called the vi'se.sa.na.
(b) The dharmin is what it is, ekaartha (竡) ,
while the dharma is applicable to many, anekaartha (
竡 ); then, in the pratij~naa of the Saa^nkhyas
already referred to, how could the intelligence be
the dharma, since it is the ekaartha, and how could
the aatman, being anekaartha, be the dharmin? The
reply is that the dharmin is not in itself the
object of the discussion between the vaadin and the
prativaadin, but, since the discussion concerns the
various dharmas which are applicable to it, that
which is the object of discussion is the dharma.
Therefore the subject is the dharmin and the
predicate is the dharma; the relation between the
two is not so fixed as that which, according to the
Vai'se.sikas, exists between drayas and gu.nas, and,
therefore, must be established by syllogism.
(c) You say that the subject is specified by the
predicate which comes after it. But, if we take the
common expression "niilotpala" "blue lotus", it is
evident that, if we say only niila, it is impossible
to understand what kind of blue object we are to
mean. Is it a cloth, a tree, a flower? If, on the
other hand, we say utpala only, then would it be
possible to know what kind of lotus we have in our
mind, blue or white or red? Therefore, when we say
"niila" we exclude any other colour, while when we
say "utpala" we exclude any other blue object; so
that it is evident that the subject and the
predicate specify each other reciprocally, and that
the specifiying attribute

p.387

precedes the specified subject. The reply is that
the subject in itself is not the object of any
discussion, but variety of opinion is possible so
far as concerns the attribute that may be predicated
of it; therefore, the first is the notion to be
specified, vi'se.sya, and the second the
characteristic which specifies.
Why is the word " prasiddha" used ? The dharmin as
well as the vi'se.sa.na are the basis of the pak.sa;
but the two are not in themselves the pak.sa and
they must be prasiddha for both, for the vaadi.n, as
well as for the prativaadin; were they not admitted
by either of them; the pak.sa would be an erroneous
one. We may give two examples; (a) if a Saa^nkhya,
discussing with a Bauddha, argues that the aatman is
the perceiver, in this case the aatman is aprasiddha
for the Buddhist, that is to say, the dharmin is
aprasiddha( aprasiddha-vi'se.ya-pak.saabhaasa, NP.,
p.3). But, if the Buddhist, discussing with a
Saa^nkhya, maintains that sound is "vinaa'sin" I "
perishable ", in this case the dharma "vinaa'sin:
cannot be admitted by the opponent
(aprasiddha-vi'se.sa.na-pak.saabhaasa, NP., p.2) .
But what we have in fact to prove is the absolute
connection (avinaabhava ぃ瞒┦ ) of the dharmin
and the dharma, which is expressed in the pak.sa as
a whole; our aim is not that of proving the
aa'srayas of the pak.sa. Were this the case, we
should prove these aa'srayas of the pak.sa. and so
the saadhana would prove another proposition.
Therefore the aa'sraya of the pak.sa must be
accepted, anuj~aata (砛 ) or prasiddha; but of
course the pak.sa, with expresses their relation,
must not be prasiddha; otherwise its result would be
nil.

Vi'si.s.tatayaa

The dharmin and the dharma are reciprocally dis-
tinct; as for instance, when I say" ruupa-skandha is
not he aatman", I do not say that the vedanaa-
skandha is not the aatman, I do not say that the
vedanaa-skandha is not the aatman, nor do I say that
the aatman is ruupa-skandha. But, as dharmin and
dharma

1 It is known that saa^nkhya is a sat-kaarya-vaada
and that it admits only change but not destruction.

P.388

are combined together, they are not such as to
exclude each other. This relation is expressed by
the law of their being vi'si.s.taavyaav.rta (畉ぃ
瞒┦ ).

Svaya.m saadhyatvenepsita.

There are four kinds of pak.sa : (a) that which is
universally accepted, as, for instance, " the eyes
see material objects. " It is evident that it cannot
be matter of discussion.
(b) That which expresses a principle already
accepted by the school to which the speakers belong,
as, for instance, if two Vai'se.sikas discuss on the
existence of the aatman.
(c) When we establish a thesis in order to demon-
strate another, as, for instance, if we say that
sound is non-eternal in order to prove that it is
anaatman.
(d) A thesis which is not entailed by some a priori
theory, 'saastra- niraapek.sa (ぃ臮阶﹙ ), but which
is freely chosen, for the purpose of proving it. The
sentences saadhyatvenepsita is meant to exclude the
saadhana, that is to say the reason and the example.
The proposition is to be proved by the reason and
the example while these, in order to prove it, must
be already proved (in order to avoid regressus in
infinitum).

Hetus triruupa.h

The first distinction that we must make concerning
a reason, hetu, is that between upttti-hehu( ネ )
and upalabdhi-hetu(or vya~njaka-hetu  ).
The utpatti-hetu is threefold : vacana-utp.-h.,
j~naana-utp.-h., artha-utp.-h
(a) vacana-utpatti-hetu.The words of the vaadin,as
the reason, etc., produce a right knowledge in the
prativaadin.
(b) J~naana-utpatti-hetu. These words are nothing
else but the effect of the knowledge of the vaadin,
and therefore this is the cause of the convincing
power of those words.
(c) artha-utpatti-hetu. "Artha" is twofold:it cor-
responds either to "logical meaning" (笵瞶 ) or to
"object" (挂 ).
In the first sense it is represented by the meaning
inherent.

P.389

in the words which are expressed; in the second
sense it indicates those objects which may produce a
corresponding knowledge in the prativaadin.

Upalabdhi-hetu is also threefold:-
(a) j~naana-upalabdhi-hetu is that intelligence
which allows the prativaadin or those who are
present at a discussion to understand the words of
the vaadin.
(b) vacana-upalabdhi-hetu are those words through
which the prativaadin can understand what is the
proposition that the vaadin wants to establish.
(c) artha-upalabdhi-hetu,that is to say,the under-
standing of the meaning expressed in the words of
the vaadin. We have therefore altogether six hetus,
of which the first three are enunciated in regard to
the vaadin and the last three in regard to the
prativaadin.

pak.sadharmatva.m sapak.se sattva.m vipak.se
caasattvam
Pak.sa-dharmatva. In this sentence the word pak.sa
means dharmin only and not the relation between a
subject and a predicate as in the definition already
given. That particular predicate which is attributed
to the dharmin is called, as we saw, dharma. This
dharma can be of two kinds: (a) asaadhaara.na (ぃ
Τ ), which is attributed to the dharmin, by the
vaadin only, but not by the praativaadin, and (b)
saadhaara.na (Τ ), that is to say, accepted by
both and representing the reason. This dharma, which
is qually accepted and which is taken as the reason,
proves that other dharma, which is not equally
recognized, to be the predicate of the subject.
These two dharmas are therefore the predicable of
the subject. So in the example " sound is
non-eternal because it is a product " the dharmin
"sound" as well as the reason are accepted by both
the disputants; but the predicate "non-eternal",
which was not accepted by the prativaadin at the
beginning of the discussion, is accepted by him as a
predicate of the dharmin in consequence of the
reason, which

P.390

he recognizes as prdicable of the subject. It is
evident that this hetu is only a dharma of the
dharmin and not of the predicate.
In Chinese the sentence pak.sa-dharmatva is trans-
lated 筂琌﹙猭┦.In Skt. there is no equivalent for
筂, which usually corresponds to vyaap, or
sarvata.h. This word, according to K'uei-chi, is
used in order to express that the hetu must cover
completely the whole extension of the subject. Were
this not the case, some part of it could remain
unproved. Sapak.se sattva (珇﹚Τ┦ ). Here the
word sapak.sa is explained as meaning: of the same
class or genus (砰摸 ). Now, they may ask, why
in this case I there not the word: completely (筂 )
as before? In order to give a reply to this
objection K'uei-chi quotes the table of the hetus
(hetu-cakra) , as it is expounded in the
Nyaaya-mukha. As I have already dealt with it in my
translation of this text I do not here repeat the
long discussion. We must only remember that
according to Di^nnaaga a hetu, which is totally
absent in the vipak.sa, but is partly present and
partly absent in the sapak.sa, is valid.
Vipak.sa designates those objects or notions in
which the saadhya cannot be found. In this way
Di^nnaaga differs from the ancient masters,
according to whom vipak.sa is either the same thing
as "contradictory to the sapak.sa" (珇笻 ), as
non-eternal is contradictory to eternal, and so on,
or is meant to express a notion different from the
proposition, as, for instance, " anaatman, du.hkha,"
etc., after having established non-eternity. But
according to Di^nnaaga the vipak.sa of the
non-eternal, for inst nce, is all that to which the
predicate "non-eternal" cannot be attributed.

Saadhyadharmasaamaanyena samaano 'rtha.h
sapak.sa.h

It is necessary to say saadhya-dharma in order to
avoid any possible misunderstanding. If one
understands it as similarity with the proposition to
be proved, then everything would be sapak.sa because
even in the vipak.sa there are qualities such as

P.391

anaatmataa and so on, which can be attributed to
sound. If one understands similarity to the dharmin
of the proposition to be proved, then many
propositions would be without sapak.sa. In the
example " sound is non-eternal, like a pot", pot is
not similar to sound. The pot, in fact, can be
broken, burnt, etc., while the same cannot happen to
sound.
Vipak.so yatra so naasti

We have already seen that the vipak.sa is repre-
sented by those objects in which the characteristics
of the saadhya cannot be found.

Tatra k.rtakatvam prayatnaanantariiyakatxa.m vaa
sapak.sa exaasti vipak.se naasty eva

These two reasons are not synonymous. But when
somebody argues that sound is non-eternal, they must
be used in relation to the particular thesis of the
eternity of sound held by the different schools of
the `Saabdikas. Among these there were two main
currents; the one maintained that sound is
originated when certain conditions are given, but
after being originated it lasts eternally. Others
held that sound is existent ab aeterno, and
therefore we cannot speak of a creation of sound,
but of a manifestation of sound under special
conditions. We have therefore the thesis of the
`sabda-utpatti and the thesis of the
`sabda-abhivyaki. If the Buddhist, refuting the
theory of the `sabda-utpatti, asserts that sound is
non-eternal, because it is a product (k.rtakatvaat),
this reason is valid. But if he uses the same reason
against the theory of the `sabda-abhivyakti, the
reason is anyataraasiddha (繦ぃΘ ), because, as
we saw, that school of the `Saabdhikas held that
sound eternally exists. Therefore t he right reason
in this case will be the other one, as the opponent
also maintains that sound may manifest itself in
consequence of some exertion. If again he refutes
the `sabda-utpatti-vaada -saying that sound is
non-eternal, because it is the consequence of some
exertion, then the pak.sa-dharma is avyaapiaka 獶 )
筂 ), because the opponent

P.392

believes that only the internal sound arises from
some exertion, but not the external one.
The word aadi(not in Tib.)is explained as including
: void, anaatman, etc.

D.r.s.taanto dvividha.h, saadharmye.na vaidharmye.na
ca, etc.
Saadharmye.na means that both the predicate of the
subject of the syllogism and the reason are
applicable to the example. In the proposition
already referred to, "sound" is the subject and
"non-eternal" and "product" are respectively the
predicate and the reason. Now these two attributes
can be equally well ascribed to the emxample "pot".
In fact "pot" is that object in which the vaadin,
the prativaadin and the pari.sat agree that the
condition of being a product and of being subject to
destructi n are coexistent.
But the ancient masters believed that the example
is not included in the reason. When a proposition is
given there must be the reason and separately the
example, which is saadharmye.na (like a pot in our
example) and vaidharmye.na (like the ether) .
According to Di^nnaga the threefold characteristic
of the reason includes the two examples; the two
examples are therefore the reason.
The followers of the ancient school object: "If it
be so, the example is not distinct from the reason,
but elucidates the meaning of the reason. Then the
syllogism is in fact reduced to two members(1); why
do you enunciate the twofold example?
Di^nnaaga replies that the example, it is true,
belongs to the reason, but it is also true that the
reason by itself indicated only the
pak.sa-dharmataa, but it cannot indicate the
sapak.sa astitvam and the vipak.se naastitvam;
therefore the separate enunciation of the two
examples in necessary, in order to show that where
there is the reason there is the pak.sa.
The opponent objects again : " As the reason only
expresses the abhidheyaartha (┮膏竡, viz.
product), why should

1 This will be maintained in fact by Dharmakiirti.

P.393

it be wrong to call only that the reason and to call
the pot or the ether the examples and not the hetu?
If you ask, what shall we obtain in this way, the
reply is that, if it be so, the example must be
enunciated as a different member of the syllogism."
Di^nnaaga replies : "Then you accept the theory of
the non-Buddhist logicians according to whom the
example is only a means or instrument (よ獽 upaaya).
If it be so, it is distinct from the reason, but not
connected with it. In this way it would assert that
there are some objects having the same qualities as
that which is to be proved, but it could not have
any probative force.
Moreover,if you think that the pot constitutes the
essence of the d.r.s.taanta then you must admit that
pot and sound cannot be compared, as the one can be
seen, burnt, and so on, while the other has none of
these qualities. But when I say: "if something is a
product it is manifestly non-eternal, as a pot", in
this way there is a similarity of meaning between
the two proprieties, i.e. "being a product" and
"non-eternal", and they rest upon objects such as a
pot, etc. and again, if reason and examp le are
distinct, the saadharmya and the vaidharmya with the
proposition to be proved cannot demonstrate in any
way the indissoluble connection (ぃ瞒┦ ) of the
hetu with the pak.sa. And therefore they cannot
prove anything. In fact, as you do not formulate the
syllogism as we do ("that which is a product is
non-eternal, " etc.), therefore the pak.sa-dharma
(product) and the predicate (non-eternal) are not
connected through the saadharmya-d.r.s.taanta(pot).
So that,as the example is defective, when you say "
like the pot", you opponent will ask "But how is the
pot non-eternal?" Then you are obliged to have
recourse to another example, and, in this way, there
is regressus in infinitum (anavasthaa, 礚絘 ).
In this way it is evident that the syllogism in
three members is valid and that there is no
necessity of admitting the upanaya and nigamana of
the ancient masters. Therefore

P.394

according to Di^nnaaga this is the real formulation
of a syllogism:--
pak.sa sound is non-eternal.
hetu because it is a product,
saadharmya-d.r.s.taanta. Whatever is a product
is non-eternal, as a pot.
vaidharmya-d.r.s.taana. Whatever is eternal is
not a product as the ether.

yan nitya.m tad ak.rtaka.m d.r.s.ta.m yathaa-
kaa.sam
The vaidharmya-d.r.s.taanta is represented by the
exclusion of the predicate (non-eternal). Therefore
in its enunciation the negation of the predicate
must come before and the negation of the reason will
come afterwards; otherwise it would be nothing else
but a re-endorsement of a thesis which represents no
object for discussion as both the vaadin and the
prativaadin will acknowledge that whatever is not a
product is eternal. On the other hand, if the
predicate is denied (non non-eternal), as it is
exclude by the vaidharmya-d.r.s.taanta, it is
evident that the hetu also does not exist (non
non-product).

nitya'sabdenaatraanityatvasyaabhaava ucyate, ak.rta-
.sabdenaapi k.rtakat-vasyaabhaava

We may have many cases.
If the proposition is negative of existence, the
saadhana (reason and example) must be negative.
(a) aatman ( not accepted by Buddhists ) is non-
existent.
(b) Because it is not perceived,
(c) As the horns of a hare.
If the proposition is affirmative of existence,
the saadhana must be affirmative.
(a) Sound is non-eternal ( the negation is only of
a quality, not of the existence of sound).
(b) Because it is a product.
(c) Like a pot.

P.395

If the proposition is negative, the existence can-
not be taken as a saadhana, because the reason
cannot rest on the non-existent and the example in
that case could not have any object to establish.
Then in this case the existence must be taken only
as vaidharmya. As, for instance:⌒
" The aatman is non-existent " because " it is not
perceived"
" Whatever is not perceived is non-existence, as
the horns of the hare";
" Whatever is not perceived is non-existent, as
the horns of the hare " ;
" Whatever is existent is perceived, as a pot."
On the contrary,if a proposition is affirmative of
existences, existence only can be employed as saad-
hana.
Therefore negation and affirmation of a existence
must be considered as the vaidharmya of each other.
Moreover, the vaidharmya-d.r.s.taanta is only
exclusive ( 綛 膏 ) ( eternal is exclusive of non-
eternity) , while the saadharmya-d.r.s.taanta is
affirmative as well exclusive.

Pak.saabhaasa.h\tadyathaa : pratyak.saviruddha, an-
umaanaviruddha, aagamaviruddha, lokaviruddha,
svavacanaviruddha, aprasiddhavi'se.sa.na,aprasiddha-
vi'se.sya, aprasiddholbhaya, prasiddhasambandha

The first five only were formulated by Di^nnaaga;
the others were added by 'Sa^nkarasvaamin.
The pratyak.sa-viruddha can be of four kinds, (a)
contradictory to one's own evidence, but not to that
of another, as, for instance, if a Vai'se.sika
discussing with a Mahaayaanist says that
saamaanya-vi'se.sa and mahaa-bhuutas are
non-perceptible by the five organs of sens. This
thesis is not only (sva)-pratyak.sa-viruddha, but
also aagama-viruddha and (prasiddha)-sambandha, as
the Buddhist does admit the same theory.
(b) Contradictory to other's evidence, but not to
one's o〖e. For instance,if Buddhist discussing with
a Vai'se.sika argues that intellance, desire, joy,
etc., are not the object of the perception of the
aatman; this is also vi'se.sa.naaprasiddha.
(c) Contradictory to the perception of both the
vaadin and the prativaadin, as the example given in
the 'saastra.

p.396

(d) The consequence is that the proposition which
is not contradictory either to the vaadin or to the
prativaadin is the right one, as "sound is
non-eternal".
Other cases are possible "⌒
(a) Partially contradictory to one's own evidence,
not contradictory to the other's evidence. For
instance, if a Vai'se.sika says "all the four
material elements can be perceived by the eye". His
thesis is that vaayu, "wind" is atomic and cannot be
perceived by the eyes. Now he says "all the four
material elements", and this is a partial
contradiction of his theory. It is also
aagama-viruddha.
(b) Partially contradictory to the other's evidence,
not to one's own. For instance, if the Buddhist
discussing with a Vai'se.sika, maintains that
"earth, water, fire cannot be perceived by the eye".
The other, on the contrary, affirms that they can be
seen in the coarse state, but not in the atomic
condition.
(c) Partially contradictory to the theories of both.
For instance, if a Vai'se.sika discussing with a
Buddhist says that ruupa, gandha, rasa, cannot be
perceived by the eyes. On the contrary both schools
admit that ruupa can be perceived.
(d) Both theories are partially non-contradictory.
For instance, if a Buddhist discussing with a
Saa^nkhya, says: The prak.rti ( ┦ ) and the
aatman, maintained by you, are subject to change,
cannot be perceived, and non-eternity are
concerned.They disagree so far as the eternity and
non-eternity are con-cerned, but they agree so far
as the non-perceptibility of both principles is
concerned. Therefore among all these various
possibilities the proposition which is contradictory
to the thesis of the opponent, or the thesis which
is contradictory to neither is correct.

Lokaviruddho yathaa 'sa'sii na candro bhavaat ,
yathaa ca 'suci
nara'sira.hkapaala.m praa.nya^ngatvaac
cha^nkha'suktivad iti

( The first sentence is not in the Sanscrit text,
but it is in the Chinese translation and in T2. It
is to be found also into the Nyaaya-mukha.)

P.397

Suppose that somebody without any further deter
mination (we shall see later on the meaning of this
determination)says, "ka'sii na candra, because it
exists, as the sun, the stars and so on "; in this
case the reason and the example may be exact, but
the proposition is contradictory to the general and
common opinion. The same is to be said concerning
the other thesis, which is attributed to the
Kaapaalins ( 挡謦 )
If the proposition to be proved and the saadhana
are determinate or specified, then there is no
mistake. So in the svato 'numaana one will introduce
his thesis saying, "according to me," etc. In this
way the error "anyataraasiddha" is avoided. In the
anumaana for another's sake one must so specify:
"according to your theory." (E.g. the Buddhist will
say to a Saa^nkhya: "The aatman which is maintained
by you is non-existent" ). In this way the viruddha-
pak.sa and other errors are avoided. In the anumaana
for one's own and another's sake the specification
will be reciprocal as regards the special meaning in
which the various notions are used. In this way
errors such as "loka-viruddha", aagama-viruddha,
etc., will be avoided.

Aprasiddhavi'se.so yathaa Bauddhasya Saa^nkhya.m
prati vinaa'sii 'sabda iti

For the Saa^nkhya the notion of indestructibility
is well established. Eight cases are possible.
(a) Vi'se.sa.na aprasiddha for the vaadin himself,
not for the opponent; e.g.if a Saa^nkhya says "Sound,
etc., are modifications of the aalayavij~naana (旅醚
)". The aal. Vij. Is admitted by some Buddhist only,
but not by the Saa^nkhya.
(b) (b) The opponent's vi'se.sa.na is apprasiddha
for the opponent, not for the vaadin; example given
in the NP.
(c) Aprasiddha for both :if a Saa^nkhya discussing
with a Buddhist says "urrpa and so on are included
among the padaarthas". The theory of the padaarthas
is accepted neither by the Saa^n. nor by the Buddh.
(d) The vi'se.sa.na is partly aprasiddha for the
vaadin, but not

P.398

for the opponent; e.g.if a Sarvaastvaadin ( 履盋 )
discussing with a Mahaayaanist says that the
bhautikas ( ┮硑︹ ) are derived from the mahaa-
bhuutas and the aalaya-vij~naana. His school does
not admit the existence or the aalaya.
(e) The vi'se.sa.na is partly aprasiddha for the
opponent, not for the vaadin; e.g. a Buddh. Discuss-
ing with a Saa^nkhya says that organs of the sense,
such as the ear, etc., are subject to destruction
and to change. The Saa^nkhya cannot accept that they
are subject to destruction.
(f) The vi'se.sa.na is partly aprasiddha for both,
e.g. a Vai'se.sika discussing with a Buddh, says
that ruupa and so on are derived from
saamaanya-vi'se.sa and from svabhaava ( ┦ ) .
Neither the Vai's. nor the Buddh. accept the
svabhaava theory. The two remaining cases are not
quoted by K'uei-chi because it is evident that they
have no error: (g) Nonaprasiddha for both: (h)
non-partly-aprasiddha for both.

Aprasiddhavi'se.syo yatha Saa^nkhyasya Bauddha.m
prati cetanaa aalmeti

Here also we have eight cases :⌒
(a) The vi'se.sya is aprasiddha for the vaadin,but
not for the opponent, e.g. a Bauddha, discussing
with a Saa^nkhya, says "the aatman is non-eternal".
It is known that no Buddhist believes in an aatman;
the theory of the specification already referred to
must be remembered.
(b) Vi'se.sya aprasiddha for the opponent, not for
the vaadin. Example as in the NP.
(c) vi'se.sya apprasiddha for both ;e.g. a Sarvaa-
stivaadin discussing with a Mahaayaanist and arguing
that the aatman exists as a substance.
(d) vi'se.sya partly aprasiddha for the vaadin,
but not for the opponent; e.g. if the Buddh.,
discussing with the Saa^nkhya, says that the aatman
and ruupa are void, without any further
specification.
(e) vi'se.sya partly aprasiddha for the opponent,
but not for the vaadin; e.g. when a Sa^nkhya,
discussing with a Buddha.,

P.399

maintains that the aatman, ruupa, etc., exist as
substance. No Buddhist school will admit that, so
far as the aatman is concerned.
(f) vi'se.sya partly aprasiddha for both; e.g. if
a Sarvaastivaadin, discussing with a Mahi'saasaka (
て场 ),says that the past ( ㄓ ) and the aatman
exist. The existence of time is accepted by both,
but not that of the aatman.(1)
K'uei-chi expounds here under the item aprasiddho-
bhayo the main principles of the Vai'se.sika system,
quoting also from the Da'sa-padaarthii. See Ui,
Vai'se.sika System, p.181.

Prasiddhasambandho yathaa 'sraava.na.h 'sabda
iti
Here also we have eight cases :-
(a) sambandha existent for the opponent, but not
for the vaadin; e.g. if a Saa^nkhya. Discussing with
a Vai'se.sika, says that karman is destructible
(vinaa'sin).
(b) sambandha existent for the vaadin, but not for
the prativaadin; the reverse of the preceding
example.
(c) sambandha for both, as in the example of NP.
(d) sambandha non-existent for both, as if a
Saa^nkhya, discussing with a Buddhist, maintains
that karman is vinaa'sin.
(e) sambandha partly existent for the opponent,
but not for the vaadin; e.g. if a Sarvaastivaadin,
discussing with a Saa^nkhya, maintains that the
aatman and the manas exist as a reality
(dravyata.h). Neither of them thinks that manas
exists as a real substance; but the Saa^nkhya holds
the reality of aatman.
(f) Sambandha partly existent for the vaadin, but
not for the opponent; e.g. if a Sarvaastivaadin,
discussing with a Mahaayaanist, argues that aatman
and paramaa.nus exist as realities. The existence of
the aatman as a reality is accepted by neither of
them, but the Sarv. accepts the reality of the
paramaa.nus.

1 So the text, but it is known that according to the
Mahii;saasakas, the past as well as the future, do
not exist. Cf. Vasumitra's treatise on the sects
(transl. By Masuda, p. 59; Walleser, Die sekten des
alten Buddhismus, p.45) ; cf. Abhidharma-ko'sa
(transl. By La Vall'ee Poussin, v, 24ff.).

p.400

(g) Sambandha partly existent for both the vaadin
and the opponent; e.g. if a Sarvaastivaadin,
discussing with a Vai'se.sika, argues that svabhaava
and sound are both non-eternal. Svabhaava is
accepted by neither of them; but the non-eternity of
sound is maintained by both.
(h) Sambandha partly non-existent for both; e.g.if
the Sarvaastivaadin, discussing with a Mahaayaanist,
says that the aatman-bhava exists dravyatah.

E.saa.m vacanaani dharmasvaruupaniraakara.namukhena
pratipaadanaa- sambhavata.h saadhanavaiphalyata's
ceti pratij~naado.sa.h

The sentence dharma...mukhena is to be referred to
the five viruddha-pak.sado.sa; svaruupa is the
dharmin of the proposition. This sentence means that
mistakes of such a kind prevent the right knowledge
from being produced in the mind of the opponent and
the pari.sat. Pratipaadanaa-sambhavata.h is to be
referred to the three aprasiddha ; saadhana-
vaiphalyata.h to the prasiddhasambandha.

Asiddhaanaikaantikaviruddhaa hetvaabhaasaa.h
Asiddha-hetu may be understood as reason not proved
in itself (asiddha eva hetu) or as a reason which
cannot prove (na saadhayati iti asiddha-hetu). A
reason is anaikaantika when it is in the sapak.sa as
well as in the vipak.sa; it is called viruddha when
it is contradictory to the proposition, but proves
the vipak.sa.

Tatra 'sabdaanityatve saadhye caak.sutvaadiity
ubhayaasiddha
Four cases :⌒
(a) Reason absolutely existent, but aprasiddha for
both (example given in the NP.).
(b) Reason absolutely non-existent and aprasiddha
for both; e.g. if a 'Saabdika, discussing with a
Buddh, says that sound is eternal, because it is
included in the eternal categories padhaarthas. The
hetu is unreal, since neither of them believe in the
existence of the padaarthas.

P.401

(c) Reason partly existent, but aprasiddha for
both; e.g. when the 'Saabdika asserts that every
sound is eternal, because it is produced by some
effort. Both the vaadin and the opponent Know that
this reason is not valid so far as the eternity of
external sound is concerned (see before).
(d) Reason partly non-existent and aprasiddha for
both; e.g. if a 'Saabdhia, discussing with a
Buddhist, says that sound is eternal because it is
included in the padaarthas and can be perceived by
the car. The first part of the reason in asiddha for
both.

k.rtakatvaad iti 'sabdaabhivyaktivaadinam praty
anyataraasiddha.h
Eight cases :-
(a) Existent, anyataraasidha for the opponent; ex-
ample as that of the NP.
(b) Existent, anyataraasiddha for the vaadin ;e.g.
if one who believes in the manifestation of sound
('sabdaabhivyakti) , discussing with a Buddhist,
argues that sound is eternal because it is a
product.
(c) Non-existent, anyataraasiddha for the opponent;
e.g. if a Vai'se.sika, discussing with a 'Saabdhika,
says that sound is non-eternal, because it is
included in the gu.na-padaartha. The 'Saabdika does
not believe in the padaarthas.
(d) Non-existent, anyataraasiddha for the vaadin;
e.g. if a 'Saabdika, discussing with a Vai'se.sika,
says that sound is eternal because it is included in
the gu.na-padaartha.
(e) Existent, partly anyataraasiddha for the op-
ponent; e.g. if a Mahaayaanist, discussing with a
'Saabdika, argues that sound is non-eternal because
it is perceived by the five senses of the Buddha,
which, according to him, possess each the totality
of sense perception. The opponent will admit only
that sound is perceived by the ear.
(f) Esistent,partly anyataraasiddha for the vaadin;
e.g. if a 'Saabdika, discussing with a Mahaayaanist,
says that sound is eternal, for the reason already
referred to.
(g) Non-existent, partly anyataraasiddha for the
opponent;

P.402

e.g.if a Vai'se.sika, discussing with a 'Saabdika,
maintains that sound is non-eternal, as it is
included in the gu.na-padaartha, and is perceived by
the ear. Perceptibility by the ear is accepted by
both. The gu.na-padaartha is a tenet peculiar to the
Vai'se.sika.
(h) Non-existent, partly anyataraasiddha for the
vaadin; e.g. if a 'Saabdika, discussing with a
Vai'se.sika, argues that sound is eternal for the
reason already referred to.
In the anumaana for the sake of another, the abso-
lute anyataraasiddha is not an error, if there is
the specification already referred to. Here
K'uei-chi gives some examples of anyataraasiddha
which may be found in some Buddhist texts, specially
because in this connection he has the opportunity to
extoll the rare ability of his master Yuan Chuang as
a dialectician. One example of anyataraasiddha is
that which may be found in the third chapter of the
Mahaayaana-sampari-graha-'Saastra, where it is
stated that all Mahaayaana suutras are the word of
Buddha (proposition), because the whole of them do
not contradict the principle that the individuum
pudgala ( 干疭霉 )is not the aatman (reason), as
the Ekottaraa-gama, etc.( 糤) also state (example).
This proposition, if confronted with anohter's
pratij~na, is anyataraasiddha, because other school,
as the Hinyaanists, do not admit that the Mahaayaana
is non-contradictory to thetheory thay the pudgala
is not the aatman, as they think that according to
Mahaayaana there is an eternally existent substance
aatman-like. Even if we suppose that Mahaayaana is
not contradictory, anyhow the reason is
anaikaantika, because there are books of other
schools as the J~n~ana-prasthaana, etc., (1) which
do not ontradict that theory, thought it is known
that they are not the word of Buddha. Therefore, in
order to avoid they are not the word Buddha,
Therefore, in order to avoid these errors, the great
Buddhist doctor Jayasena (秤瓁 ),when he was about
forty years old, propounded the following inference:
All the Mahaayaana books are the word to Buddha

1 The J~n~ana-prasthaana is the fundamental work on
the Abhidharma for the Sarvaasti-vaadins.

P.403

(proposition); becasuse they are not included among
those books whick both the disputants equally do not
recognize as the words of Buddha ( ubhayasiddha-
abuddhavacana-amantarbhaavaat) (reason) , as the
Ekottaraagama (example).
By "ubhayasiddhaabuddhavacana" he meant the teach-
ing of the heretics and of the Sarvaastivaadins
(J~naana-prasthaana and sic paadas).
Nobody could succeed in refuting this syllogism.
But when Yuan Chunang met him, he raised the fol-
lowing objection :-
The J~naana-prasthaana, according to a school of
the Hiinayaana,viz. the Sarvaastivaadins, represents
the words of the Buddha, therefore it is impossible
to say the hetu: ubhayasiddha, etc. In fact this
would mean that you accept the theory of the
Sarvvastivaadins that the J~naana-prasthaana is
"Buddhavacana". But this is against the tenets of
the Mahaayaana.
Moreover the Hiinayaanist will argue that the
syllogism expounded in that way is anaikaantika, as
it is a mere supposition of the Mahaayaanist that
the J~naana-prasthaana is not included among the
words pronounced by the Buddha. Therefore Yuan
Chuang corrected the syllogism, expounded by
Jayasena, substituting the specification "accepted
by our own system" svaanuj~naataprasiddha-
abuddhavacana).
Tatra saadhaara.na.h 'sabda.h praneyatvaan nitya
iti. Tad dhi nityaanityapak.sayo.h saadhaara.natvaad
anaikaantikam ki.m gha.tavat prameyatvaad anitya.h
'sabda aahosvid aakaa'savat prameyatvaan nitya iti.
This principle is based upon the rule that a wide
reason only can prove a wide proposition, but not a
narrow or limited one, lest it shall be
anaikaantika. In fact we may have two propositions,
a wider one and a narrower. Example of the first:
"sound is anaatman" because even beside sound,
everything, according to the Buddhist, is
anaatman.Example of the second: " sound is
non-eternal, because there are thing that are
eternal." A reason also can be wide, e.g. "because
it is known" or "because it an be perceived", etc.
It is

P.404

evident, in fact, that there is no dharma which is
not knowable, etc. Or it can be narrow; e.g.
"because it is a product, etc." because there are
things which are not products. If I establish a
narrow pak.sa (e.g. "sound is non-eternal" or "it is
eternal") and I employ a wide reason (e.g. "because
it is known, etc.") then my reason is common both to
the sapak.sa and to the vipak.sa, and therefore it
is called anaikaantika ( ぃ﹚ ). A wide reason is
valid only for a wide proposition (e.g. the
prameyatva for th e anaatmataa). A narrow reason can
prove a wide as well as a narrow proposition (e.g.
the condition of being the consequence of an effort
can prove a wide proposition such as anaatmataa,
because, as everything is anaatman, there is no
existence of a vipak.sa and it can also prove a
narrow proposition, e.g. the non-eternity of sound).
Sapak.saikade'sav.rttir vi pak.savyaapi
yathaa, etc.
In order to understand this example we must remem-
ber what we already said concerning the two theories
about the utpatti-vaada and the abhivyakti-vaada of
sound. For the first school, sound comes into
existence from a previous non-existence; it is
therefore produced, although after its production it
lasts eternally; but it is not manifested as the
consequence of some effort. On the contrary, the
second school maintains that it exists ab aeterno,
but it is actually manifested as the consequence of
some effort. Now if a 'sabda-utpatti-vaadin,
discussing with a 'sabda-abhivyakti-vaadin, employs
the argument referred to in the book, his reason is
not only aprasiddha for both, but also anaikaantika.
The reason is explained in the text.
Anitya.h 'sabda.h k.rtakatvaat (Vai'se.sika against
'sabdaabhivyakti-vaadin; see before).
Nitya.h 'sabda.h 'sraava.natvaat 'sabdatvavat
('sabda-utpatti-vaadin against Vai'se.sika only). In
fact, if this argument is used against any other
school, the reason 'sraava.natvaat would be
asaadhaara.na-anaikaantika, because the other sects
do not accept that sound has a nature of its own.
But according to the

P.405

Vai'se.sikas, sound, although a gu.na, has a nature
of its own, and this nature is the
saamaanya-vi'se.sa (sound-ness). According to the
'sabda-utpatti school sound is threefold: (a) echo,
it can be perceived by the ears but it cannot
express anything; (b) 'sabdatva, each word has an
essence of its own, which exists even beyond the
verbal expression. When some conditions are given,
it is perceived. Therefore it is not like the
'sabdatvaof the Vai'se.sikas, which is the
saamaanya-vi'se.satva; (c) verbal expression.
Both theses, although different in many points,
agree in so far as they admit that the 'sabdatva is
eternal and can be perceived by the ear.
Ubhayo.h sa.m'sayahetutvaad dvaav apy etaav eko
'naikaantika.h samuditaav eva
The proposition of the Vai'se.sika, as well as the
proposition of the 'Saabdika, is unable to produce
an absolute knowledge in the mind of the opponent;
and therefore, like the first five errors already
referred to, they are anaikaantika. In either case
the reason is possessed of the three characteristics
that any hetu must have; but it is anaikaantika,
even in the case of the Vai'se.sika, who is supposed
to employ this argument ("it is a product)", when
discussing with a 'sabda-utpatti-vaadin (cf above).
As therefore both reasons are anaikaantika, it is
out of place to recall the rule that the thesis
which is wrong or to be refuted must be expounded
before, and the thesis supposed to be right
afterwards (in this case the right one is that of
the Vai'se.sika, who holds a theory similar to that
accepted by the Buddhists, for whom also sound is
non-eternal) ; so the objection raised by some
critics against the order followed in the NP.: (a)
Vai'se.sika, (b) 'Saabdika, instead of the other:
(a) 'Saa dika, (1) (b) Vai'se.sika, followed in the
Nyaaya-mukha is invalid.
Moreover the thesis of the Vai'se.sika is support-
ed by the

1 C.f my article Buddhist logic before Rimaaga,
J.R.A.S., 1929, p.451 ff.

P.406

force of the evidence and that of the aagama and
therefore it is manifestly the right one. Nor do we
need to establish the non-eternity of sound
following the tenets of an heretical school such as
the Vai'se.sika. The Buddha, as soon as he had the
intuition of the real essence of all dharmas, saw
that sound is non-eternal and expounded this
intuition in his teachings; therefore, just in order
to show the independenced of the Buddhist tenets
from those of the other schools, the Vai'se.sikas
are quoted here b efore. And again those who do not
accept Buddhism accept at least those notions which
are considered as true by common belief. Now it is
evident that sound is produced as soon as some
circumstances are given, and after a little while it
disappears; so it is clear that the theory of the
'Saabdikas is untenable.

Tatra dharmasvaruupavipariitasaadhano, yathaa nitya.h
'sabda.h k.rtakatvaat prayatnaanantariiyakatvaad
vety aya.m hetur vipak.sa eva bhaavaad viruddha.h.
The reasons given here age two, because the schools
of the 'Saabdikas are two:-

'Sabda-utpatti-vaadin: 'Sabda-abhivyakti-vaadin:
Sound is eternal; Sound is eternal;
because it is a procuct. becasue it is the
consequence of the some
effort
The sapak.sa is the ether. The sapak.sa is the ether.
The vipak.sa is the pot. The vipak.sa is the
The characteristic of lightning and the pot,
being a product is only etc.,; the reason is com-
in the vipak.sa but pletely absent in the
absolutely absent in the sapak.sa, existent in
sapak.sa(= fourth hetu some of the vipak.sas,
of the hetucakra). as "pot" non-existent
in some other vipak.sas
as the lighting (sixth
hetu of the hetu-cakra).

Dharmavi'se.savipariitasaadhano yathaa paraarthaas
cak.suraadaya.h Sa^nghaatatvaac
chayaasanaadya^ngavi'se.savad iti aya.m hetur yatha
Paaraarthya.m cak.suraadiinaa.m saadhayati tataa
sa.mhatatvam api Parasyaatmana.h saadhayaty
ubhayatraavya-bhicaaraat.
If a Saa^nkhya wishes to establish that the aatman
is the percepient, he cannot formulate the
proposition; "the aatman

P.407

perceives and employs the organs of sense as the
eye, etc., " because the vi'se.sya (aatman) is
anyataraasiddha; the hetu "sa^nghaatatvaat" then
would be ubhayaasiddha, because for the Buddhist the
aatman is non-existent, and therefore the reason is
out of place, while for the Saa^nkhya the aatman is
not sa.mhata. The example, on the other hand, cannot
prove the saadhya. Nor would the proposition, "eyes,
etc., are necessarily for the sake of the aatman,"
be right. In fact, the vi'se.sa.na, "for the sake of
t he aatman", is asiddha, and the hetu,
"sa^nghaatatvaat," is contradictory to the dharma's
svalak.sana. Therefore as a mere expedient they try
to formulate the syllogism in another way, "eyes,
etc., must be for the sake of something else, para."
This "para" is to be referred in their mind to the
real aatman, not to the empirical ego, which is
represented by the combination of the various senses
and sense-perceptions; but they cannot clearly
express this idea lest they fall into the errors
already referred to ; so they use the undefined
"para" which may equally be referred to the
sa^nghaata-aatman (empirical ego) and to the
asa^nghaata-aatman (real aatman).
Now the real aatman preceives the various objects
through the eyes, while the empirical ego does not
use them, as it is, in fact, nothing else but the
sum as it were of the various senses and senses'
experiences. On the contrary, bed, seats, etc.,
referred to in the example are of some use to the
empirical ego, but not so much to the real aatman.
Therefore the reason sa^nghaatatvaat can prove quite
well the svalak.sa.na of the dharma (paraartha.m);
but it is contradictory to the dharma-vi'se.sa (vi .
the asa^nghaata-para that the Saa^nkhya has in his
own mind).

Dharmisvaruupavipariitasaadhano yathaa na dravya.m
na karma Na gu.no bhaava.h, etc.

For the passage see Ui, vai'se.sika Philosophy,
pp. 6, 68, and 181.

P.408

Ubhayaasiddho dvividha.h sann asa^ms ca
As to the theory asserted in the NP., that incor-
poreity cannot be attributed to the ether, as it is
non-existent, the objection may be raised; "ether is
eternally non-existent, the objection may be raised;
"ether is eternally non-existent, therefore its
essence is abhaava, so that it is amuurta". This
objection is not right, because when I say "nitya.h
'sabdo 'muurtatvaat", both the proposition and the
reason are affirmative; the proposition affirms the
the eternity and the reason affirms the prop rty of
being amuurta. Now ether is only non-existence, and
therefore it cannot be amuurta.
Another objection is possible. The tathataa held
by Mahaayaanists eternally exists and therefore is
called eternal. In the same way the ether is
eternally non-existent; so that you must admit that
it is eternal. Moreover, if it is non-existent,
certainly it is amuurta. We reply that the
pak.sa-dharma is of two kinds; (a) it excludes only
( 綛 ), but it does not affirm ( 獶 ). If I say
"the aatman is non-existent", I mean to exclude the
existence of the aatman, not to assert its
non-existence. In th s case the example also must be
exclusive, not affirmative. (b) Exclusive and
affirmative. When I say, "the aatman is eternal,"
not only do I exclude non-eternity, but I affirm
also that it has an eternal nature. The example also
then can be exclusive as well as affirmative. So, if
in the first case I say "the aatman is non-existent,
as the horn of a hare", the example, based on the
simple exclusion, is right; but if I say, "sound is
eternal, like the ether" when I discuss with
somebody who does not accept the existence of the
ether, then the example would be only exclusive and
therefore invalid.
ananvayo yatra vinaanvayena,etc.
The passage refers to the ancient masters, who ac-
cepted other members of the syllogism, beside the
reason and the d.r.s.taanta.
Pratyaksa.m kalpanaapodha.m yaj j~naanam
artharuupaadau Naamajaatyaadikalpanaarahita.m tad
aksa.m aksa.m prati Vartate iti pratyaksa.m.

P.409

j~naanam is used here in order to exclude false
knowledge, as, for instance, that of a taimirika; in
fact, this also is naama-jaatyaadikalpanaa-rahitam,
but cannot be called pratyak.sa.
The " j~naana " ( タ醇 ) of the NP. Corresponds to
abhraanta ( 礚癵睹 ) of the Mahaayaanaa bhidharma-
sa.myukta-sa^ngiiti.(1)
The sentence naama-jaaty, etc.,is explained on the
basis of the passage of the Nyaaya-mukha quoted in
the Tattva-sa^ngraha.
The pratyak.sa is of four kinds, viz. (a) pa~nca-
vij~naana-kaaya ( き醚ō ); (b) mano-vij~naana (き
種 ); (c) aatma-sa.mvedana (靡 ); (d) yogi-j~naana
(﹚ ). The theory is already referred to in the
Nyaaya-mukha.

Tasmaad yad anumeye 'rthe j~naanam utpadyate 'gnir
atra, anitya.h 'sabda iti vaa tad anumaanam
The knowledge is the upalabdhi-hetu, while the
notions "here there is fire", "sound is
non-eternal", are the result which is known. As the
cause may be different, viz. either a direct
perception or an inference, therefore the result
also is twofold. In fact the notions "here there is
fire" and "sound is eternal" are different. I may
infer the existence of fire from smoke, and
therefore in this case the perception is the cause.
But when I assert that sound is non-eternal, this
notion is the conseque ce of my being aware of the
fact that it is a product; therefore it is derived
from an inference. Now the fact of being a product
or the smoke are the remote cause, while the memory
or the activity of mind is the near

1 The definition of the pratyak.sa alluded to is 
タ癵睹竡 which corresponds to an original svayam
pratiito 'bhraanto' rthah; but cf. Mahaavyutpatti
7621. K'uei-chi takes this definition from the 馒栋
Sa.myukta-sa^ngiiti-'saastra, that is to say,from
the 瑂笷縤馒栋阶 (cf. xvi) by Sthiramati 紌.
But this definition is to found already in the 
瑂笷縤馒栋 Sa^ngiiti-saastra by Asa^nga. This fact
is worth noting as it proves that Dharmakiirti when
completing the definition of pratyak.sa as given y
Di^nnaaga followed some traditions current before
him among the ancient Yogaacaaras.

P.410

cause. Through it I remember that, wherever there is
smoke, there is fire, and that objects like a pot,
etc., are a product and non-eternal. It produces
therefore those two particular notions as a result.
These three elements together represent the essence
of the anumaana.
Ubhayatra tad eva j~naana.m phalam
adhigamaruupatvaat
Some objections are raised against the definition
on the pramaa.na. First of all, they say, we know
that a foot and so on are the measure, cloth is the
object to be measured, the knowledge of the quantity
is the result. Now, in the example already quoted,
the sentences "fire is non-eternal", etc., represent
the notion to be known, the patyak.sa and the
anumaana are the instruments, through which we know.
Which is the result? The Sarvaastivaadins on the
other hand say: the objects (vi'sayas) are to be
known, the organs of sense are the instrument that
know, the citta and caitasika-dharmas represent the
result of the pramaa.nas. But according to the
Mahaayaana, they add, the knowledge is the
instrument. Which is then the result?
The heretics maintain that the objects are to be
known, the vij~naanas know, the aatman is the result
of the knowledge. But you, they say, do not believe
in the aatman; which therefore will be the result of
the pramaa.nas?
The reply is that the very knowledge is the result.
(1) the pramaa.nas know both the svalak.sa.na
(object of the direct perception) and the
saamaanya-lak.sa.na (object of the inference).As the
knowledge derived from the pramaa.nas is in fact
nothing else but a full realization of the form of
their respective objects, in mind itself it is
called a result. The form of the external objects (
 ) is only an appearance in the mind. Owing to the
difference of the function in one and the same
citta, we dist nguish a moment or part which is
active or knows and a part that is passive or it is
known.
Or we may say that the prameya ( that which is to
be known) is an image of the mind and therefore not
being distinct from

1 For this question cf. Bhaamatii ad 'Sa^nkara, on
Brahma-suutras, II,ii,28.

P.411

the mind, is called the pramaa.na. It is object and
mind at the same moment. Or also we may say that the
knowledge is composed by three elements: (a) the
vision that represents the active moment (秖ǎだ
).(b) The consciousness, sva-samvit ( 靡だ )which
represents the result of the apprehension; as any
essence is not distinct from its function (砰ぃ瞒ノ
) therefore knowledge is called the result. It
realizes the (c)form (  ) under which the first
moment, viz. the vision manifests itself. These
three moments of the vision are the pramaa.na. But
the form also is called "pramaa.na", because it
cannot be conceived as separated from the active
moment. According to Mahaayaana from the svasa.mvit
the vision-faculty arises which grasps the objects
while the aakaara is the vij~naana derived from an
object. Knowledge is therefore a synthesis of three
movements.
Saadhanado.so nyuunatva.m
Seven cases are possible, when we have recourse to
the trairuupya of the reason as it is established by
di^nnaaga. (a) Deficiency of one lak.sa.na; three
possibilities. (b) Deficiency of two lak.sa.nas;
three possibilities. (c) Deficiency of three
lak.sa.nas; one possibility.
I.(1) Saa^nkhya against 'Saabdika: " sound is non-
eternal, because it is seen by the eyes," deficiency
concerning the pak.sa.
(2) Saa^nkhya against 'Sarvaasti-vaadin : "sound
is eternal, because it is audible," deficiency
concerns the second lak.sa.na.
(3) id. " because it can be inferred ( prameyat-
vaat)," deficiency of the third lak.sa.na.
II.(1) 'Saabdika against Buddhist, " Sound is not
the consequence of some effort, because it is seen
by the eyes"; deficiency of the first two
lak.sa.nas.
(2) Deficiency of the first and third lak.sa.na.
"The aatman is eternal, because it is not the
consequence of some effort" (against a Buddhist).

P.412

(3) Deficiency of the second and third lak.sa.na,
as the four viruddha hetus.
III. 'Saabdika against the Vai'se.sika, " sound is
eternal, because it is perceived by the eyes."

II
THE EDITION OF THE TIBETAN TEXT OF THE NYAAYAPRAVE'SA

Professor Vidhu'sekhara Shaastrin of Vi'svabharatii
has edited in the Gaekwad's Oriental Series
(vol.xxxix) the Tibetan text of the Nyaaya-prave'sa
comparing both translations with the Sanscrit
original and the Chinese translation made by Yuan
Chuang.(1)
In this comparative study⌒the first of this kind
attempted by an Indian scholar ⌒ the various
readings have been carefully noted and discussed,
and useful indexes have been added at the end of the
volume, which will facilitate the reading of the
Tibetan text. The book testifies to the diligence
and accuracy of the author. It is only to be hoped
that its example will have many followers among the
young Indian scholars. I add here some further notes
to the text.
Comparative notes ad I, 9 : vipak.sa iti; Chinese
seems to follow Skt. Iti is = Ch. ; vei 孔 must
be taken as introducing the next sentences. It
corresponds usually to Skt. Yad uta.
2, 13; Ch. ミ﹙ corresponds quite well to Skt.
Pak.sa or pratij~naa-aabhaasa.
Ad 3,3-4,Ch.Corresponds to Skt.㎝絫 samavaayi-
kaara.nam(aatmaa).
T. 2 looks like a litteral translation from the
Chinese.
Ad 2, 4-5, the example shows that T. 1 is wrong
when it reads, in 2, 16-17, prasiddhi-viruddha
instead of prasiddha-sambandha. In fact,
prasiddhi-viruddha is the same as loka-viruddha.
That a prasiddha-sambandha is a pak.saabhaasa

1 The text is based only on the editions of the
bs'Tan aggur available in India, that is to say the
Narthang redaction.

P.413

is quite evident, because a thesis like this: "sound
can be heard" cannot be the object of any
discussion, as the relation between sound and
audibility is already proved.
Ad 3,5-7; Chin. Corresponds to Skt., pratipaadana
(Θ )asambhavaat( ぃ甧 ). The difference in T. *2 is
more apparent than real; in fact, it is evident that
the translator, being unable to guess the exact
equivalent of the Chinese characters, rendered the
Chinese text in this rather ambiguous way: sgrub mi
dgos pai p'yir da^n sgrub par nus kya^n.abras bu med
pai p'yir.
Ad 3, 9, T. 1, ga^n ya^n ru^n ba la ma grub
corresponds well to anyataraasiddha; I do not see
the necessity of supposing an original
yadyogyaasiddha.
Ad 4, 8-10, Chin, is not defective at all, but
corresponds literally to Skt.
Ad 5, 4-6, Chin,  笻 may correspond to Skt.
Viruddha as well as to Skt. Vipariita.
Ad 7, 16-17, I think that the hypothesis advanced
by the editor is out of place. I should be inclined
to read:..ses pa skyes pa adir me 'an (or da^n) bum
pa mi rtag ces pa.
So far as T.2 is concerned,I think that the doubts
about it arise from the fact that the Tibetan
translator could not understand the Ch.; de dag
rtags la brten nas don rnams dpog pa yin la is an
unclever translation of Ch.パ┘┮ゑ竡,like the
following: ya^n dag pai ^nes pai ses pa da^n idan,
in which Τ has been wrongly rendered as a
possessive: da^n ldan. Moreover the reading of the
xyl. Is defective; instead of da^n med I think that
we must read der me 'O.