The Vadavidhi
By Giuseppe Tucci
The Indian Historical Quarterly
Vol.IV, No.4, 1928. pp. 630-636
p. 630
I have briefly dealt with the Vadavidhi in my
note, "On the Fragments from Dinnaga" in the Journal
of the Royal Asiatic Society, 1928, p. 368) and in my
paper, "Buddhist Logic before Dinnaga," sent to the
Oriental Congress recently held at Oxford, and which
will be shortly published in that same Journal.(l) I
have expounded in these two studies the reasons why I
think that Vidyabhusana's views about the authorship
of the Vadavidhi cannot be maintained. On the other
hand, Prof. Keith, in a very interesting article
published in the IHQ,(2) supports the theory of
Vidyabhusana against the contention of Mr. R.
Iyengar. Prof. Keith is not convinced by the
arguments expounded by the latter and thinks,
therefore, that unless new documents are found, we
have no grounds for rejecting Vidyabhusana's views.
Since the solution of the question is rather
important for the history of Buddhist logic, I think
that it is worthwhile to examine it once more in
the light of all the available material.
One Vadavidhi, as it is known, is quoted by
Uddyotakara concerning the definition of the pratijna
sadhyabhidhanam pratijna (Benares Sanskrit Series,
new edition, p. 117) Vacaspati is here silent about
the authorship of the book.
In another place Uddyotakara says: apare tu
svaparapaksayoh siddhyasiddyartham vacanam vadah (p.
150).
Vacaspati comments(3) (p.317): "vasubandhavam
laksa-
__________________
1 In the course of this note, these two papers will
be abbreviated as follows: OFD and BLBD, so NV for
Nyayavarttika, NVTT for Nyayavarttikatatparyatika, PS
for Pramanasamuccaya and PSV for
Pramanasamuccayavrtti.
2 Vasubandhu and the Vadavidhi, IHQ, vol. iv, p. 221.
3 New edition, Benares Sanskrit Series 24 (Rajesvara
Sastri ed.)
p. 631
nam dusayitum upanyasyati." ( The new edition reads
vasubandhavam instead of the saubandhavam of the
first edition).
At p. 136 we read in the NV: "avayavatraya evam
laksanenopapadite tesam trayo durvihita." This
refers, according to NVTT, p. 198, to the criticism
advanced by Vasubandhu against the definition of
the prcatijna, drstanta and hetu, a given by
Aksapada: atra Vasubandhuna pratijnadayo trayo
'vayava durvihita Aksapadalaksanenety uktam."
At last at p. 117, we read: "yad api
Vadavidhanatikayam sadhayatiti sabdasya svayam parena
ca tulyatvat svayam iti visesanam.(1)
So in NV we find only once the mention of the
Vadavidhi and without the name of its author, That is
all we know so far as Sansktit sources are concerned.
Now to whom shall we attribute this Vadavidhi?
Vidyabhusana, whose theory Prof, Keith seems
inclined to accept, assumes(2) that it is a work of
Dharmakirti called Vadanyaya and that the Vadavidha-
natika, quoted by the same Uddyotakara, is the same
as the Vadanyayatika by Vinitadeva. So Uddyotakara
should be, if not posterior, at least contemporary
with vinitadeva. The identification of the Vadavidhi
with the Vadanyaya is based upon the definition of
the pratijna as given by Uddyotakara and which is
said to appear also in the Vadanyaya.
My first objection to such a theory is this: is
it not strange that, one and the same work is known
under three different titles, viz., Vadavidhi,
Vadavidhana, Vadanyaya? This also implies that
Uddyotakara quoted the same book under two different
names just in the same page (p. 117., ll. 1 and 20).
________________
1 On the pratyaksa, cf. Randle, Fragments, p. 10.
2 History of Indian Logic, p, 320; JRAS, 1914, pp.
601-606 and Introduction to a Bilingual Index of the
Nyayabindu (BI, 1917), pp, IX-X.
p. 632
Moreover, the work of Dharmakirti is preserved in
Tibetan, and, as I already noted, its title is not,
at all Vadavidhi, but Codananyaya(l) or Codyanyaya.
Let us now proceed to discuss the definition of the
pratijna, which has been the chief argument from
which Vidyabhusana deduced the identity of the
Vadavidhi with the Vadanyaya, viz., Codananyaya.
First of all, we have in NV, sadhyabhidhanam
pratijna, but in the Codananyaya we read: dam. bcah.
ba. yan. bsgrub. bya. bstan. pai. p'yir. ro. which
corresponds to pratijnapi sadhyabhidhanat. Of course
the similarity of the definition concerning the
pratijna cannot be avoided; : it is in fact always
either sadhyabhidhana or sadhyanirdesa. The diversity
of views is concerned only with pratijna as a member
of the syllogism.
The diversity is this; according to the
Nyayasutras the pratijna is one of the five members
of the syllogism; it, represents the probandum and it
must be considered as a fundamental part of it, which
cannot be disposed of. The Buddhist logicians, on the
contrary, elaborated the doctrine of the paksa. This
is, according to the Vadavidhi, as we shall see later
on, the argument proposed in the vicarana. Now
Uddyotakara objects to the consistency of the
definition of the pratijna, as given by Vasubandhu,
with the theory of the paksa held by the same author.
If the definition of pratijna is related to the
paksa, then it is useless to say: sadhyabhidhanam.
Sadhya being the paksa itself, the definition should
run thus: "tad abhidhanam pratijna. If, on the other
hand, it is maintained that the pratijna has no
relation with the paksa, then it would be the same as
that given by the Naiyayikas; so the mistakes that
Vasubandhu finds in this, should also be present in
his own.. The commentary of Kuei Chi on
____________
1 This form has been kindly suggested to me by Prof.
W. F, Thomas. But the commentary by Santiraksita
(mdo. tse.) has Vadanyaya.
p. 633
the Nyayapravesa(1) throws further light on this
point. In fact, he says that, according to the old
masters, the syllogism is composed of two different
parts, viz., sadhya and sadhana. The first is
represented by the subject and the predicate,
dharmin, and dharma that must be proved; the second
consists in the proposition (pratijna), reason (hetu)
and example (drstanta). This theory is, in fact,
expounded in the Abhidharmasangiti and the
Prakaranarya-vaca-sastra of Asanga.(2) It is
evident, from the fragments of the Vadavidhi that
this theory was accepted by the author of this book.
Sadhya is equal to paksavicaranayam isto rthah (NV,
p.106); sadhana is equal to pratijna, hetu, drstanta.
But for Dinnaga things are different; pratijna is
abolished and paksa, viz., sadhya is substituted for
it. So also for Dharmakirti.
I must add that the passage of the Codananyaya is
not at all a definition of the pratijna, but an
incidental proposition in the course of a discussion
on the Nigrahasthanas.(3)
Prof. Keith points out that the arguments of Mr.
Iyengar are very far from being convincing, in as
much as we cannot adduce any evidence that the
definition of the
----------------------
1. The commentary of Kuei Chi on the Nyayapravesa has
been studied by me in a paper to be published
shortly.
2. Cf. BLBD where the logical theories of these books
have been expounded.
3. As regards the definition of paksa, as given in the
Vidhanatika and which is supposed to occur also in
Vinitadeva's Commentary on the Codananyaya, I must
confess that the Tib. passage has been misunderstood
and wrongly translated by Vidyabhusana, Moreover,
the word svayam is essential in the definition of
paksa as given by Dinnaga in the Nyayamukha and by
Sankarasvamin in the Nyayapravesa. As to the passage
concerning the vada also we cannot find any exact
correspondence. The Tib. supposes vadiprativadibhyam
svaparartha (don)-siddhyasiddhyartham vacanam vadah.
We do not find here that literal correspondence that
should be expected [cf. NV, 108, where the same
definition occurs once more and where we find again
paksa (Tib. p'yogs) and not artha (don)].
p. 634
pratijna, as quoted in the NV and analogous to that
of the so-called Vadanyaya (viz., Condananyaya), was
in fact contained in the Vadavidhi. Fortunately, as I
have expounded in my paper on BLBD, we are in a
position to give this evidence. In fact, the
definition of the pratijna as given in the Vadavidhi
is expressly quoted by Dinnaga in his
Pramanasamuccaya- vrtti (III, p.45, a)(1) bsgrub.
byar brjod. pa. tsam. dam. bcah. ba.: pratijna
sadhyabhidhanam (or nirdesah) eva (or matram). He
adds that sadhya is here p'yogs, paksa; and paksa, he
says, must, be understood as rnam. pear. dpyad. pai.
adod. pai. don. that is: vicaranayam isto 'rthah,
quoted by the NV already referred to. This confirms
how the statements of Kuei Chi are exact.
That the definition of the pratyaksa and of the
anumana (the former is attributed to Vasubandhu by
Vacaspati), as quoted in the NY, are really found in
the Vadavidhi, as we read in the
Pramanasamuccayavrtti, has already been pointed out
by Mr. Iyengar.
I must add that, even the definition of the
drstanta as given in the Vadavidhi (from PSV) is not
unknown to the NV; 137, tayoh sambandhanidarsanam
drstantah: de. dag. gi. abrel. pa. nes. par. ston.
pa. ni. dpe, ste.(2)
For all these reasons I think that no doubt is
any longer possible as regards the identity of the
Vadavidhi, attributed to Vasubandhu and referred to
in Dinnaga's PS and PSV, with the Vadavidhi quoted by
the NV. The Codananyaya of Dharmakirti is out of
question. Now we shall consider the problem of the
Vadavidhanatika. This
__________________
1 From the copy of the bsTan agyur of the University
of Calcutta that has been very kindly put at my
disposal for some time, In the second translation of
the same work it occurs at fol. 127 b.
2 Chap, IV, fol. 70 b.
3 As I pointed out in OFD, p. 381, no allusion to the
theory of the Nigrahasthanas as maintained by
Dharmakirti is to be found in NV, but it occurs in
NVTT as well as in the Nyayamanjari of Jayanta.
p. 635
title presupposes two different works; a mula and a
commentary on it. In the name Vadavidhana, vidhana is
a synonym of vidhi. Now, as I said in my OFD,
according to the Chinese sources, we know that
Vasubandhu wrote at least three works on logic.
論式 Lun Shih
論軌 Lun Kuei
論心 Lun Sin
There is no doubt, that the first corresponds to
the Vadavidhi; the third is not Vadakausala as
proposed by Vidyabhusana, but Vadahrdaya; the second
can be restored quite well into Vadavidhana since
Shih and Kuei are synonyms (rule, law, system, etc.).
This may be supported by the fact that, as I have
shown in BLBD, the quotation from the Vadavidhana as
given in the NV, agrees with a passage that I found
in the Abhidharmasamyuktasangiti written by
Sthiramati, the greatest disciple of Vasubandhu, who
composed the commentary on the Vijnaptimatrata
recently published by Prof. Sylvain Levi.
I am sure, therefore, that there can hardly be
any doubt that the identification of the Vadavidhi
and the Vadavidhana with the Codananyaya is no longer
tenable.
But now the question arises, who was the author
of the Vadavidhi? According to Vacaspati, the
Vadavidhi is of Vasubandhu, because when Uddyotakara
quotes passages that we find in the commentary on the
Pramanasamuccaya, as taken from the Vadavidhi, he
uses the expression, Vasubandhavalaksana. The
uncertainty arising from the reading Saubandhava of
the first edition is now, as we saw, no more existent
The Chinese sources also agree with this attribution.
Shen T'ai and Kuei Chi tell us that the Vadavidhana
and the Vadavidhi are of Vasubandhu, and this state-
ment cannot so easily be disposed of, since
p. 636
they were informed by Yuan Chwang himself. The
commentary of Kuei Chi on the Nyayapravesa clearly
shows that he was perfectly conversant with the
Buddhist logic and that he knew the Pramanasamuccaya.
The only source at our disposal that seems to deny
the attribution of the Vadavidhi to Vasubandhu is the
Pramanasamuccaya of Dinnaga. This author thinks that
that work is not by the acarya as it was believed,
Did he say so because he could not accept the
theories held by Vasubandhu, whom he celled "the
master" and in fact was obliged to refute them in his
book? Or was there really a tradition that denied the
authorship of the Vadavidhi to Vasubandhu? It is
difficult to answer these questions. But the fact
remains that among the Naiyayikas as well as among
the Buddhists who informed Yuan Chwang, (1) the
Vadavidhi was attributed to Vasubandhu, and that this
Vadavidhi existed before Dinnaga.
I must add that this Vadavidhi has nothing to do
with the Fragment of the Tarkasastra which is
preserved in Chinese and which still knows a
-five-fold syllogism as the Naiyayikas; while
Vasubandhu, and the Vadavidhi, as me know from
Chinese sources, and Uddyotakara, expounded the
doctrine of a, three-fold syllogism.(2)
______________________
1 Jinendrabuddhi, in his Visalamalavati on
Pramanasamuccaya, I, 14, states also that the
attribution of the Vadavidhi to Vasubandhu is ajig.
rten, na. rab. tu. grags. pa. loke prasiddha, but as
it contains mistakes, it cannot be attributed to the
acarya.
2 This very important text has been translated into
Sanskrit by me, and will appear very shortly in the
Baroda Oriental Series.