Vasubandhu and the Vadavidhi
By H.R. Rangaswamy Iyengar
The Indian historical Quarterly
Vol.V, No.I, 1929, pp.81-86
p. 81
In his recent article, Vasubandhu and the Vada-
vidhi.(1) Dr. Keith has been pleased to review my
position in detail and point out reasons why it
cannot be accepted as conclusive. In the course of
the article he has also examined the views of:Dr.
Ganganath Jha and Dr. H. N. Randle and found it
irresistable to conclude "that the evidence is
lamentably inadequate to overthrow the view of Dr.
Satisacandra Vidyabhusana." But on re-examination of
the question in the light of new facts, I am inclined
to believe that the theory of Dr. Vidyabhusan has not
a single fact for its support.
Uddyotakara in his Nyayavartika mentions two
treatises on logic, viz. the Vadavidhi and the
Vadavidhanatika(2) and also quotes the definition of
Pratijna found in the Vadanyaya,(3) now extant only
in Tibetan. Dr. Vidyabhusana, according to Dr. Keith,
has "reinforced this view" by holding that
Uddyotakara knew also the Vadavidhanatika which he
cites on I. 1. 33 and 41 in respect of the definition
of Paksa and Vada respectively, and that the Tibetan
version of the Vadanyayatika of Vinitadeva proves to
contain passages substantially identical with those
cited by Uddyotakara, and is therefore identical with
the Vadavidhanatika.(4) On the basis of these two
identifications the Doctor has further drawn the
conclusion that Uddyotakara and Dharmakirti are
contemporaries as is believed to have been referred
to by a pun of Subandhu in his Vasavadatta.(5)
In my previous article contributed to
JBORS(6) I took exception to the views of the Doctor
and pointed out, that the definition of Pratijna of
the Vadavidhi cited by Uddyotakara is similar to, but
not identical with that of the Vadanyaya of
Dharmakirti, and that it is not safe to establish on
bare similarity the contemporaneity of the authors of
the two works. In pleading for the view, Dr. Keith
takes substantial identity of the passages as
sufficient ground for identifying the texts and
explains away the difference in the title of the
texts as "errors in citations." But substantial
identity cannot
---------------------
1 IHQ, vol.IV, 2.
2 NV, I,i, 33 (Benares edition, p. 117).
3 dam bcah pa yin bsgrub bya bstaan pahi phyir ro
(Mdo, ce, fol. 399).
4 JRAS, 1914, pp.601-6.
5 JRAS, 1914, p.1102.
6 JBORS, xii, 587-91.
I.H.Q., MARCH, 1929
p. 82
by itself be taken as a safe ground for supporting
the identity of the texts. Any definition of Pratijna
has to be substantially identical with any other
definition of Pratijna in any treatise on logic.(1)
There ought to be literal identity as well. Both the
conditions are satisfied in the case of one of the
definitions of Pratijna quoted and criticised in the
Pramanasamuccaya of Dinnaga.(2) "Sadhyabhidhanam" is
the definition of Pratijna, which, according to
Dinnaga's comments in his Vrtti on the
Pramanasamuccaya, is alleged to be a fragment of the
Vadavidhi of Vasubandhu, (3) and is the very
definition cited by Uddyotakara as belonging to
Vadavidhi. In addition to this fragment we come
across many others, which have been noticed for
criticism by Uddyotakara in his Vartika like the
definitions of pratyaksa,(4) anumana,(5) paksa,(6)
hetu(7) and drstanta,(8) identified
----------------------
1 Cf., for instance, the definition, Sadhyani-
rdesah pratijna of the Nyayasutra with the Sadhya-
bhidhanam pratijna of the Vadavidhi.
2 PS (= Pramanasamuccaya), iii, 5: de bshin byed
brjod pa lahn.
3 PS-Vrtti:-de bshin te rigs pa can rnams la skyon
brjod pa de bshin du vtsod pa sgrub pa la yin=evam
naiyayikesu dosa uktah tatha Vadavidhav [api].
NV, p.118: yo naiyayikapratijnayam dosa uktah sa iha
prasaktah.
4 PSV, I, 15:don de las skyes rnam pa ces pa mnon
sum yin shes bya ba.
Cf.NV, 40: Apare punar varnayanti tato'rthad vijnanam
pratyaksam iti.
5 PSV, ii, 74:rtsod pa sgrub pa nas ni med na mi
hbyun bahi don mthon ba de rig pa rjes su dpag paho
shes brjod do
Cf, NV, 54: Apare tu bruvate nantariyakarthadarsanam
tadvido'numanam iti.
6 PSV : rtsod pa bsgrub par ni bsgrub bya brjod pa
tsam dam bcah ba ma yin gyi hon kyan phyogs kyi
chos bsgrub byaho phyogs gan yin pa rnam pa dbye
par hdod pahi don phyogs yin te //
Cf. NV, p.115 (NV 106)--evam vicaranayam isto'rthah
paksa ity atrapi istagrahanam anarthakam.
7 PSV,III, 36: re shig rtsod pa bsgrub par ni de
Ita bahi med na mi hbyun bahi chos ne bar bstan pa
ni gtan tshigs so snes bya ba/
Cf. NV, p.55: tadrg avinabhavi dharmopadarsanam hetur
ity anye.
8 rtsod pa sgrub pa nas de dag hbrel ba nes par
ston ni dpe ste bum pa bshin no shes brjod pa Ita
baho/
Cf. NV, p. 137: etena tayoh sambandhanidarsanam
drstanta iti pratyuktam.
p. 83
as fragments of the Vadavidhi of Vasubandhu by
Dinnaga in his Vrtti on the Pramanasamuccaya and
some of his identifications are corroborated by
Vacaspati in his tika on the Vartika of Uddyotakara.
This indicates that Uddyotakara had really access to
and made use of the Vadavidhi of Vasubandhu. Explicit
references, then, to the Vadavidhi in the Vartika of
Uddyotakara can only be to the Vadavidhi of
Vasubandhu and never to the Vadanyaya of Dharmakirti.
It is highly improbable and strange that one who had
access to the Vadavidhi would cite Vadanyaya as
Vadavidhi.
It may, however, be argued that, though Vadavidhi
and Vadanyaya are different from each other,
Vadavidhi may itself be a work of Dharmakirti. But
there is hardly any evidence to support it. Vadavidhi
is nowhere mentioned as a work of Dharmakirti. We
learn, on the other hand, from the Chinese sources,
that it is there known as Ronki and ascribed to
Vasubandhu. Dinnaga, a disciple of Vasubandhu (this
we have shown more than once) regards it as a work of
Vasubandhu. If it were a work of Dharmakirti, why
should Vacaspati, who is familiar with the works of
both Dharmakirti and Vasubandhu, ascribe the
fragments cited by Dinnaga as belonging to Vadavidhi,
to Vasubandhu and never to Dharmakirti? (1)
No doubt the question of the Vadavidhanatika
still remains unsettled. Dr. Keith complains that I
am silent on this question and believes that my
position would he strengthened by "facing the problem
at the same time". But it may be stated that the
object of my paper was to dispel the illusion created
by Dr. Vidyabhusana regarding the identity of the two
texts, the Vadavidhi and the Vadanyaya, and to
establish the authorship of Vadavidhi by Vasubandhu.
The reference: to the Vadavidhanatika by Uddyotakara
was only used as an argument against the Doctor's
conclusions, Now that Dr. Keith attaches much
importance to the problem it is indispensible to
investigate it in detail.
In criticising Dinnaga's definition of paksa(2)
Uddyotakara extracts a passage from the
Vadavidhanatika which is clearly an improved
definition of paksa(3) on those of an anonymous
writer and of a Bhadanta. The author of the Vartika
refers to the improvement on the definition
---------------------
1 NVTT, I.1,4 (p.99): tadevam pratyaksalaksanam
samarthya Vasubandhavam tavat pratyaksalaksanam
vikalpayitum upanyasyati.
2 NV, p.116.
3 NV, p.117: sadhayatiti sabdasya svayam parena ca
tulyatvat svayam iti visesanam.
p. 84
by the addition of svayarn as visesana, and proceeds
to refute the explanation and finally falls back
upon the original definition without the suggested
improvement. A few lines above Uddyotakara quotes and
criticises in his Vartika certain passages(1) of the
Vadavidhi which is a work of Vasubandhu. Next appears
the definition of paksa as cribed to 'apara'. It is
unquestionably a legitimate conclusion from the
discussions that the Vartikakara is here referring to
the definition of Vasubandhu and is further supported
by Vacaspati in his comments on the passage.(2) But
the passage in question has not been ascribed to
Vadavidhi either by Uddyotakara or by Dinnaga. It is
not therefore clear what relation the
Vadavidhanatika bears to the Vadavidhi referred to by
Uddyotakara. But this much is certain that it has
nothing to do with the Vadanyayatika of Vinitadeva.
The identification of the two texts, the Vadavi-
dhanatika and the Vadanyayatika, is beset with great
difficulties. There is at the outset the
insurmountable chronological difficulty which Dr.
Keith has not lost sight of. If the two texts are
one, not only Dharmakirti, but even Vinitadeva,
author of the Vadanyayatika, would become a
contemporary of Uddyotakara This is against all
facts and accepted conclusions. Prof. Tucci(3) has
pointed out why Dharmakirti cannot be regarded as a
contemporary of Uddyotakara. It, therefore, goes
without saying that Vinitadeva cannot be a
contemporary of Uddyotakara. The passage(4) in the
Vadanyayatika, which is
---------------------
1 NV, p.115-116.
2 NVTT, p.273 (Benares edition) : pakso yas sadhayi-
tum ista ity atrapi Vasubandhulaksane.
The Vijayanagaram edition of the text, however,
reads Subandhu for Vasubandhu. Evidently ca Subandhu
is either the printer's or the scribe's error for
Vasubandhu; the more so, because ca would be
superfluous after api. It is this simple error that
has given occasion for various explanations by Dr.
Ganganath Jha and Prof. Randle. I agree with Dr.
Keith that the ascription of the Vadavidhi to
Subandhu is purely conjectural and that there is no
justification in taking Subandhu either as a variant
or as an abhreviation of Vasubandhu as Kirti is of
Dharmakirti.
3 JRAS, April, 1928, pp, 377ff.
4 bdag nid ma yin pahi no bo ni bdag nid kyi no bo ni
gshan gyi no bo ma yin no shes bya bahi don to,
Mdo. Ze, fol.50 (according to Dr. Vidyabhusana).
p. 85
taken to be substantially identical with the one
cited by Uddyotakara from the Vadavidhanatika, may be
restored into Sanskrit as "Anatmarupam atmarupam
pararupam na bhavatity arthah," and it is, on the
face of it, absurd to identify it with the citation,
"sadhayatiti sabdasya svayam parena ca tulyatvat
svayam iti visesnam."(1)
The definition of vada from the Vadanyayatika,
(2) however, if restored into Sanskrit would read as
Vadiprativadinoh svapararthasid-dhyasiddharthavadanad
(or vacanad) vadah' and is, as Dr. Vidyabhusana
holds, substantially identical with the definition of
vada,(3) cited by Uddyotakara at I, ii, I. But it has
been clearly ascribed to Vasubandhu by Vacaspati(4)
and nowhere is it stated that it belongs to the Vada
vidhanatika. It would indeed be a too bold argument
if one were to identify the two texts,
Vadavidhanatika and Vadanyayatika, on the basis of
substantial identity of a passage and establish the
contem poraneity of their authors. The only
explanation that would be offered for substantial
identity is that the definition of vada of Vasu-
bandhu became a stock definition and appeared in the
same form in later works on vada.
The pun of Subandhu in his Vasavadatta, 'Nyayata-
ttvam iva Uddyotakarasvarupam bauddhasangitim iva
salankaram', which is taken by Dr. Vidyabhusana in
support of his contention, appears in a different
garb in some versions(5) the text, and therefore
gives the impression that the passage is
interpolated. Even if it be an integral part of the
text, it can only be taken to refer to two works
whose authors need not necessarily be contemporaries.
---------------------
1 NV, p.117.
2 rgol ba dan phyir rgol ba dag gis (Read gi) ran
dan gshan gyi don grub par byed pa dan (ma) grub
pahidon du brjod nas rtsod pa yin no// Mdo, Ze,
fol, 41 (according to Dr, Vidyabhusana).
3 Apare tu svaparapaksayoh siddhyasiddhyartham
vacanam vadah-- NV, P.150 (cf, NV, p. 121).
4 NVTT, p.317: tadevam svabhimatam vadalaksanam vya-
khyaya Vasubandhavam laksanam dusayitum upanyasyati.
Here also the Vijayanagaram edition reads: sauban-
dhavam for Vasubandhavam (p.218). Cf, also I-I-37
(P.207), and Benares edition, p. 298
5 Satkavikavyaracanam ivalankaraprasadhitam. See S.K.
De: Sanskrit Poetics, vol. I, p.20.
p. 86
The pun has been variously interpreted by scho-
lars. Prof. Levi believes that alankara in the
passage does not refer to any work of Dharmakirti.(1)
Prof. Luiders, on the authority of the discoveries in
Central Asia, contends that it alludes to the
Kalpanamanditika of Kumaralata, otherwise known
through Chinese translations as the Alankarasastra of
Asvaghosa.(2) The passage may, however, be taken to
allude either to the Sutralankara of Asvaghosa or to
the Mahayana Sutralankara of Asanga, the brother of
Vasubandhu.
From the foregoing discussion, it is clear that
the position of Vidyabhusana is untenable. The
Vadavidhi is not the Vadanyaya of Dharmakirti, but a
work of Vasubandhu and that on no ground can the
contemporaneity of Uddyotakara and Dhamakirti be
established.
----------------------
1 Bulletin de l'Ecole d'Extreme Orient, 1903, p.18.
2 Keith, History of Sanskrit Literature, Preface, viii.