Vasubandhu and the Vadavidhi
By A. Berriedale Keith
The Indian Historical Quarterly
Vol.IV, No.2, 1928.06, pp. 221-227
p. 221
In a well-known article(1) Dr. Satis Chandra
Vidyabhusana expressed the view that Uddyotakara, the
famous Nyaya commentator, knew the Vadanyaya of
Dharmakirti which he cited in his Nyayavarttika on i,
1, 33 as Vadavidhi. He 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 translation of the
vadanyayatika of Vinitadeva proves to contain
passages substantially identical with those cited by
Uddyotakara. The result of this argument is, of
course, important in that it establishes, when taken
in conjunction with the fact that Uddyotakara seems
to be known to Dharmakirti, the contemporaneity of
the two writers, who may be referred to in immediate
proximity in a pun of Subandhu's in his
Vasavadatta.(2)
To this view exception is taken in an interesting
articles(3) by H.R. Rangaswamy Iyengar, who contends
that the reference to the Vadavidhi is to a work by
the well-known Vasubandhu. He holds that the
difference of name, Vadavidhi and Vadanyaya, tells
strongly against the identification, but this can
hardly be deemed a conclusive ground; there are far
too many cases known in which works bear more than
one title, and apart from that, even in modern days
of libraries and easy access to titles, errors in
citation of books by name are not rare. Nor can we
say that the definition of pratijna cited from the
Vadavidhi "bears only a semblance of similarity to
that given by Dharmakirti in his Vadanyaya." As the
author himself proceeds to slow, a literal rendering
of the Tibetan gives us the words pratijna ca
sadhyabhidhanat which
----------------------
1. JRAS., 1914, PP. 601-6.
2. Keith, JRAS., 1914, pp. 1102 f.
3. JBORS, xii. 587-91.
p. 222
for all purposes is identical with the
sadhyabhidhanam pratijna of the citation from the
Vadavidhi. It appears, therefore, that the attempt to
disprove the reference to the Vadanyaya is
inadequate. Mr. Iyengar adduces as a further argument
the fact that on this identification of the
Vadavidhi, the Vadavidhanatika must be identified
with the work of Vinitadeva, which is objectionable
on the score of the late date of that author. He
seems not to know that Dr. Satis Chandra Vidyabhusana
definitely accepted this identification and supported
it by citation of the Tibetan renderings in the
Vadanyayatika; this is doubtless due to the fact that
this point is passed over in the History of Indian
Logic and is only set out in the article in the
Journal of the Royal Asiatic Society. As a matter of
fact, we are not in a position definitely to
determine the date of Vinitadeva from any external
evidence; that brought forward in the History of
Indian Logic(p.320) claims no special value, and
before the theory of the use of Vinitadeva by
Uddyotakara can be definitely disposed of, it will be
necessary to deal with the two citations which Dr.
Vidyabhusana claimed to identify. It must be added
that it can hardly be doubted that the Vadavidh
anatika must be intended to be a comment on the
Vadavidhi, a point which illustrates the fact that
slight variants of name without essential change of
sense may be taken for legitimate in Indian works.
Mr. Iyenger's own view would see in the Vadavidhi
a work of Vasbandhu; on the question of the
Vadavidhanatika he is silent, though obviously it is
not advisable to seek to .separate the two issues;
his position would have been enormously strengthened
had he been able to point to a commentary on the
Vadavidhi of Vasubandhu which was referred to by
Uddyotakara. Moreover it must be admitted that, as
the author very fairly points out, the Tibetan title
of the work which he has adduced would normally and
properly be rendered Vadasiddhi, which is by no means
the same as
p. 223
Vadavidhi, and that the rendering Vadavidhi rests on
the translation of the Chinese title Ronki of a work
of Vasubandhu as Vadavidhi. But, this apart, the
evidence on which the claim is made that Vasubandhu's
work is referred to is wholly indirect. Mr. Iyengar
has not adduced any definition of pratijna ascribed
to Vasubandhu's Vadavidhi (Vadasiddhi); on the
contrary, all that he can point out is that
Uddyotakara criticises two definitions of pratyaksa
and anumana, which are apparently the same as those
criticised by Dignaga in his Pramanasamuccaya (to
judge from the Tibetan version), and the comment on
that work ascribes them to the Vadavidhi, without
naming any author. On the other hand, Vacaspati in
the case of the definition of pratyaksa ascribes it
to Vasubandhu. The argument, therefore, is:
Uddyotakara cites a definition of pratijna from a
Vadavidhi; he deals with definitions of pratyaksa and
anumana, which appear apparently in the same shape in
the Tibetan version of the Pramanasamuccaya and by
its comment are ascribed to a Vadasiddhi or possibly
Vadavidhi, and one of these definitions is definitely
ascribed to Vasubandhu by Vacaspati; therefore "we
can safely conclude" that the Vadavidhi referred to
by Uddyotakara is the work of Vasubandhu. There are,
it is clear, far too many gaps in this reasoning, and
nothing convincing can be adduced unless and until
the question of the Vadavidhanatika is faced at the
same time.
When we consider that text difficulties do not
lessen. The theory of Dr.Vidyabhusana is clear, and
is supported by the passages which he cites from the
Vadanyayatika, which has definitions of paksa and of
vada corresponding to those ascribed absolutely
clearly in the first case, and with much probability
in the second, to the Vadavidhanatika. Other
authorities do not recognise that the second
reference is to that work, but unquestionably, as
will be shown below, that is a legitimate inference
from the discussion, and it is very greatly
strengthened by the coincidence of the occurrence
p. 224
of the passage in the Vadanyayatika. As regards the
first passage, there is an interesting suggestion by
Dr. Randle in his Fragments from Dinnaga that the
author may be Dignaga; unfortunately he does not
appear to know the article in the Journal of the
Royal Asiatic Society. The position would then be
that Vasubandhu wrote the Vadavidhi, and that
Dignaga corrected it in the Vadavidhanatika, and this
view would, of course, have the great advantage over
that of Mr. Iyengar in that it would solve the whole
problem and not leave it but half answered. It is,
therefore, desirable to submit this theory to a
critical investigation as far as our scanty sources
permit.
There is, of course, one strong objection to any
such view, viz., the absence of any evidence of the
writing by Dignaga of a commentary on the Vadavidhi
of Vasubandhu, assuming that Vasubandhu did write a
treatise of that name. This point is not absolutely
decisive, but it prevents us feeling any certainty
regarding the proposed explanation, even if other
matters did not tell against it. Nor is the rest of
the evidence satisfactory. Dr. Ganganath Jha(1) holds
that the Vadavidhi referred to by Uddyotakara is a
work of Subandhu, and he finds another reference to
it in the Nyayavarttika (p.157, line l7) by reading
there sastratvena ca vadavidhanam abhyupagamyate for
the vadabhidhanam of the recorded text. The amendment
may be tempting, but it is clearly illegitimate, for
as Dr. Randle (p.55, n.2) admits, the accepted text
makes sense, and, it may be added, the corruption
supposed has no obvious explanation, so that on any
sound principle of textual criticism this passage
must be ignored in this connection. The ground for
ascribing the Vadavidhi to Subandhu is thus gravely
impugned, for its justification rested(1) on the fact
that Vacaspati (p.218, line 9) ascribes the
definition of vada given by Uddyotakara (p.151, line
20) to Subandhu ( Saubandhavam laksanam) and (2) on
the amendment
---------------------
1. See his translation, i, 441 and 454 notes,
p. 225
of vadabhidhanam to vadavidhanam. If this conjecture
be laid aside, as it must properly be, the ascription
to Subandhu of the Vadavidhi is purely conjectural.
Morever, as against Dr. Ganganath Jha's view must be
set the fact that he conjectures(1) that the
definition of paksa given in the Nyayavarttika
(p.116, line 14) as yah sadhayitum istah is that of
Subandhu, But the Vadavidhanatika, as cited by
Uddyotakara, is absolutely clear in indicating that
the word svayam was contained in the definition which
it defended, and, therefore , it is most improbable
that Subandhu was the author of the Vadavidhi, if Dr.
Ganganath Jha's ascription of the definition without
svayam to him is correct. It must be added that there
seems no ground for the ascription.
But one fact emerges from this mass of
conjectures. Vacaspati definitely refers to Subandhu
the definition of vada cited anonymously in
Uddyotakara (p.151), but by Dr. Vidyabhusana ascribed
to the Vadavidhanatika and identified with a passage
in the Vadanyayatika in its Tibetan rendering. On
this identification one doubt presents itself, which
should be noted; Uddyotakara (p.124, line 9) has an
almost identical definition of vada, in which
svaparapaksasiddhyasiddhyartham in p.151 is replaced
by svaparapaksayoh siddhyasiddhyartham. It is just
possible that the Tibetan rendering could be made
applicable to this definition of p. 124 as opposed to
that of p.151, and it may be argued that the
occurrence of the phrase at p.124, in comparatively
close proximity to the citation of the
Vadavidhanatika at p. 120, is in favour of the view
that the citation at p. 124 rather than that at p.151
is from that work. However that may be, and it may be
presumed that Dr. Vidyabhusana decided against this
possiblity, though he does not mention the point,
there remains the question of who Subandhu was. Both
Dr. Ganganath Jha(2) and Dr. Vidyabhusana(3) identify
him
----------------------
1. i, 331.
2. i, 441; but compare i, 394.
3. History of Indian Logic; p.128.
p. 226
with Vasubandhu, the former, it is true, with some
hesitation. Dr. Randle(l) holds that the reasonable
supposition is that "in these passages (i.e. those in
which Saubandhavam laksanam is referred to)
Vacaspati abbreviates Vasubandhu to Subandhu, just as
he invariably speaks of Dharmakirti as Kirti". The
parallelism is not complete, for admittedly Vacaspati
(p.99, line 13) cites as Vasubandhavam pratyaksa
laksanam the definition of pratyakas given in the
Nyayavarttika (p.42, line 15), and the question,
therefore, arises why in certain cases he should use
an abbreviated form. But far more serious is the
character of the abbreviated form. That Kirti should
be used for Dharmakirti, or Hari for Bhartrhari, or
Simha for Vikramasimha or Gupta for Candragupta, and
so forth is obvious and natural, but that without
reason a man's name should be reduced from Vasubandhu
to Subandhu is extremely hard to credit, and indeed
may be dismissed as out of the question. It remains
only for those who hold this view to suggest not an
abbreviation but a variant name, and it is, to be
frank, extremely implausible to urge that the same
author should in the same work use the regular name
of a famous author, and also a name which never is
elsewhere applied to him, and which, it must be
added, is not equivalent in sense.
It must, accordingly, in my opinion be admitted
that the evidence is lamentably inadequate to
overthrow the view of Dr. Vidyabhusana.
Unquestionably on chronological grounds there is some
reason to doubt the use by Uddyotakara of Vinitadeva,
but, if the matter is to be established in any other
sense than that adopted by Dr. Vidyabhusana, new
evidence must be adduced, and this note has been
written in order to indicate the lacuna in the
existing evidence. It is hardly necessary to add that
no confirmation of the use of Subandhu for Vasubandhu
is to be found in Vamana's Kavyalamkarasutravrtti
(iii, 2, 2). If we wish to find Vasubandhu,
----------------------
1. Fragments from Dinnaga, p.26.
p. 227
there we have simply to alter the ca before Subandhu
into va, whereas no such correction is possible in
any of the passages where Vacaspati refers to
Subandhu(l) or Saubandhavam laksanam.
It may further be concluded that we have no
adequate evidence for the identification of Subandhu
with Vasubandhu, a result which is of importance
because, apart from this identification, Vasubandhu,
according to our present knowledge plays a much less
prominent part in the early history of Buddhist logic
than would be the case if we could securely(2) assign
to him the Vadavidhi, the definition of vada, and
pratijna, and assume that he was a predecessor of
Dignaga in his criticism of the views of proposition,
reason and exemplification given in the Nyayasutra.
All these things may be true, but for the time being
they are conjectures, which do not square with the
scanty evidence actually available. We do know his
definition of perception, and frankly it cannot be
said to reveal him as a profound logician. Dignaga
seems to record that he did not specialise in this
topic, though he may largely have inspired that acute
logician. Professor Stcherbatsky's suggestion(3) that
he may have adumbrated the doctrine of the affinity
of perception to inference, and so have evoked the
polemic of the Nyayasutra (ii, 1, 30), can hardly be
regarded as convincing; assuming that the passage in
question formed part of the original text, there is
nothing whatever to induce us to fix on the exact
form of the doctrine against which the Nyaya,
contended. There was unquestionably in ancient India
a vast activity of thought which is only hinted at in
the tantalising brevity and obscurity of the Sutras
of the philosophic schools, and me run serious risks
of misconstruing the facts if we seek unduly to
simplify the history of thought.
----------------------
1. p. 205, line 26
2. Fragments from Dinnaga, pp.27, 28.
3. La theorie de la connaissance et de la logique chex
les Bouddhistes tardifs, p.197, n.3. For Dignaga's
view, ibid., p. 2,