Svabhdvapratibandha Revisited

Shoryu Katsura

The aim of this papery is to clarify the significance of two key-terms of

Dharmakirti's logic, viz. svabhavapratibandha and tadatmya, which have been

recently much discussed by young Japanese scholars. I have surveyed how the
concept of logical connection was conceived by Indian logicians as the foundation

of inference and proof and how and by whom it was identified as vyapti (per-
vasion). I came to the conclusion that there were four distinct stages in the de-
velopment of the concept of logical connection in India. Namely,
Stage I : As represented by Vaisesikasutra (ca. 2 & 3 c. A. D. ) and Sastitantra (ca.

300 A. D.), where several concrete relations (sambandha), such 'as

causal relation', between probans and probandum are regarded as the

foundation of inference. A similar view is found in Asan. ga (4 c. )2).

Stage II : By the time of Vasubandhu (ca. 400-480) an abstract relation between

probans and probandum, called inevitability' (nantariyakatva/avina

bhava), was propounded as the foundation of inference.
Stage III : It was probably Dignaga (ca. 480-530) who introduced the notion of
vyapti-pervasion of probans by probandum in the universe of dis-
course'-as the foundation of the above-mentioned abstract and in-
evitable relation. The notion of vyapti was gradually accepted by

most of Indian logicians, both Buddhist and non-Buddhist.

Stage IV : Finally Dharmakirti (ca. 600-660) proposed the concept of svabhava-

pratibandha as the universal basis for inference.

Dignaga never discussed how his concept of the logical connection of inevi-
tability or vyapti was related to reality and how it could have universal appli-
cability. Dharmakirti solved these problems by introducing the concept of sva-

bhavapratibandha.

For instance, when we infer the existence of a hidden fire from the percep-
tion of smoke, Dignaga justifies the inference by pointing to the inevitable rela-
-476-
(27) Svabhavapratibandha Revisited (S. Katsura)
tion between smoke and fire which is due to the fact that the domain of smoke
is pervaded by the domain of fire in the universe of discourse3). Dharmakirti,
on the other hand, justifies the inevitable relation between smoke and fire by his
svabhavapratibandha, i. e. a universal connection between smoke and fire4>,
which is supported by the fact that in reality smoke arises from fire.
Svabhavapratibandha provides the universal basis for inference, since it ex-
presses a universal connection between two general concepts, viz. hetu-svabhava
and sadhya-svabhava. Although Steinkellner explicitly states that svabhava in
this. compound can only have the ontological meaning, viz. 'essence'5), I would
rather take it to mean 'concept', the second meaning of svabhava used in logical
contexts as pointed out by Steinkellner himself. According to Dharmakirti's
ontological conviction that everything is momentary, a relation or connection is
possible not in reality but only in the conceptual universe because only concepts,
being understood as 'exclusion of others' (anyapoha), can have the nature of the
universals' (samanya) of other systems of Indian philosophy6). Thus, svabhava
'
in svabhavapratibandha primarily denotes 'universal' as exemplified by smoke-
ness or fire-ness conceptually constructed by anyapoha ; in other words, the
concept' of smoke or fire7).
The above argument does not preclude the possibility of the meaning of 'es-
sence' for svabhava in svabhavapratibandha. Precisely because of the double
meaning of the term svabhava, svabhavapratibandha can be understood to be
related to reality. Though it primarily indicates the universal connection be-
tween two concepts in the universe of discourse, it secondarily or metaphorically
indicates the essential connection between two realities, say smoke and fire.
The reason why the universal connection between smoke and fire, (namely,
wherever there is smoke, there is fire), is observed in the universe of discourse, is
because in reality the two entities are essentially connected in that smoke arises
from fire. Similarly we can successfully infer the existence of a tree from the ex-
istence of a Simsapa tree because in reality there is one and the same essence
which is called 'tree' as well as 'Simsapa'.
Thus, because of the double meaning of the term svabhava I propose to un-
derstand the compound svabhavapratibandha as having two meanings, viz.
-475-
 Svabhavapratibandha Revisited (S. Katsura) (28)
(1) a universal connection between hetu and sadhya in the universe of dis-
course, and (2) an essential connection between two items in actual reality. The
same is true of tadutpatti and tadatmya, Dharmakirti's two fundamental anal-
yses of the universe of discourse and the foundation of svabhavapratibandha.
They represent the logical relations of causation and class inclusion as well as
the actual states of affairs in reality.
As for tadatmya, Steinkellner takes it to mean 'real identity', i. e. "the fact
that one property is in reality the same as the other property. "8) This interpre-
tation, however, captures only one side of the term tadatmya. In logical con-
texts the term can only mean that one class is included in another, as e. g.
"Simsapa is a tree. " The term t
adatman, from which tadatmya is derived, as
well as its synonyms tatsvabhava- and tadbhava-, can be used. in the sentence
form "A B-atma, " meaning "A is B."9) Thus tadatmya in logical contexts means
not merely 'identity' but 'class inclusion'. In this sense tadatmya is more or less
synonymous with vyapti. It was perhaps Stcherbatsky who first translated ta-
datmya by 'identity'. He had already grasped the double meaning of tadatmya
in his celebrated Buddhist Logic vol. I, p. 554:
"Id
entity (tadatmya); the Buddhist law means reference of two different con-
cepts to one and the same point of reality ; the concepts are identical in that sense
that the one is included in the other;...
In the universe of discourse tadutpatti represents the relation "A arises from
B," while tadatmya means the relation "A is B". In reality the former holds
when A and B are different entities, while the latter holds when A and B are
one and the same entity.
Now, according to Buddhist tradition, everything is subject to causation. A
Simsapa tree can exist because of its preceeding casual factors. Thus, tadutpatti
is, in the final analysis, the principle which governs the real world of our expe-
rience as understood by Dharmakirti. Causal relation can be determined only by
identifying 'cause' and 'effect'. Thus, tadatmya seems to be the principle which
governsDharmakirti's universe of discourse or our world of concepts. There-
fore, I agree with Richard Hayes' suggestion that there is no fundamental differ-
ence between karyahetu and svabhavahetu in Dharmakirti's logic10).
-474-
(29) Svabhavapratibandha Revisited (S. Katsura)

Conclusion: Buddhist logic up to Dignaga can be characterized by its strong

inductive nature. It was given a deductive nature by Dharmakirti's introduction
of svabhavapratibandha into the Buddhist theory of inference. This becomes
clearer when we compare Dignaga's logical proof, which consists of [pratijna, J
hetu, and drstanta, with Dharmakirti's, which consists of vyapti and paksadhar-
mata with prati j na being understood.

1) This paper is a revised English version of Katsura [1986] The Origin and
Development of the Concept of Vyapti in Indian Logic from the Carakasamhita
up to Dharmakirti- (in Japanese) Hiroshima Daigaku Bungakubu Kiy® 45,
Tokushiigo 1, pp. 96-100. See this for abbreviations. A brief English summary
appeared in Tetsugaku 38, 186.
2) In the early vada texts, such as Carakasamhita and Nyayasutra, we find no trace
of the concept of a logical connection.
3) I once suggested that Dignaga's universe of discourse consists of hierarchy of
VaiAesika-like genuses and species. See Katsura [1979b].
4) I support Steinkellner's interpretation of the compound svabhavapratibandha, "a
connection by essence/essentially (svabhavena) of a proving essential property
(svabhavasya) with an essential property to be proven (svabhave sadhye), " though
I don't agree with his interpretation of the term svabhava in this compound. See
Steinkellner [1984] p. 471, fn. 40. Svabhavapratibandha is essentially an instrumen-
tal tatpurusa compound, and commentators' understanding of the compound as
either a genitive or a locative tatpurusa expresses only a part of the whole mean-
ing of the term.
5) See Steinkellner [1984] p. 459.
6) See Hattori [1977] The Sautrantika Background of the Apoha Theory, Buddhist
thought and Asian Civilization, ed. by Kawamura & Scott, California, p. 48.
7) Cf. Durvekamisra, Dharmottarapradipa, ed. by Malvania, p. 112: yasya dharma-
sya vyavrttikalpitasya yo niyatah pratiniyatah sa eva svabhavah/
8) See Steinkellner [1984] pp. 475-476.
9) As for the interpretation of the compound tadatman, I am inclined to take it as
a bahuvrihi against Steinkellner [1984] pp. 473-476. I have the impression that
Tibetan translators do not always translate a bahuvrihi compound ending with
-atman or -svabhava by its explicit equivalent like -bdag can or -rang bzin can.
10) See Hayes [1984] On the reinterpretation of Dharmakirti's svabhavahetu, a paper
read at the Sixth World Sanskrit Conference in Philadelphia.
(Assoc. Prof., Hiroshima University, Ph. D.)

-473-