J Indian Philos (2011) 39:553–569

DOI 10.1007/s10781-011-9135-y

_
Dharmakı̄rti’s Criticism of Anityatva in the Sānkhya
Theory
Toshikazu Watanabe

Published online: 1 June 2011

Ó Springer Science+Business Media B.V. 2011

Abstract In his Pramān: aviniścaya 3, Dharmakı̄rti criticizes the view of the

_
Sānkhyas that the word anityatva (‘‘impermanence’’) means a process of transfor-
mation (parin: āma) of primordial matter (pradhāna). In this connection, he deals
with the following two explanations of transformation: (1) the disappearance
(tirodhāna) of the previous dharma of an entity (dharmin/dravya) and (2) the
cessation (nivr: tti) of the previous state (avasthā) of an entity (avasthātr: ). In
response to these explanations, he proves that whenever a transformation takes
place, the previous entity is destroyed, and therefore, impermanence does not mean
transformation, but only destruction (vināśa). His criticism is basically along the
same lines as Vasubandhu’s arguments found in the Abhidharmakośabhās: ya.
However, because of developments in the theory of transformation, Vasubandhu’s
_
criticism allows room for a retort from the Sānkhya. For this reason, Dharmakı̄rti
augments Vasubandhu’s theory in order to make it sustainable against the more
_
developed Sānkhya theory.

Keywords Dharmakı̄rti Sānkhya

_ Yuktidı̄pikā Vasubandhu Anityatva
Parin: āma Tirodhāna Avasthā

Introduction

_
Between the Sānkhya satkāryavādin and the Buddhist ks: an: ikavādin—the most
fervent asatkāryavādin—there is a fundamental disagreement about how to explain
the causal process. The former does not accept there to be production (utpatti) and
destruction (vināśa) in the process, even though he advocates the anityatva

T. Watanabe (&)
Institute for the Cultural and Intellectual History of Asia, Austrian Academy of Science,
Apostelgasse 23, 1040 Vienna, Austria
e-mail: [email protected]

123
554 T. Watanabe

(‘‘impermanence’’) of all things except the soul (purus: a) and primordial matter
(pradhāna/prakr: ti). He explains anityatva as the process of transformation (par-
in: āma) of pradhāna that continues to exist without any intrinsic changes.1 In
contrast, the latter argues that whenever any change or transformation takes place,
the previous entity is destroyed and a new entity is produced. He therefore considers
the word anityatva to mean destruction in the strict sense, i.e., destruction without
leaving any trace (niranvaya-nivr: tti, -vināśa).
Dharmakı̄rti (ca. 600–660), in his Pramān: avārttika 2.17, criticizes the anityatva
_
put forward by the Sānkhyas as a logical reason in the following argument attributed
to their school:
[Thesis] The intellect (buddhi) does not have consciousness.
[Reason] For it is impermanent.
[Example] Just as the case of a form, etc. (acetanā buddhih: , anityatvāt,
rūpādivat.)2
In the third chapter of his Pramān: aviniścaya (on k. 68; D215b2ff., P313a6ff.),
Dharmakı̄rti classifies this logical reason as ‘‘unproved’’ (asiddha) and discusses his
_
justification of this in detail. There, two different Sānkhya views on the meaning of
impermanence are dealt with. One explains the meaning of impermanence as the
disappearance (tirodhāna) of the previous property (dharma) of an entity. The other
explains it as the cessation (nivr: tti) of the previous state (avasthā) of an entity. In
_
this paper, I would like to expound on Dharmakı̄rti’s criticism of these two Sānkhya
theories and to show that his discussion is based on, but also reinforces, Vas-
_
ubandhu’s criticism of the Sānkhya theory of transformation (parin: āma).

Two Explanations of parin: āma and Vasubandhu’s Criticism

In his Abhidharmakośabhās: ya, Vasubandhu (ca. 350–430)3 refutes the Sānkhya _

notion of transformation in order to distinguish it from that of the Buddhists, that is,
1
See YD 121,20–122,2 on SK 9: ... kriyata utpadyate jāyata ity evamādir lokasya vyavahārah: pravartate
/ ... paramārthatas tu na kasyacid utpādo ’sti na vināśah: /; *Mahāvibhās: ā T. 1545 997a12–13:
; AKBh 301,1–3: vārs: agan: yavādaś caivam : dyotito bhavati—yad asty asty
eva tat, yan nāsti nāsty eva tat, asato nāsti sambhavah: , sato nāsti vināśa iti /; YSBh 186,1 on YS 4.12:
nāsty asatah: sambhavah: , na cāsti sato vināśa iti ... See Imanishi (1968, p. 642) and Yamashita (1994,
pp. 47, 57–58).
2
At the beginning of PVin 3, Dharmakı̄rti has already mentioned that the production (utpatti) or
impermanence of buddhi cannot be accepted by the Sānkhyas._ The reason for this is described in PVin 3
on k. 68. PVin 3 on k. 1cd: acetanāh: sukhādayo buddhir vā, utpatter anityatvād vā, rūpādivat.
([D187b2–3, P285b1–2]: dper na blo dang bde ba la sogs pa ni sems pa med pa yin te / skye ba can nam
mi rtag [D: rtags P] pa yin pa’i phyir gzugs la sogs pa bzhin no zhes bya ba’o zhes zer ba.) Cf. NB 3.60.
My thanks are due to Dr. Pascale Hugon for providing me the Sanskrit text of PVin 3, of which an edition
is under preparation.
_
According to the Sānkhya tenets, buddhi is impermanent and does not have consciousness. SK 10–11:
hetumad anityam avyāpi sakriyam anekam āśritam _
: lingam / sāvayavam: paratantram : vyaktam : viparı̄tam
avyaktam // trigun: am aviveki vis: ayah: sāmānyam acetanam prasavadharmi / vyaktam : tathā pradhānam :
tadviparı̄tas tathā ca pumān // See Iwata (1995, pp. 158–159, 173–174) and Inami (1995, pp. 45, 51, n.1).
The proof criticized here by Dharmakı̄rti is not, however, found in the extant treatises of the Sānkhya_
3
On the dates of Vasubandhu, see Deleanu (2006, pp. 186–194).

123
Dharmakı̄rti’s Criticism of Anityatva 555

in order to show that whenever transformation occurs, the destruction of the

previous entity and the production of a new entity must take place. He says:
AKBh159,18-22: katham _
: ca sānkhyānām: parin: āmah: / avasthitasya dravyasya
dharmāntaranivr: ttau dharmāntaraprādurbhāva iti / kaś cātra dos: ah: / sa eva
hi dharmı̄ na sam : vidyate yasyāvasthitasya dharmān: ām : parin: āmah: kalpyeta /
kaś caivam āha—dharmebhyo ‘nyo dharmı̄ti / tasyaiva tu dravyasyāny-
athı̄bhāvamātram: parin: āmah: / evam apy ayuktam / kim atrāyuktam / tad eva
cedam : na cedam: tathety apūrvais: ā vācoyuktih: (em.: vāyo yuktih: ed.) /

_
[Question:] But what is ‘‘transformation’’ for the Sānkhyas?
[Vasubandhu:] (1) [They explain it as follows:] The appearance of one property
(dharma) in an entity (dravya) that remains in existence when
another property ceases to exist.
_
[Sānkhya:] But what is the fault in this [definition]?
[Vasubandhu:] For, there can be no such property-possessor (dharmin) that
remains in existence while the transformation of its dharmas is
assumed.
_
[Sānkhya:] But who says that dharmin is different from its dharmas? Instead,
(2) transformation means the same entity becomes different (any-
athı̄bhāvamātra).
[Vasubandhu:] In this case, too, it is not correct.
_
[Sānkhya:] What is incorrect in this [definition]?
[Vasubandhu:] It is an unprecedented way of speaking to say that this is that, but
[at the same time that] this is not so.4
Here Vasubandhu considers two explanations of transformation attributed to the
_
Sānkhya: (1) In the first explanation, an entity is considered to consist of dharma
and its possessor dravya/dharmin. While a dravya/dharmin continues to exist
through the process of transformation, its dharmas cease to exist (nivr: tti). (2) In the
second explanation, on the contrary, the distinction between dharma and dravya/
dharmin is abandoned. It is asserted that through the process of transformation an
entity takes on a different form of existence, but remains the same.
The first explanation is refuted by Vasubandhu as follows: It is not possible to
assume that dharmin remains in existence when its dharmas cease to exist because,
from an ontological point of view, it is not acceptable, not only for the Buddhists but
5
_
also for most of the Sānkhyas to distinguish between dharma and dravya/dharmin.
Vasubandhu denies the second explanation by pointing out the contradiction
between the sameness and the difference of one and the same entity. However,

4
On Yaśomitra’s interpretation of this passage, see Bronkhorst (1997).
5
_
Most Sānkhyas do not accept there to be an ontological difference between dharma and dharmin/
dravya. There was, however, at least one adherent of the Sānkhya _ who did accept it, viz. Mādhava
_
(ca. 500), who was called a ‘‘destroyer of Sānkhya’’ _
(sānkhyanāśaka). See VNT: 52, 27–28: ... pūrvakān
_
kāpilān atipatya sānkhyanāśakamādhavavad _
(em.: sānkhyānām: śakamādhavavat ed.) dravyasya
vyatireke ’pi dharmakāran: atvam is: yate... On Mādhava, see Frauwallner (1953, pp. 407–408), Solomon
(1974, pp. 153–163) and Halbfass (1992, pp. 57–58).

123
556 T. Watanabe

_
in consideration of the development of the theory of transformation in the Sānkhya,
Vasubandhu’s criticisms are seen as somewhat problematic by Dharmakı̄rti.

_
Development of the Theory of Parin: āma in the Sānkhya and Problems
of Vasubandhu’s Criticism

Disappearance of Dharma

With regard to the first explanation of transformation mentioned above, in several

treatises almost the same description can be seen, as follows:6
parin: āmaś cāvasthitasya dravyasya pūrvadharmanivr: ttau dharmāntaraprav-
:rttir iti / (NBh 183,1-2 on NS 3.2.15)
avasthitasya dravyasya pūrvadharmanivr: ttau dharmāntarotpattih: parin: āma
iti / (YBh 132, 4-5 on YS 3.1.13)
parin: āmo (em.: parimān: o ed.) hi nāmāvasthitasya dravyasya dharmāntara-
nivr: ttih: dharmāntarapravr: ttiś ca / (YD 111,15-16)
Moreover, the author of the Yuktidı̄pikā (ca. 680–720)7 cites the following verse,
which seems to be taken as an authoritative explanation of transformation in the
_
Sānkhya.
[TEXT 1] jahad dharmāntaram : pūrvam upādatte yadā param /
tattvād apracyuto dharmı̄ parin: āmah: sa ucyate // (YD 111,21-22; 163,12-13)8

When the property-possessor (dharmin), without abandoning its essence

(tattvād apracyuta), relinquishes an earlier property (dharma) and subse-
quently takes on another one, it is called transformation.
In these statements, transformation is explained with the terms dharma and its
possessor dharmin or dravya. Of these, the word dharmin/dravya refers, ultimately,
to the three constituents (trigun: a) of pradhāna, i.e., sattva, rajas and tamas, which
do not undergo any intrinsic changes during the process of transformation.9 It is

6
Also in the Buddhist treatises, which were, however, strongly influenced by Vasubandhu, similar
descriptions can be seen. AD 106, 10–12: sām : khyasya tv avasthitasya dharmin: ah: svātmabhūtasya
dharmāntarasyotsargah: * svātmabhūtasya cotpādah: parin: āma iti / (*Yamashita (1994, p. 59, n.54) reads
‘‘dharmamātrasya’’ instead of ‘‘dharmāntarasya,’’ but this emendation is not needed.); VN 13,11–14:
avasthitasya dravyasya dharmāntaranivr: ttir dharmāntaraprādurbhāvaś ca parin: āmah: /; TSP 30,14–15
on TS 16: vyavasthitasya dharmin: o dharmāntaranivr: ttyā dharmāntaraprādurbhāvah: parin: āmo varn: yate.
7
See YD (introduction, pp. xxvii–xxviii). As Bronkhorst (2003) suggests, however, it is possible to
assume the author of the YD to be slightly earlier.
8
Prajñākaragupta refers to a similar explanation of transformation. PVA 297,28–29 on PV 3.230:
parin: āmaś ca tattvād apracyutasya dharmāntaraparityāgo ’paras parotpattih: /.
9
See YD 164,16–30. This part is translated in Malinar (1999, pp. 630–631). Cf. Frauwallner (1953,
pp. 390–391). For the explanation of dharma and dharmin given in the YSBh, see Chakrabarty (1951,
pp. 198–206).

123
Dharmakı̄rti’s Criticism of Anityatva 557

only their arrangement that changes through the process.10 Due to this change in
their arrangement, the entity, which consists of these three constituents, gives up a
previous dharma (property), for example a particular color, etc., and takes on a
different one. It seems that before the time of the YD this kind of explanation
of transformation was generally accepted. pffi
However, these contain a problem that is caused by the expression hā or
‘‘nivr: tti’’ (cessation) of dharma, because this means a kind of annihilation of
_
existence not only for the Buddhists but also most of the Sānkhyas, as they do not
accept there to be an ontological difference between dharmin and dharma. In order
to avoid this difficulty, the author of the YD interprets the above-mentioned verse
([TEXT 1]) in the following manner:
[TEXT 2] YD 163,14–16: yadā śaktyantarānugrahāt pūrvadharmam :
tirobhāvya svarūpād apracyuto dharmı̄ dharmāntaren: āvirbhavati tad
11
avasthānam asmākam: parin: āma ity ucyate /

When the property-possessor, after its previous property has disappeared

(tirobhāvya) due to its receiving (anugraha) another power without aban-
doning its nature (svarūpa), appears with another property, then we call such a
situation transformation.
pffi pffi
By relying on the notion pofffi disappearance (tiras dhā, tiras bhū and their
derivatives) instead of using hā or the word ‘‘nivr: tti,’’ it can be said that in the
process of transformation no annihilation or destruction takes place.12
This replacing the notion of destruction with that of disappearance can be traced
back to the explanation of the cosmic cycle presented by the followers of

10
The view that differences between cause and effect are due to the different arrangement of the three
constituents can be traced back to the following fragment of the S: as: :titantra. S: T (Frauwallner (1958,
p. 125)): ādhyātmikāh: kāryātmakā bhedāh: śabdasparśarasarūpagandhāh: pañca trayān: ām : sukhad-
uh: khamohānām : sanniveśamātram*. (*Of the four sources of Frauwallner’s reconstruction of the S: T, the
expression ‘‘sanniveśamātram’’ is found in NĀA 12,18. However, the other three suggest the reading
‘‘sanniveśaviśes: āh: ’’. See PST: (D194a3–4, P220b3): ... sgra dang reg pa dang gzugs dang ro dang dri ste
lnga po rnams ni / bde ba dang sdug bsngal dang rmongs pa rnams te gsum po rnams kyi bkod pa’i khyad
par ro /; PST: (D196b6–7, P223b3) ’o na sgra dang reg bya dang gzugs dang ro dang dri ste lnga po
rnams ni bde ba dang sdug bsngal dang rmongs pa ste gsum po rnams kyi bkod pa’i khyad par te /; NĀA
314,8–9: sāmānyapūrvakān: ām : ca bhedānām ityādy ekajātisamanvayapradarśanārthasukhāditrigun: ai-
kajātisamanvayam : kāryātmakānām : tatsanniveśaviśes: atvam: paks: ı̄kr: tya ... And also see NC 265,6–266,2:
yāni tair ārabdhāni śarı̄rādı̄ny ādhyātmikāni bhūtādı̄ni vaikārikārabdhāni cendriyān: i trayān: ām : sukh-
aduh: khamohānām : sanniveśaviśes: āh: sukhādimayā eva, tathā pr: thivyādayas tanmayakāran: ārabdhatvāt /).
The author of the YD seems to follow this Vārs: agan: ya’s view. YD 109,13–19: na hi nah: kāran: ād
arthāntarabhūtam : kāryam utpadyata ity abhyupagamah: / kim : tarhi / viśvātmakānām : sattvarajastamasām
apagataviśes: āh: sanmātralaks: an: opacayāh: pratinivr: ttaparin: āmavyāpārāh: paramavibhāgam upas-
amprāptāh: sūks: māh: śaktayah: / tāsām adhikārasāmarthyād upajātaparin: āmavyāpārān: ām : sanmātrānuk-
ramen: a pracayam upasampadyamānānām : sanniveśaviśes: amātram : vyaktam /
11
Almost the same explanation of parin: āma as in the Sānkhya’s _ reply to the asatkāryavādin is found in
YD 121,4–6 on SK 9: sādhanānugr: hı̄tasya dharmin: o dharmāntarasyāvirbhāvah: pūrvasya ca tirobhāvah:
parin: āmah: / na cāvirbhāvatirobhāvāv utpattinirodhau /
12
The development of the explanation of transformation found in the YD is pointed out by Frauwallner
(1953, pp. 389–391) and Muroya (1996, pp. 49–50).

123
558 T. Watanabe

Vārs: agan: ya (Vārs: agan: āh: ).13 In Sānkhyakārikā

_ 10, it is stated that all things except
purus: a and pradhāna are impermanent.14 In his commentary on this kārikā,
replying to the objection that as long as satkāryavāda is held to be true, the
impermanence of things cannot be justified,15 the author of the YD sets forth his
view by citing the followers of Vārs: agan: ya as follows:
[TEXT 3] YD 128,20-129,2 on SK 10: kāran: ānām : tu yah: parasparasam: sargāt
sam : sthānaviśes: aparigrahah : , tasya virodhiśaktyantarāvirbhāvād vyaktis
tirodhı̄yata ity etad vināśaśabdena vivaks: itam / tathā ca vārs: agan: āh:
pat:hanti—tad etat trailokyam : vyakter apaiti, na sattvāt/apetam apy asti
vināśapratis: edhāt/ sam
: sargāc cāsya sauks: myam
: sauks: myāc cānupalabdhih: /
tasmād vyaktyapagamo vināśah: / ...16

However, the following is intended by the word vināśa: The manifestation

(vyakti) of [the effect] that takes a special arrangement through the mutual
connection of its causes disappears (tirodhı̄yate) upon the appearance of an-
other incompatible power. In the same way, the followers of Vārs: agan: ya
(Vārs: agan: āh: ) say that the entire threefold world withdraws (apaiti) [only]
from its manifestation, but not from its reality (sattva). Even when [its
manifestation] has withdrawn, [it still] exists because destruction is negated
[by us]. And because of its merger (sam : sarga) [with prakr: ti], [the world] is
subtle; and because of [its] subtleness, it cannot be perceived. Therefore,
destruction is the disappearance (apagama) of manifestation.
According to the followers of Vārs: agan: ya, destruction (vināśa) means the dis-
appearance of its manifestation but not the cessation of its existence. An object
evolves out of its cause and then dissolves into its cause after having been manifest
for some time. Even though it disappears from manifestation, it still exists in its
cause. It is highly plausible that the author of the YD derives the notion of disap-
pearance from this statement of the followers of Vārs: agan: ya and then applies it to
the interpretation of TEXT 1 in order to avoid the problem caused by the expression

13
On the name of Vārs: agan: ya and Vārs: agan: āh: , see Chakravarti (1951, pp. 135–142), Wezler (1985a,
p. 14, n. 6) and Wezler (1992, p. 288).
14
See above footnote 2.
15
YD 128,14: āha: anityatvānupapattih: satkāryavādābhyupagamāt /
16
A slightly different version of this passage, but without the name, is found in the NBh as an example of
siddhāntaviruddha and in the YSBh, in contrast, as siddhānta. NBh 43,11–13 on NS 1.2.6: yathā so ’yam :
vikāro vyakter apaiti nityatvapratis: edhāt /... apeto ’pi vikāro ’sti vināśapratis: edhāt /; YSBh 127,8–128,2
on YS 3.13: tad etat trailokyam : vyakter apaiti nityatvapratis: edhāt / apetam apy asti vināśapratis: edhāt /
sam: sargāc cāsya sauks: myam, sauks: myāc cānupalabdhir iti / See Halbfass (1992, pp. 59; 66, n. 46).

123
Dharmakı̄rti’s Criticism of Anityatva 559

pffi
hā or ‘‘nivr: tti’’ (cessation) of dharma.17 Thanks to this notion of disappearance,
the explanation of transformation based on the division between dharma and
dharmin is strengthened. If this new explanation is taken into consideration, Vas-
ubandhu’s criticism might be refuted.18

Difference of State

From the second explanation of transformation found in the AKBh, what is known
_
is only that this Sānkhya does not accept the difference between dharma and
dharmin. But it is not clear how he explains the process of transformation. As long
as this point is not made clear, Vasubandhu’s criticism cannot be regarded as a
decisive objection. Of course, the possibility is not denied that at the time of
Vasubandhu the explanation of transformation without the distinction between
_
dharma and dharmin had not yet been fully established in the Sānkhya. In any case,
_
by the time of Dharmakı̄rti, one Sānkhya group explains the transformation by using

17
Before YD and AKBh, the notion of disappearance (and appearance), in connection with the Sānkhya _
theory, was already mentioned in the *Mahāvibhās: ā (T.1545 996c14ff.). There, just after referring to the
theory of transformation ( ), —the theory which states that there is neither destruction nor
production of things, but only appearance and disappearance take place—is presented. A similar idea is
mentioned in the Yogācārabhūmi as abhivyaktivāda. The author of YBhū regards both hetupahasadvāda,
i.e. satkāryavāda, and abhivyaktivāda as the theory propounded by Vārs: agan: ya. See Imanishi
(1968, pp. 642–643). For the text and translation of the relevant part of the YBhū, see Wezler (1985a,
pp. 10–12).
Furthermore, in another part of the *Mahāvibhās: ā (T.1545,1003c18–1004a2) the theory that the
process of transformation consists of appearance and disappearance but not of production and destruction
is criticized by two Buddhist teachers, i.e., Vasumitra ( ) and some Bhadanta ( ). However, except
for the explanation of transformation criticized by Bhadanta, the explanations of transformation criticized
by the author of YBhū and Vasumitra do not presuppose the analysis of an entity into property (dharma)
and property-possessor (dharmin). In contrast, in the explanation of transformation criticized by Bha-
danta, an entity is considered to consist of the entity itself, its characteristic ( /*laks: an: a?), and state
( /*avasthā?). There, disappearance or appearance is regarded as a state of an entity. (I will discuss
Bhadanta’s criticism later in note 19.) Considering this, it seems slightly curious that Vasubandhu does
not mention the notion of disappearance in his criticism, despite his familiarity with the *Mahāvibhās: ā.
18
As Frauwallner (1953, p. 390) indicates, the author of the YD seems to think that the explanation of
transformation by the disappearance of dharma is not enough to give an answer to Vasubandhu’s
criticism. He replies to the criticism with the help of the notion of the whole (avayavin) and its parts
(avayava). See YD 164,1–5: ucyate—na, senādivad vyavasthānopapatteh: / tad yathā senāngebhyo _
’nanyatvam _
: senāyāh: / na ca senāngavināśe senāvināśah: / tathā tantubhyo nānyah: pat:ah: bauddhānām
:
sam: yogāvayavavipratis: edhāt / na ca pat:avināśe tantuvināśah: / tatra yad uktam : dharmavināśe dharmi-
vināśa ity etad ayuktam / Considering his refutation of the existence of the whole (YD 112,12–115,12),
the author of the YD might be of the opinion that dharma, as in the case of the whole, is not a real entity
but just a fictional construction. On the criticism of the existence of the whole in the YD, see Motegi
(1994).

123
560 T. Watanabe

the term ‘‘state’’ (avasthā) instead of dharma.19 This Sānkhya

_ group considers the
term ‘‘state,’’ unlike the term dharma, to describe a more transitory and provisional
condition of an entity and therefore the change of the state, or, even the destruction
of the state, does not affect the continuing existence of that entity. With the help of
_
this term, the Sānkhyas are able to maintain that one and the same entity, which
again consists of the three constituents (trigun: a), becomes different by changing its
_
state in the process of transformation. Thus, this Sānkhya theory could counter
Vasubandhu’s criticism.

Dharmakrti’s Criticism

_
In response to these developments of the theory of transformation in the Sānkhya
school, Dharmakı̄rti, by refuting two crucial notions that had appeared, tries to
support Vasbandhu’s criticism of transformation.20

19
In the YD, causal relationships are explained by the change of state. YD 62,9–10 on SK 3: tasmin
pratyākhyāte gun: ānām evāvasthāntarāpeks: ah: kāryakāran: abhāvah: /; YD 150,1–3 on SK 15: ihāsmākam
:
kāryakāran: ayor arthāntaratvānabhyupagamād gun: ānām avasthāntaram evāvasthāntarāpeks: am :
kāryakāran: aśabdavācyatām
: labhate / See also NV 334,12–335,7 on NS 3.1.5, where the Sānkhya _
explains changes in buddhi by changes in its state.
As mentioned in the above note 17, *Mahāvibhās: ā introduces Bhadanta’s criticism of the theory of
transformation. *Mahāvibhās: ā T.1545 1003c25–1004a2:

‘‘[Moreover,] Bhadanta says [as follows]: (1) It is obviously observed in the world
that when their causes get together, constituent factors ( /*dharma) occur and that when their causes are
set apart, constituent factors come to an end at that time. There is no such difference in the case of what
disappears and appears. Therefore, it is understood that transformation does not consist of disappearance
and appearance and that there are production and destruction of the entity itself [in the process of
transformation]. (2) Furthermore, if, when a constituent factor transforms, the characteristic ( /
*laks: an: a?) of a previous [constituent factor] is different from that of the next one, then the [previous
constituent factor] itself ( ), too, must be different [from the next one] because the characteristic is
identical with the [constituent factor] itself. If a constituent factor were permanent, then, even though its
state ( /*avasthā?) such as being appearance or being disappearance differs, its characteristic would
not differ. Therefore, it is understood that there are production and destruction of the [constituent factor]
itself [in the process of transformation].’’
In Bhadanta’s second criticism, a constituent factor ( ), i.e., an entity is considered to consist of the
entity itself ( ), characteristic ( ), and state ( ). According to Bhadanta, disappearance or appear-
ance of an entity is regarded as the difference of its state. He thinks, perhaps, that only the change of the
characteristic means the transformation of an entity but not the change of the state. As a result, he accepts
_
the difference of state. Therefore, his criticism is not crucial for the Sānkhya who explains transformation
by the difference of state.
20
Also in his later work Vādanyāya (VN 13,9–15,12), Dharmakı̄rti criticizes two explanations of
transformation in the context of the criticism of satkāryavāda. Compared with PVin 3, the structure of his
criticism in the VN is much closer to that of Vasubandhu. When he criticizes the explanation, in which an
entity is considered to consist of dharma and dharmin, he mentions almost the same definition of
transformation found in the AKBh (see above note 6). In the VN, however, the notion of disappearance
of dharma is not examined. Moreover, even though both the VN and PVin 3 deal with the theory of
transformation explained with the notion of state, the manner of the criticism in the VN is different from
that of in PVin 3.

123
Dharmakı̄rti’s Criticism of Anityatva 561

In his PVin 3, the logical reason anityatva (‘‘impermanence’’) put forward by the
_
Sānkhya for the proposition that the intellect (buddhi) does not have consciousness
(acetana) is classified as a fallacious reason, i.e., unproved (asiddha). Explaining
_
this, Dharmakı̄rti criticizes two Sānkhya theories: The former explains the imper-
manence by the disappearance of a property (dharma) of an entity and the latter
explains it by the difference of the state (avasthā) of an entity. Those who proclaim
these theories are called, according to Dharmottara, abhivyaktivādin (or vyak-
tivādin) and avasthāntaravādin (or avasthāvādin), respectively.21

Criticism of the Notion of Disappearance

_
In criticising these Sānkhya theories, Dharmakı̄rti presupposes that the logical
reason put forward by the proponent must be accepted by both the proponent and the
opponent,22 in other words, that the logical reason must provide a common
understanding of a given state of affairs to both sides of the debate. From this point,
Dharmakı̄rti begins an examination of the meaning of ‘‘impermanence’’ for the
23
_
Sānkhya.
tatrāpi hi śabda eva kevalah: siddhah: , nārthah: / na hi pare ’pracyutātmana
upalayanam anityatām icchanti/24

21
See PVinT: (Ms114a5–7; D125b3–5; P147a5–8): sānkhyah _ : kaścid abhivyaktivādı̄—śak-
tirūpen: āvasthitah: sarvo ’rtho vyajyata iti. anyas tv avasthāntaravādı̄—sarvātmanā hi sarve bhāvāh:
pradhānaprasevake vyavasthitā na dr: śyante. pratyayavaśāt tu prasevakān nis: kāsitā iva dr: śyā bhavanti.
tato na vyaktir nivartate śaktiś cāvatis: thate, api tv avasthāntaraprāptih: kevalā bhavatı̄ti. ‘‘One sort of
_
Sānkhya is an abhivyaktivādin [and he is of the following opinion:] Everything that remains in existence
in the form of a power is manifested. However, the other [Sānkhya], _ i.e., the avasthāntaravādin [is of the
following opinion]: All entities, indeed, are settled in the bag of pradhāna entirely, and they are not
visible. But they become visible by the force of certain causal conditions, being, as it were, turned out of
the bag. Therefore, [in the process of transformation] there is neither the cessation of the manifestation [of
phenomena], nor remaining power, instead, [the entity] merely reaches another state.’’
As Dharmottara’s subsequent remarks suggest, it seems that there must be a close relationship between
this avasthāntaravādin and Vasumitra’s theory, i.e., avasthānyathika mentioned in AKBh, etc. as the
most authentic doctrine of the Vaibhās: ikas, in order to explain the difference of the present dharma from
the past and future dharma. But a discussion of this is beyond the scope of this paper. See PVinT:
(Ms114a8; D125b6; P147b1–2): etasmim : s tu nirākr: te bhavatu vaibhās: ikapaks: anirākaran: am, na tu tam
evoddiśyāyam _
: granthah: pravr: ttah: . tasyāpi sānkhyasya ...
22
See PVin 3 on k.67: tasmāt pratipādyapratipādakayor aprasiddhasandigdhadharmisambandhā-
nvaya-vyatirekā dharmā hetvābhāsāh: / ([D215b2, P313a5–6]: de’i phyir chos can dang ’brel pa dang rjes
su ’gro ba dang ldog pa dang [D: om. (dang) P] bsgrub par bya ba dang sgrub par byed pa dag la ma
grub pa’am the tshom za ba’i chos ni gtan tshigs ltar snang ba yin no //)
23
Uddyotakara also criticizes the Sānkhyas’_ notion of disappearance (tirobhāva). According to him, the
existence of the effect (e.g. a piece of cloth) is distinct from that of the cause (e.g. the threads it is made
of) because they appear differently, i.e., what causes the difference must be produced. See NV 460,7–12:
apares: ām : dharmāntarāvirbhāvatirobhāvāv iti / ... ye ’pi sam : sthānaviśes: en: āvasthitān tantūn pat:a iti
varn: ayanti, tān prati sādhanam, prāg upalabdhikālāt sam : sthānaviśes: aśūnyās tantavah: tatkāran: atvāt
turyādivat / etena kāryātmanāvatis: :thanta iti vyākhyātam / tathā dharmāntarāvirbhāvatirobhāvāv iti /
24
PVin 3 (D215b7–216a1, P313b4): der yang sgra tsam ’ba’ zhig grub kyi don ni ma yin no // nyams pa
med pa’i bdag nyid nye bar zhi ba ni mi rtag pa nyid du gzhan dag mi ’dod do //

123
562 T. Watanabe

For, with regard to them (i.e., buddhi, etc., the topic of the thesis) too, it is only
the word that is established [for both the proponent and the opponent], but not
the state of affairs. For others (i.e., Buddhists) do not regard impermanence as
a thing’s resting [on pradhāna] (upalayana) without abandoning its nature.
_
Here, Dharmakı̄rti ascribes the following theory to the Sānkhya: When a thing,
even though it continues to exist, goes back to pradhāna and becomes invisible,25 it
_
is considered to be impermanent. This concept of impermanence of the Sānkhya, of
course, cannot be adopted by Dharmakı̄rti, who does not accept the existence of
_
pradhāna. In response, therefore, the Sānkhya explains the transformation without
using pradhāna, namely, by using the notion of disappearance.
nanv asty eva tirodhānam/na vai paras tad anityatvam āha, kim
: tarhi
vināśam/26

_
[Sānkhya:] There must be [a common thing that is understood through the word
_
‘‘impermanence’’ by both the Sānkhya and the Buddhist, and it is]
disappearance (tirodhāna).
[Buddhist:] The other (i.e., Buddhist) does not call it impermanence, but [calls]
destruction [impermanence].
Dharmakı̄rti denies that disappearance means impermanence because Buddhists
_
only accept its meaning to be destruction. In opposition to this, the Sānkhya tries to
explain that if the meaning of impermanence is taken to be disappearance, imper-
manence is accepted as a proper logical reason for both the Sānkhya _ and the
Buddhist because there is a common feature between destruction and disappearance.
nanu vinas: :tatirohitayor dvayor apy avyaktis tulyety asty eva sāmānyam / atha
keyam avyaktih: /adr: śyātmatā/nanv anena laks: an: ena pradhānapurus: ādayo
’py anityāh: prasajanti/27

_
[Sānkhya:] _
For both [the Sānkhya and the Buddhist], non-manifestation (avyakti)
is equally [accepted] between what has been destroyed and what has
disappeared. Therefore there must be a common thing (sāmānya).

25
Dharmottara does not give much information about the meaning of ‘‘upalayana’’. PVinT: (Ms113a6–7;
D124b1, P145b8–146a1): yasmān na pare saugatā apracyutarūpasyopalayanam adr: śyātmatvam ani-
tyatām icchanti. To understand the meaning, the following verse from Kumārila’s Ślokavārttika is
informative. ŚV ātmavāda k.30: na cāvasthāntarotpāde pūrvātyantam
: vinaśyati / uttarānugun: ārthā tu
sāmānyātmani lı̄yate // See Uno (1996, pp. 109, n. 20–21).
26
PVin 3 (D216a2–3, P313b6–7): bskal (D: skal P) ba nyid yod pa ma yin nam zhe na / de la ni gzhan
dag mi rtag par mi brjod do // ’o na ci zhe na / ’jig pa la’o //
27
PVin 3 (D216a3–4, P313b7–8): gal te zhig pa dang bskal ba dag mi gsal ba nyid du ’dra (D: rung P)
ba’i phyir gnyi ga la yang spyi yod pa nyid ma yin nam zhe na / ci ste mi gsal ba zhes bya ba ’di ci (D:
om. (ci) P) zhig / mi mthong ba’i (D: ba P) bdag nyid do zhe na / mtshan nyid ’dis (D: ’di P) gtso bo dang
skyes bu la sogs pa yang mi rtag par yang thal bar ’gyur ro //

123
Dharmakı̄rti’s Criticism of Anityatva 563

[Buddhist:] Then what is this ‘‘non-manifestation?’’

_
[Sānkhya:] It is imperceptibility (adr: śyātmatā).
[Buddhist:] According to this definition, even pradhāna and purus: a, etc.28 would
be impermanent.
_
Here, this Sānkhya accounts for the commonality between destruction and disap-
pearance in terms of non-manifestation (avyakti). To this, Dharmakı̄rti replies by
examining the meaning of non-manifestation. If this non-manifestation were to
_
mean imperceptibility (adr: śyātmatā), then pradhāna and purus: a, against the Sān-
khya’s tenets, would be impermanent because they are not perceptible. To avoid this
_
difficulty, the Sānkhya gives a slightly modified explanation.
prāg anyadharman: o ‘vyaktir anityateti cet / anyadharmā ca prāg
apracyutātmeti ca suvyāhr: tam /29

_
[Sānkhya:] Impermanence means the non-manifestation of a thing that previously
possessed a different property (i.e., perceptibility) [, but not mere non-
manifestation].
[Buddhist:] But it is very nice to say that [one and the same thing] previously
possessed a different property and [at the same time] it has not yet
abandoned its nature.
With this explanation of impermanence, as provided here by the Sānkhya,_ purus: a
and pradhāna cannot be regarded as impermanent because they are always
imperceptible. However, because it is unacceptable for Dharmakı̄rti that one and the
same thing possesses two mutually incompatible properties at the same time,30 there
are only two alternatives: the previous dharmin and the next one are mutually
distinct things, that is, the previous dharmin is destroyed; or the property (dharma)
of the previous dharmin is destroyed. Since, from an ontological point of view, there
is no difference between a dharmin and its property dharma, it follows in both cases
that the previous entity has been destroyed.

28
According to Dharmottara, this ‘‘ādi’’ refers to imperceptible things in a particular state (av-
asthāviśes: a). Perhaps it means radically inaccessible things (atyantaparoks: a) that can never be ascer-
tained whether they are impermanent or not. PVinT: (Ms114a2; D125a6, P147a1): ādigrahan: ād
avasthāviśes: ā adr: śyātmāno ’nityā prāpnuvanti. On the other hand, Prajñākaragupta seems to regard this
‘‘ādi’’ as referring to absolute nonexistence, such as a rabbit’s horn. PVA 46,8–9: nanu (Ms-B: nanu (na)
ed.) tirobhāvo vinas: :tānabhivyaktayos (em. (cf. zhig pa dang mngon par mi gsal ba dag T):
vinas: :tānabhivyakyos ed., vinas: :tābhivyaktayos Ms-B) tulya eva. ko ’yam : tirobhāvah: (ed.: ko yan ti-
robhāv(o) v(i)nas: :tābhivyaktatā Ms-B). adr: śyātmatā. nanu śaśavis: ān: ādı̄nām
: ca pradhānānām ani-
_
tyatāprāptih: . It is worth noting that Prajñākaragupta criticizes the Sānkhya’s notion of disappearance in
accordance with Dharmakı̄rti’s argument in PVin 3. See PVA 46,8–14 on PV 2.17.
29
PVin 3 (D216a4, P313b8–314a1): sngar chos gzhan yin pa mi gsal ba mi rtag pa yin no zhe na / sngar
chos gzhan yang yin la nyams pa med pa’i bdag nyid kyang yin no // zhes legs par tha snyad byas so //
30
See PVSV 20,21–22: ayam eva khalu bhedo bhedahetur vā bhāvānām : viruddhadharmādhyāsah:
kāran: abhedaś ca / (For a translation and its annotation, see Gillon and Hayes (2008, pp. 352, 393–395));
PVin 2, pp. 89,14–90,1: ayam : hi bhedo bhedahetur vā bhāvānām : yad uta viruddhadharmādhyāsah:
kāran: abhedaś ca /

123
564 T. Watanabe

Criticism of the Notion of State

_
Another Sānkhya, however, raises an objection against this criticism brought forth
by Dharmakı̄rti. After considering an entity to consist of state (avasthā) and state-
possessor (avasthātr: ) instead of dharma and dharmin in order to eliminate the
_
ontological problem caused by the latter two terms, this Sānkhya explains imper-
manence of things by the change of its states.31 It is maintained that the state-
possessor continues to exist even if its states cease to exist or are destroyed.
Accordingly, it can be claimed that a remaining entity as well as some kind of
destruction exist. Dharmakı̄rti begins his refutation of this view with a cross-
examination of the meaning of ‘‘state.’’
avasthā nivartate, nāvasthāteti cet / keyam avasthā / yeyam uda-
kadhāran: ādyarthakriyāyām upanidhı̄yate, yām ayam : jantur adhyaks: am
adhyavasyati ghat:o ’yam iti, tannivr: ttāv anivr: ttāv apy aparasyāsya sarv-
asyābhāvāt / evam: tarhi saivāvasthā ghat:o ’stu, yathoktalaks: an: atvād asya / sa
ca nivr: tta ity anivr: tto ’vasthātāvasthitah: paryanuyojyah: /san vā, na ghat:o
bhāvikah: , atallaks: an: atvāt /32

_
[Sānkhya:] [In the process of transformation, only its] state (avasthā) ceases to
exist, but not the state-possessor (avasthātr: ).
[Buddhist:] What is this ‘‘state?’’
_
[Sānkhya:] It is what is employed for a purposeful activity (arthakriyā) such as
holding water. [Moreover,] this [ordinary] person determines it to be
visible in the form ‘‘this is a pot.’’ For, when it ceases to exist, though
the other (i.e., the state-possessor) does not cease to exist, neither of
these (i.e., purposeful activity and the determination of the object)
would occur.

31
In this connection, it should be noted that Kumārila Bhat::ta (ca. 600–660) also uses the notion of state
(avasthā) and state-possessor (avasthāvat) when he explains the transformation of the eternal soul
(ātman). According to him, one and the same soul, maintaining its existence, can change from the agent
of an action (kartr: ) into the experiencer of its fruits (bhoktr: ) by taking the latter state. In the sense that the
soul is transformed from one state into another, the soul can be called impermanent (anitya). ŚV
ātmavāda 22–23: nānityaśabdavācyatvam ātmano vinivāryate / vikriyāmātravācitve na hy ucchedo ’sya
tāvatā // syātām atyantanāśe ’sya kr: tanāśākr: tāgamau / na tv avasthāntaraprāptau loke bālayuvādivat //
It is seen that there is a great similarity between this view of Kumārila and that of the Sānkhya _
criticized here by Dharmakı̄rti. However, as we will see below, Dharmakı̄rti’s criticism presupposes the
_
Sānkhya’s theory of three constituents (trigun: a). Therefore, it is likely that here he does not intend to
criticize Kumārila’s view of the soul, even if he could refute it. For Kumārila’s explanation of the
transformation of the soul, see Uno (1996,1999). The latter points out the influence of the Jaina theory of
many-sidedness (anekāntavāda) on Kumārila’s theory.
32
PVin 3 (D216a5–7, P314a2–5): gnas skabs ldog gi gnas skabs can ni ma yin no zhe na / gnas skabs
zhes bya ba ’di ci (D: om. (ci) P) zhig / chu ’dzin pa la sogs (D: ... sogs pa P) don byed pa nye bar bsgrub
pa gang yin pa ste / ’di bum pa’o zhes ’jig rten pa ’di gang la mngon sum du lhag par zhen pa ’di yin te/de
log na gzhan ma log kyang ’di thams cad med pa’i phyir ro zhe na / de lta (D: om. (lta) P) na ni ’o na
gnas skabs de nyid bum pa yin te (D: no P) / de ni ji skad bshad pa’i mtshan nyid (D: ... nyid can P) yin
pa’i phyir ro // de yang log pa’i phyir gnas skabs can ma log par gnas pa ni brgal zhing brtag par bya ba
yin no // yod kyang bum pa’i dngos por ni mi ’gyur te / de’i mtshan nyid med pa’i phyir ro //.

123
Dharmakı̄rti’s Criticism of Anityatva 565

[Buddhist:] If so, this very ‘‘state’’ would be the pot, because it (i.e., the pot) has
the above-mentioned characteristic (i.e., being employed for a pur-
poseful activity and bringing about the determination). And it (i.e., the
pot as a state) has already ceased to exist. Therefore, [the existence of]
the remaining state-possessor, which has not yet ceased [to exist],
must be questioned.33 Or if it exists, the pot [as a state-possessor]
would not be real because it would not have such a characteristic.
_
According to this Sānkhya, the ‘‘state’’ is characterized as something that accom-
plishes purposeful activity and brings about the determination that this is the object
of the cognition.34 For Dharmakı̄rti, however, this means that the state can be
considered real,35 because only it has causal efficacy (arthakriyāśakti) and not the
state-possessor. Therefore, the state-possessor would lose the reason for its exis-
tence and not be accepted as a real entity even though it exists in one form or
another. As a result, it must be acknowledged that all real phenomena, whether
internal or external, are characterized as causal efficacies and undergo destruction.
_
In response to this criticism, the Sānkhya claims that the reality of an entity is not
found in phenomena, but in something that is not manifest as a phenomenon but
exists behind phenomena. This entity is pradhāna. This idea enables the Sānkhya_ to
insist that the difference in causal efficacy affects only the appearance of an entity,
not its identity.
vastv ekam evāvasthāntarāveśād bhedadr: s: :tir iti cet/sa evāvasthābhedo
vastubhedalaks: an: am
: kim
: nes: yate/evam
: hi sukhādı̄nām asandigdho bhedo
bhavati /36

_
[Sānkhya:] Due to its taking on a different state, one and the same entity (i.e., the
state-possessor) looks different [from its previous form].
[Buddhist:] Why is it not accepted that this very difference of state is the char-
acteristic of the distinctness of the entities? For, in this same manner,
the [mutual] differences among pleasure (sukha), etc. are undoubtedly
_
[accepted in the Sānkhya system].
Since all manifested phenomena, such as pots, etc., are transitory existences, the
_
Sānkhya holds that only pradhāna, i.e., the non-manifest (avyakta), can be accepted
as an unchangeable real entity. Therefore, it is the only thing that can serve as a
state-possessor. When a state-possessor connects with different states, it changes its
appearance and in each case manifests different causal efficacies. Hence it seems to
33
PVinT: (Ms114b6–7; D126a6, P148a2–3): tasmād anivr: tto yo ’vasthātānyas tasyāvasthāyāh: , sa par-
yanuyojyah: sattvam : prati—katham : punar asāv arthakriyayā pramān: ena vā vinā vidyamāno bhaved iti.
34
From SK 9, it is possible to derive the view that phenomena have the power to accomplish or produce
their effects. SK 9: asadakaran: ād upādānagrahan: āt sarvasambhavābhāvāt / śaktasya śakyakaran: āt
kāran: abhāvāc ca satkāryam // And see also YSBh 132,8 on YS 3.14: yogyatāvacchinnā dharmin: ah: śaktir
eva dharmah: / On this interpretation of YSBh, see Chakravarti (1951, pp. 198–200).
35
PV 1.165ab: sa pāramārthiko bhāvo ya evārthakriyāks: amah: / etc.
36
PVin 3 (D216a7–216b1, P314a5–6): dngos po gcig nyid gnas skabs gzhan dang ’brel pa’i phyir / tha
dad par mngon pa (D: par P) yin no zhe na / gnas skabs tha dad pa de nyid dngos po tha dad pa’i mtshan
nyid du ci ste mi ’dod / de lta yin dang bde ba la sogs pa rnam pa tha dad par ’gyur bar gdon mi za’o //.

123
566 T. Watanabe

be different things. However, the difference of the appearance does not mean that
the state-possessor is different.
Dharmakı̄rti, in turn, refutes this by pointing out its inconsistency with the
_
Sānkhya’s _
own tenet. For the Sānkhya, pradhāna consists of the three constitu-
ents—sattva, rajas and tamas. They are, respectively, of the nature of pleasure
(sukha), pain (duh: kha) and delusion (moha), and are entirely different from one
another. Moreover, they are described as having mutually distinct purposes
(artha).37 These sattva, etc. are, in non-Sānkhya
_ treatises including the works of
Dharmakı̄rti, often mentioned as being identical to pleasure, etc.38 And here too,
pleasure, etc. are referred to as constituents of pradhāna. Therefore, it can be said
that pleasure, etc. are regarded as being different entities as well as having different
powers,39 i.e., causal efficacy. From this, Dharmakı̄rti derives the necessary rela-
tionship between a difference of causal efficacy and difference of entity. In this part
_
of the PVin, however, the Sānkhya insists that the state-possessor/entity remains one
and the same, even though its states, which are associated with state-possessor and
characterized as causal efficacy, are different. If this were the case, then the dif-
ferences among the three constituents could not be maintained because they have
different causal efficacies. Therefore, as long as the Sānkhya_ maintains that there
are differences among the three constituents, he has to accept that a difference of
states means a difference of the state-possessor/entity, i.e., that the previous entity is
destroyed and a new entity is produced in the process of transformation. In this way,
_
Dharmakı̄rti shows that the Sānkhya’s explanation of transformation relying on the
notion of state is in conflict with his own tenets.

37
See SK 12: prı̄tyaprı̄tivis: ādātmakāh: prakāśapravr: ttiniyamārthāh: / anyonyābhibhavāśrayajananami-
thunavr: ttayaś ca gun: āh: //; GBh 15,4 on SK 12: tatra prı̄tyātmakam : sattvam, prı̄tih: sukham
: tadātmakam
iti / and STK 52,10 on SK 12: prı̄tih: sukham : prı̄tyātmakah: sattvagun: ah: .
38
In Dharmakı̄rti’s works, the following examples can be found. PVin 3 on k.60 (D210a4–5, P307b6–7):
sa hi dharmı̄ pradhānalaks: an: a eko nityah: sukhādyātmako ’nyo veti yathākathañcid api viśes: itas tat-
svabhāvah: sādhito bhavati / (=PVSV 94,11-13); VN 66,15: tatraikā prakr: tih: sukhaduh: khamohah: /. Also
‘‘sukhādi’’ in PV 4.144-145 (tathaiva dharmin: o ’py atra sādhyatvāt kevalasya na / yady evam atra bādhā
syāt nānyānutpādyaśaktikah: // sakr: c chabdādyahetutvāt sukhādir iti pūrvavat / virodhitā bhaved atra
hetur aikāntiko yadi //) can be understood as referring to pradhāna. See Tillemans (2000, pp. 205–207).
Note that Dharmakı̄rti does not necessarily use pleasure, etc. to describe pradhāna. He also uses sattva,
etc. See PVin 3 on k.32 (D208b5–6, P298a8–298b1): tac copalabhyamadhyarūpam pūrvāparayoh: kot:yor
astı̄ti bruvān: ah: padārthavyavasthām : bādhate / sattvarajastamasām : caitanyasya caivam : paraspa-
rarūpavivekena vyavasthāyogāt /
Kamalaśı̄la explicitly identifies sattva, etc. with sukha, etc. when explaining one of the five vı̄ta-
arguments put forth by the Sānkhya _ (probably Vārs: agan: ya) to prove the existence of pradhāna. TSP
27,1–7 on TS 14: itaś cāsti pradhānam, bhedānām anvayadarśanāt /... sukhaduh: khamohādijātisaman-
vitam : cedam : vyaktam upalabhyate/kutah: / prasādatāpadainyādikāryopalabdheh: /tathā hi prasādalā-
_
ghavābhis: vangoddhars : aprı̄tayah: sattvasya kāryam / sukham iti ca sattvam evocyate / tāpaśos: abhedast-
ambhodvegāpadvegā rajasah: kāryam / rajaś ca duh: kham / dainyāvaran: asādanādhvam : sabı̄bhatsagau-
ravān: i tasmasah: kāryam / tamaś ca mohaśabdenocyate /
Some instances showing the interchangeability of sattva, etc. and sukha, etc. can be found in Jaina
literature, i.e., the Nayacakra and its commentary, the Nyāyāgamānusārin: ı̄, as has been pointed out by
Wezler (1985b, p. 6, 28 n. 27).
39
See GBh 15, 6 on SK 12: tathā prakāśapravr: ttiniyamārthāh: / arthaśabdah: sāmarthyavācı̄.

123
Dharmakı̄rti’s Criticism of Anityatva 567

Conclusion

Vasubandhu, in his AKBh, examines the Sānkhya’s _ two explanations of transfor-

mation. One considers an entity to consist of dharma and dharmin and explains the
process of transformation as the cessation (nivr: tti) of the previous dharma and
appearance of the next dharma in a persisting dharmin. In contrast, the other does
not accept a distinction between dharma and dharmin, and claims that transfor-
mation means that the same entity becomes different. Vasubandhu criticizes both
explanations, pointing out, with regard to the former case, that there is no onto-
logical distinction between dharma and dharmin, and, in the latter case, that
sameness and difference cannot be maintained in one and the same entity. However,
by the time of Dharmakı̄rti, the theory of transformation had been developed in the
_
Sānkhya school. The defects of the explanation of transformation in the AKBh, i.e.,
the cessation of dharma and an entity’s becoming different, have been superseded
by the notion of (a) the disappearance (tirodhāna/tirobhāva) of dharma and (b) the
cessation of the state (avasthā) of an entity, respectively. Therefore, Vasubandhu’s
criticism would allow room for a retort from this more developed theory.
In discussing the meaning of anityatva propounded by the Sānkhyas _ as a logical
reason, Dharmakı̄rti closely examines these two crucial notions and criticizes both
of them. With regard to the notion of disappearance of dharma, his criticism is not
directly leveled against it, but against disappearance itself. In opposition to the
_
Sānkhya’s explanation that disappearance is a changing of the dharma, i.e., per-
ceptibility of things changing into imperceptibility, he shows that it is impossible to
assume two contradictory dharmas in one and the same entity. Therefore, whenever
something disappears, the previous dharma or dharmin itself is destroyed. As a
result, because, from an ontological point of view, dharma cannot be separated from
dharmin, the disappearance of an entity entails its destruction. Concerning the
_
second notion, he points out the contradiction between the Sānkhya’s description of
_
the state (avasthā) and his theory of the three constituents (trigun: a). The Sānkhya
identifies the state of an entity with a causal efficacy, while he insists that the three
constituents are different from each other and at the same time they have different
causal efficacies. Therefore, insofar as he holds the theory of the three constituents,
he has to accept that a difference of state necessarily leads to a difference of the
entity itself and means the destruction of the previous entity. In this way, Dhar-
makı̄rti shows that the word ‘‘impermanence’’ means destruction, and hence the
_
Sānkhya, who does not accept the destruction of entities (dharmin/avasthātr: ),
cannot put forward ‘‘impermanence’’ as a logical reason because, in a proof, the
logical reason must be accepted by both proponent and opponent.
_
Dharmakı̄rti’s criticism of anityatva in the Sānkhya theory basically follows the
same lines as Vasubandhu’s arguments against the Sānkhya _ theory of transforma-
tion. However, Dharmakı̄rti supplies some important additions in order to make
Vasubandhu’s theory sustainable against the more developed Sānkhya _ theory. It can
be said that Vasubandhu’s theory is revived with the help of Dharmakı̄rti’s
modifications.

123
568 T. Watanabe

Acknowledgments I would like to thank Prof. Mark Siderits and Prof. Kiyotaka Yoshimizu for
valuable suggestions. I would also like to thank Ms. Peck-Kubaczek for correcting my English. Work on
this paper has been generously supported by the Austrian Science Fund (FWF) in the framework of the
FWF project P21050-G15 (‘‘Tradition und Wandel in der indischen buddhistischen Logik’’).

References

AD: P. S. Jaini (Ed.). (1959). Abhidharmadı̄pa with Vibhāshāprabhāvr: tti. Tibetan Sanskrit Works Series 4.
Patna: K.P. Jayaswal Research Institute.
AKBh: P. Pradhan (Ed.). (1967). Abhidharmakośabhās: ya of Vasubandhu. Tibetan Sanskrit Works Series 8.
Patna: K.P. Jayaswal Research Institute.
Bronkhorst, J. (1997). Sām : khya in the Abhidharmakośa Bhās: ya. Journal of Indian Philosophy, 25,
303–400.
Bronkhorst, J. (2003). Review of Yuktidı̄pikā. Zeitschrift der Deutschen Morgenländischen Gesellschaft,
153-1, 242–247.
Chakravarti, P. (1951). Origin and development of the Sām : khya system of thought. Calcutta: Metropolitan
Printing and Publishing House.
D: Derge edition of Tibetan tripitaka.
Deleanu, F. (2006). The Chapter on the Mundane Path (Laukikamārga) in the Śrāvakabhūmi. A trilingual
edition (Sanskrit, Tibetan, Chinese), annotated translation, and introductory study (2 Vols.). Tokyo:
The International Institute for Buddhist Studies of the International College for Postgraduate Bud-
dhist Studies.
Frauwallner, E. (1953). Geschichte der indischen Philosophie. I. Band: Die Philosophie des Veda und des
Epos der Buddha und der Jina das Samkhya und das klassische Yoga-System. Salzburg: Otto Müller
Verlag.
Frauwallner, E. (1958). Die Erkenntnislehre des klassischen Sām : khya-Systems. Wiener Zeitschrift für die
Kunde Süd- und Ostasiens, 2, 84–139.
GBh: H. D. Sharma (Ed.). (1933). The Sām : khya-Kārikā: Īśvara Kr: :sn: a’s Memorable Verses on Sām
: khya
philosophy, with the commentary of Gaud: apādācārya. Poona: The Oriental Book Agency.
Gillon, B., & Hayes, R. (2008). Dharmakı̄rti on the role of causation in inference as presented in
Pramān: avārttika Svopajñavr: tti 11–38. Journal of Indian Philosophy, 36, 335–404.
Halbfass, W. (1992). On being and what there is: Classical Vaiśes: ika and the history of Indian Ontology.
Albany: SUNY Press.
Imanishi, J. (1968). On parin: āma. Journal of Indian and Buddhist Studies, 16-2, 641–649 (in Japanese).
Inami, M. (1995). A study of the Pramān: asiddhi chapter of Pramān: avārttika (4). Bulletin of Shimane
International College, 2, 43–67 (in Japanese).
Iwata, T. (1995). Pramān: aviniścaya III (1) (Die Definition des parārthānumāna). Wiener Zeitschrift für
die Kunde Südasiens, 39, 151–179.
*Mahāvibhās: ā ( ): Translated by Hsüan-tang. Taisho 25 (1545).
Malinar, A. (1999). Prakr: ti as sāmānya. Asiatische Studien Études Asiatiques, 53-1, 619–643.
Motegi, S. (1994). Some Vaiśes: ika thoughts referred to in the Yuktidı̄pikā. Asiatische Studien Études
Asiatiques, 48-2, 807–817.
Ms-B: S. Watanabe (Ed.). (1998). Sanskrit manuscripts of Prajñākaragupta’s Pramān: avārttikabhās: yam.
Patna-Narita: Bihar Research Society and Naritasan Institute for Buddhist Studies.
Muroya, Y. (1996). Parin: āma and Abhivyakti in the Sām : khya argument for Satkāryavāda. Studies in the
History of Indian Thought, 8, 42–63 (in Japanese).
NĀA: Muni Jambuvijayaji (Ed.). (1966). Dvādaśāram : Nayacakram of Ācārya Śrı̄ Mallavādi
Ks: amāśraman: a with the commentary Nyāyāgamānusārin: ı̄ of Śrı̄ Sim : hasūri Gan: i Vādi
Ks: amāśraman: a, Part I (1–4 Aras). Bhavnagar: Sri Jain Atmanand Sabha, 1966.
NB: D. Malvania (Ed.). (1971). Pan: d: ita Durveka Miśra’s Dharmottarapradı̄pa [Being a sub-commentary
on Dharmottara’s Nyāyabindut:ı̄kā, a commentary on Dharmakı̄rti’s Nyāyabindu]. TSWS 2 (2nd ed.).
Patna: Kashi Prasad Jayaswal Research Institute.
NBh: A. Thakur (Ed.). (1997a). Gautamı̄yanyāyadarśana with Bhās: ya of Vātsyāyana. New Delhi: Indian
Council of Philosophical Research.
NC: See NĀA.
NV: A. Thakur (Ed.). (1997b). Nyāyabhās: yavārttika of Bhāradvāja Uddyotakara. New Delhi: Indian
Council of Philosophical Research.

123
Dharmakı̄rti’s Criticism of Anityatva 569

P: Peking edition of Tibetan tripitaka.

PST: : D4268, P5766.
PV 2: See Inami (1995).
PVA: R. Sānkr _ : tyāyana (Ed.). (1953). Pramān: avārtikabhās: yam or Vārtikālankārah
_ : of Prajñākaragupta.
Patna: Kashi Prasad Jayaswal Research Institute.
PVin 2: E. Steinkellner (Ed.). (2007). Dharmakı̄rti’s Pramān: aviniścaya: Chapters 1 and 2. Beijing-
Vienna: China Tibetology Publishing House/Austrian Academy of Sciences Press.
PVin 3: D4211, P5710.
PVinT: : D4228, P5727.
PVSV: R. Gnoli (Ed.). (1960). The Pramān: avārttikam of Dharmakı̄rti: The first chapter with the auto-
commentary. Serie Orientale Roma XXIII. Roma.
SK: See GBh.
Solomon, E. (1974). The Commentaries of the Sām : khya Kārikā—A Study. Ahmedabad: Gujarat
University.
S: T: See Frauwallner (1958).
STK: R. Chandra (Ed.). (1935). Vācaspati Miśra’s Sām : khyatattvakaumudı̄ with Sām : khyakārikā of
Īśvarakr: s: n: a and the Indroduction to the Sām : khyatattvavilāsa of Raghunātha, etc. etc. Calcutta
Sanskrit Series 15. Calcutta: Metropolitan Printing and Publishing House.
ŚV: See Uno (1996, 1999).
Tillemans, T. (2000). Dharmakı̄rti’s Pramān: avārttika: An annotated translation of the fourth chapter
(parārthānumāna) (Vol. 1, k. 1–148). Wien: Verlag der Österreichschen Akademie der Wissens-
chaften.
TSP: S. D. Shastri (Ed.). (1968). Tattvasangraha_ of Ācārya Shāntaraks: ita with the commentary Pañjikā of
Shri Kamalshı̄la. Bauddha Bharati Series 1. Varanasi: Bauddha Bharati.
Uno, T. (1996). The Eternity of the Soul (ātman) in Kumārila Bhat::ta. Tetsugaku, 48, 96–110 (in Japa-
nese).
Uno, T. (1999). Ontological Affinity between the Jainas and the Mı̄mām : sakas Viewed by Buddhist
Logicians. In S. Katsura (Ed.), Dharmakı̄rti’s Thought and Its Impact on Indian and Tibetan Phi-
losophy (pp. 419–431). Wien: Verlag der österreichschen Akademie der Wissenschaften.
VN: M. T. Much (Ed.). (1991). Dharmakı̄rtis Vādanyāyah: , Teil I Sanskrit-Text. Wien: Verlag der
Österreichschen Akademie der Wissenschaften.
VNT: : R. Sānkr _ : tyāyana (Ed.). (1935–1936). Dharmakı̄rti’s VĀDANYĀYA: With the commentary of
Śāntaraks: ita. Appendix to J.B.O.R.S, Vols. 21 & 22. Patna.
Wezler, A. (1985a). A note on Vārs: agan: ya and the Yogācārabhūmi. Journal of the Asiatic Society, 27-2,
1–17.
Wezler, A. (1985b). A note on Mahābhās: ya II 366.26: Gun: asam : drāvo dravyam. In Buddhism and its
relation to other religions: Essays in honour of Dr. Shozen Kumoi on his seventieth birthday
(pp. 1–33). Kyoto: Heirakuji Shoten.
Wezler, A. (1992). Paralipomena zum Sarvasarvātmakatvavāda II: On the Sarvasarvātmakatvavāda and
its relation to the Vr: ks: āyurveda. Studien zur Indologie und Iranistik, 16/17, 287–315.
Yamashita, K. (1994). Pātañjala Yoga philosophy with reference to Buddhism. Calcutta: Firma KLM
Private Limited.
YD: A. Wezler, & S. Motegi (Eds.). (1998). Yuktidı̄pikā: The Most Significant Commentary on the
Sām : khyakārikā (Vol. I). Alt- und Neu-Indische Studien 44. Stuttgart: Franz Steiner Verlag.
YSBh: K. Ś. Āgaśe (Ed.). (1904). Vācaspatimiśraviracitat:ı̄kāsam : valitavyāsabhās: yasametāni Pātañja-
layogasūtrān: i. Ānandāśrama Sanskrit Series 47. Poona: ānandāśrama.

123