Reverberations of Dharmakirti’s Philosophy

by Birgit Kellner | 2020 | 264,305 words

This page relates ‘Dharmakirti and His Successors on Ashrayasiddha and Prasangaviparyaya’ of the study on the philosophy of Dharmakirti (6th century) and his predecessor Dignaga (5th century). This collection of articles reflects philosophical currents in India, China and Tibet during their time and investigates the Buddhist theories of Pramana (“instruments of trustworthy awareness”).

Dharmakīrti and His Successors on Āśrayāsiddha and Prasaṅgaviparyaya

(By Toshikazu Watanabe)

Work on this paper has been generously supported by the Austrian Science Fund (FWF) in the framework of the FWF project P23196–G15 “Buddhist literature and its context” and by the 2014 Bukkyo Dendo Kyokai Fellowship for Japanese scholars for my project “Acceptance of logical thought in Mahāyāna Buddhism and its development.” I would like to thank Ms. Katharine Apostle for correcting my English and my colleague Dr. Pascale Hugon for giving me invaluable suggestions.

Introduction

For Buddhist philosophers the logical fallacy called āśrayāsiddha (a pseudo-logical reason whose basis is not established) is an intractable problem when dealing with something whose existence is not accepted by Buddhists as the subject of a thesis of their own, such as the Sāṅkhya’s pradhāna (primordial matter) or the Vaiśeṣika’s eternal ākāśa (space) and so on. Dignāga (ca. 480–540), the founder of the Buddhist logico-epistemological school, presented different approaches to this issue in his two works, the Nyāyamukha and the Pramāṇasamuccaya. In his earlier work, the Nyāyamukha, when proving the non-existence of the Sāṅkhya’s pradhāna through the logical reason “not being perceived,” he permitted pradhāna to be placed as the subject of the thesis by seeing pradhāna as a conceptually constructed object (kalpita), in order to allow the logical reason to be a property of the subject of the thesis (pakṣadharma). However, in the Pramāṇasamuccaya he does not use this method, but addresses the problem in a different way. According to his new method, the Buddhist proponent can put forth the Sāṅkhya’s pradhāna as the subject of the thesis when he formulates a reductio ad absurdum kind of argument (prasaṅga) to refute (dūṣaṇa/ parihāna) the adversary’s tenets. For the sake of convenience I shall refer to this technique for avoiding the problem of āśrayāsiddha as the “method of hypothetical assumption.”

As Tom Tillemans has pointed out,[1] Dharmakīrti (ca. 600–660 or 550–650?) also provides two ways to deal with this issue: The first is based on Dignāga’s solution found in the Nyāyamukha. In his Svavṛtti on Pramāṇavārttika 1.205–206 Dharmakīrti explains the word kalpita mentioned in the NMu as referring to an image that appears in the cognition and that it thus can exist as the subject of the thesis. This first method is designated by Tillemans as the “principle of conceptual subjects.” The second method is mentioned in PV 4.136–148, where Dharmakīrti comments on the word svadharmiṇi in Dignāga’s definition of a thesis (pakṣa) given in Pramāṇasamuccaya 3.2. According to this second method, when the Buddhist proponent places ākāśa, for example, as the subject of a thesis, what is intended by the word ākāśa is not the eternal entity postulated by the Vaiśeṣika opponent, but any other entity called ākāśa whose existence is recognized not only by the Buddhist proponent, but also by ordinary people. Tillemans calls this “the method of paraphrase.”

In the third chapter of the Pramāṇaviniścaya, Dharmakīrti also adopts Dignāga’s second method as presented in the PS, i.e., the “method of hypothetical assumption.” As a result, a total of three methods–two from Dignāga and one developed by Dharmakīrti–coexist within Dharmakīrti’s system of logic in order to prevent the problem of āśrayāsiddha. How then does he harmonize these three, especially the two methods adopted from Dignāga? If the “principle of conceptual subjects” is applied, the “method of hypothetical assumption” seems to be no longer necessary. Moreover, it is reported by Takashi Iwata that Dharmakīrti’s commentators, such as Dharmottara and Prajñākaragupta, have views on prasaṅga and prasaṅgaviparyaya that are different from how they are discussed by Dharmakīrti in PVin 3. The difference of their opinions depends on which of Dharmakīrti’s methods they ascribe more importance to.

The aim of this paper is, therefore, to examine Dharmakīrti’s explanation of how to avoid the problem of āśrayāsiddha and to clarify its historical development from Dignāga to Dharmakīrti and on to his successors.

1. Dignāga

1.1 The Nyāyamukha: the principle of conceptual subjects

In the NMu, Dignāga explains the general rule that in an inference–be it an inference-for-oneself (svārthānumāna) or an inference-for-others (parārthānumāna)–one property of the subject (i.e., sādhya-dharma/a property to be proved) is known through another property of the same subject (i.e., sādhana-dharma/a proving property). In this explanation, Dignāga refers to an opponent (probably a Sāṅkhya) who insists that this rule cannot cover every instance because in some cases the subject (dharmin) itself is proved to be existent or non-existent.

As an example this opponent gives the following reasoning, which uses pradhāna as the subject of the thesis:[2]

(Prayoga 1)

[Thesis:] [pradhāna] does not exist;

[Reason:] because it is not perceived. (Cf. PVSV 105,15: na santi pradhānā-dayaḥ, anupalabdheḥ.)[3]

To this, Dignāga gives the following answer:

When the non-existence [of pradhāna] is proved [on account of its not being perceived], “non-perception” is a property of a conceptually constructed object (i.e., pradhāna) (假安立不可得法/*kalpitasyānupalabdhir dharmaḥ); hence, there is also no error of [proving] the subject of the thesis [with the logical reason].[4]

Here, in order to ensure the pakṣadharmatva for the logical reason “non-perception,” Dignāga gives a certain status of existence to pradhāna by seeing it as a conceptual construction (kalpita). In this way, any pseudo-entity can be accepted as a substratum of the logical reason, and the fallacy of āśrayāsiddha can be thereby avoided. This strategy, following Tillemans, is called the “principle of conceptual subjects.” At the time of the NMu this was the only means for avoiding the problem of āśrayāsiddha. In this connection, it should be noted that Dignāga shows here that using this principle enables Buddhists to put pradhāna as a subject of the thesis in a proper proof (sādhana), i.e., a proof that is put forth by Buddhist proponents themselves, even though they do not accept such pseudo-entities in reality.

1.2 The Pramāṇasamuccaya: the method of hypothetical assumption

In the third chapter of the Pramāṇasamuccaya, while Dignāga deals with a reasoning that has the same thesis as the (Prayoga 1), i.e., nāsti pradhāna, he discusses it in a different context, i.e., a discussion about the reductio ad absurdum kind of argument (prasaṅga).

PS(V) 3.16cd–17ab:[5]

evaṃ kecid anyaprasaṅgaṃ kathayanti–nāsti pradhā-nam, bhoktṛbhogyayor guṇāguṇatvaprasaṅgāt … iti. sa katham āvīta iti cet. na hy ayam āvītaḥ. yasmāt

prasaṅgo ’pakṣadharmatvād anyo hetupratijñayoḥ //16//
doṣoktyā dūṣaṇaṃ jñātaṃ pūrvatropagame sati /

[Objection:] Like [the Sāṅkhyas], some [Buddhists] also speak of another reductio ad absurdum reasoning (prasaṅga) [as follows:]

(Prayoga 2) It is not the case that pradhāna exists, because [if the pervasion (vyāpti) you presuppose in your proof of the existence of pradhāna were to be accepted,][6] it would follow that both the enjoyer (i.e., puruṣa) and that which

is to be enjoyed (i.e., vyakti) consist of [three] constituents (guṇa), or do not consist of [three] constituents.

… Why is this an āvīta reasoning? [Of course it is not.] [Answer:] Actually, this is not an āvīta reasoning because,

since [its logical reason is] not the property of the subject of the thesis, the re-ductio ad absurdum reasoning is different [from the āvīta reasoning]. [Instead, this reductio ad absurdum reasoning is] understood as a refutation (dūṣaṇa) because [it] points out the failure of the [adversary’s] logical reason or of his thesis after accepting [them].

Just before this passage, Dignāga rejects the Sāṅkhya view that in the āvīta reasoning–a type of prasaṅga argument–the logical reason can prove the Sāṅkhya’s proposition without its being a property of the thesis (pakṣadharmatva), and he explained that if the pakṣadharmatva of the logical reason in the vīta reasoning for proving the existence of pradhāna is accepted (at least) by the Sāṅkhya proponent, then the logical reason in the āvīta reasoning can also be regarded as possessing pakṣadharmatva because it can be reduced to the logical reason in the vīta reasoning.[7]

The Sāṅkhya raises an objection against this explanation: In spite of the fact that for the Buddhists the (Prayoga 2) must consist in sound reasoning, the logical reason cannot possess pakṣadharmatva because the existence of the subject of the thesis in (Prayoga 2), i.e., pradhāna, is not accepted by Buddhists. As a result, it cannot be held that every sound logical reason must be a property of the subject of the thesis (pakṣadharma).

In his reply, Dignāga, unlike in the case of the NMu, does not try to ensure the pakṣa-dharmatva of the logical reason in the (Prayoga 2). He rather agrees that the logical reason does not possess pakṣadharmatva. This does not mean, however, that he relinquishes the soundness of the (Prayoga 2). He is able to do this by distinguishing the logical reason in a proper proof (sādhana) from the logical reason in a refutation (dūṣaṇa). While the former should possess pakṣadharmatva (and a vyāpti relation with the property to be proved), the latter does not. According to Dignāga, a refutation is simply a means of denying the sound-ness of the adversary’s reasoning by pointing out the undesired consequences which occur when the adversary’s thesis or logical reason (which necessarily includes a vyāpti relation) is accepted.[8] And since in the refutation what is to be investigated is the logical consequence derived from the adversary’s thesis or reasoning, the existence or non-existence of the subject of the thesis does not become a topic of discussion. Therefore, the problem of āśrayāsiddha does not occur in the refutation. Since the prasaṅga argument is formulated after hypothetically assuming the opponent’s system of philosophy, this method can be called the “method of hypothetical assumption.”[9]

2. Dharmakīrti

2.1 Dharmakīrti’s version of the principle of conceptual subject–PV 1 and PVin 3

In PV 1.205–212 and his own commentary on these verses, Dharmakīrti explains Dignāga’s previously mentioned (Prayoga 1) (“pradhāna does not exist because it is not perceived”) presented in the NMu, developing Dignāga’s principle of conceptual subjects from the point of view of the theory of apoha. In the following verses, he explains why in the (Prayoga 1) the fallacy of āśrayāsiddha does not occur.

PV 1.205–206 = PVin 3.53–54:

anādivāsanodbhūtavikalpapariniṣṭhitaḥ /
śabdārthas trividho dharmo bhāvābhāvobhayāśrayaḥ //205//
tasmin bhāvānupādāne sādhye’syānupalambhanam /
tathā hetur na tasyaivābhāvaḥ śabdaprayogataḥ //206//

The verbal object (śabdārtha), which is completely derived from conceptu-alisation proceeding from beginningless karmic tendencies, is a dharma of three kinds: based on something existent, something non-existent or both.[10] When this [verbal object, such as pradhāna, etc.], which is without any existent substratum, is being proven, then the non-perception of this as being in such a way [i.e. as existing externally] is the logical reason. The non-existence of this very [śabdārtha] itself is not, for we do use words [like ‘pradhāna,’ etc.]. (Tillemans 1999: 176)

Following Dignāga’s description in the NMu, Dharmakīrti here explains that the subject of the thesis pradhāna is a conceptual construction and that this pradhāna does not have any external basis.

But why can this conceptual construction be accepted as a proper subject of the thesis? This is explained as follows:

PVSV 105,26–27 = PVin 3 68,1–3:

vaktuḥ śrotuś ca tadvikalpabhājaḥ, yathā-pratibhāsivastupratipādanasamīhāprayogāt, tadākāravikalpajananāc ca.

And both speaker and listener share such a conceptual cognition (i.e., a con-ceptual cognition coming from a beginningless imprint) because [the former] uses [words] according to the intention to convey a thing as it appears [to his conceptual cognition] and because [the latter, by hearing the speaker’s words,] brings about the conceptual cognition which has the same form [as that of the speaker].

Although the conceptually constructed pradhāna does not have an external basis, it exists as an image appearing to the cognition, that is, it has an internal basis. Moreover, since this internal image is considered to be common to the speaker and listener,[11] i.e, the proponent and opponent, it can be a proper subject of the thesis.

In this way, Dharmakīrti develops Dignāga’s “principle of conceptual subjects” further by providing an ontological basis. He first advocated this view in his PVSV, and in his later work it is included in PVin 3 without any changes.

2.2 Method of paraphrase–PV 4

Dharmakīrti’s second method for avoiding the problem of āśrayāsiddha is, in the words of Tillemans, the “method of paraphrase.” This method is employed by Dharmakīrti in his PV 4.136–148, where he comments on the word svadharmiṇi that appears in Dignāga’s definition of the thesis in PS 3.2. An example dealt with in the relevant passage of PV 4 is a Buddhist proof of the impermanence of space and so forth (khādi or vyomādi) through the logical reason “not producing sound, etc., all at once” (sakṛc chabdādyahetutva).[12] The Vaiśeṣika opponent raises the objection that if this reasoning is held to be true and is aimed at negating the Vaiśeṣika understanding of space as an eternal entity, it then follows that, because this reasoning negates the nature of space as understood by the Vaiśeṣika and hence the existence of that space cannot be accepted,[13] the logical reason is categorized as dharmisvarūpaviparītasādhana. Dharmakīrti replies to this objection by introducing the distinction between svadharmin (the subject intended by the proponent) and kevaladharmin (a thing that is not related to the subject intended by the proponent). In the case of the reasoning of impermanence of space, the subject put forth by the Buddhist proponent is not eternal space as postulated by the Vaiśeṣika opponent, but a certain entity called “space” whose existence is widely accepted by ordinary people. Therefore, even though the existence of eternal space is negated by the reasoning, no invalidation of the Buddhist proponent’s thesis can occur.[14] Although Dharmakīrti himself does not mention here the problem of āśrayāsiddha directly, this method is applicable to our relevant problem. Hence the logical reason “not producing sound, etc., all at once” is accepted to be a property of the subject of the thesis because the svadharmin of this reasoning is a certain entity called “space.”

2.3 Dharmakīrti’s version of the method of hypothetical assumption–PVin 3

In PVin 3, Dharmakīrti speaks about prasaṅga arguments formulated by relying on what is accepted only by the opponent. There, in order to avoid the problem of āśrayāsiddha, he also adopts the method of hypothetical assumption presented by Dignāga in PS 3.

Dharmakīrti begins his discussion as follows:

PVin 3 4,4–9:

yas tu paraparikalpitaiḥ prasaṅgaḥ, yathā deśakālāvasthāviśe-ṣaniyataikadravyasaṃsargāvyavacchinnasvabhāvāntaravirahād anekavṛtter ekasya na deśādiviśeṣavatānyena yogaḥ, tathābhūtasvabhāvasya virodhād bhinnadeśādiyogena, sa ekadharmopagame’paradharmopagamasandarśa-nārthaḥ. tadanabhyupagame cobhayanivṛttiḥ, vivekasya kartum aśakyatvāt, tasyānyatra vastutaḥ pratibandhāt.

On the other hand, a reductio ad absurdum kind of argument (prasaṅga) [is formulated] by means of [things] postulated by the opponent, as for instance: (Prayoga 3)

[Thesis:] A single entity (eka), [although] it is [regarded by the opponent as] occurring in several things, is not united with others which have differences in terms of place, etc.

[Reason:] because it is devoid of another essential property which is not ob-structed (avyavacchinna) by the unification (saṃsarga) with a single substance restricted by a particular place, time and state.

For it is incompatible that a thing that has such a nature (i.e., singularity) is united with those which are different in place, etc.

The purpose of such [an argument] (saḥ) is to show that when one property X is accepted, the other property Y is [also] accepted. To the contrary, when Y is not accepted, both are negated because [Y] cannot be distinguished [from X]. This is because Y is in reality bound to the other (i.e., X).

I will discuss the entire structure of this (Prayoga 3) below. What I would like to focus on here is the passage beginning with the relative pronoun saḥ. In this part Dharmakīrti explains the purpose of the prasaṅga argument. According to him, the prasaṅga argument is set forth in order to show that a vyāpti relation between the logical reason and the property to be proved is established.

To this explanation, however, a Nyāya-Vaiśeṣika opponent raises an objection by regard-ing this (Prayoga 3) as a proper proof (sādhana). If (Prayoga 3) proves the non-existence of a single entity, i.e., a universal (sāmānya) which is taken by the Nyāya-Vaiśeṣika opponent as occurring in several things, then the logical reason of (Prayoga 3) cannot escape the fallacy of āśrayāsiddha, because for the Buddhist proponent the existence of the subject, i.e., the universal, cannot be accepted as real existence. Or if, relying on the Nyāya-Vaiśeṣika theory, the Buddhist proponent accepts the existence of the universal as the subject of the thesis, then, because the consequence of this (Prayoga 3) is incompatible with the Nyāya-Vaiśeṣika tenets, the invalidation of the thesis would occur.

To this objection, Dharmakīrti replies as follows:

PVin 3 5,1–2:

nanu tathāpy asiddhir hetoḥ pratijñāyāś cābhyupetādibādhā, svayam abhyupagamāntarāvasthānāt. na, parīkṣākāle kasyacid anabhyupaga-māt.

[Objection:] Even then, the logical reason is unestablished and the thesis is invalidated by that which is accepted [by the proponent], etc., because [the proponent] bases himself on a different acceptance.

[Dharmakīrti:] This is not the case. For, at the time of the critical examination (parīkṣā) [of dogmatic ideas], any particular [dogmatic views] are not accepted [as its basis].

According to Dharmakīrti, a prasaṅga argument is used for the critical examination of a dogmatically accepted notion and any critical examination should be done without relying on particular tenets. In other words, a critical examination only concerns the logical consequence that is necessarily derived from certain characteristics postulated by the opponent as belonging to the subject of the thesis. Therefore, in a prasaṅga argument, the ontological status of the subject of the thesis is left out of consideration.

As a result, the purpose of a prasaṅga argument is merely to show the vyāpti relation.

PVin 3 5,7–8:

nāpy asiddhyādayaḥ, yady evam idam api syān na vobhayam iti dharmayoḥ sambandhopadarśanāt. ekāntaparigrahe syād eṣa doṣaḥ.

Moreover, there is no [fault of] the unestablished etc. because [in the prasaṅga argument] the relationship between two properties is shown as follows: if X were the case, then Y would also follow; or [otherwise, if the latter is not accepted], then both [X and Y] could not exist. [But if] the firm conclusion (ekānta) is grasped [through the prasaṅga argument], such a fault would occur.

As far as the purpose of the prasaṅga argument is restricted to show the vyāpti relation, there is no fault of āśrayāsiddha because in this case the first characteristics of a proper logical reason, i.e., pakṣadharmatva, is not needed. But if the prasaṅga argument is set forth for proving a property with regard to the subject of the thesis, then since the existence of the subject of the thesis is not accepted by the proponent himself, the fallacy of āśrayāsiddha cannot be avoided.

In this way, although he basically follows Dignāga’s method of hypothetical assumption, Dharmakīrti, by restricting the purpose of the prasaṅga argument to present the vyāpti rela-tion, can successfully explain the reason why the problem of pakṣadharmatva is eliminated in the case of the prasaṅga argument.

2.4 Dharmakīrti’s theory of prasaṅga and the relationship between his three methods for avoiding āśrayāsiddha

Just after the passage in PVin 3 cited above, Dharmakīrti seems to link the principle of the conceptual subject or the method of paraphrase to the prasaṅga argument.

PVin 3 5,8–6,1:

na vā sati hetau, yuktiprāptasyāvaśyaṃ parigrahārhatvāt. asati tu hetau maulasya hetor vyāpyavyāpakabhāvasādhanaprakāra eṣaḥ, na viparyayasādhanam, hetor apramāṇatvāt.

Or no [fault would occur] when the proper logical reason exists [in the prasaṅga argument] because what is obtained by rational reasoning is nec-essarily grasped [through such a logical reason]. … On the contrary, when the proper logical reason does not exist [in the prasaṅga argument], this (i.e., the prasaṅga argument) is a type of [reasoning] which proves, for the original logical reason (i.e., the reason in the (Prayoga 3)), the relationship between what is pervaded and what pervades. [But it is] not a proof of the opposite [of the opponent’s view] because [its] logical reason is not [established by] valid cognition.

In the first sentence of this passage, Dharmakīrti suggests that there is a possibility of the transformation from the prasaṅga argument to a proper proof (sādhana).

When the vyāpti relation between the logical reason and the property to be proved is acknowledged, one can turn the reasoning into a proper proof (sādhana), if both the proponent and the opponent accept the following two things:

1) the existence of the subject of the thesis,

2) the logical reason’s being a property of the subject.

If one of these two conditions is not fulfilled, the reasoning remains a prasaṅga argument that does not prove anything with regard to the subject, but simply demonstrates the vyāpti relation. But if these two conditions are fulfilled, the reasoning becomes a proper proof (sādhana) and hence no fault of asiddha occurs.

To fulfil the first of these two characteristics, the principle of conceptual subject or the method of paraphrase must be used. When we take (Prayoga 3) as an example, the method of paraphrase has probably been applied. That is, when the subject of the thesis “a single entity” (eka) is not taken to mean the universal (sāmānya) which occurs in several things, as the Nyāya-Vaiśeṣika opponent insists, but just a single entity like a form (rūpa), then the existence of the subject of the thesis is acceptable for both the Buddhist proponent and the Nyāya-Vaiśeṣika opponent. Therefore, it can be said that in Dharmakīrti’s system of logic, the principle of conceptual subject and the method of paraphrase play an important role also for the conversion from the prasaṅga argument to a proper proof.

In this connection, the relationship between the principle of conceptual subject and the method of paraphrase must be considered. If we compare these two, it is obvious that the former has a wider range of application. To wit, when the proponent proclaims the non-existence of the subject (i.e., “X does not exist”), the latter cannot be applied to the subject because in the case of the method of paraphrase the subject “X” is paraphrased by a thing that is accepted by everybody as existing externally. Since the principle of conceptual subject alone seems enough to avoid the problem of āśrayāsiddha, why then does Dharmakīrti add the method of paraphrase? Although Dharmakīrti does not address this problem, it is likely that he restricts the application of the principle of conceptual subject to negating the existence of metaphysical things such as pradhāna, etc. in the form of “pradhāna does not exist,” using the method of paraphrase as widely as possible. Otherwise the inference-for-others (parārthānumāna) might lose touch with external objects.

3. Dharmakīrti and his successors on prasaṅgaviparyaya and another way to avoid the problem of āśrayāsiddha

As I have mentioned above, a prasaṅga argument is formulated by the proponent on the basis of the opponent’s assumption with regard to some subject. Therefore, even after the existence of the subject of the thesis is accepted, the second condition for avoiding the problem of asiddha, i.e., the logical reason’s being a property of the subject, should not be fulfilled. However, in the case of (Prayoga 3) the logical reason “being devoid of another essential property which is not obstructed (avyavacchinna) by the unification (saṃsarga) with a single substance restricted by a particular place, time and state,” i.e., “being devoid of multiplicity” (*anekatvaviraha) is accepted as a property of the subject of the thesis, i.e., a single entity, by both the Buddhist proponent and the Nyāya-Vaiśeṣika opponent. Therefore, it is understood that this (Prayoga 3) is more than just a normal prasaṅga argument. That is, this prasaṅga argument has already undergone some alterations. This is the process, I believe, that one can understand as Dharmakīrti’s formulation of prasaṅgaviparyaya.

Let’s present the vyāpti relation of property “A” by the property “B” as “A → B,” and describe the establishment of the vyāpti with regard to a subject “S” as “S: A → B.”

Although Dharmakīrti himself did not explain the structure of prasaṅga and prasaṅgaviparyaya, we can describe it, following his commentators, as follows (here the sign “¬” means negation):

prasaṅga S: A → B
prasaṅgaviparyaya S: ¬B → ¬A

If this description is applied to the (Prayoga 3), the main structure of the (Prasaṅga 3) would be expressed as follows:

(Prayoga 3–1) eka : *anekatvavirahanānyena yoga

This structure of (Prayoga 3) corresponds perfectly to the prasaṅgaviparyaya described by Dharmottara, if the expressions anyena yoga (i.e., no negation of nānyena yoga) and anekavṛtti are regarded as having the same meaning:[15]

(Dharmottara’s prasaṅga) sāmānya : anekavṛttitvaanekatva
(Dharmottara’s prasaṅgaviparyaya) sāmānya : anekatvavirahanānekavṛttitva

Moreover, Prajñākaragupta also gives a similar interpretation of (Prayoga 3) as his second interpretation:[16]

(Prajñākaragupta’s prasaṅga) sāmānya [= rūpādi]: vyāpitvanaikavyaktiniṣṭha-tayopalambha
(Prajñākaragupta’s prasaṅgaviparyaya) sāmānya [= rūpādi]: ekavyaktiniṣṭhatayo-palambhaavyāpitva

Although these two commentators differ in how they express the logical reason in the prasaṅgaviparyay, i.e., Dharmottara uses a negative expression and Prajñākaragupta uses an affirmative expression, the contents conveyed by both are roughly the same. Therefore, it can be said that both commentators understand Dharmakīrti’s (Prayoga 3–1) as an example of prasaṅgaviparyaya.

When we take their interpretations into consideration, it can be seen that Dharmakīrti’s (Prayoga 3) consists of the following prasaṅga and prasaṅgaviparyaya:

(Dharmakīrti’s prasaṅga) eka : anekavṛtti → [*anekatva]
(Dharmakīrti’s prasaṅgaviparyaya) eka : *anekatvavirahanānyena yoga (= Prayoga 3–1)

That is, (Prayoga 3) describes the following process of reasoning:

Step 1 (prasaṅga): Following the Nyāya-Vaiśeṣika view, the property anekavṛtti is assumed with reference to the subject eka, and from it anekatva is logically derived although it is not stated in (Prayoga 3).

Step 2 (prasaṅgaviparyaya): But it is not the case that this anekatva is ac-cepted by both the Buddhist proponent and the Nyāya-Vaiśeṣika opponent, and it should be negated because it contradicts the subject eka. Therefore, anekatvaviraha is assumed to be a property of the subject, as is accepted by both. From this nānyena yoga, which for the opponent is an undesired consequence, is proved.

Despite the fact that the main structure of the (Prayoga 3) is a prasaṅgaviparyaya, which is in fact a proper proof (sādhana), Dharmakīrti presents this (Prayoga 3) as an example of a prasaṅga argument. Perhaps the reason for this is that he thinks prasaṅgaviparyaya can also be called prasaṅga in that both bring undesired consequences for the opponent.

Be that as it may, forming a prasaṅgaviparyaya is a method for ensuring that the logical reason is a property of the subject of the thesis, enabling Dharmakīrti, sometimes in connection with the principle of conceptual subject or the method of paraphrase, to transform a prasaṅga argument into a proper proof.

Now I would like to mention briefly how Dharmakīrti’s successors try to solve the problem of āśrayāsiddha. In commenting on (Prayoga 3), Dharmottara presents another way of avoiding the problem of āśrayāsiddha. According to him, if the logical reason consists of non-existence (abhāva), it is established in the subject of the thesis, as for example the universal, whose existence is not accepted by the Buddhist proponent.[17] This fourth way of avoiding the problem can be described as the “method of simple negation.”[18] As earlier research has shown, this method is also adopted by Devendrabuddhi and Śākyabuddhi in their commentaries on PV 4.136–148.[19]

Dharmottara presents this view on the basis of the following statement of Dharmakīrti:

PV 1.26ab (≈ PVin 2 100,8):

tasmād vaidharmyadṛṣṭānte neṣṭo’vaśyam ihā-śrayaḥ /

Therefore, a (factual) basis (āśraya) [i.e., a real locus] in the case of the dissimilar example is not assumed to be necessary here [in the tradition that follows Dignāga]. (Steinkellner 2004: 236)

Since Dharmakīrti’s statement is based on Dignāga’s explanation of vaidharmyadṛṣṭānta in PS(V) 4.3bcd,[20] the “method of simple negation” can be traced back to Dignāga. However, since neither Dignāga nor Dharmakīrti mention this method in the context in question, it might be said that in some way this method was invented by the commentators to solve the problem of āśrayāsiddha. In contrast to Dharmottara, etc., Prajñākaragupta, when explaining (Prayoga 3) in his commentary on PV 4.12, does not adopt this method; instead, he adopts the method of paraphrase, that is, he paraphrases the subject of the thesis “universal” (sāmānya) as “form, etc.” (rūpādi).[21] In his commentary on PV 4.136–148, he uses the principle of conceptual subjects and then, as a second interpretation, again employs the method of paraphrase.[22] It is thus likely that he chooses the method of paraphrase as his final view. Therefore, as far as the problem of āśrayāsiddha is concerned, Prajñākaragupta’s view, when compared to that of the other commentators, is more similar to Dharmakīrti’s.

4. Concluding remarks

As seen above, whereas in the NMu Dignāga avoided the problem of āśrayāsiddha by applying the principle of conceptual subject, in the PS he avoids the same problem by employing the method of hypothetical assumption. This raises the question: Why doesn’t Dignāga use the strategy employed in the NMu, i.e., the principle of conceptual subjects, in the PS as well? It is likely that the difference in his two works of how the thesis is defined has played a role in his shift of position.

The two definitions run as follows:

NMu:

tatra svayam / sādhyatvenepsitaḥ pakṣo viruddhārthānirākṛtaḥ //[23]

PS 3.2:

svarūpeṇaiva nirdeśyaḥ svayam iṣṭo’nirākṛtaḥ / pratyakṣārthānumā-nāptaprasiddhena svadharmiṇi //[24]

Comparing these two, one of the most significant differences is the word svadharmin in PS 3.2d. This word seems to stipulate that, in the proponent’s own thesis, not only the property to be proved, but also the subject of the thesis (svadharmin) should not be opposed. Therefore, once a thesis is defined in this way, a Buddhist proponent cannot place pseudo-entities, such as pradhāna, into the position of the subject of his own thesis, even if he intends to prove the non-existence of such a thing.

But this does not answer the question of why Dignāga does not repeat the principle of conceptual subjects in the PS, for he could have also defined the thesis without using the term svadharmin. Unfortunately, no decisive clue can be found to solve this question.

However, some possible reasons can be raised.

(i) Weakness of the NMu’s argument against the Sāṅkhya’s proof

In the NMu, even though the pakṣadharmatva of the logical reason in (Prayoga 1) is ensured by employing the principle of the conceptual subject, the Sāṅkhya does not accept the logical reason as necessarily negating the existence of pradhāna since for the Sāṅkhya, pradhāna is, by definition, not perceived, but is to be inferred.

For this reason, Dignāga may have thought that to negate pradhāna, it is easier to refute the Sāṅkhya’s proof than to explain (Prayoga 1) more convincingly to the Sāṅkhya.

(ii) Avoiding unnecessary expansion of his ontological framework

In the NMu, Dignāga posits a kind of existential status on pradhāna by using the principle of conceptual subjects. But in the PS, he provisionally accepts the Sāṅkhya’s tenet of the existence of pradhāna (along with the vyāpti relation). Basing himself on this provisionally accepted tenet, he points out the defectiveness of the Sāṅkhya’s proofs. He may have thought this strategy to be more preferable because, unlike the case of the principle of conceptual subjects, it does not need any expansion of his own ontological framework. It is likely that one or both of these reasons made Dignāga shift his position.

Unlike in the case of Dignāga, Dharmakīrti is able to overcome these two weak points by his new anupalabdhi theory and apoha theory. Therefore, the principle of conceptual subjects again becomes available as a way to avoid the problem of āśrayāsiddha. He not only inherits two methods from Dignāga, but also adds a third, i.e., the method of paraphrase. Then he tries to connect these three methods in the discussion of prasaṅga and prasaṅgaviparyaya. When the prasaṅga argument is used for refuting the opponent’s view, thanks to the method of hypothetical assumption, the problem of āśrayāsiddha does not occur. But when the proponent wants to formulate a proper proof, i.e., prasaṅgaviparyaya, from the prasaṅga argument, the method of paraphrase or the principle of conceptual subject is employed in order to avoid the fallacy of āśrayāsiddha in the case of the proponent not accepting the existence of the subject of the thesis. After that, the contrapositive (viparyaya) of the vyāpti relation in the prasaṅga argument is made in order to avoid the logical reason being an asiddha. As I have shown elsewhere, Dignāga has already indicated the possibility of reformulating a prasaṅga argument into prasaṅgaviparyaya.[25] It is, however, Dharmakīrti who establishes this theory by using the methods for avoiding the problem of āśrayāsiddha.

Note

After having finished my paper, I received Yoshimizu Chizuko’s article “Dharmakīrti’s Statement of Consequence (prasaṅga) in the Third Chapter of the Pramāṇaviniścaya” (Yoshimizu 2016). She takes, contrary to my understanding, the (Prayoga 3) to be a prasaṅga argument. The difference between our interpretations comes from how we understand the role of anekavṛtti in the (Prayoga 3). While Prof. Yoshimizu takes both eka and anekavṛtti as being properties of the subject of the thesis (Yoshimizu 2016: 1248), I understand eka as the subject of the (Prayoga 3) and regard anekavṛtti as just a subsidiary element of eka that has been added to explain the opponent’s assumption. In the passage discussed above, anekavṛtti is presented in apposition to eka. However, as Prof. Yoshimizu has also pointed out (Yoshimizu 2016: 1253, note 5), Dharmakīrti refers to almost the same argument with (Prayoga 3) in the subsequent part without referring anekavṛtti and gives only eka as its subject. See PVin 3 6,6–7: ekasya tu yathoktasvabhāvāntaravirahopagamād eva bhinnadeśādibhir yogābhāvaḥ.

References and abbreviations

D Derge edition of Tibetan Tripiṭaka.

Iwata 1993 T. Iwata, Prasaṅga und Prasaṅgaviparyaya bei Dharmakīrti und seinen Kom-mentatoren. Vienna 1993.

Iwata 1999 T. Iwata, On the interpretations of the subject (dharmin) of the inference negating invariable entities in Dharmakīrtian logic. In: Dharmakīrtis Thought and Its Impact on Indian and Tibetan Philosophy, ed. S. Katsura. Vienna 1999, 156–172.

K PS(V) translated by Kaṇakavarman and Dad pa’i shes rab.

Kataoka 2010 K. Kataoka, Mittsu no apoha setsu: Dharmottara ni itaru model no hensen. South Asian Classical Studies 5 (2010) 251–284.

Katsura 1977 S. Katsura, Inmyōshōrimonron Kenkyū (1). Hiroshima Daigaku Bungakubu Kiyō 37 (1977) 106–126.

Katsura 1978 S. Katsura, Inmyōshōrimonron Kenkyū (2). Hiroshima Daigaku Bungakubu Kiyō 38 (1978) 110–130.

Katsura 1981 S. Katsura, Inmyōshōrimonron Kenkyū (4). Hiroshima Daigaku Bungakubu Kiyō 41 (1981) 62–82.

Katsura 2004 S. Katsura, Pakṣa, Sapakṣa and Asapakṣa in Dignāga’s Logic. Hōrin 11 (2004) 119–128.

Katsura 2009 S. Katsura, Rediscovering Dignāga through Jinendrabuddhi. In: Sanskrit manuscripts in China, ed. E. Steinkellner. Beijing 2009, 153–166.

Kitagawa 1965 H. Kitagawa, Indo koten ronrigaku no kenkyū: Jinna (Dignāga) no taikei. Tokyo 1965.

NMu Nyāyamukha. See Katsura 1977, 1978 and 1981.

P Peking edition of Tibetan Tripiṭaka.

PS(V) Pramāṇasamuccaya(vṛtti). See n. 6.

PSṬ Jinendrabuddhi’s Pramāṇasamuccayaṭīkā. See n. 6. PV 1 Pramāṇavārttika, chapter 1, ed. R. Gnoli. Roma 1960. PV 4 Pramāṇavārttika, chapter 4. See Tillemans 2000.

PVin 3 Pramāṇaviniścaya, chapter 3, ed. P. Hugon and T. Tomabechi. Beijing/Vienna 2011.

PVinṬ Dharmottara’s Pramāṇaviniścayaṭīkā.

PVSV See PV 1.

Steinkellner 2004 E. Steinkellner, The Early Dharmakīrti on the Purpose of Example. In: The role of the example (dṛṣṭānta) in classical Indian logic, ed. S. Katsura and E. Steinkellner. Vienna 2004, 225–250.

T Taishō Shinshū Daizōkyō. Tokyo 1924–1932.

Tillemans 1999 T. J. F. Tillemans, Scripture, Logic, Language: Essays on Dharmakīrti and his Tibetan Successors. Boston 1999.

Tillemans 2000 T. J. F. Tillemans, Dharmakīrtis Pramāṇavārttika: An annotated trans-lation of the fourth chapter, vol. 1 (kk. 1–148). Vienna 2000.

V PS(V) translated by Vasudararakṣita and Seng rgyal.

Watanabe 2013 T. Watanabe, Dignāga on Āvīta and Prasaṅga. Journal of Indian and Buddhist Studies 61:3 (2013) 1229–1235.

Yao 2009 Z. Yao, Empty Subject Terms in Buddhist Logic: Dignāga and his Chinese Commentators. Journal of Indian Philosophy 37:4 (2009) 383–398.

Yoshimizu 2013 Ch. Yoshimizu, Reasoning-for-others in Candrakīrti’s Madhyamaka thought. Journal of the International Association of Buddhist Studies 35:1–2 (2013) 413–444.

Yoshimizu 2016 Ch. Yoshimizu, Dharmakīrti’s Statement of Consequence (prasaṅga) in the Third Chapter of the Pramāṇaviniścaya. Journal of Indian and Buddhist Studies 64:3 (2016) 1246–1254.

Footnotes and references:

[1]:

See Tillemans 1999: 174–176.

[2]:

Just prior to (Prayoga 1), Dignāga refers to another reasoning (prayoga) that also has pradhāna as the subject of the thesis: [Thesis:] pradhāna exists (asti pradhānam); [Reason:] because in various individuals homology (anvaya) [concerning the three constituents, i.e., pleasure (sukha), pain (duḥkha), and confusion (moha)] is observed (bhedānām anvayadarśanāt). (NMu [Katsura 1978: 110] T. 1628 1b29: 有成立最勝爲有。現見別物有總類故) His answer to this is: The thesis should be formulated as “The various individuals certainly possess one and the same cause [i.e., pradhāna].” [If so,] the [existence of] pradhāna is not established [directly]; hence, there is no error [of the violation of the above-mentioned general rule]. (NMu [Katsura 1978: 110] T. 1628 1c1–2: 此中但立別物定有一因爲宗、不立最勝、故無此失。) Dignāga criticizes this prayoga by pointing out that its thesis is not formulated properly. The subject of the thesis should be “various individuals,” since otherwise the logical reason “homology” cannot be a pakṣadharma. Although earlier studies (Tillemans 1999: 175 and Yao 2009: 386–387) regard this explanation by Dignāga as similar to the method of paraphrase applied by Dharmakīrti in PV 4.136–148, there is a fundamental difference. While Dharmakīrti’s method, as will be explained below, § 2.2, just rephrases the subject of the thesis pradhāna into “pleasure, etc.,” the existence of which is accepted by the Buddhists, Dignāga’s method mentioned above changes the subject from pradhāna to “various individuals” in order to give pakṣadharmatva to the logical reason “homology.” Thus Dignāga’s critique is not aimed at avoiding the problem of āśrayāsiddha.

[3]:

See NMu (Katsura 1978: 110) T. 1628 1c1: 或立爲無。不可得故。

[4]:

See NMu (Katsura 1978: 110) T. 1628 1c2–4: 若立爲無、亦假安立不可得法、是故亦無有有法過。

[5]:

Words in roman typeface are attested in Jinendrabuddhi’s commentary (PSṬ) or in fragments, whereas those in italics have been reconstructed. The Sanskrit reconstruction of the PS and PSV used in this paper is a result of the PSṬ seminar organized and led by Prof. Shōryū Katsura at Ryukoku University. Through his kindness, I have been able to use their reconstruction. As for the number of kārikās in PS 3, it is to be noted that recently the editorial team of the PSṬ 3 has adopted a different way of counting them than was hitherto widely adopted by scholars, e.g., Kitagawa 1965 and Katsura 2009, etc. While the earlier counting regards k.7 as consisting of only a half verse (i.e., 7ab), the new counting includes the following half verse into k.7. That is, in the new counting k.7 consists of 7ab and 8ab of the earlier counting. Therefore, the new number of the kārikās in PS 3 shifts back by half from k.7. In this paper I use the new way of counting. The reading of k. 16cd–17ab I have adopted here is a modified version of that presented in Katsura 2009: 158; there it reads as follows: prasaṅgo’pakṣadharmatvāt pūrvatropagame sati / hetupratijñayos teṣāṃ (or tatra) doṣoktyā duṣaṇaṃ gatam // Cf. K (P129b3–6): de ltar na’ga’ zhig gtso bo ni yod pa ma yin te / longs spyod pa po dang longs spyod par bya ba dag yon tan dang yon tan med pa’i (yon tan med pa’i em. [see P129b8–130a1]: yon tan can gyi P) rang bzhin du thal ba’i phyirzhes thal bar’gyur ba gzhan brjod par byed pa de ci ltar bsal te’ongs pa ma yin zhe na / ’di ni bsal te’ongs pa ma yin te / gang gi phyir / thal’gyur phyogs chos can min phyir // sngon du khas blangs yod na ni // rtags dang dam bca’ gzhan dag la // skyon brjod sun’byin shes par bya //; V (D44b1–3, P47a8–47b2): yang’dir thal ba gzhan brjod pa gtso bor yod pa ma yin te / longs spyod pa po dang longs spyad par bya ba dag yon tan dang yon tan can nyid du thal bar’gyur ba’i phyir ro zhes pazhes bya ba de ji ltar bsal te’ongs pa yin’di ni bsal te’ongs pa ma yin te / gang gi phyir / thal’gyur phyogs chos can min phyir // khas blangs sngon du song ba las // de gzhan (de gzhan P Vṛtti: de bzhin D Vṛtti, D Kārikā) rtags dang dam bca’ yi (yi D Vṛtti, P Vṛtti: yis D Kārikā) // skyon brjod sun’byin du shes bya //

[6]:

Dignāga refutes the Sāṅkhya’s proof of the existence of pradhāna through the following process: First the Sāṅkhya puts forth (Prayoga A) for proving the existence of pradhāna. (Prayoga A) Pradhāna exists because in various individual things homology (anvaya) [concerning three constituents, i.e., sukha, duḥkha and moha] is observed. (asti pradhānam, bhedānām anvayadarśanāt.) (2) In order to ensure the pakṣadharmatva of the logical reason “homology,” this (Prayoga A) is rewritten as (Prayoga A’). Various individual things possess one and the same cause because homology is observed. (bhedānām ekakāraṇatvam, anvayadarśanāt.) (3) The Sāṅkhya adds the following reasoning in order to prove (Prayoga A’), because otherwise the opponent, such as Buddhists, does not accept that pleasure etc. are commonly shared by various individual things. (Prayoga B) In various individual things homology is observed because they have common effects such as joy (prīti). (4) Dignāga’s refutation: if (Prayoga B) is held to be true, the following vyāpti relation must be accepted. [Vyāpti:] Whatever brings about common effects, such as joy, etc., has a homology concerning three constituents. (yasya prītyādi kāryaṃ dṛṣṭam, tasya guṇānvayam.) (5) However, because even a soul (puruṣa), which cannot be accepted in the Sāṅkhya system as consisting of three constituents, can bring about an effect such as joy if it is used as an object of meditation by adherents of the Sāṅkhya, the logical reason “having common effects” in (Prayoga B) would be inconclusive (anaikāntika). (6) Suppose that, on the contrary, the Sāṅkhya does not accept the vyāpti relation in order to avoid the undesired consequence. In this case, because denying the validity of the vyāpti relation results in the denial of the soundness of (Prayoga B), (Prayoga B) does not establish the pakṣadharmatva of the logical reason in (Prayoga A)/(Prayoga A’). And, as a result, the logical reason in (Prayoga A)/(Prayoga A’) does not prove the existence of pradhāna.

[7]:

See Watanabe 2013.

[8]:

See PS 3.13cd–14ab: hetupratijñādvāreṇa yatrāniṣṭiḥ prasajyate // taddvāreṇa prayogāt* sa parihāra itīṣyate / (*sbyor phyir V: thal ba’i phyir K. Following K, Yoshimizu 2013: 433 adopts prasaṅgāt.)

[9]:

This method corresponds to “the principle of propositional attitude,” as it has been called by Zhihua Yao. In his article (Yao 2009: 393–396) he states that this principle was developed by Chinese Buddhist logicians, such as Kuiji, and is merely implied by Dignāga and the Indian Buddhists. However, as we have seen, this principle can clearly be traced back to Dignāga’s description in PS 3.

[10]:

As Shinya Moriyama has pointed out in his presentation at the 17th IABS conference (21 Aug. 2014), Dharmapāla (T. 1571 188b12–14) also applies the same triple typology. The difference between Dharmapāla and Dharmakīrti is that the former applies it to the classification of the logical reason, but the latter to the conceptual construction in general.

[11]:

According to Dharmakīrti, although in reality the internal image held by the speaker must differ from the internal image held by the listener, these two are regarded as the same. For this fundamental problem of Dharmakīrti’s theory of apoha, see Kataoka 2010: 267–269.

[12]:

See PV 4.141abcd’: yathā parair anutpādyāpūrvarūpaṃ na khādikam / sakṛc chabdādyahetutvād ityukte. “For example, when [the Buddhist] states that space, etc. do not have a novel nature unproduced by other [conditions] because they are not causes for [producing their qualities such as] sound, etc. all at once …” (Tr. Tillemans 2000: 202.)

[13]:

See PV 4.141d’–142ab: prāha dūṣakaḥ // tadvad vastusvabhāvo’san dharmī vyomādir ity api / “…then the [Vaiśeṣika] adversary might say that in a similar way the subject, space, etc., would also not have the nature of a real entity.” (Tr. Tillemans 2000: 202.)

[14]:

See PV 4.142cd: naivam iṣṭasya sādhyasya bādhā kācana vidyate // “In this fashion [even though the subject is invalidated], there is in fact no invalidation of the intended [proposition] to be proved (sādhya) at all.” (Tr. Tillemans 2000: 202.)

[15]:

See Iwata 1993: 50.

[16]:

See Iwata 1993: 72.

[17]:

PVinṬ (Ms 6a2–3): sa ca vyāpakābhāvaḥ sāmānyādāv abhāve siddha eva; ibid. (Ms 7a5–6): anekatvā-bhāvamātraṃ cāsaty api sāmānye siddham. See Iwata 1993: 54–56, Iwata 1999: 161.

[18]:

Tillemans 1999 does not distinguish this method from the “principle of conceptual subjects.” However, as Yao (2009: 391) has pointed out, these are two different things.

[19]:

Tillemans 1999: 173–174 and Iwata 1999: 160–162.

[20]:

PS(V) 4.3bcd: sādhye’sati tv asat / (4.3b) vaidharmyeṇa tv avaśyam sādhyābhāve tasyaiva hetor abhāvo darśaitavyaḥ. paryudāso niṣedhaś ca tathā sati vilakṣaṇe //3// tathā ca sati pūrvatra paryudāsaḥ, aparatra tu prasajyapratiṣedha ity uktam. evaṃ ca nityānabhyupagame’pi vaidharmyadṛṣṭāntaḥ siddhaḥ. Cf. NMu §5.1 (Katsura 1981: 63): 由是雖対不立実有太虚空等、而得顕示無有宗処無因義成。

[21]:

See Iwata 1993: 69–73.

[22]:

See Tillemans 1999: 177–180, Iwata 1993: 63–81, and Iwata 1999: 163–164.

[23]:

NMu (Katsura 1977: 109) T. 1628 1a8–9: 是中、唯隨自意樂爲所成立説名宗。非彼相違義能遣。“Of them (viz. thesis, reason and example), a valid thesis (pakṣa) is [precisely] one which is intended (īpsita) by [the proponent] himself as something to be proved (sādhya) and which is not opposed (anirākṛta) by incompatible states of affairs.” (Tr. Katsura 2004: 119.)

[24]:

“[A valid thesis] is one which is intended (iṣṭā) by [the proponent] himself (svayam) as something to be stated (nirdeśya) in its proper form alone (svarūpeṇaiva) [i.e., as something to be proved (sādhya)]. With regard to [the proponent’s] own subject (svadharmin), it is not opposed (anirākṛta) by perceptible objects (pratyakṣārtha), by inference (anumāna), by authorities (āpta) or by what is commonly recognized (prasiddha).” (Tr. Tillemans 2000: 4–5.)

[25]:

See Watanabe 2013.

Like what you read? Consider supporting this website: