Unestablished Subjects in Arguments by Consequence

[Full title: Universals, Demons’ Pots, and Demons’ Permanent Pots: Phya pa Chos kyi seng ge on Unestablished Subjects in Arguments by Consequence by Pascale Hugon]

The work on this paper has been generously supported by the Austrian Science Fund (FWF) in the context of the project P23422-G15 “Early bKa’ gdams pa scholasticism.” I wish to thank my colleague Toshikazu Watanabe for stimulating discussions on the topic, and Katharine Apostle for improving my English.

1. Introduction

According to Dharmakīrti, prasaṅga arguments–i.e., arguments that rely on a logical reason that is merely accepted by the opponent to draw an unacceptable conclusion–are legitimate means of proof provided that they are indicative of a correct, i.e., triply characterized, logical reason on which the reverse form of the prasaṅga (the prasaṅgaviparyaya) is based.^[1] The relevant passage of the Pramāṇaviniścaya (hereafter: PVin) has been subject to conflicting interpretations by Indian commentators who notably disagree about the nature of the example that Dharmakīrti gives in this context–namely, whether it is meant to illustrate a prasaṅga argument or its reverse form–and about the status to ascribe to the subject in this example–i.e., whether the subject “universal” is a real universal (as upheld by the intended addressee of the argument) or a non-entity (as accepted by the Buddhist proponent).^[2] The second point is crucial for the acceptance of the prasaṅga as a probative argument insofar as the validity of the reverse form would be threatened by the non-existence of the subject. Indeed, the non-existence of the subject would prevent the fulfillment of a required characteristic of the logical reason, being a property of the subject (pakṣadharmatva)–a fallacy known as that of the “unestablished basis” (āśrayāsiddha).

Whereas this fallacy allows the Buddhists to discard proofs by opponents upholding, for instance, the reality of space (ākāśa), primordial nature (pradhāna) or universals (sāmānya), it also threatens their own arguments aimed at the refutation of such pseudo-entities. The logical reason in the Buddhist proof would indeed also fail to satisfy the requirement of pakṣadharmatva.^[3] The status of the subject poses a similar problem in the Madhyamaka proof establishing the emptiness of all things,^[4] and also impacts arguments for the establishment of the negative entailment of logical reasons, typically in the proof of momentariness. Various answers to the problem of the unestablished basis have been developed in the Buddhist tradition.^[5] The question of the subject in the context of prasaṅga arguments was taken up in particular by Dharmottara and Prajñākaragupta.^[6]

The present paper deals with the views on this topic of the Tibetan logician Phya pa Chos kyi seng ge (1109–1169).^[7] Phya pa, as can be expected, was influenced by the discussions of Indian thinkers but he nevertheless came up with his own creative method to deal with this issue. The examination of his views also sheds light on later Tibetan developments that obviously heavily borrowed from his works. In what follows I examine Phya pa’s commen-tary on the prasaṅga passage of Dharmakīrti’s Pramāṇaviniścaya (’Od zer 149a5–150b3) to establish his take on the question of the nature of Dharmakīrti’s example and on the status of the subject. I then situate his interpretation within the framework of his overall theory of arguments by consequence (thal’gyur)–of which we find two parallel versions in his commentary on the PVin (’Od zer 145a2–149a5) (preceding his commentary on the prasaṅga passage) and in his epistemological summa (Mun sel 83a1–95b5)–and relate this issue to discussions relevant to the problem of the “unestablished basis” (āśrayāsiddha) in the broader context of inference.

Before turning to Phya pa’s view, it will be useful to provide some references to the solutions developed by Indian thinkers to the problem of the unestablished basis.

Some methods developed by Indian thinkers to solve problems caused by unestablished subjects

I refer the reader to Watanabe’s paper in this volume for a detailed discussion of the solutions developed by Dignāga and Dharmakīrti and their link to the context of prasaṅga, as well as for further references to primary and secondary literature. I list below (adopting Watanabe’s terminology) some solutions that will be mentioned in my discussion of Phya pa’s position:

The ‘Method of Conceptual Subject:’ The actual or intended subject is the conceptual representation of the entity in question. This method is adopted by Dharmakīrti in the passage of PVin 3 dealing with the refutation of the Sāṃkhya’s primordial nature (to which I will refer as the “pradhāna passage”).^[8] The core idea can be traced to Dignāga’s discussion on the refutation of pradhāna in the Nyāyamukha, where he describes the subject as “imagined” (kalpita). Dharmakīrti makes clear in what way one can prove the non-existence of pradhāna without negating the subject in a way that would make the proof liable to a fallacy: what is being proven is not that the subject qua concept is inexistent, but that this concept does not have a real substratum.^[9] The concept itself is well established for both debaters and not liable to refutation.^[10] It is not actually the concept that is targeted by the discussion but the substratum (the entity) which, unlike the concept, is apt to fulfill some human goal.

The ‘Method of Paraphrase:’ the reasoning is reformulated in such a way that the actual subject is acceptable to the proponent. This idea, found in PV 4.136–148 in the commentary on the word svadharmiṇi in Dignāga’s definition of the thesis in PS 3.2, is linked to the distinction between nominal subject (kevaladharmin) and actual subject (svadharmin) when discussing the fallacy of the non-established basis as a fallacy of the thesis.^[11] I will refer to this context as the “svadharmin passage.”

The ‘Method of Simple Negation:’ the attribution of negative properties does not require the subject to exist, so no fault occurs as long as the properties attributed to the non-existent subject are simple negations (prasajyapratiṣedha). This method was adopted by Dharmottara in the context of prasaṅga and by Devendrabuddhi and Śākyabuddhi when discussing the svadharmin passage. Prajñākaragupta criticized this position, arguing that it does not resolve the fallacy of the unestablished subject because an inexistent subject cannot be qualified by any property whatsoever.

Iwata (1999) discerns two other methods found in the Madhyamaka works of Śāntara-kṣita and Kamalaśīla along with the Method of the Conceptual Subject and the Method of Simple Negation; they consist in taking the subject to be existent (i) as a convention (vyavahāra) and (ii) as an appearance. In Śāntarakṣita’s Madhyamakālaṅkāra(/vṛtti) and in Kamalaśīla’s Madhyamakālaṅkārapañjikā one finds the idea that the verbal designations of the elements of the inference are based on a subject that appears in the cognition of both disputants–for instance for the subject “sound,” the appearance in the auditive cognition upon hearing a sound.^[12] In these texts the authors restrict the notion of “appearance” to the context of perception and do not deal with non-existent subjects such as pradhāna. The issue is mainly to avoid that the subject becomes unestablished when it is proven to be essenceless.

Kamalaśīla addresses the issue of the inexistent subject in his Madhyamakāloka. In this work he mainly invokes the Method of Simple Negation to deal with such subjects.^[13] But he also draws a parallel between the previous idea of positing as the subject a commonly shared appearance and the refutation of pradhāna in which one posits as a subject “something existing in the mind” (blo la yod pa nyid).^[14] Actually, the idea that the subject can be a mental appearance is well represented in the pradhāna passage of the PVin. Notably, in his prose commentary to PVin 3.53–54 Dharmakīrti reformulates the notion of the concept qua “object of words” (śabdārtha) in terms of “the object that appears to conceptual thought,” an appearance that emerges upon hearing a given word.^[15] This method is thus akin to the Method of Conceptual Subject but emphasizes the mental appearance of the concept rather than the nature of conceptual construction. I will refer to it in terms of ‘Method of Conceptual Appearance.’ As mentioned above, Dharmakīrti clearly indicated that the subject in this form is established for both disputants and not liable to refutation. Among Madhyamaka authors, Jñānagarbha similarly points out that appearances are not negated insofar as they are established to be experienced.^[16]

1. Phya pa on the prasaṅga passage of the Pramāṇaviniścaya

1.1. Phya pa’s interpretation of the example in the prasaṅga passage of the Pramāṇa-viniścaya

In the prasaṅga passage of the PVin, Dharmakīrti introduces with the words “for example” (yathā) an argument against Naiyāyika and Vaiśeṣika opponents who accept the reality of universals. This argument is complex enough that one cannot avoid, for its very translation, relying on a commentarial interpretation.^[17] My tentative translation follows here the un-derstanding of the role of various portions of the Sanskrit sentence by Dharmottara, an interpretation which is also adopted by Phya pa.^[18]

a→A single thing that occurs in many←a, b→because it is devoid of another essential property which is not characterized [as] mixed (or: which is not separate, due to being mixed) with a single instance determined in location, time and status,←b c→cannot be connected with something else that differs in location, etc.←c

(PVin 3 4,4–7:^b→deśakālāvasthāviśeṣaniyataikadravyasaṃsargāvyavacchi-nnasvabhāvāntaravirahād←b a→anekavṛtter ekasya←a c→na deśādiviśeṣavatā-nyena yogaḥ←c)

Phya pa notably follows the broad lines of Dharmottara’s interpretation on the following points:

1. This example is taken to illustrate the reverse form of a prasaṅga (i.e., a prasaṅgavi-paryaya);^[19]

2. The portion marked as “^a→…^←a” represents the subject, that marked as “^b→…^←b” represents the logical reason of the prasaṅgaviparyaya, and that marked as “^c→…^←c” the property to be proven of the prasaṅgaviparyaya;

3. The type of the logical reason of the prasaṅgaviparyaya consists in the non-appre-hension of the pervader (vyāpakānupalabdhi);

4. The type of the logical reason of the prasaṅga is an essential property (svabhāva).

When commenting on this passage of the PVin, Dharmottara and Phya pa also similarly reformulate the subject, logical reason and property to be proven (respectively, the derived conclusion), as indicated in the schematic representation provided below.^[20]

According to them, the above passage should be understood as the prasaṅgaviparyaya:
A universal does not occur in many, because it is not multiple.
which is induced by the prasaṅga (not stated by Dharmakīrti):
Because (according to you) a universal occurs in many, it follows that it is multiple.

		Dharmottara	Phya pa
prasaṅga
Subject	universal	sāmānyādi	spyi cig
Reason	occurrence in many	anekavṛttitva	gsal ba du ma dang’brel pa
Derived property	multiplicity	anekatva	du ma
Dharmakīrti’s example (= prasaṅgaviparyaya)
Subject	universal	sāmānya	spyi cig
Reason	non-multiplicity	na…aneka = anekatvābhāvamātra, nānātvābhāva	du mas stong pa
Property to be proven	non-occurrence in many	*na anekavṛttitva, aneka-vṛttitvasyābhāva	du ma dang’brel pas stong, gsal ba du ma dang ma’brel ba

1.2 Phya pa on the subject “universal” in Dharmakīrti’s example

A key issue in Dharmakīrti’s example is the nature of the subject “universal.” Indeed, the Buddhist proponent does not accept the reality of universals and the non-existence of the subject would threaten the validity of the logical reason in the reverse form: the fault of “non-establishment of the basis” (āśrayāsiddha) could be invoked. The case under consideration is typical of the drawbacks of the fallacy of the unestablished basis, which Buddhist scholars have attempted to palliate through various methods.

In the prasaṅga passage of the PVin the question of the subject arises indirectly in an objection pointing out that the logical reason of the prasaṅga is unestablished (asiddha), i.e., it does not satisfy the first characteristic of a correct reason–being a property of the subject (pakṣadharmatva). This, is, according to Dharmottara, because the Buddhist proponent does not accept that “a universal occurs in many” insofar as he does not accept that universals exist in the first place.^[21] Although here the non-establishment of the logical reason is a consequence of the non-existence of the subject, the argument appears to be hinting at a more general issue: in any proper prasaṅga, whether the subject is accepted to be existent by both debaters or not, the qualification of the subject by the logical reason is only accepted by the opponent and corresponds to the opposite of what the proponent wishes to establish.^[22] Accordingly, in my understanding, Dharmakīrti’s answer does not touch on the way to deal with an inexistent subject, but first of all invokes the indeterminacy that should prevail in an inquiry pertaining to a given subject. In the given argument, the “universal” should not be taken either as “the Buddhists’ universal” or “the Naiyāyika/Vaiśeṣika’s universal.” Further, Dharmakīrti points to the hypothetical nature of the pakṣadharmatva in the prasaṅga –it is not something that is required to be established by a valid cognition as it is in a direct proof.

The discussion in this context revolves solely around the prasaṅga.^[23] The question of the subject in the reverse form is not addressed by Dharmakīrti, thus leaving the interpreters free to opt for their preferred scenario on this point. Dharmottara applied the Method of Simple Negation in this context (Iwata 1997a: 430). While Phya pa follows the major lines of Dharmottara’s interpretation of the prasaṅga passage and, like Dharmottara, identifies the logical reason and the property to be proven in the prasaṅgaviparyaya as simple negations (using the formulation “void of…”), he does not invoke this as a solution to the potential problem of the unestablished subject. His solution is not influenced by Prajñākaragupta either.^[24]

Phya pa’s solution, when commenting on the prasaṅga passage, is to appeal conjointly to the Method of Conceptual Subject and the Method of Conceptual Appearance, both of which find support in the pradhāna passage of the PVin.

The Method of Conceptual Appearance stands out notably in his commentary on the passage of the PVin referred to above discussing a potential non-establishment of the logical reason:

When one investigates the nature of things, at the time of an initial examination of philosophical tenets, one has not [yet] accepted the Buddhist philosophical tenets and one does not accept the non-Buddhist tenets. Therefore one does not posit as a basis a universal that is either an entity or a non-entity. The mental appearance is the basis. Therefore, even though [the universal] is not established as an entity for the Buddhist and is not established as a non-entity for the non-Buddhist, this mere mental appearance is established for both.^[25]

As for the Method of Conceptual Subject, it is visible in Phya pa’s reference to the subject as the “object of a word” (sgra don, Skt. śabdārtha), i.e., the concept,^[26] on the basis of which the characteristic of the logical reason being a property of the subject (pakṣadharmatva) is established:

Our [i.e., the Buddhist’s] position is correct: That the pakṣadharmatva of [the logical reason] ‘void of multiplicity’ for the object of the word [as it is] accepted is established by reflexive awareness [is explained by Dharmakīrti] in the passage “But the singular thing…” (PVin 6,6–7: ekasya tu…).^[27]

Phya pa’s commentary also refers the subject “the singular thing” as an object of the word for which the property of “appearing as singular” can be established by reflexive awareness.^[28]

Using this joint method Phya pa is able to guarantee the legitimacy of Dharmakīrti’s example in spite of the problematic status of the subject “universal:” whether universals exist in reality or not, debaters who use the term sāmānya have a conceptual representation, the appearance of which can be established by reflexive awareness and cannot be refuted. The Method of Conceptual Appearance and/or Conceptual Subject is successful when dealing with cases where the author wants to avoid the fallacy of the unestablished basis. But such a method appears to have far-reaching consequences. Indeed, if all verbal expressions

generate conceptual representations and just any conceptual representation may qualify as being “established” insofar as its appearance can be established by reflexive awareness, there should be no unestablished bases, and hence no occasion for the fallacy of āśrayāsiddha to arise.^[29] However, as I discuss in what follows, Phya pa preserves this fallacy in the context of inference-for-oneself, and the requirement that the subject must be established also plays a role in the identification of probative consequences (i.e., instances of arguments by consequence whose features correspond to Dharmakīrti’s understanding of a legitimate prasaṅga).

2. Unestablished subjects in Phya pa’s theory of argumentation by con-sequence

2.1 Phya pa’s classification of arguments by consequence

Phya pa’s theory of arguments by consequence includes a much broader range of arguments than the type of prasaṅga taken into account by Dharmakīrti as a legitimate argument. According to Phya pa, arguments by consequence include any argument of the form “Because S is P, it follows that it is Q.”^[30] A distinction is then made between genuine (rnal ma) arguments by consequence and fallacious ones (ltar snang). Fallacious consequences are to be understood as “non-pertinent.” They are arguments to which the opponent is able to retort that he does not accept the premises–he does not accept that S is P and/or that everything that is P is Q, in which case he is not bound to accept the conclusion “therefore S is Q”–or is able to retort that he accepts the conclusion (S is Q), which the proponent intended to be ‘absurd’ or at least unacceptable for the opponent. If the opponent is unable to retort, the consequence is qualified as “genuine.” Genuine consequences are then divided into probative and non-probative consequences. Only the first type, in which Dharmakīrti’s prasaṅga is to be included, amounts to an inference-for-others that proves something. The logical reason of its reverse form satisfies the triple characteristic (trairūpya)–put shortly, the qualification of the subject (pakṣadharmatva) and the pervasion (vyāpti) are ascertained by a valid cognition.^[31]

2.2 The definition of a probative consequence

Phya pa gives as a definition of a probative consequence:

The indication of a logical reason pertaining to a subject in the context such that the pervasion is determined by a valid cognition and the explicit conclusion is eliminated by a valid cognition.^[32]

This definition brings two requirements to the fore. For a consequence of the form “Because S is P, it follows that it is Q” to be probative,

(a) the pervasion of the consequence (P is pervaded by Q) must be established by a valid cognition;

(b) the derived conclusion (S is Q) must be eliminated by a valid cognition.^[33]

These requirements guarantee that the reverse form of the consequence, which has the form “Since S is notQ, it is notP,” is a correct proof: (a) guarantees that its pervasion (notQ is pervaded by notP) is established, (b) guarantees that its logical reason qualifies the subject (S is notQ).

There is no explicit requirement pertaining to the establishment of the subject in this definition nor is there any mention of this issue in Phya pa’s gloss of the individual terms. Only in a later section of the Mun sel does one finds the mention that the expression used in this definition for the subject, i.e., “property-possessor” (chos can), implies that the basis is established by a valid cognition.^[34] This requirement is otherwise highlighted in Phya pa’s classification of arguments by consequence.^[35]

Why is the establishment of the subject required? From the point of view of the reverse form, it is required in order to avoid the fallacy of the “unestablished subject,” which would affect the characteristic of pakṣadharmatva of the “root logical reason.” From the point of view of the consequence, the non-establishment of the subject does not affect its qualification by the logical reason, because the premise “S is P” merely needs to be accepted by the opponent.^[36] But it can affect the negation of the conclusion of the consequence, so that the requirement (b) present in the definition of a probative consequence will not be fulfilled.

2.3 Exmple of a consequence failing to be probative due to an unestablished subject

In his classification of consequences Phya pa adduces the following example to illustrate the category of arguments by consequence that fail to be probative because of the non-establishment of the subject:

(1) “Because an ultimate entity is produced, it follows that it is imperma-nent.”

The reverse form of this consequence would be:

(1’) “Since an ultimate entity is not impermanent, it is not produced”

According to Phya pa, the pakṣadharmatva of (1’) fails to be established. Indeed, one cannot establish that “an ultimate entity is not impermanent” due to the non-establishment of the subject “ultimate entity.” This is in line with his take on unestablished subjects when discussing inference-for-oneself. There Phya pa lists seven cases where the “qualification of the subject” (phyogs chos) is not fulfilled for the logical reason.^[37]

The first three are due to the subject (or basis):

i. Unfounded basis (gzhi gtan med pa), for instance, ‘ultimate sound’ or a ‘self consist-ing in a pudgala;’^[38]

ii. Basis not ascertained by valid cognition (gzhi tshad mas ma nges pa), for instance, ‘a demon’s pot’ or ‘a nymph’s song;’

iii. Basis with regard to which there is no desire to know (gzhi la shes’dod med pa), for instance, ‘sound’ when proving audibility.

The first two are relevant to the fallacy of āśrayāsiddha (even though Phya pa does not use the corresponding term gzhi ma grub here). Phya pa does not explain in detail the distinction between (i) and (ii), but a passage in the discussion pertaining to demons’ pots in the context of arguments by consequence suggests that instances of the second category are things that may exist, but whose existence cannot be known to (human)^[39] cognizers.^[40] The subject of the consequence (1) under consideration (“ultimate entity”) would fit into the first category: “unfounded basis” by analogy with the example ‘ultimate sound.’ In the section on inference, the subject is posited as something that the proponent wants to prove to be impermanent via the logical reason ‘produced.’^[41] In the reverse form of the consequence (1’), however, both the property to be proven and the logical reason are simple negations. This does not prevent the non-establishment of pakṣadharmatva: Phya pa rejects the possibility that an unestablished subject may be qualified by any property whatsoever, even a simple negation. We can note here a major difference to Dharmottara and other proponents of the Method of Simple Negation.

If the non-establishment of the subject “ultimate entity” affects the establishment of pakṣadharmatva, it does not prevent the elimination of the conclusion of the consequence by a valid cognition. According to Phya pa, it is possible to refute that “ultimate entity is impermanent.” This is achieved by way of negating the conjunction of ultimate entity and impermanent (mi rtag pa don dam pa’i dngos po dang tshogs pa khegs). To do so, one considers the fact that what is produced is pervaded by impermanence, which is itself pervaded by emptiness. There can thus be no connection between impermanence and something outside the range of what is empty, such as an ultimate entity; hence one can negate the association of these two.^[42] But this does not amount to establishing pakṣadharmatva. Phya pa holds (contra to some of his predecessors) that pakṣadharmatva must consist in the attribution of a property (here consisting in a simple negation) to a basis. In dealing with this example, he thus acknowledges a difference between pakṣadharmatva and the elimination of the conclusion.^[43]

The distinction amounts to the one we may draw in terms of “internal negation” and “external negation” when considering the form of the elimination of the conclusion and that of the pakṣadharmatva:

Consequence: Because S is P, it follows that it is Q	Elimination of the conclusion = it is not the case that S is Q
Reverse form: Since S is not Q, it is not P	Pakṣadharmatva = S is not Q

The distinction mostly goes unnoticed in the formulation of the examples in Tibetan: apart from the context where Phya pa speaks of “negation of the conjunction,” the formu-lation of the elimination of the conclusion in examples is not usually distinguished from that of the pakṣadharmatva. And the distinction is not actually crucial since apart from cases involving a non-existent subject, the two features are either both ascertained or both not ascertained by a valid cognition. For instance, a hill is either a thing-with-smoke or a thing-without-smoke. By eliminating, for instance by a perception, the conclusion “there is no smoke on the hill” that derives from the acceptance that there is no fire, one also warrants the establishment that “there is smoke on the hill.”

2.4 Comparison of four examples of consequences with potentially unestablished sub-jects

It would obviously be problematic if Dharmakīrti’s example were to fall into the same category as case (1) due to the nature of the subject “universal.” We have already seen in § 1.2 that the Method of Conceptual Appearance would enable Phya pa to ‘save’ this case. In what follows, I will consider two further examples of consequences with problematic subjects in an attempt to establish whether there is a criterion that directs the application or non-application of this method.

Altogether, there are four examples of consequences in Phya pa’s excursus that are relevant to the question of the subject. The first is case (1) considered above; the fourth corresponds to Dharmakīrti’s example. I list below their constitutive elements and those of their putative reverse form. Note that all four examples are genuine consequences–arguments to which the opponent is unable to retort. It must therefore be granted that the opponent accepts the qualification of the subject by the logical reason in the consequence and does not agree with the derived property that comes to be ascribed to the subject.

In all four cases the pervasion must be taken to be established by a valid cognition for both debaters.

	1	2	3	4
Consequence
Subject	ultimate entity	demon’s pot	demon’s permanent pot	singular universal
Reason	produced	existent	existent	linked with multiple instances
Derived property	impermanent	impermanent	impermanent	multiple
Reverse form
Subject	ultimate entity	demon’s pot	demon’s permanent pot	singular universal
Reason	void of imper-manence	void of impermanence	void of impermanence	void of multiplicity
Property to be proven	void of being produced	void of existence	void of existence	void of link with multiple instances

At first sight these four cases look very similar:

1. They are all consequences that a Buddhist could put forward to refute an opponent and/or negate in a proactive way the attribution of a property.

2. In all four cases the subject appears to be non-established. In particular, in Phya pa’s discussion of pakṣadharmatva in the inference-for-oneself, ‘demon’s pot,’ the subject of (2), is mentioned as an instance of a basis that is not established by a valid cognition, and ‘ultimate entity,’ the subject of (1), is akin to the example illustrating an unfounded basis. Note that the distinction between these two types of unestablished basis does not play a role in the section on arguments by consequence: Phya pa does not differentiate in this context between bases that are termed “not determined by valid cognition” (tshad mas nges pa med pa), “unfounded” (gtan myed pa), or simply “not established” (gzhi ma grub).

3. The logical reason and the derived property of the consequence are positive properties, whereas the logical reason and the property to be proven of the reverse form have the form of a simple negation. As we have seen in the analysis of case (1), this is not a feature that justifies the application of the property to an unestablished subject.

4. Hence in all four cases one can expect that pakṣadharmatva is not established, so that the reverse form is not correct.

But Phya pa treats these four cases very differently: (3) and (4) are held to be probative consequences, while (1) and (2) are held to be non-probative due to a failure to establish pakṣadharmatva in the reverse form. Further, (1) and (2) differ as to the possibility of eliminating the conclusion of the consequence.

	Elimination of the conclusion		Establishment of the pakṣadharmatva of the reverse form
1	Rejection of the conjunction of im-permanence and ultimate entity	✓	Establishment that an ultimate entity is void of impermanence	✕
2	Refutation that the demon’s pot is impermanent	✕	Establishment that a demon’s pot is void of impermanence	✕
3	Refutation that the demon’s perma-nent pot is impermanent	✓	Establishment that a demon’s permanent pot is void of impermanence	✓
4	Refutation that the singular universal is multiple	✓	Establishment that a singular universal is void of multiplicity	✓

The difference between the four cases in terms of elimination of the conclusion and establishment of the pakṣadharmatva of the reverse form can be summarized as follows:

It is understandable that Phya pa would want (4) to be probative since it is the case discussed in the prasaṅga passage of the PVin. But what distinguishes it from (1) and (2)? Can the method adopted for (4) in the commentary on the prasaṅga passage not also be applied to (1) and (2)? And further, what distinguishes (2) from (3)?

(2) “Because the demon’s pot is (according to you) existent, it follows that it is impermanent.”

The second example (2) appears in two contexts in the Mun sel (there is no parallel for the second context in the’Od zer). In the first Phya pa uses this example to refute the definition of probative consequence given by other scholars in terms of “qualification of the subject merely accepted by the opponent and pervasion established by a valid cognition,”^[44] in the second he argues that his own definition of probative consequence (which he inherits from rNgog Blo ldan shes rab) does not allow for the inclusion of this case.^[45]

The main argument in both discussions revolves around the fact that the conclusion of the consequence “the demon’s pot is impermanent” cannot be eliminated by a valid cognition.^[46] As we have seen in case (1), the non-establishment of the subject does not necessarily prevent the elimination of the conclusion–there may be a way to negate the conjunction of the subject and the derived property. But no similar solution is offered here.

The failure to satisfy the feature of elimination of the conclusion disqualifies this example from being probative according to Phya pa’s definition, which includes it as a necessary criterion. For other scholars who do not include it in the definition, this failure may threaten the consequence to become fallacious. Indeed, unless the opponent is convinced of the permanence of demons’ pots,^[47] he is likely to retort that he accepts the derived conclusion that the demon’s pot is impermanent. The consequence would hence not even be genuine.

As a final blow, Phya pa points out that the logical reason of the reverse form would be unestablished (rtags ma grub). Although it is not specified explicitly what makes the reason unestablished, the fallacy of the non-established basis is a likely candidate as it was in (1).

(3) “Because the demon’s permanent pot is existent, it follows that it is impermanent.”

(4) “Because a singular universal is linked with multiple instances, it fol-lows that it is multiple.”

Demons’ permanent pots were probably not a significant topic of debate in philosophical circles but universals certainly were. Nevertheless, case (3) is the key to understanding the difference between cases where the nature of the subject is problematic from those where it is not. On the one hand, its form and constitutive elements are considered by Phya pa to be perfectly parallel to the fourth application–the results obtained from the analysis of this example thus apply to the example from Dharmakīrti’s prasaṅga passage.^[48] On the other hand, it only differs from (2) in one aspect, namely the subject is “demon’s permanent pot” rather than “demon’s pot”–a detail that turns out to be of major importance.

The argument by consequence about the demon’s permanent pot is introduced as a potential counter-example to Phya pa’s definition of a probative consequence. Phya pa’s definition, as we have seen, includes the requirement of “elimination of the conclusion by a valid cognition,” which is a necessary (but not sufficient) condition for the establishment of the pakṣadharmatva in the reverse form. Phya pa invoked this criterion to exclude case (2) from the range of probative consequences. A (hypothetical) opponent now argues that cases (3) and (4) should be excluded for the same reason: the subject being unestablished, there can be no elimination of the conclusion by a valid cognition.^[49] Phya pa thus sets out to show how the conclusion can nevertheless be eliminated in case (3), and hence in (4).

The problem had been acknowledged by earlier scholars–in particular rNgog Blo ldan shes rab–who came up with a solution and answered various objections.^[50] But Phya pa does not agree with this solution and also claims that these scholars did not satisfactorily answer the objections addressed to them. The presentation of his own solution to the problem is thus to be understood against the background of this earlier debate: in addition to presenting his own account of a valid cognition able to eliminate the conclusion, Phya pa must show that his solution is not liable to the objections addressed to the view of his predecessors.

Leaving the details aside, let me attempt here to provide a summary of this long and complex discussion (Mun sel 91b3–95b5,’Od zer 147a3–148a4):

According to Phya pa’s predecessors, to eliminate a conclusion–“S is Q”–a valid cognition must grasp the contradiction between the derived property (Q) ascribed to the subject and a property that is part of the subject (S), either by definition or through the formulation of the subject (for instance “singular” in the case of the subject “singular universal” or “permanent” in the case of the subject “demon’s permanent pot”). When the subject is not established, the contradiction must simply be ascertained on the basis of another instance, and can then be applied to the problematic subject. For instance, the contradiction between “permanent” and “impermanent” can be established on the basis of a pot, and then applied to the case of the demon’s permanent pot.

Phya pa’s solution is to appeal to appearances. The subject of (3) is “just what appears as a demon’s permanent pot,” that of (4) is “just what appears as a singular universal.” This appearance is established by reflexive awareness. Thus, technically speaking, the subject is no longer unestablished. The appeal to another basis to ascertain the contradiction (which was the solution advocated by Phya pa’s predecessors) is not needed. The ascertainment of contradiction can be made on the basis of the appearance: The reflexive awareness that establishes an appearance as X also establishes “aspects” or “features” (rnam pa) of this appearance in such a way that the establishment of a feature induces the rejection of the direct or indirect contrary feature (see § 3 for more details).

Thus in the case of the demon’s permanent pot (3), the formulation of the subject in terms of “demon’s permanent pot” generates a conceptual cognition which reflexive awareness can establish to be an “appearance as a demon’s permanent pot” involving, among other things, the feature of “permanence.” The establishment of the feature of “permanence” in this appearance allows one to reject the feature of “impermanence.” Thereby the conclusion of the consequence, “the demon’s permanent pot is impermanent” is eliminated, and the qualification of the demon’s permanent pot as “void of impermanence” is established. Hence both the elimination of the conclusion of the consequence and the pakṣadharmatva of the reverse form are established by the same valid cognition: reflexive awareness. The same method can be applied in case (3): the mental appearance of the subject “singular universal” allows for the establishment of the feature of “singularity,” which rejects “multiplicity” and establishes the “void of multiplicity.”

3. Method of conceptual appearance and ascertainment of features

In dealing with cases (3) and (4), Phya pa thus relies on his version of the Method of Conceptual Subject which highlights the mental appearance. This Method of Conceptual Appearance allows him to remove the potential failure of non-establishment of the subject in the reverse form. Further, it allows the establishment of the pakṣadharmatva and the elimination of the conclusion thanks to an extension brought to the function of reflexive awareness: not only does it establish that there is a mental appearance (by which the subject is no longer unestablished), it also establishes features (rnam pa) that are an integral part of this mental appearance.

The notion of “feature” (rnam pa) that comes into play here appears to be related to a distinction between appearances of properties “with a form” (rnam ldan) and “without a form” (rnam med) (as I translated it in a previous article)^[51] invoked by Phya pa in his discussion on perception to explain why opposite superimpositions can be eliminated by perception itself for some properties of a perceived object but not for others. For instance, in the case of the perception of an impermanent blue object, the superimposition as “non-blue” can be eliminated but not the superimposition “non-impermanent.” This, according to Phya pa, because the property “blue” is “endowed with a form” (rnam ldan) whereas the property “impermanence” cannot be ascertained as such because it is “without a form” (rnam med). One could, as in the present discussion, speak of properties being “featured” or “not featured” in the mental appearance. Phya pa also applies this distinction in’Od zer when commenting on the passage of PVin 1 related to reflexive awareness, including the cognition of “pleasure, etc.”^[52] It is in this context also that the terms are already found to be used by rNgog Blo ldan shes rab. The question of a possible Indian source for this terminology remains unanswered so far.^[53]

The description of the exact process for eliminating the conclusion of the consequence and establishing the pakṣadharmatva of the reverse form based on these features involves first the establishment of a given cognition (1 in the table below). Based on the reciprocal link between mind (blo) and the object of mind (yul), the establishment of the cognition enables the establishment of its apprehended object as being endowed with the features that were represented in the cognition (2). Then, the ascertainment of features of what appears enables the negation of opposite features (3a). Phya pa adds that it also enables what I have translated in the table below as “the negation of the [object being] true with regard to the opposite property.”(3b) Expressions such as mi rtag par bden pa (“true with regard to impermanence”) and du mar bden pa (“true with regard to multiplicity”) must be understood in the framework of Phya pa’s characterization of objects and cognitions as it is found in the section of Mun sel on apprehended objects (see in particular Mun sel 1b7–1b9): objects (yul) are characterized as “true” (bden pa) or “false” (brdzun pa), and the cognitions that apprehend them (blo) as, respectively, “non-erroneous” (ma’khrul ba) and “erroneous” (’khrul ba).^[54] In subsequent discussions, Phya pa builds on this initial correspondence between the characterization of objects and cognitions, notably by adding specifications to the characterization as “true:” an object is “true” in a specific way (or with regard to a given property) when the cognition to which it appears as such is non-erroneous, and vice versa. Its being “true” in this specific way prevents its being true in the opposite way. For instance, when debating whether two things X and Y are distinct or identical, Phya pa considers whether the cognition to which they appear as distinct is erroneous or non-erroneous. In the latter case, the object (i.e., X and Y) is qualified as “true with regard to being distinct” (tha dad du bden pa), and its being “true with regard to being one” is negated (gcig du bden pa khegs).^[55] The “negation of the feature x” (x kyi rnam pa khegs pa) (3a) and the “negation of [the object being] true with regard to x” (x+loc. bden pa khegs pa) (3b) are also mentioned together by Phya pa in other contexts as following from the establishment of the feature opposite to x.^[56]

Although Phya pa is not explicit on the issue in the context under consideration, it would appear that step (3b) supports step (4a)–the elimination of the acceptance that the object has the property whose feature was negated–in view of the relation between the characterization of the object as true and of the mind that cognizes it as non-erroneous: by negating that the object is true with regard to property x, one also negates that the mind that apprehends this object as having the property x is non-erroneous.

Step (4b)–the establishment of the pakṣadharmatva of the reverse form–is to be understood as deriving from 3a/3b as well insofar as the logical reason is expressed in the form “void of property x” (e.g., void of impermanence, void of multiplicity), an explicit negation which echoes the “negation of the feature x” (3a) and the “negation of the object being true with regard to x” (3b) rather than the establishment of the object appearing as being positively qualified by the opposite property (e.g., permanence, unicity) (2).

Thus in the case of the demon’s permanent pot and, respectively, of the singular universal, we find the following steps:^[57]

1	Experience, by reflexive awareness, that conceptual cognition (i.e., the mind) is something that apprehends a demon’s permanent pot	Experience, by reflexive awareness, that conceptual cognition is something that apprehends a singular universal
2	The object is established to be the ap-pearance as a (demon’s) permanent pot (i.e., it has the feature of a permanent pot)	The object is established to have the fea-ture of a singular universal
3a	Negation of the feature of imperma-nence	Negation of the feature of multiplicity
3b	Negation of [the object being] true with regard to impermanence	Negation of [the object being] true with regard to multiplicity
4a	Elimination of the acceptance of the de-rived consequence that [the demon’s per-manent pot] would be impermanent	Elimination of the acceptance of the de-rived consequence that [the singular uni-versal] is multiple
4b	Establishment of the pakṣadharmatva of [the reason] void of impermanence	Establishment of the pakṣadharmatva of [the reason] void of multiplicity

4. Other applications of the method of conceptual appearance in Phya pa’s works

The method adopted by Phya pa in his commentary on the prasaṅga passage and in his own theory of arguments by consequence is also applied in his works in other contexts.

One of those is linked with Dharmakīrti’s discussion in the svadharmin passage, which, as pointed out in the introduction, is one of the two main contexts in which Dharmakīrti deals with cases where he wants to rescue arguments from the fallacy of the unestablished basis.^[58]

In the svadharmin passage, Dharmakīrti discusses the Buddhist refutation of the Vaiśeṣika’s permanent space and the Vaiśeṣika’s objection that the Buddhist’s argument would be faulty because by refuting the existence of the subject the Buddhist would invali-date his own thesis.^[59] This passage has no direct equivalent in the PVin. But Dharmakīrti hints to this issue at the end of the discussion on the word iṣṭa in the definition of the thesis (PVin 3 18,9–19,1). It is when commenting on this part of the PVin that Phya pa introduces in his commentary an excursus on logical reasons “refuting the nature of the subject” (chos can gyi ngo bo’gog pa, Skt. dharmisvarūpaviparītasādhana) (’Od zer 157b6–159b7). This excursus has an equivalent in the Mun sel within Phya pa’s discussion of the four categories of contradictory reasons.^[60]

Note that in Phya pa’s discussion, the argument under consideration is not the Buddhist’s refutation of the Vaiśeṣika’s permanent space, but the Vaiśeṣika’s proof of the permanence of space by the logical reason “void of being produced.” Phya pa’s intention is to establish that this proof is subject to the fault of “refuting the nature of the subject” all the while arguing that the fallacy of the unestablished basis does not apply.^[61] Phya pa invokes in this discussion the Method of Conceptual Appearance: the subject is “the (mere) appearance as space” (nam mkhar snang pa/nam mkhar snang pa tsam), that is established by reflexive awareness and for which the property “void of being produced” can be established.^[62] Hence pakṣadharmatva holds. But if the logical reason is thus not faulty by reason of non-establishment, it is, however, faulty in that it is contradictory. This is because, according to Phya pa, the Vaiśeṣika proponent of the proof additionally applies to the mental appearance as space an intentional determination as “real space” via an erroneous cognition. Due to the incompatibility of the property to be proven, “void of impermanence,” with a subject assumed to have such a real nature, the proof ends up refuting the nature of the subject.^[63]

A further passage worth considering occurs in the context of the establishment of pervasion in the proof of momentariness by the logical reason “produced” or “existent.”^[64] This involves a secondary inference proving that what is permanent (or not momentary) is not produced or not existent via the reason “void of instantaneous or gradual causal efficacy.” Phya pa identifies the subject in this secondary inference as “the simple negation ‘void of momentariness’” (skad cig gis stong pa’i med dgag). This simple negation is to be established for a verbal object (sgra don) or, as Phya pa rephrases, a concept (don spyi).^[65] More precisely, Phya pa invokes the appearance of a concept that has the feature of void of momentariness (Mun sel 63a9: skad cig mas stong pa’i rnam pa can gyi don spyi). Only in such a case can the void of momentariness be ascertained by mere experience (myong pa)–this amounts to reflexive awareness.

The description of the process is the same as in the case of the demon’s permanent pot and the singular universal: experience establishes what the apprehending mind is like–in this case, it apprehends something that is not momentary–and thereby what the object of the apprehending mind is like–the object has the feature of being non-momentary. This in turns leads to the negation of the aspect of momentariness and the negation of the object as being “true with regard to momentariness,” and to the establishment of the simple negation “void of momentariness.”^[66]

Phya pa specifies that the experience by which one establishes the property “void of momentariness” does not simultaneously establish or eliminate the property to be proven, i.e., the property “void of existence.” This is because to establish that the object (yul) is existent or non-existent, one must be able to establish that the mind (blo) is erroneous or non-erroneous (see above and n. 54)–but this property is not something that is featured (rnam med) in the appearance and hence cannot be established by reflexive awareness. In other words, introspection does not allow one to ascertain whether the mental appearance as “something void of momentariness” is erroneous or non-erroneous with regard to the properties of existence or void of existence, and therefore whether the object of such a cognition is unreal or real.^[67]

Conclusion: When does the fault of the unestablished subject actually apply?

As mentioned at the end of § 2, the adoption of the Method of Conceptual Subject or Method of Conceptual Appearance seems to leave no room for unestablished bases–anything can be established as a mental appearance since there is no restriction as to what can be said and thought about. And if there are no unestablished bases, the fault of āśrayāsiddha would never take place. Logical reasons would never lack pakṣadharmatva because of the lack of establishment of the basis. Still, unestablished bases (unfounded or not ascertained by valid cognition) are listed by Phya pa among the reasons for the absence of pakṣadharmatva in inference and come into play in rejecting the fulfillment of the pakṣadharmatva of the reverse form of some arguments by consequence (such as case (1) and case (2) considered above). Failure of pakṣadharmatva is not exclusively linked with the attribution of a positive property to a non-existent subject. Indeed, the pakṣadharmatva of the reverse form of consequences (1) and (2) is faulty although the logical reason is a simple negation. One could suggest that this fault is restricted to proofs formulated by opponents of the Buddhist, while Buddhist proofs can be ‘saved’ by the application of the Method of Conceptual Appearance. However, we have seen in § 4 that Phya pa also applies this method to argue that the Vaiśeṣika’s proof of the permanence of space is not subject to the fault of āśrayāsiddha.

Is there another criterion that distinguishes cases where the Method of Conceptual Appearance applies and those where the fault of āśrayāsiddha is brought forward?^[68] Or should one, as uncharitable as it may seem, conclude that the fault of the unestablished basis is advocated whenever it is convenient for the author to do so, whereas in other cases the Method of Conceptual Appearance is called to the rescue? Before jumping to such a conclusion, let us consider what could be achieved by the application of this method in the cases where Phya pa invokes this fallacy.

Consequence (2) is relevant in this regard because its formulation only differs from con-sequence (3) by a small detail regarding the subject, but while the argument by consequence about the “demon’s permanent pot” (3) is probative, that about the “demon’s pot” (2) is not: it is possible, according to Phya pa, to establish that “the demon’s permanent pot is void of impermanence” but not to establish that “the demon’s pot is void of impermanence.” What happens if the Method of Conceptual Appearance is applied to consequence (2)? It would indeed remove the potential fallacy of āśrayāsiddha in the pakṣadharmatva, i.e., the establishment that “the demon’s pot is void of impermanence.” However, consequence (2) would remain non-probative (and even possibly not genuine) because this would still not enable one to establish that the property “void of impermanence” qualifies the demon’s pot or to eliminate the connection between the demon’s pot and impermanence. Indeed, the establishment of these two features in consequence (3) is enabled by the fact that reflexive awareness could establish the aspect of “permanence” on the basis of the mental appearance as a “demon’s permanent pot.” Consequence (2) is different in this regard: the feature of “permanence” cannot be established by reflexive awareness because it is not an integral part of the mental appearance of a demon’s pot. This is because it was not explicated in the verbal formulation “demon’s pot” and is also not a feature associated by definition with demons’ pots.

The appearance as a demon’s pot can also not be ascertained to have the feature of impermanence. One must not be fooled here by the potential impact of the word “pot:” it only allows for the ascertainment of the feature of “being a pot” (one could thus refute the claim that “the demon’s pot is not a pot”), but not of the features of “impermanence,” “existence,” or “causal efficacy” that one typically associates with pots–an association that is erroneous because it fails to take into account that the category of “pots” is not restricted to real pots (it also includes dream pots, etc., that are not real).

It seems that the same analysis could be applied to the other cases where the fault of āśrayāsiddha is invoked, that is, the pakṣadharmatva of the reverse form of consequence (1)–“ultimate entity is void of impermanence”–and the examples adduced in the chapter on inference–“ultimate sound is produced” or “demon’s pot is produced.” In all these cases the Method of Conceptual Appearance would certainly remove the fault of āśrayāsiddha but would still not enable the establishment of the pakṣadharmatva because the mental appearance would not have the explicit features that would be relevant in this regard.

One could in conclusion propose that the problem of āśrayāsiddha has been transformed into a failure of the establishment of the pakṣadharmatva that has nothing to do with the ontological status of the subject, but is a matter of the features of the mental appearance that can or cannot be ascertained by reflexive awareness. However, such an interpretation of the examples where āśrayāsiddha is invoked is not suggested by Phya pa. One could deplore the lack of a unifying theory on this point or, more charitably, envisage that Phya pa is applying a principle of economy in the identification of fallacies.

References and abbreviations

bKa’ gdams gsung’bum bKa’ gdams gsung’bum phyogs sgrig thengs dang po/gnyis pa/gsum pa/bzhi pa, ed. dPal brtsegs bod yig dpe rnying zhib’jug khang. Vols. 1–30, Chengdu 2006; vols. 31–60, Chengdu 2007; vols. 61–90, Chengdu 2009; vols. 91-120, Chengdu 2015.

D sDe dge Tibetan Tripiṭaka bsTan’gyur –preserved at the Faculty of Letters, University of Tokyo, ed. J. Takasaki, Z. Yamaguchi, and Y. Ejima. Tokyo 1981–1984.

dBu ma snang bshad Phya pa Chos kyi seng ge, dBu ma snang ba’i’grel pa rgya cher bshad pa. See bKa’ gdams gsung’bum, vol. 6, 266–428.

dKa’ gnas rNgog Blo ldan shes rab, Tshad ma rnam nges kyi dka’ ba’i gnas rnam par bshad pa, ed. Sun Wenjing. Qinghai 1994.

Hugon 2011 P. Hugon, Phya pa Chos kyi seng ge’s Views on Perception. In: Religion and Logic in Buddhist Philosophical Analysis. Proceedings of the Fourth International Dharmakīrti Conference. Vienna, August 23–27, 2005, ed. H. Krasser, H. Lasic, E. Franco, and B. Kellner. Vienna 2011, 159–176.

Hugon 2013 P. Hugon, Phya pa Chos kyi seng ge on Argumentation by Consequence (thal’gyur): The Nature, Function, and Form of Consequence Statements. Journal of Indian Philosophy 41.6 (2013) 671–702 (DOI 10.1007/s10781-013-9205-4).

Hugon 2016 P. Hugon, Phya pa Chos kyi seng ge on Argumentation by Consequence (thal’gyur) (2): The analysis of the correspondence between a consequence and its reverse form and the thirteenfold typology of consequences. Journal of the International Association of Buddhist Studies 39 (2016) 51–113 (DOI: 10.2143/ JIABS.39.0.3200523).

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

Iwata 1997a T. Iwata, On Prasaṅgaviparyaya in Dharmakīrti’s Tradition–Prajñākara-gupta and gTsang nag pa. In: Tibetan Studies. Proceedings of the Seventh Seminar of the IATS, Graz 1995, ed. H. Krasser, M. Torsten Much, E. Steinkellner, and H. Tauscher; gen. ed. E. Steinkellner. Vienna 1997, 427–437.

Iwata 1997b T. Iwata, Pramānaviniścaya III (2). Wiener Zeitschrift für die Kunde Süd-asiens 41 (1997) 207–231.

Iwata 1999 T. Iwata, On The Interpretations of The Subject (dharmin) of The Inference Negating Invariable Entities in Dharmakīrtian Logic. In: Dharmakīrti’s Thought and Its Impact on Indian and Tibetan Philosophy. Proceedings of the Third International Dharmakīrti Conference. Hiroshima, November 4–6, 1997, ed. Sh. Katsura. Vienna 1999, 155–172.

Keira 2004 R. Keira, Mādhyamika and Epistemology: A study of Kamalaśīla’s Method for Proving the Voidness of All Dharmas. Vienna 2004.

Kobayashi 1987 M. Kobayashi, Mujishosei-ronsho to shoe-fujo (āśrayāsiddha) no mondai. Bunka 50.3/4 (1987) 41–60.

Moriyama 2019 Sh. Moriyama, On Dharmisvarūpaviparītasādhana. Wiener Zeitschrift für die Kunde Südasiens 56–57 (2019) 37–50.

Mun sel Phya pa Chos kyi seng ge, Tshad ma yid kyi mun pa sel pa. See bKa’ gdams gsung’bum, vol. 8, 434–626.

N Narthang (snar thang) edition of the Buddhist canon in Tib.

NB Dharmakīrti, Nyāyabindu: Nyāyabindu with Dharmottara’s Nyāyabinduṭīkā and Du-rveka Miśra’s Dharmottarapradīpa, ed. Dalsukh Bhai Malvania. Patna 1955 [2^nd ed. 1971].

’Od zer Phya pa Chos kyi seng ge, Tshad ma rnam par nges pa’i’grel bshad yi ge dang rigs pa’i gnad la’jug pa’i shes rab kyi’od zer. See bKa’ gdams gsung’bum, vol. 8, 35–427.

P The Tibetan Tripiṭaka, Peking edition, ed. D.T. Suzuki. Tokyo/Kyoto 1955–1961.

PS Dignāga, Pramāṇasamuccaya, Tibetan translation in D4203, P5700.

PV Dharmakīrti, Pramāṇavārttika.

PV 1 Chap. 1 and Pramāṇavārttikasvavṛtti: The Pramāṇavārttikam of Dharmakīrti: The First Chapter with the Autocommentary, ed. R. Gnoli. Rome 1960.

PV 2, 3, 4 Chap. 2–4 (pramāṇasiddhi, pratyakṣa, parārthānumāna), ed. Y. Miyasaka. Acta Indologica 2 (1972).

PVin Dharmakīrti, Pramāṇaviniścaya.

PVin 1/2 Chap. 1/2, ed. E. Steinkellner. Beijing/Vienna 2007.

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

PVin-Tib Tib. translation of the third chapter of the Pramāṇaviniścaya. D4211 ce 187a6–230a7; P5710 ce 285a7–329b1; N3701 ce 299a6–347a7.

PVinṬ-Skt Dharmottara, Pramāṇaviniścayaṭīkā, copy of the manuscript (see the intro-duction to PVin 1 & 2 and PVin 3 for the details).

PVinṬ-Tib Dharmottara, Pramāṇaviniścayaṭīkā, Tibetan translation in D4227, P5727. Pascale Hugon 153

PVSV Dharmakīrti, Pramāṇavārttikasvavṛtti. See PV 1.

Tachikawa 1971 M. Tachikawa, A Sixth-Century Manual of Indian Logic. (A Translation of the Nyāyapraveśa). Journal of Indian Philosophy 1 (1971) 111–145.

Tani 1987 T. Tani, The Problem of Interpretation on Pramāṇaviniścaya III vv. 1–3–with the text and a translation. The Bulletin of the Kochi National College of Technology 26 (1987) 1–16.

Tillemans 1982 T. Tillemans, The ‘Neither One Nor Many’ Argument For Śūnyatā, and Its Tibetan interpretations: Background information and source materials. Études de Lettres, University of Lausanne 3 (1982) 103–128.

Tillemans 1984 T. Tillemans, Two Tibetan Texts on The ‘Neither One Nor Many’ Argu-ment For Śūnyatā. Journal of Indian Philosophy 12.4 (1984) 357–388.

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

Tillemans 2000 T. Tillemans, Dharmakīrti’s Pramāṇavārttika. An annotated translation of the fourth chapter (parārthānumāna). Vol. 1 (k. 1–148). Vienna 2000.

Tillemans and Lopez 1998 T. Tillemans and D. Lopez, What Can One Reasonably Say About Nonexistence? A Tibetan work on the problem of āśrayāsiddha. Journal of Indian Philosophy 26 (1998) 99–129.

Tshad ma bsdus pa Anonymous (attributed to Klong chen Rab’byams pa), Tshad ma’i de kho na nyid bsdus pa, ed. Padma tshul khrims. Chengdu 2000.

Yoshimizu and Nemoto 2013 Ch. Yoshimizu and H. Nemoto, Zhang Thang sag pa’Byung gnas ye shes, dBu ma tshig gsal gyi ti ka, Part I, folios 1a–26a3 on Candra-kīrti’s Prasannapadā ad Mūlamadhyamakakārikā 1 1. Tokyo 2013.

Footnotes and references:

[1]:

Dharmakīrti’s short mention of such arguments in PV 4.12 (on which see Tillemans 2000: 21–24) is supplemented by a longer discussion in the PVin (PVin 3 4,4–6,12; the references to the Tibetan translation of this discussion in the sDe dge, Peking and sNar Thang bstan’gyur are: D188a5–189a5; P286a5–287a5; N300a6–301b1).

[2]:

For an in-depth study see Iwata 1993. The main points appear in a summarized form in English in Iwata 1997a.

[3]:

In NB 3.65 dharmyasiddha (“non-establishment of the property-possessor”) is listed as one type of fallacy of the logical reason. An additional issue with such proofs is that the Buddhist proponent could be accused of annulling his own thesis when negating the subject.

[4]:

On Madhyamaka proofs see, for instance, Tillemans 1982 and 1984 and Iwata 1999. Kobayashi 1987 (in Japanese) exposes mainly Kamalaśīla’s view in the Madhyamakāloka (the relevant section of the Madhyamakāloka is translated into English in Keira 2004).

[5]:

For a panorama see Iwata 1999 and Watanabe in this volume. For the views of specific authors see also Tillemans 1999, chap. 8, Tillemans and Lopez 1998, and the articles mentioned in n. 4.

[6]:

In addition to Iwata 1993 and 1997a, see Watanabe’s paper in this volume, which discusses the various solutions to the problem of āśrayāsiddha developed by Dignāga and Dharmakīrti and their connection with prasaṅga arguments. I summarize some of them below.

[7]:

Since the surfacing of eighteen of his works, the views of this thinker have been the subject of an increasing number of publications, so that a formal introduction may no longer be needed. For a recapitulation of the information available on his life and works the reader may refer to my “Compiled information on the life and works of Phya pa Chos kyi seng ge and bibliographical ressources,” which is available online at https://www.oeaw.ac.at/ikga/forschung/tibetologie/materialien/materials-for-the-study-of-phya-pa-chos-kyi-seng-ge-1109-1169/.

[8]:

PVin 3 67,4–70,6; Tib. P306a6, D208b3. This passage is parallel to PVSV 105,15–107,14. In both texts it is part of the anupalabdhi section. Verses PVin 3.53–54 = PV 1.205–206 are, however, frequently cited in discussions pertaining to the apoha theory.

[9]:

Cf. PVin 3.54 = PV 1.206 and the auto-commentary. The property to be ascertained is thus not bhāva or abhāva but bhāvopādānatva/bhāvānupādānatva (cf. PVin 3 68,6–7: kim ayam pradhānaśabdapratibhāsy artho bhāvopādāno na veti).

[10]:

Cf. PVin 3 68,7–8: tasya bhāvānupādānatve sādhye sa ca pratyātmavedyatvād apratikṣepārho’rtho dharmī and 69,5: na tu punar atrāyam eva śabdavikalpapratibhāsy artho’pahnūyate.

[11]:

See Tillemans 2000: 194–198, Tillemans 1999: chap. 8, and Tillemans and Lopez 1998.

[12]:

See Iwata 1999: 165.

[13]:

See Iwata 1999: 167–168. The relevant section of the Madhyamakāloka is translated in Keira 2004: 120–152.

[14]:

The latter view is presented in the form of an objection by an Alīkākāravādin, but is accepted by Kamalaśīla. See the translation and notes in Keira 2004: 142–145 (objection) and 146–148 (reply).

[15]:

See PVin 3 67,12–68,1 (ad 3.53): anādikālavāsanāprabhavavikalpapratibhāsinam arthaṃ (Tib. rtog pa la snang ba’i don); PVin 3 69,5: śabdavikalpapratibhāsy artho (Tib. sgra’i rnam par rtog pa la snang ba’i don); PVin 3 68,6–7: pradhānaśabdapratibhāsy artho (Tib. gtso bo’i sgra las snang ba’i don).

[16]:

See, for instance, the passage of the Satyadvayavibhaṅgakārikā and Satyadvayavibhaṅgavṛtti translated in Keira 2004: 35–37.

[17]:

A degree of interpretation is already present in the Tibetan translation by rNgog Blo ldan shes rab and Parahitabhadra, which was the reference for Tibetan interpreters who relied on the Tibetan version of the PVin. In this regard one can note that the Tibetan version known to Phya pa, which one can partially reconstruct from his direct quotes of the text, occasionally differs from the version preserved in the canonical collections. The absence of corresponding direct quotes by rNgog Blo ldan shes rab leaves open the question of the original translation and whether it was Phya pa or the editors of the canon (or maybe both) who adopted a slightly modified version of rNgog Blo ldan shes rab’s translation.

[18]:

Translations of the prasaṅga passage of the PVin can be found in Tani 1987 (in English) and Iwata 1997b (in German). See also Watanabe’s translation and his discussion in this volume. For the details of Dharmottara’s interpretation see Iwata 1993: 50 and Iwata 1997a: 429–430. I will present the details of Phya pa’s interpretation in a forthcoming paper.

[19]:

Iwata (1993: 50) notes that this interpretation is shared by Jñānaśrībhadra, Bu ston and rGyal tshab Dar ma rin chen. This interpretation qualifies as counter-intuitive insofar as the topic of discussion in this passage of the PVin is prasaṅga. One would thus expect the example introduced by the expression “for example” to illustrate a prasaṅga, not a prasaṅgaviparyaya. The option that Dharmakīrti exemplifies a prasaṅga is considered by Prajñākaragupta and, reportedly, by Vinītadeva and Śāntabhadra. The ascription to these two authors of alternative interpretations (which are discussed by Dharmottara) of Dharmakīrti’s example as a prasaṅga is made by Bu ston (Iwata 1993: 51) and found also in an anonymous interlinear note in the manuscript of Zhang Thang sag pa’s dBu ma tshig gsal gyi ti ka (Yoshimizu and Nemoto 2013: 58). Phya pa ascribes the first alternative interpretation to Vinītadeva, the second to Prajñākaragupta.

[20]:

Dharmottara’s reformulation of the prasaṅgaviparyaya is as schematized in Iwata 1993: 42. The reformulation of its logical reason as na cānekam is found in PVinṬ-Skt 7a5. In PVinṬ-Skt 6b3 Dharmottara speaks of the logical reason and the property to be proven in terms of “the absence of the pervader ‘multiplicity’” (vyāpakanānātvābhāva) and the “absence of the pervaded property ‘occurrence in many’” (vyāpyasyānekavṛttitvasyābhāva). For Dharmottara’s formulation of the prasaṅga, see PVinṬ-Skt 6a1: sāmānyāder anekavṛttitva iṣṭe’nekatvam aniṣṭam āsaṃjyate and PVinṬ-Skt 7a5: yad anekadeśādivṛtti tad anekaṃ tathā ca sāmānyam ity anekaṃ syān. For Phya pa’s reformulation see’Od zer 149a6–7: spyi gcig gsal ba du ma dang’brel par kun brtags pas du mar thal zhes bsgrub pa thal bar bsgrub pa’o // des rang rgyud gang’phen zhe na / du mas stong pas du ma dang’brel pas stong zhes pa khyab byed mi dmigs pa’phen te /

[21]:

See PVin 3 5,1–6,1: nanu tathāpy asiddhir hetoḥ… and PVinṬ-Skt 8a6: bauddho hi sāmānyābhāvavāde sthitaḥ prasaṅgasādhanam āha / tadāśrayāsiddher asiddham anekavṛttitvaṃ nāma hetuḥ /

[22]:

For instance, in the proof of the presence of fire on a hill where smoke is perceived via the prasaṅga “if there were (as you claim) no fire on the hill, there would be no smoke,” the status of the ‘hill’ is not controversial, but the “absence of fire on the hill” is not established for the proponent, who precisely aims at establishing the opposite.

[23]:

Obviously it was a concern to ascribe legitimacy to an argument that had the same form as an inference-for-others but whose logical reason did not satisfy the requirements of the former.

[24]:

While it is doubtful that Phya pa would have been well acquainted with Prajñākaragupta’s commentary on the PV on a first hand basis, he might have known about his position via rNgog Blo ldan shes rab’s commentarial work on the Pramāṇavārttika cum Pramāṇavārttikālaṅkāra or rNgog Blo ldan shes rab’s discussion of prasaṅga in his larger PVin commentary (both works are currently not available to us). rNgog Blo ldan shes rab’s explanation of the difficult points of the PVin (dKa’ gnas) does not offer a word commentary on the prasaṅga passage, but it explicitly enjoins the reader to follow Dharmottara’s interpretation and reject all others (dKa’ gnas 393). It is thus certain that rNgog Blo ldan shes rab’s position did not match that of Prajñākaragupta. In his 1997a paper (433–435) Iwata evokes the possibility that Prajñākaragupta influenced the Tibetan classification of arguments by way of a consequence, in which a distinction is made between arguments in which the subject is non-established–i.e., those that are “non-probative”–and arguments in which it is established, i.e., that qualify as probative. This suggestion, however, must be rejected. The classification in question, whether that made by gTsang nag pa (considered in Iwata’s paper) or Phya pa (see below § 2.1), does not distinguish between two ways of dealing with the subject in a single given argument in the way Prajñākaragupta does, but rather distinguishes those arguments in which the subject is established by a valid cognition (for instance the subject “hill” when refuting the absence of fire) from those in which it is not (gTsang nag pa does not give an example–see § 2.3 for Phya pa’s example).

[25]:

’Od zer 150b2–3: shes bya’i gnas lugs dpyad nas (read: na) grub mtha’ gzod tshol ba’i dus su sangs rgyas pa’i grub mtha khas blangs pa myed la / mu stegs pa’i grub mtha khas myi len pas spyi dngos po dang dngos med gzhir ma bkod de blo snang gzhi’ yin pas dngos por sangs rgyas pa la ma grub la dngos myed du mu stegs pa la ma grub kyang blo snang tsam de gnyi’ ga la grub po zhes pa’o //

[26]:

One may note that Phya pa does not use the term “concept” (don spyi) in this context. Cf. n. 65.

[27]:

’Od zer 150a9: rang gi lugs la’thad pa ni khas blangs pa’i sgra don la du mas stong pa’i phyogs chos rang rig pa’i shugs la grub pa […] gcig la ni zhes pa […].

[28]:

’Od zer 149a7: chos can ni du ma la yod pa’i gcig po ste gcig du snang pa’i chos sgra don la yod par rang rig pa’i shugs la grub pa’o / The establishment of the properties of “unicity” and “void of multiplicity” by reflexive awareness evoked in these passages stands out in a more detailed way in the portion of Phya pa’s excursus on arguments by consequence dealt with in § 3.

[29]:

This is not to say that other fallacies would not arise, allowing the Buddhist to criticize his opponent’s proof.

[30]:

For an overview of Phya pa’s theory of argumentation by consequence see Hugon 2013. As I make clear in this paper (675–676), the form “Because S is P, it follows that it is Q” is a reference to a consequence statement, which should not be confused with the actual consequence statement, that typically should have the form “all that is P is Q, S is P.”

[31]:

This classification of consequences and the various issues linked with it are analyzed in details in Hugon 2016.

[32]:

Mun sel 90a9–b1,’Od zer 146b6: skabs su bab pa’i chos can la dngos kyi dam bca’ (Mun sel sgrub bya, to be emended to dam bca’) la tshad mas bsal pa dang / khyab pa tshad mas nges pa’i rtags ston pa’o //

[33]:

Note that this definition does not include any explicit requirement pertaining to the qualification of the subject by the logical reason in the consequence (i.e., to the premise “S is P”). Phya pa disagrees in this regard with some of his predecessors. See below n. 44. But according to Phya pa’s gloss of the terms of his definition, this feature must be understood from the expression “subject in the context” (skabs su bab pa’i chos can), which refers to a subject (S) such that the opponent entertains the erroneous belief that S is P.

[34]:

Mun sel 91a7–8: chos can zhes pas gzhi la lta bu tshad mas grub par bstan “With the word ‘property-possessor’ one indicates that a basis such as ‘a mountain pass’ is established by a valid cognition.”

[35]:

For the details see Hugon 2016.

[36]:

Phya pa seldom uses the term phyogs chos (the Tibetan rendering of pakṣadharmatva) for this feature in the argument by consequence. He just speaks of “the nature of the reason being established” or “the reason being established.” See Hugon 2013: 679.

[37]:

See Mun sel 46a3–5,’Od zer 84b6–8.

[38]:

In Mun sel 28.1a6, the “unfounded definitional basis” (mtshan gzhi gtan med pa) is similarly responsible for the fault of “impossible definiens” (mi srid pa’i mtshan nyid). It is exemplified by the definitional basis “permanent pot” to which someone applies the definiens “causally active” to define it as real.

[39]:

Possibly demons can see their own pots and nymphs can hear their own songs.

[40]:

Cf.’Od zer 147a4: rtag pa’i sha za’i bum pa dang spyi gcig po’i dngos po chos can du byed na gzhi de dag myed pa dang yod kyang shin du lkog du gyur pas de la’jug pa’i tshad mas dam bca’ la bsal pa myed pa bden pa zhig na /

[41]:

This is the case for the other examples of (i) and (ii) also. The formal applications mentioned here are (i) “Ultimate sound is impermanent because it is produced” (don dam pa’i sgra byas pa’i phyir mi rtag), or “A self consisting in a pudgala is impermanent because it is produced” (gang zag gi bdag gzhir byas te byas pas myi rtag pa bsgrub pa lta bu’o); (ii) “A demon’s pot is impermanent because it is produced” (sha za’i bum pa byas pa’i phyir mi rtag), or “a nymph’s song is impermanent because it is produced” (dri za’i dbyangs chos can du byas te byas pas myi rtag pa bsgrub pa lta bu’o).

[42]:

Mun sel 85b4–5: dang po ni don dam pa’i dngos po byas pa yin no zhes smra ba la don dam pa’i dngos po byas pa’i phyir mi rtag par thal lo zhes brjod pa na byas pa la mi rtag pas khyab pa tshad mas nges la mi rtag pa la stong pa nyid kyis khyab pa nges pas mi rtag pa stong pa nyid kho na la yod par nges pas don dam pa’i dngos po dang tshogs pa khegs pas dam bca’ la tshad mas bsal yang gzhi don dam pa’i dngos po nyid med pas de la mi rtag pas stong pas byas pas stong zhes pa’phen mi nus pa yin no //

[43]:

Curiously, the question of the subject is not pointed out in Mun sel in the discussion on “correspondence” (gnad cig) preceding the classification of arguments by consequence, where Phya pa establishes that the negation of the conclusion of the consequence and the pakṣadharmatva of the reverse form are either both established or both not established. In’Od zer it appears in an interlinear note that, to my opinion, represents a later addition to the text. See Hugon 2016 for the details.

[44]:

Mun sel 90a7–9;’Od zer 146b4–6. Their definition, he argues, would lead them to include this case among “probative consequences,” whereas it cannot be so. The discussion has a precedent in rNgog Blo ldan shes rab’s dKa’ gnas 385.

[45]:

Mun sel 91b3.

[46]:

The formulation in Mun sel is sha za’i bum pa mi rtag pa la tshad ma’i gnod pa med pa; in’Od zer on the other hand, the formulation matches rather that of the pakṣadharmatva of the reverse form: myi rtag pas stong par tshad mas ma nges pa.

[47]:

The specification pertaining to what the opponent accepts is only found in’Od zer.

[48]:

In the list of consequences that induce a reverse form in which the logical reason is of a different type than the logical reason of the consequence, case (3) illustrates a consequence based on a logical reason qua essential property, the reverse form of which is based on the non-apprehension of the pervader, like (4) (see § 1.1). Cf. Mun sel 86a6: rtag pa’i sha za’i bum pa yod pa’i phyir mi rtag par thal zhes pa rang bzhin gyi rtags kyi thal’gyur gyis mi rtag pas stong pas na yod pas stong zhes pa khyab byed mi dmigs pa’i rang rgyud’phen ste gnyis so //

[49]:

’Od zer 147a3–4: gzhi rtag pa’i sha za’i bum pa dang spyi gcig po nyid ma grub pas gzhi de la’jug pa’i tshad ma’i bsal pa myed pa bsgrub pa’i thal ba mtshan nyid myed pas de la ma khyab po zhe na. The objection in Mun sel is more precise. For (3) it states that there is no elimination by valid cognition because the basis is completely inaccessible (Mun sel 91b4: gzhi shin du lkog du gyur pas dam bca’ la tshad ma’i bsal pa med pas grub (em: sgrub) pa’i thal ba de la ma khyab). For (4), it examines two options: the conclusion is eliminated by the opponent’s belief, or it is eliminated by establishing by valid cognition that there is no such thing as a universal. In the former case the corresponding member of the reverse form, pakṣadharmatva, will not be established by valid cognition; in the latter case the intended thesis of the proponent would be likewise established, hence making the argument under consideration useless. This is because negating the universal itself allows one to eliminate the connection between the universal and multiplicity (i.e., to eliminate the conclusion), but also to eliminate the connection between the universal and the link with multiple instances (i.e., which is the intended thesis of the proponent). This long version of the objection runs parallel to the objection in dKa’ gnas 386.

[50]:

A large portion of the dKa’ gnas (386–393) is devoted to demonstrating that cases (3) and (4) qualify as probative and in particular that their conclusion is invalidated by valid cognition (tshad mas gnod pa). rNgog Blo ldan shes rab presents his version of a valid cognition that enables the elimination of the conclusion in spite of the subject’s lack of establishment and defends it against various objections. There is little doubt that Phya pa is referring to rNgog Blo ldan shes rab in the Mun sel when he reports the view of “previous scholars” (sngon gyi mkhas pa dag), objections to this view, and the answers of these scholars, referred to as “the greatest of the greatest” (che ba’i che ba rnams). Phya pa’s presentation of these previous views, of the objections and answers, are recognizably those found in the dKa’ gnas even though the form and wording differ. If not to rNgog Blo ldan shes rab himself, Phya pa must be referring to a faithful epigone of the latter. But note that in the Tshad ma bsdus pa (354), it is rNgog Blo ldan shes rab who is associated with this view and not a later author.

[51]:

See Hugon 2011. I avoided the translation “aspect” for rnam pa in view of Phya pa non-aspectualist standpoint on cognition.

[52]:

’Od zer 59a2 ad PVin 1.19d etc.

[53]:

See Hugon 2011: 169, n. 46 for a possible source.

[54]:

“True” and “false” correspond in this context to the characterization of the objects as, respectively, “real” (dngos po) and “unreal” (dngos med). For instance, hallucinated objects are “false” and the non-conceptual cognitions in which they appear are “erroneous.” But in other contexts, such as the definition of valid cognition in terms of “the understanding of a true object” (bden pa rtogs pa), “true” takes up the meaning of “non-opposed” (gnod pa med pa), and “true object” is not paired with “non-erroneous cognition,” but with “valid cognition.” See Hugon 2011.

[55]:

See for instance Mun sel 3a6, where Phya pa discusses the case of the appearance of “white” and “visual consciousness” in parallel with the case of the appearance of “pleasure” and “suffering.”

[56]:

See for instance Mun sel 62b8 and the passage cited in n. 66 for the feature of “momentariness” (skad cig gis stong pa’i rnam pa yongs gcod la grub pas / skad cig ma’i rnam pa khegs pas skad cig mar bden pa’ang khegs pa yin no //), 63b8 for the feature of “multiplicity” (du mar bden pa), 65b9 for the features of “momentariness” and “reality” (skad cig dang dngos por bden pa).

[57]:

Mun sel 95a3–4: rang rig pas rtog pa rtag pa’i sha za’i bum pa’dzin pa dang spyi gcig du’dzin pa tsam myong pa na yul yang rtag pa’i bum par snang par grub pas mi rtag pa’i rnam pa khegs te mi rtag par bden pa’ang khegs pas mi rtag par thal ba’dod pa la des sel te mi rtag pas stong pa’i phyogs chos kyang rang rig pa’i shugs la grub la / yul spyi gcig gi rnam par grub pa na’ang du ma’i rnam pa khegs te du mar bden pa’ang khegs pas du mar thal ba’dod pa’ang rang rig pa’i shugs la bsal la du mas stong pa’i phyogs chos kyang rang rig pa’i shugs la grub pa yin no //

[58]:

The second context is the pradhāna passage, discussed both in the PV and the PVin, which was Phya pa’s main source. But Phya pa does not supplement his commentary on the pradhāna passage with an excursus, and the refutation of pradhāna does not come forth in his Mun sel.

[59]:

PV 4.141–143 (see Tillemans 2000: 202–205).

[60]:

The four categories listed in Mun sel 54a8 are: chos kyi ngo bo’gog pa; chos kyi khyad par’gog pa; chos can gyi khyad par’gog pa; chos can gyi ngo bo’gog pa. This distinction can be traced to the Pramāṇasamuccaya (3.27 = Nyāyamukha 9), and the terminology to Nyāyapraveśa 3.2.3, where one finds the terms: dharmisvarūpa°, dharmasvarūpa°, dharmiviśeṣa°, dharmaviśeṣaviparītasādhana (see Tachikawa 1971: 125–126). Note that Dharmakīrti does not identify the fallacy occurring in the refutation of the permanence of space in terms of dharmisvarūpariparītasādhana. On Dharmakīrti’s silence on this fallacy see Moriyama 2019.

[61]:

The proof of the permanence of space is given as an example of dharmisvabhāvaviparītasādhana by Dharmottara (PVinṬ-Skt 24a5, PVinṬ-Tib D26a1). As Moriyama 2019 reveals, the Vaiśeṣika proof is also given as an example of dharmisvabhāvaviparītasādhana by Jinendrabuddhi and Dharmapāla–both argue that the fault of the unestablished subject does not apply in this case. Before presenting his own view Phya pa makes an extensive presentation and refutation of another view ascribed to Kamalaśīla etc. In Kamalaśīla’s works (notably Madhyamakāloka and Madhyamakālaṅkārapañjikā) one finds a discussion which may have been the source of Phya pa’s presentation of Kamalaśīla’s position. Yet Kamalaśīla does not discuss the faulty Vaiśeṣika proof, but the Buddhist refutation of a permanent space, which is the case discussed by Dharmakīrti in the svadharmin passage, a case that escapes the fault of dharmisvabhāvaviparītasādhana as well as the fault of āśrayāsiddha. Kamalaśīla appeals in this regard to the Method of Simple Negation (see Iwata 1999: 167).

[62]:

Mun sel 55a4,’Od zer 158a9.

[63]:

By invoking the “intentional determination” (zhen pa) Phya pa identifies a subject that corresponds to Dharmakīrti’s svadharmin.

[64]:

Starting at Mun sel 62b5.

[65]:

This rephrasing is worth noting since, as indicated in n. 26, Phya pa does not use the term “concept” (don spyi) when applying the Method of Conceptual Appearance in his commentary on the prasaṅga passage. He does not use the term either when commenting on the pradhāna passage in spite of a visible influence of the pradhāna passage of the PVin 3 pertaining to Phya pa’s characterization of the don spyi, something I intend to come back to on another occasion. On the other hand, in his commentary on the Madhyamakāloka Phya pa glosses Kamalaśīla’s identification of the subject in the refutation of pradhāna as “something imagined that exists in the mind” (nye bar brtags pa blo la yod pa) in terms of “the feature of pradhāna that is present in the appearance of the concept to conceptual thought” (dBu ma snang bshad 38b4: rtog pa la don spyi snang pa la yod pa’i gtso’ bo’i rnam par tsam chos can yin no zhes bya pa’o).

[66]:

Mun sel 63a9–63b1: skad cig mas stong pa’i rnam pa can gyi don spyi la skad cig mas dben par nges pa yin te / blo skad cig ma ma yin pa’dzin par myong pa na yul skad cig ma ma yin pa’i rnam par grub la skad cig gi rnam pa khegs pa na skad cig du bden pa’ang myong pa’i shugs la khegs pas skad cig gis stong pa’i med dgag grub pas gzhi’i ldog pa grub la /

[67]:

The inability to ascertain the erroneous or non-erroneous character of mind by reflexive awareness– and thereby whether the corresponding object is false or true–is discussed by Phya pa in Mun sel 36a7–8, with the conclusion that the establishment that an object is true and that the corresponding mind has a true object, is achieved by the cognition of an object (don rig) other than mind itself.

[68]:

Asking a similar question about Dharmakīrti, Tillemans (2000: 197) concluded that the fault of āśra-yāsiddha can be invoked unproblematically unless it would involve self-refutation for the Buddhist proponent. Only in the latter case is the method distinguishing svadharmin and keveladharmin worth applying. This conclusion, however, concerns the fault of āśrayāsiddha as a fault of the thesis and leaves open the question of āśrayāsiddha as a fault of the logical reason.