Articles by alphabetic order
 Ā Ī Ñ Ś Ū Ö Ō
1 2 3 4 5 6 7 8 9 0

How to Distinguish between Non-existing Entities? Dharmakīrti and Prajñākaragupta on Universals as Objects of Knowledge

From Tibetan Buddhist Encyclopedia
Jump to navigation Jump to search



The paper concerns the very beginning of the chapter on perception in Dharmakīrti’s Pramāṇavārttika. It focuses on an apparent contradiction in Dignāga’s and Dhamrakīrti’s epistemology, namely, that universals do not exist and yet aare considered to be objects of valid inferences. Thus, universals have to be distinguished from other non-existing entities, such as objects of illusions. The paper elucidates Dharmakīrti’s theory with the help of his great commentator Prajñākaragupta. Keywords: Buddhism, logic, philosophy, Dharmakīrti, Prajñākaragupta, Buddhist epistemology

This paper concerns the very beginning of the chapter on perception in Dharmakīrti’s Pramāṇavārttika, which was conceived as a commentary on the second verse of the 1 The following paper is one of the results of research project FR 2531/2 funded by the German Research Foundation (DFG), which has been conducted at the Institute of Indology and Central Asian Studies, University of Leipzig since February 2008: A Study on the Buddhist Theory of Perception Based on Dharmakīrti’s Pramāṇavārttika, Chapter III: A Critical Edition, Translation and Study of Pramāṇavārttika III 1–122. I would like to take this opportunity to thank the DFG for its generous grant. The project is part of an international effort to translate the Pramāṇavārttika, Dharmakīrti’s magnum opus, into English; more specifically, three teams are working on the chapter dealing with

perception. The Leipzig team aims at a translation of the first 122 verses, John Taber together with graduate students of the Philosophy Department, University of New Mexico, Albuquerque, will translate verses 123–240, and Birgit Kellner, Heidelberg University, will translate selected verses of the remaining part. I would like to thank very warmly Miyako Notake, my collaborator in the above-mentioned project. Very special thanks are also due to my wife Karin Preisendanz for her most helpful comments.

Pramāṇasamuccaya. In this verse, Dignāga states that there are only two means of knowledge or instruments of valid cognition (pramāṇas), namely perception and inference, because the objects of these means of knowledge have two characteristics: Perception has the particular as its object, and inference the universal.2 However, with the same breath Dignāga also states that the universal is only a conceptual construction and that as such it does not exist in reality. “Conceptual construction is the

association of a name, genus, and so forth [with the thing perceived which results in a verbal designation of the thing].”3 Therefore, the question immediately arises: How could something non-existing be an object of a means of knowledge? Dignāga neither raises nor answers this question. This is indeed one of the greatest mysteries of the Pramāṇasamuccaya, that is, not only what Dignāga’s answer to this question would have been, but also why he saw no need even to raise the question. If this is such an obvious problem, as it seems to us, how could it be that Dignāga had nothing to say about it? I am not going to propose an answer to this question. As far as I can see, it cannot be answered with our present knowledge of Dignāga’s work. What I propose to do instead is to try to explain, very briefly, Dharmakīrti’s answer to this question. Further, as

indicated in the title of the paper, I will dwell, again very briefly, on a specific aspect of this answer, namely, the issue of what distinguishes non-existing universals from other non-existing entities, such as the object of illusion, for instance, that there are two moons in the sky or floaters in the clear air. Dharmakīrti begins his comments on Dignāga’s verse by proposing four criteria for the distinction between particulars and universals. The first and most important of them is that particulars are capable of efficient action, and universals are not. And already in the first verse of the chapter Dharmakīrti states that floaters are not objects, that is, not real objects because they are not determined as real objects (artha),4 presumably by people in the context of their everyday practice.5 This statement, albeit not an argument, is already a clarification and an indication towards where things are going: illusory objects are not real objects, which implies that universals are. A word about the word artha. The Sanskrit word artha literally means “aim,” “goal,” “purpose,” “telos,” and a most commonly derived meaning is “referent.” For instance,

2 Cf. PS I, p. 1.15-20 (I quote the verse together with the Vṛtti): pratyakṣam anumānaṃ ca pramāṇe. te dve eva, yasmāl lakṣaṇadvayaṃ prameyaṃ. na hi svasāmānyalakṣaṇābhyām anyat prameyam asti. svalakṣaṇaviṣayaṃ ca pratyakṣaṃ sāmānyalakṣaṇaviṣayam anumānam iti pratipādayiṣyāmaḥ. 3 PS I, p. 2.8: nāmajātyādiyojanā. The translation quotes Hattori 1968: 25. 4 PV 3.1: pramāṇaṃ dvividhaṃ meyadvaividhyāc chaktyaśaktitaḥ | arthakriyāyāṃ keśādir nārtho ’narthādhimokṣataḥ || The PV verses quoted here follow our forthcoming edition. For the reading pramāṇaṃ dvividhaṃ meyadvaividhyāt we follow Inami et al. 2002. See also Dreyfus 1997: 297–298 and n. 53, Dunne 2004: 391, Hugon 2011: 369. 5 This is how Manorathanandin and Jinendrabuddhi interpret the statement. It could also be understood, of course, as determined by means of knowledge or philosophical investigation. However, the exact implication is of no consequence for us here.

in a linguistic context, it is often used to refer to the referent of a word, in which case it is habitually translated as “meaning.” In an epistemological context, the word is used to refer to the referent of a cognition or an awareness (jñāna, vijñāna, etc.), and unless one presupposes an idealistic or illusionistic metaphysics, it refers to the cognition’s counterpart in reality, to a real thing that corresponds to the cognition’s content or to the image that appears in the cognition. The common German translation of

artha as Gegenstand conveys this meaning of artha well: something that stands against or towards something else. To come back to our question, how can the universal, which does not exist, be the referent, the objective counterpart, of a valid cognition? And inversely, if the universal is a real object, how could it not be capable of efficient action? After all, the criterion of true existence (paramārthasattva) is the capacity to act efficiently.6 Dharmakīrti is a very laconic, at times obscure writer, and to understand his arguments we often have to rely on his commentators. The most important of his commentators is Prajñākaragupta who lived in the second half of the 8th century (750–810

according to Ono) and was one of the most brilliant philosophers in the Buddhist epistemological tradition. In the following, I will rely on his great (both in size and importance) commentary the Pramāṇavārttikālaṅkāra. In his comment on the above-mentioned verse (PV 3.1), Prajñākaragupta explains the negation in the statement about the non-capacity of the universal to produce efficient action as an implicative negation, referring thereby to the well known division of negations as implicative and non-implicative

(paryudāsapratiṣedha and prasajyapratiṣedha). A nonimplicative negation is merely a negation, whereas an implicative negation actually implies a positive statement by means of exclusion. For instance, when we say that someone is not at home, we usually imply that she is outside; when we say that someone does not eat during the day, we imply that she eats at night. So what does the implicative negation mean in our case? It means that the universal does not produce efficient action directly, but that it produces it indirectly. An illusory object like a floater produces no efficient action, neither directly nor indirectly, and is therefore not a real object neither directly nor indirectly (PVA 169.25-26):

6 PV 3.3: arthakriyāsamarthaṃ yat tad atra paramārthasat | anyat saṃvṛtisat prokte te svasāmānyalakṣaṇe || This verse has been often translated and discussed many times, for instance in Dunne 2004: 292. Yoshimizu (2007: 1050) deals with the Sautrāntika background of this distinction, and Moriyama (1991: 202, 206) points out Jñānagarbha’s and Haribhadra’s response and criticism of it. (Dharmakīrti’s tenet is, of course, an anathema to the Madhyamakas who consider causally arisen entities unreal.) For Go ram pa’s comment on PV 3.2a see Yoshimizu 2007: 1067–1068 and for Gyel tsap’s Dreyfus 1997: 117. See also Ono 1999: 313 on the relation of this verse to the sattvānumāna; on its relation to Abhidharma see Katsura 2011: 275–276. The first half of the verse is repeated in PVin II 55ab; cf. also PV I 166ab. See also Lindner 1984: 152–153.

pāraṃparyeṇa śaktir evāśaktiḥ paryudāsavṛttyā, na śaktyabhāvaḥ. keśādes tu taimirikopalabdhasya śābdādyupalabdhasya ca nārthatattvaṃ sākṣāt pāraṃparyeṇa vā śaktatvenāpratīte. “Non-capacity is nothing but capacity in an indirect way by the [negation’s] function of exclusion, [it is] not absence of capacity [as such]. The hairs, etc., however, which are perceived by a person who suffers from timira, and something that is perceived by a verbal [[[cognition]]] and so on, do not possess reality as objects because they are not perceived as capable, neither directly nor indirectly.” However, with this we do not solve the problem, but only take it elsewhere. If the universal does not exist, how can it produce an efficient action, even indirectly? In PV 3.7, Dharmakīrti returns to the distinction between universals and illusory objects. Floaters and so forth are not universals because they are not determined as real objects (artha).7 And Prajñākaragupta explains that the universal is certainly (eva) a real object (artha)

inasmuch as a real object which is a particular by nature is in this case being perceived in the form of a universal, that is, in a general form.8 The somewhat puzzled opponent then asks: If the real object is a particular, then – since in this case it causes an efficient action – how is it a universal? And if it does not cause an efficient action, how could it be a real object (artha)? For something is a real object (artha) by the fact that it causes an efficient action. And one and the same thing cannot cause and not cause an efficient action.9 Prajñākaragupta answers that this is not impossible. The thing under discussion is a universal as well as a real object inasmuch as it is incapable of efficient action directly, but is capable indirectly; thus is capable and incapable at the same time. The universal is a real object because it is determined as a real object. From the point of view of absolute reality, the universal is not a real object; the status of a real object is ascribed to it metaphorically because of a mental transposition of the characteristics of a particular to it, because it indirectly has the effect of a real object.10

7 PV 3.7: keśādayo na sāmānyam anarthābhiniveśataḥ | jñeyatvena grahād doṣo nābhāveṣu prasajyate || This verse is translated in Dunne 2004: 394, and briefly discussed in Meindersma 1991: 171 and Hugon 2011: 373. For the second half of the verse see also below. 8 PVA 189.23: artha eva hi sāmānyarūpatayā svalakṣaṇātmā pratīyamānaḥ sāmānyam. Here, Prajñākaragupta alludes, of course, to PV 3.54cd: tasya svapararūpābhyāṃ gater meyadvayaṃ matam || 9 PVA 189.23f.: yady asau svalakṣaṇam arthakriyākāritvāt kathaṃ sāmānyam anarthakriyākāritve ca katham arthaḥ. arthakriyākāritvenārthatvāt. na ca dvayam ekatrārthakriyākāritvam itarac ca. 10 PVA 189.25-27: naitad durghaṭam. sākṣād arthakriyāśaktaṃ pāramparyeṇa śaktitaḥ | sāmānyaṃ ca tad arthaś ca śaktaṃ cāśaktam eva ca ||173|| sāmānyam arthādhyavasāyenārthaḥ. paramārthato nārtha eva. upacaritam arthatvam asya paramparayārthakāryatvāt (anyārthatvāt).* *Sāṅkṛtyāyana’s note 3 thereon: “T na bhāṣāntare.” I am not quite sure what he means; perhaps that anyārthatvāt has no equivalent in the Tibetan.

Another problem arises in this connection: If universals are real objects, what about referents of negative assertions? An example for a negative assertion would be: “Here on the table there is no bottle.” Now the referent of this statement, namely, the absence of a bottle, is considered by Dharmakīrti to be a universal, but surely, an absence cannot be a real object (artha). On this point there seems to be a difference in terminology between Dharmakīrti’s commentators. For Prajñākaragupta absences are arthas,11 whereas for Manorathanandin they are not, yet they are conventionally determined to be objects of awareness, which in this context probably means objects of valid cognition (jñeya).12 The difference in terminology among the commentators may be explained as being due to a shift in Dharmakīrti’s terminology. Hugon has suggested recently (2011: 375) that Dharmakīrti considered the universal to be artha in the Pramāṇavārttika, but abandoned this position in the Pramāṇaviniścaya. Thus, we are dealing with three kinds of

phenomena: illusory objects, universals such as ‘being a bottle,’ and special universals such as absences or non-existences of things which are the results and referents of inferences that something does not exist. The latter are also distinct from illusory objects because they are correctly cognized.13 Therefore, at least according to Prajñākaragupta, absences too are real objects/things (artha), and that for two reasons: First, inasmuch as they are in the final analysis other things, which have efficient action; for instance, the object of the apprehension of the absence of a bottle on the table is in the final analysis nothing but the table. Second, and here Prajñākaragupta

relies on the literal meaning of the word artha, absences serve a purpose, for instance, dispel fear or doubt about something by proving that it does not exist. PVA 190.10-14: abhāvo ʼpi hi kenacid ākāreṇa jñeya eva. anyabhāva evābhāva iti siddhāntaḥ, sa cārthakriyākāryeṇa. atha vāsāv api pramāṇena paricchidyata eva. prayojanavattvāc cāsāv arthaḥ. abhāvenāpi hi prayojanam eva niḥśaṅkagamanārthatvāt. tatas tatrāpy arthādhimokṣa iti, artha evābhāvaḥ svena rūpeṇa, na punaḥ keśādir, bhrāntavijñeyatayā paricchedāt. “Absence too is indeed nothing but an object of [valid] cognition with a certain form. The established tenet is that absence/non-being is nothing but the presence of something

else/another being. And it is [correctly known (jñeya)] through its result, which is efficient action. Or rather, this [absence] too is indeed determined by a means of knowledge. And because it fulfils a purpose, this [absence] is a real object. For a purpose is also [fulfilled] by an absence because its purpose is fearless going [and so forth]. Therefore, it too is determined as a real object. Consequently, absence is indeed a real object sui generis (svena rūpeṇa), but floaters and so forth are not [real

objects] because they are determined as erroneous objects of cognition.” Dharmakīrti’s universal is of course not the eternal entity maintained by the realists. It is not an entity distinct (pṛthak) from the particular; it also does have the particular 11 Cf. PVA 190.10-14 quoted below. 12 PVV 115.13-14: yady apy abhāve ʼrtharūpatāyā (read –tayā?) adhyavasāyo nāsti jñeyatayā tv asti lokasya. 13 PV 3.7cd (quoted above): jñeyatvena grahād doṣo nābhāveṣu prasajyate ||. jñeyatva clearly has to be understood here as prameyatva.

as its support (āśraya, cf. PV 3.25cd). It has no svabhāva (cf. PV 3.27b, 28) because it cannot be established as different from the particulars.14 According to Prajñākaragupta, if the universal would be different from the particular, it is not connected to the particular and thus not perceived, and therefore lacks own nature.15 For him, this also distinguishes the universal from non-existing things: Non-existing things do not have dependence relationship with individual things, but universals do. Prajñākaragupta explains the difference between universals and non-existent things in their relationship to individual things as follows. PVA 200.5-7: vastusāmānyavādino ’pi gotvaṃ karkādikasya kasmān na bhavati. apekṣābhāvāt. na hi karkādivyaktyā tad vyajyate. evaṃ tarhi śābaleyādiṣv apekṣābhāvād eva kharaviṣāṇādikasya na tatsāmānyarūpatā. gotvasya tu tadāśrayaṇena bhāvāt tatra ca pravartanād bhavaty apekṣālakṣaṇaḥ sambandha iti na doṣaḥ. “Why is it – also for the [opponent] who holds that universals are real – that [[[Wikipedia:individual|individual]]]

white horses and the like do not possess cow-ness? [Opponent:] Because there is no relationship of dependence [between them]; for it (i.e., cow-ness) is not manifested by individuals such as white horses. [Proponent:] If so, because [[[non-existent]] things] such as the horn of a donkey do not at all depend on [[[Wikipedia:individuals|individuals]]] such as spotted cow (śābaleya), they (i.e., non-existent things) do not have the form of their universals. Cow-ness, on the other hand, has a relationship characterized by dependence [on such individuals as the spotted cows] because it exists as being supported by them, and also because one is prompted to act towards them. Therefore, there is no fault [in our theory].” To go back to Dharmakīrti, the universal only appears to have an own nature due to an error that is caused by repeated experience of individuals and that has no

beginning: tasyāṃ rūpāvabhāso yas tattvenārthasya vā grahaḥ | bhrāntiḥ sānādikālīnadarśanābhyāsanirmitā || PV 3.29 “The appearance of a form/own nature in respect to a universal, or the apprehension of an object as that (i.e., as having the nature of a universal),16 is an error that is caused by repeated experience that has no beginning.”17 Although the cognition of a universal is valid inasmuch as it leads to an efficient action, it is also false: In the cognition of a universal one perceives things that are distinct from 14 There is an extensive literature on Dharmakīrti’s theory of the universal which relies mainly on the more detailed exposition in the first chapter of the PV.

See, for instance, Dreyfus 1997, chapter 7 and Dunne 2004: 113–144. 15 According to the PVV on 27ab, the universal cannot be expressed as identical with or different from the individual things. In fact, it cannot even be expressed as identical with or different from other universals (PV 3.33cd). The impossibility of being expressed as identical with or different from something real (tattvānyatva) is occasionally used by Dharmakīrti as a criterion for unreality. 16 This interpretation follows Manorathanandin (PVV 133.16: tattvena jātisvabhāvena). Prajñākaragupta’s interpretation of tattvena is more complex and possibly offers two readings of the text: tattvena and tattve na. For lack of space, it cannot be discussed here. 17 See also Dunne 2004: 141.

each other in a general form, and this form is false, for one perceives an external thing in a form that it does not have. The general form of the universal spils (abhiplava), so to speak, onto the individuals.18 Consequently, the universal has a somewhat ambiguous status: It is said to be a real object (artha), but is not a real thing (vastu)19 (PV 3.33a) and has no own nature (niḥsvabhāva, arūpa) (PV 3.32c); its cognition is false (bhrānta) and yet it is valid (pramā). Dharmakīrti takes up this topic again in 3.50 and

the following verses: The universal as such, i.e., as distinct from the individuals, is never perceived, and is thus incapable of producing even its own cognition. For this reason it is said to lack own nature because this incapacity is the characteristic sign of unreal things (avastu).20 Although unreal, the universal is divided into three kinds, depending on whether it has an existing thing as its support, non-existing thing, or both.21 Here too the commentators differ in their explanations. According to

Manorathanandin, who follows Devendrabuddhi, the division refers to the extension of the universals. For instance, the universal “being produced” (kṛtakatva), which is an inferential sign, has existing things, such as the visible (rūpa), as its support; non perception (anupalabdhi) of something that is in principle visible and non having been

produced (anutpattimattva), etc., extends to non-existing things; these too are inferential signs. Non-perception in general (anupalabdhimātra) extends to both non-existing perceptible objects and existing imperceptible objects; similarly the universal “being knowable” (jñeyatva) extends to both since both existing and non-existing things can be known. However, Manorathanandin does not mention that these universals serve as universal signs,22 and I am not aware of inferences in which they do. Prajñākaragupta suggests

two alternative explanations of the above division of universals. The text of the second is somewhat corrupt, but the explanation seems to be the same as the one proposed by Devendrabuddhi and Manorathanandin. In the first interpretation, however, the three kinds of universals correspond to the three types of valid reasons (svabhāvahetu, anupalabdhihetu and kāryahetu). A conceptual cognition that has something existing as its support is one that has arisen from a reason that is its own nature (svabhāvahetu). A conceptual construction that has non-existing thing as its support has arisen from a reason consisting in non-perception (anupalabdhihetu), and

18 Cf. PV 3.31: sāmānyabuddhau sāmānyenārūpāyām apīkṣaṇāt | arthabhrāntir apīṣyeta sāmānyaṃ sāpy abhiplavāt || 19 The terminology is indeed confusing; in Sanskrit philosophical texts artha (in the sense of object) is usually considered to be vastu. 20 PV 3.50d: tad (i.e., jñānamātrārthakaraṇam) dhy avastuṣu lakṣaṇam. See further PVSV 84.10: tasmāt sarvaṃ sāmānyam anarthakriyāyogyatvād avastu. See also Dunne 2004: 86. 21 PV 3.51cd: sāmānyaṃ trividhaṃ tac ca bhāvābhāvobhayāśrayāt ||. On Go ram pa’s response to this verse, see Dreyfus 1997: 157. See also the parallel passages in PVSV discussed in Steinkellner 1971: 195–197 and Hugon 2011: 373–374. 22 PVV 131.2-5: kiṃcid bhāvopādānaṃ sāmānyaṃ yathā rūpādīn bhāvān āśritya kṛtakādiśabdavācyaṃ liṅgam. abhāvopādānaṃ yathopalabdhilakṣaṇaprāptasyāsato ʼnupalabdhir anutpattimattvādi ca. tad dhy abhāvāśrayaṃ sāmānyaṃ liṅgam. ubhayāśrayam anupalabdhimātraṃ jñeyatvādi ca bhāvābhāvasādhāraṇatvāt.

the one that is supported by both existing and non-existing things from a reason that is an effect (kāryahetu).23 Prajñākaragupta’s two interpretations are not mutually exclusive, but their difference is considerable: according to the first interpretation, the scope of the statement about the triple division is limited to universals insofar as they occur in inferences, whereas according to the second interpretation, the division of universals is a general one (which also includes, of course, universals

used as reasons of inference).24 The fact that the universal has an existing thing as its support or basis does not mean that its cognition has an existing thing as its object (PV 3.52). If this were the case, the sense faculties would be superfluous. If the universal, which is the referent of a word, were real, the cognition of a word such as “peach” would have the same effect as seeing or tasting a peach. Thus, one would not need the sense faculties in order to experience a particular peach. Furthermore, it is

clear that the cognition of a universal does not refer to existing things because one can speak about past and future things, e.g., when one says, ‘there was a pot,’ ‘I will make a pot,’ and so on (PV 3.52d).25 In the final analysis, the object to be cognized is only the particular (PV 3.53d: meyaṃ tv ekaṃ svalakṣaṇam). If one says that there are two kinds of valid cognition, this is only because the particular is cognized in two different forms, with its own form and with a general form (PV 3.54cd: tasya svapararūpābhyāṃ gater meyadvayaṃ matam).26 Dharmakīrti has made up a nice example to illustrate the difference between universals and illusory objects (or in fact among different universals). Even though all universals are unreal, and even though all cognitions of universals are false, certain cognitions of universals lead to efficient action and other do not. Imagine two persons, the one mistaking the light of a lamp falling through a keyhole27 for a jewel, the other mistaking the glow of the jewel for the jewel itself. They both run to fetch the desired object, but only one of them will get the jewel.28

23 PVA 212.8-9: tathā hi bhāvopādāno vikalpaḥ svabhāvahetor udayī. abhāvopādāno ʼnupalabdheḥ. ubhayopādāno vikalpaḥ kāryahetor iti. 24 The threefold division of universals appears also in PV 1.205-206: anādivāsanodbhūtavikalpapariniṣṭhitaḥ | śabdārthas trividho dharmo bhāvābhāvobhayāśrayaḥ || 205|| tasmin bhāvānupādāne sādhye ʼsyānupalambhanam tathā hetur na tasyaivābhāvaḥ śabdaprayogataḥ || 206|| See also PVin III 53. Verse 205 is translated in Steinkellner 1971: 189, n. 43, see also Tillemans 1999: 176. In neither of these contexts I can discern a threefold division of universals according to the three kinds of reason. 25 One has to note, however, that universals and particulars are not

co-extensive with things that are the objects of a word or not respectively. As Prajñākaragupta clearly puts it, the universal is an object only of a word, but it is not the case that whatever is the object of a word is a universal. And what is not the object of a word is not thereby a particular. Rather, the particular is not, i.e., never, the object of a word. See PVA 170.20-21: śabdasyaiva viṣayaḥ sāmānyam. na tu śabdasya yo viṣayas tat sāmānyam eva. aviṣayo ʼpi na svalakṣaṇam evāpi tv aviṣaya eva svalakṣaṇam ity artho vivakṣitaḥ. I do not know to which further objects of words Prajñākaragupta refers here. 26 See Katsura 1991: 136 and 2011: 276, Hugon 2011: 370. 27 Cf. NBhū 197.22-23: … kuñcikāvivaraniḥsṛtāyāṃ prabhāyām … 28 See PV 3.57. This verse has been translated many times, most recently in McCrea 2011: 321. It is repeated by Dharmakīrti in PVin

The example is illuminating, but not unproblematic from a systemic point of view. Śākyabuddhi raises an objection that the cognition in the second case, the one of the glow of the jewel that leads to the obtainment of the jewel, would amount to a third type of means of knowledge, since it is neither perception nor inference. It is not perception because it is erroneous and conceptual. Nor is it an inference because it does not arise from an inferential sign (liṅga). Nor is it a cognition that is not pramāṇa because it

agrees with reality (vastusaṃvāda) inasmuch as it leads to efficient action. Śākyabuddhi’s reply is that the cognition in question is in fact an inference. The general characteristic of inference, he says, is this: Inference is the cognition of an object beyond the range of the senses (parokṣa) from its connection to another object. And in the present case, the erroneous cognition of the jewel arises because of the connection of the sense to the light of jewel and because the latter (i.e. the light of the jewel) is the result of the jewel. Therefore, because the cognition of the jewel arises from an inferential sign that is a result of that jewel (kāryaliṅgajatva), it is nothing but

inference. First a non-erroneous visual cognition of the light of the jewel arises. Through this it is apprehended that it has the nature of an inferential sign that consists in a result (kāryaliṅgasvarūpa) because only the particular is an inferential sign…29 However, the reasoning is not very convincing. Although one can establish a concomitance in the form wherever there is light of a jewel, there is a jewel, the putative inference happens unconsciously. One would similarly have to assume that the erroneous cognition

of fire with regard to smoke is actually an inference of fire from smoke. Whatever the case may be, a further distinction between universals seems necessary. In the examples mentioned above, the universals that rely on both existing and nonexisting things such as “being known” (jñeyatva), and even those that rely exclusively on non-existing things (e.g., the non-perception of non-existing, but in principle perceptible things), agree somehow with reality.30 They could be said to cognize particulars in an indirect way and in a form that is not the particulars’ own (pararūpa, cf. PV 3.54cd quoted above). There are, however, universals, (or could these be concepts without universals?31)

Sanskrit fragment in Steinkellner 1981: 291: tadā pratyakṣānumānavyatiriktaṃ tṛtīyam idaṃ pramāṇam āpatitam. tathā hi maṇiprabhāyāṃ maṇibuddher (read –buddhir) na pratyakṣam, bhrāntatvāt savikalpakatvāc ca. pratyakṣaṃ tv etadviparītam. nāpy anumānam, aliṅgajatvāt. na cāpramāṇaṃ vastusaṃvādāt. atrocyate. anumānam evaitat. tathā hy anumānasya sāmānyalakṣaṇam anantaraṃ sthāpayiṣyate: parokṣārthasyānyasaṃbandhāt pratipattir anumānam iti. iha ca maṇau maṇiprabhāsambandhāt tatkāryatvāt tasyām <maṇi>prabhāyāṃ maṇibhrāntir utpadyate. tataḥ kāryaliṅgajatvād anumānam eva. tathā hi maṇiprabhāyām ādāv abhrāntam eva cakṣurvijñānam upajāyate, tena ca kāryaliṅgasvarūpam adhigatam. yataḥ

svalakṣaṇam eva liṅgam … 30 A different interpretation is found in Steinkellner 1971: 195, n. 59. According to Steinkellner the third kind of universals refers to metaphysical errors such as primordial matter (prakṛti); see also Vetter 1968. Note that both Steinkellner and Vetter refer to the PV 1.205 (not PV 3.51) and that the example of pradhāna is suggested by Karṇakagomin. However, the formulation in the two verses is too similar to warrant a different interpretation. Interestingly, Manorathanandin on PV 1.206b suggests that pradhāna and īśvara rely both on existing and non-existing things: ubhayopādānaḥ pradhāneśvarādiḥ. This does not seem to make sense and perhaps the text is corrupt here. 31 Steinkellner (1971: 195ff.) speaks of “Vorstellung ohne Begriff”. However, this distinction does not amount to the postulation of a conceptual construction without a universal; rather, if I understand him correctly, only universals that rely on existing things are termed “Begriff”.

that have no connection at all to any particulars, for instance, the concept of primordial matter (prakṛti, pradhāna) mistakenly maintained by the Sāṃkhyas, the concept of God (īśvara) postulated by the Nyāya-Vaiśeṣikas, or simply the famous son of a barren woman.32 Such fictitious entities, inasmuch as they have no counterpart in reality, cannot exclude anything from any existing particular. Dharmakīrti (PVSV 92.21, see also Steinkellner 1971: 197) calls these conceptual constructions the “chief” or “principal

distinctions” (mukhyaviveka). I assume that viveka here is interchangeable with apoha33: the “chief distinction” must be the exclusion of everything that exists from something that does not exist.34 Basically, however, these conceptual constructions operate just as those that rely on existing things, and in fact they are the clearest cases of exclusion. For even though there could be a doubt whether conceptual constructions that refer to existing things do not have a referent in reality after all, in the case of conceptual constructions referring to non-existing things such doubt does not arise in the first place.35

32 This example (vandhyāsuta) is used by Śāntarakṣita in the parallel passage TS 1201. 33 Steinkellner (1971: 197) provides two interpretations of viveka; the one, which I follow here, as “Sonderung,” the other, in his translation of PVSV 92.21, as “Fehlen” (absence). In the latter case, he seems to follow Frauwallner 1933: 94 (= Kl. Schr. 449). 34 The question may arise as to whether two different concepts of illusory entities, which equally exclude all things, would have the same meaning (apoha). For instance, the

notions of prakṛti and puruṣa, both being completely non-existent and thus excluding all existing things (at least according to the Buddhists), would have to be synonymous. This, however, is not the case because the exclusion applies not only to individuals/particulars but also to all other concepts. Thus, the exclusion has a double function; the distinction between these two functions roughly corresponds to the one between sense and reference, connotation and denotation, intention and extension, etc. This double

function is also seen, of course, in concepts/exclusions that rely on existing things, e.g., “being a product” and “being impermanent”; see the discussion in Tillemans 1986. I owe this reference to the kindness of Pascale Hugon. The twofold function of apoha can be contrasted with the inability to distinguish universals from their particulars and from other universals. According to Dharmakīrti, the unversals as conceived by the Nyāya-Vaiśeṣikas cannot be said to be either different from or identical to (their)

particulars or other universals. It is this inexpressibility which testifies to their unreality; cf. PV 3.32cd-33: niḥsvabhāvatayāvācyaṃ kutaścid vacanān matam || yadi vastu na vastūnām avācyatvaṃ kathaṃcana | naiva vācyam upādānabhedād bhedopacārataḥ || “Because [the universal] lacks its own nature, it cannot be expressed [as being identical to or different from the particulars]. If it is thought to be a real thing because it is expressed [as being different] from some [other universal], [it should also be possible to express it as being identical to or different from the particulars, because] real things are in no way inexpressible. [A universal] can certainly not be expressed [even as

being identical to or different from other universals] because one speaks about a difference [between universals] only figuratively on the basis of the difference between [their] substrata.” The inexpressibility of identity and difference is used by Dharmakīrti several times as a criterion for unreality, e.g., PVSV 144.11, 148.5, PV 3.25-27, VN 14.11, 15.6. 35 PV I 185: rūpābhāvād abhāvasya śabdā rūpābhidhāyinaḥ | nāśaṅkyā eva siddhās te ʼto vyavacchedavācakāḥ || “Because a non-existing thing has no form (i.e., own nature), words [that denote non-existing things] cannot at all be suspected to designate [an existing] form. Therefore, it is established for you [that words] express exclusion.”

our world is populated by different kinds of non-existing things. There are illusory objects (conceptual36 and non-conceptual), which include universals, which again include absences. The universals are subdivided into two groups, depending on whether they lead to efficient action or not. Although different examples are mentioned and discussed, Dharmakīrti does not seem to suggest different mechanisms for misconceptions that appertain to everyday life such as the light of a lamp mistook for a jewel, metaphysical

errors such as the concept of primordial matter and God, or misconceptions involving logical impossibilities such as sons of barren women. The former group of universals, i.e., those that lead to efficient action, is further divided into three groups depending on whether they rest on existing things, non existing things, or both. These universals, which are objects of knowledge, are, to be sure, unreal, and yet they are not completely unreal like other objects of illusion, for they are somehow and indirectly connected to real objects. They can be said to be the erroneous aspects through which one cognizes real things.

Abbreviations and Bibliography Dreyfus 1997: George Dreyfus, Recognizing Reality. Albany 1997. Dunne 2004: John Dunne, Foundations of Dharmakīrti’s Philosophy. Boston 2004. Frauwallner 1933: Erich Frauwallner, „Beiträge zur Apohalehre. I. Dharmakīrti. Übersetzung (Fortsetzung)“. WZKM 40, 1933: 51–94. Hattori 1968: Masaaki Hattori, Dignāga, On Perception. Cambridge, Mass. 1968. Hugon 2011: Pascale Hugon, “Is Dharmakīrti Grabbing the Rabbit by the Horns? A Reassesment of the Scope of prameya in Dharmakīrtian Epistemology”. Journal of Indian Philosophy 39, 2011: 367–389. Inami et al. 2002: Masahiro Inami, Miyako Notake, Keijin Hayashi, Shinya Moriyama, “Prajñākaragupta’s View on the Two Kinds of Valid Cognitions and the Two Kinds of Objects”. Nanto Bukkyō 81, 2002: 1–53 [in Japanese]. Katsura 1991: Shoryu Katsura, “Dignāga and Dharmakirti on apoha”. In:

Ernst Steinkellner (ed.), Studies in the Buddhist Epistemological Tradition. Wien 1991: 129–146. Katsura 2011: Shoryu Katsura, “From Abhidharma to Dharmakīrti – With a special reference to the concept of svabhāva”. In: Helmut Krasser et al. (eds), Religion and Logic in Buddhist Philosophical Analysis. Wien 2011: 271–280. Lindner 1984: Christian Lindner, “Marginalia to Dharmakīrti’s Pramāṇaviniścaya”. WZKS 1984: 149–175. McCrea 2011: Laurence McCrea, “Prajñākaragupta on the pramāṇas and their objects”. In: H. Krasser

et al. (eds), Religion and Logic in Buddhist Philosophical Analysis. Wien 2011: 319–328. Meindersma 1991: Tabe E. Meindersma, “A Brief Reference to apoha Theory in the Section on paralokasiddhi in Pramaṇavārttika II”. In: Ernst Steinkellner (ed.), Studies in the Buddhist Epistemological Tradition. Wien 1991: 169–174. Cf. also TS 1202: rūpābhāvād abhāvānāṃ śabdā jātyādivācakāḥ | nāśaṅkyā eva siddhās te nirbhāsasyaiva sūcakāḥ || 36 Every universal is a conceptual construction, but if, as Prajñākaragupta stated, there are some objects of words that are not universals, does that mean that not every conceptual construction is a universal? See note 25 above.

Moriyama 1991: Shinya Moriyama, “The Later Madhyamaka and Dharmakīrti”. In: Ernst Steinkellner (ed.), Studies in the Buddhist Epistemological Tradition. Wien 1991: 199–210. NBhū: Nyāyabhūṣaṇa of Bhāsarvajña. Ed. Svāmī Yogīndrānanda. Vārāṇāsī 1968. Ono 1999: Motoi Ono, “Dharmakīrti on asādhāraṇānaikāntika”. In. Shoryu Katsura (ed.), Dharmakīrti’s Thought and its Impact on Indian and Tibetan Philosophy. Wien 1999: 301–318. PS I: Dignāga’s Pramāṇasamuccaya, Chapter 1. A hypothetical reconstruction of the Sanskrit text with the help of the two Tibetan translations on the basis of the hitherto known Sanskrit fragments and the linguistic materials gained from Jinendrabuddhi’s Ṭīkā. By Ernst Steinkellner: Mat/dignaga_PS_1.pdf PV 3: Text in: Hiromasa Tosaki, Bukkyō-ninshikiron no kenkyū. 2 vols. Tokyo: Daitō shuppansha, 1979–1985. PVA: Rāhula Sāṅkṛtyāyana, Pramāṇavārttikabhāshyam or Vārttikālaṅkāraḥ of Prajñākaragupta: Being a Commentary on Dharmakīrti’s Pramāṇavārttikam. Patna: K.P. Jayaswal

Research Institute, 1953. PVin: Pramāṇaviniścaya Chapter 1 and 2 in: Ernst Steinkellner ed. Dharmakīrti’s Pramāṇaviniścaya, Chapter 1 and 2. Beijing/Vienna 2007. PVV: Rāhula Sāṅkṛtyāyana, ed. Dharmakīrti’s Pramāṇavārttika with a commentary by Manorathanandin. Appendix to JBORS 24-26, 1938–1940. PVSV: Raignero Gnoli, The Pramāṇavarttikam of Dharmakīrti, the first chapter with the autocommentary. Roma 1960. Steinkellner 1971: Ernst Steinkellner, “Wirklichkeit und Begriff bei Dharmakīrti”. WZKS 1971: 179–211. Steinkellner 1981: Ernst Steinkellner, “Philological Remarks on Śākyamati’s Pramāṇavārttikaṭīkā”. In: Studien zum Jinismus und Buddhismus, ed. K. Bruhn and A. Wezler, Wiesbaden 1981: 283–295. Tillemans 1986: Tom J.F. Tillemans, “Identity and Referential Opacity in Tibetan Buddhist Apoha Theory”. In: B.K. Matilal and R.D. Evans (eds.), Buddhist Logic and Epistemology. Dordrecht/Boston/ Lancaster/Tokyo 1986: 207–227. Tillemans 1999: Tom J.F. Tillemans, Scripture, Logic, Language. Essays on Dharmakīrti and his Tibetan

Successors. Boston 1999. TS: Tattvasaṅgraha of Shāntarakṣita. Ed. Dwarikadas Shastri. Varanasi 1981. Vetter 1968: Tilmann Vetter, “Das Problem des metaphysischen Beweises in der logischerkenntnistheoretischen Periode der indischen Philosophie“. ZDMG 118, 1968: 351–356. Yoshimizu 2007: Chizuko Yoshimizu, “Causal efficacy and spatiotemporal restriction: An analytical study of the Sautrāntika philosophy”. In: Pramāṇakīrtiḥ. Ed. Birgit Kellner et al. Wien 2007. Vol. 2: 1049