Skip to main content
SpringerLink
Log in

When Did Svatantra Inference Gain Its Autonomy? Śāntarakṣita and Kamalaśīla as Sources for a Tibetan Distinction

  • Published:
Journal of Indian Philosophy Aims and scope Submit manuscript

Abstract

This article examines Śāntarakṣita’s and Kamalaśīla’s understandings of svatantra and prasaṅga proofs in the attempt to clarify how and why Tibetan Prāsaṅgikas came to portray svatantra inference as an instance of the very thing Madhyamaka rejects. The article proceeds in four parts. (1) A brief comparison of Patsap Nyimadrak’s (pa tshab nyi ma grags, c. 1055–1145) portrayal of svatantra inference with Bhāviveka’s and Candrakīrti’s employment of this expression shows that Patsap expanded the meaning of it, charging its users with embracing a realism at odds with Madhyamaka emptiness. (2) Patsap’s arguments against Śāntarakṣita’s “neither one nor many” proof lead us consider the latter’s and Kamalaśīla’s understanding of svatantra and prasaṅga proofs, which turns out to be consistent with, if more sophisticated than, Bhāviveka’s and Candrakīrti’s usage. (3) An exploration of Kamalaśīla’s development of the proof of the naturelessness of all things, real and imagined, calls into question the svatantra-prasaṅga distinction but still leaves us without a source for Patsap’s characterization. (4) A consideration of Śāntarakṣita’s and Kamalaśīla’s use of svatantra to characterize Mīmāṃsā conceptions of an authorless Veda and “intrinsic validity,” as well as their own view of “self-awareness,” offers potential sources for characterizing (indeed, misunderstanding) valid cognition and, by extension, inference as “autonomous.”

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

Notes

  1. Patsap’s extant texts provide the earliest presently known use of Svātantrika (rang rgyud du smra ba’i dbu ma pa) and Prāsaṅgika (thal ʼgyur smra ba’i dbu ma) as subschools of Madhyamaka; see his Lamp Illuminating the Treatise (dbu ma rtsa ba’i shes rab gyi ti ga | bstan bcos sgron ma gsal bar byed pa), 39.i.6–39.i.7, among other places. Dreyfus and Tsering 2010 have pointed out that Patsap claims in this text to report his Kaśmiri teacher Hasumati’s (Mahāsumati) view, and so Kaśmiris may have already made the Svātantrika—Prāsaṅgika distinction prior to Tibetans adopting it. Patsap’s disciples refer to variations of this distinction, most notably Khutön Dodébar (khu ston mdo sde ʼbar), in a text that Apple (2018) has identified as likely authored by him and that includes a lengthy discussion of Svātantrika and Prāsaṅgika methods of realizing and demonstrating emptiness. Khutön’s text refers to Patsap’s discussion of the types of prasaṅgas, a discussion found in Patsap’s Explanation of the Difficult Points of the Prasannapadā (Tshig gsal ba’i dka’ ba bshad pa), 148, and so Khutön’s text must post-date that of his teacher and collaborator Patsap. Apple’s conclusion (2018, p. 942) that a Svātantrika-Prāsaṅgika distinction was “based on the study of Candrakīrti’s Madhyamakāvatāra” rather than the Prasannapadā would seem to go too far; rather, knowledge of the Prasannapadā allowed Tibetans to read a distinction stemming from Candrakīrti’s mature work into his earlier composition.

  2. These four occasions have been discussed in Yotsuya (1999, pp. 48–54). Yonezawa (2019) points out that in at least three of these usages, svatantra is construed adverbially (rang dbang du) and so, properly speaking, does not modify anumāna at all, but modifies an accompanying verb—in short, Bhāviveka tells us that Nāgārjuna stated an inference svatantra-ly, or “dependent on his own [position].”

  3. MacDonald, In Clear Words, vol. 1, 145.4–146.1: na ca mādhyamikasya svataḥ svatantram anumānaṃ kartuṃ yuktaṃ pakṣāntarābhyupagamābhāvāt ||). The context of these remarks, Nāgārjuna’s rejection of arising from the four alternatives (catuṣkoti), demands that we understand Candrakīrti here as rejecting that Mādhyamikas hold any “other positions” concerning arising. Candrakīrti’s rather straightforward statement here soon gives way to complexity concerning just what form of inference can be used for Madhyamaka pursuits; see Oetke (2006), a lengthy analysis of MacDonald (2003), which in turn is a response to Oetke (2003), which criticizes MacDonald (2000).

  4. Again, Bhāviveka did not call for Buddhapālita to state a svatantra inference; Candrakīrti’s rebuttal to Bhāviveka’s demand for an inference termed the requested inference svatantra. Candrakīrti, like Bhāviveka, referred to Nāgārjuna’s denial of the four alternatives of arising “theses” (pratijñā). For instance, in explaining that in Nāgārjuna’s opening stanza jātu has a temporal sense while kvacana has a locative sense, Candrakīrti points out that we are to understand the statement as “Not at all do any things arise from self, ever, anywhere” and notes “The [other] three theses are composed similarly” (MacDonald, In Clear Words, vol. 1, 139.1-139.2:) naiva svata utpannā jātu vidyante bhāvāḥ kvacana kecana | evaṃ pratijñātrayam api yojyam || On the other hand, Candrakīrti supports his denial of svatantra inference with references to Nāgārjuna’s Vigrahavyāvartanī 29 rejection of a thesis (pratijñā) and Āryadeva’s Catuḥśataka XVI.25 rejection of a “position” (pakṣa). Clearly, Candrakīrti has given his later interpreters much to consider in just what kinds of theses are problematic and admissable for Mādhyamikas.

  5. Pa tshab, sGron ma gsal bar byed pa, 44.i.9-44.i.10: rang rgyud ni chos can lartogs pa tshad mas grub pa’i sgo nas rang gi ’dod pa sgrub pa yin pa la |

  6. Zhang Thang sag pa, dBu ma tshig gsal gyi ti ka (Yoshimizu and Nemoto, eds.), 64: rang rgyud ni sgrub bya mtshan nyid dang ldan pa la tshul gsum rgol phyir rgol gnyis ka’i tshad mas grub pa zhig gis sgrub pa’o ||

  7. Yoshimizu (2010, pp. 448–452). Zhang notes, “Use of a svatantra reason entails a fully qualified probandum; since Mādhyamikas have no probandum that could be entailed by it, it is not reasonable to state a svatantra reason”; dBu ma tshig gsal gyi ti ka (Yoshimizu and Nemoto, eds.), 64: rang rgyud kyi he du byed pa la sgrub bya mtshan nyid dang ldan pas khyab la | dbu ma pa la khyab byed du gyur pa’i sgrub bya med pa’i phyir rang rgyud kyi he du brjod par mi rigs so ||

  8. Pa tshab, Tshig gsal ba’i dka’ ba bshad pa, 153.ii.6–153.ii.7: rang rgyud chos can lartsogs pa dngos po la grub pa dgos la | and sGron ma gsal bar byed pa, 42.ii.1-42.ii.2: de lta bu’i khyab pa grub po zer na khyed dbu ma pa’i ʼdod pa nyams te | khyed chos thamd rang bzhin med par ʼdod pa la gzhi gang la khyab pa grub pa de rang yang dag gi rang bzhin grub pas so || The latter statement comes in a Substantialist’s critique of Svātantrika, which is followed by a Svātantrika rebuttal, and then Patsap’s own Prāsaṅgika critique of Svātantrika. The Substantialist’s complaints against Svātantrika are amplified in the Prāsaṅgika attack. Both point out the incompatibility of inference and emptiness; only the conclusions differ: for the Substantialist, the incompatibility necessitates the falsity of emptiness, while for the Prāsaṅgika, the incompatibility portends the failures of inference.

  9. Pa tshab, sGron ma gsal bar byed pa, 40.ii.7: tshad ma la don yod par grub kyi rang bzhin med pa ma yin no |, 41.i.4–41.i.5: tshad mas yod par grub kyi med par grub pa ma yin no | and 42.ii.10-43.i.1: rang rgyud las ma gtogs pa’i tshad ma med pas so ||. These statements come from Patsap’s Substantialist but are again supported when he writes from his own Prāsaṅgika standpoint.

  10. Pa tshab, sGron ma gsal bar byed pa, 47.i.6: snang tshod pa’i dngos po lartsogs pa nyid yang dag gi rngos por byas nas | and Tshig gsal ba’i dka’ ba bshad pa, 153.ii.4: don dam gyi tha snyad ʼjig rten pa la byed pa. Another of Patsap’s disciples, Mabja Jangchup Tsöndrü (rma bya byang chub brtson ʼgrus, d. 1185), likewise pointed to valid cognition’s claimed factual status. As Doctor’s (2010: 430-431) work shows, Mabja speaks of “objectively gained valid cognition” to denote the purportedly factual nature of valid cognition’s conclusions, in which an “unmistaken mode of apprehension” yields knowledge of things as they truly are, without superimpositions, “abiding in reality” (dngos po la gnas pa); rMa bya Byang chub brtson ʼgrus, dBu ma rigs pa’i tshogs kyi rgyan De kho na nyid snang ba, 8b.7: shes bya rang gis sgro ma btags su dngos po la gnas pa sngar ma rtogs pa 1 myong pa dang mi ʼkhrul pa’i rtags las ji ltar gnas pa ltar rtogs pa la dngos po’i stobs kyis zhugs pa ’am ʼdzin stangs mi ʼkhrul ba’i tshad ma zhes bya la |

  11. MacDonald, In Clear Words, vol. 1, 175.1–175.3: yasmād yadaivotpādapratiṣedho ’tra sādhyadharmo ’bhipretaḥ | tadaiva dharmiṇas tadādhārasya viparyāsamātrāsāditātmabhāvasya pracyutiḥ svayam evānenāṅgīkṛtā | MacDonald (vol. 2, 109–110) translates, “[T]his very [one, namely, Bhāviveka] himself has accepted that precisely when the negation of arising is intended here as the property to be proved, there is the loss of the property possessor, the locus of that [property], whose [ascribed] existence has been procured through sheer error.”

  12. Tibetan Gélukpa (dge lugs pa) scholars read the problem of a “mutually appearing inferential subject” (chos can mthun snang) as a key piece of evidence for the inferiority of Bhāviveka’s view, seeing in his use of svatantra inference a false belief in the reality of the things that might serve as inferential subjects; see Hopkins (1989). On Candrakīrti’s view of perception, framed as a rejection of “the given,” that would have bearing on the impossibility of a Mādhyamika and non-Mādhyamika to perceive the same inferential subject, see Tillemans (1990, vol. 1, pp. 41–53). On early Tibetan interpretations of the problem of emptiness, when placed as the property-to-be-proved, undermining the inferential subject, see Vose (2015).

  13. For Jamyang Zhépa’s (ʼjam dbyangs bzhad pa) definition of svatantra, see Hopkins (1996 [1983]), the Tibetan text appearing after p. 910, labelled p. 3: rang dbang du grub pa dang rang bzhin gyis grub pa dang rang rgyud don gcig de lta bu’i tshul gsum pa’i rtags sbyor ni rang rgyud kyi rjes dpag yin pas, translated on pp. 584–585: “‘Existing under its own power,’ ‘existing inherently,’ and ‘autonomous’ are synonyms. Thus, the application of a reason [that is, a syllogism] the three aspects of which exist inherently is an autonomous inference.”

  14. Hopkins (1996 [1983], p. 848, n. 497): “Jam-ȳang-shay-ba here equates svatantra with terms meaning existing under the object’s own power, and thus the term has been translated as ‘autonomous’. However, most Western scholars and many non-Ge-luk-ba Tibetan scholars have take the term just to mean a syllogistic statement that the stater himself supports and thus is in his own continuum (rang rgyud).” Hopkins (pp. 846–847, n. 496) also notes that Stcherbatsky’s “independent argument” and Sprung’s “self-contained argument” do not capture the Gélukpa understanding.

  15. As we will see below, in discussing if and when prasaṅga arguments can become valid inferences/proofs, the epistemological tradition following Dharmakīrti speaks of valid arguments as having “actual reasons” (maula hetu) and does not use the expression svatantra to denote these. Some writing about this debate, however, interject the term svatantra into the discussion.

  16. Tillemans (1983, p. 312), drawing on Lakatos (1970). Tillemans there analyzed Tsong kha pa’s reading of Śāntarakṣita’s Madhyamakālaṃkāra, in which Tsong kha pa argues that Śāntarakṣita’s Madhyamaka ascribes “natures” on the conventional level.

  17. Pa tshab, sGron ma gsal bar byed pa, 43.i.8–43.i.11: chos can ji ltar snang ba tsam po gang gis kyang kyad par du ma byas pa tha snyad pa’i tshad mas grub pa de la gtan tshig 1 dang du ma bcad pa tsam yang tha snyad pa’i tshad mar mngon 3 ede pas grub | khyab pa yang gzhi gang zhig gi steng du yang dag tu yod pa la yang dag gi gcig dang du mas khyab pa’i grub* pa’i khyab byed | yang dag gi gcig dang du ma snang ba tsam po la ʼgags yang dag gi rang bzhin bkag** nas rang bzhin med pa grub ste | * Text reads khyab pa’irub, contracting the two syllables. ** Text reads b__ with two letters illegible (a fault of the reproduction, not the manuscript); bsal would also suit.

  18. Throughout, references to the Tattvasaṃgraha and –pañjikā are to Dwarikadas Shastri’s edition, except where noted. Tattvasaṃgraha 1424 (510.6–510.7: hetudharmapratītiś ca tatpratītir ato matā | tatpratītiḥ svatantrā ʼsti na tu kācid ihāparā ||) does use the term svatantra (in the negative) in relation to an inferential cognition, noting that it does not arise “independently” of a cause. Śāntarakṣita here considers a challenge to the claim that reasons come in only three varieties. His opponent lists a range of cases where one thing can be inferred from another in which the relationship between the two do not obviously fit into any of the three categories of valid reason (nature, effect, and non-observation). Among the cases that the challenger adduces is the arising of star clusters in the night sky: the regularity of celestial patterns allows one to infer that when one star cluster is visible, another—while not yet visible—is just below the horizon. Śāntarakṣita answers that this case really is an effect reason: from the observation of one star cluster, one infers its cause, which is the same “causal complex” (sāmagrī) that produces the soon-to-be-visible star cluster. The inferential cognition of that latter star cluster is not independent (svatantra) of or separate (apara) from the inference of the underlying cause of both star clusters. This denial that a particular inferential cognition is “independent” functions to rule out any other category of reason; Śāntarakṣita and Kamalaśīla affirm that inference depends on, or arises from, only three kinds of reasons. This is, of course, a very different proposition from the characterization of inference-for-other or proof as svatantra.

  19. Tattvasaṃgraha 304 (147.7–147.8): svatantre sādhane mate v. prasaṅgo; 3370–3371 (1070.5–1071.2): svatantraṃ sādhanaṃ matam v. prasaṅgasādhanaṃ; 3603cd–3606ab (1118.6–1119.1): svatantraṃ sādhanaṃ bhavet v. prasaṅgasādhanenedam aniṣṭaṃ codyate yadi and prasaṅgasādhane.

  20. Tattvasaṃgraha 613, 253.1: svātantryeṇa prasaṅgena sādhanaṃ yat pravarttate | Śāntarakṣita here quotes the Naiyāyika Śaṅkarasvāmin. For a study of the Tattvasaṃgraha’s use of quotes from Naiyāyika authors, see Marks (2019).

  21. Tattvasaṃgrahapañjikā ad 613, 253.12, rendering Śaṅkarasvāmin’s quote: svātantryeṇa prasaṅgamukhena vā yat sādhanaṃ kriyate.

  22. Tattvasaṃgrahapañjikā ad 3371, 1070.22: svātantryeṇa sādhanaṃ … prasaṅgasādhanaṃ and ad 3601–3604, 1118.12: etac ca prasaṅgasādhanaṃ mayoktaṃ na svātantryeṇety …

  23. The Tibetan translations of Kamalaśīla’s Madhyamakālaṃkārapañjikā and Madhyamakāloka lack a parallel grammar between the two kinds of logic, likely reflecting the Tattvasaṃgrahapañjikā’s use of the instrumental for svatantra and a compound for prasaṅga: the Madhyamakālaṃkārapañjikā and Madhyamakāloka contrast rang gi rgyud kyis sgrub pa with thal bar 'gyur bar sgrub pa and thal bar sgrub. We will examine those texts’ usage in the following section.

  24. The following investigations into the Tattvasaṃgraha’s treatment of the Buddha’s omniscience and the momentariness or non-existence of the entities constituting central non-Buddhist beliefs are not intended as in any way thorough treatments of these topics. Rather, my interest in these passages lies in what they tell us about Śāntarakṣita’s and Kamalaśīla’s understandings of svatantra and prasaṅga proofs.

  25. McClintock (2010, pp. 152–156) points out that Tattvasaṃgraha 3127–3245 detail Kumārila’s argument against omniscience, while stanzas 3246–3260 sketch Sāmaṭa’s and Yajñaṭa’s case against it. She analyzes the Mīmāṃsā argument and Śāntarakṣita’s rebuttal at 165–187. Stanzas 3358–3372 take up the charge that omniscience and “being a speaker” (vaktṛtva) are incompatible and provide Śāntarakṣita’s response. McClintock (p. 355, n. 770) treats stanzas 3596–3599, in which Śāntarakṣita offers one solution to the quandry: the Buddha might have conceptuality, but these are not afflictive conceptions and so do not stain him, as he knows that they are unreal.

  26. Tattvasaṃgraha 3370–3372, 1070.5–1071.4: ucyate yadi vaktṛtvaṃ svatantraṃ sādhanaṃ matam | tadānīm āśrayāsiddhaḥ sandigdhāsiddhatā ʼtha vā || asya cārthasya sandehāt sandigdhāsiddhatā sthirā | prasaṅgasādhanaṃ tasmāt tvayā vaktavyam īdṛśam || tatra cāgamamātreṇa siddho dharmaḥ prakāśyate | na tu tadbhāvasiddhyarthaṃ jñāpakaṃ vidyate param ||

  27. Tattvasaṃgrahapañjikā, 1070.21–1070.23: atra vikalpadvayaṃ kadācid vaktṛtvaṃ svātantryeṇa sādhanaṃ vābhipretaṃ bhavet prasaṅgasādhanaṃ vā | tatrādye pakṣe viśeṣeṇāśrayo na siddha ity āśrayāsiddhatā hetoḥ |

  28. Tattvasaṃgrahapañjikā, 1070.23–1070.25: atha sāmānyenāśrayo vivakṣitaḥ tathāpi yāvat prativādinaṃ prati pramāṇena vaktṛtvaṃ sādhyate tāvat sandigdhāsiddhatā ya eva tūbhayaniścitavācī sa eva sādhanam iti nyāyāt |

  29. Yotsuya (1999, p. 93, footnote 72) identifies a passage in Bhāviveka’s Prajñāpradīpa (182b.5) that explains that the inferential subject must be taken generically (spyir), without the specifics of either party’s assertions about it: tha snyad du bdag spyir khas blangs pa’i khyad par ma grags pa sel ba’i phyir skyon med do || We saw above Candrakīrti’s rejection of this possibility when emptiness is put forward as the property-to-be-proved. We will examine Kamalaśīla’s views and their position in the Dharmakīrti tradition below.

  30. Tattvasaṃgrahapañjikā, 1071.15–1071.18: tasmān mā bhūd ayaṃ doṣa iti prasaṅgasādhanam aṅgīkarttavyaṃ tvayā | tatrāpi prasaṅgasādhane ya evāvicāraramaṇīyatayā’ ʼgamamātrāt parasya prasiddho dharmaḥ sa eva sādhanatvena prakāśanīyaḥ parasparavirodhodbhāvanāya na tv asau pramāṇena sādhanīyaḥ niṣprayojanatvāt |

  31. Kamalaśīla’s use of the term here might serve to strengthen the argument that the Tattvasaṃgraha, or at least Kamalaśīla’s –pañjikā, was written from a Madhyamaka perspective.

  32. Tattvasaṃgrahapañjikā, 1071.18–1071.19: na ca vaktṛtvaṃ parasyāgamamātreṇa prasiddham ity ubhayathā ʼprasiddhatā hetoḥ ||

  33. Tattvasaṃgraha 3603cd–3604, 1118.6–1118.8: prasaṅgasādhanenedam aniṣṭaṃ codyate yadi || na ced vaktṛtvam iṣyeta nāgamopagamo bhavet | tatpraṇetāgameṣṭau tu tasya vaktṛtvam iṣyatām ||

  34. Tattvasaṃgraha 3605, 1118.9–1118.10: yady evam īdṛśo nyāyaḥ prasiddho nyāyavādinām | prasaṅgasādhane dharmaḥ śraddhāmātrāt parair mataḥ ||

  35. Tattvasaṃgrahapañjikā, 1118.25–1119.13: ya eva hi dharmaḥ pareṇāgamamātrād abhyupetaḥ sa eva prasaṅgasādhane ʼbhidhīyata iti nyāyaḥ |

  36. Tattvasaṃgraha 3606ab, 1119.1: yuktiprasiddhatāyāṃ ca svatantraṃ sādhanaṃ bhavet |

  37. Tattvasaṃgraha 3607–3608, 1119.3–1119.6: sambhārāvedhatas tasya puṃsaś cintāmaṇer iva | niḥsaranti yathākāmaṃ kuṭyādibhyo ʼpi deśanāḥ || ādhipatyaprapattyā ʼtaḥ praṇetā so ʼbhidhīyate | vikalpānugataṃ tasya na vaktṛtvaṃ prasajyate || See McClintock (2010, pp. 354–359) for a discussion of Śāntarakṣita’s and Kamalaśīla’s use of this metaphor, along with references to its sūtra sources and use in Candrakīrti’s Prasannapadā and Prajñākaramati’s Bodhicaryāvatārapañjikā. McClintock (pp. 357–358) frames Tattvasaṃgraha 3601–3604 as Śāntarakṣita’s answer to Sāmaṭa’s and Yajñaṭa’s objections to the Buddhist notion that insentient objects can preach the dharma; she labels his answer “a rather cheap trick.”

  38. Tattvasaṃgraha 304 finds Śāntarakṣita asking his Sāṃkhya opponent whether the attempt to prove the existence of primordial matter (pradhāna) endowed with an “intellect that has an insentient nature” (acetanātmikā buddhiḥ), given that for Sāṃkhya only puruṣa possesses sentience, is intended as a svatantra or prasaṅga proof. Tattvasaṃgraha 614 has the Naiyāyika Śaṅkarasvāmin asking whether Śāntarakṣita’s argument against “substance” (dravya, one of the sixteen padārthas, the existence of which Buddhists reject) is a proof by svatantra or by prasaṅga. Kamalaśīla’s additional use of this terminology, in comment to Tattvasaṃgraha 392-394, will be examined below.

  39. Pramāṇaviniścaya Chapter 3 (Hugon and Tomabechi, eds.), 4.4-4.8: paraparikalpitaiḥ prasaṅgaḥ … sa ekadharmopagame ʼparadharmopagamasandarśanārthaḥ | tadanabhyupagame cobhayanivṛttiḥ | and Pramāṇavārttika IV.12bcd (Tillemans, ed. and trans.), 21: parakalpitaiḥ | prasaṅgo dvayasambandhād ekābhāve ʼnyahānaye || “An [absurd] consequence is [drawn] by means of the other’s conceptual constructs; since the [consequence’s] two terms are necessarily connected, it serves to negate the second term in the absence of the first.”

  40. Iwata (1993, pp. 23–27) and (1997, pp. 427–428).

  41. Pramāṇaviniścaya Chapter 3 (Hugon and Tomabechi, eds.), 5.11–6.1: asati tu hetau maulasya hetor vyāpyavyāpakabhāvasādhanaprakāra eṣaḥ | na viparyayasādhanam, hetor apramāṇatvāt || Iwata’s analysis, translation, and Tibetan text of the first part of this statement is found in 1997, p. 428.

  42. I thank Pascale Hugon for providing Sanskrit passages from the diplomatic edition of the third chapter of Dharmottara’s Pramāṇaviniścayaṭīkā that she prepared in collaboration with Takashi Iwata and Toshikazu Watanabe, which will soon be published in the Sanskrit Texts from the Tibetan Autonomous Region series. Here, we read: paraparikalpitaiś ca yaḥ prasaṅgaḥ na sa svatantraḥ* karttavyaḥ hetor asiddhatvāt api tu prasaṅgaviparyayaniṣṭha iti darśayitum. *The editors suggest that svatantraḥ may be emended to svatantrataḥ, which I adopt here. This emendation agrees with the Tibetan, which reads (sde dge edition, Toh. 4227, 8a.2-8a.3): gzhan gyis kun brtags pas thal bar sgrub pa gang yin pa ni rang rgyud du bya ba ma yin te | gtan tshigs ma grub pa’i phyir ro || ʼon kyang thal ba bzlog pa’i mthar thug pa yin no zhes bstan pa'i phyir ro || Iwata (1993, p. 47) renders rang rgyud du “vom eigenen Standpunkt.” The passage comments on the autocommentary to Pramāṇaviniścaya III.2 (Hugon and Tomabechi, eds.), 4.4: paraparikalpitaiś prasaṅgaḥ …

  43. Dharmottara, Pramāṇaviniścayaṭīkā, 10b.7–11a.1: thal bar ʼgyur ba’i sgo nas kyang gnod par byed pa’i tshad ma ni ʼphen pa yin pa des na ʼdi ni sgrub par byed pa’i rnam pa gzhan yin no || The Sanskrit reads prasaṅgadvāreṇāpi tv ākṣipyate dhābakam ity anyo yaṃ sādhanaprakāraḥ which Hugon emends to prasaṅgadvāreṇāpi tv ākṣipyate bādhakam ity anyo ʼyaṃ sādhanaprakāraḥ. Note the Tibetan gnod par byed pa’i tshad ma (bādhakapramāṇa) for bādhaka.

  44. Dharmottara, Pramāṇaviniścayaṭīkā, 6b.4–6b.5: khyab par byed pa med pa de ni spyi la sogs pa med pa la yang grub pa nyid do || de’i phyir mngon par ʼdod pa go bar byed pa yin pa’i phyir de nyid ʼdir rtsa ba’i gtan tshigs su brjod pa yin gyi | See also Iwata (1993, pp. 55–56) and (1997, p. 430, notes 7 and 8). The Sanskrit manuscript is here problematic; the editors read sa ca vyāpākābhāvaḥ sā(mā)••••••••(bh)••••• (eva) tato bhipre(tasya gakatv)āt sa evā(tra) maulo hetur ucyate.

  45. Dharmottara, Pramāṇaviniścayaṭīkā, 8a.2: du ma med pa tsam nyid ni spyi med pa yang grub pa yin no || See also Iwata (1997, p. 430). The Sanskrit reads ʼnekatvābhāvamātraṃ cāsaty api sāmānye siddham.

  46. In this, Dharmottara is in line with Devendrabuddhi, who understands non-existent subjects to be valid bases of reasons and properties to be proven that are “mere exclusions” (rnam par gcod pa tsam) and Śākyabuddhi, who specifies that “mere exclusion” means “absolute negation” (med par dgag pa tsam). See Iwata (1999, pp. 160–162), who further notes that Śākyabuddhi accepts that subjects imputed by the opponent will have a conventional (vyavahāra) existence, whereas Dharmottara denies them any sort of existence.

  47. Prajñākaragupta, Pramāṇavārttikabhāṣya, ad IV.12, Sāṅkṛityāyana, ed., 481.25–481.26 with corrections from Iwata (1993, pp. 101–102): yadi vyāpi sāmānyaṃ vyāpitayaivādhyakṣeṇa pramāṇena gṛhyeta | na khalu viparītagrahaṇaṃ pramāṇam |

  48. Prajñākaragupta, Pramāṇavārttikabhāṣya, ad IV.12, Sāṅkṛityāyana, ed., 482.21–482.22 with corrections from Iwata 1993: 103: prasaṅgasādhanam etat | na tv ayaṃ maulo hetuḥ | yadi tvayā sāmānyaṃ vyāpy abhyupagamyate pramāṇena tat pratipattavyaṃ pratyakṣatadādinā | and Sāṅkṛityāyana, ed., 481.28–481.30 with corrections from Iwata (1993, p. 102): yady ayaṃ sthitapakṣaḥ syād bhaved dharmyasiddhidoṣaḥ parasparavirodhaprakaṭanaparam evaitat … anumānaṃ ca na vinā dṛṣṭāntam avatarati | Iwata (1993, pp. 67–68 and 1997, p. 431) adds Yamāri’s comment on this passage: ʼdi ltar thal ba dang bzlog pa gnyis kyis ni gzhan gyis yongs su brtags pa la ʼgal ba rjod pa tsam zhig byed par zad kyi | ʼdi gnyis las sangs rgyas pa’i grub pa’i mtha’i sgrub par byed pa ni ma yin no ||

  49. Prajñākaragupta, Pramāṇavārttikabhāṣya, ad IV.12, Sāṅkṛityāyana, ed., 483.6: virahamātrakasya hetutve tasya kenacid asambandhāt kathaṃ sādhyasādhanam | and 483.23–483.24: tasmān nātra maulo hetuḥ | sāmānyasya dharmiṇaḥ siddher abhāvāt |

  50. Prajñākaragupta, Pramāṇavārttikabhāṣya, ad IV.12, Sāṅkṛityāyana, ed., 482.30–482.31: viparyayas tu maulo hetuḥ | prasaṅgas tu maulasya hetor vyāptisādhanārtha eva | and 483.25: rūpādaya eva sāmānyam iti pakṣe nedaṃ prasaṅgasādhanam | pratyakṣeṇa teṣām avyāpitāpratīteḥ | tasmād ubhayavādisiddha eva hetur na tv anyatarasya prativādina eva | Manorathanandin makes a similar point, likewise accepting that a contraposed consequence formed on a real subject utilizes a maula hetu, one having a valid pakṣadharmatā; see Tillemans (2000, p. 22).

  51. Iwata (1999, p. 163) highlights this alternative that Prajñākaragupta offers in comment to Pramāṇavārttika IV.141abc (Pramāṇavārttikabhāṣya, Sāṅkṛityāyana, ed., 505.12–505.13): vikalpapariniṣṭhite dharmiṇi sādhyasādhanād vastubhūtas tu dharmī na tasya |

  52. Iwata (1993: 68) quotes Yamāri’s comment on Prajñākaragupta’s usage of maula: rtsa ba zhes bya ba ni gzhan gyis khas blangs pa la mi ltos pa’o || Iwata (1997, pp. 432–433) introduces the term “svatantra inference” into Prajñākaragupta’s discussion, writing, “in the case of the existent subject only the reason of prasaṅgaviparyaya is to be fundamental and this prasaṅgaviparyaya can be an autonomous [svatantra] inference.” Prajñākaragupta, however, does not use the term svatantra.

  53. Tattvasaṃgraha 392–393ab, 180.5–181.1: tathā hi santo ye nāma te sarve kṣaṇabhaṅginaḥ | tadyathā saṃskṛtā bhāvās tathā siddhā anantaram || santaś cāmī tvayeṣyante vyomakāleśvarādayaḥ |

  54. Tattvasaṃgrahapañjikā, 180.18–180.22: prayogaḥ yat sat tat sarvaṃ kṣaṇikam yathā samanantaraṃ pratipāditāḥ kṣaṇikāḥ padārthāḥ | santaś ca bhavatā vyomādayo bhāvā iṣyante iti svabhāvahetuḥ | … tvayeṣyanta ity anena prasaṅgasādhanam etad iti darśayati | anyathā hetor anyatarāsiddhatā syāt |

  55. Tattvasaṃgraha 393cd, 181.2: kṣaṇikatvaviyoge tu na sattaiṣāṃ prasajyate || and Tattvasaṃgrahapañjikā, 180.23: kathaṃ punar asya hetor vyāptiḥ siddhā ity āha kṣaṇikatvaviyoge tv ity ādi |

  56. Tattvasaṃgraha 394, 181.3–181.4: krameṇa yugapac cāpi yasmād arthkriyākṛtaḥ | na bhavanti sthirā bhāvā niḥsatvās te tato matāḥ || and Tattvasaṃgrahapañjikā, 180.24–181.6: arthakriyāsāmarthyalakṣaṇam iha sattvaṃ hetutveneṣṭam tac ca kṣaṇikatvanivṛttau nivarttate | tathā hi bhāvā arthakriyāṃ kurvantaḥ krameṇa vā kurvīran yaugapadyena vā | na hi kramayaugapadyābhyām anyaḥ prakāraḥ saṃbhavati tayor anyo ʼnyavyavacchedarūpatvāt | etac ca pratyakṣata eva prasiddham | Kamalaśīla further describes successive and simultaneous production as “mutually exclusive contradictories”: 181.13–181.15: pratyakṣapramāṇāvasita evānayor anyonyavṛttiparihārasthitalakṣaṇo virodhaḥ | tena tṛtīyarāśyabhāvāt kramayaugapadyābhyām arthakriyā vyāptā | I thank one of the anonymous reviewers for highlighting the similarity between Kamalaśīla’s wording here and Dharmakīrti’s in Nyāyabindu III.75 and ad Pramāṇaviniścaya III.52 (Hugon and Tomabechi, eds., 65.6): anyonyaparihārasthitalakṣaṇatayā.

  57. Tattvasaṃgrahapañjikā, 181.15–181.16: sā ca sthireṣu bhāveṣu svavyāpakanivṛttau nivarttamānā tallakṣaṇaṃ sattvaṃ nivarttayatīti siddhā vyāptiḥ | For Dharmakīrti’s discussion of how the entailment in this argument is established, a discussion that Kamalaśīla clearly relies on, see Steinkellner (1991, particularly 319–320).

  58. My account of the procedure whereby sādhyaviparyaye bādhakapramāṇa establishes the entailments of nature reasons relies on Steinkellner 1991, which examines Dharmakīrti’s explications in the Vādanyāya and Hetubindu.

  59. See particularly Steinkellner (1991, pp. 319–320).

  60. Steinkellner (1991, p. 320) translates Vādanyāya 8.5: tatra sāmarthyaṃ kramākramayogena vyāptaṃ siddham, prakārāntarābhāvāt as “In this case the capability (for causal efficiency) is proven as pervaded by the possibility (to produce an effort) successively or simultaneously, because there is no other way (of producing).” As we saw just above, Kamalaśīla appeals simply to perception.

  61. Tattvasaṃgrahapañjikā, 182.9–182.10: na hy asmābhiḥ svātantryeṇa pramāṇatayā vyatirekasādhinyā asyā vyāpakānupalabdheḥ prayogaḥ kriyate | kiṃ tarhi prasaṅgāpādanaṃ paraṃ prati kriyate |

  62. Steinkellner (1991, p. 318), drawing on Vādanyāya 9.10: viruddhapratyupasthāpanāt.

  63. Steinkellner (1991: 318).

  64. Steinkellner’s discussion (1991, pp. 319–320) of Dharmakīrti’s example of the process of establishing a nature reason in the Vādanyāya points out that “the valid cognition that negates the logical reason in the contradictory opposite of the property to be proved” in the sattvānumāna (such as “sound is impermanent because of existing/being a product”) shows that “Where (causal efficiency) is not possible successively or simultaneously, that is incapable for every (effect); and this (impossibility) is extant in a non-momentary (thing)”; yatra kramayaugapadyāyogaḥ, na tasya kvacit sāmarthyam, asti cākṣaṇike sa iti.

  65. For an incisive discussion on Indian and Tibetan views on whether entailments must have “existential commitment,” that is, be instantiated by real things, see Tillemans (1999, pp. 12–17).

  66. Madhyamakālaṃkāra 1 (Ichigō, ed.), 22: bdag dang gzhan smra’i dngos ’di dag || yang dag tu na gcig pa dang || du ma’i rang bzhin bral ba’i phyir || rang bzhin med de gzugs brnyan bzhin || Ichigō cites the Sanskrit of this stanza as quoted in Prajñākaramati’s Bodhicaryāvatārapañjikā (173.17–173.18): niḥsvabhāvā amī bhāvās tattvataḥ svaparoditāḥ | ekānekasvabhāvena viyogāt pratibimbavat ||

  67. Kamalaśīla, Madhyamakālaṃkārapañjikā (Ichigō, ed.), 27.1–27.7: 'o na 'di ci thal bar 'gyur bar sgrub pa yin nam | 'on te rang gi rgyud kyis sgrub pa yin | gal te dang po yin par rtog na ni de’i tshe gtan tshigs ma grub pa yin te | gzhan dag dngos po thams cad la gcig dang du ma dang bral ba khas mi len pa’i phyir ro || ji ste gnyis pa yin no zhe na ni | de lta na yang shin tu ma grub pa yin te | bdag nyid gzhi khas mi len pa’i phyir dang | phyir rgol ba la yang rang gi ngo bo ma grub pa’i phyir ro snyam du dogs pa la gtan tshigs ʼdi ma grub po snyam du ma sems shig ces bya ba smos so || The underlined represents Śāntarakṣita’s words introducing his second stanza.

  68. Kamalaśīla, Madhyamakālaṃkārapañjikā (Ichigō, ed.), 27.9–27.11: re zhig gzhan dag gis brtags pa nam mkha' la sogs pa'i dngos po ma grags pa'i rang gi ngo bo gang dag yin pa de la ni thal bar bsgrub pa kho na yin la | grags pa’i rang gi ngo bo rnams la ni gnyi ga ltar yang nyes pa med de | Kamalaśīla provides this account immediately after introducing Śāntarakṣita’s second stanza, which he portrays as responding to the question of whether the first stanza is a prasaṅga or svatantra proof. In the second stanza, Śāntarakṣita rejects the notion that a permanent cause could produce successive effects and, in his autocommentary to this stanza, takes the Sāṃkhya doctrine of prakṛti as his example of a “permanent cause.”

  69. Madhyamakāloka (Keira, ed.), §2, p. 253: de la ʼdir thal bar sgrub na gtan tshigs ma grub pa yang ma yin te | ʼdi ltar ji ste pha rol po dag gis dngos po rnams gcig dang du ma dang bral bar khas ma blangs su zin kyang | ʼon kyang des khyab pa’i chos khas blangs pa’i phyir shugs kyis ni de yang khas blangs pa kho na yin te |

  70. Madhyamakāloka (Keira, ed.), §3, p. 253.

  71. Madhyamakālaṃkāra stanzas 2–15, Ichigō, ed., 22–66; Madhyamakāloka, Keira ed., §3–13, pp. 253–255.

  72. This reflects Śāntarakṣita’s program of “relying on” a Yogācāra perspective to show the naturelessness of external objects, before turning to a Madhyamaka perspective to show the naturelessness of consciousness itself, as Madhyamakālaṃkāra 92 states (Ichigō, ed., 294): sems tsam la ni brten nas su || phyi rol dngos med shes par bya || tshul ʼdir brten nas de la yang || shin tu bdag med shes par bya ||

  73. Madhyamakālaṃkāra stanzas 19–60, Ichigō, ed., 74–166; Madhyamakāloka, Keira ed., §15–23, pp. 255–257.

  74. Madhyamakālaṃkāra stanza 76 (Ichigō, ed.), 252: gzhung gis bskyed pa’i bye brag gi || chos can spangs nas mkhas pa dang || bud med byis pa’i bar dag la || grags par gyur pa’i dngos rnams la ||

  75. Madhyamakālaṃkāravṛtti (Ichigō, ed.), 256.1–256.5: rjes su dpag pa dang rjes su dpag par bya ba’i tha snyad thams cad ni … mkhas pa dang bud med dang byis pa’i bar kyi mig dang rna ba la sogs pa’i shes pa la snang ba’i ngang can gyi phyogs sgra la sogs pa’i chos can la brten nas ʼjug go | Śāntarakṣita goes on to say that were it otherwise, if subjects particular to one party’s tenet system were used, reasons—like “smoke” and “exists”—could not prove predicates, like “fire” and “impermanent” due to having unestablished bases. And so we must understand “place” and “sound” as the subjects of these particular inferences. In a cognate passage that Ichigō (p.255) provides from Haribhadra’s Abhisamayālaṃkārāloka, we find the expression pratibhāsamānaṃ dharmiṇam, which may be the original of snang ba’i ngang can … chos can. The Tibetan translation (sde dge edition, Toh. 3791, 238a.3–238a.7) of the Abhisamayālaṃkārāloka passage reads chos can snang bzhin pa and simply chos can snang ba. The Abhisamayālaṃkārāloka was translated by Rin chen bzang po and Subhāṣita and later revised by Rin chen bzang po and Atiśa, while the Madhyamakālaṃkāra texts were translated by Ye shes sde and Śīlendrabodhi; this could explain the differences in translation if indeed the Sanskrit was the same. Blumenthal (2004: 156) reads snang ba’i ngang can in Madhyamakālaṃkāra stanza 78a as “possess the taste of appearances” and suggests rasa as the Sanskrit original of ngang. Tillemans translates this phrase “as they appear” (2003, p. 119).

  76. Commenting on Madhyamakālaṃkāra stanza 76, Kamalaśīla identifies Śāntarakṣita’s “treatises” as tenet systems (Madhyamakālaṃkārapañjikā, [Ichigō, ed.], 253.12: gzhung ni grub pa’i mtha’o ||), in contrast with subjects that are renowned to all, including, as he explains, subjects that appear to the six consciousnesses (253.15–253.16: grags par gyur pa rnam shes drug par bsdu’o ||). In comment to the following stanza, Kamalaśīla notes that the latter can be engaged by all disputants (Madhyamakālaṃkāra 77b [Ichigō, ed., 252]: ma lus yang dag ʼjug par ʼgyur || Kamalaśīla explains [253.17] ma lus pa ni rgol ba thams cad kyi’o ||).

  77. Madhyamakāloka (Keira, ed.), §24, p. 257: de nyid kyi phyir rang gi rgyud kyis bsgrub pa la yang gtan tshigs ma grub pa nyid ma yin te | ji skad bshad pa’i tshul gyis rgol ba dang phyir rgol ba gnyi ga la yang chos thams cad gcig dang du ma dang bral ba tsam du grub pa’i phyir ro ||

  78. Madhyamakālaṃkāravṛtti (Ichigō, ed.), 252.4–252.12. The second prong of the opponent’s challenge first suggests that the Mādhyamika could choose not to state a reason that would establish that all phenomena lack nature, in which case, the claim of naturelessness would still not be established. The presence of a reason (gtan tshigs yod), the opponent concludes, would undermine the claim of naturelessness. I read “the presence of a reason” to mean “the reason being a property of the subject,” which the opponent interprets as implying the real nature of the subject.

  79. Madhyamakālaṃkāra 78 (Ichigō, ed.), 254: bdag ni snang ba’i ngang can gyi || dngos po dgag par mi byed de || de lta bas na sgrub pa dang || bsgrub bya gzhag pa ’khrugs pa med ||

  80. Madhyamakāloka (Keira, ed.), §4, p. 233: ʼjig rten na ji ltar grags pa’i chos can kho na la gnas te | rab tu smra ba rnams de kho na nyid gtan la dbab pa’i phyir rjes su dpag pa la sogs pa’i tha snyad ʼjug par byed de | Kamalaśīla cites Pramāṇavārttika I.85–86, which makes this same point; see Keira 125–126 and n.202.

  81. Madhyamakāloka (Keira, ed.), §8, p. 234: ʼon kyang sgro btags pa’i chos rnam par gcad pa sgrub pa tsam zhig brjod par ʼdod pa de la ni ma grub pa nyid la sogs pa’i nyes pa brjod pa tha snyad du yang dngos por gyur pa’i chos can mi dgos te | Keira (2004: 128, n.207) notes that ma grub pa nyid la sogs pa’i nyes pa brjod pa must be taken as a bahuvrīhi compound, referring to the reason as “that for which the faults such as being unestablished would be pointed out.”

  82. Madhyamakāloka, (Keira ed.), §25, p. 257: rnam par bcad pa tsam zhig yin pa la ni bsgrub par bya ba dang | sgrub pa dang rjes su mthun pa’i chos can dngos por gyur pa ma yin pa rigs pa nyid do zhes sngar bstan zin to || des na rang dang gzhan gyi gzhung lugs la gnas pa rnams kyis nye bar brtags pa’i chos can la yang bsgrub pa’i chos rnam par bcad pa tsam gyis ji skad bshad pa gnyi ga la grub pa nyid do || “[the reason] just as explained” here refers to the fact that Kamalaśīla’s discussion at this point references an explanation given much earlier in the text (see Keira’s edition, §6–12, pp. 234–237), in which he more elaborately defends his inferential procedure from the charge of having an unestablished basis (āśrayāsiddha). My reading of the progression of Kamalaśīla’s argument in the present section differs slightly from Keira’s, judging by the outline that he provides to his translation. He reads Kamalaśīla’s case for how the “neither one nor many” reason is not unestablished to proceed in four parts: a defense of the reason as a prasaṅga proof, a defense of the reason as a svatantra proof, a statement that the locus of the reason is not unestablished, and a discussion of how the reason is established to be a property of the locus (2004, pp. 179–202). I treat Keira’s third and fourth sections as elaborations of Kamalaśīla’s defense of svatantra proofs—and so as subsections of that movement. The key passage is the beginning of Kamalaśīla’s defense of the svatantra proof, in which he states, “Thus, it is not the case that the reason is not established as a svatantra proof either, because all phenomena are established as just devoid of one and many for both proponent and opponent, proceeding just as explained” (Madhyamakāloka, [Keira ed.], §24, p. 257: de nyid kyi phyir rang gi rgyud kyis bsgrub pa la yang gtan tshigs ma grub pa nyid ma yin te | ji skad bshad pa’i tshul gyis rgol ba dang phyir rgol ba gnyi ga la yang chos thams cad gcig dang du ma dang bral ba tsam du grub pa’i phyir ro ||). I take the passages that follow, on how the locus is established, to be an elaboration of “proceeding just as explained.” That is, Kamalaśīla explains how the reason is established for both parties, such that there is no fault of a reason having an unestablished basis—which would only be a fault for a svatantra proof, not a prasaṅga proof.

  83. Iwata (1999, p. 168) notes the similarities between Kamalaśīla’s, Dharmottara’s, Devendrabuddhi’s, and Śākyabuddhi’s perspectives on this point.

  84. For Kamalaśīla’s portrayal of the two versions of Yogācāra, “proponents of [cognition] having aspects” (sākāravāda), often called “proponents of true aspects” (satyākāravāda), and “proponents of [cognition] without aspects” (nirākāravāda), often called “proponents of false aspects” (alīkākāravāda), in his Tattvasaṃgrahapañjikā, see Funayama (2007).

  85. Madhyamakāloka, (Keira ed.), §20, p. 238: ji ste thog ma med pa’i rang gi sa bon yongs su smin pa las yang dag par byung ba’i rnam par rtog pas yongs su bsgrubs pa | byis pa rnams kyis phyi rol dang dngos po nyid du nye bar brtags pa blo la yod pa nyid chos can yin te | Tillemans and Lopez (1999, pp. 271–272, n.13) point to the origins of this idea in Dignāga’s Nyāyamukha.

  86. Madhyamakāloka, (Keira ed.), §20, p. 239: de la phyi rol dang gtso bo la sogs pa’i ngo bo nyid dgag pa tshad mas sgrub par byed kyi | de nyid dgag pa’i phyir gtan tshigs sbyor ba ma yin no ||

  87. Madhyamakāloka, (Keira ed.), §20, p. 239: rjes su dpag par bya ba dang | rjes su dpag pa la sogs pa’i tha snyad ʼdi thams cad ni blo la yod pa’i chos can kho na la brten nas ʼjug pa nyid de | rnam pa gzhan mi srid pa’i phyir ro zhes bya bar ʼdod na | These passages (Keira ed., §20–21, pp. 238–239) are also translated in Tillemans and Lopez (1999, pp. 277–280, n. 24).

  88. Madhyamakāloka, (Keira ed.), §21–22, pp. 239–240: ji ltar khyed don dam pa la ʼjug par bya ba’i phyir gtso bo la sogs pa dgag par kun brtags pa’i chos can kho na la bsgrub par bya ba dang | sgrub pa’i sems pa rgyas par byed pa de bzhin du kho bo cag kyang gzugs dang sgra la sogs pa grags pa dag la | yod pa dang med pa la sogs pa’i ngo bo nyid du sgro btags pa dgag par byis pa rnams la de dag sgyu ma dang | smig rgyu dang | rmi lam dang | gzugs brnyan dang mtshungs pa nyid du ston par byed do || … kho bo cag yang dag pa nyid du kun brtags pa’i dngos po thams cad la de rab tu bsgrub pa’i phyir yang dag pa la nus pa med pa la sogs pa smos pa.

  89. Iwata (1999, p. 168) suggests that Kamalaśīla’s allowance of imputed inferential subjects appearing to conceptuality echoes Śākyabuddhi’s notion that non-existent subjects exist as “verbal conventions” (vyavahāra). One can see Iwata’s point, given that for Śākyabuddhi (following Dharmakīrti), “conventional existence” and “conceptual existence” amount to the same thing—ultimate existents are momentary entities that form the cause-and-effect world. Kamalaśīla’s distinction is subtler, as he denies even the conventional existence of these imaginary objects but accepts their appearance to conceptuality (vikalpa).

  90. Madhyamakāloka, (Keira ed.), §28–32, pp. 257–258. I translate dbang pos mi sod pa as “not accessible to the sense faculties” and lkog tu ma gyur pa (aparokṣa), literally “not hidden,” as “perceptible.” One wonders if Kamalaśīla’s distinction between perceptible and non-perceptible subjects in part led later generations to speak sometimes of the “common establishment” of the inferential subject (chos can mthun grub), which extends back to Dignāga, and other times (becoming the more frequent locution) of the subject’s “common appearance” (chos can mthun snang).

  91. Madhyamakāloka, (Keira ed.), §30, p. 258: dngos po de lta bu’i dgag pa la thams cad du chos can brtags pa kho na yin pa’i phyir te | de kho na’i chos mi dmigs pa yin no zhes sngar bstan zin to ||

  92. Madhyamakāloka, (Keira ed.), §32, p. 258: chos can gzugs la sogs pa lkog tu ma gyur pa gang dag yin pa de dag kyang ngo bo nyid gcig pa dang bral ba mngon sum kho nar gyur pa yin te | ʼdi dag rtag tu rnam pa sna tshogs su snang ba nyid kyis bye brag med par ngo bo nyid gcig par so sor myong ba med pa’i phyir ro || ngo bo nyid gcig dang bral ba grub na ngo bo nyid du ma dang bral ba yang mngon sum kho nar gurb pa yin te | du ma nyid ni gcig tshogs pa’i ngo bo yin pa’i phyir ro || In holding that the negation of “many” follows directly from the negation of “one,” Kamalaśīla follows Śāntarakṣita’s Madhyamakālaṃkāra stanza 61cd and autocommentary (Ichigō 1985: 172): gang la gcig nyid yod min pa || de la du ma nyid kyang med | … gang gcig pa’i rang bzhin du mi ’thad pa de du ma’i bdag nyid du khas blangs pa ni rigs pa ma yin pa nyid de | ’di ltar du ma ni gcig bsags pa’i mtshan nyid do | gcig med na de yang med de | shing la sogs pa med na nags tshal la sogs pa med pa bzhin no || (“Anything that does not have a singular [nature] also lacks a multiple [nature].” … “‘Many’ is defined as a collection of ‘ones.’ If there are no ‘ones,’ that, too, does not exist, just as if there are no trees and so forth, there is no forest and so forth.”)

  93. The Buddhist Idealist’s point concerning inferential subjects would still hold: inference is a conceptual process that must utilize the concept of say, sound, rather than a particular sound. However, even in the common Buddhist inference, “sound is impermanent because of being a product,” debaters would recognize that the concept “sound” corresponds to actual sounds, as the concept “sound” is not (in the Buddhist conception) a product nor impermanent.

  94. Madhyamakāloka, (Keira ed.), §13, p. 237: thal bar sgrub pa tshad ma ma yin pa las ji ltar ʼdi dag gcig tu med pa’i tha snyad ʼgrub par ʼgyur te | de ni tshad ma la rag las pa’i phyir ro ||

  95. For an overview of Śāntarakṣita’s and Kamalaśīla’s reliance on Dharmakīrti’s pramāṇa theory and conceptions of perception and inference, see McClintock (2010, pp. 73–85).

  96. On Dharmakīrti and his commentators’ views of anupalabdhi, see Kellner (1997a, 2001).

  97. Madhyamakāloka, (Keira ed.), §36, pp. 258–259. See Kellner (2004) for the prospects in Dharmakīrti’s view of the perceptual process as, under particular circumstances, having the ability to cause determination, a perspective that Dharmottara fully embraced.

  98. Madhyamakāloka, (Keira ed.), §39, p. 259: dgnos po rnams rta’i rva lta bur khas blangs pa ni ma yin no || ʼo na ci zhe na | sgyu ma la sogs pa’i ngo bo nyid lta bur yin te | de dag kyang sgyu ma la sogs pa bzhin du mngon sum kho na’o ||

  99. A concept itself—a permanent thing in the Buddhist conception—could not be a cause, but a moment of conceptual cognition could cause a moment of determination. The causal chain of inference will form one target for Patsap’s critique (in both his sGron ma gsal bar byed pa and his Tshig gsal ba’i dka’ ba bshad pa) of svatantra inference. I treat his analysis of this problem in a forthcoming monograph on Madhyamaka reasoning.

  100. Kamalaśīla defends the use of absolute negations as inferential reasons and predicates in Madhyamakāloka (Keira ed.), §44–60, pp. 261–265. Keira’s translation is on pp. 207–216.

  101. Keira (2004, pp. 47–86).

  102. See Kellner (2001, pp. 499–501) for this discussion in Dharmakīrti’s Hetubindu (ed. Steinkellner 1967, 21.20–22.2).

  103. Keira (2004, pp. 68–85).

  104. Keira (2004, pp. 64–68), corresponding to his edition of Madhyamakāloka §11–12, 228–229, beginning gang yang don gzhan gyis dben pa’i phyogs la sogs pa ʼga’ zhig mngon sum gyis bzung nas de las gzhan pas stong pa mngon sum gyis rtogs par bya bar ʼgyur zhes bya ba la sogs pa smras pa de yang rigs pa ma yin te |

  105. Madhyamakāloka, (Keira ed.), §1, p. 228: dngos po rnams kyi gcig gis gcig stong pa nyid gang yin pa de ni don dam pa ma yin gyi | ʼon kyang tha snyad kyi bden ba la brten pa yin no ||

  106. See Vose (2011, p. 116), which translates and explicates Tauscher (1999, p. 94) and Hugon (2015).

  107. See MacDonald, In Clear Words, vol. 1, 152.6–157.1 and vol. 2, 75–81.

  108. The following search for plausible sources of Patsap’s notion of svatantra valid cognition draws on the Tattvasaṃgraha’s rejection of Vedic authority, intrinsic validity, and absence (abhāva) as a valid cognition, all components of Mīmāṃsā doctrine that are treated over three separate chapters of the work. This investigation is not intended as an exhaustive account of these three distinct chapters, two of which are the longest in the overall work, but rather an exploration of how Śāntarakṣita and Kamalaśīla use the expression svatantra in contexts where cognition, its validity, and its sources are at stake.

  109. Tattvasaṃgraha 5ab, 3.1: svatantraśrutiniḥsaṅgo jagaddhitavidhitsayā | and Tattvasaṃgrahapañjikā 18.20–18.22: svatantrā śrutiḥ svataḥ pramāṇabhūto vedaḥ nityaṃ vacanam iti yāvat | tasyā niḥsaṅgaḥ nirāsthaḥ tannirapekṣa eva sākṣād darśī pratītyasamutpādaṃ gaditavān ity arthaḥ |

  110. Tattvasaṃgrahapañjikā, 18.25: ayaṃ ca śrutiparīkṣāyāḥ svatantraprāmāṇyaparīkṣāyāś copakṣepaḥ |

  111. See Verpoorten (1994) for a detailed outline of this 24th chapter, including identifications of Śāntarakṣita’s many references to Kumārila’s Ślokavārttika. McClintock (2010, pp. 102–105) characterizes this chapter as beginning the second “half” of the Tattvasaṃgraha, marking a transition from the progressively higher levels of analysis in the first 23 chapters to a specifically polemical assault on Mīmāṃsā in chapters 24 and 25, culminating with a defense of the Buddha’s omniscience in chapter 26.

  112. Tattvasaṃgraha 2424ab, 805.5: śruteḥ svatantrataiṣā hi puṃvyāpārānapekṣaṇāt | referencing the Mīmāṃsā claim at Tattvasaṃgraha 2095ab, 717.1: aprāmāṇyanivṛttyarthā vedasyāpauruṣeyatā |

  113. Tattvasaṃgraha 2411–2413 argues that an absence of pramāṇa establishing the Veda’s author does not prove that is has no author, just like any book whose author is no longer recalled; 2424–2428 argue that even a lack of proof for the authored status of the Veda would not prove that it is authorless, only that the matter is inconclusive. Kamalaśīla’s comments on Tattvasaṃgraha 5 (18.23–18.24) quote a sūtra source suggesting an author of the Veda and its mantras, namely, the great sages: yathoktaṃ bhagavatā ity ete ānanda paurāṇā maharṣayo vedānāṃ karttāro mantrāṇāṃ pravarttayitāraḥ iti |

  114. Tattvasaṃgraha 2375, 792.1–792.2: svatantrasya ca vijñānajanakatve sati sphuṭam | prāmāṇyam api naivāsya sambhāvyaṃ puruṣekṣaṇāt || Kamalaśīla glosses this with “It is not reasonable for that which has the power to produce knowledge to depend on people” (792.11–792.12: śaktasya hi na puruṣāpekṣayā jñānajanakatvaṃ yuktam iti pratipāditam |).

  115. Tattvasaṃgraha 2358–2361, 788.3–789.2: atha satyārthavijñānajanmaśaktaḥ svataḥ sthitaḥ | vedo nāro nirāśaṃsaḥ satyārtho ʼyam ato mataḥ || yady evaṃ sarvadā jñānaṃ nairantaryeṇa tadbhavet | sadāvasthitahetutvād tadyathābhimate kṣaṇe || ekavijñānakāle vā tajjanyaṃ sakalaṃ bhavet | śaktaṃ hetutayā yadvad tadvijñānaṃ vivakṣitam || tataḥ paramato jñānajanmaśaktiparikṣayāt | na nityaḥ syād ayaṃ vedaḥ śaktau vā dhīḥ punar bhavet ||

  116. Tattvasaṃgraha 2365–2367, 790.1–790.6: na hi tāvat sthito ʼpy eṣa jñānaṃ vedaḥ karoti naḥ | yāvan na puruṣair eṣa dīpabhūtaiḥ prakāśitaḥ || tataś cāpauruṣeyatvaṃ bhūtārthajñānakāraṇam | na kalpyaṃ jñānam etad dhi puṃvyākhyānāt pravarttate || saty apy eṣā nirarthā ʼto vedasyāpauruṣeyatā | yadiṣṭaṃ phalam asyā hi jñānaṃ tat puruṣāśritam ||

  117. Tattvasaṃgrahapañjikā, 790.20–790.21: yathāvasthita evārthaḥ puruṣaiḥ prakāśyate nāpūrvaḥ kriyate | apūrvakaraṇe hi svātantryam eṣām abhyupagataṃ syāt |

  118. Tattvasaṃgraha 2369, 790.9–790.10: svatantrāḥ puruṣāś ceha vede vyākhyāṃ yathāruci | kurvāṇāḥ pratibaddhuṃ te śakyante naiva kenacit || This bears traces of Dharmakīrti’s denial of the Veda’s fixed meanings, in which he argues that there can be no necessary connection (niyama) between Vedic words and meanings, as the conventions of language function “arbitrarily” (svātantryam icchāyā). See Pramāṇavārttika I.317–329 (Gnoli ed., 166–173) and Pramāṇaviniścaya II.35–50 (Steinkellner ed., 70–72); this section is translated in Eltschinger et al. (2012, pp. 44–61). The final stanza reads: yatra svātantryam icchāyā niyamo nāma tatra kaḥ | dyotayet tena saṃketo neṣṭām evāsya yogyatām || Eltschinger, Krasser, and Taber translate (p. 61): “In this [convention], where there is arbitrariness, how can there be necessity? Thus, the convention cannot bring to light just a desired capacity [of the Vedic word and no other].”

  119. Tattvasaṃgrahapañjikā, 791.9–791.12: na vedārthasyātīndriyārthasya kaścit svātantryeṇa vijñātā naro ʼbhyupagato yo vedārtham ākhyāsyati | tathā hi vedārthaparijñānadvāreṇātīndriyārthadarśitvam asya na svātantryeṇa vedārthaparijñānaṃ na ca nātīndriyārthadarśitvam antareṇa iti vyaktam avatarati nitarām itaretarāśrayatvam iti darśayan nāha na cātīndriyadṛk ity ādi | With na cātīndriyadṛk, Kamalaśīla introduces Tattvasaṃgraha 2371.

  120. See Taber (1992, p. 207) for Kumārila’s Ślokavārttika II.47–48: svataḥ sarvapramāṇānāṃ prāmāṇyam iti gamyatām | na hi svato ʼsatī śaktiḥ kartum anyena śakyate || ātmalābhe ca bhāvānāṃ kāraṇāpekṣatā bhavet | labdhātmanāṃ svakāryeṣu pravṛttiḥ svayam eva tu || Taber translates, “The validity of all valid cognitions is to be understood as intrinsic, since a potency not existing intrinsically cannot be brought about by something else. And [in general] things depend on [other] causes in arising, but once they exist they exercise their functions by themselves.”

  121. Tattvasaṃgraha 2849, 917.3-917.4: janane hi svatantrāṇāṃ prāmāṇyārthaviniściteḥ | svahetunirapekṣāṇāṃ teṣāṃ vṛttir ghaṭādivat || Taber (1992: 210-211) shows that this characterization of intrinsic validity aligns with Pārthasārathimiśra’s interpretation of Ślokavārttika II.47-48, including Pārthasārathimiśra’s comparison of valid cognition to a pot. Given that Pārthasārathimiśra (10th century) postdates Śāntarakṣita (725-788), the former could well draw on an earlier interpretation of Kumārila.

  122. This is made clear in the following two stanzas, Tattvasaṃgraha 2850–2851.

  123. Taber (1992, p. 211).

  124. Taber (1992, p. 211).

  125. Tattvasaṃgraha 2921, 932.9–932.10: yat tu jñānaṃ tvayā ʼpīṣṭaṃ janmān antaram asthiram | labdhātmano ʼsataḥ paścād vyāpāras tasya kīdṛśaḥ || Kamalaśīla (932.24) glosses vyāpāra with sāmarthya.

  126. Tattvasaṃgrahapañjikā, 933.7–933.8: svārthaparicchedātmakam astīti ced na jñānaparyāyatvād asyātmānam eva karotīti.

  127. Tattvasaṃgraha 2811ab quotes Ślokavārttika II.47ab, reading svataḥ sarvapramāṇānāṃ prāmāṇyam iti gṛhyatām | with gṛhyatām replacing gamyatām. Tattvasaṃgraha 2812 has the opponent explain śakti as inherently existing and then quotes Ślokavārttika II.47cd: meyabodhādike śaktis teṣāṃ svābhāvikī sthitā | na hi svato ʼsatī śaktiḥ kartum anyena śakyate || with pāryate replacing śakyate. Tattvasaṃgraha 2813 has the opponent introduce the independent character of validity: anapekṣatvam evaikaṃ prāmāṇyasya nibandhanam | tad eva hi vināśyeta sāpekṣatve samāśrite || The first part of this chapter (stanzas 2810–2845) has been critically edited and translated into Japanese in Ishimura (2018).

  128. Tattvasaṃgrahapañjikā, 905.21–906.9. Śāntarakṣita treats “difference” in stanzas 2816–2823, “non–difference,” “both,” and “neither” in stanzas 2824–2825.

  129. Tattvasaṃgraha 2817, 905.7–905.8: iṣṭakāryasamarthaṃ hi svarūpaṃ śaktir ucyate | tasya bhāvātmanābhāve bhāvo na syāt sa kārakaḥ || Tattvasaṃgrahapañjikā, 906.11–906.12: kāryakaraṇasamarthā hi svabhāvaśaktiḥ tasya ca svabhāvasya bhāvātmatāyā abhāve sati sa bhāvaḥ kārako na syāt |

  130. Tattvasaṃgrahapañjikā, 906.10: sarvapadārthānām avyatiriktaiva śaktir iti bahuśaḥ pratipāditatvāt | and 906.14: athāvyatirikteti pakṣaḥ … Ishimura (2018, p. 111, n. 34) locates Kamalaśīla’s reference to the extensive establishment of this point as Tattvasaṃgrahapañjikā ad 2514, 827.15–827.20.

  131. Tattvasaṃgraha 2818–2819, 906.1–906.4: sā cānityedṛśī śaktiḥ svahetubalabhāvinī | svābhāvikī pramāṇānāṃ yuṣmābhiḥ katham iṣyate || svābhāvikyāṃ hi śaktau syān nityatāhetutātha vā | pramāṇānāṃ ca tādātmyān nityatāhetute dhruvam ||

  132. Tattvasaṃgraha 2820–2821 avers that valid cognitions having an inherent power would not depend on causes and so would either always exist or always not exist, and the effects that depend on those causeless valid cognitions would likewise be constant; since the nature and effects of valid cognitions are occassional, their power must not be inherent (906.8: śaktir yuktā svābhāvikī na vaḥ ||).

  133. Tattvasaṃgrahapañjikā, 907.15–907.18 (in the lead up to stanza 2824) notes that “the position of both” would be susceptible to the faults of a power being both different and not different from a valid cognition, while “the position of neither” is impossible owing to the fact that the negation of one would establish the other. At 906.9, Kamalaśīla noted that if the two were different, no connection (sambandha) would be established between them. And so it would be plausible that he considered “the position of different” to have already been dispatched. However, stanza 2824 addresses “separateness” (pṛthaktvam) and “having a nature of both” (ubhayātmatvam), to show that in either scenario cognitions would be permanent due to their “association” (saṅgata) with a permanent power. In comment, Kamalaśīla (907.20–907.22) notes that “the position of having a nature of neither” is included in this fault.

  134. Tattvasaṃgraha 2827, 908.5–908.6: tad atra na vivādo naḥ ko hy anaṃśasya vastunaḥ | svahetor upajātasya śaktiṃ paścāt prakalpayet ||

  135. Tattvasaṃgraha 2831, 909.3–909.4: na hi teṣām avasthānāṃ parastād asti yena te | pratyayāntarataḥ śaktiṃ labheran kutracit phale ||

  136. Tattvasaṃgraha 2831, 909.6–909.7: etāvat tu vadanty atra sudhiyaḥ saugatā ime | jñāne kvacit sthitāpy eṣā na boddhuṃ śakyate svataḥ || Tattvasaṃgrahapañjikā, 909.14–909.15: boddhum iti niścetum | svata iti vijñānasvarūpād anubhavamātrād* anapekṣitottarakālabhāvikāryasaṃvādāt || * Ishimura (2018: 125) prefers the manuscript reading of vijñānasvarūpānubhavamātrād

  137. Tattvasaṃgraha 2834, 909.9–909.10: apramāṇe ʼpi yenaitat keśapāśādidarśane | vidyate ʼnubhavātmatvaṃ vispaṣṭākārabhāsini ||

  138. Tattvasaṃgraha 2835, 910.1–910.2: tasmād arthakriyājñānam anyad vā samapekṣyate* | niścayāyaiva na tv asyā ādhānāya viṣādivat || * Ishimura (2018: 128) prefers the manuscript reading of samapekṣate. In comment to Tattvasaṃgraha 2944 (938.7–938.8: yenaikaiḥ svata eveti proccair niyama ucyate | kiñcit svato ʼnyataḥ kiñcit paraiś cāniyamo mataḥ ||), Kamalaśīla will explain that certain kinds of cognitions, including cognitions of causal efficacy, themselves come to have an intrinsic validity—acquired through their repeated reliability; Tattvasaṃgrahapañjikā, 938.19–938.21: parair iti bauddhaiḥ | taiḥ kiñcit svataḥ pramāṇam iṣṭam yathā svasaṃvedanapratyakṣaṃ yogijñānam arthakriyājñānam anumānam abhyāsavac ca pratyakṣam tad dhi svata eva niścīyate | This passage has been examined in Steinkellner (1992, p. 259); Krasser (2003:, p. 163), and Shida (2007, p. 31).

  139. Tattvasaṃgrahapañjikā, 910.12–910.13, in comment on Tattvasaṃgraha 2835b: anyad veti hetuśuddhijñānam | Śāntarakṣita addresses validity stemming from a cognition’s causes in Tattvasaṃgraha 3001, 954.5–954.6: tathā bodhātmakatvena buddheḥ prāptā ʼpramāṇatā | yathārthajñānahetūtthaguṇajñānād apodyate ||

  140. Tattvasaṃgraha 2926, 934.4–934.5: ataś ca śakyate vaktuṃ svataḥ eva na vartate | paścāt pramā svakāryeṣu nairūpyād gaganābjavat ||

  141. This section of the Tattvasaṃgraha, stanzas 1647–1690, has been edited, translated, and studied in Kellner (1997b), where, among other things, Śāntarakṣita’s quotations of Kumārila are analyzed. Taber 2001 is a review article of Kellner’s work that offers some supplements to it.

  142. Tattvasaṃgraha 1680, 583.9–584.1: jñānarūpaviviktaśca so ʼbhāvo gamyate katham | tadgocarapramābhāvād evaṃ na tv anavasthitiḥ ||

  143. Tattvasaṃgraha 1681, 584.2–584.3: vastvabhāvāt pramāṇasya pramābhāvāc ca vastunaḥ | nāstitā yadi gamyeta bhaved anyonyasaṃśrayaḥ ||

  144. Tattvasaṃgraha 1682, 584.4–584.5: tasmād ekasya yā dṛṣṭiḥ saivānyādṛṣṭir ucyate | sā ca svatantrasaṃsiddhā svarūpeṇājaḍatvataḥ || Taber (2001, p. 81) translates the latter half of the stanza as “And that [perception of something different] is known of itself, insofar as it is by nature conscious.”

  145. Dharmakīrti’s treatments of anupalabdhi span Pramāṇavārttika IV.265–274 and Pramāṇaviniścaya III.40–47 (Hugon and Tomabechi, eds., 56.4–60.8). His argument against absence as an independent pramāṇa follows his presentation of anupalabdhi in each text.

  146. Dharmakīrti notes that non-apprehension is established by self-awareness in Pramāṇaviniścaya III, 60.5 and 60.10: tasmād anupalambho ʼyaṃ pratyakṣeṇaiva sidhyati || and sa ca svasaṃvedanapratyakṣasiddhaḥ || and also in Pramāṇavārttika IV.274ab: tasmād anupalambho ʼyaṃ svayaṃ pratyakṣato gataḥ. For more on the role of reflexive cognition in grounding a valid cognition, see Kellner (2011).

  147. Tattvasaṃgrahapañjikā ad 1682cd, 584.21–585.7: sā cety ādi | sā caikasya dṛṣṭiḥ svayam eva svatantrā siddhā nānyata iti nānavasthāprasaṅgaḥ | kasmāt | svarūpeṇa prakṛtyaiva ajaḍarūpatvāt prakāśasvabhāvatvāt |

  148. Williams (1998) and Arnold (2005) discuss the advancements Śāntarakṣita’s conception of svasaṃvedana/svasaṃvitti offers over that of Dignāga. Kellner (2011) analyzes Dignāga’s and Dharmakīrti’s respective strategies for establishing self-awareness and avoiding infinite regress.

  149. Tattvasaṃgraha 1999, 682.3–682.4: vijñānaṃ jaḍarūpebhyo vyāvṛttam upajāyate | iyam evātmasaṃvittir asya yā ʼjaḍarūpatā || This chapter of the Tattvasaṃgraha has been edited, translated, and studied in Saccone (2018).

  150. Tattvasaṃgrahapañjikā ad 1999, Saccone ed., 174: na hi grāhyagrāhakabhāvenātmasaṃvedanam abhipretam, kiṃ tarhi svayaṃprakṛtyā prakāśātmatayā nabhastalavartyālokavat || Saccone emends Dwarikadas Shastri’s (682.20–682.21) grāhakabhāvenātmasaṃvedanam to grāhyagrāhakabhāvenātmasaṃvedanam on the basis of the Tibetan translation. I follow her edition here.

  151. Arnold (2005, pp. 95–96).

  152. Śāntarakṣita’s further characterization of self-awareness in Tattvasaṃgraha 2011ab (685.5) highlights that cognition “does not depend on another knower to cognize its own nature”; svarūpavedanāyānyad vedakaṃ na vyapekṣate. See also Saccone’s (2018, pp. 179 and 264) edition and translation.

  153. Arnold (2005) shows that Candrakīrti’s criticisms of Dignāga’s conception of self-awareness would be blunted by Śāntarakṣita, whose appeal to the nature of “cognition as cognition” Arnold (p. 97) suggests we understand as advocating “simply the intentional structure that constitutively characterizes any token of the type ‘cognition’.”

  154. McClintock (2010, pp. 216–219) discusses Śāntarakṣita’s and Kamalaśīla’s appeals to the nature of cognition in order to demonstrate that seeing things correctly—including seeing the impermanence of things—is natural to cognition, whereas seeing things incorrectly is adventitious. This is part of their broader argument in the Tattvasaṃgraha’s “Analysis of Supersensible Perception” (which Kamalaśīla also terms “Analysis of Omniscience”; see McClintock 2010: 104, n. 286) for omniscience—that all cognitive faults can be eradicated, allowing the true nature of consciousness to manifest. Taber (2015–2018, pp. 90–92) examines Kamalaśīla’s reliance (in comment to Tattvasaṃgraha 3337, which appeals to cognition’s nature of knowing correctly) on Dharmakīrti’s Pramāṇavārttika II: 206–208, suggesting that those Pramāṇavārttika stanzas echo a Mīmāṃsā conception of intrinsic validity. That is to say, Dharmakīrti’s and Kamalaśīla’s arguments concerning the perfectibility of mind, which turn on the very nature of mind, seem to espouse intrinsic validity. Taber (p. 94) points out, however, that Dharmakīrti’s and Kamalaśīla’s comments do not address valid cognition, but simply cognition itself, and so should not be taken as an examination of validity. It would be a stretch, then, to read Śāntarakṣita’s and Kamalaśīla’s analysis of intrinsic validity into their discussions of cognition’s luminous nature. As I note below, it appears that Patsap makes just such a stretch.

  155. In a larger study, I investigate Patsap’s use of a Dharmakīrti stanza (Pramāṇavārttika III.82 = Pramāṇaviniścaya II.3) to show the incompatibility of emptiness with the causal process that produces inferential cognition. I also explore Jayānanda’s Madhyamakāvatāraṭīkā’s comments to Madhyamakāvatāra VI.171–178 in which he accuses Svātantrikas of adopting a version of “production from other” in their use of inference to give rise to inferential realization of emptiness.

References

  • Apple, J. B. (2018). Khu lo tsā ba’s Treatise: Distinguishing the Svātantrika/*Prāsaṅgika difference in early twelfth century Tibet. Journal of Indian Philosophy, 46(5), 935–981.

    Article  Google Scholar 

  • Arnold, D. (2005). Is Svasaṃvitti transcendental? A tentative reconstruction following Śāntarakṣita. Asian Philosophy, 15(1), 77–111.

    Article  Google Scholar 

  • Blumenthal, J. (2004). The ornament of the middle way: A study of the Madhyamaka Thought of Śāntarakṣita. Ithaca, NY: Snow Lion Publications.

    Google Scholar 

  • Chapa Chökyi Senggé (Phya pa Chos kyi seng ge). dBu ma shar gsum gyi stong thun. In Tauscher 1999.

  • Collected Works of the Kadampas:

  • bKaʼ gdams gsung ʼbum phyogs bsgrigs, edited by dPal brtsegs bod yig dpe rnying zhib ʼjug khang. Chengdu: Si khron mi rigs dpe skrun khang, 2006–2015.

  • Dharmakīrti. Hetubindu. In Steinkellner 1967 and Steinkellner 2016.

  • Dharmakīrti. Pramāṇavārttika I. In Gnoli 1960.

  • Dharmakīrti. Pramāṇavārttika IV. In Tillemans 2000.

  • Dharmakīrti. Pramāṇaviniścaya I and II. In Steinkellner 2007.

  • Dharmakīrti. Pramāṇaviniścaya III. In Hugon and Tomabechi 2011.

  • Dharmottara. Pramāṇaviniścayaṭīkā (Tibetan translation). bstan ʼgyur, sde dge edition, Tohoku 4227 and 4229. tshad ma, vols. tshe and dze.

  • Dharmottara. Pramāṇaviniścayaṭīkā III (Sanskrit). Edited by Pascale Hugon, Takashi Iwata, and Toshikazu Watanabe. Vienna and Beijing: Austrian Academy of Sciences Press and China Tibetology Publishing House, forthcoming.

  • Dignāga. Pramāṇasamuccaya and Pramāṇasammucayavṛtti. bstan ʼgyur, sde dge edition, Tohoku 4203 and 4204. tshad ma, vol. ce.

  • Doctor, T. (2010). In pursuit of transparent means of knowledge: The Madhyamaka project of rMa bya Byang chub brtson ’grus. Journal of the International Association of Buddhist Studies, 32(1–2), 419–441.

    Google Scholar 

  • Doctor, T. (2014). Reason and experience in Tibetan Buddhism: Mabja Jangchub Tsöndrü and the traditions of the middle way. New York: Routledge.

    Book  Google Scholar 

  • Dreyfus, G. B. J., & Tsering, D. (2010). Pa tshab and the origin of Prāsaṅgika. Journal of the International Association of Buddhist Studies, 32(1–2), 387–417.

    Google Scholar 

  • Dwarikadas Shastri, S. (1968). Tattvasaṅgraha of Ācārya Shāntarakṣita with the Commentary ‘Pañjikā’ of Shri Kamalashīla (2 Vols.). Varanasi: Bauddha Bharati.

  • Eltschinger, V., Krasser, H., & Taber, J. (2012). Can the Veda Speak? Dharmakīrti against Mīmāṃsā Exegetics and Vedic authority. An annotated translation of PVSV 164,24–176,16. Vienna: Verlag der Österreichischen Akademie der Wissenschaften.

  • Funayama, T. (2007). Kamalaśīla’s distinction between the two sub-schools of Yogācāra. A provisional survey. In B. Kellner (Ed.), Pramāṇakīrtiḥ: Papers dedicated to Ernst Steinkellner on the occasion of his 70th birthday (pp. 187–202). Vienna: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien.

  • Gnoli, R. (1960). The Pramāṇavārttikam of Dharmakīrti: The first chapter with the autocommentary. Rome: Istituto Italiano per il Medio ed Estremo Oriente.

  • Haribhadra. Abhisamayālaṃkārālokā. In Wogihara 1932.

  • Hopkins, J. (1989). A Tibetan delineation of different views of emptiness in the Indian Middle Way School: Dzong-ka-ba’s two interpretations of the Locus Classicus in Chandrakīrti’s Clear Words showing Bhāvaviveka’s assertion of commonly appearing subjects and inherent existence. Tibet Journal, 14(1), 10–43.

    Google Scholar 

  • Hopkins, J. (1996). Meditation on emptiness. Boston: Wisdom Publications (originally, London: Wisdom Publications, 1983).

  • Hugon, P. (2015). Proving emptiness: The epistemological background for the ‘neither one nor many’ argument and the nature of its probandum in Phya pa Chos kyi seng ge’s works. Journal of Buddhist Philosophy, 1, 58–94.

    Article  Google Scholar 

  • Hugon, P., & Tomabechi, T. (2011). Dharmakīrti’s Pramāṇaviniścaya (Chap. 3). Vienna, Beijing: Austrian Academy of Sciences Press and China Tibetology Publishing House.

    Google Scholar 

  • Ichigō, M. (1985). Madhyamakālaṃkāra of Śāntarakṣita with his own commentary or Vṛtti and with the subcommentary or Pañjikā of Kamalaśīla. Kyoto: Buneido.

    Google Scholar 

  • Ishimura, S. (2018). A study of the Svataḥprāmāṇyaparīkṣā of the Tattvasaṃgraha (1): Arguments about pramāṇa’s innate capacity. The Annals of the Research Project Center for the Comparative Study of Logic, 15, 91–158.

    Google Scholar 

  • Iwata, T. (1993). Prasaṅga und Prasaṅgaviparyaya bei Dharmakīrti und Seinen Kommentatoren. Vienna: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien.

    Google Scholar 

  • Iwata, T. (1997). On prasaṅgaviparyaya in Dharmakīrti’s tradition—Prajñākaragupta and gTsaṅ nag pa. In H. Krasser, M. T. Much, E. Steinkellner, & H. Tauscher (Eds.), Tibetan studies: Proceedings of the 7th seminar of the International Association for Tibetan studies (pp. 427–437). Vienna: Verlag der Österreichischen Akademie der Wissenschaften.

  • Iwata, T. (1999). On the interpretations of the subject (dharmin) of the inference negating invariable entities in Dharmakīrtian logic. In S. Katsura (Ed.), Dharmakīrti’s thought and its impact on Indian and Tibetan philosophy: Proceedings of the third international Dharmakīrti conference, Hiroshima, November 4–6, 1997 (pp. 155–172). Vienna: Verlag der Österreichischen Akademie der Wissenschaften.

  • Jayānanda. Madhyamakāvatāraṭīkā (dBu ma la ’jug pa’i ’grel bshad). bstan ʼgyur, sde dge edition, Tohoku 3870; dbu ma, vol. ya. Peking edition, 5271; vol. 99 (dbu ma’i ’grel, vol. ra).

  • Jha, G. (1937). The Tattvasaṅgraha of Śāntarakṣīta with the commentary of Kamalaśīla. Baroda: Oriental Institute.

    Google Scholar 

  • Kamalaśīla. Madhyamakālaṃkārapañjikā. In Ichigō 1985.

  • Kamalaśīla. Madhyamakāloka. In Keira 2004.

  • Kamalaśīla. Tattvasaṃgrahapañjikā. In Dwarikadas Shastri 1968.

  • Katsura, S. (2016). A report on the study of Sanskrit manuscript of the Pramāṇasamuccayaṭīkā Chapter 4: Recovering the example section of the Nyāyamukha. Journal of Indian and Buddhist Studies, 64(3), 195–203.

  • Keira, R. (2004). Mādhyamika and epistemology: A study of Kamalaśīla’s method for proving the voidness of all dharmas. Vienna: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien.

    Google Scholar 

  • Kellner, B. (1997a). Non-cognition (anupalabdhi)—Perception or inference? The views of Dharmottara and Jñānaśrīmitra. Tetsugaku, 49, 121–134.

    Google Scholar 

  • Kellner, B. (1997b). Nichts bleibt nichts: Die Buddhistische Zurückweisung von Kumārila’s Abhāvapramāṇa; Übersetzung und Interpretation von Śāntarakṣita’s Tattvasaṅgraha vv. 1647–1690 mit Kamalaśīla’s Tattvasaṅgrahapañjikā sowie Ansätzs und Arbeitshypothese zur Geschichte negativer Erkenntnis in der indischen Philosophie. Vienna: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien.

  • Kellner, B. (2001). Negation—Failure or success? Remarks on an allegedly characteristic trait of Dharmakīrti’s Anupalabdhi-theory. Journal of Indian Philosophy 29: 495–517.

  • Kellner, B. (2004). Why infer and not just look? Dharmakīrti on the psychology of inferential processes. In S. Katsura & E. Steinkellner (Eds.), The role of the example (dṛṣṭānta) in classical Indian Logic (pp. 1–51). Vienna: Arbeitskreis für Tibetische und Buddhistische Studien.

  • Kellner, B. (2011). Self-awareness (svasaṃvedana) and infinite regresses: A comparison of arguments by Dignāga and Dharmakīrti. Journal of Indian Philosophy, 39, 411–426.

    Article  Google Scholar 

  • Khutön Dodébar (Khu ston mDo sde ʼbar). sLob dpon gyi lugs kyis bden pa gnyis ʼchad pa. In Collected works of the Kadampa (Vol. 64, pp. 3–18).

  • Krasser, H. (2003). On the ascertainment of validity in the Buddhist epistemological tradition. Journal of Indian Philosophy, 31, 161–184.

    Article  Google Scholar 

  • Krishnamacharya, E. (1926). Tattvasaṅgraha of Śāntarakṣita with the commentary of Kamalaśīla. Baroda: Central Library.

    Google Scholar 

  • Lakatos, I. (1970). History of science and its rational reconstructions. Proceedings of the Biennial Meeting of the Philosophy of Science Association. Boston Studies in the Philosophy of Science, 8, 91–136.

    Google Scholar 

  • Mabja Jangchup Tsöndrü (rMa bya Byang chub brtson ʼgrus). dBu ma rigs pa’i tshogs kyi rgyan De kho na nyid snang ba. In Collected works of the Kadampa (Vol. 13, pp. 754–820).

  • MacDonald, A. (2000). The Prasannapadā: More manuscripts from Nepal. Wiener Zeitschrift für die Kunde Südasiens, 44, 165–181.

    Google Scholar 

  • MacDonald, A. (2003). Interpreting Prasannapadā 19.3-7 in context: A response to Claus Oetke. Wiener Zeitschrift für die Kunde Südasiens, 47, 143–195.

    Google Scholar 

  • MacDonald, A. (2015). In clear words: The Prasannapadā (Chap. 1, 2 Vols.). Vienna: Verlag der Österreichischen Akademie der Wissenschaften.

  • Marks, J. M. (2019). Playfighting: Encountering Aviddhakarṇa and Bhāvivikta in Śāntarakṣita’s Tattvasaṃgraha and Kamalaśīla’s Pañjikā. Ph.D. Dissertation, University of California, Berkeley.

  • McClintock, S. L. (2010). Omniscience and the rhetoric of reason: Śāntarakṣita and Kamalaśīla on rationality, argumentation, and religious authority. Boston: Wisdom Publications.

    Google Scholar 

  • Oetke, C. (2003). Prasannapadā 19,3-7 and its context. Wiener Zeitschrift für die Kunde Südasiens, 47, 111–142.

    Google Scholar 

  • Oetke, C. (2006). Logic matters in the Prasannapadā: A study on reasoning and proof in metaphysics. Stockholm: Stockholm University.

    Google Scholar 

  • Patsap Nyimadrak (Pa tshab Nyi ma grags). dBu ma rtsa ba’i shes rab kyi ti ka bstan bcos sgron ma gsal bar byed pa. In bKa’ gdams gsung ’bum, vol. 11, 29-132.

  • Patsap Nyimadrak. Tshig gsal ba’i dka’ ba bshad pa. In bKa’ gdams gsung ’bum (Vol. 11, pp. 137–203).

  • Prajñākaragupta. Pramāṇavārttikabhāṣya. In Sāṅkṛityāyana 1943.

  • Saccone, M. S. (2018). On the nature of things: A Buddhist debate on cognitions and their object. With a manuscript description by Paolo Giunta. Vienna: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien.

  • Sāṅkṛityāyana, R. (1943). Pramāṇavārtikabhāshyam or Vārtikālaṅkāraḥ of Prajñākaragupta: Being a commentary on Dharmakīrti’s Pramāṇavārtikam. Patna: Kashi Prasad Jayaswal Research Institute.

    Google Scholar 

  • Śāntarakṣita. Madhyamakālaṃkāra. In Ichigō 1985.

  • Śāntarakṣita. Tattvasaṃgraha. In Dwarikadas Shastri 1968.

  • Shida, T. (2007). Udayana’s critique of the intrinsic theory of validity: With respect to the origination of the validity. Journal of Indian and Buddhist Studies, 55(3), 28–33.

    Google Scholar 

  • Steinkellner, E. (1967). Dharmakīrti’s Hetubindu, Teil 1. Tibetischer Text und rekonstruierter Sanskrit-Text. Vienna: Hermann Böhlaus Nachf.

  • Steinkellner, E. (1991). The Logic of the Svabhāvahetu in Dharmakīrti’s Vādanyāya. In E. Steinkellner (Ed.), Studies in the Buddhist epistemological tradition: Proceedings of the second Dharmakīrti conference, Vienna, June 11–16, 1989 (pp. 311–324). Vienna: Verlag der Österreichischen Akademie der Wissenschaften.

  • Steinkellner, E. (1992). Early Tibetan ideas on the ascertainment of validity (nges byed kyi tshad ma). In S. Ihara & Z. Yamaguchi (Eds.), Tibetan studies: Proceedings of the 5th seminar of the international association for Tibetan studies. Monograph series of Naritasan Institute for Buddhist Studies: Occasional Papers 2 (pp. 257–273). Narita: Narita-san Shinshō-ji.

  • Steinkellner, E. (2007). Dharmakīrti’s Pramāṇaviniścaya (Chaps. 1 and 2). Vienna, Beijing: Austrian Academy of Sciences Press and China Tibetology Publishing House.

  • Steinkellner, E. (2016). Dharmakīrti’s Hetubindu: Critically edited by Ernst Steinkellner on the basis of preparatory work by Helmut Krasser with a translation of the Gilgit Fragment by Klaus Wille. Beijing, Vienna: China Tibetology Publishing House and Austrian Academy of Sciences Press.

    Google Scholar 

  • Taber, J. (1992). What did Kumārila Bhaṭṭa mean by Svataḥ Prāmāṇya? Journal of the American Oriental Society, 112(2), 204–221.

    Article  Google Scholar 

  • Taber, J. (2001). Much Ado about nothing: Kumārila, Śāntarakṣita, and Dharmakīrti on the cognition of non-being. Journal of the American Oriental Society, 121(1), 72–88.

    Article  Google Scholar 

  • Taber, J. (2015–2018). Dharmakīrti, svataḥ prāmāṇyam, and awakening. Wiener Zeitschrift für die Kunde Südasiens, 56–57, 77–98.

  • Tauscher, H. (1999). Phya pa chos kyi seṅ ge: dbu ma śar gsum gyi stoṅ thun. Vienna: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien.

    Google Scholar 

  • Tillemans, T. J. F. (1983). The ‘neither One nor many’ Argument for Śūnyatā and its Tibetan interpretations. In E. Steinkellner & H. Tauscher (Eds.), Contributions on Tibetan and Buddhist religion and philosophy (pp. 305–320). Vienna: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien.

    Google Scholar 

  • Tillemans, T. J. F. (1990). Materials for the study of Āryadeva, Dharmapāla and Candrakīrti. The Catuḥśataka of Āryadeva, chapters XII and XIII, with the commentaries of Dharmapāla and Candrakīrti: Introduction, translation, Sanskrit, Tibetan, and Chinese Texts, Notes (2 Vols.). Vienna: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien.

  • Tillemans, T. J. F. (1999). Scripture, logic, language: Essays on Dharmakīrti and his Tibetan successors. Boston: Wisdom Publications.

    Google Scholar 

  • Tillemans, T. J. F. (2000). Dharmakīrti’s Pramāṇavārttika: An annotated translation of the fourth chapter (parārthānumāna) (Vol. 1, k. 1–148). Vienna: Verlag der Österreichischen Akademie der Wissenschaften.

  • Tillemans, T. J. F. (2003). Metaphysics for Mādhyamikas. In G. B. J. Dreyfus & S. L. McClintock (Eds.), The Svātantrika-Prāsaṅgika distinction: What difference does a difference make? (pp. 93–123). Boston: Wisdom Publications.

  • Tillemans, T. J. F. (2004). What are Mādhyamikas refuting? Śāntarakṣita, Kamalaśīla et alii on Superimpositions (samāropa). In S. Hino & T. Wada (Eds.), Three mountains and seven rivers: Prof. Musashi Tachikawa’s felicitation volume (pp. 225–237). Delhi: Motilal Banarsidass.

  • Tillemans, T. J. F. (2016). Prasaṅga and proof by contradiction in Bhāviveka, Candrakīrti, and Dharmakīrti. In How do Mādhyamikas think? and other essays on the Buddhist Philosophy of the middle (pp. 95–110). Somerville, MA: Wisdom Publications.

  • Tillemans, T. J. F., & Lopez, D. S. (1999). What can one reasonably say about non-existence? In Scripture, logic, language: Essays on Dharmakīrti and his Tibetan Successors (pp. 247–284). Boston: Wisdom Publications.

  • Verpoorten, J.-M. (1994). The 24th chapter of the Tattvasaṁgraha: Refutation of the Mīmāṁsā doctrine of Vedāpauruṣeyatva. In R. C. Dwivedi (Ed.), Studies in Mīmāṃsā: Dr. Mandan Mishra felicitation volume (pp. 117–130). Delhi: Motilal Banarsidass.

  • Vose, K. (2015). Do Mādhyamikas see what the rest of us see? Early bKa’ gdams pas on ‘commonly appearing subjects’ (chos can mthun snang ba). Journal of Buddhist Philosophy, 1, 95–135.

    Article  Google Scholar 

  • Williams, P. (1998). The reflexive nature of awareness: A Tibetan Madhyamaka Defence. Richmond: Curzon Press.

    Google Scholar 

  • Wogihara, U. (1932). Abhisamayālaṃkārālokā Prajñāpāramitāvyākhyā: The work of Hari-bhadra. Tokyo: Sankibo Buddhist Book Store.

    Google Scholar 

  • Yonezawa, Y. (2019). On svatantram anumānam. Journal of Indian and Buddhist Studies, 67(3), 82–88.

    Google Scholar 

  • Yoshimizu, C. (2010). Źaṅ Thaṅ sag pa on theses (dam bcaʼ, pratijñā) in Madhyamaka thought. Journal of the International Association of Buddhist Studies, 32(1–2), 443–467.

    Google Scholar 

  • Yoshimizu, C., & Nemoto, H. (2013). Zhang Thang sag pa ʼByung gnas ye shes. dBu ma tshig gsal gyi ti ka. Part I: folios 1a-26a3 on Candrakīrti’s Prasannapadā ad Mūlamadhyamakakārikā 1.1. Tokyo: Toyo Bunko.

  • Yotsuya, K. (1999). The critique of Svatantra reasoning by Candrakīrti and Tsong-kha-pa: A study of philosophical proof according to two Prāsaṅgika Madhyamaka traditions of India and Tibet. Tibetan and Indo-Tibetan Studies 8. Stuttgart: Franz Steiner Verlag.

  • Zhang Tangsakpa (Zhang thang sag pa ʼByung gnas ye shes). dBu ma tshig gsal gyi ti ka. In Yoshimizu and Nemoto 2013.

Download references

Acknowledgements

I thank Pascale Hugon for sharing passages from the Sanskrit edition of Chapter III of Dharmottara’s Pramāṇaviniścayaṭīkā that she, Takashi Iwata, and Toshikazu Watanabe have prepared for publication in the Sanskrit Texts from the Tibetan Autonomous Region series. I thank two anonymous reviewers of this article for their extensive, detailed, and insightful comments, which have greatly improved the present work. They are, of course, blameless for the remaining faults.

Author information

Authors and Affiliations

  1. Religious Studies Department, College of William and Mary, Williamsburg, VA, USA

    Kevin Vose

Authors
  1. Kevin Vose
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Kevin Vose.

Ethics declarations

Conflict of interest

The sole author states that there are no conflicts of interest.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Vose, K. When Did Svatantra Inference Gain Its Autonomy? Śāntarakṣita and Kamalaśīla as Sources for a Tibetan Distinction. J Indian Philos 48, 703–750 (2020). https://doi.org/10.1007/s10781-020-09436-z

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10781-020-09436-z

Keywords

Search

Navigation