1

in the Upaya-kauśalya-hṛdaya-śāstra 1 (20 kinds), the Tarka-sastra 2 (22 or 27 kinds), and Dig-nāga's Hetuvidya-nyāya-dvāra-śāstra3 (22 kinds). "Unfitness to be argued with" is also mentioned in these three works, and the Tarka-sastra perfectly agrees with N.S. in this point, though the explanations sometimes agree with, sometimes differ from, those of Vatsyāyana. The agreement evidently shows that the "unfitness to be argued with" in this work was derived from N.S.

3. The date of. N.S. has been discussed by Jacobi and Suali, as stated above. The former places the date in 200-500 A.D., while the latter places it in 300-350. The basis which Jacobi took for his argument is the fact that N.S. attacks the theory of Sunya-vāda, but the Vijñāna

vada is not alluded to.

Nagarjuna, discussing pramana in his Vigraha

1 No. 1257, translated by Kekaya in 472 A.D.

2 No. 1252, translated by Paramartha in 550 A.D.

3 No. 1223 = No. 1224, translated by Yuan Chwang and by I-tsing. Ch. iii, Nigraha-sthāna, p. 72a. Dharma-gupta, a Buddhist of Southern India, came to China in 590 A.D. and died in 619 A.D. (Nanjio's App. II, 131). According to his Life (No. 1493, p. 926, No. 1485, p. 66b) he read the work in Sha-lö (a province of Chinese Turkestan), when on the way to China. What he read consisted of 2,000 slokas. But the extant work consists of about 330 slokas. Some of the oldest Catalogues, Nos. 1604, 1609, mention that the book is in 2 vols., but the extant one is in 1 vol. And the beginning of the work clearly shows that it is only a concluding part of the original. It has only three sections in a chapter, apparently the last, called Paripṛcchā, (1) wrong refutation", (2) "true refutation (of jāti)", and (3) nigraha-sthāna. The work is sometimes ascribed to Vasu-bandhu, but this is doubtful. Paramartha commented on it, but the commentary (3 vols.) has been lost (No. 1504, pp. 56b, 92a, No. 1609, p. 11la, No. 1483, p. 77b). He also translated the Nigraha-sthāna-śāstra (1 vol.), the Pariprechaśāstra (1 vol.), and the Can-shwo- (or lun-) tão-li-lun (1 vol.); and he wrote a commentary (5 vols.) on the last, called the Explanation of the Can-lun. All have been lost. Can-shwo- (or lun-) tão-li literally means "correctly (truly) explaining reason (or reasoning)", which may have been the translation of Nyāya. Yuan Chwang translates Nyāya by "true reason (or reasoning)" (Can-li). If so, the Can-shwo- (or lun-) tãoli-lun may have been a translation of N.S. That the commentary consisted of 5 vols. suggests the five adhyāyas of N.S.

2

vyāvartani,1 says: "If a pramāņa can establish objects, the pramāņa needs also to be established by another pramana. What sort of pramāņa can establish the pramāņa? If (you say that) the pramāņa is established without another, your argument is incoherent and defective, so that another reason should be offered. If (you say that) the pramāņa is compared to the light of a lamp, which illuminates both itself and other things-so that the pramana can establish itself and other objects, your argument is a mistake, because the light does not illuminate itself like a pot in the dark, and illuminating itself is inconsistent with illuminating other things."

4

vā

N.S. says, pramāṇataḥ siddheḥ pramāṇānāṁ pramāṇaantara-siddhi-prasangaḥ,5 tad-vinivṛtter va pramāņasiddhi-vat tat-siddhiḥ, na pradīpa-prakāśa-siddhi-vat tat-siddheḥ, and kvacit tu nivṛtti-darśanād anivṛttidarśanac ca kvacid anekantaḥ8 (2, 2, 17-19). If we

1 No. 1251, pp. 13b, 18a-b. See Madhyamaka-vṛtti (Prasanna-padā, Bibl. Bud., No. 4), pp. 16, 30, 56 (n.), 69 (n.).

2 See note 6.

3 According to the commentary "itself" means "light itself". 4 According to the commentary, if the light of a lamp can illuminate itself, the light must exist as being dark before it has been illuminated. A pot is dark in a dark room, before it has been illuminated. Just so in the case of illuminating itself. But the light is not dark; consequently it cannot illuminate itself.

Ballantyne: "Since it is by Proofs (pramāņa) that the existence of Proofs is established, the existence of other Proofs presents itself (for demonstration)."

Ballantyne: "Or in the absence thereof (-i.e. of Proof,—since Proof may, in virtue of itself, be Proof), then, just as Proof is established (independently), so may this (viz. right knowledge, independently of any cause of it) be established." The translator of the Vigrahavyāvartanī may have understood -vā pramāņa-siddhi-vat as -vā-a-pramāņasiddhi-vat.

7 Ballantyne: "It is not so (that an endless series of Proofs of Proofs are required-), because it (viz. Proof) really is, just as the light of a lump is."

8 This last sūtra is not reckoned as a sūtra by the editor of the Nyāya-bhāṣya (Bibl. Ind.) or by the Nyāya-sūtra-vṛtti of Viśva-nātha ; but Vācaspati-miśra enumerates it as a sūtra in his Nyāya-sūcī-nibandha (App. in the Nyāya-vārttika, Bibl. Ind.).

compare the latter with the former, we can hardly deny the relationship between the two passages. The Vigrahavyāvartanī is principally engaged in refuting the realistic explanation of pramāņa and prameya (objects of pramāņa) and establishing the author's idealistic system. Pramāņa's are enumerated as four: perception (pratyakṣa), inference (anumana), analogy (upamana), and trustworthy instruction 1 (sabda). The refutation in this work of the relation between pramāņṇa and prameya, and the answer and the attack by N.S. on the Sunya-vāda, may lead to the supposition that the sutra was composed about the same time as Nagarjuna and Deva, in order to protect the realistic standpoint against their attack. If such be the case, the date of N.S. may be placed in 150-250 a.d.

1

4. As stated above, N.S. was commented upon by an unknown author before Vātsyāyana. The next extract from Pingala's commentary on the Madhyamika-śāstra seems to represent an explanation of inference earlier than Vatsyāyana's Bhāṣya. But it cannot be ascertained whether Pingala quoted from the commentary by the unknown author or not.

"If you argue that the existence of self is established by pramana, this is not possible, because the four sorts of pramāņa cannot support it. The four sorts are, namely, 1, perception (pratyakṣa); 2, inference (anumāna), e.g. we infer the existence of fire from smoke; 3, analogy (upamāna), e.g. for people who have not as yet seen brass, brass is compared to gold; and 4, trustworthy instruction (apta-vacana sabda), e.g. in assertions like "there exist hell (naraka), heaven (svarga), and the Uttara-kurus (a kind of paradise)', which are altogether unascertainable

=

1 The Upāya-kauśalya - hṛdaya-śāstra also enumerates four, Asanga enumerates three, pratyakṣa, anumāna, and śabda (No. 1170, pp. 65a–68b), while Dig-naga admits only two, pratyakṣa and anumāna. Dig-nāga says, “śabda and upamāna are contained in pratyakṣa and anumāna” (No. 1223 No. 1224, p. 3a = p. 8a). After Dig-nāga all Buddhists admit only two.

=

2

by perception, but are believed upon the authority of trustworthy instruction.1 The existence of self can never be proved by any sort of pramāņa; that is to say, existence is not proved by perception or by inference. Inference is reasoning concerning an object which has not yet been experienced from an object which has been seen, e.g. if a man has ever seen that, where smoke is, there is fire, he can afterwards infer the existence of fire from seeing smoke. This is not applicable to the proof of self. If you argue, 'we can prove the existence of self from seeing the five skandhas, because we have seen that self conjoins with the five skandhas. This is also supported by the following sorts of reasoning. There are three sorts of inference, pūrvavat, seṣavat, and sāmānyato dṛṣṭa. The purvavat is the reasoning in which a man has formerly known that where smoke is there is fire, and afterwards infers the existence of fire, as in the former cases (pūrvavat), from seeing smoke.3 The sesavat is the reasoning in which a man, having known that one grain of rice has been cooked, can infer from that that all other grains in the cooking kettle have also been cooked. The sāmānyato dṛṣṭa is the reasoning in which a man, having seen that another person has started from one place and arrived at another, and having hence known that he is locomotive, can infer from that that the sun has

1 Cf. Gauḍa-pada-bhāṣya on S. K., v. 4.

2 Literally, "if a man has once seen that fire has smoke, afterwards he can infer the existence of fire from seeing smoke." The first part is not strictly logical, but it may be justified if the author means the first experience, which is followed by the knowledge of the concomitant relation between fire and smoke, when the experience is repeated.

This is the second explanation of the pūrvavat in Vatsyāyana's Bhāṣya. The Bhāṣya says, “pūrvavad iti yatra yathā pūrvam pratyakṣabhūtayor anyataradarśanena-anyatarasya-apratyakṣasya-anumānam, yatha dhūmena-agnir iti" (p. 13).

This example is not found in the Bhāṣya. A similar example is used in the Gauḍa-pada-bhāṣya on S. K., v. 5, "seṣavad yathā samudrād ekaṁ jalapalam lavaṇam āsādya śeṣasya-apy asti lavaṇabhāva iti." This has the same meaning as the above. See notes on ch. i, 3, 12 (cognition).

motion, though not seen, because the sun moves from the east to the west.1 In like manner pain, pleasure, aversion, desire, and cognition, etc., are known to have a substratum, just as a people has the king as supporter'2-this is not possible. The reason is that in the procedure of the samanyato drsta we infer that the sun has motion, because we have observed its change of place; but in the case of self its existence cannot possibly be inferred from the five skandhas, because the conjunction of self with the skandhas has not been observed. In the same way existence is not proved by trustworthy instruction, because trustworthy instruction has its origin in perception. Consequently, self cannot be concluded to be something existent." 3

Vātsyāyana explains the three sorts of inference in two ways. According to the first of these the pūrvavat is reasoning from cause to effect, i.e. from the present to the future, and the seṣavat is reasoning from effect to cause, i.e. from the present to the past, while the sāmānyato dṛsta occurs only in the present. This explanation was followed by the unknown commentator on S.K. in the Chinese translation. The second explanation is more logical, and almost agrees with Pingala's, as stated above. The purvavat in the second explanation is after all the same as the sesavat in the first, as Vātsyāyana suggests,5 including the purvavat in the first explanation. The samanyato drsta in both the first and the second explanation is the same reasoning as Pingala

1 This is the first explanation of sāmānyato dṛṣṭa in the Bhāṣya The same explanation is found in the Gauḍa-pada-bhāṣya and Śabarasvamin's Bhāṣya on M.S. (p. 10).

66

2 The Bhāṣya by Vatsyāyana, yatha-iccha-adibhir ātmā icchaādayo guṇāḥ guṇāś ca dravya-samsthānāḥ tad yad eṣāṁ sthānaṁ sa ātmā-iti" (p. 14).

3 No. 1197, ch. xviii, p. 436.

4 Vātsyāyana says, trikala-yukta artha anumanena grhyante" (p. 14).

5 See n. 3, p. 87.