Published in the Indian Culture, Volume 1, 1984, pp. 263-273 All the schools of Indian philosophy have discussed the problem of the various sources of knowledge (pramanas) whereby the truths regarding the ultimate reality can be tested. The discussion of the doctrine of pramana is therefore regarded as an indispensable preliminary to the main problems of philosophy. It is quite in the fitness of things that we must make sure of the correctness of the means of knowledge (prameya) are cognised by some instrument of knowledge. As the Maitryupanisad [1] has it, the object of knowledge (prameya) are cognised by some instrument of Knowledge. Isvarakrsna in his Samkhyakarika [2] has similarly observed, that the knowable are determined by the various sources of knowledge (prameyasiddih pramanaddhi) which, according to him, are perception, inference, and authority. It has been a common dictum with Indian philosophers that the truth of what is to be known depends on the means of knowing it (manadhina meyasiddhih). The Jain philosopher Umasvati in his Tattvartha-dhigamasutra [3] says: the truth of the ultimate reality is to be known by means of pramana and naya. In the Nyayabindu, Dharmakirti [4] introduces the problems of right cognition with the statement that introduces the problems of right cognition with the statement that inasmuch as right knowledge is the condition precedent of all objects of human desire, it is to be expounded.
Thus the problem of pramana, though mainly associated with Nyaya school, has come to be treated in every system of Indian philosophy.
The appropriateness of these discussions on pramana has been adequately appreciated by Prof. Max Miller who in his preface to ‘The Six Systems of Indian Philosophy’ [5] observes as follows:-
‘Such an examination of the authorities of human knowledge (pramanas) ought, of course, to form the introduction to every system of philosophy. How much useless controversy would have been avoided, particularly among Jewish, Mahomedan, and Christian philosophers, if a proper place had been assigned in limine to the question of what constitutes our legitimate or our only possible channels of knowledge, whether perception, inference, revelation or anything else!’
There is, however, much difference of opinion with regard to the various problems connected with the pramanas among the different schools of philosophy. The Buddhist logicians have reduced their problems of pramanas to four and have accordingly discussed them under these four heads: 1. Number (Samkhya), 2. Nature (laksana), 3. Object (Gocara), 4. result (phala). [6]
Let us take up the first problem and see how the Buddhists differ from the rival schools. The number of pramanas by the different schools varies from one to eight. Varadaraja, a later Brahmanic logician, has summed up the different views in the following karikas:-
Pratyaksam anumanam syad upamanam tathagamah
Pramanam pravibhajyaivam aksapadena laksitam
Pratyaksamekam carvakah kanadasugatau punah
Anumanamca taccatha samkhyah sabdam ca te api.
Nyayaikadesino pyevam upamanam ca kecana
Arthyapattya sahaitani catvaryaha prabhakarah
Abhavasasthanyetani Bhatta vedantina statha
Sambhavaitihyayuktani tani pauranika jaguh. [7]
Aksapada in his Nyaya system has divided pramanas into four: pratyaksa (perception), anumana (inference), upamana (analogy), and agama or sabda (authority). The Vasesikas admit two, i.e., pratyaksa and anumana. Besides these two, pratyaksa and anumana, the Samkhya as well as the Yoga [8] accept one more, i.e. agama or sabda (authority). Some school [9] of Nyaya accepts these three, pratyasha, anumana and sabda, while others take upamana (analogy) along with them. Prabhakara, one Mimamsa teacher, admits these (presumption). Bhatta, another Mimamsa teacher, accepts abhava (negation) besides the five above, i.e. pratyaksa, anumana, upamana, sabda, and arthapatti. And the Pauranikas admit sambhava (probability) and atitihya (tradition) along with above six, i.e. they accept altogether eight pramanas which are as follows : I. pratyaksa, II. anumana, III. Upamana, IV. Sabda, V. arthapatti, VI. Abhava, VII. Sambhava and VIII. aitihya. There are some other minor pramanas as pratibha [10] (intuitive knowledge) and cesta, the first being admitted in some form or other by most of the schools of Indian philosophy and the latter by the Tantrikas.
But in several schools attempts have been made due to reduce the number of these various pramanas to some modest one, say, two, three or four as the case may be. Nyaya has refused to accept more than four pranamans and has accordingly attempted to include aitihya, arthapatti, sambhava, and abhava into one or other of the four pramanas i.e. pratyaksha, anumana, upamana, and sabda admitted by it. Similarly Prasastapada in his Padarthadharamasamgraha, [11] has submitted all the pramanas under pratyaksa and anumana.
The pre-Dinnaga Buddhist teachers who went upon the old school of Brahmanic logic accepted sometimes four and sometimes three pramanas. [12] But from Dinnaga onwards the Buddhist logicians rejected upamana and sabda and accepted pratyaksa and anumana as the only valid sources of knowledge. Dinnaga is his Pramanasauccaya has repudiated the view that upamana and sabda are separate pramanas distinct from pratyasha and anumana. [13] Santaraksita, a later Buddhist teacher is his encyclopaedic work Tattvasamgraha (=TS) has taken great pains to refute the acceptance of the various pramans other than pratyaksa and anumana and has included some of these in either of these two pramanas, rejecting what he does not regard as valid.
The Buddhist logicians have thus reduced the number of pramanas to two I. pratyaksa and II. Anumana. Like them the Jain logicians have accepted two-fold pramana but their classification is different. According to them pramanas are divided into two I. pratyaksa and II. Paroksa. Now paroksa has been subdivided into 1. Smriti (memory), 2. Pratyabhijna (recognition) 3. Uha (the instrument of knowledge of concomitance), 4. Anumana (inference) 5. Agama (scriptural authority) [14]
The Buddhist have, however, some principle of division forthe two-fold pramana. All objects have two characteristics, I. self-characteristics (svalaksana) and 2. Common characteristic (sananyalaksana). Corresponding to these two there are two pramanas, pratyaksa and anumana. Pratyaksa pramana takes cognisance if the svalaksana, while anumana pramana takes cognisance of samanyalaksana. It is also said, that prameyas or the object of knowledge are either perceptible (pratyaksa) or imperceptible (paroksa) and besides these two kinds no third is possible and these two kinds of prameyas two pramanas are necessitated, pratyaksa pramana for perceptible objects and anumana pramana for imperceptible objects. [15]
By paroksa Buddhist logicians mean samanyalaksana or the class characteristics of a thing. On looking at a cow when we say this is a cow, this judgement though apparently arising from perception (pratyaksa) is not really a case of perceptual knowledge (pratyaksa pramana), for the class notion or the concept of the cow is, according to the Buddhists as opposed to the Naiyayikas, is not the result of perception proper. We can perceive this cow or that cow but cowness we cannot perceive. But to express a judgement this is cow, we require the concept of cow to associate with what we directly perceive as the particular configuration and colour - (sannivesopadhivarnatmaka = svalaksana) [16] of the cow we are looking at. So the judgement, ‘this is cow’, does not arise from perception alone: it is paroksa (imperceptible), involving as it does the samanyalaksna. Now we see that by these two terms in the other schools of Indian philosophy. Santaraksita observes pratyaksha and anumana, as recognised by others (like the Vaisesikas) for the purpose of apprehending the upadhis (categories) i.e. dravya (substance), guna( attribute), kriya (actions) and the like have not been properly defined and they are to be defined correctly from a Buddhist point of view. [17]
The Buddhist position on the twofold division of pramanas has been introduced as purvapaksa in the Nyaymanjari of Jayanta Bhatta who, considering the numerous quotation not to be ordinarily found in Brahmanic or Jain works, seems to have a first-hand acquaintance with the Buddhist texts and it may be summarised as follows [18].
Objects of knowledge admit of a dichotomic division on the basis of the two mutually exclusive characteristics, pratyaksa (perceptible) and paroksa (non-perceptible), or svalaksana (having the unique character) and samanyalaksana (having the general or common character). As these two contradictory characteristics make up the entire world of objects, no third is possible and we have it so on the authority of perception knowledge. There is no denying the fact that the perception of a blue thing, while making known to us what is blue, differentiates the blue from what is not blue, the latter not being presented in the perception. What is presented in the perception of the blue is the blue and all else is not-blue. Now that a thing can be either blue or not-blue, the perception of a blue precludes the possibility of anything besides blue and not-blue. So it has been said that the perception of a thing determines its existence of any third entity besides what it is and what it is not. [19] Thus the perception of a thing, while establishing its perceptibility, proves the non-perceptibility of all else and does away with the possibility of any characteristics besides perceptibility and non-perceptibility. Thus perception proves that an object of knowledge can be either perceptible (pratyaksa) or imperceptible (paroksa), or in other words it is either svalaksana or samanyalaksana.
Now it is argued that perception, as understood by the Buddhist logicians, is to take perception of a thing to give rise to any other knowledge such as distinguishing the thing perceived from what is not and establishing the non-existence of any third entity besides what is and what is not. Now the objection can be met, if recourse be has to the pragmatic efficiency of the vikalpas or the conceptual cognitions that follow in the trail of the perception of the things-in-themselves cannot. So the threefold aspects of perception that have been stated above are with the conceptual cognitions that arise immediately after it.
Jayanta Bhatta proves the hollowness of the above argument and his reply may be summarised as follows:-
Indeterminate perception of the Buddhist cannot connect any cognition either with the past or with the future and, except taking cognizance of what is directly presented, is incapable of proving the negation of any other thing. The vikalpas which have been requisitioned for the purpose are mere imaginary construction according to the Buddhists and therefore their testimony is of little use. Again, in perception we perceive a thing but we do not perceive that is the perceptible. It has been pertinently pointed out by Bhatta [20] that audibility (sravanata) cannot be cognised by perception. It can, on the contrary, be ascertained only by means of inference on the observation of its relation in presence and absence. Now a sound is seen to be perceived by a deaf person. The sense of hearing, too, being imperceptible is inferred only on the cognition of sound and the percipient’s reaction towards it. And the uniform absence of this sound and the percipient’s reaction on the part of a deaf person leads to the inference of his deafness. So audibility or for matter of that perceptibility as a quality can only be inferred by virtue of the relations which a sense-object is seen to bear to a sense-organ. The latter, again, is capable being ascertained only on the evidence of a person’s reaction towards a sense-object, and is thus a matter of inference in its turn. That being the case, it is impossible to except that perceptual knowledge should give direct evidence of the perceptibility or otherwise of an object. The whole process shown above involves a tedious chain of inferences and when the Buddhist puts all thus knowledge to the credit of perception he is guilty of a grievous misinterpretation of the real nature of perception.
So baseless is the thesis of the Buddhist that by perception we can determine that an object of knowledge must possess either of these two mutually exclusive characteristics- pratyaksa and paroksa, or, in other words, svalaksana and samanyalaksana.
The twofold division of pramana has been also criticized by the Madhyamikas whose arguments against the neo-logicians of Dinnaga’s school may be summed up as follows [21].
Logician: You say entities do not originate. But is this assumption based on any pramanaor not? If the answer be in affirmative, you must state the number, the definition, and the content of pramana and also state how a pramana derives its validity. If your assertion be not based on any pramana, it has no value, since nothing can be established except by a pramana. If a statement can be valid without a pramana, one is as much valid as it’s opposite.
Madhyamika: A pramana is necessitated only in the case where one has got something definite to determine (niscyaya). One who has something undermined will go in for its determination. (But now that we have nothing to determine the above objection is wholly out of place.) We can do without pramanas and are, therefore, under no obligation to state their number, definition, etc.
Logician: You profess to make no definite statement whatsoever. But when you say entities do not originate out of themselves, or of some external source or of both combined or from any cause whatever (na svato napi parato na dvabhyam napahetutah bhava bhavanti), it sounds like a definite assertion.
Madyamika: This definite statement is only with references to the common run of people and their line of thinking and not to the saints. [22]
Logician: It then follows that he saints do not believe in or attach importance to arguments (upapatti).
Madhyamika: How can it be known that they do not resort to arguments? As for the ultimate truth (paramartha), they remain silent. [23]
Logician: If the saints do not have recourse to arguments, how do they manage to explain the ultimate truth to ordinary people?
Madhyamika: For the purpose of instruction to ordinary people, the saints do not bring in any arguments of their own but provisionally accepting only those arguments that are held by the people at large, they teach the people. [24]
Thus form the Madhyamika standpoint Candrakirtti condemns logic for the determination of the ultimate truth, but he cannot altogether invalidate logic so far as it is connected with our empirical life. His point seems to be that if logic is intended for our ordinary life, it is proper that the logic of the realist school of philosophy (i.e. the Naiyayikas) is to be preferred as it well accords with and is suited to the popular way of thinking.
Candrakirtti further argues and refutes Dinnaga’s theory of twofold (pramanas) with reference to the two objects of knowledge - the svalakasana (the unique, the particular characteristics of a thing) and the samanyalaksana (the general characteristics of a thing). Now it can be argued that the theory of svalaksana and samanyalaksana) instead of the two, svalaksana and samanyalaksana, and consequently a third source of knowledge is needed. If it’s denied that there is nothing apart from svalaksana and samanyalasana, then the two would remain without any substratum, [25] i.e. anything to be characterised.
Nagarjuna says: In the absence of a characteristic (laksana) that which is to be characterised (laksya) becomes impossible. And, again, in the absence of the latter, the former becomes impossible. [26] Even on the interpretation of laksana as that which is characterized [27], the difficulty cannot be avoided (a thing will have to be characterised by itself, which is impossible). According to this new interpretation, a characteristic must be instrumentally related to what is characterised.
Again, if Svalksana be thought to be characterised by self-awareness (svasamvitti), this self-awareness will require another self-awareness and thus the difficulty of the infinite regress arises. It is also to be pointed out that the theory of self-awareness is not acceptable. Consciousness cannot apprehend its own self. The edge of sword cannot cut the edge of the same sword. The tip of a finger cannot touch the tip of the self-same finger. [28]
The Madhyamika asks, what is the relation between ‘what is characterized’ and ‘the character’. If the two be entirely distinct, they will cease to be what they are, as they cannot remain dissociated will cease to be what they are, as they cannot remains without the former. If the two be identical, neither of them exists.’ [29]
Being thus confronted with a dilemma that upholders of the svalaksana theory would reply that the relation between what is characterised and the characters is unspeakable, [30] it being similar to what the Madhyamikas regard as the true nature of Reality. This is Madhyamikas reject on the ground that a thing said to be unspeakable, when it is incapable of any dichotomic division, but unspeakable, when it is proved that we cannot independently cognize - this is the characteristic, this is the thing characterized, -- we conclude that both are unreal’. [31]
Taking pranama as the instrument of knowledge, the question arises ‘Who is the knower, of whom pramana is an instrument?’. If the reality of the pranamas be admitted, the Madhyamika says, the difficulty of positing a cognizer cannot be avoided, but the theory of a cognizer would be against
All schools of Buddhism not excluding the neo-logicians of Dinnaga’s school. Again, svalaksana theory presupposes something of which it is the characteristics. The reply that it is similar to the expressions as the body of a statue or the head of Rahu (who, being without the trunk, exists in head only), is not convincing. The definition of perception according to the svalaksana theory is too narrow, as it does not apply to every day expression as the pitcher is perceived (ghatah pratyakshah) and all these cannot be ignored or held wrong.
The svalaksana theory of perception is claimed to be based upon the scriptural statement that a man with the visual consciousness apprehends blue not that it is blue.[32] This passage, the Madhyamika contends, is not intended to give a definition of perception that is meant to indicate that the knowledge through the five sense organs is unconscious (jada). The point of the Madyamika seems to be this: From the higher standpoint the theory of the pranamas, be it four or be it two, is untenable. The nirvikalpa or svalaksana is as much unreal as the savikalpa or samanyalaksana. So if the defects of the empirical logic with four pramanas should be accepted.
In the Brahmanic system pratyaksa has been placed on a higher rank as a valid source of knowledge (pramana-srestha or jyestha) probably on the ground that perception gives us an immediate knowledge of things, and also supplies the data of inference. As the Nyayasutra has it, inference is preceded by a perceptual knowledge (atha tatpurvakam trividham-anumanam, etc.). This view has been rejected by the Buddhist logicians. As pramana, bothpratyaksa and anumana are equally important and valid and the both are equally helpful in making us reach an object by virtue of their unfailing correspondence with the object. [33]
So there is no reason why pratyaksa should be ranked higher than other pramanas. They are all equally important in their respective places. [34]
Footnotes
1. Pramanema prameyasyopalabdhik. Maitryupanisad (A.S.B), 6, 14.
2. Drtam anumanam aptavacanam ca sarvapramanasiddhatvat trividham pramanam istan prameyasiddhih pramanaddhi. Samkhyakarika, 4.
3. Pramanamyabhyam tattvarthadhigamah. Tattvarthadhigamasutra, II.
4. Samyaginanapurvika sarvapurusarthasiddhiriti tad vyutpadyate. Nyayabindu, I.I.
5. P. XIII. Also ef. Ibid, p.XII
6. Caturvidha catra vipratipattih, Samkhya-laksana-gocara-phala-visaya, p. 6. Also tatar pramane svarupaphalagocarasamkhyasu paresam vipratipattis caturvidha. Cf. Tattvasamgraha (G.O.S), p 366. For similar viewsof the Jaina logicians, see Siddharsigani’s tika on the Nyayavatara (ed. By Dr. Vaidya, 1928 p.8)
7. Tarkikaraksa (Reprint from the Pandit), pp. 55-56. There is a similar Karika, probably by some Jaina philosopher, referred to in the tippana of Deva-bhadra on the Nyayavatara, ed., by Dr. Vaidya, p.9):
Carvako dhyaksamekan, sugatanabhujau sanumanam sasabdam
Taddvaitam paromarsah sahitam upamaya tattrayam caksapadah.
Arthapattya prabhakrd vadati sa nikhilam manyate Bhatta etat.
Sabhavam dve pramane jinapatisamaye sppastato’ spastatasca.
8. Pratyaksanumanagamah pramanani. Yogusutra, 1, 7.
9. Bhasarvajna in his Nyayasara enumerates three pramamas, viz. pratyaksa, anumana, and agama. C.f. tattrividhan pramanam pratyaksam anumanam agamasceti. Nyayasara (Bib. Indica), p.2 Vide History and bibliography of Nyya-Vaisesika Literature by Gopinath Kaviraja in the Saravati Bhavana studies, pp. 81-82.
10. See the learned and informing article by Gopinatha Kaviraj on the doctrine of pratibha in the Annuls of Bhandarkar Oriental Institute.
11. Prasastapada Bhasya, with Kandali (Viziannagram Sanskrit Series), p.220.
12. Nagaarjuna -four pramanas. Asanga-three (rejects upamana).
Three pramanas, pratyaksa, anumana, and agama are also referred to by Sthiramati in his commentary on Vasubandhu’s Trimsika, p. 26. Sutta-Udaharana the earlier Buddhist controversies in the Kathavatthu.
13. History of Indian Logic, p.287.
14. These five are not, as they say, independent pramanas, but are only varieties of paroksa in which they have been included like four sub-classes of indriyavijana, manovijnana, svasamvedana, and yogijnana, all subsumed in pratyaksajinana:
Yathaiva hi pratyaksalaksanasamgrhitanindriyajnanamanasa-svasamvedana-yogijnanani saugatanam na pratyaksad atirieyante, tathaiva hi paroksalaksanaksiptani smrtyadini na mulapramanasamkhyaparipanthiniti…Pramanamimamsa, p.50
15. Na pratyaksaroksabhyam meyasyanyasya sambhavah tasmat
Prameyadvitvena pramanadvitvam isyate.
Quoted in the Saddarsanasammuccaya, I, 2:
Compare Dinnaga in the Pramanasamuccaya
Mnon sum dan ni rjes dpag tshad ma dag ni mtshan nid gnis
Gzal bya de la rab sbyor phyir tshad ma gzan ni yod ma yin.
Pratyksa and anumana re apprehended and besides these two no other pramana is necessary. Also, pratyaksasca paroksasca dvidhaivanho vyavasthitah. T.S, 1701, et. Seq.
16. Nyayabindutika (B.B), p. 7.2.
17. Pratyasam anumanamca yadupadhi-prasiddhaye
Parair uktam na tat siddham…. T.S., 1213
18. Nyayamanjari (=NM),Vizianagram Sanskrit Series, pp.28ff.
19. Taduktam tat paricchinatti anyad vyavacchinatti trtiyaprakarabhavamca sucayaty ekapramanavyaparah, NM, p.29. 13-14
20. Na hi sravanata nama pratyaksenavagamyate
Sanvayavyatirekabhyam gamyate uadhiradisu.
Slokavarttika, Anumanapariccheda, 60-61. Quoted with a little variation in NM, p.32.
21. Madhyamakavrtti (bib. Buddhica), pp. 55. 11-75.13. Buddhist Conception of Nirvana-Stcherbatsky, pp. 135-164
22. Niscitam idam vakvam loksaya svaprasiddhyaivo naryanam, ibid, p. 57. 5ff.
23. Paramartho hy aryanam tusnimbhavah, ibid., p.57.8.
24. Na khalvarya lokasamvyavaharenopapattim varnayanti kintu lokaia eva ya prasiddhopapattistam paravabodhartham abhyupetya tayiva lokam bodhayanti, ibid, p.57. Ioff.
25. Tada laksanamapi nirasrayam, ibid., p.59.9.
26. Laksanasampravrthau ca na laksyam uapadyate
Laksasyanupattan ca laksanasyapyasambhavah, ibid., p. 59. 10-11
27. Krtyaluto bahulam iti karmani lutam krtva laksyate tad iti laksanam, ibid, p.60, I-2
28. Na ca cittam cittam anupasyati. Tad yathapi nama tayaivasidhaaraya saivasidhara na sakyate chettum na tenaivangulayagrena tad evangulyagram sakyate sprastum, ibid., pp 62-63.
29. Laksyallaksanam anyascet syattallaksyam alaksanam
Tayorabhavo nanyatve vispastam kathitam tvaya, ibid. p. 64. 5-6
30. Avacyataya siddhir bhavisyati, ibid., p. 64. 10
31. Buddhist Conception of Nirvana, p. 148
32. Caksurvijnanasangi nilam vijanati natu nilamiti, ibid., p. 74, 11, 8-9. See my paper on the Sources of Buddhist Logic from the traditional point of view. IHQ., IX,2.
33. Dharmottora says in course of his comment on the Nyayabindu Sutra I, 33, pratyaksam anumanamca:-
Cakarah pratyaksanumanayo stulyabalatvam samuccinoti. Yatharthavinabhavitivad artham prapayat pratyaksam pramanam tadvad arthavinabhavitvadanumanam api paricchinnam artham prapayat pramanam iti.
An older Buddhist logician like Nagarjuna is, however, at one with the Brahmanic philosophers in holding pratyaksa to be superior to other pramanas owing to their being independent on pratyaksa. Cf. the Upayahrdaya:-
Caturviham pramanam pratyaksam anumanam agamasceti. Catursu pramanesu pratyaksam srestham kutah punah pratyaksam sretham iti cet aparesam trayanam pramananam pratyakso pajivakatvacchraisthyam.
Upayahrdaya (reconstructed from Chinese by Prof. Tucci; Pre-Dinnaga Buddhist Texts on logic, p-13).
34. But it is to be noted in this connection that according to the Dinnaga’s school of logic the object of perception is real while that of inference is unreal or imaginary. Cf. NBT (Bib. Bud.), p.7.12ff-
Bhrantam hy anumanam. Svapratibhase’ narthe’ ethdyavasayena pravrttatvat. The Jaina logicians too, the Buddhists, refuse to attach any superiority to pratyasa. It cannot be said, they argue, that praatyaksa always precedes all other pramanas, for on the contrary it is sometimes seen that pratyaksa is preceded by anumana and agama; we first inferentially from smoke or authoritatively from the information of any trustworthy person (apta) and then going to the proper place, perceive fire.
Commenting on the Sutra, ‘pratyaksam paroksam ca’, Hemacandra says in his Pramanamimamsa (I.I.II), -
Cakarah svavisaye dvayostulyabalatvakhyapanarthah. Tena yadahuh sakalapramanajyestham pratyaksam iti tadapastam. Pratyaksamitipurvakatvaditarapramananam tasya jyesthateti cet, na oratyakksapi pramanantarapurvakatvopalabdheh. Lingadapto padesadvavahnyadikam avagamya pravrttasya tadvisayapratyasotpatteh.’
Published in the Indian Culture, Volume 1, 1984, pp. 263-273