Pramana-varttika, ( Sanskrit: “Commentary on Valid Knowledge”) perhaps the foremost work on Buddhist logic and epistemology, written in the 7th century. The Pramana-varttika is the chief work of Dharmakirti, originally a southern Indian Brahman.
The Pramana-varttika is written in about 2,000 stanzas of mnemonic verse, and its four chapters deal, respectively, with inference, valid knowledge, sense perception, and syllogism. Composed as comprehensive commentaries on an earlier work by Dignaga, Dharmakirti’s treatises in turn stimulated a great number of commentaries and have become the standard works in their field, especially in Tibet.
It hardly needs to be stated that the question of Pramaa.na, i. e. the source and proof of knowledge-in all its four aspects of Number (Sa"nkhyaa), Nature (Svaruupa), Object (Arth or Vi.saya) and the Result (phala); constitutes the heart and soul of Epistemology as we have it in various systems of Indian philosophy. I intend drawing through this short paper, the attention of the scholars to the question of the number of Pramaa.nas as pertains to the Buddhist Epistemology. While doing so I am fully conscious of the first reaction, they may well evince. It is likely to be said, "What is there in this question ? It is already settled. The Buddhist accept two Pramaa.nas Pratyak.sa (perception) and Anumaana (inference) and the fact is testified not only by the modern historians of Indian Philosophy but also by such well-known authors of compendia on Indian Thought as Maadhavaacaarya and Haribhadra Suuri."
I dare challenge the view, throwing thus overboard the heavy burden of tradition, emboldened by an oft-quoted advice of the Buddha, as we find it recorded in A"nguttara-Nikaaya. "Believe nothing", the Tathaagata is attributed to have said, "because a so-called wise man said it. Believe nothing because the belief is generally held. Believe nothing because it is written in ancient books. Believe nothing because it is said to be of a divine origin. Believe nothing because someone else believes it. But believe only what you yourself judge to be true."
Contrary to the general belief, what I judge to be true is that Buddhist Epistemology, in general, does not accept Anumaana (Inference) as an independent source and proof of knowledge and the only school of Buddhist Philosophy that seems to do so fails to show it as such and succeeds merely in depicting Anumaana to be an instrument dealing with connected, directed and reflective, but abstract thought, at best.
I, for one, am not able to understand as to why should not we check our tendency to lump together schools of thought that differ widely. Covered, as they are, by the general name of Buddhist Philosophy, it is alright for us, to seek in them the common and essential. But, in the process, we must not gloss over the important differences and distinctions. Because, if we do so, we are likely to confuse issues. To me, it seems an all the more unpardonable a mistake when knowing fully well the various distinctions and differences among various schools of thought one attributes something to the totality of them what may legitimately or illegitimately belong to only one of them.
To my mind, Maadhavaacaarya, as also all those who seem to follow him in this, appears to commit precisely this serious mistake. For, he is not only aware of the division of Buddhist Philosophy into four important points of departure with the names Vaibhaa.sika, etc. but also takes note of their epistemological differences. Further, he does indicate very clearly that only one of the schools, namely, the Sautraantika alone could, if at all, accept 'Anumaana as an additional Pramaa.na to Pratyak.sa to gain knowledge of the external objects, in particular, and that too by implication as his actual words are: "According to the Sautraantikas the external objects are not graspable by Pratyak.sa (Sautraantikena pratyak.sa graahyortho na bahirmata.h)". In the light of this, one is left wondering as to what makes Maadhavaacaarya attribute the view to the Buddhists, in general, that they accept two .
It is true, the Sautraantika school is distinguished from the. Vaibhaa.sika in being the upholder of Baahyaanumeyavaada as against the Baahyapratyak.savaada view of the latter. But that is all that we have of the Sautraantika position. I am unaware if there is any detailed treatment of Anumaana in all its (already mentioned) aspects by any thinker claiming to be a Sautraantika. Scholars are yet not sure to which of the four well-known schools of Buddhism, namely, Vaibhaa.sika, Sautraantika, Yogaacaara and Maadhyamika do Dharmakiirti, the celebrated author of the well-known epistemological treatise 'Nyaayabindu' and Dharmottara, the accredited commentator of the work belong .
May be, they, along with Di"nnaaga, the author of such important works on Buddhist Epistemology as Pramaa.nasamuccaya, Nyaayaprave'sa, and lambanapariik.sa, etc. subscribe to a school independent of all the four. Their school is known, roughly, by the name of the school of Buddhist logic. It is to this school that 'Saantarak.sita, the author of Tattvasa"ngraha and Kamala'siila, the commentator of the work also seem to belong. Despite having a common bond, pressed by their metaphysical considerations, the great logicians, seem to pull in different directions. If Di"nnaaga is keen to take a neutral stand and wants his logic to be an instrument fit for the use of both Realists and Idealists and is ready to drop the character 'non-illusiveness (abhraantam)' from his definition of pratyak.sa to that effect , then 'Saantarak.sita and Kamala'siila do lean towards the Yogaacaara position and Dharmakiirti and Dharmottara towards a position somewhat between the Sautraantika and the Vaibhaa.sika . T.R.V. Murti's remark seems to be apt in this context when he observes, "there is no such a thing as a neutral logic which every philosopher accepts or has to accept ".
No doubt, they (the Buddhist Logicians) all, apparently accept both Pratyak.sa and Anumaana as Pramaa.nas. But, somehow, caught in the web of their own reasoning, guided or misguided by the consequences of accepting a certain definition of Pratyak.sa (Pratyak.sa"m Kalpanaapo¯ ham), they find it extremely difficult to maintain even Savikalpa pratyak.sa (determinate Perception) to be an instrument of knowledge what to say of the Anumaana "which is" in the words of'Satkari Mookerjee, "invariably associated with ideal constructions (Kalpanaa) ".
It is hard for me to agree with Mookerjee when he suggests, that Di"nnaaga's omission to put pratyk.sa under the rubric of "valid knowledge"(Samyagj~naana) was a trifling matter unduly made capital of by Uddyotkara  and so was the fact of not caring, to supply the word J~naana in his definition of Pratyak.sa on the part of 'Saantarak.sita . For, the explanation of Kamalasiila on 'Saantarak.sita's behalf and declaring not only Pratyak.sa but also Anumaana to be 'valid knowledge'-dvividha"m Samyagj~naanam  and Pratyak.samanumaana~nca on the part of Dharmakiirti  by way of amending the view of Di"nnaaga are of little avail when, despite Dharmottara's clever logical tight-rope-walking the School of Buddhist Logicians is unable to come out of Logical difficulties of its own making.
The significance of Di"nnaaga's omission and 'Saantarak.sita's lapse, I am sure, could not have been lost to Mookerjee and he should not have ridiculed the Brahmanical critics on the point had it been kept by him in his mind that nirvikalpa pratyak.sa-which alone is revalatory of knowledge-was of little practical use and Savikalpa was invalid because of Vikalpa. More so, when he is not only aware of the situation but also takes note of it elsewhere in his brilliant work on Buddhist Philosophy . "Unless and until it is determined" observes he about nirvikalpa pratyak.sa "as such the experience is as good as non-existent (Asatkalpa), because it can not lead to any activity and so there is no acquisition of anything". Despite referring to a distinction drawn by Dharmottara between Vikalpa (reflective thought) that arises in the trail of perception and generated under its influence (Pratyak.sabalotpanna) and the one that represents pure imagination without any touch with external reality and mentioning the same to be of vital importance according to him (Dharmottara). Mookerjee could not avoid the inescapable conclusion about Savikalpa Pratyak.sa as he states, "the contention that perceptual knowledge together with Vikalpa should be held as valid testimony therefore falls to the ground"'therefore' in the statement refers to his argument a part of which runs thus, "the purely subjective character of this reflective process, which is necessary for the interpretation of Perception, does not in any way detract from or add to the evidentiary value of perceptual knowledge."
Dharmottara is blowing hot and cold when he asserts that there is a distinction of vital importance between two varieties of Vikalpa and also insists at the same time that 'both the varieties are equally unreliable and invalid by their very constitution' . This alone is enough to reach the conclusion drawn by Mookerjee in a somewhat roundabout manner.
Summing up, I may say that as far as I am concerned I do not think it can be validly affirmed that the Buddhists accept Anumaana as a source and proof of knowledge. This, however, does not mean that the way Anumaana has been conceived by the Buddhists has no utility of its own. Anumaana, understood the Buddhists way has immense value as an instrument of purely formal logic.
1. Cf. Nyaayabindu '.Tikaa (Bibliotheca Indica edition) p. 7;
M. Hiriyanna, Outlines of Indian Philosophy, p. 188 ff;
3. Sarvadar'sanasa"ngraha, Ch. II, verse 36 (edited by V. S. Sharma Rishi), p. 102.
4. Sa¯ ddar'sanasamuccaya, verse 9.
5. The same as supra note 3 (2nd line of the verse)
6. Ibid verses 37, 38, pp. 102-3.
7. Ibid verse 37, second line.
13. Ibid p. 274 (See fn. 2 for Uddyokara's criticism).
16. Nyaayabindu (supra note 8), pp. 7, 8.
17. Op, cit. (Supra note 12), pp. 343-45.
18. Ibid p. 343.
19. Ibid p. 344,
21. Ibid p. 343.