When Buddhism came to Tibet in the eighth century, Tibetan Buddhist philosophers continued the epistemological programs of Dignāga (sixth century) and Dharmakīrti (seventh century). Dharmakīrti’s Pramāṇavārttika and his later Pramāṇaviniścaya attracted particular exegetical and philosophical attention. The great translator Ngog Lotsawa Loden sherab (Rngog Lo tsa ba Blo ldan shes rab) (1059?–1109?), was among the first Tibetan philosophers fully to come to terms with Dharmakīrti’s writings and a good number of his major Indian commentaries, especially those by Dharmottara (c. 740–800) and Prajñākaragupta (c. 800). Ngog Lotsawa, then, became the fountainhead of subsequent Tibetan studies in Buddhist logic and epistemology. Nearly a century later, a major sea change occurred with the advent of the 1219 Tsema rigpay ter (Tshad ma rigs pa’i gter), Treasury of Epistemic Reasoning, and autocommentary, of Sakya Paṇdḍita Künga gyeltsen (Sa skya Paṇḍita Kun dga’ rgyal mtshan) (1182–1251). Written in reaction to the exegetical traditions rooted in Ngog Lotsawa’s contributions, this work gave rise to a vast commentarial literature. The Ngog tradition philosophers were mainly interested in Dharmakīrti’s PramāṇAviniścaya, in contrast to the Sakya philosophers, who were primarily focused on the Pramāṇavārttika. Sakya Paṇḍita’s Treasury of Epistemic Reasoning—with some two dozen commentaries, making it one of the most frequently commented-on Tibetan treatises—was the result of his disaffiliation from the Ngog tradition’s interpretations of Dignāga and Dharmakīrti and his intention to let their texts speak for themselves with a minimum of theoretical interference.
Tibetan Buddhist epistemologists were concerned to distinguish between nonauthoritative (tsema mayin, tshad ma ma yin pa or tsemin, tshad min) and authoritative (tsema, tshad ma) means of cognitive access to the external world and our own inner states. That is, they were interested in distinguishing those means that result in unjustified or false belief and those that lead to genuine knowledge. They took Dharmakīrti’s Pramāṇavārttika • to provide the definitive account of knowledge, according to which there are two definitions of authoritative cognition: it provides nondeceptive (milu wa, mi bslu ba) access and new awareness (sar, gsar). Tibetan commentators, following their Indian predecessors, recognized that these definitions were not obviously coreferential. Both of these were grounded in a thoroughgoing pragmatism, anchoring knowledge to an individual’s successful activity.
The Ngog tradition distinguishes seven epistemic categories. The first five of these seven are the nonauthoritative and the last two the authoritative means of cognitive access. Ngog Lotsawa lists the nonauthoritative states in the following order:1
1. Nonascertainment of what is apparently present
(nangla ma ngepa, snang la ma nges pa)
2. Determinative cognition
(cepay yülcen, bcad pa’i yul can [i.e., ceshay, bcad shes])
3. Erroneous cognition
(logpay shepa, log pa’i shes pa [i.e., logshe, log shes])
4. Supposition
(yeece, yid dpyad) [read: cö, dpyod]
5. Doubt
(tetsom, the tshom)
But he concludes:
the first two are instances of noncognition (matogpa, ma rtogs pa); the middle two are instances of misconception (logpar togpa, log par rtog pa); the last is doubt.
He thus disagrees with Dharmottara, who, he says, argued that
Doubt is included in the category of an erroneous cognition, because it is an apprehension of the nature of both a thing and the absence of a thing. And since also the three kinds of supposition—without a reason, with a wrong reason, and with a real but unsettled reason—were included in the essence of doubt, they are forms of an erroneous cognition.2
Ngog argues that this is wrong, “because doubt has not fully determined the nature of both,” that is, whether something is or is not the case.
The origin of this classification of nonauthoritative cognition is uncertain, and the first, second and fourth of his listing are probably Tibetan, rather than Indian, in origin. The first two and the fourth of this pentad are rejected by Sakya Paṇḍita in the second chapter of the Treasury of Epistemic Reasoning on philosophical grounds. He argues that the correct classification of such states is the following triad:3
1. Noncognition
2. Misconception
3. Doubt
An important commentator on the Treasury’s autocommentary, Lowo Kenchen Sonam lundrub (Glo bo Mkhan chen Bsod nams lhun grub) (1456–1532), sums up the results of Sakya Paṇḍita’s critique of these in his work of 1482. He writes in a summarizing verse:4
Supposition is not different from doubt.
The nonascertainment of what is apparently present involves all forms of sensation.
Because a determinative cognition qua a nonauthoritative means of knowledge
is a noncognition,
There are no subdivisions and other enumerations.
Problems with this and related issues are many, and things do get complicated when we bear in mind that much of the dispute regarding this matter hinges on a complex network of interrelated views on ontology, sensation, inference and concept formation, to name but a few. It is therefore difficult to determine the degree to which this dispute is purely epistemological, as opposed to being grounded in ontological differences.5
…as for my own position, three rubrics:
1. The general defining characteristic of the nonauthoritative means of cognition.
2. The typological summary of the number of the nonauthoritative means of cognition.
3. An exposition of each of the nonauthoritative means of cognition [not translated].
An unestablished infallibility is a nonauthoritative means of knowledge.6
The defining characteristic of the nonauthoritative means of knowledge is the so-called cognition where infallibility is not established; there is no error, just as when a dewlap is denied for indicating what is not a cow.7
Second, the typological summary of the number of the nonauthoritative means of knowledge:
Noncognition, misconception, and doubt:
The threefold contraries of the authoritative means of knowledge
Were distinguished on account of the way in which they engage [the object].
Were these three consolidated in view of their essence, there is one single nonauthoritative cognitive process.
The three cognitions that are such nonauthoritative means of knowledge and the three faults of a logical argument have the same core of reasoning; as has been stated:8
As for the thesis of an inference, one proves the certainty of the argument being without force, indecisive, what is not desired, and what is desired.9
Just as what is to be proved by an inference has no force due to the grounding superstratum being unestablished, so noncognition, too, has no force wherever the mind engages an object. Just as uncertainty brings about indecision as to whether something is or is not X, so also doubt makes one apprehensive about extreme positions other than what is physically present. Just as a contradiction establishes what is undesirable, so also misconception makes one cognize what is not the case. Just as an authoritative argument establishes the desired objective, so the authoritative means of knowledge, too, cognize the object. Hence, on setting forth an argument, other alternatives than the aforementioned four are not possible. And just as, since one [of the four] has no fault, it is not possible to have more than three faults, because one is the authoritative means of knowledge, it is not possible to have more than four ways in which the mind is constituted.
Were we to consolidate these on account of their essence, then, since both misconception and doubt are simply forms of noncognition, the nonauthoritative means of knowledge are consolidated into one cognition that is not an authoritative means of knowledge. This is just like when Dharmakīrti stated that both authoritative means of knowledge are consolidated into immediate self-awareness and also the two objects of the two authoritative means of knowledge are consolidated into the epistemic object, the unique particular.10 The Tibetans, having cast yonder the epistemology of noncognition, the principal one of the nonauthoritative means of knowledge, have but meaninglessly divided the nonauthoritative means of knowledge into five.
Jackson, David. (1987) The Entrance Gate for the Wise (Section III): Sa skya Paṇḍita on Indian and Tibetan Traditions of Pramāṇa and Philosophical Debate. Vienna: Arbeitskreis für Tibetische und Buddhistische Studien.
Krasser, Helmut. (1997) “Rngog Lo tsa ba on the Sahopalambhaniyama Proof in Dharmakīrti’s Pramāṇaviniścaya.” Studia Indologiczne [Aspects of Buddhism] 4: 63–87.
Lowo Kenchen Sonam lundrub. (1988) Tshad ma rigs gter gyi ‘grel pa’i rnam bshad rigs lam gsal byed. Beijing: Krung go’i bod kyi shes rig dpe skrun khang.
Sakya Paṇḍita Künga gyeltsen. (2005) Rigs gter rtsa ‘grel dpe bsdur ma. Edited by Dpal brtsegs bod yig dpe rnying zhib ‘jug khang. Chengdu, People’s Republic of China : Si khron dpe skrun tshogs pa/Si khron mi rigs dpe skrun khang.
van der Kuijp, Leonard W. J. (1983) Contributions to the Development of Tibetan Buddhist Epistemology from the Eleventh to the Thirteenth Century. Wiesbaden: Franz Steiner.
van der Kuijp, Leonard W. J. (1993) “Apropos of Some Recently Discovered Manuscripts Anent Sa skya Paṇdita’s Tshad ma rigs pa’i gter and Autocommentary.” Berliner Indologische Studien 7: 149–162.
van der Kuijp, Leonard W. J. (2003) “A Treatise on Buddhist Epistemology and Logic Attributed to Klong chen Rab ‘byams pa and Its Place in Indo-Tibetan Intellectual History.” Journal of Indian Philosophy 31: 381–437.