Tibetan philosophy

4. Epistemology, logic and philosophy of language

The major developments by Tibetans in this area are to be found in the interaction of two currents of interpretation of Dignāga and Dharmakīrti’s thought (see Dignāga; Dharmakīrti). The first current has its origin largely in gSang-phu Ne’u-thog monastery (Sangpu Nay-utok, founded in 1073), in the works of rNgog lōtsa ba Blo ldan shes rab (Ngok lodzawa Loden shayrap, 1059–1109), Phya pa Chos kyi seng ge (Chaba Chögyi sengge, 1109–69) and their disciples. Subsequently, Phya pa’s Tshad ma’i bsdus pa (Summaries of Epistemology and Logic) became the groundwork for the ‘Collected Topics’, or bsdus grwa (düra) literature, which furnished so much of the epistemology of the dGe-lugs school. From what we can glean of Phya pa’s thought from other works in his tradition and the works of his adversaries – his own works have not survived – he combined a creative intellect with what was probably a quite lacunary knowledge of the main texts of the Indian Buddhist epistemological school: the result is not unlike what we find in certain early Chinese Buddhists, namely numerous insightful interpretations and new ideas, and, at the same time, many notions which were no doubt invented and had no actual basis in Indian texts.

The gSang-phu positions were strongly criticized by Sa skya Paṇḍita, the Tibetan who probably came closest to faithfully representing the positions of later Indian epistemology and logic. In his Rigs gter (The Treasure of Reasoning), we find a repeated trenchant critique of ‘Tibetans’ or ‘Tibetans who pride themselves on being logicians’ – polemical shorthand for the gSang-phu philosophers. None the less, gSang-phu thought had a persistent appeal due to what must have been its extremely seductive appearance of subtlety and rigour – indeed, as D. Jackson (1987: 137–8) points out, Śākya mchog ldan reported that in his time, in the fifteenth century, people often felt that the ‘Summaries’ were subtle and proven correct, while the Rigs gter was ‘extremely rough’.

Central to the gSang-phu-ba versus Rigs-gter-ba debates was the problem of universals (see Universals, Indian theories of §3). The gSang-phu side (including the dGe-lugs) maintained a type of realism where at least some universals were thought to be real entities (dngos po), while Sa skya Paṇḍita himself and many of his Rigs-gter-ba followers, such as Go ram pa and Śākya mchog ldan, argued for the more usual Dharmakīrtian apoha theory, maintaining the unreality of all universals, indeed of all objects of conceptual thought. (The apoha theory of language is the Buddhist explanation of the unreality of all concepts, treating a concept of A as a mind-invented exclusion of non-A (see Nominalism, Buddhist doctrine of).) Indian Buddhism’s philosophy of language was thus considerably reinterpreted by the Phya-pa school, especially by the dGe-lugs-pa. The latter would find textual sources in later writers like Śaṅkaranandana, but while certain Indian Buddhists (for example, Dharmottara, Mokṣākaragupta and perhaps Śaṅkaranandana) were probably tending towards some form of realism on the question of universals, it is very difficult to see how the Tibetans’ rationale for realism had much, if anything, to do with the reasonings of these Indian Buddhists. The Phya-pa school’s interpretation of the Dignāgan–Dharmakīrtian version of concepts-qua-exclusion (ldog pa; vyāvṛtti), which they also applied to universals, logical reasons and all other mind-invented constructs, was to say that ‘exclusion itself’ or ‘universal itself’ (spyi kho rang) is unreal, but not all exclusions, universals, and so on, are unreal. The point probably partly turns on certain features of the Tibetan language which make this type of explanation possible. This and closely related moves, which the dGe-lugs-pas would maintain to be at the heart of ‘the difficult point of the apoha theory’ (gzhan sel dka’ gnad), had already been criticized by Sa skya Paṇḍita himself and would be devastatingly attacked by Śākya mchog ldan.

There were also numerous Tibetan developments in epistemology which genuinely contributed to clarifying key Dignāgan and Dharmakīrtian issues: for example, the Tibetan debates on the theory of truth in Indian Buddhism, on the notion of nonperception (anupalabdhi), and on the use of scriptural arguments, and Tibetan positions on the question of how to interpret Dharmakīrti’s proofs of the Buddha’s authority. We find gSang-phu-bas and Rigs-gter-bas offering (conflicting) explanations of Dignāga’s Hetucakra (The Wheel of Logical Reasons), in particular of the so-called ‘inconclusive reason which is too exclusive’ (asādhāraṇānaikāntikahetu), which is by far the most difficult fallacy to understand in Dignāga’s system of nine reasons. We also find the gSang-phu traditions, including the dGe-lugs, accepting many of the key elements of Ratnākaraśānti’s idea of ‘intrinsic entailment’ (antarvyāpti) as presented in his Antarvyāptisamarthana (The Justification of Intrinsic Entailment), even though these same thinkers still did not go so far as to accept the Antarvyāptivāda position that examples are dispensable in argumentation, and are only used for dullards (rmongs pa). There were actually very few Tibetans who considered themselves real Antarvyāptivādins, the major possible exception being the Sa-skya-pa thinker Nya dbon Kun dga’ dpal (Nyawön Güngapel, c.1300–80), who is said to have espoused this view in his now lost commentary on Dharmakīrti’s Pramāṇavārttika (Commentary on Valid Cognition).

The Phya-pa tradition, largely independently, developed a remarkably sophisticated type of debating logic, with consequences (thal ’gyur; prasaṅga) replacing the usual triply-characterized reasons (trirūpahetu) and inferences-for-others (parārthānumāna) of Indian Buddhist argumentation. This logic, which one finds elaborately developed in bsdus grwa texts, has a unique type of quantification and employs variables. The Dharmakīrtian idea of the general entailment (vyāpti, literally ‘pervasion’) between the reason and what is to be proved had presented enormous philosophical problems in Indian logic, as it was very difficult to say precisely how one was to establish or know for certain when there was vyāpti – vyāpti had to be assured by ascertaining a necessary connection (sambandha) between terms (see Inference, Indian theories of §§5–6). In this debating logic, however, the epistemological questions as to how we know that there really is vyāpti, or that there can be no counterexamples, are of curiously minor importance; they are, in any case, separated from the purely logical matter of the truth conditions for a universally quantified statement, that is, absence of counterexamples. This is a type of progress, freeing the logical and formal matters around vyāpti from the probably unsolvable epistemological issues which preoccupied the Indians. Interestingly enough, the Phya-pa school and its successors, in their classifications of various types of consequences as having twelve different types of vyāpti, went into considerable detail on the logical connections between statements in a way which was quite unknown in India. Moreover, we see here and elsewhere a marked turn towards a formal perspective: saying, for example, that such-and-such a consequence-statement will satisfy a certain kind of vyāpti if it satisfies another kind is indeed an abstraction away from content to matters of form. It is, however, difficult to argue that the rules and approach of this logic represent a propositional or term logic of the sort developed by Stoic or Aristotelian philosophers, although it is certainly not opposed to classical theorems.