Reversing Śāntarakṣita’s Argument. Or do Mādhyamikas Derive Part-whole Contradictions n All Things?

Reversing Śāntarakṣita’s Argument. Or do Mādhyamikas Derive Part-whole Contradictions in All Things?

Tom J.F. Tillemans

Indologists have significantly showed us how arguments of Indian Buddhists take on needed depth when we better understand their Brahmanical critics. Mīmāṃsaka arguments against Buddhist philosophy of language and metaphysics, for example, provide the indispensable backdrop to understanding Buddhist logicians’ ideas on scripture, meaning, universals, and abstract entities and, indeed, may ultimately remain unanswerable in some of those debates. I have long maintained that Indian Buddhist arguments are also profitably understood via Tibetan critical analyses, although the context is more complicated as Tibetan Buddhist authors very rarely present themselves as openly opposing major Indian coreligionists but instead as showing their true intentions. Here is one such thinly disguised Tibetan critique (henceforth called the “quoted passage”) of an important Indian argument about parts and wholes:

“If it were to be contradictory (‘gal ba = viruddha) for any phenomenon (chos de) to have multiple parts and to be a single thing, then singleness would not exist, and if it [i.e., singleness] were to be nonexistent, multiplicity would be nonexistent too. Then, those two being nonexistent, and given that there is no third alternative (phung = rāśi) apart from those two, there could not be anything at all. Therefore, those two [i.e., having multiple parts and being a single thing] are not contradictory, and thus it was not taught [by Śāntarakṣita and Kamalaśīla] that if one believes that God has multiple temporal stages, He would have to be non-single. And [[[Kamalaśīla]]] did not mean [that having parts entailed not being single] when he said that [the non-Buddhist opponents] have accepted an antecedent term (khyab bya = vyāpya) [viz., having multiple temporal stages] for non-singleness.” Tsong kha pa, dBu ma rgyan gyi zin bris (see appendix 1, §11).

This is surprising stuff from anyone who is a professed Buddhist Philosopher of the Middle (mādhyamika), as were Śāntarakṣita (circa 725-788 C.E.) and his disciple Kamalaśīla (circa 740-795), and as was the subtle Tibetan dGa’ ldan pa (or dGe lugs pa) thinker Tsong kha pa Blo bzang grags pa (1357-1419). Although there are many cases where Tibetans have reformulated Indian thinkers’ own positions in new and perspicacious ways, what we see here is not just a reformulation or fine tuning of an Indian position: it is an important philosophical critique and, as we shall try to show below, presents a rival view on parts and wholes and thus on the Philosophy of the Middle (madhyamaka) in general, even if it is presented as a proof of Indian authorial intent. The context is Tsong kha pa’s discussion of Śāntarakṣita’s Madhyamakālaṃkāra and Kamalaśīla’s Madhyamakāloka, and notably the famous “neither one nor many” argument (gcig du bral gyi gtan tshigs = ekānekaviyogahetu). In the words of Śāntarakṣita, this argument shows that “All things, whether promoted by Buddhists or non-Buddhists, are without any intrinsic natures (niḥsvabhāva), like a reflection, because they have in fact/really (tattvatas) neither the intrinsic nature of single things nor that of multiplicities.”

Madhyamakālaṃkāra, in keeping with that programmatic opening verse, is indeed naturally read as a long and elaborate derivation of latent contradictions in all versions of things, be they Buddhist or non-Buddhist, for Śāntarakṣita sought to show that each such putative thing would have to be in fact single or many, and yet they were always neither. There are also several other contradictions that seem to be derived from versions of things at key stages of the “neither one nor many” argument, such as the contradiction between being partite and being one single thing. However, while that is a natural reading of the Madhyamakālaṃkāra, if we follow Tsong kha pa then such a formulation of Madhyamaka mereology would not be what Śāntarakṣita and Kamalaśīla had in mind at all. In what follows, we will do a close reading of the passage quoted above from Tsong kha pa’s dBu ma rgyan gyi zin bris, his “Synopsis of the Madhyamakālaṃkāra,” in which he turns the tables on, or reverses, the very argument that Śāntarakṣita and his disciple likely had in mind. To ensure context for the passage quoted above, a translation of the whole discussion in dBu ma rgyan gyi zin bris is given in an appendix. Relevant paragraphs in the appended translation are indicated as supporting textual references for our reading.

The argumentation in which the quoted passage is situated is somewhat technical and needs a bit of explanation from the outset. The passage figures within a discussion about how a Mādhyamika can derive propositions and attribute them to opponents when those same opponents do not themselves accept them explicitly (dngos su = sākṣāt). Śāntarakṣita had warned against taking the opponent’s recalcitrance as decisive in Madhyamakālaṃkāravṛtti to verse 1: “Do not think that this [[[neither one nor many]]] reason is unestablished.” He fully recognized that opponents would vociferously contest his proof that entities are not single things but discounted that opposition as not insurmountable. It fell on Kamalaśīla to explain how.

Initially in his Madhyamakālaṃkārapañjikā and then later in more detail in his Madhyamakāloka (see §1), Kamalaśīla argued that the Mādhyamikas’ arguments bring out what their opponents “have [already] implicitly accepted” (shugs kyis khas blangs pa = sāmarthyād abhyupagata) or “have in fact (kho na = eva) accepted by implication” (shugs kyis khas blangs pa kho na). The Madhyamaka method, then, is to cite an antecedent term Φ (or “pervaded term” (khyab bya = vyāpya)) that the opponent accepts and then argue that the opponent has at least implicitly accepted the consequent term Ψ (or “the pervader” (khyab byed = vyāpaka)). In particular, the opponent explicitly accepts that God and other permanent entities act to produce different effects at different times; thus, according to Kamalaśīla, the opponent has also ipso facto implicitly accepted that God is not a single entity because He is divisible into temporal stages correlated with His different actions.

Note that “implicitly” or “by implication” (shugs kyis) is most probably to be understood as the Sanskrit term sāmarthyāt, which plays an important role in Dharmakīrti’s logic. It is clear, in both Dharmakīrti (6th-7th century C.E.) and Kamalaśīla, that this “implication” is not just a purely formal variety; the implied propositions that one accepts if one accepts a set Γ of propositions will not just be the set of logical consequences formally derivable from Γ , but also at least some of those that follow from the meaning of the propositions composing Γ. Tsong kha pa shows that Kamalaśīla’s idea here potentially leads to the problem that one’s acceptance by implication will become absurdly wide-ranging— a Materialist Cārvāka, who only explicitly accepts existence of this life, would also implicitly accept that there are past lives, because (following Buddhist Dharmakīrtian reasonings) their existence is implied by the existence of the present one (see §2).

Ultimately, however, Tsong kha pa does not take up the problem of what “by implication” precisely means. His own solution to the difficulties he sees on this logical issue is the same as his solution to the difficulties he sees in the part-whole arguments of the Madhyamakālaṃkāra: in both cases the solution is that we need to appropriately qualify terms. Instead of speaking of singleness, multiplicity, parts, wholes, etc. simpliciter, or as Tsong kha pa puts it, instead of speaking of them “in a general [unqualified] fashion” (spyir ‘chad) (see §§5, 6), we need to speak of “qualified terms” (khyad par ba) such as “real singleness” (bden pa’i gcig). He concludes the quoted passage: “[[[Kamalaśīla]]] did not mean [that having parts entailed not being single] when he said that [the non-Buddhist opponents] have accepted an antecedent term (khyab bya = vyāpya) [viz., having multiple temporal stages] for non-singleness.” What Mādhyamikas supposedly actually meant was that the terms in the ‘‘neither one nor many’’ reasoning should all be prefixed by bden pa (real) and that real singleness (bden pa’i gcig), as opposed to singleness simpliciter (gcig), implied being partless (cha med). Tsong kha pa goes further and even sees Kamalaśīla talk of implicit acceptance as needless: “Then when one has accepted that [[[God]] and the like] deploy effects having temporal stages, one has explicitly (dngos su) accepted that those things lack real singleness: it is not implicitly (shugs kyis) [accepted].” (see §§11-12). Clearly, he is a very powerful critic of Kamalaśīla here and not just a reformulator or fine-tuner.

The context being given, let’s now backtrack a bit to the quoted passage and seek a better understanding as to why Tsong kha pa’s statement that there is no contradiction in single things themselves (or single things simpliciter) having many parts should seem so surprising. First of all, it is so that many of the finest and philologically grounded twentieth century writers on Madhyamaka philosophy, like Louis de La Vallée Poussin and Jacques May, did think that logically deriving latent contradictions in non-Buddhist opponents’ would-be entities or in the world’s customary truths (saṃvṛtisatya) themselves, typically by reductio ad absurdum (prasaṅga) from the opponent’s premises, was what, in essence, Mādhyamika thinkers like Nāgārjuna, Āryadeva, Candrakīrti and others did. Reductio ad absurdum showed internal inconsistencies in putative things, and so the Madhyamaka method to counter all versions of things was to relentlessly show them to be flawed, i.e., riddled with inconsistencies, or latent internal contradictions. Part-whole arguments played a very significant role in showing precisely that.

Secondly, the common thread in most (but not all) Madhyamaka philosophies, be they traditional or modern, Indian or Tibetan, is usually a Buddhist variant upon global irrealism; it is a fictionalism about everything, or at least everything customary, coupled with an error theory to account for the genesis of the mistaken minds that think and experience some of these fictions to be real. Latent inconsistency of all that is customary is taken to imply that the customary is only appearances (ābhāsa) that are, to use the frequent phrase of Candrakīrti, Kamalaśīla and others, mṛṣāmoṣadharmaka, false and deceptive, with no entities underlying them—although it is invariably claimed that these appearances are not simply dispensable and are needed in worldly transactions (vyavahāra). A large number of the Indo-Tibetan commentators interpret the Madhyamaka idea of emptiness and no intrinsic natures anywhere as implying that a customary thing like a cart is empty of any carts, or in other words a putative cart is not in fact a cart, and hence is only a mistaken appearance of one, or is one only “customarily.” Customary entities are thus said by them to be “empty of themselves” (rang stong), with the corollary also not infrequently being drawn that there simply are no right accounts of the customary, no sources of knowledge (pramāṇa) that grasp customary truths, but only more or less widespread mistaken opinions and beliefs. To use a related Tibetan catchphrase, the would-be cart in which we supposedly ride does not really, or in fact, exist, but only “exists for mistaken minds” (blo ‘khrul ba’i ngor yod pa).

So much, for the moment, about Mādhyamikas’ bringing out the latent inconsistencies in all; we’ll come back to this and to the Buddhist global irrealism that readily results. The extraordinary feature of Śāntarakṣita’s Madhyamakālaṃkāra is that it not only shows inconsistencies in various current Buddhist and non-Buddhist putative entities—as do most other Indian Madhyamaka texts—but shows the very same inconsistency as running through all Buddhist and non-Buddhist versions of things, now or to come. Indeed, while Nāgārjuna (2nd-3rd century C.E), in the chapters of his Mūlamadhyamakakārikā seems to show how going second century versions of things have their respective and different inconsistencies, Śāntarakṣita shows that there is one recurring inconsistency that all putative things have and will have, namely, that x’s having multiple parts and x’s being itself a single thing are contradictory. This along with other premises leads to the further omnipresent contradiction that anything existent would have to be a single thing or a multiplicity and yet can be neither.

Here is what I take to be the most plausible and natural way that the argument from Śāntarakṣita could be fleshed out, a way that Tsong kha pa nevertheless pointedly said could not have been what Śāntarakṣita had in mind. For the moment, we’ll remain uncommitted as to exactly what Śāntarakṣita, his philosophical mentors and disciples themselves had in mind—a more detailed discussion will come later—and give this argument as a clear starting point for analysis.

(a) Suppose that there are things.

(b) All things are either single things or multiplicities;

(c) All things are wholes;

(d) All wholes have multiple parts;

(e) Nothing has multiple parts and is itself single; (f) Nothing is a single thing;

(g) There are no multiplicities (as multiplicities are aggregates of singles);

(h) Nothing is either a single thing or a multiplicity.

(i) If there was something, it would be both either a single thing or a multiplicity and also neither a single thing nor a multiplicity.

(j) There isn’t anything—only at most mistaken appearances of things.

What is surprising in Tsong kha pa is that he turns the tables on this seemingly plausible version of the “neither one nor many” argument and says that wholes having multiple parts does not embody a latent contradiction, precisely because, if it did, an argument like (a)-(j) would be unanswerable: we would prove the absurd conclusion that there would just be nothing at all. Indeed, for Tsong kha pa, (j) is an unacceptable absurdity; as he puts it, “There could not be anything at all” (gang yang mi srid par ‘gyur). He therefore denies the truth of (j), turns the tables on the entire argument and says that one of the earlier steps is thus a false statement. The culprit is (e), for it is not true, for Tsong kha pa at least, that nothing can have multiple parts and be single–the two properties, as the quoted passage makes clear, are instead compatible.

The move is astute. Indeed, turning the tables, or reversing an argument, is a familiar move in East-West argumentation. One may acknowledge that a set of premises entails a conclusion by a seemingly valid argument, refuse to accept that the conclusion is true, and so infer that at least one of those premises is false, or alternatively infer that the argument from true premises to false conclusion is not valid after all. Reversing the argument is, for example, what convinced atheists usually do when faced with mind-numbingly complex theological arguments, or what working scientists do when presented with a theory and observations that together entail very unlikely conclusions. It is what skeptical people routinely do. In Tsong kha pa’s case, he accepted Śāntarakṣita’s argument as validly leading to the unacceptable conclusion that nothing would in fact exist at all and then denied the truth of (e). Another way to put it is that Tsong kha pa’s argument is a typical type of reductio from a counterfactual conditional, one with the protasis “if-clause” needing the English

subjunctive. Indeed, when Tsong kha pa says chos de cha du ma dang bcas pa dang gcig yin pa ‘gal na ...“If it were to be contradictory (‘gal ba = viruddha) for any phenomenon to have multiple parts and to be a single thing” (see §11), the statement is best not translated in the simple indicative, for his point, as we see in the rest of the passage, is that it is not in fact contradictory at all for wholes to be single things and yet have multiple parts. What then are we to make of Tsong kha pa’s own conclusion in the quoted passage, viz., “Therefore, those two [i.e., having multiple parts and being a single thing] are not contradictory, and thus it was not taught [by Śāntarakṣita and Kamalaśīla] that if one believes that God has multiple temporal stages, He would have to be non-single”? So, granted that Tsong kha pa is himself convinced that being partite and single are not contradictory, were they so also for Śāntarakṣita and his disciple Kamalaśīla? Were they contradictory for mainstream Madhyamaka? It is, finally, high time to get clearer on what Śāntarakṣita himself probably had in mind. The issues are exegetically and philosophically tangled. Before we try to show that Tsong kha pa, in a philosophically

interesting, subtle, and important way, imposed an alien thought on Śāntarakṣita, let’s first adopt a slightly backhanded approach asking, “Why would one ever think that Tsong kha pa’s own exegesis of the Indian “neither one nor many” argument was actually right and accounted for what Śāntarakṣita thought?” There are some considerations that might, prima facie at least, seem to be in favor of his exegesis of the mens auctoris. For Tsong kha pa, as we mentioned earlier, the “neither one nor many” argument turns on an appropriately strongly circumscribed, or qualified (khyad par ba) reason that things are not truly/really single (bden pa’i gcig) nor really multiple (bden pa’i du ma) rather than a reason which is just stated in a general, unqualified manner (spyir ‘chad) (see §6) – having multiple parts is thus contradictory with being really single, but not with singleness simpliciter. This introduction of various qualifiers, like bden par (satyatas, “truly”), don dam par (paramārthatas “in fact”, “ultimately”, “really”), rang bzhin gyis (svabhāvena “by its intrinsic nature”), rdzas su (dravyatas “substantially”) and some others, is a well-known feature of Tsong kha pa’s Philosophy of the Middle; he sees it as the way to make sense both of Śāntarakṣita’s one-many argument and Kamalaśīla’s account of implicit acceptance. Significantly too, Śāntarakṣita’s programmatic verse in the Madhyamakālaṃkāra clearly has the word tattvatas (yang dag tu), which is a qualifier in the same semantic circle as satyatas, paramārthatas, svabhāvena, dravyatas, etc. So, such qualifiers are not only important to Tsong kha pa, but also prominent in Indian Madhyamaka, in one way or another. Moreover, whether we are dealing with Śāntarakṣita, Kamalaśīla, Jñānagarbha, or Tsong kha pa there is a consensus that Mādhyamika refutations target “superimpositions” (samāropa, adhyāropa) of truth or reality, but not ordinary, customary things, or appearances. It is sometimes thought that the Prāsaṅgika Mādhyamika Candrakīrti is somehow different on this score. That is not easily supported. The methodology of leaving customary entities, or appearances, untouched while refuting projections of reality, or “superimpositions,” is best seen as just mainstream Indo-Tibetan Madhyamaka.

Do these points of convergence show that Tsong kha pa, with his qualified “neither one nor many” argument targeting superimpositions, simply had Śāntarakṣita dead right and that the argument we laid out in (a)-(j) is indeed both not the actual “neither one nor many” argument and not Śāntarakṣita’s thought? I don’t think so. The key term that remains to unpack is tattvatas and its equivalents, which we had translated as “in fact” or (equivalently) “really.” The mere fact of targeting superimpositions while leaving the customary somehow unscathed is itself common and uncontroversial, but precisely how that superimposition is to be interpreted, what role the qualifier plays, and the status of the customary that remains is where significant divergences may be found. So, let’s look in a bit of detail at what Śāntarakṣita and Kamalaśīla meant by the qualifier. Śāntarakṣita had a heavy debt to the epistemological school of Dharmakīrti et al.—indeed in his major work, the Tattvasaṃgraha, he was certainly a full-fledged member—so that the epistemologists’ use of tattvatas, paramārthatas, etc. is most likely also his. We’ll look at some examples and then try to generalize.

In his Tattvasaṃgraha verse 200, Śāntarakṣita argues against the non-Buddhist Nyāya school’s example of the “furrowed brow of a dancing girl” (narttakībhrūlatābhaṅga), that they use to illustrate their idea of self—the Naiyāyika Uddyotakara had said that just as several people each think they have seen the same girl’s beckoning brow because the cause of their perceptions is a single entity, so too a given person’s various apprehensions of colours, shape, etc., are all recalled as perceptions belonging to one and the same person, because the self that is an underlying condition of these perceptions is a single entity. Śāntarakṣita replies:

“The furrowed brow of a dancing girl is not really (paramārthatas) a single [thing] at all, as it is an aggregate of many atoms. Its singleness is [just] imagined (kalpita).”

Further on, in his chapter on semantic theories (Śabdārthaparīkṣā), Śāntarakṣita takes up the requirement of the grammarians that meanings to univocal words (śabdārtha) be single entities; the universal denoted by a word must be one and the same permanent thing present in a multitude. He accepts this requirement in his own apoha (exclusion) theory of meaning but maintains that the opponent cannot ridicule him as holding the same view as a non-Buddhist: what sets the Buddhist apart from the non-Buddhist is that word meanings and universals (taken as apoha) are not real, but only commonly imagined (kalpita). Tattvasaṃgraha 1200:

“Singleness, permanence and the like are imagined (kalpita), but not real (tāttvika).” To this Kamalaśīla comments in the Tattvasaṃgrahapañjikā: “If we had said that singleness and so forth were real (pāramārthika), then there would have been reason to ridicule us. Since indeed our teacher has said they are fictitious (kālpanika) in keeping with mistaken [common] conceptions (bhrāntipratipatti), how could there result any reason for scholarly ridicule?” Finally, let’s look at Śāntarakṣita’s and Kamalaśīla’s characterization of customary truth (saṃvṛtisatya) in Madhyamakālaṃkara 63 and Madhyamakālaṃkārapañjikā thereupon:

“Thus, such entities bear only saṃvṛti characters.”

Kamalaśīla’s Pañjikā:

“Because minds that are mistaken veil the real status (de kho na nyid) of entities, all these mistaken minds are called saṃvṛti (customary, obscured). Something that exists for that [type of mind], because it exists with a nature that is imagined (brtags pa = kalpita) due to the thought of a mistaken mind, is saṃvṛti.”

The recurring contrast in Śāntarakṣita’s and Kamalaśīla’s works, whether it is with regard to notions of the self, universals, or customary entities, is the contrast between reality and appearance (snang ba = ābhāsa), or in other words, what is in fact/really so, and what is only mistakenly imagined (kalpita) or thought to be so but is not so. In the case of imagined things, they may be in keeping with commonly held conceptions—in which case, they may be treated as customary truths—, but those conceptions are nonetheless bhrāntipratipatti “mistaken conceptions.” What is more, this reality-appearance, or real-imaginary contrast is not just idiosyncratic to Śāntarakṣita: it is a common way of using pāramārthika, tattvatas, etc. amongst other members of the Dharmakīrtian school. Thus, for example, when someone like the eleventh century Kashmiri commentator Manorathanandin characterizes the reason (hetu) in a reductio ad absurdum (prasaṅga) as “not a real reason ” (pāramārthiko hetu)”, he is saying no more, no less, than that it isn’t one because it does not satisfy the criteria– na tv ayaṃ pāramārthiko hetus trairūpyābhāvāt “This, however, is not a real reason, because it lacks the triple characterization.” It is at most thought to be one. When Dharmakīrti’s early commentator Devendrabuddhi (6th-7th century C.E.) grants that scriptural passages are don dam par rtog pa de’i rten ma yin pa “not bases for really (don dam par) understanding,” he means something similar: they are certainly accepted as such by Buddhists and others, but fail to lead to understanding because they don’t in fact lead to inferences (anumāna)—scripturally based inferences for Dharmakīrti and Devendrabuddhi are not inferences in fact, or stricto sensu, because words lack the requisite certainty (niścaya) of a connection with their objects.

I think that the textual data shows that whether we take Śāntarakṣita, most other Indian Mādhyamikas, or Buddhist epistemologists like Dharmakīrti or Devendrabuddhi, “x is in fact Φ” or “x is really Φ”, “x is intrinsically Φ” and other such equivalently qualified formulations can best be taken as expressing a commitment to the truth of “x is Φ.” “Really Φ” or “in fact, Φ”, etc. thus emphasize the speaker’s truth-claim that x having Φ-ness is so, and not just a fiction accepted in a certain context of ignorance, other peoples’ ideas, a going story, error-ridden common opinions, or convenient make-believe and white lies. (A literal versus metaphorical contrast seems also relevant here and indeed is something we find explicit in thinkers like Sthiramati. ) Equally, to say that “x is not in fact/really Φ” usually means that x fails to pass muster to have Φ-ness, that it only seems to have Φ-ness, or is commonly or metaphorically said to have it in keeping with worldly epistemic practices and opinions, but doesn’t. To put things another way, x has Φ-ness and x in fact/really has Φ-ness have the same truth conditions; we have a simple equivalence: x has Φ-ness if and only if x in fact/really has Φ-ness. If x only customarily has Φ-ness, or is imagined having it, or only appears to have it, then x doesn’t have it. We shall call a use of the qualifier that obeys the above equivalence, a “weak use.”

The direct consequence of using qualifiers in this weak way – a way close to how “in fact” or “really” are used in ordinary English discourse—is that it would make no difference whether we formulated the “neither one nor many” reasoning in unqualified terms like (a)-(j) or with the qualifier “in fact/really” figuring on each line; the equivalence enables us to go from one way to the other. (In fact, this weak use of qualifiers is so weak that it collapses, for all intents and purposes, into what we called in Tillemans 2018 an “unqualified Madhyamaka.”) The key mainstream Madhyamaka contrast is then not between “x is a single thing” and “x is in fact/really a single thing”—they are equivalent—, but between “x is a single thing” and “x (only) appears mistakenly to be a single thing”. Louis de La Vallée Poussin’s depiction of Madhyamaka philosophy as showing l’inexistence métaphysique des choses et l’irréalité de l’expérience (see n. 6

above) would not be far off the mark. Equally, the Tibetan rang stong pa, and even those Tibetan Mādhyamikas who saw customary entities as being only “existent for mistaken minds” (blo ‘khrul ba’i ngor yod pa), were also pursuing a quite convincing and similar type of exegesis of India in saying that carts are empty of carts, and hence are not carts, only wrongly appearing to be so. Finally, targeting superimpositions and leaving the customary unscathed means that one refutes commitment to something being so, but does not challenge the fact that it appears to be so to mistaken individuals. All these Madhyamaka philosophies turn on the contrast between what is, or is in fact/really, and what only erroneously appears to be. Mainstream Madhyamaka generally uses qualifiers in the weak way described above, which results in that contrast.

Tsong kha pa, by contrast, has a strong use of qualifiers, which is not the same as the weak and is part of a very different Madhyamaka philosophy, as the equivalence between “x has Φ-ness” and “x really has Φ-ness” does not hold. In particular, the key contrast in Madhyamaka thought for Tsong kha pa is not between commonly accepted mistaken appearances of Φ-ness, on the one hand, and Φ-ness, or equivalently in fact/really Φ-ness, on the other. Rather, it is between Φ-ness (i.e., customary truths), on the one hand, and in fact/really Φ-ness, on the other. As we have done in some earlier publications (Tillemans 2016, chapter 2; 2018), we could highlight Tsong kha pa’s strong use of the qualifier by putting “really” in capitals, i.e., “REALLY”. The point would be that while using “really”, “in fact” and the like in the weak way would collapse into the assertion that such and such a state of affairs is the case, and doesn’t just appear to be so, “REALLY” would be a much stronger assertion, one that does not collapse in that fashion but means that things would have to be what they are independently of all other factors. Mādhyamikas, in short, supposedly know that customary things are so (in a way that does not unpack as just “appearing-to-be-so-to-the-mistaken”), but not REALLY so because they are thoroughly dependent phenomena.

What then is badly wrong, for Tsong kha pa, in saying that customary entities are just commonly accepted mistaken appearances? Although many Tibetan thinkers, like the Jo nang pa and others, held that customary things are all just fictions, or mistaken appearances, that seem to exist to the ignorant, but don’t, for Tsong kha pa global error and a resultant all-encompassing fictionalism are anathema: there are pramāṇas (tshad ma, sources of knowledge) that establish customary existence; it has properties that can be argued about and have a significant objective status, as they are not just seeming-properties only so because people ignorantly think they are. We find the following in his polemic, in Drang nges legs bshad snying po, against the Jo nang pa and other thinkers’ idea of customary truth being only blo ‘khrul ba’i ngor yod pa “existent for mistaken minds”:

“You can’t say, ‘There is no problem in establishing customary existents as being just what exists for the mistakes (‘khrul ba’i ngor yod pa tsam) that have come down to us unanalysed from beginningless [time].’ Suppose that were so. Then the objects as determined (zhen yul = adhyavaseya) when one apprehends things to be permanent, i.e., as being the same previously and subsequently, as well as those apprehended by the innate self-grasping that holds persons and phenomena to be intrinsically established (rang gi mtshan nyid kyis grub pa) would [all absurdly] end up existing customarily (tha snyad du yod par ‘gyur).” And a few lines further he states the linchpin of his own position:

“It is because things that customarily exist have to be established by pramāṇas.”

In short, customary existence of x, or more generally, x having Φ-ness, do not unpack as simply that x mistakenly seems to exist (to all or most of us?) or that x mistakenly seems to have Φ-ness. If they did unpack in this way, there would be no correctives to widespread ideas of customary existence and properties, no possibility of reform, and, indeed, no normativity to truth at all. This critique of no-pramāṇa views is a recurrent theme in the philosophy of Tsong kha pa and his dGe lugs pa followers and is subtle, notwithstanding the fact that certain prominent dGe lugs pa thinkers like ‘Jam dbyangs bzhad pa (1648-1721) go into some rather predictable hyperbole against those who get the point wrong. It is polemically argued that misguided individuals who see the customary as being false conceptual discriminations and thus best transcended would sabotage the Buddhist path and be comparable to the Chinese Heshang (Hvashang) Mohoyen who, in his debate at bSam yas with Kamalaśīla at the end of the eighth century, advocated a Buddhist path consisting in transcendence of discriminations by the abandonment of all conceptual thinking.

Buddhist fire and brimstone polemics aside, ‘Jam dbyangs bzhad pa’s composite picture of no-pramāṇa Madhyamaka, customary truth as mistaken appearances, and the Heshang-like abandonment of conceptual thinking is not just pure calumny. There are serious philosophical problems if one tries to eliminate reasoned deliberation about normative matters such as what people should think there is in favor of just acquiescing in what they do mistakenly think. A viable stance on truth requires more than just talk about what people, de facto, wrongly think is true; and an active pursuit of such truth demands refined conceptual thinking about how customary things are and should be, not just how they seem to be to the benighted. Interestingly enough, Kamalaśīla, to his credit, did recognize the need for pramāṇas concerning the customary and was not satisfied with simply substituting widespread current opinion for truth; Śāntarakṣita probably thought the same. As I have discussed elsewhere (Tillemans 2016, chapter 2), in texts like the Sarvadharma¬niḥsvabhāvatāsiddhi Kamalaśīla did argue for the need for pramāṇas and normativity, and he did so for reasons that were, in the end, not much different from those of Drang nges legs bshad snying po. His problem, however, is essentially the following: if he is indeed the Buddhist global irrealist that I have been making him out to be, it is doubtful that he will be able to meet his own needs. Tsong kha pa was not an advocate of panfictionalism and global error theories and that gives him a better chance at a meaningful pursuit of customary existence and truth.

It is time to draw some corollaries. Given Śāntarakṣita’s weak use of qualifiers that collapse “x is in fact/ really Φ” and “x is Φ”, it is therefore clear that (a)-(j) is indeed Śāntarakṣita’s “neither one nor many” reasoning. Moreover, we also have a response to the question as to whether mainstream Mādhyamikas show latent contradictions in the putative entities themselves. They do. (For a semi-serious illustrative parable of what it is like for there to be latent contradictions in everything, see Appendix II.) For Tsong kha pa, on the other hand, Madhyamaka does not show latent contradictions in the putative entities themselves: it only shows contradictions in the ideas of REAL entities or REAL properties. His Madhyamaka thought is thus quite significantly different from that of most of its Indian forerunners and sources. Finally, it should be said that his own position that there is no contradiction between something having many parts and being nonetheless single makes good philosophical sense—after all, why would one say that there was such a contradiction if one accepted ordinary views as capturing truths and not just errors, white lies or useful make-believe? But such a position is not Śāntarakṣita’s nor probably that of other major Indian Mādhyamikas. It is a position that, I would maintain, is not only subtle but may provide a simple, convincing alternative to Indian Madhyamaka’s tortuous mereology.

More generally, Tsong kha pa has significantly changed the rules of the game: we now have a Madhyamaka philosophy turning on a contrast between two different sets of propositions that each could have genuine truth and justification. The point of Madhyamaka is now to separate out an ordinary type of realism (with its minimally adequate account of truth) from a metaphysics of REALISM, the latter being a seductive and needless trap. Arguably, if one’s Madhyamaka is primarily a means to non-conceptuality and transcendence, that new game, which does stress and legitimize conceptual thought, would not be attractive; one would probably want to stick with the old as it would dismiss reasoned discriminations and indeed all concepts of ordinary things as concerned with false appearances. For a philosopher who needs the tools to make reasoned discriminations about the ordinary as best she can, however, the new game is much more promising than the old.  

Appendix I: An Extract from Tsong kha pa’s dBu ma rgyan gyi zin bris.

§1. Now, in [[[Kamalaśīla’s]]] Madhyamakāloka it was explained that it is fine whether one takes [the neither one nor many argument] as an absurd consequence (thal ‘gyur = prasaṅga) or as an autonomous [[[Wikipedia:reasoning|reasoning]]] (rang rgyud = svatantra). So, let’s explain the faultless way to prove it as an absurd consequence. In that case, the Madhyamakāloka says that although the opponent has not explicitly (dngos su) accepted the logical reason [viz., that the entities in question are not single things], implicitly he has accepted it, and hence we present an explicitly accepted antecedent term (khyab bya = vyāpya, lit. “what is pervaded”) for that reason, for when he has explicitly accepted the antecedent term [e.g., producing effects various successively], then he has implicitly (shugs kyis = sāmarthyāt) accepted the consequent term (khyab byed = vyāpaka, “the pervader”)[viz., not being a single thing].

§2. In that were so [we would reply that] a Cārvāka [[[Wikipedia:Materialist|Materialist]]] would have implicitly accepted past and future lives and omniscience, for given that he has accepted consciousness in this life, that [consciousness’s existence] would imply that it was preceded by a consciousness in a previous life and given that he has accepted that the four elements exist, such existence implies that it is seen by an omniscient being. There are a lot of other sorts [of cases] like those too [where it could be shown absurdly that the opponent would implicitly accept propositions that run directly counter to his professed views]. §3. Suppose it is objected that we do not say the opponents have accepted [the lack of singleness], but it just has been accepted (khas blangs pa tsam mo). §4. [We reply:] That version of things would not be right, for if one referred to one phenomenon, then those words would implicitly refer to all the phenomena that are consequent terms (khyab byed = vyāpaka) implied by it and the mind would implicitly ascertain [those consequent terms]. §5. It might be said that he therefore explicitly accepts a duly qualified (khyad par ba) antecedent term. Let us accordingly present the antecedent term for what he has implicitly accepted, namely, not being really a single thing (bden pa’i gcig min pa),

§6. If we state the antecedent and consequent terms in a general [unqualified] fashion (spyir ‘chad), then the faults would ensue as before [i.e., the opponent would end up accepting all sorts of propositions that he explicitly disavows]. Thus, we do have to accept qualified terms. §7. Now, if we take as an illustration the [[[non-Buddhist]]] conceptions of God (dbang phyug = īśvara), then here is the way in which [[[non-Buddhist]] philosophical thinkers] accept that He is of the nature of a single (gcig = eka) [[[entity]]]. They believe that the very God who existed in the morning exists in full in the afternoon too. So, they do not believe that while certain parts of the God that exists in the morning may exist in the afternoon, other parts are however nonexistent. Rather, they believe that the entirety of parts existing earlier exists later, and that there are no parts nonexistent earlier that are parts of the God that exists later. Therefore, they do not believe that the parts of God are different beings (ngo bo tha dad) from each other. This is the way in which they hold God to be a single [[[entity]]].

§8. This [point] is not the same as [the fact] that, according to our [[[Buddhist]]] tradition, we maintain that a vase is a case of a single [[[entity]]] and believe that it exists in both the morning and the afternoon, but nonetheless do not believe that the two vases [respectively] of the morning and afternoon are one single [[[entity]]], and maintain [instead] that they are cases of multiple [entities].

§9. So much for [[[Kamalaśīla]]] saying that it is not explicitly (dngos su) accepted that God lacks the nature of being a single [[[entity]]]. §10. The antecedent term (khyab bya = vyāpya) by the acceptance of which [the non-Buddhists] have implicitly (shugs kyis) accepted that God is not such a single entity, is that this God would have multiple temporal stages that produce [various] effects such as creating happiness in the morning and misery in the afternoon. Once [those temporal stages] have been accepted, then by implication from that [proposition] they would have accepted that God was not a single [[[entity]]]. When [[[Śāntarakṣita]]] says this, the singleness at stake is a singleness where all (thams cad) the Gods that exist at multiple earlier and late times would not be different beings (ngo bo tha dad), but rather would exist as one being (ngo bo gcig). However, [[[Śāntarakṣita]]] is not talking about anything like the fact that [[[people]] correctly] accept blue to be a single [thing] while it [i.e. blue] produces a blue perception having multiple earlier and later temporal stages.

§11. If it were to be contradictory (‘gal ba = viruddha) for a phenomenon to have multiple parts and to be a single thing, then singleness would be impossible, and if it [i.e., singleness] were to be nonexistent, multiplicity would be nonexistent too. Then, those two being nonexistent, and given that there is no third alternative (phung = rāśi) apart from those two, there could not be anything at all. Therefore, those two [i.e., having multiple parts and being a single thing] are not contradictory, and thus it was not taught [by Śāntarakṣita and Kamalaśīla] that if one believes that God has multiple temporal stages, He would have to be non-single. And [[[Kamalaśīla]]] did not mean this when he said that [the non-Buddhist opponents] have accepted an antecedent term (khyab bya = vyāpya) [implying] non-singleness.

§12. So, suppose that [instead] one construes the point of things lacking real singleness (bden pa’i gcig) in terms of a singleness of something partless (cha med kyi gcig), i.e., that does not have multiple parts. Then when one has accepted that [things] deploy effects having temporal stages, one has explicitly (dngos su) accepted that those things lack real singleness: it is not implicitly (shugs kyis = sāmarthyāt) [accepted]. Via this position, one should also understand just how [opponents] have accepted the other four antecedent terms [implying] that things are not really one.

Tibetan Text of the Extract

A note on the editions consulted. I have looked at the text in the various editions of Tsong kha pa’s Collected Works (gsung ‘bum) on www.tbrc.org. There are no significant differences from the Tashilhunpo edition (“T”) and thus these editions have not been referenced here. On the other hand, in the Sarnath edition of dBu ma rgyan gyi zin bris destined for classes of Tibetan students, the editor has made two important amendments. They are listed below in notes. I think this perspicacious editor was right in making those corrections.

§1. [37. f. 7b] yang dBu ma snang bar thal ‘gyur dang rang rgyud gang du byed kyang byas pas chog par bshad pas thal bar bsgrub pa la skyon med tshul bshad na / dBu ma snang ba las pha rol pos rtags dngos su khas ma blangs kyang / shugs kyis khas blangs pa dang / de’i rgyu mtshan du rtags de’i khyab bya dngos su khas blangs pa bkod de / khyab bya dngos su khas blangs na khyab byed shugs kyis khas blangs pas so // §2. de lta na rGyang pan gyis kyang skye ba snga phyi dang / thams cad mkhyen pa khas blangs par gyur te / ‘di ltar des tshe ‘di’i rig pa khas blangs la de la ni skye ba snga ma’i rig pa sngon du ‘gro bas khyab pa’i phyir dang / ‘byung ba bzhi yod par khas blangs la / de yod pa la thams cad mkhyen pas gzigs pas khyab pa’i phyir ro // de ‘dra ba’i rigs gzhan yang mang ngo //

§3. gal te rgol ba de dag gi khas blangs zhes mi smra’i shugs kyis [38] khas blangs pa tsam mo zhe na/ §4. de lta na mi ‘thad de / chos gcig brjod pa na de la khyab byed du yod pa’i chos thams cad sgra des shugs la brjod pa dang / blo des shugs la nges par ‘gyur ba’i phyir ro //

§5. des na khyab bya khyad par ba ji ‘dra ba zhig dngos su khas blangs pas bden pa’i gcig min pa shugs kyis khas [f. 8a] blangs pa’i khyab bya ston cig ce na / §6. ‘dir khyab bya dang khyab byed spyir ‘chad na skyon snga ma ltar ‘ong bas khyad par ba la ‘dod dgos so // §7. de yang dbang phyug rtog pa la mtshon na / de gcig gi rang bzhin du khas len tshul ni / snga dro’i dus na gang yod pa’i dbang phyug de nyid ma lus pa phyi dro’i dus na’ang yod par ‘dod pas / snga dro’i dus na yod pa’i dbang phyug gi cha ‘ga’ zhig phyi dro’i dus na yod kyang / cha gzhan ‘ga’ med par yang ‘dod pa min gyi / snga dus na yod pa’i cha hril po phyi dus na yod cing / phyi dus na yod pa’i dbang [39] phyug gi cha yin pa’i cha snga dus na med pa med par ‘dod do / des na dbang phyug yin pa’i cha ngo bo phan tshun du tha dad pa mi ‘dod pa ni / dbang phyug gcig tu ‘dod lugs so //

§8. ‘di ni rang lugs kyis bum pa ni gcig gi mtshan gzhir ‘jog la / snga dro dang phyi dro’i dus gnyis su’ang yod par ‘dod kyang / snga dro’i dus kyi bum pa dang phyi dro’i dus kyi bum pa gnyis gcig tu mi ‘dod la / de gnyis du ma’i mtshan gzhir yang ‘jog pa dang mi ‘dra’o // §9. dbang phyug gcig gi rang bzhin dang bral bar dngos su khas ma blangs zhes pa de’o //

§10. gang khas blangs pas dbang phyug de de ‘dra ba’i gcig ma yin par shugs kyis khas blangs pa’i khyab bya / dbang phyug des snga dro bde ba bskyed pa dang / phyi dro sdug bsngal skyed par byed pa sogs kyi ‘bras bu bskyed pa’i rim pa du ma khas [f. 8b] blangs na / de’i shugs kyis dbang phyug gcig min par khas blangs zhes pa gcig ni / [40] dus snga phyi du ma na yod pa’i dbang phyug thams cad ngo bo tha dad min par ngo bo gcig na yod pa’i gcig yin gyi / sngon po gcig tu khas len pa dang / des sngo ‘dzin snga phyi du ma rim pa can du bskyed pa lta bu la de ltar ston pa min no //

§11. chos de cha du ma dang bcas pa dang / gcig yin pa ‘gal na gcig mi srid la de med na du ma yang med cing / de gnyis med na de gnyis las gzhan pa’i phung gsum med pas gang yang mi srid par ‘gyur bas / de gnyis ni mi ‘gal bas dbang phyug la rim pa du ma yod par ‘dod pa la gcig ma yin par ‘gyur zhes bstan pa yang min la / gcig min pa’i khyab bya khas blangs zhes pa yang min no // §12. des na bden pa’i gcig yin pa dang bral ba’i don cha du ma dang bcas pa min pa’i cha med kyi gcig la byed na / ‘bras bu rim can la nye bar sbyor bar khas blangs pa ni / bden pa’i gcig bral du dngos su khas blangs pa yin gyi [41] shugs kyis min no // phyogs ‘dis ni bden pa’i gcig min pa’i khyab bya bzhi gzhan la khas blangs tshul yang shes par bya’o //  

Appendix II. Latent Contradictions in all Things: a Borgesian Parable

There is an interesting rumour—thoroughly unsubstantiated, alas— that Jorge Luis Borges told a labyrinthine tale showing how all things would be internally contradictory and thus false appearances. Here is a fragment of what he supposedly said to the assembled literati of Buenos Aires: “For a reason that I can not explain here, we all–or perhaps, only most of us— imagine there is a mysterious village barber: he shaves all and only the men in his village that do not shave themselves. Perhaps we learned that in school; perhaps we have been reading too much Bertrand Russell; or perhaps, as Buddhists say, we have had innate tendencies to think so since beginningless time. No matter. The point is that while people imagine, or even see that such barbers are good or bad, cheap or expensive, identical with or different from their bodies, even momentary or enduring, still there aren’t any, or aren’t any in fact, and they don’t have the properties in question, because if there were to be such a barber, he would shave himself if and only if he didn’t.

“Indeed, fortunately perhaps, the village barber is an impossible barber and so doesn’t exist. Likewise for all the other customary things whose metaphysical status so troubles us, because they are riddled with their own various contradictory features, like part-whole contradictions, and are thus empty of themselves. No need to be troubled: they too are just appearances, fictions, errors, things that are mṛṣāmoṣadharmaka (false and deceptive), as they say in Sanskrit. Worldlings, unfortunately, don’t understand such non-obvious contradictions. Somos todos el barbero.” The erudite Argentine apparently returned to the Café Tortoni and said no more on the subject. He was more obsessed with tortoises and their avatars than barbers. He was, though, arguably right in saying that everything from ideas of God to those of carts, atoms, minds, and people, if latently contradictory, would be deeply barber-like.  

References

Candrakīrti. Prasannapadā Madhyamakavṛtti. Ed. L. de La Vallée Poussin. Mūlamadhyamakakārikās (Mādhyamikasūtras) de Nāgārjuna, avec le commentaire de Candrakīrti. St. Petersburg: Bibliotheca Buddhica IV, 1903-1913. Reprinted Osnabrück: Biblio Verlag, 1970. Demiéville, Paul. 1987. Le Concile de Lhasa: une controverse sur le quiétisme entre bouddhistes de l’Inde et de la Chine au VIIIe siècle de l’ère chrétienne. Paris: Collège de France, Institut des Hautes Études Chinoises. First published 1952.

Dharmakīrti. Pramāṇavārttikasvavṛtti. Edited by Raniero Gnoli, The Pramāṇavārttikam of Dharmakīrti: The First Chapter with the Autocommentary. Serie Orientale Roma 23. Rome: Istituto Italiano per il Medio ed Estremo Oriente, 1960. Ganeri, Jonardon. 2011. “Subjectivity, Selfhood and the Use of the Word ‘I’.” In Mark Siderits, Evan Thompson, and Dan Zahavi (eds). Self, No Self? Perspectives from Analytical, Phenomenological and Indian Traditions. New York: Oxford Publishing, 2011: 176-192.

Higgins, David and Martina Draszczyk. 2016. Mahāmudrā and the Middle Way. Post-classical Kagyü Discourses on Mind, Emptiness and Buddha-nature. Wiener Studien zur Tibetologie und Buddhismuskunde 90.1 and 90.2. Vienna: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien. ‘Ju Mipham = ‘Jam dbyangs phyogs las rnam rgyal rgya mtsho. 2004. Speech of Delight. Mipham’s Commentary on Śāntarakṣita’s Ornament of the Middle Way. Translated by Thomas H. Doctor. Ithaca N.Y.: Snow Lion Publications. ‘Ju Mi pham. gZhan stong khas len seng ge’i nga ro. Vol 17 in his Collected Works, i.e., Mi pham gsung ‘bum las gzhung ‘grel skor 33 volumes, Chengdu, 2008. Accessible at www.tbrc.org. Kamalaśīla. Madhyamakālaṃkārapañjikā. Ed. Ichigō 1985. Kamalaśīla. Madhyamakāloka. Degé Tibetan edition 3887 (see www.tbrc.org). Partly translated and edited in Keira 2004. Kamalaśīla. Tattvasaṃgrahapañjikā. Sanskrit edition in Krishnamacharya 1926. Keira, Ryusei 2004. Madhyamaka and Epistemology. A Study of Kamalaśīla’s Method for Proving the Voidness of all Dharmas. Wiener Studien zur Tibetologie und Buddhismuskunde 59. Vienna: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien.

Krishnamacharya, Embar. 1926. Tattvasaṅgraha of Śāntarakṣita: With the Commentary of Kamalaśīla. Ed. with an Introduction in Sanskrit. GOS, Gaekwad’s Oriental Series 30–31. Baroda: Central Library, 1926. Ichigō, Masamichi. 1985. Madhyamakālaṃkāra of Śāntarakṣita. With his own Commentary or Vṛtti and with the Subcommentary or Pañjikā of Kamalaśīla Critically Edited by Masamichi Ichigō. Kyoto Sangyo University. Kyoto: Bun’eido. ‘Jam dbyangs bzhad pa’i rdo rje Ngag dbang brtson ‘grus. Grub mtha’ = Grub pa’i mtha’i rnam bshad kun bzang zhing gyi nyi ma. Lanzhou, Gansu: Kan su’u mi rigs dpe skrun khang, 1994. Accessible at www.tbrc.org.

La Vallée Poussin, Louis de 1933. “Réflexions sur le Madhyamaka.” Mélanges chinois et bouddhiques 2 : 1-59. May, Jacques. 1979. “Chūgan.” P. 470-493 in Hōbōgirin, Dictionnaire encyclopédique du bouddhisme d’après les sources chinoises et japonaises V, Tokyo : Maison franco-japonaise. Paris : Librairie d’Amérique et d’Orient. Pad ma dkar po. Phyag chen rgyal ba’i gan mdzod = Phyag rgya chen po’i man ngag gi bshad sbyar rgyal ba’i gan mdzod. In Padma dkar po gsung ‘bum, vol. 21, 7-370. Darjeeling: Kargyud sungrab nyamso khang, 1973-1974. Accessible at www.tbrc.org. Prajñākaramati. Bodhicaryāvatārapañjikā. Ed. P.L. Vaidya, Śāntideva, Bodhicaryāvatāra, With the Commentary Pañjikā of Prajñākaramati. Buddhist Sanskrit Texts. Darbhanga, India, 1960. Russell, Bertrand 1968. Logic and Knowledge. Essays 1901-1950. Ed. by Robert C. Marsh. Fourth edition. London: George Allen and Unwin, 1968. S. = Sarnath edition.

Śāntarakṣita. Madhyamakālaṃkāra and Madhyamakālaṃkāravṛtti. Ed. Ichigō 1985. Śāntarakṣita. Tattvasaṃgraha. Ed. Krishnamacharya 1926. T. = Tashilhunpo (bKra bshis lhun po) edition of Collected Works of Tsong kha pa, or Khams gsum chos kyi rgyal po Tsong kha pa chen po’i gsung ‘bum. Geden sungrab minyam gyunphel series 79-105. Published by Ngag dbang dge legs bde mo. Delhi, 1975-1979. Tillemans, Tom J.F. 1984. “Two Tibetan Texts on the ‘Neither One nor Many’ Argument for Śūnyatā.” Journal of Indian Philosophy 12: 357-388. Tillemans, Tom J.F. 1992. “Note liminaire.” In Études bouddhiques offertes à Jacques May, ed. Johannes Bronkhorst, Katsumi Mimaki and Tom Tillemans. Asiatische Studien / Études asiatiques 46.1: 9-12.

Tillemans, Tom J.F. 1993. Persons of Authority. The sTon pa tshad ma’i skyes bur sgrub pa’i gtam of A lag sha Ngag dbang bstan dar. A Tibetan Work on the Central Religious Questions in Buddhist Epistemology. Tibetan and Indo-Tibetan Studies 5. Stuttgart: Franz Steiner Verlag. Tillemans, Tom J.F. 1999. Scripture, Logic, Language. Essays on Dharmakīrti and his Tibetan Successors. Studies in Indian and Tibetan Buddhism. Somerville, MA: Wisdom Publications.

Tillemans, Tom J.F. 2000. Dharmakīrti’s Pramāṇavārttika. An Annotated Translation of the Fourth Chapter (parārthānumāna). Volume 1 (k. 1-148). Vienna: Verlag der Österreichischen Akademie der Wissenschaften. Tillemans, Tom J.F. 2003. “Metaphysics for Mādhyamikas.” In Georges Dreyfus and Sara McClintock (eds.). The Svātantrika-Prāsaṅgika Distinction: What Difference Does a Difference Make? Studies in Indian and Tibetan Buddhism. Somerville, MA: Wisdom Publications: 93-123. Tillemans, Tom J.F. 2016. How do Mādhyamikas Think? And Other Essays on the Buddhist Philosophy of the Middle. Studies in Indian and Tibetan Buddhism. Somerville, MA: Wisdom Publications.

Tillemans, Tom J.F. 2018. “Metaphysics and Metametaphysics with Buddhism: The Lay of the Land.” In Steven M. Emmanuel (ed.), Buddhist Philosophy. A Comparative Approach. Hoboken NJ: Wiley Blackwell: 87-107.

Tsong kha pa. dBu ma rgyan gyi zin bris. T. volume ba. S. = dBu ma rgyan gyi rtsa ba dang / dbu ma rgyan gyi zin bris / dbu ma rgyan gyi brjed byang bcas. Sarnath U.P. India: Pleasure of Elegant Sayings Printing Press, 1976. Tsong kha pa. Drang nges legs bshad snying po = Drang ba dang nges pa’i don ram par phye ba’i bstan bcos legs bshad snying po. T. volume pha. S. = Dan-ne-leg-shed-nying-po. Sarnath U.P. India: Pleasure of Elegant Sayings Printing Press, 1973.