To view the diacritics on this page, you must install the Indic Times font on your machine and have a browser capable of displaying the Unicode (utf-8) character set.
Reflections on The Reflexive Nature of Awareness: A Tibetan Madhyamaka Defence. By Paul Williams. Surrey, England: Curzon Press, 1998, xix + 268 pp, ISBN: 0–7007–1030–2, $55.00.
Reviewed By Matthew T. Kapstein
Associate Professor,
Department of South Asian Languages and Civilizations and the Divinity School
The University of Chicago
The present review article discusses aspects of Paul Williams’s excellent and highly recommended book, which focuses on the question of “reflexive awareness” (Tib. rang rig, Skt. svasaṃvittiḥ, svasaṃvedana) in Tibetan Mādhyamika thought. In particular, I am concerned with his characterization of so so rang rig ye shes and its relation to Rdzogs-chen teaching, and his notions of the gzhan stong doctrine and its place in the intellectual life of Far-eastern Tibet. My critical remarks on these topics are in many respects tentative, and I would welcome correspondence about them.
The Reflexive Nature of Awareness is a companion to Williams 1998b, and continues the author’s learned and perceptive investigation of selected arguments from the ninth chapter of Śāntideva’s Bodhicaryāvatāra (BCA) and their Indian and Tibetan commentaries. In particular, Williams is concerned here with the refutation of “reflexive awareness” (Skt. svasaṃvedana, svasaṃvittiḥ, Tib. rang rig) in BCA 9.20–26 (using Vaidya’s numbering) and the debate about this that was generated in Tibet by the Rnying-ma-pa master Mi-pham rnam-rgyal rgya-mtsho (1846–1912). In his commentary on BCA 9, Mi-pham had argued—pace Rje Tsong-kha-pa (1357–1419) and his Dge-lugs-pa successors—that Śāntideva’s refutation was intended only with reference to ultimate truth (paramārthasatya, don dam bden pa) and did not preclude recourse to the concept of reflexivity in relative terms, even on the part of an adherent of Prāsaṅgika-Mādhyamika. In so arguing, he was contradicting the Dge-lugs-pa tenet that one of the eight special features of the Prāsaṅgika system is its critique of the concept of reflexive awareness in both ultimate and relative terms. (This is very clearly argued, for instance, by Tsong-kha-pa’s disciple Rgyal-tshab-rje Dar-ma rin-chen [1364–1432] in Rgyal-tshab 1985.)
In guiding the reader through the maze of conceptual and dialectical difficulties this material presents, Williams exhibits the same strengths that inform the companion volume: a determination to unpack philosophical arguments thoroughly and with great care and a keen sense that what is at issue in philosophical dispute is best exhibited by exploring the confrontation among a variety of opposing viewpoints, rather than just setting forth the doctrines propounded by a single author or school. His approach does much to enliven the study of Tibetan Buddhist philosophy and to engage the reader in the intellectual dynamic of the tradition. Williams has been interested in the questions surrounding the treatment of reflexivity in Tibetan thought for a long time now, and his first article on the subject (Williams 1983a) is usefully reprinted here as an appendix (pp. 232–246).
In the body of the book, Williams builds his account of the dispute brick-by-brick, beginning (ch. 1) with an introduction to the concept of svasaṃvedana and what Williams considers to be its two main types: self-awareness (i) appears to take an object, and is reflexive in the sense that the apparent object is a phenomenal feature of the act of awareness itself; self-awareness (ii) is proper reflexivity, awareness’s awareness of itself as awareness. The first is a concept stemming from Cittamātra epistemology, while the latter relates primarily to the question of determining the defining characteristic of consciousness, and is for all intents and purposes no different from the property of “luminosity.” The manner in which the eighth-century Indian philosopher Śāntarakṣita developed and deployed this notion of reflexivity is the subject of chapter two. In chapter three, Williams turns to the BCA itself, particularly to examine the commentator Prajñākaramati’s intentions in citing Śāntarakṣita’s discussion of reflexivity. He shows that, in the passage in question (the commentary on BCA 9.21, or 9.20 in Vaidya’s numbering), Prajñākaramati takes Śāntarakṣita to exemplify the position that Śāntideva is opposing, and argues further that the commentator is concerned to refute svasaṃvedana ultimately, leaving the world and Sanskrit grammar to legislate convention. Williams rightly suggests (p. 44) that Śāntarakṣita in fact does not wish to affirm svasaṃvedana ultimately and that Prajñākaramati uses Śāntarakṣita’s text just to illustrate the position that Śāntideva refutes. Like Williams, I do not find evidence in Śāntarakṣita’s own writing that he adhered to that position ultimately; his affirmations of Cittamātratā, considered in relation to his Mādhyamika writings, must be understood as representing only his view of relative reality. This is perhaps one of the reasons (and here I am speculating) that Tsong-kha-pa and his followers thought that Śāntideva must have been refuting svasaṃvedana both ultimately and relatively at this point. For why otherwise, they may have wondered, was Śāntarakṣita cited in the commentary as exemplifying the pūrvapakṣa? In chapter four, Williams provides a very thorough survey of the commentarial tradition on BCA 9.26, the closing verse in Śāntideva’s critique of svasaṃvedana. Williams convincingly demonstrates, I think, that the pre-Dge-lugs-pa commentators, both Indian and Tibetan, were unanimous in their view that Śāntideva’s refutation was addressed to the ultimate level, and did not pertain to conventional reality at all. The chapter very well illustrates the merits of Williams’s broad consideration of commentarial writings over and against the common tendency to treat a single author or school; for without the perspective supplied by Williams, we would have in this case no way to assess just how innovative Tsong-kha-pa and his followers really were in their approach to the interpretation of Prāsaṅgika-Mādhyamika.
All of this, in a sense, is a preamble that provides the reader with the background necessary for a thorough consideration of Mi-pham’s treatment of these topics and the response that this elicited from his critics. In the fifth chapter, Williams discusses Mi-pham’s arguments as presented in his famous commentary on BCA chapter nine, the Sher ṭīk nor bu ke ta ka. Mi-pham argues, in essence, that without accepting reflexivity in conventional terms, our conventional knowledge of our own mental states becomes inexplicable. In this way, the denial of svasaṃvedana in relative reality leads to a variety of absurd conclusions, eventually undermining our knowledge of all referents; for, if we do not in some sense know our own mental states, what knowledge can we have of their contents?
As Williams rightly argues (p. 107), Mi-pham’s work demonstrates his intimate familiarity with Dge-lugs-pa approaches to Madhyamaka thought, but at the same time markedly differs from them. His work seemed to invite Dge-lugs-pa response, and this, indeed, was forthcoming. In chapter six, Williams considers one of Mi-pham’s main Dge-lugs-pa critics, Tre bo brag dkar sprul sku Blo bzang dpal ldan bstan ’dzin. (Williams mistranscribes “Blo” as “bLo”—though the pronunciation is roughly “lo,” the ming gzhi in this case is b-.) Williams’s discussion of Tre bo brag dkar sprul sku’s work is based entirely on Mi-pham’s rebuttal, the Brgal lan nyin byed snang ba, and he is apparently unaware that Tre bo brag dkar sprul sku’s own writings are available in the collection of the Oslo University library (Kvaerne 1973). Nevertheless, Mi-pham supplies extensive verbatim citations from his opponent’s work, and these are probably adequate for Williams’s purposes. One is inclined to concur with Williams that Tre bo brag dkar sprul sku’s “attack on Mi-pham…when all is said and done seems rather disappointing” (p. 109). His most intriguing point, discussed by Williams at length (pp. 110–116), is that, were one to follow Mi-pham in accepting svasaṃvedana conventionally, it would be impossible to overturn the entailment that conventionally (tha snyad du) the three constituents of an act of consciousness—agent, object, and act—would have to be present when svasaṃvedana occurs. This is my own paraphrase of Tre bo brag dkar sprul sku’s argument, and it differs slightly in emphasis from Williams’s, though I very much agree with him that the argument is to some degree unclear. Though I find Williams’s reflections on the argument to be in most respects illuminating, I cannot concur that Tre bo brag dkar sprul sku is tacitly arguing that “if svasaṃvedana existed conventionally it would also have to exist ultimately” (p. 115). So far as I can tell, he is only saying that if it exists at all—even in only conventional terms—then it must fulfill the definition of an act of consciousness. This, I think, at least helps us to understand Mi-pham’s response, to which Williams turns in chapter seven, far and away the longest in the book (pp. 119–182). As Williams shows (pp. 126–140), Mi-pham holds against his opponent that “[t]he activity-agent-action model cannot be applied in the case of a partless unity like reflexive awareness, the very quality of consciousness itself” (p. 132).
Chapter eight is entitled “Why all the fuss?” The significance of the questions that occupied traditional Tibetan Buddhist philosophers, like that of those that exercise contemporary anglophone philosophers, is not usually self-evident except to those involved in the discourse communities concerned. Too often those of us writing on Buddhism do not seem to recall this fact, and it is one of Williams’s merits that his perspective is not so self-enclosed. He adduces four main reasons for which Mi-pham’s Dge-lugs-pa interlocutors took their stand against him on the question of the conventional reality of reflexive awareness: (i) its conventional existence might imply its inherent existence, which would indeed be anathema to the Prāsaṅgika view (pp. 186–188); (ii) the affirmation of reflexive awareness may be closely tied to positions congenial to Cittamātra (pp. 188–193); (iii) the affirmation of reflexive awareness may tacitly lend support to some varieties of the gzhan stong view, as favored among some teachers associated with the so-called “ris-med movement” in nineteenth-century Eastern Tibet (pp. 193–205); and (iv) which concerns the implications of reflexive awareness for the interpretation of “a Buddha’s direct nondual and nonconceptual omniscient awareness” (pp. 206–214). Though I concur with Williams that these are all important and interesting issues in this context, it will emerge in the following pages that I differ with aspects of Williams’s approach to the last two.
Despite my high praise for most of Williams’s book, there are a number of important points about which I must express some rather strong reservations. Williams asserts on several occasions (and also repeats in 1998b) that Mi-pham’s position on svasaṃvedana was motivated by his commitment to a particular concept of the Rdzogs-chen system, which Williams calls so so rang rig ye shes tsam, and for which he gives as a Sanskrit equivalent the term pratisvasaṃ vedanajñānamātra, translated as “a mere reflexive gnosis.” So far as I can determine, he offers us no evidence whatsoever that there is in fact such a Rdzogs-chen term or that there is any such Sanskrit term as the one that he provides. In fact, his entire basis—so far as I can determine—for positing such a term at all is a single occurrence in Mi-pham’s commentary, an occurrence that Williams has certainly misinterpreted.
Let us begin by considering Mi-pham’s text, which is found in his commentary on BCA 9.35:
de ltar na gang gi tshe dngos po dang dngos po med pa dag gang yang blo yi mdun na mi gnas pa de’i tshe/ de las gzhan bden par grub pa’i rnam pa gzhan med pas na bden ’dzin gyi dmigs pa’i gtad so mtha’ dag med par spros pa ma lus pa rab tu zhi ba yin te so so rang rig pa’i ye shes tsam gyis rab tu phye ba smra bsam brjod du med pa nam mkha’i dkyil lta bu’i mnyam pa nyid do// (Mi-pham 1994: 39).
“In that way, at which time neither entity nor nonentity abide before the intellect at all, at that time, because there is no other veridical feature, all elaborations without exception are pacified, without there being any objectified intentions involving veridical apprehensions whatever. Being disclosed by only so so rang rig pa’i ye shes, this is an equanimity that is like the sphere of space—ineffable, inconceivable, and unutterable.”
It will be immediately apparent that I disagree with Williams regarding the force of the particle tsam here. Whereas he interprets it as integral part of the compound, with the adjectival meaning “mere,” I take it to be an adverbial particle of limitation or exclusion. In its primary significance, of course, tsam is a particle signifying approximate quantity, but by extension it may be used in the senses that Williams and I suggest. My reason for doubting Williams’s interpretation in this case is just that there is no regular usage of which I am aware of so so rang rig pa’i ye shes tsam as a well-formed compound in the Rdzogs-chen tradition, though Mi-pham and other writers do use so so rang rig pa’i ye shes quite frequently. Consequently there seems to me to be no basis for adopting Williams’s understanding on the basis of a single instance. We should note, however, that the precise force of tsam in philosophical contexts does sometimes cause confusion even among learned Tibetan readers. In fact, Mi-pham chides his greatest opponent, Dpa’-ris Rab-gsal, on one occasion for just this reason. Perhaps his words are appropriately addressed also to Williams:
kho na dang tsam sogs kyi sgra ’di dag brjod ’dod dang sbyar tshul gyi dbang gis mi ldan rnam gcod dang gzhan ldan rnam gcod sogs kyi gnas skabs so sor go rgyu yod pas/ phyogs re’i u tshugs ’dis ci bya ste/ gtso bor bstan pa tsam yin zhes pas gtso bo ma yin pa gcod de sems tsam zhes pas sems las gzhan pa sems ma yin pa gcod pa bzhin no// gal chung la nan tan brtags pa khyed ni tsam sgra ’chad pa’i slob dpon tsam ni yin par mngon no// (Mipham 1994: 140–1).
“These words kho na and tsam, etc., according to the intention with which they are uttered and the manner of composition are to be understood contextually as excluding that which does not possess [the property in question], or excluding that which possesses another [property that is not in question], etc. What is to be gained by this extreme partiality [of interpretation that you have expressed]? When I said, ‘it is teaching just (tsam) what is foremost,’ it was an exclusion of what is not foremost, just as “mind only” excludes the nonmental, which is other than mind. You, who engage in forced examinations to little purpose, are clearly a mere (tsam) master of the explanation of the word ‘mere’!”
What it is most important for us to establish here, however, is just what Mi-pham intends when he introduces the expression so so rang rig pa’i ye shes in the passage under discussion. The verse upon which he is commenting in this instance, BCA 9.35, is famously regarded as a quintessential expression of Śāntideva’s realization of the Mādhyamika teaching, so it seems most unlikely that Mi-pham would have casually inserted here an allusion to a doctrine regarded as alien to Mādhyamika thought. If he is being controversial at this point, as Williams takes him to be, then one might have expected his Dge-lugs-pa opponents to criticize him for this, but so far as I can determine, his reference to so so rang rig pa’i ye shes was not thought by anyone to be an objectionable point.
In fact, it will be at once evident to many readers of the passage given above that Mi-pham, far from introducing an exotic Rdzogs-chen term into his commentary, has simply given us a paraphrase of one of the most famous of Tibetan verses, which is found in the liturgies of all the Tibetan Buddhist orders and is usually attributed to Rahulabhadra’s Prajñāpāramitāstrotra (Sher phyin bstod pa):
smra bsam brjod med shes rab pha rol phyin// ma skyes mi ’gag nam mkha’i ngo bo nyid// so so rang rig ye shes spyod yul pa// dus gsum rgyal ba’i yum la phyag ’tshal lo//
“I bow to the mother of the Jinas of the three times, Perfection of Wisdom, who is ineffable, inconceivable, unutterable. Unborn, unceasing, she is of the nature of space, And in the scope of so so rang rig ye shes. ”
I believe that I would not be far wrong in holding that all traditionally educated Tibetan Buddhists, regardless of sectarian affiliation, know this verse by heart and that none of Mi-pham’s readers, whether Rnying-ma-pa or Dge-lugs-pa, would have been inclined to see this allusion in any way as suggestive of a peculiarly Rdzogs-chen affirmation. Nevertheless, we still must enquire into just what so so rang rig ye shes might mean here, and certainly also countenance the possibility that it is a term that partisans of differing schools understand quite differently.
As mentioned above, Williams gives pratisvasaṃvedanajñānamātra as the Sanskrit term underlying so so rang rig ye shes tsam. I have dispensed already with the final element -mātra, so now what about pratisvasaṃvedanajñāna? Though Williams cites this expression on several occasions in the present book (e.g., on pp. xi, 185, 196–7; cf. also 119, 199) as well as in Williams 1998b (p. 24) so that the reader may come to accept the authority of this usage (as does Pettit 1999a, for example), Williams in fact does not provide a single citation from a Sanskrit text in justification of it. Indeed, he could not, for the term in question does not exist.
Fortunately, however, we do know just what the underlying Sanskrit is in this case. Rje Tsong-kha-pa, whose teaching Williams supposes to be at odds with Mi-pham’s positive reference to so so rang rig ye shes, quotes the third Bhāvanākrama of Kamalaśīla approvingly as follows:
de ltar gang dang gang du bsam gyis mi khyab pa la sogs pa’i tshig thos na/ de dang der thos pa dang sems pa tsam kho nas de kho na rtogs par gang dag sems pa de dag gi mngon pa’i nga rgyal dgag pa’i phyir/ chos rnams so so rang gis rig par bya ba nyid du ston par byed do/ / (Tsong-kha-pa 1985: 793).
“Thus, wherever one hears such expressions as ‘inconceivable,’ some think that one is to realize just what is just by merely hearing and thinking on those [expressions]; but in order to negate their arrogance the dharmas are taught to be so so rang gis rig par bya ba nyid.”
In Tucci’s edition of the Sanskrit, this reads:
tad evaṃ yatra yatrācintyādiprapañcaḥ śrūyate, tatra tatra śrutacintāmātreṇaiva tattvādhigamaṃ ye manyate, teṣām abhimānapratiṣedhena pratyātma vedanīyatvaṃ dharmāṇāṃ pratipādyate/ (Tucci 1971: 19).
“Thus, wherever elaborations such as ‘inconceivable’ are heard, [there are] those who think [that there may be] realization of reality just by merely hearing and thinking on those; as a negation of their arrogance the pratyātma vedanīyatvaṃ of dharmas is set forth.”
In Tibetan translations of Sanskrit texts for which we have the originals, so so rang rig is in fact the standard rendition of pratyātma-vid and its derivatives. Clearly this was a case in which Williams, not having located the actual Sanskrit, should have clearly marked his term as a hypothetical reconstruction. (Let me remark in passing that the indiscriminate use of calque translations from Tibetan into Sanskrit is a significant methodological problem that has long infected Tibetan Buddhist Studies.) But now we must ask just what pratyātma-vid means. Is it a close synonym of svasaṃvedana, in which case Williams’s mistaken reconstruction would be a matter primarily of philological interest, or does it refer to a very different concept, in which case a major issue of interpretation is involved here as well?
Because the term is well-known to occur in texts such as the Ratnagotravibhāga (for example, ch. 1, v. 9b in Johnston 1950), that are often associated with the Indian antecedents of Tibetan gzhan stong thought, it may be urged that, although the concept is by no means peculiar to the gzhan stong or Rdzogs-chen traditions, it is nevertheless closely tied to approaches to Buddhist teaching that some would characterize as affirming some sort of idealism or a substantial absolute. Tsong-kha-pa’s favorable citation of Kamalaśīla, however, counsels caution on this point. Indeed, there is very good reason to hold that pratyātma-vid has no special relationship in Indian Buddhism with Cittamātra and that the concept in question belongs even to very early Buddhism. In Majjhima Nikāya I 265 (PTS ed.), for instance, we read:
Upanītā kho me tumhe bhikkhave iminā sandiṭṭhikena dhammena akālikena ehipassikena opanayikena paccattaṃ veditabbena viññūhi.
“Monks! you have been guided by me by means of this visibly true dhamma, that is timeless, ostensible, conducive [to the goal], and to be intuited individually by the wise (paccattaṃ veditabbena viññūhi).”
Paccattaṃ veditabba is precisely equivalent to Sanskrit pratyātmaveditavya (or -vedanīya) and to Tibetan so so rang gis rig par bya ba. In all cases, it means only that the adept’s realization is intuitive, a discovery that in the final analysis she must make by and for herself. The Buddha, as it is elsewhere said, can neither wash away our taints with water, nor pull us by the hand to nirvāṇa. The term, therefore, in its original and primary signification has nothing whatever to do with epistemological theories of reflexive awareness, or with substantialist metaphysical accounts of the mind, or with gzhan stong, or with Rdzogs-chen. It may well be that certain later traditions of Buddhist philosophy and meditation appropriated the term, but they probably did so in large measure owing to its ancient resonances and not in the first instance due to any doctrinal novelty. Moreover, as the citation above of Rahulabhadra’s stotra demonstrates, the addition of the term ye shes/jñāna was by no means a Tibetan innovation, and need not be taken as much altering the basic sense of the term.
Though I certainly concur with Williams that the assessment of rang rig (=svasaṃvedana) as a type of reflexivity that may or may not be affirmed to exist relatively is a point of contention between the Dge-lugs-pa interpreters of Prāsaṇgika Mādhyamika and certain of their opponents, recourse to the canonical concept of enlightenment as so so rang gis rig par bya ba (=pratyātma-vedanīya) in itself is not. But this, of course, is not to say that all understood this concept in just the same way. We must ask, then, just what Tsong-kha-pa intends through his employment of the term in the Lam rim chen mo. The context in which the quotation from Kamalaśīla given above is found is the close of Tsong-kha-pa’s discussion of vipaśyana, contemplative insight, where he takes up objections to his account (Tsong-kha-pa 1985: 788–795). His primary concern in these passages is to refute a purely quietistic approach, which holds that the analytical comprehension of selflessless that Tsong-kha-pa champions must oppose the dawning of nonconceptual gnosis (bdag med pa’i don la so sor dpyod pa rtog pa yin pas de las rnam par mi rtog pa’i ye shes skye ba ’gal). For Tsong-kha-pa, so so rang rig ye shes is a valuable concept precisely because it underscores that “nonconceptual gnosis” is not properly conceived as a mere absence of conception; it must be preceded by a certain sort of conceptual activity, and is positively realized by those who have become contemplative virtuosi (ārya, ’phags-pa). As he affirms: “[I]nconceivability by others and the like are taught because those are to be intuitively realized among the virtuosi” (de rnams ’phags pa’i so so rang gis rig par bya ba yin pas gzhan gyis bsam gyis mi khyab pa la sogs par ston).
Interestingly, Tsong-kha-pa’s discussion in the Lam rim (with which Mi-pham was undoubtedly familiar) may help to explain why Mi-pham chose to paraphrase Rahulabhadra just at BCA 9.35; for there Śāntideva states that neither being nor non-being is apprehended. Tsong-kha-pa had invoked so so rang rig precisely to quell the misapprehension of negative predications of the absolute, such as “non-being,” as underwriting an extreme type of quietism. It seems clear, now, why it was that Mi-pham’s Dge-lugs-pa opponents did not seek to challenge him regarding this particular point.
Despite all of this, it is evident that the Tibetan terms rang rig and so so rang rig (ye shes) do resemble one another very closely, so that we cannot rule out the possibility that they may have been conflated by some. Indeed, the noted Dge-lugs-pa scholiast Se-ra rje-btsun Chos-kyi rgyal-mtshan (1469–1546) maintains that Karma-pa VIII Mi-bskyod-rdo-rje (1507–54) and the Sa-skya-pa thinker Gser-mdog Paṇ-chen Shākya-mchog-ldan (1428–1507) have done just that. He writes:
’di la rje karma pa dang chos rje gser mdog can pas rnam par mi rtog pa so sor rang gis rig par bya ba ci zhig ltar yang dag par skyes na/ zhes pas gnyis med kyi ye shes gnas lugs mthar thug tu bstan pa yin te/ lung ’dis gnas lugs mthar thug de mnyam gzhag so sor rang rig pa’i myong byar bshad pa’i phyir zhes gsungs pa ’brel yod par ye ma go ste/ rnam gnyis kyi thugs bzhed la/ rang rig pa’i ye shes kyi dngos yul la shes pa gcig las ’os med snyam du dgongs par ’dug na’ang/ so sor rang rig ye shes zhes pa rnal ’byor pa so sor rang rig par bya ba’i ye shes zhes pa’i don yin gyi gsung rab spyi ’gro nas bshad pa’i rang rig ’dzin rnam khyad par gsum ldan lta bu gtan ma yin pas ’brel med la/ de lta ma yin na/ legs ldan ’byed rjes ’brangs dang bcas pa dang/ dbu ma thal ’gyur pa dang/ bye brag smra ba sogs kyis mnyam gzhag so sor rang rig ye shes khas len kyang rang rig khas mi len pas sgrub byed de la ’brel yod par ma go lags pa ji ltar lags/ (Se-ra rje-btsun 1997: 163).
“Here, the venerable Karma-pa [Mi-bskyod-rdo-rje] and the lord of the doctrine Gser-mdog-can-pa [Shākya-mchog-ldan] have said that, due to the statement that ‘non-conceptual, individual intuition somehow has truly come into being…,’ non-dual gnosis is taught to be the culminating abiding reality; for this scriptural citation explains that culminating abiding reality to be the experiential object of an individual intuition in equipoise. But they have not at all understood the context. According to the idea of both, they think that nothing but a unique cognition can be the real object of intuitive gnosis (rang rig pa’i ye shes). Nevertheless, ‘individual intuitive gnosis’ has the meaning of gnosis that is to be intuited on the part of the individual yogin, and it is never like the reflexivity endowed with threefold subjectivity, objectivity and specificity that is explained in the scriptures in general, and which [here] has no relevance. For Bhāvaviveka and his followers, the Prāsaṅgika-Mādhyamikas, the Vaibhāṣikas and others affirm the individual intuitive gnosis in equipoise, but disavow reflexivity. If you assume it to be otherwise, how can you have failed to have understood the relevance of that proof?”
It will require further research to determine whether or not Se-ra rje-btsun’s critique of his adversaries on this point is just, but it is noteworthy that the interpretation of so so rang rig (ye shes) that he regards as correct—that is, acceptable from a Dge-lugs-pa perspective—accords rather closely with the sense in which we have seen paccattaṃ veditabba- used in the Pali canon. Of interest, too, is his insistence that the conception is encountered throughout the Indian Buddhist tradition.
The foregoing amply demonstrates, I believe, that Williams has confounded two rather different concepts that some Tibetan thinkers were eager to avoid conflating. Nevertheless, one might still urge that Mi-pham has himself conflated them, in which case Williams’s mistake about this may still lead to an acceptable conclusion in the present context. Careful consideration of Mi-pham’s own writings, however, makes it quite clear that this is not the case. For present purposes I limit myself to adducing one particularly clear statement. In his renowned textbook of Buddhist doctrine, the Mkhas ’jug, a work intended for relatively elementary pedagogy and thus stressing topics that Mi-pham thought to be essential, he writes that “the very gnosis whose nature is liberated from the phenomena of the skandhas, which are of the character of the eight aggregates of consciousness, is intuited” (rnam par shes pa tshogs brgyad kyi rang bzhin can gyi phung po’i chos las rnam par grol ba’i bdag nyid kyi ye shes nyid so so rang rig [Mi-pham 1988: 239]). Because rang rig in the sense of “reflexive awareness” (svasaṃvedana) must be counted among “the phenomena of the skandhas, which are of the character of the eight aggregates of consciousness,” it is definitionally impossible to identify it with so so rang rig, the intuition of liberated gnosis. I must add that all of the traditional Tibetan scholars I have known during a period of close to thirty years who were trained in Mi-pham’s tradition, including some who were his direct grand-disciples, have insisted that rang rig and so so rang rig (ye shes) must be carefully distinguished. Their unanimity on this point no doubt reflects the impetus of their common precursor.
Williams creates some unnecessary trouble for himself, I think, by his decision early in the book to treat vijñāna and jñāna as effectively two words for the same thing, which he translates as “consciousness” (p. xiv, n. 2) This decision, of course, flies in the face of virtually all Tibetan exegesis—whether Rnying-ma-pa, Jo-nang-pa, Dge-lugs-pa, or what have you—that insists on differentiating, never conflating the two. Certainly Williams is right to hold that there must be some consciousness-like dimension to jñāna; otherwise it would be hard to explain why a word derived from jñā- (Tibetan shes) is used here at all. But the most that can be said, I think, is that jñāna and vijñāna are related analogically: “jñāna is to a Buddha what vijñāna is to the rest of us” expresses this, but not very helpfully. (After all, flapping is to a bird what slithering is to a snake, but a subterranean dweller familiar only with serpents cannot be expected to form an adequate conception of avian flight on this basis alone.) In point of fact, the only way one can really know what a Buddha’s knowledge is like is to experience it oneself, and this one can only do by attaining Buddhahood. In this respect, buddhajñāna is truly inconceivable, and this is part of what the conception of pratyātmavedanīyatvam underscores. Translated into contemporary jargon it means: “You had to have been there.” This, of course, did not prevent Indian and Tibetan thinkers and meditators from attempting to discuss jñāna, whether speculatively or on the basis of reported contemplative experiences. What it prevented was their assuming that they could simply lump jñāna together with vijñāna and be done with it. One might well compare, in this regard, the treatment of the so-called “omni-properties” in Western theology and, above all, the puzzles generated in connection with the reflections of St. Anselm on the conceivability of God.
Among the questions requiring further exploration here, then, one that is particularly important concerns the status and understanding of jñāna in Mādhyamika contexts. Sometimes it seems the case that contemporary Western interpreters treat jñāna as a peculiarly Cittamātra topos, and this is certainly an error. Surely, anyone who actually reads Candakīrti’s Madhyamakāvatāra, and above all its autocommentary, through to the end cannot but be impressed that the Prāsaṅgika master is involved in the exegesis of jñāna—nothing could be less true than to hold that he treats jñāna as a non-Mādhyamika topos. Thus, for instance, he does not hesitate to describe the dharmakāya as ye shes kyi rang bzhin can gyi sku, the body whose nature is gnosis (Vallée Poussin 1907: 361, line 15). Tsong-kha-pa and his followers would insist, of course, that this is just a conventional locution (for example, Tsong-kha-pa 1987, p. 305), but no matter; the point here is that the discourse of jñāna is indeed part and parcel of Prāsaṅgika discourse, even if only conventionally.
Despite all that I have argued so far, I do not wish to maintain that Williams is wholly wrong in his suggestion that Mi-pham’s insistence upon affirming rang rig relatively stems from his commitment to Rdzogs-chen. As I will suggest below, Mi-pham’s overall commentarial project with respect to the Madhyamaka is informed throughout by his intention to elaborate a perspective that is well harmonized with Rdzogs-chen. In the first place, as I have elsewhere argued (Kapstein 1992, reprt. in Kapstein 2000, ch. 10), reflexivity is crucial to the process of Rdzogs-chen teaching and meditation. It seems to me quite impossible to interpret the constant emphasis in Rdzogs-chen writings on terms such as rang rig byang chub sems (perhaps “self-presencing bodhicitta”) without recourse to some concept of reflexivity. What is not required, however, is that this be just the same concept as is involved in svasaṃvedana. It may well be that, although most Tibetan authorities agree with Tsong-kha-pa and Mi-pham that svasaṃvedana and pratyātma-vid cannot be one and the same, the latter cannot be cashed out without reference to some notion of reflexivity. If this is so, then it may well be that Rdzogs-chen writers were not in fact seeking to introduce svasaṃvedana into their concept of enlightened awareness, so much as they were concerned to unpack the difficult concept of enlightened intuition, pratyātma-vid.
In Kapstein 1988, writing on Mi-pham’s epistemology, I hedged my bets on this by stating only that “it is characteristic of Rnying-ma-pa thought to find in our ordinary states of awareness (rig pa) a subtle but abiding link with the ineffable truth of enlightenment.” It is now clear to me that Mi-pham did indeed wish to preserve both the distinction between svasaṃvedana and pratyātma-vid as discussed above, while at the same time accepting the concession to reflexivity that his commitment to the Rdzogs-chen entailed. How he achieved this is something that I propose to discuss at length elsewhere. For the moment, I will just state generally that, in his Rdzogs-chen writings, Mi-pham describes the ngo sprod, the initiatory moment when the disciple is introduced to the nature of her mind, as an act of rang rig, in the sense of svasaṃvedana. When, following contemplative cultivation of what had been introduced, intuitive gnosis is disclosed, it is realized to be free from all aspects of conditioned reality, including of course svasaṃvedana. In other words, the relationship, for Mi-pham, between rang rig at the moment of the introduction and the so so rang rig of enlightenment is precisely similar to that which obtains between dpe’i ye shes (jñāna as exemplified [in an initiatory context]) and don gyi ye shes (genuine jn~āna [as realized following the cultivation of the path]) in the new tantric schools, including the Dge-lugs-pa. Though it thus seems that Mi-pham went very far in the way of harmonizing Rdzogs-chen thought with more mainstream scholastic traditions, it is now equally clear that he could not dispense with rang rig altogether and did have a positive reason to assert it conventionally.
As the disagreement between Pettit and Williams (JBE 1999) over whether Mi-pham is or is not to be characterized as a proponent of gzhan stong illustrates, the question of how best to classify Tibetan thinkers is sometimes not altogether clear, and may be somewhat contentious. It is worthwhile noting, therefore, that this was sometimes a problem within Tibetan intellectual circles too, and the present instance is a case in point. I first began to discuss Mi-pham with Tibetan scholars trained in his tradition in 1973, when I started to study Mi-pham’s writings with the late Ser-lo Mkhan-po Sangs-rgyas-bstan-’dzin, who was a great-grand-disciple of Mi-pham through both Bod-pa sprul-sku Mdo-sngags-bstan-pa’i nyi-ma and Zhe-chen Kong-sprul Rin-po-che. Since that time I have had the good fortune to have enjoyed contact with several Rnying-ma-pa, Sa-skya-pa, and Bka’-brgyud-pa scholars who similarly owed elements of their background to Mi-pham’s tradition. Over the years I have been repeatedly struck by an interesting discrepancy in the reception of Mi-pham’s Mādhyamika teaching among Tibetan authorities themselves and his views on gzhan stong in particular. On the one hand, there are those who emphasize those texts and passages in which Mi-pham speaks favorably of gzhan stong and who on this basis regard Mi-pham’s position on gzhan stong as quite similar to that of ’Jam-mgon Kong-sprul (1813–99), who was after all one of his mentors. In other words, they maintain that he did wish to affirm a “soft” variety of gzhan stong, that is, one that adopted a style of discourse markedly influenced by Dol-po-pa, but without the strong ontological claims sometimes associated with the latter’s teaching. Against this, there are others who hold Mi-pham to have adhered more closely to Prāsaṅgika-Mādhyamika, emphasizing the interpretive approach of Klong-chen-pa, rather than Tsong-kha-pa. The favorable remarks on gzhan stong, they say, were motivated by the intentions (1) to illustrate the best defense of gzhan stong for use in debate (and we need to bear in mind here Tillemans’ (1989) perceptive comments on the relationship between Tibetan debate and game-theory); and (2) to position gzhan stong in relation to rang stong as an instance of the two extremes (mtha’ gnyis) to be overcome by a right understanding of Mādhyamika thought.
Despite this interpretive discrepancy, I have never noted any active contestation between these two wings among Mi-pham’s successors. There is a broad consensus, I think, that Mi-pham’s final view is in any case that of the Rdzogs-chen teaching of Klong-chen-pa. Even those who favor a pro-gzhan stong interpretation of Mi-pham seem to agree that in the last analysis this must give way to a radical freedom from conceptual elaborations (spros bral) and that the latter expression, and not gzhan stong, surely represents Mi-pham’s preferred idiom. (I note in passing that the discrepancy we find among Mi-pham’s successors perhaps also reflects a broad discrepancy in Rdzogs-chen exegesis. Whereas some authorities on the Rdzogs-chen—the late Dudjom Rinpoche is a case in point—were very well disposed towards gzhan stong teaching—others have been disinclined to associate the Rdzogs-chen with gzhan stong at all. Thub-bstan chos-kyi grags-pa may be mentioned among the latter.)
Williams is in some sense alive to these complications, as is reflected in his long note (pp. 199–206) on Mi-pham’s relation to gzhan stong, repeated in his response to Pettit (1999). His attempt to argue there, however, that Mi-pham’s use of the expression chos nyid spros bral demonstrates an ontologically positive characterization of the absolute is really nothing more than a quite unfounded contrivance. This becomes clear when we consider Tsong-kha-pa’s remarks in the Drang nges legs bshad snying po, a text of cardinal importance for Dge-lugs-pa thought and one with which Mi-pham and his Dge-lugs-pa interlocutors were all certainly familiar:
de la dgag pa ni sgras brjod pa na tshig gis zin par dgag bya bcad pa’am de’i rnam pa blo la ’char ba na dgag bya bkag pa’i rnam pa can du dngos su shar nas rtogs par bya ba zhig ste/ dang po ni bdag med lta bu’o/ / gnyis pa ni chos nyid lta bu ste/ ’di la tshig gis zin par dgag bya bcad pa med kyang de’i don ’char ba na spros pa bcad pa’i rnam pa can du ’char ba yod do/ / (Tsong-kha-pa 1987: 517).
“Now, as for negation, it is that which is to be understood, having actually arisen as [an intellectual act] whose feature is the negation of the negatum, when there is an explicit utterance grasped verbally [through the use of a negative expression] that excludes the negatum, or when that feature occurs to the intellect. The first is like ‘not-self.’ The second is, for instance, ‘reality’ (dharmatā, chos nyid). Here, even though there is no explicit utterance grasped verbally that excludes the negatum, when its significance arises, it arises as [an intellectual act] whose feature is the exclusion of all elaborations.”
It may well be that Tsong-kha-pa does not often emphasize this way of speaking, but it is clear from the foregoing that he found it quite acceptable, for there is no trace of disparagement in his remarks. And there is no reason of which I am aware to hold that Mi-pham would have taken issue with Tsong-kha-pa about this.
A second red herring that Williams introduces in the same context (p. 200) involves the notion of nyi tshe ba’i stong pa nyid, an “ephemeral emptiness,” in Mi-pham’s references to which Williams finds a tacit assault on Tsong-kha-pa. Tsong-kha-pa, however, also uses this term, and, so far as I can determine, he uses it to mean exactly what Mi-pham does. Thus, for instance, “a conjuror’s knowledge of the falsehood of [conjured] horses and oxen is ephemeral emptiness” (sgyu ma mkhan gyis rta glang brdzun par shes pa yang nyi tshe ba’i stong pa’o [Tsong-kha-pa 1985, p. 748]). Mi-pham’s example is the emptiness of a pot (Mi-pham 1994, p. 118). For Tsong-kha-pa and Mi-pham alike, ephemeral emptiness plays a role in introducing emptiness, but it is by no means to be confounded with the realization of the absolute.
The suggestion, therefore, that the use of expressions such as chos nyid, spros bral, ye shes, nyi tshe ba’i stong pa nyid, and dbu ma chen po (see Kapstein 1995 and Pettit 1999a; Tsong-kha-pa and his successors occasionally use this as a term of praise as well, for example, in Tsong-kha-pa 1987, p. 304) automatically involves some sort of gzhan stong code is groundless. These are terms distributed throughout the writings of most traditions of Tibetan Mādhyamika thought, and only markedly tendentious interpretations of them would support Williams’s conclusions.
I should note, too, in passing that both Williams and Pettit seem to hold that the proponents of gzhan stong in Tibet wished to refute Prāsaṅgika-Mādhyamika. I cannot speak for all varieties of gzhan stong, but my studies of the Jo-nang-pa school (Kapstein 1992/3, 1993, 1997) have led me to conclude that this was not so. The Jo-nang-pas regarded the Prāsaṅgika philosophy associated with Candrakīrti not as wrong, but as incomplete, and thus—though this may strike some as counterintuitive—as ultimately capable, even in its Dge-lugs-pa interpretation, of harmonization with a Jo-nang-pa program. This harmonization was in large measure accomplished in the nineteenth century by the great Jo-nang-pa master ’Ba’-mda’ Dge-legs (1844–1904), who is sometimes depicted as a rival of Mi-pham. I have written about this at length elsewhere, however (Kapstein 1997), and interested readers may refer there.
These reflections do, however, introduce one further topic touched upon by Williams, but perhaps not considered by him in sufficient depth, that is, the religious situation in nineteenth-century Eastern Tibet. (It was, after all, Williams who once argued that “[a]lthough it should not be overemphasised, it does seem that too little attention is paid generally to the political/social context of Oriental philosophical ideas” [Williams 1983b, p. 138].) Perhaps he has taken too seriously Samuel’s (1993) depiction of an opposition between a Dge-lugs and a Ris-med synthesis. In fact eighteenth- and nineteenth-century Khams and Amdo offer plenty of evidence of positive interaction between Dge-lugs-pas, Rnying-ma-pas, Jo-nang-pas, and others, though of course there were sectarian partisans of all stripes as well (Kapstein 1989, reprt. in Kapstein 2000, ch. 8; Kapstein 1997). Again, it would be beside the point to discuss this in detail here; what needs be emphasized, however, is this: there were numerous figures, both Dge-lugs-pa and Rnying-ma-pa, who saw no particular contradiction between the Dge-lugs-pa Prāsaṅgika approach and the Rdzogs-chen, and therefore the notion that there is an special relation between Rdzogs-chen and gzhan stong is no doubt specious. Among Dge-lugs-pa adherents of the Rdzogs-chen, we may mention the renowned Mongolian commentator Bstan-dar Lha-rams-pa, as well as Mi-pham’s opponent Tre-bo Brag-dkar sprul-sku, who was himself a rdzogs chen snying thig practitioner, wrote on this topic, and enjoyed positive relations with the Bon-po Rdzogs-chen master Shar-rdza Bkra-shis-rgyal-mtshan and the latter’s disciples. The great Rdzogs-chen adept from Amdo, Zhabs-dkar Tshogs-drug-rang-grol, was educated as a Dge-lugs-pa, and indeed continued to teach the Lam rim and related materials throughout his life, while preaching the inner identity of Tsong-kha-pa and Padmasambhava (Ricard 1994). And a Rnying-ma-pa like Thub-bstan chos-kyi grags-pa has adopted such a thoroughgoing Dge-lugs-pa approach to Madhyamaka that Williams in fact has mistakenly identified him as a Dge-lugs-pa. I could go on in this way at length, but this much should be sufficient to suggest that the sectarian and doxographic boundaries were often less clear than we sometimes make them out to be. The full complexity of Eastern Tibetan religious life remains poorly studied, and general assessments here require much caution.
I would suggest, therefore, that given our present knowledge of Tibetan doctrinal history doxographic labels such as gzhan stong pa and rang stong pa are best avoided, except of course where they are used within the tradition itself. Our primary task must be to document and interpret precise concepts and arguments, and in many cases the recourse to overly broad characterizations seems only to muddy the waters. Indeed, Williams is certainly at his best when engaged in the careful analysis of dialectical details; here, his philosophical acumen really shines. His reasons for insisting on the question of whether Mi-pham is a gzhan stong pa or not are not at all clear to me, and I do not see just what this really contributes to our understanding of Mi-pham’s thought. That—in accord with his eclecticism—he admitted some aspects of gzhan stong discourse in some contexts no one would dispute, but that is a far cry from defining his general approach. Tsong-kha-pa, for instance, incorporates material derived from Śāntarakṣita and Kamalaśīla into his instructions on vipaśyana, but no one would on that account title him a Svātantrika.
Specialists in Tibetan Buddhism will be grateful to Williams for providing detailed citations from the original texts throughout and for providing in appendix one (pp. 217–230) full transcriptions of the main passages from Mi-pham’s work on which his study is primarily based. A few miscellaneous Tibetological problems may however also be noted:
(1) p. 118: ’khogs bshad “feeble explanation”
Williams admits to some difficulty in interpreting Tre bo brag dkar sprul sku’s apparently derisive characterization of Mi-pham’s work as ’khogs bshad, which he takes to refer to the teaching of one who is senile, feeble with age. There are several problems here. First, disparaging reference to a teacher’s advanced age is not at all consistent with Tibetan cultural norms. Second, Mi-pham was not old when he wrote the commentary on BCA 9. He was just thirty-two in 1878 (sa stag lo), according to the date that he gives in the colophon, and forty-three when he replied to Tre bo brag dkar sprul sku in 1889 (sa glang lo). (In general, it seems a good practice for scholars of Tibetan Buddhism to date accurately the works they study, wherever that is possible, as it readily is here.) Hence, ’khogs bshad probably cannot be understood as Williams suggests. The second problem is that the text does not read ’khogs bshad at all; this is an error in the edition that Williams has used, which reproduces an Indian tracing of the text. Tracings done in India must always be used cautiously, as they often are prepared hastily without adequate correction so that ligatures (especially) are frequently misrepresented. The correct reading here (and found in Mi-pham 1994, among other editions) is ’khyogs bshad, meaning a twisted, or convoluted, explanation (’khyog po, defined in Tibetan lexicons as drang po ma yin pa, “not straight”).
(2) p. 119 mdo tsam brjod na “just the sūtra perspective, omitting that of tantra”
This is a surprising error from a seasoned scholar like Williams. Mdo tsam brjod is a very common idiom meaning “to epitomize, set forth in brief.” It has nothing at all to do with sūtras and tantras. The misunderstanding causes Williams some confusion a few lines later in Mi-pham’s text, where he introduces the terms gnyug ma’i sems (“mind in its natural state”) and bde ba chen po (“great bliss”). Williams recognizes that these are part of the tantric lexicon, but, given his understanding that Mi-pham will avoid tantric discourse, cannot explain just what they are doing here. By contrast, Mi-pham’s point is just that, were we to limit ourselves only to ways of talking that are explicitly sanctioned in Prāsaṅgika works, we would be in the absurd position of excluding, even from our discussions of the conventional, much of the tantric vocabulary that all of the Tibetan traditions prolifically use.
(3) pp. 194–5, n. 5
Williams suggests that Mi-pham’s great disciple Mkhan-po Kun-bzang-dpal-ldan must have based his commentary on the BCA on an oral exposition by Mi-pham of the entire text. But the author himself makes perfectly clear in both his opening invocations and closing colophons that, though his comments on the ninth chapter are indeed based on Mi-pham’s work, the commentary overall derives from Dpal-sprul Rin-po-che’s teaching. Mkhan-po Kun-bzang-dpal-ldan had, in his youth, studied the BCA under the latter.
In concluding, I wish to stress that my critical remarks concern only a small portion of Williams’s book overall and that, as stated in the opening paragraphs of this review, The Reflexive Nature of Awareness is a work of real excellence. The sections in which I have disagreed with Williams I have found to be of great value, too, for Williams’s stimulating and provocative approach to the material always demands critical reflection and response. Williams’s special merit is to engage his readers in a rigorous dialogue with his sources, and by doing this so well he gives new depth and vitality to the field. Serious students of Buddhist philosophy will be grateful for this, perhaps most especially when they find themselves moved to take issue with him.
References
Johnston, E. H., ed. 1950. Ratnagotravibhāga Mahāyānottaratantraśāstra. Bihar Research Society.
Kapstein 1988. “Mi-pham’s Theory of Interpretation,” in Donald Lopez (ed.), Buddhist Hermeneutics (Honolulu: University of Hawai’i Press, 1988), pp. 149–74.
Kapstein 1989. “The Purificatory Gem and its Cleansing: A Late Tibetan Polemical Discussion of Apocryphal Texts,” History of Religions, 28/3 (February 1989): 217–44.
Kapstein 1992. “The Amnesic Monarch and the Five Mnemic Men,” in Janet Gyatso, ed., In the Mirror of Memory. Albany: SUNY Press, 1992: 239–69.
Kapstein 1992/3. Catalogue to The ’Dzam-thang Edition of the Collected Works of Kun-mkhyen Dol-po-pa Shes-rab-rgyal-mtshan. Collected and presented by Matthew Kapstein. 10 vols. in Tibetan + 1 vol. introduction and descriptive catalogue in English. New Delhi: Shedrup Books and Konchhog Lhadrepa, 1992/3.
Kapstein 1993. Introductions to Contributions to the Study of Jo-nang-pa History, Iconography and Doctrine: Selected Writings of ’Dzam-thang Mkhan-po Blo-gros-grags-pa. 2 vols. Collected and presented by Matthew Kapstein. Dharamsala: Library of Tibetan Works and Archives, 1993.
Kapstein 1995. “gDams-ngag: Tibetan Technologies of the Self,” in Roger Jackson and José Cabezón, eds., Tibetan Literature: Studies in Genre (Ithaca: Snow Lion Publications, 1995).
Kapstein 1997. “From Dol-po-pa to ’Ba’-mda’ Dge-legs: Three Jo-nang-pa Masters on the Interpretation of Prajñāpāramitā.” In Ernst Steinkellner, ed., Tibetan Studies: Proceedings of the Seventh Seminar of the International Association for Tibetan Studies. Vienna: Austrian Academy of Science, 1997, vol. 1, 457-475.
Kapstein 2000. The Tibetan Assimilation of Buddhism: Conversion, Contestation and Memory. New York: Oxford University Press.
Kvaerne 1973. Per Kvaerne, A Norwegian Traveller in Tibet. New Delhi: Bibliotheca Himalayica.
Mi-pham 1988. ’Ju Mi-pham rnam-rgyal rgya-mtsho, Mkhas ’jug. Xining: Mtsho sngon mi rigs dpe skrun khang.
Mi-pham 1994. ’Ju Mi-pham rnam-rgyal rgya-mtsho, Spyod ’jug sher ’grel ke ta ke. Chengdu: Si khron mi rigs dpe skrun khang.
Pettit, John. 1999a. Review of Williams 1998b in JBE 1999.
Pettit, John. 1999b. Mipham’s Beacon of Certainty. Boston: Wisdom Publications.
Rgyal-tshab 1985. Rgyal-tshab-rje Dar-ma rin-chen, Dbu ma rtsa ba’i dka’ gnad chen po brgyad kyi brjed byang, in Dbu ma’i lta khrid phyogs bsdebs. Sarnath: Central Institute of Higher Tibetan Studies, 1985: 154–87.
Ricard, Matthieu et al., trans. 1994. The Life of Shabkar: The Autobiography of a Tibetan Yogin. Albany: State University of New York Press.
Samuel, Geoffrey. 1993. Civilized Shamans: Buddhism in Tibetan Societies. Washington/London: Smithsonian Institution Press.
Se-ra rje-btsun Chos-kyi-rgyal-mtshan et al. 1997. Dgag lan phyogs bsgrigs. Chengdu: Si khron mi rigs dpe skrun khang.
Tillemans 1989. Tom J. F. Tillemans, “Formal and Semantic Aspects of Tibetan Buddhist Debate Logic,” Journal of Indian Philosophy 17: 265–97; reprt. in Tom J. F. Tillemans, Scripture, Logic, Language. Boston: Wisdom Publications, 1999.
Tsong-kha-pa Blo-bzang-grags-pa 1985. Byang chub lam rim che ba. Xining: Mtsho sngon mi rigs dpe skrun khang.
Tsong-kha-pa Blo-bzang-grags-pa 1987. Rje Tsong kha pa chen po’i gsung ’bum. Vol. Pha. Xining: Mtsho sngon mi rigs dpe skrun khang.
Tucci, Giuseppe. 1971. Minor Buddhist Texts, Part III. Serie Orientale Roma XLIII. Rome: Is.M.E.O.
Vallée Poussin 1907. Louis de la Vallée Poussin, Madhyamakāvatāra par Candrakīrti. Bibliotheca Buddhica IX.
Williams, Paul. 1983a. “On rang rig,” in Contributions on Tibetan and Buddhist Religion and Philosophy, ed. Ernst Steinkellner and Helmut Tauscher, Vol. 2 of the Proceedings of the Csoma de Körös Symposium held at Velm-Vienna, Austria. Vienna: Arbeitskreis fur Tibetische und Buddhistische Studien Universität Wien, 1983, pp. 321–32.
Williams, Paul. 1983b. “A Note on Some Aspects of Mi Bskyod Rdo Rje’s Critique of Dge Lugs Pa Madhyamaka,” Journal of Indian Philosophy 11.
Williams, Paul. 1998b. Altruism and Reality: Studies in the Philosophy of the Bodhicaryāvatāra. Surrey: Curzon.
Williams 1999. Response to Pettit 1999, in JBE 1999.
