1 For a conspectus of variant readings in this passage, see Ruegg, D. Seyfort, Le traité du tathāgatagarbha de Bu ston Rin chen grub (Paris, 1973), 44Google Scholar.

2 Some of these implications have been discussed in the introduction to Ruegg, Le traité du tathāgatagarbha, 37–49.

3 The two great Tibetan masters just named followed Candrakīrti who, in his Madhyamakāvatārabhāṣya vi.95, described the version of the tathāgatagarbha doctrine set forth in the Lankāvatārasūtra as being of provisional meaning (neyārtha). Sa skya paṇḍi ta's interpretation is found in his sDom gsum rab dbye, f. 9a. See Ruegg, D. Seyfort, La théorie du tathāgatagarbha et du gotra (Paris, 1969), 58Google Scholar, 394; Le traité du tathāgatagarbha, “Introduction”, 31 ff. and 111 n.

4 Concerning a different passage of the Śrīmālāsūtra, Bu ston pointed out another divergence between Ye šes sde's Tibetan version of this sūtra and rṄog's rendering of the same sūtra passage cited in the Tibetan translation of the commentary on the Ratnagotravibhāga; Bu ston considered the latter to be authoritative. See his mDzes rgyan, f. 31a3–4, with Le traité du tathāgatagarbha, 44, 132.

5 Taken by itself and without reference to the Sanskrit, Ye šes sde's ma grol bas/bar šes pa looks morphologically like an endocentric determinative (tatpuruṣa) compound; but although the possessive particle can often be used in Tibetan to translate a Skt. bahuvrīhi has not been written here, the Tibetan expression renders an exocentric possessive (bahuvrīhi) compound qualifying the kośas and can be understood as “having knowledge as not free”; and for the parallel grol bas šes pa – which renders a bahuvrīhi compound qualifying the buddha-qualities – “having knowledge as free” may be proposed. The exact intended meanings of the Tibetan expressions perhaps remain somewhat unclear. (Compare the differently used compounds vimuktijñāna-darśana in RGV iii.12 and RGVV i.94, and muktijñāna-nidarśana in RGV iii.14). As for rṄog's translation-equivalents bral šes pa and bral mi šes pa, taken by themselves and without reference to the Sanskrit expressions which they render, the first – functioning syntactically as an adjectival possessive compound qualifying the kośas even though no possessive particle can has been written – might be rendered “having knowledge as separate”, and the second expression – functioning syntactically as a possessive adjectival compound qualifying the buddha-qualities – might be rendered “not having knowledge as separate”. (For the sense of šes pa, compare, e.g., the compounds brda mi šes pa = asamayajña “not knowing convention, usage”; and smra mi šes pa “incapable of expression” opposed to smra šes pa = pravyavahārasamartha “capable of expression”, where Tib. šes pa corresponds to Skt. samartha “capable, able”. The Sanskrit text of the sūtra passage in which these Tibetan expressions are attested is not found among the Sanskrit fragments of the ŚMDSS published by Matsuda, K. in Braarvig, J. (ed.), Buddhist Manuscripts from the Schøyen Collection, Vol. 1 (Oslo, 2000)Google Scholar.

6 In Ruegg, Théorie, 358–9, for Ye šes sde's ma grol bas/bar šes pa – morphologically an endocentric determinative compound, which here functions syntactically as an adjectival possessive compound even though the possessive particle can has not been written – the rendering “savoir de non-affranchissement” was proposed. In the text the expression qualifies the enveloping kleśakośas, where there is no knowledge of freedom (viz. from these sheaths of which the tathāgatagarbha as such is, nevertheless, ultimately free and empty). And a rendering “savoir d'affranchissement” – an expression qualifying the buddha-qualities, of which the tathāgatagarbha is not empty and where there is knowledge of freedom – was proposed for his grol bas šes pa, which also functions syntactically as a possessive compound although the possessive particle can has not been written. In Le traité du tathāgatagarbhatathāgatagarbha, p. 133, the last translation was also offered for Ye šes sde's rendering (on p. 104 “savoir avec affranchissement” was tentatively given). In Ruegg, Théorie, p. 360, the compound avinirmuktajñānaguṇa = ma bral ba'i ye šes kyi yon tan can in the RGVV’s quotation from the Anūnatvāpūrṇatvanirdeśa was rendered by “à la propriété du savoir non-séparé / savoir inséparable”. When examining the relevant Sanskrit material from the RGVV in WZKS 15, 1971, 132Google Scholar, L. Schmithausen proposed for Skt. a(vinir)muktajñ(ān)a the rendering “bei denen Erkenntnis als losgelöst nicht stattfindet”, providing two interpretations of this qualification and its unnegated form: (1) “die (niemals) als [von der absoluten Wesenheit] losgelöst oder beseitigt festgestellt werden”, and (2) “deren Erkenntnis [von der Erkenntnis der absoluten Wesenheit] (nicht) losgelöst [werden kann]”. He plausibly considered that here ^ojña- has the meaning of ^ojñāna-. rṄog's and Ye šes sde's diverging Tibetan renderings were not discussed by him. Mathes, Recently, K.-D., A Direct Path to the Buddha Within (Boston, 2008), 9,Google Scholar translated the two Tibetan qualifiers in RGVV i. 154–5, pp. 76.7–8, by “recognized as something disconnected” and “[... impossible] to recognize ... as something disconnected”. Concerning the important term sarvākāravaropeta see n. 10 below.

7 The sense intended by the crucial compound avinirmuktajñānaguṇa is supported in RGVV i.44 (p. 32.3–7) by the use of the compound avinirmuktaguṇa to qualify the relation of inseparability of a lamp from its light and warmth and of a gem from its brilliance and hue.

In WZKS 15, 1971, 131–2Google Scholar, L. Schmithausen preferred connecting the qualifier in question with tathāgatadharmāḥ rather than with dharmakāyaḥ, an interpretation that would require emending Johnston's Skt. text in both places; he analysed the compound as: “(besitzt) Vorzüge, welche avinirmuktajñāna- sind”. In A Direct Path to the Buddha Within, p. 203, Mathes translated “properties and qualities, [impossible] to recognize as something disconnected”. The compound can be analysed as a bahuvrīhi containing a tatpuruṣa/karmadhāraya, meaning “having the inseparable quality of Gnosis”; rṄog's Tibetan rendering opts for this interpretation by employing the Tibetan particles kyi and can and writing ma bral ba'i ye šes kyi yon tan can. In rṄog's rendering, this compound might possibly qualify de bžin gšegs pa'i chos dag/rnams, but the Tibetan syntax is not altogether unambiguous here. The Tibetan wording of the whole sūtra quotation differs slightly between RGVV i.1 (ed. Nakamura, p. 3) and RGVV i.44 (ed. Nakamura, p. 75). In his comment on the passage of the ŚMDSS in his mDzes rgyan, f. 19a, Bu ston supplies the gloss ye šes kyis bsdus pa'i yon tan “guṇas included in Gnosis”; see Ruegg, Théorie, 360 n. 3, and Le traité de Bu ston, 104–5.

8 It is important to observe that in RGV(V) i.153 it has been stated that śraddhā (Tib. dad pa) is required for comprehending the tathāgatagarbha, or dhātu, on the level of paramārtha; an eyeless person would indeed be unable to see the blazing disc of the sun. The śraddhā required here is presumably more than just blind faith (what is known in Tibetan as blun dad “fool's faith”, i.e. the opposite of reason and intellectual analysis); rather, it could correspond to prasāda, i.e. tranquil clarity of mind and spiritual receptivity. In the RGVV, the four kinds of persons who need to rely on śraddhā are specified: worldlings, Auditors, pratyekabuddhas, and neophyte bodhisattvas. These are distinguished gnoseologically. In the early stages of practice, mental construction (kalpanā) and dichotomizing thinking (vikalpa) – features characteristic of the thinking of the speculative ratiocinator (tārkika = rtog ge pa) – only begin to be transcended. In its explanation of RGV i.153, the RGVV has stated that it is only the tathāgata who can truly and fully apprehend the tathāgatagarbha by direct and immediate cognition (sākṣātkaraṇa): it is in the ken of the Omniscient one (sarvajñaviṣaya). But the “young” bodhisattva newly set out in his Vehicle (navayānasamprasthita) does not possess this faculty of directly cognizing it; for his mind is confused with respect to Emptiness (śūnyatāvikṣiptacitta). And a fortiori an ordinary worldling (pṛthagjana) attached to the satkāyadṛṣṭi, and an Auditor (śrāvaka) and Pratyekabuddha attached to mistakenness (viparyāsa = phyin ci log, namely in respect to the true nature of tathāgatagarbha as expounded in the Mahāyānist Mahāparinirvāṇasūtra) do not possess this faculty. These four categories of persons would, therefore, need to resort to what is here termed śraddhā; and knowledge acquired through śraddhā differs from direct and immediate cognition (sākṣātkaraṇa) by being mediated through “faith”. See ŚMDSS, 134 ff.; cf. Ruegg, Théorie, 256–7, 309 ff. RGVV i.153 also refers to the ŚMDSS where the pāramitās of nitya, sukha, ātman and śubha representing a reversal (viparyaya) of the saṃskṛtalakṣaṇas have been alluded to; see Ruegg, Théorie, 362 ff.

9 For the expression śūnyatārtha, compare Nāgārjuna's Mūlamadhyamakakārikā xxiv.7, where the issue is also misunderstanding Emptiness:

Indeed, according to Nāgārjuna, when wrongly seen Emptiness destroys the unintelligent, just as a snake if wrongly held also would (xxiv.11; cf. Ratnāvalī ii.19–20, cited by Candrakīrti in his comment). According to Candrakīrti, the sense of śūnyatā is not non-existence (nāstitva, abhāva: Prasannapadā xxiv.7 and 13) – i.e. nihilism – which is but a complementary extreme (anta) to substantialist eternalism, neither of these extremes being in fact espoused by the Mādhyamika as a follower of the “middlemost” (madhyamaka). Rather, śūnyatā has the sense of absence of self-existence (niḥsvabhāvārtha, xxiv.11, p. 496.5), i.e. non-origination of entities by self-existence (svabhāvenānutpādo bhāvānām, xxiv.18). In Mūlamadhyamakakārikā xxiv.18, moreover, śūnyatā, pratītyasamutpāda and prajñaptir upādāya are co-referential; and the final verse in the commentary on the Vigrahavyāvartanī describes śūnyatā, pratītyasamutpāda and the madhyamā pratipat as one in sense (ekārtha “co-referential”). But this does not signify that the terms śūnyatā and pratītyasamutpada are strictly synonymous in the sense of being substitutable for each other; rather, it means that, if things are describable as śūnya, this is because they originate in dependence (pratītyasamutpanna) and are accordingly Empty of self-existence (svabhāvaśūnya). Their origination in dependence is the ground for describing entities as “Empty”, as explained in Nāgārjuna's Vigrahavyāvartanī, 22 ff. The apagogic/deconstructivist procedure of the Madhyamaka school thus follows a line of thinking that is distinguishable from the “constructivist” tathāgatagarbha theory of the Ratnagotravibhāga and its sources with which this paper is concerned.

10 See also RGV ii.11; and Ruegg, Théorie, 351 ff. The Tibetan expression rnam kun mchog ldan gyi stoṅ ñid is also attested with the same meaning; see Ruegg, D. Seyfort, “A Karma bKa’ brgyud work on the lineages of the Indo-Tibetan Madhyamaka”, in Gnoli, G. et al. (eds), Orientalia Iosephi Tucci memoriae dicata, Vol. 3 (Rome, 1988), 1267–8Google Scholar.

The concept of śūnyatā endowed with all excellent modes or features (sarvākāravaropetā) is not peculiar to the tathāgatagarbha literature alone. It is found, in the context of a passage on dhyāna, in the Ratnacūḍasūtra quoted in Śāntideva's Śikṣāsamuccaya xv (ed. Bendall, C., Bibliotheca Buddhica I, 1902, 272): tatra katamā sarvākāravaropetā śūnyatā/ yā na dānavikalā/ yāvan nopāyavikalā/.Google Scholar.. In ch. v of the same text, it is moreover linked, on the level of the bodhisattva's practice, with Emptiness of all reified entities (p. 117): yeyaṃ bodhisattvacaryāyā aparityāgenābhyasyamānā, abhyastā vā, sarvabhāvaśūnyatā. In the context of the major, and contentious, issues of theory and practice that later arose in Tibet in connection with the so-called Great Debate of bSam yas, Kamalaśīla's third Bhāvanākrama quotes the Ratnakūṭa where it is stated that what is to be practised is not śūnyatā alone but Emptiness endowed with all excellent modes – i.e. all salvific means – through being endowed with liberality and so forth (ed. Tucci, G., Minor Buddhist Texts, Part III (Rome, 1971), 27–8Google Scholar): sakaladānādikuśalopetatayā sarvākāravaropetaśūnyatā sevanīyeti uktam, na tu kevalā ...; Śikṣāsamuccaya iv, p. 97, as well as Kamalaśīla's first Bhāvanākrama (ed. Tucci, G., Minor Buddhist Texts, Part II (Rome, 1958), 196Google Scholar) and his second Bhāvanākrama (ed. K. Goshima, 1983, 59–61) also relate to this matter. The issue at stake has been outlined in Ruegg, D. Seyfort, Buddha-Nature, Mind and the Problem of Gradualism (London, 1989), 184Google Scholar and 190. There appears also to exist a link with the concept of sarvākāravaropetaṃ (sarvajña)jñānam. This may be ultimately connected – in the light of the interconnection between Gnosis and its object, “noesis” and “noema”, which are ultimately without duality (advaya) – with the concepts of amuktajña/amuktajñāna and avinirmuktajñana(guṇa), concepts themselves closely connected with the idea of avinirbhāga(buddha)dharmas. Stated otherwise, there appears to exist a kind of structural isomorphism between tathāgatagarbha not Empty of inseparable constituent features and śūnyatā endowed with all excellent modes, and, further, between these two factors and a form of Gnosis (jñāna = ye šes) that is endowed with all excellent modes. In the final analysis, this may be the import of what is suggested by the expression tathāgatagarbhaśūnyatārthanaya.

11 This deconstructivist/apagogic/apophatic view (lta ba = darśana) is to be distinguished from the nihilist extreme (ucchedānta), which – as the counter-position to the eternalist extreme (śāśvatānta) – is nothing but an extreme position (mtha’ = anta) that can find no place in the Madhyamaka, a way of thinking that is explicitly a “middle”; see n. 9 above.

12 In the Laṅkāvatarasūtra (ii, p. 75) itaretaraśūnyatā – a relative, or mutual, emptiness – has been described as inferior (jaghanyā = tha śal). The idea is alluded to also in Majjhimanikāya iii, 104–5. The concept is explained by the formula yad yatra nāsti tat tena śūnyam. This formula was adopted in RGVV i.154–5 (p. 76.9–10), and in Vasubandhu's Madhyāntavibhāgabhāṣya i.1, where it appears in a positive light. On “relative emptiness”, see Ruegg, Théorie, Part III, chapter VIII.

13 Guṅ thaṅ’s interpretation summarized here is found in this important scholar's explanation of Haribhadra's Śāstravṛtti (Sphutārthā) on Prajñāpāramitā thought based on the Abhisamayālaṃkāra – the ’Grel pa don gsal gyi steṅ nas rgyas ’briṅ bsdus gsum mṅon rtogs rgyan rtsa ’grel sogs mdo rgyan sbyar ba'i gzab bšad kyi zin bris sBas don gsal ba'i sgron me (bKa’ ’bum, vol. ka/4) – and in his unfinished comment on Tsoṅ kha pa's Draṅ ṅes rnam ’byed Legs bšad sñiṅ po – the Draṅ ṅes rnam ’byed kyi dka’ ’grel rtsom ’phro legs bšad sñiṅ po'i yaṅ sñiṅ (bKa’ ’bum, vol kha/4). This matter has been discussed in Ruegg, Théorie du tathāgatagarbha et du gotra, Part III, chapters VII–IX. Concerning creativity and doctrinal innovation in this connection, and also in respect to the tathāgatagarbha doctrine as something innovative in Buddhist tradition, see D. Seyfort Ruegg, “Observations on renewal, innovation and modernization in Indian and Tibetan knowledge systems” (forthcoming). There an effort is made to demarcate between novelty and rupture and what is evolving tradition and the continuing deployment of established interpretative instruments in order both to gain fresh insights and to resolve old, and continuing, problems in exegesis and hermeneutics. The latter process may be creative; it is neither “conservative” in the sense of epigonal, nor is it “conservative” in the sense of being antithetical to the renewal of tradition. But it is “conservative” in the sense that it does not jettison the traditional in favour of novelty. In Tibetan Buddhism as well as in Indian Buddhism, there is much that can be legitimately described as revitalization and creativity, but also, at the same time, as deeply rooted in the tradition with which it maintains continuity.

14 Both the apophatic/apagogic/deconstructivist – and also analytical and scholastic – approach and the cataphatic/constructivist – and sometimes even devotional/bhākta – approach to ultimate reality are found exemplified within the corpus of writings traditionally ascribed to Nāgārjuna, an important example of the cataphatic being the Dharmadhātustava and other hymns ascribed to him. The distinction between these two approaches is mirrored in the division of Nāgārjuna's works into a scholastic “Analytic Corpus” (rigs tshogs) and a “Hymnic Corpus” (bstod tshogs) adopted in the organization of the bsTsan ’gyur and in Tibetan doxography. To attempt to determine on the basis of this criterion which of these works were actually composed by the author of the Mūlamadhyamakakārikā, and which by some different “Deutero-Nāgārjuna”, tends, unfortunately, to involve circularity in argument: differences in approach to a complex matter do not necessarily and automatically imply difference of authorship. See Ruegg, D. Seyfort, “Le Dharmadhātustava de Nāgārjuna”, in Études tibétaines dédiées à la mémoire de Marcelle Lalou (Paris, 1971), 448 ff.Google Scholar; Ruegg, Literature of the Madhyamaka School of Philosophy in India (Wiesbaden, 1981), 33 ffGoogle Scholar.

15 See Ruegg, Théorie, 35–6. (Compare also the recent article by Mathes, K.-D., “The gzhan stong model of reality – some more material on its origin, transmission, and interpretation”, Journal of the International Association of Buddhist Studies 34, 2011, 187 ffGoogle Scholar.)

16 See Ruegg, D. Seyfort, “Some reflections on translating Buddhist philosophical texts from Sanskrit and Tibetan”, Asiatische Studien/Études asiatiques 46, 1992, 385–8Google Scholar; and Ruegg, “On translating the Buddhist Dharma” (forthcoming).