The origin and development of the Emptiness of Dharmas in Mahāyāna Buddhism

which imagines and then interprets them, is real.

The Mādhyamika schools are further divided by concerns over whether it is valid to propose any specific position at all about the nature of reality. One group, deriving from Bhāviveka (c.500570CE) was later labelled by the Tibetans with their equivalent of the Sanskrit “Svātantrika” (also known as the Autonomy School). They maintained that both sides must accept the truth of the propositions of an argument. Another group, deriving from Buddhapālita (c.470-540) and Candrakīrti (c.600-650) was later labelled by the Tibetans with their equivalent of the Sanskrit “Prāsaṅgika” (also known as the Consequence School). This group maintained that since they only ‘adopt’ the propositions of the opponent and then apply argument based on them in order to ‘prove’ a contradiction in the opponent’s system, there is no need to actually accept the truth of the axioms themselves.

Later still, Śāntarakṣita (c.8thCCE) synthesizes Mādhyamika and Yogācāra ideas of the emptiness of dharmas and of all things being produced by the mind, saying that the mind itself is empty of self-nature and that its products are a conventional reality.

The tradition of Śāntarakṣita and his pupil Kamalaśīla is taken to Tibet during the first phase of its Buddhist propagation (second half of 8thCCE), whilst later, Atiśa (from 1042CE) and Jayānanda (second half of 11thCCE) bring to Tibet a development of the teachings of Candrakīrti which denies even a “conventional existence” to entities. Later developments of this produce the approach found in the Tibetan Gelukba school up to the present day where the philosophical “content” derives from Candrakīrti whereas the “method” is from Dharmakīriti and Śāntarakṣita. The dates above for various persons are from “The literature of the Madhyamaka School of Philosophy in India” (Ruegg 1981) and “Mahāyāna Buddhism” (Williams 2009).

Development of the doctrine of the Emptiness of Dharmas Background One of the main doctrines which separates the Hīnayāna and Mahāyāna strands of Buddhism is their treatment of the nature of reality. Prior to the emergence of the Mahāyāna view of this, the books of the Abhidharma and Abhidhamma had been produced by the followers of the Theravādin and the Sarvāstivādin schools. These divided all the elements of existence in an exhaustive set of categories called dharmas. For example, the Abhidharmakośa and Bhāṣya of Vasubandhu (c.4thCCE), based on the Abhidharma of the Sarvāstivādins, says:

sarvaṃgraha ekena skendhenāyatena ca / dhatunā svabhāvena parabhāvaviyogataḥ // Ch. 1 v. 18. (Vasubandhu 1983, p.58) Everything (that is ) grasped with a self-nature of one of a skandha, an āyatana and a dhātu is distinct from the nature of the others. Since the Sarvāstivādins defined all their dharmas under these three categories and say each is unique, the set of categories is exhaustive. The treatment of dharmas in the Mahāyāna is taken from the Sarvāstivādin Abhidharma as reflected in Vasubandhu’s work, although of course, this was written later.

Williams divides the development of Mahāyāna into a number of doctrinal elements, of which the development of the Prajñāpāramitā Sūtras and of the Mādhyamika school may be conveniently used to denote the development of the doctrine of the Emptiness of Dharmas (Williams 2009, chap.2-3).

Origin of Prajñāpāramitā Nāgārjuna’s approach to establishing the Emptiness of Dharmas follows after the development of the Prajñāpāramitā Sūtras. These originally seem to be long lists of contradictions and paradoxes, arranged almost as a mirror to the long lists of assertions found in the Abhidharma texts. The best known of these, still chanted daily by many Tibetan, Chinese and Japanese Buddhists is the “Heart (of the perfection of wisdom) Sūtra” (Prajñāpāramitā-hṛdaya Sūtra). This verse provides a concise summary of Prajñāpāramitā ideas: . . . pañca-skandhās tāṃś-ca svabhāva-śunyān paśyati sma. v.2

iha śāriputra rūpaṃ śūnyatā śūnyataiva rūpaṃ rūpān na pṛthak śūnyatā śūnyatāyā-na pṛthag rūpaṃ yad rūpaṃ sā śūnyatā yā śūnyatā tad rūpaṃ. evam eva vedanā-saṃjña-saṃskāra-vijñānam. v.3 (Conze 1948, p.35) . . . (there were) five heaps and he saw these (as) empty of self-nature. v.2 Therefore, Śāriputra, form is emptiness, as emptiness is form. Emptiness is not apart from form, form is not apart from emptiness. Whatever is form, that is emptiness, whatever is emptiness, that is form. Just the same are feeling, perception, mental formations and consciousness. v.3.

We saw above that the Sarvāstivādins maintained that everything has a self-nature of one of a skandha, an āyatana and a dhātu, yet the Heart Sūtra says that all these are empty. As above, the five skandhas (heaps) are empty and the text goes on the say the same about the twelve āyatanas and the eighteen dhātus, denying any self-nature to each and every component of the Sarvāstivādin analysis of reality.

Conze spent a great deal of effort in classifying and analyzing the Prajñāpāramitā literature which he documented in “The Prajñāpāramitā Literature” (Conze 1978). He proposed a sequence for the development of the Prajñāpāramitā Sūtras and a dating as well (Conze 1952). He said that parts of the Aṣṭasāhasrikā Prajñāpāramitā represent the earliest development of Prajñāpāramitā, a view which has been strengthened by other authors since that time and further underpinned by recent archaeological evidence which will be outlined later. As to a possible source of these ideas, he explains:

Several authors have claimed that it probably developed among the Mahāsāṅghikas in Southern India, in the Andhra country, on the Kistnā river. Near Amarāvati and Dhānyakaṭaka, the Mahāsāṅghikas had two famous monasteries, which gave their names to the sects of the Pūrvaśailas and of the Aparaśailas. These sects are significant because 1) they had a Prajñāpāramitā in Prakrit, they 2) spoke of the dharmadhātu in the same way as the Prajñāpāramitā, and 3) their Buddhology prepared the way for that of the Prajñāpāramitā (Conze 1978, p.1). He offers three references in support of the first two points. The third point is Conze’s opinion based on his extensive readings: a) Paramārtha, given by Przyluski in Le Concile de Rajagrha (Przyluski 1926, p.364), b) ‘jam dbyangs bzhad pa (Jam-yang-shay-ba), given by Lamotte in Asiatica – Festschrift Friedrich Weller (Lamotte 1954, p.387n52),

c) Candrakīrti, also given by Lamotte in Asiatica FFW (Lamotte 1954, p.387n53). Conze has taken each of these supporting references from the work of other authors, which on inspection merit further investigation, since the idea that Prajñāpāramitā concepts arise out of the Mahāsāṅghikas has been widely repeated:

a) In Le Concile de Rajagrha, Przyluski immediately cites the Bukkyō daijii (Przyluski 1926, p.364)and Rahder (Rahder 1926, p.xxviii) as the sources of this. Rahder just cites the same section of the Bukkyō daijii.

The Bukkyō daijii quotes from又眞諦の部執異論疏 (Commentary by Paramārtha on the Samayabhedoparacanacakra). It says: 又眞諦の部執異論疏に依れば大衆部には華嚴涅槃勝鬘維摩金光明般等の 諸大乘經ゑ攝すと云ひ(Ryukoku Daigaku 1914, p.2619.b.20). Moreover; according to Paramārtha’s Samayabheda commentary, it is said that, with regard to the Mahāsāṅghikas, they had various Mahāyāna sūtras (such as) the Avataṃsaka(-sūtra), the Nirvāna(-sūtra), the Śrīmālā(-sūtra i.e. Śrīmālādevīsiṃha-nāda-sūtra), the Vimalakīrti-nirdeśa(-sūtra), the Suvarna-prabhāsa-(uttama- sūtra) and the Prajñāpāramitā(-sūtra).

This quotation suggests that the Mahāsāṅghikas had a range of well-known Mahāyāna sūtras available to them, including a Prajñāpāramitā sūtra. However, Paramārtha’s Samayabheda commentary is lost to us (Silk 2009, p.41) but portions of it were included in the三論玄義検幽集 (A deep examination of the Profound Meaning of the Three Treatises, Taisho 2300) of 中觀澄禪 (Chūkan Chōzen, 1227-1307, Japanese but writing in Chinese), so it is possible to examine this more closely. The text from which the Bukkyō daijii took the extract and paraphrased it is as follows:

即於此第二百年大衆部流出三部眞諦疏曰。 第二百年。 大衆部併度行央掘多羅國此國在王舍城北此部引花巌涅槃勝鬘維摩金光明 般若等諸大乘經於此部中有信此經者有不信此經者若不信者謗言無般若等 諸大乘經言此等經皆是人作。 非是佛説悉簡置一處還依三藏根本而執用之。 小乘弟子唯信有三藏由不親聞佛説大乘故爾。 (Takakusu & Watanabe 1924, 70.0459b09) Then, in the 200th year (after the Buddha), from among the Mahāsāṃghikas came forth three divisions, it is said in Paramārtha’s commentary: In the 200th year: The whole of the Mahāsāṃghika travelled to the Aṅguttarāpa country which is to the north of Rājagṛha and this group withdrew there. Among this group, there was belief in the Avataṃsaka(-sūtra), Nirvāṇa(-sūtra), Śrīmālā(-sūtra), Vimalakīrti-nirdeśa(-sūtra), Suvarṇa-prabhāsottama(-sūtra), Prajñāpāramitā(-sūtra) and various other Mahāyāna sūtras; there were also those with no belief in these sūtras.

As for those with no belief, they spoke slander against the Prajñāpāramitā and the various other Mahāyāna sūtras, saying that such sūtras as all those above were the works of Man. Even though it is not that the Buddha spoke all the texts, establishing them in a single occurrence and depending on the original tripiṭaka, nonetheless, they took them up. The Hīnayāna followers only believed what there is in the tripiṭaka, because of not hearing Mahāyāna Buddha-teachings in person. Placed in a fuller context, the Bukkyō daijii extract seems to present certain difficulties. Firstly, in the Samayabhedoparacanacakra of Vasumitra, on which Paramārtha is commenting, neither the Chinese (Taisho 2033, translated by Paramārtha himself) nor Tibetan (Tohoku 4138) versions make any mention of these sūtras nor to those who did and didn’t believe in them. Also, Paramārtha dates this to 200 years after the Parinirvāṇa of the Buddha, saying that these sūtras all exist and are “Mahāyāna sūtras”. He also describes those who believe in these sūtras as “Mahāyāna” and those who don’t as “Hīnayāna”. These terms do not appear in the Samayabheda either. Also none of these terms appear in the other translation of the Samayabheda into Chinese – Taisho 2031 by Xuánzàng. This give the overall impression that this passage concerning Mahāyāna sūtras

and their believers was added to the scope of the existing text by Paramārtha in his commentary. Vasumitra was a Sarvāstivādin scholar who lived four hundred years after the Parinirvāṇa of the Buddha (Lamotte 1988, p.275) – surely he would have been aware of the items that Paramārtha included in his Commentary if they had been extant at the time. b) In Asiatica – Festschrift Friedrich Weller, Lamotte refers to Le Boudhisme by Vasilief, saying:

Les Pūrva et Aparaśaila . . . possédaient, au dire du Grub-mtha ḥ tibétain, des Prajñāpāramitā en Prakrit . . . (Lamotte 1954, p.387) The Pūrva and Aparaśaila (sects) . . . had, as is said in the tibetain Grub-mtha’, Prajñāpāramitā(-sūtras) in Prakrit . . . In Le Boudhisme, Vasilief in turn says: Sie zeigen, dass ihre Lehre in dieselbe Reihe mit dem Buche der Mahāsāṃghika gehört, welches das grosse Statut (mahāvastu, གཞི་ ཆནེ་པོ) gennant wird, wo bereits von den zehn Bhūmi und den Ideen der Pāramitā’s gesprochen wird; dazu fügen sie noch, dass zwei Sekten dieser Schule: die Pūrvaçaila und die Aparaçaila die sūtra’s der Pradschnāpāramitā und andre aus der Lehre des Mahājāna, in PrakritSprache abgefasst, besassen. (Vasilief 1860, p.264) They show that their teaching belongs to the same series as the book of the Mahāsāṃghikas that is called “the big statute” (Mahāvastu, gzhin chen po) where the ten Bhumis and the ideas of Pāramitās are already being spoken; to that they add that two sects of this school: the Pūrvaśaila and the Aparaśaila possessed Prajñāpāramitā and other sūtras of the teachings of the Mahāyāna written in the Prakrit language, it is said. In the paragraph immediately before this, Vasilief refers to གབུ་མཐའ Blatt 81 (grub mtha’ sh. 81):

Die Mahājānisten rechtfertigen sich gegen diese Anklagen durch die Ungewöhnlichkeit (ཐུན་མིན) ihrer Lehre, welche in Folge davon durch die gewöhnlichen Sammler der Sutra’s nicht zugänglich gemacht werden konnte, sondern der Betheiligung von Bodhisattva’s, wie Samantabhadra u.s.w. bedurft habe. (Vasilief 1860, p.264) The Mahāyānists defend themselves against these charges by way of the unusual nature (thun min [i.e. thun mong ma yin pa]) of their teachings, which, in

consequence thereof, could not be made available to an ordinary gatherer of sūtras, but it required the involvement of Bodhisattva’s such as Samantabhadra etc. It is not possible to determine which edition of “grub mtha’“ Vasilief was using, but elsewhere in the book he makes reference to ‘jam dbyangs bzhad pa (Jam-yang Shay-ba), whose “Great Exposition of Tenets” is known as “grub mtha’ chen mo” ('Jam dbyangs bzhad pa'i rdo rje 1997). This has been translated by Hopkins in “Maps of the Profound”. Using Hopkins’s edition (Hopkins 2003, p.716), the relevant section in English can be found (Hopkins uses “collector” rather than

“gatherer”) and the corresponding section in Tibetan has been located ('Jam dbyangs bzhad pa'i rdo rje 1997, pha pt.2 111a). The section on the Pūrvaśailas and the Aparaśailas follows this (ibid., pha pt.2 113.a). Hopkins only provides references to a Taipei reprint of this work in book format not available to this author. The references in this paper have been substantially reworked to the Tibetan foliation from the bla brang bkra shis ‘khyil blockprint. Hopkins only uses

the Root Text of ‘jam dbyangs bzhad pa, interspersed that with the Word Commentary of ngag dbang dpal ldan (Nga-wang Bel-den). As a result, when Hopkins writes: . . . in the scriptures of the Pūrvashailas and Aparashailas seven scriptural collections are mentioned, consisting of the Bodhisattva scriptural collection, the scriptural collection of the Knowledge Bearers, the scriptural collection of discipline, the scriptural collection of discourses, the scriptural collection of manifest knowledge, the extensive scriptural collection, and the scriptural collection of birth stories. Hopkins does not give ‘jam dbyangs bzhad pa’s source as it is not identified in the texts he has selected. ‘Jam dbyangs bzhad pa does provide the source as a quotation in his own Word Commentary. It is from Candrakīrti’s Triśaraṇasaptati (gsum la skyabs su ‘gro ba bdum cu pa Tohoku 3971). Candrakīrti says: / nyan thos rnams kyi dbang mdzad nas /

/ sde snod gsum du ‘dod pa yin / / byang chub sems dpa’i gnas skabs kyi / / nges pa der ni yod ma yin / / ‘phags pa nub kyi rir bshad dang / / shar gyi ri pa’i nyan thos kyis / / byang chub sems dpa’i dbang byas pa / / sde snod bdun du bshad pa yin / / byang chub sems dpa’i sde snod dang / / de bzhin rig ‘dzin zhes bya dang / / mdo sde chos mngon ‘dul ba dang / / rgyas dang de bzhin skyes pa’i rabs / / de ltar sde snod bdun po ni / . . . Candrakīrti: (Rin-chen 1985, T.3971 dbu-ma gi 253a), quoted by ‘Jam dbyangs bzhad pa: ('Jam dbyangs bzhad pa'i rdo rje 1997, pha pt.2 111a.3)

As justified by [lit. after making force for] the Śrāvakas, three baskets are considered; There, the existence of a Bodhisattva is not established. For [lit. with] those who expound of the noble Aparaśailas and the Śrāvakas of Pūrvaśaila, seven baskets are said to exist, justifying [lit. making the force of] a Bodhisattva:

The Bodhisattva basket and also what is called “the Knowledge Holder” (basket) [i.e. vidyādharapiṭaka] and the sūtra (basket), the abhidharma (basket), the Vinaya (basket) and the extensive basket [i.e. vaipulyapiṭaka] and also the (basket of) “birth-stories” so there are seven baskets . . . ‘Jam dbyangs bzhad pa goes on to say:

cing spyan ras gzigs brtul zhugs kyis sher mdo don dam pa’i chos mngon par bshad cing mdo’i sde snod yang yin pa’i phyir te / ('Jam dbyangs bzhad pa'i rdo rje 1997, pha pt.2 113.a.5) Avalokitavrata explains that the Prajñāpāramitā Sūtra (is) true abhidharma, since it is also (in) the basket of sūtras. Avalokitavrata’s commentary to Bhāviveka’s Prajñāpradīpa, called Prajñāpradīpaṭīkā (shes rab sgron ma rgya cher ‘grel pa Tohoku 3859) also says: ‘di ltar dge ‘dun phal chen sde nyid kyi sde snod kyi gzhi chen po zhes bya ba’i khongs su theg pa chen po ‘di yang gtogs te / de nas sa bcu pa’i mdo dang pha rol tu phyin pa’i mtshan nyid dag ‘byung ba’i phyir dang / dge ‘dun phal chen sde nyid kyi shar gyi ri bo’i sde dang nub kyi ri bo’i sde dag las kyang ‘phral skad du / shes rab kyi pha rol tu phyin pa la sogs pa theg pa chen po’i mdo dag ‘byung ba’i phyir ro / (Rin-chen 1985, T.3859 dbu-ma za 270a.6)

So, Mahāyāna is also included among the Vinaya basket of (those of) the saṃgha of the Mahāsāṅghikas called the Lokottaravādins, because within that are the sūtra of the Ten Grounds and the defining characteristics of the (Six) Perfections, and because there are the Prajñāpāramitā and other Mahāyāna sūtras from the sects of the Pūrvaśailas and the Aparaśailas of the saṃgha of the Mahāsāṅghikas, even in everyday language [i.e. a Prakrit].

The references in Avalokitavrata’s Commentary to a Vinaya of the Lokottaravādins and to Prajñāpāramitā and other sūtras in Prakrit do not occur in Bhāviveka’s Prajñāpradīpa, nor in his Madhyamakahṛdayavṛtti-Tarkajvālā (Tohoku 3856, also known as “The Blaze of Reasoning”) which Avalokitavrata also quotes in his commentary. c) In Asiatica – Festschrift Friedrich Weller, Lamotte refers to the Mūlamadhyamakakārikās de Nāgārjuna by de la Vallée Poussin, saying of the supposed Prajñāpāramitā in Prakrit:

Dans des stances en quasi-sanskrit, dont le texte nous a été conservé par Candrakīrti, les Pūrvaśaila professaient sur le Dharmadhātu, ou l’Absolu, des vues identiques à celles du Mahāyāna. (Lamotte 1954, p.387)

In stanzas of quasi-sanskrit, whose text was preserved for us by Candrakīrti, the Pūrvashailas professed views identical to those of the Mahāyāna over the Dharmadhātu, or “the absolute” De la Vallée Poussin (de la Vallée Poussin et al. 1903, p.548) cites two verses from the Sanskrit version of Prasannapadā. However, better editions of these same verses have been produced since that time and these are presented by Harrison:

āgamasūtreṣu / ṇa viṇaṣṭa (or viṇaṭṭha?) ṇa uppaṇṇa (?) dharmmadhāusamaṃ jagaṃ / sattadhāuṃ ca daṃśesi eṣā loānuvattaṇā // tīsu adhvāsu sattāṇaṃ pakatī nopalambhatī / sattadhāuṃ ca damśesi eṣa loānuvattanetyādi // (Harrison 1982, p.225)

Harrison translates this as: Not destroyed, not produced, the world is the same as the dharmadhātu , Yet he taught the sattvadhātu; this is in conformity to the world. He does not apprehend the nature of beings in the three times, Yet he taught the sattvadhātu ; this is in conformity to the world. (Harrison 1982, p.225) So we can see that this does not really support the claim that there was a Prajñāpāramitā sūtra in Prakrit, but it does support Lamotte’s contention that Mahāyāna-like ideas did exist in Prakrit texts of the Mahāsāṅghikas. Candrakīrti also quotes these two verses in the Madhyamakāvatārabhāṣya where he attributes them to the Pūrvaśailas. (de la Vallée Poussin et al. 1903, p.458n2). It is clear that the “quasi-sanskrit” referred to by Lamotte is a form of what Edgerton calls “Buddhist hybrid Sanskrit”, albeit of a very “Prakrit” type. Harrison has identified the source of these verses, which will be discussed later. Assessment of Conze’s evidence Returning finally to Conze’s contention that the Mahāsāṅghika sects of the Pūrvaśailas and the Aparaśailas spoke of the dharmadhātu in the same way as the Prajñāpāramitā and had a Prajñāpāramitā in Prakrit, we find that his reasons seem to support the first point, but that the second point is not strongly supported by them. The best that might be said is that there were texts containing similar ideas to some of those in the Prajñāpāramitā sūtras and that these were

written in what is now called Buddhist hybrid Sanskrit, as coined by Edgerton for this type of language. However, even this position bears further consideration. The authors upon whose works these conclusions rest are all of a relatively late date compared with the period they are describing. In the Samayabheda of Vasumitra, these sects are described as existing two hundred years after the Parinirvāṇa of the Buddha (Lamotte 1988, p.282), yet Paramārtha (499-569CE), Avalokitavrata (c7-8thCCE) and Candrakīrti (c600-650CE) all lived at least 500 years later still. Each of these works is a commentary to an earlier work, yet we do not find the relevant information in those works, only in the later commentaries. However, there must be some foundation to these ideas, since Paramārtha would have had no contact with either Avalokitavrata nor Candrakīrti and Avalokitavrata was a supporter of Bhāviveka and an critic of Candrakīrti. Nonetheless, they all agree on the main points.

More recent developments Two more recent developments throw further light on the development of the Prajñāpāramitā: Paul Harrison’s work on the Chinese translations of Mahāyāna sūtras by Lokaṣema during the latter half the second century CE and Harry Falk’s transcription, analysis and carbon-14 dating of a birch-bark manuscript from Gandhāra of a section of the Aṣṭasāhasrikā (8000 line) Prajñāpāramitā sūtra in Kharoṣṭhī.

Harrison looked at a particular work translated by Lokaṣema, the Lokārnuvartana sūtra, which he has determined has many similar verses to the Māhavastu of the Lokottaravādin subsect of the Mahāsāṃghikas. Because of this and because the subject matter of the first half the sūtra itself is the transcendent nature of the Buddha, Harrison concludes that the Lokārnuvartana sūtra is a work of the Lokottaravādins (Harrison 1982, pp.212-213). The second part of the sūtra is concerned with the teaching of the Buddha. These verses are not found in the Mahāvastu and Harrison does not address them by and large. However, he does show that the two verses quoted above by Candrakīrti in Buddhist hybrid Sanskrit are in fact from this text, being verses 94 and 63. He compares these verses with those of the only other translation of this sūtra, into Tibetan by Jinamitra, Dānaśīla and Yeshes sde in the 9th CCE and the Tibetan version of the Madhyamakāvatāra of Candrakīrti where these verses are again repeated, although having been translated by Tilakakalasa, Pha tshab nyi ma grags and Kanakavarma. There is a close correspondence.

Falk has been working on a variety of Kharoṣṭhī fragments in recent years. At the 2008 conference of the International Association of Buddhist Studies, Falk described a birch-bark

manuscript found in Gandhāra and now in a private collection. It consists of parts of the Aṣṭasāhasrikā Prajñāpāramitā sūtra in Kharoṣṭhī. In a recent email exchange with this author, Falk has said that the manuscript has been carbon-14 dated “in the seventies (of) 1st century AD, with a rather narrow margin up and down”. A paper on the text and dating is expected in 2010. Although to be confirmed, the writing style of the manuscript is said to be similar to that of one of the scribes of the British Library Kharoṣṭhī fragments. One of the other Gandhari Kharoṣṭhī fragments is part of the Mahāvastu which as was said above is associated with the Lokottaravādins whilst, albeit later, when Xuánzàng visited India in the 7thCCE, he found considerable numbers of Lokottaravādin monasteries at Bāmyān, in Gandhāra (Lamotte 1988, p.541). So the Lokottaravādins show a continued presence in this area for many centuries.

These two more recent developments provide a firmer foundation for the ideas related by Conze that the Prajñāpāramitā, or at least its core doctrines, developed amongst the Mahāsāṅghikas. We also see that the dating of the manuscript of the Aṣṭasāhasrikā Prajñāpāramitā sūtra to the first century CE supports Conze’s proposal that the earliest parts of the Prajñāpāramitā sūtras might be as early as 100BCE.

Nāgārjuna’s proof Nāgārjuna (c.150-250CE) wrote his most influential work, the Mūlamadhyamakakārikā as a “proof” of the emptiness of self-nature (svabhāva) of dharmas. In this work he relentlessly and repeatedly applies logical argument based on the four-part syllogism (also called a tetra-lemma or catuṣkoṭi in Sanskrit) to prove that any proposition that there might be some self-existence in some entity or another leads to contradiction. He applies this argument across a whole range of possible circumstances, starting with “cause and effect”, ranging across the sense faculties (indriya), aggregates (skandha) and elements (dhātu) of the Abhidharma, through the functioning of Karma, the Self, the Tathāgata, the Four Noble Truths and even Nirvāṇa itself. He completes the work with a justification of the method he has adopted, that of not adopting his own specific view on existence, but by being able to demonstrate contradiction in whatever specific view anyone might ever adopt.

Nāgārjuna makes a specific defence against the accusation of nihilism in chapter 24 of the Mūlamadhyamakakārikā since it would seem that if everything is empty of self-nature, nothing “useful” remains. He outlines an attack where he is accused of implying that dependent arising, the four Nobel Truths and the Buddha and all his teachings are all non-existent. His defence against this is to explain that the Buddha’s teachings are supported by two truths: a truth of worldly convention and a truth of ultimate reality; that without understanding the distinction between the truths, the Buddha’s message cannot be understood; that, without relying on

mundane activity, ultimate reality is not explained and that without ascertaining ultimate reality, nirvāṇa is not attained (Inada 1970, pp.144-146; Kalupahana 1986, pp.326-333).

Is Nāgārjuna right? This question has led to a series of doctrines within the Mahāyāna, based on various interpretations of Nāgārjuna’s position. None specifically contradict Nāgārjuna, but they all “reinterpret” him. Three particular positions and their proponents will be examined:

1. The place of “mind” in understanding the true nature of reality – this leads to the distinction between the Mādhyamika and Yogācāra divisions of Mahāyāna Buddhism; 2. Whether one should hold any position (a view) whatsoever about the nature of true reality or only to approach it through negation of everything else. This leads to the distinction between the Svātantrika and Prāsaṅgika schools of the Mādhyamikas; 3. What the conventional nature of reality is – is it the product of a conventional mind or does it have some kind of conventional nature of its own, or is empty of even a conventional self nature. This leads to the distinction between the Prāsaṅgika and what the Tibetans later labelled with their equivalent of the Sanskrit “Yogācāra-SvātantrikaMādhyamika”. Looking at each of these in turn:

1. Yogācāra views of Emptiness and the place of “Mind”.

The foundational text of Yogācāra is generally taken to be the Saṃdhinirmocana Sūtra (“The sūtra unravelling the thought” or “the scripture on the explication of underlying meaning) and provides the doctrinal basis for the subsequent developments by Asaṅga (c310-390CE) and Vasubandhu. Some kind of date for the beginning of Yogācāra can be obtained by looking at the dates of translation of various texts into Chinese. The Saṃdhinirmocana Sūtra was first translated during 435-443CE by Guṇabhadra (Taisho 678), with later versions by Bodhiruci (Taisho 675), Paramārtha (Taisho 677) and Xuánzàng (Taisho 676). Compare this with the translation of the Aṣṭasāhasrikā (8000 line) Prajñāpāramitā by Lokaṣema (Taisho 224) in 170CE. Likewise, the first work by Asaṅga translated into Chinese was the Mahāyānasaṅgraha by Buddhaśanta (Taisho 1592) in 531CE, compared with the Mūlamadhyamakaśāstra, an edition of the Mūlamadhyamakakārikā of Nāgārjuna with a

commentary by Piṅgala, first translated in 409CE by Kumārajīva (Lancaster & Park 1980). We see that the Yogācāra arises some time after the Mādhyamika, and if we consider that the Saṃdhinirmocana Sūtra might have existed in India for up to 100 years before being taken to China, we arrive at a period coincidental with the dates of Asaṅga (Snellgrove 1987, p.103). So we see that the Yogācāra follows the Mādhyamika in time, but should it be considered a successor to the Mādhyamika, in opposition to it, or as arising as an independent doctrine. The Abhidharmasamuccya of Asaṅga and the Abhidharmakośakārikā of Vasubandhu recast the Abhidharma in Yogācāra terms, redefining the skandhas and expanding the lists of dharmas, dhatūs etc. (ibid., p.99). So there is at least a development of the Abhidharma ideas here, rather than a rejection of them as we see in Nāgārjuna. As to being a successor, there are a series of present-day descriptions of Yogācāra which do propose this. In “Mādhyamika and Yogācāra, Nagao says: The Yogācāras . . . inherited the Nāgārjunian notion of emptiness, and, when they elucidated features of yoga-praxis such as the six pāramitās, the ten bhūmis, and so on, emptiness, seems to have been the basis of their theories. (Nagao 1991, pp.51-52)

and in his translation of the Saṃdhinirmocana Sūtra, Keenan says: (is) an affirmation of the validity of Mādhyamika thought and practice, to which Yogācāra brings its theory of conscious interiority (Hsüan-tsang 2000, p.91n34)

However, others point out specific texts which place the Mādhyamika and Yogācāra points of view in opposition, for example in “Bhāviveka and his Buddhist Opponents” by Eckel, quoting Dignāga: The teaching in the Perfection of Wisdom is based on three [[[Wikipedia:identities|identities]]]: imagined, dependent and absolute. The words ‘do not exist’ rule out everything that is imagined. Examples such as illusion teach dependent [[[Wikipedia:Identity (social science)|identity]]]. The fourfold purification teaches absolute [[[Wikipedia:Identity (social science)|identity]]]. The Buddha has no other teaching in the Perfection of Wisdom. (Eckel 2008, p.67)

Bhāviveka (c500-570CE) devotes a whole chapter of the Madhyamakahṛdaya (Heart of the Middle Way) to rebutting a variety of Yogācāra doctrines, which he clearly saw in opposition the Mādhyamika. In addition, the Saṃdhinirmocana Sūtra itself says: There are [other] sentient beings who . . . even if they hear this doctrine will be unable to understand the underlying intent of my words . . . They would only be clinging to the words that express the meaning, that is: that all things most certainly have no essence and no arising, are originally quiescent, and are essentially in cessation. Consequently they take up the view of nihilism and nonexistence of all marks. . . They do form concepts about my teaching, but in their negation of meaning, they do not form concepts of its meaning. (Hsüan-tsang 2000, p.43)

In Power’s translation of the same text from Tibetan, he notes that, much later, Tsongkha-pa identifies the target of this as being Mādhyamikas (Powers 1995, p.119n13). So what do the Yogācāra have to say about what does exist? The (possibly mythical) founder of Yogācāra, as “reported” by Asaṅga in the Madhyāntavibhāga says, in the very first verse after the Homage: abhūtaparikalpo ‘sti dvayaṃ tatra na vidyate / śūnyata vidyate tvatra tasyāmapi sa vidyate // 2 (Pandeya 1971, p.9)

Imagination of the non-existent (i.e. conception of everyday experience) is real, In it, duality is not evident. But emptiness is evident there and in it that (the imagination of the non-existent) is evident.

And Vasubandhu comments on this in his bhāṣya commentary that when we hold that something is empty, we mean that one thing is empty of another (Williams 2009, p.93). Also in verse 13, reinforcing this: dvayā ‘bhāvo hyabhāvasya bhāvaḥ śūnyasya lakṣaṇam / 13a (ibid., p.36) Indeed, the non-existence of duality (and) the existence of (that) non-existence is the characteristic of emptiness.

So we see here not an “elucidation” of Mādhyamika as claimed by Nagao, but an outright doctrinal opposition – Yogācāra claiming that something really does exist and that the Mādhyamika are nihilists, whilst the Mādhyamikas claim that the Yogācāra position does not uphold emptiness as expounded in the Prajñāpāramitā sūtras and is reifying emptiness. In the works of such Mādhyamikas as Bhāviveka and Candrakīrti, we find explicit arguments against the Yogācāra view based on Nāgārjuna’s position and structured as “debates”. On the other hand, although such Yogācāra followers as Dignāga and Dharmkīrti are renowned as “logicians”, they did not produce works containing debates against Nāgārjuna’s position, although as shown above, the Yogācāra position is opposed to it in several ways. The Mādhyamika position is criticized as needing further interpretation rather than being wrong. It is particularly interesting that whereas the foundational text of the Mādhyamikas, the Mūlamadhyamakakārikā, is clearly attributed to Nāgārjuna, the equivalent text of the Yogācāras, the Saṃdhinirmocana Sūtra, which explains this need for further interpretation, is placed in the mouth of the Buddha. Perhaps this is necessary if no human Yogācāra follower felt able to contradict Nāgārjuna! 2. Svātantrika position on holding views

This division of the Mādhyamikas between the Svātantrikas and Prāsṅgikas is based on their approach to discussing the nature of conventional reality. Both sides do agree on the nature of ultimate reality as propounded by Nāgārjuna, however. In the last two verses of the Mūlamadhyamakakārikā (Ch 27 v.29-30), Nāgārjuna says: athavā sarvabhāvānāṃ śūnyatvācchāśvatādayaḥ / kva kasya katamaḥ kasmāt sambhaviṣyanti dṛṣṭayaḥ // 29 sarvadṛṣṭiprahāṇāya yaḥ saddharmam adeśayat / anukampam upādāya taṃ namasyāmi gautamaṃ // 30 (Kalupahana 1986, pp.390391)

So where, for (lit. of) whom and because of what would views such as “the eternal” arise from the empty nature of all existing things? I pay homage to Gautama who, with compassion, taught the true dharma for the abandonment of all views. Since Nāgārjuna has already discussed the two truths in chapter 24 and explained that without relying on mundane activity, ultimate reality is not explained and that the example, “the eternal”, is a “view” on the ultimate nature of reality, not the mundane

nature, it might be thought that Nāgārjuna was only referring to views about ultimate nature. On the other hand he does say “all views”. Lopez sums up the two positions well: For the Svātantrika, the two parties are debating about an attribute of an object . . . the subject of the debate must be established as commonly appearing. . . . the Prāsaṅgika . . . asserts that a commonly appearing subject does not exist. They use reasons which are renowned to others . . without having to assert the subject himself. (Lopez 1987, pp.78-79)

The Prāsaṅgika position described here was that defended by Candrakīrti and taken to Tibet by Atiśa. This unwillingness to accept any kind of views at all leads later Prāsaṅgika interpreters such as Tsong-kha-pa to a very subtle (or possibly over complex) definition of the existence of conventional objects as “merely imputed”, lacking either a self nature from their own side , nor being a mental construct either, “which even its adherents admit is a tenet notoriously difficult to understand properly” (Cozort 1998, p.43). 3. What is the conventional nature of reality?

Based on the position of not holding views, this division is between the position adopted by Śāntarakṣita, known as Yogācāra-Svātantrika-Mādhyamika and the Prāsaṅgika. In some ways it repeats the division between the Yogācāra and Mādhyamikas on the existence of entities, but accepting that they have no true existence. Their position is that conventional entities themselves have no conventional existence and are “projections” of the mind, albeit a mind with only a conventional nature. In support of this, Kamalaśīla, Śāntarakṣita’s successor, in turn reinterprets the Yogācāra doctrine of the Three Natures so that only one of them, parikalpita-svabhāva (the self nature of the imaginary entity) is an ultimate reality, the other two being interpreted as only having a conventional nature (Ruegg 1981, p.95). This allows a reinterpreted Yogācāra view of conventional reality to be compatible with the Mādhyamika view of ultimate reality. The Prāsaṅgika response to this is based on the “original” refutation of Yogācāra views about whether there is any difference between the perceived snake imagined when a piece of rope is seen in poor light, or a perceived “real” snake. The Prāsaṅgika view does require a conventional mind to impute (or label) the collections of perceived skandhas as specific objects, but differs in saying that these do have a conventional existence in that way.

Are the Prāsaṅgikas right - a contemporary assessment of Nāgārjuna’s method

In 1931, the mathematician Kurt Gödel published a paper entitled, in English, “On Formally Undecideable Propositions of the Principia Mathematica and Related Systems” (Gödel 1931). It proved that most formal systems constructed out of axioms and operations can not be proved to be free of contradiction. In fact it proved that there must be a contradiction. It can be applied to the formal system of anyone who “holds views” to show they will be unable to withstand a Nāgārjunian attack of demonstrating contradiction in their view. Their only defence against this is to prove that their view IS without contradiction, which Gödel proved to be impossible, except in a very few simple cases. It turns out that Nāgārjuna had already addressed these in his dismissal of the Ātman and of the Dharmas. Gödel’s proof demonstrates that there must be a contradiction in any such system, without the need actually to find it. Therefore, the “holding of views” is always unsafe against Nāgārjuna’s approach.

Conclusions From the origins of the emptiness of dharmas in the Prajñāpāramitā sūtras as an antidote to the lists of entities in the Abhidharma, through the deconstruction of dharmas by Nāgārjuna and his warning concerning views and with the subsequent developments by Bhāviveka, Dignāga, Candrakīrti and the many other contributors to the Mādhyamika school, we see a constant focus on the application of debate and of “Buddhist logic”. See also see that no one actually contradicts Nāgārjuna, except for the Buddha. Everyone else “re-interprets” what he says. In modern times, Buddhist logic has been shown to be able to be represented in formal mathematical logic, as in, for example, “Buddhist Formal Logic” by Chi (Chi 1969). It would seem appropriate to continue this tradition by applying one of the 20th century’s greatest developments in mathematics and logic, that of Gödel, in support of Nāgārjuna’s (non-) position.