<<

Buddhist Studies Review 25(2) 2008, 210–31 ISSN (print): 0256–2897 doi: 10.1558/bsrv.v25i2.210 ISSN (online): 1747–9681

The Date and Cause of the First Schism

Bhikkhu Sujato

Santi Forest 100 Coalmines Road, (PO Box 132), Bundanoon, NSW 2578 Australia [email protected]

ABSTRACT: The texts and inscriptions dating from the early period – roughly the fi rst 500 years after the Buddha’s passing away – do not support the conclusion that fully-fl edged existed at that time. Rather, we should think in terms of ‘sectar- ian tendencies’ that emerged as actual sects towards the end of the early period. The available sources that speak of the First Schism are best read as sectarian accounts de- picting the situation in the of their own time – roughly 100–500 CE – rather than as historical records of the pre-Aśokan era. All of the Sthavira sources mention the so-called ‘fi ve theses’ of the Mahāsaṅghika (later ascribed to a certain ‘Mahādeva’) as either the cause of the First Schism or as important doctrinal issues. And a review of relevant sources demonstrates that the ‘fi ve theses’ were probably accepted within the Mahāsaṅghika, which con- fi rms that this was likely the principal cause of the First Schism.1

INTRODUCTION

Early Buddhist history pivots around the date of the First Schism, which forever divided the Buddhist community into separate schools. Virtually all our ancient accounts agree that this schism resulted in the formation of the Mahāsaṅghikas and the Sthaviras, which formed the basis for the subsequent emergence of the ‘18 schools’.2 But here the agreement ends. Both the date and the cause of the First Schism are presented in very diff erent ways in the ancient sources. Hence

1. This paper is essentially a summary of some of the main arguments in my Sects & Sectarianism. This book, together with several additional papers, is available at http://sectsandsectarianism. googlepages.com. 2. I use the form sthavira rather than the theriya to avoid giving the unwarranted impression that this school is identical with the Theravāda of the present day, whose ances- tors I call the Mahāvihāravāsins. The Mahāvihāravāsins are of course just one of many schools descended from the Sthaviras. As has been often pointed out, the number ‘18’ is conventional, and does not represent the exact number of schools.

© Equinox Publishing Ltd 2008, Unit 6, The Village, 101 Amies Street, London SW11 2JW SUJATO THE DATE AND CAUSE OF THE FIRST SCHISM 211 one of our tasks is to examine these sources to determine whether the diff erences may be reconciled. If they cannot, we must look to indirect evidence. In this paper, I fi rst consider the four main accounts of the First Schism, which are our direct evidence for its date. Then I briefl y survey the overall fi eld of indi- rect evidence. Finally I consider the likely cause of the schism. While my discus- sion of the date of the schism is largely based on a negative method – criticizing the sources that have been used to fi x the date – the discussion of the cause of the schism is based on a positive method, using sources that I believe have so far remained unnoticed. I think this evidence is suffi ciently compelling to make it unnecessary to discuss the various proposals that have been made in the past for the cause of the schism. For the purposes of this paper, I adopt a broad distinction between the ‘early period’ – roughly the 500 years from the Buddha’s birth until the start of the Common Era – and the ‘middle period’, ending around 500 CE. I try to avoid false precision by expressing dates in the most general way that is compatible with the context. While one aspect of historical inquiry must be to specify dates as closely as possible, such attempts in this fi eld are so tenuous that I prefer not to use them as the basis for further arguments.3 Nevertheless, for the purposes of clarity I accept the ‘median chronology’ for the Buddha: thus the Parinibbāna and the First Council were roughly 413 BCE, the Second Council was around 313 BCE, and Aśoka around 277–246 BCE (Cousins 1996, 109). Acceptance of a diff erent chronology for the Buddha would not aff ect my main arguments. For the most part I will express dates as ‘After Nibbāna’ (AN); thus, for example, 100 AN = 313 BCE. Bechert has established that the emergence of sects (nikāya-bheda) is not syn- onymous with a schism in the narrow legal sense of the word (saṅgha-bheda) (Bechert 1982). I agree with this aspect of his analysis,4 and further believe that there is no evidence that any of the sects emerged due to formal saṅgha-bheda. Throughout my article, therefore, I use ‘schism’ in the general sense of a division of groups of Saṅgha, without implying a formal saṅgha-bheda.

PART 1: THE DATE

The state of play

Modern scholars typically place the First Schism between the Second Council and Aśoka. Thus Bechert refers to: ‘the fi rst schism, which must be placed before Aśoka’ (Bechert 2005b, 66). Prebish concurs: ‘Now we all know that a schism did take place around this time’ (Prebish 2005a, 237). Cousins lends his voice: ‘Even if

3. See my remarks at http://sectsandsectarianism.googlepages.com (12 May 2008) > Names & Dates at Vedisa: http://sectsandsectarianism.googlepages.com/names%26datesatvedisa. 4. But see http://sectsandsectarianism.googlepages.com (12 May 2008) > Sanghabheda vs. Nikay- abheda: http://sectsandsectarianism.googlepages.com/sanghabhedavs.nikayabheda.

© Equinox Publishing Ltd 2008 212 REVIEW it is now clear that the schism between the Mahāsaṅghikas and the Sthaviravāda is not connected with the Second Council, it cannot have been long after’ (Cousins 2005a, 104). Nakamura, citing Midzuno, Gokhale, Basham, and Ingalls, also agrees: ‘Diff erent sects were already in existence during the reign of King Aśoka’ (Nakamura 1996, 100). Lamotte is more cautious, saying: ‘The date of the schism has not been precisely established, but everything leads to the that it had been prepared over a long period and only came to a head in Aśoka’s era’ (Lamotte 1976, 292). Despite this mass of authority, I think the First Schism must be placed after Aśoka. If I am correct, a major redraft of our map of early Buddhist history is needed. We cannot cast away such a consensus of authority lightly, and so I would like to start by explaining in detail exactly why our authorities believe the fi rst schism must have been before Aśoka. But unfortunately I have yet to discover explicit and detailed justifi cation for this . Every account I have read, dating back to Oldenberg in 1879 (Oldenberg 1879, xxxii), assumes this is the case, with sketchy if any attempt at justifi cation. But such a conclusion must be based on the direct accounts of the schisms, so we should briefl y survey these fi rst. There are, of course, many other sources that we might refer to, but four in particular are central to any account. This is because, while the traditions speak of ‘18’ early schools, in fact there were four major groups of schools, which I refer to as Mahāsaṅghika, Vibhajjavāda,5 Sarvāstivāda, and Puggalavāda. The Lakṣmī has seen fi t to bestow one of her rare graces upon us by making available at least one source representing each of these four major groups of schools. The Pali Dīpavaṁsa represents the Mahāvihāra (Theravāda/Vibhajjavāda); the Sarvāstivādin ’s6 Samayabhedoparacanacakra is available in three Chinese (T no. 2031, T no. 2032, T no. 2033) and one Tibetan (see Lamotte 1976, 529–31 for details); the third list of Bhavya’s Tarkajvala (‘Bhavya III’), preserved in Tibetan, stems from the Puggalavādins (trans. Rockhill, pp. 186ff . See Lamotte 1976, 535); and the Mahāsaṅghika Śāriputraparipṛcchā survives in Chinese (T no. 1465). Each of these sources contains two parts: a ‘head’ (detailing the First Schism, when, where, why, who, etc.); and a ‘tail’ (describing the subsequent evolution of the ‘18 schools’). In each case the ‘tail’ is fairly similar, allowing for a few sectarian adjustments and the odd confusion, while the ‘heads’ are for the most part radically diff erent. Now, an obvious initial historical thesis would be that the ‘tails’ record the sectarian situation at the period the texts were composed, while the ‘heads’ record the attempts by the sects to authorize their own school within this context. Later on I will try to show how the texts confi rm this, but fi rst we must examine the

5. See Cousins (2001). I use this as a convenient term for the schools closely related to the Mahāvihāravāsins, but it should be noted that there is no evidence that the Mahāvihāravāsins themselves used this term to denote a group of schools. The Dīpavaṁsa shows their sense of radical isolation, and by ‘Vibhajjavāda’ they meant themselves alone (as shown in the texts quoted at Cousins (2001, 134–5, 136, 139, 141, etc.)). See Sects & Sectarianism, chapters 6 & 7. 6. Cousins (2005b, 53) is too sceptical in dismissing Vasumitra’s authorship.

© Equinox Publishing Ltd 2008 SUJATO THE DATE AND CAUSE OF THE FIRST SCHISM 213 opposite thesis, upon which our authorities apparently rely: that the ‘heads’ sup- ply genuine historical information regarding the date of the First Schism. Since I can’t fi nd any serious argument in favour of the pre-Aśokan dating of the First Schism, it seems I must play Māra’s advocate and try to imagine what such an argument might look like. Our four primary texts place the First Schism in relation to Aśoka; either before (Dīpavaṁsa and Bhavya III), during (Vasumitra), or after (Śāriputraparipṛcchā). The sources that place the schism before Aśoka include the Dīpavaṁsa, which stems from the more historically reliable Sri Lankan tradi- tion. Vasumitra places the event in the time of Aśoka, but this seems to be a con- fusion stemming from its use of the ‘short chronology’, which underestimates the period between the Buddha and Aśoka. The calendar date of the schism accord- ing to Vasumitra is about 100 AN, which roughly agrees with the Dīpavaṁsa (100+ AN) and Bhavya III (137 AN). Vasumitra, therefore, has the date approximately right, but following the tradition of his school, he thinks that this was the reign of Aśoka. This is because his ‘short chronology’ tradition confuses the ‘Kāḷaśoka’ of the Second Council with the famous ‘Dharmāśoka’ (Nattier & Prebish 2005, 224). The Śāriputraparipṛcchā relates the same events, but in placing the schism later has become confused in its chronology, as is evident from the corrupt state of the text (Lamotte 1976, 172). Hence, despite apparent diff erences, all the main textual sources stem from a unifi ed tradition of a pre-Aśokan schism, with the diff erences due to later chronological confusion. Presented in this way, the case seems simple enough. But under scrutiny the evidence disintegrates little by little, until we are left with a very naked emperor. One major problem with this theory is that the Aśokan edicts make no men- tion of the existence of sects. In particular, the so-called ‘Schism Edicts’ actually state that the Saṅgha is unifi ed, not schismatic.7 The scholars have have suggested various scenarios happening behind the evidence: Aśoka didn’t really know what was going on; or he was referring to a mere party dispute among the Theravādins; or he was just boasting; or he was concerned only with resolving local disputes. It is scarcely an exaggeration to say that the history of modern interpretation of the ‘Schism Edicts’ is driven by the need to explain away the non-appearance of the sects. My own suggestion is that Aśoka did not mention the sects for a very good reason: they didn’t exist.

Why the Dīpavaṁsa is wrong

The Dīpavaṁsa, the earliest Sri Lankan chronicle, is perhaps the most infl uential text in our fi eld.8 Much of this text preserves valuable historical information, yet this is not suffi cient to prove its reliability for dating the First Schism. The

7. Sāñchī: [saṁ](ghe)e*[sa]*mag(e) kate; : (sa)ma(ge)* kate* saṁghas[i]. 8. I ignore the Mahāvaṁsa, as I believe that in the sections we are concerned with it does not use any archaic sources apart from the Dīpavaṁsa itself.

© Equinox Publishing Ltd 2008 214 BUDDHIST STUDIES REVIEW

Dīpavaṁsa is the only early source that connects the schism with the events of the Second Council. Countless academic and popular accounts have constructed out of this tenuous linkage a grand narrative of the dispute between the rigorist, legalist, monastic Theravādins versus the laxist, open, lay-oriented Mahāsaṅghikas, the supposed forerunners of the Mahāyāna. The heroes of this contemporary myth are, of course, whoever you happen to like. But the connections between the Second Council and the First Schism in the Dīpavaṁsa are slender and artifi cial, and a little refl ection is enough to discredit them. First of all, the text preserves some nice little markers telling us that the entire section dealing with the schisms is an interpolation. A ‘-ti’ signifying ‘end of section’ appears between the Second Council and the ‘head’ of the schism narrative (Dv 5.29). Then just after the end of the ‘tail’, following the origina- tion of all 18 schools after the Second Council, we are told that the ‘true ascetic’ (Moggaliputtatissa) will arise 118 years in the future (Dv 5.55). The fi gure of 118 years obviously looks forward from the end of the Second Council, and hence this saying must have originally occurred at the end of the section on this. The entire schism narrative has no organic connection with the Dīpavaṁsa’s sources, and so we cannot assume that it appeared in this context any earlier than the compila- tion of the Dīpavaṁsa itself. Second, the themes of the Second Council and the supposed ‘Great Council’ of the Mahāsaṅghikas have nothing in common. The Second Council was a dispute about Vinaya, where the texts were agreed upon but the practical implementation was disputed. The ‘Great Council’ says nothing of Vinaya, and is concerned solely with matters of grammar and textual redaction (Dv 5.32–38). Third, the texts spoken of at the Second Council, which were disputed by the Mahāsaṅghikas, all belong to the later strata of the Pali canon, and no modern scholar would wish to argue that they existed by 100 AN.9 Generally they would be regarded as pertaining to the period, say, 200–400 AN. The critique of such texts by the Mahāsaṅghikas shows that they were the fi rst reliable text-critical schol- ars in Buddhism (and, incidentally, shows that concerns over textual authenticity are not a modern ‘Protestant’ imposition on the Indic tradition). Fourth, it is unlikely that a major dispute over dialect, as implied by this account, could have taken place so early and in such a central region. At that time and place, it is likely that all Buddhists used closely related dialects. The only example we have of the Vajjiputtakas’ language – the 10 points of the Second Council – show that they used a type of ‘pre-canonical’ or ‘Magadhan’ dialect, as did the old Sthaviras (Lamotte 1976, 131–132). Fifth, the Mahāsaṅghika Vinaya (T no. 1425, pg. 493a25–c11) views the events of the Second Council from the same perspective as the Sthavira . In particular, the decisive issue at the Second Council was whether a could accept money. But the rules forbidding use of money by are found in the

9. Dv 4.76: Parivāraṁ athuddhāraṁ abhidhammaṁ chappakaraṇaṁ / Paṭisambhidañca niddesaṁ ekade- sañca jātakaṁ

© Equinox Publishing Ltd 2008 SUJATO THE DATE AND CAUSE OF THE FIRST SCHISM 215

Mahāsaṅghika Vinaya just as in all the others, and so the Vajjiputtakas are accord- ingly condemned. The fact that the Mahāsaṅghika account is shorter than the others is not relevant, for the Mahāsaṅghika Vinaya regularly abbreviates nar- rative sections (Frauwallner 1956, 52, 198–207). This is a literary feature of the particular redaction translated into Chinese, and has nothing to do with sectar- ian bias.

Why the Sinhalese Vinaya Commentary is right

The Sinhalese tradition itself preserves an alternative account of this period which says nothing of the schisms or the existence of schools. This account is found pri- marily in the Sinhalese Vinaya Commentaries, available in the Pali Samantapāsādikā of , and in the Chinese of the Sudassanavinayavibhāsā.10 This narrative tells us that, after the Second Council, the Elders considered what the dangers for the sāsana would be. They used their special powers and deter- mined that problems would arise long in the future, during the reign of Aśoka. Accordingly, they fl ew to the Brahmā realm and begged the Brahmā Tissa to take , ordain as a , and protect the sāsana. While all these elaborate preparations were going on, there was no suggestion that they knew of the cas- cade of schisms that was supposedly starting to happen at the same time. In fact, this account goes on to tell the elaborate story of the Saṅgha under Aśoka, the various problems and successes of that time, the holding of the Third Council to expell non-Buddhist heretics from the Saṅgha, and the sending out of missionar- ies, all without a single mention of the existence of diff erent sects or the occur- rence of a schism. The only possible suggestion of sectarianism here is the Samantapāsādikā’s claim that Moggaliputtatissa at the Third Council recited the , which discusses at length the doctrinal disputes current in the Buddhist community. But even on the basis of the Pali alone we should recognize this as an interpola- tion. Nothing up until this point mentions disputes among Buddhists. The text has been entirely concerned with the intrusion into the Saṅgha of non-Buddhists, identifi ed as Nigaṇṭhas, Ājīvakas, Jāṭilas, and so on. This is confi rmed by com- parison with the Sudassanavinayavibhāsā, which in its otherwise nearly identical account omits all mention of the Kathāvatthu (T no. 2034, pg. 684b.12).

10. I use a reconstructed Pali title for the Chinese 善見律毘婆沙 (T no. 2034). Bapat and Hirakawa follow the Taisho editors in treating this as a translation of the Samantapāsādikā, although they note the presence of many diff erences from the existing Pali text. But Guruge is surely correct in arguing that the Sudassanavinayavibhāsā is not a translation of the Samantapāsādikā; while the two have much in common, the diff erences are too far-reaching. However, in the section we are dealing with they are very close. Where they diff er, the Sudassanavinayavibhāsā regularly shows more archaic traits. When these texts do not diff er I refer to them generically as the ‘Sinhala Vinaya Commentary’.

© Equinox Publishing Ltd 2008 216 BUDDHIST STUDIES REVIEW

How could the Sinhalese have ended up with two such contradictory accounts? The answer to this question helps us understand the perspectives of the Sinhalese redactors. One means at our disposal is to compare our sources with the few facts we can glean from the archaeological record. There are two major pieces of inscriptional evidence that derive from the early period of Indian Buddhism: the Aśokan edicts and the reliquaries at Vedisa. Strikingly, both of these confi rm the evidence found in Sinhalese Vinaya Commentary. The Vedisa inscriptions mention the names of several monks, ‘teachers of the whole Himalaya’, who the Sinhalese Vinaya Commentary says were sent as to the Himalaya soon after the ‘Third Council’ (Willis 2001). And Aśoka’s so-called ‘Schism Edicts’ mention an expulsion of corrupt , which many scholars have identifi ed with the events prior to the ‘Third Council’ (Sasaki 1989). We should also note that Moggaliputtatissa’s sending out of missionaries has often been compared with Aśoka’s sending out of Dhamma- ministers (Wynne 2003, 16–21; and that the Sri Lankan archaeological record is in general agreement with the picture of the missions (Bandaranayake 2008). The northern sources also confi rm certain details of this picture. While not retaining such a detailed account of the missions, they do mention the missions of to (Lamotte 1976, 205) and Majjhantika to Kaśmīr. The Mahāsaṅghika Vinaya also refers to the expulsion of schismatic monks, which seems to refer to the same events as the Third Council and the Schism Edicts (T 1425 440c–441a; trans. Sasaki 1989, 193–4). It is therefore clear that, legendary elaborations notwithstanding, the Sinhalese Vinaya Commentary is based on a genuine record of the Aśokan period, which must have been transmitted in Sri Lanka from early times. The agreement between the Pali and Chinese versions of this text suggests that this passage was part of the ancient Sinhalese commentarial tradition.

The Sri Lankan background

The First Schism account in the Dīpavaṁsa, however, entirely lacks corroboration from either archaeology or northern sources. Since it appears to be a late inter- polation into the Sinhalese tradition, where did it come from? The ‘tail’, detailing the 18 schools, derives from a Vasumitra-style text (Cousins 2001, 151). As for the ‘head’, it depicts a situation where the ‘Sthaviras’ are arguing with the Mahāsaṅghikas about language and textual redaction. One of the best-attested facts we know about the Mahāvihāra tradition is that they argued at length with the Mahāsaṅghikas. The Kathāvatthu details hundreds of doctrinal disputes. While the text itself tells us nothing about who these discussions involved, the com- mentary attributes each tenet to one or more of the ‘18’ schools; about half of these are the Mahāsaṅghikas and their sub-schools, mainly the Andhra communi- ties. Excavations have revealed that not only did the Mahāsaṅghika have several

© Equinox Publishing Ltd 2008 SUJATO THE DATE AND CAUSE OF THE FIRST SCHISM 217 communities in Andhra, but the Sinhalese also maintained a branch monastery there, no doubt because it was one of the major trading ports between Sri Lanka and the mainland. The Mahāsaṅghika clearly rejected the Pali Abhidhamma, and no doubt also many of the other later works of the Mahāvihāra, as detailed in the Dīpavaṁsa. In addition, we now know that the Mahāsaṅghika generally used the distinctive ‘’ dialect, which would indeed appear to anyone from a Pali background as mixed up and full of errors. Thus the situation described in the Dīpavaṁsa, while incongruous in its setting after the Second Council, perfectly depicts what we know about the situation in the middle period (say, the early centuries CE). It seems that the authors of the Dīpavaṁsa, seeking archaic authority for their own school in the face of international competition, created a myth of origins. This was not a historical account of the past based on evidence; rather, it extrap- olates the present conditions into a timeless past. We should not think of this as a cynical fabrication. It is a manifestation of cyclic time, as graceful and irrefu- table as the seasons themselves. Through all Buddhist literature, in hundreds of Jātakas, the question is put: ‘Why are things as they are now?’ And the answer is always the same: ‘Because that is how they were in the past’. And so if we want to know: ‘Why are the Mahāvihāravāsins now (in the fourth century) arguing with the Mahāsaṅghikas about texts and language?’, the answer is inevitable: ‘Because that is how it was in the past, from the very birth of these two schools; in fact, that is what must have caused the schism in the fi rst place’. The authors of the Dīpavaṁsa have merely given concrete literary form to this truth, which is inevitable precisely because it is mythic. The authors of the Dīpavaṁsa needed to root their authority in the deepest past they could, just as they rooted the origins of Sinhalese Buddhism in the refl ections of the Buddha immediately after his Awakening. The canonical texts made it clear that there was as yet no schism in the time of the Second Council, so the schism must be later than that. But then there is a problem with the Aśokan stories. The same question must have occurred to them as has troubled modern commenta- tors: Why are no schools mentioned in the time of Aśoka? While we might prefer to see Aśoka as the impartial supporter of all religious traditions, no doubt the Elders of the Mahāvihāra found it congenial to believe that he was the exclusive patron of their tradition alone. This meant that all the other seventeen schools had to be ‘schismed off ’, so to speak, and out of the way before Aśoka. Since the appear- ance of all eighteen schools would have required some time, then, the First Schism must have followed immediately after the Second Council. Such is the inexorable conclusion that follows from the Mahāvihāravāsins’ unique perspective. These two accounts of the early period refl ect two epistemological means. The Vinaya Commentary derives its Aśokan narrative from ancient tradition, remem- bered and elaborated down the years. The Dīpavaṁsa account of the First Schism is inferred according to the cyclic logic of myth. From the perspective of the Elders, each of these would be valid means of knowledge, and the contradictions, if not too blatant, could be easily ignored.

© Equinox Publishing Ltd 2008 218 BUDDHIST STUDIES REVIEW

Vasumitra

Vasumitra places the schism at the time of Aśoka, which for his short chronol- ogy is 100+ AN, and attributes the schism to a dispute on the ‘fi ve points’ at Pāṭaliputra. The similarity of the date of the First Schism in the Dīpavaṁsa and Vasumitra gives the superfi cial impression that they support each other. Since the Dīpavaṁsa has in this regard no historical credibility, however, its agreement with anything means nothing at all. Also, the two accounts are otherwise totally diff er- ent: diff erent place, diff erent issues, diff erent procedures, diff erent people. And the similarity of dates is really no more than a coincidence, stemming from the use of diff erent chronologies. For the Dīpavaṁsa, 100 AN means 118 years before Aśoka, while for Vasumitra 100 AN means in the time of Aśoka. We have seen that for the Dīpavaṁsa the important thing is not the recording of accurate calendar dates, but the establishing of sectarian authority through the charisma of Aśoka. In the same way the Sarvāstivādin mythos depends on leveraging the memory of Aśoka in favor of their own school. How so? The Mahāvibhāṣā gives a detailed account of the First Schism, which fi lls out Vasumitra’s brief account.11 This account tells how, after the evil Mahādeva prop- agated his ‘fi ve points’, the good monks (all Arahants) were originally persecuted by the king (presumably Aśoka), who tried to drown them in the Ganges; but they fl ew into the air and off to Kaśmīr. The king was mortifi ed to see this clear dem- onstration of their spiritual might, and begged forgiveness, off ering them many (‘which are still thriving today’), and indeed went so far as to off er them the whole of Kaśmīr. As is well known, the Mahāvibhāṣā was the defi nitive work of the Kaśmīr Vaibhāṣika branch of the Sarvāstivāda, who under the patronage of the great king Kaniṣka (127+ CE) tried to establish themselves as the central Buddhist ortho- doxy. But this ambitious eff ort faced a critical problem: since the Buddha never went to Kaśmīr, how can the Kaśmīris claim to be the most orthodox school? A narrative explaining the corruption in the centre and the purity of the border- ing districts would be just the ticket; and Kaniṣka must have enjoyed the implied parallels between himself and Aśoka. Just as the Mahāvihāra transmitted two accounts, one based on early period history, the other a middle period myth, so too the Sarvāstivādins, in addition to this late account, preserved a much earlier memory of the founding of Kaśmīr Buddhism by Majjhantika (see Lamotte 1976, 226). In this story, the drama stems not from confrontation with rival , but from subduing yakkhas, a heroic endeavour which permeates legends of this time. The historical dimension of such stories records the and assimilation of local sacri-

11. T no. 1545, 510c23–512, a19. Translated at http://sectsandsectarianism.googlepages.com (12 May 2008) > 5. The Three & the Five Theses: http://sectsandsectarianism.googlepages. com/5.thethreesins%26thefi vetheses. Authorship is uncertain, but tradition links Vasumitra with the compilation of the Mahāvibhāṣā.

© Equinox Publishing Ltd 2008 SUJATO THE DATE AND CAUSE OF THE FIRST SCHISM 219

fi cial into the Buddhist fold, especially with the setting up of yakkha/yakkhinī shrines inside the monasteries to function as cultic centres. Like the Mahāvihāra records, the early period story deals with the tensions between Buddhism and primitive , while the middle period story deals with tensions among schools of Buddhism. It is thus clear that during the middle period, Kaśmīr Sarvāstivādins, like the Mahāvihāravāsins, constructed a mythos which, while ostensibly telling of events in the far past, in fact represented the contemporary situation. This is, of course, generally true of all myth (as also its mirror, ). The purpose of the myth is to associate the Sarvāstivādins of Kaśmīr with the root-Sthaviras in the time of Aśoka. The calendar date is irrelevant to this mythos, and has merely been inserted to give historical fi xity to an event which, from Vasumitra’s point of view, must have happened around that time.

Bhavya III

Bhavya, writing in the seventh century, compiled several sources explaining how the schools came about. The most relevant of his sources is his third list, which is generally regarded as stemming from the Puggalavāda group of schools. Like the Dīpavaṁsa, Bhavya III places the schism before Aśoka. But the events have nothing to do with the account of the Dīpavaṁsa. Rather it attributes the schism to the ‘fi ve points’, as does Vasumitra. But due to diff erent ways of counting the years between the Buddha and Aśoka, the dating is hoplessly confused: Vasumitra places the events at Aśoka, which it says is 100+ AN; Bhavya III places the same events before Aśoka, but the date is 137 AN. In addition, circumstantial details such as the name of the king and the monks diff er in the two versions. We have little literature of the Puggalavādins, and no mythic context apart from that provided by this text itself. This makes this account harder to assess, but no doubt it was pressed into service to authorize the Puggalavāda school. We note that it is the two latest sources (Bhavya III and Dīpavaṁsa) that place the schism pre-Aśoka. It seems that as time goes on the schism date becomes earlier, a natural feature of the mythic process.

Śāriputraparipṛcchā

This text, available only in a less than satisfactory Chinese translation of about the third century CE, is framed in the style of a sūtra, with the Buddha predicting to Sāriputta the future course of the sāsana. It has been frequently cited by modern studies, but unfortunately these have relied on only a few paragraphs, divorced from the overall context (Nattier & Prebish 2005). This is perhaps because of the impression conveyed by Lamotte’s claim that the text was chronologically con- fused. Lamotte claimed that the text is:

© Equinox Publishing Ltd 2008 220 BUDDHIST STUDIES REVIEW

so obscure that it allows for the most diverse interpretations. After hav- ing narrated at length the persecution by the Śuṅga Puṣyamitra, the text, going back to the past, speaks of events which took place under a king whom it does not name, but who, from the evidence of other parallel texts which we shall quote, can be none other than Aśoka the Maurya. (Lamotte 1976, 172) But, leaving aside its obvious mythical nature and several textual problems, the Śāriputraparipṛcchā is not confused in its own chronology. Lamotte’s that the text ‘goes back to the past’ is purely a result of trying to make it agree with Vasumitra, etc., and has no basis in the text itself. The ascription of the schism to a date after Aśoka is no accident, but is inherent in the logic of the text. First the narrative speaks of the usual fi ve ‘Masters of the Dhamma’: Kassapa, Ānanda, Majjhantika, Śāṇavāsin, and Upagupta. This brings us up to the time of Aśoka. Clearly there can have been no schism so far, as the list of patriarchs is identi- cal with the (Mūla) Sarvāstivāda tradition. After Aśoka we are told of the perse- cutions under his ‘grandson’ Puṣyamitra; again, this is entirely in accord with the (Mūla) Sarvāstivāda tradition (See Lamotte 1976, 354). Because of the per- secutions, the sacred texts were removed by and taken to the heaven realms for safety. This motif, similar to the strategies used to authenticate the Pali Abhidhamma and the Mahāyāna sūtras, tries to us of the authenticity of the texts; but it betrays an unconscious recognition of the frailty of the transmis- sion. Following Puṣyamitra there was a good king, under whom the sāsana fl our- ished and the texts were accordingly returned to earth. There appeared a learned monk who undertook an ‘enlargement’ of the Vinaya. This prompted a dispute, which was settled by taking a vote mediated by the (unnamed) king. The minority voted for the new Vinaya and they became the Sthaviras; the majority voted for the old Vinaya ‘of Kassapa’ and they became the Mahāsaṅghika. After this ‘head’ the text presents the ‘tail’, detailing the emergence of all the 18 schools.12 In evaluating the chronology of the Śāriputraparipṛcchā, we should fi rst note that the text dates the emergence of Mahāsaṅghika schools from around 200 AN. In the ‘short chronology’ of the text, this brings them, say, around 150 BCE, which roughly agrees with the date of Puṣyamitra’s death, around 151 BCE. Thus the internal chronology of the text is not only coherent, it agrees with the his- torical data. In further support of the Śāriputraparipṛcchā’s chronology, we note the inter- ference of the king in the aff airs of the Saṅgha – surely a post-Aśokan motif – and the presence of written texts, taken for granted as the essential medium of trans- mission. All this confi rms that the story cannot date from before the end of the early period; more likely, the early part of the middle period.

12. Trans. at http://sectsandsectarianism.googlepages.com (12 May 2008) > 4. Monster or Saint?: http://sectsandsectarianism.googlepages.com/4.monsterorsaint%3F.

© Equinox Publishing Ltd 2008 SUJATO THE DATE AND CAUSE OF THE FIRST SCHISM 221

The only problem with the chronology is that it disagrees with other texts, espe- cially Vasumitra, on the date of the First Schism. But not only is Vasumitra, as we have seen, unreliable on this point, the events described in the Śāriputraparipṛcchā are very diff erent from the other accounts. We cannot infer that the account of the root-schism is merely a confusion of Vasumitra. It is an independent story, telling the origin of the Mahāsaṅghika from their own perspective. Can we infer anything more about the contextual meaning of this account? While we do not have such a developed mythos as seen in the case of the Mahāvihāra and Sarvāstivāda, there are a few clues available. The most striking point is the emphasis on the of the fi ve Masters of the Dhamma. Our text does not merely mention these in passing, but specifi - cally invokes the head of this lineage at the crucial moment of schism, declaring allegiance to the Vinaya ‘of Kassapa’. We are familiar with this list from (Mūla) Sarvāstivāda sources, but its appearance here shows that these patriarchs were not sectarian. Of course, this is obvious as far as Śāṇavāsin, who is remembered in all the canonical Vinayas as an Elder of the Second Council. But the inclu- sion of Upagupta suggests that even this Aśokan patriarch was not connected with any particular school. Rather than being a sectarian fi gure, he is a regional one: Śāṇavāsin and Upagupta were the local saints of Mathura. This agrees with the curious status of Upagupta within the Theravāda tradition. He is entirely unknown to the Sinhalese tradition, which would fi t the case of a Mathuran Elder whose fame only became widespread in the early centuries CE; and yet he is well known in the folk of northern Theravāda (Burma, , , etc.), suggesting he is not excluded from Sinhalese Theravāda for doctrinal reasons so much as due to the pattern of dissemination of Buddhism: Upagupta travelled to ‘Suvaṇṇabhūmi’ via the northern road, and is a relic of pre-Theravāda Buddhism (Strong, 1994). Lamotte plausibly suggests that the fame of Śāṇavāsin and Upagupta coin- cided with the rise of Mathura as a Buddhist centre (Lamotte 1976, 331), which can be dated on epigraphic evidence to the second century CE. In this period we fi nd inscriptions referring to both the Mahāsaṅghikas and the Sarvāstivādins (Lamotte 1976, 523–5). I would suggest, then, that our text was composed in Mathura around this time, as a result of a local dispute or competition between these Buddhist communities. If this hypothesis has any cogency, it enables us to further speculate as to the nature of the ‘expanded Vinaya’ rejected by the Mahāsaṅghika. By far the largest of all Vinayas is the Mūlasarvāstivāda, which has close connections with Mathura (Frauwallner 1956, 24–37). Perhaps, then, the schism was associated with the radi- cal textual revisions required to produce this text. While we can never approach certainty in such matters, this hypothesis seems to agree well with the available evidence. The Śāriputraparipṛcchā, like all our sources, seems to be referring to events of its own time and backdating them to provide mythic authority for its school. Nevertheless, its dating seems more plausible than the other accounts. For one

© Equinox Publishing Ltd 2008 222 BUDDHIST STUDIES REVIEW thing, it stands much closer to the events it describes, being probably composed 200 or 300 years after the events, as opposed to 500 or more in the case of the Dīpavaṁsa, Vasumitra, and Bhavya III. Also, mythic texts typically seek author- ity through setting events in the deepest time possible, so we should expect that the version placing the schism later would be more reasonable. In addition, the Śāriputraparipṛcchā is less polemical than the other sources, indicating a healthier and more realistic attitude, and consequently less pressure to twist events to its own perspective.

The nature of sects

So far we have dealt with the sectarian accounts of the First Schism. While these are of course essential sources, it should really come as no surprise that they aff ord us more insight into the politics of the schools than into their historical origins. If we can assume, as an initial hypothesis, that these sources leave the date of the First Schism uncertain, we must turn to the indirect evidence. Perhaps what is merely implied or even unspoken will prove a better guide to the truth. But fi rst, I think we need to consider a little more carefully what we mean by a ‘school’ or a ‘’. In the sources we have seen above, the notion of a school was taken for granted, as the texts consciously set out to demonstrate the ori- gin and nature of the schools. But in the real world things are rarely so cut and dried. In reality we know little of the actual relationships between the schools, the extent to which monks and might move between them, or even whether all Buddhists even accepted the notion of schools. As we consider the indirect evidence, we need to carefully distinguish between a ‘sect’ and ‘sectarian tendencies’. The forces that led to the division into schools are many and complex, and can be traced even back to the Buddha’s time. But diff erences, contradictions, and tensions can be accommodated for a long time in a loose decentralized organization like the Buddhist Saṅgha. No doubt there was a degree of separation and divergence felt between Buddhist Saṅghas, which gradually resulted in a schismatic situation. Thus we cannot see the existence of sectarian tendencies as proof of the existence of schools. A ‘school’ in the mature sense requires, I believe, a Saṅgha’s conscious iden- tifi cation of themselves as a ‘distinct totality’. They must see themselves as dis- tinct from other Saṅgha groups in essential aspects, and view their own system as complete, adequate for a full spiritual life. This would involve a textual tradition, devotional centres, lineage of masters, institutional support, etc. We can speak of a school when these factors are there to a suffi cient degree for a particular por- tion of the Saṅgha to agree that they themselves constitute such a ‘distinct total- ity’, which will be formalized by the adoption of a name encapsulating a certain distinctive feature of the school. Let us survey, very broadly, the sources available that likely originate in the early period, and compare this with the evidence of the middle period, seeking

© Equinox Publishing Ltd 2008 SUJATO THE DATE AND CAUSE OF THE FIRST SCHISM 223 evidence for such a ‘distinct totality’. Within each period I shall consider the archeological evidence fi rst, as that can clearly be fi xed in time. The dates of all of the textual sources are questionable, and most of them probably straddle our divide. Nevertheless, I try to assign a place as best I can.

The early period (before the Common Era)

Here our main sources are the archaeological evidence of the Aśokan inscrip- tions and the Vedisa stūpas and inscriptions, the canonical literature, especially doxographical literature (Kathāvatthu and Vijñānakāya), and the Sinhalese Vinaya Commentary (which by its defi nite links with the archaeological evidence is proved to have roots in this period). We might also include the Aśoka legends which, while lacking such distinct archaeological confi rmation as the Vinaya Commentary, nevertheless may have at least some origins in this period. The Aśokan inscriptions (circa 250 BCE) do not mention any schools or any explicit occurrence of schism. When the edicts say the Saṅgha has been ‘made unifi ed’, this suggests that there has been some confl ict, but it falls short of estab- lishing that a schism had occurred. In any case, even if there had been a schism, the edicts assert that it had been resolved. Nor do the Aśokan edicts mention any sectarian , texts, or anything else that might even hint at the existence of schools. The main sect-formative factor at work here would appear to be the geographical spread of the Saṅgha, which was to become a powerful force in the evolution of distinct sectarian identities. The inscriptions on reliquaries retreived from the stūpas in Vedisa (circa 100 BCE) mention several sectarian-formative factors, such as local saints, local insti- tutions, and the name Hemavata, which at least at some time was taken to be the name of a school. So we might here, towards the end of the early period, be witnessing our fi rst evidence of an actual school. And yet the evidence is still not defi nitive: ‘Hemavata’ may be purely a geographical term here, presaging its adoption as the name of a school. The emergence of a local is a natural progression from the geographical spread under Aśoka, and there is no proof that the Vedisa community saw itself as distinct from other Buddhist communities. We remember that Mahinda had sojourned in Vedisa a century or so earlier while on his way to Sri Lanka, and there was no sectarian suggestion at that time. The canonical suttas, most of which were compiled in this period, show no sign of the existence of schools. There are occasions in the Āgama literature where the diff erent versions seem to refl ect the doctrinal positions of the schools, but these show no more than incipient tendencies. Similarly, the canonical Vinayas, while remarkably uniform in their main structures, display considerable divergences of detail, which points to the diff erentiation of local communities. However, they do not mention actual schisms and the existence of diff erent schools. The Abhidhamma literature displays a far more developed sectarian perspective, but even here the diff erences are more in the arrangement and structure of the works

© Equinox Publishing Ltd 2008 224 BUDDHIST STUDIES REVIEW than in explicit doctrinal diff erences. Such diff erences may be understood as the diff erent opinions and perspectives of the teachers or communities responsible for these works, rather than defi nitive diff erences between schools. The doxographical Abhidhamma texts evidence sectarian-formative factors more clearly, particularly the articulation of controversial doctrines that char- acterized certain schools. But there is no explicit acknowledgement of the exist- ence of schools, with the sole exception of the mention of the Puggalavāda in the Vijñānakāya.13 Thus from these texts alone we are unable to say to what extent they refl ect either discussions within a Buddhist community that contributed to the emergence of sects, or else explicit sectarian positions. Even the Milindapañha, a paracanonical work from the end of this period, appears more as a general Buddhist work rather than a sectarian one. It evidently originated in a northern school and was adopted by the Mahāvihāra, who did not feel the need to change certain that disagreed with the supposed doctrines of the school, such as the idea that space is unconditioned. The Sinhalese Vinaya Commentary was fi nalized much later, but as we have seen its connection with the archaeological evidence proves that some portions must stem from genuine historical records. This details an extensive account of the period in question, and fi nds no reason to mention even in passing the exist- ence of any schools. Likewise the Aśokavadāna, Aśokarājasūtra, Divyavadāna, etc., give many elaborate stories of Aśoka without involving the schools. Of course these legendary works were much augmented over time, but if anything this strengthens our argument: since these texts were doubtless fi nalized in the sectarian period, there must have been a temptation to explicitly associate Aśoka with their own school. But this was not done, at least so far as I have seen. Summing up this period, there is no evidence unambiguously belonging to the early period that mentions or implies the existence of schools. We fi nd only the mention of various forces that lead to sectarian formation, never to the actual schools that resulted from these forces. This remains true even if we allow texts that are actually fi nalized later, but which probably have roots in this period.

The middle period (after the Common Era)

For this period our primary sources are the inscriptional evidence, the vari- ous schism accounts, and the śāstra/commentarial literature. The inscriptions, starting in Mathura around 100 CE, regularly mention the names of schools. The śāstras (e.g. Abhidharmakośa, etc.) and commentaries (e.g. Kathāvatthu-aṭṭhakathā, Mahāvibhāṣā, etc.) regularly mention schools by name, and discuss their doctrines. The schism accounts, as we have already discussed, are explicitly concerned with the of sectarian origins and development. Regarding the names of the

13. 補特伽羅論 (T no. 1539, p. 537, b2)

© Equinox Publishing Ltd 2008 SUJATO THE DATE AND CAUSE OF THE FIRST SCHISM 225 schools, their interrelationships, and their distinctive doctrines these sources are, all things considered, reasonably consistent, as we would expect since they are describing contemporary conditions. Throughout the middle period, then, the existence and basic nature of the schools is taken for granted and constitutes an essential component in all our sources. This stands in stark contrast with the sources from the early period. But the accounts of the origins of the schisms, already in the far distant past from their own perspective, are a mass of contradictions. Of the three schism accounts that supply us with suffi cient information (Śāriputraparipṛcchā, Vasumitra/Mahāvibhāṣā, Dīpavaṁsa), we have seen that the primary function of the accounts was not to record history but to authorize their own school. I believe this provides suffi cient reason to explain how the schools came up with their vari- ous datings of the First Schism. Of course, this does not prove that the dates in these texts are all wrong. It is quite possible and in fact very common to construct a mythology out of real events. But given the evident contradictions, I think it is sheer naïvity to use the dates given in these texts to reach any simple historical conclusions. Like all myths, they are describing the situation in their own time (a situation of sectar- ian Buddhism) and backdating that in search of archaic authorization. I conclude that various separative forces gathered momentum through the early period and manifested in the emergence of ‘schools’ towards the end of the early period, as depicted in the Śāriputraparipṛcchā (and various Chinese and Tibetan works). As the question of sectarian identity became more conscious, mythic accounts of the schisms emerged in the middle period.

PART 2: THE CAUSE

The fi ve points

This topic, too, has been dealt with at length by the scholars. Without recapping the long discussions on the topic, I believe I can contribute some textual sources which, so far as I am aware, have been hitherto overlooked, and yet have a deci- sive importance. What was the cause of the basic division? As we have seen, our texts off er us diff erent reasons: either the fi ve points (Vasumitra, Bhavya III) or textual argu- ments (Dīpavaṁsa, Śāriputraparipṛcchā).14 Of course, there were many other fac- tors, and in general perhaps the most powerful factor was simply geographical dispersal. Nevertheless, the existence of diff erent sects in the same region – as in

14. Cousins (2005b, 67), Nattier & Prebish (2005, 223) and Prebish (2005b) try to show that the cause of the schism was Vinaya rather than Dhamma. My reasons for disagreeing with them are at http://sectsandsectarianism.googlepages.com (12 May 2008) > Dhamma or Vinaya?: http:// sectsandsectarianism.googlepages.com/dhammaorvinaya%3F.

© Equinox Publishing Ltd 2008 226 BUDDHIST STUDIES REVIEW

Andhra or Mathura – shows that more than geography was involved. As always, we proceed in the knowledge that the causes were doubtless more complex and multi-facted than we can ever know. On the face of it, the fi ve points seem to off er the most plausible explanation. They are accepted as the reason by the Sarvāstivāda and Puggalavāda groups of schools; in addition, the Mahāvihāravāsin Kathāvatthu, while not attributing the schism to this cause, preserves an early polemic against the Mahāsaṅghikas based on exactly the same fi ve points. The fi ve points are as follows:15 1. An Arahant can have wet dreams (be seduced by , or have semen cov- ertly conveyed while they are asleep). 2. An Arahant can have doubts. 3. An Arahant can lack knowledge. 4. An Arahant can be led by another. 5. The path can be aroused by exclaiming: ‘What suff ering!’ The diffi culty with attributing the First Schism to these theses is that, while they are discussed and refuted among the Sthavira group of schools, the Mahāsaṅghikas themselves do not claim to adhere to these fi ve points – or do they? The most shocking thesis is obviously the fi rst one, and like other lists, such as the ‘ten points’ of the Vajjiputtaka monks refuted at the Second Council, it is likely that this one main point was augmented by others to fi ll out the list.16 Given its indelicate nature, it is unsurprising that the Mahāsaṅghikas did not choose to make a prominent display of the matter. But our salvation is at hand, for this topic clearly relates to one of the major rules included in all the Vinayas.

The Vinayas and dreams

As so often in Buddhist controversies, the problem arises because of a grey area in the canonical texts, in this case the fi rst bhikkhu saṅghādisesa. Saṅghādisesa is the second most serious class of off ence in the Vinaya. While the most serious class of off ences, the pārājikas, entail immediate and permanent expulsion from the Saṅgha, saṅghādisesa requires a period of rehabilitation involving loss of sta- tus, confession of the off ence to all bhikkhus, and similar mild but embarassing penances. The basic rule for saṅghādisesa 1 is identical in all existing pāṭimokkhas: ‘Intentional emission of semen, except in a dream, is a saṅghādisesa’. In the Pali, the background is this. First the rule was laid down simply for ‘intentional emis- sion of semen’. Then a number of bhikkhus had gone to sleep after eating delicious food, without , and had wet dreams. They were afraid they had com-

15. Discussion and various sources in Nattier & Prebish (2005, 208–13). 16. Possibly the number fi ve is intended to echo the fi ve theses of .

© Equinox Publishing Ltd 2008 SUJATO THE DATE AND CAUSE OF THE FIRST SCHISM 227 mitted an off ence. The Buddha said: ‘There is intention, but it is negligible’.17 Thus there is no off ence for a wet dream, but this is a practical concession for Vinaya purposes, not an admission that there is no ethical content to wet dreams. The point is made clear in Kathāvatthu 22.6, where the Mahāvihāravāsin specifi cally refutes the proposition (attributed by the commentary to the Uttarapāthakas) that dream is always ethically neutral. The Pali rather curiously repeats the story of the mindless, greedy monks emitting semen as a pretext for making an allowance for using a sitting cloth in order to prevent the dwelling from being soiled (Pali Vinaya I.294). Why such a cloth should be called a ‘sitting cloth’ (nisīdana) is unclear, and the use of such a small cloth rapidly proves inadequate, so the Buddha allows a sleeping-cloth ‘as large as you like’. But this passage, which appears to spring from the same origin as the saṅghādisesa story, adds some emphatic messages. Those, Ānanda, who fall asleep with mindfulness established and clearly comprehending do not emit impurity. Even those ordinary people who are free from lust for sensual pleasures, they do not emit impurity. It is impossible, Ānanda, it cannot happen, that an Arahant should emit impurity.18 The text goes on to list fi ve dangers of falling asleep unmindfully: One sleeps badly, wakes badly, has nightmares, devas don’t protect one, and one emits semen. Those who sleep mindfully may expect the corresponding fi ve benefi ts. This list of fi ve dangers/benefi ts occurs in similar contexts in the Sarvāstivāda,19 ,20 and Mahīśāsaka21 Vinayas. The Sarvāstivāda moreover adds the following: Even if a bhikkhu who is not free of greed, hatred, and delusion sleeps with unconfused mindfulness and unifi ed he will not emit semen; still less a person free from lust.22

17. Pali Vinaya III.112: ‘Atthesā, bhikkhave, cetanā; sā ca kho abbohārikā’ti. 18. Pali Vinaya III.112: ‘Ye te, Ānanda, bhikkhū upaṭṭhitassatī sampajānā niddaṁ okkamanti, tesaṁ asuci na muccati. yepi te, Ānanda, puthujjanā kāmesu vītarāgā tesampi asuci na muccati. aṭṭhānametaṁ, Ānanda, anavakāso yaṁ arahato asuci mucceyyā’ti. 19. 一者無難睡苦。二者睡易覺。三者睡無惡夢。四者睡時善神來護。五者睡覺心易入善覺觀法 (T no. 1435, p. 197, a18–20). The last is diff erent: one easily enters wholesome thoughts. 20. 1. Nightmares; 2. Not guarded by devas; 3. Mind doesn’t enter thought of Dhamma; 4. One does not gain perception of light; 5. One emits semen. (一者惡夢。二者諸天不護。三者心不入法。 四者不思惟明相。五者於夢中失精 (T no. 1428, p. 579, b25–27)) 21. 1. Nightmares; 2. Not guarded by devas; 3. Not gain perception of light; 4. No thought of Dhamma in mind; 5. Emits semen. (一者惡夢二者善神不護三者不得明想四者無覺法心 五者失不淨 (T no. 1421, p. 10, b22–24)). This is identical with the Dharmaguptaka, except items three and four are swapped. 22. 比丘有婬怒癡未離欲。不亂念一心眠。尚不失精。何況離欲人 (T no. 1435, p. 197, a20–22, also T no. 1435, p. 197, a20–22).

© Equinox Publishing Ltd 2008 228 BUDDHIST STUDIES REVIEW

The Mahīśāsaka adds a similar statement: If one who is not free from greed, hatred, and delusion goes to sleep with mind distracted and confused, they will emit semen; even if unable to be free, going to sleep with established mindfulness, one will not com- mit that fault.23 These are similar to the statements found in the Pali Vinaya, but I have found nowhere else that declares so emphatically that it is impossible for an Arahant to emit semen in a dream. The Mūlasarvāstivāda Vinaya, while preserving an identical saṅghādisesa rule, gives only a brief, formulaic origin story, and no statement that one emits after falling asleep mindlessly, although it does speak of having sensual desire while in the dream.24 This suggests that nocturnal emissions are a product of defi le- ments, but is much less explicit than the other Vinayas on this point. The whole rule is dealt with relatively briefl y; but this is typical of this section of this Vinaya, so the brevity is more likely to be a mere literary characteristic than a sectarian diff erence. Thus all the Vinayas preserve the same rule against emitting semen. With the exception of the Mūlasarvāstivāda, the Sthavira schools all contain strong admo- nitions emphasizing that wet dreams occur because one goes to sleep unmind- fully. The Mahāvihāravāsin, Sarvāstivādin, and Mahīśāsaka in addition say even an unenlightened person can prevent wet dreams by mindful sleeping, still more an enlightened one. The Mahāvihāravāsin alone explicitly declares that it is impossible for an Arahant to emit semen. In the Mahāsaṅghika Vinaya the origin story is quite diff erent to the Mahāvihāravāsin. After the initial laying down of the rule, there were two train- ees (i.e. ariyas but not Arahants) and two ordinary people who had wet dreams. They doubted and told Sāriputta, who told the Buddha. The Buddha said: Dreams are unreal, not true. If dreams were real, one who practised the holy life in my Dhamma would not fi nd liberation. But because all dreams are untrue, therefore, Sāriputta, those who practise the holy life in my Dhamma reach the end of suff ering.25 This passage excuses wet dreams by emphasizing their non-reality. While it does not concern the Arahant directly, it carefully avoids condemning wet dreams, and clearly implies that there is no ethical content to a wet dream. This refl ects an aspect of the emerging of non-substantialism, which was one aspect of the Mahāsaṅghika perspective. This doctrine, usually associated with the

23. 若未離欲恚癡散亂心眠必失不淨。雖未能離。以繫念心眠者無有是過 (T no. 1421, p. 10, b27– 29). 24. 夢中雖有情識 (T no. 1458, p. 540, b28–29). 25. 夢者虛妄不實。若夢真實。於我法中修梵行者。無有解脫。以一切夢皆不真實。是故舍利弗。 諸修梵行者於我法中得盡苦際 (T no. 1425, p. 263, a26–29)

© Equinox Publishing Ltd 2008 SUJATO THE DATE AND CAUSE OF THE FIRST SCHISM 229

Mahāyāna, is sectarian, since it opposes the realistic ontology of the Sthavira schools. It should therefore be seen as a sign of lateness in this passage, consist- ent with the fact that the passage is not shared with other Vinayas. This passage, then, would have been composed after the First Schism, quite likely in direct response to the dispute over the fi ve points. Next, in apparent contradiction to this doctrine, the text lists (and defi nes) fi ve kinds of dream: true dreams (such as the 5 dreams of the before his awakening); false dreams (when one sees in a dream what is not true when awake); unrealized dreams (having woken, one does not remember); a dream inside a dream; dreams born of thinking (one plans and imagines during the day, then dreams about it at night).26 Then the text gives us 5 causes of erections: sensual desire; excrement; urine; wind disorder; contact with non-humans.27 A similar list is found in the Pali cases for the fi rst pārājika, in the context of affi rming that an Arahant can have an erection: There are, monks, these fi ve causes of erections: lust, excrement, urine, wind, or insect bite. These are the fi ve causes for an erection. It is impos- sible, monks, it cannot happen that that bhikkhu could have an erection out of lust. Monks, that bhikkhu is an Arahant.28 The last point is interesting: in the Pali it clearly refers to ‘bites of caterpillars and little creatures’, whereas the Mahāsaṅghika speaks of ‘non-humans’, a term widely used of spirit beings, and thus inclusive of the idea of ‘conveyance by Māra’.

CONCLUSION

So the Mahāsaṅghika does not contain any statement condemning wet dreams, or attributing them to mindlessness. While the Mūlasarvāstivāda is also silent on the topic, in that case it is a mere omission, whereas the Mahāsaṅghika is deliber- ately excusing wet dreams with the late doctrine about the unreality of dreams. Similarly, they appear to have rephrased the fi ve causes of erections to allow the possibility of Māra’s involvement.

26. 者實夢。二者不實夢。三者不明了夢。四者夢中夢。五者先想而後夢 (T no. 1425, p. 263, b8– 10). 27. 身生起有五事因緣。欲心起。大行起。小行起。風患起。若非人觸起 (T no. 1425, p. 263, b20– 21). 28. Pali Vinaya III.39: ‘pañcahi, bhikkhave, ākārehi aṅgajātaṁ kammaniyaṁ hoti – rāgena, vaccena, passāvena, vātena, uccāliṅgapāṇakadaṭṭhena. imehi kho, bhikkhave, pañcahākārehi aṅgajātaṁ kammaniyaṁ hoti. aṭṭhānametaṁ, bhikkhave, anavakāso yaṁ tassa bhikkhuno rāgena aṅgajātaṁ kammaniyaṁ assa. arahaṁ so, bhikkhave, bhikkhu’. The identical list in Mahīśāsaka Vinaya saṅghādisesa 1, except sensual desire is last (T no. 1421, p. 10, b26–27).

© Equinox Publishing Ltd 2008 230 BUDDHIST STUDIES REVIEW

On this basis, we are justifi ed in seeing a sectarian divergence in this Vinaya issue. All the Vinayas are concerned about wet dreams. The Sthavira schools, with the dubious exception of the Mūlasarvāstivāda, condemn them with vary- ing degrees of stridency, while the Mahāsaṅghika are concerned to excuse them. There seems little doubt that this diff erence is connected with the root cause of the separation between the schools on the basis of the ‘fi ve points’. The Vinaya of the Mahāsaṅghikas was found in their ancient centre of Pāṭaliputra, so this is not something that can be dismissed as a regional development among the Mahāsaṅghika sub-schools. I cannot yet trace the remaining four points in the Mahāsaṅghika texts. Perhaps further evidence will be found, although since they do not relate so closely to Vinaya issues it is perhaps unlikely. In any case, I think we can defi nitely accept that the fi ve points were a key issue in the schism, prompted most likely by actual events rather than mere theoretical discussion. It is of course possible that such issues were discussed for a long time in the Buddhist community before being seen as schismatic; and we must expect that other issues played their part in the schism as well.

ABBREVIATIONS

CBETA Chinese Buddhist Electronic Text Association Dv Dīpavaṃsa, ed. and trans. H. Oldenberg (Oxford: , [1879] 2000). T Taishō VRI Vipassana Research Institute

BIBLIOGRAPHY

The following is a list of works cited. For a full bibliography, see http://sectsandsectarianism.goog- lepages.com/bibliography. Pali texts are taken from the VRI CD, but references are to the Pali Text Society editions. The exception is the Dīpavaṁsa, which I retrieve from www.sub.uni-goettingen.de/ ebene_1/fi indolo/gretil/2_pali/3_chron/dipava2u.htm (13 May 2008). (Note that as of this date the Dīpavaṁsa does not appear to be accessible from the menu on the host site www.sub.uni-goettingen. de.) Chinese texts are quoted from the CBETA CD 2007.

Bandaranayake, Senake. ‘The Pre-modern City in Sri Lanka: The “First” and “Second” Urbanization’. www.arkeologi.uu.se (13 May 2008): hwww.arkeologi.uu.se/afr/projects/BOOK/Bandaranayake/ bandaranayake.pdf Bapat, P. V. and A. Hirakawa. 1970. Shan-Chien-P’i-P’o-Sha: A Chinese Version by Sanghabhadra of Samantapāsādikā. Poona: Bhandarkar Oriental Research Institute. Bechert, Heinz. 1982. ‘The Importance of Aśoka’s so-called Schism Edict’. In Indological and Buddhist Studies, ed. L. A. Hercus, 61–8. Canberra: Faculty of Asian Studies (ANU). Bechert, Heinz. 2005a. ‘Notes on the Formation of Buddhist Sects and the Origins of Mahāyāna’. In Buddhism: Critical in , vol. II, ed. Paul Williams, 23–33. London: Routledge. Originally published in German Scholars in , vol. I, 6–18, Chowkhamba Sanskrit Series (Varanasi, 1971). Bechert, Heinz. 2005b. ‘The Date of the Buddha Reconsidered’. In Buddhism: Critical Concepts in Religious

© Equinox Publishing Ltd 2008 SUJATO THE DATE AND CAUSE OF THE FIRST SCHISM 231

Studies, vol. II, ed. Paul Williams, 64–71. London: Routledge. Originally published in Inologica Taurinensia 10 (1982): 29–36. Cousins, L. S. 1996. ‘The Dating of the Historical Buddha: A Review Article’. Journal of the Royal Asiatic Society series 3, 6(1): 57–63. Cousins, L. S. 2001. ‘On the Vibhajjavādins’. Buddhist Studies Review 18(2): 131–82. Cousins, L. S. 2005a. ‘Pali Oral Literature’. In Buddhism: Critical Concepts in Religious Studies, vol. I, ed. Paul Williams, 96–104. London: Routledge. Originally published in Buddhist Studies Ancient and Modern, eds P. Denwood and A. Piatigorsky, 1–11 (London: Curzon Press/Totowa, NJ: Barnes & Noble, 1982). Cousins, L. S. 2005b. ‘The “Five Points” and the Origins of the Buddhist Schools’. In Buddhism: Critical Concepts in Religious Studies, vol. II, ed. Paul Williams, 52–83. London: Routledge. Originally pub- lished in The Buddhist Forum Volume II, ed. T. Skorupski, 27–60 (London: School of Oriental and African Studies, 1991). Frauwallner, E. 1956. The Earliest Vinaya and the Beginnings of Buddhist Literature. Roma: Is. M. E. O. Guruge, Ananda W. P. 2005. ‘Shan-Jian-Lu-Piposha as an Authentic Source on the Early and Aśoka’. In Dhamma-Vinaya: Essays in Honor of Venerable Professor Dhammavihari (Jotiya Dhirasekera), eds Tilakaratne, Toschiichi Endo, G. A. Somaratne, and Sanath Nanayakkara, 92–110. Sri Lanka Association for Buddhist Studies (SLABS). Jawawickrama, N. A. 1986. The Inception of Discipline and Vinaya Nidāna. London: Pali Text Society. Lamotte, Etienne. 1976. History of Indian Buddhism. Paris: Peeters Press. Nakamura, Hajime. 1996. Indian Buddhism. Delhi: Motilal Barnasidass. Nattier, Jan & Charles S. Prebish. 2005. ‘Mahāsaṅghika Origins’. In Buddhism: Critical Concepts in Religious Studies, vol. II, ed. Paul Williams, 199–228. London: Routledge. Originally published in History of Religions 16 (1976–7): 237–72. Oldenberg, Hermann. 1879. The , vol. I. London: Pali Text Society. Pachow, W. 2000. A Comparative Study of the Pratimoksa. Delhi: Motilal Barnasidass. Prebish, Charles S. 2005a. ‘Review of Scholarship on ’. In Buddhism: Critical Concepts in Religious Studies, vol. I, ed. Paul Williams, 224–43. London: Routledge. Originally published in 1974, in Journal of Asian Studies 33(2) (1974): 239–54. Prebish, Charles S. 2005b. ‘Śaikṣa- Revisited’. In Buddhism: Critical Concepts in Religious Studies, vol. II, ed. Paul Williams, 186–98. London: Routledge. Originally published in History of Religions 35(3) (1996): 258–70. Rockhill, W. Woodville. 1992. The Life of the Buddha. New Delhi: Asian Educational Services. Sasaki, Shizuka. 1989. ‘Buddhist Sects in the Aśoka Period. (1) The Meaning of the Schism Edict’. Buddhist Studies (Bukkyo Kenkyu) 18: 157–76. Strong, John S. 1994. The Legend and of Upagupta. Delhi: Motilal Barnasidass. Sujato, Bhikkhu. 2007. Sects & Sectarianism. Taipei: Corporate Body of the Buddha Educational Foundation. Full text and supplementary essays at Sects & Sectarianism, http://sectsandsectari- anism.googlepages.com. Willis, Michael. 2001. ‘Buddhist Saints in Ancient Vedisa’. Journal of the Royal Asiatic Society, Series 3, 11(2): 219–28. Wynne, Alexander. 2003. ‘How Old is the Suttapitaka?’ Oxford Centre for Buddhist Studies, http:// www.ocbs.org/research.php.

© Equinox Publishing Ltd 2008