Early Buddhist Oral Tradition (3)

Anālayo, Dīrgha-āgama Studies (Dharma Drum Institute of Liberal Arts Research Series 6). Dharma Drum Publishing Corporation, Taipei, 2017.

Originally published in 2016 under the title: "Selected Madhyama-āgama Discourse Passages and Their Pāli Parallels" in the Dharma Drum Journal of Buddhist Studies, 19: 1–60.

Introduction

In this chapter I continue to study the early Buddhist oral tradition, here in particular highlighting its potential vicissitudes in the way these can be identified through comparative study. My main aim is to document the contribution Chinese Āgama passages can offer for an alternative understanding, or even for a correction, of their Pāli counterparts. For this purpose I survey some passages from the Madhyama-āgama preserved in Chinese translation.[n.1] This chapter was originally presented at a seminar on the Madhyama-āgama held at the Dharma Drum Institute of Liberal Arts in October 2015. It incorporates cases discussed in more detail in Anālayo 2011b; cf. also Anālayo 2005a and 2005b; for a survey of instances that corroborate the main point I make in this chapter cf. Bodhi 2012: 72–74. As a result of taking up examples from this collection, the present last chapter in this book in a way takes the function of leading over to my collected studies of Madhyama-āgama discourses (2012f).

The evidence I survey is quite variegated; some differences have considerable consequences for our understanding of early Buddhist thought, whereas others merely testify to the type of errors that tend to occur during oral transmission, similar to cases mentioned in the preceding chapter. When considered collectively, these instances make it in my view impossible, methodologically speaking, to consider the Pāli version as invariably the earliest and most trustworthy specimen of a particular discourse.[n.2] Here I am forced to disagree with von Hinüber 2015b: 198, who in reply to criticism I raised in Anālayo 2008d (of a hypothesis presented by von Hinüber 2008a) argues that "concentration on the Theravāda tradition is neither a 'methodological problem' (p. 114) nor a 'methodological shortcoming' (p. 122), but a methodological necessity. Only the oldest levels of the Buddhist tradition we can reach might occasionally tell something about the very early history of Buddhism." Instead, a proper appreciation of Buddhist thought and history makes a comparative study of all parallel versions indispensable. I believe that the selected Madhyama-āgama passages I present below suffice to bear out the methodological need to go beyond basing one's research on Pāli material only.[n.3] The decision to focus on Madhyama-āgama discourses simply reflects the topic of the seminar to which I contributed this study. The same potential exists for parallels preserved in the other Āgamas, in Indic language fragments, in Tibetan translation, etc. In fact von Hinüber 2015b: 197 note 4 refers to my study of the Nandakovāda-sutta, Anālayo 2010a, which clearly shows that the Saṃyukta-āgama parallel has preserved an earlier presentation than its Pāli counterpart; cf. also p. 553 note 157 below. Needless to say, the same holds even more so the other way round, in that a study of the Chinese Āgamas must take into consideration their Pāli parallels. The fact that in this chapter I focus on the potential contribution of a Chinese Āgama to a study of the Pāli discourses is simply motivated by the circumstance that this potential is not as widely recognized as the help the Pāli texts offer for understanding and studying their Chinese parallels.

My presentation follows the sequence of the discourses in the Madhyama-āgama collection and thereby can also be read as a commentary alongside the translation of the collection that is at present under preparation.[n.4] The first volume has already been published, Bingenheimer et al. 2013; the remainder has been translated and is being prepared for publication. In the case of differences that I have already examined in detail elsewhere as part of a study of the respective discourse, I just summarize my findings and refer to those publications for further details.

MĀ 9: The Seven Stages of Purification

The seven stages of purification are a central topic in the Rathavinīta-sutta and its Madhyama-āgama parallel, where they form part of a discussion between the Buddha's chief disciple Sāriputta and another eminent disciple named Puṇṇa Mantāniputta. In the Rathavinīta-sutta, the question asked by Sāriputta to get the discussion started takes the following form:

Venerable friend, is the holy life lived under our Blessed One?[n.5] MN 24 at MN I 147,16: bhagavati no, āvuso, brahmacariyaṃ vussatī ti?

The form this question takes is somewhat unexpected. It makes little sense for a Buddhist monk to ask another Buddhist monk whether the holy life or brahmacariya, a standard reference to the monastic life, is being lived under the Buddha. The Madhyama-āgama version of the question, with which Sāriputta begins the discussion, reads as follows:

Venerable friend, are you practising the holy life under the recluse Gotama?[n.6] MĀ 9 at T I 430b26: 賢者從沙門瞿曇修梵行耶? This question has already been translated by Minh Chau 1964/1991: 144, without discussing this indeed very minor difference in formulation from MN 24.

Although the question in the Madhyama-āgama version as such reads more meaningfully, it is unusual for a disciple of the Buddha to be depicted as referring to his teacher as "the recluse Gotama". In such contexts in the early discourses the respectful expression bhagavant would be the appropriate form of reference. That the expression found in the Madhyama-āgama version is not simply the result of an error by the translator can be seen from another parallel preserved in Sanskrit fragments, which contains the word "recluse", śramaṇa, in a context paralleling one of Sāriputta's questions about the purpose of the holy life under the Buddha.[n.7] SHT VI 1329 B1, Bechert and Wille 1989: 84 (identified by J. Chung and M. Schmidt in Bechert and Wille 2000: 201). The version of the question in EĀ 39.10 at T II 734b18 instead employs the expression "Blessed One"; also here Sāriputta addresses Mantāniputta by his name. This does not seem a particularly meaningful presentation, because later the same discourse reports that, at the end of their discussion, Sāriputta asked for Mantāniputta's name, EĀ 39.10 at T II 735a19: 汝今為名何等? (a query also reported in the parallels). This shows that the wording of the passage at T II 734b18 is not reliable, as it conflicts with the remainder of the same discourse.

Close inspection of the narrative setting indicates that this is more likely to be the original reading. The parallel versions agree in reporting that Puṇṇā was surprised to find out, at the end of their discussion, that the person with whom he had been conversing was the Buddha's chief disciple Sāriputta. This part reads as follows in the Madhyama-āgama version:

Now I have been discussing with a disciple of the Blessed One without knowing it, I have been discussing with the second most respected one without knowing it,[n.8] In MN 24 at MN I 150,27 Puṇṇa calls Sāriputta "the disciple who is comparable to the teacher", satthukappa sāvaka. Although this does involve high praise, it does not put him on the same level as the Buddha; Sāriputta remains merely a disciple. This conforms to the way he is presented in other early discourses, pace Karashima 2015a: 186f, who proposes that "in the Suttanipāta … disciples, such as Sāriputta/Śāriputra, are designated as buddhas. Śākyamuni proclaimed that anybody, who follows his teachings and his practices together with its mode of living, can become a buddha", and then states that "much later … the hierarchy of lay Buddhists, disciples, pratyekabuddhas and Śākyamuni Buddha was formed." Karashima does not provide a more specific reference and I have not been able to identify a passage in the Suttanipāta that presents Sāriputta as a Buddha or even as completely on a par with the Buddha. In fact, whereas the hierarchy mentioned by Karashima is consistently reflected in the early discourses extant from various transmission lineages, the idea of becoming a Buddha is clearly a later development; cf. in more detail Anālayo 2010b. I have been discussing with the general of the Dharma without knowing it, I have been discussing with the disciple who keeps the wheel of Dharma rolling without knowing it. If I had known this was the venerable Sāriputta, I would not have been able to say a single sentence in reply, let alone discuss in such depth.[n.9] MĀ 9 at T I 431b27 to 431c1. The parallel passages are MN 24 at MN I 150,27 and EĀ 39.10 at T II 735b1; cf. also SHT II 163b R6, Waldschmidt et al. 1968: 16.

Another point relevant to the narrative background against which the above episode is probably best read emerges from an episode found in the different Vinayas, according to which the typical Buddhist way of sewing up robes would have been decided only at an already somewhat evolved stage in the development of Buddhist monasticism. [n.10] According to the episode in question, the Buddha told Ānanda that from now on the robes should be sewn together conforming to the pattern of paddy fields; cf. the Dharmaguptaka Vinaya, T 1428 at T XXII 855a20, the Mahāsāṅghika Vinaya, T 1425 at T XXII 454c27 (where the instruction is given not to Ānanda, but to unnamed monks), the Mahīśāsaka Vinaya, T 1421 at T XXII 137a22, the Mūlasarvāstivāda Vinaya, Dutt 1984b: 50,11, the Sarvāstivāda Vinaya, T 1435 at T XXIII 194c25, and the Theravāda Vinaya, Vin I 287,7. As Ānanda is on record for becoming the Buddha's personal attendant only about twenty years after the Buddha's awakening (cf. Th 1041ff), it would follow that the regulation concerning the way to sew Buddhist robes should not be reckoned as belonging to an early stage in the history of the Buddhist monastic order. Considering the present discourse from the viewpoint of this depiction, it would follow that during earlier times members of the monastic community would presumably not have been easily recognizable by their outer appearance as Buddhist monks, in that they would supposedly have dressed in the way used in general among recluses and wanderers roaming the Ganges valley. The appropriateness of such a reading can be seen from yet another episode found in the Mūlasarvāstivāda and Sarvāstivāda Vinayas, according to which King Bimbisāra worshipped a heterodox practitioner, mistaking him to be a Buddhist monk.[n.11] Cf. Dutt 1984b: 49,3 and T 1435 at T XXIII 194c12.

If the same perspective should hold for the narrative setting of the Rathavinīta-sutta, it would follow that Sāriputta can be visualized as someone who need not have been immediately recognizable as a Buddhist monk. On this assumption, in order not to reveal his identity prematurely by betraying that he is a disciple of the Buddha, it would indeed make sense for Sāriputta to be shown as using the expression "the recluse Gotama".[n.12] Cf. in more detail Anālayo 2012g: 70–77. If this should indeed be the more original formulation, it could easily have happened, in the course of transmission, that the apparent inappropriateness of Sāriputta using such an expression resulted in it being replaced by the term bhagavant.

This in itself minor variation carries further significance, since the narrative context to which it points also makes it fairly probable that the topic of discussion chosen by Sāriputta would have been one of general interest among ancient Indian wanderers and recluses. Had he been shown instead to broach a topic characteristic of Buddhist thought, then in the narrative setting of the discourse this would have also run the risk of prematurely revealing his identity as a disciple of the Buddha.

Now the main doctrinal topic of the Rathavinīta-sutta, the seven stages of purification, occurs in fact only in this discourse and its parallels, and as part of a set of nine purifications in the Dasuttara-sutta and its parallels.[n.13] DN 34 at DN III 288,16 and its parallels in Sanskrit fragments, Schlingloff 1962: 18, DĀ 10 at T I 56a23, and T 13 at T I 238c25. Only with the Visuddhimagga do the seven stages of purification come to the foreground as the chief framework for progress on the path.

In the Rathavinīta-sutta and in the Dasuttara-sutta the seven stages lead up to liberation, but do not include liberation itself. Liberation instead constitutes a separate stage of purification mentioned only in the Dasuttara-sutta, placed after the set of seven stages.

Yet in his Visuddhimagga Buddhaghosa assigns the four levels of awakening to the seventh purification.[n.14] Vism 672,4: sotāpattimaggo sakadāgāmimaggo anāgāmimaggo arahattamaggo ti imesu pana catusu maggesu ñāṇaṃ ñāṇadassanavisuddhi nāma; cf. in more detail Anālayo 2005c and 2009g: 8–11. In reply to my suggestions in these two papers, Endo 2015: 55 argues that the circumstance that the discussion in MN 24 was not reported to the Buddha implies that "the topic of the seven stages of purification was a well-known subject among the Buddhists in ancient times." Yet the reporting of a discussion to the Buddha usually happens when a disciple is uncertain whether he or she has represented the Buddha's position correctly in a debate-type situation. This is not the case in the present instance. Endo 2015: 68 also comments that "the scheme of 'seven stages of purification', it is contended, was commonly practiced and aspired to among the various contemplative and philosophical traditions in ancient India. But this contention does not supply the reason why the Buddhists should adopt this scheme." In Anālayo 2009g I suggested that the reason for adoption of the scheme could be related to Buddhaghosa's compilation of the Visuddhimagga in competition with the Vimuttimagga. Endo 2015: 70 also argues that it can "be inferred that Buddhaghosa did not adopt the scheme of sevenfold purification (satta-visuddhi) as the structural scaffolding of the Visuddhimagga, but that by reducing and incorporating the last two items of 'paññā' and 'vimutti' of the 'nine factors of exertion for purity' (nava pārisuddhi-padhāniyaṅga) into the category of 'ñāṇadassana' … the list of purifications finally became 'seven' in the Visuddhimagga." This suggestion does not seem to work, as the terminology used by Buddhaghosa corresponds to the visuddhi terminology in MN 24, not to the expression pārisuddhi-padhāniyaṅga used in DN 34. Moreover, Endo's suggestion does not solve the problem, since Buddhaghosa does include under the seventh purification what does not belong there, whether this seventh purification is taken from the scheme in MN 24 or from the scheme in DN 34. This does not square with the implications of the seventh purification either in the Rathavinīta-sutta or in the Dasuttara-sutta, where the seventh stage leads up to but differs from arrival at the final goal. This in turn gives the impression that even Buddhaghosa himself was not sufficiently familiar with the significance of the seven stages of purification.[n.15] The present case would conform to a pattern noted by von Hinüber 2015a: 354 in regard to the commentators Buddhaghosa and Dhammapāla, who "did not always work with the same care and concentration. Consequently, oversights and weakness in systematization help to trace the material they had at hand, and to detect their method of approach." In this way, an in itself rather minor difference in the Madhyama-āgama version provides a significant perspective on the main teaching in the Rathavinīta-sutta, which acquired a position of central importance in later Theravāda exegesis.

MĀ 15: Karma and Its Fruit

A topic closely related to the theme of purification is karma. The Karajakāya-sutta among the Tens of the Aṅguttara-nikāya begins broaching this topic with a statement on karma that reflects Jain thought,[n.16] AN 10.208 at AN V 299,14: na tvevāhaṃ, bhikkhave, sañcetanikānaṃ kammānaṃ katānaṃ upacitānaṃ appaṭisaṃviditvā (Be: appaṭisaṃveditvā) dukkhass'antakiriyaṃ vadāmi; "yet, monks, I do not say that there is a making of an end of dukkha without having experienced [the fruits of] intentional deeds that have been undertaken and accumulated." According to, e.g., the Jain Dasaveyāliya-sutta, Lalwani 1973a: 212,1, liberation will be attained only when one's former bad deeds have been experienced, wherefore it is impossible to reach liberation without karmic retribution being either experienced or else expiated through asceticism. a statement not found in its Madhyama-āgama parallel and presumably the result of an error in transmission.[n.17] MĀ 15 at T I 437b27 instead points out that unintentionally performed deeds do not entail karmic retribution; cf. also D 4094 ju 236b3 or Q 5595 tu 270a5. The Karajakāya-sutta also appears to have lost an exposition of the ten unwholesome courses of action, found in the Madhyama-āgama parallel, which would have originally led to its inclusion among the Tens of the Aṅguttara-nikāya. Clearly the Karajakāya-sutta has suffered from errors during its transmission that can be corrected by recourse to its Madhyama-āgama parallel (as well as a parallel in the Abhidharmakośopāyikā-ṭīkā).[n.18] Cf. in more detail Anālayo 2009e and Martini 2012.

MĀ 16: The Kālāmas and Doubt

According to another discourse in the Aṅguttara-nikāya, the Buddha told the Kālāmas that it is appropriate for them to have doubts about various views. In the Madhyama-āgama parallel, however, he rather told them not to have doubts.[n.19] AN 3.65 at AN I 189,6 and MĀ 16 at T I 438c12; the same type of difference recurs between SN 42.13 at SN IV 350,15 and MĀ 20 at T I 447a22.

Although in this case it is not easy to decide which of the two versions has preserved the earlier reading, the difference is significant in so far as the advice to the Kālāmas given in the Pāli version is not supported by its parallel and thus its claim to represent early Buddhist thought does not stand on as firm a ground as it would have if the same position had also been reflected in its parallel.[n.20] According to Bodhi 2012: 73f, "in contemporary Buddhist circles it has become almost de rigueur to regard the Kālāma Sutta as the essential Buddhist text … held up as proof that the Buddha anticipated Western empiricism, free inquiry, and the scientific method, that he endorsed the personal determination of truth … it has become one of the most commonly quoted Buddhist texts, offered as the key to convince those with modernist leanings that the Buddha was their forerunner." In other words, discussions of the advice to the Kālāmas need to keep in mind the possibility that the Pāli version could be the result of a change that occurred during transmission and that its testimony regarding the early Buddhist attitude towards free enquiry and the resolution of doubt is comparatively less certain than other discourses on this topic that are supported by their parallels.

MĀ 19: A Critique of the Jains

Returning to the topic of the Jains, a minor difference between a Pāli discourse and its Madhyama-āgama parallel that is also related to Jain thought can be found in the case of the Devadaha-sutta and its parallel, which report the Buddha examining the fruitlessness of Jain asceticism. In the Pāli version the Buddha announces ten grounds for censuring the Jains.[n.21] MN 101 at MN II 222,2: dasa sahadhammikā vādānuvādā; on the expression sahadhammiko vādānuvādo cf. also Alsdorf 1959. Following this announcement, however, he specifies only five. These are that pain may be caused by:

one's former deeds (1),
a creator god (2),
the company one keeps (3),
one's type of life form (4),
one's present practice (5).

In each of these five cases the Jains are to be censured for experiencing such pain. Then the Buddha points out that even in the absence of each of these five, the Jains are to be censured. Yet the absence of these five does not provide any real ground for censure, let alone five more grounds. Thus the exposition does not contain what the initial announcement promises.

In the Madhyama-āgama parallel the exposition of grounds for censure works through a similar set of five topics, although in a different sequence. Here, too, the Jains are to be censured for their experience of pain, if this experience is attributable to:

former deeds (1),
the company one keeps (2),
one's destiny (3),
one's view (4),
a supreme god (5).

In the Madhyama-āgama version this exposition does not result in an internal inconsistency, however, as here the Buddha announces at the outset that he will deliver five grounds for criticism.[n.22] MĀ 19 at T I 443c18: 得五詰責.

This gives the impression that an error in textual transmission has affected the Devadaha-sutta, resulting in the incorrect count of ten grounds for criticism. Recourse to the Madhyama-āgama version helps to clarify this situation.

MĀ 29: Right View and the End of dukkha

Moving from the fruitlessness of Jain asceticism to the fruits of the Buddhist path, the Sammādiṭṭhi-sutta concludes expositions of arrival at right view, a reference to stream-entry, by mentioning the making of an end of dukkha.[n.23] MN 9 at MN I 47,22: so sabbaso rāgānusayaṃ pahāya paṭighānusayaṃ paṭivinodetvā asmī ti diṭṭhimānānusayaṃ samūhanitvā avijjaṃ pahāya vijjaṃ uppādetvā diṭṭhe va dhamme dukkhass'antakaro (Ce: antaṅkaro) hoti. ettāvatā pi kho, āvuso, ariyasāvako sammādiṭṭhi hoti; "by completely abandoning the underlying tendency to lust, having abolished the underlying tendency to irritation, having exterminated the underlying tendency to the conceited view 'I am', abandoning ignorance and having given rise to knowledge, he here and now is one who makes an end of dukkha; to that extent, friend, a noble disciple is of right view." This formulation continues for subsequent expositions of right view in the discourse. This statement seems out of context, since to make an end of dukkha represents full awakening.

The Sammādiṭṭhi-sutta follows each reference to making an end of dukkha by declaring that "to that extent", ettāvatā, a noble disciple is endowed with right view and has gained perfect confidence in the teaching. Such right view and perfect confidence are already gained with stream-entry, at which stage the making of an end of dukkha is yet to be accomplished. Hence the expression "to that extent" does not tally with the content of the passage to which it refers.[n.24] Ps I 197,24 records a discussion between the rehearsing monks on the significance of the present passage. Evidently they also had difficulties reconciling the statements made in this passage with the main theme of the discourse.

The Madhyama-āgama version and other parallels, however, describe right view without any comparable statement regarding the overcoming of dukkha,[n.25] MĀ 29 at T I 461c8 and its parallels S 474 folio 16R7, Tripāṭhī 1962: 51, and SĀ 344 at T II 94b24; cf. also Anālayo 2011e: 20. making it safe to conclude that the passage in the Sammādiṭṭhi-sutta is the result of an error in transmission.

MĀ 32: The Bodhisattva Already Supreme at Birth

Shifting from the right view gained with stream-entry to qualities of the Buddha, the Acchariyabbhutadhamma-sutta presents a listing of the Buddha's marvellous qualities, one of which is that the Buddha-to-be, right after being born, made the following proclamation:

I am highest in the world, I am best in the world, I am supreme in the world, this is my last birth, there will be no more renewed becoming.[n.26] MN 123 at MN III 123,21: aggo'ham asmi lokassa, seṭṭho'ham asmi lokassa, jeṭṭho'ham asmi lokassa (Be and Se have the last two in the opposite sequence), ayam antimā jāti, natthi dāni punabbhavo ti.

For the bodhisattva Gotama to make such a proclamation right after his birth implies that he was inevitably destined to awakening, that his struggle to reach awakening by various methods was anyway destined to be successful. This is the case to such an extent that already as a baby he can proclaim to have reached the transcendence of future birth that is reached when full awakening is attained.[n.27] Silk 2003: 864 points out that, according to this proclamation of supremacy, the bodhisattva "is virtually fully awakened ('enlightened') from the moment of his birth", highlighting the resulting contrast where "the infant, upon his birth, knows everything; the young man he becomes knows nothing."

This passage reflects a mature stage in the apotheosis of the Buddha, whereby the qualities that according to other discourses were the outcome of his struggle for awakening came to be considered as already in his possession at the time when he was born.

No such proclamation is recorded in the Madhyama-āgama parallel, confirming the impression that the above passage in the Pāli version reflects distinctly late developments.[n.28] As already noted by Nakamura 1980/1999: 18, "the verse claimed to have been proclaimed by the Buddha at his birth was composed very late." Nevertheless, the Madhyama-āgama version also offers a distinctly late indication, which comes at the outset of the listing of marvellous qualities of the Buddha. This reads as follows:

The Blessed One, at the time of the Buddha Kassapa, made his initial vow to realize Buddhahood.[n.29] MĀ 32 at T I 469c24: 世尊迦葉佛時始願佛道. Minh Chau 1964/1991: 160f has already translated this passage and also summarized the statement quoted above from MN 123. In his comparison he mentions, in relation to the first, that this vow is usually associated with Dīpaṅkara Buddha. In relation to the second, he notes that the Pāli version additionally mentions that the bodhisattva stood on his own feet and faced north.

No such statement is found in the Acchariyabbhutadhamma-sutta. Together with the apotheosis of the Buddha-to-be evident in the Pāli version, this evidently late statement in the Madhyama-āgama about a time in the past when the future Buddha Gotama presumably took the vow to attain Buddhahood can be seen to reflect the beginnings of what was to become the bodhisattva ideal.[n.30] Cf. in more detail Anālayo 2010b.

Even though my survey in this chapter focuses on cases where the Madhyama-āgama version provides significant indications regarding its Pāli counterpart, the present case reflects the fact, already mentioned in the introduction, that the same potential certainly applies the other way around, in that the Pāli discourses often offer significant indications regarding their Āgama parallels. In fact the Madhyama-āgama contains some distinctly late discourses that have no Pāli parallel.

One example is a discourse that works through a series of topics that form the scaffolding of the Dharmaskandha.[n.31] MĀ 86 at T I 562a19; cf. Watanabe 1983/1996: 54 and Anālayo 2014c: 41–44. Incidentally this discourse and MĀ 127 are strong indicators for the school affiliation of the Madhyama-āgama, on which cf. also Anālayo 2008g: 7, 2012f: 516–521 (in reply to Chung and Fukita 2011) and 2017c (in reply to Chung 2015 and 2017). Yet another example is the last discourse in the collection. This could only have come into being in the written medium, since giving the full texts of all of its abbreviations would make it longer than the remainder of the Madhyama-āgama collection.[n.32] MĀ 222 at T I 805c10; cf. Anālayo 2014c: 44–47. Still another example is a listing of noble beings that refers to an arahant who can fall away from his attainment, a notion that reflects a schoolspecific development in the conception of what becoming an arahant implies.[n.33] MĀ 127 at T I 616a18; for the counterpart expression in a quotation of this discourse in the Abhidharmakośavyākhyā cf. Wogihara 1936: 566,34.

MĀ 34: Acela Kassapa's Repeated Going Forth

The conception of an arahant is also the topic of my next example from the Bakkula-sutta, which reports that the marvellous qualities of the monk Bakkula inspired the ascetic Acela Kassapa to go forth.[n.34] MN 124 at MN III 127,14. Having gone forth he in due course became an arahant. The Madhyama-āgama parallel does not give the name of Bakkula's visitor and, even though he clearly was delighted on hearing about the marvellous qualities of Bakkula, he did not go forth, nor is there any indication that he reached any level of awakening.[n.35] MĀ 34 at T I 475c13 just reports that he rejoiced together with the monks who were present. Minh Chau 1964/1991: 75 already noted that only the Pāli version reports his going forth and attaining of arahantship.

Such variations regarding the effect of a discourse on its listeners are a recurrent feature that emerges from comparative studies of parallel versions. What makes this case of particular interest is the circumstance that Acela Kassapa's going forth recurs in other Pāli discourses. A discourse in the Dīgha-nikāya reports that Acela Kassapa went forth after hearing a discourse by the Buddha on asceticism.[n.36] DN 8 at DN I 176,29. In this case the Dīrgha-āgama parallel agrees that Acela Kassapa went forth. [n.37] DĀ 25 at T I 104c11, which at T I 102c26 introduces him as 倮形梵志姓迦葉; listed in Akanuma 1930/1994: 4 as corresponding to the name Acela Kassapa. According to both versions, after going forth he became an arahant.

A discourse in the Saṃyutta-nikāya also reports that Acela Kassapa went forth and eventually became an arahant, this time after hearing a different discourse by the Buddha on the topic of what causes pleasure and pain.[n.38] SN 12.17 at SN II 21,26. According to three parallels to this discourse, however, Acela Kassapa instead attained streamentry during the discourse, did not go forth, and was soon after killed by a cow.[n.39] SĀ 302 at T II 86b3, T 499 at T XIV 768c20, and fragment S 474 folio 14R2, Tripāṭhī 1962: 47; cf. also SHT V 1133a A3, Sander and Waldschmidt 1985: 128.

In another discourse in the same Saṃyutta-nikāya Acela Kassapa once more goes forth and eventually becomes an arahant, this time after being inspired by a meeting with the householder Citta.[n.40] SN 41.9 at SN IV 302,9; von Hinüber 1997: 68 already noted this instance and MN 124 as separate occasions where Acela Kassapa is on record for having gone forth. A parallel in the Saṃyukta-āgama reports the same happy outcome, with the difference that it does not provide the name of the one who was so inspired by the householder Citta as to go forth as a Buddhist monk.[n.41] SĀ 573 at T II 152b24.

In sum, according to the Pāli discourses Acela Kassapa went forth and became an arahant on these four occasions:

DN 8 after meeting the Buddha (discourse on asceticism),
MN 124 after meeting Bakkula,
SN 12.7 after meeting the Buddha (discourse on pleasure etc.),
SN 41.9 after meeting Citta.

Although the same person can go forth on several occasions, it would not be possible for the same person to become an arahant each time.[n.42] In fact Malalasekera 1937/1995: 26 finds it difficult to reconcile the different reports of Acela Kassapa's going forth. Nor does it seem particularly probable that four different persons by the same name of Acela Kassapa went forth and became arahants. This makes it possible that Acela Kassapa's going forth in the Bakkula-sutta is a case of integrating material that stems from an originally different context. Judging from the situation in the Chinese Āgama parallels, the Dīgha-nikāya discourse in which he gets to hear a discourse from the Buddha about asceticism could well have been the place of origin of the depiction of his going forth and eventual attainment of full awakening.

MĀ 59: The Marks of a Buddha

Another Dīgha-nikāya discourse, the Lakkhaṇa-sutta, lists the thirty-two marks that adorn the body of a mahāpurisa and then proclaims that one endowed with these marks will become either a wheel-turning king or a Buddha. The Madhyama-āgama parallel concludes after a similar exposition.[n.43] MĀ 59 at T I 494b6, corresponding to DN 30 at DN III 145,17. The Lakkhaṇa-sutta, however, continues with a detailed exposition of what karmic deeds done in the past will lead to a particular mark and in what way these karmic deeds relate to qualities of a wheel-turning king or a Buddha. This exposition appears to be a commentarial type of presentation that has become part of the Pāli discourse.[n.44] Cf. in more detail Anālayo 2017a: 107ff.

MĀ 70: The Future Buddha Maitreya

Proceeding from the past of the Buddha Gotama to the future, the advent of the next Buddha Maitreya is not reported at all in the Madhyama-āgama parallel to the Cakkavatti-(sīhanāda-)sutta.[n.45] DN 26 at DN III 76,1 (cf. also DĀ 6 at T I 41c29); for the description of the future wheel-turning king under whose reign in DN 26 the arising of the next Buddha takes place cf. MĀ 70 at T I 524b29. The absence of any such reference in MĀ 70 has already been noted by Karashima et al. 2000: 310f note 121 (cf. Karashima 2013: 178; owing to my ignorance of Japanese I am not able to consult Karashima et al. 2000). The addition of sīhanāda to the title is not found in the Asian editions. A closer study of the narrative context makes it safe to conclude that the motif of the future Buddha is a later addition.[n.46] Cf. in more detail Anālayo 2010b: 107–113 and above p. 349ff.

MĀ 79: The Radiance of devas

Moving from Buddhas to devas, the Anuruddha-sutta and its Madhyama-āgama parallel illustrate the radiance of a congregation of devas with the example of several lamps that have been brought into a room, as a result of which their radiance becomes unified and it is no longer possible to distinguish the individual radiance of each lamp.

When the devas disperse again, however, their individual radiance can be recognized. To illustrate this, the Anuruddha-sutta describes removing the lamps from the house in which they had been placed.[n.47] MN 127 at MN III 148,5: seyyathāpi, gahapati, puriso tāni sambahulāni telappadīpāni tamhā gharā nīhareyya. This illustration seems to be slightly incomplete. Although removing the lamps would indeed stop the merging of their radiance, one would expect some additional specification in the simile to illustrate the radiance of individual devas.

The Madhyama-āgama parallel furnishes such additional information. It describes that, after being removed from the room in which they had been together, the lamps are put into separate rooms.[n.48] MĀ 79 at T I 550b15: 猶如有人從一室中出眾多燈, 分著諸室. This helps to appreciate the point the simile is meant to illustrate.

MĀ 85: The Conceit of Being from a High Family

Another difference of similarly minor significance can be seen in the Sappurisa-sutta, which in agreement with its parallels examines various occasions for the arising of conceit. The first such occasion in the parallel versions concerns someone who comes from a high-ranking family.[n.49] MĀ 85 at T I 561a26: 豪貴族, T 48 at T I 837c29: 大姓, and EĀ 17.9 at T II 585a23: 豪族. This seems indeed a potential source of conceit, particularly in the hierarchical society in ancient India.

The Sappurisa-sutta covers the same topic in four different ways, all of which in one way or another refer to a superior type of family. Each of these receives a separate and full treatment as a distinct occasion for conceit. The four are:

the uccākula, the "high family",
the mahākula, the "great family",
the mahābhogakula, "the family of great wealth",
the uḷārabhogakula, "the family of outstanding wealth".[n.50] MN 113 at MN III 37,15.

Although there could indeed be a difference between a family of high social standing and a family that has much wealth, the four types of family described in the Sappurisa-sutta do not seem to result in four distinct occasions for the arising of conceit.

The Madhyama-āgama parallel as well as other parallels have only a single reference to conceit based on one's family, after which they continue with quite different qualities that could arouse conceit. This suggests that perhaps during the course of oral transmission this part of the Sappurisa-sutta underwent expansion through attracting synonyms. Eventually these synonyms would have become independent qualities, resulting in its present listing of four types of family.

MĀ 97: The Three Types of Craving

Shifting from conceit to craving, the Mahānidāna-sutta refers to three types of craving: craving for sensual pleasures, craving for becoming, and craving for non-becoming, kāmataṇhā, bhavataṇhā, and vibhavataṇhā. In its ensuing exposition it then refers back to "those two states", ime dve dhammā.[n.51] DN 15 at DN II 61,33: ime dve dhammā dvayena vedanāya ekasamosaraṇā bhavanti. Rhys Davids 1910: 58 translated this as "these two aspects [of craving] from being dual become united through the sensation" and Walshe 1987: 225 as "these two things become united in one by feeling." This reference is puzzling. The commentary understands this to refer to the difference between primordial craving and craving arisen during action.[n.52] Sv II 500,18: ime dve dhammā ti vaṭṭamūlataṇhā ca samudācārataṇhā cā ti.

A similar reference to two dharmas occurs also in the Madhyama-āgama parallel, with the decisive difference that here this reference is preceded by indeed mentioning only two types of craving, craving for sensual pleasures and for becoming.[n.53] MĀ 97 at T I 579b22: 欲愛及有愛, 此二法因覺; this case has already been noted by Choong 2000: 166 note 78. Two further parallels preserved in Chinese agree with the Madhyama-āgama, in that they speak of two types of craving, namely for sensual pleasure and for becoming, and then refer back to these as "two states". [n.54] T 14 at T I 243a19: 欲愛亦有愛, 是二皆痛相會 and T 52 at T I 845a9: 欲愛, 有愛, 由此二法生諸過失. Another parallel in DĀ 13 at T I 61a22 just refers to craving at the present juncture, although at an earlier point at T I 60c13 it does mention three types of craving: 欲愛, 有愛, 無有愛者. Sanskrit fragment parallels have also preserved references to the two types of craving or else to the two dharmas.[n.55] SHT III 822R7f, Waldschmidt et al. 1971: 40: dve ānanda tṛṣṇe, kāmatṛ[ṣṇ]ā bhavatṛṣṇā ca (identified as a parallel to DN 15 by Tang Huyen in Bechert and Wille 1995: 268) and Or 15009/140Ar3, Kudo 2009: 191: dvau dharmau dvayena vedanā.

This reflects a recurrent difference between Pāli (and Dīrgha-āgama) discourses on the one hand and Madhyama-āgama (and Saṃyukta-āgama) discourses on the other hand.[n.56] In regard to a reference to the three types of craving in EĀ 49.5 at T II 797c8: 欲愛, 有愛, 無有愛, it needs to be kept in mind that this discourse appears to be the result of a reworking of parts of the Ekottarika-āgama collection in China; cf. Anālayo 2014/2015: 65. This makes it uncertain whether a reference to three types of craving was already present in the Indic original of the collection. The fact that in the present context the Pāli version has preserved a reference to two states would lend support to the hypothesis that this passage originally only mentioned two types of craving. If that should indeed be the case, then the idea of craving for non-becoming would be a later addition at least in the case of this discourse.[n.57] The suggestion by Choong 2016: 37 note 19 that variations between listings of either two or three types of craving imply that both are late, in that "this implies that in early Buddhism craving may have meant simply craving for the six sense objects" is unconvincing. Variations like the one observed in the present instance point to the possibility of a gradual evolution from two to three cravings; they do not imply that both are late.

MĀ 98: The Four Noble Truths and Satipaṭṭhāna

According to the teaching of the four noble truths, craving is the chief culprit for the arising of dukkha. The exposition of contemplation of dharmas in the Mahāsatipaṭṭhāna-sutta of the Dīgha-nikāya has a long exposition on these four noble truths, a topic not found at all in its Madhyama-āgama parallel.[n.58] DN 22 at DN II 304,22 to 313,27 and MĀ 98 at T I 584a14 (here and below for EĀ 12.1, the reference is to the beginning of the exposition of contemplation of dharmas). The same holds for a parallel in the Ekottarika-āgama; in fact even the Satipaṭṭhāna-sutta of the Majjhima-nikāya, otherwise similar, only has a short reference to the four noble truths.[n.59] EĀ 12.1 at T II 569a18 and MN 10 at MN I 62,21 to 62,24. This makes it safe to conclude that the long exposition in the Dīgha-nikāya discourse reflects a later development.[n.60] Cf. in more detail Anālayo 2014c: 91–100.

MĀ 101: Thought Control and Awakening

Proceeding from satipaṭṭhāna to thoughts, early Buddhist meditative theory recognizes a variety of unwholesome thoughts that need to be overcome in order to progress to awakening. For the purpose of dealing with such types of thought, the Vitakkasaṇṭhāna-sutta and its Madhyama-āgama parallel describe five methods that are to be employed in order to prevent unwholesome thoughts from remaining in the mind. This exposition being completed, the Vitakkasaṇṭhāna-sutta concludes its presentation with the following statement:

Whatever thought he will wish, that thought he will think; whatever thought he will not wish, that thought he will not think.

He has cut off craving, done away with the fetters, and made an end of dukkha through rightly comprehending conceit.[n.61] MN 20 at MN I 122,3: yaṃ vitakkaṃ ākaṅkhissati taṃ vitakkaṃ vitakkessati, yaṃ vitakkaṃ nākaṅkhissati na taṃ vitakkaṃ vitakkessati, acchecchi taṇhaṃ, vāvattayi saṃyojanaṃ (Be and Se: vivattayi saññojanaṃ), sammā mānābhisamayā antamakāsi dukkhassā ti.

This reads as if mere control of one's thoughts results in full awakening. Such a conclusion is not easily reconciled with what other early Buddhist discourses indicate regarding requirements for progress to full awakening. The corresponding passage in the Madhyama-āgama parallel proceeds as follows:

He thinks what he wishes to think and does not think what he does not [wish] to think. If a monk thinks what he wishes to think and does not think what he does not wish to think, then such a monk is reckoned as thinking according to his intentions and as having mastery over the courses of thought.[n.62] MĀ 101 at T I 589a6: 欲念則念, 不念則不念. 若比丘欲念則念, 不欲念則不念者, 是謂比丘隨意諸念, 自在諸念跡. The fact that MĀ 101 does not have a reference to cutting off craving has already been pointed out by Minh Chau 1964/1991: 244.

With this the Madhyama-āgama discourse concludes. The absence of any reference to making an end of dukkha is certainly more pertinent to the context. In fact closer inspection shows that in the Vitakkasaṇṭhāna-sutta the making of an end of dukkha is formulated in the past tense, unlike the mastery over one's thoughts, which is in the future tense. To stand in a meaningful relationship to each other, the use of the two tenses should be the reverse. This confirms the impression that also suggests itself from the Madhyama-āgama parallel, in that the reference to cutting off craving and making an end of dukkha are spurious elements in the Vitakkasaṇṭhāna-sutta. That is, it seems that in this case a passage not belonging to the discourse has come to be part of it at some point during its transmission.

Another difference of minor relevance concerns the title of the discourse. The stilling of thought, vitakkasaṇṭhāna, is only one of the five methods that both discourses present for the purpose of cultivating the higher mind.[n.63] The title of the discourse is given in MN 20 at MN I 122,8. The title of the Madhyama-āgama discourse is rather "On the Higher Mind". This expresses the content of the whole exposition, whereas the Pāli title reflects only one of its aspects.[n.64] MĀ 101 at T I 588a3: 增上心經.

MĀ 102: A Simile out of Context

Continuing with the topic of thoughts, the Dvedhāvitakka-sutta and its Madhyama-āgama parallel report how the Buddha, in the period before his awakening, handled the arising of unwholesome thoughts. Whenever an unwholesome thought arose, the future Buddha would quickly dispel it, as he was aware of the danger inherent in such thoughts.

The Madhyama-āgama discourse compares this to a cowherd who will stop the cows from straying into the ripe crop, as he knows that he will incur trouble if he does not prevent them from eating the crop.[n.65] MĀ 102 at T I 589a27. After having explained this approach to thought and illustrated it with the cowherd simile, in the Madhyama-āgama discourse the Buddha turns to the general nature of the mind, explaining that whatever one frequently thinks about will eventually lead to a corresponding inclination of the mind.

The Dvedhāvitakka-sutta presents the same topics in a different sequence. It first takes up the dispelling of unwholesome thought, then describes how the mind follows the course set by whatever one frequently thinks about, and only after that brings in the simile of the cowherd.[n.66] MN 19 at MN I 115,29. Its sequence is thus as follows:

dispelling of unwholesome thought,
frequent thinking leads to mental inclination,
cowherd simile.

Here the Madhyama-āgama discourse presents a more straightforward sequence, since the purpose of the cowherd simile is to illustrate the need to take firm action and have fear of unwanted consequences, not to exemplify that frequent thoughts lead to a mental inclination. Therefore the simile of the cowherd finds its best placing right after the exposition of unwholesome thoughts, as an illustration of this exposition, the placing it has in the Madhyama-āgama discourse. In contrast, the placing of the simile in the Pāli version seems to be the result of a shift of the passage from its proper location. Such shifting of a passage within a text is in fact a recurrent feature of orally transmitted literature.[n.67] For several examples cf. Anālayo 2011b: 874–876.

MĀ 104: Conversion and Its Consequences

Another topic related to the period before the Buddha's awakening is asceticism. The Udumbarika(-sīhanāda)-sutta reports a detailed exposition to a group of wanderers on this topic of asceticism,[n.68] The addition of sīhanāda to the title is not found in the Asian editions. evidently suited to their interest. At the end of his talk the Buddha invited the wanderers to join his ranks, stating that they should not think he wanted them to abandon the recitation of their rules, uddesa,[n.69] Regarding this term cf., e.g., the phrase ekuddeso in the definition of saṃvāsa at Vin III 28,20. Upasak 1975: 43 defines uddesa as "recitation (particularly of the Pātimohkha (sic))"; Cone 2001: 428 lists, besides "recitation" and "brief statement", etc., also "that to which one refers, point of authority" as a possible meaning of uddesa. or their way of life, ājīvā. Instead, they will be free to follow their own uddesa and ājīvā.[n.70] DN 25 at DN III 56,14: yo eva vo uddeso, so eva vo uddeso hotu … yo eva vo ājīvo so eva vo ājīvo hotu (Ce: yo eva te uddeso, so eva te uddeso hotu … yo eva te ājīvo so eva te ājīvo hotu, Se: yo evaṃ te uddeso, soyeva te uddeso hotu … yo ca te ājīvo soyeva te ājīvo hotu).

This is unexpected, since according to what tradition reckons to have been the first sermon delivered by the recently awakened Buddha, to implement his teaching requires a middle path of practice avoiding the two extremes, one of which is asceticism.[n.71] For a study of the Dhammacakkappavattana-sutta in the light of its Chinese parallels cf. Anālayo 2012a and 2013a. This rather gives the impression that ascetics joining the Buddhist order would be required to give up their uddesa and ājīvā, instead of which they would have to adopt the type of rules observed by other Buddhist monastics and their type of livelihood.

The Madhyama-āgama parallel in fact does not have a reference to uddesa or ājīvā. It agrees with the Pāli version in stating that the Buddha was not motivated by desire for disciples,[n.72] Before mentioning the desire for disciples, MĀ 104 at T I 595b18 lists as its first topic the desire to be the teacher. This seems redundant, as this topic is implicitly covered by a reference to desire for disciples, found in both versions. after which the next two reasons given are that the Buddha was also not desirous of offerings or fame.[n.73] MĀ 104 at T I 595b22 and 595b24 lists the suspicions that "the recluse Gotama teaches the Dharma because he is desirous of offerings", 沙門瞿曇貪供養故說法, and "the recluse Gotama teaches the Dharma because he is desirous of fame", 沙門瞿曇貪稱譽故說法. The final reason in the Madhyama-āgama version takes the form of a reference to the wholesome teachings of these wanderers which the Buddha supposedly is going to destroy, probably corresponding to the last of the four reasons given in the Pāli version, according to which the wanderers might suspect that the Buddha wants to establish them in things that they consider unwholesome.[n.74] MĀ 104 at T I 595b26 and DN 25 at DN III 56,19.

In this way, between the reference to desiring disciples and this last reference, shared by the two versions, the Madhyama-āgama version presents a natural continuity of the theme of desire by mentioning offerings and fame, whereas the Pāli version instead brings in uddesa and ājīvā. To my knowledge the present passage in the Udumbarika(-sīhanāda)-sutta is unique in suggesting that wanderers could join the Buddhist order without needing to adopt the rules and form of livelihood incumbent on Buddhist monastics.[n.75] Walshe 1987: 600 note 777 concludes on this part of DN 25 that "the extreme tolerance of Buddhism is shown here. This can be quoted to those who, wishing to practise, e.g. Buddhist meditation, are worried about their prior allegiance to another faith. But see DN 29.4" (in fact the passage quoted, DN 29 at DN III 119,3, stands in direct contrast to what he concludes from DN 25). Brekke 1996: 19 sees in the same passage more a defensive spirit, as "one gets the impression … that the Buddha is used to accusations about ruthless missionary activity among members of other sects." Freiberger 2000: 126 suggests two possible interpretations of the passage in DN 25: a) the Buddha indeed invited the wanderers to continue with their way of life, etc., b) the passage is a rhetorical means for the purposes of conversion. In this respect the Madhyama-āgama version has preserved a presentation that concords better with what other texts convey.

MĀ 111: The Cause and Result of Feeling

Proceeding from asceticism to feeling, a discourse among the Sixes of the Aṅguttara-nikāya and its Madhyama-āgama parallel agree in presenting contact as what causes feeling, but differ when it comes to the result (vipāka) of feeling. Whereas the Madhyama-āgama refers to craving, a presentation clearly in line with the standard exposition of dependent arising (paṭicca samuppāda), the Pāli discourse rather speaks of the production of individual existence as the result of one's merit or demerit.[n.76] MĀ 111 at T I 599c12 and AN 6.63 at AN III 411,23.

The same is already mentioned in the preceding passage in the Pāli version, which takes up the topic of the result of sensual pleasures. The Aṅguttara-nikāya discourse and its Madhyama-āgama parallel agree that sensual pleasures indeed result in a form of existence related to one's merit and demerit.[n.77] MĀ 111 at T I 600a14 and AN 6.63 at AN III 412,25 (the sequence of topics differs in the two versions). This makes it safe to conclude that the Aṅguttara-nikāya discourse would have suffered from an error in transmission whereby a passage originally only meant to explain the results of sensual pleasures was accidentally also applied to the results of feeling.[n.78] The same type of error during transmission can also be seen in MĀ 111 at T I 600b11 and 600b14, which seems to have mixed up the result and the diversity of dukkha, in contrast to the preferable presentation in AN 6.63 at AN III 416,14, according to which the diversity of dukkha is to be seen in its different manifestations (strong or weak, etc.) and the result of dukkha in being afflicted.

MĀ 112: Wholesome and Unwholesome Qualities

Feelings and their relation to craving are of course central in the arising of either wholesome or else unwholesome qualities. The distinction between these two is taken up in another discourse among the Sixes of the Aṅguttara-nikāya and its Madhyama-āgama parallel, which describe different individuals according to the wholesome or unwholesome quality of their mind.[n.79] A difference in relation to their respective introduction to this exposition is similar to a difference between MN 136 at MN III 209,12 and its parallel MĀ 171 at T I 707a19, which I discussed in Anālayo 2011b: 778. In the present case, too, MĀ 112 at T I 601a15 has the complete sentence, whereas AN 6.62 at AN III 404,5 seems truncated. The second case taken up in the Aṅguttara-nikāya version concerns someone who manifests wholesome qualities but the roots of unwholesomeness are still in the mind and will eventually lead to the decline of that person. The second case in the Madhyama-āgama parallel is someone who manifests unwholesome qualities but still has the roots of wholesomeness, which will eventually also be cut off.

The two versions agree in illustrating their respective second case with good seeds sown in a place where they cannot grow.[n.80] In MĀ 112 at T I 601b29 the seeds will not grow because they do not receive timely rain; in AN 6.62 at AN III 405,20 they will not grow because they have been put on a rock. The simile seems appropriate for illustrating wholesome roots that will not lead to wholesome states, as is the case of the Madhyama-āgama version, not unwholesome roots as in the Pāli version, for which spoiled seeds would be more appropriate.

This can in fact be seen in relation to the case of someone who is thoroughly unwholesome, in which case both versions speak of spoiled seeds. In the Madhyama-āgama version's illustration of this thoroughly unwholesome case the spoiled seeds are sown in a barren field without timely water, whereas in the Aṅguttara-nikāya discourse the spoiled seeds are sown in a well-prepared field.[n.81] MĀ 112 at T I 601c14 and AN 6.62 at AN III 406,14. As an illustration of the utterly hopeless case of someone completely immersed in unwholesomeness the Madhyama-āgama illustration appears to fit better. In sum, it seems fair to assume that the similes in the Aṅguttara-nikāya discourse have suffered from an error in transmission.

Another puzzling aspect in the Aṅguttara-nikāya discourse is a reference in yet another simile, which mentions midnight as the time for the meal.[n.82] AN 6.62 at AN III 407,19: aḍḍharattaṃ (Be: addharattaṃ) bhattakālasamaye. Based on the commentarial gloss on bhattakālasamaye at Mp III 406,7 as rājakulānaṃ bhattakālasaṅkhāte samaye, Bodhi 2012: 1767 note 1408 reasons that "perhaps in the Buddha's time the royal court ended the day with a midnight meal." The Madhyama-āgama parallel agrees with the Aṅguttara-nikāya version in regard to the preceding part of the simile, which refers to the rising of the sun. It differs in so far as it next describes that the sun has risen higher and mealtime has arrived.[n.83] MĀ 112 at T I 601a28: "the sun has risen high and the time of the meal has arrived", 日轉昇上, 至于食時; although the wording seems clearly preferable, the placing of this simile in MĀ 112 is less apt. This is a more natural phrasing of the simile and the reference to a meal at midnight could be the result of some confusion in the transmission of this part of the Aṅguttara-nikāya discourse.

MĀ 113: Two Parallels – One Discourse

Proceeding with another example that involves the Aṅguttara-nikāya, a discourse in the Madhyama-āgama actually has two parallels in this collection, found among the Eights and the Tens of the Aṅguttara-nikāya. The parallel versions agree in presenting a series of short questions and their replies. The difference between the two Aṅguttara-nikāya versions is that the one among the Tens has two additional questions, which query what all things culminate in and what constitutes the consummation of all things. The reply is that all things culminate in the deathless and all things have Nirvāṇa as their consummation.[n.84] AN 10.58 at AN V 107,10; neither topics is covered in the otherwise similar AN 8.83 at AN IV 339,4. As already pointed out by Bodhi (2012: 1845 note 2069), these two replies "seem to be synonymous".

The Madhyama-āgama parallel agrees with its Aṅguttara-nikāya counterparts in mentioning that wisdom is supreme among all things, whose essence is liberation, which are the last two qualities in the listing given in the Aṅguttara-nikāya discourse found among the Eights and the two qualities that in the discourse among the Tens lead up to the references to the deathless and Nirvāṇa.

The Madhyama-āgama version just mentions a single item at this juncture, which is Nirvāṇa as the completion of all things.[n.85] MĀ 113 at T I 602c16: 涅槃為諸法訖. Another parallel preserved as an individual translation, which shows several differences compared to MĀ 113 and thus seems to be reflecting a different lineage of transmission, agrees in also mentioning only a single item, namely Nirvāṇa as the completion of all things; cf. T 59 at T I 855c16. The Madhyama-āgama discourse then continues with an exposition that parallels the next discourses among the Tens of the Aṅguttara-nikāya.[n.86] This has already been pointed out by Bodhi 2012: 1845 note 2069. This gives the impression that perhaps during oral transmission the last reference to Nirvāṇa came to be doubled by adding the synonymous deathless and in some way the remainder of the text also became a separate discourse. In fact this separate discourse starts off with the phrase "therefore monks, you should train like this", tasmā ti ha bhikkhave, evam sikkhitabbaṃ, which does hang a little in the air.[n.87] AN 10.59 at AN V 107,14. Such a phrase would serve naturally as a connector to something described earlier, as it does in the Madhyama-āgama version.[n.88] MĀ 113 at T I 602c16 continues after the reference to Nirvāṇa as the completion of all things with the phrase "therefore, monks, you should train like this", 是故比丘, 當如是學, to introduce the remainder of its exposition. A comparable transition can also be found at this juncture in the other parallel, T 59 at T I 855c21: "monks, train like this", 諸比丘, 當學是. The same is considerably less natural as an introductory phrase for a new exposition.

MĀ 116: The Buddha's Refusal to Ordain Women

Another discourse among the Eights in the Aṅguttara-nikāya reports that the Buddha, on being requested to grant the going forth to his fostermother and thereby found an order of nuns, flatly refused.[n.89] AN 8.51 at AN IV 274,7. According to the Madhyama-āgama parallel, however, he offered her the alternative of shaving off her hair and donning robes to live a form of celibate life,[n.90] MĀ 116 at T I 605a17. presumably in the more protected environment at home instead of wandering around freely in India.

Given that a number of other discourses consider the existence of an order of nuns an integral aspect of the teaching of the Buddha, the Madhyama-āgama version seems to have preserved an earlier version of the Buddha's reaction to the request of his fostermother, in that the point at issue would have been how her going forth should take place, not whether to grant it to her at all.[n.91] Cf. in more detail Anālayo 2011d and 2016c.

MĀ 128: Giving to Outsiders

Still continuing with the Aṅguttara-nikāya, a discourse among the Fives of this collection has a verse according to which fools give to those outside (of the Buddha's dispensation).[n.92] AN 5.179 at AN III 214,14: bālā ca avijānantā, dummedhā assutāvino, bahiddhā denti (Be: dadanti) dānāni. This stands in direct contrast to a discourse among the Threes of the same Aṅguttara-nikāya, which in agreement with several parallels reports that the Buddha quite explicitly denied that he had ever stated that gifts should only be given to those in his dispensation and not to others.[n.93] AN 3.57 at AN I 161,13 and its parallels SĀ 95 at T II 26a14, SĀ2 261 at T II 465c11, and SĀ3 2 at T II 493b22. Another contrastive example is the report in the Upāli-sutta and its parallel that the Buddha encouraged the recently converted Upāli to continue giving alms to his former teachers, the Jains.[n.94] MN 56 at MN I 379,16 and MĀ 133 at T I 630b8.

The Madhyama-āgama parallel to the discourse from the Fives of the Aṅguttara-nikāya has a different version of the respective verse. This version instead proclaims that giving to a fool will bring little fruit.[n.95] MĀ 128 at T I 617b10: "to a fool who knows nothing, who has no wisdom and has learned (literally: 'heard') nothing, giving to him will yield little fruit", 愚癡無所知, 無慧無所聞, 施彼得果少. Such a formulation of the verse implies no negative assessment of outsiders as recipients of gifts and concords with the presentation of the fruitfulness of gifts found elsewhere in the discourses. This leaves open the possibility that the present version of the verse in the discourse from the Fives of the Aṅguttara-nikāya is the result of a later change. In fact the Dakkhiṇāvibhaṅga-sutta and a range of parallels explicitly describe the worthiness of a gift made to an outsider free from lust,[n.96] MN 142 at MN III 255,8, Bajaur Kharoṣṭhī fragment 1 line 20, MĀ 180 at T I 722b18, and D 4094 ju 255b6 or Q 5595 tu 291a7; cf. also Tocharian fragment YQ 1.20 1/2a8, Ji et al. 1998: 182, and Uighur fragment folio 9a2700, Geng and Klimkeit 1988: 204, which refer to a seer and thus implicitly to an outsider. A reference to this exposition in EĀ 23.1 at T II 609b15 also refers to a seer free from lust. Another parallel, T 84 at T I 903c29, does not explicitly qualify the recipient to be an outsider or a seer and thus just speaks of someone free from lust. thereby confirming that the presentation in the Aṅguttara-nikāya discourse is out of keeping with the position taken in general in early Buddhist discourse.

The Dakkhiṇāvibhaṅga-sutta itself has another noteworthy instance, also related to the worthiness of making gifts. The Pāli version states that a gift received by an immoral and evil monk will be of incalculable merit, as long as it is given to the community as a whole.[n.97] MN 142 at MN III 256,5.

The Madhyama-āgama discourse instead just speaks of a gift to a monk who is not energetic.[n.98] MĀ 180 at T I 722b1; cf. in more detail Anālayo 2011b: 817. In the case of the Dakkhiṇāvibhaṅga-sutta the suggestion that an immoral monk is a worthy recipient of gifts stands in contradiction to the remainder of the same discourse, which highlights the importance of the ethical behaviour of the recipient (as well as of the giver) of a gift. This clearly accords better with the position taken in general in the early discourses, making it fair to assume that the incalculable merit accrued by a gift received by an immoral and evil monk is the result of a later change in the Pāli discourse.

MĀ 153: Similes out of Context

Moving from gifts to sensual pleasures in general, my next example involves similes, found in the Māgandiya-sutta as part of a detailed examination of the drawbacks of sensual pleasures. The Madhyama-āgama parallel to the Māgandiya-sutta begins with a simile describing a leper who scratches and cauterizes his wounds over a fire.[n.99] MĀ 153 at T I 671b25. Although the leper experiences momentary pleasure, his scratching and cauterizing results in the wounds becoming infected and his condition becoming worse. This illustrates indulgence in sensual pleasures, which likewise yields momentary gratification but results in worsening one's mental condition, as the tendency to crave for sensual pleasure is strengthened.

The next simile describes a king or minister who has easy access to sensual pleasures, yet is unable to experience mental peace if he is not free from desire for sensual pleasures. This expands on the point made above, in that indulgence in sensuality does not lead to true inner peace of mind.

The Madhyama-āgama discourse next describes how someone healed from leprosy would not find it attractive to see another leper, who is still sick, scratching and cauterizing his wounds. Yet another simile depicts how someone healed from leprosy would resist with all his strength being forcefully dragged close to a fire. This simile provides a contrast to the earlier situation mentioned in the first simile, when he was still sick and therefore attracted to fire in order to cauterize his wounds. The last two similes of seeing another leper and of being dragged close to fire illustrate how one who has overcome sensual desire no longer delights in what earlier, when still under the influence of sensual desire, seemed so attractive. The sequence of the similes in the Madhyama-āgama is thus as follows:

leper scratches and cauterizes wounds,
king not free from sensual desire,
healed leper sees other leper scratching and cauterizing wounds,
healed leper dragged forcefully close to fire.

Whereas the Madhyama-āgama presentation follows an inherent logic in its presentation, the Māgandiya-sutta has the following sequence:[n.100] MN 75 at MN I 506,6.

healed leper sees other leper scratching and cauterizing wounds,
healed leper dragged forcefully close to fire,
leper scratches and cauterizes wounds,
king not free from sensual desire.

This sequence appears jumbled. The point made by the two similes of the healed leper comes out fully only if the situation of the sick leper worsening his own condition by scratching and cauterizing his wounds has already been described. Moreover, the sequence in the Māgandiya-sutta lacks the gradual build-up that can be observed on reading the Madhyama-āgama presentation, where the culmination point comes with the image of struggling to avoid being dragged towards the fire that earlier was eagerly sought. In contrast, in the Pāli version the simile of the king comes almost as an anticlimax. This makes it possible that the Māgandiya-sutta has suffered from a positional alteration of the four similes, which can be corrected with the help of its Madhyama-āgama parallel.

MĀ 154: Reasons for Pasenadi's Respect

Shifting from similes to the topic of caste, as part of a discussion of claims to caste superiority by brahmins the Aggañña-sutta and its parallels describe the reasons why King Pasenadi pays respect to the Buddha. According to several Pāli editions, the king thinks that the Buddha is of superior birth to him, a reading followed by translators of this passage into German and English.[n.101] DN 27 at DN III 84,3: nanu sujāto samaṇo gotamo, dujjāto 'ham asmi (same in Ce and Se). Neumann 1912/2004: 478: "'ist denn nicht', sagt er, 'der Asket Gotamo wohlgeboren? Unwohlgeboren bin ich'"; Franke 1913a: 276: "der Samaṇa Gotama ist ja doch hochwohlgeboren, ich bin (ihm gegenüber) von niederer Herkunft"; Rhys Davids and Rhys Davids 1921: 80: "for he thinks: Is not the Samaṇa Gotama well born? Then I am not well born"; Walshe 1987: 409: "thinking: 'If the ascetic Gotama is well-born, I am ill-born.'" This reading fails to make sense in its context, since the whole point of the passage is to undermine claims of superiority based on birth.

In the Madhyama-āgama parallel, as well as several other parallels extant in Chinese and Tibetan, the corresponding passage instead indicates that it is not because of superiority of birth that the king pays respect to the Buddha.[n.102] MĀ 154 at T II 674b8: "the king does not think like this … 'The recluse Gotama is from a superior clan, I am from an inferior clan'", 王不如是意 … 沙門瞿曇種族極高, 我種族下. Similar statements are found in the parallels DĀ 5 at T I 37b23: "he does not think: 'The recluse Gotama has gone forth from an excellent clan, my clan is inferior'", 彼不念言: 沙門瞿曇出於豪族, 我姓卑下, and T 10 at T I 218a20, according to which "it is not because the king thinks that the recluse Gotama is from a superior clan, yet the king does not come from a superior clan", 其王不以沙門瞿曇是高勝族, 王亦不起高勝族意; cf. also D 4094 ju 192a5 or Q 5595 tu 219b1. SHT V 1583 V1, Sander and Waldschmidt 1985: 264, has preserved a reference to being duvarṇa. This fits the context, making it clear that the king's respect is not related to any birth superiority, but is rather due to regard for the Dharma taught by the Buddha. This concords with the reading of the Pāli passage found in the Burmese edition.[n.103] The Be edition reads: na naṃ sujāto samaṇo gotamo, dujjāto 'ham asmi. Thus in this case the Chinese versions help to decide which variant reading is the most appropriate one.

A comparable case can be found in the Dhammacetiya-sutta, which in agreement with its parallels reports that King Pasenadi described several reasons why he respected the Buddha. Reasons given in the Pāli version are that they are both from the warrior clan, both from Kosala, and both eighty years old.[n.104] MN 89 at MN II 124,16. Besides confirming that the presentation in most editions of the Aggañña-sutta even conflicts with other Pāli discourses, where the king and the Buddha are presented as being of the same clan, the present passage is yet another instance where the topic of Pasenadi's respect can be considered from the viewpoint of the relevant Madhyama-āgama parallel. The problem with the presentation in the Dhammacetiya-sutta is that the mere fact of being from the same clan and country, and of the same age, is not easily conceived as a reason to inspire respect.

Besides mentioning the fact that the two are from the same clan and Kosala country, as well as of the same age, the Madhyama-āgama parallel highlights that, whereas Pasenadi is the king of the country, the Buddha is the king of Dharma.[n.105] MĀ 213 at T I 797b12: 我亦國王, 世尊亦法王. A similar distinction is made in a parallel in the Kṣudrakavastu of the Mūlasarvāstivāda Vinaya; cf. Anālayo 2011b: 518 note 367. This marks a difference that would indeed provide a meaningful reason for respect, making it reasonable to assume that some such statement might have been lost in the Dhammacetiya-sutta.

MĀ 158: Brahmins of Different Types

Continuing from brahminical claims to caste superiority to the theme of differences among brahmins, a discourse found in the Fives of the Aṅguttara-nikāya and its Madhyama-āgama parallel describe brahmins of different types. In the Pāli version the first type of brahmin, having completed his period of celibacy and study of forty-eight years, goes begging so as to make an offering to his teacher.[n.106] AN 5.192 at AN III 225,3. Bodhi 2012: 1742f note 1185 comments on the impression conveyed by AN 5.192 that it "suggests, contrary to a common assumption, that during the Buddha's time brahmins were not obliged to marry and adopt the life of a householder … that in this period some brahmins, after completing their training, chose to renounce secular life even in their youth and maintained their renunciant status throughout their lives." Then he goes forth, practises the four brahmavihāras, and is reborn in the Brahmā world. In the Madhyama-āgama parallel this type of brahmin also completes his period of celibacy and study of forty-eight years, and goes begging so as to make an offering to his teacher. The description continues by just mentioning his cultivation of the four brahmavihāras, without any reference to going forth.[n.107] MĀ 158 at T I 680c24.

This squares better with what we know about brahminical practices in ancient India. Perhaps during oral transmission the reference to the forty-eight years of celibacy, brahmacariya, and to begging, bhikkhācariya, might have led to the interpolation of the standard description of going forth.

Another noteworthy difference can be seen in relation to the next type of brahmin described in the Aṅguttara-nikāya discourse. After having completed the period of study and made an offering, this type of brahmin gets married, then goes forth, cultivates the four absorptions, and is reborn in a heavenly world.[n.108] AN 5.192 at AN III 227,1. This brahmin is considered in the Aṅguttara-nikāya discourse to be like a deva, in contrast to the previous type who is like a Brahmā. Yet the type of rebirth suiting a brahmin who is like a deva would be one of the heavenly realms of the sense-sphere, not the heavenly realms of the form-sphere or Brahmā world to be expected for one who cultivates the four absorptions.

The Madhyama-āgama parallel in fact does not describe any meditation practice in the case of this type of brahmin (nor any going forth), but just reports that he cultivates good bodily, verbal, and mental conduct, as a result of which he is reborn in a heavenly world.[n.109] MĀ 158 at T I 681a10. This seems more appropriate for a description of a brahmin who is like a deva.

MĀ 163: Three Satipaṭṭhānas

Turning from brahmins to a form of mindfulness associated specifically with the Buddha, the Saḷāyatanavibhaṅga-sutta describes three satipaṭṭhānas practised by the Buddha when teaching disciples. According to its presentation, when the disciples do not listen, the Buddha is not satisfied.[n.110] MN 137 at MN III 221,10. But when all of them listen, he is satisfied. In the Madhyama-āgama parallel the Buddha remains unaffected by the behaviour of his disciples.[n.111] MĀ 163 at T I 693c29. This presentation receives support from a range of other sources and may well be closer to the original intention behind the presentation of these three satipaṭṭhānas.[n.112] Cf. in more detail Anālayo 2013f: 240–243.

MĀ 164: A Summary without Its Exposition

Continuing with the topic of the Buddha's teaching activities, my next example comes from the Uddesavibhaṅga-sutta. The beginning part of the discourse reports the Buddha's announcement that he will give a summary and its exposition:

Monks, I will teach you a summary and its exposition.[n.113] MN 138 at MN III 223,5: uddesavibhaṅgaṃ vo, bhikkhave, desissāmi (Be: desessāmi).

Nevertheless, according to the Uddesavibhaṅga-sutta the Buddha only taught a summary and then withdrew into his dwelling. This problem does not arise in the Madhyama-āgama version, as here the Buddha does not make such an announcement. Instead he introduces his teaching as follows:

I will teach you the Dharma that is sublime in its beginning, sublime in its middle, and sublime in its end, with its meaning and phrasing.[n.114] MĀ 164 at T I 694b16: 我當為汝說法, 初妙, 中妙, 竟亦妙, 有義有文.

Hence in the Madhyama-āgama version for the Buddha to teach only in brief and then withdraw does not create a comparable inconsistency. The problem observable in the Uddesavibhaṅga-sutta concords with a pattern for discourses in this chapter of the Majjhima-nikāya of referring to a summary and its exposition in ways that do not fit the remainder of the discourse.[n.115] Cf. in more detail Anālayo 2008g and 2010g: 48f. In the present case this is particularly noteworthy, since the initial proclamation about giving a summary, uddesa, and its exposition, vibhaṅga, corresponds to the title of the Pāli discourse. [n.116] The title is found in MN 138 at MN III 229,32. The Madhyama-āgama discourse instead has the title "Analytical Contemplation of Dharmas".[n.117] MĀ 164 at T I 694b13: 分別觀法.

MĀ 179: The Path-factors of an Arahant

Shifting from the Buddha to the arahant, the Samaṇamaṇḍikā-sutta briefly refers to the ten path factors of an arahant at the outset of its exposition, but has the corresponding exposition only at a considerably later point. In this way the standard procedure of following an announcement of a particular topic directly with its exposition is not followed.

Judging from its Madhyama-āgama parallel, the relationship between these two pieces of text has been lost during oral transmission, resulting in a displacing of the introductory statement to its present position.[n.118] MN 78 at MN II 25,18 (introduction) and 28,34 (exposition), compared to both being found consecutively in MĀ 179 at T I 721c14. This is one of several aspects of the Pāli discourse that can be clarified with the help of its Madhyama-āgama parallel.[n.119] Cf. Anālayo 2009f and 2011b: 424–431.

MĀ 181: Impossibilities for Women

Returning to the Buddha, one of a set of five impossibilities in the Bahudhātuka-sutta stipulates that a woman cannot be a Buddha. The whole set of five impossibilities is absent from the Madhyama-āgama parallel.[n.120] MN 115 at MN III 65,24 (cf. also T 776 at T XVII 713b20, T 1537 at T XXVI 502b16, D 297 sha 300b1 or Q 963 lu 329a5, and D 4094 ju 32a3 or Q 5595 tu 35a4); the exposition of various impossibilities, which does not include any impossibility for women, can be found in MĀ 181 at T I 723c26. A study of the context shows this dictum to be probably a later addition to the Bahudhātuka-sutta.[n.121] Cf. Kajiyama 1982: 57 and in more detail Anālayo 2009b.

MĀ 184: Mahāmoggallāna and Abhidharma

Shifting again from the Buddha to an arahant, my next case concerns Mahāmoggallāna in the Mahāgosiṅga-sutta. In agreement with its parallels, the Mahāgosiṅga-sutta reports an occasion when several eminent monks extolled their special qualities or abilities in a poetic contest on a moonlit night. In the Mahāgosiṅga-sutta, after Ānanda, Revata, Anuruddha, and Mahākassapa have expressed themselves in line with the way their personality is depicted elsewhere in the early discourses, Mahāmoggallāna comes out with the following statement that supposedly illustrates his special and characteristic abilities:

Two monks talk Abhidharma talk. They ask each other a question. On asking each other a question, they respond and do not founder, and their talk on the Dharma goes forward.[n.122] MN 32 at MN I 214,24: dve bhikkhū abhidhammakathaṃ kathenti, te aññamaññaṃ pañhaṃ (Se does not have pañhaṃ) pucchanti, aññamaññassa pañhaṃ puṭṭhā vissajjenti, no ca saṃsādenti, dhammī ca nesaṃ kathā pavattanī (Be: pavattinī) hoti.

Instead of such ability at discussing the "higher Dharma" (abhidharma),[n.123] On the significance of the term abhidharma in such contexts cf. Anālayo 2014c: 69–79. the early discourses usually represent it as characteristic of Mahāmoggallāna to have supernormal abilities.[n.124] As already pointed out by Horner 1941: 309, "Moggallāna is chiefly famed for his psychic powers and there is little reason to suppose him to have had gifts of an abhidhamma nature or we should have heard more about them." This is indeed what the Madhyama-āgama parallel to the Mahāgosiṅga-sutta records as his statement on this occasion, which reads as follows:

A monk has great supernormal power, great and mighty virtue, great merit, great and mighty power, mastery of immeasurable supernormal powers.[n.125] MĀ 184 at T I 727c15: 比丘有大如意足, 有大威德, 有大福祐, 有大威神, 自在無量如意足.

The Madhyama-āgama version of Mahāmoggallāna's statement continues by describing various feats that according to early Buddhist thought can be performed by someone endowed with such supernormal powers, such as walking on water, levitation, etc.[n.126] On levitation cf. above p. 281ff. In a similar vein, parallels to the Mahāgosiṅga-sutta preserved in the form of Sanskrit fragments, in a discourse in the Ekottarika-āgama, and in a discourse extant as an individual translation into Chinese consider supernormal powers to be characteristic of Mahāmoggallāna.[n.127] SHT V 1346 V2, Sander and Waldschmidt 1985: 233, introduces mahāmau [dga](lyāyan), and the next line has preserved bhikṣūr-bhavati mahardhiko ma[h], followed by parts of the standard description of the exercise of supernormal abilities in the next lines. In EĀ 37.3 at T II 711a18 Mahāmoggallāna similarly refers to a monk endowed with supernormal ability, 比丘有大神足, and then describes various feats. The same is the case for the statement attributed to him in T 154 at T III 81b29: 比丘得大神足.

The Madhyama-āgama discourse also has a statement on Abhidharma type of talk, but in its presentation this is instead characteristic of the monk Mahākaccāna. The relevant part proceeds in this way:

Two monks who are Dharma teachers discuss with each other the profound Abhidharma. Whatever matter they ask about, they understand it well and know it completely. They reply without hesitation and expound the Dharma eloquently.[n.128] MĀ 184 at T I 727b23: 二比丘法師共論甚深阿毘曇. 彼所問事善解悉知, 答亦無礙, 說法辯捷. Already Anesaki 1908: 57 had pointed out that in MĀ 184 "besides the other theras there is Mahākaccāna. He plays the role of Moggallāna in Pāli and the utterances of Moggallāna which we find in Chinese are wanting in Pāli."

To associate such a statement with Mahākaccāna is more in line with the way the discourses represent him elsewhere than to attribute this statement to Mahāmoggallāna.[n.129] Minh Chau 1964/1991: 76 concludes that "the Chinese version here is more faithful and more reliable as Mahākaccāna is well known as being expert in Dharma discussion and Moggallāna is well known as being expert in supernatural powers." This reasonable conclusion is based on the criterion of coherence. The relevance of coherence for assessing early Buddhist discourse material in general has been called into question by Sujato 2011: 213, who argues instead that contradiction is a feature "found in Buddhism, for Sāriputta is the disciple who is 'like the Teacher', yet the Buddha is 'without a counterpart'. The Indic mind is simply less concerned with contradiction than the European … this is true even in the sphere of logic. For Aristotle, any proposition was either true (A) or false (not-A) … but the Indians had a four-fold scheme: A; not-A; A & not-A; neither A nor not-A." Neither example is a case of actual contradiction. Sāriputta is only qualified as a disciple comparable to the teacher, not as his equal; cf. above p. 504f note 8. The tetralemma is compatible with the principle of coherence; cf. Anālayo 2011b: xxvii. The early discourses in fact explicitly indicate that consistency was regarded as a criterion of truth. Examples are debate situations, where the Buddha is shown to employ the criterion of consistency when countering incoherent statements made by an opponent; cf. Anālayo 2010e: 49 note 27. Jayatilleke 1963/1980: 334 concludes that "in the Nikāyas, consistency is regarded as a criterion of truth." Vetter 1988: ix notes that the doctrine of "ancient Buddhism … does not give any reason for proposing … inconsistency as a characteristic of it"; cf. also Watanabe 1983/1996: 74f. Reat 1996/1998: 34 reasons that "one must assume that the historical Buddha's teachings were coherent … when they were given. Therefore any reconstruction of these teachings should reveal a coherent framework of doctrine." Tilakaratne 2000a: 14 points out that the mahāpadesas are based on the premise "that what is called Dhamma and Vinaya is characterised by internal consistence and coherence". This makes it safe to conclude that an error in oral transmission led to a loss of Mahākaccāna in the Pāli version and to a statement originally attributed to him becoming instead associated with Mahāmoggallāna.

MĀ 187: Loss of One of the Purities of an Arahant

Continuing with the topic of an arahant, the Chabbisodhana-sutta expounds five purities of an arahant, even though its title speaks of six.[n.130] The title is found in MN 112 at MN III 37,5. By having recourse to the Madhyama-āgama parallel the missing sixth purity can be restored.[n.131] The missing purity concerns the nutriments, expounded in MĀ 187 at T I 732b18. In this case a whole portion of a discourse was apparently lost during the transmission from India to Sri Lanka, as the commentarial tradition preserves a reference to reciters from India who were still aware of the nature of the missing purity.[n.132] Ps IV 94,23; cf. in more detail Anālayo 2008b.

MĀ 189: Supramundane Path Factors

Turning from textual loss to its addition, a case of substantial addition can be identified in the exposition of the path factors in the Mahācattārīsaka-sutta, whose presentation of supramundane path factors is without a counterpart in its Madhyama-āgama parallel.[n.133] The supramundane exposition in the case of right intention can be found in MN 117 at MN III 73,11; for the treatment of this same path factor cf. MĀ 189 at T I 736a1 and D 4094 nyu 44b7 or Q 5595 thu 84b2. The absence of the supramundane path treatment from MĀ 189 has already been noted by Meisig 1987b: 226; for a more detailed discussion cf. Anālayo 2010d. The evident addition of the supramundane dimension in the Pāli version reflects the influence of evolving Abhidharma thought.[n.134] Cf. in more detail Anālayo 2014c.

MĀ 190: Emptiness and Neither-perception-nor-nonperception

Another apparent addition can be found in the Cūḷasuññata-sutta, which describes a gradual meditative descent into emptiness that employs the perceptions related to the first three immaterial attainments. Whereas its Madhyama-āgama parallel proceeds from the perception of nothingness to signlessness, the Cūḷasuññata-sutta has as an interim step the attainment of neitherperception-nor-non-perception.[n.135] MĀ 190 at T I 737c3 (emending 無想 to 無相; cf. also Anālayo 2012d: 30 note 16) and MN 121 at MN III 107,10; cf. also Skilling 1994: 172,5. This appears to be the result of an error in transmission, where the fact that usually in other contexts a reference to the three immaterial attainments will be followed by the fourth would have led to its intrusion in the present context.[n.136] Cf. in more detail Anālayo 2015e: 134–136.

MĀ 192: Loss of a Transition

Returning from textual addition to its loss, the Laṭukikopama-sutta reports that the monk Udāyin has just described an event from the past. Udāyin concludes by expressing his appreciation for the rules on restraint with food which the Buddha had promulgated. This then motivates the Buddha to describe foolish people who are not willing to follow his rules on restraint. The transition to this topic in the Laṭukikopama-sutta takes the form of the Buddha replying to Udāyin's expression of appreciation in the following manner:

In the same way, Udāyin, there are some foolish people …[n.137] MN 66 at MN I 449,10: evam eva pan' udāyi, idh' ekacce moghapurisā …

The use of the expression "in the same way", evam eva, does not fit the present context well, since the foolish people whom the Buddha is about to describe are the opposite of Udāyin. Therefore it does not seem meaningful to introduce them with an expression that conveys similarity. In the Madhyama-āgama parallel, the corresponding transition to a description of foolish people, who are unwilling to follow the Buddha's injunctions, proceeds as follows:

The Blessed One praised him: "It is well, it is well, Udāyin, you are now not like those foolish people. Those foolish people, on being taught by me …"[n.138] MĀ 192 at T I 741b23: 世尊歎曰: 善哉, 善哉, 烏陀夷, 汝今不爾如彼癡人. 彼愚癡人, 我為其說 …

This confirms the impression that some sort of transition might have been lost in the Laṭukikopama-sutta.

The Pāli passage translated above is not preceded by an indication that a change of speaker has taken place. Only the circumstance that the actual speech is addressed to Udāyin makes it clear that at this point he is no longer the speaker, but rather the one to whom something is being said.

Perhaps an explicit indication of the change of speaker, as found in the Madhyama-āgama discourse, was lost in the Laṭukikopama-sutta together with an introductory reference by the Buddha to foolish people that stand in contrast to Udāyin. Some such reference seems required to provide a reference point to which the expression "in the same way" could refer.

MĀ 198: Satipaṭṭhāna and the Second Absorption

Another apparent loss of text involves the Dantabhūmi-sutta, which describes a progression of practice where cultivation of satipaṭṭhāna meditation free from thoughts directly leads to attainment of the second absorption.[n.139] MN 125 at MN III 136,26. Consultation of its Madhyama-āgama parallel shows that here the first absorption is taken into account.[n.140] MĀ 198 at T I 758b25, which gives the introductory formula for the first absorption and then abbreviates up to the attainment of the fourth, a standard procedure in the collection to represent attainment of all four absorptions. This makes it fairly safe to assume that the reference to being free from vitakka in the description of the preceding satipaṭṭhāna practice accidently caused a loss of the description of the first absorption during oral transmission by directly continuing with the second absorption, whose standard description refers to being free from vitakka.[n.141] This is one of several findings that emerge from a comparative study of this discourse; cf. Anālayo 2006b and 2011b: 717–722.

MĀ 201: The Tathāgata and the Gradual Path

Moving again from loss of text to its addition, my next case comes from the Mahātaṇhāsaṅkhaya-sutta. The relevant passage in this discourse and its Madhyama-āgama parallel begins with a reference to the arising of the Tathāgata in the world. The Madhyama-āgama discourse continues by describing detachment towards sensory experience and being established in mindfulness of the body.[n.142] MĀ 201 at T I 769c14. The transition from the arising of a Tathāgata to this exposition is not entirely smooth, however, as the passage just refers to the one who is detached as "he", 彼, thereby lacking any explicit indication that this should be understood to refer to a disciple of the Tathāgata.

The Mahātaṇhāsaṅkhaya-sutta similarly refers to detachment towards sensory experience and mindfulness of the body,[n.143] MN 38 at MN I 270,9. with the difference that between this reference and the remark on the arising of a Tathāgata it provides a detailed account of the gradual path of practice.[n.144] On the gradual path of practice cf. in more detail Anālayo 2016d. As a result, it deals with the issue of how to handle sensory experience twice, once in the form of sense-restraint as part of the gradual path and again in the passage just mentioned, which is the one that is also found in the Madhyama-āgama parallel.

The resultant sequence of topics in this part of the Mahātaṇhāsaṅkhaya-sutta is as follows:

arising of a Tathāgata,
going forth,
moral restraint,
contentment,
sense-restraint,
clear comprehension of bodily activities,
removal of the five hindrances,
attainment of the four absorptions,
detachment and mindfulness of the body.

In this way the last reference to detachment in relation to the senses and mindfulness of the body comes in the Mahātaṇhāsaṅkhaya-sutta after the gradual path has been completed up to the successful attainment of the four absorptions. This is contrary to the usual sequence of practice depicted in the early discourses, where mindfulness of the body and detachment in relation to sensory experiences are pre-conditions for the attainment of absorption and therefore mentioned earlier. In the scheme above, both should come in the place now occupied by sense-restraint and by clear comprehension of bodily activities.

Taken together these features suggest that a reference to the arising of a Tathāgata has misled the reciters of the Pāli version to insert a full account of the gradual path where this was not originally required. Such an error could happen easily, since the standard description of the arising of a Tathāgata does often function as an introduction to an account of the gradual path.

MĀ 204: Brahmā's Intervention

The arising of a Tathāgata was according to tradition due to the famous intervention of Brahmā to convince the recently awakened Buddha to teach his discovery to mankind. This intervention is not mentioned at all in the Madhyama-āgama parallel to the Ariyapariyesanā-sutta, [n.145] Brahmā's intervention is reported in MN 26 at MN I 168,18; the parallel MĀ 204 at T I 777a18 continues directly from the Buddha's awakening to his reflection on whom he might teach first. The absence of this episode in MĀ 204 has already been noted by Minh Chau 1964/1991: 155. which leaves open the possibility that this episode could be a later addition.[n.146] Cf. in more detail Anālayo 2011a.

MĀ 207: The Path of Practice

Proceeding from the Buddha's decision to teach to a description of his qualities, the Mahāsakuludāyi-sutta and its Madhyama-āgama parallel agree in listing five qualities of the Buddha. Under the fifth of these, the Pāli version offers a long and detailed survey of various aspects of the path of practice, absent from its Madhyama-āgama parallel.[n.147] MN 77 at MN II 11,3 to 22,15 and thus over eleven pages compared to just two lines of text in its parallel MĀ 207 at T I 783b15 to 783b17. Closer inspection of this list and its context makes it fairly probable that this is a case of expansion on the part of the Mahāsakuludāyi-sutta.[n.148] Cf. in more detail Anālayo 2009c: 155–158.

MĀ 208: The Gradual Path Again

Another example related to the description of the path of practice can be observed in the Cūḷasakuludāyi-sutta. In this case the Pāli discourse and its Madhyama-āgama parallel agree in presenting an account of the gradual path, but this account occurs in the two versions at different places within the discourse. The narrative context is a teaching to a non-Buddhist wanderer on how to reach an entirely pleasant world, which the Buddha shows to be possible through absorption attainment.

The Madhyama-āgama version leads up to a description of the attainment of absorption with an account of the gradual path of training. [n.149] MĀ 208 at T I 785c24 abbreviates the account, a standard practice in the early discourses. The gradual-path account is usually spelled out in full only on its first occurrence in the collection, subsequent occurrences being abbreviated. This is a meaningful placement, since the account serves to indicate to the non-Buddhist wanderer in what manner progress to absorption attainment takes place according to early Buddhist thought.

In the Cūḷasakuludāyi-sutta the account of the gradual path instead comes after the Buddha has explained to the non-Buddhist wanderer that attainment of the third absorption is the way to reach an entirely pleasant world, and also after he has further made it clear that full realization of an entirely pleasant world requires attainment of the fourth absorption. After this clarification, the Buddha indicates to the non-Buddhist wanderer that Buddhist monks practise for the sake of states that are still superior and more refined. The wanderer then asks:

Venerable sir, what are the superior and more refined states for the sake of realizing which the monks practise the holy life under the Blessed One?[n.150] MN 79 at MN II 38,1: katame pana te, bhante, dhammā uttaritarā ca paṇītatarā ca yesaṃ sacchikiriyāhetu bhikkhū bhagavati brahmacariyaṃ carantī ti?

It is in reply to this question that in the Cūḷasakuludāyi-sutta the Buddha delivers the gradual-path account. When he comes to the part of the gradual-path account that describes the successive attainment of the absorptions, the Buddha indicates after each such attainment that this is superior and more refined. This serves to relate each of these aspects of the gradual-path account to the question posed by the non-Buddhist wanderer. In the case of the first absorption, the statement reads:

He dwells having attained the first absorption. Udāyin, this is also a superior and more refined state for the sake of realizing which monks practise the holy life under me.[n.151] MN 79 at MN II 38,8: paṭhamaṃ jhānaṃ (Ee: paṭhamajjhānaṃ) upasampajja viharati. ayam pi kho, udāyi, dhammo uttaritaro ca paṇītataro ca yassa sacchikiriyāhetu bhikkhū mayi brahmacariyaṃ caranti.

In this way, the first absorption becomes a state superior and more refined than the fourth absorption mentioned before. The same inconsistency continues with the other levels of absorption that form part of the gradual-path account. Since the first to third absorptions are obviously inferior and less refined than the fourth, it seems fair to conclude that some shifting in place of the gradualpath account has resulted in the presentation now found in the Cūḷasakuludāyi-sutta. The placement of the same account in its Madhyama-āgama parallel in contrast is clearly preferable.

MĀ 209: The Peak of Sensual Pleasure

Moving from the pleasure of absorption to sensual pleasure, in the Vekhanassa-sutta the Buddha makes a somewhat cryptic statement that appears to extol the highest type of sensual pleasure, which the commentary then takes to be a reference to Nirvāṇa.[n.152] MN 80 at MN II 43,4 and Ps III 277,19. In the Madhyama-āgama parallel a comparable statement involves merely an acknowledgement that the sense-objects of sensual pleasures appear to provide supreme happiness when contrasted to the sense-objects of undesirable sense-experiences.[n.153] MĀ 209 at T I 787a21; for a more detailed discussion of this and the next points cf. Anālayo 2007b and 2011b: 437–440.

Another puzzling aspect of the Vekhanassa-sutta is its report that the Buddha agreed with what appears to be an unjustified criticism raised by his visitor, whereas in the Madhyama-āgama parallel the Buddha does not agree with the criticism.[n.154] MN 80 at MN II 44,3 and MĀ 209 at T I 787b28. The Madhyama-āgama version also helps to understand better a simile made in both versions that illustrates bondage with the example of an infant, and it also provides a more natural conclusion to the discourse. Whereas in the Vekhanassa-sutta the wanderer Vekhanassa takes refuge as a lay disciple, in the Madhyama-āgama parallel he goes forth under the Buddha,[n.155] MN 80 at MN II 44,20 and MĀ 209 at T I 788a1. a considerably more natural choice for someone who has already been living as a wanderer. In sum, in the case of this discourse several aspects of the Pāli version can be corrected or improved by having recourse to its Madhyama-āgama parallel.

MĀ 214: Censure of the Buddha

Turning from going forth under the Buddha to his being censured by others, the Bāhitika-sutta and its Madhyama-āgama parallel report a meeting between King Pasenadi and Ānanda, in which the former tried to ascertain whether the Buddha would engage in the type of conduct that would be censured by other recluses and brahmins. The editions of the Pāli discourse differ on whether the king qualified these recluses and brahmins as wise, or whether this qualification was only introduced by Ānanda in his reply that the Buddha would not engage in such conduct; according to the Madhyama-āgama parallel this qualification was only introduced by Ānanda.[n.156] The king's enquiry in MN 88 at MN II 113,33 and in Se already qualifies the recluses and brahmins as wise, a qualification not found in his enquiry in Be and Ce and in the parallel MĀ 214 at T I 798a13; for a more detailed discussion cf. Anālayo 2007c and 2011b: 508.

Since in both versions the king expresses his appreciation for the finer distinction introduced in this way, which implicitly sets aside censure by those who are not wise and only takes up justified types of censure, the reading in the Madhyama-āgama version, and those Pāli editions that correspond to it, is clearly the preferable one.

Conclusion

The selected passages from the Madhyama-āgama surveyed above confirm in my view the methodological requirement mentioned at the beginning of this chapter, namely the need to proceed beyond the Pāli material when studying early Buddhist thought and history.[n.157] Regarding the hypothesis by von Hinüber 2008a, mentioned above p. 501f notes 2 and 3, as far as I am able to see for purposes of historical reconstruction the Pāli discourses in the four Nikāyas are better not used on their own, without taking into account their parallels preserved in the Chinese Āgamas, etc. (by "parallels" I mean early discourse parallels, not later texts such as, e.g., the Mahāyāna Mahāparinirvāṇa-sūtra, mentioned by von Hinüber 2015b: 198), as well as whatever other sources of information we can marshal for this purpose. An illustrative example is the Nandakovāda-sutta, whose relevance to the topic of the existence of nuns during the Buddha's lifetime I already mentioned in Anālayo 2008d: 117, and which I discussed in more detail in Anālayo 2010a. Comparative study shows that the more negative presentation of the nuns in the Pāli discourse is with high probability the result of later change. Thus when von Hinüber 2008a: 24 notes that in this Pāli discourse the Buddha "does not even talk to her … as if Mahāpajāpatī Gotamī would not exist", then I contend that this should be treated as historical information about the attitude of the reciters of the Pāli discourse, rather than as a source of historical information about the actual situation on the ground at the time of the Buddha. In sum, in my view it is indeed a "methodological problem" and a "methodological shortcoming" when one relies on the four Pāli Nikāyas on their own for purposes of historical reconstruction of what happened on the ground, contrary to the assessment by von Hinüber 2015b: 199 that "this methodological virtue is the only possible way to access the early Buddhist tradition." The cases surveyed range from very minor issues to at times significantly different presentations in the Madhyama-āgama discourses, which offer indications that help to improve our understanding of the relevant Pāli passage or discourse and thus must be taken into account in any attempt at historical reconstruction of early Buddhist thought.

The findings in the present chapter also complement the position I argued in the first chapter regarding the historical value of the 'early' discourses. In that chapter I concentrated in particular on the Pāli discourses, reflecting the fact that their value or lack of value for reconstructing the early stages in the history of Buddhist thought has received considerable attention by other scholars. Given that the examples in the present chapter show several instances where an Āgama collection preserved in Chinese translation appears to have preserved an earlier version of a particular passage when compared with the respective Pāli discourse, it seems reasonable to extend the conclusion drawn in the first chapter to the 'early' discourses in general. In other words, it seems to me safe to conclude that the Pāli discourses and their Āgama parallels (with the exception of the Ekottarika-āgama, where there is clear evidence that the collection underwent substantial change in China) can indeed be considered as textual testimonies to trends in the early stages in the development of Buddhist thought.

A proper appreciation of such thought requires a systematic reading of the early discourses, as opposed to focussing on isolated instances only. Such a systematic reading involves evaluating any particular passage against the background of relevant passages found elsewhere and based on a clear appreciation of the oral nature of these texts. Due to their prolonged oral transmission, the possibility of accidental errors has to be taken into account, the occurrence of which can at times be identified with the help of comparative study of the extant parallel versions.