Introduction

In the present appendix, I examine the question of the school affiliation of the Madhyama-āgama, based on a recent contribution to this topic by Chung (2011) and Fukita, followed by turning to the discussion in Bingenheimer (2011) of the expression 奉 行. Then I take up some of the conclusions drawn by Thich Minh Chau in his comparative study of the Majjhima-nikāya and the Madhyama-āgama. Thus the points taken up for examination are:

• The hypothesis that the Madhyama-āgama stems from a Sarvāstivāda line of transmission.
• The significance of the expression 奉行 found at the conclusion of Madhyama-āgama discourses.
• The question of the degree to which the roles of Mahākassapa and Sāriputta in the discourses collections of the Sarvāstivāda and Theravāda traditions reflect the influence of the respective schools.
• The hypothesis that descriptions of the Buddha taking a rest in front of the monks were deliberately omitted by the Theravādins.
• The significance of the absence of Chinese Āgama parallels to Pāli discourses, in particular the relation of the lack of a Chinese version of the Jīvaka-sutta to vegetarianism.
• The unusual conclusion of the Mūlapariyāya-sutta, which reports that the monks did not delight in the discourse, and the probable reasons why this conclusion is not found in the Madhyama-āgama version.
• The supposition that the Buddha's instruction to his son Rāhula in the Madhyama-āgama expresses decisively different perspectives on the significance of morality held by the Sarvāstivāda and the Theravāda traditions.
• The question of how far the usage of the terms thera in the Pāli Nikāyas and Chinese Āgamas should be understood as expressing sectarian concerns.

---

The School Affiliation of the Madhyama-āgama

The general consensus among scholars so far considers the Madhyama-āgama to have been transmitted by reciters of the Sarvāstivāda tradition.[n.1] Cf., e.g., Lü 1963: 242, Waldschmidt 1980: 136, Enomoto 1984, Mayeda 1985: 98, Enomoto 1986: 21, Minh Chau 1991: 27 and Oberlies 2003: 48. In a recent publication, Chung and Fukita provide several arguments that support this school identification, demonstrating also that the alternative of attributing the collection to the Dharmaguptaka tradition is rather unconvincing.[n.2] Chung 2011: 29–32.

Their conclusion, Chung (2011: 33), is nevertheless that "while we respect the prevailing opinions on this issue, we have considerable difficulty in supporting the view that the Zhong-ahanjing [Madhyama-āgama] derives from the Sarvāstivāda school, at least insofar as we define the term 'school' as a community of Buddhist monks who share a common canonical tradition".

The evidence they have surveyed suggests the existence of at least two distinct Āgama traditions within the Sarvāstivāda and/or Mūlasarvāstivāda traditions.[n.3] Chung 2011: 27f notes that it seems as if "the Zhong-ahanjing [Madhyamaāgama] and the Shisong-lü [Sarvāstivāda Vinaya] derive from a common tradition. Apart from the fact that there are certain incongruities with regard to some Vinaya rules, and that questions concerning the language of the underlying Indic original of the Zhong-ahanjing remain to be solved, it is possible that these two texts stem from two closely related branches of a Buddhist school. Assuming that the Zhong-ahanjing stems from a Sarvāstivāda branch, it would be logical to suppose that there were at least two Āgama traditions or bhāṇaka (reciter) lineages of the Madhyamāgama within the Sarvāstivāda school. These either differed from each other from the start or shared a common origin but went through different developmental processes ... the transmission of the Āgama texts within these respective traditions in different locations might have resulted in sub-traditions with regional variants in diction, or with variants conditioned by the times ... if that is indeed the case, then it would mean that within the Sarvāstivāda school there existed not only two or more Vinaya traditions, but also more than a single Āgama tradition". It seems indeed quite possible that there were not only different Vinayas within the Sarvāstivāda and/or Mūlasarvāstivāda traditions,[n.4] On the relationship between Sarvāstivāda and Mūlasarvāstivāda cf. Enomoto 2000, Skilling 2002 and Wynne 2008. but also distinct Āgama transmission lineages, although certainly exhibiting less pronounced differences between each other than what can be found between the respective Vinayas.[n.5] This would explain the observation by Chung 2011: 17 that "there are divergences between the sūtras in the Za-ahan-jing [Saṃyukta-āgama] and their counterparts in the Zhong-ajanjing [Madhyama-āgama], but these are not to a degree substantial enough to warrant an assumption that the Zhong-ahanjing derives from a Buddhist school which had no direct relation to the school from which the Za-ahanjing originates". The same holds for other observations mentioned by Chung 2011: 23 and 26, such as structural differences between the Chinese Madhyama-āgama and the Madhyama-āgama reflected in Śamathadeva's compendium of discourse quotations in the Abhidharmakośabhāṣya as well as the Mūlasarvāstivāda Vinaya, or variations in listings of the 32 marks of a great man in MĀ 59 as against such listings in Sanskrit fragments of the Mahāvadāna-sūtra and the Saṅghabhedavastu of the Mūlasarvāstivāda Vinaya; cf. also the table in Chung 2011: 185ff. All this fits the assumption that the Madhyama-āgama stems from a Sarvāstivāda reciter lineage, whereas the Saṃyukta-āgama preserved in Chinese, Śamathadeva's compendium and obviously the Mūlasarvāstivāda Vinaya stem from the Mūlasarvāstivāda tradition(s).

In view of the fluctuating nature of oral transmission that for centuries has had a determining influence on the early discourse collections, it would only be natural if followers of various branches of a particular school, quite probably separated from each other geographically, do not preserve the texts in precisely the same manner, with the same errors and variations, etc.

The research by Chung and Fukita also shows that, in spite of a considerable degree of general correspondence,[n.6] Cf., e.g., Chung 2011: 18, who regarding MĀ 37 at T I 478b13 to 479c10 and its counterpart in the Sarvāstivāda Vinaya, T 1435 at T XXIII 239b7 to 240a19, concludes that "these two versions are so similar as to suggest derivation from a common source". Again, comparing SĀ 592 at T II 157b18 to 158b23 and two equivalents in the Mūlasarvāstivāda Vinaya, Gnoli 1977: 166,16 to 171,1 and Gnoli 1978b: 14,13 to 18,13, with the second third of MĀ 28 at T I 459c9 to 460c1 and its counterpart in the Sarvāstivāda Vinaya, T 1435 at T XXIII 243c20 to 244b23, Chung 2011: 19 comments that "these last two texts [i.e., MĀ 28 and the Sarvāstivāda Vinaya] are more closely related to each other than to the first three texts [i.e., SĀ 592 and the two passages in the Mūlasarvāstivāda Vinaya]". Cf. also above page 307. occasionally passages in the Madhyama-āgama do not exactly match their counterparts in the Sarvāstivāda Vinaya.[n.7] Regarding a reference to a discourse in the Sarvāstivāda Vinaya, T 1435 at T XXIII 301a27, which de Jong 1988: 12 had identified as referring to MĀ 26, Chung 2011: 19 comments that "the cited passage cannot be located therein [i.e. in MĀ 26]". Now the instructions attributed to the Buddha in T 1435 at T XXIII 301a17 begin with a description of the proper conduct of a forest monk that clearly stands in relation to what the preceding tale had described, where a forest monk is unable to give proper replies to those visiting the place where he is staying. This is not a quote from MĀ 26, where the setting is different, as here a forest monk comes to visit other monks. These instructions are then followed by some additional indications at T XXIII 301a23, according to which a forest monk should also be knowledgeable in the discourses, the Vinaya and the Abhidharma, as well as in the absorptions and the stages of awakening. Following these indications the reference to the discourse comes in the form: "as described in detail in the Goṇiṣāda-sūtra", T 1435 at T XXIII 301a27: 如 瞿尼沙修多羅中廣說. Similar qualities are stipulated in MĀ 26 at T I 455c13, which describes that a forest monk should be knowledgeable in Vinaya and Abhidharma, in the immaterial attainments and in the destruction of the influxes. Thus, while the terminology employed does not correspond exactly (no explicit mention of the 'discourses', reference to absorptions as against immaterial attainments and to stages of awakening instead of destruction of influxes), the thematic similarity makes it clear that this need not be considered a mismatch. The point at stake does not appear to be an exact quotation from the discourse, but rather to indicate that the topics broached have received a more detailed treatment in MĀ 26. Thus this Vinaya passage could well have come into being in a reciter tradition whose Madhyama-āgama collection contained a discourse corresponding to MĀ 26. In addition to this case, another discourse quotation in T 1435 at T XXIII 410c4 refers to the 跋陀婆羅經. Chung 2011: 19 comments that this reference "is similar in name to Sūtra 194 of the Zhongahanjing, the Batuoheli-jing 跋陀和利經. In content, however, this sūtra can hardly be the source of the citation in the Shisong-lü". Thus perhaps this reference is not to MĀ 194, but to another discourse found elsewhere. This is perhaps not surprising, since given the nature of oral transmission it cannot be expected that the discourses and the Vinaya of any particular school match each other without the least variation.

An illustrative example in this respect can be found in the count of precepts for monks given in several discourses in the Aṅguttara-nikāya as "over 150 rules".[n.8] AN 3.83 at AN I 230,17: sādhikaṃ ... diyaḍḍhasikkhāpadasataṃ, an expression found again in AN 3.85 at A I 231,18, AN 3.86 at A I 232,33 and AN 3.87 at A I 234,11; cf. also, e.g., Dutt 1924/1996: 75f, Law 1933: 21, Bhagvat 1939: 64, Pachow 1955: 8f, Misra 1972: 33 and Dhirasekera 1982/2007: 145. This reference does not correspond too well to the number of rules for monks according to the Theravāda Vinaya tradition, which amount to 227 rules.[n.9] The Theravāda pātimokkha comprises 4 pārājika, 13 saṅghādisesa, 2 aniyata, 30 nissagiya pācittiya, 92 pācittiya, 4 pāṭidesanīya, 75 sekhiya and 7 adhikaraṇasamatha, resulting in a total count of 227. Such a difference would obviously not be sufficient ground for concluding that the Theravāda affiliation of the Aṅguttara-nikāya is doubtful. A more natural conclusion, following a suggestion given in the Pāli commentary, would be that at the time of the coming into being of these discourses not all of the rules recorded in the Vinaya had been promulgated.[n.10] Mp II 346,30: tasmiṃ samaye paññattasikkhāpadān' eva sandhāy' etaṃ vuttaṃ.

The above variation regarding the overall count of rules is not the only such instance found when comparing the Theravāda Vinaya to the Pāli discourses. Another example occurs in an examination of seven ways of settling litigations (adhikaraṇa samatha) in the Sāmagāma-sutta, which takes these up in a sequence that differs from the way they are listed in the Pāli Vinaya.[n.11] MN 104 at MN II 247,19 has decision by majority (yebhuyyasikā) as the second item in its exposition, whereas Vin IV 207,5 has decision by majority in fifth position. This difference seems to be an inconsistency within the discourse itself, as MN 104 at MN II 247,8 gives a listing that corresponds to the sequence in Vin IV 207,5; cf. also Anālayo 2011a: 607 note 112 and for a survey of such inconsistencies in the case of the Madhyama-āgama Chung 2011: 33 note 113.

Such divergences affect not only the count or sequence of rules and regulations. Even the Buddha's reflections right after having reached awakening – presumably a topic of considerable significance for his disciples – are reported differently in the Theravāda Vinaya and in the Udāna.[n.12] Ud 1.1–3 at Ud 1,6 report that the recently awakened Buddha contemplated dependent arising during the first watch of the night, dependent cessation during the second watch of the night, and both during the last watch of the night. According to Vin I 1,5 he contemplated both during each of the three watches. Judging from the stanzas that accompany the prose in Ud and Vin, especially the reference to khaya in the second stanza, the Ud account is the more original; cf. also Anālayo 2008b: 379.

Yet another example can be found in relation to the Raṭṭhapāla-sutta. The Theravāda Vinaya reports an exchange in verse between Raṭṭhapāla and his father that clearly belongs to an episode reported in the Raṭṭhapāla-sutta, when Raṭṭhapāla visits his home town after going forth and does not receive any alms at his family's house.[n.13] Vin III 148,30. Raṭṭhapāla's visit to his former home is reported in MN 82 at MN II 61,26; cf. also Anālayo 2011a: 457f. The Pāli discourse has none of these stanzas.

These few instances may already suffice to show that some degree of variation is natural in the case of the type of oral material that we are dealing with. Here it also needs to be kept in mind that the Pāli discourses are not a product of the Theravādins, just as the Madhyama-āgama is not a product of the Sarvāstivādins. Instead, these are collections of discourses transmitted orally for long periods by successive generations of Indian reciters, of which only the last generations could be considered members of a particular school.

In sum, it is only natural that some minor divergences between the texts transmitted within a particular school occur. The nature of the transmission of the early texts therefore makes it unavoidable that school affiliations cannot be settled beyond any doubt. This in turn implies that, in a discussion of the question of school affiliation, an examination of textual variations needs to be combined with an evaluation of the impact of doctrinal views on the transmitted material.[n.14] For an example cf. above p. 437f. Yet, Chung and Fukita do not seem to have taken doctrinal influences into consideration at all.

While the survey by Chung and Fukita does offer a welcome reminder that the Sarvāstivāda affiliation of the Madhyama-āgama is just a hypothesis, as far as I can see it remains clearly the most probable and convincing hypothesis.

---

The Conclusion of Madhyama-āgama Discourses

A character couplet found regularly at the conclusion of discourses in the Madhyama-āgama and other Āgamas is 奉行, where it occurs together with a reference to the listeners delighting in what has been said.[n.15] MĀ 145 at T I 654b27 confirms that the object of such delighting and 奉行 is what has just been said, 我等一切歡喜奉行彼比丘所說. MĀ 131 at T I 622c12 indicates that 奉行 and delighting should be considered as separate activities, as in the context of a five character stanza they are divided by "and": 歡喜而 奉行. The corresponding concluding formula in Pāli discourses only speaks of the listeners delighting (abhinandati, at times together with anumodati, which conveys the same meaning). This raises the question if there is an Indic counterpart to 奉行.

In a recent study of 奉行, Bingenheimer (2011: 53) suggest the expression renders an Indic original √dhṛ.[n.16] Bingenheimer 2011: 53 bases his suggestion on a passage in the Āyuḥparyanta-sūtra, Matsumura 1989: 82,27 (§36): śrutvā bhikṣavo dhārayiṣyanti, on which he comments that "for this passage, which is not the closing formula, we have a parallel in both the ZA [Saṃyukta-āgama] and the BZA ... in the ZA dhārayiṣyanti is rendered fengxing". While the expression 奉行 does not occur in the Chinese parallel to the above passage from the Āyuḥparyanta-sūtra, T 759 at T XVII 603a26, which reads: 我等樂聞, 受持憶念, it occurs at the end of the text, T 759 at T XVII 604a21, which unfortunately has no counterpart in the Sanskrit fragment. Counterparts to the above type of expression in the Saṃyukta-āgama (= ZA) show some variations (such as 當受持奉行 and (當)如說奉行), but in those instances where I have been able to identify a Sanskrit or Pāli parallel, the expression corresponding to śrutvā bhikṣavo dhārayiṣyanti or sutvā bhikkhū dhāressantī ti is 諸比丘聞已, 當受奉行; cf. SĀ 75 at T II 19c2 and SN 22.58 at SN III 66,9, SĀ 282 at T II 78b8 and MN 152 at MN III 299,2, SĀ 291 at T II 82b8 and SN 12.66 at SN II 107,22, SĀ 294 at T II 84a6 and SN 12.19 at SN II 24,16, SĀ 684 at T II 186c5 and a Sanskrit fragment parallel, together with the Pāli counterpart, in Chung 2009: 9, SĀ 813 at T II 208c20 and SN 54.10 at SN V 323,3, SĀ 1136 at T II 299c16 and 300a6 and SN 16.3 at SN II 198,16 and 199,7. In these cases, dhāressanti is already rendered by 當受, thus 奉行 appears to be without a counterpart. Bingenheimer (2011: 56) then concludes that "the intention of the Indian original was to express the semantics of 'retain/remember/uphold' rather than 'practice/fulfill/act upon'".

Since in the discourses translated above I rather rendered 奉行 as "receive respectfully", in what follows I briefly explain my choice. Given that the implications of this character couplet may well differ with various translators, I will focus on its usage in the Madhyama-āgama. Now in the only case where I have been able to locate an Indic counterpart to 奉行 in the Madhyama-āgama, the parallel reads samādāya vartati / vattati, conveying the sense of "undertaking and practising".[n.17] This occurs in the context of describing the Vajjians well undertaking and practising the Dharma set forth by their ancestors, MĀ 142 at T I 648c17: 舊 跋耆法善奉行, which has as its parallel DN 16 at DN II 74,13: vajjidhamme samādāya vattantī ti and Sanskrit fragment M 372 V5, Waldschmidt 1950: 66: [vṛ]jidharmaṃ samā[da](ya vartante); cf. also the Tibetan parallel in Waldschmidt 1951: 111. The same principle is then applied to the monks who should undertake and practice the precepts set forth to them by the Buddha, MĀ 142 at T I 649b18: 戒善奉行者, DN 16 at DN II 77,5: sikkhāpadesu samādāya vattissanti, and fragment S 380 (= no. 254) V1, Waldschmidt 1950: 84: (si)k- ṣāṃ samādāy(a vartiṣyante). Looking at 奉行 from the viewpoint of the meaning suggested by the context, the sense of undertaking something or practising seems indeed relevant for a few other occurrences in the Madhyama-āgama, where the question at stake is putting into practice a meditative instruction received earlier.[n.18] MĀ 8 at T I 429b15, MĀ 130 at T I 619c4 and MĀ 160 at T I 684a14: 奉行法 者. On this expression, Bingenheimer 2011: 55 comments that "the passage, which appears six times in T.26, is always followed by 修四梵室 where 'practice' is clearly expressed by xiu, 修, and the meaning 'remember' for fengxing therefore seems preferable". It seems to me that this is not invariably the case, cf., e.g., MĀ 8 at T I 429b15: 諸弟子等有不具足奉行法者, 彼命終已, 或生 四王天, 或生三十三天, where the passage in question is not followed by another term that stands for "practice", so that the meaning of 奉行 in such contexts does indeed seem to be "to practice".

However, the sense of undertaking or practising does not fit contexts in the same collection where 奉行 occurs together with delighting, as is the case for the standard ending of a discourse. An example would be the recurrent use of 奉行 as part of a discussion where, on receiving a reply to a query, the questioner reacts by delighting and 奉行.[n.19] Besides MĀ 187 quoted in the note below, cf., e.g., MĀ 29 at T I 461c11, MĀ 134 at T I 635a22, MĀ 172 at T I 709a28, MĀ 201 at T I 788a25 and MĀ 211 at T I 790b22. There is little scope for the questioner to put things into practice while being in the midst of a discussion.

At first sight one might think that perhaps here the sense of remembering falls into place, since in an oral setting this is precisely what is required from the questioner, namely that he keeps in mind what he has just heard. A closer inspection, however, shows that this does not always work. A case in point is the parallel to the Chabbisodhana-sutta (MN 112) translated above.

The discourse indicates that, when a monk claims to be an arahant, the other monks should first of all approve of it, rejoice in it and undertake 奉行, followed by then investigating this claim.[n.20] MĀ 187 at T I 732a25: 汝等聞之, 當善然可, 歡喜奉行; cf. above page 225. In the case of the first instance of 奉行 in relation to the bare fact that the monk in question claims to be an arahant, there would seem to be little need to make any special effort to remember this. The point of the description thus seems to be that the other monks should first of all react in an appreciative and respectful manner, before undertaking a closer investigation. The same sense of appreciation and respectful reception would also fit the standard conclusion of a discourse.

In its use in the Madhyama-āgama, then, when 奉行 occurs together with delighting I adopt as a provisional rendering "receive respectfully",[n.21] According to Hirakawa 1997: 338, 奉行 can render pratipatti, for which Monier-Williams 1899/1999: 667 gives as a possible meaning "respectful reception". which as far as I can see seems to work for most occurrences of 奉行. Depending on context, such respectful reception can be understood to imply at least to some degree also remembering what one has heard and/or putting it into practice as the natural consequence of a respectful reception of an instruction. I realize that this is not ideal, but for the time being hopefully this will work until further research has clarified the meaning of 奉行.

---

The Role of Mahākassapa and Sāriputta

In his groundbreaking comparative study, Minh Chau (1991) examines various aspects of the Madhyama-āgama that, in his view, make it probable that this collection was transmitted by the Sarvāstivāda tradition. As mentioned above, I think his identification is probably correct. However, not all of the arguments that he proposes serve the purpose of establishing this conclusion.

Having highlighted Mahākassapa's status as "the first and foremost ācariya of the Sarvāstivāda sect, while Sāriputta is considered to be the first ācariya of the Theravāda sect", Minh Chau (1991: 21) notes that the Anupada-sutta, which praises Sāriputta as the disciple who is born from the Blessed One's mouth and who keeps rolling the wheel of Dharma, does not have a Chinese parallel. He then concludes that "as the Buddha ... extolled elder Sāriputta so highly, the Sarvāstivādins might drop this sutta from their Canon or it might be interpolated by the Theravādins in their Canon".

Yet, another Madhyama-āgama discourse also states that Sāriputta kept rolling the wheel of Dharma.[n.22] MĀ 121 at T I 610b9: 舍梨子, 我所轉法輪, 汝復能轉. Even the long praise of Sāriputta's wisdom, found at the beginning of the Anupada-sutta, recurs in similar words in several Madhyama-āgama discourses.[n.23] MN 111 at MN III 25,5; cf. MĀ 27 at T I 458b15, MĀ 28 at T I 461b10, MĀ 31 at T I 467b10 and MĀ 121 at T I 610b5. Thus the Madhyama-āgama does not seem to be consistent in underrating Sāriputta's status.

Moreover, a discourse in the Saṃyutta-nikāya reckons Mahākassapa as having been born from the Blessed One's mouth, employing exactly the same terms used to describe Sāriputta in the Anupada-sutta.[n.24] SN 16.11 at SN II 221,22: bhagavato putto oraso mukhato jāto dhammajo dhammanimmito dhammadāyādo, which corresponds word for word to the praises of Sāriputta in MN 111 at MN III 29,11. Thus this Pāli discourse shows quite a positive attitude towards Mahākassapa. Conversely, other Pāli discourses even present a critical perspective in regard to Sāriputta. One example is the Cātumā-sutta of the Majjhima-nikāya, which records the Buddha explicitly voicing criticism of Sāriputta.[n.25] MN 67 at MN I 459,19; a criticism also reported in its parallel EĀ 45.2 at T II 771b6.

A particularly telling case is the Dhānañjāni-sutta and its Madhyama-āgama parallel. Both discourses report that Sāriputta had given a teaching to the dying brahmin Dhānañjāni due to which the latter was reborn in a Brahmā world. The two versions differ, however, in their evaluation of Sāriputta's teaching. According to the Madhyama-āgama version, the Buddha praised Sāriputta's wisdom and explained to the assembled monks that Sāriputta had successfully taught Dhānañjāni the path to the Brahmā world, adding that if he had continued to teach the brahmin further, Dhānañjāni would have understood the Dharma.[n.26] MĀ 27 at T I 458b17: 舍梨子比丘成就實慧, 此舍梨子比丘教化梵志陀然, 為說梵天法來, 若復上化者, 速知法如法.

The Majjhima-nikāya version, however, gives a rather different evaluation. It points out that Sāriputta had left after establishing Dhānañjāni in the "inferior Brahmā world", even though "there still was more to be done by him".[n.27] MN 97 at MN II 195,20: sati uttarikaraṇīye, hīne brahmaloke patiṭṭhāpetvā (Ce: uttariṃ karaṇīye), a description repeated at MN II 195,25 and MN II 196,1. This criticism is voiced three times in the Pāli discourse. At first the reciters themselves express this criticism as part of their narration of what happened.[n.28] For the reciters to include this criticism in their narrative description is noteworthy, since in general the narrative voice in the discourses tends just to relate circumstances without explicitly evaluating them. Then the Buddha conveys the same disapproval to the assembled monks when he sees Sāriputta approaching them. Once Sāriputta has arrived, the Buddha directly asks him why he had left after establishing Dhānañjāni only in the inferior Brahmā world, even though there was still more for him to do. Thus the present example, taken on its own, would give the impression that the Theravāda tradition had a less positive attitude towards Sāriputta than the Sarvāstivāda reciters.

The above variations reveal the lack of uniformity in the way the discourse collections of the Sarvāstivāda and Theravāda schools present Sāriputta and Mahākassapa. Thus the role assumed by these two disciples in various discourses need not be an expression of the influence of the school that transmitted the respective discourse. That is, the esteem in which Mahākassapa and Sāriputta were respectively held by the Sarvāstivāda and Theravāda schools may not have had such far-reaching consequences as to lead to the wholesale omission of discourses.

---

The Buddha's Deportment in the Dhammadāyāda-sutta

Thich Minh Chau points out that in the Madhyama-āgama parallel to the Dhammadāyāda-sutta the Buddha takes a rest by lying down in front of the monks, whereas the same is not recorded in the Majjhima-nikāya version.[n.29] MĀ 88 at T I 570b22 and MN 3 at MN I 13,34. He also notes that a similar description is found in another Madhyama-āgama discourse, which does not have a Pāli counterpart. From this he infers that "that the Pāli compilers were not happy over these attitudes of the Buddha and dropped" the corresponding passage in the former case and the whole discourse in the latter case (Minh Chau 1991: 30). In contrast, for the Sarvāstivādins "the dignified way which the Buddha adopted in lying down in front of the monks had nothing questionable" which "justified the Sarvāstivādins in preserving these details".

The Dhammadāyāda-sutta has another parallel in the Ekottarika-āgama, which also does not describe the Buddha lying down. Instead, it agrees with the Majjhima-nikāya version that the Buddha just retired to his dwelling.[n.30] EĀ 18.3 at T II 588a25. Moreover, several Pāli discourses do describe the Buddha lying down to take a rest in front of the monks.[n.31] DN 33 at DN III 209,18, MN 53 at MN I 354,25, SN 35.202 at SN IV 184,8 and AN 10.67 at AN V 123,1. Minh Chau 1991: 30 seems to have been aware of such instances, as he notes that "the Theravādins referred to these details very rarely", thus he did not base his argument on assuming a total absence of such descriptions in Pāli discourses. This makes it improbable that the absence of such a description in the Dhammadāyāda-sutta and the lack of a Pāli parallel to the other Madhyama-āgama discourse could be an expression of a wish of the compilers of the Pāli canon to erase such descriptions.

---

Vegetarianism and the Jīvaka-sutta

The hypothesis that certain discourses were dropped under the pressure of sectarian influence comes up again in relation to the Jīvaka-sutta. Minh Chau (1991: 31f) reasons that "the dropping from all the Chinese Āgamas of the Pāli sutta N 55, Jīvakasutta, in which the Buddha was reported to allow the monks to take three kinds of meat, confirms the Sarvāstivāda's attitude against meat-eating". "This difference in attitude towards meat-eating in the two versions serves to explain why now-a-days meat-eating is allowed for the monks of countries following the Theravāda tradition, while the monks of such countries as China, Korea, Vietnam observe vegetarianism faithfully".

Yet, to take the absence of a parallel to the Jīvaka-sutta to be an expression of Sarvāstivāda vegetarianism would not fit too well with the Sarvāstivāda Vinaya, which mentions the three instances when meat should not be consumed in the same way as the Jīvaka-sutta; both indicate that, apart from the stipulated conditions, meat-eating is allowable.[n.32] T 1435 at T XXIII 190b9 and T XXIII 264c27, corresponding to the description in MN 55 at MN I 369,4. Moreover, T 1435 at T XXIII 91b21 includes meat and fish in a listing of allowable food, and T 1435 at T XXIII 190b14 agrees with Vin I 238,8 on reckoning meat consumption apart from the prohibited three instances as pure, 三種淨, tikoṭiparisuddha; the same is the case for the Mūlasarvāstivāda Vinaya; cf. T 1458 at T XXIV 570a15 (also in Dutt 1984: 236,17); cf. also Waldschmidt 1939/1967: 104f, Prasad 1979, Schmithausen 2005b: 188f and Heirman 2006: 60. Had vegetarianism in the Sarvāstivāda tradition been strong enough to lead to a suppression of the Jīvaka-sutta, one would expect it to have exerted its influence also on the Vinaya.

In fact, the early texts of the different Buddhist schools seem to be in general agreement that meat-eating is allowable. The only recorded instance of vegetarianism being advocated appears to come from the schismatic Devadatta. [n.33] For a survey of the points raised by Devadatta according to the different Vinayas cf., e.g., Bareau 1991: 108 and Mukherjee 1966: 76f. According to Ruegg (1980: 236f), it seems probable that "vegetarianism became established in Buddhism ... in close connection with ... the tathāgatagarbha doctrine", which would make it unlikely that the idea of vegetarianism could have exerted a determining influence on the compilation of the Āgama collections.

Independent of the exact time at which vegetarianism became a tenet of general importance in the history of Buddhism, the absence of a parallel to the Jīvaka-sutta in the Chinese Āgamas would not be related to such issues, as among Sanskrit fragments of the Sarvāstivāda/Mūlasarvāstivāda Dīrgha-āgama a counterpart to the Jīvaka-sutta has been found.[n.34] Cf. the survey in Hartmann 2004: 127 and the uddāna in Hartmann 2002: 138; cf. also fragment SHT VI 1525 V1 to R2, which according to the identification in Bechert 2004: 439 belongs to a version of the Jīvaka-sutta. These fragments have preserved the passage on the conditions for the consumption of meat that Thich Minh Chau assumes to be the reason for the absence of a version of this discourse in the Sarvāstivāda tradition.[n.35] DĀ (Skt) fragment 433v (I am indebted to Jens-Uwe Hartmann for kindly providing me with a preliminary draft transliteration of this fragment).

Thus the absence of a parallel to the Jīvaka-sutta among the discourses found in the Chinese Āgamas does not seem be related to ideological issues, but rather appears to be an outcome of the circumstance that the four Āgamas belong to different schools. The Madhyama-āgama collection preserved in Chinese does not have a parallel to the Jīvaka-sutta because the Sarvāstivādins would have had a version of this discourse in their Dīrgha-āgama collection, a collection not preserved in Chinese. The Dīrghaāgama preserved in Chinese translation is generally held to belong to the Dharmaguptaka tradition.[n.36] Cf. above note 62 page 35. This collection does not include a version of the Jīvaka-sutta, possibly because this discourse was found in another Āgama of the Dharmaguptaka tradition, perhaps in its Madhyama-āgama, similar to the placing given to this discourse in the Theravāda tradition.

The Jīvaka-sutta is not the only such instance of Pāli discourses that are without a counterpart in the Chinese Āgamas. Even in the same chapter of the Majjhima-nikāya, the Gahapati-vagga, several more such cases can be found.

One such case is the Apaṇṇaka-sutta, parts of which have similarly been preserved in Sanskrit fragments.[n.37] For a survey of the fragment parallels to MN 60 cf. Anālayo 2011a: 339 note 147. The same is the case for the Kandaraka-sutta, MN 51, though here it is uncertain if the relevant Sanskrit fragments are parallels to MN 51 or to one of the other Pāli discourses that expound the same theme; cf. the survey of fragment parallels in Anālayo 2011a: 309 note 2. Another example is the Kukkuravatika-sutta, which has no known parallel either in the Chinese Āgamas or in Sanskrit fragments. Nevertheless, an exposition that seems to stem from a version of this discourse can be found in Śamathadeva's compendium of discourse quotations in the Abhidharmakośabhāṣya, extant in Tibetan.[n.38] D 4094 ju 268b2 or Q 5595 tu 235a3, translated in Skilling 1979.

Moreover, the Saṅgītiparyāya's comment on the four types of action treated in the Kukkuravatika-sutta explicitly refers to the protagonist of this discourse, Puṇṇa, who was observing the ascetic practice of behaving like a cow.[n.39] T 1536 at T XXVI 396a8 notes that the Buddha gave this exposition of the four types of action to 圓滿, the observer of the cow conduct, 牛戒, corresponding to Puṇṇo, the govatiko, mentioned in MN 57 at MN I 387,9. There can be little doubt that this passage has a version of the Kukkuravatika-sutta in mind.

A last example, still from the same Gahapati-vagga of the Majjhima-nikāya, is the Abhayarājakumāra-sutta, which has no known parallel either in the Chinese Āgamas or in Śamathadeva's compendium. Nevertheless, parts of this discourse have been preserved in a discourse quotation in the *Mahāprajñāpāramitā- (upadeśa-)śāstra.[n.40] T 1509 at T XXV 321b15 to b25. On an as yet unpublished Sanskrit fragment parallel cf. Hartmann 1992: 28.

Thus the Gahapati-vagga of the Majjhima-nikāya has a number of discourses that are without counterpart in the Chinese Āgamas, yet Sanskrit fragments or discourse quotations preserved in Chinese or Tibetan sources document the existence of parallel versions. This goes to show that the absence of a discourse from the Āgamas may simply be due to differences in the distribution of discourses among the four discourse collections transmitted by various Buddhist schools. Since the four Āgamas translated into Chinese do not stem from a single school, such variations in distribution have affected the range of discourses preserved in Chinese translation.

---

The Conclusion of the Mūlapariyāya-sutta

In relation to the unusual concluding section of the Mūlapariyāya-sutta, according to which the monks did not delight in the Buddha's exposition, Minh Chau (1991: 204) notes that the discourses identified by Akanuma (1929/1990: 163) as the parallel versions – a discourse in the Madhyama-āgama and an individual translation – instead employ the standard conclusion to a discourse, namely, that the monks were delighted with the Buddha's exposition.[n.41] MĀ 106 at T I 596c14: 彼諸比丘聞佛所說, 歡喜奉行 and T 56 at T I 851b21: 彼諸比丘聞世尊所說, 歡喜而樂. Thich Minh Chau then reasons that the negation na "might have been used to earmark its [the discourse's] expunging from the Pāli Tipiṭaka, but the later Pāli compilers forgot to do so". Yet, that the monks did not delight in the discourse is also recorded in an Ekottarika-āgama parallel to the Mūlapariyāya-sutta.[n.42] EĀ 44.6 at T II 766b15: "at that time, the monks did not accept that teaching [given to them by the Buddha]", 是時諸比丘不受其教.

This rather unusual ending was also known to the Pāli commentators, who provide an explanation for it. They report that the Buddha had preached this discourse to humble the pride of a group of five hundred monks.[n.43] Ps I 56,8; cf. also Jā 245 at Jā II 259,14. The commentary continues by indicating that on a later occasion the same group of monks received a short discourse by the Buddha, at the conclusion of which all of them became arahants; cf. AN 3.123 at AN I 276,23. EĀ 44.6 at T II 766b18 continues with an additional instruction to the monks on practising meditation in seclusion, as a result of which the monks do delight in what the Buddha has told them. The monks did not delight in the discourse, the commentary explains, because they were unable to understand what the Buddha had taught them. On this explanation, their unusual reaction could indeed have been part of the discourse right from its outset.[n.44] According to an alternative interpretation suggested by Bodhi 1980/1992: 20, Thanissaro 2002: 156 and Ñāṇananda 2005: 286, the monks might also have been unable to delight in this penetrative discourse because they did understand its thorough undermining of various concepts to which they were still attached.

The Pāli editions of the Mūlapariyāya-sutta do in fact vary, as according to the PTS edition the monks did delight in the exposition given by the Buddha on this occasion.[n.45] While the Ee edition at MN I 6,24 reports that the monks delighted, according to Be, Ce and Se they did not delight in the exposition given by the Buddha. This difference between texts within the Theravāda tradition suggests an explanation that might also apply to the parallels to the Mūlapariyāya-sutta found in the Madhyama-āgama and in an individual translation, in that the natural levelling tendency of oral transmission or else the influence of editors or translators led to a 'correction' of the concluding section in accordance with the standard phrase employed at the end of all other discourses: the monks delighted in what the Buddha said.

In other words, the lack of delight of the monks as the more unusual reading, lectio difficilior, is in this case likely to be the more original version and there seems to be no need to assume that it expresses a wish to erase the discourse from the Majjhima-nikāya collection.

---

The Instruction to Rāhula

In relation to the Ambalaṭṭhikārāhulovāda-sutta, Minh Chau (1991: 34) notes that the instructions given by the Buddha to his son Rāhula in the Chinese version differ from its Pāli counterpart. The Madhyama-āgama instructions refer to a bodily action that is "pure" but at the same time is unwholesome and results in dukkha, while a bodily action that is "not pure" is wholesome and does not result in dukkha.[n.46] MĀ 14 at T I 436c11: 彼身業淨 ... 不善與苦果受於苦報, and 436c14: 彼 身業不淨 ... 善與樂果受於樂報. The corresponding Majjhima-nikāya passage does not envisage that a bodily action could be simultaneously pure and unwholesome, but simply describes an unwholesome bodily deed as one that results in affliction.[n.47] MN 61 at MN I 415,29: akusalaṃ idaṃ kāyakammaṃ dukkhudrayaṃ dukkhavipākaṃ.

In an attempt to make sense of the Madhyama-āgama passage, Thich Minh Chau interprets the expression "pure" (淨) to mean "permissible" and comments that "by the term permissible, the Chinese means that it is approved by the Scriptures or it does not infringe the monastic rules".[n.48] In fact Hirakawa 1997: 727 lists kalpika and kalpa as possible equivalents for 淨; cf. also Bapat 1970: 223, who notes that the expression akappiya in Sp II 289,24 has as its counterpart 不淨 in T 1462 at T XXIV 727c23. He then concludes that "in the Sarvāstivāda definition there is wide scope for a liberal attitude towards the interpretation of the monastic regulations. We might find ... in this difference an explanation of the rigid orthodoxy of the Theravāda monks who up to now-a-days would frown on any attempt to relax or to interpret differently the established rules, while the Sarvāstivāda and its offshoot schools adopted a more or less liberal attitude toward the observance of monastic rules".

A problem with applying this interpretation, however, is that, according to the instruction for a past bodily action in the Madhyama-āgama discourse, a "permissible" (淨) bodily deed should be confessed, while a "not permissible" (不淨) bodily deed leads to the arising of joy.[n.49] MĀ 14 at T I 436c27. This makes little sense, since for a "permissible" bodily deed there would be no need for confession. In fact, the idea to give precedence to other's welfare over the requirements of moral conduct belongs to a later phase of Buddhist thought and does not seem to be reflected in the early discourses.

In addition to the Madhyama-āgama and Majjhima-nikāya versions, the instructions to Rāhula can be found in several other texts: in the Mūlasarvāstivada Vinayavibhaṅga, preserved in Chinese and Tibetan,[n.50] T 1442 at T XXIII 761a11: 是不善事, 是苦惡業, 能於未來感苦異熟 and D 3 cha 217a5 or Q 1032 je 201a6: bdag dang gzhan la gnod par ʼgyur ba mi dge ba (D: baʼi) sdug bsngal ʼbyung ba rnam par smin pa sdug bsngal ba yin. in a discourse quotation in the Vyākhyāyukti- ṭīkā, preserved in Tibetan,[n.51] D 4069 si 200b3 or Q 5570 i 71a5: gnod pa dang ldan pa mi dge ba sdug bsngal ʼbyung ba dang. and in a paraphrase in the Śrāvakabhūmi.[n.52] T 1579 at T XXX 405b5: 自損及以損他是不善, Shukla 1973: 55,16: vyābhādhikaṃ ... ātmano vā parasya vā akuśalaṃ. All of these parallels agree with the Majjhima-nikāya discourse that the instructions are to refrain from a bodily deed that is harmful and unwholesome, without envisaging that such a deed could be reckoned as pure.

Thus the most straightforward explanation of this part of the Madhyama-āgama version would be to assume the occurrence of a translation error. Such an error could have happened due to misinterpreting a sandhi in the Indic original, assuming that a particular word has, or else does not have, the negative prefix a-.

This type of error does, in fact, occur in another Madhyamaāgama discourse, which describes the behaviour of a monk who is investigated for some misdeed. According to the Madhyamaāgama description, this monk might either display anger and say that he wants to act in accordance with the wishes of the community that investigates him, or else he might not display anger and say that he does not want to act in accordance with the wishes of the community.[n.53] MĀ 194 at T I 748b28 and c18: 瞋恚憎嫉, 發怒廣惡 ... 作如是說, 我今 當作令眾歡喜而可意, and 不瞋恚憎嫉, 發怒廣惡 ... 不如是說, 我今當作 令眾歡喜而可意. According to the Majjhima-nikāya parallel, however, the monk who displays anger is also unwilling to act in accordance with the wishes of the community, while the one who does not display anger wants to act in accordance with the wishes of the community.[n.54] MN 65 at MN I 442,31 and 443,10.

This is clearly the more logical presentation and in this case, too, the mistaking of a sandhi by the translator(s) of the Madhyama-āgama appears to be the most probable reason for the wording in the Chinese discourse.[n.55] Karashima 1992: 263 notes the occurrence of this type of error in Dharmarakṣa's translation of the Saddharmapuṇḍarīka-sūtra; Bapat 1970: lix points out occurrences of this type of error in the Chinese counterpart to the Samantapāsādikā, T 1462. This shows that such mistakes were not uncommon. Thus, rather than reflecting decisively different perspectives on the significance of morality held by the Sarvāstivāda and the Theravāda traditions, the present Madhyama-āgama passage may simply be the result of a translation error.

---

The Term Thera

Minh Chau (1991: 33) quotes Anesaki (1901: 897) to the effect that the term thera "in the Chinese Āgamas ... occurs, as far as I know, only thrice – the title Āyushman being used otherwise. Can we conclude that the Chinese version is derived from traditions dating from an age in the history of Buddhism when the authority of Theras was not yet solidified, or it descended from a school antagonistic to the authority of the orthodox Theras?" This suggestion by Anesaki seems to be incorrect, as a counterpart to thera can be found in numerous Āgama discourses, which use the expression 長老, literally "elder", often additionally qualified as 上尊, "senior".[n.56] Occurrences in the Madhyama-āgama alone are: MĀ 22 at T I 450a21, MĀ 26 at T I 455c7, MĀ 33 at T I 471c29, MĀ 38 at T I 481a9, MĀ 39 at T I 481b16, MĀ 82 at T I 557c25, MĀ 83 at T I 559b27 (in the discourse's title), MĀ 84 at T I 560b25, MĀ 88 at T I 569c26, MĀ 116 at T I 606c25, MĀ 142 at T I 649b20, and MĀ 196 at T I 754b16; cf. in more detail Deeg 2012: 130ff. Thus there seems to be no reason to assume that references to thera express sectarian concerns.

---

Conclusion

In sum, it seems to me reasonable to assume that the Madhyama-āgama was transmitted by Sarvāstivāda reciters; and for the conclusion to Madhyama-āgama discourses I provisionally adopt the translation "receive respectfully".

The above surveyed differences between the Madhyama-āgama and the Majjhima-nikāya in regard to the depiction of chief disciples or the Buddha's deportment do not appear to reflect sectarian influences. Sectarian concerns need also not be responsible for the absence of a parallel to the Jīvaka-sutta in the Madhyama-āgama, for the way the reaction of the listening monks is reported in the Mūlapariyāya-sutta, or for how the Madhyama-āgama records the instructions given by the Buddha to his son Rāhula.

When evaluating the above surveyed suggestions by Thich Minh Chau, I think it needs to be kept in mind that, at the time when he was undertaking his research, the study of the nature and dynamics of oral tradition was still in its beginnings. Nowadays, awareness of the impact of orality on the actual shape of the discourses found in the Pāli Nikāyas and the Chinese Āgamas is constantly growing and it is becoming increasingly evident that a good number of variations are best understood in the light of the dynamics and vicissitudes of oral transmission.[n.57] On these dynamics cf. Anālayo 2011a: 14–22 and 855–891.

Thus when Minh Chau (1991: 75) contrasts an "unmistakenly common source of the Chinese and the Pāli versions" to "differences in the details ... due either to the compilers' choice or to the characteristics of the schools they represented", it becomes clear that for him variations are necessarily the outcome of conscious manipulation by editors and compilers, influenced by the ideas and dogmas of the particular Buddhist school to which they belong. Therefore the oral nature of the material with which he was working was naturally not at the forefront of his attention.

Moreover, the digitalisation of the canonical texts facilitates checking particular passages or expressions to a considerable degree. With the computer technology nowadays at our disposal, it is a matter of seconds to test certain hypotheses that in earlier days could only be evaluated by extensive and time-consuming research.

Thus my criticism above mainly reflects the shift of perspective that has resulted from the progress of Buddhist studies and the advances of computer technology made during the more than five decades that have gone by since Thich Minh Chau originally did his research (completed in 1961). This does not affect the value of his work as a lasting inspiration for those of us who are engaged in comparative studies between discourses from the Pāli Nikāyas and the Chinese Āgamas, with their considerable potential to reveal the common core and the traces of change that affected the legacy of early Buddhist literature.