|BuddhaSasana Home Page||English Section|
Reflections on Comparative Āgama Studies
Center for Buddhist Studies, University of Hamburg, Germany
The present paper offers a few methodological reflections on comparative studies between the discourses found in the Chinese Āgamas and their parallels in Pāli, Sanskrit and Tibetan. The issues taken up are: the impact of oral transmission on this material; the notion of a parallel and difficulties in applying this notion; the advantage of approaching the category of a parallel with the help of the Buddhist four-fold logic; and the potential of comparative studies.
A clear appreciation of the characteristics and dynamics of early Buddhist oral transmission is a prerequisite for a comparative study of the legacy of discourses preserved in the Āgamas and Nikāyas.2 These discourses in fact quite explicitly introduce themselves as products of oral transmission, as they invariably employ the standard beginning: "Thus have I heard" (evaṃ me sutaṃ 如是我聞). The oral transmission of the discourses, reflected in this standard formulation, may well be as old as Buddhism itself. According to the different Vinayas, soon after his awakening the Buddha sent his first monk disciples out to teach others.3 For these disciples to engage in teaching activities, one would expect them to have taken some teachings along that they might use to explain the Dharma, teachings they could then pass on to their disciples.4
This body of oral material appears to have been heterogeneous from its very outset, as is in fact the case for oral tradition in general.5 According to Davidson (1992, 293):
During the more than forty years of the Buddha’s teaching career, there were many monks acting as authoritative teachers of the doctrine throughout the kingdom of Magadha and its border areas. They would cross paths with the master from time to time and receive new information as his doctrine and teaching style developed. They would also receive new information from one another during the fortnightly congregations, the summer rains retreats, and whenever they met as their mendicant paths crossed. After forty years of their obtaining new information through such contact, we may be certain that, by the death of the Buddha, the process of receiving new ‘teachings of the teacher’ (śāstuḥ śāsanaṃ) had become a well-accepted practice. The network of instruction was thus established, and doubtless most of the monks realized that much of what the Buddha had said during his lengthy career remained unknown to them personally. They therefore kept the network alive to obtain instruction committed to other bhikṣus. In my opinion, this was the beginning of the continuing cross-fertilization of scripture and doctrine which was the hallmark of Indian Buddhism.
The same element of heterogeneity can also be seen in the account given in the different Vinayas of the so-called first council, which differ even in regard to such basic aspects as the sequence of the four Āgamas or Nikāyas.6 The lack of homogeneity of the early Buddhist oral tradition even finds a quite explicit expression in an episode recorded in the traditional account of the formalization of this oral tradition into a canonical body. According to this episode, the monk "Purāṇa rejected the consensual understanding of the Buddha’s teaching and preferred instead to transmit it as Purāṇa himself had heard it" (Cox2004, 502).
Not only did the Buddhist schools differ on the sequence of the Āgamas, but they also allocated discourses in considerably different ways to their respective Āgama collections. As the Āgamas translated into Chinese stem from different Buddhist schools, the net result of this
is that at times a discourse may not have been preserved in Chinese translation because of the way it was assigned to a particular Āgama by different reciter traditions.
A case example is the Jīvaka-sutta of the Majjhima-nikāya,7 which has not been preserved in Chinese translation. As the discourse treats the issue of meat-eating, one might at first sight be tempted to attribute this absence to an act of conscious elimination by reciters who favoured vegetarianism.8 Yet, among the Sanskrit fragments discovered in CentralAsia parts of a version of the Jīvaka-sutta have been preserved, which appears to belong to a (Mūla-)Sarvāstivāda Dīrgha-āgama collection.9
This shows that the absence of a Chinese translation of the Jīvaka-sutta need not be related to any conscious act of elimination, but may simply be an outcome of the ways in which discourses were allocated to the four Āgamas by different schools. That is, in the Sarvāstivāda traditions this discourse would have been allocated to the Dīrgha-āgama, which explains why it is not found in the Madhyama-āgama translated into Chinese, a collection that probably stems from a Sarvāstivāda tradition.10 The Dīrgha-āgama translated into Chinese, which appears to be of Dharmaguptaka provenance,11 also does not have a version of the Jīvaka-sutta, quite possibly because the reciter tradition to which this collection belonged had allocated their version of the Jīvaka-sutta to another Āgama, perhaps to their Madhyama-āgama.
Thus conclusions based on the absence of a parallel to a discourse need to be treated with considerable circumspection, as mere absence of a parallel could be due to the circumstance that the Āgamas preserved in Chinese stem from different schools.
The Criteria for Recognizing Parallels
The basis for examining variations found between different versions of a discourse is the identification of parallels. The notion of a parallel intends to express that the versions under examination, though differing to some extent, are sufficiently similar to make it probable that they stem from the same occasion.
As a basis for considering two (or more) discourses as ‘parallels’, the information given on the location of a discourse, for example, is not a very reliable criterion, owing to the lack of concern for historical details prevalent in ancient India.12 This lack of concern for the correctness of the location of a discourse finds its explicit expression in a passage in the Mahāsāṅghika and (Mūla-) Sarvāstivāda Vinayas, according to which a reciter who has forgotten the site where a discourse was spoken should just allocate it to one of the main places where the Buddha used to stay 13. The ancient Indian indifference in regard to ‘historical’ details evident from these instructions stands in contrast to the much greater care with which the doctrinal teachings in the discourses were transmitted.14
Not only the information given on the setting of a discourse, but often also the titles of discourses appear to be unreliable criteria for determining ‘parallels’, as titles can be highly variable between otherwise similar versions of a discourse. An example can be found in the Sanskrit version of the (Mūla-)Sarvāstivāda Bhaiṣajyavastu, which refers to five Madhyamaāgama discourses, each time specifying the chapter where the discourses are to be found.15 Each of these references corresponds to the location of the equivalent discourse in the Chinese Madhyama-āgama.16 Though these locations accord with the indications given in the Bhaiṣajyavastu, all five discourses have titles that differ from the titles mentioned in the Bhaiṣajyavastu.17 From the perspective of oral transmission, this indicates that the title of a discourse was still open to modification at a time when this discourse had already been allocated to a particular discourse collection.
A similar conclusion can also be drawn based on examining the twelfth discourse of the Majjhima-nikāya, which concludes with the Buddha giving this exposition the title "hair raising instruction".18 The title "hair-raising" recurs also in a Jātaka tale that parallels the beginning part of this discourse, and in a considerable range of later Pāli works.19 These occurrences indicate that this discourse must have been known for quite some time under this title. A reference to this discourse in the Karmavibhaṅgopadeśa also employs the title "hairraising",20 and a Chinese parallel has the title "discourse that raises the bodily hairs out of joy".21 Yet, the present title of the Pāli version is Mahāsīhanāda-sutta, the "greater discourse on the lion’s roar", or perhaps the "discourse on the great lion’s roar".22
Thus in order to determine ‘parallels’, the information given in the discourse’s introduction, such as the location of the discourse, and even the discourse’s title, are not always reliable. The main criteria should rather be the teachings given in the discourse itself, in the sense of its doctrinal presentation, the similes and illustrations used, etc. As Williams (1970, 166-167) points out: the oral transmission of teaching is generally more conservative than that of narrative material ... [so that] it is likely that the teaching is a more accurate reflection of the oldest tradition than the narrative.
Difficulties in Applying the Concept of a "Parallel"
When attempting to apply the concept of a ‘parallel’, difficulties can arise when two discourses have only part of their presentation in common. A case in point can be found in the eighth discourse of the twenty-fourth chapter of the Ekottarika-āgama, the長壽經.23 The first part of this discourse describes the quarrel among the Kosambī monks and thus parallels the first part of the 長壽王本起經 in the Madhyama-āgama and its Pāli counterpart, the Upakkilesa-sutta.24
The Ekottarika-āgama長壽經 then continues with a description of a visit paid by the Buddha to Anuruddha and his friends which parallels the description of such a visit given in the 牛角娑羅林經 in the Madhyama-āgama and its Pāli counterpart, the Cūlagosiṅga-sutta.25 Thus, while the first part of this Ekottarika-āgama discourse parallels the 長壽王本起經 and the Upakkilesa-sutta, the second part of the same Ekottarika-āgama discourse is rather a parallel to the 牛角娑羅林經 and the Cūlagosiṅga-sutta.
Another example from the same collection is the seventh discourse in the forty-ninth chapter of the Ekottarika-āgama. The first part of this discourse agrees with the跋陀和利經 of the Madhyama-āgama and the Bhaddāli-sutta of the Majjhima-nikāya in describing how a monk publicly refused to obey the Buddha’s instruction to eat only a single meal per day.26 The Ekottarika-āgama discourse then turns to a different event by describing how the sight of a monk who went begging during a stormy night caused fear to a woman, an event described in the 加樓烏陀夷經 of the Madhyama-āgama and the Laṭukikopama-sutta of the Majjhimanikāya.27 After this account, the Ekottarika-āgama discourse continues by relating the events related to the monk who had refused to eat only a single meal per day, events described also in the 跋陀和利經 and the Bhaddāli-sutta. Thus in this case the first and the last section of this Ekottarika-āgama discourse parallel the 跋陀和利經 and the Bhaddāli-sutta, while its middle section parallels the 加樓烏陀夷經 and the Laṭukikopama-sutta.
A third example is the eighth discourse in the fiftieth chapter of the Ekottarika-āgama. Similar to the牟犁破群那經 of the Madhyama-āgama and the Kakacūpama-sutta of the Majjhima-nikāya, this Ekottarika-āgama discourse describes a monk who lived in excessively close association with nuns.28 According to the Ekottarika-āgama account, the same monk also proclaimed the mistaken belief that according to the 阿梨吒經 of the Madhyama-āgama and the Alagaddūpama-sutta of the Majjhima-nikāya was held by another monk and led to quite a different exposition by the Buddha, so that the later part of this Ekottarika-āgama discourse parallels the 阿梨吒經 and the Alagaddūpama-sutta.29
In each of these three cases, the Ekottarika-āgama versions parallel only parts of the expositions found in their Madhyama-āgama and Majjhima-nikāya counterparts. Hence it would not seem appropriate to consider the Ekottarika-āgama discourses as parallels, nor would it appear correct to treat these discourses as if they were not parallels at all.
Another type of problem with the concept of a ‘parallel’ can be seen in the case of the意行經 of the Madhyama-āgama, which Akanuma (1990, 169) reckons a parallel to the Saṅkhāruppatti-sutta of the Majjhima-nikāya.30 Closer inspection shows that these two discourses differ considerably, as the Madhyama-āgama discourse describes how jhāna practice leads to rebirth in the Brahma worlds and the immaterial realms, while the Pāli version describes how formulating a mental aspiration can lead to a range of rebirths, covering rebirth in human families, in the sensuous heavens, the Brahmā worlds and the immaterial realms. That is, the two discourses differ not only in the range of rebirths they describe, but also in regard to the principle responsible for such rebirth. While in the 意行經 the principle responsible for rebirth is meditation practice, in the Saṅkhāruppatti-sutta it is the power of making an aspiration. Notably, such aspiration is based on a set of qualities that do not even mention the development of concentration. This makes it improbable that the two expositions go back to the same occasion.
Another similar case is the想經 in the Madhyama-āgama, which Akanuma (1990, 163) reckons aparallel to the Mūlapariyāya-sutta of the Majjhima-nikāya.31 While the Mūlapariyāyasutta describes the perceptual reactions of worldlings, disciples in higher trainings, arahants and a Tathāgata, the 想經 discusses two types of recluses/ Brahmins and the Buddha. Other differences are that the Pāli discourse includes Nibbāna in the range of objects for such perceptual reactions and concludes by noting that the listening monks were not pleased with the Buddha’s exposition,32 a rather unusual conclusion to a discourse. While the 想經 disagrees in all these respects, a discourse in the Ekottarika-āgama, the 一切諸法之本經,33 agrees with the Mūlapariyāya-sutta in examining worldlings, disciples in higher trainings, arahants and a Tathāgata; it also includes Nibbāna in its range of objects; and it also reports that the monks did not delight in the Buddha’s exposition.34 The agreement between the Ekottarika-āgama 一切諸法之本經 and the Mūlapariyāya-sutta suggests that the considerably different presentation given in the Madhyama-āgama 想經, found also in an individual translation,35 probably goes back to a different occasion.
Thus these two discourses from the Madhyama-āgama, the意行經 and the 想經, do not really fit the idea of a ‘parallel’, since in spite of considerable similarity in the theme of their exposition with the corresponding Majjhima-nikāya discourses, the differences make it improbable that the Chinese and Pāli versions derive from the same original occasion.
These cases show that the distinction between what can be reckoned as a ‘parallel’ and what cannot be reckoned as a ‘parallel’ do not fully capture the actual situation. Perhaps at this point the basic mode underlying this dual distinction needs to be questioned, in order to attempt developing a better approach. In fact, while the dual mode of distinction is central to Western logical thought, ancient Indian logic also knows a four-fold type of logic, the so-called tetralemma, a mode of thought that is a recurrent theme in the Pāli Nikāyas and the Chinese Āgamas.36
The Four-fold Logic
The four-fold logic expands the dual notion of affirmation and negation by adding two further possibilities. In addition to ‘yes’ and ‘no’, according to the four-fold logic there are also the possibilities of ‘both yes and no’ and of ‘neither yes nor no’. In the language of the discourses, these four modes are:
1. "is", hoti,有;
2. "is not", na hoti, 無;
3. "is and is not", hoti ca na ca hoti, 有無;
4. "neither is nor is not", n’eva hoti na na hoti, 非有非無.
In this way, instead of the simple ‘black’or ‘white’type of distinction inherent in the dual logic of affirmation and negation, the four-fold logic envisages the possibility that there could be ‘both black and white’, namely different shades of grey, and the possibility that there could be ‘neither black nor white’, namely colours like yellow, red, or blue, etc.
The mode of thought underlying the four-fold logic not only makes its appearance in a set of questions that according to the discourses the Buddha would set aside as irrelevant, but also is a recurring theme in the early Buddhist analysis of the nature of reality. Instances are, for example, when four types of persons are expounded by listing one who torments himself, one who torments others, one who torments both, and one who torments neither.37 The same mode of thought can also be used to distinguish four modes of action as dark action, bright action, dark-and-bright action, and neither-dark-nor-bright action.38
The more open situation that results from the perspective offered by the four-fold logic might also be apt in comparative Āgama studies, since placing two or more versions of a discourses side by side usually shows various degrees of similarity and difference. Therefore, the dual mode that distinguishes only between ‘same’ and ‘different’ does not always appropriately reflect the complexity of the actual situation and at times runs the risk of obscuring rather than clarifying.
Applied to the present situation, the three examples from the Ekottarika-āgama fall under the third of the four propositions in the four-fold logic, ‘both parallel and not parallel’. From the perspective of their Madhyama-āgama or Majjhima-nikāya counterparts, each of these Ekottarika-āgama discourses is indeed both a parallel (in the case of the part that does correspond to the other versions) and not a parallel (in the case of the part that does not correspond). That is, they are ‘partial parallels’.
In the case of the two examples from the Madhyama-āgama, the fourth logical possibility would suit the occasion best, namely ‘neither parallel nor not parallel’. The discourses in the two cases listed above seem to differ to such an extent that they probably do not go back to the same occasion and hence do not qualify as ‘parallels’. Nevertheless, the Chinese and Pāli versions are so closely related that the possibility that they might stem from the same occasion cannot be totally excluded and a comparative study would need to take them into account, so that they also do not fully fit the category of being ‘not parallels’. Being ‘neither parallels nor not parallels’, each of these two Madhyama-āgama discourses is merely ‘related’ to the respective Majjhima-nikāya discourse.
In using the concept ‘related’, it needs to keep in mind that in order to remain a workable concept only such discourses should be included that do show some very specific and characteristic similarity to each other. Discourses that simply share the same general topic, as for example an examination of the five aggregates, should on that account not be included in this category, in order to avoid that the category becomes impracticable.
Hence, by applying the four-fold logic to the recognition of parallels, the additional categories of a ‘partial parallel’ and a ‘related’ version can be developed, which help to more adequately reflect situations that do not fit easily into the concept of being either a ‘parallel’ or ‘not a parallel’. The employ the four-fold logic in this way is simply an attempt to have recourse to a category, taken from the early Buddhist teachings themselves, in order to approach the study of early Buddhist literature. That is, this attempt does not intend to make a pronouncement on the significance of the four-fold logic as such, but only assumes that modes of thought taken from the cultural and philosophical background of early Buddhism may be helpful to analyze the patterns that have resulted from early Buddhist oral transmission.
The approach that results from having recourse to the fourfold logic would thus yield altogether four concepts for assessing the degree of relatedness between two discourses. The first of these would be a ‘parallel’, a term that designates discourses which share over half of the exposition given within the body of the respective texts and which show such a degree of similarity as to make it probable that they stem from the same occasion. Determining similarity here needs to rely more on the import of the passages in question than on the totalcorrespondence of the respective wording, since due to the prolonged period of oral transmission quite similar instructions can at times show variations in the actual wording. Strong factors for determining similarities are proper names that occur in the respective discourses but are found only rarely elsewhere, unusual development of ideas, use of similar imageries and similes, and an unexpected turn of events.
Once these similarities become less to such a degree that substantial sections found in one version are absent from the other, the concept of a ‘partial parallel’ would come into place. The idea of a ‘partial parallel’would reflect the fact that, even though the two discourses in question no longer fit the idea of being full-fledged ‘parallels’, they nevertheless have sufficient in common so as to give the impression that they may well stem from the same occasion. That is, there is at least ‘partial’ overlap between the versions in question, in that considerable sections ofthediscourses‘parallel’each other,yetatthesametimesizableportionsofthediscoursesdo ‘not parallel’ each other. This would correspond to the option suggested by the four-fold logic of being ‘both parallel and not parallel’.
When, however, a comparison of two discourses gives the impression that these two treatments probably do not seem to stem from the same occasions, yet they still have some elements in common that are sufficiently unique as to make it possible that they could be derivatives of the same original, then the uncertainty involved in such a case could best be captured by treating the versions in question as being ‘related’ to each other. This would then reflect the logical possibility of being ‘neither parallel nor not parallel’, in that the similarities and the difference are such that neither the concept of ‘parallel’ nor theconceptof ‘not parallel’ fit the discourse as a whole, or substantial sections of it. Once, however, even such relatedness is not found and the discourses under comparison clearly do not go back to the same occasion, then these are simply ‘not parallels’.
Evidently, to assign such categories involves a certain degree of subjectivity and the point of employing this fourfold distinction would not be to create watertight conceptual compartments into which any given case in comparative studies has to fit neatly. Instead, having recourse to the fourfold logic is simply an attempt to show that with the help of a category taken from the early Buddhist teachings, the complexity of the relationship between early Buddhist texts can better be taken into account.
It remains to be seen, how far this approach can be of use in regard to the next level of comparison which, based on having identified two discourses as parallels, proceeds to identify the similarities and differences between such parallels.
A possible basis for such identification of similarities and differences is the employment of "mark-up" that encodes such information in a digital version of the discourses in question. Bingenheimer (2003, 374) demonstrates the possibilities of such mark-up with the example of the箭經 of the Saṃyukta-āgama and its parallel, the Sallattena-sutta of the Saṃyuttanikāya.39 Now while the Saṃyukta-āgama version mentions the location where the discourse was spoken, the Pāli version does not give any location. Bingenheimer (2003, 376) comments that in this case: it is for the encoder / interpreter to decide if the absence of the stereotypical formula ‘while the Buddha stayed in Rājagaha ... ‘ in the Pāli version is enough to say that the setting is different ... since the location is not explicitly different in the Pāli, but only absent, I would in this case opt for ‘same setting’.
This could well be a good example for the fourth logical possibility, in the sense that the setting given in the two discourses is ‘neither the same nor different’. And the third logical possibility would fit any instance where the two versions make the same point in considerably different words, thus being a case of ‘both same and different’.
Thus it seems as if approaching a comparative study of Āgama discourses from the perspective of the four-fold logic may help to develop a set of categories that reflect the situation betterthanusingthedual mode of logicwhichsimplydistinguishes betweena‘parallel’and its opposite, or which decides if the sections of two (or more) parallels are ‘same’ or ‘different’.
The Potential of Comparative Studies
Similar to the broadening of perspective that results from applying the four-fold logic to the notion of a parallel, comparative studies in general have a considerable potential to lead to a broadening of perspective in regard to early Buddhism. In addition to clarifying transmission errors, a comparative study of the discourses preserved by different reciter traditions considerably broadens one’s perspective on the early teachings, drawing attention to the common core that is found in versions of a discourse that may at times vary considerably in details. In short, comparative studies can confirm essentials and clarify details.
At the same time, such comparative studies offer a precise tool for investigating the early stages of development of Buddhist thought, as variations detected in this way can reveal the influence of changing viewpoints and opinions on the transmitted material. Differences between various versions of a discourse probably reflect different stages in the organic process of their oral transmission and can help to detect the gradual aggregation of material built on an originally much simpler exposition. In this way, comparative studies provide a methodologically sound approach for assessing the originality of certain aspects of the teachings and for identifying later influences.
In contrast, from a methodological perspective it appears to be considerably less sound to make pronouncements on a discourse or certain aspects of it that are not based on either disagreements with other discourses or variations in regard to parallel versions.
The lack of methodological soundness in such an approach can perhaps best be demonstrated by considering some of the opinions voiced by Buddhist scholars, who have attempted to unearth an original Buddhism that differs decisively from what can be found in the Pāli Nikāyas and the Chinese Āgamas. Among the various hypotheses offered so far, some suggest that the most ancient form of Buddhism did not hold all phenomena to be devoid of a self.40 Others conclude that the five aggregates were absent from the early teachings. 41 Others again see the four stages of awakening as constituting merely a later elaboration of the original teachings. 42
This list of scholarly suggestions could well be continued further and a net result of combining these different hypotheses would result in close to nothing of the early teachings being left, showing the drawbacks of attempting to ‘understand’ early Buddhism merely through subjective opinions that are not backed up by substantial internal inconsistencies or variations between parallel versions. Other scholars have in fact voiced their disagreement with such approaches.
Thus Bareau explains that the idea of a conspiracy among the early reciters to introduce fundamental changes in the early teaching arose at a time when scholarship was aware only of the Pāli canon. By now, however, comparative studies of the parallel material preserved in Chinese and Sanskrit have shown that this hypothesis is not convincing. He also notes that is it inconceivable that the first generation of disciples should have consciously altered the teachings, since during the later stages of Buddhist history, which saw endless time spent in discussion on minor points of Buddhist doctrine among the different schools, criticism would most certainly have been raised in regard to such a move if it had taken place. He concludes that the Buddhism lived and preached by the Buddha and his disciples is, in its main lines, what has been preserved in the early canonical texts.43
Concerning the theory of a conspiracy by the early monks to alter the teachings, Olden-berg wonders where to find the redactors and shrewd forgers that would have been able to undertake such a task without betraying themselves a hundred or a thousand times. According to him, what we find in the early Buddhist texts has grown of itself, out of its own roots, rather than being a secondary redesign of a literature that originally could have been of a completely different appearance.44 According to de Jong (1993, 21 and 25): It would be hypercritical to assert that nothing can be said about the doctrine of earliest Buddhism ... the basic ideas of Buddhism as found in the canonical writings could very well have been proclaimed by him [the Buddha], transmitted and developed by his disciples and, finally, codified in fixed formulas.
In a similar vein, Frauwallner voices disagreement with those who treat the canonical texts as totally unreliable and believe that nothing certain can be said about the teachings of the Buddha. He points out that the transmitted texts cannot be deemed unreliable merely because they are not confirmed by external proofs. Those who nevertheless wish to reject the value of such material would, according to him, also have an obligation to explain how this material has come into being.45 Lamotte (1988, 156 and 639) comments that
Any attempt to reconstruct a ‘pre-canonical’ Buddhism deviating from the consensus between the Āgamas and Nikāyas can only end in subjective hypotheses ... in order to appreciate early Buddhism, the only valid evidence - or indication -which we possess is the basic agreement between the Nikāyas on the one hand and the Āgamas on the other. This evidence or indication carries more weight than academic hypotheses put forward after an interval of twenty-five centuries.
Here comparative Āgama studies offer an important tool for confirming and expanding the position taken by these scholars, opening up a methodologically sound approach to an assessment of the early Buddhist teachings and to a delineation of their early stages of development. Such an assessment and delineation of the early Buddhist teachings would be relevant to all
Buddhist traditions, as it reveals their common starting point and at the same time shows the beginnings of tendencies that, in one way or another, have influenced the development of each of the Buddhist schools.
A central task in relation to such comparative studies is exploring the riches of the Chinese Āgamas, which have not yet received the attention they deserve, at least when compared with the attention given to the Pāli Nikāyas. This task has now been facilitated to a remarkable degree through the digitalisation of the Chinese canon by the Chinese Buddhist Electronic Text Association, CBETA. The impact of this tool offered to the scholarly world is considerable and later generations may well distinguish between comparative studies of Buddhism undertaken during the pre-CBETA and the post-CBETA periods, similar to the distinction drawn between the Chinese translations undertaken before or after Kumārajīva. With such tools placed at our disposal, what remains to be done is to move ahead in this field of research, whose potential for improving our knowledge of early Buddhism can hardly be overestimated.
(Chinese and Pāli sources are cited according to the Taishō and PTS editions, giving first the discourse by number and then its location by volume, page and line; Tibetan sources are cited by location in the Derge edition.)
Akanuma, Chizen. 1990. The Comparative Catalogue of Chinese Āgamas & Pāli Nikāyas. Nagoya: Hajinkaku-Shobō, 1929. Reprint, Delhi: Sri Satguru (page references are to reprint edition).
Anālayo. 2007. Oral Dimensions of Pāli Discourses: Pericopes, other Mnemonic Techniques, and the Oral Performance Context. Canadian Journal of Buddhist Studies 3:5-33.
Anesaki, Masaharu. 1908. The Four Buddhist Āgamas in Chinese, A Concordance of their Parts and of the Corresponding Counterparts in the Pāli Nikāyas. Transactions of the Asiatic Society of Japan 35(3):1-149.
Bareau, André. 1966. L’Origine du Dīrgha-Āgama Traduit en Chinois par Buddhayaśas. Essays Offered to G.H. Luce by his Colleagues and Friends in Honour of his Seventy-fifth Birthday. Ed. B. Shin et al. Switzerland: Artibus Asiae. 49-58.
Bareau, André. 1974. Le Parinirvāṇa du Buddha et la Naissance de la Religion Bouddhique. Bulletin de l’École Française d’Extrême Orient 61:275-299
Bingenheimer, Marcus. 2003. Issues in the Use of Electronic Markup for the Comparative Analysis of Āgama Literature. Chung-Hwa Buddhist Journal 7:361-384.
Brough, John. 2001. The Gāndhārī Dharmapada. London: Oxford University Press, 1962. Reprint, Delhi: Motilal Banarsidass (page references are to reprint edition).
Coward, Harold. 1986. Oral and Written Texts in Buddhism. The Adyar Library Bulletin 50:299-313.
Cox, Collett. 2004. Mainstream Buddhist Schools. Encyclopedia of Buddhism. Ed. R.E. Bus-well. New York: Macmillan. 2:501-507.
Davidson, Ronald M. 1992. An Introduction to the Standards of Scriptural Authenticity in Indian Buddhism. Chinese Buddhist Apocrypha. Honolulu: University of Hawaii Press, 1990. Ed. R.E. Buswell. Reprint, Delhi: Sri Satguru (page references are to reprint edition). 291-325.
de Jong, J.W. 1993. The Beginnings of Buddhism. The Eastern Buddhist 26(2):11-30.
Demiéville, Paul. 1951. A propos du Concile de Vaiśālī. T’oung Pao 40:239-296.
Dutt, Nalinaksha, ed. 1984. Gilgit Manuscripts, Mūlasarvāstivāda Vinayavastu. 3(1). Delhi: Sri Satguru. Enomoto, Fumio. 1984. The Formation and Development of the Sarvāstivāda Scriptures. Proceedings of the Thirty-First International Congress of Human Sciences in Asia and North Africa. Ed. Y. Tatsuro. Tokyo: Tōhō Gakkai. 197-198.
Frauwallner, Erich. 1953. Geschichte der Indischen Philosophie. Salzburg: Müller. 1.
Gombrich, Richard F. 1990. How the Mahāyāna Began. The Buddhist Forum 1:21-30.
Hartmann, Jens-Uwe. 2002. Further Remarks on the New Manuscript of the Dīrghāgama. Journal of the International College for Advanced Buddhist Studies 5:133-150.
Horner, I.B. 1934. The Four Ways and the Four Fruits in Pāli Buddhism. Indian Historical Quarterly 785-796.
Horner, I.B. 1979. The Early Buddhist Theory of Man Perfected. London: Williams & Norgate, 1936. Reprint, Delhi: Oriental Books (page references are to reprint edition).
Jayatilleke, K.N. 1980. Early Buddhist Theory of Knowledge. London: Allen & Unwin, 1963. Reprint, Delhi: Motilal Banarsidass (page references are to reprint edition).
Lamotte, Étienne. 1988. History of Indian Buddhism. From the Origins to the Śaka Era. S. Webb-Boin, trans. Original title: Histoire du Bouddhisme Indien. Louvain: Pulications Universitaires, 1985. Reprint, Louvain-la-Neuve: Institut Orientaliste (page references are to reprint edition).
Lévi, Sylvain, ed. and trans. 1932. Mahākarmavibhaṅga. Paris: Ernest Leroux.
Lü, Cheng. 1963. Āgama. Encyclopaedia of Buddhism. Ed. G.P. Malalasekera. Sri Lanka: Department of Buddhist Affairs. 1(2):241-244.
Mayeda [=Maeda], Egaku. 1985. Japanese Studies on the Schools of the Chinese Āgamas. Zur Schulzugehörigkeit von Werken der Hīnayāna-Literatur, Erster Teil. Ed. H. Bechert. Göttingen: Vandenhoeck & Ruprecht. 1:94-103.
Minh Chau, Thich. 1991. The Chinese Madhyama Āgama and the Pāli Majjhima Nikāya. Delhi: Motilal Banarsidass.
Oldenberg, Hermann. 1898. Buddhistische Studien. Zeitschrift der Deutschen Morgen ländischen Gesellschaft 52:613-694.
Pradhan, P., ed. 1967. Abhidharmakośabhāṣya. Patna: K.P. Jayaswal Research Institute.
Prasad, Chandra Shekhar. 1993. The Chinese Āgamas vis-à-vis the Sarvāstivāda Tradition. Buddhist Studies Review 10:45-56.
Przyluski, Jean. 1926. Le Concile de Rājagṛha, Introduction a l’Histoire des Canons et des Sectes Bouddhiques. Paris: Paul Geuthner.
Rhys Davids C.A.F. 1978. The Birth of Indian Psychology and its Development in Buddhism. London: Luzac & Co., 1936. Reprint, Delhi: Oriental Books (page references are to reprint edition).
Scharfe, Hartmut. 2002. Education in Ancient India. Leiden: E.J. Brill.
Schopen, Gregory. 2004a. Marking Time in Buddhist Monasteries: On Calendars, Clocks, and Some Liturgical Practices. Buddhist Monks and Business Matters, Still More Papers on Monastic Buddhism in India. Honolulu: University of Hawai’i Press. 260-284.
Schopen, Gregory. 2004b. If you can’t Remember, How to Make it up, Some Monastic Rules for Redacting Canonical Texts. Buddhist Monks and Business Matters, Still More Papers on Monastic Buddhism in India. Honolulu: University of Hawai’i Press. 395-407.
Skjærvø, Prods Oktor. 2003. Fragments of the Ratnakūṭa-sūtra (Kāśyapaparivarta) in Khotanese. Religious themes and texts of pre-islamic Iran and Central Asia: Studies in honour of Professor Gherardo Gnoli on the occasion of his 65th birthday on 6th December 2002. Ed. C.G. Cereti. Wiesbaden: Reichert. 411-424.
Skilling, Peter. 1979. Discourse on the Four Kinds of Karma. The Journal of Religious Studies 7(1): 86-91.
Skilling, Peter, ed. 1997. Mahāsūtras: Great Discourses of the Buddha. Oxford: Pali Text Society. 2.
Senart, Émile, ed. 1897. Le Mahāvastu. Paris: Imprimerie Nationale. 3.
Stache-Rosen, V. 1968: Dogmatische Begriffsreihen im Älteren Buddhismus II, Berlin: Akademie Verlag.
Vansina, Jan. 1985. Oral Tradition as History. Madison: University of Wisconsin Press.
Vorobyova-Desyatovskaya M.I. et al., ed. 2002. The Kāśyapaparivarta, Romanized Text and Facsimiles. Tokyo: Soka University.
Waldschmidt, Ernst, ed. and trans. 1980. Central Asian Sūtra Fragments and their Relation to the Chinese Āgamas. The Language of the Earliest Buddhist Tradition. Ed. H. Bechert. Göttingen: Vandenhoeck & Ruprecht. 136-174.
Williams, Raymond B. 1970. Historical Criticism of a Buddhist Scripture: The Mahāparinibbāna Sutta. Journal of the American Academy of Religion 38(2):156-167.
Yin-shun. 1983.原始佛教聖典之集成 [The Compilation of the Early Buddhist Canon]. Taipei: 正聞出版社, 1962. Reprint, Taipei: 正聞出版社 (page references are to reprint edition).
1I am indebted to Bhikkhu Bodhi, Rod Bucknell and Ken Su for comments on a draft of this article.
2For a more detailed study of oral characteristics of the early discourses cf. Anālayo (2007).
3This account can be found in the Dharmaguptaka Vinaya at T 1428, 793a7; in the Mahāvastu of the Mahāsāṅghika tradition in Senart (1897, 415,8); in the Mahīśāsaka Vinaya at T 1421, 108a7; in the (Mūla-)Sarvāstivāda Vinaya at T 1450, 130a20; in the Sarvāstivāda Vinaya at T 1440, 511a12; and in the Theravāda Vinaya at Vin I 21,1; cf. also SN 4:5 at SN I 105,24 (or SN2 1 1 at SN2 I 236,10) and its parallel SĀ 1096 at T 99 288b3.
4Cf. also Gombrich (1990, 25).
5Vansina (1985, 161) explains that "at any moment in time the [oral] corpus of any community is in fact not totally homogeneous".
6The Mahāsāṅghika Vinaya at T 1425, 491c16 and the Mahīśāsaka Vinaya at T 1421, 191a24 agree with the Theravāda tradition on the order Dīrgha, Madhyama, Saṃyukta, Ekottarika (Vin II 287,27 does not list the order explicitly, though the same appears to be implicit). The Dharmaguptaka Vinaya at T 1428, 968b19 lists the four collections in the order Dīrgha, Madhyama, Ekottarika, Saṃyukta. The (Mūla-)Sarvāstivāda Vinaya at T 1451, 407b27 lists the four collections in the order Saṃyukta, Dīrgha, Madhyama, Ekottarika; an order that appears to also underlie the presentation in the Sarvāstivāda Vinaya at T 1435, 448b13, though this Vinaya does not list the four collections and therefore does not explicitly specify their order. For a more detailed study that takes into account additional records of the first council cf. Przyluski 1926; on the order of the four collections in other works cf. also Mayeda (1985, 96).
7MN 55 at MN I 368-371.
8Minh Chau (1991, 31) concludes that "the dropping from all the Chinese Āgamas of the Pāli sutta No , Jīvakasutta, in which the Buddha was reported to allow the monks to take three kinds of meat, confirms the Sarvāstivadin’s attitude against meat-eating".
9Hartmann (2002, 138).
10Lü (1963, 242); Mayeda (1985, 98); Minh Chau (1991, 27); Waldschmidt (1980, 136) and Yin-shun (1983, 703). According to Enomoto (1984, 198), the Madhyama-āgama translated into Chinese probably represents the earliest of three versions of this discourse collection, the second of the three being the version partly preserved in some of the Central Asian Sanskrit fragments, while the third version underlies sūtra quotations in later works.
11Bareau (1966, 50); Brough (2001, 50); Demiéville (1951, 252-253); Lü (1963, 242); Mayeda (1985, 97); Prasad (1993, 50); Waldschmidt (1980, 136); Yin-shun (1983, 720).
12Coward (1986, 305) explains that "the early Buddhists shared ... the Indian indifference to historical details. Historical events surrounding a text are judged to be unimportant in relation to the unchanging truth the text contains". Gombrich (1990, 22) comments that "from the religious point of view this is perfectly understandable: the narrative framework of the sayings is not relevant to salvation".
13T 1425, 497a6; T 1451, 328c15; and T 1458, 575b29; the corresponding passage in the Tibetan (Mūla-)Sarvāstivāda Vinaya (found at D ’dul ba, da 39b3) together with a discussion in Schopen (2004b, 395-407); cf. also Schopen (2004a, 283 note 59).
14Scharfe (2002, 25; note 93) comments on the Vinaya instructions regarding the location of a discourse that "it is worth noting that no such ‘creativeness’ was allowed where the contents of the lesson is concerned".
15The Bhaiṣajyavastu in Dutt (1984, 93,10) refers to the Māndhātṛ-sūtra found in the Rājasaṃyuktanipāta; (98, 15) to the Velāma-sūtra in the Brāhmaṇanipāta; (111, 20) to the Mahādeva-sūtra; (112, 19) to the Nimi-sūtra in the Rājasaṃyuktanipāta; and (217,12) to the Nandīpāla-sūtra in the Rājasaṃyuktanikāya. The Chinese counterparts references T 1448, 56b11 (Māndhātṛ); T 1448, 57b13 (Velāma); T 1448, 58c1 (Mahādeva); and T 1448, 58c16 (Nimi), however, mostly refer to the subject matter without explicitly giving the discourse title; and only specify the chapter of the Madhyama-āgama collection in the case of the Māndhātṛ-sūtra, T 1448, 56b11 (the chapter in which the Mahādeva-sūtra is found is given at T 1448, 30b8).
16Waldschmidt (1980, 142-144) identifies the following counterparts: the Māndhātṛ-sūtra corresponds to MĀ 60 at T 26, 494b-496a; the Velāma-sūtra corresponds to MĀ 155 at T 26, 677a-678a; the Mahādeva-sūtra and the Nimi-sūtra correspond to MĀ 67 at T 26, 511c-515b; and the Nandīpāla-sūtra corresponds to MĀ 63 at T 26, 499a-503a. MĀ 60, MĀ 63 and MĀ 67 are indeed found in the 王相應品, and MĀ 155 occurs in the 梵志品.
17The counterpart to the Māndhātṛ-sūtra, MĀ 60 at T 26, 494b9, has the title 四洲, "four continents". The counterpart to the Velāma-sūtra, MĀ 155 at T 26, 677a8, has the title 須達哆, corresponding to "*Sudatta". The counterpart to the Mahādeva-sūtra and the Nimi-sūtra, MĀ 7 at T 2 , 11c21, has the title 大天奈林, "*Mahādeva's mango-grove". The counterpart to the Nandīpāla-sūtra, MĀ 63 at T 26, 499a9, has the title, 鞞婆陵, corresponding to its location at Vebhaḷiṅga or Vaibhiḍiṃgyā. Notably, Śamathadeva's commentary on the Abhidharmakośa gives the title of MĀ 67 just as Lha chen po, cf. D mngon pa, ju 76b2, thereby agreeing with the (Mūla-)Sarvāstivāda Vinaya (and the Karmavibhaṅgopadeśa in Lévi 1932, 161,14). In the case of MĀ 63, however, Śamathadeva’s commentary uses the name of the location as the discourse’s title, thereby agreeing with the Sarvāstivāda Madhyama-āgama collection against the (Mūla-)Sarvāstivāda Vinaya, cf. D mngon pa, ju 244b4 and Skilling (1997, 279-285).
18MN 12 at MN I 83,2 : Lomahaṃsanapariyāya. The title Lomahaṃsa occurs also in the summary verse (uddāna) of the Burmese and Siamese editions, Be-MN I 172, and Se-MN I 2 7,17, even though these editions use Mahāsīhanāda as the title, cf. Be-MN I 97,14 and Se-MN I 137,1.
19This is the Lomahaṃsa-jātaka at Jā I 389. In Mil 396, Nāgasena quotes part of the Mahāsīhanāda-sutta, to which he refers as the Lomahaṃsanapariyāya. Another reference to the Lomahaṃsanapariyāya occurs in Ppk-a 104. Several Pāli works speak of the same discourse as the Lomahaṃsa(na)-sutta: Sv I 179; It-a I 109; Be-Mp-ṭ II 256; Be-Sp-ṭ I 334; Be-Abhidhān-ṭ 504; and Be-Sīlkkh-abh-ṭ II 74. Cp 102 entitles a set of verses concerned with the bodhisattva’s ascetic practices as Mahālomahaṃsacariyā, the "great hair raising conduct". As MN 12 treats the bodhisattva’s ascetic practices, this title may well be related to the present discourse.
20Lévi (1932, 158, 11): Romaharṣaṇīya sūtra.
21T 757, 591c11: 身毛喜豎經.
22MN 12 at MN I 83,28: Mahāsīhanādasuttaṃ; according to the subcommentary, Be-Psṭ II 40, the title Mahāsīhanāda was accorded to this discourse by the reciting elders, saṅgītikāramahāthera.
23Taking the title from the summary verse at T 125, 630b16, cf. also Anesaki (1908, 143).
24EĀ 24.8 at T 125, 626b-629a parallels MĀ 79 at T 125, 532c-535c and MN 128 at MN III 152-154.
25EĀ 24.8 at T 125, 629a-630a parallels MĀ 185 at T 26, 729b-731a and MN 31 at MN I 205-211.
26EĀ 49.7 at T 125, 800b-c parallels MĀ 194 T 26, 746b-c and MN 65 at MN I 437-438.
27EĀ 49.7 at T 125, 800c-801b parallels MĀ 192 at T 26, 741b and MN 66 at MN I 448-449, after which it returns to the events described in MĀ 194 and MN 65.
28EĀ 50.8 at T 125, 812c parallels MĀ 193 at T 26, 744a and MN 21 at MN I 122
29EĀ 50.8 at T 125, 812c-813b parallels MĀ 200 at T 26, 763b-764b and MN 22 at MN I 130-134.
30MĀ 168 at T 26, 700b-701b and MN 120 at MN III 99-103.
31MĀ 106 at T 26, 596b-c and MN 1 at MN I 1-6.
32Be-MN I 8,19, Ce-MN I 18,9 and at Se-MN I 11,6 report that the monks did not delight in what the Buddha said, whereas the PTS edition at MN I 6,24 reports that the monks did delight in the discourse.
33Title taken from T 125, 766a7, the summary verse at T 125, 769b6 reads 法之本.
34EĀ 44.6 at T 125, 766b15. A somewhat similar episode, with a group of monks being unable to appreciate a teaching given by the Buddha, can be found in the Kāśyapaparivarta, cf. the Sanskrit edition by Vorobyova-Desyatovskaya (2002, 48), folio 69 verso line 2-3 (§ 138); Chinese versions in T 310, 637b13; T 350, 193b15; T 351, 199b26; T 352, 214c21; a Khotanese version in Skjærvø (2003, 417); and a Tibetan version at D dkon brtsegs, cha 146b6.
35T 56 at T 26, 851a-b.
36For a detailed study of the implications of this four-fold logic cf. Jayatilleke (1980, 339-350).
37DN 33 at DN III 232,22, with a Sanskrit fragment counterpart edited in Stache-Rosen (1968, 122); cf. also the Saṅgītiparyāya, T 1536, 406a7.
38MN 57 at MN I 389,21, cf. also the Abhidharmakośabhāṣya 4:60 in Pradhan (1967, 235,1) its translations T 1558, 83b18 and T 1559, 239b24, and a Tibetan counterpart in Skilling (1979).
39SĀ 470 at T99, 119c-120b and SN 36.6 at SN IV 207-210; title for SĀ 470 taken from Anesaki (1908, 95).
40Horner (1979, 41).
41Rhys Davids (1978, 193).
42Horner (1934, 787).
43Bareau (1974, 280): "le Bouddhisme vécu et prêché par le Buddha et ses premiers disciples est donc bien, dans se grandes lignes tout au moins, celui que nous trouvons décrit et enseigné dans les textes canoniques antiques qui nous sont parvenue en sanskrit, en pāli ou en traduction chinoise".
44Oldenberg (1898, 674-675): "aber wo gab es denn die raffinierten Fälscher, wo gab es die Redaktoren ... dass sie einer solchen Arbeit gewachsen gewesen wären, ohne sich hundert- und tausendmal zu verraten ... was wir hier vor uns haben, ist so, wie wir es sehen, aus sich selbst, aus seinen eigenen Wurzeln erwachsen; es ist nicht eine sekundäre ... Umgestaltung einer Literatur, welche ursprünglich ein ganz anderes Aussehen gezeigt haben könnte."
45Frauwallner (1953, 465): "ebenso wenig kann ich mich aber auch der Auffassung anschließen, welche die kanonische Überlieferung des Buddhismus für vollkommen unglaubwürdig hält und ... meint, daß es aussichtslos sei, über die Lehre des Buddha selbst irgendetwas Sicheres ermitteln zu wollen ... überliefertes Quellenmaterial ist noch nicht unglaubwürdig, wenn die äußere Bezeugung fehlt ... wer sie [die kanonischen Texte] aber trotzdem verwirft, darf sich nicht auf die bloße Verneinung beschränken, sondern hat die Pflicht, auch ihr Zustandekommen zu erklären und zu begründen."
Source: Chung-Hwa Buddhist Journal (2008, 21:3-21)