This attempted reconstruction of the early Buddhist path to enlightenment is based entirely on sources preserved in written texts. The principal sources used are the Pali suttas and their Chinese parallels, but occasional reference will also be made to relevant Sanskrit and Tibetan texts. The present chapter begins by surveying and evaluating these sources, briefly in the case of the Pali and Sanskrit, more extensively for the less familiar Chinese. Then follows a discussion of difficulties encountered in establishing parallels between Pali and Chinese suttas and of what can be revealed by comparing them. Glimpses are thereby provided into the analytical and interpretative methods that will be applied in the remainder of the book. The chapter ends with brief accounts of what we know about how Chinese and Pali sources were transmitted and became available to scholars.
As much of this discussion of sources and methods is of a technical nature, it will probably be of greater interest to scholars than to general readers – particularly the longish sections on Chinese sources and Pali–Chinese parallels. General readers may therefore prefer to skip either or both of these sections – or indeed the whole chapter and go directly to “The Stepwise Training” (Chapter 4).
Pali Sources
Despite the importance I have attached to Chinese materials, this study draws, in the first instance, on Pali sources. Of the three piṭakas recognised in the “Pali canon”, the Sutta-piṭaka or “Basket of Discourses” is self-evidently the most relevant one here. The Vinaya-piṭaka or “Basket of Discipline”, insofar as it is relevant, has already been made good use of in the previous chapter and will only rarely be mentioned henceforth. The third basket, the Abhidhamma-piṭaka, with its systematising presentation of doctrinal material, represents for the most part a period subsequent to the early splits in the Sthavira transmission. Consequently, it too has limited relevance here. A partial exception is the Vibhaṅga, the second of the seven Pali Abhidhamma books, on which more will be said later in this chapter.
Within the Pali Sutta-piṭaka, the basic corpus for this project is the first four nikāyas: the Dīgha-, Majjhima-, Saṃyutta-, and Aṅguttara-nikāyas (here abbreviated DN, MN, SN, AN), the collections of Long, Middle-Length, Connected, and Numerical Discourses respectively. The very diverse fifth nikāya, the Khuddaka-nikāya or Minor Collection, will be drawn on only occasionally. Two of the Khuddaka’s fifteen books, namely Suttanipāta and Dhammapada, are widely considered to belong to the earliest stratum of the canon, a feature that might seem to make them particularly appropriate sources for a study of early Buddhism. However, because they have relatively little explicitly doctrinal content, these two books will actually be drawn on only to a limited extent.Footnote 1
Early and Late in the Sutta-piṭaka
Despite the traditional claim that the four main nikāyas were put together at the First Council, their component suttas are recognisably of various different ages. Material of various ages is also sometimes discernible even within an individual sutta. In this study, aimed at identifying the teachings of early Buddhism, matters of relative chronology are of particular significance and need to be taken into account.
A wide-ranging preliminary attempt to distinguish between earlier and later material in the Pali Sutta-piṭaka is presented by G. C. Pande in his Studies in the Origins of Buddhism (1983). While the findings of Pande’s survey are clearly relevant for the present project, they are not as directly and easily applicable as might be expected, for several reasons.
Some of the criteria by which Pande judges sutta material to be historically “early” or “late” are not self-evidently well founded. Most open to question is his generalisation that a more complex account of a doctrine is likely to be historically later than a simpler account. For example, Pande considers that an account of the grades of “noble ones” (ariya-puggalā) that recognises five types of non-returner (anāgāmi) is likely to be relatively late; the more familiar account, which does not subdivide the non-returner category in this way, is likely to be earlier.Footnote 2 Thus, a sutta or portion of a sutta containing a more-than-usually complex account of a piece of doctrine would be assigned to the “late” category in Pande’s classification.
To apply this as a general principle, however, would be to go too far. Admittedly, the findings presented in later chapters will often support the generalisation that the more complex or detailed version of a given doctrine has probably developed out of the simpler version – for example, through scholastic elaboration; however, they will also reveal cases where it is fairly clear that the simpler version has developed out of the more complex – for example, through unintended deletion of material. In the present study the task of deciding which is the earliest form of each piece of doctrine will entail much more than an assessment of relative complexity. Accordingly, any of Pande’s judgements that are based on this criterion alone will be treated with caution.
More reliable, I would argue, are those of Pande’s judgements that are based on a sutta’s structure. A surely uncontroversial example is the last discourse in the Long Collection, the Dasuttara-sutta or “Discourse with Increasing Decades” (DN34), which Pande deems “very late”.Footnote 3 This discourse (which is spoken by the senior monk Sāriputta, rather than by the Buddha himself) has a remarkably regular and neat appearance. In it one hundred doctrinal sets are grouped, according to the number of items they contain (from one to ten) into ten decades, which are then presented in ascending numerical order, with the items of each decade listed according to a fixed set of ten attributes or categories. These features certainly suggest that the Dasuttara-sutta is an artificial product of monastic scholasticism. Pande proposes that it may have developed out of the Saṅgīti-sutta, “Discourse on Group Recitation” (DN33), which immediately precedes it in DN and is also spoken by Sāriputta.Footnote 4 But, although its structure indicates that the Dasuttara is a relatively late compilation, it does not follow that all of the one hundred doctrinal sets listed in this sutta are also late creations. Indeed, many of these sets are well documented in other, evidently earlier, less artificial-looking suttas.
Another instructive example is provided by material contained in the Sāriputta-saṃyutta, the section of SN that is “Connected with Sāriputta”.Footnote 5 The first nine of the ten suttas making up this saṃyutta are collectively identified by Pande as late.Footnote 6 In keeping with a pattern typical of SN, the first sutta begins with “On one occasion the Venerable Sāriputta was dwelling at Sāvatthī, in Jeta’s Grove, Anāthapiṇḍika’s Park”, but the eight that follow it begin with just the terse “At Sāvatthī”. These nine suttas deal, in succession, with the stages of jhāna, concentrative meditation. The first of the nine deals with the first jhāna, the second with the second jhāna, and so on through the four jhānas, the four āruppas (non-material concentrations), and nirodha (cessation) – nine stages in all, in nine consecutive suttas. The description of each meditation stage is embedded in an identically repeated account of the circumstances in which the teaching is delivered: Sāriputta, having completed his alms-round in Sāvatthī and eaten his meal, spends the rest of the day meditating in a nearby grove; in the late afternoon, on being asked by Ānanda why his countenance is so calm and clear, he explains that this is due to his detached abiding in the meditation stage in question – the first jhāna, the second jhāna, and so on.
These nine suttas are presented as a record of nine nearly identical encounters between Sāriputta and Ānanda on nine separate occasions, the only variation being that in each sutta Sāriputta reports being one stage further ahead in the jhāna series than he was in the preceding sutta. As Pande points out, this is not a typical saṃyutta. Typically a saṃyutta of SN comprises a varied lot of discourses, said to have been delivered at various different locations and in various different circumstances. But these nine discourses do not possess the features of a typical saṃyutta. In a typical saṃyutta the suttas have a common topic or other shared feature that justifies the editorial process of bringing them together under a single heading. Examples of typical saṃyuttas are: Khandha-saṃyutta, “Connected with the Aggregates” (SN22); Kassapa-saṃyutta, “Connected with [the senior monk] Kassapa” (SN16); and Opamma-saṃyutta, “Connected with Similes” (SN20). In a typical saṃyutta it would not matter if the suttas were arranged in a different sequence because each of them is self-contained. The Sāriputta-saṃyutta, however, although it is indeed connected with Sāriputta, possesses none of the other features typical of a saṃyutta of SN.
Pande is right, therefore, in observing that this set of nine suttas was “clearly composed as a whole”.Footnote 7 A likely explanation is that it developed out of a single original sutta through unwarranted application of the framing narrative to each of the nine paragraphs describing the meditation stages. The nine pseudo-suttas thus produced were kept together as a set by being identified (together with an unrelated tenth suttaFootnote 8) as constituting a saṃyutta.
While it is appropriate to apply the label “late” to the existing structure of the Sāriputta-saṃyutta, it would not necessarily be appropriate to apply the same label to the doctrinal message contained within that structure. The series of nine meditation stages is described in virtually identical wording at many other locations in the Pali Sutta-piṭaka. The earliness or lateness of this account of meditation practice must be judged on criteria that have little to do with the structure of the Sāriputta-saṃyutta.
Examination reveals that the phenomenon just described is not uncommon. There are several cases, scattered through SN, where a group of consecutive suttas (usually ten of them together portrayed as constituting a saṃyutta) shows clear signs of being derived from what was formerly a single sutta.Footnote 9 The self-evident artificiality of such a piece of text is an unmistakable sign of lateness. Whether the doctrinal message contained in that text is correspondingly late is a separate issue – though lateness of structure does frequently correlate with lateness of content.
Another, more conspicuous type of artificiality found in SN is the “Ganges repetition series”. This is a kind of refrain which converts one sutta into forty-eight pseudo-suttas within the Magga-saṃyutta (Connected with the Path) and then produces a similar effect in seven following saṃyuttas.Footnote 10 Here again the manifest lateness of the superimposed structure does not disqualify the affected material from serving as a source in the search for early Buddhism. Rather, it signals that this late structural feature, once recognised as such, can justifiably be bracketed out of the subsequent discussion of the content.
Much artificiality is found in the Numerical Collection (Aṅguttara-nikāya) as well, particularly in the Section on Ones (Ekaka-nipāta). Pande cites a good example contained in the fifth and sixth decades (vaggas) of this section.Footnote 11 He points out that the ten so-called suttas making up the fifth decade plus the first two of the sixth decade (i.e., AN1.5.1–10 plus AN1.6.1–2) together constitute a single genuine sutta with a consistent topic: the directed and pellucid mind. Looking beyond the artificial division into twelve “suttas”, and the even more artificial separation of the first ten from the remaining two (designed to produce regular decades), Pande concludes, with good reason, that the underlying unitary set of twelve pieces of text is “early”.
A simpler and more typical example is the first decade of the Ones (AN1.1.1–10). In the first of its ten suttas the Buddha says: “Monks, I know of no other single visible form that so enslaves a man’s mind as the visible form of a woman …”; in the next four suttas he says the same of the sound, odour, taste, and tactile experience of a woman; and in the remaining five he repeats it all with the words “man” and “woman” interchanged. That is, what the text identifies as a decade of suttas (a vagga) actually has the characteristics of a single sutta consisting of ten closely related paragraphs. Clearly, this decade has resulted from breaking up a single sutta in order to produce “suttas” featuring the required number, One. More natural would have been to locate that source sutta in the Fives, or perhaps in the Tens; indeed, inspection of the Fives reveals a sutta that may well have been that inferred source.Footnote 12
Looking further in the AN Section on Ones (Ekaka-nipāta), one is led to conclude that this entire section is an artificial creation. The PTS sutta numbering indicates that the total number of suttas in the Ones is 535, and the traditional numbering used in native-script editions puts the number at 627;Footnote 13 it can be shown, however, that the underlying source suttas number just 28.Footnote 14 In the Twos this kind of artificiality (designed to produce suttas featuring the number Two) is again in evidence, though on a much smaller scale, and in the remainder of the collection it appears only occasionally. The seemingly intentional creation of so many “suttas” featuring the number One may have been motivated by a wish to provide the Pali Numerical Collection with a Section on Ones that it had formerly lacked. In any case, the implications are as with the artificial saṃyuttas in SN: in interpreting any such contrived piece of text, one begins by inferring the relevant doctrinal statement as it was in the underlying source sutta.Footnote 15
The phenomena just considered reveal an ambiguity in the notion of sutta. On the most immediate and most superficial level, a sutta is a visually discrete piece of text that typically begins with “Thus have I heard” or “At Sāvatthī”. On a deeper level, however, a sutta is a putative record of a single entire discourse that was delivered on some occasion by the Buddha or a specified disciple, was reviewed and approved at the First Council, and ended up being preserved more or less faithfully in one or more versions of the Sutta-piṭaka. As a consequence of the phenomena described above, a sutta in the latter sense (unitary discourse) may sometimes be represented in the nikāyas by a set of two or more suttas in the former sense (discrete piece of text).
On a slightly different theme, Pande points out that one can often discern earlier and later layers within a single sutta. A likely example of this emerges when the “Greater Discourse on the Foundations of Mindfulness” (Mahā-satipaṭṭhāna-sutta, DN22) is compared with the “Discourse on the Foundations of Mindfulness” (Satipaṭṭhāna-sutta, MN10) and its two parallels in Chinese (MA98 and EA12.1). As mentioned in Chapter 1, the Greater (Mahā) version (DN22) differs from the other (MN10) in including a lengthy and detailed exposition of the four noble truths. The inclusion of this exposition is likely to be a relatively late development, and this needs to be taken into account in any attempt to infer what was in the ancestral description of mindfulness practice. Doing so does not necessarily imply, however, that the added exposition of the four truths is itself a late creation. Much the same account of the four truths is found in the free-standing “Discourse on Analysis of the Truths” (Saccavibhaṅga-sutta, MN141), which indicates that the expansion of the “Discourse on the Foundations of Mindfulness” to yield the “Greater” version may have consisted in appending to it the main content of the “Analysis of the Truths”. The earliness or lateness of the “Analysis of the Truths” itself is a separate issue.
The many cases cited by Pande show that classification as early or late may be applicable to entire suttas or to sections within suttas. The underlying notion that different portions of the Sutta-piṭaka date from different periods is clearly fundamental to the present project. However, the presence of a late section within a sutta does not necessarily render that entire sutta unreliable as a source of data. A noteworthy implication of this is that the high level of structural artificiality found in SN and AN does not make those two collections less reliable overall than the more narratively structured DN and MN.
Sanskrit Sources
In the previous chapter mention was made of the progressive translation of the non-Pali Buddhist canons into Sanskrit. Of the many Sanskrit Sūtra-piṭakas that presumably resulted from this movement, what now remains is mostly fragmentary manuscripts recovered from archaeological sites along the Silk Road and adjacent regions. These range in size from a large semi-intact Collection of Long Discourses (Dīrghāgama) from the Gilgit region of Pakistan down to fragments bearing no more than a few lines of text and known only by their catalogue numbers.Footnote 16 Large pieces of sutta material are also preserved in the remains of the Sanskrit Mūla-Sarvāstivāda Vinaya.Footnote 17
These various Sanskrit materials are probably later than, or at best contemporaneous with, the Indic source-texts from which some of the Chinese āgama translations were made. In the present study the main Sanskrit text utilised is the above-mentioned Collection of Long Discourses, hereafter denoted by “DAs”.Footnote 18 This text will now be briefly described.
DAs is believed to belong to either the Sarvāstivāda or the Mūla-Sarvāstivāda, the schools supposedly represented by the Sarvāstivāda Vinaya and Mūla-Sarvāstivāda Vinaya respectively. Because of continuing uncertainty about attributions to these two closely related schools, I adopt the common practice of referring to DAs and related āgama texts as belonging to the “[Mūla-]Sarvāstivāda”.Footnote 19 For typographic convenience, and with no intention of taking sides on this issue, I shall make it a practice, throughout the remainder of this book, to write “Sarvāstivāda” as shorthand for “[Mūla-]Sarvāstivāda”, except when referring to the two Vinayas (as in the previous chapter).
Despite the incomplete state of the DAs manuscript, the structure of this collection is well known, thanks to the presence in it of mnemonic verses (uddānas), which list the component suttas like a table of contents. Whereas the Pali DN has thirty-four suttas, DAs has forty-seven. In both texts the component suttas are grouped into three sections; however, a clear correspondence exists for just one of these three. The first section of DN, called the Section on Moral Discipline (Sīla-kkhandha-vagga) is matched by the third section of DAs, which bears essentially the same title (Śīlaskandha-nipāta). The DN version of this section comprises thirteen suttas (DN1–13), of which all but the first contain a detailed account of the stepwise training (anupubba-sikkhā). The Sanskrit version of this section (DAs25–47) contains parallels for those thirteen Pali suttas, together with a further ten suttas most of which are too poorly preserved for their contents to be known with certainty. Nevertheless, the Sanskrit account of the stepwise training contained in this section of DAs is well preserved and has been subjected to close analysis by various scholars.Footnote 20 It will be utilised in the next chapter. The remaining two sections of DAs contain a varied lot of discourses, half of which have their Pali parallels in MN rather than in DN.
For practical reasons this study makes no use of Sanskrit sutta quotations preserved in treatises such as the Abhidharmakośa.Footnote 21 Also neglected for practical reasons (chiefly, my own linguistic limitations) is the corpus of sutta quotations contained in Śamathadeva’s Abhidharmakośa-ṭīkā (or Upāyikā), originally in Sanskrit but now extant only in Tibetan translation.Footnote 22 The same applies for the suttas preserved in the Tibetan Kanjur as individual translations from Sanskrit originals.Footnote 23
Chinese Sources
From about the first century ce, when Buddhist pilgrims began travelling between India and China, many assorted Buddhist texts in various Indic languages found their way to China and were translated into Chinese. Most of them presumably arrived there as manuscripts, but some are reported to have made the journey in memorised oral form.Footnote 24 The selecting of the Indic source-texts for this purpose was a little haphazard, and the translating of them was done unsystematically, by many different individuals, over a period of about a thousand years.Footnote 25
The ever-growing corpus of Buddhist texts in Chinese to which this process contributed was preserved and transmitted within China, at first by hand-copying of individual texts and later – from the tenth century onward – by wood-block printing of the entire corpus. From early Chinese catalogues and other sources it is known that some of the translated texts had already been lost, either totally or partially, before the first full printed edition was produced.
Most of the resulting body of Chinese texts belongs to the Mahāyāna. However, there is also a substantial portion representing the so-called Hīnayāna. It includes four entire sutta collections comparable to the four main Pali nikāyas but representing three different schools and generally referred to as āgamas. Besides these there are three smaller (and apparently incomplete) sutta collections and many individual suttas paralleling ones already contained in the āgamas. The most systematic and authoritative edition of this corpus of āgama material is that contained in the first two volumes of the Taishō edition of the Chinese Tripiṭaka, which were published in Japan in 1924.Footnote 26
The four Chinese āgamas are usually referred to by the corresponding Sanskrit titles: Dīrghāgama, Madhyamāgama, Saṃyuktāgama, and Ekottar(ik)āgama (henceforth abbreviated as DAc, MA, SA, and EA). They correspond more or less closely to the Pali DN, MN, SN, and AN respectively. Basic information on these four Chinese collections – serial number in the Taishō edition, probable sectarian affiliation, and year of translation – is shown in Table 3.1. A remarkable fact thereby revealed is that the four translations were done within a period of little more than fifty years.Footnote 27
Table 3.1 The four āgamas in Chinese
Taishō number and title | School | Year ce |
|---|---|---|
T1 or DAc, Dīrghāgama | Dharmaguptaka | 412 |
T26 or MA, Madhyamāgama | Sarvāstivāda | 397 |
T99 or SA, Saṃyuktāgama | Sarvāstivāda | 435 |
T125 or EA, Ekottarikāgama | Mahāsāṅghika | 384 |
In editions earlier than the Taishō these four are found together as a discrete group within a section labelled “Hīnayāna Sūtras”, which in turn is located near the middle of the multi-volume Tripiṭaka.Footnote 28 In the Taishō edition, however, they are located in volume i (DAc and MA) and volume ii (SA and EA), where they are separated, each from the next, by other related texts, as will now be summarised.
Following the Dīrghāgama (DAc), which is T1, are translations of twenty-four individual suttas numbered T2 to T25. Each of these is demonstrably a parallel to one or another of the thirty suttas that make up DAc. For example, T2, T3, and T4 are parallels to DAc1, the first sutta of T1. Each of these four parallel versions (DAc1, T2, T3, T4) is found to correspond further to DN14, the Pali Mahāpadāna-sutta or “Great Lineage Discourse”.
Much the same situation exists with the other three āgama texts. The Madhyamāgama (MA or T26) is followed by seventy-two such individual translations (T27 to T98); the Saṃyuktāgama (SA or T99) is followed (but not immediately) by twenty-three (T102 to T124); and the Ekottarikāgama (EA or T125) is followed by twenty-six (T126 to T151).
Three further texts in Taishō volume ii remain to be introduced. T100, which immediately follows SA (T99), is titled “Other Saṃyuktāgama translation”. It is a translation of a sutta collection corresponding to roughly the last quarter of T99.Footnote 29 The next item again, T101, is a collection of twenty-seven suttas, possibly the translated remnants of an early, formerly complete, Saṃyuktāgama.Footnote 30 Finally, T150A, oddly inserted near the end of the group that follows EA, is a set of forty-seven suttas arranged numerically in the manner of EA and AN but limited mainly to the Threes and Fours.Footnote 31
These 4 complete āgamas, 3 incomplete sutta collections, and 145 individual suttas constitute the main corpus of Chinese sutta materials for this project. Each of the four complete āgama texts includes, at its beginning, information on the name of the translator (or perhaps of the translation team leader) and the period and location of the translation. Nothing is said about the language of the Indic source-text, the location from which that text was obtained, or the school with which it is thought to be affiliated. The three incomplete collections include no such information at all;Footnote 32 of the individual suttas, some provide no information while others provide “information” that is often unreliable.Footnote 33
For each of the four āgamas researchers have inferred the language of the source-text, with varying degrees of certainty, on the basis of phonetic transcriptions of Indic terms, proper names, and other such clues. Current scholarly opinion is that the source-text for the DAc translation was in the Gāndhārī language; that for MA was in some Prakrit, possibly again Gāndhārī; and the one used for SA was in Sanskrit. The language of the EA source-text remains uncertain but appears likely again to have been some Prakrit.
Identifying the school affiliation of each āgama amounts to linking it with one or another of the five extant Chinese Vinayas. This has been achieved mainly by considering various doctrinal details or other features characterising the different schools.Footnote 34 Current opinion identifies DAc with the Dharmaguptaka school, MA and SA with the Sarvāstivāda, and EA with some Mahāsāṅghika school.Footnote 35
Of the three incomplete sutta collections, T100 (“Other Saṃyuktāgama translation”) is demonstrably a close relative of T99 (i.e., of SA, the (nearly) complete Saṃyuktāgama translation); the two are descended from a near common ancestor within the Sarvāstivāda line.Footnote 36 At the same time, there is some evidence that, while the source-text for T99 was in Sanskrit, the source for T100 may have been in a Prakrit such as Gāndhārī.Footnote 37
T101 (partial Saṃyukta collection) and T150A (partial Ekottarika collection) have both been tentatively identified as Sarvāstivādin.Footnote 38
For the individual suttas, attempts to identify the original language and/or the school affiliation have yielded mixed results. Most worthy of mention here is the claim by Mizuno Kōgen that twenty-four of the seventy-four individual suttas associated with MA are the scattered remains of an earlier, formerly complete, translation of the same Middle-Length Collection done by Zhu Fonian – which, if Mizuno’s reasoning is soundly based, would also make them Sarvāstivādin.Footnote 39
Whereas the Pali nikāyas represent a single Buddhist school, the Chinese āgamas probably represent, though incompletely, three different schools. This sectarian diversity is advantageous for the present project, given the emphasis on external (inter-sectarian) comparison as a research method. To provide further background for such comparative work, I now give a brief account of each of the Chinese āgamas, together with some observations on how they match up with their Pali counterparts. The emphasis is on broad structural patterns of correspondence. The four āgamas are considered in the conventional Pali order, which is also the order of their appearance in the Taishō edition: DAc, MA, SA, EA.
The Collection of Long Discourses, DAc
Alongside the Pali DN and the Sarvāstivādin DAs, researchers have at their disposal a third version of the Collection of Long Discourses: the Dharmaguptaka version in Chinese (DAc or T1). This version comprises thirty suttas grouped in four sections, as against the Pali version’s thirty-four suttas in three sections. However, the fourth section of DAc consists of a single large text titled “Record of the World”, which has been aptly described as “a Buddhist purāṇa” and is in many ways totally out of place in this collection.Footnote 40 It appears to represent an accidental addition from some unknown source. As in DN and DAs, one of the (originally three) sections of DAc is composed almost entirely of suttas that incorporate a detailed account of the stepwise training. All but three of the twelve relevant DN suttas have parallels here.Footnote 41 In its remaining two sections DAc has parallels for all but three of the suttas in the corresponding sections of DN.Footnote 42
From this summary it will be apparent that there is a much higher degree of correspondence between DAc (Dharmaguptaka) and DN (Pali) than there is between either of these two and DAs (Sarvāstivāda). This correlates with the pattern of similarity and difference noted (in the previous chapter) in the corresponding Vinaya accounts of the First Council: the Pali and Dharmaguptaka are closely similar, while the Sarvastivāda is in a class by itself.
The Collection of Middle-Length Discourses, MA
For the Pali Majjhima-nikāya with its 152 suttas, the Chinese counterpart is the Sarvāstivādin Madhyamāgama (MA or T26), with 222 suttas.
The much larger size of MA relative to MN is related to the fact that 77 of the 222 MA suttas have their Pali parallels not in MN but rather in AN. This feature of MA is most conspicuously evident in its first decade (suttas MA1 to MA10), which is titled “Section with Sets of Seven”. All these ten suttas contain sets of seven items, and most of them have parallels in the Sevens of the Pali AN and/or the Sevens of EA (the Mahāsāṅghika Numerical Collection, discussed below).Footnote 43 Other smaller numerical groupings are found scattered throughout MA.
The Middle-Length Collection in Chinese (MA) belongs to the same school as the Long Collection in Sanskrit (DAs). This is indicated by the complementarity of these two collections as regards their sutta composition: no sutta in MA has a parallel in DAs, and vice versa.Footnote 44 In contrast to this situation, seven suttas in MA have parallels in DAc, a sure sign that MA and DAc belong to two different schools (Sarvāstivāda and Dharmaguptaka respectively).Footnote 45 In the Pali Sutta-piṭaka, complementarity similarly exists between MN and DN – though the pairing of the “Discourse on the Foundations of Mindfulness” (MN10) and its “Greater” counterpart (DN22) does constitute a partial exception.Footnote 46
The Collection of Connected Discourses, SA
The Sarvāstivādin Saṃyuktāgama (SA or T99) comprises 1362 suttas, according to the Taishō numbering.Footnote 47 Like the other āgamas, the SA text includes headings that indicate a division into “scrolls” or “fascicles”,Footnote 48 a reflection of the traditional Chinese manuscript format. In the case of SA the scrolls, of which there are fifty, deserve mention for two reasons. First, two of them, nos. 23 and 25, actually belong to an unrelated text, having apparently been inserted to fill gaps caused by loss of two of the original fifty scrolls early in the transmission of the translated text; that is, the extant SA is not quite complete.Footnote 49 Second, thirteen of the fifty scrolls are demonstrably out of sequence. Yet another shortcoming of SA is that it has virtually no headings to indicate the grouping of the suttas into saṃyuttas.Footnote 50 Both the disarrangement of the scroll sequence and the failure to mark the saṃyutta boundaries have been rectified by a succession of researchers, though this is not reflected in the Taishō edition.Footnote 51 In both cases this restoration of the lost information reveals that the version of SA whose Chinese translation was completed in 397 ce had a structure very different from the Pali SN. Counterparts to the individual saṃyuttas of SN can, in many cases, be identified; however, their sequence is totally different. This indicates that the two collections, though arranged on much the same principle (the saṃyutta principle), developed largely independently of each other.
The Collection of Numerical Discourses, EA
The Chinese Ekottarikāgama (EA or T125) is based on the same numerical principle as the Pali Aṅguttara-nikāya and, as in AN, its numerical sections (nipāta) run from Ones to Elevens. The eleven sections are not marked by headings in the text but are easily identified.Footnote 52
As regards both its sectarian affiliation and the language of the source-text, EA remains a source of academic uncertainty. An often-mentioned basis for regarding EA as Mahāsāṅghika is the occurrence of references to Mahāyāna concepts in some of its suttas. Perhaps more telling, however, is the fact that, despite sharing the numerical principle of arrangement with AN, as regards component suttas EA has a low level of commonality with AN. A relatively small proportion of the suttas of EA have parallels in AN.Footnote 53
The split between the Mahāsāṅghikas (claimed to be represented in EA) and the Sthaviras (represented in all of the other extant nikāyas/āgamas) was the first of the many splits that progressively fragmented the Buddhist Saṅgha. In view of the early separation into these two main Saṅgha branches, it is to be expected that a sutta collection on the Mahāsāṅghika branch and the corresponding collection on the Sthavira branch (i.e., EA and AN), should have less in common than any such pair of collections on sub-branches of the Sthavira branch (e.g., DAc and DN, or DAs and DN).Footnote 54 It follows that the observed low level of commonality between EA and AN is consistent with the claim that EA is Mahāsāṅghika.Footnote 55
In seeming contradiction to the above considerations, EA bears one striking resemblance to AN. Its first ten decades, which together account for the greater part of the Ones, closely resemble certain portions of the Ones of AN. They have the same artificial appearance, being self-evidently products of the same stratagem of subdividing formerly unitary discourses. For example, the first ten “suttas” of EA (EA2.1–10) correspond to ten consecutive “suttas” in the Ones of AN (AN1.20.93–102); both of these parallel decades are based on a set of ten “recollections” (anussati).Footnote 56 After the initial ten decades of artificial-looking suttas, the EA Section on Ones continues for a further three decades: EA12.1–14.10. Here, however, and hereafter to the end of the collection, the suttas have a completely “normal” appearance. Given this sharp line of demarcation, and given the patent artificiality of the initial block of ten decades, it appears likely that this block was a late addition to EA, reproduced from some now-lost close cousin of AN. This phenomenon is consistent with the proposition, justifiably advanced by some scholars, that the EA text includes materials drawn from several different sources.Footnote 57
From the above observation regarding the first ten decades of EA, it would follow that at some past time the collection began at sutta EA12.1, which is the “Discourse on the Foundations of Mindfulness”. This sutta earns its place at the beginning of the ones from its opening statement: “There is one entrance-way…”.Footnote 58
Abhidhamma Sources
The Chinese corpus includes partial reflections of the Pali Abhidhamma. Like the Pali tradition, the Sarvāstivāda has an Abhidhamma of seven books; one of those seven, called Dharmaskandha (“Dhamma-group”), bears a distinct resemblance to the Pali book called Vibhaṅga (Analysis).Footnote 59 The Vibhaṅga and Dharmaskandha have been shown, independently by both Yinshun and Frauwallner, to have enough in common to justify the inference that they are divergent derivatives of a single ancestral text.Footnote 60 And there is a third work, titled Śāriputra-Abhidharma and evidently belonging to the Dharmaguptaka school, that is also considered to be derived in part from this ancestral text. This lost common ancestor, dubbed “Original Vibhaṅga” by Bronkhorst,Footnote 61 therefore appears to belong to the period before the Sthavira tradition began to split up.
In addition to their early origin, the Vibhaṅga and its two sibling Abhidhamma texts in Chinese share a characteristic that makes them useful here. They are organised into sections according to doctrinal topics, and each such section begins with a quotation from a sutta that deals with the topic in question. For example, there is a section on the four foundations of mindfulness, and it begins with what is clearly meant to be a quotation from some version of the “Discourse on the Foundations of Mindfulness” (Satipaṭṭhāna-sutta). The wording of such quotations is sometimes not entirely identical with that of the apparent source suttas preserved in the Sutta-piṭaka. We therefore have, in these quotations, variant sutta versions that can be used for comparative work. Each section of the Vibhaṅga, Dharmaskandha, and Śāriputra-Abhidharma also contains Abhidhamma-style explanations, but as these clearly represent sectarian developments, they will be disregarded here. The sutta quotations, however, are deemed to fall within the corpus of relevant source material.Footnote 62
The Notion of Sutta Parallels
The above discussion of sutta materials in Pali, Sanskrit, and Chinese has entailed frequent reference to “parallels”, cases where a sutta in one collection (nikāya/āgama) “matches” a sutta in another such collection – or matches a sutta preserved outside of any known collection. The notion of Pali–Chinese sutta parallels is implicit in the three Pali–Chinese comparisons discussed in Chapter 1. The most elaborate of these compared the Pali “Discourse on the Foundations of Mindfulness” (MN10) with a discourse from the Sarvāstivādin Middle-Length Collection (MA98) and another from the Mahāsāṅghika Numerical Collection (EA12.1). This comparison showed how the three texts matched up, even with respect to details of no evident doctrinal importance.Footnote 63 It is natural, therefore, to conclude that these three texts derive from a single ancestral version of the discourse. In this book such parallel relationships are represented thus: MN10 = MA98 = EA12.1.
In his 1908 article “The four Buddhist āgamas in Chinese”, Anesaki Masaharu presents extensive tables of Chinese-to-Pali sutta correspondences covering all four āgamas. These have been superseded in Akanuma’s Comparative Catalogue of Chinese Āgamas and Pali Nikāyas (1929), which has the merit of referencing to the Taishō edition (though regrettably without using the Taishō sutta numbers for SA) and of including tables for parallels in the reverse direction, Pali to Chinese.Footnote 64 Akanuma’s Comparative Catalogue remains the principal reference work for identifying Pali–Chinese parallels, despite a significant number of errors and omissions. Improved updated alternatives are gradually becoming available.Footnote 65
In the case of large suttas the identification of parallels is usually straightforward and uncontroversial. Doubts and difficulties are more likely to arise with small suttas, most commonly in the Connected and Numerical collections. The smaller the sutta, the less the amount of content on which correspondence might be established, and the greater the number of similar suttas that need to be considered as possible parallels. In such cases differences of opinion may arise, or no parallel may be found at all.Footnote 66 The identification of Pali–Chinese sutta parallels is a work in progress. Consequently, some comparisons presented here require reappraisal of Akanuma’s correspondences.
The partly mixed nature of the Chinese āgama corpus means that there is often more than one Chinese parallel for a single Pali sutta. The “Discourse on the Foundations of Mindfulness” (just considered) is an example of this. Another is the “Discourse on Brahmā’s Net” (Brahmajāla-sutta), the first sutta in the Pali DN, which has Chinese parallels not only in sutta no. 21 of DAc but also in one of the twenty-four individually translated suttas that are located after DAc in the Taishō edition, namely T21. We therefore have this relationship: DN1 = DAc21 = T21.
If a sutta in one of the Pali nikāyas has a parallel within the Chinese āgamas, then that parallel is most often found in the corresponding āgama, though at a different location within it. For example, the Pali parallel to a sutta in the Saṃyuktāgama is likely to be in the Saṃyutta-nikāya. Sometimes, however, the Chinese parallel to a Pali sutta is located in a different āgama. For example, for the “Discourse on Fear and Dread” of the Pali Majjhima-nikāya (Bhayabherava-sutta, MN4) the only known Chinese parallel is located not in MA but in EA; thus: MN4 = EA31.1. That is to say, the various schools do not always agree regarding the distribution of the suttas among the four collections.Footnote 67
The “Great Discourse on Causation” (Mahānidāna-sutta) is remarkable in having a particularly large number of parallels. The Pali version, DN15, has one Chinese parallel in the Dharmaguptaka DAc13, another in the Sarvāstivādin MA97, and two more among the individual translations that are of uncertain affiliation: T14 and T52. This situation has the advantage of providing historical depth. Regardless of which school it represents, the T14 version of this discourse, which was translated by Ān Shìgāo in the second century ce, represents an earlier stage in the development of the doctrine of conditioned arising than does the T52 version, translated by Shīhù in the tenth century.
At the other extreme is the situation where a sutta has no known parallel. An example is SA320, for which no parallel has been found anywhere. This raises the suspicion that this sutta may be a late development unique to the Sarvāstivāda. That this is likely to be the case is indicated by its content, which is mainly on the distinctively Sarvāstivādin notion that “All exists” (Skt.: sarvam asti).Footnote 68 SA320 is likely to have developed within the Sarvāstivāda at a time after that school had separated from the Sthaviras.
The above examples show that identification of Pali–Chinese parallels can sometimes reveal how suttas have developed differently in the different traditions. They also show that the process of identifying parallels can sometimes be complicated and its outcomes open to question. A noteworthy point is that the process of identifying sutta parallels is not very different from the technique of comparing sutta parallels, which I have called “external comparison”. Identifying parallels emphasises points of agreement in order to justify the claim that, despite their differences, these are versions of “the same sutta”. The technique of external comparison emphasises differences in order to show how one or another of parallel suttas may have changed over time.
Issues Relating to Pali–Chinese Sutta Comparison
Comparing a Pali sutta with its Chinese parallel(s) usually entails, among other things, recognising how the doctrinal terms contained in the Chinese version match up with the doctrinal terms of the Pali version.Footnote 69 For example, it may entail recognising that the Chinese term 識 shì, which in everyday usage can mean “to know, recognise, discern; opinion, view”, and so on, is being used to represent the Pali/Sanskrit term viññāṇa/vijñāna, usually translated “consciousness”. What particular psychological function or state viññāṇa/vijñāna ultimately refers to is a “second-order question”, which need not intrude into the process. The primary task is to recognise which Indic term corresponds to which Chinese term. In effect, one is seeking to restore, as far as the case requires, the Indic text from which the Chinese translation was made. Some scholars have actually attempted to restore in their entirety the Sanskrit originals of certain Chinese āgama texts.Footnote 70 For the present purpose, however, whole translation is unnecessary, since identification requires recovery of just the skeleton of a text’s doctrinal message.
Identifying the right Pali counterpart for a Chinese translation term can sometimes be difficult. One problem is inconsistency from one translator to another. This is well illustrated in the following example from the above-mentioned four Chinese parallels to the Pali “Great Discourse on Causation” (Mahānidāna-sutta, DN15). In naming the links in the chain of Conditioned Arising the four translators generally agree with one another: yet in some instances they do not. For the seventh link, which in Pali is vedanā (feeling), and the ninth link, Pali upādāna (clinging), they employ the Chinese words shown in Table 3.2.
Table 3.2 Chinese translation terms for vedanā and upādāna
DN15 | T14 | MA97 | DAc13 | T52 |
|---|---|---|---|---|
7. vedanā (feeling) | 痛tòng | 覺jué | 受 shòu | 受 shòu |
9. upādāna (clinging) | 受shòu | 受shòu | 取 qǔ | 取 qǔ |
Note: The translators are: for T14, An Shigao (second century); for MA, Saṅghadeva (397 ce); for DAc, Buddhayaśas (412 ce); and for T52, Dānapāla (tenth century). The locations are: T14\Ti242a14–16, 243a19–22; MA97\Ti578b27–c03, 579b22–25; DA13\Ti60b25–29; T52\Ti844c08, 845a13–18.
Here the Chinese word 受shòu (“receive, experience”) is intertextually ambiguous: it represents vedanā in two of the translations and upādāna in the other two. The researcher needs to allow for this when comparing the four Chinese texts with the Pali. The task is to discover which Pali term the Chinese word corresponds to, given the immediate context and what is known of the particular translator’s conventions. It is not just a matter of looking up the Chinese word in a dictionary.
The further one looks, the greater this problem becomes. One of the three Chinese counterparts for vedanā shown in Table 3.2 is the word 覺 jué. Elsewhere in works by the same translator (Saṅghadeva) this word corresponds to Pali vitakka (“initial thought”, as a factor in jhāna, concentrative meditation) or it may correspond to bodhi (“wisdom, awakening”). Within his own translation corpus, a translator is usually consistent in how he represents any given Indic term; for example, one can be sure that vedanā and upādāna will be represented by 覺 jué and 受 shòu respectively throughout MA. However, the converse is not true. As shown by the case of 覺 jué, even within one āgama a given Chinese word may correspond to several different Pali/Sanskrit terms.Footnote 71 Fortunately, such cases can usually be resolved by considering the context.
Consistency in a translator’s rendering of Indic terms is generally to be recognised as a virtue, but it can actually be a hindrance to comprehension if the Indic term is itself ambiguous. An example is the multi-purpose Pali/Sanskrit term dhamma/dharma. In Chinese this word is consistently translated as 法fǎ “law, method, nature”, even in situations where it is functioning as an adjective-forming suffix. For example, the combination老法 lǎo-fǎ, which looks like “old law”, is more likely to mean “subject to aging”, since it corresponds to Pali jarā-dhamma.
Sometimes the Chinese translators have evidently made errors in interpreting the Indic text. For example, SA contains frequent references to “another monk” (異比丘 yì bǐqiū) in contexts that clearly require the meaning “a certain monk”. The explanation is likely to be that the translator misread Sanskrit anyatara, “a certain”, as anyatra, “another”.
In some cases an obvious error is attributable not to mistranslation but rather to subsequent miscopying of the translated Chinese text. An exceptionally frequent example of this is confusion of the character 想xiǎng (saññā, perception) with the formally and phonetically similar character 相xiàng (nimitta, appearance) – an error that naturally has serious semantic consequences.Footnote 72
Another, more localised example of this phenomenon is found in the description of the third jhāna (stage of concentration), as it appears in a dozen suttas of MA. In the Pali the description includes the passage upekkhako satimā sukhavihārī ti, and in most occurrences in the Taishō MA this appears, reasonably correctly, as a Chinese phrase meaning “equanimous, mindful, a pleasant abode”. However, in a number of cases the meaning represented is instead “equanimous, mindful, a pleasant abode, empty”. Given the emphasis on the notion of “emptiness” in some traditions, this looks like an interesting doctrinal difference. However, close checking, supported by the Taishō critical apparatus, reveals that this discrepancy has come about as follows.
The Indic phrase “a pleasant abode” was competently rendered by the MA translators as 樂 住 室 lè zhù shì, literally “pleasure-reside-dwelling”, and this reading has been preserved in a number of cases, not only in the Taishō edition but also in the earlier “three editions” (Song, Yuan, and Ming). Often, however, the third character 室 shì was miscopied as the superficially similar 空 kōng, which means “empty”, thus yielding “pleasure-reside-empty”.Footnote 73
Although fundamentally different, translation errors and copyist errors call for much the same treatment by the researcher: one can use them to infer the Indic original and to discern how the presumed error could have occurred.
The Chinese Scriptural Transmission
When reading the Romanised PTS edition of the Pali Sutta-piṭaka, or the Taishō edition of its Chinese counterpart, one can easily forget how different the situation would have been a millennium ago for people reading the same suttas from rare and partly damaged manuscripts. Let us look first at the Chinese situation.
In the history of the Chinese Tripiṭaka the production of the first xylograph edition (the Kaibao edition of 983 ce) was a crucial watershed. Beginning from that time multiple uniform copies of the entire known corpus would have become widely available. Also, subsequent editions (numbering, up to the present, two dozen or more) could be based upon it, in a continuing cumulative process of filling gaps and resolving discrepancies. From editors’ notes and other sources one learns that certain portions of the āgama text that were missing from the first edition (Kaibao) and/or the second (Qidan) were inserted in later editions. This process made up for earlier loss, though only partially and sometimes very imperfectly.
For the Sutta-piṭaka the best examples of this phenomenon are to be found, not surprisingly, in the three incomplete āgamas, T100, T101, and T150A. In the case of the incomplete “Other Saṃyuktāgama Translation” (the present T100) the compiler of the second Korean edition (completed in 1251 ce) tells us, in notes still preserved in the Taishō edition, how he dealt with substantial gaps in the source-texts he was using.Footnote 74 He found that the Kaibao version of this text lacked the Brahmā-saṃyutta, which however was present in the Qidan version;Footnote 75 he also found that the Qidan version lacked a block of eleven suttas from within the Brāhmaṇa-saṃyutta, which was present in the Kaibao.Footnote 76 He therefore copied the Brahmā-saṃyutta from the Qidan into the Kaibao, thereby producing a more complete version of the “Other Saṃyuktāgama Translation” for the new Tripiṭaka edition he was compiling. It is this augmented, but still incomplete, version that is preserved in the present T100.Footnote 77
In the case of the Incomplete Numerical Collection (the present T150A), Harrison (Reference Harrison, Kieffer-Pülz and Hartmann1997) demonstrates that what is now the last section of this collection (T150A suttas 31–47) was missing from the Kaibao and Qidan editions but present in the later “three editions”. This indicates that, at some time between the Qidan and the earliest of the three editions, the missing portion had been found (presumably in some old manuscript) and reunited with the rest of the T150A collection. Unfortunately, the editors responsible simply joined the two pieces head to tail, not recognising that the newly found piece properly belonged near the middle of the collection rather than at the end.Footnote 78 It was left to present-day scholars to sort out the resulting confusion.
Again, in the case of the anthology of twenty-seven SA discourses (now known as “T101”), other findings by Harrison (Reference Harrison2002), based mainly on relevant catalogues, indicate that the suttas of this collection were formerly arranged in a sequence rather different from the present one. It appears that, at some early time, the manuscript containing the anthology was accidentally divided into pieces, that were later re-joined – but they were re-joined in the wrong sequence, one sutta being lost in the process and three others added.Footnote 79 As Harrison observes, the fact that this could happen indicates that the tenuous line of descent of this text “passed at one point through a single manuscript copy”.Footnote 80
The four āgamas fared better but did not entirely escape such mishaps. In the case of SA, the lost scrolls numbered 23 and 25 (out of fifty) were never found.Footnote 81 This suggests that only a few copies of the SA translation existed.Footnote 82 Already by the tenth century, five centuries after this translation was produced, a complete copy of it could not be found, or pieced together, by the compilers of the Kaibao edition or of any subsequent one.
Prior to the production of that first printed edition individual manuscripts of individual suttas or sutta collections were, no doubt, held in scattered monastery libraries, whose collective holdings came to be listed in catalogues from time to time.Footnote 83 It was during this period (up to the tenth century) that rearrangement or partial loss of sutta collections would have occurred.
In short, for the first five centuries following the translation of the āgamas into Chinese, the preservation and transmission of the resulting texts were precarious and uncertain. Already in that period some Chinese Buddhist monks had become concerned that the Dhamma transmission might die out. The ultimate “Ending of the Dhamma” (mofa), said to have been foretold by the Buddha, was of particular concern to a certain seventh-century monk named Jingwan. To ensure the Dhamma’s survival in China, Jingwan initiated a project to have the principal Buddhist scriptures carved in stone inside a group of caves associated with his temple.Footnote 84 His project, funded by a succession of devout lay supporters, continued down to the thirteenth century, by which time the growing corpus of “stone suttas” had come to include all four āgamas. These carved āgama copies are not, however, of any value for this study of early Buddhism, because all of them date from later than the production of the Kaibao edition. In keeping with the general trend in China, Jingwan had given priority to texts of the Dasheng or “Great Vehicle” (Mahāyāna). The texts of the less popular Xiaosheng or “Small Vehicle” (Hīnayāna) had to wait their turn. To judge from their dates, the carved āgamas were very probably based on some block-printed edition.
Turning attention now from the collections as units to the details of their contents, let us consider the widely held perception that, as regards preservation of the wording of the Teaching, EA is significantly less “reliable” than the other Chinese āgamas. This perception is based on tangible evidence of internal inconsistency. The most striking example is variation in listings of the noble eightfold path. In the entire EA the eight path stages are listed in twenty-three different suttas, and at most of these occurrences they are in the familiar sequence: right view, thought, speech, action, livelihood, effort, mindfulness, and concentration. However, on seven occasions they are listed in one or another of five different sequences – for example, with right action and right livelihood interchanged.Footnote 85
Are these five genuinely different formulations of the eightfold path? Or do they represent errors introduced during hand-copying of the Chinese text – or perhaps even in producing the Chinese translation in the first place? The preface to EA describes how the translation team worked. Dharmanandi recited the source-text in the original language as he recalled it; Zhu Fonian put it orally into Chinese; and a scribe wrote down Zhu Fonian’s words. This conversion of the message from oral recitation in the source language to oral dictation in the target language would clearly have been conducive to error.
It appears that Zhu Fonian and his team employed a similar procedure when working on the first MA translation (later largely lost); that is, they relied on an orally transmitted version. The team led by Saṅghadeva, which produced the second MA translation (preserved as T26), worked from a manuscript source-text, surely a much more reliable method. Nevertheless, it remains the case that the translation of a sutta text, whether oral or written, from an Indic language into the totally dissimilar Chinese is likely to have been a particularly weak link in the chain of transmission – a point at which the Dhamma message was in danger of becoming distorted.
The Pali Scriptural Transmission
Moving back now from the Chinese transmission to the Pali, one sees a very different set of circumstances. Once written down (in Lanka in the first century bce), the suttas are likely to have changed much more slowly than before, but change they almost certainly did. To inhibit further change and to resolve existing discrepancies were probably among the objectives of the commentarial work undertaken by Buddhaghosa five centuries later. The learned monk’s authoritative Pali commentaries, based on earlier ones in Sinhala, would have determined, in each case of doubt, which reading was the “correct” one – correct, at least, according to the orthodoxy of the Mahāvihāra, the large, influential monastery that supported Buddhaghosa and his work. The effect would have been to freeze the Pali canon in that authorised form.Footnote 86 It is noteworthy that Buddhaghosa was virtually a contemporary of the translators who produced the Chinese versions of the āgamas: the Pali nikāyas and the Chinese āgamas were frozen at much the same time.
Despite Buddhaghosa’s efforts, the texts of the Pali canon did continue to change as the centuries passed. This is apparent from the critical apparatus to the PTS editions, which notes numerous variant readings found in source manuscripts in a range of regional scripts such as Sinhalese, Burmese, and Siamese. Many of these variants are non-significant orthographic alternatives. Some, however, represent substantial errors that distort the message. Even in modern printed editions in regional scripts such errors are not uncommon.Footnote 87
In recent times a move to eliminate all such anomalies from the Pali Tipiṭaka was initiated with the convening of the so-called Fifth Council in Mandalay, Myanmar, in 1871 to celebrate the carving of a standardised version of the entire text in Burmese script on 729 marble slabs in the Kuthodaw Pagoda.Footnote 88 The process was subsequently internationalised at the Sixth Council, held in Yangon from 1954 to 1956, at which the participating monks are said to have come from eight different countries.Footnote 89 Tangible current outcomes of it are the World Tipiṭaka Edition (accessible on-line),Footnote 90 as well as the Vipassanā Research Institute’s Chaṭṭha Saṅgāyana CD-ROM, both containing the revised standardised version and making no mention of variant readings.
For research projects such as the present one, the textual uniformity thus achieved is not at all advantageous. For example, in these standardised editions one finds that the “Discourse on the Foundations of Mindfulness” in MN has been replaced by a replica of the “Greater Discourse on the Foundations of Mindfulness” from DN. That is, the old MN10 is lost, and the new MN10 is identical with DN22.Footnote 91 The effect is to conceal important data on the historical development of the teachings on mindfulness. The change is instructive, however, given that a perceived need to standardise and unify is likely to have persisted throughout the entire history of that vaguely defined entity called “the Pali Canon”.Footnote 92
The texts of both the Pali and the Chinese “canons” are now freely available in electronic form. This development has greatly facilitated research into early Buddhism, and the present project has benefited from it. However, the new technology brings with it new uncertainties. In addition to the obvious likelihood of a new generation of copyist errors associated with the process of data entry, there exists the possibility that over-enthusiastic editors have sought to “improve” the text in various subtle ways. Evidence of both these forms of textual corruption occasionally presents itself. This means that for serious work on the scriptural transmission the old block-printed or hand-copied versions remain indispensable.
This concludes the first part of this book, providing “Background and Context” for the analysis that follows.
