The Structure of the Sagātha-Vagga of the Samyutta Nikaya.pdf

Buddhist Studies Review 24(1) 2007, 7–34

ISSN (print): 0256–2897

doi: 10.1558/bsrv.v24i1.7

ISSN (online): 1747–9681

The Structure of the Sagātha-Vagga

of the Saṃyutta-Nikāya

RODERICK S. BUCKNELL

School of History, Philosophy, Religion, and Classics, The University of Queensland

rodbucknell@yahoo.com

ABSTRACT: The meaning of its title, ‘Section with Verses’, may appear suﬃ cient to

explain why the Sagātha-vagga was identi ﬁ ed as a discrete entity within the Saṃyutta-

nikāya . However, this article looks beyond that simple explanation, to discover whether

some other rationale may underlie this grouping of saṃyutta s. It examines evidence that

the compiling of the Sagātha-vagga was probably based on a familiar, although doctrinally

marginal, piece of Buddhist teaching, namely the ‘eight Assemblies’.

The Sagātha-vagga (Section with Verses), the ﬁ rst of the ﬁ ve vagga s 1 of the Pāli

Saṃyutta-nikāya (SN), comprises 271 short sutta s grouped according to topic into

eleven saṃyutta s. These eleven topics/ saṃyutta s are: Devatā s (gods), Devaputta s

(sons of gods), Kosala (Pasenadi’s kingdom), Māra (the Evil One), Bhikkhunī s (nuns),

Brahmā s (higher gods, who tend to think of themselves as creators), Brāhmaṇa s

(priests), Vaṅgīsa (a senior monk), Vana s (forest spirits), Yakkha s ( ﬁ erce spirits) and

Sakka (Indra, king of the gods). The Devatā-saṃyutta is exceptionally large, with

eighty-one sutta s; the remaining saṃyutta s range in size from ten sutta s to thirty. 2

Stylistically, the Sagātha-vagga is a very natural grouping; every one of its com-

ponent sutta s contains at least one gāthā (piece of verse), usually embedded within

a prose framework. 3 This stylistic uniformity may appear suﬃ cient to explain

why these eleven saṃyutta s were brought together as a discrete section within

SN. Yet this seemingly reasonable explanation for the compiling of the Sagātha-

vagga is problematic. The ﬁ rst problem is that the sagāthā form (verses embedded

1. Here and henceforth the term vagga refers to a grouping of about ten saṃyutta s, not ten sutta s.

The ambiguity of the term vagga is mentioned by Feer (SN I viii) and Bodhi (2000, 22). In my

references ‘SN I’ denotes Léon Feer’s edition of the ﬁ rst vagga of the Saṃyutta-nikāya (PTS, 1884;

repr. 1991). The new edition by G. A. Somaratne (PTS, 1998) does not diﬀ er signi ﬁ cantly for the

purposes of this discussion of the vagga ’s broad structure. Useful background on SN is provided

by von Hinüber (1996, 35–8) and Norman (1983, 49–54).

2. See tables by Feer (SN I ix–x), and by Bodhi (2000, 24).

3. At MN-a II 106 and Vin-a I 28, Buddhaghosa identi ﬁ es the Sagātha-vagga with geyya-aṅga , the

second of the nine recognised classes of text ( navaṅga-dhamma ; e.g. MN I 133–4 ~ T I 764a 25 – 6

and AN II 103 ~ T II 635a 17–19 ).

BUDDHIST STUDIES REVIEW

in a prose framework) is common elsewhere in the Sutta-piṭaka . The 112 sutta s of

the Itivuttaka are all in sagāthā form; 4 the Aṅguttara-nikāya contains no fewer than

174 sutta s in sagāthā form, seventy of them conspicuously grouped at the begin-

ning of the Fours (AN II 1–76); and within SN itself the Bhikkhu-saṃyutta , located

not in the Sagātha-vagga but in the Nidāna-vagga (SN II 273–85), is predominantly

(ten sutta s out of twelve) in sagāthā form.

Conversely, there are saṃyutta s in the Sagātha-vagga whose presence there

seems super ﬂ uous. All but two of the ﬁ fty-one gāthā s in the Vaṅgīsa-saṃyutta

are also present in the Theragāthā , and in precisely the same sequence. 5 Also, the

gāthā s in the ﬁ rst eight of the ten sutta s that make up the Bhikkhunī-saṃyutta rep-

licate ones found in the Therīgāthā , although sometimes imperfectly and in a dif-

ferent sequence. 6 It appears likely that at least these two saṃyutta s were created

intentionally out of existing material by providing each of the selected verses

with a narrative introduction-commentary. 7

In short, a large amount of material in sagāthā form was excluded from the

Sagātha-vagga , while other material in sagāthā form has the appearance of having

been assembled into saṃyutta s speci ﬁ cally for inclusion in this vagga . Such facts

indicate that this literary form does not, by itself, account fully for the identity of

the Sagātha-vagga as a discrete entity within SN. Accordingly, this article has the

aim of discovering whether some other rationale may underlie the grouping of

these eleven saṃyutta s ( Devatā to Sakka ) into a recognized vagga . 8 It will be argued,

on the basis of a wide range of evidence, that the existing Sagātha-vagga derives

from an earlier collection whose structure was based on a familiar, although doc-

trinally marginal, piece of Buddhist teaching.

SOURCES AND METHOD

The principal method employed here is comparison of the Pāli Sagātha-vagga with

its two counterparts in the Chinese Tripiṭaka , contained in Taishō no. 99 and no.

100 (hereafter written ‘T99’ and ‘T100’). 9

4. The stylistic closeness of the Itivuttaka to the Sagātha-vagga is noted by von Hinüber (1996, 47).

5 . Theragāthā verses 1209–57, 1261–2. For the detailed correspondences see Bodhi (2000, 1978); also

Ishigami (1966, 245–9), which, despite its title, has little in common with the present article.

6. Versesin Bhikkhunī-saṃyutta sutta s 1–8 correspond to Therīgāthā verses 57–8, 60–61, 59 & 142,

139–40, 230–33, 191, 197–8 & 200–201, 183–5 respectively, according to Bodhi (2000, 1976) and

(with slight variations) Ishigami (1966, 231–3).

7. This is how it is put by Feer (SN I xvi) and by Bodhi (2000, 70). A similar process of adding prose

introductions to existing verses is suggested by von Hinüber (1996, 46) for the Udāna .

8. I am not aware of any previous attempt to do this. A good foundation for such a study is pro-

vided by Yìnshùn (1983, I xxxi–xxxii), and I rely heavily on it in this article.

9 T99 is at T II 1–373; T100 is at T II 374–492. For a brief overview of these, see Bucknell (2006).

Numerous portions of the Sagātha-vagga are preserved in Sanskrit (e.g. Enomoto 1994); however,

they are too fragmentary and scattered to be of use in this discussion of structure. The entire

Vana-saṃyutta , together with a few other scattered sutta s, has been identi ﬁ ed in Gāndhāran

BUCKNELL THE STRUCTURE OF THE SAGĀTHA-VAGGA OF THE SAṂYUTTA-NIKĀYA

T99, titled 雜阿含經 ( Zá āhán jīng ) ‘Diverse Āgama ’, comprises 1362 sutta s, 10 of

which about 1050 correspond to sutta s in the Pāli Sutta-piṭaka , most of them in

SN. 11 T99 is recognized as a translation of a lost Sanskrit Saṃyuktāgama , belonging

to either the Sarvāstivāda or the Mūlasarvāstivāda. 12 A statement following the

Chinese title attributes the translation to an Indian monk, Guṇabhadra, during

the (Liú) Sòng dynasty (420–79 CE). 13 Whereas the Pāli Sagātha-vagga is located at

the beginning of SN, the corresponding section of T99, comprising 309 sutta s, is

located at the end of the Zá āhán jīng . 14

T100, titled 別譯雜阿含經 ( Bié yì Zá āhán jīng ) ‘Other Translation of the Diverse

Āgama ’, was done by an unknown translator and its sectarian aﬃ nities are unclear. 15

The text appears to be incomplete; it contains only 364 sutta s, corresponding to

about the last quarter of T99. The greater part of it (280 sutta s) corresponds closely

to the Sagātha-vagga of T99 and less closely to the Pāli Sagātha-vagga . 16

remains (Mark Allon, personal communication). Translations into Western languages exist for

only a few sutta s or gāthā s of T99 and T100, and for some Sanskrit fragments. See especially

the English translation of the Bhikkhu-saṃyutta of T100 (scroll 1 = sutta s 1–22) by Bingenheimer

(2006), which is the ﬁ rst installment of a planned full translation of T100.

10. This ﬁ gure (also that for T100, below) is as indicated by the numbering of the sutta s in the

Taishō edition. The actual number of sutta s is unclear, as it is also for SN (see Saigusa 1978,

613–69). In the present article, glosses of Chinese terms and section titles, if not in English,

are usually given in Pāli; e.g. sutta . This is done for ease of comparison with the Pāli texts, and

despite the fact that the language of the source text for each of the Chinese versions is likely to

have been some form of Sanskrit. Transcriptions of Chinese are in Hànyǔ pīnyīn .

11 . Sutt a correspondences between the Chinese and Pāli versions of Sagātha-vagga are set out in

Anesaki (1905, 34–7); Akanuma (1929, 31–2, 62–4, 90–92, 94–100, 102–19, 172–91, 204); Taishō

(1924–34, supp. vol. 1, 166–7, 170–71, 174–9); and Fóguāng (1983, IV, 53–72). Only Taishō and

Fóguāng use the Taishō sutta numbers. Only Anesaki (1905) has a table of Chinese-to-Pāli cor-

respondences for T100: his ‘Text β’, which, however, follows the ‘Chinese arrangement’ dis-

cussed below. Only Akanuma gives Pāli-to-Chinese correspondences. And only Anesaki (1905)

and Fóguāng recognize the restored scroll sequence for T99 discussed below.

12. For scholarly opinion on the likely background of the source text, see the summary provided

by Glass (2006, 22–5). For the Sarvāstivāda attribution, see Mayeda (1985); for the Mūla-

Sarvāstivāda, see Enomoto (1980; 1986, 23).

13. T II 1a2. Recent research dates the translation to 435–6 CE and raises questions about

Guṇabhadra’s role in the translation team; see the review by Glass (2006, 7, 20–25).

14. In T99 the Sagātha-vagga comprises the sutta s numbered 88–102, 576–603, 995–1022, 1062–120,

1145–63, 1178–240, 1267–362. This discontinuous distribution, seemingly at odds with the

stated location of the vagga ‘at the end’, is discussed below.

15. A summary of the various published opinions is oﬀ ered by Bingenheimer (2006, 22), who him-

self supports attribution to the (Mūla-)Sarvāstivāda.

16. In T100 the Sagātha-vagga comprises sutta s 1–110, 132–42, 161–89, 214–329, 351–64. A few

Sagātha-vagga sutta s contained in T99 are not represented in T100, and vice versa. In the

extreme case, of the twenty sutta s making up the Māra-saṃyutta in T99 (nos. 1084–103) only the

ﬁ rst ten have counterparts in T100 (nos. 23–32). Within short sections of text, T100 agrees very

closely with T99 as regards the sutta sequence; see Bingenheimer (2006, 23–5, table). Another

incomplete Chinese translation of a Saṃyuktāgama , Taishō no. 101 at T II 493–8, contains only

twenty-seven sutta s, eight of them (nos. 1–5, 21, 25, 26) from the Sagātha-vagga , so cannot be

made use of in this study.

BUDDHIST STUDIES REVIEW

Also useful as sources for this study are the Mūlasarvāstivāda-vinaya and the

Yogācārabhūmi-śāstra . The Mūlasarvāstivāda-vinaya contains a list of the vag-

ga s of the Saṃyuktāgama as known to the Mūlasarvāstivādin tradition (T XXIV

407b 21–8 ). As for the Yogācārabhūmi-śāstra , one of its ﬁ ve divisions, the Vastu-

saṅgrahaṇī , is a partial commentary on a text that clearly resembled fairly closely

the Saṃyuktāgama version preserved in Chinese as T99. This commentary does

not cover the Sagātha-vagga but is of value here in containing a further listing of

the vagga s of the Saṃyuktāgama . 17

T99, the complete Saṃyuktāgama translation by Guṇabhadra, contains only

fragmentary indications of a division into vagga s, and none for saṃyutta s. Instead,

it is divided mechanically into ﬁ fty equal-sized ‘scrolls’ ( 卷 juàn ), a purely Chinese

development. 18 It has long been known that the existing text is to some extent

in disarray; some of the ﬁ fty scrolls must have been accidentally interchanged. 19

In the restored ‘correct’ sequence, the section of T99 that contains the Sagātha-

vagga comprises the thirteen scrolls that now bear the numbers 38, 39, 40, 46, 42,

4, 44, 45, 36, 22, 48, 49, 50. 20

Although most of the vagga titles (and subtitles) have been lost from T99, it is

clear from those that remain, and from indications given in the Yogācārabhūmi-

śāstra and the Mūlasarvāstivāda-vinaya , that the text translated by Guṇabhadra

was divided into the following seven vagga s:

1. Khandha-vagga (Section on the Aggregates)

2. Saḷāyatana-vagga (Section on the Sense-bases)

3. Nidāna-vagga (Section on Causation)

4. Sāvakabhāsita-vagga (Section Spoken by Disciples)

5. Magga-vagga (Section on the Path)

6. Buddhabhāsita-vagga (Section Spoken by the Buddha)

7. Sagātha-vagga (Section with Verses) 21

17. The Chinese translation of Vastu-saṅgrahaṇī is at T XXX 772–882, the list being at 772c 11–15 ; the

Tibetan is Peking no. 5540 in vol. 111, 121–218, the list being at page 121, folio 143b 1–5 .

18. Although now printed in modern book format, the Chinese canon continues to indicate the old

division into scrolls (often called ‘fascicles’).

19. As demonstrated by Anesaki (1908); Lǚ (1923); Mayeda (1964, 654–7); Yìnshùn (1983, I i–lxxiv,

esp. xli–liii); Mukai (1985); and others. For a summary, see Glass (2006, 25–30).

20. This explains the discontinuous distribution mentioned in n.14, above. According to the most

recent research of Yìnshùn (1983, I xlv–xlix) and Mukai (1985, 18), the correct scroll sequence

for the entire T99 is: 1, 10, 3, 2, 5–9, 43, 11, 13, 12, 14–21, [23], 31, 24, [25], 26–30, 41, 32–5, 47,

37–40, 46, 42, 4, 44, 45, 36, 22, 48–50. Scrolls 23 and 25 belong not to the Saṃyuktāgama but to

the unrelated Aśokāvadāna , apparently having been used inappropriately to ﬁ ll gaps created

by accidental loss of two scrolls. The inferred original sequence of the scrolls is adopted in the

Fóguāng edition of 1983.

21. A heading ‘2. Saḷāyatana-vagga ’ appears in the Yuán and Míng editions of the Chinese Tripiṭaka ,

although not in the Taishō edition (see n.8 to T II 49b 3 ). The Nidāna-vagga of the Saṃyuktāgama

is identi ﬁ ed by title at the beginnings of only the 4th and 5th of its ﬁ ve scrolls: ‘3. Nidāna-

vagga , Part 4’ and ‘3. Nidāna-vagga , Part 5’ (T II 108c 27 , 116c 9 ). The Saṃyuktāgama counterpart

BUCKNELL THE STRUCTURE OF THE SAGĀTHA-VAGGA OF THE SAṂYUTTA-NIKĀYA

The Sagātha-vagga is the last of these seven. In the Yogācārabhūmi-śāstra it is

actually called by a diﬀ erent title, discussed below. Four of the seven vagga titles

in the above list, namely Khandha , Saḷāyatana , Nidāna , and Sagātha , match with

titles in the Pāli SN. A ﬁ fth, Magga-vagga , clearly corresponds (on the basis of the

vagga ’s content) to the Pāli title, Mahā-vagga . The Pāli SN lacks counterparts for

the two remaining titles, Sāvakabhāsita-vagga (Section Spoken by Disciples) and

Buddhabhāsita-vagga or Tathāgatabhāsita-vagga (Section Spoken by the Buddha). 22

Most of the saṃyutta s that make up these two extra vagga s in T99 do exist in SN

but are scattered throughout its second to ﬁ fth vagga s. The Sāvakabhāsita-vagga

comprises six saṃyutta s whose component sutta s are spoken by disciples ( sāvaka )

rather than by the Buddha, and which therefore form a natural group. 23 The

Buddhabhāsita-vagga comprises about nine saṃyutta s 24 whose component sutta s

are spoken by the Buddha but are for some reason set apart from the remaining

vagga s.

In the case of T100, the incomplete ‘Other Translation’ of the Saṃyuktāgama ,

an even more serious disarrangement of the text is known to have occurred. In

this respect the Taishō edition of the Tripiṭaka (compiled by Japanese scholars

in the 1920s), along with the thirteenth-century Korean edition on which it is

directly based, disagrees with some at least of the editions produced in China. 25

of the Pāli Mahā-vagga is marked at its beginning: ‘5. Magga-vagga , Part 1’ (T II 170c 27 ). Between

Nidāna- and Magga-vagga s is a heading ‘4. Sāvakabhāsita-vagga ’ (T II 126a 3 ). (These ﬁ ve extant

vagga headings are cited by Anesaki [1908, 70].) This scarcity of vagga titles helps explain how

T99 could become disarranged: not one of the twelve transposed scrolls bears a title. The pres-

ent scroll numbers on these twelve are taken to be secondary additions; the sutta numbers in

T99 date only from the compiling of the Taishō edition (1924). There exist variations in the

sequence of listing the seven vagga s. The Vinaya list, at T XXIV 407b 21–8 , has Buddhabhāsita

before Magga rather than after it. The Chinese Yogācārabhūmi-śāstra list, at T XXX 772c 11–15 , has

Buddhabhāsita and Sāvakabhāsita at the beginning, ahead of Khandha ; but the Tibetan, at Peking

vol. 111, p. 121, f. 143b 1–5 , agrees with the Vinaya sequence.

22. Their distribution is shown in Mukai (1985, 18). A Sanskrit Saṃyuktāgama fragment described

by Hosoda (1989, 541) has the word Buddhabhāṣita as a ‘running header’ on the folio. It is

unclear whether there is any connection with the terminology of the Pāli Vinaya (Vin IV 15, 9–

10 ), where the Dhamma is equated with four categories: Buddhabhāsita, Sāvakabhāsita, Isibhāsita ,

and Devatābhāsita .

23. The six are: Sāriputta, Moggallāna, Anuruddha, Mahākaccāna, Ānanda, Citta .

24. The saṃyutta s are Mahākassapa, Gāmaṇī, Anamatagga , and others. The count of nine is approxi-

mate because of uncertainty regarding the boundaries of a few seeming saṃyutta s that lack evi-

dent counterparts in SN. The reason these saṃyutta s are grouped as a separate vagga is perhaps

that they mostly do not deal with speci ﬁ c Dhamma topics. The title, Buddhabhāsita , ‘Spoken by

the Buddha’, is odd, since it seems equally applicable to all vagga s other than Sāvakabhāsita-

vagga .

25. Just which editions have the Taishō arrangement and which the ‘Chinese’ arrangement is a

question for future research. Editions having the Chinese arrangement include those num-

bered 29, 35, 39, 43 in the list by Grönbold (1984, 24–5). (The Taishō edition and its Korean ante-

cedent are Grönbold’s nos. 49 and 32, respectively.) How T100 corresponds with the Chinese

arrangement is shown in tabular form at Fójiào Dàzàngjīng XXV 867–8 (Grönbold’s no. 48). How

the Sagātha-vagga section of T100 corresponds with that of the reconstituted T99, and with the

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: