The Pali Saṃyuttanikāya, rather like the Aṅguttaranikāya, is an attempt to make a systematic classification of a large, but indeterminate, number of often quite brief texts, culled to some extent from older material, and artificially multiplied. Léon Feer's edition of Saṃyuttanikāya appeared in five volumes, 1884–98, with indexes by Mrs Rhys Davids, 1904. A less than satisfactory re-edition of Book I appeared in 1998, using fresh manuscript evidence. A corrected version of Book V has now been typeset by William Pruitt in consultation with K. R. Norman. A preface describes it twice, inadvertently, as a corrected reprint of Book I, but one of its comments refers indeed to a passage in Book V (the page reference given for its other comment is evidently misprinted).
The pagination of the original has wisely been retained, but a superfluous and somewhat inconsistent attempt to avoid splitting words has often resulted in unattractive gaps between words and spaced-out lettering. The improvements are largely cosmetic, and not, it must be said, entirely consistent. Splitting of dukkhass’ antaṃ, etc., and hyphenation of such as anu-d-eva and na-y-idaṃ are certainly helpful; but one may wonder what is gained by anomalously splitting sammā-ājīvassa. Given that kathañ c’ Ānando and pī ti have now been split, it is surprising that katamesaṃ, idhāhaṃ, khvāhaṃ have not become katam’ esaṃ, idh’ āhaṃ, khv āhaṃ. The inconsistent use of m/ṃ has been eradicated (but not, e.g., in the first relevant occurrence, p. 1, line 13, where dhammānam samāpattiyā subsists). At p. 142, line 16, saṃghāṭi patta- has correctly been compounded: but it is not clear whether the resulting saṃghāṭI-patta- is merely a misprint or intended as an unsignalled emendation to saṃghāṭī-. Outside the text, misprints have tended to subsist: the rubrics Kūṭāgara, Paññāsam-catuttham, and SammapadhānaSaṃyutta remain at pp. 452, 495 and 498. Instead of correcting two wrong page numbers at p. 493 f. (for the Rāhula and Lakkhaṇa Saṃyuttas), a third error has been introduced into the pagination of the Kassapa Saṃyutta.
In the matter of abbreviations, one would have welcomed more assistance. Indication of Feer's own abbreviations has been disguised by retaining (as at p. 299) the type ni◦ kho ā◦ (nisinno kho āyasmā), while tacitly replacing antaṃ◦ ◦etad with antaṃ … etad. The former type serves little purpose, especially when (as in the case of sati◦ bhā◦ bahu◦ at p. 304) the abbreviation is sporadic and the words are more frequently written in full. The latter type too might have been dispensed with (at the expense of an occasional supernumerary line, without disturbing the pagination): only six words are involved in the above instance.
The system used by Feer “for indicating which manuscripts use pe, la, or pa for abridgments” has been deemed too complicated and has been dropped in favour of “… pe …”. Simple italicization would, however, have sufficed to clarify their occurrences, without thus masking his distinction between abridgments in S[inhalese] manuscripts (pe), those in B[urmese] manuscripts (la, pa), and those common to both. There may be a case, however, for following the PTS translation in filling out those abbreviations which Feer implies are found only in one source: at p. 142, his āmantesi || [pe] abbreviates by less than a dozen words. An abbreviation which is common to both sources (with la [pe] in Feer) is confusingly asymmetric: (…) kāye kāyānupassī viharati (…) vedanāsu … pe … citte … pe … dhammesu dhammānupassī viharati. Full restoration is pointless, but it would have been helpful to supply [vedanānupassī] and [cittānupassī] to match dhammānupassī. The translation could solve this problem by suppressing the asymmetric dhammā-.
Feer's choice of readings has not been altered. In particular, the volume retains the title Mahāvagga, although the colophon mahāvaggasaṃyuttaṃ niṭṭhitaṃ appears only in B manuscripts. This awkward use of the term vagga, consistent only in B, to label the five Books, as well as their minor components, is doubtless spurious and would have been better ignored. The relevant wording varies as between Feer's S and B manuscripts, and seems likely to reflect a misunderstanding of vagguddānaṃ “list of minor Vaggas” as vaggass’ uddānaṃ “index to the major Vagga”. The effect, apparently in the first instance, is to extend the application of the term Vagga, not to an entire Book, but to the first major component of a Book. Following the final minor Vagga and its list of Suttas (tatr’ uddānaṃ), we have, listing the minor Vaggas of each Book's major component, the ambiguous term vagguddānaṃ in B at II, 130 (Nidāna), in S at III, 52, 157, 188 (the three Paññāsakas of Khandha), and in both B and S at IV, 204 (the four Paññāsakas of Saḷāyatana). We find this interpreted as vaggass’ uddānaṃ in S at II, 130, paññāsakassa vaggass’ uddānaṃ in B at III, 52, thus transferring its reference to the initial major component and rendering superfluous the term nipāta that designates the Paññāsakas at III, 188 and IV, 204. (For equation of vagga with nipāta, see v. Hinüber, A Handbook of Pāli Literature, 1996, 39, n.) Feer adopted the reading vaggassa for Books II and III, despite the fact that only vagguddānaṃ was attested for IV. The further labelling of entire Books as Vaggas presumably stems from the ambiguity of the rubric khandhavaggassa nipātaṃ, applied in S at III, 188 to the final Paññāsaka (nipāta) of the initial component (khandhavagga), but inducing both S and B at III, 279 to apply khandhavagga to the Book as a whole. As regards the complication of issues, B colophons begin where S colophons leave off.
The final table of contents for the entire Nikāya is reprinted, retaining all Feer's invented rubrics. The term saṃyutta is generally thought to denote “thematic group”, but this notion has been arrived at by applying to a large part of the text a designation saṃyutta which Feer “did not find in any of the MS. at my disposal” (II, viii). On the contrary, a distinction is basically made in the colophons of Books II–IV between their initial major components, Nidāna, Khandha, Saḷāyatana, and their subsequent Saṃyuttas. This implies that, e.g., “Nidānasaṃyutta” originally meant more naturally “(text) appended to the Nidāna”. The initial major component of Book II bears the designation Nidāna alone (II, 130, 10); subsequent components are separately numbered as Saṃyuttas: the first, the fifth, the seventh (this has samattaṃ in place of sattamaṃ, as Feer suggested), and the eighth in S, but only the third in B. An apparently shared confusion has induced B to number the second also as tatiyaṃ, and S to number the third as catutthaṃ. That this should be, as Feer believed, the relic of an original numeration that included the initial component is unlikely, given that Nidāna, Khandha, and Saḷāyatana lack the rubric Saṃyutta. In Book III also, the numbering of the Saṃyuttas in S (III, 234, 240, 249, 279) consistently excludes the initial component, the *ti-paññāsaka Khandha. [Erasure of ṇi in khandhavaggassa nipātaṃ … tī <ṇi> paññāsakaṃ of S1 (III, 188) is borne out by *catu-paññāsaka for the initial Four Paññāsakas of Book IV (with catu <ttha> paññāsake, but also catupaññāsakā ete in B (IV, 204); S has †kela† after nipātaṃ, possibly for tena (vuttaṃ); Feer prints -pāṇṇāsake, etc.] The Saṃyuttas of Book IV remain unnumbered.
Books I and V have no distinctly larger first component. All their segments are appended Saṃyuttas, numbered serially, and only B has supplied a separate title “Mahāvagga” for the Saṃyuttas of Book V. All these data were described by Feer, but he chose to generalize the inflated terminology of B, so that in his “new recension” (v. Hinüber, op.cit., 36) the initial Nidāna of Book II has become the “Nidānasaṃyutta” of the “Nidānavaggasaṃyutta” (although even B retains the respective labels Nidāna and Nidānavaggo saṃyuttako, before duplicating the latter with Nidānavaggasaṃyutta at II, 286).
This publication is not the radical re-appraisal of the manuscript evidence that one might desire: and it is doubtful whether even Feer's painstaking apparatus is sufficiently lucid for such a purpose. Some effort to remove a number of typographical errors and inconsistencies was certainly overdue, however, and the opportunity has usefully been taken to add line-numbering and to conform to more modern norms of transliteration.
