The Suvarnabhasa Sutra

8/3/2019 The Suvarnabhasa Sutra

1/6

77

In 1913, only four years after returning (October1909) from his monumental expedition into Cen-tral Asia (19061908),1 Paul Pelliot published aKhotanese text for the first, and last, time from

among his numerous manuscript findings fromDunhuang. The manuscript, in pohi format, P3513,2 is a collection of Buddhist texts,3

amongwhich is found the de

an

a

confession chapterof the S

u

tra of Golden Light in Late Khotanese.Pelliot published the first quarter (the introduc-tory prose 3.01010 and verses 3.124 in Skjrvsedition, to which references of the Suvar

abh

a

sa

text are made hereafter) in transcription and trans-lation accompanied by a commentary (only tothe introductory prose). It was Sylvain Lvi whofirst noticed in 1910 the presence of a part of the

Suvar

abh

a

sa as well as the Bhadracary

a

de

an

a

among the 84 folios of P 3513,

4

and, although thework on the

Suvar

abh

a

sa

text was conducted inclose collaboration with Robert Gauthiot, it isPelliot alone, as it is explicitly stated on p. 94,who was responsible for the decipherment, theseparation of words, and the literal interpreta-tion. Gauthiots part of the work was expected tobe published as a second part (ibid.), which washowever never to be realized.

Seen from the present perspective, Pelliotswork of 1913, when the study of the Khotaneselanguage was in its infancy, is naturally full of

shortcomings. On the other hand, consideringwhat was known of the language at that time,and in the context of contemporary works byothers, the high quality of Pelliots paper is un-mistakable.

5

Apart from the manuscripts broughtfrom Dunhuang by himself, of which referenceis made only to P 3513 and another batch of fo-lios P 3510, he had at his disposal two articles by

Hoernle of 1910 and 1911, and two works by ErnstLeumann of 1908 and 1912. Hoernles 1911 paperobviously benefited others including Pelliot andGauthiot by presenting for the first time in fac-

simile some manuscripts from Dunhuang wheresyllabaries are given as scribal exercises of thecursive Br

a

hm

i

of similar types to P 3513, andby discussing their reading, Hoernle has beencredited with having deciphered

Khotanese forthe reason that he pointed out its similarity tosome Iranian Pamir languages.

6

However, deci-pherment in the case of the Khotanese languagewas a gradual process. Hoernles role at the initialstages of this process can more adequately becharacterized as publishing, often with facsimileplates, the materials he had at hand with a lim-ited degree of success as regards interpretation.His identification of what he called an unknownlanguage of Eastern Turkestan with Iranian isbut superficial.

7

In 1901 he compared a numberof words with Sanskrit, Persian and Ghalchahdialects (of the Pamir) without discriminatingindigenous words from Indic loans. In a series ofarticles in 19101911 dealing with abundant newmaterial from Steins Second Expedition (19061908) he hardly goes into the details of the geneticaffiliation of the language except for a commenton another unknown language, i.e. Tocharian(1910, 1299f.). His interest was simply not there.

Moreover, Hoernle consistently read (1911, 460,and passim) the very frequent ligature tt- as nt-even in an initial position.

8

Pelliot (p. 102f.), onthe other hand, agreeing with Konow and Leu-mann, preferred the reading tt

-. As with Konow(1912a, 1130), the main argument for Pelliot wasthe comparison with the demonstrative pronounsof Old Iranian (which was to be discussed by

Paul Pelliot and the De

an

a

-parivarta

of the Suvar

abh

a

sa-s

u

tra

*

H I R O S H I K U M A M O T O

university of tokyo


2/6

k u m a m o t o

: Paul Pelliot and the De

an

a

-parivarta

of the Suvar

abh

a

sa-s

u

tra

78

Gauthiot in the unpublished second part). Alreadyin a lecture given in December 1911 (Pelliot1912), Pelliot maintained, against Leumann (whofirst called it Sprache II

, then Nordarisch

) andwith Konow (1912b, 553), that the language wasIranian, assez aberrante, assez use, beaucoupplus loigne des autres langues iraniennes que lesogdien par exemple, mais dont les caractris-tiques ne sont pas douteuses (1034).

9

In the text of the De

an

a

-parivarta

printed byPelliot, we notice that in a great many cases Pel-liot read the vowel sign -

i

where later editions

10

preferred -

. This relates to a detached vowel signabove the right shoulder of an ak

ara. The firstattempt to distinguish it from the sign -

i

whichis attached on the top center or left of the ak

arawas made by Bailey (1938a, 521; 1938b, 592), whoused -

i

(underlinedi

) and explained that it was anatural development from the double dots (dier-

esis) found in formal script. The next year, how-ever, Bailey abandoned the transcription -i andadopted - instead (1940, 365), which, with theblessing of Konow, has been standard practiceever since. Another difference in transcriptionconcerned the anusv

a

ra. Pelliot, like most othersat the time, transcribed all the anusv

a

ras as -

m

%

.The use of the so-called unetymological anus-v

a

ra, which is extremely frequent in Late Khota-nese, indicated not by the letter -

, but by asubscript tail (Polish ogonek

, Lithuaniannosine

%

)as in -

a

4

for -

a

, was first used by Leumann 1912,59. With Leumann it was an extension of the

scheme he employed to distinguish vowels inmanuscript readings and metrically required vow-els in Old Khotanese verse texts. This methodof transcription which violates the principle ofdiplomatic reproduction never found favor withKonow. However, Bailey first adopted it in 1938a,530, continued it in Khotanese Texts

(1945, ix),and it has since become standard usage. Lastly,Pelliot attached the Dative-Ablative ending -

jsa

of certain types of declension to the precedingword, unlike Leumann and Konow, both of whomused a hyphen, or Bailey, who consistently tran-scribed it as a separate word. Pelliots method

unfortunately did not find a following, but it isentirely justifiable, since -

jsa

behaves just like aninflectional ending rather than a clitic.

A small number of misreadings

11

of the Br

a

hm

i

script could not be avoided. The ak

ara l

a

istwice read aslo

in 61r1 (3.4)

il

a

m

%

lokas and61r3 (3.5)

ama-l

a

m

%

Yama-loka.

12

The vowelsign in question is clearly -

a

. The reason for the

misreading may be sought in the facsimile plateof the syllabary Ch. 0046

13

published by Hoernle1911, plate IV (opp. p. 458), wherele

is followedby twol

a

s. The independent (initial) vowel let-ter o

, which Hoernle 1910, 1295, transcribes as

wa

, in accordance with its use in Tocharian, isrendered by Pelliot (91, n. 2) as o5 and o5 (affectde la marque de la longue). Pelliots proposal,which has something to recommend it for dis-tinguishing the independent form from the vowelsign attached to a consonant letter, did not how-ever find a following, while the reading of o-(and au- with the a-matra), and not wa, wasargued for by Konow 1916a, 217, and was gen-erally accepted.

Apart from these minor points, there are a fewmisreadings which at first sight seem quite seri-ous (see fig. 1). In 60v4 (3.3) Pelliots text has

paduaam%am% daesta with the translation en

avant (?) se trouvant (?) ils taient assis. The firstword, which must be paa before, in front,14

has a slightly different shape here from the sameword in 63r2 (3.18), where Pelliot has the correctreading. However, the vowel sign -u below theakara da is normally (and always in this manu-script) the hook type as in 61r4 (3.5) dukha mis-ery, and not the wedge type as in 60v3 (3.2)khuas, like. Another apparent misreading poses aproblem. It concerns 63r2 (3.18)nucura harsh(Pelliot violent), a loan word from Skt.nihurahard, severe. Leumann (1920, 58) corrected it to

nuhura, a form found 16 times in the manu-

script E (= The Book of Zambasta). Bailey in KT1(1945) followed Pelliot, but in the second edition(1969) changed it to nuhura like Leumann.However, if we compare the word in 63r2 to 65v2(3.38)nuura (both editions of KT1), the differ-ence in the second akara is clear. Pelliots read-ing in 63r2 was correct, and the correction byLeumann and Bailey (1969) amounts to an edito-rial emendation, with the implication that thescribe inadvertently extended the stroke in themiddle part to the left, since the existence of adoublet in -c- due to phonetic change is unlikely.Somewhat unfortunately for Pelliot, a rather rare

ligature -hy- in 64r1 ahyau jsa (3.24) throughdeceit led him to read the word as avyaujsawhich he could not understand. A comparisoncan be made with the shape of vye was in 60v1(3.010). Leumann, who did not go as far as thispart in 1920, 58, would no doubt have had thecorrect reading as the word occurs seven times inthe manuscript E. See ahyau in Z 24.260.


3/6

79

k u m a m o t o : Paul Pelliot and the Deana-parivarta of the Suvarabhasa-sutra

In morphology, Pelliot failed to recognize theabstract suffix -aua in dukhittauina (3.5, 24)through poverty and tsam% ttauijsa (= tsam% -ttaui jsa) (3.20) through wealth, separatingttauina, ttauijsa as a word of unknown mean-ing.15 It was Konow who, almost 20 years later(1932, 62), first recognized this suffix, deriving itfrom *-a-van-ya- and comparing it with the Sog-dian suffix -wny. Another example of an unrecog-nized grammatical form was the present middleparticiple ana lying in 60v2 (3.1) aha4phia4na lying undisturbed. Pelliot read aham%phi-

iam%na as one word with the translation ntantpas puis (?) as it appears to correspond to theSanskrit atandritena not tired. His questionmark no doubt resulted from his inability toanalyze the form. Leumann (1920, 59) was able tocorrect the reading with the translation unange-griffen . . . liegend. Both words, hapha- ver-wirrt and the present participle ana- wereknown to him from the manuscript E (Book ofZambasta). Similarly, Pelliot did not recognisethe word haara in haara baysa the formerbuddhas, 60v2 (3.1), and took baysa as the voc-ative singular, while Leumann (ibid.) was able to

translate haara as einstige, since the olderform hatara- frher, einstig was known tohim from the manuscript E. Many further suchcases could be mentioned which demonstratehow severely handicapped Pelliot was by the lackof material available to him before 1913. In addi-tion the only Sanskrit printed edition was notori-ously inadequate (Das and astri 1898, described

by Nobel (1933, 573), as eine der schlechtestenAusgaben, die jemals von einem indischen Textgemacht worden sind)16 and the Chinese ver-sion by Yijing (cf. Nobel 1948) as a rule does notgive word-to-word correspondences.

With all these shortcomings one might get theimpression that Pelliots attempt at understand-ing the newly discovered text was a failure. Infact it was not. Even when he questioned themeaning of individual words, he nevertheless inmost cases correctly understood the basic struc-ture of the sentence, and, most importantly, he

meticulously distinguished what could be knownfrom what was uncertain. In this respect it is un-fortunate that he was only able to publish a com-mentary on the initial prose part, in which hegives lengthy discussions of important words. Inan explanation of the passage ttye ivi byuthat night having turned into dawn (60r2 (3.06);see Skjrv 2004, vol. 2, 109), Pelliot (12021),following the Sanskrit tasya ratrya atyayenaat the end of that night, reluctantly translatedthe phrase as cette nuit sacheva, and then un-successfully tried to connect byu dawnedwith the verb bud- to know, be awakened (with

the form bvade they will know) on the anal-ogy of pyu heard besidepvade they willhear. All the forms alluded to by Pelliot are LateKhotanese, and we now know that the analogyfails in Old Khotanese. It emerges from the dis-cussion of the word, however, that the Khotaneseversion is much closer to Yijings Chinese zh tian xia

7 o y

7 quand fut arrive laurore

pa

a

pa

a

dukha khu n

u

cura n

uu

ra(60v4) (63r2) (61r4) (60v3) (63r2) (65v2)

a

hyau vye

a

hyau(64r1) (60v1 (Z 24.260)

Fig. 1. Problematic readings.


4/6

k u m a m o t o

: Paul Pelliot and the De

an

a

-parivarta

of the Suvar

abh

a

sa-s

u

tra

80

than the Sanskrit (all manuscripts). It has beenpointed out (Skjrv 2004, vol. 1, lxv) that theKhotanese version was made from the sameCentral Asian recension that Yijing used around700 c.e.

This is one of the clearest points thatsupport the view.

17

When Pelliot published his article in 1913, notext publication of Khotanese existed in themodern sense, in which the manuscript readingis presented in transcription, accompanied by atranslation in a modern language with a com-mentary which discusses philological and othermatters. In this sense Pelliots work was the veryfirst. Before him, Hoernle (1910, 128493) hadpublished a small portion of the Vajracchedik

a

and the Aparimit

a

yu

-s

u

tra

in Late Khotanese,both found at Dunhuang during the second ex-pedition of Aurel Stein. Here each word of theKhotanese text is accompanied by the Sanskrit

word below. However, since the Khotanese version isnot a word-for-word translation of the Sanskrit inexactly the same sequence, this method leavesmany Khotanese words unexplained. Some paleo-graphical and grammatical points are mentionedin the footnotes, which are rather rudimentarycompared to Pelliots. Leumann (1912, 7783) pub-lished the same texts as Hoernle 1910, and hereagain the text was presented in the same way withsome improvements on the reading. Leumannsown editio princeps

of the Sanskrit-Khotanese textof the Adhyardha

atik

a

Praj

a

p

a

ramit

a (1912,9299) has only the transcription of the text. It

was not until Konow (1916a, 1916b, 1916c) pub-lished the two texts, the Vajracchedika and theAparimitayu-sutra,in toto together with a gram-matical sketch of the language and a completeglossary, that a model for the publication of suchtexts was established. Although Konows worktook many years in preparation, possibly laid asidewhile Hoernle was assembling the contributionsto the volume,18 the honor of the first publicationof a Khotanese text in the sense discussed abovebelongs to Pelliot.

We can only speculate as to why Pelliot did notcontinue his work on Khotanese. The publication

of Konows major work (1916a, 1916b, 1916c), orthe inaccessiblity of older, and grammaticallyricher material being worked upon by Leumannmay have been part of the reason. Gauthiotsdeath in 1916 could also have been a blow. Butabove all he had, with the abundant materialavailable to him, more than a lifetimes work of

Sinological and Central Asian studies. It was aquarter century later, in 1937, before anotherscholar, H. W. Bailey, started to work again onthe Pelliot Khotanese manuscripts.

Notes

* I dedicate this short article to Professor Skjrv towhom I am particularly grateful for making his mon-umental 2004 edition of the Suvarabhasa-sutraun-fortunately still difficult to obtainavailable to me indigital format.

1. Duyvendak 2001, xix. Pelliot reached Beijing inOctober, 1908, not in December, 1910 as Lieu (n.d.) hasit. In the meantime, having given a number of lecturesin 1909 and 1910 (Walravens 2001, 2224, nos. 136,148, 150), he had published, together with douardChavannes, Un trait manichen retrouv en Chine,traduit et annot, JA (1911), 499617; (1913), 99199;

261394.2. Bibliothque nationale, Pelliot chinois 3513.Digital images of this manuscript are available onhttp://idp.bl.uk (search under Pelliot chinois 3513).

3. See Dresden 1977, 75.4. JA 1910, 62627; Pelliot 1913, 94 n. 3.5. Pelliots work is not mentioned in Emmerick

1992 either among the pioneers of the decipherment(p. 6) or in the section dealing with the Suvarabhasa(p. 33f.). Lieu (n.d.) does not mention it either.

6. Emmerick 1992, 6; Skjrv 2002, lxix; Sims-Williams 2004, 419.

7. Hoernle 1901, 3236.8. Only some years later (Hoernle 1916bd) did he

choose to follow Konow (1912a, 112930; 1916a, 217)and Leumann (1912, 38).9. This view was shared, independently from each

other, by Lders 1913 and Reichelt 1913, among others.10. Skjrv 2004, vol. 2, 4044, notes all the impor-

tant differences in Pelliot and later editions (Leumann1920, 5758, and both editions of Baileys KhotaneseTexts I).

11. 61r3 (3.5) syama- for ysama- is likely to be atypo. 61r4 (3.6) paham% mejsa by the sounding for

pa-, 61v4 (3.8) bvam% de they will know for bvadi,62v3 (3.15) ttriyaum% among animals for -u aremisreadings.

12. Corrected by Leumann 1920, 57. Leumann ap-

pears not to have seen the manuscript himself. He alsoomitted the last five verses (3.2024) published by Pel-liot without explanation. Leumann moreover restoredall the forms of baysa- Buddha and baysusti- bo-dhi in the verse part to balysa- and balysusti- for met-rical reasons (p. 56), but not in the prose part. He thusproduces a Late Khotanese text with a mixture of typi-cally Old Khotanese forms.


5/6

81


13. = IOL Khot S. 30. See Skjrv 2002, 536.14. So in Bailey 1945 and all subsequent editions,

but Leumann 1920 haspadua.15. Leumann, 1920, 57, read 61r2 dukhittauina as

one word, translating it (p. 60) as Unglcklichsein.16. For example in verse 4 (p. 9) both the second and

fourth padas are defective and make no sense. While

Pelliot partially managed to get the correct reading ofthe Khotanese version and its meaning, in the fourthpada he readkva ysira fornvaysira would issue andtranslated extrmement with a question mark. Thefact that Leumann (1920, 57) corrected Pelliot tohvata

ysira with an arbitrary reconstruction of the Sanskrit(p. 60) suggests again that he proposed all the correc-tions to Pelliot without seeing the manuscript.

17. See also uvara noble in verse 4 discussed inthe footnote above. Yijing has ` mio excellentwhich is absent from all the known Sanskrit versions.On the other hand the absence of verse 12 from theKhotanese version alone (Nobel 1937, xxvii) may sim-ply be a lapsus in this single manuscript.

18. Pelliot, 1913, 91, had seen the announcement ofthe forthcoming publication in Konow 1912c, but thebook took three more years to come out.

References

Bailey 1938a H. W. Bailey. Hvatanica III.BSOS

9.3:52143.Bailey 1938b . Vajra-praj

a

-p

a

ramit

a.

ZDMG

92.23:57993, 6056.

Bailey 1940 . R

a

ma [, I]. BSOS

10.2:36576.

Bailey 1945 . Khotanese Texts.

Vol. 1. Cambridge (2nd ed.1969: as Khotanese Texts IIII

,Cambridge) = KT

1.D

a

s and

a

str

i

1898 arat Chandra D

a

s and aratChandra

a

str

i

. Suvar

aPrabh

. Fasc. 1. Calcutta.Dresden 1977 M. J. Dresden. Khotanese

(Saka) Manuscripts: A Provi-sional Handlist. In Varia1976,

ActIr

12, 2785.Thran.

Duyvendak 2001 J. J. L. Duyvendak. PaulPelliot: (May 28th 1878

October 26th 1945). InWalravens 2001, xiiixiv(originally in Toung Pao 38[19471948], 115).

Emmerick 1992 R. E. Emmerick.A Guide tothe Literature of Khotan

. 2nded. Tokyo.

Hoernle 1901 A. F. R. Hoernle. A Report ofthe British Collection of An-tiquities from Central Asia.JASB 70, pt. 1, Extra no. 1.

Hoernle 1910 . The Unknown Lan-guages of Eastern Turkestan[. I]. JRAS

, 12831300.

Hoernle 1911 . The UnknownLanguages of EasternTurkestan. II. JRAS

, 44777.Hoernle 1916a . Manuscript Remains

of Buddhist Literature Foundin Eastern Turkestan

. Oxford.Hoernle 1916b . A Bilingual Fragment

in Chinese-Khotanese. InHoernle 1916a, 38799.

Hoernle 1916c . A Bilingual Fragmentin Tibetan-Khotanese. InHoernle 1916a, 400404.

Hoernle 1916d . Khotanese Vocabu-lary. In Hoernle 1916a, 4059.

Konow 1912a S. Konow. Zwei Handschrift-enbltter in der alten arischenLiteratursprache ausChinesisch-Turkistan.SPAW

, 112739.Konow 1912b . Review of Ernst

Leumann, Zur nordarischenSprache und Literatur: Vorbe-merkungen und vier Aufstzemit Glossar.Gttingische

gelehrte Anzeigen

,

55165.Konow 1912c . Vedic dasyu, to

c

ridah. In Festschrift VilhelmThomsen

, 9697. Leipzig.

Konow 1916a . The Vajracchedik

a

inthe Old Khotanese Version ofEastern Turkestan: Stein M[S].,Ch. 00275. In Hoernle 1916,21488.

Konow 1916b . The Aparimit

a

yu

S

u

tra: The Old Khotanese Ver-sion together with the SanskritText and the Tibetan Transla-tion. Stein MS., Ch. xlvi. 0013b. In Hoernle 1916, 289329.

Konow 1916c . Vocabulary toVajracchedik

a

and Aparimit-

a

yu

S

u

tra. In Hoernle 1916,

33056.Konow 1932 . Saka Studies

. Oslo:Oslo Etnografiske Museum(= Bulletin 5).

Leumann 1908 E. Leumann. ber die ein-heimischen Sprachen vonOstturkestan im frhern


6/6


82

Mittelalter: Zweiter Teil.ZDMG 62:83110.

Leumann 1912 . Zur nordarischenSprache und Literatur Vorbe-merkungen und vier Aufstzemit Glossar. Straburg.

Leumann 1920 . Buddhistische

Literatur. Nordarisch undDeutsch. I. Teil: Nebenstcke(=Abhandlungen fr dieKunde des Morgenlandes, 15,Bd. 2). Leipzig (repr. 1966Nendeln, Liechtenstein).

Lieu n.d. S. Lieu. PELLIOT, Paul.(A preliminary version of anEIr article seen as of 2008 at:http://www.iranica.com/articles/sup/Pelliot.html.)

Lders 1913 H. Lders. Die akas und dienordarische Sprach. SPAW,40627.

Noble 1933 J. Nobel. Rez. von Nanjio, Bun-yiu: The SuvaraprabhasaSutra, OLZ 89:57275.

Nobel 1937 . Suvarabhasot-tamasutra. Das Goldglanz-Sutra: Ein Sanskrittext desMahayana-Buddhismus.Leipzig.

Nobel 1948 . Suvaraprabhasot-tamasutra. Das Goldglanz-Sutra: Ein Sanskrittext desMahayana-Buddhismus /I-Tsings chinesische Versionund ihre tibetische Uberset-

zung. 2 Bde. Leiden.Pelliot 1912 P. Pelliot. Les influencesiraniennes en Asie Centrale

et en Extrme-Orient. RevuedHistoire et de Littrature

religieuses, n.s. 3.2:97119.Pelliot 1913 . Un fragment du

Suvaraprabhasasutra eniranien oriental: Textetranscrit, traduction et com-

mentaire. Mmoires de laSocit de Linguistique deParis 18:89125.

Reichelt 1913 H. Reichelt. Das Nord-arische. IndogermanischesJahrbuch 1:2036.

Sims-Williams 2004 U. Sims-Williams.HOERNLE, AugustusFrederic Rudolf. EIr 12.4,41820.

Skjrv 2002 P. O. Skjrv. KhotaneseManuscripts from ChineseTurkestan in the British Li-brary: A Complete Catalogue

with Texts and Translations.London.Skjrv 2004 . This Most Excellent

Shine of Gold, King of Kingsof Sutras: The KhotaneseSuvarabhasottamasutra.2 vols. Cambridge,Massachusetts.

Stal-Holstein 1908 Baron A. von Stal-Holstein.Tocharisch und dieSprache II. Bulletin de lAca-dmie Impriale des Sciencesde St.-Ptersbourg, 136772.

Walravens 2001 H. Walravens. Paul Pelliot

(18781945): His Life andWorksa Bibliography.Bloomington, Indiana.

Documents

The Suvarnabhasa Sutra