A F R A G M E N T OF T H E S A N S K R I T SUMUKHADH.4RA.

Ni*

G. M. BONGARD-LEVIN,

by M.I. VOROBYEVA-DESYATOVSKAYA,

Moscow
E. N. TYOMKIN

Leningrad Leningrad

Among the Central Asian documents from the collection o f N . F. Petrovsky kept in the Manuscript Collection of the Leningrad Branch of the Institute of Asia, USSR Academy of Sciences, under the code S I ~ there is a fragment of a manuscript written in Br~hmi. The fragment is a sheet of yellowish paper, 65.5 x 6 centimetres. 5.5 cm. from the left edge recto there is a circle in black ink, 2 cm. in diameter, with a hole in the centre evidently for a cord. The text is written in black ink, 4 lines on each page. The script is Upright Central Asian Calligraphic Br~hmi, the size of characters is 1 cm. (width) x 0.5 cm. (height). The spacing is 1 cm. On the left margin recto the number of the sheet is preserved - 6. The date of the document is unknown. It was acquired by N. F. Petrovsky, Russian consul in Kashgar at the beginning of the 20th century. The whole collection was passed by him as a gift to the Asiatic Museum of the Russian Academy of Sciences. The text of the document reads as follows :1 recto 1. 2. 3. 4. 1. 2. 3. 4. husavatipusavati, hilihili, yath~tbhaya, yath~tgniyatha param,yathabhaya.myath~tvajreyath~hrdayaro. iyam.vajrap~i.'sumukhanamadh~tra.nitath~tgatenabhas.it~sarvasatv~n~t[ma]rth~t[ya]k~ru~yatay~ahamapibh~.si.sy~tmiyathataipfirvakaisamyaksam, bu[d]dh[aiverso rbh~.sit~dhi~t.hit~tcasatyapratijfienavyavasth~pita, athakhalubhagav~[to]rnako~[n]mah~tpuru~alak~a .n~traw vabhhsapr~durbhfi[t a]tay~tcara~m[y]asaro,codit~buddh~tbhagavatasvakasvakaibuddhaks.etrasar~3prasthit~tsahebuddhak~etrai.h : I I : : ]1

9 The authors express their sincere gratitude to Professor H. W. Bailey for his very valuable notes on this article. 1 The text here is not divided into words. We followed only the line division and punctuation of the original. The marks of division at the end of the text are the same as in the manuscript.

A FRAGMENT OF THE SANSKRIT

Sumukhadhdran.~

151

The content of the text and above all the name of the dhgra.ni, occurring in the second and third lines of the text recto give a sound basis for considering this fragment part of the Sanskrit text Sumukhadhdra.nL Acad. S. F. Oldenburg was the first to suppose that this document is part of the Sumukhadhdran.L On the paper in which this document was wrapped there was written by S. F. Oldenburg - Sumukhadhdra.nL This text is of considerable interest, for it supplies us with a fragment of the Sanskrit original which was up to the present time regarded as lost and preserved only in the Tibetan, Chinese and Khotanese-Saka translations. The discovery of the Sanskrit original is of particular importance because it helps to understand the whole of the Khotanese-Saka version better and, in particular, to determine the meanings of those Saka words which are not found in other Khotanese-Saka texts. The creation of the text of Vajrapd!n. isumukhadhdra.ni may be referred to the first centuries A.D., i.e. to the period when Vajray~na, the achievement of nirvg.na through dhgran.i (magic formulas) began to take shape. The Sanskrit name of the dhgran.i is "Excellent Gate of Vajraphn.i".-" The discovery of the fragment of the Sanskrit text gives us the opportunity to collate the four versions of the Vajrapdn. isumukhadhdran, l, namely, the Sanskrit, the Khotanese-Saka, the Tibetan and the Chinese ones. The Sanskrit text should naturally be regarded as original, the Saka, Tibetan and Chinese as translations. But the analysis and collation of the texts shows that they were not merely translations of the original now found but rather reveal the existence of at least three different versions of the source in question from which they originate. The Tibetan and Chinese versions can be dated with rather close approximation. The four Chinese versions date back to the 4-5th centuries A.D. The first mention of the Chinese versions is found in the catalogues of the Eastern Chin but the fifth one was translated by Bodhiruei in the beginning of the 8th century A.D. (N 1139).3 There is a considerable divergence between all translations but Bodhiruci's translation is closest to the Sanskrit text which is at our disposal. The Tibetan translation of the dh~ran.i is called 'phags-pa sgo-bzah-po
Vajrapg .ni - name o f a B u d d h a or Bodhisattva, see BHS, p. 467. H r b r g i r i n , Dictionnaire Encyclopddique du Bouddhisme d'aprks les sources chinoises etjaponaises. Fascicule annexe: N 1137, 1138a, b; 1139, 1140. N 1137 (Chanfangpien t'o lo ni king); N 1138a (Kin kang pi mi chart men t'o lo ni king); N 1138b (Kin kang p i mi chart men t' o lo ni tcheou king); N 1139 (Hou ming f a men chen tcheou king); N 1140 (Yen cheou miao men t'o lo ni king). Bodhiruci's translation is in TaishO, vol. XX, pp. 585b(2)-589.

152

BONGARD-LEVIN, VOROBYEVA-DESYATOVSKAYA, TYOMKIN

~es bya-ba'i gzuhs (the literal translation of the Sanskrit name) and is part of the Kanjur. 4 Jinamitra, Dfinagila and Ye-ges-sde who lived in the 9th century A.D. are indicated as translators. 5 The Saka text6 is written ill the so called "late-Khotanese" language and can be dated to the 9-10th century A.D. 7 The collation of the Sanskrit text with the above mentioned translations reveals that neither of them (except for Bodhiruci's version) originates from the discovered Sanskrit version. This is demonstrated above all by the dharan.is themselves being different in all the above-mentioned texts. Closest to the Sanskrit version are the Tibetan text and the Chinese translation made by Bodhiruci. The Saka translation, as compared with the Sanskrit and Tibetan ones, is fuller mainly due to the introduction of some descriptive details. At the same time the Sanskrit, the Tibetan, the Saka versions and the Chinese of the T'ang period form one group which clearly opposes other Chinese translations. Bringing together the chronological data on the translations of the given text and the results of textological analysis, we are justified in supposing that in the first centuries A.D. (earlier than the 4th century, that is before the Chinese translation was made) the original Sanskrit version of the dh~ran.i was created (Sanskrit version I), on which the Chinese translation bases itself- King kang pi mi chan men t' o Io ni king (N 1138a). Simultaneously, there evidently appeared the shorter Sanskrit version (Sanskrit version II), which gave birth to another Chinese translation (N 1138). The supposition that the short Chinese translation is a summary of the complete one can hardly be accepted for the Chinese translations differ not only in their length but also by the texts of the magic formulas (dhara.ni) themselves. The Sanskrit versions I and H which can be reconstructed on the basis of the Chinese translations did not survive. The Sanskrit manuscript that we now possess comes from Central Asia judging by the hand-writing and the place of acquisition - Kashgar. The
4 See Kanjur, Derge edition, rgyud-'bum, ba, f. 47b-52a; the fragment in question is on f. 48a. 5 In the Tibetan chronicles the activity of the translators Jinamitra, D~.nagila and Ye-ges-sde is referred to the reign of Ralpacan (815-839 A.D.). See, for example, E. ObermiUer, Bu-ston. History of Buddhism, pt. 2: The History of Buddhism in lndia and Tibet (Heidelberg, 1932), p. 196. 6 Published in H. W. Bailey, Khotanese Buddhist Texts (London, 1951), p. 135-143; the fragment under consideration is on p. 137. 7 See H. W. Bailey, "Indo-Iranica, II", BSOAS, vol. XIII (1949), p. 138 ; M. J. Dresden, "The J~takastava or 'Praise of the Buddha's former births'", Transactions of the American philosophical society (Philadelphia), New Series, vol. 45, part 5 (1955), p. 404 and others.

A FRAGMENT OF THE SANSKRIT

Sumukhadhdra.ni

153

A FRAGMENT OF THE SANSKRIT

Sumukhadhdra.n?

155

version which it contains is closer to the Tibetan, Saka and Chinese one of the T'ang period. The date of the compilation of the Tibetan translation being exactly known, and the possibility of approximately determining the time when the Saka one appeared, allow us to suppose that the discovered Sanskrit version (III) became known in Central Asia no later than the 8-9th century A.D. or maybe even earlier. This supposition is backed up by the existence of the Chinese translation, belonging to the beginning of the 8th century A.D., which was evidently made from the Sanskrit version found in Kashgar. For the textological analysis of the above mentioned four versions the table below is provided in which the Sanskrit text divided into lines according to the division in the manuscript serves as a basis for comparison. Every line of the Sanskrit text is counterposed by the corresponding passages from the Tibetan and Saka texts (and for mantras also the Chinese ones). It is necessary to point out that Bernhard Karlgren a bases his reconstruction of T'ang phonetics on the classic literature of the T'ang period, above all on poetry. However, the dh~ran.is and the transcription of the Sanskrit untranslated names and terms in the translations of Buddhist texts made at the time of the T'ang dynasty (Hsiian-tsang, I-ching, Bodhiruci etc.) convince us that the translators transcribed the words most probably on the basis of the Northern dialect of the regions of Loyang, Ch'angan, Tunhuang, the phonetics of which is almost completely identical with that of the modern northern dialects. That's why we will first give the transcription of the text of the dh~ran, i using characters adopted for modern p'u t'ung hua. The cases of the supposed deviations from modern p'u t'ung hua which we draw from the dh~ran.i texts in Chinese are indicated in brackets; huso(sa)pati, puso(sa)-

pati, puso(sa)pati, si(hi)lisi(hi)li, ye(ya)t'a 9 she(che)ya, ye(ya)t' aach'i(k'i)ni, ye(ya) t' apo(pa)lan( lam )che( cha), ye(ya) t' apayen(yam ), lo ya(ye)t' apachelan(/am), ye(ya)t' ahelito(ta)yen(yam), n
8 B. Karlgren, "Grammata Serica Recensa', Bulletin of the Museum of Far Eastern

Antiquities (Stockholm), vol. 29 (1957).

9 Here as well as henceforward it is clear from the text of the dhhra.ni that the hieroglyph t'a must be read long. ao This sign read chien, but we presume, judging by the Sanskrit text, that this is a mistake. It is used instead of the sign yen ("Salt") similar to it in shape and having the reading which corresponds to the Sanskrit one. 11 The authors express their sincere gratitude to L. N. Menshikov and L. G. Gerzenberg whose help in the analysis of the Chinese and Saka texts was invaluable. The remarks on the transcription of the Chinese dh~rani and translations of the Chinese versions have been made by L. N. Menshikov.

Lines of Sanskrit original with corresponding parts of Saka and Tibetan Text
Lines

Sanskrit

Saka
.s~'mi .sfi' vajrrap~l.la suma-

Tibetan
lag-ha rdo-rje 'di-ni sgo bzafipo

2 (end) iya.m vajrap~n.isumu-

kha n~na dhfira.ni tathft- kha n4ma d~r~fi~t [~ ma.mgatena bh~ifft sarvasatv~ng- drr~n/i pata cu pa.d~.jsyau avam~yyau] gyastyau ba'[ma]rth~[ya] ysyau jsa b hvata ~ bi~nu sarvasatvgnu kgmwatayh ~ aham api bh~- mu'gdi' pracaina aysi vafia .si.syfimi yathatai a pfirvakai p~tca' hv~.fiim~i ~ khu ra tv samyaksam, bu[d]dh[ai]dftrfifi pa(lf~mjsya gyasta ba'ysa

ies bya-ba'i gzufiste ] de b~inggegs pa-rnams-kyi[s] thugsrjes sems-can-rnams-kyi dondu bgad-do I ji-ltar sflon-gyi de-b~in-g~egspa yafi-dag-par rdzogs-pa'i safls-rgyas-rna.ms-kyis gsufls~in byin~-gyis brlabs-pa daft I bden-pa'ithugs-dam-gyisrnam-par b2ag b~in-du flayafl 'chad-do I

ye1"8o

r bhfi.sitfidhi.s.thitfi ca satya- [5..sa'.na-vajsama vyacha sarva- de-has bcom-ldan-'das-kyis pratijfiena vyavasthfipitfi r atha dharma] hvfi.md~i ide u khvi ba.s.tyS..mda]de ~ u khvi hi.thi khalu bhagavfi[to]g prattifia vara vistS.md~iide tti mi ttya bfidfi [~akyamuga~i gyastS~n/i] gyast/i ba'ys~i

rnhko~fi[n] h mah~purusalaksanfi[d] i ra~mi~ pramukt~ sarvabuddhaksetresu [rag]my[~]-

urfii jsa [vh ham.dama ttina] mahfipurusa-lak.san, fi [hudaht~lfi gt~aina] vasve [pattava.mci] bg'y~i pagS, ve sfi' mi bi~e ysama~a .mdai vira harbigvft buddha-k.setru~, bfi'y~n/i hivya

skyes-bu 'chen-t3o'i mtshan smin-mtsham-skyi mdzodspu-nas 'od-zer gyur te I safls-rgyas-kyi iifl thams-caddu 'od-zer-gyi

vabhftsa k pradfirbhfit [ta]yfd h a r r ~ m a cira himya ~ cara~m[y~] ~asa.mcoditfin bud- [ttyau bS?yyau jsa nfi harbi- snafi-ba des bskul-bas raft-raft~ye ysamaw bierS, bud- gi safls-rgyas-kyi iifi-dagdh~ bhagavata ~ svaka = dha-k.s~trufi gyasta ba'ysa] nas chas-te ] hajavi.sya himya ttina aye.sthflna~ tti mi hid

svakai buddhak.setrasa .mpra- hivi buddha-k.s~tdina vh.st.a 'jig-rten-gyi khams mi-mjed sthitgq sahe buddhak.setrail) I] pastgta [tti gyasta ba'ysa 'dir byon-nas ttira ku mara ttifia] sahelovadeta 5,ta. ,.

~For the notes to this table, see p. 159.

Translation Sanskrit text Saka text Tibetan text Chinese text Buddha said to the Bodhisattva-maMsattva - "The Hand of Vajra": "This

This dMral)i by the This dhfiran.i by the This dh~ra~i by name of Excellent gate of name of Vajrap~tni- name of Vajrap~n.a VajraNai sumukha (Excellent sumukha, gate of Vajrap~.ni)

is uttered by the tatM- the stanzas of magic by the tathS.gatas for divine incantation, dhftgata for the sake of all formulas, was uttered the sake of all living ram preserving life, "The Excelthe living beings [and] by the former innu- beings is uttered merable Buddhas. out of compassion; I shall also utter it just as by the former complaely enlightenedBuddhas From compassion to all the living beings I shall now say it again [also] just as the former divine Buddhas out of compassion. Just as by the former t a M gatas samyaksarpbuddhas it was lent Gate" was already uttered [and] protected by the former tathagatas samyaksam,buddhas.

it was said and produced magically and by true claim established. Then by the Bhagavat

arhats who conceived all dharmas, proclaimed and produced magically and bytrueclaim established this dhfirarg. Then at that time divine Buddha Sgkyamtmi from the circle of hair between the eye-brows and from other marks [of thel great man emitted a shining ray. All over the world, over all Buddha-fields

said and blessed and by true claim established so do I proclaim it. Then the Bhagavat

They proclaimed a vow and established it truly. I at present also for the benefit and out of compassion

from the circle of hair between the eye-brows - the mark of a great man - was emitted a ray (of light) and over a l l the Buddhafields I the light of the ray

from the circle of hair between the eye-brows - the mark of a great man - emitted a ray of light. Over all the Buddha-fields the brilliance of this

for all beings proclaim a great vow and estabfish it truly. Therefore I say this dMran.L Then the Bhagavat emitted light from the circle between the eye-brows - the mark of a great man. This light illuminated all the Buddha-fields.

spread. And urged by the fight of this ray fight spread. Urged by that ray the blessed spread. Of these rays this brilliance of t h e all over the world in all light, Buddhas each the Buddha-fields the Buddhas became incited by this magic transformation. And then from each from their Buddha- Buddha-field so many they came from each fields to the world of Buddhas went there to Buddha-field to that that world of people. part of the Universe people departed. where people (live).

By the magical power of the Buddha all the tath~,gatas in the Buddha-fields saw this light and each of them at the same time went from his field to the world of human beings (sahd lokadMtu).

158

BONGARD-LEVIN, VOROBYEVA-DESYATOVSKAYA, TYOMKIN

I n the table below we suggest the possible reconstructed variant of the reading o f the given dhgran.i in the T ' a n g period.
Sanskrit Saka Tibetan Chinese

husavati pusavati hilihili yathftbhaya yathagni yathftpara .m yathftbhayam, yathftvajre yath~h.rdaya.m __

hasavate phusavate pusavatP z hili hell yathft vajrri yath~ agni yath~ paranacam yathftbhayam, yathft param,ca TM yath~ vajrram yathg hyaday~arn 1. __

husavati mupavati hilihili yath~padzuya yathftagni yathfiparantsa yathgnala yathgtvadzran, a yath~h.rdaya ~s sv~h~d 6

husapati pusapati, pusapati 1~ hili hili yat'ftcheya yat'fiak'ini yat'gpalamcha yat'ftpayam yat'~pachelam yat'fihelitayam

The table shows that magic formulas in different versions coincide and diverge in a way which makes it impossible to reconstruct the original Sanskrit variant. The tables on pp. 156-157 present the text divided into parts corresponding to the lines of the Sanskrit original with comments on the discrepancies between the Sanskrit, Saka and Tibetan text, and translation o f the text.

13 Two words instead of one in the Sanskrit text.

13 Para .mcais repeated here. In the Sanskrit text it is absent.

1~ The reading of the Khotanese-Saka version suggested by ProL H. Bailey should possibly be changed to h~rdayq.m. The underline sign for ya and r in Upright Central Asian Brfthmi is almost the same and its variations often depend on the peculiarities of the hand-writing ( ~ ya and,~.r). However, the possibility of Skr. hrdaya.m changing into hyadaya.m in the "late Khotanese-Saka" language should be considered, for the latter in a position before a consonant often proves unstable. See H. W. Bailey, BSOAS, vol. X, p. 917; Asia Major, N,N.S., I, p. 38; TPhS, 1952 (1953), p. 57; M. J. Dresden, "The Jgttakastava", p. 408. 1~ Inseparably from the following word. 1~ This word is present only in the Tibetan text.

A FRAGMENT OF THE SANSKRIT Sumukhadhdra.n7

159

Square brackets [ ] mark those parts of the Saka text which have no corresponding part in the Sanskrit one. Individual Saka and Tibetan words and forms unparalleled by the Sanskrit ones are explained in the notes. b To the Skr. tathdgatena bhd.sitd "[dh~ra.ni] is said by the tath~gata [sg.]" the Saka gyastyau ba'ysyau jsa hvata and the Tibetan de-b.~in-gAegs-pa-rnams kyi thugs-rjes... b~ad-do - "[dh~ra.ni] is said by the tatMgatas [pl.]" correspond. e Skr. arthdya kdrul.~yatayd "out of compassion; for the benefit of" exactly corresponds to the Saka mu'Adi' praeaina. a We leave here the reading of the manuscript. The given text is notable for the absence of the final visarga, which is possible in BHS. e Due to the rules of Tibetan syntax the words aham api bhd.si.sydmi "I shall also say" when translated into Tibetan were placed at the end of the sentence na),a~i'ehad-do; the corresponding Sanskrit to the part of the Tibetan sentence from the words byin-gyis to b~in-du is found in the next line. The Sanskrit text helps reveal the meaning of the Saka ba.s.tydznda. Evidently, it corresponds to Skr. adhi.st.hita "established, founded". The Saka word ddr~ii (see the preceding line) "in dMra.ni" has no correspondence in the Sanskrit text, but the agreement in the Sanskrit text is evidently directed to this word, because the participial forms bhdsitd, adhi.st.hitd and vyavasthdpitd are Nora. sg. fern. The Saka ddr~ii should evidently be regarded as Loc. sg. fern., for the subject here is gyasta ba'ysa Nora. pl. "buddhas". One must read bhagavd[to]r.na. h In the manuscript [gt]r.nakoAd[d] with d in (trnd. Evidently, a mistake of the scribe. i In the manuscript ~ The sense demands here Abl. sg. In classical Sanskrit it is ~ dd. May be it is possible to have ~ in BHS with d (see Edgerton, 1.8.46-48; most examples are concerning Mah~vastu). Visarga is absent in the manuscript. It is possible in BHS. k Skr. avabhdsa "light, brilliance" corresponds to Saka harru~dma. Skr.pr6durbMtta "spread over, manifested" is rendered by Saka eira "himya" and Tibetan byu~ "appeared". See H. W. Bailey, "Saka Miscellany", Indo-Iraniea, Mdlanges prdsentOs Georg Morgenstierne (1964), pp. 8-9. We suggest the reconstruction [ta]yd. m In the manuscript only ra~md is written. n Skr. sa.meoditd[.h] meaning "urged, incited", ef. Pali eodeti "to urge, incite, exhort", corresponds to Saka hajavLsya. The form of hajav(sya (appears in combination with himya) shows it to be probably an adjective with the meaning "incited" (cf. hajua "wise"). The Tibetan text proves this meaning to be correct: bskul-bas "urged, incited". o The text is characterised by the absence of the final visarga. It is possible in BHS, so we leave the text without changes. P Saka ayest.hana (Skr. adhis.t.hdna, Pali adhit.t.hdna) means here, evidently, magic, supernatural appearance, transformation. (Tibetan byin gyis brlabs-pa exactly corresponds to S a n s ~ i t and Saka.) Cf. Skr. adhiti.sthati, Edgerton, II, pp. 12-13. q Probably Skr. san.lprasthitd[.h] Nom. pl. agrees with Buddha[hi. The Saka text is more complete than the Sanskrit one. In the Sanskrit text - "to the world of human beings (sahe) the blessed Buddhas went"; in the Saka - "departed (pastdta) as many Buddhas as here to this world of people (Sahelova-deta) came (6ta)". In the Saka text the Sanskrit words and their Saka equivalents are doubled: mahdpuru.sa-laksa.nii hudahttnii gttnaina; pattavamei bd'yii. The translation does not reflect this duplication.