Dharmagupta's Unfinished Translation of the "Diamond-Cleaver" "(Vajracchedikā-

Prajñāpāramitā-sūtra)"
Author(s): Stefano Zacchetti
Source: T'oung Pao, Second Series, Vol. 82, Fasc. 1/3 (1996), pp. 137-152
Published by: BRILL
Stable URL: http://www.jstor.org/stable/4528687
Accessed: 19/11/2009 00:55

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=bap.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

BRILL is collaborating with JSTOR to digitize, preserve and extend access to T'oung Pao.

http://www.jstor.org
 DHARMAGUPTA'SUN}INISHED TRANSLATIONOF
( VAJRACEDI>-
THE DIAMOND-(;lHAVER
PRAJNAPA-SUTRA) 1
BY

STEFANOZACCHETTI
Universita degli Studi di Venezia-CaFoscari

List of Abbreviations
JGNDJ = Dharmagupta (transl.), Jingang nengduanbanruoboluomi
jingiWldzl*n£Xtt , T. 238 (vol. VIII p. 766c-771c)
S = EdwardConze's edition of the Vafracchedika-Prajnaparamita-sutra
(in: Conze 1957 p. 27-63)
tr = EdwardConze's translation of the Vajracchedika-Prajnaparamita-
sutra(in: Conze 1957 p. 65-92)
CSZiJ= Sengyou f$, Chusanzangjiji, th=i T. 2 145
Pet;

KYL= Zhisheng tZ, Kaiyuanshijiaolu, W3Xwtk; T. 2154

XGSZ = Daoxuan An, Xugaosengzhuan,gHfz; T. 2060
T = J. Takakusu and K. Watanabe (eds.), Taishoshinshudaizokyo,
Tokyo, 1922-1933 AdEfrfXi

Part I: The techniqueof ChineseBuddhisttranslations:an outlineof its

evolutionfrom thesecondto theseventhcenturyA.D.
One of the most interesting aspects of the history of Chinese
Buddhism is the progressive development of a highly specialized
technique of translation.
Since the earliest pioneering translations of the Late Han dy-
nasty, through the whole history of Chinese Buddhism, the ren-
dering of Buddhist scriptures was due to the collective work of
translation teams.2

1 I want to express my sincerest gratitude to Professor Tilmann Vetter and

Professor Erik Zurcher of Leiden University, who took the care of reading the
first draft of the present article, suggesting many corrections and improvements.
This article is a development of a part of my graduate thesis, The ChineseTransta-
tions of the Diamond Sutra (in Italian), Universita degli Studi di Venezia, unpub-
lished thesis, academic year 1992-1993. Research in the Netherlands during the
19961995 academic year was made possible by a grant from Venice University.
2 See Zurcher 1959, p. 31.

(C)EJ. Brill, Leiden, 1996 T'oung Pao LXXX:1I

138 STEFANO ZACCHETTI

The first documents describing one of these teams are two an-
cient colophons3 dating back to the Late Han dynasty.
The most detailed of the two4 says:
<<Banzhousanmeijing on the 8th day of the second month of the
second year of Guanghe era [179 A.D.], at Luoyang, the Indian
Bodhisattva5 Zhu Shuofo recited [it] (chu 1); at that time the
Yuezhi6 Bodhisattva [Loka]ksema was translator (chuan yan zhe
*=X) and transmitted [the translation] to Meng Fu, named
Yuanshi from Luoang in Henan, [and to] Zhang Lian, named
Shaoan, who, assisting the [two] Bodhisattva, wrote down [the
translation] (bi shou *) ... >>
As far as we can induce from the sources we possess (especially
colophons and prefaces collected in chapters VI-IX of CSZJJ), this
tri-functional structure of translation teams (i.e. 1. the main trans-
mitter who recited the original text; 2. the interpreter, who orally
translated it; 3. the scribes, who wrote down the translation)7,

3 1)Colophon of the Dao xingjing Ait (T. 224; the earliest Chinese transla-
tion of Astasahasrzka-Prajnaparamita suitra);in CSZIJ p. 47c 5-9. 2) Colophon of the
Banzhou sanmei jing Q4 $`If (T. 417/418; the earliest Chinese translation of
Pratyutpanna-samaidhisutra); in CSZIJ p. 48c 10-15. The two translations were
probably carried out at the same time by the same team of translators, led by
Lokaksema and Zhu Shuofo *M4 (see Zurcher 1959, n. 67 p. 330). The two
documents contained in CSZJJare in fact later copies of Han originals (the first
dates back to 255 A.D., the second to 208 A.D.; see Zurcher 1959, ibidem).
4 CSZJJ p. 48c 10-13. In quoting this passage, I follow the text emended and
punctuated by Tang Yongtong (1938 p. 68). The colophon also mentions two
later revisions of the text, carried out in 198 and 208 A.D. The other colophon
hardly differs from this one. See also Paul Harrison, The Samadhi of DirectEncounter
with the Buddhas of the Present:An AnnotatedEnglish Translation of the Pratyutpanna-Bud-
dha-Sammukhavasthita-Samadhi-Suitra, Tokyo, 1990, p. 259-260.
5 The term "Bodhisattva" was widely used as an honorific title in early Chi-
nese Buddhism, especially referring to foreign translators. See Tang Yongtong
1938 p. 102.
6 Indoscythian.
7 A main variant of this scheme happened when the main transmitter's com-
mand of Chinese enabled him directly to supply the translation (see Fuchs 1930
p. 87). The first explicit mention of this way of translating can be found in Kang
Senghui's t* preface to the Fa jing jing 9,W- (T. 322), which was translated
during the Han dynasty by An Xuan Wt and Yan Fotiao BR (see CSZJJ p. 46c
5-6). Many great transmitters used to translate without resorting to a translator:
such as Dharmaraksa (see Zurcher 1959 p. 69; a list of sources on Dharmaraksa's
translations can be found in Zurcher 1959, n. 221 p. 343; actually for the first text
produced by him Dharmaraksa had to resort to a translator: see below, note 8,
item 1); Sanighadeva (see CSZIJ p. 72b 23-24; p. 73a 23-24); Kumarajiva (see his
biography, CSZIJ p. lOlb 17; Sengrui's 1R85preface to Da pin jing *
 6) 382 A.D. Mohebanrao chao jing t;#fi; Dharmapriya, Buddharaksa;

DHARMAGUPTA S UNFINISHED TRANSLATION 139

hardly changed in the three centuries following the Han dynasty.8

Such a persistence of this particular organization is not mean-
ingless: surprisinglyenough, in this early phase of Buddhist trans-

CSZiJp. 53b 5); Buddhabhadra (CSZiJp. 61a 65; but compare with note 8 below,
item 14); Dharmaksema (CSZiJp. 59c 22-23; p. 64c 27-28); and Paramartha(see,
for instance, the preface of Abhidharmakosa T. 1559, vol. 29, p. 161b 12).
8 The following are the other documents (prefaces and colophons) of CSZtJ
in which this organization of translation teams is mentioned (I shall list the date
of each translation, its title, the names of the main translator and that of the in-
terpreter;page numbers only refer to the passages of these documents which spe-
cificallydeal with the subject.)
1) 266 AtD. Xa zherl tiarlzi jing gkt<i+#£; Dharmaraksa;An Mrenhui t<,#
and Bo YuanxinM 7EM;colophon p. 48b 22-26.
2) 291 A.D. Fangguangjing9t; Wuchaluo SE>M,; Zhu Shulan ^*0512; colo-
phon p. 47c 16-20.
3) 328 or, more probably, 373 A.D. Sho?llengyanjing ttt? Zhi Shilun
t**? Bo Yan Mi; colophon p. 49b 19-29 (see also Tang Yongtong 1938 p. 389-
391).
4) 379 A.D. (>) Biqiani dafie H6fi:figt.t,;Dharmadhl; Zhu Fonian ,;;
three documents (an initial 'inote9')a colophon and a final colophon); p. 81c 21-
22 (compare with CSZiJp. lOa 2S29).
5) 382 A.D. (?) Si Ahan muchaoE3Wt4g;; Kumarabodhi; Zhu Fonian and
Buddharak,sa; anonymous preface p. 64c 13-15 (see TangYongtong 1938 p. 224).

Dao An's A:t preface p. 52b 19-20.

7) 383 A.D. Abhidharmazzibhasa; Sanghabhadra, Dharmanandin (who wrote
down the Sanskrittext); Buddharaks. a; Dao An's preface p. 73c 3-8.
8) 383 A.D. Abhidharmasara; Sanghadesra;Zhu Fonian; Dao An's preface p.
72a 27-29.
9) 384 A.D. Poxamiji gtt; Sanghabhadra7Dharmanandin and Sang-
hadeva;Zhu Fonian7anonymous preface p. 71c 28-72a 5.
10) 384 A.D. Ekottaragama; Dharmanandin; Zhu Fonian; I)ao An's preface
p. 64b 10-14 (see also colophon of Sengiialuocha ji p. 71c 1-2).
1 1) 384 A.D. Sengyialuosha ji bgRt; Sanghabhadra;Buddharaks,a; colo-
phon p. 71b 25-271there is also an anonymous preface p. 71b 17-20 (slightly dis-
cordant: it states that Zhu Fonian was interpreter).
12) 397 A.D. Madhyamagama; Sangharaksa;Sanghadeva;Daocils XEMpreface
p. 64a 13-15^
13) 413 A.D. Dirghagama; Buddhayasas;Zhu Fonian; Sengzhao's * preface
p. 63c 16-18.
14) 417 A.D. Mahaparanirvana siitra;Buddhabhadra;Baoyun t; colophon
p. 60b 9-11.
15) 434 A.D. Za Apitanxin tk*.b; Sanghavarman;Baoyun; Jiaojing's
postface p. 74c 5-7.
16) 436 A.D. Sramaladevasimhanada sitra; Gunabhadra; Baoyun; Huiguan's
preface p. 67b 2-5.
140 STEFANO ZACCHETTI

lations, command of both languages (that of the Indian original,

and Chinese) does not seem to have been an essential require-
ment for the main transmitters (those names normally attached to
the translated texts in bibliographical sources). Instead, what re-
ally was deemed crucial was their doctrinal insight into the texts to
be translated. In fact, from the very beginning of the Sui dynasty,
exegesis played an important role in the work of translation, dur-
ing which the texts were normally commented upon by the for-
eign master presiding over the team.9
From the beginning of the fifth century A.D., particularly with
the great Indo-kuchean master Kumarajiva (350-413 A.D.), trans-
lation teams greatly increased in size, turning into assemblies that
were attended by great numbers of monks and lay devotees,10 and
Buddhist scriptures, before being translated, were deeply anal-
yzed and debated.11
However, toward the beginning of the sixth century A.D., there
gradually emerged the tendency to establish small-sized teams of
selected translators, under the direct control of the imperial
court. This phenomenon, which in some respects anticipated the
structure of Tang dynasty teams, was particularly evident in the
translations produced in the last decades of the Northern Wei dy-

9 See Tang Yongtong 1938 p. 296; Wang Wenyan 1984 p. 131-139; Zurcher
1959 p. 31.
10 Starting with Kumarajiva, the sources quite regularly record the number of
persons assisting the translations, and during the fifth century it is said often to
have reached several hundreds, and sometimes thousands; see for instance CSZIJ
p. 58a 9; 58b 12-13; 79c 2.
1 See Cao Shibang 1990 p. 100. The sources dealing with fifth-century transla-
tions sometimes mention this kind of group exegesis. See, for example,
Dharmaksema's biography (CSZJJ p. 103b 1-2), which states that during his trans-
lations "... Several hundred monks and laymen raised all kind of questions ..."; in
Huiguan's preface to the Fahua zongyaojing MY' 0;- (CSZJJ p. 57b 6-8), it is said
that "... In the Great Temple of Chang'an gathered more than two thousand
sramanas experts in exegesis from the four directions ..." to assist Kumarajiva's
translation of Saddharmapundarikasitra, which was "carefully scrutinized [by
Kumarajiva] with all [the present]" (yu zhong xiang jiu ORP). It seems that
those debates had the very important function of establishing (ding 2, a term of-
ten occurring in these sources) the definitive text of the translation, as we can
clearly see in Huiguan's preface to Srimazladevisimhanada suitra (CSZJJ p. 67b 3-5):
"... The foreign sramana Gunabhadra, holding in his hands the Sanskrit correct
text, recited the Sanskrit sounds ... Shi Baoyun and other monks of straight con-
duct, more than one hundred, checked the sounds and investigated the meanings
in order to establish the text [of the translation] (kao yin xiang yi yi ding jue wen
t~SH--iXtA 2m 3)
 DHARMAGUPTA'S UNFINISHED TRANSLATION 141

nasty (386-534 A.D.),12 when Buddhism had become the state reli-
gion.
With the rise of the Sui dynasty in 581 A.D. the imperial control
of the Buddhist church increased. This was also reflected in the
organization of translations: translation teams were directly associ-
ated with the imperial court, and the assistants of foreign masters
were carefully selected.13 Also the internal organization of the
teams apears to have evolved further in Sui times: the sources be-
gan to record a somewhat more complex division of labour14
which, after further improvement, became the distinctive feature
of Tang dynasty "new translations".15

12 The most representative among Northern Wei translators were Bodhiruci,

Buddhasanta and Ratnamati (see Bagchi 1927 p. 246-260), all active during the
first half of the sixth century A.D.. As Wang Wenyan (1984, p. 149) says, it is note-
worthy that in some cases the translators worked inside the imperial palace,
which very probably means that only a few specialists (monks or high-ranking offi-
cials) were admitted to do the work of translation. In fact, according to some
sources, as Cuiguan's preface to the DasabhuLmikasutra Adstra(translated in 508-511
by Bodhiruci, Ratnamati and Buddhasanta), T. 1522, vol. 26 p. 123b 1-5, and Tan-
ning's preface to the Samdhinirmocanasitra (translated in 514 by Bodhiruci), T.
675, vol. 16 p. 655a 16-b4, only "about ten persons" took part in the work of trans-
lation. Some translations carried out almost during the same period in Southern
China also share these features. See, for instance, the biography of Safighabhara
(active in 506-524 A.D. during the Liang dynasty), in XGSZ p. 426a 15-16.
13 When Yangdi (605-617 A.D.), the second Sui emperor, moved his capital to
Luoyang, he also founded, inside the "Park of the Upper Grove" (Shanglin yuan
j:t9), an "Office for translating the sutras" (Fan jing guan A ), "... selecting
persons of outstanding talent" (souju qiaoxiu 4* ). See Dharmagupta's biogra-
phy in XGSZ p. 435c 7-8. See also Wang Wenyan 1984 p. 150-151; Cao Shibang
1990 p. 105.
14 The most detailed description of Sui translation work is reported in
Jinagupta's (or, more probably, Jianagupta's) biography, XGSZ p. 434 a 26-b3
(see also E. Chavannes, 'Jinagupta (528-605 ap.J.C.)", in: T'oungPao 6, 1905, p.
332-356; part. p. 349-351). Jnianagupta himself acted as the main transmitter
("principal specialist" 7jG ), assisted by Dharmagupta; two brothers, the laymen
Gao Tiannu and Gao Heren, were translators; ten eminent monks "supervised the
work of translation and examined the original meaning"; two other monks,
Mingmu and Yancong, "compared again [the translation with] the Sanskrit text
(chong duifan ben14*)".
15 On Tang translation teams, see Cao Shibang 1990 p. 105-107. For example,
in the team led by Xuanzang in 645 A.D. (see his biography by Huili and
Yancong, Da ci'en si sanzangfashi zhuan ., b+ 1f4 T. 2053, vol. 50, p. 253c
19-254a 5) there were: 1) twelve monks in charge of "checking the meaning [of
the translated text]" (zhengyi ); 2) nine who "composed the text [of the trans-
lation]" (zhui wen g&l); 3) one expert in the "study of [Chinese] characters"
 STEFANO ZACCHETTI
142
The Sui dynasty thus was a
crucial point in the history of
lation technique. What we trans-
can ascertain about this
mainly based upon external evolution is
etc.). There is, however, a sources (biographies,
notable exception - a document colophons
allowsus a closer look at this which
subject.
Part II: Dharmagupta's
JGNDJ
The translaiion of
maguptal6 (JGNDJ)17isVafracchedika-Prajnaparamita-sutra
generally considered the most
by Dhar-
amongthe six Chinese
versions19of this Mahayanatext. literall8
goesfar beyond the normal In fact it
concept of literal translation: it
resemblesan interlinear version of fairly
the Sanskrit text. By way of
troduction,a short quotation will in-
bear evidence to this
statement:

(zi
xue t): 4) one who checked that
the original Sanskrittext was
fanyufanwen
ttR.). The main functions correct (zheng
was
carried out through the were the first two, and the
continuous cooperation of these two translation
16 Dharmagupta's groups.
biography is in XGSZ p. 434c
Chavannes 1903, p. 439-440; and Bagchi 23-435c 16. See also
122. 1938 p. 464-467; Ui Hakuju
1979, p. 112-
17In KYL(p. 552b 26) this
translation is quoted with a
duange
bantuoboruomijing iNlJ*JX different title: Jingang
18 Conze (1957 p. 4) t@£gt E.
says that"... Dharmagupta's
out
to reproduce the Sanskrit Chinese translation ... sets
text with great literal
Bougault,"Les paradoxes de la fidelity". See also Guy
coupure", Vajracchedika: une connexion qui
Cahiersd'etudeschinoises,8
(1989), p. 45-63;see p. 45. opere une
19Dharmagupta's
translation is the fourth in
by
thoseof Kumarajlva(T. 235), chronological order, preceded
followed Bodhiruci (T. 236), Paramartha
by those of Xuanzang (T. (T. 237), and
point 220 IX) and Yljing (T.239).
of view, JGNDJ represents From the textual
a late, rather expanded
Vafracchedika.
For instance, inJGNDJ a recension of the
of the Sutra, can alreadybe puzzling passage, not present in the earlier
versions
the found, which states, among
Tathagatais a synonym of other things, that
dharmoccheda(JGNDJp. 770 a 9-12;
present
in Xuanzang's translation S p. 48; it is also
T. 220 p. 983c 27-29). It is
that
Conze,in his edition of interesting to note
Vafracchedika,
"a addition which has crept into theomitted this phrase, holding that it was
later
see
alsosome critical remarks by text after 800 AD" (Conze
Schopen, 1957, p. 6;
97). In in L.O. Gomez andJ.A.
fact,JGNDJ provides a much earlier terminus Silk 1989, p.
ante quemof this addition.
 DHARMAGUPTA'S UNFINISHED TRANSLATION 143

(Sanskrit text)20 (JGNDJ)

Sacet Subhuite
if o Subhuti
tesam bodhisattvanamm
of these Bodhisattvas
mahasattvanam __1
Mahasattvas
dharma-samjnadpravarteta 2
dharma-notion could take place
sa eva tesam
that indeed of them
atma-graho bhavet ... 23
self-seizing would be

Without the Sanskrit text it would probably be impossible to

decipher this puzzling translation. Moreover, this by no means
seems to have been Dharmagupta's usual style of translating.24
The explanation of the peculiar features of JGNDJ is to be
found in biographical and bibliographical sources: no catalogue
quotes Vajracchedika-Prajiinparamita-sutra among Dharmagupta's
translations,25 and its first mention, in KYL (not in the catalogue

20 The Sanskrit text is quoted in E. Conze's edition (Conze 1957

p. 27-63;
hereafter S); page and line number are indicated. From this work (p. 65-92) I
have borrowed the translation of the quoted passages given in the footnotes. I
have also added to the Sanskrit examples a word-for-word translation which, I
hope, will help readers to checkJGNDJ quotations.
21 This straightforward rendering of pravrt- (here: "to take place", "to occur"),
clearly based on the fundamental meaning of the root vrt- ("to turn" etc.), is actu-
ally not peculiar to Dharmagupta alone. See for instance the corresponding pas-
sage in the Xuanzang translation (T. 220 p. 980c 22-23).
22 S p. 31, 20-22 (tr. p. 69: "If, o Subhuti, in these Bodhi-beings, these great be-
ings, a perception of a dharma could take place, that would be with them a seiz-
ing on a self ...")
23 767b 26-28.
24 KYL (p. 551b 21-c5) attributes nine translations to Dharmagupta: T. 310
(vol. 11 p. 571b-592c; part of the Maharatnakuta-suitra); T. 415 (vol. 13, p. 830a-
871c); T. 716 (vol. 16, p. 830-837); T. 449 (vol. 14, p. 401-404); T. 1510 (vol. 25, p.
766b-781a); T. 1660 (vol. 32, p. 517a-541b); T. 1596 (vol. 31, p. 271a-321a); T. 1652
(vol. 32, p. 482-486); T. 25 (vol. 1, p. 365a-420a). Though rather literal, all these
texts basically are written in the standard language of Buddhist translations.
25 See Da Tang neidian lu ;k*p T. 2149, p. 280a; KYL p. 551b-c; Da Zhou
kanding zhongjing mulu *MU5ItIgA T. 2153 p. 382a-b. See also Ono Gemmyo
(ed.), Bussho kaisetsu daijiten, Tokyo 1968-1975 (2d ed.), vol. III p. 499b; Bagchi
1938 p. 467.
which normally transcriptions are used in Buddhist textsn such as wenzhe for

144 STEFANO ZACCHETTI

bllt in the short biographyof Oharmagupta,p. 552b 26-28),says:

';Previously [Dharma]guptatranslatedthe LafracGhedika-Prajnapara-
mita-sutra,one juanand Pu lejin$26 fifteen juan, [which] could not
be polished.27At the time of the Zhengusurpers28[the two trans-
lations]were abandoned,and there was not sufficienttitne to re-
sTisethem. Nowadaystheirvariouspartsare preservedin the capi-
tal5'.29
JGNDJthus is an unfinished translation30which by a rare
chance, has been includedinto the BuddhistCanonand handed
downto the presentday.
If we take a closer look at this unique documentScomparingit
with the Sanskritoriginaltext7we can perceive,in spite of its ap-
parent incoherence, an underlyingrigorousformativeprinciple:
almostwithout exception, everysingle word of the Sanskrittext
has been translatedinto Chinese,31yet strictlykeeping the same

26 In a note following the biograph,v(p. 552b 29-cl), Zhisheng, the author of

KYL,states that this Pa le jing was a translation of the Lalatasistara,which already
had been translated into Chinese by Dharmaraksain 308 A.D. (see CSZDp. 48b
28; Pu yaojirtgi T. 186) and, later, in 683 A.D., during the Tang dynastyf was
translated once more by Divakara (Fangdeng da z11uangyanjing t#XZgk T. 187;
see Bagchi 1938 p. 50+505).
27 Lian fu **; perhaps 4;polished"is too vague, and fu § must be taken in the
sense of 4'toreverse [the text]", i.e. to arrange the syntax of the translation ac-
cording to Chinese grammar (see below).
28 WeiZheng^ refers to the ephemeral dynastyfounded at Luoyang by M;rang
Shichong X.1:3E(?-621 A.D.) after the fall of the Sui empire, and suppressed by
the Tang army in 621 A.D. See Sui shu 85 (ed. Zhonghua shuju, Beijing 1973, vol.
3, p. 1898);Jiu Tangshu 1 (ed. Zhonghua shuju, Beijing 1975, vol. 1, p. 11-12).
'i9 t1 X.+SiP9IJ*Jt§kt # > NciE tE > X*§ ' ^zi +
N:1$> #gtE\. It is noteworthy that this passage is quoted s7erbatimfrom
Dharmagupta'sbiography in XGSZ (p. 435c 17-18), the only difference being that
in the latter no mention is made of the ltafracchedika,but only of the Lalitavistara.
Since Luoyang was the center of the Zheng dynasty, this passage of KYLfits in
with an anonymous note following the title of JGNDJ in the Taisho edition (p.
766c), urhichstates that this translation dates back to the Daye era (605-617 A.D.)
of the Sui dynasty,just when Dharmagupta was active at Luoyang (see XGSZ p.
435cSand Chavannes1902, p. 440).
30 See also BTangWenyan 1984, p. 333. For a general description of JGNDJ,
see Ui Hakuju 1979>p. 122-127.
31 Generally speaking, the terminology of the JGDNJ is not too far from that
of the former translations of the Vayrac(;hedika.Nevertheless it shows a certain ten-
dency to make larger use of translation rather than of phonetic transcription of
Sanskrit terms. Thus sometimes we find in JGDNJ odd renderings of words for

Srarasta (p. 766c 1), probably based on an etymology deris7edfrom the root sru-
 DHARMAGUPTA S UNFINISHED TRANSLATION 145

position as in the Sanskrit phrase. This way of translating has pro-

duced an odd, obscure mixture of Chinese vocabulary and San-
skrit syntax. On the other hand, no word without an exact corre-
spondence in the Sanskrit text was added to the translation. Thus,
in order to maintain some basic distinctions expressed in the San-
skrit text, Dharmagupta sometimes had to use Chinese mor-
phemes as postpositional affixes (suffixes or even endings). In the
following examples I shall confine myself to some regularly recur-
ring phenomena.32
a) The form verb + ying 1E is systematically used in JGNDJ to
translate Sanskrit gerundives.33

Examples
1) ... Na sa Subhitte
not that o Subhuti

Bodhisattvo
Bodhisattva

vaktavyah... 34 35
is to be called

2) ... Nalaksana-sampadaa . .
not from mark-accomplishment

Tathagato
[the] Tathagata

(see M. Monier Williams, A Sanskrit-EnglishDictionary, Oxford University Press,

1899, p. 1098). It is interesting to note that Dharmagupta himself, in other "nor-
mal" translations, made use of a transcription to render Sravasti (Sheluoboxidi
t-MERtf); see for example T. 716 (p. 830b 1), which was translated by
Dharmagupta and his team at Luoyang in 606-607 A.D. (see its preface, p. 837b
10-12), more or less at the same time asJGNDJ. See also Ui Hakuju 1979 p. 125.
32 I shall not mention other usages in JGNDJ which, while probably based on
the same principle, normally occur in Chinese Buddhist translations, like -deng W
used as a pronominal plural ending (See Zurcher 1977, p. 184-185 and n. 24) or -
yi E in post-verbal position as a rendering of Sanskrit past participles or gerunds.
See Zhu Qingzhi, "Hanyi Fodian yuwen zhong de yuandian yingxiang chutan"
&&itmbtl3gFwiM1g, Zhongguoyuwen,1993/5, p. 379-385; see p. 381-382 and
p. 380.
33 See: p. 767a 3-4; 767a 8-9; 767a 10; 767a 14; 767a 16-17; 767a 19-21; 767b 1-3;
767b 6; 768a 25-27; 768c 7-8; 769a 16-22; 769b 14; 769c 16-23; 770a 24; 770a 29;
770c 3-4; 770c 7-8; 771a 9; 771a 16-17; 771a 19-20; 771a 23; 771a 27; 771b 4; 771c 3-
4.
34 S p. 29, 5-6 (tr. p. 66: "... He is not to be called a Bodhi-being...").
35 767a 17.
146 STEFANO ZACCHETTI

drastavyah ...36 37
is to be seen

b) In a similar way, Sanskrit verbs in the future tense are mostly

rendered by the form verb + dang a .38

Examples
3) ... Eka-citta-prasadamapi
One-thought-faith even

pratilapsyante ... 39 40
[they] will obtain

4) ... Te
paribhphtaa
....
They despised

bhavisyanti ... 41 42
will be

c) One of the most striking features of JGNDJ is the use of the

personal pronoun wo R after verbs, as a first person singular end-
ing.43

Examples
5) ... Katham cainam
How this too

dhrayami ... 45
[I] bear [in mind]

36
S p. 30, 8-9 (tr. p. 67: "... Not by the possession of his marks can the
Tathagata be seen").
37 767b 2-3.
38 See: 767a 7-8; 767b 21; 767b 23; 768b 20-21; 768c 19-20; 768c 23-27; 769a 3;
769b 3-4; 768b 6; 769b 15-16; 769b 18-19; 769b 25-29; 769c 1; 769c 7-8; 769c 14;
770a 3; 770a 7. Sometimes, however, we can find in JGNDJ dang ' in its normal
preverbal position; see S p. 30, 16: ... kecit sattva bhavisyanti ... ("... There will be
some beings ..."), andJGNDJ p. 767b 7-8: po you zhongshengdangyou M4*CR4.
39 S p. 31, 9-10 (tr. p. 68: "... Will find even one single thought of serene
faith
40
767b 21.
41 S p. 45, 1 (tr. p. 80: "... they will be
humbled ...").
42 769b 28-29.

43 See: 767c 6; 768a 22; 768b 7; 768b 24; 768c 22; 770a 23; 771a 17; 771a 23.
44 S p. 37, 21-24 (tr. p. 74: "... How should I bear it in mind?").
45 768b 24.
 DHARMAGUPTA S UNFINISHED TRANSLATION 147

6) ...Aham
I
ksetravyuthadn
field-arrangements

.......46~
ni.spadayi.sya-mi-iti ~~~~4 46
;X7
will accomplish unquote

It seems that no attempt has been made to render other end-

ings of Sanskrit conjugation in this way, using postponed pro-
nouns, except for one phrase:48

.Y,
7) ... Yathavadasi ...49 &50
as you say

d) Sanskrit locatives, even absolute, are often (though not regu-

larly) translated as substantive + -zhong 43,51 without any preposi-
tion.

Examples
8) ... Tasu Ganiganadasu...52 53
in those Ganges rivers

9) ... Imesv
these

sfutrantapadesu
evamrfupesu
of such a form sutra-words
54
....?55
bh&syama-nesu
being spoken

46 S
p. 35, 22 (tr. p. 72: "... I will create harmonious Buddha-fields ...").
47 768a 22. Note that in this case the future tense has not been rendered into
Chinese.
48 This phrase also occurs one more time in the Sanskrit text (S p. 28, 8) with-

out correspondence inJGNDJ, and another time in the latter (769a 2) but not in
the former.
49 S p. 56, 8 (tr. p. 88: "... As you say...").
50 771a 18.

51 See: 766c 1; 766c 6; 767b 9-10; 767b 12-13; 767b 15-16; 767b 20-21; 768b 4-5;

768b 6; 768b 17; 768b 21; 768c 20; 768c 22; 769a 3; 769a 9-12; 769a 14; 769c 6-7;
769c 13-14; 770b 17; 770b 19; 770c 17; 771b 7.
52 S p. 36, 19 ("... In those Ganges rivers ...").
53 768b 7-8.
54 S p. 30, 18-20 (tr. p. 68: "... When these very words of the sultra are being
taught ...").
55 767b 9-10.
148 STEFANO ZACCHETTI

e) In rendering verbs of saying, the indirect object is regularly

denoted by noun + bian &56 as in:

10) ... Bhagavan

The Lord

ayusmantam Subhutim
to "Life-possessing" Subhuti

etad avocat ... 57 g58

in this manner said

f) Adjective + guo A functions as a rendering of Sanskrit compa-

ratives.59
Examples
11) ... bahutaramr
More abundant

punya-skandham ...60 61
merit-amount

12) ... Asamkhyeyaihkalpair

During incalculable aeons

asamkhyeyataraih .. .62 63
more incalculable

Other peculiarities ofJGNDJ are due to the fact that apparently

all words of the Sanskrit text must have a corresponding term in
the translation. Thus some Chinese words are used in a quite unu-
sual way. Very common (recurring almost in every line) is, for in-
stance, the curious use of ruo X to translate (besides, more nor-
mally, sacetand va-) the relative pronoun yah.

Examples
13) Evam vastu-patito
So object-fallen

56 See: 766c 11; 767a 4; 767a 9; 767b 4; 767b 7; 767c 4; 768b 23; 768b 25; 768c

16; 769a 1; 769c 17; 769c 26; 770c 15; 771a 22.
57 S p. 28, 7 (tr. p. 66: "... The lord said to the venerable Subhuiti ...").
58
767c 3-4.
59 See: 767c 21; 768b 16; 768c 14; 769b 10; 769c 2; 769c 3; 771b 8; 771c 10.
60 S 58, 10-11 (tr p. 90: "... A greater heap of merit ...").
61 771b 8.

62
S p. 45, 7-8 (tr p. 81: "... During incalculable, quite incalculable aeons...").
63 769c 2.
 DHARMAGUPTA S UNFINISHED TRANSLATION 149

Bodhisattvo drastavyo
Bodhisattva is to be seen

yo
who

vastu-patito64
[having been] object-fallen

danam parityajati65 66
a gift gives

14) ... Na-asti dharmo

there is not a dharma

yas
which

Tathagatena desitah ... 68

by the Tathagata has been taught

Besides the features listed above (those which occur fre-

quently) no other "suffixes" or "endings" can actually be detected
in JGNDJ. This suggests that Dharmagupta is very unlikely to have
had the intention of systematically representing Sanskrit morphol-
ogy in Chinese. Those strange phenomena ofJGNDJ seem, on the
contrary, to have a more practical function: in fact he tried to
translate in this way only those morphological features which
mutatis mutandis (i.e. arranging the text according Chinese stand-
ard syntax) were normally represented in Chinese Buddhist trans-

64 This expression, which has been translated by Conze as "fallen among

things", is actually rather obscure. The translations by Kumarajiva (T. 235 p. 750b
29c 1) and Bodhiruci (T. 236, p. 754c 18; verbatim followed by Yijing T. 239, p.
773c 9), and Paramartha's more interpretative translation point at a
gnoseological meaning of vastu (= "object of perception"). Bodhiruci renders this
phrase as tj It& : a ffifT$ ("The Bodhisattva [who] devotes himself to giving
gifts, having his mind based on objects").
65 S p. 42, 15-17 (tr. p. 78: 'just so should be viewed a Bodhisattva who has
fallen among things, and who, fallen among things, renounces a gift"). It should
be noticed that all the Chinese translators of Vajracchedikahere interpret danam
parityajati as "[the Bodhisattva] gives a gift". In fact parityaj- seems to involve, in
this context, the idea of "renouncing (some goods in order to give) a gift". I am
indebted to Prof. Vetter for this suggestion.
66 769a 28-29.
67
S p. 32, 13-14 (tr. p. 69: "... There is no dharma which has been demon-
strated by the Tathagata .
68 767c 7-8.
150 STEFANO ZACCHETTI

lations. This is the case with ying f and dang X, which are often
used, in the normal pre-verbal position, to render the Sanskrit ge-
rundive and future tense.69 Hence "suffixes" and "endings" of
JGNDJ only had the very practical aim of preserving some linguis-
tic distinctions with respect to the final version of the translation.
Perhaps this could explain why no use of the "ending" -wo R has
been made, when the Sanskrit text itself has a first person pro-
noun, thus allowing Dharmagupta to insert R in its normal
preverbal position, without sacrificing the principle of word-to-
word translation (though with some exceptions: see above, exam-
ple n. 6: but compare p. 770a 29).

For example:
15) ... Tesam
Of them

aham

nanabhazvamcitta-dharam
the various thought-streams

prajanami 70 71
know

Another possible instance of "economy" of linguistic distinc-

tions made by JGNDJ is the use of shi UI and shi tit, which, when
69 See, for example, S p. 28, 3-5: the gerundive ... katham ... statavyam ...

How ... should stay ...") is translated inJGNDJ (p. 767a 2-3) as ... yunhe ... zhuying...
( fl ying.unhe
and as zhu ( fIW ) by all other Chinese translators of
Vajracchedika (see: T. 235 p. 749c 9, T. 239 p. 752c 12-13; T. 237 p. 762a 15; T.
220 p. 980a 15; T. 239 (p. 772a 5-6).
70 S p. 51, 17 (tr. p. 85: " ... Of them I know, in my wisdom, the manyfold
trends of thought").
71 770b 20. Theoretically one cannot exclude that, in this case, the absence
of the -wo R ending is due to variant reading in the original Sanskrit text of
JGNDJ: in fact the Gilgit manuscript of the Vajracchedikaat this point has a wrong
variant reading (Jadnyas,which does not agree with the subject aham, instead of
Prajeanami;see G. Schopen, "The Manuscript of Vajracchedika found at Gilgit", in
L.O. Gomez and J.A. Silk 1989, p. 103 and n. 17 p. 138) which, if it was also
present in Dharmagupta's own original text, given his way of translating the
JGNDJ, would have resulted in this rendering zhi 0. But this seems unlikely to be
the case: JGNDJ is based on an original text which often diverges from the Gilgit
manuscript (see also above, n. 19), and it is thus hard to believe that it could
share with the Gilgit manuscript a reading which, as Schopen says, is "clearly
wrong". After all translation, no matter how mechanical, always is a somehow
more conscious action than mere textual transmission.
 DHARMAGUPTA S UNFINISHED TRANSLATION 151

they directly involve a certain idea of temporal location, are nor-

mally used without a -zhong + ending,72 even when they render
Sanskrit locatives, as in:

16) ... ekasminsamaye... 73 74

at one time

Many features of JGNDJ, odd as they may seem, are explained,

as Wang Wenyan has pointed out,75 by a passage of the Buddhist
chronicle Fozu tongji ieE76 T 2035 (p. 398b; see also Fuchs
1930, p. 101-102), which describes the organization of the Song
dynasty "Translation Bureau" (yi jing yuan glg-),77 actually a di-
rect offshoot of the Sui and Tang translation teams.78
After having listed 1) the "director of translation" (yizhu 4t),
2) those who "checked the meaning [of the Sanskrit text]" (zheng
yi 1), 3) those who "checked the [Sanskrit] text" (zheng wen
i), and 4) those who "transliterated [the Sanskrit text into Chi-
nese characters]" (shu zi
J), the text says:
fifth were the scribes (bi shou 1), [who] translated Sanskrit
transcriptions into Chinese words; sixth [those who] composed the
text (zhui wen MSC),79 turning the Chinese characters in order to
form meaningful sentences ....80
A following note provides a further explanation:

"For instance, [the text drawn up by] scribes would say: zhaojian
wuyun bi zixing kong jian ci 8 Now [i.e.

72 See: p. 766c 3-4; 766c 7; 766c 9; 767b 8; 767b 12; 767b

14; 769a 8-9; 769a 11-
13; 769b 7-8; 769c 2; 769c 6; 770a 4; 770a 7. Some exceptions (mostly absolute
locatives) are: 767b 9; 769c 6-7; 769c 14; 770c 17.
73 Sp. 27, 1.
74 766c. 1.
75 Wang Wenyan 1984, p. 333.
76 By Zhipan , (13th century A.D.).
77 See T. 2035 p. 398a. The 0M was established in 982 A.D. at the capital
Kaifeng.
78 See Fuchs 1930 p. 103, and Cao Shibang 1990, p. 106-107.
79 Fuchs 1930 p. 101 renders it as "Textzusammensetzer".
80 T. 2035 p. 398b
12-14. L:EJh! EMM1Of* M91-CP a
81
Quotation from Prajnaparamitd-sitra; very likely renders the phrase:
vyavalokayatisma pancaskandhas tamsca svabhzvasunyan pasyati sma (quoted from E.
Conze, Thirty years of Buddhist studies, London 1967, p. 150); "He [i.e.
Avalokitesvara] examined the five skandhas and saw they were empty of own-be-
ing". Note how this sentence resembles those of JGNDJ!
 STEFANO ZACCHETTI
152
after having been
i> %i:M*a. On thecomposed] it says: zEaojian wayun jie kong
ted [in comparison wholeS the Sanskritword order is
with Chinese] ... so often inver-
need to be reversed in characters and sentences
order to be in
guage".82 accordance with Chinese lan-

JGNDJ clearly proves that

readyfully developed as this technique of
early translation was al-
as in the Sui dynasty.
tionhas been stopped The transla-
exactly at the stage which
calledbi shou ;S2. in Fozu tongi is
Therefore the Sanskrit
totallytranslated into
Chinese (even to a terminology has been
larger extent than in
normalcompleted
isstill that of the translations, see above, n. 31), while the syntax
original text.
Rather than a deliberately
constitutesa unique superliteral translation,JGNDJ
document for the history of thus
dhist translation technique. Chinese Bud-

BIBLIOGRAPHY
Bagchi
P.C. 1927, Le canon
bouddhique
Bagshi
P.C. 1938, Le canon en Chine,
tome Ier,Paris
bouddhique
Cao
Shibang WH:gl 990,
en Chine,
tome IIeme,Paris
Zhongguo Fojiao yijing
Chavannes
E. 1903, ;'Note sur shi lunji
, Taibei
en Chine avant divers ouvrages relatifs@g"FEt -
a l'Inde qui furent
l'epoque T'ang",B.E.F.E.O. 3, p. 430-441 publies
Conze
E. 1957
duction Vajracchedika-Prajnaparamita-sutra edited and
and glossawy, Roma, ISMEO, Serie translated with intro-
Edgerton F. 1953, Buddhist orientale vol. XIII
HSb77d Sanshrit
Haven Grammar and Dictionary, 2 vols., New
FuchsW 1930 "Zur
technischen Organisation der
tischer Schriften ins Ubersetzungen Buddhis-
GomesL.O. and J.A. Silk Chinesische", Asia Major
(eds) 1989, Studies in the
6^ p. 8s103

Mahayana Literature of the Great

Buddhist Texts, Ann Arbor Vehicle: Three
Tang
Yongtongj%,lb 1938, Han Wei
liangJin
Ft, Shanghai nanbei chao Fofiao shi,
iwF#1k#J
Ui
Hakuju FtSbX 1979, Daijo Butten no
Wang
Wenyan:ESig 1984, Fodian Kenkyu
y<***tOXfft,, 2nd ed., Tokyo
YuLiming *2> 1993, Fodian wenxian
Hanyi zAi yanjiu
2wt,, Taibei
yuyan
Zurcher
E. 1959, The Buddhist Ct<; f « ,
Chengdu, Bashu shushe
Conquest of China,
Zurcher
E. 1977, '4LateHan Leiden, EJ. Brill
Vernacular Elements in the
Earliest Buddhist Trans-
177-203
lations",JournaloftheChineseLanguageTeachersAssociation, 13,3 (Oct. 1977),p.

822035
T p. 398b 1615:
tH tEei; >M>E:ZE84ttZ
+tk**- -agEStt t t1XWt+29.