Discerning
the Original
Language
of the Tibetan
Versions
of Mahayana
Sutras
From a Simple Mistake in the lDem kar ma Regarding the
Ajatasatrukaukrtyavinodhanasutra
MIYAZAKI Tensho
1 Introduction
An excellent historiographer may sometimes make a mistake. Once the error is
compiled in an influential catalogue, it can long be accepted as historical fact,
in-deed until it is noticed and corrected. The account of the
Ajatasatrukaukrtyavinodh-anasutra (AjKV) given under the caption of "Mahayana sutras translated from
Chinese" in the lDem kar ma (Dk),1) one of the oldest and the most authoritative
Tibetan catalogues of Buddhist canonical scriptures, is most likely to be one such
case.
What strikes us first about the understanding of the origin of the Tibetan
ver-sion of the AjKV is the discrepancy found between the Dk and the 'Pang than ma
(Pt).2) The two are almost contemporary (around the first half of the 9th century),
but unlike the Dk, the Pt classifies the sutra among "sutras translated from
Indian."3) Since there seems no doubt that both catalogues refer to the same extant
AjKV,
composed of five fascicles, the truth will lie either in the account of the Dk or
in that of the Pt. To examine which account is more credible, this paper will
dis-cuss the subject from the following three viewpoints: first, what the colophon
sug-gests, second, what Chinese materials suggest, and third, the use of relatives in
Tibetan translations.
2 What the Colophon Suggests
The colophon of the AjKV mentions two scholars as those engaged in the translation
of the sutra: Manjusrigarbha, an Indian monk, and Ratnaraksita, a Tibetan monk
re-ferred to in the Sgra sbyor bsam po gnyis pa.4) Neither scholar leaves any sign of
having translated material from Chinese scriptures. Given that specialists well
versed in Chinese Buddhist scriptures were obviously needed in the projects
involv-ing translation of sutras from Chinese, the extant Tibetan version is therefore clearly
not taken from a Chinese original. 3 What Chinese Materials Suggest
Since each of the extant three Chinese versions of the AjKV, i.e. the Asheshi wang jing阿 闇 世 王 經(T626), the Wenshuzhili puchao sanmei jing文 殊 師 利 普 超 三 昧 經 (T627), and the Weicengyou zhengfa jing未 曾 有 正 法 經(T628), shows no literal
agree-ment with the extant Tibetan version, we are left with the possibility that either the Dk is erroneous or, if the record is in fact correct, that the text mentioned in the Dk could represent a missing Chinese version of the sutra. The Kaiyuan shijiao lu開 元 釋 教 録(T2154,730 CE), in fact, argues that"the AjKV has three extant versions and three missing versions."5) With whatever little certainty we attach to this source, it cannot be completely denied that the sutra referred to by the Dk may have been among these three missing versions. In addition, there is another possibility that a specific text translated into local, vernacular Chinese, thereby escaping the attention of the compiler of the Chinese catalogue,6) might have been made use of by Tibetan scholars and listed in the Dk.
However, how should we examine such a possibility that "an extant text" could have been translated from "a missing text?" The examination of the description in the Dk conducted thus far has involved no more than circumstantial observations, and we will remain in a fog unless we change viewpoints. We will now move on to the third, and final, perspective that points to "the style of translation" as direct evi-dence of the original form of the sutra.
4 The Use of Relatives in Tibetan Translations
One of the most effective indices with regard to "style" to distinguish Sanskrit from Chinese is the use of relatives; Sanskrit has highly developed relative structures that are completely lacking in Chinese. Tibetan translators, while originally unfamiliar with the use of relatives but later introducing them through the process of transla-tion from Sanskrit, felt the right to adopt a different style of translatransla-tion in terms of
a relative structure in accordance with the stylistic differences of the original texts. The use of relatives can effectively function as the method for discerning the origi-nal of Tibetan translations.7) However, given the complete lack of philological stud-ies on this theme, a couple of examples will be shown by quoting from the Mahayana sutras before proceeding to examine the case of the AjKV
(1) the Purnapariprccha and the Suvarnaprabhasa
Two Mahayana sutras are regarded as having been translated from Chinese: the
*Purnapariprccha (D61, PurP)8) and one of the Suvarnaprabhasas (D555, Suvch>tib).9)
In these sutras, while a certain kind of relatives proved to have originated from
Chinese does appear, another attested to in Sanskrit never does.
(a) The expression "gang yin pa de ('di) ni," corresponding to the expressions "ming
Yi"名 為,"ji shi"即 是,etc., as shown below, is an example fbr the case in which a style that had originally been devised for the purpose of translating from Indian originals was applied at a later time to translation from Chinese tinged with the nu-ance of "paraphrase" or "emphasis."
・de lta bu'i dmigs pa med pa'i byams pa gang yin pa de ni chos kyi byams pa zhes bya'o //de lta bu'i chos kyi byams pa gang yin pa de ni sangs rgyas kyi byams pa zhes bya'o// 如 是 無 所 得 慈 名 爲 法 慈 。如 是 法 慈 名 爲 佛 慈 。10)
・yang dag pa'i ngo bo nyid gang yin pa de ni de bzhin nyid yin la de bzhin nyid kyi ngo bo nyid gang yin pa de ni de bzhin gshegs pa yin yongs mya ngan las'das pa zhes bya'o// 實 性 體 者 即 是 眞 如,眞 如 性 者 即 是 如 來,名 爲 涅 槃 。11)
(b) The PurP and the Suvch>tib
include a relative clause of place, "ga la ba der,"
cor-responding to "yena-tena"
in Sanskrit. Here I present an example from each text.
・de nas bras ze zhig to spul to/rgyal po de ga la ba der son nas 時 即 化 作 婆 羅 門,往 詣 王 所12)
・de dag gzhon nu tham cad (…) phyi dro'i dus na bcom ldan 'das gang na ba der lhags te/
如 是 等 人(…)各 於 哺 時 往 詣 佛 所 頂 禮 佛 足,13)
(c) The verses of Suvch>tib
include relative pronouns, which are used to meet the
re-striction of syllables in verses.
/sems can gang dag skyabs dang dbung gnyen med//mgon med sdug bsdal song bar
gyur ba dag//ma 'ongs skal ba kun du de dag la//mgod dang skyabs dang dpud gnyen
gyur par shog/
若有衆無救護 長夜輪迴受衆苦 我於來世作歸依 令彼常得安隠樂14)
(d)
In stark
contrast
to
the
use
of relatives
in
the
examples
mentioned
in(a),
(b)and
(c),
many kinds
of relatives
do not appear
in the
Suvch>tib
and the
Pure.
As one of
these
examples,
we enumerate
here
"relative
clause
of time",
e.g. "gang
gyi
tshe-de'i tshe" or "nam zhig-tshe-de'i tshe," common phrases corresponding to "yada-tada" in translations from Sanskrit originals. Probably for the reason that this struc-ture, when translated, is difficult to retain in Chinese, the lack of its appearance can be used as an index that suggests the original should be Chinese. An ideal example for this case can be drawn from the Suv, which has two Tibetan translations, one from Chinese and another from Sanskrit, as shown below.
Skt yadd samudrika navah sayantrah sapatakikah/ sthalam aruhya gaccheyus tada dhatur bhavisyati//20//
Skt>Tib/nam zhig rgya mtsho'i gru bo the//'khrul 'khor dang ni gyor mor bcas/ /thang la byung ste 'gro ba na//de'i tshe ring bsrel yod par gyur/
T655假 令大 船舶 盛満 諸財 實 能令 陸地行 方 求佛舎 利
Ch>Tib/gal to gru chen thams cad las//nor dang rin chen rab bkang ste/ /thang bu skam la 'gro nus na//bcom ldan ring bsrel btsal bar bya/15)
(2) AjKV
In the extant Tibetan translation of the AjKV, there are many kinds of relative
struc-tures, one of which is "the relative clause of time" as mentioned in (d). It will
suf-fice to quote here an example from Sanskrit fragments fortunately preserved in the
Schoyen Collections.16)
Skt yada ca maitreyena
bodhisatvena
bodhih prapta bhavisyati
tatra esa punar eva tatas
sahayam lokadhatau upapadyisyati akhyatavi(…)
Tib gang gi tshe na byang chub sems pa sems dpa' then po byams pa byang chub mngon par rdzogs par 'tshang rgya 'gyur bar 'gyur pa de'i tshe yang(…)byang chub sems pa sems pa chen po mi gyo ba zhes bya bar 'gyur ro//17)
5 Conclusions
The conclusions reached above were that, first, the colophon suggests that the trans-lators of the extant AjKV had no involvement in the projects translating from Chinese texts, second, the existing Chinese AjKV shows no literal agreement with the Tibetan versions, and third, the distinctive condition of Chinese originals lack-ing the clause of time is not satisfied by the extant AjKV Thus, the classification employed by the Pt that must be correct and it is highly likely that the compiler of the Dk made a simple mistake in recording the name of the sutra.
1) Dk: M. Lalou [1953] "Les Textes Bouddhiques au Temps du roi Khri-sron-lde-bcan," Journal Asiatique #241-3, pp.313-353, S. Yoshimura [1950] "The Denkar-Ma, An oldest Catalogue of the Tibetan Buddhist Canons." AjKV Lalou [1953] No. 257, Yoshimura [1950] No. 256. 2) Pt: E. Kawagoe [2005], Mar chang 'Phang than ma. 3) Kawagoe [2005] No. 74 4) M. Ishikawa [1990] A Critical Edition of the Sgra sbyor bam po gnyis pa, Studia Tibctica #18 p.1. 5)三 存 三 欠(T2154 629b) 6) D. Ueyama [1968] "A Study on the Life and Works of Fa-ch'eng (Hgo Chos grub), a Translator of Buddhist Texts at Tun-hunang under the Tibetan Rule. Part II" (in Japanese), Journal of Oriental Studies #39, pp.119-222. 7) S. Harada [1985] "Toban Yakkyo Shi"吐 蕃 訳 経 史Tonko kogo kenkyu and Ueyama [1968] etc. 8) PurP is not listed under the translated texts from Chinese in any catalogue, but it is regarded as the retranslated text from Da baojijing Fulouna hui大 宝 積 經 富 樓 那 會(T310(17)) as the previous studies indicates (e.g. B. Sakurabe [1930] "Sei-zoyaku Daihoshakukyo no kenkyu"西 蔵 訳 大 宝 積 経 の 研 究and Ueyama [1968], etc.). 9) Suvch>tib was translated from Jin guangming zuisheng wang jing金 光 明 最 勝 王 經(T665) by Fa Cheng法 成. The other two Tibetan versions of Suv (D556 and D557) were translat-ed from Indian, and the Sanskrit version of Suv (Suvskt) survives. Suvskt: Nobel, J [19371 Su-varnaprabhasottamasiitra. 10) D61 173a5-6=T.310(17) 436b15-16. Other examples: D61 184a4-5=T310(17) 440c15-16; D61 188b7=T310(17) 442b22, etc. The examples of gang yin pa 'di ni: D61 170b3-4=T310(17) 435b8-9; D61 173a4-5=T310(17) 436b14, etc. 11) D555 30b6=T665 407b2-4. Other examples: D555 34b5-6=T665 408b26-28; D555 39b3=T665 410a19-20; D555 79b5-6=T665 425b22-24, etc. 12) D61 212b2=T310(17) 452a26-27. Another example: D61 177a1-2=T310(17) 438a24-25. 13) D555 21b1--3=T665 403c3-5. Other examples: D555 26b5=T665 405c9-10, etc. 14) D555 74a2-3=T665 423b1--2. Other examples: D555 46b4-6=T665 413a11-14; D555 72b4=T665 422c5-6, etc. 15) Su-Vskt p.17 3-4 D556 159b5=T655 406b24-25=D555 28b6. 16) AjKVskt: Harrison and Hart-mann [2000] "Ajatasatrukaukrtyavinodhanasutra," Manuscripts in the Schoyen Collection II: Buddhist manuscripts, vol. I & Harrison and Hartmann [2002] "Another fragment of the Ajatasatrukaukrtyavinodhanasutra," Manuscripts in the Schayen Collection III: Buddhist manuscripts, vol. II 17) AjKVskt No. 12c=P882 273b3-4=D216 261b3-4. Another example: AjKVskt No. 13c=P882 274b4-6=D216 262b13-4.
〈keyword〉 Ajatasatrukaukrtyavinodhanasutra, lDem kar ma, 'Pang than ma, the use of relatives in Tibetan translations
(Graduate Student, Tokyo University)
