Introduction
In his complete catalogue of texts and translations of the Khotanese manuscripts in the British Library, Professor Prods Oktor Skjœrvø not only transliterated all the Khotanese texts but also provided translations in English. Using all means available to him, he attempted to identify the texts as much as he could.Footnote 1 Since Professor Skjœrvø published his catalogue in 2002, not much identification of the remaining unidentified Buddhist texts has been made. In 2004 Professor Yoshida Yutaka announced his discovery of four fragments (Or. 6402B/2.1+3, Or. 6402B/2.4, IOL Khot 172/4 and IOL Khot 172/5) of Raśmivimalaviśuddhaprabhādhāraṇī in the British Library,Footnote 2 for which he found Chinese parallels in the Taishō canon (Taishō shinshu daizōkyō the modern Buddhist canon, hereafter abbreviated as T., which has been familiar to the scholars in the field). This discovery supplemented his earlier identification of a fragment (SI M 44.1) from the same text in the St Petersburg Malov collection when he wrote a book review for Saka Documents Texts Volume III in 1997. Professor Yoshida noted that the fragment (SI M 44.1) reminds us of the crucial role of the Buddhist stūpa-worship practice in Khotan.Footnote 3 Still, many fragments in Khotanese in the British Library have remained unidentified. In this article, I will identify three more fragments in the British Library which have parallels in the Chinese Buddhist canon. I will also explore the ways in which we can relate these newly identified Khotanese fragments and their Chinese parallels in the context of medieval Buddhist history. I hope that the brief discussion here can cast new light on our understanding of the connections between Khotanese Buddhism and Chinese Buddhism.
As Professor Skjœrvø noted, his translations are meant to facilitate other scholars’ access to these materials. Indeed, thanks to Skjœrvø's generosity in offering English translations for all the Khotanese texts, I am grateful, as a scholar who has no knowledge of the Khotanese language, to have had access to these valuable resources which assisted my identification of several fragments. My identifications are based on his translations. First of all, I would like to offer my identifications of three more fragments, including IOL Khot 25/4,Footnote 4 IOL Khot 147/5 (H. 147 NS 106) Footnote 5 and Khot missing frags. 3.Footnote 6 The first fragment IOL Khot 25/4 is part of Agrapradīpadhāraṇī, which has a parallel in the Chinese version titled Dongfang zuisheng dengwang rulai jing 東 方 最 勝 燈 王 如 來 經 in T. no. 1354, vol. 21, translated by the Gandharan monk Jñānagupta (Ch. Shenajueduo 闍 那 崛 多, 524–600). The second fragment IOL Khot 147/5 (H. 147 NS 106) from a Khotanese text corresponds to the Chinese version in T. no. 300, Da fangguang fo Huayanjing busiyi jingjiefen 大 方 廣 佛 華 嚴 經 不 思 議 佛 境 界 分 (Mahāvaipulya-buddha-Avataṃsaka-sūtra-acintya-visaya-pradesa), translated in 689 by the Khotanese monk Devendraprajña (Ch. Tiyunbore 提 雲 般 若) the fourth Chinese text translated by this monk for which we now know that a Khotanese version exists. Khot missing frags. 3 can be identified as Hastikakṣyā (or more accurately Hastikakṣya-nāma-mahāyāna-sūtra, Tib. Glaṅ-poḥi rtsal in Tōhoku 207, Ch. Foshuo wu xiwang jing 佛 說 無 希 望 經 in T. no. 813, translated by 竺 法 護, 265–313; and Foshuo xiangye jing 佛 說 象 腋 經 T. no. 814) and its Chinese parallel is found in T. no. 814, a text translated by a Tocharian monk from Kashmir, Dharmamitra (Ch. Tanmomiduo 曇 摩 蜜 多, 356–442) in the Qihuan Temple 祇 洹 寺 in Jianye (now Nanjing) in 441.Footnote 7 In the following discussion, besides offering my English translations of the Chinese parallels, I will also elaborate on the context of Buddhist history in which we can build the connections between these fragments and their Chinese parallels.
The Khotanese Fragment IOL Khot 25/4
As I have noted above, this fragment is from Agrapradīpadhāraṇī and the Chinese parallel is in T. no. 1354. Skjœrvø's catalogue has provided sufficient information for us to understand where this fragment originally came from. This fragment is currently preserved in the Stein collection in the British library. The original number of this fragment, Kha. i. 135a4, indicates that it was collected in Khadaliq by Aurel Stein during his second expedition. As Stein remarked in his meticulous archaeological report Serindia, in September 1906 he left Khotan for the Domoko oasis where he excavated at Khadaliq and uncovered numerous manuscripts at the ruined shrine Kha.i. Skjœrvø also notes that there might be a similar manuscript numbered SI P 51.1 (pl. 29) in the St Petersburg Petrovky Collection. However, for this fragment, we cannot find the corresponding passage in T. no 1354. It is unlikely that it belongs to the same text as IOL Khot 25/4. The Khotanese fragment IOL Khot 25/4 was first published as a preliminary transcription by Harold W. Bailey in the appendix to his work Khotanese Texts V (1963, pp. 381–382), but Bailey did not offer a translation. Skjœrvø's reading is slightly different from Bailey's, but he offers an English translation that sets the basis for my identification.
I will cite Skjœrvø's transliteration and translation of this fragment first and offer English translations for the sections from its Chinese parallel in the Taishō canon. Then I will discuss the similarities and differences between the Khotanese version and the Chinese one. According to Skjœrvø's catalogue, this Khotanese fragment reads as follows (p. 224):
a 1 /śt- te kamalu hā nä netä gyastu balysu x /
2 dāra?]jñinu pata hvāñīmä uysnorā[ṇu (Bailey's transliteration does not have -ṇu)
3 / na buljsiye käḍäna ttīśä käḍäna. dätä käḍäna /
4 / nu käḍäna. ai'metätu (this word has “independent” ai-, i.e., -e vowel mark on top of independent e-)
b 1 / ttauñāte kye tta nāmi ye t- n- /
2 / -ắ o yi ūśu nāsāte. O yi va x /
3 / ttītä vā manyuśrī alysānai ā[ysanna (Bailey's transliteration stops at ā-)
4 pra]hoṣṭe hvarandau ysānū śaṃdya västāte. gyast[ānu (Bailey's transliteration is limited the section not beyond] [)
Skjœrvø translates this fragment as follows:
a 1 . . . he *bowed the head toward the Lord Buddha . . .
2 . . . I speak the padas of the *dhāraṇīs [for the benefit of] beings . . .
3 . . . on account of praise, on account of splendor, on account of look(s) . . .
4 . . . on account of . . . the dwelling . . .
b 1 . . . fattens . . . whose name was this . . .
2 . . . or will take his strength or . . . his . . .
3 . . . Then the youth Mañjuśrī [rose from the] *seat, draped
4 [his garment over the left shoulder], placed the right knee on the ground, [and spoke to the Buddha, Lord of] Lords . . .
I translate the similar section (T. vol. 21, no. 1354: 869a18–869b12) in the Chinese version of Agrapradīpadhāraṇī as follows and I have italicised the paralleled lines:
At that time Ajita Bodhisattva 阿 逸 多 菩 薩 rose from the seat, draped his garment over the left shoulder, placed the right knee on the ground, and spoke to the Buddha [偏 袒 右 肩] 右 膝 著 地, [而 白 佛 言] 世 尊, “The World-Honoured One, I also speak the padas of the dhāraṇīs for the benefit of beings 我 亦 說 陀 羅 尼 章 句, 為 眾 生 [受 利]故, for achieving progress, for improving merits, for improving prabhāva, rupa, and yaśas, for improving bāla [得 增 益 故] 增 功 德 故, 增 威 力 增 色 [增 名 聞 故], for following practice and receiving joy, for following the practice and receiving the assisting joy, for concentration without disturbance, for no killing (Ch. bushahai 不 殺 害, Skt. pāṇātipātāveramaṇi), and for protection, and therefore I speak the following spell:
. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . ..(I omit a section of spells here).
At that time, the Buddha told the Elder Ānanda, “You receive these padas of the dhāraṇīs and often withhold, read and recite, explain to the others, preach and transmit, and disseminate it. Ānanda, the Buddha appeared in this world, which is difficult to encounter. Listening to these dhāraṇīs is even many times more difficult than [encountering the Buddha]. If there are good men and good women who can also receive and withhold these dhāraṇīs, and after having received they can withhold, read, and recite (Ch. shou chi du song 受 持 讀 誦), and they can preach, transmit, and explain it to the others, they will be able to know what will happen over the next fourteen ages in future. Ānanda, these padas of the dhāraṇīs were spoken by the 8,400 billions of Buddhas in the past. If someone slanders (Ch. feibang 誹 謗, Skt. apavāda) these dhāraṇīs and does not believe and practice them, this person also slanders those Buddhas in the past. He betrays and abandons (Ch. beishe 背 捨) the Buddhas. Or if some beings (Ch. zhonglei 衆 類) take the vitality (Ch. jingqi 精 氣) from someone among us (Ch. moujia 某 甲) 若 有 眾 類 奪 我 某 甲 精 氣,Footnote 8 or do not return it after having taken it, this person also acquires this guilt. At that time, Mañjuśrī, the youth (Ch. tongzi 童 子, literarily the boy) 爾 時 文 殊 師 利 童 子, rose from the seat 即 從 座 起 整 衣 服, draped his garment over the left shoulder, placed the right knee on the ground 偏 袒 右 肩 右 膝 著 地, and spoke to the Buddha, “The World-Honoured One (Ch. shizun 世 尊), I also speak the padas of the dhāraṇīs for the benefit of beings . . .”
Although the Khotanese text is very fragmentary, several lines in this text can be found in their corresponding sentences in the Chinese text T. no. 1354. Clearly the Khotanese fragment can be identified as the same text as the Chinese version. Looking at the picture of this Khotanese fragment in the database of the International Dunhuang Project, it would seem that the dhāraṇī section is missing. Compared to its Chinese parallel, this Khotanese manuscript is a very fragmentary piece. I am not sure why in the Khotanese text one line appears as “whose name was this. . .”; this seems not to correspond to the lines in the Chinese version on Buddhas.
In Chinese Buddhist literature there are five Chinese versions (T. no. 1351, 1352, 1353, 1354, and 1355) for Agrapradīpadhāraṇī, including two attributed to Jñānagupta, T. no. 1353 (Dongfang zuisheng dengwang tuoluoni jing 東 方 最 勝 燈 王 陀 羅 尼 經) and T. no. 1354. T. no. 1353 was preserved in the Korean canon, which has been omitted from the Song, Yuan, and Ming editions of Chinese Buddhist canon. Although the Taishō canon has incorporated this title and listed it as T. no. 1353, this title does not appear in any Buddhist catalogue in traditional China. I suspect that it was a version wrongly attributed to Jñānagupta. It could be a version translated between the Eastern Jin Dynasty and Jñānagupta's version. Concerning the sentences on Mañjuśrī in the Khotanese version and in T. no. 1354, they are the same. In T. no. 1353, Mañjuśrī has an epithet as “the royal prince, Bodhisattva, Mahāsattva 法 王 子 菩 薩 摩 訶 薩”, and when he spoke to the Buddha, he also joined his palms and crossed his hands (hezhang chashou 合 掌 叉 手), which is very different from the description of his behaviour in the Khotanese version and T. no. 1354. Another difference is that the role of Ajita Bodhisattva in T. no. 1354 was replaced by Maitreya in T. no. 1353. Yet we cannot tell what it would be in the Khotanese fragment.
Now let me briefly note other Chinese translations. The earliest extant translation in T. no. 1351 (Chiju shenzhou jing 持 句 神 咒 經) was a short text translated by Zhi Qian 支 謙 in the third century, which does not even mention Mañjuśrī. T. no. 1352 (Tuolinnibo jing 陀 鄰 尼 缽 經) was translated by Zhu Tanwu Lan 竺 曇 無 蘭 in the Eastern Jin Dynasty. We cannot find Mañjuśrī in this version either. This version can also be found in T. no. 1336, but the latter was attributed to the edition from the Liang dynasty, with an unknown translator. In terms of content, although both T. no. 1351 and T. no. 1352 have three sections, they are slightly different. However, some vocabulary in T. no. 1352 has been altered from that in T. no. 1351 but can be found in T. no. 1353 and no. 1354. For instance, in T. no. 1351, the phrase “the padas of the dhāraṇīs” appeared as “the spells of holding verses” (chijuzhou 持 句 呪), yet in T. no. 1352, it appeared as “dhāraṇī padas” (tuolinnibo 陀 隣 尼 鉢), which is an exact transliteration. In T. no. 1353, it became “dhāraṇī verse” (tuoluoni ju 陀 羅 尼 句) ,half transliteration and half translation, which in T. no. 1354, became “dhāraṇī chapters and verses” (tuoluoni zhangju 陀 羅 尼 章 句). However, the most striking difference between T. no. 1351 and T. no. 1352 is that in the second section, T. no. 1351 said that the padas of the dhāraṇīs were spoken by 8.4 billions of Buddhas, but in T. no. 1352 this was altered to 8,400 billions of Buddhas, which apparently was inherited by T. no. 1354. Interestingly, T. no. 1353 also used 8.4 billions of Buddhas. The last version T. no. 1355 (Sheng zuishang dengming rulai tuoluoni jing 聖 最 上 燈 明 如 來 陀 羅 尼 經) was translated by Shihu 施 護 in the Song dynasty.Footnote 9 This version looks different from the Khotanese version because the sentence referring to Mañjuśrī's action indicates that Mañjuśrī joined his palms with respect and spoke to the Buddha, which appeared in T. no. 1353. But in this Song version, the title for Mañjuśrī is same as that in T. no. 1254.
Jñānagupta was not a stranger to Khotan. He actually stayed in Khotan for a long time. According to his biography and the catalogues compiled in the Tang dynasty, he translated the version in T. vol. 21, no. 1354 during the Kaihuang period in the Sui dynasty. His biography in the Continued Biographies of Eminent Monks (Xu gaoseng zhuan 續 高 僧 傳) says that he originally was from a Kṣatriya family in Gandhara in Northern India. He was ordained by a Khotanese master Jinayaśas
那 耶 舍. He travelled with his Dharma friends around many areas of Central Asia, including Khotan. He was stopped in Khotan by the rain in summer and the snow in winter. Later on he passed by Shanzhou 鄯 州 (now Shanshan) and arrived in Chang'an in the Wucheng period (559–560) of the Northern Zhou dynasty. He went to Yizhou 益 州 (now Chengdu) with Yuwen Jian 宇 文 儉 for a while and became the head of the monastic order in Yizhou, residing at the Zengyuan Temple 增 淵 寺. When Yuwen Yong came to power, he ordered Jñānagupta back to Chang'an and asked him to obey the Confucian rites. But Jñānagupta refused his order and left Chang'an for the Turkic Empire via Ganzhou 甘 州 (now Zhangye) in 574. In 584, he returned to Central China and was ordered to translate Buddhist texts in the Imperial Library by the Sui Emperor Wendi, who was willing to promote Buddhism These texts were mostly the Sanskrit texts brought back from the Western Regions in the late Northern Qi dynasty by eleven monks, including Baoxian, Daosui, Zhizhou, Sengwei, Fabao, Sengtan, Zhizhao, Sengü and so forth. From 587 to 592, Jñānagupta also helped Narendrayaśas 那 連 提 耶 舍 to translate texts in the Guangji Temple. In 592, Jñānagupta then began his translation enterprise at the newly established Great Xingshan Temple 大 興 善 寺 in Chang'an. T. no. 1354 was translated during his stay at the Great Xingshan Temple. Although the exact year for the translation of this text was not attested in any catalogue, it first appeared in the Catalogue of Assembled Scriptures (Zhongjing mulu 衆 經 目 錄) compiled by Fajing 法 經 in 594. Thus it should be safe to say that it was translated between 592 and 594.
The Khotanese Fragment IOL Khot 147/5 (H. 147 NS 106)
The second fragment I discuss here is the fragment numbered IOL Khot 147/5 (H. 147 NS 106) in the British Library. According to Skjœrvø's catalogue, A. F. R. Hoernle received a parcel on 22 December 1905 from J. H. Marshall who sent it on 24 November 1905.Footnote 10 This parcel is numbered H. 147 in the Hoernle collection, which includes 181 fragments according to Hoernle's record. Marshall noted that these fragments were “some ancient manuscripts from Kashgar”. Hoernle labelled them “Non-Sanskritic MSS large fragments”. On 24 April 1906, Hoernle sent Sten Konow 136 fragments, including this one. But two or three weeks later, Konow returned them. On 28 October 1907, Hoernle sent them to Ernst Leumann. In December 1918, the Librarian of the India Office F. W. Thomas established the Hoernle collection in Oxford.Footnote 11
IOL Khot. 147/5 (H. 147 NS 106) was first published by Ernst Leumann in 1920 (see his book, p. 111, as Skjœrvø has noted), and later in 1963 Harold W. Bailey published it again in his Indo-Scythian Studies: Khotanese Texts V (KT V).Footnote 12 Almost forty years later, in 2002, for the first time Skjœrvø made the text available in English in his catalogue. Based on his translation, I have discovered that this fragment has a Chinese parallel in Da fangguang fo Huayan busiyi fo jingjie fen 大 方 廣 佛 華 嚴 經 不 思 議 佛 境 界 分 (tentatively its Sanskrit title can be reconstructed as Mahāvaipulya-buddha-Avataṃsaka-sūtra-acintya-visaya-pradesa). This text was translated into Chinese by the Khotanese monk Devendraprajña (Tiyun Bore 提 雲 般 若) in the Wei Kingdom Eastern Temple 魏 國 東 寺 in Luoyang in the first year of the Yongchang period (689). The Khotanese fragment IOL Khot 147/5 may have been written in Middle Khotanese, which roughly was used from the seventh to the eighth century. If so, then its date was close to the Chinese version. In this time framework, one might suspect that they were based on the same Sanskrit version that was circulated in Khotan and Central China.
Devendraprajña's Buddhist learning was featured with dhāraṇī and dhyana. He perhaps arrived in China in 688 and visited the Empress Wu Zetian in Luoyang. Wu Zetian appointed him as a translator in the Wei Kingdom Eastern Temple in 689. From 689 to 691, he translated six texts, including one treatise (T. no. 1626, Mahāyānadharmadhātunirviśeṣaśāstra) and five sūtras (T. no. 300, no. 306, no. 694. Tathāgatapratibimbapratiṣṭhānuśaṃsā, no. 1346. Sarvabuddhāṅgavatīdhāraṇī, no. 1397. Jñānolkādhāraṇī). Footnote 13 All these texts were incorporated into the Buddhist canon when Mingquan 明 佺 compiled the catalogue for the Buddhist canon in the Wu Zetian Zhou period, and they appeared in the Taishō canon. Only T. no. 300 was translated in 689; the other five texts were translated in 691. T. no. 306, Da fangguangfo huayanjing xiucifen 大 方 廣 佛 花 嚴 經 修 慈 分, the Maitrī-bhāvanā-prakaraṇa has been recognised as the parallel version of the Khotanese manuscript of the Book of Zambasta, Chapter three.Footnote 14 T. no. 694, Tathāgatapratibimbapratiṣṭhānuśaṃsā (Ch. Dacheng zaoxiang gongde jing 大 乘 造 像 功 德 經) actually also has a Khotanese parallel in Chapter 23 of the Book of Zambasta, which was discovered by Inokuchi.Footnote 15 T. no 1397, Jñānolkādhāraṇī (Ch. Zhijü tuoluoni jing 智 炬 陀 羅 尼 經) also has Khotanese fragments, including one manuscript SI P 3 in the St Petersburg collection, three fragments in the British Library, IOL Khot 156/2 (H. 142 NS 82), IOL Khot 158/4 (H. 143 MBL 22) and IOL Knot 204/6 (H. 149).Footnote 16 In addition to T. no. 300, the text I just identified here, we know that at least four texts translated by Devendraprajña have parallel Khotanese versions. We currently do not have a general catalogue for all Chinese Buddhist manuscripts from Dunhuang. I have checked the revised edition of the Index to the General Catalogue of Dunhuang Manuscripts compiled by the Dunhuang Research Academy and published in 2000,Footnote 17 but so far none of these four texts have been found in this catalogue, which is worth noting while discussing the Buddhist connection between Dunhuang and Khotan. It seems that Devendraprajña's texts were not circulated in Dunhuang, but the reason for this remains puzzling.
Devendraprajña's translation Da fangguangfo huayan busiyi fo jingjie fen is not the unique Chinese translation of this text. Later in 700 (the first year of the Jiushi period), another Khotanese monk, Śikṣānanda 實 叉 難 陀 (652–710) retranslated this text in the Shouji Temple in Luoyang. However, if we compare the two Chinese translations with the Khotanese fragment of this sūtra, Devendraprajña's translation is closer to the Khotanese version. In the following I will translate the corresponding section in Devendraprajña's translations, and show how it is close to the Khotanese version.
Here I first cite Skjœrvø's transliteration and translation and then offer my English translation of the Chinese parallel. Skjœrvø gives the Khotanese fragment as follows:
Fol. 17.
r 1 väna jägye vātco jiyāre. Väno ṣkogyānu dyāre 60 1 väna pätāyemi [jsa] dātu gaṃbhīru rraṣṭu hvatā
2 ndä. Väna tsūmate jsa tsutāndä vino viñānä bvāre 2 vina ātme jīvitä satvi vina poṣi pudgali
3 sajñe vina padīmākä paysānāki buvāki tsūki hvarāki. Vino skandhānu ne ską-
4 ndha haṣṭūsu dāta dvāsu āyatana biśi ttä jsa ne īndä ne vā vina ttānä ttäna cu ta{tva-}-
5 tvatu ttyo biśyo hamaṃgu biśi sarvadharma u balysa śo näjsaḍu samu kho marīca
v 1 biśi sajñamātre lovä biśi sajñmātrā balysa biśi sajñmātrā kṣetra samu kho ṣei sajñe jsa date 60 6
2 viñāni pracai rūva hamatä karä rūta ne īndi viñāni balysa niruddhä ttäna vino rūti diyāñä 7 cu ṣä samu saṃ-
3 ñe jsa date cu rro hamatä harbiśi balysa. Dro-mase niśtä samu kho ātāśä hamaṃgu 60 8
4 saña ttū balysu padaṃdātä hamatu ttatvatu niśti. Trāmu ysamaśandā draya samu hīviñe saṃñe jsa ’īndi
5 kvī vikalpīmi balysä date. Ku ne kar-ju ne date trāmu ttätä harbiśi dharma samu hīvī a[ysmuī] īñä 70 cvī citra-mātra
Skjœrvø offers his English translation as follows:
(r1) . . . without destruction they are destroyed. Without the skandhas they appear. 61.
Without speaking they have spoken the correct, profound law. (r2) Without going they have gone. Without vijñāna they know. (6)2.
Without soul, life, being, without poṣa, person, (r3) saṃjñā, without maker, realiser, knower, goer, eater. (63)
Without the skandhas there are no skandhas, (r4) 18 dhātus, 12 āyatanas. All those are not with them, nor without them. (64)
Because (r5) in reality all dharmas are equal to all those. And the Buddha is of one kind. (65)
Just like a mirage (v1) all the world is just saṃjñā, all the Buddhas are just saṃjñā, all the fields are just saṃjñā. Through saṃjñā it appears just like this. 66
(v2) Because of vijñāna there are forms. Forms themselves are not at all. Vijñāna, O Buddha, is obstructed. Therefore it should be viewed as without form. (6)7.
Whatever (v3) appears only by saṃjñā and what (are) all the buddhas themselves – there is no difference even the size of a hair. It is just equal to space. 68.
(v4) saṃjñā has made this Buddha. By itself nothing is in reality. Thus the three earths: they are only by one's own saṃjñā. (69)
When I construct him (vikalpaya) the Buddha appears, when I don't he does not appear at all. All these dharmas (are) only on account of one's own *mind. 70.
What . . . its nothing but *citta. . .
The passage from Devendraprajña's Chinese translation (T. vol. 10, no. 300: 908b) is the following:
[All Tathāgatas], are of not birth as birth, not extinction as extinction, which are different from all phenomenal changes. Their Dharma nature is profound, which is of not speaking as being able to speak, not disappearing as disappearance, not coming as coming, not vijñāna as vijñāna, not my saṃjñā, not the saṃjñā of living beings, not the saṃjñā of sentient beings, not the saṃjñā of fostering beings, not the saṃjñā of pudgala (Ch. buteqie 補 特 伽), not the saṃjñā of eaters, not skandhas, not non skandhas, and having aggregate-alike marks. 18 dhātus (Ch. shiba jie 十 八 界) and 12 āyatanas (Ch. shi'er chu 十 二 處) do not exit after all and they are not different from each other. All Buddhas and all dharmas are impartial and equal. They are same to one principle, just like the Sun flames (Ch. yangyan 陽 焰). All sentient beings, All Tathāgatas, and all Buddha-lands (Ch. fotu 佛 土, alternatives: Buddha-realm, Buddha-sphere; Skt. buddhakṣetra), are not separated from the saṃjñā. Like this saṃjñā appears, and the vijñāna originates from rupa. The svabhāva (self-nature) is like this, which has no birth at all. Tathāgata blows up his vijñāna, thus this is different from rupa. You should know from your observance that all dharmas are not separated from the saṃjñā. Then you will recognise Tathāgata. Even for the size of a hair's tip (Ch. maoduan 毛 端) there is no difference. Like śūnya gagana are impartial and equal. If I differentiate them, then the Buddha appears before me. If I do not, nothing can be seen. Saṃjñā can become the Buddha, there is nothing beyond saṃjñā. Like this three dhātus, and all dharmas, are not separated from mind.
The passage from Śikṣānanda's Chinese translation (T. vol. 10, no. 301: 911b) is slightly different from that in Devendraprajña's version. In terms of the differences among all three versions in Chinese and Khotanese, the most striking point is that both Devendraprajña's translation and the Khotanese version (r4) introduced the ideas of eighteen dhātus and twelve āyatanas, which were absent in Śikṣānanda's Chinese translation. The second difference lies in the lines on all dharmas in three dhātus in Devendraprajña's version, yet in Śikṣānanda's version it appears as all Buddhas in three ages (the past, present and future). The Khotanese fragment does not mention Buddha at all in this line (v4). It would appear that Devendraprajña's translation and the Khotanese version were based on the same original Sanskrit version. In this passage of Devendraprajña's Chinese version the phrase “pingdeng pingdeng 平 平 等 等 ” (impartial and equal) appears twice, which corresponds the Khotanese word hamaṃgu. In chapter three of the Book of Zambasta, it also appears and corresponds to the same Chinese translation in Devendraprajña's version T. no. 306.Footnote 18
The Khotanese Fragment Khot “missing” frags. 3
The last fragment I have identified is Khot “missing” frags. 3. This fragment was originally from the Godfrey 3 collection, as Skjœrvø notes; most of the manuscripts in Godfrey collection were from Dandan Uiliq. In fact, this fragment was once thought to be missing from the British Library collection but was discovered to be in the Bibliothèque Nationale. As Ursula Sims-Williams recently told me, it was originally purchased by Captain Godfrey on behalf of the Government of India. On 19 January 1898, it was given by Hoernle to the French scientist H. Dauvergne, “in the name of Captain S. Godfrey with permission of the government of India”. Later this fragment was rediscovered in the collection of the Musée Guimet and later it was transferred to the Bibliothèque Nationale, becoming part of the Pelliot collection.Footnote 19
I discovered that this Khotanese fragment has a Chinese parallel in Hastikakṣya-nāma-mahāyāna-sūtra. This text has two Chinese versions, translated by Dharmamitra and Dharmarakṣa, with two different titles: The Sūtra spoken by the Buddha on the Armpit of the Elephant (Foshuo xiangye jing 佛 說 象 腋 經, T. no. 814, translated in 441) and The Sūtra spoken by the Buddha on the Hopeless (Foshuo wu xiwang jing 佛 說 無 希 望 經, T. no. 813, translated between the second year of Taishi and the first year of Jianxing, the Western Jin dynasty, 266–313) respectively. Some catalogues in the Chinese canon also indicate that T. no. 814 was also called The Sūtra spoken by the Buddha on the Steps of the Elephant (Foshuo xiangbu jing 佛 說 象 步 經).
Again, I cite Skjœrvø's transliteration and translation first, and then offer my own English translation of Dharmamitra's Chinese version. Skjœrvø gives this Khotanese fragment as follows:
a 1 [. . .. . .. . .. . .. . .. . .] ra x x x nāma bodhisatväna 2 caturanda-saṃkusumittä nāma bodhisatväna 3
2 [. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .] x x x x pradīpä nāma bodhisatväna 4 anantāvalokitta-sarvaviṣayasa
3 [. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .nā]ma bodhisatväna 5 merukūṭātyudgattä nāma bodhisatväna 6 abhi
4 [. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . ..] saṃjananä nāma bodhisatväna 7 vimala-bāhu-apratihata-raśmi
5 [. . .. . .. . .. . .. . .. . .. . .. . .bo]dhisatväna 8 sarvasatvāśaya-viniścitārtha-pāragata-ggāmä nāma bodhisa-
b 1 [tväna 9 . . .. . .. . .. . .. . .. . .. . . ..]dṛḍha-prrāptä nāma bodhisatvina 10 sarvaruta-ravitayomastara-vy-
2 [. . .. . ..] nāma bodhisatväna 11 brahmasvara-tteja-atyudgattä nāma bodhisatväna 12 kīrtisva
3 [. . .. . .. . .. . .. . .. . .. . .. . .] nana-apratihata-buddhi nāma bodhisatvina 13 sarvakula-ratna-saṃca-
4 [ya . . .. . .. . .nāma] bodhisatväna 14 u maṃjuśrīna alysānaina 15 ttyau paḍauysyau kṣaṣṭä yse-
5 [ryau . . .. . .. . .bodhi]sa[tvy]au [tt]ī[tä] mai āśirī śśāriputrä palśārä haḍai samāhānäna panatä
(a1) . . . with the bodhisattva called . . . 2; with the bodhisattva called Caturanda-saṃkusumita 3;
(a2) . . . with the bodhisattva called –pradīpa 4; with the bodhisattva called Anantāvalokita-sarvaviṣaya-sa-. . .5; (a3) with the bodhisattva called Merukūṭātyudgata 6; (a4) with the bodhisattva called Abhi- . . . saṃjana 7; with the bodhisattva called Vimala-bāhu-apratihata-raśmi- . . . 8; (a5) with the bodhisattva called Sarvasatvāśaya-viniścitārtha-pāragata-ggāma. [9]; with the bodhisattva called . . . (b1) -dṛḍha-prrāpta 10; with the bodhisattva called Sarvaruta-ravitayomastara-vy-. . .11; (b2) with the bodhisattva called Brahmasvara-tteja-atyudgatta 12; with the bodhisattva called Kīrtisva[ra]- (b3) . . . -nana-apratihata-buddhi 13; with the bodhisattva [called] Sarvakula-ratna-saṃca[ya]- . . . 14; with the bodhisattva called Mañjuśrī. 15; with those first sixty thousand . . . bodhisattvas.
(b5) then the reverend Śāriputra rose from the meditation in the evening of the day.
The parallel section in Dharmamitra's version (T. no. 814: 781c6–15) is actually the opening scene in this sūtra. We find that in Rājagṛha the Buddha was about to preach Hastikakṣyā to a great assembly of monks and bodhisattvas. I translate this section as follows:
Thus have I heard, at that time the Buddha was in Rājagṛha, Gṛdhrakūṭa, together with the great assembly of five hundred monks, and 60,000 bodhisattvas. Each of these bodhisattvas is well known to the assembly. They have attained (Ch. de 得, like de in dedao 得 道, attained enlightenment) dhāraṇīs, joyfully speaking without obstacles, preaching the Dharma [which is] non-dualist (Ch. shuofa wuer 說 法 無 二), and they have achieved inconceivable supernatural power. Their names are: Vīrya-anupama/nirupama-mati Bodhisattva, Kruṣṭa-tejaḥ-sāra Bodhisattva, Catur-ratna-puṣpa-vistara Bodhisattva, Mahā-vidyut-dīpa Bodhisattva, *Ananta-avalokita-sarva-deśa-atikrama Bodhisattva, *Acalendra-kūṭa-ābha-atiadbhuta Bodhisattva, *Abhinanda-saṃtoṣa-ja Bodhisattva, *Bāhu-vimala-prabhā-samudghāta Bodhisattva, *Sattva-cintā-sūrata-artha-pāramita Bodhisattva, *Vajropama-dhīra/dhruva-prāpta *udāhāra Bodhisattva, *Sarva-ghoṣa-nāda-atikrama Bodhisattva, *Brahma-ghoṣa/svara-kīrti/tejas-atiadbhuta *kīrti-ghoṣa-bhaya-gata Bodhisattva, *Sarva-kuśala-mūla-ratna-saṃcaya Bodhisattva.Footnote 20 Mañjuśrī the youth with first 60,000 bodhisattvas together. At that time the great virtue (Ch. dade 大 德, perhaps from Skt. bhadanta) Śāriputra rose in the evening of the day from the meditation and came to visit the Buddha's place.
If we compare the Khotanese and Chinese versions, they are very close. If we take the Tibetan version into account, it is likely that the Tibetan translation is based on a source text that is different from the one for both the Khotanese and Chinese translations. Professor Jan Nattier kindly helped to read the Tibetan version for me. Footnote 21 According to her note, the Tibetan title is The Mahāyāna Sūtra Entitled “The Skill of the Elephant”, which is different from the title of the Chinese version in T. no. 814. She also translated the Tibetan version as follows and provided the vocabulary notes:
Thus have I heard at one time: He Who has Passed Away Victorious (sic, the standard equivalent of bhagavat “blessed one”; Tib. bcom ldan ‘das) was staying on Vulture Peak (Gṛdhrakūṭa) Mountain, together with a great saṁgha of bhikṣus (dge slong “virtuous beggar”, the standard equivalent of bhikṣu), and with very famous (abhijñāta, Tib. mngon par grags pa) bodhisattvas. They [i.e., the bodhisattvas] were ones who had attained the dhāraṇīs, [had] unobstructed eloquence (pratibhāna, Tib. spobs), explained the non-dual (advaya, Tib. gnyis su med pa) Dharma, and possessed the inconceivable power to work wonders (ṛddhi-vikurvaṇa, Tib. rdzu ’phrul rnam par ’phrul ba) – sixty thousand of them. [Their names] were as follows. . .. . .Then the venerable (āyuṣmat) Śāradvatīputra arose in the evening from meditative seclusion (pratisaṁlayana) and went to where the Buddha was.
In terms of vocabulary, Dharmarakṣa's earlier translation (T. no. 813) in the fourth century departs from Dharmamitra's version and is also different from the Khotanese version. Dharmamitra translated the last sentence in this section on Śāriputra in a different way. He uses “the worthy (xianzhe 賢 者)” to address Śāriputra. According to Nattier's note, “the worthy” is very common among Dharmarakṣa's translation corpus for the Sanskrit word āyuṣmat, which corresponds to tshe dang ldan pa in Tibetan. In his version, Śāriputra was said to practise meditation with a single mind in a solitary place (在 於 獨 處 一 心 禪 思), rose from a silent sitting (yenzuo 燕 坐, a highly concentrated meditation, Skt. pratisaṁlayana, Tib. nang du yang dag ‘jog pa), and went to visit the Buddha's place. This expression does not specify when Śāriputra departed to see the Buddha. Both Dharmamitra's version and the Khotanese one indicate the evening was when Śāriputra went to see the Buddha.
Dharmamitra plays a crucial role in the Buddhist history of Dunhuang. When he first arrived in Dunhuang, he constructed a meditation pavilion in the empty wilderness, planted a great number of trees and opened a huge garden, which attracted numerous followers who came to learn meditation techniques with him. In early theYuanjia period (early fifth century) he moved to Jianye (now Nanjing), the capital city of the Liu Song dynasty, where he focused his translation interests on mediation texts and dhāraṇīs (Sengyou 僧 祐, Chu sanzang jiji 出 三 藏 記 集, juan 14, T. vol. 55, no. 2145: 105a–b; Huijiao 慧 皎, Gaoseng zhuan 高 僧 傳, juan 3, T. vol. 50, no. 2059: 342c–343b).
However, if we look at the vast number of Buddhist manuscripts recovered from the cave library in Dunhuang, it is striking that there is one manuscript from Dharmarakṣa's in the Stein collection (S. 4645), but nothing from Dharmamitra. Although Dharmamitra was an important figure in the history of Buddhism in Dunhuang, his translation of Hastikakṣyā did not gain currency in Dunhuang, perhaps because it was made in South China. What we can conclude is that both Dharmamitra's translation and the Khotanese translation might have been based on the same Sanskrit version. Fortunately, in the British Library there is a Sanskrit fragment H.150.vii.18 that has already been identified as Hastikakṣyā by TesshinWatanabe, has the official number Or.15009/672 and whose photograph is now available in the database of the International Dunhuang Project at the British Library.Footnote 22 Watanabe did not detail his identification. Liu Zhen and I have found its corresponding part in the Chinese canon as T. vol. 17, no. 814, p. 782b20–24, and b25–28.Footnote 23
On Khotanese Raśmivimalaviśuddhaprabhādhāraṇī and its Chinese Parallel
The Khotanese version of Raśmivimalaviśuddhaprabhādhāraṇī was discovered by Yoshida Yutaka. He first recognised a fragment in the Petrovsky collection in St Petersburg and later he identified four fragments (Or. 6402B/2.1+3, Or. 6402B/2.4, IOL Khot 172/4, IOL Khot 172/5) preserved in the British Library. Footnote 24 Among these fragments is Or. 6402B/2.1, which can be dated to around 786, sometime later than the two Chinese translations. What interests me about this version is that the text Raśmivimalaviśuddhaprabhādhāraṇī has two Chinese versions with different titles: one translated in 704 by a Tocharian monk Mitraśānta (Mituoshan 彌 陀 山, 690–704 in Chang'an) but also translated several years earlier by a Khotanese monk, Śikṣānanda.Footnote 25 Although Śikṣānanda's version of Raśmivimalaviśuddhaprabhādhāraṇī is not extant, the title has survived in many catalogues compiled since the Tang dynasty. This translation seems to have been lost or forgotten soon after it was made. As the Kaiyuan catalogue indicated (Zhisheng, Kaiyuan shijiao lu 開 元 釋 教 錄 The Catalogue of Buddhist Canon in the Kaiyuan Period, T. vol. 55, no. 2154: 566a.), Śikṣānanda translated nineteen texts: fourteen of them have survived and the other five have been lost. This catalogue was compiled in 730, but before this year, Śikṣānanda's version of Raśmivimalaviśuddhaprabhādhāraṇī was not in circulation at all. With regard to the Dunhuang manuscripts, we have many of them (S. 1634 and S. 4156 in the Stein collection, P. 3916i in the Pelliot collection, and Bei 7379–7385 in the Beijing Library collection, as well as some fragments in other collections) with Mitraśānta's translation but no fragments of Śikṣānanda's version, which might also indicate that Śikṣānanda's was not circulated at all. Śikṣānanda's version of Raśmivimalaviśuddhaprabhādhāraṇī was translated together with the Chinese translation T. no. 301 in the Shouji Temple, in 700 (Zhisheng, Xu gujin yijing tuji, T. vol. 55, no. 2152: 369c). The Kaiyuan catalogue notes that the text for T. no. 301 was also lost, but it actually survived as now we can see it in T. no. 301. In short, Mitraśānta's translation is found to be the only Chinese parallel to the Khotanese version. From the perspective of Chinese Buddhist literature, even the extant title of Śikṣānanda's version is helpful to our discussion here.
In terms of the title, Mitraśānta and Śikṣānanda's translations only disagree with each other over two characters. The former uses Wugou jingguang da tuoluo ni jing 無 垢 淨 光 大 陀 羅 尼 經 (T. no. 1024), referring to The Great Dhāraṇī Sūtra of No Impurity and Pure Light, while the latter uses Ligou jingguang tuoluoni jing 離 垢 淨 光 陀 羅 尼 經, referring to “The Dhāraṇī Sūtra of detaching from impurity and pure light”. If we trace the history of these different words in currently available Chinese Buddhist literature, new light is shed on how they travelled from one translator to another. They are not common and popular words such as “Buddha” or “Bodhisattva” which are too frequent in the Chinese Buddhist canon to trace out a line. The Chinese word ligou 離 垢 (detaching impurity) was a popular word; apparently it appeared as early as in An Shigao's translation. However, if we look at ligou jingguang 離 垢 淨 光 (detaching impurity and pure light), this first appeared in current Chinese Buddhist literature as part of an epithet of the Buddha in Tathāgatamahākaruṇānirdeśa (T. no. 398, Ch. Da'ai jing 大 哀 經), translated by Dharmarakṣa in 310 and later in Avataṃsakasūtra (T. no. 278, Ch. Huayan jing 華 嚴 經), translated by Buddhabhadra 佛 馱 跋 陀 羅 (359–429) in 408. Śikṣānanda's version of Avataṃsakasūtra (T. no. 279) must have borrowed this phrase from Buddhabhadra's version. This phrase appears in both Buddhabhadra and Śikṣānanda's versions, though in different places. While Śikṣānanda and his team members were translating Avataṃsakasūtra, they must have consulted the previous Chinese versions, including Buddhabhadra's version. There is no reason why they did not look into Buddhabhadra's version. Buddhabadra's version of Avataṃsakasūtra has also used the phrase wugou jingguang (no impurity and pure light), which we see in the title of Mitraśānta's translation. But Śikṣānanda's version does not adopt this phrase. So even though he was aware of this phrase from Buddhabhardra's version, he did not adopt it. The reason must be because he preferred the expression“detaching impurity”; “detaching impurity” is also attested by his adoption of this phrase in the title of his translation of Raśmivimalaviśuddhaprabhādhāraṇī. His choice was not the preference of his era; his translation of Raśmivimalaviśuddhaprabhādhāraṇī was lost very soon after he translated it.
Although a lot of Chinese Buddhist texts have been lost through the ages, we know at least that Śikṣānanda's translation of Raśmivimalaviśuddhaprabhādhāraṇī used ligou jingguang rather than wugou jingguang in its title. It is worth noting that the Great Sūtra on Maitreya's Becoming the Buddha 彌 勒 大 成 佛 經 (T. no. 456) translated by Kumārajīva 鳩 摩 羅 什 (350–409) in 401 was seemingly the earliest text using this phrase that we can find in the corpus of Chinese Buddhist texts. Buddhabadra's version of Avataṃsakasūtra might have taken this phrase from Kumārajīva's translation T. no. 456. Later, other Chinese translations, such as Vimaladattāparipṛcchā (T. no. 339, Ch. De wugou nü jing 得 無 垢 女 經)Footnote 26 and Vikurvaṇarājaparipṛcchā (T. no. 421, Ch. Fenxunwang wen jing 奮 迅 王 問 經), both translated by Gautama Prajñāruci 瞿 曇 般 若 流 支 in the sixth century, also adopted this phrase. For the text T. no. 339, Dharmarakṣa has a different translation: Ligou shi nü jing 離 垢 施 女 經, T. no. 338, which is another example showing the difference between the titles using ligou 離 垢 (detaching impurity) and wugou 無 垢 (no impurity).
Interestingly, Mitraśānta and Śikṣānanda, along with Fazang, were working together side by side. According to Fazang's commentary, right after Śikṣānanda finished his translation of Avataṃsakasūtra in 698 in Luoyang, he was ordered to translate Laṅkāvatārasūtra (T. no. 672, Ch. Dacheng ru lengqie jing 大 乘 入 楞 伽 經). Soon he was invited to reside at the Qingchan Temple in Chang'an. His incomplete translation was left behind. In 702, Mitraśānta was ordered with Fazang and others to resume the translation of this text, because Mitraśānta was said to be an expert in Laṅkāvatārasūtra and Abhidharmakośaśāstra and had spent twenty-five years in India, according to Fazang's Commentary notes on the Laṅkāvatārasūtra (Ru Lengqie xuanyi 入 楞 伽 玄 義, T. vol. 39, no: 1790: 430b). Fazang also noted that, after Mitraśānta finished the work on Laṅkāvatārasūtra in 705, he continued his collaboration with Fazang on the translation of Raśmivimalaviśuddhaprabhādhāraṇī (Dacheng qixin lun shu kewen 大 乘 起 信 論 疏 科 文, Zokuzokyō 續 藏 經, vol. 45, no. 767). Nevertheless, in contrast to the leading roles played by Śikṣānanda and Fazang in the Buddhist community during Wu Zetian's era, Mitraśānta was a marginal figure.
Conclusions
To sum up the discussions above: first of all, we now have three more Khotanese Buddhist texts, including Agrapradīpadhāraṇī, Mahāvaipulya-buddha-Avataṃsaka-sūtra-acintya-visaya-pradesa, and Hastikakṣya-nāma-mahāyāna-sūtra, and we know they have their Chinese parallels. Second, we know that the Khotanese monk Devendraprajña was particularly important in terms of the connection between Khotanese Buddhism and Chinese Buddhism. At least four of his Chinese translations have Khotanese parallels. Jñānagupta is another interesting translator, for at least one of his Chinese translations T. no. 1354 has parallels with the Khotanese fragments (Agrapradīpadhāraṇī, IOL Khot 25/4) and two of his translations (T. no. 431; T. no. 416, translated in 595) have parallels with the Sanskrit fragments recovered from the Khotan area (Aṣṭabudhakasūtra, H. 143. SB. 98; Bhadrapālasūtra, H. 143, SA.3).Footnote 27 Third, dhāraṇī texts were very popular in Khotan: the dhāraṇī texts in Khotanese include Aparimistāyuḥsūtra (or Aparimistāyuḥdhāraṇīsūtra), Agrapradīpadhāraṇī, Anantamukhanirhāradhāraṇīsūtra, Raśmivimalaviśuddhaprabhādhāraṇī, Jñānolkadhāraṇī, Avalokiteśvaradhāraṇī, and Amṛtaprabhadhāraṇī, and some other Sanskrit dhāraṇī texts such as Tathāgatoṣṇīṣa, Sitātapatra, Sumukha-sūtra, and so forth.Footnote 28 Although the newly identified Hastikakṣya-nāma-mahāyāna-sūtra does not have dhāraṇī in its title, it does emphasize dhāraṇī practice.
