________________
122
THE INDIAN ANTIQUARY.
[APRIL, 1881.
up of vriksh and a). Thus a name for a thing is been which the Chinese translators tried to render formed by mixing the parts together, according to into their language, and we may thus succeed in the rules of the book, which consists of more than occasionally correcting the text as handed down twenty sentences (or feet of sloka). Unddi is to us in Sanskrit MSS. nearly the same as the above, with a few differ- But here a very curious phenomenon presents ences, such as what is full in the one is mentioned itself. There are mistakes in the Sanskrit text in brief in the other, and vice versd."
of our MSS. which it is easy to correct, parMr. Kasawars now informs me that Mancha ticularly when they occur in metrical passages. may be meant for manda, possibly for mandaka, For instance, in the Lalita-vistara (ed. Calc. p. but I do not see that even this would help us | 543, 1. 8) we read: much. Mand means to adorn, manda is used for Chakshur anityam adhruvan tatha brotaghranan cream on milk, also for gruel, but all this, even if jihvdpi, we admitted the meaning of mixing, would not Kaya-mana-duḥkha andma api riktaşvabhavayield us a technical name for the formation of búnych. words by means of joining a suffix with a root. Here the metre shows clearly that we must omit At all events, I have never met with mand, or jihrápi in the first, and kdya-mana in the second, any of its derivatives, in that technical sense. line. They are additions, and very natural addiI thought at one time that manda might be meant tions, to the original text. But when we take for Mandúka, because the Mândukeyas were famous Divakara's translation, the Fang-kwang-ta-chwangby their grammatical works (see History of Ancient yan-king, which was made about A.D. 685, we find Sanskrit Literature, p. 146), and one of these might both jihvdpi," also the tongue," and kdya-mana possibly have been used by I-tsing when study; "body and mind," reproduced, and we find exactly ing the Kridanta chapter. But I do not think the same in the far later Tibetan version. this likely, even if, as I am told, the Chinese trans- In the same chapter (p. 527), after Upaka had literation should admit of it.
asked Bhagavat how he could bear witness of But while we must leave this point unsettled, we himself, and claim for himself the names of Arban are able to identify another title-namely, Juni or and Jina, Buddha answers : Chuni, given as the name of Patafijali's Mahabhd- Jind hi madrid jñeya ye prapta dóravakshayam shya. Mr. Beal informed me that this might be Jild me pdpaka dharmds tenopajino hy aham. read Charni, and Chárni, a general name for com- Here the last pada is clearly wrong in metre and mentary, as in Jitakalpa-chúrni, a Prakrit commen- matter. There is no such word as upajina, and tary on the Jitakalpastra of the Jainas, &c., is the PAli version of the same verse (Mahdvagga, more especially the name of Patasijali's commentary, vol. I, p. 8) shows that the Sansksit text must Patasijali himself being called Chargiksit. have been tenopaka jino hy aham, the sense being :
There is every reason to hope that a more "Those who like me have reached the destruction accurate study of the Buddhist Chinese literature
of all frailties are to be known as Jinas; all evil will be of great help in determining the age of a dispositions have been conquered by me, therefore, number of Sanskrit works the dates of which are o Upaka, I am a Jina, & conqueror." at present floating about between several centuries. Here, again, there is no trace of the vocative And there is another advantage likely to accrue Upaka, a Upaka; in Divakara's translation, and, from that study which has not yet been pointed whatever the Chinese translator may have had out, and to which I should like to call the attention before him, it could hardly have been tenopaka both of Chinese and Sansksit scholars.
jino hy aham. When we have literal translations of Sanskrit This shows how little assistance we can hope texts, these translations help us, not only to fix for from existing Sansksit MSS. towards a rethe date of the Sansksit originals, but also to storation of corrupt passages in the Lalitavistara. determine the ancient readings of the Sansksit There are few Sansksit MSS. as old as the Tibetan texts. Of course there are translations and trans- translation: none as old as Divakara's Chinese lations, and we know now that the translation of & version. Yet, what seem to be palpable blunders Life of Buddha ascribed to K & 6yapa Matanga must have existed when these translations were and Chu-falan (76 A.D.) does not prove, 88 made. What hope, then, is there of our finding → Stanislas Julien thought, that this was a transla- medela for these wounds from existing Sanskrit tion of our Lalita-vistara (sce Selected Essays, vol. MSS., unless they come from totally different ii, p. 191). But when we have to deal with literal localities, and had branched off from the general translations, some of them so literal or mot-a-mot stream before the seventh century of our era as to defy all rules of Chinese syntax, then we are
F. Max MÜLLER able to find out what the Sanskrit text must have Oxford: Feb. 6,1881.
