________________ 302 REVIEWS The present edition is based upon six Sanskrit manuscripts, two versions of the Tibetan translation (Peking and Narthang) and two Chinese translations. Moreover, the editor has added a long introduction which occupies the greater part of volume one (pp. 7-250). Undoubtedly, this edition marks a great progress upon the editio princeps, but I am afraid that it is not free from imperfections. As long as no other manuscript materials become available, it will probably be impossible to establish an entirely satisfactory Sanskrit text. Nevertheless, with the help of the materials at present available, a better text could have been established. The main reason for the defects of this edition must be sought in the fact that the editor has not sufficiently taken into account the grammatical particularities of the text. He devotes a section of his introduction to the "Method of presentation of the text" (pp. 33-58) and discusses in it peculiarities of the spelling of words in manuscripts and differences between the Sanskrit text and the Tibetan and Chinese versions, especially as regards proper names, but little is said about the grammatical features of the text. The editar merely observes that on account of the flexibility of Buddhist Sanskrit grammar, he has refrained from correcting any grammatical peculiarities in the Sanskrit manuscripts. As examples he quotes abhisamskarsit instead of abhisamaskarsit (!), samanupasyamah instead of samanupasyavah, pradaksinikrtva, sajjikstva and irregularities of samdhi before initial r. This statement and the editorial practice of the editor make it clear that, in his opinion, a text, written in Buddhist Hybrid Sanskrit, may show almost any grammatical or lexicographical irregularity. Especially since the publication of Edgerton's Grammar and Dictionary of Buddhist Hybrid Sanskrit, there is often a tendency to select the most aberrant readings of the manuscripts and to justify them with references to Edgerton's work.a Dr. Yamada does not refer to Edgerton, but his edition shows clearly his conviction that any anomaly, described in Edgerton's Grammar, is admissible in the Karunapundarika. Insufficient attention is being paid to Edgerton's classification of BHS (=Buddhist Hybrid Sanskrit Grammar) into three classes. Each editor of a BHS text has to determine to which of the three classes his text belongs. In an important article on "The language of the Buddhist Sanskrit texts" (BSOAS, xvi, 1954, pp. 351-371) John Brough has declared that the immediate task for the future is the closer delineation of the various forms and styles of Buddhist writings in Sanskrit, and a detailed grammatical analysis of each type. In his article he presents specimens of nine distinct styles. For the time being, it will perhaps be better to adhere to the three classes set up by Edgerton. The Karunapundarika clearly falls within the second class in which the verses are hybridized, but the prose has relatively few signs of Middle Indian phonology and morphology. The prose of the KP (= Karunapundarika) is very rich in BHS words. Although most of them have been recorded in Edgerton's dictionary, it would have been the task of the editor to give an index of BHS words. Below. I list a number of BHS words with references to the text without trying to be complete with regard to the words selected and to the references given. A second list contains words, which are absent from Edgerton's Dictionary and a few others worthy of note. The Sanskrit text of the KP has been badly transmitted. This is especially true of the first part. All six manuscripts used by the editor are recent Nepalese manuscripts, written by scribes who had only a vague knowledge of Sanskrit. They are certainly responsible for such anomalies as confusions of number, gender and case. The prose of the KP is written in a fairly correct Sanskrit as is obvious from the passages which have suffered less in the course of transmission. Undoubtedly, in many places it is impossible to establish a correct text, because the manuscripts are too corrupt. However, in other cases it is quite well possible to correct the readings of the manuscripts with the help of the Tibetan and Chinese translations. The editor shows an exaggerated confidence in the correctness of the manuscripts. Even when the manuscripts omit an 2 Cf. my review of Ratna Handurukande's edition of the Manicudavadana, IIJ, XII, p. 140-143.
