一切經音義 25.471

25.471

乞灑（二合兩字合為一聲。此一字不同眾例也。已上三十四字名為字母。野字囉字已下九字是歸本之聲從外向內如上所音梵字並依中天音旨翻之。只為古譯不分明更加訛謬疑誤後學。此經是北涼小國玄始四年。歲次乙卯。當東晉義熈十一年。曇無讖法師於姑臧依龜茲國胡本文字翻譯此經遂與中天音旨不同取捨差別言十四音者錯之甚矣。誤除暗惡兩聲。錯取魯留盧婁為數所以言其十四未審如何用此翻字龜茲與中天相去隔遠又不承師訓未解用中天文字所以乖違故有斯錯。哀哉已經三百八十餘年竟無一人能正此失。昔先賢道安法師符秦帝師東晉國德有言曰。譯經有五失三不易也。斯言審諦誠如所論智人遠見明矣。以此觀之失亦過於此說。慧琳幼年亦曾稟受安西學士稱誦書學龜茲國悉談文字。實亦不曾用魯留盧婁翻字。亦不除暗惡二聲。即今見有龜茲字母梵夾仍存。亦只用十二音取暗惡為聲翻一切字。不知何人作此妄說。改易常規謬言十四音。甚無義理其實四字）。

乞灑 (*kṣa* [i.e., *kṣa*]: two combined, the two characters merge into one sound. This single character is unlike the common rules. The thirty-four characters above are called the mātṛkā [consonants]. The nine characters from *ya* and *ra* downwards are sounds that return to the source, proceeding from the outer towards the inner. As for the sounds of the Sanskrit characters given above, they are all translated according to the phonetic standards of Central India. It is merely because ancient translations were not clear and added errors and fallacies, misleading later students. This sūtra was translated in the fourth year of the Xuanshi era of the small Northern Liang state—the cyclical year *yimao*, corresponding to the eleventh year of the Yixi era of the Eastern Jin—by Dharma Master Dharmakṣema in Guzang, based on a barbarian text from the state of Kucha. Consequently, it differed from the phonetic principles of Central India in its selections and rejections. To speak of fourteen vowels is an egregious error! They mistakenly omitted the two sounds *aṃ* and *aḥ*, and mistakenly took *ṛ*, *ṝ*, *ḷ*, and *ḹ* as part of the count, which is why they spoke of fourteen. It is not known how they used these to translate characters. Kucha is separated by a great distance from Central India, and without having received the instructions of a master, they did not understand how to use Central Indian script, which caused discrepancies and gave rise to these errors. Alas, more than 380 years have passed, and yet not a single person has been able to correct this deficiency! Formerly, the ancient worthy Dharma Master Dao'an, Imperial Preceptor to the Fu Qin and a man of virtue of the Eastern Jin state, had a saying: 'In translating the scriptures there are five losses and three difficulties.' These words are exact and true, just as he reasoned; the foresight of the wise is evident! Viewed from this perspective, the losses even exceed that discourse. When I, Huilin, was young, I also received instruction from scholars of Anxi, reciting and studying the Siddham script of Kucha. In truth, they never used *ṛ*, *ṝ*, *ḷ*, and *ḹ* to translate characters, nor did they omit the two sounds *aṃ* and *aḥ*. Even now, the Sanskrit palm-leaf manuscripts of the Kuchean alphabet that survive only use twelve vowels, taking *aṃ* and *aḥ* as sounds to translate all characters. I know not who fabricated this baseless theory, altering the standard rules and erroneously asserting fourteen vowels; it is entirely devoid of principle. In reality, these four characters [viz. *ṛ*, *ṝ*, *ḷ*, *ḹ*] are:)