《法句經》（T210）的初譯偈頌與後譯偈頌

Early and Late Translations in the Faju Jing (T210)

近代學者對T210《法句經》源頭文本（sourcetext）所隸屬的「部派」有諸多探討，例如印順法師、黃懺華、魏查理（Charles Willemen）等數位學者認為「初譯」即「核心26品」，其源頭文本為巴利《法句經》；呂澂則主張其出自「化地末宗」，或者如法光法師（Bhikkhu Dhammajoti）持保留態度，認為「『核心26品』中至少有些偈頌不是出自巴利《法句經》，而且有些偈頌的譯詞顯示其印度語系文本（Indic text）的語言可能既不是巴利（Pāli），也不是梵文（Sanskrit）」。屈大成則主張：「初譯出自赤銅鍱部或法藏部或化地部。……全本為赤銅鍱部併合有部。」 但是，支謙〈《法句經》序〉描述《法句經》（T210）經由「初譯」、「後譯」及「編訂」三個步驟而成。如果未能辨識出「初譯偈頌」與「後譯偈頌」，並就兩組的特性深入探討，就無法解答諸如「隸屬的部派」、「源頭文本的語言」、「重譯偈頌」、「不同語言版本的系譜關係」等等議題，因此應先探索此一辨識方法。 本文提出「梵巴偈頌的分章差異」、「梵巴偈頌的用字差異與對單字的詮釋差異」、「誤譯的偈頌」、「重譯偈頌」與「核心26品的後綴偈頌」等五種方法來辨識前譯與後譯偈頌，希望能夠藉此作為討論平台，以激發其他辨識方法的增補或評判。

In recent years, scholars have discussed the question of the sectarian affiliation of the Dharmapada source text used for the translation of the T210 Faju Jing. Scholars such as Yin Shun, Huang Chanhua, and Charles Willemen argue that the earliest section, namely the core 26 chapters, are based on the Pali Dhammapada. Lu Cheng on the other hand believed that it belonged to the Mahīśāsaka school. Bhikkhu Dhammajoti showed how a number of the verses do not have any parallels in the Pali version, and that the translation reveals that the language of the original text was probably neither Pali nor Sanskrit. Qu Dacheng argued that“the early translation belongs to the Tāmraparṇīya, Dharmaguptaka, or Mahīśāsaka school. […] The whole text belongs to the Tāmraparṇīya school joined with the Sarvāstivāda.” Yet, in his preface Zhi Qian tells us that the translation went through three phases: early translation, subsequent translation, and compilation. For this reason, the questions of the sectarian affiliation and language of the source text, together with duplicate verses and genealogical issues among different versions of the Dharmapadas, cannot be adequately addressed without clearly distinguishing the contents belonging to the early translation and those belonging to the subsequent translation phase. It is therefore necessary to first devise a method or methods for identifying the contents of these two parts of this text. In this article I shall propose five principles for this work: comparing chapter divisions to the Pali and Sanskrit versions; comparing translation word use and word interpretation to the Pali and Sanskrit; considering mistranslated verses; considering duplicate verses; looking at the verses appended to each of the 26 core chapters. My wish with this framework is to foster interest in this work of developing methodologies for this type of textual analysis.