「大悲咒」漢梵對音研究

Research on the Chinese to Sanskrit Translation of the Great Compassion Mantra

本文以唐代伽梵達摩譯的「千手千眼觀世音菩薩廣大圓滿無礙大悲心陀羅尼經大悲神咒」（大悲咒）為材料，透過漢梵對音的研究，瞭解唐代的語音狀況。本文發現在聲母方面多數是整齊對譯，清濁混用是在條件下產生的現象，全濁音可以對譯送氣音與不送氣音。韻母方面，三等介音、四等介音、唇音的合口介音通常沒有譯出。主要元音部分，有多韻對譯同一梵語元音的狀況。韻尾部分，開尾韻、陽聲韻幾乎整齊對譯，入聲的t尾比k尾容易弱化，弱化的時間也比較早。整體來說，聲母的對譯比韻母還要嚴整，許多語音變異符合「同化」、「異化」、「脫落」、「插入」、「弱化」的演變規則。

This study uses the Tang dynasty Bhagavad-dharma translation of the Great Compassion Mantra (Mahā-Kārunika-citta-dhārani) to examine the transliteration of Buddhist scripture with the objective of understanding the pronunciation of Tang dynasty Chinese.We find that for most initial consonants, the translation is very regular, with the mixing of voiced and unvoiced initials produced only under specific conditions. A voiced stop can he translated as either aspirated or unaspirated. For finals, the majority of third and fourth division head vowels and labial closed head vowels are not translated. Different Chinese characters are used to translate identical Sanskrit main vowels, while most tail vowels are translated quite regularly, based on the end rhyme. The entering tone tail vowel [t] is more easily weakened compared to [k], and this weakening occurs earlier. As a whole, the translations of the initials are more regular than those of the finals, and much of the language variation conforms to the evolutionary rules of assimilation, dissimilation, deletion, insertion, and weakening.