初擬語言田野調查逐字稿標記原則：以噶瑪蘭語自然對話語料為例

Preliminary Draft: The Principles of Transcribing Language Data Field Recordings with kebalan Vernacular Conversation as an Example

本研究紀要嘗試初步擬訂南島語自然對話語料逐字稿的標記原則。本文所刊載之語料為筆者近期於花蓮縣豐濱鄉立德部落進行田野調查所得成果。有別於先前研究的語料蒐集方法，本文收錄受訪者之自然對話，因自然對話體裁能夠呈現噶瑪蘭族語日常生活使用現況。本文所收錄的語料中可見源自阿美語及閩南語的借詞使用、噶瑪蘭語與華語的語碼轉換等，著實呈現台灣多語使用現況。因此，透過什麼方式能夠使逐字稿的轉寫，詳細標記語音以及非語音的資訊，是迫切需要進行整理說明的。雖然音檔轉寫成逐字稿、整理分析所需工時甚鉅，但語料採集紀錄對瀕危語言的保存而言，極為重要。本文分享筆者初擬之逐字稿標記原則，希望能夠拋磚引玉，在今後台灣原住民族語自然對話轉寫工作上，作為撰寫逐字稿之參考。

This paper hopes to facilitate the organization of language fieldwork by describing the principles of transcribing vernacular conversation in Austronesian languages in Taiwan. It takes as an example conversational data of the kebalan language collected during the author’s recent fieldwork in the Lide community, Fengbin Township, Hualien County. Unlike corpora of other types of texts, the corpus of this research project comprises audio recordings of vernacular conversation, which document the current daily use of kebalan (also known as Kavalan) in this community. The findings show substantial use of loanwords from Amis and Hokkien and of code-switching between kebalan and Mandarin. These findings reflect Taiwan’s multilingual context and have important implications for the organization of methods used to transcribe detailed phonetic and non-phonetic information. Although transcribing, collating, and analyzing audio files is time-consuming, audio records and their transcription are extremely important for the preservation of endangered languages. This article shares a preliminary draft of an outline of the principles of transcribing such data, which is intended to serve as a reference for future work that involves transcribing vernacular conversations in Taiwanese indigenous languages.