建構台灣日語學習者語言資料庫－以LARP at SCU為例

「學習者語言」的研究課題日受關注，在台灣學生於日文系四年接受正規體制教育下日文能力如何形成如何成長，尤其是日語教學領域學者共同關心的焦點。LARP at SCU (Language Acquisition Research Project at Soochow University)即基於此理念所進行的4年研究計畫，為了建構縱斷式的語言成長資料庫，前後有90名教師、研究生參與採集工作，協助接受採集「學習者語言」的是2003年入學的學生37名。方式是一個月一次寫作文，並接受日文訪談，計畫持續至2007年5月學生畢業前，2007年5月16日已完成了33次。語料採取之順序為(1)①書寫初稿的作文（題目統一、約600字内）(2)朗讀作文（録音）(3)教師與學生一對一訪談，給予回饋（用回想法讓學生描述産出表現時遣詞用句的選択。也確認作文内容，並採自我發現之訂正法，自己修正自己）(4)④書寫第二稿之作文(5)教師將第一二次作文與訪談內容文字化，做成電子檔案管理，共同建構語言資料庫。在Windows OS環境中，可用emeditor軟體檢索。隨研究目的亦可在Linux OS環境下，利用「形態素解析程式」計算形態素，或抽出共現結構。本計畫學習者語言的文字化尚在進行中，語言資料庫逐步完成。透過此追蹤式記錄的学習者言語，以習得研究方法處理，可看出台灣日文系學生的日語能力発達過程，例如依音聲、語彙、文法、談話等領域，細分種種小項目，提供探討台灣學生語言能力形成過程，習得順序，而藉此研究結果，期待下階段開發更合理有效的指導法，促進学習效果，本篇為建構此語料庫之過程報告。

關鍵字

LARP ；學習者語言；語料庫；縱斷式語料；習得途徑

並列摘要

Learners' language has attracted increasing attentions and it has become the focus for Japanese language teachers who are interested in Taiwanese students' language acquisition process through 4 year colleague education. In order to better understand this topic, we propose LARP at SCU (Language Acquisition Research Project at Soochow University). It is based on a longitudinal corpus study with which 90 instructors and graduate students are involved, assisting with data analyses and collection. Our subjects are 37 students admitted in 2003 and the project will continue till they graduate in 2007. They have been asked to write once a month with follow-up interviews. We have conducted data collection for 33 times so far and completed the following items: 1) first drafts; 2) recorded recitations; 3) one on one interview with teacher's feedback; 4) second drafts; 5) transferred drafts into electronic files to make the corpus. We can search these electronic files by the soft like Emeditor in Windows OS, and calculate the morphemes or extract the collocations by using the part-of-speech and morphological analyzer ”MeCab” in Linux OS. Though the corpus is still under construction, we hope to complete it step by step and expect to follow up learners language development such as vocabulary, grammar, conversation and their acquisition sequence. Our findings should make significant contributions to language learning and teaching.