憑藉語料分析制定古代漢語「可靠」詞彙表

Corpus-driven Creation of a Reliable Learner's Vocabulary for Classical Chinese

古代漢語教材大多沒有現代漢語教材、拉丁文教材以及其他古代語言教材的水平與先進的教學法。此外，從統計方面觀察時，使用現有的古代漢語教材的學生要學習使用的詞彙似乎並非以實踐為導向。這是因為詞彙選擇以課文的選擇為主，並不以生詞的使用為主。同時，課本詞彙數量之大、解釋之繁瑣會令人望而止步。筆者計畫憑藉自己的教學經驗，以現代、高效、可靠、便於掌握的詞彙和語法為框架，為外國漢學學生編寫一本新的古代漢語教材。為此，有必要定義分析古代漢語字符頻率的框架。鑒於上述目的，中文「古代漢語」一詞定義為春秋戰國到漢末之書面語。此論文說明編寫這個時期的代表性的語料的標準，並討論語料內容與篇幅 。Michael Loewe 先生於1993年出版的參考書目Early Chinese Texts中介紹的六十四個文本可以作為一個適合漢學學生的代表性語料 。此語料總共約四百萬字，用字一萬兩千五百四十六。通過運行PERL字符頻率分析腳本可以發現：認識六百個常用字的人大約能夠看懂「古代漢語」的百分之八十。也就是說，學 一千一百零五個字能使學生讀懂先秦兩漢文章的百分之九十。論文結尾處附有語料原文列表以及字符頻率分析字數表。

The quality of teaching materials for Classical Chinese often lags behind that of teaching materials available for modern vernacular Chinese, more widely studied ancient languages like Latin and also behind current developments in didactics. At the same time, the learner's vocabulary of a textbook can be a deterrent because of its sheer volume and the way it is presented and organized. The author plans to compile a new Classical Chinese textbook for students of Sinology, based on his teaching experiences and an efficient, reliable learner's vocabulary as well as a grammatical analysis presented in a modern, student-friendly way. To do so, it is necessary first to define the framework for a character frequency analysis of the classical Chinese language. The term Classical Chinese is defined for this purpose as the written language of the time from the Spring and Autumn period until the end of the Han dynasty. The criteria used to compile a representative corpus for this period are summarized, discussing the considerations about content and size of the corpus. The 64 texts presented in Michael Loewe's Early Chinese Texts: A Bibliographical Guide (1993) are found to be a representative corpus for students of Sinology. The corpus thus created consists of a total of around four million characters, containing 12,546 tokens. By running a PERL script character frequency analysis of the corpus, it is found that by learning only the 600 most common characters, a “Classical Chinese” text coverage of 80 % can be reached. A number of 1,105 most common characters would make for an adequate working vocabulary, enabling students to read 90 % of an Ancient Chinese text. The tables contain a list of the corpus source texts and the detailed word counts of the frequency analysis itself.