賽夏族語詞典編輯的語義分析問題

Semantic Analysis and Dictionary Compiling: A Case Study of Saisiyat

本文從語義學出發討論我們在編輯賽夏族語詞典遭遇到的問題，包含詞彙釋義、詞彙之間的語義關係，如近義詞辨析、多詞義之處理及呈現等。我們發現認知語義學有關詞彙語義的研究可以協助我們解決以上的問題，在詞彙定義或釋義方面，認知語義學採用的百科知識觀點（encyclopedic approach），有助我們定義表達特殊文化經驗和涵意的詞彙。詞彙關係方面，近義詞辨析上，除了語義特徵之外，我們發現還必須採用句法方面的次類畫分（subcategorization）與選擇限制（selectional restriction）。多詞義方面，我們發現從語義延伸的觀點分析，可以釐清語義延伸的方向，有助於決定是否分為不同之義項，以及多個義項之間的呈現順序。

This paper deals with the problems on semantic analysis arising from the compilation of a Saisiyat-Chinese dictionary. Issues discussed include how to achieve accuracy in word definition and how to catch the semantic relationships between words. For definition, it is found that cross-language translation may lead to inaccurate definition that is either too wide, too narrow or containing false elements(Goddard 1998). It is thus proposed that explanation rather than translation is employed in defining a word. For dealing with near synonyms and polysemys, it is found that results from the studies on grammaticalization and semantic extension in Cognitive Semantics can shed light on how to determine the relationship between different meanings of the same word.