以日語解讀教科書、論說文、漫畫為調查對象之量化詞彙調查研究－以不同文本類別間的詞彙特徵為主要分析範圍

A Statistical Word Count Research on Japanese Textbook Reading, Editorial Texts, and Manga: Lexical Characteristics among Different Text Types as Main Scope of Analysis

本文以日語讀解教科書、論說文、漫畫為調查對象，佐以量化分析的手法探討上述三種不同文本類別間的詞彙使用特徵。具體研究方法主要從語種、詞性、高頻詞、相似度等觀點檢討三者問詞彙使用的異同。其中，利用「相似度」所進行的分析顯示，讀解教科書與論說文的詞彙最為相似，讀解教科書與漫畫間的相似度次之，論說文與漫畫間的相似度最低。此外，本文進一步刪除其中級程度的日語學習者已習得的詞彙，再次計算三種文類間的相似度後，發現論說文與漫畫間的相似度呈現逆轉，高於讀解教科書與漫畫間的相似度。究其原因，應是刪除掉既習得的詞彙後，漫畫本身所具有的資料解說性特質才得以一舉浮現，此特質也因此反映在其結果上。

This study is a research on Japanese textbook reading, editorial texts, and manga, with the approach of quantitative analysis, to explore the characteristics of lexical usages among the three text types. The physical research method is mainly from the perspectives of origin type, part of speech, high-frequency word, and similarity to explore the similarities and dissimilarities of lexical usages among the three. The result, after carrying out the analysis by applying ”similarity”, reveals that lexicons are the most similar in textbook reading and editorial texts, less similar in textbook reading and manga, and the least similar in editorial texts and manga. Moreover, it is found that the similarities between editorial texts and manga represents quite the reverse, and they have more similarities than those between textbook reading and manga after the learned lexicons of intermediate level are further omitted in this study and the similarities are calculated among the three text types again. As we look into the cause, that the characteristic of data explanation manga itself possesses gets to appear instantly and entirely should be due to the omission of learned lexicons, thus the characteristic is reflected on its result.