CTLJ誤用辭典之開發與分析

本研究介紹以CTLJ為基礎開發之誤用辭典系統，並分析誤用辭典中各學習階段之誤用率，提供日語教學者之參考，以減少學習者之誤用。本誤用辭典系統主要以Perl和HTML開發，提供查詢學習者之誤用字及訂正字，統計誤用出現頻率，俾便助於日語教育、日語研究、或以中文進行對照分析為目的而編纂。使用者可利用系統提供之關鍵字、學習時數、詞性與誤用型別等作為查詢條件，亦可將查詢結果下載。CTLJ已建有53,865組誤用標誌（error tag），誤用辭典係將上述之誤用標誌涵蓋的部分納入予以開發建置，其中包含誤用之句子數總計為53,497，誤用之詞素累計達73,371、個別詞素也有9,122例等。經分析結果得知學習時數和誤用傾向之間並無一定比例，資料顯示學習時數遞增而誤用遞減之情況僅出現於部分之學習時數。此外，誤用辭典中每個學習階段之誤用詞類樣態不一，然而，動詞誤用頻率高過其他詞類為所有學習時數之共同現象，是以減少動詞誤用為日語教學者後續優先目標。

關鍵字

CTLJ ；誤用辭典；語料庫；品詞；檢索功能

並列摘要

This paper presents a misused Japanese dictionary system based on CTLJ, and investigates how the dictionary system can be used to analyze the misuse rate of the learners of Japanese at each level of their learning This system was developed by using Perl and HTML with the objective of enabling the learners of Japanese to access misused words or corrections, misuse frequencies, in order to facilitate their language learning Japanese language studies, and contrastive analysis of Japanese and Chinese. This system also provides key words, learning periods, parts of speech, and misuse types. Users in addition can download Excel attachments regarding their respective search results. Since the CTLJ already includes 53,865 error tags, the proposed dictionary system is constructed by incorporating the rich resource of the error tags. In the system, the sentence level misuses amount to 53,497, and the total number of the morphemes in misused words is 73,371, which further falls into 9,122 individual types. Analytical results reveals that misused tendency is not proportional to learning hours, and the fact that misuses decrease as learning hours increase only appears in parts of learning hours. Aside from that, misuse words and types in the dictionary system are obviously different in each learning phase, however, it is common that the misuse rate of the verb is higher than other word types for all learning hours. Therefore, the reduction of verb misuse would be the future priority goal for Japanese educators.