Analysis of Japanese Language Learners' Dictation Errors

本研究试图从学习者口述错误的角度来分析日语听力理解的难点。听写并不一定反映听力理解，即拼写错误并不一定意味着对内容的误解，反之亦然。然而，学习者写的听写很有价值，因为它以直观的形式显示了一些听力困难。学习者被提供了6种大约4-5分钟长的音频资源，以便他们可以写下来，听他们想听的次数。总的投稿数据为173条。这些听写错误的关系型数据库有望对日语教学有所帮助，尽管本研究没有对此进行探讨。1)听写错误分散在意义单位或语法单位上，使数据库变得复杂。2)从语音的角度来看，错误出现在元音、清浊音、特殊Mora和r/t/k音的组合中。3)从语法角度来看，错误出现在助词和其他听觉表达中，如toritate-joshi，表示体态和语气的句末表达，以及使用toiu、yoona等修饰表达。

This study attempts to analyze the difficulty of JFL listening comprehension through learners' dication errors. Dictation does not necessarily reflect aural comprehension as it is, i.e.wrong spelling does not necessarily mean misunderstanding of contents and vice versa. However, dictation written by learners is valuable as it shows some listening difficulties in a visible shape.Six kinds of audio resources of about 4-5minutes length were given to learners so that they could write them down listening them as many times as they wanted. The number of total dication data is 173. A relational data-base of these dictation errors is expected to be useful for Japanese language teaching, though this has not been explored in this study. Instead, data is printed out so that the trends of errors can be seen.The following issues are discussed.1) Dictation errors spread over meaning units, or grammatical units, making database complex.2) From the phonetic standpoint, errors are found in the combination of vowels, devoiced sounds, special mora, and r/t/k sounds in particular.3) From the grammatical standpoint, errors are found in the particles, and other aural expressions such as toritate-joshi, sentence-final expressions that indicate aspect and mood, and modifying expressions using 'toiu', 'yoona' etc.