The purpose of this study was to examine negative phonetic grades assigned to speeches by Korean students of Japanese. In particular, the study focused on auditory and acoustic-phonetic characteristics in terms of how they are affected by a pause, and whether such characteristics were used for evaluation by native Japanese speakers. For speeches that received negative evaluations, we edited the length of pauses thought to have affected the speech evaluation so that they were the same length as those made by a native Japanese speaker in order to determine whether the edited data would increase the evaluation grade.

Higher marks were achieved when pauses were positioned after sentences, conjunctive particles, the topic marker “wa,” and adnominal modifiers. These results indicate that adjusting pause length may increase speech quality as evaluated by native Japanese speakers.

As there is no [ʥ] phoneme in the Korean language, it is said that many Korean Japanese learners cannot distinguish between syllables of the z series (“za-gyô on”) and those of the j series (“ja-gyô on”). This fact has been revealed mainly by studies using hearing tests.

The purpose of this paper is to clarify how the learners distinct z series sounds from j series sounds using ERPs (Event Related Potentials).

In our experiment, first, we conducted a hearing test to analyze the ability of Japanese native speakers and Korean Japanese learners to distinguish between [dsa] and [ʥa]. Next, the subjects listened to a set of Japanese natural speech sounds /za/ and /ja/ played through ERP equipment. The ERPs were recorded, and later analyzed in order to investigate how well the informants had recognized the sounds.

These results suggest (i) that even the Korean learners of Japanese who obtained relatively low scores in the hearing tests were able to distinguish the sounds through the differences in their acoustic qualities unconsciously, and (ii) that the other Korean students who obtained perfect scores like their Japanese counterparts used different strategies to discriminate these sounds.