Abstract
Confusion between the Japanese stop /d/ and flap /r/ happens frequently to Taiwanese learners of Japanese, though it also happens to Japanese native speakers at times. In order to discover the difference in the recognition of these sounds between speakers of Japanese and Taiwanese, I used an identification task to measure categories of recognition, using 'closure duration', 'spike strength', and 'spike (transition) duration' as variables. Comparing both groups of speakers' recognition boundaries and the boundaries of identification at the 80%-level, many divergences were found. The result shows that Japanese native speakers achieved categorical recognition by using 'closure duration' as a major cue, but Taiwanese learners used in order of frequency, 'closure duration', 'spike duration', and 'spike strength' as cues to identify /d/ and /r/.
