English spoken word recognition for Japanese native speakers: A comparison with online and offline processing

Abstract

The aim of this study was to investigate the mechanisms of English (L2) spoken word recognition for Japanese native speakers by two experiments using a cloze test. In Experiment 1, 90 Japanese university students were divided into three groups based on their listening level. Then they took a listening test of a CNN material. The correct rates and error patterns were analyzed. The result showed that the correct rates for three groups were less than 50%, however, the upper group took the highest correct rate and the lowest number of blanks, extra phoneme insertions and phoneme recognition errors. In Experiment 2, 32 Japanese university students were divided into two groups, and one group took a cloze test without listening, and the other took the cloze test with listening. The results of correct number showed that the similar tendency between two groups. This may indicate that listening to L2 is difficult, as much research has suggested, but if Japanese speakers continue to learn English listening, they can better develop recognition of the spoken word more accurately and the inferring skill from top down information such as grammar and semantics may play a role for listening in the second language.