[PS-2.21] The Relationship between Working Memory and L2 Simultaneous Oral Reproduction (Content Shadowing) Processing

Yamauchi, Y. 1 , Kunikoshi, A. 2 , Nishikawa, M. 3 , Kawamura, A. 1 , Husky, K. 1 & Minematsu, N. 4

1 Tokyo International University
2 McRoberts B.V.
3 Tokai University
4 University of Tokyo

Since capacity of working memory (WM) is limited, the more advanced L2 learners become, the more effectively they use WM. Content shadowing (CS) requires L2 learners to decode auditory information while storing main points of target passages and simultaneously reproduce messages orally. CS cognitive load is so high that WM effective use is expected to be highly related to CS performances. However, because of difficulty in objectively testing WM and CS performances, little research has been done so far. This study aimed to investigate the relationship between CS performances and WM effective use by developing an on-line WM test based on backward digit span (BDS), and an automatic CS assessment system was employed. Sixty-six Japanese EFL learners listened to and memorized sequences of three to ten numbers, choosing the correct option out of four which identified the same numbers printed backwards on PC monitors. They then performed CS while listening to English passages never presented before. The assessment system using latest speech technology GOP automatically analyzed their shadowed speech and showed numerical scores. Statistical analysis revealed WM effective use had significantly high correlation with CS and overall proficiency. Thus CS could facilitate L2 processing becoming more effective, accelerated, and automatized.