Japanese language-oriented English sentence structure (JLOS) is a kind of English sentence structure that Japanese students often employ when thinking in Japanese rather than English. JLOS is hypothesized to have the following elements: illogical, not collocated, not harmonious, and incorrect preposition. The research questions are (1) what type of weak points does Japanese students’ writing present? and (2) does the multivariate analysis model with online learning facilitate in helping students’ writing? The sample size is 58 university students in the Kanto area. The data were collected in 2017 and 2018. POSA was utilized to evaluate whether the structure was significant in Japanese English writing (JLOS). A 3-step model was proposed to improve students’ JLOS writing.

The exiting type of katakana alongside the English lyrics for Karaoke is fardifferent from its phonetic realization of the original song phrase. Under these circumstances, we proposed a new katakana system which approaches the real rhythm and phonetic realization of the original English songs. In so doing, we attempted to reconcile the trade-offs among the phonetic reproducibility, visibility, and affinity so that Japanese learners or singers may have less cognitive load and entertain themselves to have a feeling of self-efficacy. We also report an experiment to verify the effects of our new katakana system on the improvement of pronunciation through singing exercise. The results of the evaluations by five raters in terms of six phonetic evaluation items show that singing the target song along with the new katakana five times leads to better phonetic performance in most of the evaluation items. The results of the questionnaire also support the effectiveness and validity of our new katakana system, suggesting that it may help learners sing or read aloud English phrases and sentences more fluently with greater confidence and motivation