An Open Information Structure Approach to Stepwise Japanese–English Translation Exercises: Designing and Evaluating Syntactic Rule Learning via Manipulable External Representations

This study proposes a stepwise Japanese-English translation exercise system that operationalizes syntactic rulelearning through manipulable external representations, based on the Open Information Structure Approach. The proposedmodel decomposes the translation process from Japanese to English into a series of intermediate representations—such asseparated sentences, Japanese word order, role-based structures, English word order, and the final English sentence—andpresents each stage as an external representation that learners can reconstruct by manipulating cards. In doing so, thesystem exposes the underlying information structure of the translation process in a machine-readable and learner-operableform, allowing learners to experience syntactic transformation as a sequence of explicit operations rather than as an opaque,one-shot conversion. The system was implemented and used in a 90-minute classroom session with 113 first-year studentsat a college of technology (three classes). Learners engaged in stepwise reconstruction tasks and subsequently completedword-ordering exercises as well as self-evaluation items regarding their understanding of syntactic rules and translationprocedures. The results showed that, particularly for lower-achieving students, performance on English word-ordering tasksimproved with a large effect size (r ≥ .5). For higher-achieving students, self-evaluations indicated significant gains in theirability to explain syntactic rules and translation strategies. These findings suggest that designing learning activities as themanipulation of explicitly structured external representations, following the Open Information Structure Approach, canenhance both procedural skills and metacognitive awareness in syntactic rule learning. The study also discusses implicationsfor extending this approach to other language pairs and to structural learning tasks in different subject domains.