Building and analyzing an English/Japanese comparable corpus to seek pedagogical insights

在日本和美国大学的合作研究中，我们建立并分析了一个可比较的语料库，该语料库由四种类型的子语料库组成：英语学习者（EL）、英语母语者（ENS）、日语学习者（JL）和日语母语者（JNS）。目标是提高话语层面的认识，重点是实体一致性。在中心理论框架内的语料库分析结果（Grosz et al., 1995）为我们提供了一些教学见解，教师可以在EFL/JFL课堂上使用这些见解。这包括两种语言之间的异同，以及学习者的特定倾向，比如一个被称为CENTER的集中实体的过渡类型，以及它们与CENTER形式选择（如名词、代词）的关系。

In this collaborative research between Japanese and U.S. universities, we built and analyzed a comparable corpus which consists of four types of sub-corpora: English learner (EL), English native speaker (ENS), Japanese learner (JL) and Japanese native speaker (JNS). The goal was to increase discourse-level awareness with a focus on entity coherence. The corpus analysis results, within the framework of Centering Theory (Grosz et al., 1995), provided us with some pedagogical insights that teachers can utilize back in the EFL/JFL classrooms. These include similarities and differences between the two languages and learner-specific tendencies, in terms of transition types of a focused entity called CENTER and their relation to CENTER form choices (e.g., nouns, pronouns).