Entity Coherence in Comparable Learner Corpora: Seeking Pedagogical Insights

Entity Coherence in Comparable Learner Corpora: Seeking Pedagogical Insights Mitsuko Yamura-Takei a, Miho Fujiwara b, and Etsuko Yoshida c a Faculty of Law, Hiroshima Shudo University, 1-1-1 Ozukahigashi, Asaminami-ku, Hiroshima 731-3195, JAPAN takeim@shudo-u.ac.jp b Department of Japanese & Chinese, Willamette University, 900 State St. Salem, Oregon 97301 USA mfujiwar@willamette.edu c Faculty of Humanities and Social Sciences, Mie University, 1577 Kurima-machiya-cho, Tsu, Mie, 514-8507, JAPAN tantan@human.mie-u.ac.jp Abstract. This paper describes an ongoing collaborative project, between Japanese and U.S. universities, that aims to build, analyze and use comparable learner corpora in an attempt to promote discourse-level proficiency in foreign language learning contexts. The focus is placed on discourse coherence created by reference to nominal and clausal entities. The corpus analysis results, within the framework of Centering Theory (Grosz et al., 1995), will be presented, along with some pedagogical insights that teachers can utilize, back in the EFL/JFL classrooms where the data was initially collected. Keywords: learner corpora, entity coherence, event reference, centering theory 1. Introduction Discourse competence, one of the four principal components of communicative competence put forward by Canale (1983), is defined as the ability to connect sentences in stretches of discourse and to form a meaningful whole out of a series of utterances. Unlike “hard rules” at morphological and syntactic levels, many discourse phenomena are usually governed by “principles” or “preferences.” Therefore, it is hard to provide explicit and systematic instructions. While discourse-level research and practice have been attracting attention (cf. McCarthy, 1991), it is undeniable that a focus on sentence-level form tends to be emphasized in many foreign language teaching contexts. Therefore, it is crucial to develop effective ways for learners to better understand the concept of discourse coherence and to utilize relevant cohesive devices. Corpus-based research in the past two decades has unquestionably had a considerable impact on various linguistic sub-disciplines. Foreign language pedagogy is not an exception. Corpora come in various shapes. For instance, learner corpora are structured collections of language produced by language learners. Comparable corpora are two (or more) sets of corpora in different languages designed along the same lines. Learner corpora can provide information about how learner production differs from a target model and thus can inform the field of second language acquisition (SLA), for which a comparable corpus of native speaker texts is required. In an attempt to integrate these two pedagogical challenges and promises, we have designed the collaborative corpus project between Japanese and U.S. universities. In this paper, we will  The work reported in this paper was partially supported by Hiroshima Shudo University Research Grant (2010). Copyright 2010 by Mitsuko Yamura-Takei, Miho Fujiwara and Etsuko Yoshida. PACLIC 24 Proceedings 779 first describe the project and then present our corpus analysis results, with a focus on entity- coherence that is characterized in the centering framework (Grosz et al., 1995). Special attention will be paid to event reference found in the data. 2. Corpus Project Our ongoing collaborative corpus project, between Japanese and U.S. universities, aims to (1) build, (2) analyze and …

Content truncated to 500 words