日本大学生産工学部　研究報告B（文系）第51巻

3/60

─ 1 ─１．はじめにコーパス言語学や自然言語処理の技術は，外国語のデータ駆動型学習（data-driven learning: DDL）に応用されてきた。DDLは，学習者にコンコーダンスと呼ばれるツールを使ってコーパス（電子化されたテキスト）からターゲット語を検索させ，パソコン画面に表示された豊富な言語使用例を観察してことばの規則を発見させる外国語教授法である（Aston, 2001）１）。例えば，図１（前置詞out ofの検索結果）のような複数の例文を観察して，out ofの前後に共通して現れている品詞のパターン（例えば，動詞に続いている，後ろに場所を表す名詞句が続く）やout ofが使われている場面や状況などの共研究ノート日本大学生産工学部研究報告Ｂ2018　年　６　月　第　51　巻教育用例文コーパスSCoRE第４次開発中條清美＊，濱田　彰＊＊，若松弘子＊＊＊，小林雄一郎＊＊，横田賢司＊， 内山将夫＊＊＊＊，赤瀬川史朗＊＊＊＊＊，ジョンソン・ミシェル＊＊， 西垣知佳子＊＊＊＊＊＊The Fourth Phase of Development of the Sentence Corpus of Remedial English （SCoRE）Kiyomi CHUJO＊, Akira HAMADA＊＊, Hiroko WAKAMATSU ＊＊＊, Yuichiro KOBAYASHI＊＊, Kenji YOKOTA＊, Masao UTIYAMA＊＊＊＊, Shiro AKASEGAWA＊＊＊＊＊, Michelle JOHNSON＊＊, and Chikako NISHIGAKI＊＊＊＊＊＊The Sentence Corpus of Remedial English (SCoRE) is a free, web-based corpus and concordancer created for beginner level L2 students, teachers and material writers. The purpose of this paper is to introduce the fourth phase development and organization of this specially-developed SCoRE program. The details are as follows: (1) all of the 10,113 English sentences previously created were reviewed, and as a result, 1,254 sentences were revised so that SCoRE currently contains 22 categorized grammar items with 10,459 level-specic, semi-authentic sentences written to satisfy particular pedagogical considerations and fair use for copyright issues. In addition, 5,217 beginner-level sentences containing 14 topics were created and will be added to the next version of SCoRE to expand its topical coverage; (2) an access log-recording system was added; and (3) a more convenient, handy concordance tool, mobile SCoRE (m-SCoRE), has been created and added for using SCoRE on a smartphone and/or a tablet.Keywords:Educational Corpus, Data-Driven Learning, Sentence Corpus, Mobile Concordancer, Remedial English　　　　　　＊日本大学生産工学部教養・基礎科学系教授　　　　　＊＊日本大学生産工学部教養・基礎科学系助教　　　　＊＊＊茨城工業高等専門学校非常勤講師　　　＊＊＊＊情報通信研究機構主任研究員　　＊＊＊＊＊Lago言語研究所代表　＊＊＊＊＊＊千葉大学教育学部教授