Developing a database of native speaker sample sentences for EFL learners providing an appropriateness judgment service for user-provided sentences.

这个项目的主要目的是建立一个不受版权保护的样本句子数据库，其中包含由训练有素的母语人士创建的基本词汇，供日本大学水平的英语学习者使用。该数据库包括一个用户友好的浏览功能。最初的计划是增加另一项服务，通过这项服务，用户可以询问母语人士对他们感兴趣的特定英语句子的适当性的判断。该数据库至少可以在以下三个领域为英语教学做出贡献：(1)英语教师和材料开发人员可以在课堂活动或材料开发中使用这些样句，而不会有侵犯版权的潜在风险；(2)由于浏览功能，英语学习者可以很容易地获得适合自己水平的样句，从而获得大量真实的输入；(3)研究EFL和第二语言习得的学者可以通过调查EFL学习者使用现有系统的模式，以及在有和没有使用系统的班级之间进行比较实验，来更深入地了解这种系统的影响。虽然句子适当性判断的功能尚未实现，但该项目已经成功地创建了一个由五个模块组成的数据库：基于JACET 8000个单词的基本词汇样句，高错误概率句子，每个500个单词的模型段落，模型文章，以及由学习者生成的具有母语人士更正的句子组成的平行语料库。开发了以并行语料库为主，同时也适用于其他组件的浏览系统。除了这五个主要组成部分外，由于探索将原始系统扩展到多语言环境的可能性，系统中还增加了专门用于特定副词-形容词模式的多语言样句组成部分。项目组积极参与相关研究活动，在各类学术会议上发表演讲，发表论文多篇。此外，该团队还邀请了版权、平行语料库、语料库语言学、生成语法、英语写作、语用学等方面的专家，举办了多次讲座会议。此外，该小组在过去三年中每年夏天都组织了一个关于面向语料库语言学的应用和统计的讲习班。少

The primary purpose of this project was to build a copyright-free database of sample sentences containing essential lexical items created by well-trained native speakers, for use by Japanese college level EFL learners. The database includes a user-friendly browsing function. The original plan intended to add another service, through which users could inquire about native speakers'judgment on the appropriateness of particular English sentences in which they are interested.The database could contribute to EFL teaching and learning in at least the three following domains : (1) EFL teachers and material developers could use those sample sentences in their classroom activities or material development without potential risk of copyright violation, (2) EFL learners could have easily access sample sentences appropriate to their level, and consequently receive a large amount of authentic input, owing to the browsing function, and (3) researchers of EFL and second language acquisition could obta … More in insight on the effects of the systems of this kind by investigating EFL learners'patterns of accessing current system and by conducting comparative experiments between classes with and without access to the system.Though the function for sentence appropriateness judgment has not been actualized yet, the project has succeeded in creating a database consisting of five modules : Sample sentences for essential lexical items based on JACET 8000 words, sentences with high probability of errors, model paragraphs of 500 words each, model essays, and a parallel corpus composed of learner-produced sentences with native speaker corrections. The browsing system, chiefly for the parallel corpus yet applicable to the other components, has also been developed. In addition to these five major components, a multilingual sample sentence component specialized for a particular adverb-adjective pattern has also been added to the system as a result of the exploration of possibly extending the original system to a multilingual environment.The project team has been very active in related research activities including presentations at various academic conferences and publishing many papers. In addition, the team has held many lecture conferences by inviting experts on the issues of copyright, parallel corpus, corpus linguistics, generative grammar, EFL writing, pragmatics, and so forth. Moreover, the team has organized a workshop on corpus linguistics-oriented applications and statistics every summer for the last three years. Less