Error Analysis of medical research articles by Japanese learner and its application to instruction based on learner corpus data

本研究旨在通过建立一个由日本医学博士生撰写的学习者语料库，并将其与英语本族语语料库进行比较，为日本EFL（英语作为外语），特别是EMP（医学英语）学习者的学术写作奠定理论基础，并建立有效的教学方法。我们收集了少量的学习者数据，并编制了一个小型语料库，名为：医学模拟实验学生摘要（Students' Abstracts of Medical Mock Experiment，简称MERS）语料库。学生们被要求就一个给定的模拟医学实验制作一份至少200个单词的英文摘要。摘要必须按照IMRD（引言、材料和方法、结果和讨论）结构编写，但必须作为一个有意义的、连贯的段落定稿，而不是作为单独的章节。实验的主题是幽门螺杆菌感染，并用日语简要解释了实验的概要。每一份手稿都以电子方式收集和汇编成文本文件。使用语料库软件AntConc对数据进行分析。我们在两个国际会议上介绍了研究成果。结果表明：1）有能力写出语法正确的句子的学习者并不一定会使用风格恰当或有体裁意识的表达方式; 2）学生的体裁意识与准确性水平无关。根据准确性和体裁意识的不同，学习者可以大致分为四类。我们的结论是，从研究生阶段开始教授学术写作已经为时已晚，应该鼓励日语学习者在早期阶段，也就是大学生阶段，掌握基本的学术写作技能。

The purpose of this research is to both ground theory and establish effective teaching methods of academic writing for Japanese EFL (English as a Foreign Language), particularly EMP (English for Medical Purposes) learners, by creating a learner corpus database written by Japanese medical PhD students, and comparing it with an English native speaker corpus. We collected a small amount of learner data and compiled a small corpus database entitled: Students’ Abstracts of Medical Mock Experiment (SAME) corpus. The students were required to produce an English abstract of at least 200 words on a given mock medical experiment. The abstract was required to be written according to the IMRD (Introduction, Materials and Methods, Results and Discussions) structure, but had to be finalized as a single meaningful, cohesive paragraph, not as separate sections. The topic of the experiment was a Helicobacter pylori infection and the outline of the experiment was briefly explained in Japanese. Each manuscript was collected and compiled electronically as a text file. The data was analyzed using the corpus software AntConc. We have presented the research results at two international conferences. The results indicated that 1) learners who have the ability to write grammatically correct sentences do not necessarily use stylistically appropriate or genre conscious expressions, and 2)students awareness of genre was not correlated to level of accuracy. The learners can be categorized roughly into four groups according to their levels of accuracy and genre awareness. We concluded that it would be too late to start teaching academic writing from a graduate level, and that Japanese learners should be encouraged to acquire basic academic writing skills at an earlier stage, that is, as undergraduates.