Errors on subject verb agreement made by Japanese learners' English : Effect of syntactic context

20世纪80年代和90年代的二语习得研究表明，学习者所知道的不仅仅是输入中所能得到的，也不仅仅是通过一般的学习策略所能学到的。另一方面，为什么一些特定的项目是非常难以获得的问题并没有吸引研究人员，直到90年代末。研究人员试图回答这个问题，但他们的论点通常依赖于自然主义的数据，这并不完全适合的目的。本研究进行了一系列控制句法语境的实验。结果表明，错误的比例随句法语境的不同而不同，这种差异并不完全归因于主语和动词之间的线性距离。在第一项研究中（技术报告第一章），我指出了理论框架的重要性，并建议语言学理论是最合适的。在第二项研究（第2章）中，提供了三组实验数据。前两个是通过口头翻译任务收集的产出数据，另一个是通过自定进度的阅读任务收集的理解数据。研究三（第三章）进行了另一项口语翻译任务。第四项研究（第4章）收集了事件相关脑电位（ERP）。在第五项研究（第五章）中，我们收集了大量的数据来研究动词的屈折变化系统，这次的材料是英语过去时态的屈折变化，而不是第三人称单数-s。正如报告中所讨论的，我们需要进行进一步的实验，其中一些变量，如语音，句法和话语，是受控的。一些相关的项目，如be动词和助动词have/has在本研究中没有涉及，因此，当然需要进一步的研究。

Studies in second language acquisition in 1980s and 1990s revealed that learners know more than what is available in input and what can be learned with general learning strategies. On the other hand, the question why some specific items are extremely difficult to acquire did not attract researchers until late 90s. Researchers have tried to answer this question since, but their arguments generally rely on naturalistic data, which were not completely suitable to the purpose. This study has carried out a series of experiments in which syntactic contexts were controlled. The results show that the proportion of errors varies in accordance with certain syntactic contexts, and that this variation is not totally attributable to the linear distance between the subject and the verb.The whole study is divided into five. In the first study (Chapter 1 in the technical report), I pointed out the importance of the theoretical framework and suggested that the linguistic theory is the most appropriate. In the second study (Chapter 2), three sets of experimental data were presented. The first two were production data collected with an oral translation task, and the other was comprehension data with a self-paced reading task. In the third study (Chapter 3), another oral translation task was carried out. In the fourth study (Chapter 4), event-related brain potentials (ERP) were collected. In the fifth study (Chapter 5), data were collected to investigate the system of verb inflections, and this time, the material was English past tense inflection instead of 3rd person singular -s.As discussed in the report, we need to carry out further experiments, where a number of variables, such as phonology, syntax, and discourse, are controlled. Some relevant items, such as be-verb and auxiliary have/has, were not dealt with in this study, and hence further research is certainly required.