英语母语者不同学术文体的计量研究  

A Quantitative Study of Different Academic Genres of Native English Speakers

作　　者：周婉婷 

机构地区：[1]浙江财经大学外国语学院,浙江 杭州

出　　处：《现代语言学》2024年第2期843-854,共12页Modern Linguistics

摘　　要：文体的句法复杂性一直受到学者们的关注,依存距离是评估句法复杂性的指标之一。以往的研究发现,依存距离、句子长度和文体三者密切相关。然而,很少有研究系统地考察英语母语者不同学术文体下,依存距离和句子长度的分布情况。本研究基于LONCESS语料库,采用定量方法研究了议论文和散文的句长分布和依存距离分布,以及两者之间的关系。结果表明:1) 两种文体的句长分布遵循不同的分布模型。2) 两种文体的依存距离分布受体裁的影响有一定的规律性。3) 关于句长和平均依存距离之间的关系,线性回归模型能较好地拟合议论文体,复合曲线模型能较好地拟合散文文体。总之,研究结果表明,两种文体的句子长度和依存距离的概率分布遵循长尾效应,表明人类的工作记忆是有限的。The syntactic complexity of genres has always been of interest to scholars, and dependency distance is one of the indicators for assessing syntactic complexity. Previous studies have found that de-pendency distance, sentence length and genre are closely related. However, few studies have sys-tematically examined the distribution of dependency distance and sentence length in different genres of native English speakers. In this study, based on the LONCESS corpus, a quantitative method was used to investigate the distribution of sentence length and the distribution of depend-ency distance in argumentative and prose genres, as well as the relationship between the two. The results show that: 1) The sentence length distributions of the two genres follow different distribu-tion models. 2) The dependency distance distributions of the two genres are affected by the genres with a certain regularity. 3) Regarding the relationship between the sentence length and the aver-age dependency distance, the linear regression model fits the argumentative genre better, and the composite curve model fits the literary genre better. In conclusion, the results show that the proba-bility distributions of sentence length and dependency distance for both genres follow the long-tail effect, indicating that human working memory is limited.

关 键 词：定量研究 依存距离 句长 文体 句法复杂度 

分 类 号：G63[文化科学—教育学]