基于依存树库的语言计量特征对比分析——以乔治·华盛顿和唐纳德·特朗普就职演说为例

Contrastive Analysis of Linguistic Quantitative Characteristics Based on Dependency Treebank:A Case Study of Inauguration Speech by George Washington and Donald Trump

基于语料库研究方法,以依存语法为理论框架,分别从依存距离分布和主语、宾语、定语以及状语的词类构成上,对美国首任总统乔治·华盛顿和现任总统唐纳德·特朗普的就职演说进行了量化对比分析,发现唐纳德·特朗普演讲词比乔治·华盛顿演讲词的依存距离更小,文本更容易理解;乔治·华盛顿演讲词中Wh-词出现在主语中的频率远远高于唐纳德·特朗普顿演讲词;乔治·华盛顿倾向于使用形容词作定语,而唐纳德·特朗普倾向于使用形容词性物主代词作定语;此外,副词在乔治·华盛顿和唐纳德·特朗普演讲词中充当状语的功能上并不完全占优势。因此,对语言进行计量特征对比分析有助于对文本内容的理解以及对文本复杂度的判断。

Based on corpus and the Dependency Grammar, this study makes a contrastive analysis on the distribution of dependency distance and the correspondence between dependency relation and word class (subject, object, attribute and adverbial) of the inauguration speech by George Washington and Donald Trump. It suggests that dependency distance of George Washington’s speech is much greater than Donald Trump’s, which indicates that George Washington’s speech is more difficult to understand than Donald Trump’s. Besides, the frequency of Wh-words as subject in George Washington’s speech is far higher than Donald Trump ’s. What’s more, George Washington tends to use adjectives as attribute, while Donald Trump tends to use possessive pronoun as attribute. This paper also finds adverbs as adverbials in George Washington’s and Donald Trump’s speeches are not entirely overwhelming. Therefore, this paper points out comparative analysis of the quantitative features of language contributes to the understanding of the content and the judgment of the text complexity.