名词分布是人类语言的不变量吗?--以德语书面语中名词分布为例  

Is the Distribution of Nouns an Invariant in Human Languages?--An Investigation Based on Written German Corpora

作　　者：李媛[1] 段庭辉 刘海涛[1] Li Yuan;Duan Tinghui;Liu Haitao(School of International Studies,Zhejiang University,Hangzhou 310058,China;Institute for Germanic Linguistics,Friedrich Schiller University Jena,Jena 07745,Germany)

机构地区：[1]浙江大学外国语言文化与国际交流学院,浙江杭州310058 [2]耶拿大学日耳曼语言学系,图林根耶拿07745

出　　处：《浙江大学学报（人文社会科学版）》2019年第6期39-48,共10页Journal of Zhejiang University:Humanities and Social Sciences

摘　　要：此前对人类自然语言中词类分布的研究显示,不同语言中名词所占比例相对固定。德语中名词所占比例是否也符合这一普遍规律?通过对三个大型德语语料库进行研究发现:首先,德语书面语中的名词占比约为38%,尽管德语复合名词比例高、名词化结构多,但其名词占比同英语以及其他语言中的名词占比大致相当,从而进一步证实了人类自然语言中名词占比具有普遍规律这一结论;其次,不同文体中名词及其各子类的占比有所差异,而这一差异由文体特征决定,并且具有跨语言的相似性;最后,时间因素与文体类型均对名词各个子类占比有显著影响,但名词总体占比未受二者影响。综上,可以进一步证实名词分布是人类语言的不变量这一结论。Hudson indicates that the proportion of nouns in written English is about 37%.Since then,many other languages haven been studied in this respect,finding out that the proportion of nouns in all human languages is an invariant.German and English have differences in word formation,though they both belong to the West Germanic language subfamily.As for nouns,on the one hand,German has a larger proportion of compound nouns,resulting in intensive information,thus the total quantity of its nouns could be relatively smaller than that of other languages;on the other hand,nominalized structures are common in German,which may cause a larger proportion of nouns in comparison with other languages.Does German conform to the universal law of language?We try to answer this question based on three large-scale corpora of German:The DWDS-Kernkorpus consists of texts of different genres from the 20 th Century and has more than 100 million words in total;The Deutsches Textarchiv(DTA)is a diachronic corpus of written German and contains about 150 million words from texts of the same genres as DWDS-Kernkorpus;The TüBa-D/Z treebank is a German newspaper corpus with more than 1.5 million words,containing 3 644 mainstream newspaper articles of Die Tageszeitung from 1989 to 1999.In order to make the results comparable,we adopted the same classification criteria for nouns and the part-of-speech tag sets suggested by Hudson.The result shows that the proportion of nouns in all three corpora of written German is about 38%.Thus,the above-mentioned hypothesis is corroborated.Furthermore,we studied the relationship between the proportions of nouns in different genres.Differences exist between different genres in terms of the proportions of subclasses of nouns including common nouns,proper nouns and pronouns.While common nouns are larger in proportion in informational texts,imaginative texts have a larger proportion of pronouns.This result also complies with that of Hudson.Little work has previously been conducted with the diachronic development o

关 键 词：德语 名词分布 语料库 计量特征 文体 历时变化 

分 类 号：H33[语言文字—德语]