Study On The Discourse Structure Of Biographies In Historical Records

Different from the textual linguistics research,the standpoint of this paper is the semantic syntactic category(Li Baojia,2001/2007;Qiu Xuemei,2011/2013).The paper chooses 69 texts from Historical Records(about 2,520,000 words)as corpus material to make analysis of in-sentence structure of ancient Chinese Language.Here we introduce the research steps.Firstly,we make the tagged corpus which is suitable for the further analysis according to the theory of Chinese topic-comment structure syntax.The study establishes a discourse structure information corpus annotation excerpts from Historical Records.Then on the one hand,we find out the factors that are devoted to the formation of the distribution features by the statistical results;on the other hand,we find out those special cases which existed theory can?t explain and make them assemble.Finally,we make discourse structure analysis of our statistical and analytical results,then revise the existed theory while comparing our discovery with modern Chinese topic-comment structure to built the new topic-comment structure adapted to ancient Chinese Language.Our discourse structure statistical results can be stated as follow: the text chosen from Historical Records have 20055 tagged sentences.The distribution of subject-predicate sentence has an absolute advantage.The total quantity is 13015,occupying 64.9%.In interior classification,shidong type occupies 71% in subject-predicate sentences,the other four types(shiyuan type,shiren type,shiyan type,shigan type)occupy 29%.The percentage of topic-comment sentences is 34.9%.In interior classification,the distribution of zhuangtai type,shuoshi type and suoduan type have obvious advantage,which occupies over 10% respectively.Jiyong type and huansuan type in shuliang category,gongju type,cailiao type,yongchu type and yingdui in shiyong category,guanshe type and zhendui type in yingdui category do not exist.The rest topic-comment sentences have a large number types and low existing rate.The other special types(question,exclamation and independent element)has a extremely low existing rate.The especial parts of our statistical results can be stated as follow: the subject-predicate sentence in chosen text has a large percentage;topic-comment sentence is less balanced and multiple in distribution;jishu type,gusuan type and suogui type have an advantage;pingjia type have obvious differences from modern language.From our opinion,the special forms of historical records and special discourse structure of ancient Chinese result in the unique distribution features.After analyzing the distribution features of tagged sentences,we revise the existed discourse structure theory to built the topic-comment structure adapted to Historical Records.The paper partly reflects the discourse structure types and distribution features of ancient Chinese language(Old period).On basis of topic-comment structure adapted to Historical Records built by us,we can get close to the overview of discourse structure types and distribution features of ancient Chinese language(Old period)step by step if we make analysis of more historical texts.