Preliminary Study On Building Small Spoken Corpus Of Pingxiang Dialect

Based on the fieldwork of recording spantaneous discourse data of Pingxiang dialects in Pingxiang City's urban distric and suburb countrysides nearby in Jiangxi Province in China,this dissertion is devoted to make a basic transcription of discourse data with the software tool EXMARalDA in order to build a small speech corpus of Pingxiang dialect,and make the description and quantitative analysis of speech phenomena and linguistic features,such as word frequency and coocurrence.The main contents of this dissertion are as follows:In Chapter One,I make an outlines of Pingxiang's human geography and the Chinese dialect distribution,and also paraphrase the theme of this dissertion and the research methods.In Chapter Two,this paper systematically combs and reviews the relevant research on the theories,rules,and transliteration of the spoken corpus at home and abroad.In Chapter Three,this paper expounds the genre of natural discourse and the classification of subject matter and the corpus genre and subject selection of Pingxiang dialect oral corpus?the methods and steps of corpus collections researches?ethical issues in the practice of corpus collections Researches?data quality control problem of corpus collections researches.In Chapter Four,the DT2 transliteration rules are used to mark the natural discourse of Pingxiang dialect.Then I expounds the method and procedure of transliteration annotation and how to deal with the phenomena in discourse corpus.In Chapter Five,Using software to build a corpus and makeing a preliminary statistical analysis of the discourse phenomenon and the dialect vocabulary in the natural discourse of Pingxiang dialect.It is revealed that in the spoken language of Pingxiang dialect,the speech features are the most widely distributed,such as long,small pauses,pauses and overlapping of words.And reflected in daily verbal communication,4% of high-frequency words can cover 80% of words.