摘要
文章以政府工作报告、讲话稿、访谈录、电视剧对白等语料为考察对象,根据语体语法理论,运用语距定律(体原子),从语料中提取了词长、嵌偶词词频、合偶词词频、词汇复杂性(甲乙丙丁四级词汇)、动态助词词频、平均小句长、陈述句比例等10项评价语体正式度的特征指标,利用熵值法对各项指标赋权,构建能够实现汉语语体正式度自动测量的模型,利用盲测样本对模型进行了有效性检验。利用测量模型对历年国务院政府工作报告的语体正式度进行测评,结果发现:政府工作报告的语体正式度相对稳定,并逐渐向正式度更高的方向发展。
This paper makes surveys on government reports,speech texts,interview scripts,TV play dialogues,etc.From these linguistic materials,according to Stylistic-register Grammar Theory and Law of Register-distance(Register-Atom),up to ten features are extracted to evaluate the stylistic-register formality,including word length,frequency of monosyllabic words in disyllabic template,frequency of disyllabic couplets,lexical sophistication(four levels:A,B,C,D),frequency of dynamic auxiliaries,average length of short sentences,proportion of declarative sentences.These features are all empowered using the Entropy method,to establish the model capable of automatically measuring the formality of stylistic-register in Chinese.Its validity is also verified in blind test specimens.Finally,the model is employed to measure and evaluate over-the-years government reports of the State Council in terms of the formality of stylistic-register.The results show that the formality stays relatively stable while gradually turning higher.
作者
王用源
陈宇豪
WANG Yongyuan;CHEN Yuhao
出处
《语言教学与研究》
北大核心
2023年第4期79-90,共12页
Language Teaching and Linguistic Studies
关键词
语体语法
语体特征
体原子
语体正式度
自动测量
stylistic-register grammar
stylistic-register feature
register-atom
stylisticregister formality
automatic measurement