中文动词及分类研究:中文动词词汇语义网的构建及应用  

Chinese Verbs and Their Categorization: Construction and Application of the Semantic Network of Chinese Verbs

作　　者：刘美君[1] 万明瑜 Liu Mei-Chun;Wan Mingyu

机构地区：[1]香港城市大学翻译及语言学系

出　　处：《辞书研究》2019年第2期42-60,I0002,共20页Lexicographical Studies

摘　　要：随着人工智能的日益发展,语言学成为"产""业""学"界寻求合作及突破的新契机。其中语言学语义资源的构建及标注问题成为了当前的一大热点及难点。文章针对中文动词语义分类问题,从理论研究、语义网构建及实践应用三方面进行了全面的探讨和分析。理论研究上,文章以"框架为本,构式为用"的研究方法为基石,依循框架语义和构式语法以区分动词和构式之间的"形-义"搭配,形成"格式塔"(Gestalt)般互补。语义网构建上,以语言学分析为基础,语料实证为依归,通过"框架元素"与"定义性构式"来定义动词属性,使语料兼有词汇表征、框架阶层及语义标注等信息。语义知识库目前包含"沟通""认知""感知""情绪""评价""社会互动""自动"和"致使移动"八大类框架动词,已有效运用于多种基于语义及事件框架的中文自然语言处理任务,包括中文语义自动消歧,自动语义角色标注,事件框架甄别及故事自动生成。With the growth of AI,linguistics has become a new field in which researchers in different areas can cooperate and make breakthroughs.The construction and labeling of semantic resources have thus turned into a hot topic,as well as a difficult point to overcome.This article discusses Chinese verbs and their semantic categorization from three perspectives,namely,theoretical bases,semantic network construction,and application.The theory used for this research is framework semantics and particularly Gestalt;the construction of the semantic network is based on linguistic analysis and language data.The semantic knowledge base currently consists of eight types of verbs and has been effectively used in many kinds of Chinese natural language processing tasks,including automatic semantic disambiguation,automatic semantic role labeling,event identification,and automatic story generation.

关 键 词：框架语义 构式语法 动词分类 语义标注 自然语言处理 

分 类 号：H146.2[语言文字—汉语]