借助Sketch Engine和WordNet进行隐喻的概念模型和语义映射分析——以目标域TIME为例

发布时间：2018-10-26 21:17

【摘要】：本文利用新一代语料库检索程序Sketch Engine的词汇特性素描(Word Sketch)功能和语义知识库WordNet的聚类功能,以大型英语语料库BNC为数据来源,以TIME为例对隐喻的概念模型和语义映射进行研究,探索隐喻"整体系统性模式"的语料库研究方法。研究发现:1)Sketch Engine的词汇特性描绘功能在辨别隐喻表达时非常有效,大大减轻手工检索的工作量且准确度较高;2)WordNet的聚类功能在判断源域或目标域的概念和范畴,对概念隐喻的语义进行归类时非常有效;3)同一目标域的若干隐喻表达中,动词性隐喻所占的比率最高,而很多隐喻理论关注的名词性隐喻并不能代表自然语言中隐喻的多样性用法;4)概念隐喻也具有上下义关系,可以通过树形结构反映这种关系。文章最后指出两种工具存在的一些缺陷。
[Abstract]:This paper makes use of the lexical characteristics of the new generation corpus retrieval program Sketch Engine and the clustering function of semantic knowledge base WordNet, and takes BNC as the data source. Taking TIME as an example, this paper studies the conceptual model and semantic mapping of metaphor, and explores the corpus research method of metaphor "global systematic mode". The results show that: (1) the lexical characterization function of) Sketch Engine is very effective in identifying metaphorical expressions, which greatly reduces the workload of manual retrieval and has higher accuracy; 2) the clustering function of WordNet is very effective in judging the concepts and categories of source or target domains and classifying the semantics of conceptual metaphors; 3) among the several metaphors in the same target domain, the proportion of verb metaphors is the highest, while many metaphor theories focus on nominal metaphors, which can not represent the diversity of metaphors in natural languages; 4) Conceptual metaphor also has the relation of upper and lower meanings, which can be reflected by tree structure. Finally, some defects of the two tools are pointed out.
【作者单位】：暨南大学外国语学院;华南师范大学外国语言文化学院;
【基金】：国家社会科学基金项目“隐喻性范畴化的实现机制研究”(项目编号:10BYY002) 广东省高校优秀青年创新人才培养计划资助项目(育苗工程项目)“基于Sketch Engine-WordNet的隐喻语义网络研究”(项目编号:wym11023)的阶段性成果
【分类号】：H05

【参考文献】