摘要
在现代汉语基层词研究中,完形性是提取基层词的根本标准,词频通常只被视为基层词分级的工具。但本研究发现,在受到客观限制,完形性标准不能充分发挥作用时,利用词义范畴不平衡的特点,"相对词频"可以成功定位具有完形性的准基层词,并通过属性验证和异常值分析,实现对基层词的提取。经检验,该定位法比以往研究中的提取方法更为准确、客观、高效,还可以通过异常值检测到义类词典的范畴划分偏误。
In the research on modern Chinese Basic-level Vocabulary （BLV）, gestalt is regarded as the basic criterion for the extraction of BLV, while word frequency is only used for the classification of BLV. However, it is found that given the imbalance in semantic category, BLV can be positioned and then extracted with the ＂ relative frequency＂ through feature confirmation and analysis of abnormal value when gestalt cannot fully play its role due to objective constraints. After inspection, the corpus-based frequency location method has proved to be more accurate, objective and efficient than other extraction methods in previous studies. Furthermore, the errors in categorical classification in the dictionaries of semantic category can be detected from abnormal values.
出处
《语言文字应用》
CSSCI
北大核心
2014年第4期77-84,共8页
Applied Linguistics
关键词
文本语料库
现代汉语
基层词
词频
定位法
text corpus
modern Chinese
Basic-level Vocabulary
frequency
location method
