副词“几乎”的分布验证与语义提取

On the distributional verification and semantic extraction of the adverb jihu

本文以语义语法为理论基础,在精确定位副词"几乎"句法分布并正反验证其同现成分语义类型的基础上,把有界极性敏感副词"几乎"的语法意义界定为"客观估测极近极界值"。首先,从语义性质、语义指向、语义特征、语义演变四个方面综述前人关于"几乎"的基本观点,并指出三点不足:句法分布描写缺乏精确定位、语义指向概括缺乏系统分类、语义特征提取缺乏正反验证。其次,基于CCL语料库的调查,精确定位副词"几乎"的句子功能、句法位置和关联成分。第三,根据"几乎"的句法分布,并结合句法特征和逻辑关系准确概括"几乎"语义所指向的四种语义类型:存量满空值、时量高低值、比量异同值、能量偏常值。第四,根据"几乎"的语义指向,提取"几乎"的语法意义,并正反验证"几乎"的语义特征:[+客观性]、[+估测性]、[+界极性]、[+近极性]。最后,阐明副词研究应该加强句法分布的精确定位、语义指向的准确分类、语义特征的正反验证,以此为近义副词的辨析奠定基础。

Based on the theory of semantic grammar, this paper accurately locates the syntax distribution of the adverb jihu （几乎） and verifies the semantic types of its co--occurrences both positively and negatively, which pave the way to define the grammatical meaning of this polarity-- sensitive adverb as ＂to objectively estimate the degree of limitation nearness to the bounded extre mum＂. First, this paper summarizes the previous basic viewpoints on jihu from four aspects： se- mantic properties, semantic orientation, semantic features and semantic evolution, based on which it points out the following three shortcomings： syntactic distribution description lacks accurate loca- ting, semantic generalization lacks systematic classification and semantic feature extraction lacks positive and negative verification. Second, this paper explores the syntax function, distribution and relative components of jihu based on the corpus of CCL. Third, according to the syntactic distribu tion and combining the syntactic features and logical relations, this paper generalizes four semantic types of jihu： the value of storage, temporal length, the difference and sameness of comparison, and normality of ability. Forth, based on the semantic orientation analysis, this paper extracts the grammatical meaning of jihu and verifies its semantic features both positively and negatively, that is, objectivity, estimation, limitations and nearness. And finally, it argues that the studies on ad- verbs are supposed to focus on the syntactic distribution, the semantic classification and the verifica- tion of semantic features to distinguish those synonymic adverbs.