The Analysis of Chinese Sentence Semantic Chunk Share Based on HNC Theory

The Analysis of Chinese Sentence Semantic Chunk Share Based on HNC Theory Quan Zhang1, Chen Wu1,2, Xiangfeng Wei1 1 The Institute of Acoustics, CAS, Beijing 100080 P.R.C. 2 The Graduate School, CAS, Beijing 100039 P.R.C. E-mail:zhq{at}mail.ioa.ac.cn Abstract. This paper analyses anaphora from the view of concept association vein. It probes into the problem of sentence semantic chunk share based on the sentence category expression of HNC theory. Firstly, the chunks share is divided into two kinds, one is sharing in a sentence where there is chunk extension into sentence in the sentence, and the other is sharing between two sentences. Secondly, the anticipation knowledge of sharing in one sentence is acquired by deduction from the definition of sentence category which chunk extends into sentence. Then, the classification of sharing between two sentences and corpus investigation are presented, the distribution data of the share is counted, and preliminary analysis for the data is put forward. The sentence analysis processing can benefit from the result. Keywords: anaphora, HNC theory, sentence, sentence category, semantic chunk share. 1 Introduction In Chinese expression, the sentence often omits its components in the unambiguous precondition. As to the specific sentence, it is ellipsis. But the omitted content usually appears in the context. The phenomena can be classified into anaphora, and be called Zero type anaphora. It constitutes the main content of the anaphora study with the pronoun anaphora and noun anaphora. Many linguistic investigations focus on this field [1]. However from the view of the Hierarchical Network of Concepts (HNC in short) theory, the phenomena can be realized as the share of semantic chunk. In order to establish the natural language interactive engine which simulates the intelligence of human being's brain, HNC frames a space for language concepts which stresses the concept association vein, which constructs a digitized symbol system for natural language [2]. Furthermore, HNC defines some key notions, such as sentence, and formalizes them for convenience for computer processing. In general, a sentence is one linguistic unit with relatively complete content. HNC deepens this definition, and puts forward the notion of sentence category, which can express the concrete concept association knowledge for a sentence. The main chunks configuration in one sentence category provides the certain restriction for the semantic integrity by the sentence category knowledge. The expression of HNC semantic chunks in a sentence category [2] is shown as follows: Expression 1. The expression of HNC semantic chunks in a sentence category. SC=GBK1+EK+GBKm(m=2-3) SCR=SC+fKm Note: where, SC is the mathematical expression for the sentence category space, and the SCR is the SC with auxiliary semantic chunks. In the HNC sentence category space, the type and quantity of the main chunk in a sentence depend on the concept association vein of the sentence, and the number of the main chunk is no more than 4; the auxiliary semantic chunks occur in the sentence according to necessity of the sentence meaning, they depend on the concept association vein of the sentence feebly. 175 In the view of …

Content truncated to 500 words