<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-1117">
  <Title>A Large-Scale Semantic Structure for Chinese Sentences</Title>
  <Section position="3" start_page="0" end_page="0" type="intro">
    <SectionTitle>
2 Theoretical Framework and Case Study
</SectionTitle>
    <Paragraph position="0"> The basic assumption of Frame Semantics (Fillmore 1976;1977; Fillmore &amp; Atkins 1992; Petruck 1996) as it applies to the description of lexical meanings is that each word (in a given meaning) evokes a particular frame and possibly profiles some element or aspect of that frame. By being linked to frames, each word is directly connected with other words in its frame(s). where word dependence association are needed from surface syntactic structures which actually reflect the grammatical relationship to the deep semantics structure whereby semantic content are put into natural language. The meaning of a word, in most cases, is best demonstrated by reference to a semantic network. Referential meaning on its own is insufficient. Word meaning would include the other dimensions concerning the structure and function of words. Unlike English, in which there are two major types of evidence that help to determine the syntactic structure of a phrase or sentence: morphological information and distributional information (such as word order) , in Chinese the lack of conclusive morphological cues makes ambiguity analyses for one sentence more likely. Moreover, most Chinese sentences order are very flexible. Phrase omission, word movement, ellipsis and binding also make it difficult to characterize their grammatical relation. So the semantic information provides important clues for Chinese sentence analyse. We have to rely on semantic knowledge to guide role assignment.</Paragraph>
    <Paragraph position="1"> Thus, we propose a method allowing a syntactic and semantic-based analysis of sequences and relationship of semantic items to obtain the common distribution of the relationship order.</Paragraph>
  </Section>
class="xml-element"></Paper>