Building Chinese Discourse Corpus with Connective-driven Dependency Tree Structure

Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 2105–2114, October 25-29, 2014, Doha, Qatar. c⃝2014 Association for Computational Linguistics Building Chinese Discourse Corpus with Connective-driven Dependency Tree Structure Yancui Li1,2 Wenhe Feng2 Jing Sun1 Fang Kong1 Guodong Zhou1 1Natural Language Processing Lab, School of Computer Science and Technology, Soochow University, Suzhou 215006, China 2Henan Institute of Science and Technology, Xinxiang 453003, China {yancuili, wenhefeng}@gmail.com {20104027009, kongfang, gdzhou}@suda.edu.cn Abstract In this paper, we propose a Connective- driven Dependency Tree (CDT) scheme to represent the discourse rhetorical structure in Chinese language, with ele- mentary discourse units as leaf nodes and connectives as non-leaf nodes, large- ly motivated by the Penn Discourse Treebank and the Rhetorical Structure Theory. In particular, connectives are employed to directly represent the hier- archy of the tree structure and the rhetor- ical relation of a discourse, while the nu- clei of discourse units are globally de- termined with reference to the depend- ency theory. Guided by the CDT scheme, we manually annotate a Chinese Dis- course Treebank (CDTB) of 500 docu- ments. Preliminary evaluation justifies the appropriateness of the CDT scheme to Chinese discourse analysis and the usefulness of our manually annotated CDTB corpus. 1 Introduction It is well-known that interpretation of a text re- quires understanding of its rhetorical relation hierarchy since discourse units rarely exist in isolation. Such discourse structure is fundamen- tal to many text-based applications, such as summarization (Marcu, 2000) and question- answering (Verberne et al., 2007). Due to the wide and potential use of discourse structure, constructing discourse resources has been at- tracting more and more attention in recent years. In comparison with English, there are much fewer discourse resources for Chinese which largely restricts the researches in Chinese dis- course analysis. The general notion of discourse structure mainly consists of discourse unit, connective, structure, relation and nuclearity. However, pre- vious studies on discourse failed to fully express these kinds of information. For example, the Rhetorical Structure Theory (RST) (Mann and Thompson, 1988) represents a discourse as a tree with phrases or clauses as elementary dis- course units (EDUs). However, RST ignores the importance of connectives to a great extent. Fig- ure 1 gives an example tree structure with four EDUs (e1-e4). In comparison, Penn Discourse Treebank (PDTB) (Prasad et al., 2008) adopts the predicate-argument view of discourse rela- tion, with discourse connective as predicate and two text spans as its arguments. Example (1) shows an explicit reason relation signaled by the discourse connective “particularly if” and an implicit result relation represented by the insert- ed discourse connective “so”, with Arg1 in ital- ics and Arg2 in bold. However, as a connective and its arguments are determined in a local con- textual window, it is normally difficult to deduce a complete discourse structure from such a con- nective-argument scheme. In this sense, the PDTB at best only provides a partial solution to the discourse structure. [Catching up with commercial competitors in retail banking and financial services,] e1 …

Content truncated to 500 words