Using Markedness Principle for Abstraction of Dependency Relations of Natural Languages


  • Ju-Ri Kim College of Convergence & Liberal Arts, WonKwang University, City Iksan, JeonBuk, 54538, Korea


Markedness, dependency relations, natural language processing (NLP), syntactic analysis, subcategorization, syntax graph


Background/Objectives: There is no attempt to investigate the relationships between dependency and markedness even though the syntactic roles in language are decided by dependency relations and markers. The main objective of this study was to understand markedness beyond syntactical tables and propose a syntax graph with various syntax structures to verify the relationship between dependence of proposed markers. Methods/Statistical analysis: The methodology involved enquiry into the origin and development of dependency relations, starting from their definition, abstraction, and usage in syntactic structures through graphical presentations. Findings: This study revealed that dependency relations denoted by the markers can be classified into two types of dependencies according to their syntactic functions: implicit and explicit. Eventually, this markedness can be presented through syntax graphs of various syntactic structures to validate the functions of the markers and dependency relations. Improvements/Applications: This paper presents a reasonable method to define dependency relations based on the markedness and the valence theory of the predicate. This approach provides a systematic view to define dependency relations for natural language processing.  The implementation of syntax graphs is a future research project.