A Tibetan Perspective on the Common Knowledge System of HowNet
HowNet is a fundamental resource of knowledge bases in China,demonstrating promising results in se-mantic similarity computation and word embedding representation in both Chinese and English.However,current re-search still lacks valuable exploration into the rationality of HowNet's common knowledge system itself and its cross-linguistic adaptability.As an ergative language,Tibetan serves as a good perspective to examine the rationality of the knowledge system of HowNet due to its significant differences from Chinese and English.In this paper,it is revealed that the cross-language adaptability of the HowNet system requires further refinement via the analysis of the semantic denotative properties of Tibetan case markers,controllable-uncontrollable relationship of verbs and se-mantic classification of verbs.The architecture data of the HowNet knowledge system are refined accordingly.It is suggested to apply the prototype theory and the controllable-uncontrollable relationship on Tibetan to enhance the HowNet knowledge system.
万方数据
维普
