Research on the Sememe-Based Knowledge Representation Integrating Tibetan Features

The sememe, as the smallest semantic unit, is widely used in knowledge bases to represent world knowledge. HowNet is one of the most well-known sememe-based knowledge bases. In theory, the HowNet sememe system is applicable to all human languages and capable of expressing all world knowledge. However, this claim has yet to be verified. This paper uses specific instances to highlight issues with the HowNet sememe system when applied to Tibetan, such as vague definitions, unclear features, and the lack of sufficient sememes to describe various linguistic phenomena. The paper integrates the Tibetan controllable–uncontrollable relationship into the HowNet sememe system, refining its definitions and features, and subsequently modifies the HowNet knowledge dictionary based on this integration.