A Metaphor Comprehension Method Based on Culture-Related Hierarchical Semantic Model

摘要

Usually a metaphor is encoded with rich cultural connotation, which signifies that culture plays a key factor in truly comprehending a metaphorically-used utterance. Given that, we developed a culture-related hierarchical semantic model to perform metaphor comprehension. Based on the character of a metaphor, to better represent context and background knowledge, we embedded word-level, attribute-level, perception-level, and context-level information into the model. Moreover, in the attribute-level, a culture mapping is developed to better use cultural information. We use a random walk algorithm to search for the most reasonable comprehension results. The model was tested in a nominal Chinese metaphor corpus. The results show the effectiveness of the model and demonstrate its advantages in understanding cultural metaphors.