Feature interaction in metaphor aptness: the impact of topic-and-vehicle applicable features and semantic distances.

Abstract

This study utilizes the metaphor aptness rating task (how appropriate a metaphor was in the given context) and the feature arrangement task (participants arrange each of the presented features so that words with similar meanings are close together) to investigate the relationships between the semantic distances of topic (e.g., "an anxiety"), vehicle (e.g., "a dense fog"), features (e.g., "dark," "isn't clear," and "unclear"), and metaphor aptness (e.g., "an anxiety is a dense fog"). The results revealed that the number of topic-and-vehicle applicable features affected the metaphor aptness ratings. Furthermore, the results implied that the closer the semantic distances between features, the higher the metaphor aptness rating. These results suggest that not only the number of topic-and-vehicle applicable features but also the semantic integration and coherence among features play an important role in metaphor aptness. Furthermore, this study demonstrates that the feature arrangement task provides a non-rating-based method for evaluating participants' similarity judgments at the feature level, offering a complementary approach to traditional rating-based measures of metaphor processing.