System of Generating Japanese Sound Symbolic Expressions Using Genetic Algorithm

• Published in: Human-Computer Interaction. Interaction Modalities and Techniques pp. 126 - 134
• Main Authors: Shimizu, Yuichiro, Nakamura, Tetsuaki, Sakamoto, Maki
• Format: Book Chapter
• Language: English; Japanese
• Published: Berlin, Heidelberg Springer Berlin Heidelberg 2013
• Series: Lecture Notes in Computer Science
• Subjects: Auditory Imagery; Category Score; Genetic Algorithm; Sensory Experience; Sound Symbolism
• ISBN: 9783642393297; 3642393292
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-642-39330-3_14

Japanese has a large number of sound symbolic words, onomatopoeia, which associates between sounds and sensory experiences. According to previous studies, a quantification of relationship between phonemes and images enables to predict the images evoked by onomatopoeia and to estimate meanings of onomatopoeia. In this study, we applied the quantification method and developed a system for generating Japanese onomatopoeias using genetic algorithm (GA). Our method uses 90 SD scales for expressing various impressions and genes for genetic algorithm which denote each phonological symbol in Japanese. Through genetic algorithm, the system generates and proposes onomatopoeias appropriate for impressions inputted by users. From the evaluation of our system, impressions of onomatopoeias generated by our method were similar to inputted impressions to generate onomatopoeias.