Effects of discrepancy between vocal emotion and the emotional meaning of speech on identifying the speaker’s emotions

Abstract

This study investigates the effects of a discrepancy between vocal emotion and emotion contained in the literal meaning of speech on the identification of speaker’s emotions. Except measuring response time, few studies have examined this discrepancy between the emotional meaning of speech and vocal emotion. In the current study, two stimulus conditions were prepared: the “congruent condition,” where the emotional meaning of speech was compatible with the tone of voice and the “incongruent condition,” where the emotional meaning of speech was discrepant to the vocal emotions. Four native Japanese actors (two males and two females) spoke short sentences emotionally. We used a repeated-measures design with the emotional meaning of speech (happiness/sadness) and the tone of voice (neutral/happy/sad) as two factors. Using a 5-point scale, 31 participants were required to identify the speaker’s emotions in a forced-choice task with seven alternatives and report the degree of their confidence for every response. Results indicated that identification of speaker’s emotion was successful even in the incongruent condition and that the degree of confidence differed little except the identification of sad voices. The results suggest that Japanese participants infer the speaker’s intended emotion even if it is incongruent to the emotional meaning of speech.