A cross-cultural investigation of the vocal correlates of emotion

Abstract

Universal and culture-specific properties of the vocal communication of human emotion are investigated in this balanced study focussing on encoding and decoding of Happy, Sad, Angry, Fearful and Calm by English and Japanese participants (eight female encoders for each culture, and eight female and eight male decoders for each culture). Previous methodologies and findings are compared. This investigation is novel in the design of symmetrical procedures to facilitate cross-cultural comparison of results of decoding tests and acoustic analysis; a simulation/self-induction method was used in which participants from both cultures produced, as far as possible, the same pseudo-utterances. All emotions were distinguished beyond chance irrespective of culture, except for Japanese participants’ decoding of English Fearful, which was decoded at a level borderline with chance. Angry and Sad were well-recognised, both in-group and cross-culturally and Happy was identified well in-group. Confusions between emotions tended to follow dimensional lines of arousal or valence. Acoustic analysis found significant distinctions between all emotions for each culture, except between the two low arousal emotions Sad and Calm. Evidence of ‘In-Group Advantage’ was found for English decoding of Happy, Fearful and Calm and for Japanese decoding of Happy; there is support for previous evidence of East/West cultural differences in display rules. A novel concept is suggested for the finding that Japanese decoders identified Happy, Sad and Angry more reliably from English than from Japanese expressions. Whilst duration, fundamental frequency and intensity all contributed to distinctions between emotions for English, only measures of fundamental frequency were found to significantly distinguish emotions in Japanese. Acoustic cues tended to be less salient in Japanese than in English when compared to expected cues for high and low arousal emotions. In addition, new evidence was found of cross-cultural influence of vowel quality upon emotion recognition.