Emotional Vocalizations Are Recognized Across Cultures Regardless of the Valence of Distractors

Commentary The question of whether emotional expressions have universal meanings has been at the center of heated debates for many decades (e. and Barrett (2014) took issue with our study of emotional vocalizations, in which we claimed to show that some nonverbal sounds reliably communicate emotional states across cultures (Sauter, Eisner, Ekman, & Scott, 2010). In their Study 2, Gendron et al. replicated part of our study (Himba listeners hearing British vocalizations), but analyzed their data in terms of the valance and arousal of the distractors. They found that recognition was not better than chance for most types of distractors and concluded that emotional vocalizations signal valence, but not specific emotions. We welcome their attempted partial repli-cation of our work. Their article has impelled us to reanalyze our original data in more depth. However, given the results of our reanalysis, together with alternative explanations for the results Gendron et al. obtained, we find that their conclusion does not hold. participants listened to a series of brief emotion stories. On each trial, they heard two vocalizations: one target (consistent with the story) and one distractor. Our original results, analyzed by emotion but not distractor type, showed cross-cultural recognition of six emotions: Participants selected target vocalizations more often than would be expected by chance. However, three of the four positive emotions included in our study were not reliably recognized by Himba participants listening to British sounds. A fair test of whether recognition performance exceeds chance levels with different types of dis-tractors should include only those emotion categories that participants can recognize at better-than-chance levels , because we made no claims that the other types of vocalizations communicate specific emotional states across cultures. The fact that Gendron et al. (2014) included these emotions in their analysis resulted in a considerable reduction in overall accuracy (as shown by the particularly low accuracy they found for positive emotions). Our reanalysis therefore excluded target vocalizations of the culture-specific signals of triumph, pleasure, and relief. In our original study, we varied distractors systematically , according to whether their perceived valence was the same as or different from that of the target (valence ratings taken from Sauter, Eisner, Calder, & Scott, 2010). For each emotion, every participant performed two trials with distractors having the same valence as the target and two trials with distractors having the valence opposite that of the target. 1 We reanalyzed the data from …