Cultural Variation in Emotion Perception Is Real: A Response to Sauter, Eisner, Ekman, and Scott (2015)

We recently published results of two experiments (Gendron, Roberson, van der Vyver, & Barrett, 2014a) showing that emotions were not universally perceived in vocalizations. We traveled to remote locations in Namibia and sampled participants from the Himba cultural group. Participants in Study 1 listened to vocalizations and freely labeled them. Participants in Study 2 followed a forcedchoice procedure similar to that reported by Sauter, Eisner, Ekman and Scott (2010); on each trial, they had to choose which of two vocalizations (a target and a foil) matched a brief emotional story (embedded with emotion words). Refining Sauter et al.’s finding that Himba perceivers performed above chance in choosing vocalizations to match story and word cues, we observed that Himba individuals correctly perceived vocalizations according to their positivity or negativity (i.e., valence), but not their presumed universal emotion categories (e.g., anger or fear). Our finding replicated the results of our free-labeling study (Study 1), as well as other data showing that Himba participants perceive valence, but not Western discrete emotion categories, in posed facial expressions (Gendron, Roberson, van der Vyver, & Barrett, 2014b). In their Commentary, Sauter, Eisner, Ekman, and Scott (2015) report that a reanalysis of their data rules out valence as an alternative to their interpretation that their findings support universal emotion perception. They also suggest two reasons why we did not replicate their findings. We appreciate the opportunity to respond and offer three suggestions of our own that question their conclusions and details of the experimental method that they have now elaborated on.