In their commentary, Sauter et al. claim that we (Gendron, Roberson, van der Vyver & Barrett, 2014) failed to replicate their findings of universal emotion perception, originally published in the Proceedings of the National Academy (Sauter, Eisner, Ekman, & Scott, 2010), because we (1) included non-universal positive emotion categories in our analysis and (2) did not use rigorous manipulation checks. We show that (1) we fail to find universal emotion perception even for negative emotion categories and (2) the manipulation checks that Sauter et al. now elaborate on in their commentary likely taught Himba participants the Western emotion categories needed to produce the performance they observed. We conclude that free-labeling experiments (such as the one we used in Gendron et al., 2014, Study 1) provide a better test of cross-cultural emotion perception.