Audio only. No video. There is evidence that humans have adaptations to avoid outgroup members who potentially harbor novel pathogens. However, intergroup contact can produce fitness costs (e.g., violence and disease), or benefits (e.g., trade, mates, and technologies), which suggests that it would be beneficial to possess an adaptation that enables the accurate tracking of group memberships. We predicted that accurate group tracking is accomplished through cognitively differentiating between social groups, and that this differentiation would be calibrated to the potential risk of infection. Consistent with this hypothesis, we found that increases in pathogen disgust are associated with increases in perceived similarity to ingroup-accented speakers and perceived dissimilarity from outgroup-accented speakers, particularly after exposure to pathogenic stimuli. Further, the effect of pathogen disgust on the accuracy of social categorization was mediated by intergroup differentiation. In this talk I present evidence for this group tracking hypothesis for accents, as well as recent evidence that extends the hypothesis to perceptions of similarity to religious groups, and to female judgments of the sexual attractiveness of ingroup- over outgroup-accented male speakers.