Discovering social groups via latent structural learning
Abstract: Social groups exist in every society. One crucial process to the functioning of social groups is social categorization—being able to recognize in- and out-group members (i.e., “us” and “them”). Past work suggests that we can do this via dyadic similarity—specific, static, salient features we share directly with others (e.g., race). Yet, in our increasingly diverse societies where explicit cues such as skin tone may not be consistently reliable signals of group membership, how can we identify in- and out-group members? I present studies suggesting that, in contrast to dyadic similarity accounts, a domain-general model of latent structure learning may be a more comprehensive explanation of social categorization. I also examine the neural correlates of both models and discuss more recent work that applies this novel account to larger observational datasets.
About Us:
Closed captioning and/or ASL interpretation are provided for all ICN talks.
Events are free and open to the public unless otherwise noted.