Mental representation of the locations and identities of multiple hidden agents or objects by a bonobo

Humans are adept at navigating the social world in part because we flexibly map the locations and identities of agents around us. While field studies suggest primates can track individual conspecifics, controlled experiments are needed to determine the complexity of this capacity and isolate the underlying representations. Across five object-choice tasks, we show that our closest relative, a bonobo (Kanzi), can concurrently track the locations and identities of multiple (specifically, two) hidden agents (Experiment 1), that this capacity deploys mental representations rather than tracking agents' last observed locations (Experiment 2), and that these representations can integrate visual or auditory signatures of identity (Experiment 3). Finally, we show that this bonobo performs similarly on an analogous multiple-object tracking invisible displacement task (Experiments 4-5), consistent with multiple agent- and object-tracking potentially recruiting common representational machinery. This work uncovers the rich representations of the social world that are shared by humans and other apes.