Skip to content

Author

Sam G. B. Roberts

1 paper indexed here

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

#reinforcement learning Open access Sep 2026

Inferential learning explains association between self-relevance cues and overlap in chimpanzee visual bodily gesture repertoires

We investigated whether the association of self-relevance cues with overlap in visual bodily gestures is better explained by inferential learning or experience-based reinforcement of gesture–outcome contingencies. Inferential accounts predict that self-relevance cues facilitate convergence by supporting goal inference in complex socio-ecological situations, in secondary contexts and responses. In contrast, experience-based accounts predict that self-relevance cues facilitate repeated exposure, promoting convergence through reinforcement of stable gesture–outcome associations in primary contexts. Using generalized linear mixed models and social network analyses of 12 wild East African chimpanzees, we examined relationships among self-relevance cues, social relationships, socio-ecological conditions, contextual specificity, and gestural repertoire overlap. Overall, the results are more consistent with inferential accounts than reinforcement accounts. Although reciprocated social bonds provided opportunities for reinforcement learning, they did not predict repertoire overlap, indicating exposure is insufficient for convergence. Three findings supported inferential account. First, self-relevance cues accompanied heterogeneous gestures in socially complex and uncertain situations, including interactions with non-kin, larger parties, and weaker bonds. Second, self-relevance cues such as vocalizations, manual gestures, and response waiting were associated with secondary contexts and responses rather than with primary contexts expected under reinforcement learning. Third, repertoire convergence was predicted by gestures in secondary contexts, whereas gestures in primary contexts did not predict convergence. Consistent with inferential account, these findings suggest that self-relevance cues may facilitate inference of communicative goals under uncertainty, potentially supporting flexible alignment of gestural repertoires in dynamic fission–fusion societies, explaining how gestural convergence and divergence coexist, making gesture culture likely. Understanding how communicative repertoires converge among dyad partners is central to explaining the evolution of social complexity. We tested whether overlap in wild chimpanzee visual bodily gestures is better explained by experience-based reinforcement of gesture–outcome associations, or by inferential processes to interpret communicative goals. Analyses of 12 wild chimpanzees showed that self-relevance cues accompanied heterogeneous gestures in complex social conditions, predicted secondary rather than primary communicative contexts, and were linked to greater repertoire overlap only when gestures occurred in secondary contexts. Opportunities for reinforcement learning through reciprocated social bonds did not predict convergence. These findings are more consistent with the interpretation that self-relevance cues facilitate inference of communicative goals than simply increasing exposure to signals and outcomes. Communicative convergence in fission–fusion societies may emerge through flexible goal attribution under uncertainty, providing a mechanism through which socially complex species can maintain divergence and convergence in gestural communication.

Anna Ilona Roberts, Piotr Sorokowski, Sam G. B. Roberts · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.