Network-based diffusion analysis demonstrates that a novel tool-use behavior, “moss-sponging”, spread via social learning in a wild East-African chimpanzee community.

International audienceLanguage's intentional nature has been highlighted as a crucial feature distinguishing it from other communication systems. Specifically, language is often thought to depend on highly structured intentional action and mutual mindreading by a communicator and recipient. Whilst similar abilities in animals can shed light on the evolution of intentionality, they remain challenging to detect unambiguously. We revisit animal intentional communication and suggest that progress in identifying analogous capacities has been complicated by (i) the assumption that intentional (that is, voluntary) production of communicative acts requires mental-state attribution, and (ii) variation in approaches investigating communication across sensory modalities. To move forward, we argue that a framework fusing research across modalities and species is required. We structure intentional communication into a series of requirements, each of which can be operationalised, investigated empirically, and must be met for purposive, intentionally communicative acts to be demonstrated. Our unified approach helps elucidate the distribution of animal intentional communication and subsequently serves to clarify what is meant by attributions of intentional communication in animals and humans

Population and group-specific behavioral differences have been taken as evidence for animal cultures, a notion that remains controversial. Skeptics argue that ecological or genetic factors, rather than social learning, provide a more parsimonious explanation. Work with captive chimpanzees has addressed this criticism by showing that experimentally created traditions can be transmitted through social learning. Recent fieldwork further suggests that ecological and genetic factors are insufficient to explain the behavioral differences seen between communities, but the data are only observational. Here, we present the results of a field experiment that compared the performance of chimpanzees (P. t. schweinfurthii) from two Ugandan communities, Kanyawara and Sonso, on an identical task in the physical domain-extracting honey from holes drilled into horizontal logs. Kanyawara chimpanzees, who occasionally use sticks to acquire honey, spontaneously manufactured sticks to extract the experimentally provided honey. In contrast, Sonso chimpanzees, who possess a considerable leaf technology but no food-related stick use, relied on their fingers, but some also produced leaf sponges to access the honey. Our results indicate that, when genetic and environmental factors are controlled, wild chimpanzees rely on their cultural knowledge to solve a novel task.

The notion of animal culture, defined as socially transmitted community-specific behaviour patterns, remains controversial, notably because the definition relies on surface behaviours without addressing underlying cognitive processes. In contrast, human cultures are the product of socially acquired ideas that shape how individuals interact with their environment. We conducted field experiments with two culturally distinct chimpanzee communities in Uganda, which revealed significant differences in how individuals considered the affording parts of an experimentally provided tool to extract honey from a standardised cavity. Firstly, individuals of the two communities found different functional parts of the tool salient, suggesting that they experienced a cultural bias in their cognition. Secondly, when the alternative function was made more salient, chimpanzees were unable to learn it, suggesting that prior cultural background can interfere with new learning. Culture appears to shape how chimpanzees see the world, suggesting that a cognitive component underlies the observed behavioural patterns.