Back

Rule-based sequences in sooty mangabey vocal communication

Le Floch, A.; Girard-Buttoz, C.; Souha Azaiez, T.; Bande, N.; Wittig, R. M.; Moran, S.; Zuberbuehler, K.; Crockford, C.

2025-07-10 animal behavior and cognition
10.1101/2025.07.07.663462 bioRxiv
Show abstract

Investigating how non-human animals produce call sequences provides key insights into the evolutionary origins of meaning in vocal communication, including syntax. Many species combine calls into structured sequences, often following specific rules, yet most studies focus on only a few sequences per species. This limits our understanding of their ability to combine calls and convey meaning through sequences. Our study addresses this gap by documenting the vocal sequence repertoire and underlying rules of sooty mangabeys (Cercocebus atys), a West African monkey species. We collected and analysed 1,672 recordings from two wild groups in the Tai National Park, Ivory Coast. Sooty mangabeys frequently combine calls but rely on a limited set of sequences. Within these, we identified rules of call ordering, recurrence and hierarchical structuring. Notably, they produced hierarchically structured sequences using only two call types, a previously unreported system in animal communication. Our findings suggest that sooty mangabeys use both structured and unstructured sequences, likely encoding different types of information. While context of production remains essential for determining meaning, our results highlight how a whole-repertoire approach reveals the full range of rule-based sequences and their potential for meaning expansion beyond the number of individual calls available in the repertoire.

Published in Royal Society Open Science (predicted rank #2) · training set

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.