Speaker
Description
Vocal accommodation, the process by which individuals make their vocalizations more similar to those of social partners, facilitates communication and interactions between individuals. This convergence can occur at multiple levels, however, most studies have focused on within-call structural changes and whether mammals also converge in call sequence structure remains largely unknown. We tested for both structural and syntactic convergence in the call sequences of common marmosets (Callithrix jacchus) during pair formation. Six adult individuals were recorded before pairing and again four months after random assignment into three breeding pairs. For each session, we extracted phee sequences and classified each call as produced to a partner or to a non-partner. Acoustic similarity between partners was quantified using three approaches: traditional spectral parameters, Mel-frequency cepstral coefficients, and dynamic time warping alignment costs. Syntactic similarity was assessed with four independent measures capturing transition probabilities, bigram frequencies, repeat-length distributions, and local alignment scores. After four months together, pairs showed clear syntactic convergence: distances in sequence structure decreased in non-partner repertoire (LMM, p<0.001). In contrast, spectral distances did not decline, suggesting that marmosets modulate sequence organization more readily than call structure. This study provides the first evidence of syntactic vocal convergence in a primate, revealing that social bonds can shape not only the acoustic form of calls but also their arrangement, highlighting an under-recognized dimension of mammalian vocal flexibility.
