Parental feedback shapes vocal flexibility in marmoset monkeys Callithrix jacchus
vocal flexibility, marmoset monkey, social context, development, reinforcement, vocal interaction
One of life's most fascinating spectacles is observing the vocal relationship between babies and their parents. The baby produces cries and moans from birth, and parents learn the function of the different types of sounds of the newborn with weeks of interaction. Months later, the child begins forming its first syllables, and parents respond with adult words, interpreting the context. For example, when a caregiver hears a sound like “ma,” she often repeats the word “mother.” The same pattern occurs as the baby learns to say “father,” the names of relatives, and objects in its environment. This vocal interaction helps the child change vocalization to production in a specific context – vocal flexibility. Although researchers have studied this topic extensively in humans, we still know little about why children of the same age show such stark differences in vocal flexibility. Our closest relatives, nonhuman primates (NHPs), may help us understand how human vocal fluidity develops. Yet scientists still poorly understand how vocal flexibility emerges in NHPs. One reason is that primate infants rarely vocalize with their parents, and many once believed that vocal variability did not change during development. However, a new model of primate vocal learning is emerging: the marmoset (Callithrix jacchus). This New World primate species produces over 11 distinct vocalizations in specific contexts and relies on parental feedback to learn them. Our study tested whether parental feedback could shape vocal flexibility by reinforcing trill calls (normally made with visual contact) and phee calls (normally made without visual contact) in opposite contexts. We studied nine marmosets (four twin pairs + one individual), splitting them into a control group (feedback in natural contexts) and a switch group (feedback in opposite contexts). The experiment used contingent playback—automated vocal reinforcement for trills or phees in contexts with or without visual contact. We recorded daily before, during, and after playback. A neural network (TweetyNet) classified the offspring’s calls, and statistical analyses compared call probability, proportion, and duration. Results: the switch group produced more trills (19.56% vs. 11.71%; p < .001) and fewer phees (42.56% vs. 52.54%; p = 0.012) between P30–P50, confirming feedback’s effect. This work is the first to show that parental feedback guides vocal flexibility in nonhuman primates, shaping calls for specific contexts. Marmosets display plasticity similar to humans and birds, challenging the idea of primates having rigid vocal development. Cooperative breeding systems (like those of marmosets and humans) may hold clues to how complex communication evolved.