Interspecific and intraspecific responses to synthesized pygmy marmoset vocalizations

Abstract

Four different types of trill vocalizations used by pygmy marmosets are acoustically similar to one another and occur in different contexts. Synthetic replicas of these vocalizations and variants along each of the acoustic continua separating them have been produced and played to both pygmy marmosets and humans. The marmosets responded to a broad range of variations as equivalent to natural vocalizations, but responded quite differently to variants of some calls used in different contexts. This was analogous to the ‘categorical perception’ that humans show with speech sounds. Humans made finer discriminations among all of the synthesized marmoset sounds and showed no perceptual boundaries between them. Thus humans differ from marmosets in their perceptual responses to these sounds. The tolerance of broad variations in vocalizations coupled with sharp boundaries between functionally different sounds is a mechanism that could ensure an accurate reception of a signal by a recipient despite variations due to specific individuals or environmental noise.