Abstract
An important step in the study of animal vocal communication is to elucidate the species' vocal repertoire: the set of call types produced by that species, each with biologically distinct production mechanisms, functions and meaning. The goal of the present study was to characterize the vocal repertoire of domestic cats, Felis catus, while investigating function and meaning through human perception. In Study 1, we subjected a sample of 212 cat vocalizations to an unsupervised fuzzy clustering algorithm based on 10 acoustic parameters to identify acoustically distinct call categories. In Study 2, 237 human listeners classified 50 cat vocalizations into one of seven call type categories such as ‘chirp’, ‘meow’ and ‘yowl’ and rated their perceptions of the vocalizers' emotional arousal (intensity of emotion) and valence (positivity or negativity of emotion). The clustering solution identified four call categories with distinct acoustic identities but substantial gradation, corresponding roughly to listener classification as tonal meows, chirps, noisy meows and shrieks/trills. Listeners provided significantly different classifications and emotion ratings to sounds based on cluster membership. While listeners classified some calls based on consistent acoustic features (e.g. chirps), others were less consistent (e.g. yowls), suggesting folk call categories do not necessarily reflect distinct acoustic classes or biological call types. Emotion ratings were correlated with call duration and interactions between duration, noisiness and mean fundamental frequency. Together these results demonstrate the significance of acoustic variation between and within call types in human perception of heterospecific vocalizations, supporting a model in which listeners base emotion attribution in part on the call type. The present study illustrates the promise of unsupervised clustering for minimizing bias and promoting understanding of cat vocal communication.