Harsh is large: nonlinear vocal phenomena lower voice pitch and exaggerate body size.

Andrey Anikin,David Reby,Katarzyna Pisanski,Mathilde Massenet

doi:10.1098/rspb.2021.0872

Abstract

A lion's roar, a dog's bark, an angry yell in a pub brawl: what do these vocalizations have in common? They all sound harsh due to nonlinear vocal phenomena (NLP)—deviations from regular voice production, hypothesized to lower perceived voice pitch and thereby exaggerate the apparent body size of the vocalizer. To test this yet uncorroborated hypothesis, we synthesized human nonverbal vocalizations, such as roars, groans and screams, with and without NLP (amplitude modulation, subharmonics and chaos). We then measured their effects on nearly 700 listeners' perceptions of three psychoacoustic (pitch, timbre, roughness) and three ecological (body size, formidability, aggression) characteristics. In an explicit rating task, all NLP lowered perceived voice pitch, increased voice darkness and roughness, and caused vocalizers to sound larger, more formidable and more aggressive. Key results were replicated in an implicit associations test, suggesting that the ‘harsh is large’ bias will arise in ecologically relevant confrontational contexts that involve a rapid, and largely implicit, evaluation of the opponent's size. In sum, nonlinearities in human vocalizations can flexibly communicate both formidability and intention to attack, suggesting they are not a mere byproduct of loud vocalizing, but rather an informative acoustic signal well suited for intimidating potential opponents.

Highlights

Differences in relative body size are of crucial importance for social animals, and displays of size and dominance are among the most common and important messages conveyed by animal signals [1,2]
As described by Morton nearly 50 years ago [13] and confirmed in numerous later studies, a rough or harsh voice quality associated with nonlinear vocal phenomena (NLP) is often found in aggressive calls of animals, and is perceived as aversive in human vocalizations [33,34]
To explain the function of this harsh vocal quality, it has been hypothesized that NLP lower perceived voice pitch and make the vocalizer appear larger [17], with obvious potential benefits for vocalizers

Summary

Introduction

Differences in relative body size are of crucial importance for social animals, and displays of size and dominance are among the most common and important messages conveyed by animal signals [1,2]. To the best of our knowledge, there is no previous empirical work testing whether amplitude modulation and chaos affect perceived voice pitch, nor whether any NLP affect perceived body size To address these important questions, we capitalized on the newly available technology for synthesizing natural-sounding vocalizations with a very precisely controlled voice quality, making it possible to create convincing examples of vocalizations with or without NLP (amplitude modulation, subharmonics and chaos) while keeping constant all other acoustic characteristics such as pitch and formants [41]. While the effect of NLP on perceived voice pitch was our main psychoacoustic outcome of interest, changes in tone brightness, rather than pitch per se, may contribute to the perception of physical largeness, especially in the case of deterministic chaos Roughness is another important psychoacoustic characteristic that increases in the presence of NLP and effectively conveys aversive affective states such as fear [28] and aggression [33]. The results of these three experiments provide convincing evidence that NLP make the voice sound lower, darker and rougher, which exaggerates the speaker’s apparent size and conveys aggression

Methods

Results

Discussion