Calendrier

• Gauthier Dubois

dans le cadre DILIS

Réunion d'information afin de présenter le déroulé et organisation de la visite HCERES prévue le 12 Janvier 2026.

• Rosa Illàn Castillo

dans le cadre COVALI

Lien Zoom : https://cnrs.zoom.us/j/94012669665?pwd=aZTbbFeKmHeWEHV8b2bhH4nyLfR6lT.1
ID de réunion : 940 1266 9665
Code secret : P75iPL

Journées de Fin de projet autour du Multilinguisme, Diversité et Contact de langues

Programme

• Dan Dediu (ICREA and Universitat de Barcelona)
• John P. Huelsenbeck (UC Berkeley)

dans le cadre DILIS

This seminar follows an unusual format, in two phases:

15:00 - 16h15: Talk by Dan Dediu (ICREA, Universitat de Barcelona)

16:30-17:30 Meeting with John. P. Huelsenbeck (UC Berkeley)

The workshop will be hosted by our lab's own Matthias Urban, who will share his expertise from being an associate editor for two academic journals. The workshop will focus on the procedure after submission to a journal, from the editor's "behind-the-scenes" perspective, for a better understanding of how this process works. Matthias will present, followed by ample time for questions and discussion

Human vocal behaviour extends beyond speech. We produce gleeful laughter, moans of pleasure, guttural sounds of disgust, and other nonverbal vocalizations that share acoustics with other animals. While spontaneous vocalizations represent ancient emotional communication, humans can control these vocalizations volitionally, expressing them independently from external and internal stimuli. This rare capacity manifests daily in social interactions, from polite laughter to feigned pleasure. This thesis investigates how evolutionary pressures favored vocal control by examining both acoustic forms and listeners’ perceptions.
Following a comprehensive literature review (Chapter I) that examines the evolutionary foundations of vocal control through neural, acoustic, and behavioral perspectives on both production and perception, I report three empirical studies. These studies explore how humans use volitional vocalizations to exploit emotional and social perception in listeners. The first study (Chapter II) demonstrates how speakers voluntarily modulate laughter to encode contexts like amusement, sarcasm, or nervousness through specific acoustics, such as louder, longer laughs for amusement versus a less rough sound for sarcasm. Listeners recognize these contexts above chance, though confusions follow predictable patterns: similar-context laughs cluster together, suggesting laughter operates as graded rather than categorical signals. The following study (Chapter III) explores this gradation through emotional dimensions: valence, arousal, and authenticity. These interrelate highly: more positive laughs are perceived as more aroused and authentic. Nevertheless, valence emerged as a primary information, and similar-valence contexts show the highest confusion rates.
Building on these findings, the third study (Chapter IV) pushed the question of vocal control beyond laughter to test whether humans can inhibit and even reverse honest affective expressions. Using food odors to trigger genuine hedonic reactions, speakers produced honest, attenuated, exaggerated, or opposite vocalizations (pleasure when disgusted, disgust when pleased). Speakers successfully deceived listeners by modulating similar acoustic parameters to those encoding honest expressions. However, opposite vocalizations, especially faking disgust, were judged less authentic, revealing asymmetrical control across affective states.
Across studies, authenticity perception is central to navigating deceptive signals, though acoustic cues vary by vocalization type. Laughter authenticity correlates with pitch variability and breathiness, while hedonic authenticity is associated with harsh onsets for disgust and sustained voicing for pleasure. This specificity suggests specialized detection mechanisms adapted to each expression's physiological constraints. These results shed light on how speakers can deliberately alter their vocal signals along an emotional gradient in human nonverbal communication to influence listeners' perceptions. The capacity for vocal control transforms these signals from honest indicators of internal states into flexible tools for social interaction. At the same time, listeners maintain sensitivity to authenticity by detecting subtle acoustic deviations from spontaneous productions. These findings underscore the evolutionary arms race of deceptive communication: as speakers evolve greater volitional control to deceive, listeners co-evolve enhanced sensitivity to subtle acoustic deviations that betray inauthenticity.
To conclude, understanding these mechanisms of vocal control in nonverbal vocalizations illuminates not only the complexity of human communication through an evolutionary lens, but also opens broader perspectives, from detecting and treating pathological emotional expression in clinical settings, to technical applications as synthesizing naturalistic emotional voices to even understanding emotional performance in the arts.