Bootstrapping Vocal Communication Systems in Virtual Reality iconicity and compositionality in vocalization
Keywords
Loading...
Authors
Issue Date
2024
Language
en
Document type
Journal Title
Journal ISSN
Volume Title
Publisher
Title
ISSN
Volume
Issue
Startpage
Endpage
DOI
Abstract
Understanding the origins of modern human communication is central to language evolution research, and a major focus within the field has been on determining which modality — vocal, visual-gestural, or a combination of both— was primary in the development of communication. Some work suggests an advantage of gestural communication due to its high affordance for iconicity that can aid meaning grounding and learning. However, recent research reveals that human speech and vocalization have more affordance for iconicity than previously thought. At the same time, iconicity can slow down the development of compositional structure, yet cognitive constraints favor compositionality for efficiency. My Research Master’s thesis explores the development of iconicity and compositionality in novel vocal communication systems created by pairs of participants communicating in a VR environment. Over repeated interaction, dyads produced and guessed labels for a set of unfamiliar stimuli, that differed in shape, size, movement type, and speed. Results indicate that participant vocalizations can be iconically motivated through modulations in pitch, loudness, duration, speech rate, and number of syllables to convey stimuli features. Additionally, all pairs developed some level of compositionality, which increased over time. Moreover, pairs' communicative accuracy improved with increasing compositionality. Together, the results suggest that repeated dyadic communication can yield vocal systems that are both iconically motivated and compositional.
Description
Citation
Supervisor
Faculty
Faculteit der Letteren