Fellbaum, ChristianeRosenberg, Gillian2025-08-112025-08-112025-04-18https://theses-dissertations.princeton.edu/handle/88435/dsp01ht24wn89pHumans have a preference for minimizing effort in many physical and cognitive tasks. A number of studies have argued that the preference for reduced effort also manifests in language-related tasks. Minimized effort in communication is one part of the larger concept of communicative efficiency, a phenomenon describing linguistic pressure for least effort possible relative to highest communicative success possible. Iterated learning experiments — where individuals sequentially learn and transmit signals — have shown that communicative efficiency can also emerge in artificial communication systems. However, such experiments typically use homogeneous groups like college students. In reality, language exchange involves diverse individuals, each bringing unique cognitive, physical, and social biases. In this paper, we investigate whether a preference for reduced effort (as one part of communicative efficiency) manifests in non-verbal communication systems with participants of greater demographic variation. We use anonymized data from an iterated learning experiment conducted by the FLESH team at Leibniz Institute for General Linguistics. In their experiment, 60 individuals (ages 9-80) learned and taught five gestures representing concepts within three separate chains (i.e., 20 individuals per chain). Using computer-vision-based motion tracking, we extracted features measuring effort — gesture space, temporal variability, and sample entropy —for each gesture as it progressed through the chains. Mirroring previous findings, we hypothesized that gestures at the end of the chains will be less effortful (e.g., reduced gesture space). We also anticipated that participant demographics (i.e., age and extroversion) affect gesture evolution. Our results show no reliable relationship between generation number and the effort measures or between the demographics (age, extroversion, and gender) and the effort measures. However, we did find slight differences in effort measures depending on the gesture concept, indicating that the signal being transmitted may affect the way it is transmitted. These findings highlight the complex, nonlinear nature of transmission in human communication and remind us that individuals are not perfect learners and teachers, nor passive transmitters of information. Individuals inject their own cognitive, physical, perceptual, or expressive tendencies which influence the way they engage in communication. This knowledge is relevant for any field that seeks to understand how individuals interact with one another and how they disseminate information from one to another.en-USPrinceton University Senior Theses