MLD at Evolang 2026

Researchers from the Multimodal Language Department (MLD) presented new work at Evolang 2026, contributing insights into how communication systems emerge and evolve through multimodal interaction.

Lois Dona: Facial Signals in Emerging Communication Systems

Lois Dona presented her work titled “Facial visibility and the emergence of communication systems: Insights from behavioral measures and facial signals.”

Her research explores how facial signals contribute to the development of new communication systems—an area that has received limited attention in experimental studies of language emergence. While early human communication is inherently face-to-face and multimodal, many experimental paradigms tend to overlook the role of facial cues such as eyebrow movements.

To address this gap, the study manipulated facial visibility in a communication game and examined its impact on communicative behavior. The analyses focused on three key aspects: communicative success, convergence between participants, and the role of eyebrow movements.

The findings show that both communicative success and convergence improved over time, regardless of whether faces were visible. However, greater variability in convergence was observed when facial signals were not available. Importantly, eyebrow movements appeared to function as signals of uncertainty and were associated with longer response times. Both of these factors were linked to how participants adapted and refined their linguistic labels during interaction.

Together, these results suggest that participants can flexibly rely on different cues—such as facial signals or timing—depending on the communicative context. The study highlights the importance of integrating facial behavior into models of language emergence and multimodal communication.

 Full abstract: https://evolangconf.github.io/2026/proceedings/paper.html?nr=121

Anita Slonimska: Iconicity in Language Evolution

Anita Slonimska presented a talk titled “Optimized, not lost: The evolutionary trajectory of iconicity as a linguistic property.”

This study builds on recent findings in sign language research suggesting that iconicity might evolve in linguistic systems to support communicative efficiency through flexible and simultaneous structures (Slonimska et al., 2022).

To test this idea, the study examined the iconic strategies used by 23 hearing participants using silent gesture and compared them with those of Italian Sign Language signers when conveying events of varying complexity in a director–matcher task.

The analysis focused on how event complexity influenced iconic strategies used by silent gesturers and how they compared to the iconic strategies used by signers.

Results showed both similarities and differences across both groups, revealing nuanced insights into the evolutionary trajectory of iconicity. 

Taken together, the study shows that rather than eroding or being merely maintained, iconicity is optimized in language evolution, enabling more efficient and flexible encoding of information.

Full abstract: https://evolang.org/2026/proceedings/paper.html?nr=68

About Evolang

The Evolang conference is a leading interdisciplinary venue bringing together researchers from linguistics, cognitive science, anthropology, and artificial intelligence to explore the origins and evolution of language. The conference features cutting-edge work on communication systems, multimodal interaction, and the cognitive and social foundations of language.