: Uncanny Language

The goal of this project is to investigate the influence of perceived expectations and an artificial system's spoken language sophistication on its social interaction with a human participant. Is it advantageous for a robot to have conversational human-like speech with all its discourse particles and disfluencies when its voice quality is mechanical? Or is the key to balance the mechanical only with some human-like qualities? This follows on work on investigating the balance of function and form in designing a socially-capable artificial system. In particular, the project explores whether an artificial system's visual design, actuator capabilities and communication sophistication support Mori's "Uncanny Valley" hypothesis.

Pursuing these questions required the following techniques and expertise

• Designing effective linguistic stimuli
• Designing and conducting psychological experiments using a task-oriented scenario.
• Several engineered system platforms (with speech system, vision and control capabilities)

Human-System Interaction constitutes a multi-disciplinary research arena encompassing work from roboticists, psychologists, sociologists, cognitive scientists, linguists and communication experts. This is indicative of the complexity of social influences. The Uncanny Language project thus involves a multi-disciplinary collaboration with

• Nikolinka Collier (discourse linguist, University College Dublin, Ireland)
• Eva Jacobus (programmer, Media Lab Europe, Dublin, Ireland)
• Gina Joue (cognitive scientist, University College Dublin)
• Frauke Zeller (linguist, Universität Kassel, Germany)