*LOGO* Conversational Interaction Studies

In an ongoing series of studies, we would like to investigate features of "natural" and "ideal" conversational interactions involving augmented communicators, with the goal of developing AAC designs that realize these conversational features.



General Motivations

Why are we performing studies on conversational interaction? Two reasons: to identify some of the interactive features that normally take place, and to ensure that our assumptions when designing AAC systems are consistent with these features.
Interactive Features in Natural Conversation
Conversational utterances are rarely produced in isolation. Conversation partners co-construct utterances by asking each other questions, offering suggestions, and making corrections.

  • What AAC design will most effectively facilitate interactive participation in a conversation?
  • Would incorporating some of these interactive features into the AAC system be beneficial to the augmented communicator?
  • Validation of System Design Assumptions
    Inherent in the design of our AAC systems are a number of assumptions. For instance, Compansion currently assumes that the user will input all important content words in the correct order for the sentence.

    We need to verify and validate our assumptions, and investigate the natural patterns of input to the AAC system. If our assumptions are inaccurate, systems may need to be modified to correctly interpret conversational input.

    [NLP] [NLI] [ASEL]
    Lisa Michaud -- michaud@asel.udel.edu
    Last modified: Fri Jan 30 15:47:31 EST 1998