Panelist: Diane Litman Title: Cooperation in Dialogue Systems from Eliza to Alexa The knowledge-based dialogue research that was pursued largely in the 1970s and 1980s focused on incorporating desirable aspects of human-human cooperative dialogue into human-machine communication. For example, computational models based on Speech Acts and Gricean Maxims were proposed to enhance a system's interpretation and generation capabilities with inference (e.g., by recognizing user indirect speech acts or by anticipating what a generated system response might incorrectly convey). While such research was largely abandoned for more practical approaches, "There is much to be gained from recognizing not just what was said, but why; from identifying conclusions people naturally draw from both what has been said and hasn't" (Workshop CFP). How cooperative are current dialogue systems from this 1980s perspective, and how cooperative should they be?