The Pragmatics of Indirect Commands in Collaborative Discourse

Matthew Lamm,Mihail Eric
DOI: https://doi.org/10.48550/arXiv.1705.03454
2017-09-05
Abstract:Today's artificial assistants are typically prompted to perform tasks through direct, imperative commands such as \emph{Set a timer} or \emph{Pick up the box}. However, to progress toward more natural exchanges between humans and these assistants, it is important to understand the way non-imperative utterances can indirectly elicit action of an addressee. In this paper, we investigate command types in the setting of a grounded, collaborative game. We focus on a less understood family of utterances for eliciting agent action, locatives like \emph{The chair is in the other room}, and demonstrate how these utterances indirectly command in specific game state contexts. Our work shows that models with domain-specific grounding can effectively realize the pragmatic reasoning that is necessary for more robust natural language interaction.
Computation and Language,Artificial Intelligence
What problem does this paper attempt to address?