Feedback Relevance Spaces: A Dynamic Syntax account of dialogue data

Abstract:

Dialogue is co-constructed by multiple interlocutors with frequent feedback demonstrating whether utterances are taken as understood. This feedback does not just occur at the ends of turns -- grounding occurs incrementally, before a complete proposition has been processed. Here, we present a corpus study of acknowledgements and clarifications, and describe how the Dynamic Syntax (DS) model of 'feedback relevance spaces' accounts for feedback placement in low-level, semantic processing terms. The model trivially accounts for 85% of cases where feedback occurs at FRSs, but also describes how it can be integrated or interpreted at non-FRSs using the predictive, incremental and interactive nature of DS.

Research areas:

• Dialogue
• Repair
• Dialogue modelling
• Dynamic Syntax
• IncReD

Year:

2019

Type of Publication:

In Proceedings

Book title:

The third Dynamic Syntax conference