Towards a computational model of frame of reference alignment in dialogue

Abstract:

We are interested in how participants align the frame of reference of spatial descriptions such as "to the left of the green cup" in dyadic text dialogues in English and Swedish when they perceive a scene from different locations. How do they identify if a misalignment has occurred, and what strategies do they use to get back on track? We show that there is no general preference of FoR in dialogue but the choice is related to the communicative acts of particular dialogue games (a sequence of dialogue moves centred towards a particular goal). There is also evidence that participants align their FoR locally over a sequence of turns, but not globally; at points of misunderstanding it may be prudent to shift FoR to get the conversation back on track. We isolate several conversational games where the dynamics of the FoR assignment appears to be linked to other properties of interaction between the agents, for example whether they are focusing on a particular part of the scene or whether they are identifying individual objects scattered around the scene. It follows that alignment is consistently used as a strategy but there are other factors that trigger changes in FoR.

Research areas:

• Dialogue
• Alignment
• Online text chat

Year:

2017

Type of Publication:

In Proceedings

Book title:

Proceedings of the 7th Joint Action Meeting (JAM)

Address:

London, UK