It is argued in this paper that a multimodal analysis of turn-taking, one of the core areas of conversation analytic research, is needed and has to integrate gaze as one of the most central resources for allocating turns, and that new technologies are available that can provide a solid and reliable empirical foundation for this analysis. On the basis of eye-tracking data of spontaneous conversations, it is shown that gaze is the most ubiquitous next-speaker-selection technique. It can function alone or enhance other techniques. I also discuss the interrelationship between the strength for sequential projection and the choice of next-speaker-selection techniques by a current speaker. The appropriate consideration of gaze leads to a revision of the turn-taking model in that it reduces the domain of self-selection and expands that of the current-speaker-selects-next sub-rule. It also has consequences for the analysis of “simultaneous starts”.

本文认为，需要多模式的转弯分析，这是会话分析研究的核心领域之一，并且必须将注视整合为分配转弯的最主要资源之一，并且有可用的新技术可以为该分析提供坚实而可靠的经验基础。根据自发对话的眼动数据，可以看出凝视是最普遍的旁听者选择技术。它可以单独运行，也可以增强其他技术。我还将讨论顺序投影的强度与当前发言人选择下一个发言人选择技术之间的相互关系。对注视的适当考虑导致对转弯模型的修改，因为它减少了自我选择的范围，并扩展了当前讲话者选择下一个子规则的范围。这对于“同时启动”的分析也有影响。