Abstract: Conversation analysts treat conversation as social action; this sociological orientation of CA has an implication that it does not exclude non-verbal actions from CA’s remit since its foundation. This article discusses the theoretical and methodological succession from ethnomethodology, through CA to CA-based multimodal interaction research and their interrelationships, focusing on strengths and issues in applying sequential analysis to multimodal interaction research. On the positive side, as a fundamental procedure and method in CA, sequential analysis provides a way of identifying the participants’ treatment and understanding of each other’s actions, a proof procedure intrinsic to the data. It thus enables in-depth, fine-grained investigations into the miscroscopic interactional processes between co-present participants on a moment-by-moment basis. Besides, this article reviews the ways CA-based multimodal interaction analysts deal with their data, using Christian Heath’s transcription scheme as a typical example. Following CA’s strict and clearly-articulated transciption tradition, CA-based multimodal transcription practice inherits every bit of these merits. One problem associated with sequential analysis in multimodal interaction research, especially during data transcription, is that it sometimes poses a challenge to analysts in terms of how to identify clear-cut actions within multimodal sequences, due to the complexities arising from simultaneity in multimodal interactions. Some solutions such as Charles Goodwin’s conceptualization on embodied participation framework or suggestions such as Sigrid Norris’ notions of modal density, etc., have been mentioned as a way out or possible solutions for further research.

Keywords: conversation analysis; ethnomethodology; sequence organization; multimodal interaction

摘要： 会话分析（CA）将会话视为一种社会行为。这种社会学研究取向决定了CA从发轫之始便没有将非语言交际行为排除在研究视野之外。本文讨论了从民俗方法学到CA再到基于CA的多模态交际研究奠基阶段的学理传承和发展，着重探讨序列分析在多模态交际研究中的优势、问题以及未来的发展方向。序列组织是会话组织的核心原则。在基于会话分析的多模态交际研究中，序列分析的优势在于它为研究者理解和处理交际行为提供了一套实用方法和一种基于语料本身的验证程序。存在的问题是：不同于会话组织中相对严格的序列性，多模态交际中除了序列性外还有共时性，这给序列中交际行为起始点的切分造成了一定难度。此外，本文还探讨了CA多模态交际学者处理语料的方法。CA多模态交际转写沿续了CA转写的严格性、精细性和清晰度，其中Christian Heath的转写体例最为典型。最后，我们提出，Charles Goodwin的具身参与框架和Sigrid Norris的模态密度的概念有利于解决序列分析的共时性问题。

关键词: 会话分析, 民俗方法学, 序列组织, 多模态交际