Drawing on a descriptive and language-oriented approach to argumentation, this paper explores the multimodal dimension of argumentation in talk-in-interaction by considering the various resources used by an opponent to refer to and position themselves in relation to the target of their opposition, namely the adverse position and/or the person who expressed it. More specifically, it studies how speakers exploit multimodal strategies in order to both maintain their discourse at a high level of generality (orientation to context-independency) and guarantee the indexicality of the position taken in the interaction and the disagreement (orientation to context-dependency). The analysis is based on two data collections documenting settings where all participants are temporally and spatially co-present: (i) a video-recorded corpus of Swiss French public debates and (ii) a video-recorded corpus of New Zealand English management meetings. Examining the role of multimodal orchestration of choices in gaze direction, deictic gestures, and speech in establishing different positions in argumentative events such as public debates or management meetings reveals specific contextual features of the activity types, participation frameworks, and sociolinguistic backgrounds involved in an argument.

借助描述性和面向语言的论证方法，本文通过考虑对手用来指称自己并相对于反对者目标定位自己的各种资源，探索了互动对话中论证的多模态维度，即不利地位和/或表达不利立场的人。更具体地说，它研究说话者如何利用多模式策略，以既保持其话语的高度普遍性（针对上下文无关性）又保证在交互和分歧中所持立场的索引性（针对上下文相关性） ）。该分析基于记录设置的两个数据集合，其中所有参与者在时间和空间上都共同存在：（i）瑞士法语公开辩论的录像全集，以及（ii）新西兰英语管理会议的录像全集。研究选择模式在注视方向，指示手势和言语的多模式编排在辩论事件（例如公开辩论或管理会议）中确立不同立场时的作用，揭示了辩论所涉及的活动类型，参与框架和社会语言背景的特定上下文特征。