Mind - Merging

We focus on highly multimodal depictions of multi-party meetings (US Air Force war gaming sessions). The data we code are eclectic—chains of linguistic co-reference, gaze deployments, ‘F-formations’ (a category from Kendon), and parses of turn management, floor control, coalition formation and conflict. The concept of a hyperphrase ties all the threads together; the idea of a growth point is the ultimate theoretical unit to which it is all tethered, with the hyperphrase describing various surface ways that growth points, the cognitive units communication aims to share, materialize during ongoing interactions. The project is relevant to a comment Krauss & Pardo (2004) made regarding a BBS contribution by Pickering & Garrod (2004), namely, that while communication plausibly involves the alignment of speakers and their cognitive states, a move to reduce this to mechanistic priming excludes the reflective processes in dialogue. Our emphasis in this paper is on ‘floor control’ in multiparty discourse. The approach is broadly psycholinguistic, a perspective that includes turn management, turn exchange and coordination; how to recognize the dominant speaker even when he or she is not speaking; and a theory of all this. The data to be examined comprise multimodal depictions of 5-party meetings (US Air Force war gaming sessions). Multiparty discourse can be studied in various ways, e.g., as signals of turn taking intentions, marking the next ‘projected’ turn unit and its content, and still others. We adopt a perspective that emphasizes how speakers coordinate their individual cognitive states as they exchange turns while acknowledging and maintaining the dominant speaker’s status. This goal is similar to Pickering & Garrod’s interactive alignment account of dialogue (2004), but we add gesture, gaze, posture, F-formations (Kendon 1990) and several levels of coreferential chains—all to be explained below. We adopt a theoretical position agreeing with their portrayal of dialogue as ‘alignment’ and of this alignment as automatic, in the sense of not draining mental resources, but not the type of ‘mechanistic’ (priming) account of it they advocate (cf. Krauss & Pardo 2004 for other qualms). The theory we are following is described in the next section. Alignment in this 1 Now at Department of Linguistics, Rice University, Houston, TX 2 Now at Kushiro Public University, Kushiro, Japan. 3 Now at Department of Cognitive Science, Case Western University, Cleveland, OH. 4 Now in the Peace Corps, serving in Jordan. 5 At the Spoken Language Processing Lab, School of Electrical & Computer Engineering, Purdue University, West Lafayette, IN. 6 At Center for Human-Computer Interaction, Virginia Tech, Blacksburg, VA.