Abstract
Abstract The COVID-19 pandemic has reshaped academic communication, accelerating the move toward digital and video-mediated formats. This paper reports a multimodal genre analysis of virtual Three Minute Thesis (3MT) presentations, a popular yet underexplored academic speech genre in the COVID era. The study examines 33 virtual 3MT videos from Malaysia’s national finals (2021–2023), delivered by PhD students in Engineering & Technology, Science, and Social Science. Drawing on Swales’s move analysis (1990, 2004) and Sigrid Norris’s Multimodal Interaction Analysis (2011, 2019), the research investigates embodied semiotic resources (i.e., gestures, postures, facial expressions, and slide reference) across rhetorical moves. Seven moves emerged from the corpus, including three obligatory (Topic Generalization, Research Development, Discussion) and four optional (Audience Orientation, Literature Review, Research Introduction, Closure). Overall, presenters favored open posture for establishing rapport and engagement, complemented by beat gestures for rhythmic emphasis, while iconic/deictic gestures aided clarity. Facial expressions were relatively infrequent (3.41%) and primarily co-occurred with greeting and closing formulas to functionally establish rapport, while metaphoric gestures and explicit slide references were also used sparingly. Embodied semiotic activity peaked during Research Development (Move 5), followed by Topic Generalization (Move 2) and Discussion (Move 6), and was more restrained in Audience Orientation (Move 1) and Closure (Move 7). Disciplinary differences emerged in mode preferences, length of mode patterns, and delivery focus. Theoretically, the study reconceptualizes move structures not purely as linguistic constructs, but as multimodal networks where embodied resources carry structural weight; practically, it offers concrete, actionable pedagogical insights for 3MT training and spoken academic discourse instruction.