Centring humanity in academic writing: Learning diaries as windows to student experience

Abstract

While recasts are widely recognized as beneficial for the development of second language (L2) (Li, 2010), their effectiveness in synchronous computer mediated communication (SCMC), particularly videoconferencing, remains underexplored. A key affordance of videoconferencing is its capacity to deliver multimodal recasts, operationalized in this study as oral reformulations
accompanied by synchronized on-screen captions (i.e., captioned oral recasts). This format draws on the sensory–modality framework of working memory (Baddeley et al., 2009), which highlights the benefits of distributing input across the auditory and visual channels. When aligned precisely, this dual input can enhance processing efficiency and reduce cognitive load (Mayer, 2014; Sweller, 2005). Empirical studies further support these claims. Reading-while-listening research shows that audio–text combinations foster consistent eye movements and deeper processing (Hui & Godfroid, 2025). Similarly, captioning research shows that textual support can enhance learners’ attention to grammatical forms, aid speech segmentation, and improve comprehension and retention (e.g., Gass et al., 2019; Lee & Révész, 2020; Reynolds et al., 2022). Multimodal recasts also combine the strengths of written and oral feedback. Written input offers visual salience and persistence—a record learners can reread (e.g., Smith, 2004; Yilmaz, 2012)—whereas oral feedback provides
immediacy, prosodic cues, and turn adjacency, facilitating real-time comparison between the learner’s output and the target form (Leeman, 2003; Long, 2007). Each modality, however, has limitations: oral recasts may be misinterpreted as confirmation, and text-based feedback often lacks precise temporal alignment with the utterance it addresses. The hybrid nature of multimodal recasts mitigates these weaknesses, producing feedback that is both perceptually salient and interactionally contingent. Despite these theoretical and empirical foundations, direct comparisons of multimodal versus unimodal recasts in video-based interaction are scarce (Jung & Révész, 2024). Consequently, little is known about how feedback modality shapes learners’ visual attention to the target forms, and how such attentional engagement relates to subsequent grammatical gains during real-time video-based tasks. This study addresses these gaps by investigating whether multimodal recasts promote more focused visual attention and facilitate greater grammatical development than oral-only or text-only recasts, and whether the link between attention and learning varies depending on the feedback modality.