Instruction for students classified as English Learners (EL-classified) in the US has been dominated by formalist orientations to language that focus instruction on acquiring decontextualized linguistic forms often as a prerequisite to mainstream classroom learning (Valdés et al., 2014). Many language scholars argue that these formalist orientations to language and segregating EL-classified learners from mainstream learning is insufficient in promoting meaningful language and content learning and limits these students’ participation in deeper disciplinary learning (Kibler et al., 2021). In contrast, sociocultural and ecologically-informed scholarship proposes more action-based orientations to language which position all learners to co-construct meaning together in scaffolded, dialogic activity across the curriculum (Lantolf & Thorne, 2006; van Lier, 2004). This dissertation project explores the potential of Collaborative Video Analysis (CVA) in Pre-Service Teachers’ (PSTs’) university teacher education classrooms to target PST language noticing and orientations towards these action-based orientations for Multilingual Learners (MLs) in their disciplines. This qualitative study follows five PSTs pursuing History-Social Science secondary teaching credentials and their Course Instructor across six CVA sessions in a US university-based teacher preparation program. In these sessions, participants worked together to narrate, re-narrate, and re-envision videos of PST teaching with a focus on noticing student language use and imagining more action-based language supports. I applied an ethnomethodological approach to collecting and analyzing interactional and individual data including video recordings and participant observations during the CVA sessions and interviews and written reflections from participants. Applying cross-event discourse analysis (Wortham & Reyes, 2015), I traced how interactive moments within and across the six CVA sessions afforded or constrained participants’ orientations to student language use and language scaffolding toward action-based orientations to language. The cross-event analyses suggest that the structured, collaborative discourse around classroom videos contributed to three shifts variably evident in participants’ orientations to student language use and scaffolding toward more action-based orientations to language. Findings demonstrate the methodological utility of cross-event discourse analysis to examine video-embedded teacher learning. More importantly, the findings highlight the potential of CVA to support wider efforts to prepare PSTs to work with EL-classified students and MLs in their disciplines.