Multimodal Orchestration in TikTok Vocabulary Videos: A Qualitative Analysis of Vocabulary Meaning Making

Abstract

The use of TikTok for short-form vocabulary instruction has become increasingly prevalent; however, empirical work examining the instructional design of these videos, particularly how multimodal elements are combined to construct vocabulary meaning, remains limited. This study uses qualitative multimodal discourse analysis to investigate how vocabulary meaning is constructed across ten English vocabulary videos (five institutional and five creator-driven). A theoretically informed coding frame was developed to analyse five semiotic modes (linguistic, visual, auditory, embodied, and spatial) through timestamped co-occurrence analysis in ATLAS.ti. The findings suggest that the meaning of vocabulary in these videos is represented through the coordinated orchestration of multiple modes rather than through isolated modal inputs. Patterns of co-occurrence point to recurring configurations consistent with signalling, redundancy, and attentional guidance. Temporal density, defined as the concentration of multimodal cues within brief instructional intervals, emerged as a salient design feature. Creator-driven videos exhibited higher multimodal density and performance convergence than institutional videos. This study contributes to multimodality research by foregrounding temporal orchestration as an important analytic dimension in short-form instructional videos.