Modern Social Science Research

From the perspective of metadiscourse theory and multimodal discourse analysis, this paper examines the composition, functions, and identity construction mechanisms of visual metadiscourse in social media, taking Xiaohongshu as a case study. Characterized by immediacy, interactivity, and multimodality, social media offers a fertile ground for users to employ visual resources in discourse organization, intentional regulation, and identity shaping. Drawing on De Groot et al.’s (2016) classification of visual metadiscourse and adapting it to platform-specific features, this study categorizes visual metadiscourse into image-text and video-text types, further subdivided into layout-oriented and information-oriented subcategories. It systematically analyzes their modal features and discursive realizations. The findings reveal that visual metadiscourse, through strategies such as composition, annotation, attitude markers, self-mention, and engagement markers, not only enhances information delivery and cognitive guidance but also dynamically constructs user identities in interaction, adapting to diverse social contexts. This study provides theoretical insights into multimodal communication and identity practices in the digital age and lays a foundation for future cross-cultural and cross-platform comparative research on visual metadiscourse.

Social media; Visual metadiscourse; Multimodal discourse analysis; Xiaohongshu