Projecting sustainability: A corpus-informed multimodal analysis of transparency in airline YouTube videos

In the aviation sector, CSR transparency is increasingly framed as a resource for building stakeholder trust, yet visibility does not necessarily entail verifiability or responsibility (Schnackenberg & Tomlinson, 2016). In digital contexts, transparency operates along a continuum: claims may look clear and accountable while remaining only partially accessible or weakly evidenced. While corpus-based research has examined disclosure and opacity in textual genres and caption-centred social media, far less attention has been paid to how transparency is performed in audiovisual institutional communication. This paper addresses this gap by analysing environmental commitments in airline YouTube videos and examining how credibility is negotiated through language, image, voice, editing and evidential cues. The dataset consists of climate-related videos (2023–2025) from four airlines (ITA Airways, British Airways, Delta Air Lines and China Southern Airlines). Methodologically, the study adopts a corpus-informed multimodal discourse approach informed by Systemic Functional Linguistics and social semiotics (Kress & van Leeuwen, 2006). Videos are segmented into analytically coherent sequences and coded for the co-articulation of verbal and audiovisual resources, focusing on (i) thematic focus (e.g., values, projects, targets), (ii) agency and responsibility framing, (iii) future-projection cues (targets, timelines, modal patterning), and (iv) evidential anchoring and clarity work realised through voiceover, on-screen text, imagery, editing, numerical overlays, and standards or certification markers. Coding is supported by light quantification (frequency counts and cue distribution across sequences) to compare how evidential cues are staged across the video’s sequence structure. Semiotic processes of iconization, recursivity and erasure (Irvine & Gal, 2000) are used to interpret how cross-modal repetition, symbolic condensation and strategic backgrounding reinforce, qualify or constrain evidential support. Findings from the coded patterns are synthesised through a previously developed Soft–Semi-Hard–Hard continuum (Notari, under review), grounded in signalling-based distinctions of evidential cost (Spence, 1973). This mapping captures sequence-level CSR transparency as symbolic projection, partially checkable support, or externally validated evidence. In conclusion, the continuum provides an operational bridge between textual approaches to disclosure and multimodal analyses of audiovisual transparency, offering a diagnostic basis for identifying where evidential support is strengthened through multimodal staging and where transparency remains primarily performative.