<p>This article investigates how a still image can constrain argument reconstruction without being reduced to verbal paraphrase or opened to unconstrained interpretation. It examines Théodore Géricault’s <i>Le Radeau de la Méduse</i> (<i>The Raft of the Medusa</i>) as a demanding case for visual argumentation: a historically situated public image with weak explicit argumentative markers. I propose a burden-sensitive procedure linking optical description, depicted configuration, contextual uptake, reconstructable premises, warrants, and critical questions. A reproducible analysis of a digital reproduction makes selected descriptive commitments inspectable. The analysis shows Findings reveal that rescue cues are visually salient, yet remain inherently bound to physical suffering, exposure, and delayed recognition. The painting is therefore reconstructed as supporting a qualified visual argument: rescue is perceptible as a possibility but not secured as narrative or moral closure. The study’s methodological contribution demonstrateshow visual argument analysis can localise criticisable points in the passage from visual description to argumentative reconstruction.</p>