Abstract
Mise-en-scène is traditionally analyzed through semiotic and representational frameworks that decode setting as a system of signs, an approach that struggles to explain the perceptual, pre-symbolic dimensions of aesthetic experience. It remains unclear how the perceptual structure of a filmed environment generates aesthetic experience without symbolic decoding. This article addresses that gap by proposing that Gibson’s ecological psychology, particularly the concept of affordance, can bridge the perceptual structure of mise-en-scène and the aesthetic experience it evokes. It introduces the construct of an aesthetic affordance landscape and suggests that the suspension of practical action in cinematic spectatorship transforms ordinary affordances into aesthetic ones. To illustrate this framework, the paper presents a structured, multi-step prompting methodology for multimodal large language models that guides reasoning from visible surface properties to affordances to aesthetic experience. This methodology is applied to an in-depth case analysis of surfaces in a single frame from Tarkovsky’s Stalker. The analysis illustrates the promise of multimodal LLMs for understanding cinematic experience. The paper contributes a new theoretical construct for film and aesthetic theory and a reproducible, AI-assisted methodology for ecological film analysis that complements existing semiotic, phenomenological, and cognitive approaches.