Stateful Viewers is an art and research project that simulates a visitor moving through a gallery. A vision-language model encounters images one at a time, carrying memory, attention, and affect forward so that each encounter shapes the next.
Rather than treating images as independent inputs, the project approaches viewing as a continuous, cumulative experience.
Stateful Viewers draws on reception theory, phenomenology, and aesthetic psychology. It models a viewer’s perceptual stance prior to viewing, separates expressive voice from image content, and treats emotional response as something that unfolds over time.
The system operationalizes qualitative theories of aesthetic experience within a structured generative framework, without reducing experience to numerical scores or fixed emotion labels.
Where affective computing often asks what emotion is present, Stateful Viewers asks what it is like to experience this image, having already experienced the previous ones.
Before entering the gallery, the system generates a viewer profile and a corresponding reflection style. The profile represents a temporary inner stance rather than a fixed personality.
It includes:
Derived from the profile, the reflection style shapes how experience is expressed, independently of the images themselves.
It influences:
Each reflection:
The internal state is qualitative rather than numerical, expressed through dimensions such as:
All images used in Stateful Viewers are original photographs by Frédéric Bénard (1990-2026), presented both as artworks and as material for exploring machine-mediated perception.