Visual composition stripped of recognizable objects or narrative anchors — abstracted forms, color fields, and line predominate. Creates alienation effect essential to modernist cinema.
You're sitting in the editing room and suddenly realize: the shot actually shows nothing concrete. No face, no action, no story — just color, texture, perhaps a window frame or the edge of an object deliberately kept out of frame. This is non-objectivity: a visual composition that has removed its narrative anchor points and instead foregrounds abstract qualities — color space, geometry, texture.
In practical filmmaking, this doesn't happen by accident. Antonioni deliberately worked towards it when he showed long takes of empty spaces, industrial landscapes, or just sky and water. The camera lingers, seeking no dramatic focal point. This creates a psychological void that confronts the viewer with themselves — not with an external story. Tarkovsky, in his later works, similarly and consistently utilized the absence of objectivity: long takes of water, forest, diffused light. The focus is on the visual experience, on stillness and tension through what is not seen.
For practical work, this means: lighting becomes the protagonist. You don't need objects with meaning, but very precise control over brightness, color temperature, and shadow. Composition works through lines and planes instead of objects. In editing, such shots act as pauses — they slow down the pace and create emotional space. Many instinctively avoid this because it generates impatience in the audience. That's precisely the point. This visual language only works if the entire montage is designed for it.
The counterpart to non-objectivity is classic narrative visual composition — where every object serves as a plot element. Non-objectivity, on the other hand, is the admission that a film doesn't just have to tell, but can also be spatially and emotionally. Difficult to implement, but when it works, it stays with the viewer longer than any perfectly staged scene with a clear plot.