Two shots placed next to each other produce a meaning that neither contains. Lev Kuleshov demonstrated it around 1918, and it is the reason film editing works at all.
Kuleshov took a single neutral close-up of the actor Ivan Mosjoukine and intercut it with three different shots: a bowl of soup, a girl in a coffin, and a woman on a divan.
Audiences reportedly praised the actor's range. They saw hunger, then grief, then desire. The face was identical in all three; the expression they described was produced by the shot beside it.
The original film does not survive, and the accounts come mainly from Kuleshov's own later writing and from Vsevolod Pudovkin, told decades afterward. The details vary between tellings, and the audience response may well have been improved in the retelling.
The historical uncertainty about the original does not affect the phenomenon, which has been reproduced under controlled conditions.
Modern replications using neutral faces from validated expression databases, paired with emotionally loaded context shots, consistently find that viewers attribute emotion to the neutral face consistent with the context. The effect is robust and does not require the viewer to be naive about it.
Neuroimaging work has found that the context shot alters activity in regions involved in emotion attribution when the neutral face is subsequently viewed, which indicates the effect operates in perception rather than as a considered inference after the fact.
Its magnitude varies with the neutrality of the face and the strength of the context, and it is weaker than the strongest versions of Kuleshov's claim, but its existence is not in question.
Kuleshov's workshop in Moscow, teaching during a period when film stock was almost unobtainable, treated editing as the medium's defining operation. If meaning is created between shots, then the shots themselves are raw material and the cut is where the work happens.
Sergei Eisenstein took this furthest. His theory of montage held that meaning arises from collision between shots rather than from their smooth linkage, and he classified types of collision by rhythm, by graphic quality, by tone, and by intellectual content. Battleship Potemkin and October apply it deliberately, sometimes cutting to images from outside the story entirely to make an argument.
Pudovkin took a gentler view, treating editing as construction, building a scene from details the way a sentence is built from words. The disagreement between them is one of the founding arguments of film theory.
The effect is the working assumption of nearly all narrative film, whether or not the term is used.
A reaction shot works because it is placed after something. Directors routinely instruct actors to do less in close-up, because the surrounding shots will supply what the audience reads on the face. Suspense is manufactured by cutting between a character and something they cannot see. Comic timing is often entirely a matter of how long a neutral reaction is held.
The same principle governs documentary, where an interview answer placed beside chosen footage acquires a meaning the speaker did not intend. That is an editorial power with obvious ethical weight, and it is the same mechanism Kuleshov demonstrated with a bowl of soup.
The effect is a demonstration that a face is not read in isolation. Emotion perception, which has often been studied by showing isolated expressions on a screen, turns out to depend heavily on context, and a growing body of psychological work argues that context is not a modifier of expression reading but a constituent of it.
A film experiment from 1918, made because there was no film stock to waste, anticipated that conclusion by most of a century.
