Uncanny Valley

The uncanny valley is the hypothesis that as a robot, animated character, or synthetic voice becomes more humanlike, an audience's affinity for it rises — until it approaches human closely without fully arriving, at which point affinity drops sharply into unease, recovering only once the likeness is indistinguishable from a real person.

Masahiro Mori proposed the effect in a 1970 essay, plotted as a curve of familiarity against human likeness with a sharp dip positioned just before the graph reaches actual human beings. Mori's paper was speculative rather than experimental: no data underlies the original curve, and he reasoned from intuition about prosthetic hands and corpses, not measurement.

Movement was in the original proposal, not a later addition to it. Mori plots a second curve for moving objects and argues that motion amplifies the peaks and valleys of the graph: a still prosthetic hand is merely odd, and the same hand in motion is unsettling. A static image can look convincingly human at a lower fidelity than a moving one, because motion exposes timing and tracking cues a near-human render tends to get wrong.

Decades of attempts to establish the curve empirically have not settled the matter, which makes Falsifiability worth invoking directly: some experiments find a dip near the predicted place, others find none, and results vary sharply by category — robots, dolls, avatars, and voices don't behave alike. Competing accounts of the mechanism coexist: one ties the unease to categorical ambiguity, a face resisting sorting into either "human" or "not," straining the process Prototype Theory describes; another ties it to mortality salience, a not-quite-alive face resembling a corpse closely enough to trigger the response Terror Management Theory describes elsewhere.

The practical response designers converged on regardless is to route around the valley rather than cross it: stylise deliberately, keep proportions mismatched to human on purpose, or push all the way to photoreal instead of stalling in between. This underlies most character Motion Design for synthetic performers and much of what Game Feel treats as intentional stylisation.