you don't need a body to recognize a body · Johansson, 1973
Thirteen dots. No face, no outline, no color, no shape. Just white circles against a dark field. Yet within about 200 milliseconds, most observers perceive a person walking.
Each dot tracks a single joint: shoulders, elbows, wrists, hips, knees, ankles, and head. In isolation, any one dot is indistinguishable from any other oscillating point. Together, arranged in their biologically correct spatial relationships, they trigger the immediate perception of a walking human being.
Press scramble. Each dot is repositioned at a random location on the canvas. But the motion of each dot — its trajectory, its timing, its amplitude — is preserved exactly. The individual kinematics are unchanged. Only the spatial arrangement has been disrupted.
The walker disappears. What remain are thirteen bouncing points with no apparent structure.
This is the critical finding. The information that produces the perception of a walking person is not in any individual dot. It is in the relationships between the dots — specifically, in how those relationships match a spatial template the visual system holds for articulated biological motion. Destroy the configuration, and the person vanishes, even though nothing about any single dot has changed.
The visual system maintains a specialized prior for biological motion. The extrastriate body area (EBA) and superior temporal sulcus (STS) respond selectively to point-light walkers and not to spatially scrambled controls, and this discrimination begins before conscious recognition. The response is not to motion per se — it is to this kind of motion, organized in this spatial pattern.
The invert mode flips the figure upside-down. The motion information is mathematically identical. But recognition slows and degrades. The prior appears to be tuned specifically for upright biological motion, a bias that sits below the level where deliberate effort can fully compensate.
From the same thirteen points, biological motion perception also extracts gender, emotional state (depressed gaits have measurably different kinematics from elated ones), the identity of known individuals, whether the figure is walking toward or away from the observer, and — in studies using motion capture of participants' own gait — whether the walker is oneself. All from points in motion, no body required.
Johansson, G. (1973). Visual perception of biological motion and a model for its analysis. Perception & Psychophysics, 14(2), 201–211.
Cutting, J. E., & Kozlowski, L. T. (1977). Recognizing friends by their walk. Bulletin of the Psychonomic Society, 9(5), 353–356.