Light a room with one warm lamp and everything behaves: your hand casts a single soft shadow. Now switch on a second lamp in a saturated color — a red LED strip, a Christmas light — and the wall grows two shadows: one tinted red where only the warm lamp's light reaches, and, a hand's width away, one that is unmistakably greenish-cyan. Nothing in the room is cyan. No lamp is cyan. The second shadow is a color your visual system computed and then attributed to the wall, and it is one of the most direct demonstrations you can set up at home that color is an estimate, not a readout. Daylight runs the same machinery on a grander scale — it is why snow shadows go blue on a clear afternoon — but two lamps on one wall compress the whole meteorology into something you can reach.
The effect is old enough to have a literary history. Goethe documented colored shadows obsessively in 1810 as evidence against a purely mechanical account of color, and the phenomenon has been part of vision science's demonstration repertoire ever since — a standard lab exercise in chromatic contrast, and a puzzle philosophers of perception still cite when they argue that shadows are seen objects.
The arithmetic of two lamps
Start with the physical part, which is simple. With two sources lighting a white wall, any point on the wall receives light from both, mixed. Where your hand blocks the red lamp, the wall behind it is lit only by the warm lamp — that region is genuinely dimmer and warm-tinted, and it casts the reddish shadow. Where your hand blocks the warm lamp, the wall is lit only by the red lamp — that shadow is genuinely red. So far, no surprises: the two shadows are just maps of which lamp reaches where.
The cyan is the surprise. Consider what that second region — the one receiving only warm light — actually is. Warm white is slightly weaker in blue than in red, but it is still broad-spectrum white light. The wall patch is white-lit, slightly warm. Almost no one sees it that way. Seen against a surround dominated by red-plus-warm mixtures, that neutral patch appears greenish, sometimes outright cyan, and the effect strengthens the more saturated your red lamp is.
The shadow that is not really cyan
The standard explanation runs through color constancy — the visual system's habit of estimating the surface color of things while discounting the light illuminating them. It does not compute each patch in isolation; it compares across the scene, tracking how the illumination itself shifts, and assigns each region the surface color that best explains the whole arrangement. Vision researchers describe constancy as a comparative process built precisely on such contextual judgments, which is why a sheet of paper stays white under a sunset and the same neutral patch, ringed by red light, gets assigned a greenish surface. The estimated color is the one that would "explain" the patch if the overall light were red-shifted — so the visual system hands you a green.
This is a close cousin of the afterimage, but not the same animal. An afterimage is fatigue in retinal cells after staring at one color, which is why looking away from a bright image plants its complement in your vision. Colored shadows need no staring, no adaptation time — the cyan appears instantly and holds as long as the lamps do, because it is a judgment, not a burn-in.
Try it tonight
The demo rewards precision. Use a white or very light wall, one ordinary warm lamp for the base light, and one strongly colored source — deep red works best — a meter or two apart. Hold your hand midway and watch the two shadows form; move your hand toward the red lamp and the cyan shadow sharpens and spreads. Vary the red lamp's saturation if you can, and the cyan follows: weaker red, weaker cyan. That coupling is the tell. A physical cyan would hold still while you changed the lamps; an estimated cyan tracks the context, which is the signature of constancy doing the work.
Observers differ in how strongly they report the tint, and the effect softens on textured or non-white walls, since the system pools whatever information the scene offers. Neither caveat removes the lesson: what changed when you moved your hand was not the light's wavelength at any single point but the comparison set your visual system was consulting.
What the two shadows teach
A shadow is subtraction — a hole cut in the illumination by an object. What this demo shows is that the hole does not stay empty: the visual system fills it with a color computed to keep the scene coherent, in this case by pairing red illumination with its complementary counterweight. Opposing hues that neutralize each other in mixture behave as paired descriptors in perception, which is the relationship Complementary Color lets you probe directly — find the hue that cancels a given one, and you are tracing the same axes the shadow estimate used. Goethe's instinct was not entirely wrong, even if his physics was: the cyan on your wall is real as an experience, consistent across observers, and located nowhere in the lamps. Two shadows, two lessons: one is where the light went. The other is where the seeing happens.
The reddish shadow is genuinely lit by the red lamp alone. The "cyan" shadow is lit by the warm lamp alone — the tint is added by comparison in the visual system.