Opponent Process Theory

Opponent process theory holds that colour reaches the brain not as three independent cone signals but as opposing pairs: an achromatic light-dark channel, a red-green channel, and a blue-yellow channel. Ewald Hering proposed it in the nineteenth century against the trichromatic account; both proved right at different stages — three cone types sample the spectrum, then retinal circuitry differences those signals into opponent channels before the cortex.

The theory earns its keep by predicting what cannot be seen. There is no reddish green and no yellowish blue, because one channel cannot take both of its values at once, while reddish blue and yellowish green are unremarkable. Magenta is the clearest case: no wavelength produces it and it is absent from a rainbow — see Dispersion. It is what the red-green channel reports when the long and short cones respond and the middle ones do not, a colour built rather than found.

Afterimages follow directly. Stare at a saturated patch, fatigue one side of a channel, then look at white, and the channel rests at the opposite pole. Pushing this further yields chimerical colours — stygian blue, saturated yet darker than black, and self-luminous red, which appears to glow — by adapting a channel and then presenting an otherwise impossible stimulus. They demonstrate the encoding, not new pigments.

For design the theory sits underneath perceptual colour models rather than competing with them. The axes of OKLCH and its relatives are opponent axes with lightness separated out, which is why moving along one leaves the others perceptually intact and why Color Contrast arithmetic works on lightness alone. It also explains why a palette rebuilt for Dark Mode cannot simply be inverted: adaptation shifts where the channels rest. Human vision encodes no polarization channel at all, unlike many insects — see Polarized Light.