The History of Color Science from Newton to Hering

The argument from Newton to Hering is not a parade of disconnected discoveries. Each era answers a different question about the same stimulus: Newton separates wavelengths, Young and Maxwell reduce spectra to three channels, Hering recodes those channels into perceptual oppositions, and the CIE turns the result into a standard observer.

For two thousand years, almost everyone agreed on the wrong answer. From Aristotle in the 4th century BCE to Robert Hooke in 1665, color was understood as a mixture of light and darkness. Black and white were primal; reds, greens, and blues were interpolations between them. White light, in this view, was pure and simple, and color appeared when light was somehow contaminated by shadow.

A few thinkers pushed against this. Ibn al-Haytham (Alhazen), writing in Cairo around 1021, produced the first serious mechanical theory of vision: light travels from objects into the eye, not the other way around. He carefully separated the act of seeing from the property of seeing color, but the question of what color was, physically, stayed tangled in the light-plus-darkness assumption.

Isaac Newton was 23, sent home from Cambridge by an outbreak of plague, when he conducted the experiment that would reorganize optics. In a darkened room at his family home in Woolsthorpe, he let a thin beam of sunlight through a hole in a window shutter and positioned a glass prism in its path. The wall opposite showed not a colored fringe but a full, elongated band of color - from red at one end to violet at the other.

The crucial move came next. Newton placed a second prism, oriented opposite to the first, to see whether it would "uncolor" the light. It did: the spectrum recombined into white. He had shown that white light is composed of all the spectral colors, separable by refraction and reassemblable into white again. This experimentum crucis - "crucial experiment" - appeared in his 1672 paper to the Royal Society and, more fully, in Opticks (1704).

Newton went further: he labelled the spectrum red, orange, yellow, green, blue, indigo, violet - seven hues, chosen partly for symbolic reasons (seven notes of the diatonic scale, seven days of the week). Indigo barely earns a category to most modern eyes, but it survives in the ROYGBIV mnemonic that schoolchildren still recite three centuries later.

In 1810, more than a century after Newton, the German poet and naturalist Johann Wolfgang von Goethe published Zur Farbenlehre (Theory of Colours), explicitly intended to refute Newton. Goethe peered through prisms at edges and white surfaces and observed something Newton's pure-beam setup hid: that color seems to arise at the boundary between light and dark, not from the light alone.

Goethe was, physically speaking, wrong - and modern color science treats Newton's spectral composition as foundational. But Goethe was right about something equally important. The experience of color depends not just on the spectrum entering the eye, but on surround, contrast, expectation, and attention. His careful documentation of physiological colors, afterimages, simultaneous contrast, and colored shadows anticipated nearly a century of perceptual research.

If the spectrum is continuous, why do most color-mixing experiments seem to behave as if there are only a handful of "primaries"? The English polymath Thomas Young - the same Young who decoded part of the Rosetta Stone and ran the double-slit interference experiment - proposed an answer in a 1802 lecture to the Royal Society.

Young reasoned that the retina could not possibly contain a separate receptor for every wavelength. A small number of receptor types, each broadly tuned, would be enough to encode an entire continuous spectrum by the ratio of their responses. He suggested three types, tuned roughly to red, green, and violet. The trichromatic hypothesis was born.

Young's theory remained essentially philosophical for half a century. The Scottish physicist James Clerk Maxwell turned it into something operational. In a series of experiments between 1855 and 1861, Maxwell built a "color box" with adjustable apertures that let observers match a test light by mixing three primaries. His results could be plotted inside a triangle whose corners were the three primaries - the first true chromaticity diagram.

In 1861 Maxwell also produced what is generally regarded as the first color photograph: three separate black-and-white plates of a tartan ribbon, taken through red, green, and blue filters, then projected back through the same filters and superimposed. The composite was imperfect (his "red" filter was largely ultraviolet, and the plates' sensitivities were wrong), but the principle - three independent channels combined additively - was the principle every color image system has used ever since.

Hermann von Helmholtz was a physiologist and physicist whose Handbuch der physiologischen Optik (1856-1867) consolidated the trichromatic theory into a complete physiological framework. He synthesized Young's three-receptor hypothesis with Maxwell's color-matching data and a careful experimental study of the eye itself - the optics of the cornea and lens, the structure of the retina, the limits of acuity, the behavior of accommodation. The combined theory is still called Young-Helmholtz trichromacy.

Crucially, Helmholtz also articulated the modern view of perception as unconscious inference: the visual system, given ambiguous data, builds the most likely interpretation of the world consistent with its priors. That single idea anticipates not only opponent-process and constancy research, but also predictive-coding accounts of vision that dominate computational neuroscience today.

Trichromatic theory had a problem. It accounted for the receptor stage beautifully, but the conscious experience of color did not feel like three independent channels. It felt like three opposing axes: red versus green, blue versus yellow, and bright versus dark. There is no such thing as a reddish-green or a bluish-yellow. These pairs cancel each other in perception.

Ewald Hering, a German physiologist, formalized this in 1872 as the opponent process theory. Hering's model had three opponent channels: red-green, blue-yellow, and black-white. Each channel could be activated in one direction (say, red) or the other (green), but never both at once. The theory explained afterimages (a fatigued opponent channel rebounds opposite), unique hues (the endpoints of the axes), and the perceptual structure of color naming across languages.

For decades the trichromatic and opponent camps argued bitterly. The opponent side pointed at phenomenology; the trichromatic side pointed at receptors. The argument felt zero-sum. In fact, both sides were describing different layers of the same system.

In the 20th century, two-stage models showed that Young-Helmholtz and Hering were not in competition - they described different layers of the same system. The retina has three cone types (Young-Helmholtz). Their signals are immediately recombined by horizontal and bipolar cells into opponent channels (Hering) before any information leaves the eye. Both theories are right; they describe stage one and stage two of the same biological pipeline.

The synthesis is elegant: the brain receives three cone signals, transforms them into L − M (red-green), (L + M) − S (blue-yellow), and L + M + S (luminance). Three channels in, three channels out, but the bases have rotated. This rotation is why color appearance feels like three opposing axes rather than three independent reds, greens, and blues.

While physicists and physiologists argued about receptors and channels, painters, dyers, and educators needed something practical: a way to name and order colors that matched how people perceived them. The early 20th century produced several rival systems.

The most influential was Albert Munsell's, first published in 1905. Munsell organized color along three independent perceptual axes: hue (the position around the color wheel), value (lightness from black to white), and chroma (intensity from neutral to saturated). His Atlas of the Munsell Color System became a standard reference for soil scientists, geologists, forensic examiners, and museum conservators - all of whom still use Munsell notations today.

By the late 1920s, the field had everything it needed to define color numerically: cone physiology, opponent processing, color-matching experiments, and stable perceptual systems like Munsell. The remaining problem was political: a paint company in London and a textile factory in Lyon needed a shared coordinate system to agree on what "the same color" meant.

In 1931, the Commission internationale de l'éclairage (CIE) convened at Cambridge and published two foundational standards: the 1931 2° standard observer - an averaged set of color-matching functions distilled from experiments by Wright (1928-29) and Guild (1931) - and the CIE XYZ tristimulus space, a positive-coordinate transform of Maxwell-style matching designed for practical computation. Every modern color space, ICC profile, ΔE formula, and display gamut is defined relative to these CIE primitives.

The story does not stop at 1931. Two important next steps unfolded over the next half-century. First, CIELAB (1976) gave the field a coordinate system designed for perceptual uniformity - colors that differ by the same amount in Lab feel like they differ by the same amount visually. Second, color appearance models (CIECAM97, CIECAM02, CAM16, CAM16-UCS) added explicit treatments of surround, adaptation, luminance level, and background - bringing Goethe's old observations back into the quantitative fold.

In parallel, Edwin Land's Retinex work in the 1970s showed how color constancy could emerge from simple local computations - inspiring decades of computer-vision algorithms. Specialized standards proliferated: sRGB for screens (1996), Rec.2020 for HDR video (2012), Oklab for modern interactive color picking (2020). Each one is an answer to a problem the previous era had left unsolved.

Scrub the rail. Each dot marks a published moment in the history of color science. Click or tap a year, and the card below explains what changed - and what came next.

Six questions on the historical spine. Wrong answers come with brief explanations and a pointer back to the relevant section.