Color

"The rays are not coloured." --- Isaac Newton

Salient Points

Natural misperception: Color is an intrinsic property of objects.
Additional misperception: Color is in the reflected light. (Newton: "The rays are not colored." See Hardin article (Word)
Further misperception: Colors map to specific wavelengths

Notes from Evan Thompson

"If we regard the eye as an imperfect detector, then we are committed to the idea that improved detectors would respond with increasing accuracy to the wavelengths and intensities of light. In the limit case, we would have the perfect detector of the spectral composition of light. Such a device, however, would be totally useless for visual perception: illumination varies constantly from moment to moment and place to place. Since the light reflected from objects is a function of illumination, it too varies constantly. If we were sensitive to these variations, our visual experience would have no stability whatsoever, for there would be no basis for generating a set of perceptual color categories which would remain relatively constant through such variations."

"Hence I think that tastes, odors, colors and so on are no more than mere names so far as the object in which we place them is concerned, and that they reside only in the consciousness. Hence if the living creatures were removed, all these qualities would be wiped away and annihilated. But since we have imposed upon them special names, distinct from those of the other and real qualities mentioned previously [shape, size, location], we wish to believe that they really exist as actually different from those." -- Galileo (CV p19)

"Think for a moment what would happen if, on a sunny day, you took an apple out of the bright light of an open field and viewed it in the shade of a low, leafy tree. In the shade, the ambient light would look somewhat green, and yet the apple would continue to look red. Or to take another example, if you placed a white piece of paper in the shade and a black sheet of paper in the sunlight, the black sheet would continue to look black, even though it would now be reflecting considerably more light than the white sheet. In each of these situations, the spectral power distribution of the light reaching the eyes changes dramatically, and yet the colours we preceive are relatively constant. This phenomenon is known as approximate color constancy. (CV p43)

"... the problem is to assign colours to a scene that are insensitive to changes in the ambient light. The ambient light in the scene is the product of the spectral power distribution of the incident illumination and surface spectral reflectance (the percentage of light at each wavelength throughout the spectrum that a given surface reflects). Consequently, as the illumination varies, so too does the light that reaches the eye. But the color of the incident light that a surface reflects, which depends on the physical structure of the surface, does not change. These two variables of illumination and spectral reflectance are confounded to the eye." (CV p43)

"The statement that tetrachromatic color vision systems possess 'ternary' colors cannot be overemphasized. It means that there are colors that are percieved as 'red-green-blue' at the the same time! This is impossible for us to imagine. For the goldfish or turtle we must expect four classes of such colors that are located in the four planes of the tetrahedron: 'red-green-blue', 'red-UV-green', 'green-UV-blue', and 'blue-UV-red'. -- C. Neumeyer (CV p159)

"Although the origin of the human cone visual pigments are ancient, trichromacy is not the norm. Many vertebrate animals have more complex retinas possessing not only four or even five visual pigments, but also oil drop inclusions. Among mammals, however, oil droplets are absent, retinas with three cone visual pigments are found only in primates, and the primitive condition, still exemplified by most modern mammals, is dichromatic. The conclusion that has generally been drawn is that mammalian color vision is evolutionarily degenerate. During much of their evolutionary history, which occurred prior to the extinction of the dinosaurs, mammals were most probably small and had nocturnal habits. As a result, their capacity for photopic vision declined considerably. In this context, humans and Old World monkeys are something of an exception, for they seemed to have secondarily evolved trichromatic vision with the adoption of diurnal habits. In contrast, the evolutionarily more ancient visual systems of fishes, amphibians, reptiles, and birds appear to be largely tetrachromatic." (CV p192)

Additive color mixture: study of which wavelengths can be substituted for each other without changing the percieved color. "For example, at an average light level, a spectral stimulus of 580nm will appear to have a yellow hue. This hue can be matched of combinations of 590nm and 570nm, 550 and 610, 540 and 630, 540 and 670, or many other stimulus combinations." (CV p52)

Principle of Invariance: The signal generated by a receptor has only one dimension of response resulting from its state of excitation, not from the wavelengths of the exciting stimulus. For this reason the neural signals that a single type of photo receptor generates in the presence of light of a given wavelength can always be duplicated by a stimulus of a different wavelength at the appropriate intensity. (CV p55) Stated differently, "The nerve impulse is always exactly the same — 'light struck me' — no matter which specific wavelengths caused the stimulation. The cone output will increase as the light becomes more intense, but the output does not change qualitatively from one wavelength to another." (http://handprint.com/HP/WCL/color1.html)

Different hue boundaries in different animals: "In an experiement to determine whether and how pigeons group spectral stimuli into hue categories, Wright and Cummings (1971) found that pigeons treat wavelengths to either side of 540nm as falling into different hue categories, whereas humans do not. As Jacobs notes in his discussion of this experiement: 'Among other things, this result strongly emphasizes how misleading it may be to use human hue designations to describe color vision in non-human species.'" (http://handprint.com/HP/WCL/color1.html)

Evidence is that pigeons are tetrachromatic or perhaps pentachromatic; goldfish and turtles are tetrachromats.

"Consider, for example, the six primaries red, green, yellow, blue, black and white and the phenomenal structure that color exhibits in virtue of their relations. There is nothing in the lightwaves themselves that accounts for this sixfold structure: the spectrum is rather a physical continuum." (CV p64)

"Spatial antagonism is fundamental in the operation of the visual system and implies that the visual system responds primarily to contrasts rather than absolute magnitudes. In particular, it responds to contrasts across the boundaries of objects rather than the overall light reflected from their surfaces. For this reason, spatial antagonism is thought to play a considerable role in achromatic perception, for whether an area is perceived as black, grey, or white depends on contrasts in the light intensity between it and adjacent areas." (CV p58-59)

"The honeybee and other animals detect objects using three photoreceptor systems but may never see the colors of these objects if the input from these receptors is not mixed to allow for color contrast, the essential feature of color perception in higher animals ... A color receptor must be able to distinguish how much each of three cone mechanisms is activated by the object." - Gouras (CV p55)

"Indeed, most philosophers who defend versions of the received view (for example, McGinn 1983; Peacocke 1983, 1984; Nagel 1986) do not bother to concern themselves with the scientific study of color and color vision....Those contemporary philosophers who do concern themselves with the scientific study of color vision all abandon, in one ] way or another, the recieved view: David R. Hilbert (1987) employs computational color vision to defend objectivism; C.L. Hardin (1988) employs neurophysiology and psychphysics to defend a nondispositionalist version of subjectivism; and Jonathan Westphal (1987) employs colorimetry ... and a Goethean conception of phenomenal color science to establish real definitions of color that are simultaneously physical and phenomenal." (CV p2-3)

"The evolution of color vision is intimately linked to the evolution of color on the surface of the earth. It may go without saying that, in a world without color, animals would have no use for color vision; but it does need saying that in a world without animals that possessed color vision there would be very little color. The variegated colors which characterize the earth's surface (and make the earth perhaps the most colorful planet in the universe) are in the main organic colors, carried by the tissues of plants and animals -- and most of these life-born colors have been designed in the course of evolution to be seen ... the most striking colors of nature, those of flowers and fruits, the plumage of birds, the gaudy fishes of the coral reef, are all 'deliberate' evolutionary creations which have been selected to act as visual signals carrying messages to those who have the eyes to see them. The pigments which impart visible color to the petals of a dandelion or a robin's breast are there for no other purpose." -- Nicholas Humphrey (CV p2-3)

Misc. Notes

Color perception believed to evolve to help detect ripe fruit

Color is our perceptual response to a very narrow span (less than 1%) of the total electromagnetic radiation emitted by the sun, from approximately 400 nanometers to 700nm for humans. The likely extent of animal light sensitivity between 300nm to 800nm.

The figure shows the visible spectrum roughly as it appears in sunlight reflected from a diffraction grating (such as a compact disc), which produces an equal spacing of light wavelengths. Notice the very gradual falloff in luminosity at the near infrared (IR) end of the spectrum, and the relatively sharper falloff toward ultraviolet (UV). In fact, the boundary on the infrared side varies with light intensity: infrared up to 1000nm is visible if the light source is strong enough and viewed in complete darkness.

Antagonistic Colors: Red/Green (There are no reddish-greens or greenish-reds) and Yellow/Blue (CV p46)

Black and white: "Black can be defined as the visual impression experienced in directions from which no visible light reaches the eye. (This makes a contrast with whiteness, the impression of any combination of colors of light that equally stimulates all three types of color-sensitive visual receptors.) Pigments that absorb light rather than reflect it back to the eye "look black". A black pigment can, however, result from a combination of several pigments that collectively absorb all colors. If appropriate proportions of three primary pigments are mixed, the result reflects so little light as to be called "black".

"White and black are also colors (in particular, black is not the absence of color but a distinct color experience), and their perception is closely related to the amount of light or luminance in the visual environment."
(Bruce MacEvoy, http://handprint.com/HP/WCL/color1.html)

"What an organism perceives as "white light" is actually a combination of all wavelengths that that organism is sensitive to. For humans, that basically corresponds to combinations of the three wavelengths of light we are most sensitive to. Goldfish, however, besides having red, green, and blue pigments like humans (or long-, medium, and short-wave sensitive pigments) have an additional pigment that responds in the ultraviolet range. From a previous round of behavioral tests, Neumeyer was able to determine a mixture of primaries that would stimulate appear as white light with a UV component."
(Mike Siuta, http://instruct1.cit.cornell.edu/courses/bionb424/students2004/mas262/behavior.htm)

"Each single wavelength of the spectrum, seen by itself at sufficient brightness in a dark surround, creates the perception of a recognizable hue. Yet light itself has no color, because the apparent color caused by the same light wavelengths can change depending on the context in which the light is viewed. For example, long wavelength or "red" light can, depending on context, appear red, scarlet, crimson, pink, maroon, brown, gray or even black! This color relativity means that color is a sensation in the mind, not an objective feature of the physical world."
(MacEvoy, http://handprint.com/HP/WCL/color1.html) Newton found: "White is not the acme of all color phenomena but a murky mixture, no purer than dust or dirt, that can be produced in many different ways. All these ideas went against the grain of Aristotelian theory and well beyond the dyer's lore that three 'primary' colors defined color mixing."
(MacEvoy, http://handprint.com/HP/WCL/color2.html)

Problems with simplistic ideas of black = lack of color, and white = all colors:

White can be formed by only two (antagonistic) colors
White light does not contain purple (is purple a color?)
Black paper in shade example (see above)
Other colors, e.g. red, can be made to look black in certain contexts

Newton to be the first to realize and explicitly state that color perception cannot be attributed directly to the wavelengths of light: color is in the mind, not in the world.

The only wavelengths of light that can be reflected by a paint mixture are the wavelengths that both paints reflected before they were mixed.

Newton to explore the effects of various color mixtures. He found that three or often just two wavelengths from almost anywhere in the spectrum could mix to make "white" light;

'Newton's parallel experiments with pigments and lights misled many readers into thinking that pigments and lights mix in the same way, which only confused 18th century "color theorists".'
( MacEvoy, http://handprint.com/HP/WCL/color2.html)

Visible Wavelengths
(potentially misleading)

Additive Color Mixing

Subtractive Color Mixing

The most precise research on cone responses to light has been done on a suprisingly small number of subjects, but even these studies show very large variations across individuals in color vision. Although we may want to draw universal "color theory" conclusions from the biological structure of color vision, individual variability warns us that generalizations about color vision -- from the ability to distinguish or describe hues to the emotional associations that colors arouse -- should be made very gently. Look with your own eyes, and not the eyes of color science or "color theory," to discover the emotional significance and esthetic principles for your art.

RYB is a historical set of subtractive primary colors. Nowadays, we know that this set is incorrect, but it continues to be in common use in art. In it, green is a mixture of blue and yellow, yellow is the complement of violet and orange is the complement of blue. Today, scientists know the true set is CMY, which uses cyan as opposed to blue and magenta as opposed to red. (http://www.answers.com/library/Wikipedia-cid-1036189017)

"The most obvious (and common) explanation for this peak sensitivity is simply that vision is tuned to the wavelengths where sunlight is brightest (most abundant) at the surface of the earth.... It does appear that this is part of the explanation for our visual span, though it is not the whole story.... Avoiding harmful or useless light is equally important.... At wavelengths much below 450nm (near UV), light energy is so high that it rather quickly destroys photopigment molecules and, over decades, yellows the clear lens. Many birds and insects have evolved to see UV (ultraviolet) wavelengths, but they have relatively short life spans and die before the UV damage becomes significant.... At the other extreme, wavelengths above 750nm are essentially heat, and heat is optically nonspecific about object attributes: in infrared film, a human face has the same "color" as a hot iron skillet. Worse, the visual system's heat sensitivity would have to be shielded from the heat radiated by the animal's own body and fluctuations in external temperatures. These problems make long wavelength energy more trouble than it is worth."
( MacEvoy, http://handprint.com/HP/WCL/color1.html)

"Humans with dichromatic vision or colorblindness have considerable difficulty distinguishing between reds, yellows and greens, and between blue greens and purples, especially across areas that also vary in lightness."
( MacEvoy, http://handprint.com/HP/WCL/color1.html)

"If at any time I speak of Light and Rays as coloured or endued with Colours, I would be understood to speak not philosophically [scientifically] and properly, but grossly, and according to such Conceptions as vulgar [uneducated] People in seeing all these Experiments would be apt to frame. For the Rays to speak properly are not coloured. In them there is nothing else than a certain Power and Disposition to stir up a Sensation of this or that Colour. ... So Colours in the Object are nothing but a Disposition to reflect this or that sort of Rays more copiously than the rest; in the Rays they are nothing but their Dispositions to propagate this or that motion into the Sensorium; and in the Sensorium they are Sensations of those Motions under the Forms of Colours."
(Isaac Newton, Opticks Book One, Part II, MacEvoy, http://handprint.com/HP/WCL/color1.html)

Newton's Hue Circle
The original "color circle" (1704) superimposed on the spectral and extraspectral hues (there is no "magenta" or "purple" light in the spectrum); applies only to light, not pigments

Newton's Confusions (partial): "I believe these controversies were fed by the following confusions:

Newton endorsed the idea of "primary" colors, because the solar spectrum seems to divide into a handful of dominant spectral hues, despite the continuous gradations possible between their angles of refraction, despite the fact that "primaries" are irrelevant to his color circle explanation of light color mixtures, and despite the fact that no spectral hue was more or less important than any other in these mixtures.
With the help of a sharp eyed assistant, Newton identified seven "primary" colors in the spectrum, in contradiction to the traditional choice of three "primaries" — red, yellow and blue.
Few people recognized the difference between additive and subtractive color mixing — and this included Newton himself, who used three or four colored pigments to make a "white" (gray) powdered mixture.
Lacking an awareness of the difference between lights and paints, the color of paint was assumed identical to the "color" of light reflected from the paint — yellow paints reflected yellow light, green paints reflected green light, etc. (This 18th century falsehood is still taught today.)

( MacEvoy, http://handprint.com/HP/WCL/color6.html#confusions)

Primary Colors: The problem with refuting the idea of primary colors is that it is a moving target, with a vast number of shifting ideas of what the primaries are, how many of them exist (usually between 3 and 7), and whether the primaries are actual colors seen in nature or only imaginary colors. It takes some patience, but MacEvoy has a very fine online site which addresses the problems with primary colors here:
http://handprint.com/HP/WCL/color17.html#primaries And here:
http://handprint.com/HP/WCL/color6.html#imaginary

MacEvoy: "I've explained elsewhere why the concept of "primary" colors is only a useful fiction, and why "primary" colors must be either imaginary or imperfect. A limited number of "primary" colors can be a cost effective method of color production (in video, printing or painting) or a mathematically elegant way to predict additive color mixtures and apparent color matches from an idealized retina. In those situations, as shortcuts or fictions adapted to solve specific problems, "primary" colors are harmless because they are specific solutions to a specific problem. The flaw in artistic "color theory" is that the reasoning goes in the other direction: the various limited, practical solutions that use "primary" colors have led "color theorists" to leap into the imaginary realm of "pure" color, declare that "primary" colors are real, and use "primary" colors to explain every aspect of color experience. The result is that "primary" colors degenerate into the rigid thinking that is inadequate to describe color facts and too limiting to guide artistic intuitions. In fact, subtractive color mixing can never be precise or geometrically simple and, in the abstract, doesn't even exist!"
( MacEvoy, http://handprint.com/HP/WCL/color6.html)

"The perceptual problem confronted by the eye begins with the sheer range of light sensitivity possible with our eyes: a piece of white paper viewed in direct noon sunlight is roughly 6.3 million times brighter than the same paper viewed under starlight! In fact, the eye is capable of making lightness discriminations across a 10 billion fold change in light levels. This vastly exceeds the range possible with any film or photosensitive material ..."
( MacEvoy, http://handprint.com/HP/WCL/color2.html)

Other Animals

"All primates — monkeys, apes and humans — acquired a second set of contrasting receptor cells: the R and G cones. These evolved from a genetic alteration in the mammalian Y cones, which created two receptor pigments that differ only slightly in molecular structure." (I.e. rise of trichromatic vision)
( MacEvoy, http://handprint.com/HP/WCL/color1.html)

Color Realism

David Hilbert acknowledges his view, that objects have color, is the minority view. "We defend the minority view that physical objects (for instance, tomatoes, radishes, and rubies) are colored, and that colors are physical properties, specifically types of reflectance." http://www.bbsonline.org/Preprints/Byrne/Referees/

Questions for people who don't believe the science

If there is one "correct" color and black is the absence of color, why does the black page in the sunlight still look black and a white page in the shade look white when the former is reflecting much more light?
Why are there antagonistic colors? Two wavelengths don't cancel each other out -- there are still two wavelengths hitting the eyes. And if this is "incorrect" the *all* percieved colors except precise primaries must be incorrect (and those incorrect in other animals).
What is the correct color for light beyong the visible spectrum? None? But increased intensity can make light in some (UV) of these areas visible, as can removal of then lens during a cataract operation. Does that imply there is a "correct" color for all wavelengths?
Consider aliens that can read our minds. "Picture a large blue dot," they say; then "Oh, so that is 'green.'" Then you look at an apple and tell them it is red. "Oh," reply the aliens, "to us they look blue." Are the humans correct and the aliens incorrect?
Consider a mutation in a human being so that he sees "red" apples as blue, "green" leaves as red, and "purple" clouds as yellow. Is that person seeing them incorrectly? Now suppose over a million years that person's mutation has overtaken the species. Is blue now the "correct" color for apples? For that matter, before humans evolved their current color scheme, where our ancestors seeing things incorrect colors?
"The statement that tetrachromatic color vision systems possess 'ternary' colors cannot be overemphasized. It means that there are colors that are percieved as 'red-green-blue' at the the same time! This is impossible for us to imagine. For the goldfish or turtle we must expect four classes of such colors that are located in the four planes of the tetrahedron: 'red-green-blue', 'red-UV-green', 'green-UV-blue', and 'blue-UV-red'. -- C. Neumeyer
Check to see that they understand that flowers would not have colors if there were no animals (not that colors would not be perceived, but that they would not have evolved in the first place)

Dispensing with false dichotomies in the defintion of "color" - If the pool table suddenly appears to be a different hue, we now say it looks "red." So regardless of the underlying reality (and whether there is one "correct" color), the way we use typically use "red" is to denote a color experience, not a specific wavelength or some other abstruse definition.

Common themes of confusion:

Confusing continua of nature for discrete entities
Anthropocentrism - particularly assuming the universe reflects human values and/or that how humans experience nature is the sole or correct experience
"Chronocentrism" - bias that we are at an especially correct moment (e.g. in the evolution of eyesight)

Links

Very good article on color, for artists
Testing color perception in animals