The announcement that researchers uncovered a ‘new color’ is a welcome break from the year’s tragic news cycle. Of course, having spent over a decade learning and teaching about the psychophysics of color and colorimetry, the BBC headline triggered my skepticism. I skimmed the media regurgitation and went straight to the study itself (Fong, et al., 2025). I encourage you to read the original publication regardless of your feelings about absorbing scientific sources. There is also an excellent video explanation by the researchers that is part of the supplementary materials.
The concept, execution, and results of this experiment are truly remarkable. First, I want to take a moment to appreciate the technological and methodological feat in this study. Second, it is worth pausing to muse over the question of whether this is a ‘novel color’. If Isaac Newton was alive and subjected to this experiment, would he expand ROYGBIV to ROYGOBIV to include the newly minted color ‘olo’?
The Experiment
All cross-discipline research requires many individuals with differing expertise across multiple institutions. The three core authors (J. Fong, H.K. Doyle, and C. Wang) are affiliated with the Department of Electrical Engineering & Computer Sciences, University of California, Berkeley. They worked in collaboration with two colleagues from the same department along with six collaborators from UC, Berkeley’s Herbert Wertheim School of Optometry & Vision Science. The total contributors are rounded out by two authors from the Department of Ophthalmology, University of Washington School of Medicine, Seattle.
The color display ‘Oz’ warrants superlatives because it allows test subjects to view and match colors from micro-doses of laser beams projected onto individual cones. This complex technology allowed the researchers to stimulate only one category of cone - either the long, medium, or short cones (L, M, S cones) with different spectra. (While popularly called red, green, and blue cones, experts would prefer to avoid color labels at the physiological level.)
The difficulty of stimulating a single cone arises from the nearly constant movements of the eye and the fact that a consistent stimulus diminishes to a baseline response due to neural adaptation. Eye movements were (mostly) negated by tracking the participant’s eye movements and utilizing adaptive optics to continually direct the laser light to remain incident on the cone type of choice. Neural adaptation is warded off by exposing the subject to a complex field of color at specific intervals.
The entire purpose of the Oz display is in pursuit of J. Fong’s thesis proposal on ‘How to See Impossible Colors (Fong, 2021). Traditional color studies and matching experiments stimulate multiple photoreceptors across a chosen field of view. This is understandable, the human color experience is broadband and composite. That is, the spectra of light and objects in the natural world encompass a broad range of wavelengths (or electron volts) and the physiological response is complicated by the unequal spectral sensitivity of the cones. The diagram below of cone sensitivity helps illustrate the problem:
The spectral sensitivity of the three cone types in the human eye. Notice the broad overlap between M and L cones compared to the relative isolation of the S cone. Good luck trying to locate wavelengths where only a single cone is active! The parsing of the visible spectrum by the three cones is very different than the choice of color filters used in the photographic arts, which default to equitable divisions of the spectrum.
The methodology and results of this experiment are a way to get around the fact that “spectra information is contained in the ratios, rather than the absolute magnitudes” of cone response (Koenderink, 2010, p. 125). To appreciate the complexity of the retina’s receptors, consider that the M and L cones overlap by 93%. There is a region of the spectrum where all three cones dynamically change sensitivity - in the range from approximately 470 and 520nm (Koenderink, 2010, pp. 126-127). This is precisely why Fong, et al.’s experiment focused on this region of human color response - to uncover the experience that occurs when the response of the M cone in the cyan bandwidth is free from the weight of its two partners.
The technological feat of stimulating a single cone type is commendable enough, but the apparatus also incorporates color matching capabilities. The entire experiment is a demonstration of the harmonious interaction of electrical engineering, physiology, psychophysics, and colorimetry.
In the Light of Standard Colorimetry
The results of this experiment break with the conventional basis of colorimetric experiments. Colorimetry traditionally focuses on stimulating a region of the retina - not individual cones. Moreover, colorimetry focuses on the visual stimulus “when viewed by an observer with normal color vision, under the same observing conditions” (Wyszecki & Stiles, 2000, p. 117). The ‘Novel color…’ experiment produces a visual experience quite unlike anything one would find in a color matching experiment, or in nature for that matter!
At the same time, Fong, et al.’s experimental methods still rely upon the colorimetric paradigm for its analytic basis. (To read further about the colorimetric paradigm I recommend Koenderink’s seminal Color for the Sciences, 2010, pp. 99-101. This is simply the fact that a given beam of light can be matched by the combination of three ‘primaries’ in some proportion. This can be described mathematically as; an unknown beam x is a perceptual match to the combination of three primary colors {RGB} in proportions u, v, w. Or, x = uR+vG+wB. This is called the Trichromatic Generalization by Wyszecki & Styles (2000, pp. 117-119).) Due to olo’s high level of saturation, the addition of white is necessary to bring this color back within the gamut of the three color-matching lights. The elegant mathematical power of colorimetry is still capable of handling ‘out-of-gamut’ colors, or in this case, ‘out-of-common-human-experience’ colors.
For anyone who has participated in color matching experiments, or is conversant with its methodology, the region of cyan (green-blue) is of particular interest because the test color cannot be matched by the RGB primary colors alone. (To experience the ambiguity in this region, check out the website https://ismy.blue/.) Instead, red must be introduced to the test color to facilitate a match. This idiosyncrasy is represented mathematically as a ‘negative red’. While there is no such thing as a ‘negative color’, the colorimetric equations are capable of accounting for this by changing the colorimetric equation from x = -uR+vG+wB to x+uR = vG+wB. This ‘negative red’ can be seen in the graph of the rgb color matching functions (r_bar(λ), g_bar(λ), b_bar(λ)) between the wavelengths of 436 and 546nm (Wyszecki & Stiles, 2000, p. 124). You can view a diagram of the a rgb color matching functions here. The discovery of a ‘novel color’ in this region of cyan is, therefore, not a surprise because this region of color stimulus requires special treatment. In some sense, the saturated blue-green ‘olo’ occurs because there is no ‘red’ signal from the L cone to dampen color quality in the region of cyan.
Another interesting facet of the experiment is how the gamut of human vision is enlarged by Oz stimulating single cones. (See Figure 2, B in ‘Novel color…’) Compare the perceptual shifts in the central blue and green wavelengths to reds and blue/violet. Red and blue/violet regions of human color vision are not extended as much as green/cyan. Clearly, the bandwidth where all three cone sensitivities change together is a prime region to search for a ‘novel’ or ‘impossible’ colors because the crosstalk between all three signals is significant. Notice that towards the extant of the visible spectrum, one cone is doing all the heavy lifting thereby providing a saturated color that accords well with conventional experience.
On a personal note, I am pleased to see LMS color space diagrammed using Maxwell’s triangle. Every semester I make my students reproduce Maxwell’s color mixing wheel experiment and for their homework plot the color matches on triangular graph paper. I always suspect they harbor the opinion that this assignment is born from a character flaw of mine that indulges in cruel colorimetric obscurantism. Actually, this 19th century experiment is the best way to understand the formal basis of colorimetry. For the best explanation of Maxwell’s methods and mathematics see Judd, 1961.
Left) Four color mixtures matched and plotted using J.C. Maxwell’s methods. Since Maxwell’s triangle is a chromaticity diagram (graphing hue and saturation only), the addition of black to the central color can be ignored. However, notice that the addition of white is necessary to desaturate the yellow paper being matched - it is an ‘out-of-gamut’ color. However, the degree of desaturation can be inverted to find its location outside the triangle. The determination of ‘olo’ outside of the normal human gamut follows the same principle. Right) An industrious student attempting to find an achromatic mixture of RGB color papers. Obviously, the rolling shutter of a cellphone does little justice to the superposed mixture of colors on the wheel.
What Constitutes a Novel Color?
The dimension of this study I find compelling regards whether olo is truly a ‘novel color’. Two important caveats before proceeding further. First, entering the terrain of philosophical speculation does not alter in any way the engineering/psychophysiological/colorimetric conclusions in their experiment. Musing over linguistico-philosophical terms does nothing to diminish the originality, technological acumen, and results. Secondly, the researchers have every right to name this color whatever the hell they desire.
Fong, et al. report that:
…color matching confirms that our attempt at stimulating only M cones displays a color that lies beyond the natural human gamut. We name this new color “olo,” with the ideal version of olo defined as pure M activation. Subjects report that olo in our prototype system appears blue-green of unprecedented saturation, when viewed relative to a neutral gray background. Subjects find that they must desaturate olo by adding white light before they can achieve a color match with the closest monochromatic light, which lies on the boundary of the gamut, unequivocal proof that olo lies beyond the gamut.
When I first saw the saturated cyan color sample within the BBC article the word that crashed into my frontal cortex was ‘grue’! This blue-green color is a favorite discussion topic in philosophy classes. However, the epistemic debate surrounding grue is not an argument involving ‘never-before-seen’ colors despite its obvious conflation of the blue and green color terms. Instead, it is a riddle about the nature of induction that was proposed by philosopher Nelson Goodman (Goodman, 1955, pp. 73-80). The liminal boundary between green and blue that Goodman coined has nothing to do with perception but involves a problem in determining the properties of an object (in this case a color - either green or blue) when adding time-dependency to its conditions. Nonetheless, the decision to name the color olo was a missed opportunity to finally give grue a psychophysiological basis!
My copy of D. Stalker’s (ed.) Grue: The New Riddle of Induction. I cannot confirm nor deny the occurrence or non-occurrence of my blurting “Hello! How do you grue?!?” when I saw this on the shelf at the Strand bookstore.
The problem at stake is what constitutes a ‘novel’ or ‘impossible’ color. A brief survey of the literature (between my bookshelf and Bobst library resources) quickly confirmed that the language within this niche of color psychophysics is thorny at best.
J. Fong provides a definition in his graduate thesis with the hypothesis that “‘impossible’ relative activations [between cones] will be perceived as impossible colors (Fong, 2021, p. 6). This is quite logical, because the experience of olo will never occur in conventional experiences. The ‘impossibility’ here is understood as a physiological circumstance - that of having one cone stimulated. However, the use of ‘novel’, as in ‘new’ falls within trickier territory. Professor Ren Ng, interviewed in the BBC article, gives a definition in the form of an analogy. He asks us to imagine that we routinely experience pink until encountering ‘red’ for the first time (Khalil, 2025). This new experience of the same hue, but in a higher saturation, is what constitutes a ‘novel’ color.
I believe that J. Fong’s definition more precisely explains the novelty of olo. At the same time, I am realizing that a codification of what constitutes a ‘novel’, ‘impossible’, ‘forbidden’ and ‘imaginary’ color is paramount in this terrain of color experience. I hope to circle back on this matter in a future post, but for the moment permit me to randomly muse on this problem.
“I cannot learn the colour unless I can see it; but I cannot learn it without language either. I know it because I know the language…I can remember the sensation I had, just as I can remember the color I saw. I feel the same sensation, and that is the same color. But identity - the sameness - comes from the language” (Rhees, 1954, p. 81).
As described by J. Fong, olo is a ‘blue-green of unprecedented saturation’’ (Fong, et al., 2025). The linguistic nature of this description is what piques my interest. His explanation of this color experience relies on two of three color properties - hue and saturation, but not brightness. What is the threshold for describing a new color? Is it novel if there is a substantial change to only one of its three properties? Does it require changes to two or even all three dimensions?
Consider an alternate case where the introduction of a new color term proves useful but does nothing to give rise to a ‘novel color’. Pretend that for years you have been peeling and eating the soft flesh of a reddish-yellow fruit. Depending on the varietal of the fruit and its ripeness you directly refer to some specimens as red and others yellow. One day, a stranger visiting from a foreign land says that you can describe your color experience as orange. Did you experience a never-before-seen color? Does orange have a special existence outside of red and yellow? Or is it quite reasonable to recognize orange as an experience of yellowish-red and reddish-yellow. As philosopher C.L. Hardin recognizes, “…orange is internally related to red and yellow” (2001, p. 292). (Appeals to orange in Newton’s spectrum will not save you in this argument. Newton introduced orange to fit a predetermined schema of seven colors (Shapiro, 1987, p. 190).)
Professor Ng’s analogy of the ‘saturated red’ putting an end to the famine of ‘pink’ reminded me of a seminal study in the evolution of color terms by Brent Berlin and Paul Kay (1969/1999). These two researchers (one in anthropology and the other in linguistics) established a ‘temporal-evolutionary ordering’ to color language that is shown below. Notice that the identification of color experiences begins with spectral colors: red, green, yellow, and blue. What are we to make of ‘brown’, which typically emerges later? Is this a ‘novel color’? Not really, because it can be described through the previous terms - brown is yellow + black or red + black. In other words, brown is a dark red or dark yellow and therefore a modification of preexisting color experiences. While the experience, and the new color term associated with this experience, may be novel to you, it still exists within the boundaries of perception and language. Notice that all the color terms at the far right of their diagram can be explained by the previous colors. If you find yourself in uncharted territory hearing about puce, ecru, chartreuse, or gamboge from a member of the color literati, simply ask them to explain and they will oblige to stoop to perfectly ordinary descriptions.
The temporal-evolutionary ordering of color terms as proposed by Berlin & Kay (1969/1999, p. 4, §1).
If modifications to the properties of saturation and brightness do not constitute a ‘novel color', this leaves only hue. Philosopher C.L. Hardin explains the importance of the hue dimension when explaining the “internalist” color stance of biologist and philosopher Bernard Harrison (Hardin, 1988, p. 144).
If there are to be qualities different from these, qualities that could count as colors, they must resemble red, etc.- the “standard” colors. And if they were to resemble the standard colors, it could only be with respect to hue, brightness (or lightness), and saturation. But color space is closed under these relations; if there were, for example, a novel hue, it would have to fit on the hue loop, and it could not do so without introducing a discontinuity on that loop. There can therefore be no novel hue, and, without a novel hue, there can be no novel colors. (A hueless color is achromatic, and the existence of black, white, and gray is presumably not at issue.) The skeptic “is claiming both that his array is qualitatively radically disparate from the colour array, and that it is at the same time qualitatively continuous with it” (Harrison, 1973, p. 127).
Here is the crux of the dilemma in claiming a ‘novel color’; describing it using the qualities of color experience seems to bring it within the bounds of experience. Or, at least, this is the argument that Rhees and Harrison are making. The description of olo as a “a highly saturated blue-green” seems to describe it perfectly well and evokes both an intellectual understanding through color theory along with my imagination. So what truly constitutes a ‘novel’ color?
Consider the situation if the subjects in this study all claimed, “I saw a new color and I am at a loss to explain it,” the ramifications scientifically and philosophically would be tremendous. This reminded me to look back at the “…study that vision researchers did not talk about - the Crazy Old Aunt in the Attic of Vision” (Billock & Tsou, 2010, p. 73). Billock and Tsou are referring to a 1980s experiment where a bipartitioned field of red and green stripes were stabilized on the retina of a group of subjects (Crane & Piantanida, 1983). All subjects experienced the disappearance of a boundary between the red and green stripes. Within this totality a fraction observed the boundary dissolve into an array of red and green dots and another fraction saw it blend into islands of the two colors. What is more confounding is the remaining fraction experienced a ‘reddish green’. This is a contradiction because the psychological color primaries, arrayed as they are along orthogonal axes of red to green and blue to yellow, forbid the mixture of these opponent pairs. The researchers describe a situation in which ‘some observers indicated that although what they were viewing was a color (that is, the field was not achromatic), they were unable to name or describe the color. One of these observers was an artist with a large color vocabulary” (Crane & Piantanida, 1983, p. 1079). Crane and Piantanida proceeded to test the opponent colors of yellow and blue and induced a ‘yellowish blue’ in their subjects. This experiment was reproduced in the 21st century by Billock, Gleason, and Tsou at the U.S. Air Force Research Laboratory with similar reports (Billock, Gleason, & Tsou, 2001). I have not been able to identify if it has been performed since.
A New Dimension in Understanding Color
In my (non-expert opinion) I believe that Fong, et al.'s experiment has uncovered a ‘novel color experience.’ As I would define it, the experience of the test subjects is of an ‘unprecedented’ and unique modification to one attribute of cyan. I would reserve the term ‘novel color’ for a truly inexplicable experience such as ‘reddish green’ or ‘yellowish blue’. Despite my linguistic equivocations, I would not turn down the opportunity to see olo if I had the chance! And for that matter, neither would Newton. (Although, we would actually need Newton’s assistant whose “Eyes for distinguishing Colours were more critical” than his own (Newton, 1730/1979, p. 126).)
In the fecund landscape of philosophical speculation, the problem of describing unusual color experiences is often treated through hypothetical conversations about color with alien lifeforms. Could they see color experiences beyond the capabilities permitted by our carbon-based building blocks? What would unfold as we stood pointing to objects and teaching each other our color vocabulary? Would there be agreement or disagreement? If so, could they even explain their phenomenological experience? Could an alien see and describe grue?
Ultimately, we don’t need to invoke aliens to explore the psychophysical and philosophical aspects of color. Lifeforms experiencing a radically different color palette live and work amongst us - those with color vision anomalies. This is where I would recommend keeping an eye on the J. Fong. The fact that “Oz can be programmed to probe the plasticity of human color vision” opens numerous avenues of research. The study ends with the tantalizing proposition that Oz can “flexibly probe neural plasticity to boosting color dimensionality in humans, such as attempting to elicit full trichromatic color vision in a red-green colorblind person or eliciting tetrachromacy in a human trichromat” (Fong, et al., 2025). Perhaps it is time to revisit the experiments which produced reddish-green and yellowish-blue. As the name of this apparatus suggests, perhaps color research and philosophy are at the same inflection point in the Wizard of Oz where Dorothy exit sepia-drenched Kansas and enter a Technicolor kingdom.
References:
Berlin, B., & Kay, P. (1999). Basic Color Terms: Their Universality and Evolution. CSLI Publications. (Originally published 1969).
Billock, V. A., & Tsou, B. H. (2010). Seeing Forbidden Colors. Scientific American. 302(2), 72-77.
Billock, V. A., Gleason, G. A., & Tsou, B. H. (2001, October). Perception of Forbidden Colors in Retinally Stabilized Equiluminant Images: An Indication of Soft-wiredd Cortical Color Dependency? Journal of the Optical Society of America A. 18, 84-89.
Crane, H. D., & Piantanida, T. P. (1983, September). On Seeing Reddish Green and Yellowish Blue. Science, 221(9), 1078-1080. DOI:10.1126/science.221.4615.1078
Fong, J. (2021). “How to See Impossible Colors: First Steps Toward the Oz Vision Display.” [Thesis, University of California, Berkeley].
Fong, J., et al. (2025). Novel color via stimulation of individual photoreceptors at population scale. Science Advances, 11(16). DOI:10.1126/sciadv.adu1052
Goodman, H.N. (1955). Fact, Fiction, and Forecast. Harvard University Press.
Hardin, C.L. (1988). Color for Philosophers:: Unweaving the Rainbow. Hackett Publishing Company
Hardin, C.L. (2001). Reinverting the Spectrum. Readings on Color. Volume 1: The Philosophy of Color (A. Byrne & D.R. Hilbert, Eds.). A Bradford Book.
Harrison, B. (1973). Form and Content. Blackwell Publishing Company.
Judd, D. B. (1961). Maxwell and Modern Colorimetry. The Journal of Photographic Science, 9(6), 341-352.
Khalil, H. (2025, April 19). Scientists claim to have discovered “new colour” no one has seen before. BBC News. https://www.bbc.com/news/articles/clyq0n3em41o
Konderink, J.J. (2010). Color for the Sciences. The MIT Press.
Newton, I. (1979). Opticks or A Treatise of the Reflections, Refractions, Inflections & Colours of Light. Dover Publications, Inc. (Originally published 1730)
Rhees, R. (1954). Symposium: Can There Be a Private Language? Proceedings of the Aristotelian Society, Supplementary Volumes, 28, 63–94. http://www.jstor.org/stable/4106594
Shapiro, A. E. (1987). The Spectre of Newton’s “Spectrum.” From ancient omens to statistical mechanics: Essays on the exact sciences presented to Asger Aaboe (Acta historica scientiarum naturalium et medicinalium (J.L. Berggren, Ed.). University Library.
Wyszecki, G., & Stiles, W.S. (2000). Color Science: Concepts and Methods, Quantitative Data and Formulae, 2nd ed. John-Wiley & Sons, Inc.