The Meaning of Color: A Documented History of Pigment, and an Unsettled Science of Naming

Color is a physical phenomenon and a biological one before it is a cultural one. Ordinary human color vision rests on three classes of cone photoreceptor, conventionally labeled S, M and L for their short, medium and long wavelength sensitivities, and the genes encoding the three cone pigments were identified and sequenced in 1986 by Jeremy Nathans, Darcy Thomas and David Hogness. One consequence of that architecture matters more than any other here, and it has a name: metamerism. Two lights or surfaces with completely different spectral compositions can produce identical color percepts, because the eye reduces the whole incoming spectrum to three cone responses before anything else happens. Color is not a property of light that the eye reads off. It is a three-number summary the eye computes and the brain interprets.

Two more mechanisms follow from that, and both are ordinary enough to check without equipment. Rod cells give monochromatic vision in low light and cones need brighter conditions, which is why color drains out of the world at dusk. And the brain compensates for the illumination it is working under: a sheet of paper is judged white under a tungsten bulb and under noon sunlight, although the light physically reaching the eye from it is very different in the two cases. That compensation is called color constancy. It is the reason the world looks stable, and it is doing far more work than it usually gets credit for.

The clearest modern demonstration that color is a computation rather than a reading is the photograph of a dress that went viral in February 2015. Rosa Lafer-Sousa, Katherine Hermann and Bevil Conway surveyed 1,401 respondents and found the population split three ways: most saw white and gold, many saw blue and black, and a smaller group saw blue and brown, with individual reports stable over time. Their account is that the image is ambiguous about its own illumination, so observers unconsciously discount either a bluish or a yellowish light source and arrive at different answers from identical pixels. It is color constancy caught in the act of disagreeing with itself across observers. The photograph is copyrighted and is not reproduced here, which costs the reader nothing, because the finding is about what happens inside the observer rather than about anything visible in the file.

Variation between people goes further than that, although not nearly as far as the popular version has it. Tetrachromacy, a fourth functional cone class, is common in birds and reptiles and rare in humans. Women carrying an anomalous long-wavelength opsin gene may express four cone types, but expressing four is not the same as using four. Jordan, Deeb, Bosten and Mollon tested twenty-four such carriers in 2010 and found exactly one who behaved as a functional tetrachromat on a demanding color-matching task. The capacity is real. Its behavioral expression is rare even among the people who carry the genetic substrate for it, and an article that reports the substrate without the behavioral result has reported half a finding.

At Blombos Cave on the southern Cape coast of South Africa, excavators recovered two abalone shells used as containers for a deliberately compounded ochre-based mixture, together with grindstones, hammerstones and bone used to work it, in a layer dated to about 100,000 years ago. Christopher Henshilwood, Francesco d'Errico and their colleagues published the find in Science in 2011. What it is, exactly, is the oldest known pigment-making workshop: not a lump of ochre but a recipe, a toolkit and a container, which together imply planning, storage and transmitted know-how. Our own file U_2_01 calls Blombos the earliest pigment use. That is a different and much larger claim, the paper does not make it, and ochre use is asserted at considerably older African sites on a weaker evidential footing. The superlative printed here is the one the excavation supports and no more.

For most of recorded history the only serious source of lapis lazuli was one district: Sar-e-Sang and its neighbors in Badakhshan, in the Hindu Kush of what is now northeastern Afghanistan. From there the stone traveled into Mesopotamia and Egypt in one direction and the Indus valley in the other. The blue in Tutankhamun's inlays and in the Standard of Ur came down the same road. Nothing about that is mysterious and all of it is expensive: a single geological accident in one mountain range was the source, for thousands of years, and every gram of it had to be carried out of those mountains. One honest note on this paragraph specifically. It rests on our own file U_2_01, and the research behind this article resolved no external source for the Badakhshan trade route itself, which is why it is named again in the source notes at the foot of this page. The material claims that follow, about ultramarine as a pigment and where it has been identified, do not depend on it.

Egyptian blue also does something its makers had no way of detecting. Excited by visible light it emits strongly in the near infrared, with an exceptionally high quantum yield and a long emission lifetime, a property characterized by Accorsi and colleagues in 2009. Conservators exploit it. A photograph taken through an infrared filter makes traces of Egyptian blue glow bright against everything around them, so painted decoration invisible to the eye after three thousand years of weathering can be recovered from stone that looks bare, including on Greek and Roman sculpture long assumed to have been left white. A pigment chosen for how it looked turns out to have left a signature in a band nobody could see.

Tyrian purple was priced above gold by weight in the Roman world, and the direction of that comparison is defensible. The Edict of Diocletian, the empire's schedule of maximum prices issued in 301 CE, prices purple-dyed silk by the pound at a level far above the same weight of gold, and Roman law progressively restricted the deepest purples to the emperor. What this page will not print is the multiple. The research behind this article could not resolve it: figures for gold in the same Edict differ between reconstructions, so the resulting ratio circulates as anything from roughly twice to more than twenty times, and popular sources routinely substitute purple-dyed silk, a finished luxury textile, for raw dye, which is a different commodity being priced. Direction is a claim the evidence supports. A number is not, so there is no number here.

Every version of the thousands-of-snails-per-gram statement in circulation descends from a single experiment, and it is worth naming rather than repeating as folklore. In 1909 the chemist Paul Friedlaender processed a large quantity of Murex brandaris in order to isolate and identify the dye, and reported a yield in the region of 1.4 grams of pure dye from about 12,000 snails, which is on the order of eight or nine thousand animals per gram. That is the paper that established the compound as a dibromoindigo. Two things about how it reaches you here. The figure belongs to Friedlaender's experiment and to the material and methods he used, not to murex dyeing in general. And the research behind this article did not read his original German text, so the numbers themselves are carried at second hand and are given attributed rather than flat.

The emperors-only version of purple is also too simple. In 2021 Naama Sukenik and colleagues reported wool textile fragments dyed with true molluscan purple from Slaves' Hill in the Timna Valley, Israel, radiocarbon-dated to around 1000 BCE, with chemical analysis identifying the molluscan dye compounds directly in the fibers. They are the earliest known examples of textile actually dyed with royal purple found in the southern Levant, and they come not from a palace but from a copper-smelting camp in the desert. Whoever wore purple there, it was not a Roman emperor.

The most technically remarkable pigment in this section is not from the Mediterranean at all. Maya blue is neither a mineral nor a dye but a hybrid of the two: indigo from the anil plant bonded into the channels of palygorskite, a fibrous clay. The bond is what makes it extraordinary. Ordinary organic blues fade, and the Maya lowlands are hot and wet, yet Maya blue survives on murals, pottery and figurines after more than a thousand years and resists acids, alkalis and solvents that destroy most pigments. Electron microscopy reported in Science in 1996 characterized it as an ancient nanostructured material. In 2008 the excavation of a bowl from the Sacred Cenote at Chichen Itza yielded the first direct evidence of how it was made, with indigo, palygorskite and copal incense burned together as part of a ritual, which means the manufacturing process and the ceremony were the same event.

Natural ultramarine was the most expensive pigment in the European painter's cupboard, and Renaissance patrons wrote it into contracts by grade and by price per ounce, because it was not a thing that could be left to the painter's discretion. That much is safe to say. What this page will not do is put a ratio on it, for the same reason it printed no ratio for purple: the claim more expensive than gold circulates for ultramarine in as many incompatible forms as it does for Tyrian purple, and the research behind this article found no single reliable figure. One indirect anchor is real and is worth having. When the Societe d'Encouragement pour l'Industrie Nationale offered a prize in 1824 for a synthetic substitute, it set as a condition of the award that the product cost no more than three hundred francs per kilogram, a ceiling that only makes sense against a natural article costing very much more.

In 2019 a team analyzing the dental calculus of a woman buried around 1000 to 1200 CE at a small religious community in Dalheim, Germany, found lapis lazuli particles embedded in the hardened plaque on her teeth. The most economical explanation the authors offer, among several they consider, is that she painted with ultramarine and pointed her brush with her lips. It is an inference and it is presented here as one. What the particles establish on their own is narrower and still remarkable: a woman in that community had the single most expensive pigment available in medieval Europe in her mouth, in a period whose manuscript production is overwhelmingly credited to men.

The monopoly of the Badakhshan mines ended in a chemistry laboratory. In 1824 the Societe d'Encouragement pour l'Industrie Nationale in Paris offered six thousand francs for a synthetic ultramarine costing no more than three hundred francs a kilogram. Nothing adequate was submitted for four years. On 4 February 1828 the prize went to Jean-Baptiste Guimet, who had worked out a process in 1826 and kept it secret; Christian Gottlob Gmelin at Tubingen published a different route about a month later and claimed priority, and the two men argued about it for years. Two dates in that paragraph are load-bearing and two are not: 1824 and 1828 come from the historical study cited below, while the 1826 date for Guimet's process and the exact interval before Gmelin's publication rest on the standard account rather than on a document read for this article. What is not in doubt is the consequence. Within a generation the color that had been rationed by the ounce could be bought by the pound, and the whole economics of blue collapsed.

Mummy brown was a real artists' pigment: a warm transparent brown made by grinding up Egyptian mummies, human and animal, sold by London colormen into the twentieth century. Sally Woodcock's 1996 study for The Conservator is the standard scholarly treatment and traces the trade through the colormen's own records. The story everyone knows about it needs its source attached rather than being repeated as fact. Edward Burne-Jones, on discovering what was in the tube, is said to have taken it into the garden and buried it; that account comes from Georgiana Burne-Jones's 1904 memoir of her husband, which is a memoir and is cited here as one.

Two of the century's most popular greens were arsenic compounds. Scheele's green, copper arsenite, was described by Carl Wilhelm Scheele in 1775. Emerald green, also sold as Paris green and Schweinfurt green, is copper acetoarsenite and dates from about 1814. They were brilliant, cheap and everywhere: in paint, in fabric, in artificial flowers, in children's toys, and above all in wallpaper. Chemical analysis of Morris and Company's own floral wallpapers confirmed that the foliage was printed with arsenical greens, which is a useful detail for anyone inclined to think of the arts-and-crafts interior as the safe alternative to the industrial one.

The harm was real and the mechanism most often given for it is contested, and dropping either half makes the paragraph false in one direction or the other. Arsenical pigments genuinely sickened and killed people through direct handling, dust and ingestion, particularly among the women and children who made artificial flowers and dyed fabrics; James Whorton's history of the period documents the scale of it. The specific and endlessly repeated story is a different claim: that damp wallpaper mold converted the arsenic into a lethal gas which killed sleepers in their beds and possibly Napoleon in his. William Cullen and Ronald Bentley argued directly in 2005 that the toxicity attributed to trimethylarsine, the gas in question, is an urban myth, because it is far less acutely toxic than the story requires. So the poisoning happened. The most cinematic route to it is the one under dispute.

Brent Berlin and Paul Kay published Basic Color Terms: Their Universality and Evolution with the University of California Press in 1969. Their method combined direct elicitation from native speakers of twenty languages, carried out largely with speakers available around the San Francisco Bay Area, with published descriptions of a further seventy-eight languages, for the ninety-eight-language total our own file reports. The book proposed that basic color terms are not arbitrary across languages, and that they enter a language in a constrained order. This paragraph is a statement about what was published, by whom, and how. Whether the proposal is true is the rest of this section.

The sequence they proposed runs: black and white first, then red, then green and yellow, then blue, then brown, then purple, pink, orange and gray in no fixed order among themselves. Stated as an implicational hierarchy rather than as a historical narrative, it predicts that a language with a basic term for blue will also have one for red, and never the reverse. That is the claim in its exact form, and stating it exactly matters, because the version that circulates is usually a story about languages evolving up a ladder, which is a stronger and more vulnerable thing to assert than an implicational pattern.

The framework did not stand as published. Paul Kay and Luisa Maffi drew in 1999 on the World Color Survey, a data set covering 110 unwritten languages, and replaced the original rigid ladder with a looser account driven by partition and composite categories, in which languages carve the color space by successive divisions rather than by adding fixed terms in a fixed order. That revision is often left out of popular accounts, which tend to cite the 1969 ladder as though the intervening thirty years had not happened.

English has eleven basic color terms and that number is not a universal maximum. Russian carries twelve, because goluboy, light blue, and siniy, dark blue, function as two separate basic terms rather than as light and dark shades of one, in the same way English treats red and pink as two terms rather than as light and dark red. Galina Paramei's 2005 paper is the dedicated argument that the two Russian terms meet the criteria for basic-term status. Keep that pair in mind: they return in section 07 as the most precisely measured effect on this page.

The universals claim has been attacked on method, on sampling and on premise, and the attack is serious rather than fringe. Saunders and van Brakel argued in a 1997 target article in Behavioral and Brain Sciences that the constraints Berlin and Kay reported are artifacts of the elicitation apparatus itself, in particular of using a fixed Munsell chip array that presupposes the hue-centered organization it then claims to discover. Stephen Levinson's 2000 study of Yeli Dnye, spoken on Rossel Island in Papua New Guinea, argued that the language has no basic color terms in the Berlin and Kay sense at all, its color vocabulary being small, largely derived from object names, and resistant to the elicitation the theory requires. And Debi Roberson, Ian Davies and Jules Davidoff's 2000 study of the Berinmo of Papua New Guinea, and Roberson, Davidoff, Davies and Laura Shapiro's 2005 study of the Himba of northern Namibia, reported categorical-perception boundaries that track each language's own color divisions rather than any universal set.

The universals side answered, and the answer is statistical rather than absolute, which is the single most important thing to understand about the current state of this argument. Paul Kay and Terry Regier's 2003 analysis of the World Color Survey reported that color naming across its 110 languages clusters far more tightly than chance would produce. Terry Regier, Kay and Richard Cook argued in 2005 that best-example choices for color terms cluster near the same regions of color space across unwritten languages. Regier, Kay and Naveen Khetarpal argued in 2007 that attested color vocabularies sit near optimal partitions of a perceptual color space, so that the regularity is a consequence of efficient division rather than of an innate list of terms. Kay and Regier replied directly on Berinmo the same year. The upshot is that a strict universal sequence is not established, while a statistical universal tendency in where languages place their boundaries is well supported.

A 2017 study reframed the question as one about communication rather than perception, and it is the most useful recent contribution here. Edward Gibson and colleagues found that across many languages warm colors, the reds and oranges and yellows, are communicated more efficiently than cool ones, and argued that the asymmetry follows from what color is used for: objects that matter to people tend to be warm-colored against cooler backgrounds, so warm regions of the space get finer and more agreed-upon labels. On that reading a language's color vocabulary reflects its speakers' visual environment and their needs, which explains a regularity across languages without requiring an innate list of terms at all.

The honest state of the question, sixty years on. Berlin and Kay's book changed the field and its central intuition survived: languages do not divide the spectrum at random. Its specific claims did not survive intact. The fixed evolutionary sequence is not established and is contradicted by at least one carefully documented language. What is supported is weaker and considerably more interesting: color vocabularies cluster far more than chance allows, plausibly because efficient division of a shared perceptual space and shared communicative pressures push independent languages toward similar solutions. Anyone who tells you the sequence is a settled law of language is overstating it. Anyone who tells you color naming is purely arbitrary is overstating it in the other direction. This page does not settle it, because it is not settled.

In 1858 William Ewart Gladstone, then a Member of Parliament and later four times Prime Minister, published a three-volume study of Homer whose third volume contains a long section on Homer's use of color. Gladstone had counted, and what he found is genuinely striking: Homer's color vocabulary is small, it is applied in ways that make no sense as hue, with the sea wine-looking and oxen wine-looking and honey green and sheep violet, and there is essentially no word doing the work English gives to blue. From that observation Gladstone inferred that the color sense of the Homeric Greeks was only partially developed. Hugo Magnus extended the inference in 1877 into a full evolutionary hypothesis: that color vision itself had developed within recorded history. The counting was real. The inference was not warranted, and the next paragraph is why.

Refused: that ancient Greeks, or Homer, or any historical population, could not see blue, had undeveloped color vision, or were in any sense physiologically color-blind. There is no evidence for it and there is evidence against it. The hypothesis was tested in its own century and it failed. W. H. R. Rivers, working on the Cambridge Anthropological Expedition to the Torres Straits in 1898 and publishing in 1901, administered color-vision tests to people whose languages had few color terms and found their discrimination normal. Nothing since has revived it. Human color vision is a property of the three cone opsins described in section 01, and there is no evidence of a change in those genes over the two and a half thousand years since Homer was written down. The perceptual story survives entirely as internet folklore, and it is worth noticing that it survives by being flattering in a particular way: it invites a modern reader to believe they can see something a Greek could not.

The defensible version is about words, not eyes, and it is better than the folklore version rather than a consolation prize for it. Ancient Greek color vocabulary was organized around axes English does not foreground: brightness, darkness, saturation, wetness, and the way a surface glitters or moves. Two Greek words cover much of what English calls blue. Kyaneos names a dark shade running into black, and glaukos a pale blue-gray with a gleam to it, the word behind Athena's epithet glaukopis. Neither is a hue label in the modern sense. A Greek could see the sea perfectly well while having no single word that meant its hue and nothing else. That is a claim about how a language cuts up experience, which is precisely the claim the entire color-terms literature in section 05 is about, and it is the reason Gladstone's counting still matters.

The best-known demonstration that a language's color categories affect performance is Jonathan Winawer and colleagues' 2007 study of Russian and English speakers, and it is worth reporting in full because the full version is smaller and more interesting than the headline. Participants saw three blue squares and pressed a key to say which of the lower two matched the upper one. Russian speakers were faster when the two lower squares fell on opposite sides of the goluboy and siniy line than when both fell inside one category. The size of that advantage, in the hardest trials where the two blues were perceptually close, was 124 milliseconds: 1,288 milliseconds within category against 1,164 across it. English speakers, tested on the identical squares, showed no such advantage in any condition. A concurrent verbal task wiped the Russian advantage out, turning the 124 milliseconds into minus 64, while a concurrent spatial task did not, leaving 109 milliseconds. That last contrast is the substance of the finding: it suggests the language is being recruited online, during the task, rather than having permanently reshaped the eye. One hundred and twenty-four milliseconds in a speeded button-press is a real effect and an invisible one in ordinary life. It is not that Russian speakers see more blues.

Debi Roberson and colleagues worked with the Himba of northern Namibia and the Berinmo of Papua New Guinea, both of whom have color vocabularies that group parts of what English calls green and blue under one term. Across recognition, similarity-judgment and category-learning tasks, both groups showed categorical-perception effects aligned to their own language's boundaries rather than to English ones. Read the dependent variable carefully, because everything downstream turns on it: what the studies measured throughout was speed of response and rate of error. At no point did they report a participant unable to tell two colors apart. A separate note is owed on the citation itself. Our own file U_2_01 cites this material through an entry that merges the 2000 Berinmo paper's title with the 2005 Himba paper's year, and gives a volume, issue and page range that match neither of them. Both papers are cited separately and correctly below.

Refused: the version of the Himba experiment that circulates online, in which participants instantly spot one odd green square among eleven near-identical greens but cannot find a blue square among greens at all. It is not a published finding. It was staged for a 2011 BBC Horizon program. Debi Roberson, one of the researchers whose work the program drew on, wrote of the actual literature that "the differences found have been in the speed of identification of the oddball, and sometimes in a greater number of errors. I'm not aware of any finding ... where a participant has failed to spot an oddball." Jules Davidoff confirmed that the televised stimuli "were modeled on those used by Paul Kay and others in experiments on other groups; but in all of the relevant experiments, the dependent measure was reaction time." Both statements come from correspondence published by Mark Liberman on Language Log, the University of Pennsylvania linguistics blog, on 17 March 2015, and that page is linked below. This is the cleanest correction available in the whole subject: a real effect, real studies, and a famous distortion refuted in the words of the people whose work it claims to report.

So the framing our own file uses is the right one and is carried forward here without softening it. The strong form of the Sapir-Whorf hypothesis, linguistic determinism, holds that language determines what can be perceived or thought, and is largely rejected. The weak form, linguistic relativity, holds that language influences perception and cognition, and in the color domain it is supported by studies of the kind above. The file's own closing hedge is carried forward exactly as it stands: effect sizes are small, and perceptual categories may be shaped by both biology and language. A hundred and twenty-four milliseconds is what supported looks like here. Anyone reporting linguistic relativity in the color domain as a dramatic reshaping of experience is reporting something the measurements do not contain.

The Roman Catholic liturgical year is color-coded, and the code is old: white for Christmas and Easter and for the feasts of confessors and virgins, red for Pentecost and for martyrs and the Holy Cross, green for Ordinary Time, violet for Advent and Lent, and black for funerals, now optional. Our own file dates the codification to 1198 and attributes it to Pope Innocent III, and both halves of that are imprecise. The program is set out in De sacro altaris mysterio, Book I chapter 65, a treatise written in the 1190s by Lotario dei Segni before his election as Innocent III, that is, while he was still a cardinal. And the Council of Trent did not legislate a color sequence; the rubrics were fixed in their familiar form by the Roman Missal of 1570 that followed the Council. One honest limit on this paragraph: the research behind this article did not locate a peer-reviewed article specifically dating the treatise. The pre-election composition date is standard in the liturgical literature, and the primary text sits in Migne's Patrologia Latina volume 217, for which no stable public locator was resolved, so it is named here rather than linked.

Two comparisons are worth a sentence each and no more than that, and both come from our own file with no external source resolved for either, which is stated here rather than left for a reader to discover. Eastern Orthodox practice is less standardized than the Roman but follows a comparable symbolic logic. Buddhist iconography works with a five-color set, white, blue, red, yellow and green, associated with the five Buddhas and the five wisdoms. Neither carries any weight in this article's conclusions, and no section is built on either.

Chinese five-phase cosmology assigns a color to each of the five phases: wood green, fire red, earth yellow, metal white, water black. The scheme runs through medicine, siting, court ritual and political legitimacy rather than sitting in a book about art. Its most visible consequence for a Western reader is that white, not black, is the color of mourning: white hemp garments at a funeral, where a European tradition would wear black. Our own file frames that as simply the opposite of Western convention, and the framing is worth keeping loose, because mourning color in Chinese practice is graded by the mourner's relationship to the dead rather than being one uniform color for everybody in the room.

The seven-chakra rainbow, red at the root through violet at the crown, is not an ancient Indian scheme. Classical Sanskrit sources including the sixteenth-century Sat-Cakra-Nirupana do assign colors to chakras, but not this sequence and not on this logic. The rainbow mapping is a Western construction, assembled through Theosophical writing and reaching its familiar modern form in the twentieth century, with Christopher Hills's Nuclear Evolution of 1977 the usual point of consolidation. Kurt Leland's book-length history traces the assembly step by step, from Blavatsky through Leadbeater, Steiner, Bailey, Jung and Esalen to the 1970s, and notes that the version now in universal use came together as recently as 1977. The tell is sitting in plain sight: the sequence happens to match the Western spectral order, which is not something a sixteenth-century Sanskrit text would have any reason to reproduce.

Set two of these systems side by side and the article's central observation makes itself without commentary. In Chinese practice red carries luck, prosperity and protection: red envelopes of money at New Year and at weddings, red as the traditional color of a bride's dress, red on doorways and lanterns. In the same tradition white, not black, is the color of mourning, and a European tradition would wear black to the same funeral. Two traditions, one set of wavelengths, opposite assignments, and nothing in the physics to arbitrate between them. One scoping note. The mourning-color material is sourced below; the observations about red rest on our own file alone, with no external source resolved for them, so they are offered here as a widely known cultural observation and not as a finding.

Michel Eugene Chevreul, director of dyeing at the Gobelins tapestry works, was asked why certain dyed yarns looked duller than they should, and found the problem was optical rather than chemical: adjacent colors change each other's appearance. He published the law of simultaneous contrast in 1839, as De la loi du contraste simultane des couleurs. A gray patch on a red ground looks greenish; the same gray on a blue ground looks yellowish. He systematized complementary relationships into a chromatic circle and showed that juxtaposing complementaries makes both look more saturated. Ogden Rood's Modern Chromatics of 1879 added the distinction between additive mixing of light, which brightens, and subtractive mixing of pigment, which darkens, which is why broken strokes of pure color read as more luminous than the same colors mixed on a palette. This material comes from our second file, U_2_19, and only this layer of it: what the painters did with it is that file's subject and not this article's. The research behind this article resolved no identifier for Chevreul's or Rood's own works, so neither man's book is linked below, and the source notes say so.

Goethe published Zur Farbenlehre in 1810 as a direct attack on Newton, arguing that color arises where light meets darkness through a turbid medium rather than being decomposable into a spectrum. As physics he was wrong and Newton was right, and nothing here softens that. As phenomenology he recorded a great deal worth recording: colored shadows, afterimages, simultaneous contrast, the psychological weight of particular colors. Schopenhauer took the theory up and modified it, and its influence runs through Steiner, Wittgenstein's Remarks on Colour and Albers's Interaction of Color. It is the clearest case in this article of a book being scientifically false and intellectually productive at the same time, which is a combination our tier system is built to say out loud rather than to resolve.

Whether color-emotion associations are universal is open. Domicele Jonauskaite and colleagues surveyed more than four thousand participants across thirty nations in 2020 and found substantial cross-national agreement in which emotions people attach to which colors, together with real variation: the association of yellow with joy, for instance, appeared in most but not all of the samples. The pattern that emerged is that the similarity between two nations' color-emotion maps tracks how close their languages and their geography are, which is what you would expect if the associations are partly shared and partly transmitted rather than either purely innate or purely arbitrary. Whether the shared part reflects biology, a shared environment such as blue sky and red blood, or cultural diffusion, is unresolved. The participant total here is given as more than four thousand rather than as an exact figure, because the exact figure was carried from the standard description of the paper rather than from a table read for this article.

Experimental color psychology has had a hard decade, and two of its most publicized findings are now doubtful. Russell Hill and Robert Barton reported in Nature in 2005 that competitors assigned red in four Olympic combat sports won more often than those assigned blue; the effect has since been contested on analysis and on generality. The romantic-red finding, that men rate women as more attractive against a red background, was subjected to two pre-registered replications by Gabrielle Lehmann and Robert Calin-Jageman in 2017, which found effects far smaller than the original and not clearly distinguishable from zero. Andrew Elliot's own 2015 review of the field is candid that the literature was built on small samples and thin theory and that the robust findings are few. That review is by a researcher inside the field rather than a critic outside it, which is what makes it worth the space.

Refused as fact: the marketing folklore of color. That red increases appetite. That blue suppresses it. That blue is calming. That a particular shade of pink pacifies violent prisoners. That a specific color on a button raises conversion. These circulate as established psychology and they are not. Some rest on single small studies that did not replicate, and some on no published study at all. What is true, and worth reporting as true, is that these claims are widely made and widely believed, and that whole design practices have been built on them. The honest summary is the one from inside the field: its robust results are few, and the confident claims in circulation are mostly not among them.

Synaesthesia, in which a stimulus in one modality reliably produces an experience in another, includes the color cases: seeing letters, numbers, days of the week or musical notes as colored. The prevalence figure our own file gives, around four percent, has a real study behind it. Julia Simner and colleagues in 2006 tested 500 people in the first study to screen a large population without relying on self-referred volunteers, and put the overall prevalence of synaesthesia at 4.4 percent, with about one percent for the grapheme-color form specifically. Ramachandran and Hubbard proposed in 2001 that the condition reflects excess cross-activation between adjacent cortical areas, and demonstrated psychophysically that synaesthetic color can affect perceptual grouping, which argues that the experience is perceptual rather than merely associative. The subject has its own page in this library, synaesthesia, and the fuller picture, including why prevalence figures disagree with each other, belongs there rather than here.

One further idea is carried here exactly as our own file carries it, which is to say with its own refutation in the same breath. U_2_01 floats the hypothesis that synaesthesia may be a survival of an ancestral perceptual mode more common in earlier humans, and states in the same sentence that there is no direct evidence for the evolutionary claim. That is the whole of it. The genetics of synaesthesia are under active study; the relic-perception story is a speculation with nothing under it, and it appears on this page only because leaving it out would misrepresent what the file we are working from actually says.

Refused, and our own file refuses these too, which is worth noting because much of this page has been spent correcting that file and here it is right. That chromotherapy, shining colored light on the body, cures disease: no controlled evidence supports it. That ancient Egyptians operated color-healing chambers inside the pyramids: no archaeological support of any kind. That aura colors provide reliable medical diagnostic information: no. And the shape of that refusal is worth stating plainly rather than leaving as a list. Color genuinely carries enormous cultural and psychological weight, which is exactly why claims of physical healing power attach to it so readily and so persistently. The meaning is real. The medicine is not.