Basic Color Terms
Things of This or That Nature
My original opener/thought/theme for this blog was
This cerulean month has just about done me, so this edition’s going to be a bit of housekeeping. Or maybe a thorough brain-dewebbing with an inadequate tool—what I’m going to try to accomplish is to establish what it is that I think I’m talking about when it comes to the semantic phenomenon I’ve been calling materialism.
but instead I’m going to not do that. As I poked and prodded and circled that idea, trying to find a way in, I realized that I kept implicitly a) conflating two concepts that are most likely—but possibly not—distinct; and b) letting my English demon whisper in my ear.
(A) is a problem best explained with horses. In Latin (and this was going to be the central discussion of this blog), there are a few colors that are strictly reserved for horses. One of these for example is spadix. Spadix never refers to anything but horses; like the English word ‘roan’, it’s ontologically saddled to the members of genus Equus. However, it’s also the case that there’s a larger set of color terms, encompassing both horse-exclusive color terms (like spadix) and terms that are applied to horses but also other things. Which means that there’s yet another set of color terms, which are the horse excluding words—terms, that is, that are never applied to horses. What kept creeping into my mind and generally gumming up the works was
(B) my English demon, which is what I call the unwelcome tendency to apply English-language concepts to Latin. The imp kept insisting, in a very sub rosa sort of way, that there were groups of Latin words (‘the whites’, say), some of which could be horse words and some of which could not.
The problem is that is no Latin ROY G BIV, no set of canonical color terms containing branching sub-categories—none at least that was ever recorded in any surviving text. But it nonetheless remains possible that color terms were conceived of in vague categories. Gellius’ gathering of the rufi colores1 and his referring to both rufus and viridis as ‘simplices’2 indicates that one person, at least, and as many as three (Gellius and his debaters, Fronto and Favorinus) were thinking of at least two colors in these terms. That we don’t have any other explicitly named categories here or elsewhere isn’t necessarily a problem. Synonyms, substitutions, definitions, and poetic variations can tell us a lot. But first I need to honest with myself about why I’m even looking in the first place.
That Damned Berlin and Kay
In 1969 Brent Berlin and Paul Kay published a very short book called Basic Color Terms: Their Universality and Evolution. Even if you’ve never heard of it, you’ve heard of it. You know how if a language has just two color terms, they’ll be the terms for ‘light’ and ‘dark’, and if three the terms for ‘light’, ‘dark’, and ‘red’, and the fourth will be ‘green’/’yellow’, and so on? That’s Berlin and Kay.
B&K’s fundamental assertion is, like it says on the tin, that basic color terms are universal across languages. The spectrum of linguistic color may seem so vast as to be infinite, but in fact, if you get people in a room and ask them to sort Munsell chips like with like, you’ll find that we’re not so different. I don’t have Basic Color Terms in front of me, but a subsequent recent project by Kay and Richard Cook outlines the original experimental conditions neatly:
First, without the stimulus palette present, the major color terms of the collaborator’s native language were elicited by questioning that was designed to find the smallest number of simple words with which the speaker could name any color (basic color terms). Once this set of terms was established, the collaborator was asked to perform two tasks. In the naming task the stimulus palette was placed before the speaker, and for each color term t, a piece of clear acetate was placed over the stimulus board, and the collaborator was asked to indicate, with a grease pencil on the acetate sheet, all the chips that he or she could call t. In the focus task the stimulus palette was shown as before, and the collaborator was asked to indicate the best example(s) of t for each basic color term t.
What they ultimately theorized, as can be seen in this chart, is that linguistic color terminology ‘expands’ in a universally predictable order. All languages with two basic color terms have the same two basic color terms (‘light’ and ‘dark’); all languages with three, the same three; and so on, more or less, up to the maximum of 12 in Russian.
The thesis was shelled to shit almost immediately, especially by anthropologists. All B&K’s ‘collaborators’ were bilingual3 English speakers, and all of them lived and were tested in the Bay Area except for one group of Tzeltal speakers4. How could there not have been cultural and linguistic contamination by the experimenters that rendered the whole dataset worthless? Kay eventually made tactical adjustments to the methodology, and in his World Color Survey directed that
no preliminary interview was administered to establish a set of basic color terms, and in the naming task the 330 individual color stimuli were shown to each cooperating speaker, one by one, according to a fixed, pseudorandom order, and a name elicited for each (in contrast with the B&K procedure of presenting the entire stimulus palette at once in eliciting naming responses). Fieldworkers were instructed to urge observers to respond with short names…
In the decades between their original book and Kay’s World Color Survey, B&K would further adduce, as Saunders summarizes, that
[t]here are six Fundamental Neural Response Categories (FNRs): black, white, red, yellow, green, blue. Each is identical with a primary semantic category and can be named by the English words. More than six BCTs in a language (stage V and beyond) result from fuzzy intersections of FNRs, leading to derived BCTs. Less than six BCTs is the result of composite categories based on fuzzy unions of FNRs, the most common being GRUE (GREEN + BLUE).5
Now, up above I put ‘expands’ in scare quotes, but—and here’s the crucial point—B&K did not. Half of their theory was ‘universality’; its other half is ‘evolution’. Their contention was that, if all languages with three color terms have the same color terms, and all languages with eleven have the same color terms, and the languages with between three and eleven color terms have (sort of, give or take) the same color terms for each same number of terms, then these categories must perforce represent developmental stages. The beefed-out languages are more evolved; the shrimpy two-term languages are least evolved. Generally speaking, the most and least evolved languages are located exactly where you’d expect.
Unsurprisingly, it’s students of dead languages, like me, who have the least patience for this shit. We have evidence of the evolution of color terms: it’s in the court cases, poetry, novels, histories, studies, graffiti, tombstones, shopping lists, and bills of sale which make up the corpus of dozens of languages either extinct or in fossilized earlier forms. We can trace the appearance, disappearance, replacement, and development of color terms in and across languages over the course of millennia. What these color terms ‘actually mean’ is the subject of tremendous controversy and debate, hence the existence of this blog. But by every standard of interpretation attempted thusfar, virtually nothing that B&K’s evolutionary theory asserts is correct.
This has resulted in a great deal of polite, academic whaling. Bradley, a Latinist, calls B&K’s theory ‘universally criticised’ (27); Clarke, a Hellenist, has an almost cruel little aside about what one finds when “you scratch the surface” of their book (131-132); and Lyons excruciatingly dismantles more or less every part of their theory, by inquiring as to which, exactly, of Greek’s many terms for e.g. ‘green’ is the basic one. (If I recall correctly, Warburton, a copticist, also gets his claws in). Kay’s defense against this decimation is to talk about synchronic data drawn from modern languages; his only specific rebuttal is the single sentence “Philological reconstructions of data on extinct languages… are unlikely to cast more than hazy light on the matter” (Kay 1999 p. 745). It’s embarrassing, but at least they have the evolutionary psychologists on their side.6
We’ll Tell You What to Tell Us
The universality part is probably wrong too.7 But the thing that particularly galls and fascinates me is neither the universality proposition nor the clearly bankrupt evolutionary hypothesis; rather, it’s their original definition of a Basic Color Term: the poisoned fruit from which this whole forest grew.
According to them, a basic color term is8:
monolexemic (one word rather than a phrase)
not restricted to a narrow class of objects
salient (the first word that ‘leaps’ to mind when you ask for a list of color terms)
productive (you can make new words, like ishs, from it)
ideally not a word derived from a material (like golden, ashen, &c.)
not a recent loanword
not complex
So color terms like ‘sky-colored’, ‘leaf-colored’, ‘blood-colored’, &c are out, on counts 1 and 7; even if those were a language’s primary ways of talking about the colors of things, they wouldn’t count as basic color terms. It’s probably not escaping anyone’s notice that this is an excellent set of criteria for distinguishing the canonical color list of English—Berlin & Kay’s native language—from the rest of its many color terms.
Why does English (and why do many modern European languages) have canonical colors? Baker et al (298) trace the tendency back to Aristotle, or rather to an Aristotleian tendency in the early modern period. But it’s these seven colors because of Newton, and it’s these seven colors because, as the man himself said:
Cum igitur colores iuxta medium constipatiores sunt, ita ut inter flavum et rubeum, iuxta & inter caeruleum & purpureum, quasi triente maius intersit intervallum quam inter viridem & flavum ... ; quo imago elegantius in partes inter se proportionales distinguatur, in numerum quinque insigniorum colorum duos alios, citreum scilicet inter rubeum et flavum, ac indicum inter caeruleum ac violaceum, asciscere convenit.9
Since therefore the colors are more crowded near the middle, such that between yellow and red, and likewise between blue and purple, there exists an interval roughly 1/3 larger than that between green and yellow ... ; so in order for the image to be divided more elegantly into proportional parts, it was fitting to admit to the five more notable colors two others: orange, naturally, between red and yellow, and indigo between blue and violet.
He just liked how it looked.
Anthropologists and linguists10 love to point out how incredibly narrow B&K’s conception of both culture and language is; even in his recent work, Kay imprecates his fieldworkers to “urge observers to respond with short names” when possible. But why should brevity be a requisite quality of a color term? Probably he’s trying to avoid descriptions like ‘the color of, you know when it’s winter, and the sun’s just gone down, but there are still a few rays shooting up over the horizon, and you’ve just spoken to the man you love for what you have realized will be the last time, and the light catches the tongue of an icicle on the edge of a leaf, and a terrible fear consumes you? That color.’ But if that’s the description that’s the description! Here we’ve stumbled upon our One Easy Trick for establishing universals: tell people what they can and can’t say.
But we don’t need to flee to the edges of the modern world to find languages that upend B&K’s theory: the imperial color of the imperial language of the Imperium sans phrase will do quite nicely. Purpureus, by any measure one of the most culturally significant words of any kind in Latin, could not, by B&K’s criteria, be considered a basic color term: it’s nonproductive (like most Latin words); it’s derived, and was understood by its speakers to be derived, from a material (purpura); while not a ‘recent’ loanword, it does come from Greek, and its referent material was felt to have a distinctly foreign origin. It’s also far from the ‘last’ word to appear in Latin, which B&K’s evolutionary schema would predict if we dare to translate it as ‘purple’—in fact it appears to predate the preferred Latin term for, and again I apologize for flippantly translating, ‘black’.
Terms of Color Basics
Still, the question of, if not ‘basic’ terms, at least general categories of concordances remains interesting to me.
If we want to talk about categories, though, what should their basis be? I’ve already pointed out how there were a discreet set of terms for hair colors. Should the operational categories be things like ‘hair colors’, ‘horses’, ‘clothes’, &c? Or should the categories be chromatic (‘the pinks’) and the colors differentiated by their contextual restrictions—what I’ve been calling their ontological limits? The temptation to say ‘both’, post, and go to sleep is deeply alluring.
The great difficulty and enormous advantage of working with dead people is that you can’t bully your collaborators. They’ve said what they’ve said sine ira et studio. The only thing I can do is apply the methodology below and hope I find something. If I don’t, then, fair play to the ghosts: ‘nothing’ is still a result, a lesson Paul Kay has never been able to learn.
What I’m Going to Do
If it turns out there are conceptualized but unnamed categories, that’d be very significant. Various modern authors have claimed that ‘reds’, ‘whites’,11 and ‘yellows’12 were culturally salient categories for Romans and/or ancient Latin speakers. If that’s the case, it’s remarkable that not all of these categories were named by ancient writers. So what I’ll do is begin by addressing claimed categories, ancient and modern, and then see if I can divine any others.
I’ll test if these categories have any weight to them by looking at first at loci with lots of color terms in one place: definition of one color in terms of others like we see in Gellius or Servius (just because I think they’re incorrect analysts doesn’t mean they’re bad informants), or passages where an author varies terms for the sake of emphasis or simple variety. If the same thing, in the same place, is defined by multiple terms, those terms are probably closely related. It’ll also be useful to cross-reference objects by the terms used for them: if a thing can be a few different colors in largely similar contexts, those terms probably have some connection. And eventually I’ll see if that all feels terribly wrong, and then I’ll just do something else.
Biblio
Baker, Tawrin, et al. “Introduction: Early Modern Color Worlds.” Early Science and Medicine, vol. 20, no. 4-6, 7 Dec. 2015, pp. 289–307.
Berlin, Brent, and Kay, Paul. Basic Color Terms: Their Universality and Evolution. United Kingdom, University of California Press, 1991.
Bradley, M. Colour and meaning in ancient Rome. New York, Cambridge Univ Press, 2011
Clarke, Michael. “The semantics of colour in the early greek word-hoard”. Cleland, Liza, et al. Colour in the Ancient Mediterranean World. British Archaeological Reports Limited, 2004. pp. 130-138
Conklin, Harold C. “Hanunóo Color Categories.” Southwestern Journal of Anthropology, vol. 11, no. 4, Dec. 1955, pp. 339–344.
Dana, Francis Marion. The Ritual Significance of Yellow among the Romans. 17 Jan. 2007. Accessed 19 Jan. 2025.
Houstoun, R.A., “Newton and the Colours of the Spectrum.” Science Progress. Vol. 12, No. 46, Oct 1917, pp. 250-264
Kay, Paul and Maffi, Luisa. “Color Appearance and the Emergence and Evolution of Basic Color Lexicons” American Anthropologist. Vol. 101, No. 4, Dec 1999, pp. 743-760
Levinson, Stephen C. “Yélî Dnye and the Theory of Basic Color Terms.” Journal of Linguistic Anthropology. Vol. 10, No. 1, June 2000, pp. 3-55
Lyons, John. “The Vocabulary of Color with Particular Reference to Ancient Greek and Classical Latin.” In: Borg, Alexander (ed.). The Language of Color in the Mediterranean: An Anthology on Linguistic and Ethnographic Aspects of Color Terms. United Kingdom, Almqvist & Wiksell International, 1999. 38-72.
Newton, Isaac. Lectiones Opticae. 1729.
---. Opticks : Or, a Treatise of the Reflections, Refractions, Inflections and Colours of Light by Sir Isaac Newton. London, Bell, 1931.
Saunders, Barbara. “Revisiting Basic Color Terms.” Journal of the Royal Anthropological Institute, vol. 6, no. 1, Mar. 2000, pp. 81–99.
Warburton, D. 2004. "The terminology of ancient Egyptian colours in context". In: L. Cleland, K. Stears, G. Davies (eds.). Colour in the ancient Mediterranean world. Oxford, Hedges. BAR International Series, 1267: 126–130.
Aulus Gellius Noctes Atticae 2.26.5
Ibid 2.26.4. Gellius’ term is, I weakly suspect, a translation of Aristotle’s ‘hapla’.
BCT 12
BCT 7
Saunders 82
Saunders 81. I don’t entirely know how she meant it, but if I gave an instant’s satisfaction to an evolutionary psychologist I’d never leave the house again.
Bradley (29-30) points out that the aspects of the Sudanese Dinka’s color system that were so appealing affirmative to B&K were not universal terms but restricted to cows; Conklin (341 n12) notes the difficulties presented by the lack of a Hanunó'o equivalent to the English word ‘color’ and adds (342) a long discursus on several conceptual categories that influence color (?) terms in that language, such as light/dark, wet/dry, indelible/faded, attractive/démodé, &c. My favorite line from Kay’s outline of the World Color Survey is the admission that “a small number of other languages…have in fact been found to violate universal tendencies.” Me when my universals are violated &c &c
BCT 6
Newton Lectiones Opticae 243-244. (I translate the terms without caution based on Newton’s English work). Houstoun (260) points out something in the Lectiones Opticae that Newton also alluded to in his English-language Opticks (186)—the appealingly musical chime of the number seven. He also notes (259) that based on Newton’s writing ‘sea-green’ is as canonical a color as orange or indigo.
Conklin and Levinson, to take two among many
Thomas, for example
Dana, for example