From: Mark Odegard
Message: 294
Date: 1999-11-20
Proto-Indo-European probably had what Berlin and Kay classify as a Stage II color system, which includes only black, white and red; Berlin and Kay use the term 'row' for red, as the semantic space occupied by 'red' in such a system is not that you find in larger systems. 'Row' stands for red-yellow. Homeric Greek would seem to also be Stage II, or perhaps, transition to Stage III, where either yellow or blue-green (what Berlin and Kay call 'grue') are added to the system. Stage IV has both grue and yellow. Stage V fully discriminates green and blue. Stage VI adds brown. Stage VII adds pink, purple, orange and gray. It needs to be noted that different languages break up color differently. Russian, so I understand, breaks up blue and green differently than English. Russian blue is not English blue. At the earlier stages, the distinction between black and white also needs to be widened. The semantic space would seem closer to that conveyed by English 'light' and 'dark'. After *h1reud, erythros in Greek (which Piotr has covered), the next IE color term is *ghel, which is ancestral to the English words yellow, gold. There is also the term *h1elu, 'dull red'.Piotr responds:I wouldn't be that dogmatic about Berlin & Kay's classification. Their generalisations have a physiological basis and are borne our by many languages, but they cannot aspire to absolute validity. What colour terms a language has depends to some degree on universals of colour perception and semantic evolution, but also on cultural factors such as the everyday needs of the speakers of that language. Grey is apparently a luxury that you can't afford until you climb to stage VII, but if there's a lot of grey around you and the need to distinguish various greyish hues is vitally important to you, you may develop terms for shades of grey even if your grass is still grue. PIE seems to have had some "grey" words, such as *kas-no- 'whitish-grey, the colour of a hare's fur'. And of course it is very likely that some colour terms which existed in PIE are accidentally unattested.
First, an extract from my ever-present Encyclopedia of Indo-European Culture. This is from the article "Color" (p. 113), by Martin Huld (edited by Mallory).
The term primary as used here should not be confused with primary spectral colors (red, blue, and yellow) which refer to the physical nature of color blending. A primary term is monolexemic; its meaning is not deducible by analyzing the meaning of its parts, e.g., 'lemon-colored' or 'reddish' are not primary color terms. It is also primary because it cannot be subsumed under another color catagory, e.g., orange is a primary term because speakers of English regard it as a color between yellow and red, while crimson is regarded as 'a kind of red', that is it can be defined as a variety of a more inclusive term. Secondary color terms, those employed only for restricted objects, e.g., roan or bay (for animals) or brunette (for hair), and those defined in terms of other colors, e.g., scarlet ( = a kind of red) or those derived from natural objects as a sub-shade, e.g., turquoise (a kind of blue) are all excluded from primary. However, colors have been known to shift classes, thus 'orange' was originally a shade of yellow derived from the fruit, but is now widely regarded as a distinct color.
This exactly describes my understanding of how one defines 'colors' in terms of human language. The stage VII colors (black, white, red, green, yellow, blue, brown, purple, pink and orange) are all quite distinct to me, and cannot be classed with any other color. For myself, I would add something like beige/tan/khaki to the list, and probably navy as well. These are the primary English color words; other languages have their own words, and often divide up the colors differently (I gather that orange is a peculiarly English-language color).PIE probably had 'hare-colored' (or more likely 'as a hare', 'of a hare', hare-like), but did it have 'gray'? It would seem the Mycenaeans who dealt in purple cloth (Murex dye) did not have a word for this color. Murex, I gather, produces what English-speakers regard as a blue, though it's usually referred to as a purple.
If you describe a color as healthy-leaf-colored, you are not speaking of a primary color, but comparing what you are describing to something else. Green stands alone; it is not compared to anything else.
Huld picks through the problems of what stage PIE must be assigned to in the Berlin-Kay system. Stage II seems the most probable, but we nonethless have the solidly-attested *ghel, which pops up in the daughters as either yellow or green. We also have the equally solidly attested *kueh1 (circumflex over k, with a symbol that turns u into w under the u), which gives us green (in English, it gives us 'hue') and would describe Berlin and Kay's 'grue'. This could make PIE transitional stage IV, but asserting such opens all sorts of problems for Homeric Greek, requiring us to say it was somehow 'degenerate' in this regard. Once they acquire a primary color word, languages don't seem to lose them; "in general, color systems to not descend to a lower stage over time." (pp. 114-115)
Me? Dogmatic? Nah. I just know (and parrot) what I read. This stuff makes a lot of sense. It especially agrees with my own sensibilities about how you define colors.
Mark Odegard.