30 December 2010

Color and Language

What could be more fundamental than primary colors? You learn them in pre-school. Almost everyone on Earth (apart from the colorblind) has three different kinds of cone cells which analyze photon wave length (it actually takes three or four photons for a cell in your eye to register) and communicates that color to your brain.

But, for all practical purposes, photon wave length is a continuous property. To communicate we must assign words to ranges of photon wave length. And, it turns out, there is not a universal way to translate the photon frequencies that our eyes see into words. One of the more notable examples is that not all languages have separate primary color words for blue and green.

Neither Basque, nor the Celtic and Scandinavian languages made that distinction until in archaic texts although it exists now, and some of those languages divide the blue-green colors at the boundary between blue and green linguistically into primary colors in a different way. Russian, Italian and Hebrew all have three primary colors in the part of the color spectrum that English allocates to blue and green. Many African languages, and archaic forms of East Asian languages do not have a single primary color word that includes blue and green in the same color.

The process of distinguishing more kinds of colors linguistically does tend to follow a common pattern in the great majority of languages, however. Languages tend to distinguish blue from green before they distinguish red from pink, for example. Also, it appears that languages tend to gravitate towards making more rather than fewer primary color distinctions over time, and that finer color distinction seems to be a frequently linguistically borrowed language feature.

The examples above are not random. One point that they illustrate is the fact that there are a wide variety of different patterns of color naming even between relatively closely related language families in the European branch of the Indo-European languages. Two of the most closely related Indo-European language families, for example, are the Celtic languages and the Italic languages, which share many isoglosses (i.e. they share the same linguistic feature in areas where related languages can be grouped into two or more categories based on their linguistic features). Yet, Celtic languages and Italic languages, which probably started to diverge from each other linguistically in the Iron Age (i.e. after 1200 BCE), have different words for describing colors.

This in notable because commonly used, simple words of the kinds one learns in pre-school tend to be the least prone to random linguistic drift or evolution. The less often a word is used, the more likely it is the evolve over time. In the Indo-European languages, for example, the root words for core simple vocabulary tend to be much more likely to match proto-Indo-European terms than advanced words that are used less frequently. Yet, primary color words are very common and used frequently.


Maju said...

Yes, pretty interesting. I'd say it is very common in languages to have a clear black-white and red-blue distinction, and notice that the red-blue dichotomy pretty much describes the rainbow, and with either yellow (pigments) or green (light) you have a the full palette.

I understand that in Basque it's not clear if yellow (hori, also meaning "that") existed in archaic times or not. Similarly there are Basque-etymology words for brown (nabar) and grey (uher) but their antiquity and original meaning is not too clear. But if it had green, pink, etc. they were lost and replaced by Romance or Latin loanwords.

As for Celtic and Italic, I am not too sure they are particularly close: their similitude may have more to do with sprachbund within the series of shared cultures: Corded Ware, Bell Beaker, Tumuli and Urnfield. But even if they had a common phylogenetic origin, gradually diverging along time, the mothers may have at several times spoken other languages (substrate influence) and that may be why you see them so different.

For example I was raised in Spanish language and in my family only one grandfather (no grandmother) could speak Basque fluently but even these grandmothers, grand-aunts and great-grandmothers used Basque words that were "kindergarten words" for me.

Still I'm checking and it seems that red follows IE etymology in Brythonic Celtic and Scottish, even if there are alternate words. Not sure about blue, black or white because these are more complicated.

I presume that red is a more emotionally intense color because even in Basque (gorri) it seems to resemble the kind of vibrant idea that also exists in IE: the only thing they have in common is the letter R but that seems enough. This sound R is present also in Arabic (hamraa', humr), Azeri/Turkish (k/qirmizi). However of non-West Eurasian languages I can only spot Khmer (grahorm) and some Austronesian languages using R in this word, so maybe it's a West Eurasian peculiarity (with some exceptions too: Georgian, Finnic, some IE languages, including most Indo-Aryan ones).

Andrew Oh-Willeke said...

The fascinating thing, of course, is that one would expect these words to go all the way back to PIE, thousands of years earlier, and would think that such a fundamental concept was more linguistically universal than it actually seems to be.

Maju said...

What I tried to suggest is that IE expanded with men and these "kindergarten" stuff are usually learned from women. That's one of the ways languages change and I'd say that one of the less well understood. There's too much emphasis in what we can understand rather easily (phylogenetic regular or at least recognizable patterns) but we there's many other ways a language can change: mostly adoption (from substrate or contact languages) and innovation (invention, more or less capricious and chaotic).