THE COLOUR OF WORDS

The human eye sees more than a million colours. We name only a few and these vary from language to language. This exhibition explores the colours some of these words describe: the white of a eucalyptus tree and the white of a clay pit, the colour of ripe bamboo. The images are made using artist’s pigments, each of which has a name in a specialist vocabulary of about 600 names¹. By contrast, there are languages which leave much of the colour space unnamed, or where colour is inseparable from other characteristics, such as smoothness or brilliance.

In Yucatec (Mexico), naming colour entails talking about shape, size and texture. In Hanunóo (Philippines), the conversation is about ripeness and desiccation, colour fastness and fading, while in Mursi (Ethiopia), cattle colours lie behind all colour terms. In 8^th century Japan, the word ao could be used for all cool colours: the blue of the sky, the green of bamboo, black earth and a dappled grey horse. 17th Century English had around 50 different words for yellow.

Among linguists, the ‘universalists’  maintain that there is a common pattern to the way  languages name colours, while linguistic 'relativists' argue  that the naming of colour can only be understood in the context of the culture and the structure of the language.  Many take some position between these two.

This exhibition offers no answers to these fundamental questions about how similar and how different we are in what we name and what we see, or if what we can name affects what we can comprehend. I have simply tried to translate some of the research back into colour and share my enjoyment of the complexity of the questions.

_{¹Siddall, R., Pigment Reference Collection, Personal Collection of Dr Ruth Siddall. London: University College London.}

What we see and what we name

The squares around the edge of the image are painted using  a range of pigments, lightened with Titanium White towards the pupil of the eye, which is Mars Black.

Top, left to right: Indigo, Antwerp Blue, Ultramarine, Cobalt, Cerulean, Ultramarine Violet, Cobalt Violet Light, Mars Violet,  Caput Mortuum, Quinacridone Magenta, Alizarin Crimson, Coral Red, Cadmium Vermilion, Red Oxide, Venetian Red, Red Oxide (repeated), Potters Pink, Light Pink, Mid Pink, Dark Pink.

Bottom, left to right: Chrome Yellow Orange, Cadmium Yellow Orange, French Yellow, Lead Red Minium, Ochre,  Cadmium Yellow Deep, Cadmium Yellow Mid, Cadmium Yellow Lemon,  Naples Yellow, Yellow Green, Terre Verte, Cobalt Green Light, Cobalt Turquoise, Viridian, Phthalo Mono Green,  Sienna, Burnt Sienna, Umber, Burnt Umber, Vandyke Brown.

Right-hand edge: The 11 basic colour terms  of Brent Berlin and Paul Kay, 1969¹. This model, central to the debate about colour naming, proposed that all languages name the whole colour space using at most 11 basic colour terms. They challenged the linguistic relativist position that ‘…each language performs the coding of experience into sound in a unique manner. Hence, each language is semantically arbitrary relative to every other language’ (1969, p.1), claiming ‘...colour words translate too easily among various pairs of unrelated languages for the extreme linguistic relativity theory to be valid’ (1969, p.2).

According to Berlin and Kay, all languages have from 2 to 11 basic colour terms, acquired in a set pattern over a long period of linguistic evolution. In the first stage of dividing the colour space, only black and white (brightness and darkness) are named. In the final stage, there are between 8 and 11 colour terms, added in this order: black, white, red, yellow and green (in either order), blue, brown and finally one or more of orange, pink, grey and purple. Basic Colour Terms were defined as: (a) not subsumed under any other term, e.g., not ‘scarlet’; (b ) terms which are morphologically simple, e.g., not 'pinky orange'; (c) terms which are not restricted by context, e.g., 'blonde'; and (d) terms which are in frequent use, e.g., not ‘lilac’.

Their research, looking at 98 languages, was followed in 2009 by the publication of the World Colour Survey², which collected comprehensive colour-naming data from an average of 24 speakers of each of 110 unwritten languages.  The researchers  used over 300 coloured chips (Munsell chips), which speakers were asked to identify by name. Their ‘universalist’ findings were challenged by linguistic 'relativists', who considered that the research methodology imposed a Western perspective, ignoring cultural and linguistic context; that the ‘evolutionary‘ framing reflected bias; and that the coloured chips were an inadequate representation of the natural world.

_{¹Berlin, B., and P. Kay (1969).  Basic Color Terms: Their Universality and Evolution. Berkeley: University of California Press.}

_{² Kay, Berlin, Maffi, Merrifield, Cook (2009). The World Color Survey. Stanford: Center for the Study of Language and Information.]}

Yellow

Defined in the OED as being ‘of a colour intermediate between orange and green in the spectrum; of the colour of the yolk of an egg, ripe lemons, daffodils, sunflowers, etc.’, the etymology of ‘yellow’ links it with diverse colour words in other Indo-European languages. Among them are Welsh †gell bay, brown, Lithuanian želvas greenish, Avestan zairi yellow, Persian zar gold, classical Latin holus vegetable, the English words gall and gold, and the ancient Greek cloros.

‘… the adjective χλωρός, cloros, usually translated 'green', refers to both wood and sea water, but also to sand, people, cheese, fish, flowers, fruits, gold, blood and tears (Liddell, Scott & Jones 1968 sv). In fact, this use suggests a range that goes from pale green to greenish yellow to yellow, and involves more or less any pale color. The explanation lies in its proto-Indo-European root * ghlo-, ghel-2 variant of * ǵʰloū- sparkle, shine, connected with  χλοερός, cloeròs, green, and χλόη,klon, the green of  new growth. Words such as yellow,  gold, gleam, , gloaming (twilight) came from this root * ghlo-, ghel-2.‘¹

Among the many other words linked to ghel-2 in the Indo -European Lexicon are the Old Icelandic word glámr ‘moon’, the Old Irish word glan ‘clean’, Old English glæd‘glad’, Middle English colera‘ bile’, Danish ‘gulerod’ carrot , Swedish ‘glöd’ ember, and dhelpër, the Albanian word for fox, literally ‘the yellow one’.²

_{¹ Busatta, S. The Perception of Color and The Meaning of Brilliance Among Archaic and Ancient Populations.  Antrocom Online Journal of Anthropology, vol.10 n.2-ISSN 1973-2880 (2014), pp. 311-312.}

_{²Indo European Lexicon, The University of Texas at Austin, Linguistics  Research Centre (online).}

Purple

‘Originally, purple derived from shells (Purpura) found on the coast of the eastern Mediterranean Sea. The animals gathered into shoals in spring time; abrasion produced a milky white fluid from which purple dye was obtained. When the shells were broken, the white substance oozed out. Upon exposure to the air and light this substance passed through a series of colours: first citron-yellow, then greenish yellow, then green, and finally, purple or scarlet. The juice obtained from Murex brandaris, a kind of Purpura, changed photochemically into a deep blue-violet, but that of Murex trunculus, another kind of Purpura,  gave  a scarlet red hue (Forbes, 1964: 114-122; Gipper, 1964). Mixing shells in various ratios and stopping the photochemical process at different points produced yellow, blue, green, red and violet. According to OED, in the middle ages purple applied vaguely to various shades of red but now it applies to 'mixture of red and blue in various proportions'. The purple dye industry goes back into pre­ classical period. However, its heyday was reached during the classical period, and the Greeks applied the term’ porphureos’ to cover all these hues.’¹

Etymologically  Greek ‘porphoreos ‘, the origin of ‘purple’, probably links with ‘Phoenicia’.  As early as mid-14^th century BCE this was the centre of the manufacturing of Tyrian Purple, a dye also known by the Greeks as Phoenician red, Phoenician purple, royal purple and imperial purple, worn as a sign of high status because of its costly production and colour fastness.²

_{¹McNeill, N. B., Colour and Colour Terminology, Journal of Linguistics, Vol. 8, No. 1 (Feb., 1972), pp. 21-33.  Cambridge: Cambridge University Press}

_{²Veropoulidou, R. The Tyrian Purple, a “royal” dye. In P. Adam-Velen & E. Stefani (Eds.), Greeks and Phoenicians at the Mediterranean Crossroads (2012), pp.103-106.  Thessaloniki: Archaeological Museum of Thessaloniki.}

White

Martu Wangka means 'Aboriginal language' and it is spoken by about 800-1,000 Martu people in and around the Gibson and Great Sandy Desert area of Western Australia.¹

Martu Wangka speakers participated in the World Colour Survey and were found, unusually, to have no basic colour term for white. The two basic colour terms they have are black and red.

‘There are three Martu Wangka terms that mean ‘white’ in Marsh (1992). The word piirl-piirlpade notes both ‘white’ and ‘white soil’ from a claypan. The latter is a naturally occurring cavity in the landscape made up of clay-like soil that retains water after rain. The word nyumpurlpa refers to the ‘white (of gum trees)’, while ngintarlpa translates to ‘white (of a dog).’ (Lenarcic p.158). In the World Colour Survey responses ‘twelve different terms in all were used to denote ‘white,’ including karntawarra (which supposedly has the general meaning ‘yellow ochre).’ (Lenarcic p.168)

‘… this runs counter to what would be expected due to the Berlin and Kay (1969) colour encoding sequence. Following this the first terms to emerge in any lexicon should be ones that equate to the categories ‘macro-black’ and ‘macro-white’ (these being analogous to ‘dark’ and ‘light.’)

Hargrave, the researcher looking at the data in 1982, marked a similarity with other Western Desert Dialects. Observing that other Western Desert dialects also had no single term for the colour white, he ‘alluded to the possibility that desert nomads may have a greater perceptual connection to texture, which may include the surface shine possessed by objects.’ (Lenarcic p.179)

_{¹ https://www.wangkamaya.org.au}

_{²Lenarcic, J. (2010). The Salience of the Hues: Colour Cognition from an Indigenous Australian Perspective. DPhil thesis,pp.158-170, RMIT University}

AO: Blue sky, green bamboo black earth and a dappled grey horse

Following on from the Basic Colour Terms model of the colour space divided by hue, researchers have suggested other perceptual criteria that may influence the colour vocabulary. The cool / warm contrast is one of these. In a recent study of 100 languages, cognitive scientists from MIT found that there are more colour terms for warm than for cool colours.  Their study suggests that this may be because warm-coloured objects stand out against a cooler background. Because there are more colour terms for warm colours, they were also labelled more consistently by speakers of all their sample languages.¹

Looking at 8^th Century Japanese texts, James Stanlaw found the word ‘ao’ used to describe: ‘green hills’, ‘green frog’,  ‘green bamboo’, ‘green peas/beans’, ‘blue sky’, ‘black (Artic) fox’, ‘black ox’, ‘black earth’, ‘black horse’ and ‘the kingfisher’s blue dress’ referring to his green wings.  He adds that one of the texts, the ‘Manyoo-shuu explicitly even says that ao was a white color with a bit of grey added … Thus, it appears that ao is the exemplar COOL term.’²

It is also the colour of the ‘green’ traffic lights in Japan:

‘The first traffic lights in Japan were installed in Hibiya in Tokyo in May 1930. These were made in U.S.A. They were soon followed by Japanese made traffic lights which were installed in Kyoto in December of that year. According to the law at that time the colours of the traffic signals were officially deemed to be ‘aka’ (red), ‘ki-iro’(amber) and ‘midori’ (green). It was the custom, however, to refer to the green light as ‘ao’ in accordance with the names of the 3 primary colours found in artists’ paints ‘aka’,‘ki-iro’ and ‘ao’.  According to the Japanese Traffic Bureau (2003, personal communication), the Japanese people felt no resistance calling the green light ‘ao’ as ‘green things could traditionally be described using this term’. Eventually, in recognition of the fact that the custom of referring to the ‘go’ light as ‘ao’ was universal across the nation, the Japanese government, in 1956, changed the law so that ‘ao’ became the official name of the colour. The colour itself remained unchanged, however.³

_{¹ Trafton, A.  Analyzing the language of color: Cognitive scientists find that people can more easily communicate warmer colors than cool ones. MIT News Office, September 18, 2017, online.}

_{²Stanlaw J.M., Japanese color terms, from 400 CE to the present: Literature, orthography, and language contact in light of current cognitive theory. In R.E. MacLaury, G.V. Paramei, D. Dedrick (Eds), Anthropology of Color: Interdisciplinary multilevel modelling (2007), pp.295-318.  Amsterdam/Philadelphia: John Benjamins Publishing Company.}

_{³Conlan, F. (2005). Searching for the semantic boundaries of the Japanese colour term 'ao',  p.135.  https://ro.ecu.edu.au/theses/58}

_{This Thesis is posted at Research Online.}

Red

‘The Maya language of Yucatan has only five basic color terms (Jeek' 'black',  i:ak' red, pink, orange, rust colored', k'a.an' yellow, orange', sak' white', and ya.?as 'green'), but they appear in seventy-five compound stems that discriminate semantically among variables other than hue, including brightness, saturation, relative size and discreteness, opacity, and texture.’ (Brickner p.283) ¹

Using the examples Brickner gives  of red, combining  these stems with the basic colour term can describe four nuances of saturation, from (cak-hep'-e'1en),  very red, brown, purple to cak-puk'-e'1en 'faded red, pink’. Other stems combine with the colour term to describe dullness and brightness, cak-ha'gleaming red (expanse, wall)', cak-pak'-e?en 'red (expanse, wall)’. These last two terms can only apply to red objects that are relatively smooth and flat. There are other compound terms that apply to objects that can be piled or clustered, and that are many, like cak-ya?p'-e?en' red (many fruits on tree; many tomatoes)'.

And then there is texture: wet  or dry, opaque or translucent, soft or hard, and smooth, lumpy, or prickly. Watery colours are compounded with the adjective meaning ‘bloody’. There is translucent red in contrast to opaque red, and finally stems that refer to texture: cak-'lol-e'len

'red (tender tips of young leaves; skin of young person affected by sun)', cak-p'os-e?en 'red, flushed', cak-yul-e'len 'raw red (nose from cold; ischial callosities of monkeys)' which is compounded with a stem meaning ‘blunt’, implying that it is rubbed raw.

Brickner concludes that while Basic Colour Terms may be sufficient as far as hue is concerned  it ignores other variables salient to colour naming. ‘In part, this is a result of the limitations imposed by the use of Munsell color chips, or similar stimuli, which by their nature are opaque, small, smooth, and flat. ….’ (Brickner p.299)

_{¹ Brickner, V.A., Color and Texture in the Maya Language of Yucatan, Anthropological Linguistics, Vol. 41, No. 3 (Fall, 1999), pp. 283-307.  Published by: The Trustees of Indiana University on behalf of Anthropological Linguistics.}

The colour of freshly cut bamboo

In his study of the colour categories in Hanunóo, a language spoken on the island of Mindoro, Philippines, the ethnobotanist Harold Conklin found an unfamiliar colour vocabulary which challenged the Basic Colour Term model.

1. mabiru: relative darkness (of shade of color); blackness (black)

2. malagti: relative lightness (or tint of color); whiteness (white)

3. marara: relative presence of red; redness (red)

4. malatuy: relative presence of light greenness; greenness (green) (Conklin p.341)

‘In general terms, mabiru includes the range usually covered in English by black, violet, indigo, blue, dark green, dark gray, and deep shades of other colors and mixtures; malagti, white and very light tints of other colors and mixtures; marara, maroon, red, orange, yellow, and mixtures in which these qualities are seen to predominate; malatuy, light green, and mixtures of green, yellow, and light brown. All color terms can be reduced to one of these four but none of the four is reducible.’ (Conklin p.343)

Within these categories he names the  most typical colour in each category as (more or less)  black, white, orange-red, and leaf-green.

He goes on to explain these unfamiliar colour categories as follows:

‘First, there is the opposition between light and dark, obvious in the contrasted ranges of meaning of lagti and biru. Second, there is an opposition between dryness or desiccation and wetness or freshness (succulence) in visible components of the natural environment which are reflected in the term rara and latuy respectively. This distinction is of particular significance in terms of plant life. Almost all living plant types possess some fresh, succulent, and often ‘greenish’ parts. To eat any kind of raw, uncooked food, particularly fresh fruits or vegetables, is known as pag–laty–un (< latuy). A shiny, wet, brown-colored section of newly-cut bamboo is malatuy (not marara). Dried-out or matured plant material such as certain kinds of yellowed bamboo or hardened kernels of mature or parched corn are marara. To become desiccated, to lose all moisture, is known as mamara (< para, ‘desiccation’)… A third opposition, dividing the two already suggested, is that of deep, unfading, indelible, and hence often more desired material as against pale, weak, faded, bleached, or ‘colorless’ substance, a distinction contrasting mabiru and maarara with malagti and malatuy.’ (Conklin p.342)

‘In short, we have seen that the apparent complexity of the Hanunóo color system can be reduced at the most generalized level to four basic terms which are associated with lightness, darkness, wetness, and dryness. ’ (Conklin p.343)

Conklin, H.C.  Hanunóo Color Categories. Journal of Anthropological Research,  Autumn, 1986, Vol. 42, No. 3,  Approaches to Culture and Society (Autumn, 1986), pp. 441-446. The University of Chicago Press.

Cattle coloured

Among the Mursi, an Ethiopian tribe of around 5,000 people in the  Lower Omo Valley, all colour terms are interpreted through the colour of cattle. This colour naming system is an example of a colour model, ‘in which the entire colour vocabulary of a language is interpreted through the colour set of an entity or phenomenon which is well-known in that society.’¹

 ‘A native Mursi speaker, in classifying any natural or artificial colour stimulus … will use no term that does not specify a recognised cattle colour. Similarly, there is no pattern, whether naturally occurring, or artificially produced, that a Mursi speaker will not designate by means of a term which he also uses, habitually, to discriminate between the colour configurations of cattle.’ (Turton p.321) He adds that the Mursi have no difficulty in naming colours that diverge from the colours found in cattle, but that the ’differences between cattle colours, then, provide the Mursi with an efficient model for classifying all the objects of the environment in terms of colour.’ (Turton p.328).²

Turton provides a list of Mursi colour terms, translated as black, white, red, yellow, green-blue, pink, brown, grey, light orange and bluish violet and the parallel cattle terms to which they apply.  For example, chag,I, designating  slate grey, bluish-grey or ash-coloured cattle, would also be the colour term for any natural or artificial blue-green stimulus

_{¹Biggam, C.  (2012). The Semantics of Colour: A Historical Approach, p.47.  Cambridge, Cambridge University Press,}

_{²Turton, D. There's No Such Beast: Cattle and Colour Naming Among the Mursi.  Man, New Series, Vol. 15, No. 2 (June 1980), pp. 320-338. Royal Anthropological Institute of Great Britain and Ireland.}

With thanks to:

Endangered Languages Archive (ELAR), SOAS University of London and Arts Council England