CHAPTER 4

COLOR TASKS

by Hugh M. Lewis

The color tasks were clearly projective and rank order in nature. The problem with the projective nature of these tasks seems to be that no valid "universal" ascriptions of associations can be given to the choice of colors. We really do not know what the choice of red over blue, or of the distaste of black or brown may represent, even in our own culture, much less within any other such as the Chinese. Colors do seem to represent moods to some extent, as well as some general sense of anxiety, anger, gregariousness, shyness and depression, and it is possible that the orders in which colors are selected or arranged may relate somehow to configurations of personality. But there appears few if any hard and fast associations between specific "colors" and specific feelings. This question is explored in greater detail in other color-related tasks.

The color tasks included an adaptation of the Luscher 8 color ranking task, a 12 color version of the same task, an ethnosemantic "grid", association and pilesort task using the same 12 colors, and an adaptation of the "Color Pyramid Technique" using 12 colors matched as close as possible to the 12 color Luscher. An 8 color triads task utilizing the same colors was afterward constructed and administered to a small American sample.

The method of analysis of these tasks involved the creation of frequency histograms for each of the colors of the different samples, the creation of correlation matrices from these histograms, 3-dimensional bar graphs and multidimensional scaling diagrams. Afterward Chi-square tests were run upon salient differences between the samples.

The nature of the design of the Luscher tasks leads to two successive rank orders of the same cards. These two rank orders are then split to create two sub-samples. In the Luscher 8 and Luscher 12 tasks, the colored cards are laid in a pyramid and half square in front of the informant on an interview board that gave uniformity of local ground. The informant was asked to choose their most favorite colors, the card was taken up, and the rank order of selection was recorded. This procedure was then repeated, with the interviewee being told to make the second selections independent of the order of the first.

For the Luscher 8 for all sub-samples the most salient color is black in the 8th or final position. The larger the sample size, the more salient black as the last choice became. The second most salient color is purple for the 1st and 2nd position, a saliency which is greater for women than for men, and for children than for adults. For men, purple is replaced by red as the most frequent color of first choice. For all the sub-samples, red emerges as the second most frequent color as a first and second choice. Yellow emerges as the most frequent colors in the third, fourth and fifth ranks. For the male sub-samples, both colors are more strongly salient in the second and third positions. Green and Blue compete for the intermediate positions of 3, 4, and 5, and are followed by Grey and Brown respectively which compete for the 6th and 7th positions. Brown is the most salient 7th color in three of the four sub-samples, except for the women for whom it is grey.

Thus it appears in the Luscher eight that the most salient color rank order is purple, red, yellow, green, blue, grey, brown and black, competing with juxtapositions between red and purple, yellow green and blue, grey and brown.

In the control group which consisted mostly of young Chew Jetty females, purple is clearly the most salient first choice, followed by black as a last choice. Green is the second most salient color, competing with yellow and blue which follow. Red, grey and brown emerge in the 5th, 6th and 7th positions.

The greatest differences between male and female samples is the saliency of red in the first and second position for the males and the saliency of yellow and green in the second and third positions for the females.

For the Luscher 12, black as the last choice remains the most salient color and purple remains the most salient color first choice, followed by violet, and then red and pink, green and yellow and orange, then blue and grey and finally brown and white. For the reference group, white and green emerge as the most common second choice colors, followed by violet and blue. Orange and yellow occupy intermediate positions, followed by grey, pink and brown as the 9th, 10th and 11th colors. For the men, red and pink are the first and second most common choices, followed by grey, yellow and purple, green and blue.

The bipolar ambivalence, or tendency to cluster at both ends, of some colors must be taken into account, as well as the relative flatness or even distribution across the landscape. Overall, there were few such curves in the color. Most colors are represented by fairly strong peaks. Red appears strongly ambivalent, especially for men more than women or children. For men and boys the flatest color is grey.

It is interesting to compare these color saliency landscapes with a color grid done in written English utilizing only the names of the colors. Though only small in number, the sample shows red, followed by black and brown, pink, grey and orange as the least favorite color and purple, followed by yellow, then blue, green and black and white as the most favorite. Strong intermediate colors are green and blue, grey, brown and yellow.

A more ethnosemantic sort of task was given with the same 12 luscher cards in an effort to get at some common color associations and values. The subjects were asked to choose one card. No mention was made of it being a favorite card. Of 21 subjects, the most salient first choice was red (33.3%), followed by purple (16.6%). The last three choices were brown, grey and black. Pink and orange both have peaks at the eighth color position.

The subjects were asked to identify the colored cards along certain basic dimensions of contrast that came out from the interviews as somehow important in differentiating the colors. These dimensions were male and female, pretty and ugly, light and dark, good and bad, dirty and clean, and dangerous and safe. For each category an intermediate category emerged that was considered "both male and female" and "sometimes dirty and sometimes clean"

In the category of male (ta por) /female (char bor), the most female colors were purple and pink (both 78%), followed by red (76%) and then violet (54.5%). It is interesting that among these three are the most salient first choice colors. The most male colors are grey and brown (59% each) and then followed by blue and green (54.5% and 50.%) respectively. Strongly "fe/male" colors included white (57%), black (50%) and orange (42.8%). Yellow is the only remaining color and is more evenly distributed between the three categories, though it is more strongly "female" (42.8%) than either male (23.8%) or inbetween (33.3%)

Similar types of patterns emerge for the other dimensions. "Pretty" (Swie) colors are purple, yellow, violet, red, pink, orange, white and green in order of decreasing frequency. "Ugly" (boh swie) colors were black, brown and grey. It is interesting to note that there were no significant "pretty/ugly" colors. Overall, colors were prettier than they were uglier.

Light (chien) colors included purple, pink white and violet. Dark colors (tang) were red, black, brown, blue, yellow, green, orange. Inbetween colors were minor, but violet emerges as the most ambivalent, being equally distributed between dark and light categories.

Good (ho) and Bad (boh ho) also emerged as individuals consistently made value judgment about different colors. Good colors were purple (90.9%), orange (95.2%), pink (82.6%), red (90%) blue (81.8%) and green (85.7%), all with a high level of agreement. Bad colors were black (73.7%) and grey (50%), while there were no significant inbetween colors. Two colors, grey and brown, were strongly ambivalent between the two categories (47.6% each).

Dirty (la sam, or "boh chin ki") colors were brown (71%) and black (33.3%), while clean colors (chin ki) were white (100%), pink and purple (92.8% each), orange and violet (85.7% each), green and yellow (78.5%) and red (76.9%). Ambivalent colors included grey and blue.

Dangerous (Hong hiam) colors included black (77.8%) and red (62.5%). "Safe"(boh hong hiam) included violet (100%), white (100%), purple (95%), pink (90.4%), green and grey (85%), orange (84.2%), blue (78.9%) and yellow (83.3%). Again, there were no significant inbetween colors.

Table Appendix 1-1. Correlations between Semantic Dimensions of 12 colors.

The most salient dimensions associated with the 12 colors are "safeness," "goodness", "prettiness" and "femaleness", followed by "cleanness and lightness". These are all positive associations. The following table demonstrates the positive correlations between the six positive dimensions, the negative correlations between these and the six negative dimensions, the the positive correlations between the negative dimensions of "dangerousness", "badness", "ugliness", "maleness", "dirtiness" and "darkness."

From this table it is apparent that the highest positive correlations are between the dimensions of goodness and prettiness, prettiness and cleanness and cleanness and goodness, as well as between badness, dirtiness and ugliness, as well as with maleness and dirtiness. High negative correlations are between prettiness and badness and dirtiness, and between goodness and dirtiness and ugliness.

In the same task, the terms for the colors were also elicited. The following colors show uniform (100%) agreement in color terminology: red ("ang"), blue ("lam"), white ("pek"), black ("oo"). Of 22 people, one person referred to the green card as "hai che" or "sea green" whereas most of the others called it "che" (95.45%). The purple card was referred to as "lamina" by 19 of 23 people (82.6%) while one person didn't know the name, one person referred to it as "chien lamina" or light purple, and one person called it "chien lam" or light blue and one referred to is tas "chui lam" which can be translated as "watered (down) blue". Yellow was referred to as "ooi" by 20 of 22 people (90%). One person referred to it as "sunkist" in reference to the orange, and the other as "kam" or orange. The color orange was called "kam" by 20 of 22 people (90%) and referred to as "chui ooi" or "watery yellow" by one person and as "tang kam" or "dark orange" by another. Pink was referred to as "chui ang" or "watery red" by 20 people (90%) and as "chien ang" by one person and "kam ang" or orangish red by another. Brown was called "chocolate" by 15 of 22 people (72.7%), perhaps reflecting a borrowed word, as "coco" by 3 people (13.6%), as "lamina" or "purple" by another, and as "tua te sek" or "liver colored" or interpreted as "maroon" by 3 people (13.6%). Grey was referred to as "hoay hu" (literally "cement or fire fish") by 20 of 22 people (90%), English "grey" by one person, and as "oo" (black) by another. Violet appeared to be the most ambiguous category. Nine persons referred to it as "lamina" or purple. Four persons called it "tang lamina" (dark purple) and four "chui ang" (watery red). One person each referred to it as "tang lam" (dark blue), "chui lamina," (watery purple), "chien lamina" (light purple), "chui lam" or watery blue, "sian ka ang" (literarily "off red" or referred to as "chien ang" or "tang ang") and as "tua ta sek" or "liver color".

It appears that violet especially is not a clear-cut color category, and the color seems mostly unfamiliar to the Jetty Chinese, except that I have noticed it on some of their standards and small flags that they fly. The ambivalence of the color violet, as it spans the categories of red, blue and purple, as well as the "dark/light" terms with which it is evaluated, suggests that this category is new or emergent for this culture.

Lightness and darkness appear to be very basic dimensions by which colors are analyzed and evaluated, and though it is not the most salient dimension by which all the colors are thought of, it does appear to be related and perhaps underlie basic distinctions between the other dimensions.

Semantic associations of each color were also elicited from subjects. There appeared a great deal of variability in pattern of response, in the number of associations made drawn on internalized or memory associations. At some point this resource became exhausted, then individuals would begin looking around their immediate context in order to find things of the same color. This was a common pattern with almost everyone thus interviewed. A couple of highly nervous children appeared to immediately begin searching their contexts for informational clues to the color, not relying upon memory process very much.

The red color was associated with temples (28.5%), shirts (23.8%), blood and red light (on boats and street stoplight) (19%), house and cars (14.3% each), fire, apple, postboxes, lipstick, paper, flower (9.5%), and dresses, music box, pencil box, shoes, plastic bags, incense burner, sampan, lantern, cat's eyes, street light, candy, skirt, bush, stool, paint, boat, motorcycle, brake lights, bicycles, bamboo pole, roof, clothes, rose, men's and women's clothes, lipstick, pen, and ice bucket (4.7% each). One person didn't know any associations. The average number of associations was 3.12 per person, with a range of 7 and a mode of 3.

Elicitations of semantic associations also yielded the following comments. Red is a lucky color. The Heart is bad if it is red--red is a bad color. If red is not dangerous then it is good. If you see red be careful--danger. Women like red--after cooking something and it turns red then real good. A lot of things are red--red is a warning color for danger. All Gods are honored with red--it is a spirit color.

Associations with blue included blue plastic buckets common on the Jetty (38%), motorboats painted blue (33.3%), shirts (23.8%), sky and pen, (19% each); clothes and pants (14.3%); the sea, houses, balls, shoes, bicycles, paint, and bags (9.5%); books, T-shirts, sampans, watches, ships, material, eggplant ("Brinjal"), lights, ribbons, motorcycles, helmets, cars, sapphire, cap, paper, water, pencil, and work clothes (4.8% each). In regard to the last item, "if clothing, blue hides dirt. For welding blue is the color." The average number of associations per person were 3.6 per person with a mode of 2 and a range of 9.

The color grey is associated with shirts, cars, and cement (19% each); houses (14.3%);floors, bamboo, ashes, boats (9.5%); and T-shirts, marble, dresses, motorboats, sampans, balls, shoes, hair, body, paper, material, table, pen, pencil, oil drums and pants (4.8% each). 3 people didn't know any associations (14.3%). Grey is a common color, it is associated with "a lot of things" If you mix black and white it becomes grey. Grey and black "fit together." "Grey is a dirty color for me." The average number of associations per person is 2.5, while the mode is 2 and the range is 8.

The color orange is associated with oranges ("chiam", usually "sunkist") (81.8%), shirts (18%), bicycles (9%), traffic lights, fire, t-shirts, table, shoes, paper, magic ink, material, skin, umbrellas, persimmons, tins, lights, Holland's soccer Jerseys, water tanks, ginger (4.5% each). Two people didn't know any associations and the average number of associations per person was 2.05 with a range of 5. It appears that not many things in the Jetty world are orange, and orange are not more common. The color term "orange" is used almost interchangeably with "sunkist" to refer to oranges.

Pink is associated with clothes (23.8%), chairs, bikes and chinese cakes ("mee ku") (14.2% each), material, houses, lipstick, and buckets (9.5%); and dresses, balls, shoes, tables, water buckets, paper, watercolor paints, pencil boxes, pens, plastics, women's clothes, material curtains, heart, shop, strawberries, koay (glutinous rice cakes), paint, snooker pin balls (4.7% each). The average number of associations per person was 2.35 with a range of 7 and a mode of 1.

Yellow is associated with umbrellas and shirts (19% each), lights (14%), oranges and houseplants (9.5% each) and with road signs, leaves, pencil boxes, dresses, bicycles, balls, houses, shorts, shoes, boards, paper, plastics, boats, bags, milk powder, koay, cakes, corn, bananas, Gods, plastic trim, flowers, clothes, cushions, pineapples, hats, socks, mee (noodles), the Shell logo, the ferry boat, Brazilian Soccer teams, sports shirts posters, poster letters, and poster colors (4.7% each). Two people didn't know any associations, and one simply said "a lot of things". One person told me that "darker yellow is prettier." The average number of associations is 2.67 per person, with a range of 7 and a mode of 2.

Purple is associated with clothes (27%), shirts (13.6%), t-shirts, pants, bikes, shoes, purple containers, material and cars (9%) and dresses, balls, ribbon, socks, plastics, umbrellas, pants, candy, drinks, paper, cakes, flowers, lamp, mangosteens, balloons, vehicles, cap, hat, stone, ribbons, "Brinjal" (eggplant) paint, colored pencils, pub lights, and the park (4.5% each). The average number of associations is 2.78 per person with a mode of 2 and a range of 7.

Green is associated with trees and leaves (22.7% each), traffic lights (14.3%), houses, sampans, shirts, shoes, apples, paint, grass, motorcycles, leafy vegetables ("chai") and cars (9%) and also with the forest, blackboards, pens, motorboats, bicycles, Durians, doors, paper, table, tins, containers, daun pandan (sweet, aromatic leaves used in making pandan bread and koay), bananas, material, tiles, pineapple, cucumber, clothes, pants, t-shirts, jade, park, poster and adultery (4.5%). "If one sees a tree that is green it is living. Green things are alive. Green is pretty." "Adultery--one thing green for Chinese." The average number of associations is 2.36 per person with a range of 6 and a mode of 1.

Violet is assocated with paint, shirts, colored pencils, watches and bicycles (8.3% each) and with candles, shirts, dresses, pants, tiles, roofs, houses, crayons, shoes, paper, "Brinjal", material, clothes, bracelets, flowers, umbrellas, bags and kueh (koay, or glutinous rice cakes) (4.1% each) That violet is not a common color and is an inherently ambiguous color for the Jetty people might be reflected in the fact that only 17 of 24 people gave any clear associations with the color and that 6 people said they didn't know any associations (25%). One person said "a lot of things" and left it at that. Another person tried to qualify the color and define it, "like blue but brighter, "purple" but not sure. After a while blue changes and becomes that color." Another person referred to violet as an "Indian color, gaudy--nothing." The average number of associations was 2.88 per person, with a range of 7 and a mode of 3.

Brown is associated with chocolate (22.7%) which it is called, with shoes, skin, and houses (18%), with colored pencils, the body, pens, joss sticks, and paint (9%) and with dogs, hair, bicycles, balls, bags, boats, kop peng (coffee with ice and condensed milk) paper, candy, pants, material, Sarsi soda, Brinjal, milo, clothing, socks, cupboards, wood, car, blankets, carbon paper, glass holders and coco (4.5%). Brown "can take dirt, hide it." The average number of associations is 2.88 with a mode of 3 and a range of 7.

Black is associated with hair (52.4%), shirts (28.6%), pants (23.8%)

bags and shoes (19%), tires and charcoal (14.3% each), boats, pens, eyes, death, bicycles, and motorcycle seats (9.5% each) and with socks, t-shirts, plastic bags, motorboats, sampans, houses, pencils, ships, paints, tar, tar, hair bands, motorcycles, shorts, clothes, trousers, sotong ink (octopus ink), people, monk's cloaks, cars and hand carts (4.8%). Three people said "a lot of things." "For Chinese, black is O.K.". "Old people "soi" black (bad luck, evil). Old people don't like it. Young people like to wear black. It is good to have red and pink with it. It is the color of death." "Black means death in the house for the family, but when you wear it its an easy color. It's dangerous because its dark. If you wear black pants any color shirt will go with it." The average number of associations is 3.2 per person, with a range of 9 and a mode of 3.

White is associated with shirts (36.4%), with paper (31.8%), house paint (22.7%), houses (18%), pants (13.6%), motorboats, ocean liners, shoes, t-shirts, fans, cars, dresses and boats (9%) and with uniforms, cotton, wool, the sky, dresses, lamps, the sea, cruisers, eggs, bags, plastics, tins, material, tiles, tissue, the sun, clothes, shoes, cigarettes, buildings, cushions, balconies, walls, lamplights and clouds (4.5%). Four people said "a lot of things." It was qualified with "White is clear". "If you use white as paint in the house, it's clear, clean, light. Most coffee shops usewhite as the base paint for walls. People like it." The average number of associations per person was 2.95 with a mode of 1 and a range of 7.

The same task also included at the end a "top-down" pile sort with the twelve color cards. Overall, the predominant trend was to subdivide the piles into two equal sized groups, from 12 to 2 piles of 6, and then from 6 to two piles of 3 cards each, then one card would be separated from the 3 card piles. Only two people did not follow this pattern, one grouping in odd sized piles of 8 and 4 cards, and one who seemed unable to do a top-down sort with the cards, and instead ended up doing a bottom up sort joining two cards at a time together. At the end of this informant's sort, a bottom-down configuration could be inferred.

The analysis of the pile sorts proceeded from left to right, and from grouped to ungrouped at the bottom of the tree. Red at each level was taken as an anchor point for the identification of the different colors.

The results of the pile sorts reveal that there were four overall levels of pile sorts for 19 of 20 subjects, with only 1 subject completing a five level sort. At the first level, red was most frequently associated with yellow (70%), blue (65%), green (60%), brown (55%), orange (50%), grey and pink (45% each), black and violet (40% each), purple (30%), and least with white (25%). The second cut in the first level included in order of greatest frequency white (75%), purple (70%), black and violet (60% each), orange (50%), brown (45%), green (40%), blue (35%) and, finally, yellow (30%). Orange was equally associated with both cuts.

On the second level in four groups of three (except one), the first two cuts coming from the first cut on level one, and the second two from the second cut on level one. Red was associated yellow (40%), brown (35%), green and blue (30%), black (25%), orange, violet and grey (15%), purple and white (10%) and no association with pink. In the second cut was pink (40%), blue and orange (35% each), grey and green (30%), purple, yellow and violet, (25%), brown (20%), black (15%) and white (10%). In the third cut was blue, white, pink (35% each), green (30%), orange, brown, violet and grey (25% each) and purple, yellow and black (20% each). In the fourth cut was purple (45%), white (40%), black and violet (35%), grey and orange (25%), pink and brown (20%), yellow (15%), and green (10%). For one sort that entailed six rather than seven groups at this level, the fifth pile was white and grey and the sixth one was pink and black.

On the third level there were 8 subsorts, with four single cards and 4 subgroups. Single cards associated with the first cut on the second level include yellow (35%), black and grey (15%), blue, green and brown 910%), and purple, orange, white and violet (5% each). Associated most strongly with red in the first cut on the third level, associated with the first cut on the second level, were brown, blue and green (15% each), orange, black and violet (10% each) and purple, yellow and white (5% each, or one time). No other associations were made with red at this level. Single cards associated with the second cut on the second level was orange and grey (20% each), blue, pink and brown (15% each), green, yellow and black (10% each). In the second cut associated with the second cut on the second level was purple and violet (25% each), blue (20%), green, pink and white (15%), and yellow, orange, brown and grey (10% each). The single cards associated with the third cut on the second level was blue and brown (20% each), green and pink (15% each), yellow and grey (10%) and orange, black and violet (5% each). The third cut on the third level associated with the third cut on the second level include purple (30%), green (25%), orange, white and violet (20% each), blue (15%), yellow, pink and brown (10% each) and black (5%). The fourth card associated with the fourth cut on the second level includes grey (30%), pink and white (15%), purple, yellow, black and violet (10%) and orange and brown (5% each). The fourth cut associated with the fourth cut on the second level includes white and black (30%), orange (20%), violet and purple (15%), pink and brown (10%) and green, yellow and grey (5%). The lone set of cards from the fifth and sixth cuts of the second level include violet, purple, white, grey, pink, and black (5% each).

The four cuts of the on the third level separate out on the fourth level into their individual cards. Only one individual carried a sort to the next, fifth level, and this was due to failure to separate on the third level the first and second cuts, and include separating a cut of yellow and orange with a card of green from the second cut on the third level, and a cut of white and grey and a single card of purple from the first cut of the third level.

There is surprising consistency in these pile sorts, as well as a tendency to draw out red, and yellow from purple and and violet in the first cuts. The following correlations by color reflect the different the degree of association at different levels of the pilesorts.

Table Appendix 1-2. Associations of the 12 colors across the levels of the pilesort

Analysis of one other color task may shed light on the grouping of colors. A 12 color version of the "color pyramid test" was given by matching the 12 colors of the Luscher cards with colored felt. Informants were asked to construct pretty and then ugly pyramids from these colored squares of felt. Each pyramid was symmetrically balanced, requiring fifteen patches of felt to complete. Thus there were from top down five rows of 1, 2, 3,4, and 5 squares on each level, respectively, and a total of fifteen such patches per pyramid. This technique, though complex and time consuming to prepare and administer, provides a useful and interesting way of corroborating the Luscher related rank order tasks.

Correlation scores for both colors and levels for five sub-samples were run, for men, women, girls, boys and the control-reference groups. A number of chi square tests can be run on different levels between groups and pretty and ugly pyramids. For example, purple and black in the first three positions for the girls' pretty and ugly pyramids reveal a chi square of 24.46, significant far above the .001 level of significance. Also, average frequency scores for colors and for levels of each sample can be calculated, and then correlated between the different sub-samples, as well as between pretty and ugly pyramids.