Does the emotional context influence the recollection of color ? To answer this question, in an abstract painting with 12 colors, area was the same size, also there were four texts, each of wich concerned different emotional connotations (fear, anger, joy, and sadness) and was related to the life of an imaginary painter. Items were show each of 142 children age 12 and up, one painting associated with one of four stories. Then, they organized the colors according to the subjective place that they thought there occupied on the painting they had just seen (the paintings were removed before the question). Our results demonstrate that the emotional context does influence the recollection of color.
Everyone has an intuitive knowledge of the relation between colors and emotions. The French language has many color-related expressions such as "black as anger", "red as shame", "green as fear". Meerum Terwogt and Hoeksma (1995) point out that this pattern is similar in English, e.g. "purple with rage" or "green with envy." Thus, it seems reasonable to think that colors have an affective positive or negative value which depends on personal as well as social factors.
It is possible to study the link between colors and emotions. Meerum, Terwogt and Hoeksma (1995) developed the idea that in a homogeneous population, favorite colors will readily be associated with pleasant emotions. Although the hypothesis is justifiable, it poses problems because of the effects of favorite color change with age (Meerum, Terwogt & Hoeksma, 1995), gender (Silver, McCulley, Chambliss, Charles, Smith, Waddell, and Winfield, 1988; Silver & Ferrante, 1995) and cultural origin (Silver et al., 1988). The links between emotions and colors may be difficult to characterize given these factors. Mood could influence perception and judgment of color. In fact, Vorsobin and Zhidkin (1980) showed that the preference for some colors such as red, orange, or yellow increases for children 5 to 7 years old when they are in a happy situation.
Colors also can apparently influence the mood of the subject. A study by Deribere (1983) indicated that a vivid green color reduces aggressive behavior and nervous tension. Shauss (1979), studying prison inhabitants, found that a pink color in their cells decreased the aggressive behavior of the inmates. To explain this, he refered to Gerard (1958) who observed that subjects placed in front of a red or blue light for a short time showed a signifcant decrease in blood pressure and speed of electrodermal response, respiration, and blink reflex. All these physiological reactions may be directly related to changes in internal mood state.
These ties between colors and emotions cannot be understood apart from cognitive processes that govern them. These relations affect treatment of information and memory. Thus, Bornstein (1976) has developed the idea that colors are at the root of some memory activations because they have an important place in the setting-up of a memory network. He shows that after memorization of a picture with a particular color, the presentation of this color activates the salience of the figure. During the encoding, a link is created between the two memories.
Concerning encoding, Allen (1983) had showned that the "name" of a color and the "idea" of a color do not have the same representations in our memory. The recall of the color is better than the recall of its name. This phenomenon has been explained in terms of double encoding (Paivio, 1969, 1971). It gives colors a double status on the mnemonic levels: colors are susceptible to simultaneous verbal encoding or as image, although the visual encoding seems to be more important.
One observation shows that the treatments of objects and colors are not necessarily correlated. In fact, while patients with visual agnosia can memorize color identification, they cannot memorize objects (Wasserman, 1993).
At the neuropsychological level, it has been shown that for "split-brain" subjects, the memory of color is located in the left hemisphere (Levy & Trevarthen, 1981) or at least that this hemisphere is more important for memory of colors. Sonnier and Dow (1985) located in this hemisphere the memory of emotion, which in relation to the memory of colors must play a key role in learning quality of an environment.
Experimental evidence of color-emotion links and color-memory links exists. Moreover, it has long been known that memory and emotions interact. Organizational theories of memory take note of these ideas (Bower, 1981; Beck, 1974). Therefore, the associations of memory, colors, and emotions are worth exploring.
Our research explores how emotional context affects the salience of color. We placed a colored item in four different emotional contexts. Then, we tested the way this colored item was remembered to observe the influence of emotional context.
Subjects. The subjects were 68 girls and 76 boys (age 12 on the average) attending four different French junior high schools. None of the children were color blind. They were randomly separated in four groupes of 36 children xxxx.
Materials. The colored material in this research was an abstract painting, generatd by computer software created specially for this research project. This program treats color distribution as one image and makes it possible to create an abstract picture without meaning, composed of 41 color patches (9.5 cm by 14 cm).
The patches were not simple geometrical forms. All are different, and they do not resemble any namable outlines. It is necessary to control the forms, because, as Wilton (1989) showed, colors can strongly affect the memorization of common geometrical shapes.
Blue, red, yellow, orange, green, purple, pink, beige, gray, brown, white, and black were chosen because French children can identify and differentiate these colors without difficulty (Hautek_ete-Sence, 1990).
In the painting, each color occupied the same surface area. We verified with judges in a neutral emotional context that no single color of the painting seems larger in surface area than others. To control the possible effects of central or lateral position, we made random counterbalanced arrangements of the colors, producing a total of four paintings. The configuration of the patches is the same in all the paintings, but the several colors do not occupy the same patches in each painting (Figure 1).
|Insert figure 1|
To place the painting in a particular emotional context (joy, sadness, fear, anger), we wrote four verbal texts. The four emotions have been chosen with reference to Plutchik's theory (1980) because they possess certain features. According to Plutchik, joy and anger are emotions that a subject manifests in his behavior, while sadness and fear are more internal.
Each text is a "dictionary biography" and relates the emotional life of an imaginary painter to the chosen emotion. All texts were written around a principal emotion, which was named only once. Each text contains several sentences, and each describes, in 150 words (in French), important moments involving strong emotions in the life of the imaginary artist.
Each text has been tested with children of the same age as the subjects of our experiment to verify that it would be understood without difficulty and with recognition of the intended emotion.
Procedure. Subjects, in groups of 12, were tested in the classroom. A researcher announced to them that they would be subject to a memory test involving a painting and a text.
Each subject received a reproduction of the painting (one of the four arrangements, randomly assigned) and after a couple of minutes, one of the four texts was read by a researcher. The researcher read the text twice, slowly, clearly and in an appropriate emotional tone. So, for each subject, just one text and one arrengement was presented.
The subjects were told that they held in their hands a reproduction of a real painting in a museum of their region and that the text read to them explained the painting.
Then, we took back the paintings and the children were asked to "think" of the author and his work for 5 minutes.
After this time, we gave them a pack of 12 cards (12 cm x 5 cm), reproducing the 12 colors present in the painting. Subjects had to classify the cards according to the importance they thought these had in the painting. We asked them not to classify two cards in the same rank.
The children's responses were coded: the first card chosen by each child was considered to indicate the first color seen in the picture and noted 12, the second 11, etc.
With regard to the position of the colors in the painting, no statistical difference was observed in the responses given by the subjects. We concluded that the place of the colors in the painting did not have any effect on the memorization of the colors.
No statistic effect of sex was discovered. This confirms results of Hautek_ete-Sence (1986), who has noted that when subjects are asked to associate colors with emotions, a difference can be shown between responses due to age but not sex.
Table 1 shows the mean obtained for each color as a function of the emotional context in which the painting was placed. Table 2 is summarizes the analysis of variance to test for the principal effect of the emotional context for each color.
Emotional context had a significant effect only for blue, orange, green and purple. We made a further examination of contrasts between the emotions for each color. Following this analysis, significant results are more numerous and involve seven colors:
Blue: the mean in the joy situation is the lowest (m = 4.25). The contrasts of joy vs. fear (F(1-108) = 7.09; p < .01), joy/sadness (F1-108=5.13 ; p < .05) and joy/sadness-fear-anger (F(1-108) = 7.47; p < .01) gave significant results. It seems that blue occupied a more important place in the subjects' experience of the painting when it was experienced in a "joy" context.
Yellow: the only significant contrast is joy-anger/sadness-fear (F(1-108) = 3.91; p < .05). The lowest averages are obtained in the joy situations (m = 4.80) and anger (m = 4.74). It seems that in joy and anger situations, the subjects remember yellow as taking a more important place than in the sadness and fear situations.
Orange: the mean for this color in the joy situation (m = 4.88) is the lowest, while in the anger situation (m = 6.85) it is the highest. The contrasts joy/fear ((f1-108) = 4.91; p < .05), joy/anger (F(1-108) = 9.80; p < .02) and joy/sadness-fear-anger (F(1-108) = 8.11; p < .01) are significant, as is the contrast anger/joy-sadness-fear (F(1-108) = 4.48; p < .05). Therefore, we can conclude that the orange color was very often associated with joy and less often with anger. If the same subject could see the painting in the four situations, each time we asked what space is occupied by the orange color, he or she would say that this space is more important in the joy situation and less in the anger situation. The emotional context set by the text was important.
Green: the mean for green is lowest in the fear situation (m = 5.45), and the contrast fear/joy-sadness-anger (F(1-108) = 6.51; p < .02) is significant. Moreover, the average of the responses given for green in the anger situations (m = 7.42) is the most important and the contrast anger-sadness-fear is significant too (F(1-108) = 6.30; p < .02). We conclude that in the representation of the painting the green color was ranked higher if the associated emotional context was fear, and given a less important place if the associated emotional context was anger.
Purple: for purple the contrasts sadness-fear/joy-anger (F(1-108) = 12.3; p < .01) and fear/joy (F(1-108) = 4.51; p < .05) are significant. The averages in the sadness situations (m = 4.72) and fear (m = 4.91) are lower than the averages of the joy (m = 6.80) and anger (m = 7.11) situations. It seems that in the cognitive representation of the painting made by the subjects, purple occupied less space if the associated emotional context was joy or anger, and more space if the associated emotional context was sadness or fear.
White: the lowest average for white is the one obtained in the anger situation (m = 6.28). Only the joy/anger contrast (F(1-108) = 4.35; p < .05) and anger/joy-sadness-fear contrast (F(1-108) = 4.20; p < .05) are significant. Therefore, white is seen by the subjects as occupying more space in the picture when the associated emotional context is anger.
Black: the results are the same for black as for white. The lowest average is given in the anger situation (m = 6.08). Only the contrasts anger/joy (F(1-108) = 6.13; p < .02) and anger/joy-sadness-fear (F(1-108) = 4.37; p < .05) are significant.
Based on the results of this experiment, our analyses tend to confirm the general hypothesis that we initially proposed: the emotional context of a painting can influence the salience of colors recollected by the children.
We infer that this salience is established both by information from "emotions" and from the "colors" seen. The primary goal of our study was to show the existence of this link, not to clarify the meaning or origin of the link.
Now that the evidence is in place, we can ask a number of questions. Certainly, between the presentation of information to the senses and its recollection, various cognitive processes take place. The experimental procedure that we followed does not permit us to affirm that one or another process is at the origin of the link between emotions and colors. We can only make hypotheses.
First, we propose that the effect of the emotional context on the salience of colors in the painting is due to processes of information selection, which depend strongly on the subject?s attention.
Clearly there exists an acquired cultural tie that associates some colors to particular emotions (Hautek_ete-Sence, 1986; Meerum Terwogt & Hoeksma, 1995). In the presentation of our painting, our emotionally-toned texts may have focused the audience's attention and guided their processes of selection to favor some colors over others.
It is also possible that the interpretation of the painting does not directly depend on the selection of information or on the attention of the subject, but rather on the encoding of information. If the visual form is to be encoded from short-term memory, these processes depend on a working memory module, whose capacity is limited in two ways: by the number of items that it can contain and by the duration of processing (Frick, 1988; Morris, 1989). Of course, the number of items contained and their treatment in this module depends partly on selective and attentional processes.
Some colors may have received more thorough processing, creating a particular configuration among the diverse stocks of information. These configurations may not include emotional information because emotion is not necessarily present in the assumed processing of the memory configuration.
But the observed effect of emotional context could also occur in information retrieval. As the subjects of our experience did not know beforehand what task they would confront, they could carry out their ranking of colors as a function of the interpretations of the painting made a posteriori.
In this case, the cognitive processes responsible for the results that we obtained would not be linked to memorization of the painting presented in some emotional context. Rather, subjects, making their sequential ranking, might ask themselves a question like: "What are the colors that I would use if I wanted to explain the message that the artist wished to put across?"
The cognitive processes used in this way would not, therefore, be directly linked with the material that we used, but to "emotion-color" links experienced previously by the subject, principally as a result of cultural learning.
On this point, the controversy can be resolved. If the subjects were acting this way, their responses should have been the same in our experiment as if we were asking them to estimate the degree of relation between each of the proposed colors and emotions. Now, the degree of relation between colors and emotions has been tested by Meerum, Terwogt and Hoeksma (1995) on 11-year- old subjects, and by Hautek_ete-Sence (1990) on subjects aged 12 (the average age of subject in our experiments). Their results are very different from the ones that we obtained.
Meerum Terwogt and Hoeksma (1995) used only six of the 12 colors on our study. Therefore, we choose to consider only the study of Hautek_ete-Sence (1990), which used all our 12 colors.
For example, we note that in this study, the colors most strongly associated with "anger" are black and gray, and the colors less associated with "anger" are pink, yellow and orange. For the same population, but with a different procedure, our results show that the red, the purple, the blue and the green are strongly associated to "anger", while white, gray, black and beige are less strongly associated with "anger."
For each emotion that we have tested, results are different from the findings of Hautek_ete-Sence (1990). Table 3 displays these differences.
These differences are explained by the tasks that the subjects were asked to complete in the two experiments. For a color to be placed in a particular emotional context is therefore quite different from judging an emotional weight for that color.
In this way we can show the influence of the emotional context on the salience of seven of the colors in our study.
The procedures of our experiment provide for future research, material whose significance has been tested. Our hypothesis could be verified only with material specially created to limit the experimental bias. Now this material exists, and it will be possible to use it in numerous different procedures.
Csikszentmihalyi and Schiefele (1992) affirm that it is generally admitted that science and art are the two symbolic systems which convey knowledge. Of course, the nature of knowledge coming from these two sources is not the same. Nevertheless, in our study we find both of them inducing meaning.
The author wishes to thank Val?rie Vienne and Monique Billaut.
Allen, C. K. (1983). Short-term memory for colors and color names. Psychological Reports, 53, 579-582.
Beck, A. T. (1974). The development of depression: A cognitive model. In, R. Friedman & M. Katz (Eds.), The Psychology of Depression, Contemporary Theory and Research. Washington, DC: Windson.
Bornstein, M. H. (1976). Name codes and color memory. American Journal of Psychology, 89(2), 269-279.
Bower, G. H. (1981). Mood and memory. American Psychologist, 36, 129-148.
Csikszentmihalyi, M. & Schiefele, U. (1992). Arts education, human development, and the quality of experience. In, B. Reimer & R. A. Smith (Eds.), Arts in education. Ninety-first yearbook of the National Society for the Study of Education (pp. 169-191). Chicago: University of Chicago Press.
Deribere, M. (1983). La couleur. Paris: PUF.
Frick, R. W. (1988). Issues of representation and limited capacity in the visuospatial sketch pad. British Journal of Psychology, 79, 289-308.
Gerard, R. M. (1958). Differential effects of coloured lights on psycho-physiological functions. Los Angeles: University of California. In Schauss 1979.
Hautek_ete-Sence, D. (1986). Couleurs et Emotions. (unpublished manuscript).
Hautek_ete-Sence, D. (1990). Impact Emotionnel de la Couleur Appliqu?e ? l'Emballage. CPCIA Communication, Nov 90, Paris.
Levy, J. & Trevarthen, C. (1981). Color-matching, color-naming and color-memory in split-brain patients. Neuropsychologia, 19(4), 523-541.
Meerum Terwogt, M. & Hoeksma, J. B. (1995). Colors and emotions: Preferences and combinations. The Journal of General Psychology, 122(1), 5-17.
Morris, N. (1989). Spatial monitoring in visual working memory. British Journal of Psychology, 80, 333-349.
Paivio, A. (1969). Mental imagery in associative learning and memory. Psychological Review, 76, 341-363.
Paivio, A. (1971). Imagery and verbal processs. New-York: Holt, Rinehart & Winston.
Plutchik, R. (1980). Emotions: A psycho-evolutionary synthesys. New-York: Harper & Row.
Schauss, A. G. (1979). Tranquilizing effect of color reduces aggressive behavior and potential violence. Orthomolecular Psychiatry, 8(4), 218-221.
Silver, N. C. & Ferrante, R. (1995). Sex differences in color preferences among elderly sample. Perceptual and Motor Skills, 80, 920-922.
Silver, N. C., McCulley, W. L., Chambliss, L. N., Charles, C. M., Smith, A. A., Waddell, W. M., & Winfield, E. B. (1988). Sex and racial differences in color and number preferences. Perceptual and Motor Skills, 66, 295-299.
Sonnier, I. L. & Dow, M. G. (1985). The right hemisphere: Seat of emotion colors. Education, 105(4), 373-375.
Vorsobin, V. N. & Zhidkin, V. N. (1980). A study of color preference in preschool children who are experiencing positive and negative emotions. Voprosy Psikhologii, 3, 121-124.
Wasserman, E. A. (1993). Does the brain contain color modules?. American Journal of Psychology, 106, 466-470.
Wilton, R. N. (1989). The structure of memory: Evidence concerning the recall of surface and background colour of shapes. The Quaterly Journal of Experimental Psychology Human Experimental Psychology, 41(3-A), 579-598.
Joy situation text
"Life is delightful. Let's make the most of it." This is Alberto Papagai's cry of joy. Alberto, born in 1952 in a happy sunny village in the Southern Italy, says he's so contented, he wouldn?t leave his village for love or money. Alberto is a happy man, always having a good time for him, everything is source of pleasure and laughter, and each day brings a new happiness wich one must take the most of. More over in each of his paintings is found all the joy that he wants to share with us. This painting entitled "The most beautiful instant" represents the tears of joy that he shed the day his little girl came into being.
Sadness situation text
Louise Lalune (1940-1991), born in Lille, was nicknamed "Melancholy" by those that knew her. It is often said that her life was just a series of lamentable tragedies. She lost her family in a fire whilst still young, then was brought up in a mediocre orphanage. Pitifully, she descended into misery and boredom, a deplorable world for someone who had so much talent. This work, entitled "Complaint and Sadness" represent the vision that she had of a doleful, morose world in which she had no reason to live. Louise Lalune finished her unhappy existence alone in a retirement home where, as she said herself, "Time seemed too long".
Fear situation text
Louis Tatinghen (1947-1990) lived as a child in a sunny Alpine village. Before turning seven, he was lost for over 3 days and 3 nights in a thick forest. From then on, his dreams were no more than deplorable nightmares, filled with anguish and terrifying monsters. In all his works, Louis Tatinghen tried to represent his fears and his scary nights that he couldn?t overcome. At 33 years old, Louis Tatinghen commited suicide for reasons we can only guess at since he'd just finished this painting called "Last Terror" wich is doubtless only a pale reflection of a nightmare more terrifying than all the others.
Anger situation text
Jack McNooler, born in Belfast in 1947 is today considered the most committed artist in the struggle for the Republican cause in Northern Ireland. Aged 15, the day his father was gunned down by a British soldier, Jack took up arms and kpomed the I.R.A. He abandoned everything and fought with fury and vengeance. Aged 23, he was betrayed by one of his own and arrested for a crime he did not commit. Full of rage, he escaped and left Ireland although he did not give up the struggle. He has found refuge in the United States from where he continues to speak out against the British occupation of his country. This painting called "Storm in the Head" is, according to Jack, the memory of his powerless fury.
SUMMARY: Does emotional context influence the salience of color? We created an abstract painting with 12 colored areas, each area the same size. We also created four brief texts, each ostensibly the biography of an imaginary artist, each life involving different emotional connotations (fear, anger, joy, and sadness). To 142 children, aged 12 and older, we showed one of the four stories and then the painting. We then removed the painting. We asked the subjects to rank the painting's colors according to how important the subjects thought they were in the painting they had just seen. Our results suggest that emotional context does in fact influence the salience of color.