Color words are more than a clever way to sell crayons. A study by British researchers suggests that color words in a given language shape human perception of color, perhaps explaining why some native English-speaking children, familiar with the rainbow of colors in the Crayola 64-pack, actually can tell "rust" from "brick" and "moss" from "sage," while children who grow up speaking languages with fewer color names lump such hues together. The research on English children and children in seminomadic Namibian tribes appeared in the December issue of the Journal of Experimental Psychology: General (Vol. 133, No. 4). Cognitive psychologist Akira Miyake, PhD, of the University of Toronto, says that the study "addresses an age-old question: To what extent does language shape or even determine the way we think?"

The answer appears to be, quite a bit, says Curtis Hardin, PhD, a social psychologist at Brooklyn College. He says, "The strongest finding is that with increasing linguistic competence, the correlation between memory and language scheme increased, while the correlation between memory and perceptual space decreased."

Miyake adds that these findings support the Sapir-Whorf hypothesis that language shapes or determines thought. "It had been out of favor, but the work of [this research team] has rekindled psychologists' interest," he notes.

A rose by any other name

The study tracked color naming, comprehension and memory in two populations over three years. Researchers led by Debi Roberson, PhD, of the University of Essex, compared young English children with children of the seminomadic, cattle-herding Himba tribe in northern Namibia, a country on Africa's southwest coast. At the study's start, the English children included 32 3-year-olds and 36 4-year-olds; of these, 28 became the longitudinal sample, completing all six tests. The Himba children included 42 3-year-olds and 27 4-year-olds; of these, 63 became the longitudinal sample.

Roberson and her colleagues first wanted to test two opposed hypotheses of color categorization--universalist (the idea that we categorize and remember colors the same way around the world, in keeping with the structure of our visual system) versus relativist (the idea that color perception depends upon culture and language). They did this by observing how children acquired color categories over time in two different languages and cultures. Second, the researchers hoped to systematically compare how the two groups of children learned by tracking over time how they mentally organized color as well as how they named and comprehended color terms.

Across cultures, the children acquired color terms the same way: They gradually and with some effort moved from an uncategorized organization of color, based on a continuum of perceptual similarity, to structured categories that varied across languages and cultures. Over time, language wielded increasing influence on how children categorized and remembered colors.

In short, the range of stimuli that for Himba speakers comes to be categorized as "serandu" would be categorized in English as red, orange or pink. As another example, Himba children come to use one word, "zoozu," to embrace a variety of dark colors that English speakers would call dark blue, dark green, dark brown, dark purple, dark red or black.

Roberson and her colleagues explain that different languages have differing numbers of "basic color terms." English has 11 such terms, the same as in many of the world's major languages, and Himba has five, each of which covers a broader range of colors.

In the study, testers used sets of 22 colored squares representing the English categories: the 11 basic terms of black, white, gray, red, green, blue, yellow, pink, orange, purple and brown, and 11 more colors that fit halfway between the basic colors (for example, blue-green).

An experimenter tested the English children in Witham, England, at first at home and then in school. Himba children were tested in their home village by an experimenter through an interpreter. At the sixth time of testing, 34 Himba children had attended a new mobile school for between six and 12 months; the remaining children never went to school.

Testers visited the children twice a year for three years. During each of the six sessions, the children completed four tasks of listing color terms, naming colors, comprehension and memory. First, the experimenter asked each child to "tell me all the colors that you know" (either in English or Himba).

Second, the experimenter showed each child all 22 tiles, one at a time, and asked, "What color is this?"

Third, the experimenter laid out all 22 tiles in random order in front of each child, and asked him or her to, "Show me a red [serandu] one." After a selection, the experimenter asked the child, "Is there another red [serandu] one?" This continued until the child responded negatively. Children were considered to know a term when they both correctly named the focal tile and pointed to it when asked.

Fourth, the experimenter laid the array of 22 tiles in random order in front of the child and then covered them with a cloth. The experimenter showed the child a tile from an identical set for five seconds and told the child that the same colored tile was hiding under the cloth. To test color memory, the experimenter then took away the cue tile, exposed the 22-color array, and asked the child to find the color that he or she had just seen.

True colors

Not only has no evidence emerged to link the 11 basic English colors to the visual system, but the English-Himba data support the theory that color terms are learned relative to language and culture.

First, for children who didn't know color terms at the start of the study, the pattern of memory errors in both languages was very similar. Crucially, their mistakes were based on perceptual distances between colors rather than a given set of predetermined categories, arguing against an innate origin for the 11 basic color terms of English. The authors write that an 11-color organization may have become common because it efficiently serves cultures with a greater need to communicate more precisely. Still, they write, "even if [it] were found to be optimal and eventually adopted by all cultures, it need not be innate."

Second, the children in both cultures didn't acquire color terms in any particular, predictable order--such as the universalist idea that the primary colors of red, blue, green and yellow are learned first.

Third, the authors say that as both Himba and English children started learning their cultures' color terms, the link between color memory and color language increased. Their rapid perceptual divergence once they acquired color terms strongly suggests that cognitive color categories are learned rather than innate, according to the authors.

The study also spotlights the power of psychological research conducted outside the lab, notes Barbara Malt, PhD, a cognitive psychologist who studies language and thought and also chairs the psychology department at Lehigh University.

"To do this kind of cross-cultural work at all requires a rather heroic effort, [which] psychologists have traditionally left to linguists and anthropologists," says Malt. "I hope that [this study] will inspire more cognitive and developmental psychologists to go into the field and pursue these kinds of comparisons, which are the only way to really find out which aspects of perception and cognition are universal and which are culture or language specific."

Rachel Adelson is a science writer in Raleigh, N.C.