Proof Positive That People See
        Colors With the Tongue

        By HENRY FOUNTAIN

              When it comes to colors, the English language is at
              no loss for words. Interior decorators and paint
        manufacturers, among others, are constantly coming up
        with ever more fanciful color names.

        To linguists, however, all the sample chips in a paint
        store can be categorized by 11 English words. While
        one decorator's buttercup breeze may be another's
        desert bouquet, to a linguist they are both yellow.

        But even 11 is a lot compared to the Berinmo, a small
        tribe of hunter-gatherers that lives along the Sepik
        River in Papua New Guinea. The Berinmo language
        categorizes colors with just five words.

        This makes the tribe a good subject for studying a
        linguistic concept that first gained wide currency in the
        1950's. Called the linguistic relativity hypothesis, it
        argues that humans see the world less with their eyes
        than with their language. Each language filters
        experience differently, and thus there are no universal
        principles of meaning. (In the most famous explanation
        of the hypothesis, Benjamin Lee Whorf, one of its
        proponents, argued that an Eskimo could never have
        just one word for snow, because of the importance that
        different types of snow have for Eskimo culture.)

        Following the hypothesis, if one people categorizes
        color differently from another, they should perceive it
        differently as well.

        That's not what researchers in the 1970's found with
        another remote tribe, in a study that sent linguistic
        relativity into something of a tailspin. The tribe, the
        Dani, who live about 200 miles from the Berinmo, have
        even fewer words for colors -- two. Yet in tests, their
        ability to remember colors was much like that of
        English speakers. The results seemed to indicate that
        there were some kind of universal color categories that
        transcended language.

        Now, three psychologists in Britain have repeated the
        Dani experiments with the Berinmo, and have come to
        the opposite conclusion.

        In the study, which is described in the current issue of
        the journal Nature, the Berinmo were shown samples
        from a standard 160-color chart and asked to categorize
        them. In addition to having fewer categories than
        English speakers, the categories are different. English
        speakers, for example, draw a distinction between blue
        and green. Berinmo do not, but they draw a distinction
        within what English speakers would consider yellow,
        with the word "nol" on one side and "wor" on the other.

        The critical part of the study came when the researchers
        asked the Berinmo to remember colors, by showing
        them one sample, waiting a short time, and then asking
        them to match the first color from two similar
        alternatives. Sometimes the two choices came from
        within the same color category, and sometimes not.

        The researchers found that the Berinmo were much
        better at matching colors across their "nol" and "wor"
        boundary than across English blue and green categories
        (after having been shown the blue-green distinction).
        And English speakers, given the same tests, were
        similarly good at blue-green matches and poor at
        matches across the Berinmo categories.

        By showing that color perception is dependent upon
        categorization through language, the results support the
        idea of linguistic relativity, said one of the authors of
        the study, Debi Roberson of the University of London.

        "Berinmo color vision is the same as ours," said Dr.
        Roberson, who came to know the tribe during nine
        months in the jungle. "If they are asked to identify a
        single color from a group of colors, they would do it in
        the same way as you or I.

        "But say you have three colors, and call two of them
        blue and one green," she continued. "We would see
        them as being more similar because we call them by the
        same name. Our linguistic categories affect the way we
        perceive the world."