{"id":10016,"date":"2017-09-18T00:21:08","date_gmt":"2017-09-18T00:21:08","guid":{"rendered":"http:\/\/www.lifeandnews.com\/articles\/?p=10016"},"modified":"2017-09-19T00:25:53","modified_gmt":"2017-09-19T00:25:53","slug":"languages-dont-all-have-the-same-number-of-terms-for-colors-scientists-have-a-new-theory-why","status":"publish","type":"post","link":"https:\/\/www.lifeandnews.com\/articles\/languages-dont-all-have-the-same-number-of-terms-for-colors-scientists-have-a-new-theory-why\/","title":{"rendered":"Languages don’t all have the same number of terms for colors \u2013 scientists have a new theory why"},"content":{"rendered":"

Ted Gibson<\/a>, Massachusetts Institute of Technology<\/a><\/em> and Bevil R. Conway<\/a>, National Institutes of Health<\/a><\/em><\/span><\/p>\n

People with standard vision can see millions of distinct colors<\/a>. But human language categorizes these into a small set of words. In an industrialized culture, most people get by with 11 color words: black, white, red, green, yellow, blue, brown, orange, pink, purple and gray. That\u2019s what we have in American English.<\/p>\n

Maybe if you\u2019re an artist or an interior designer, you know specific meanings for as many as 50 or 100 different words for colors \u2013 like turquoise, amber, indigo or taupe. But this is still a tiny fraction of the colors that we can distinguish.<\/p>\n

Interestingly, the ways that languages categorize color vary widely. Nonindustrialized cultures typically have far fewer words for colors than industrialized cultures. So while English has 11 words that everyone knows, the Papua-New Guinean language Berinmo has only five, and the Bolivian Amazonian language Tsimane\u2019 has only three words that everyone knows, corresponding to black, white and red.<\/p>\n

The goal of our project was to understand why cultures vary so much<\/a> in their color word usage.<\/p>\n

Is it about which colors stand out the most?<\/h2>\n

The most widely accepted explanation for the differences goes back to two linguists, Brent Berlin and Paul Kay<\/a>. In their early work in the 1960s, they gathered color-naming data from 20 languages. They observed some commonalities among sets of color terms across languages: If a language had only two terms, they were always black and white; if there was a third, it was red; the fourth and fifth were always green and yellow (in either order); the sixth was blue; the seventh was brown; and so on.<\/p>\n

Based on this order, Berlin and Kay argued that certain colors were more salient. They suggested that cultures start by naming the most salient colors, bringing in new terms one at a time, in order. So black and white are the most salient, then red, and so on.<\/p>\n

While this approach seemed promising, there are several problems with this innate vision-based theory.<\/p>\n

Berlin, Kay and their colleagues went on to gather a much larger data set<\/a>, from 110 nonindustrialized languages. Their original generalization isn\u2019t as clear in this larger data set: there are many exceptions, which Kay and his colleagues have tried to explain in a more complicated vision-based theory.<\/p>\n

What\u2019s more, this nativist theory doesn\u2019t address why industrialization, which introduced reliable, stable and standardized colors on a large scale, causes more color words to be introduced. The visual systems of people across cultures are the same: in this model, industrialization should make no difference on color categorization, which was clearly not the case.<\/p>\n

How do you describe this color?<\/h2>\n

Our research<\/a> groups<\/a> therefore explored a completely different idea: Perhaps color words are developed for efficient communication. Consider the task of simply naming a color chip from some set of colors. In our study, we used 80 color chips, selected from Munsell colors<\/a> to be evenly spaced across the color grid. Each pair of neighboring colors is the same distance apart in terms of how different they appear. The speaker\u2019s task is to simply label the color with a word (\u201cred,\u201d \u201cblue\u201d and so on).<\/p>\n

\n \"\"<\/a>
\n Participants had to communicate one of the 80 color chip choices from across the color grid.<\/span>
\n Richard Futrell and Edward Gibson<\/span>, CC BY<\/a><\/span>
\n <\/figcaption><\/figure>\n

To evaluate the communication-based idea, we need to think of color-naming in simple communication terms, which can be formalized by information theory<\/a>. Suppose the color I select at random is N4. I choose a word to label the color that I picked. Maybe the word I choose is \u201cblue.\u201d If I had picked A3, I would have never said \u201cblue.\u201d And if I had picked M3, maybe I would have said \u201cblue,\u201d maybe \u201cgreen\u201d or something else.<\/p>\n

Now in this thought experiment, you as a listener are trying to guess which physical color I meant. You can choose a whole set of color chips that you think corresponds to my color \u201cblue.\u201d Maybe you pick a set of 12 color chips corresponding to all those in columns M, N and O. I say yes, because my chip is in fact one of those. Then you split your set in half and guess again.<\/p>\n

The number of guesses it takes the ideal listener to zero in on my color chip based on the color word I used is a simple score for the chip. We can calculate this score \u2013 the number of guesses or \u201cbits\u201d \u2013 using some simple math from the way in which many people label the colors in a simple color-labeling task. Using these scores, we can now rank the colors across the grid, in any language. <\/p>\n

In English, it turns out that people can convey the warm colors \u2013 reds, oranges and yellows \u2013 more efficiently (with fewer guesses) than the cool colors \u2013 blues and greens. You can see this in the color grid: There are fewer competitors for what might be labeled \u201cred,\u201d \u201corange\u201d or \u201cyellow\u201d than there are colors that would be labeled \u201cblue\u201d or \u201cgreen.\u201d This is true in spite of the fact that the grid itself is perceptually more or less uniform: The colors were selected to completely cover the most saturated colors of the Munsell color space, and each pair of neighboring colors looks equally close, no matter where they are on the grid.<\/p>\n

We found that this generalization is true in every language in the entire World Color Survey (110 languages) and in three more that we did detailed experiments on: English, Spanish and Tsimane\u2019.<\/p>\n

\n \"\"<\/a>
\n Each row orders the color chips for one language: Colors farther left are easier to communicate, those farther to the right are harder to communicate.<\/span>
\n Richard Futrell<\/span>, CC BY<\/a><\/span>
\n <\/figcaption><\/figure>\n

It\u2019s clear in a visual representation, where each row is an ordering of the color chips for a particular language. The left-to-right ordering is from easiest to communicate (fewest guesses needed to get the right color) to hardest to communicate.<\/p>\n

The diagram shows that all languages have roughly the same order, with the warm colors on the left (easy to communicate) and the cool ones on the right (harder to communicate). This generalization occurs in spite of the fact that languages near the bottom of the figure have few terms that people use consistently, while languages near the top (like English and Spanish) have many terms that most people use consistently.<\/p>\n

We name the colors of things we want to talk about<\/h2>\n

In addition to discovering this remarkable universal across languages, we also wanted to find out what causes it. Recall that our idea is that maybe we introduce words into a language when there is something that we want to talk about. So perhaps this effect arises because objects \u2013 the things we want to talk about \u2013 tend to be warm-colored.<\/p>\n

We evaluated this hypothesis in a database of 20,000 photographs of objects that people at Microsoft had decided contained objects, as distinct from backgrounds. (This data set<\/a> is available to train and test computer vision systems that are trying to learn to identify objects.) Our colleagues then determined the specific boundaries of the object in each image and where the background was. <\/p>\n

We mapped the colors in the images onto our set of 80 colors across the color space. It turned out that indeed objects are more likely to be warm-colored, while backgrounds are cool-colored. If an image\u2019s pixel fell within an object, it was more likely to correspond to a color that was easier to communicate. Objects\u2019 colors tended to fall further to the left on our ranked ordering of communicative efficiency.<\/p>\n

When you think about it, this doesn\u2019t seem so surprising after all. Backgrounds are sky, water, grass, trees: all cool-colored. The objects that we want to talk about are warm-colored: people, animals, berries, fruits and so on.<\/p>\n

Our hypothesis also easily explains why more color terms come into a language with industrialization. With increases in technology come improved ways of purifying pigments and making new ones, as well as new color displays. So we can make objects that differ based only on color \u2013 for instance, the new iPhone comes in \u201crose gold\u201d and \u201cgold\u201d<\/a> \u2013 which makes color-naming even more useful.<\/p>\n

\"TheSo contrary to the earlier nativist visual salience hypothesis, the communication hypothesis helped identify a true cross-linguistic universal \u2013 warm colors are easier to communicate than cool ones \u2013 and it easily explains the cross-cultural differences in color terms. It also explains why color words often come into a language not as color words but as object or substance labels. For instance, \u201corange\u201d comes from the fruit; \u201cred\u201d comes from Sanskrit for blood. In short, we label things that we want to talk about.<\/p>\n

Ted Gibson<\/a>, Professor of Cognitive Science, Massachusetts Institute of Technology<\/a><\/em> and Bevil R. Conway<\/a>, Investigator at the National Eye Institute’s Sensation, Cognition, Action Unit, National Institutes of Health<\/a><\/em><\/span><\/p>\n

This article was originally published on The Conversation<\/a>. Read the original article<\/a>.<\/p>\n","protected":false},"excerpt":{"rendered":"

Ted Gibson, Massachusetts Institute of Technology and Bevil R. Conway, National Institutes of Health People with standard vision can see millions of distinct colors. But human language categorizes these into a small set of words. In an industrialized culture, most people get by with 11 color words: black, white, red, green, yellow, blue, brown, orange, […]<\/p>\n","protected":false},"author":44,"featured_media":10017,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[8],"tags":[3170,3172,149,2755,737,3169,3171],"_links":{"self":[{"href":"https:\/\/www.lifeandnews.com\/articles\/wp-json\/wp\/v2\/posts\/10016"}],"collection":[{"href":"https:\/\/www.lifeandnews.com\/articles\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.lifeandnews.com\/articles\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.lifeandnews.com\/articles\/wp-json\/wp\/v2\/users\/44"}],"replies":[{"embeddable":true,"href":"https:\/\/www.lifeandnews.com\/articles\/wp-json\/wp\/v2\/comments?post=10016"}],"version-history":[{"count":1,"href":"https:\/\/www.lifeandnews.com\/articles\/wp-json\/wp\/v2\/posts\/10016\/revisions"}],"predecessor-version":[{"id":10018,"href":"https:\/\/www.lifeandnews.com\/articles\/wp-json\/wp\/v2\/posts\/10016\/revisions\/10018"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.lifeandnews.com\/articles\/wp-json\/wp\/v2\/media\/10017"}],"wp:attachment":[{"href":"https:\/\/www.lifeandnews.com\/articles\/wp-json\/wp\/v2\/media?parent=10016"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.lifeandnews.com\/articles\/wp-json\/wp\/v2\/categories?post=10016"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.lifeandnews.com\/articles\/wp-json\/wp\/v2\/tags?post=10016"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}