Skip

Do you see what I see?
August 30, 2011 3:31 AM   Subscribe

How language affects our perception of colour...(SYTL) more on the 'linguistic relativity hypothesis' here and here
posted by Rufus T. Firefly (52 comments total) 28 users marked this as a favorite

 
Related?
posted by ShutterBun at 3:44 AM on August 30, 2011


Damn. An almost double :(
posted by Rufus T. Firefly at 3:46 AM on August 30, 2011


Yeah, I thought so too, at first. Just another group arriving at the same conclusion.
posted by ShutterBun at 3:53 AM on August 30, 2011


I'm colorblind. Can barely tell blues and purples apart, same problem with green and brown, and some browns and reds. The green light on a traffic light looks so white to me, that on various occasions I've sat looking at a green light wondering why there was a hole where the green light should be (in fact, there were clouds in the sky behind it).

We had a lot of lawn growing up and dad would ask me to go put the sprinkler on the brown parts. I could tell some patches of grass looked different from others, but often dad would be a bit bewildered by my placements. Later in life, this lead to problems with dressing professionally - it can be pretty hard to match a shirt and tie to a suit if you don't know what colors you're working with - I used to take two ties to work with me, until I learned to memorize labels, patterns, textures, etc.. Funny thing was I was working in the fashion industry and always thought I had to keep my disability secret from bosses and colleagues until my dad pointed out it was probably protected and I couldn't be discriminated against for it.

Perhaps most amusing, however, was high school biology class. Remember when you had to color in those drawings of the human organs or reptile's nervous system or what have you with the colored pencils? "Could you pass me the green pencil..." lead to a lot of brown "mistakes."
posted by allkindsoftime at 3:56 AM on August 30, 2011 [6 favorites]


Are the relationships in our brain between colors syntactical ones? If they can be described as syntactical, then would what we've seen be evidence for semantics being derived from deep syntactical structures?
posted by jwhite1979 at 4:12 AM on August 30, 2011


my dad pointed out it was probably protected and I couldn't be discriminated against for it.

You might want to check into that before coming out to your employers. A colorblind person of my acquaintance recently had a futile discussion with HR and they told him he was NOT protected.
posted by DU at 4:25 AM on August 30, 2011


On the one hand, there is nothing in the physics of light (e.g., in facts about surface spectral reflectances) that suggests drawing boundaries between colors at one place rather than another; in this sense our segmentations of the spectrum are arbitrary. On the one hand, it was well known that different languages had color terms that segmented the color spectrum at different places. So since nothing in the physics of color could determine how humans thought about color, it seemed natural to hypothesis that color cognition followed the grooves laid down by color language.

This seems ridiculous. Nothing in the physics of light? I get that there is some good argument over whether or not indigo and violent are distinct components of white light, but the rest of the visible color spectrum (red, orange, yellow, green, blue) is segmented as a simple matter of physics and nothing else.

It does not seem to stretch credibility to assume rainbows existed before language.
posted by three blind mice at 5:10 AM on August 30, 2011


Not-entirely-coherent defense of Sapir-Whorf in T-minus 3... 2... 1...
posted by stavrosthewonderchicken at 5:11 AM on August 30, 2011 [2 favorites]


I think the idea is that they only appear "segmented" that way because of the way the photoreceptors in our eyes work, ie tuned to the frequencies we call "blue," "green," and "red." But theoretically you could divide them up differently. You could also be culturally attuned to smaller differences within frequencies (like noting the difference between "salmon" and "coral" or just considering them to both be "red").
posted by rikschell at 5:18 AM on August 30, 2011 [1 favorite]


the rest of the visible color spectrum (red, orange, yellow, green, blue) is segmented as a simple matter of physics and nothing else.

That's kinda the point of the study. If there were no verbal differentiation between Yellow and Blue, how would someone describe green? If all you know are "Yellow, blue and red," it's gonna be tough.

On the other hand, while previous studies have demonstrated that understanding/comprehension of color is dependent on language, I'm still clinging to the notion that the opposite is more likely: that colors were first observed, then described (since when has language followed any other path?)

Test results be as they may, this whole line of inquiry strikes me as a case of "Eskimos have (umpteen) words for snow."
posted by ShutterBun at 5:28 AM on August 30, 2011 [1 favorite]


how would someone describe green?

See that rainbow? See those bands of light? See the fourth one from the left? That's green.
posted by three blind mice at 5:38 AM on August 30, 2011 [1 favorite]


If all you know are "Yellow, blue and red," it's gonna be tough.

"The shade of yellow or blue, whichever you want to call it, that appears exactly between the yellowest yellow and the bluest blue in a rainbow."
posted by DU at 5:40 AM on August 30, 2011


@ThreeBlindMice I know it's hard to imagine that colour doesn't exist in physics, but it's really true. What we see as colour is an interpretation of the world, just as what we hear as words are an interpretation of vibrations in the air. Many of the colours we see do no have any counterpart in the electromagnetic spectrum - there's no brown in the rainbow for instance. Also the spectrum is linear, whereas, as even you point out, we see clusters of colours - a part of the spectrum called orange and a part called red, for instance. Some of these colour clusters seem universal across cultures and all people with normal vision, and some aren't. Part of the ways that colours cluster are probably due to the antagonistic relationship of the colour receptors in our eyes (our eyes don't just "add up" the red, green and blue components of light as is commonly imagined - stimulating one colour receptor will suppress others), but as you can see in the video clip linked above it seems that other colour clusters are culturally influenced. So the rainbow, as you see it, almost certainly didn't exist outside culture. Certainly it seems as if the Himba tribe, at least, see a very different rainbow than I do.
posted by silence at 5:40 AM on August 30, 2011


You can't say "fourth band from the left" if you can/do not distinguish between yellow and green. Thus my solution which merely interpolates between two "best examples" of named colors.
posted by DU at 5:50 AM on August 30, 2011


It's a double. Same show, same researchers (Roberson, Davidoff), same Namibian tribe (the Himba), posted by a different YouTube user.
posted by nangar at 5:58 AM on August 30, 2011


I know almost nothing about how difficult it is to conduct these experiments in a tribal environment such as that one, but the video suggests that the experiments aren't (can't be?) too well-controlled.
posted by stroke_count at 6:07 AM on August 30, 2011


If there were no verbal differentiation between Yellow and Blue, how would someone describe green?

"It's blue like the leaves on a tree."
posted by Nomyte at 6:20 AM on August 30, 2011 [1 favorite]


"The shade of yellow or blue, whichever you want to call it, that appears exactly between the yellowest yellow and the bluest blue in a rainbow."

That's kinda begging the question, since you're essentially verbally directing someone to look at a particular color, and defining it for them with a new word.

Probably the best comparison I can give (to what the experimenters were trying to demonstrate) would be to describing musical genres.

Think back 20 years ago, when most (crappy) record stores had 4 or 5 categories of music. Rock, Country, Jazz, Classical, and Easy Listening (for the sake of example) A new song comes along, and your friends ask you to describe it. Given these limited categories, you're apt to lump it in with some other music that, while related, is quite a bit different than the song you're talking about.

Now fast forward a few decades, where a new culture has established a "language" which can categorize dozens (hundreds?) of different musical styles from each other. Not only can you more accurately describe different styles of music to others, but you yourself may come to understand the differences between genres, and hence comprehend the music differently.

(I'm guessing this is kinda how the term "Umami" came to be recognized.)
posted by ShutterBun at 6:28 AM on August 30, 2011 [1 favorite]


It seems far more likely to me that basic color distinctions are a pre-linguistic evolutionary survival trait. Don't eat the fruit until it turns a ripe color, otherwise you will get an upset stomach and die. As for the hypothesis that language then retroactively influences more nuanced human perception - sure, isn't that obvious?
posted by jet_manifesto at 6:31 AM on August 30, 2011


That's kinda begging the question, since you're essentially verbally directing someone to look at a particular color, and defining it for them with a new word.

How is that begging the question? The "verbal directions" were from known landmarks and the location from there is precisely given.
posted by DU at 6:40 AM on August 30, 2011


> This seems ridiculous. Nothing in the physics of light?

It's about segmentation, not whether you can see color or not. Color is a continuum. There's not much besides convention that tells us we have to pick a particular point in the continuum and say wave lengths a little shorter are "green" and ones just a bit longer are "yellow."

The Himba can see rainbows, but the way the way they'd describe it, it has three bands: serandu (red & orange). dumbu (yellow & yellow-green), and borou (blue and other shades of green that aren't dumbu). They can see the gradations, but they label them differently.
posted by nangar at 6:42 AM on August 30, 2011


Do people actually see bands in rainbows? They're continua.
posted by Jpfed at 6:43 AM on August 30, 2011


How is that begging the question?

Expecting somebody to see a "band" consisting of colors labeled as shades of green in English is probably not going to work very well, unless they know ahead of time that they're shade of green.

If you're trying to teach someone the English word "green" you can point to things that are green and "say we call this green," "this dark green," "this is light green" etc., but that's a bit different. You're not asking them to read your mind.
posted by nangar at 6:57 AM on August 30, 2011


"See that rainbow? See those bands of light? See the fourth one from the left? That's green."

HAHAHAHA. No silly, it's the 12th one from the left!
posted by iamkimiam at 6:58 AM on August 30, 2011


Metafilter:

- There are limitations to your perception of the universe due to the workings of your mind and sensors.

- That can't be! I don't percieve any limitation!!
posted by CautionToTheWind at 7:31 AM on August 30, 2011 [6 favorites]


Do people actually see bands in rainbows? They're continua.

Wow. I guess I am really out of place in this thread because light is not "a continua" but a quanta. Photons - particles of light - do not assume a continua of states, but are restricted to certain discrete, quantum values.

See: The Photoelectic Effect.

White (visible) light is composed of discrete wavelengths of light that you can separate with a prism. A rainbow does not contain different shades/hues of blue, green, etc. - it contains one and only one shade of each.
posted by three blind mice at 7:33 AM on August 30, 2011


Re the physics of light:

The light impinging on a particular patch of your retina can be described as a very quickly varying function of electric and magnetic disturbances over time- even an apparently steady color is characterized by very quickly fluctuating electromagnetic fields.

It's a pain in the ass to describe these fluctuations over time, so we'd rather describe the signal differently: as a superposition of different frequencies of fluctuation, each with an amplitude. We can imagine a function that takes a frequency and gives you the amplitude with which the electromagnetic field is fluctuating at that frequency. That function is called the "spectrum" of the light. The spectrum of a steady color does not change over time.

When the retina is exposed to a particular spectrum of light, the rods and each of the on-average three kinds of cone (called photoreceptors) have their own response to that spectrum. We characterize the response of a photoreceptor by its "absorption spectrum"- a function that takes a frequency and gives you the strength that the photoreceptor's response to a standard intensity of illumination just at that frequency. Any particular photoreceptor responds at least a little to pretty much every frequency, but each kind has different peak frequencies that they really respond more intensely to. The three cones that most people have (some people have four!) each have peaks in different areas of the spectrum. One responds to short wavelengths (which we see as blue), another to medium (which we see as green), and another to long wavelengths (which we see as red).

To get the response of a photoreceptor to a particular spectrum of light, we sum the responses the photoreceptor has to every frequency* present in the original spectrum. The response for one frequency* is the product of the light's amplitude at that frequency and the photoreceptor's responsiveness to that frequency. We add* all those products up to get the total response** of the photoreceptor to the light.

Ok, so now we have the responses of individual photoreceptors in this patch of retina. Input from each of the kinds of cone- short, medium, and long wavelength cones- are combined to give us an overall impression of the color.

Note that our cones are basically giving us information in the format of redness, greenness, and blueness. Why are most people so crappy at mixing colors in terms of varying amounts of red, green, and blue, then? Because red, green, and blue are a low-level code that we don't have direct access to. Almost immediately after our photoreceptors gather this RGB information, it is transformed by various sums and differences into a different format.

The new format is called "color opponencies". There are three color opponencies- three axes along which our thalamus and the lower areas of visual cortex process color.

Red (R - G) vs Green (G - R)
Blue (B - R - G) vs. Yellow (R + G - B)
Black (-R - G - B) vs. White (R + G + B)

The bolded R, G, and B refer to the responses of the cones sensitive to long, medium, and short wavelengths respectively.

Now, note that almost every quantity in this process up until this point has been effectively continuous. The light's amplitude at any of a continuum of frequencies can take on any value***; the cones' responses, while discrete from moment to moment****, are averaged out over time to be continuous. The color opponency values are mere sums and differences of continuous values, so they are also continuous.

If people are perceiving colors as discrete, then whatever quantizes that perception into chunks must be happening later than the early visual cortex.

*There are infinitely many frequencies forming a continuum. You can't really "add up" the response for every frequency; you add up the responses to frequency intervals, and you make the intervals smaller and smaller until by the magic of calculus you get an integral.

**This whole process of taking a sum of the products of two vectors- in this case, vectors with infinitely-many values- is called "taking the inner product", a useful concept from linear algebra

***Yes, yes, the light is technically quantized- and people can under the right conditions even observe a single photon- but if you're talking about rainbows, there are so many photons around that it's effectively continuous.

****Turns out any given photoreceptor can only respond once in a while, and it takes a while to reset itself. The "strength" of response is really about the probability of response of one receptor per unit time, which gets averaged a bit over time and other receptors' responses nearby in space.
posted by Jpfed at 7:38 AM on August 30, 2011 [10 favorites]


White (visible) light is composed of discrete wavelengths of light that you can separate with a prism. A rainbow does not contain different shades/hues of blue, green, etc. - it contains one and only one shade of each.

Either I am spectacularly ignorant about the physical world, or this is a spectacularly uncharitable construal. Are you suggesting that scattered light in a rainbow contains no more than seven distinct frequencies? Or are you making the more or less irrelevant point that light frequencies are distinct, but on a level of such fine granularity that it becomes more convenient to talk about a continuous color spectrum?
posted by Nomyte at 7:43 AM on August 30, 2011


@ThreeBlindMice - firstly, the bands in a rainbow are certainly much cruder than the quantum effect on wavelength. But more fundamentally you seem to have missed the point - wavelength!=colour.
The colour you perceive is only distantly related to the wavelength of the light entering your eye. Firtsly, there are many colours that do not have any corresponding wavelength of pure light (as I said before - there's no brown in the rainbow). Secondly, you can perceive two objects to be the same colour even if the light from them is wildly different wavelengths, and conversely you can perceive two objects as being different colours even when the light coming from them is the same. Colour happens in your head. That's why optical illusions like these are interesting.
posted by silence at 7:45 AM on August 30, 2011


three blind mice- if you perceive colors as not being on a continuum, that is unrelated to the quantization of light into photons. If you're receiving millions of photons at a time, there are still millions of physically-distinguishable levels of stimulation present in that signal- a far cry from the very few psychologically-distinguishable bands of color that people are claiming to perceive in a rainbow. If those people are accurately reporting their experience, their brains must be throwing out a lot of information- in a process that we have no reason to believe would be guided by the quantization of light.
posted by Jpfed at 7:46 AM on August 30, 2011


Thanks, Jpfed.
posted by nangar at 7:48 AM on August 30, 2011


I know it's hard to imagine that colour doesn't exist in physics, but it's really true. What we see as colour is an interpretation of the world, just as what we hear as words are an interpretation of vibrations in the air.

And physics is also interpretation of the world: All aboard the marry-go-round!
posted by quoquo at 8:08 AM on August 30, 2011


Is it possible that westerners typically use an RGB model to describe color, while the Himba are using more of a HSV/HSL model, which gives emphasis to lightness and saturation?

I'm also curious as to why the researchers in the video immediately jump to language as the cause for the Himba's inability to distinguish different colors (that I assume had similar lightness/saturation values). Couldn't it also be possible that, as a fairly isolated genetic group, they all had similar color 'blindness' and the language followed?
posted by rh at 8:09 AM on August 30, 2011


Well, and the rainbow problem really comes back down to how do we define the bands. If your language puts the frets between red, yellow, blue and purple, then saying "the 4th band" lands you at purple. Not to mention the problems with ego-centric vs. geo-spatial reference clash on the phrase "from the left". I mean, it's a rainbow! It more or less starts at some ground and goes to some point in the sky. So is it in front of the mountain? Above it? To left of it, but to the right of me? And which direction do you start counting? Maybe by fourth band, you meant red, if I even found the friggin' rainbow.
posted by iamkimiam at 8:10 AM on August 30, 2011


And don't even get me started on counting...
posted by iamkimiam at 8:11 AM on August 30, 2011


allkindsoftime: I'm colorblind. Can barely tell blues and purples apart. . . Later in life, this lead to problems with dressing professionally. . . I was working in the fashion industry and always thought I had to keep my disability secret. . .

My high school physics instructor was and is mono-chromatically colorblind. No cones.

This was during the post-Woodstock "Peacock Revolution" in men's fashion, and he had a hell of a time with all that stuff. We could tell on what days his wife worked because she wouldn't be around in the A.M. to nix his more egregious color combinations. Blues Brothers works better for him than Carnaby Street.

I know that for his army hitch back in the 1950s and for some years after, he was an electronics technician. Used to repair radios, TVs, and such. What I don't know is how the hell he dealt with the resistor color code. But as far as I know, he never blew anything up and did manage to keep it a secret from the army and most employers.
 
posted by Herodios at 8:27 AM on August 30, 2011


According to the embroidery thread manufacturers DMC, there are 464 different colours. If you ever find a cross-stitch chart with very subtle shading and colour changes, you will find there is indeed a difference between 'khaki' and 'light brown-green'. I have done the unpicking of stitches to prove it.
posted by mippy at 8:45 AM on August 30, 2011


In the previous thread nangar makes a very interesting comment:


Hey, remember ROYGBIV? There's two words for blue in there.

Last time I looked, the French Wikipedia article on Indigo was mostly a rant about the "myth of the seventh color" and how it was "invented" by Isaac Newton to "make them coincide with ... other culturally important heptads."

Somebody needs to go tell the people on Chinese Wikipedia that. They made an entire category page for it.

So if you eliminate Newton's interpolated indigo, that leaves 6 colors in the spectrum, and adding black and white makes 8.

But most of us have 3 distinct populations of cones, and the total number of combinations of 3 things just happens to be 8 (see jepfed"s excellent comment above).

I don't know whether it works to assign each color in the spectrum to a particular combination of cones, but I do find the possibility intriguing.

I'd particularly like to ask tetrachromats how many colors they see in the spectrum, because the total number of combinations of 4 things is 16. It's a great pity so few of them can talk.
posted by jamjam at 8:53 AM on August 30, 2011


Is it possible that westerners typically use an RGB model to describe color

No.

This model is emphatically not universal even in the cultural 'west', it is relatively new and isolated to representing color electronically via video. Until the 1980s (or thereabouts) most people experienced and described color in terms of pigment mixing, not video: the primary colors were red, blue, and yellow. If you poll people over 40 who don't work with video, you'll find that they still are. Only those few who worked in color television or computers would even have been aware of the RGB model back then.

The RGB color model has saturated (heh) our popular culture recently, but the map is not the landscape. It can be replaced, and probably will.
 
posted by Herodios at 8:57 AM on August 30, 2011


jamjam: the combinations of RGB work out like this:

0 0 0 : black
0 0 1 : blue
0 1 0 : green
0 1 1 : cyan
1 0 0 : red
1 0 1 : magenta
1 1 0 : yellow
1 1 1 : white

So no, it doesn't really map to the color words in ROYGBIV.
posted by Jpfed at 9:07 AM on August 30, 2011 [3 favorites]


Herodios- it's kind of amazing that Helmholtz was able to come up with the RGB model at all, given how counterintuitive it is.
posted by Jpfed at 9:08 AM on August 30, 2011


> Couldn't it also be possible that, as a fairly isolated genetic group, they all had similar color 'blindness' and the language followed?

The video doesn't mention it, but they did actually check for color blindness when they did the study. People who were color blind (either Himba or English) weren't included in the study. (That's fairly standard in this kind of research for pretty obvious reasons.)

Himba's system of classifying colors isn't actually very unusual. A lot of languages use systems of five basic colors similar to Himba's.

Debbi Roberson at the University of Essex, the lead researcher in the Himba study, has published a lot of research on this topic. A good bit of it's available in pdf in on her website.

Is it possible that westerners typically use an RGB model to describe color ...

Echoing Herodios. We don't. Web designers know RGB colors, since HTML uses them. But hardly anybody else does.
posted by nangar at 9:21 AM on August 30, 2011


I'm going to go ahead and link to the Wickuhpedia page for Berlin and Kay's Basic Color Terms again.

Most discussion of this sort end up there at some point anyway.
 
posted by Herodios at 9:50 AM on August 30, 2011 [1 favorite]


> Is it possible that westerners typically use an RGB model to describe color ...

Echoing Herodios. We don't. Web designers know RGB colors, since HTML uses them. But hardly anybody else does.


That wasn't quite my point. I was getting more that the fact that an RGB (or CMYK or whatever) system uses the hue as the primary differentiatior of a color. That's how we name colors in English at least - red, blue, yellow, orange - with modifiers to describe about saturation, intenisty, lightness - light red, dark yellow, etc. From the named examples given, it seems that the Himba people are describing colors first around lightness and saturation, which I'd describe as more like a HSV or HSL model.

The video doesn't mention it, but they did actually check for color blindness when they did the study.

In the video, it wasn't simply that they didn't have distinct words to describe two different colors of similar saturation/intensity, it was that they couldn't even point to the one that was different. That seems to me to be the very definition of color blindness, i.e. "the inability to distinguish two colors that others can perceive as different".

So perhaps they have one form of color blindness, while I have a different form. We're all unique snowflakes.
posted by rh at 10:32 AM on August 30, 2011


The video is a double from the prior thread, both taken from a recent BBC documentary (Still on iPlayer for UKers).
posted by Gordafarin at 10:49 AM on August 30, 2011


rh- I think you're using words differently than others. Color spaces like RGB and HSV don't have a "primary" way of differentiating colors; they have some number (in these cases, three) of equally-important dimensions.

Color words correspond to volumes in color spaces. We might describe those color words in terms of the boundaries of those volumes in a particular color space.

If a color word's volume's width along a particular dimension is small, then we could say that that dimension is "important" to what that color captures, or alternatively, that this color word "emphasizes" that dimension. Interpreted in this way, I disagree that RGB has any particular importance for Westerners' color words.

I think it's actually far easier to describe the volumes of English color words using HSV/HSL than RGB; with HSV/HSL, "green" for example is any color for which Hue lies between 75 and 160 degrees (and I stop it there mostly because cyan is an important color for me; others might use a different boundary), so long as the saturation is high enough not to be called "gray" and the value/luminance/whatever is between what we call "white" and "black". In RGB, "green" is this weird polyhedron that would be kind of a pain in the butt to describe (it can have a little red, but the amount of red it can have and still be green is dependent on how much green it has; it can also have a little blue; and it has to have some minimum amount of green in order to not be black).

I'm guessing what you really mean to say is that lightness and saturation are not terribly important to English color words, and on that I agree. But hue is really important. That does not at all suggest that RGB is how we're thinking about color; it rather suggests that one of the color spaces that incorporates hue (like HSV or HSL) is.
posted by Jpfed at 10:54 AM on August 30, 2011


That wasn't quite my point.

OK. Got you. Some languages do identify color mostly or entirely in terms of lightness, darkness or brightness. But this isn't really the case for Himba. The term vapa covers most pale colors as well as white, and zoozu covers most dark colors as well as black. However, the remaining basic color terms in Himba, serandu, dumbu and burou distinguish hue, basically corresponding to long medium and short wavelengths (or combinations approximating them).

... it wasn't simply that they didn't have distinct words to describe two different colors of similar saturation/intensity, it was that they couldn't even point to the one that was different. That seems to me to be the very definition of color blindness ...

You missed that works both ways. Presented with a set slightly different colors, English speakers can easily pick out the one that's different enough to be a shade of blue rather than green, Himba speakers have a hard time with this if this if they're all shades of burou. However, given a similar set of very similar color watches Himba speakers can easily pick out the one that's just yellowish enough to be a shade of dumbu and not burou. In this case, English speakers have a hard time - they're all practically identical shades of green. (The last version of this thread linked to a mock-up of this.)

It's worth noting that the color used in this example is cyan. It's really on the boarder of the colors we consider to be shades of blue. If it was a tad darker or greener, we'd call it blueish green instead of pale greenish blue. In the green set, they're also all almost exactly the same, but one of them sticks out to Himba speakers because it's dumbu.

The Himba aren't colorblind, and most of us aren't either. But apparently practice with a certain system of classifying colors does make certain distinctions more salient, though it only shows up if you're dealing with very slight differences. (The effect is a lot stronger if you're asking people to match colors from memory, something Roberson and other researchers have also looked at.)
posted by nangar at 11:55 AM on August 30, 2011


In the video, it wasn't simply that they didn't have distinct words to describe two different colors of similar saturation/intensity, it was that they couldn't even point to the one that was different. That seems to me to be the very definition of color blindness, i.e. "the inability to distinguish two colors that others can perceive as different".

The point of the study is that language partly establishes the ability to distinguish colors. Based on what I've seen studying language, I believe it. For instance, Japanese infants readily distinguish the difference between English L and English R, but by adulthood many Japanese speakers cannot hear it. Likewise, many adult English speakers cannot hear the difference between Japanese "su" and "tsu". Both phoneme pairs are "sounds that others can perceive as different" -- no native to either language would confuse them -- but if the distinction does not exist in a given adult's language, they may initially have trouble even hearing them as different sounds, much less with reproducing the difference. Likewise, some objects which are "green" in English are "blue" in Japanese, because the language originally had just one word for blue/green, just as many languages do. Traditionally, green is a shade of blue in Japanese, just as pink is a shade of red in English. This makes blue a larger category in Japanese than it is in English, and certain classes of objects (mainly vegetation) are still described as "blue" even though the same color anywhere else is "green". "A blue stoplight" is as correct as "a blue sky". To an English speaker, this is just plain wrong, if not an indicator of color-blindness; to a Japanese speaker a "green light" is also wrong ("yes, it looks green but we say blue").

I wouldn't say that either group of people have any form of "phoneme blindness" or "color blindness", though. How can you be blind to a distinction that doesn't exist in your culture? Are we all color-blind with respect to serandu, dumbu, and borou? And if so, why should other-culture-color-blindness matter, since everyone on Earth has it?

In short: we've all been trained since early childhood to distinguish colors, sounds, and other input in certain ways, and not in others. Distinctions that matter change the brain... and that means that all people have "the inability to distinguish two colors that others can perceive as different", once you find "others" who are far enough outside their own culture.
posted by vorfeed at 12:02 PM on August 30, 2011


vorfeed, I'm no expert on the science of the eye/brain, but as a Japanese speaker, I can tell you that yes, the "go" traffic signal is "blue", linguistically, but everyone knows that it's actually green. I don't think is a question of perception, but rather accepted naming convention. It's like in English, people say...I can't think of a good example with color, but "hang up the phone". Now we don't actually imagine 'hanging up' the phone anymore, but we still use the term. Historically, we used to do that, hang it on a hook on the wall, but now the term just means 'end a phone call'. Similarly, historically, the signal lights were blue, and the term remains, but no one actually sees them as blue. They are green.

As anecdotal evidence, today I asked six people what color the "go" signal is...all said "blue". I then said, "linguistically, or visually?" Every single one immediately replied, "oh, of course visually, it's green."

So anyway...although I have no doubt that language can affect the way people talk about their perceptions, I'm highly skeptical that language alters the actual perception in the brain.
posted by jet_manifesto at 6:06 AM on August 31, 2011 [1 favorite]


Even in English, we call traffic lights green, when most of them are really cyan!

Note that depictions of traffic lights show them as green, but most photos show them as cyan.
posted by Jpfed at 7:03 AM on August 31, 2011


Single Yink TouLube!
posted by kcds at 7:34 AM on August 31, 2011


vorfeed, I'm no expert on the science of the eye/brain, but as a Japanese speaker, I can tell you that yes, the "go" traffic signal is "blue", linguistically, but everyone knows that it's actually green.

Yes. Like I clearly said: "yes, it looks green but we say blue". My point was that the linguistic boundaries for "blue" are different in Japanese than they are in English, not that Japanese people can't see green. Of course, the Himba can also "see green", just as we can "see dumbu"... we simply have no specific word for it, just as the Japanese didn't have one for green hundreds of years ago.

Studies like this one suggest that linguistics matter, though. No non-colorblind English speaker would ever describe a frog or a lawn as "blue" when asked to describe the color they saw, yet roughly 20% of Japanese respondents did so. The study concludes that "the data clearly indicates that the perceived level of suitability of these terms ["ao" and "midori"] is considerably referent dependent and that, furthermore, for certain referents (such as no. 31, cheese) vastly different choices of descriptors can be employed". This is not true in English -- English color words rarely change depending on the object being referenced. Similarly, colors have connotations in Japanese ("an association between the concepts of being ‘kirei' (attractive/appealing) and being ‘ao’", "it is generally referents which have life for which the connotative use of ‘ao’ as a descriptor of green things is permitted") which do not exist in English, and which seem to influence the color descriptions used. Note that not all of the native Japanese speakers in this study made a distinction between calling streetlights blue and their actual color, either -- even though this was mentioned, a significant portion of people claimed that the light appeared blue (or "both blue and green") rather than green.

As for the idea that language doesn't alter actual perception in the brain -- did you read the paper I linked to earlier? There's plenty of evidence that language does alter the actual perception of phonemes in the brain. I don't think color perception is much more of a stretch.
posted by vorfeed at 2:08 PM on August 31, 2011


« Older Adam West had the right to remain silent ... a...   |   The Gospel of You Newer »


This thread has been archived and is closed to new comments



Post