Recently there was a schism on the internet between people about the color of a dress.  That led to all kinds of scientific articles about how we perceive color differently, memes about white and gold things and black and blue things, and finally, eventually, this.

Screen Shot 2015-03-02 at 4.13.41 PM

That is from an article on LinkedIn, written by a woman named Diana Derval, who claims to be an expert in neuromarketing, whatever the fuck that is.  The title of the article looks like this:

Nothing says legitimate scientific knowledge like a winky face with its tongue out.
Nothing says legitimate scientific knowledge like a winky face with its tongue out.

This is already almost a hundred percent incorrect, but in order to explain why, I need to give you a little anatomy lesson.

Vision starts in the eye.  There are three sets of cells in the eye called “cones” and one set called “rods.”  Rods only have one kind of light-sensitive pigment in them, which means they can only tell how much light is coming in, not what color it is.  They are far more sensitive than cone cells and are almost entirely responsible for low-light vision, but have little to no role in color vision.

The majority of people have three cones, called L, M, and S for the long-, medium-, and short-wavelength light they detect.  After the pigments pick up light, they are sent to the brain along three channels, one for each color.  L corresponds with red, M with green, and S with blue.

Roughly one in sixteen men is what’s called red-green colorblind, which is a slightly misleading term.  The correct term is anomalous trichromacy, meaning they have two fully functioning sets of cone cells instead of three.  The S (blue) set is fine, but either the M (green) has sensitivity shifted toward the red portion of the spectrum, or the L (red) has shifted toward the green.  Importantly, though, the brain does not know that this has happened.  This is a genetic condition that affects the eyes, but not the color-sensing portion of the brain.  The brain assumes that each cone is sending it the correct color and builds images accordingly.

I, for example, am deuteranomalous.  I have a perfectly functional set of S cones and a perfectly functional set of L cones, but my M cones are shifted toward the L end of the spectrum.  This means, theoretically, that I am less sensitive to green light than a person with normal vision, but I can’t tell.  As far as my brain is concerned, the signals are coming through fine.

Here’s an example.  Imagine a gray square, composed of equal parts blue, red, and green light.  Then you turn up the red and blue light, making the gray square a sort of dull magenta color.  To you, that square is now magenta. My stupid deformed M cones, however, are detecting red when they shouldn’t be, so they detect the increase in red light as well.  They report back to the brain that the green levels have gone up, when they haven’t. My brain is now getting signals that all three channels of light have increased in magnitude and the square is now a brighter shade of gray.  It’s not.  It’s pink.  But I can’t tell.  Want to see that in action?

Screen Shot 2015-03-02 at 5.04.32 PM

This is a graphic to test my particular flavor of color blindness.  My coworker assures me that the sky inside the circle is pink, but I can’t tell because my dumbfuck M cells think that the increase in red and blue is an increase in all three colors, which cancels out.  I can tell that it’s not exactly the same as the other sky, but it’s more of a texture than a color.  She tells me that the grass in the circle is yellower (because red light was added to the existing green), but I can’t tell for the same reason.  The gist of it is that if something is pure green, it looks paler to me.  If you add red to something, I can’t tell.  Dark blue and purple are a nightmare. Traffic lights look very pale, almost blue.  Dull greens look brown because I can’t see the green part.  And so on.

This brings me back around to tetrachromacy, or the presence of four sets of cone cells.  A Dutch researcher in the 1940s noticed that the mothers and daughters of deuteranomalous men like myself all had normal color vision.  He knew that the genes responsible for cone cells came from sex chromosomes, which left two possible explanations.  If the mutated M cells came exclusively from the father, all fathers and sons of deuteranomalous men would have the same condition, which wasn’t the case.  If they came in equal part from the mother, then deuteranomaly would be similarly present in women, which wasn’t the case.  He concluded, therefore, that the mothers and daughters of deuteranomalous men must have a fourth set of cells, giving them three functional ones and one mutant.  He hypothesized that women with four functional sets of cells might exist, but it wasn’t the point of his research so he didn’t look into it.

This has been a long wall of text. Here is a cat being friends with a horse.

Fast forward to 1980, when two researchers became intrigued by the idea of four-coned women.  They knew that anomalous trichromacy was common, which meant that four-coned women must be common as well.  They sought out the mothers and daughters of colorblind men and had them take a color matching test.  In such a test, the subject mixes levels of red and green light to match the yellow light provided.  Colorblind men will have to add more of either red or green to compensate for their defective cones, and people with normal vision will be able to match the colors correctly.  People with four cones, theoretically, would be able to tell the difference between true yellow light and light made by mixing red and green, and would therefore be unable to make a match.  That wasn’t the case.  The researchers found plenty of women with four sets of cones, but none of them had more sensitive color vision than the average trichromat.

In 2007, one of the researchers tried a different technique.  She flashed three colored circles in front of her subjects’ eyes.  A trichromat would have been unable to tell them apart, but a tetrachromat should have been able to recognize that one of the circles was actually a very subtle mix of red and green, rather than a solid yellow color.  Only one woman was able to pass the test.  Which brings me to my point (1100 words later):

If two researchers who dedicated their careers to the task were only able to find one functional tetrachromat in 27 years, do you really think that a test on LinkedIn written by a marketing professor is going to help?

Obviously, the answer is no.  But there’s more bullshit here.  First, the title.

25% of people are tetrachromats

Lies.  It’s something like 12% of women, which is 6% of people — and is probably more common in women of Northern European ancestry, so that number’s even lower worldwide — and it’s so rare that women with four cells can actually use them that we can’t even put a number to it.  Only two women in history have ever been empirically confirmed as functional tetrachromats.

and see colors as they are

That’s a preposterous thing to say.  Everyone’s cones see slightly differently already due to genetic variation, so theoretically the same wavelength of light looks infinitesimally different to every single person.  The only reason color blindness is a thing is that color blind people can’t distinguish between certain colors, not that they’re seeing them wrong.  Sure, you can empirically say that a certain LED bulb emits light at a wavelength of 581nm, but what does that look like?  No one can really say for sure.  There is no such thing as colors “as they are.”

You see less than 20 color nuances: you are a dichromats, like dogs, which means you have 2 types of cones only. You are likely to wear black, beige, and blue. 25% of the population is dichromat.

Horseshit.  Dichromacy affects less than 3% of males and .03% of women.  That’s about 1.5% of the general population.

You see between 33 and 39 colors: you are a tetrachromat, like bees

No part of that is true.  Firstly, you can’t diagnose tetrachromacy on a computer screen AT ALL because computer screens are made up of combinations of only three different colors of light.  It is literally not possible for an LED computer screen to generate the kind of nuance that distinguishes tetrachromats from trichromats.  Secondly, bees see in ultraviolet, meaning their extra color vision is in a wavelength that no human being (or even mammal)* has ever seen.  Being a tetrachromat in the visible spectrum does not mean you can see what bees see.  And thirdly, BEES ARE NOT TETRACHROMATS. Bees are trichromats, with cones in what we might call the green-yellow, blue, and UV portions of the spectrum.  They still only have three cones.

It is highly probable that people who have an additional 4th cone do not get tricked by blue/black or white/gold dresses, no matter the background light

Die in a multicolored fire.  Let me say this one more time: that dress is a photo on a computer screen, taken on a digital camera.  The screen on your computer is only capable of generating three wavelengths of light, and all others are projected as mixtures of those three.  The sensor in your camera only records three wavelengths of light (because that’s what you see in), and all others are a mixture of those three.  ANY COLOR IN THE WORLD THAT IS NOT A SPECIFIC WAVELENGTH OF RED, GREEN, OR BLUE is shoehorned into a combination of those three by your eyes, your brain, your camera, and your screen.

This graphic is like testing your depth perception with one eye closed***.  It is fundamentally impossible.  It is stupid, insulting, and worst of all, it is popular.  Stop it immediately.

*[EDIT 03/10: Turns out a whole shitload of mammals can see in UV, but humans are not among them.]**
**[EDIT OF THE EDIT 03/12: Turns out that some humans can see in UV (a condition called aphakia) some of the time.  UV is blocked by the lens, so if the lens is fucked up due to surgical removal, a perforating wound or ulcer, or congenital anomaly, UV light can hit the retina.  It is thought that Claude Monet was aphakic because he painted flowers in a way that a person with normal vision shouldn’t have been able to see.]
***[EDIT 03/12: It is possible to percieve depth to a small extent with one eye based on perspective, size comparison, motion, and the flexing of the lens in the eye to adjust focal length.  TESTING depth perception using one eye would still be dumb.]

174 Thoughts

  1. you are wrong!!! the 3rd cone person sees less than 30 colors you know? you are a “borderline tetrachromat!”

  2. I was tested at UMKC. They gave me two different color mixtures and asked me which one contained more red, more blue, more green. Even to the researchers, the colors appeared the same and had they not mixed them, they wouldn’t have known the difference. Out of 24 barrels of color mixtures, I accurately distinguished between all of them. I am male therefore, genetically, I couldn’t have tetrachromacy, but decided to test for a fourth cone anyways… No surprise, I only have 3 in both eyes, but why am I able to see the different colors you would expect only a tetrachromat to see? Nobody can explain why I see color and light the way I do without a 4th cone, please help.
    P.S. I see color the same in the dark, I thought it was normal until now… What about this?

    1. You may have three colour sensors but they may not be in the normal range. That would give you different discrimination, so you would expect to be better at some parts of the spectrum than a typical male but you could be worse at others. It is theoretically possible to determine the exact colour sensitivities you have but in practice, as far as I am aware, it is not offered. It would involve analysing your metamers; it is clear from what you say that your metamers are very different to those of the researchers.

      Perceiving colour in the dark seems to be unrelated. The cones lose their sensitivity at low light levels and you rely on rods, so your vision should become monochromatic.

      Unless (stupid idea) your cones are somehow providing a fourth colour in the blue-green range and you are perceiving it as a colour for some unimaginable reason which does not make sense. Then you might perceive that single colour at night. But that’s pretty far fetched. Cones saturate in daylight so it’s really unlikely.

      You need more lab work.

    1. Absolutely incorrect. An RGB display can SIMULATE any color by combining the light from three sub-pixels. The exact color gamut of each sub-pixel varies, but an RGB displaay cannot produce “true” yellow light, for example. It combines red and green to do so. You can even open a yellow image on your phone and then put a drop of water on the screen, which magnifies the sub-pixel structure, and clearly see that the color is only red and green.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s