By that logic, white, black, grey or brown or pink does not exist either. The article confuses human color perception with wavelengths. It is not that purple doesn't exist, it is just that a rainbow does not contain all perceptible colors because it only contains colors represented by a single wavelength. Just like our ears can perceive chords as different from a single "average" note, our eyes can perceive combination of multiple wavelengths.
It doesn't require any pop-psychological explanation - you can easily make an electronic device which detect the color purple. It just needs to be sensitive to more than a single wavelength.
> Whilst we can see violet and blue next to each other, there's no purple. Because purple doesn't exist.
Back in reality, "purple" and "violet" are synonymous.
The article also makes the completely unsupported assumption that the color you "should" be seeing when you look at a mixture of wavelengths is the color of the average wavelength of that mixture. That's how it decides that purple "should be" green. But that's nonsense.
> When you see red and green for example, the L and M cones both fire, and your brain interprets the result as something near both. And what's near both is yellow. So even though there's no yellow light entering your eye, your brain imagines the result as yellow, because both cones are being stimulated in similar amounts.
This gets the causality backwards. When you see red and green, your L and M cones both fire. And you interpret that as yellow, because yellow is what you perceive for that pattern of cone activation. To the brain, cone activation patterns are all there is. The question "what single wavelength would best approximate this activation pattern" is not even part of the model the brain is working with. But, obviously, if there is a single wavelength that approximates the activation pattern of a mixture, you'll perceive those two things as being similarly colored. Is it necessary for the single wavelength to be the arithmetic average of the wavelengths in the mixture? No.
Except when they’re not. People who have to think and talk about color on a regular basis, such as artists, sort colors into much more detail than people who don’t.
Have a look at the Wikipedia pages for purple (https://en.wikipedia.org/wiki/Purple) and violet (https://en.wikipedia.org/wiki/Violet_(color)). The image examples given for purple range widely, from a photo of some grapes I would call more blue than purple, to a dress that I would call magenta at best. And a color swatch that is somewhere in the middle of that range.
Meanwhile the page for violet leads with a specific wavelength, and a set of photos that are all very close to the swatch above them. This particular color could fit into the range of colors given as examples of “purple”, towards the bluish end of the range.
"Purple" is perhaps ambiguous, but it is true that magenta (a reddish purple) does not exist in the rainbow. But you are correct about the rest; the article is bs.
No, when you see purple you are not seeing something your brain "thinks should be green".
It's true that purple is a non-spectral colour. The rest of what that article says about purple is nonsense. (Perhaps motivated by the analogy they want to make with what they say about "customer experiences"?)
https://toughcompetent.com/blog/purple-doesnt-exist/