4 ms·
Nope. Wavelength is not a useful proxy for color, because the way we detect color is based on intensity differences between different light detector responses,
by jacobolus 11y ago
Nope. Wavelength is not a useful proxy for color, because the way we detect color is based on intensity differences between different light detector responses, rather than wavelength per se. Trying to compute a specific “wavelength” value for a particular blue or green color is misguided. As Newton wrote in 1730, “For the Rays to speak properly are not coloured.” That is, color is a perceptual effect, created in the eye and brain, not an inherent property of objects or light.
In the human retina, the responses from the 3 types of cone cells (daylight light detectors) are immediately combined into a lightness signal (M + L), and two color signals, “red–green” (L – M) and “yellow–blue” (L + M – S), where {L, M, S} are the long-wavelength-, medium-wavelength-, and short-wavelength-sensitive cone cells, respectively. The basic L/M/S signals never make it past this first level of processing in the eye; all the brain sees are the combined signals.
The perceptual geometry of human color vision therefore has a built-in notion of red, yellow, green, and blue as “unique hues”. There’s a fair amount of variance from person to person, but anyone with trichromatic vision will be able to pick out some particular hue of “blue with no green or red in it”, “green with no blue or yellow in it”, etc.
You’re right that some languages apparently didn’t have any particular need to distinguish blue from green. I could believe that the introduction of blue dyes made this distinction useful. There’s been a lot of study of the topic of color naming, but I’m not sure there’s a definitive answer on that point.
- ececconi 11y agoPretty cool. I was going by a book I read a couple years ago, but could be totally wrong. What you mentioned is very interesting and seems to contradict their findings.