How many distinct colors would tetrachromatic birds have?

Question

I'm creating a world where the primary conscious species are essentially just ravens. Birds, in general, have tetrachromatic sight (extending in the UV range), meaning they see 4 base colors. My question is, how many of these colors would be distinct? Obviously it would depend on the language (languages in our world have different base colors), but in general how many colors would they consider exist?

I want to start with combinations of the four base colors, but even doing that with the three base colors of humans, it still isn't clear. Like, just doing combinations where a base color is fully on or fully off, you get some colors like #0FF ("turquoise") which in English is usually just considered a shade of blue. This reveals a question under that question, namely: what colors could reasonably be seen as "shades" of another color?

In short, I'm mainly just asking what colors (or words for them) would be commonly used. Of course they would be able to see millions of different colors, but they would only distinguish between a few ranges of them, so I'm mainly asking what ranges are reasonable?

Answer 1

Insufficient data. The number of minimally-distinguishable colors that a creature can distinguish is totally separate from the dimensionality of their color space. Humans can distinguish tens of millions of colors; birds can do the same, and if having a fourth color receptor type makes that number larger... well, that doesn't really have any practical consequence. The number is already so large that it might as well be infinite.

However, they can see completely different kinds of colors. See Ways of Coloring: Comparative Color Vision as a Case Study for Cognitive Science, Describing Non-human Vision, and What Are Tetrachromatic Colors Like?.

Exactly how the colors are organized would depend on the relative overlap in response curves between each of their four color receptor types, but they would have four physical primaries and 6 perceptual primary hues corresponding to the endpoints of 3 opponent axes output by retinal preprocessing. We can generically label those the R-G axis, the Y-B axis, and the P-Q axis.

Binary combinations of those basic hues with their non-opponents results in 12 maximally-distinct secondary colors (R+Y, R+B, R+P, R+Q, G+Y, G+B, G+P, G+Q, Y+P, Y+Q, B+P and B+Q).

So far, that is analogous to the structure of trichromatic vision, just with more boxes, but ternary combinations of non-opponent primary hues produce 8 extremal instances of an entirely new kind of hue--tertiary colors--not found in the perceptual structure of trichromatic color space (R+Y+P, R+Y+Q, R+B+P, R+B+Q, G+Y+P, G+yY+Q,G+B+P, G+B+Q). Additionally, there is not merely one non-spectral secondary color (magenta) in the fully-saturated hue space, but 3--and in general, that number will correspond to however many pairs of non-spectrally-adjacent sensor types there are (which actually works out to the sequence of triangular numbers!)

If we assume without loss of generality that R, G, B, and Q are the physiological primaries (note that the spectral locations of y and p depend on the decorrelation output for a specific set of 4 receptors with species-specific sensitivities), then the non-spectral secondaries are R+B, R+Q, and G+Q. All of the tertiary colors are non-spectral.

That covers all of the maximally-saturated hues, which exist on a color sphere rather than a color wheel. Those hues can still be modified by varying levels of saturation and luminosity, just as in trichromatic vision.

How those 26 basic hues would grouped into named categories is entirely dependent on language and culture.

Answer 2

It isn't clear, is it?

The Munsell colour atlas has 330 colours. These are not individually named colours, though they do have a unique code for each. But they are a fair representation of the colours we can see as being significantly different from the others.

A crude approach might be to say 330 is almost 343 which is 7x7x7. We can see about 7 shades in RGB if the space was cubic. It isn't - the Munsell space sticks out in the yellow and blue directions - but let's go with this for now. We might guess a tetrachromat might have roughly 7x7x7x7 = 2401 colours to a first approximation.

Is this right? A mantis shrimp has 16 separate channels from IR to UV. The daphnia water-flea has 48. Do they see 7^14 and 7^48 colours? This is unlikely. The mantis shrimp is probably trying to see camouflaged prey, and the exact nature of the mismatch is not important, while the water-flea is probably looking for the ideal yellow-green water where food is.

If you were a tetrachromat human - and there are people who claim to be just that - then this would make it almost impossible to match a colour using pigments designed for trichromatic. Instead of making you a super-artist, it would probably make matching a hue much harder, particularly if you were using pigments designed for tetrachromats.

If we had our extra channel in the infra-red, then we would see few new colours. Most things that reflect or absorb red do the same for infra-red. Kodak Aerochrome film was sensitive to green, red, and infra-red but prints only show a limited palette. There are no bright greens in the print corresponding to a sharp peak in the red.

If we have the extra channel in the ultra-violet then there may be more shades to see. Flowers have patterns in the UV that insects use. There may be more real colours in the infra-red.

Suppose the tetrachromat space has possibly 7x7x7x7 Munsell hues. However, we find it harder to match colours as a tetrachromat, so maybe the number of importantly different colours is only 5x5x5x5 = 625. That would be twice as many colours as ourselves. We would be able to see the stripes on a flower that the insects see. We would not see thousands of extra colours.

Is this reasonable? We do have four vision channels, if we include the night vision rod cells. There are some vision modes, such as mezopic vision that do integrate the cones and the rods. And the rod peak does fit nicely between our blue and our green sensitivity peaks. But if we have normal vision, our vision has never taken advantage of this. Our red channel allows us to see tigers in a forest, but their camouflage works well for deer which do not have this extra channel. We have never adapted to use the blue-green peak that our rods already have, which suggests there is little advantage to our survival to do so.

My guess is your tetrachromats will be better at spotting camouflage, but they will be less able to match colours. They may see the same number of salient colours as we do, or twice as many, or some figure in between.

It's a guess, but I hope it is not a bad one.

Answer 3

More than us but that is all you can say

the exact line between colors are largely cultural, as is how many colors there are. But although the number is cultural the pattern of which colors they have follows a fairly consistent pattern in humans, black and white, then red then the colors common in nature then rarer colors. Birds will have a much larger variety of colors even if the exact pattern is difficult to predict. To them nature has more colors so in general they should have more colors than us even in names. But there is no way to give exact numbers because how many colors get names in cultural and historical.

Answer 4

As many as they need.

Other answers have pointed out that the number of "basic" colours are very much cultural. The reason for this is that people have had different needs. If someone don't need to tell the difference between blue and turquoise, then they will lump them together as "basically the same colour".

If, on the other hand, there is a food crop that is blue, and a common poisonous plant that is turquoise, they you can be sure these are considered completely different colours.

So with the ravens. If they need colours, they will distinguish colours, if they don't they won't.

Tetrachromaticacy doesn't really enter into it.

Answer 5

Artists (as opposed to physicists), tend to describe three primary colours red, yellow, and blue. Equal mixtures of these produce three secondary colours - orange (RY), green (YB), and purple (BR). Add black and white and that provides enough colours for a simple description of many things (for example if you are describing a car then red, blue, green, yellow, black, white, and, occasionally, orange or purple would usually be sufficient for descriptve puroposes).

A tetrachromic artist might similarly describe four primary colours (W,X,Y and Z), which can be mixed in pairs to produce six secondary colours (WX, WY, WZ, XY, XZ, and YZ), and mixed in triples to produce four 'tertiary' colours (WXY, WXZ, WYZ, and XYZ). Thus a total of 4+6+4=14 colours. So 14 simple descriptive colours plus black and white might be sufficient.

Of course in practice, even us trichromics will further subdivide into pink and brown and teal and beige and gray etc., to describe the million odd shades we can distinguish. But with too much subdivision people start to disagree what colour something actually is. Also, the languages of some societies dont have simple words that distinguish six colours - for example they use the same word for blue and green, so your ravens might forgo the tertiary shades and just rely on 10 descriptors.