Thanks for thinking of me! I've just finished writing out a few thoughts which I'll repost here: "Most people would probably read Kevin Loria's provocative title as claiming that human colour vision changed in modern times to enable the perception of blue for the first time, which would be quite mad (our blue - yellow opponent signal is as old as colour vision itself!). But up until he starts talking about the Himba claims (see previous post), Loria is only saying that most ancient languages didn't have a term for "blue", which is partly right. Ancient Greek does have a term for dark blue - "kuanos" or "kyanos", and classical Latin adopted "cyaneus" from Greek and added other blues including "caeruleus" (sky blue) and "lividus" (bruise-coloured or "black and blue"). But like many ancient and medieval colour terms, these words refer to a combination of hue, lightness and/or colour-strength properties rather than hue alone. So it's true that Greek and Latin don't have a general term covering all blues, just as we don't have a general English name covering all colours of an orange hue. Looking at my polished pine desk I see it as orange in hue because I am used to thinking in terms of the Munsell dimensions of hue, value and chroma, but most people would call it brown and would have to think for a while to decide what Munsell hue it is. I expect that, not being clear on the concept of hue, a Roman might similarly have had to think for a while to see that cyaneus, caeruleus and lividus had something in common, but that doesn't mean that he saw these colours differently in the usual sense of the word. "

Thanks for the link to the critique of the Himba experiments! As this recent discussion points out, the experiments to date have still not been published in any form: http://languagelog.ldc.upenn.edu/nll/?p=17970

Yes, as indicated below, insects see ultraviolet. At the other extreme marine mammals have only one cone type, I think.

A small minority of women are have four cone types, the usual L, M and S, plus a polymorphic variation of either the L or the M, but this doesn't necessarily mean that the retina is wired up to take advantage of the extra cone type. A recent investigation of 24 such women found that 23 had normal colour vision but one saw differences between pairs of stimuli that could be matched by all other observers.

Well, that's very exciting!

Perhaps length would be at the objective extreme, but even there we at least sometimes need to make a distinction between physical and perceived length, for example in the T illusion.

I see the gap between spectral distribution and colour as being at the other extreme, in that there are two gaps of subjectivity involved. With just three cone cell types we can't know the spectral distribution of a light, only some limited qualities of it (dominant wavelength and purity as defined by an opponent visual system), so many different spectral distributions look identical to us. And then these limited qualities that we do detect are tagged for us with entirely subjective experiences whose relationships have no basis in the linear spectrum.

In between you have perceptions that correlate more directly (if non-linearly) with the physical stimulus, such as sound and vibration rate, or warmth and molecular motion.

No, not a lurker, in fact I'm not really sure where I am yet! I have a Google Alert set for "Color Wheel" that brought me here.

I think there's quite a lot of fiction in the quote below about Newton, but it's getting too late here for a detailed response just now. I have a few of my own thoughts about Newton's circle here: http://www.huevaluechroma.com/071.php#Newton_s_hue_system

The answer is physiological. According to the zone theory of colour vision, colour does not reside in the linear spectrum but is created by the visual system in the form of two opponent signals, red vs green and yellow vs blue. With their four possible combinations these create a circular range of hues: red, red and yellow, yellow, yellow and green, green, green and blue, blue, blue and red. When all wavelengths are present the red vs green and yellow vs blue opponent signals both cancel out, and we see the light as white: http://www.huevaluechroma.com/036.php