As both dmd and rp began to answer this then balked at it (then came back and had another go... they're gone again now), I am chary of attempting it myself; but perhaps unlike them, I don't feel I need any detailed knowledge of rods and cones, of how the optic nerves and the visual cortex work. (Gilbert Ryle said this better, but I can't find the quote.)

This node contains two topics. (i) whether colours are real; (ii) whether people's actual sensations of colours are different. I'll answer them separately.

What on earth does it mean to say colour (or a colour) "exists"? Or "doesn't exist?" What on earth are you trying to say about it? Unicorns don't exist—if you search the world you won't find any. The lost poems of Gallus don't exist—they used to but the manuscripts have been destroyed by the ravages of time. Phlogiston doesn't exist—it was a hypothetical substance, in a theory no longer held. Here are three good examples where we say X doesn't exist—and 'exist' in each case has a different sense. Unicorns don't exist and dodos don't either, but in different ways.

How can you apply this to colours? In science fiction, you might describe an alien race as seeing in a different part of the spectrum—but no, those wavelengths exist too, even if we can't see them. You wouldn't say the bee-colours the fiction-writer describes "don't really exist". What doesn't exist is the giant robot bee that can see them.

A colour intermediate between red and green perhaps? Other than yellow...? A colour that's simultaneously red and green all over? That can't exist. That's what set Wittgenstein back onto philosophy. But here we're screwing up our minds trying to imagine queer non-standard colours so that we can say they don't exist, i.e. that you couldn't ever really see such a colour.

Unlike the ones you can see: red, green, blue, white, fawn, ... It simply makes no sense whatsoever to look at something you can see and say "I can't see that". If it's blue, and you're not colour-blind, and you've got enough daylight, and your head isn't in a photographer's black hood (etc.), you can see it. It's there. It's blue.

But "blue exists"? "blue doesn't exist"? Both nonsense. "Exist" isn't something you say about colours.

The second question is of the existence of qualia (Latin plural of quale "what sort of?"). Some philosophers believe that as well as everything that goes on in the brain—all the photons hitting rhodopsin in the retina, the firings of the optic nerves, the activation of different regions of cortex to process the incoming message "something blue out there"—in addition to all this, they think, there is something extra: qualia: "what blue looks like"? "How it feels to see blue"?

So, according to them, a brain could undergo all the pings and bloops that indicate "blue" but somewhere deep inside a little detached Cartesian spirit within the brain, some undetectable something lights up as "yellow".


We concede that this curiously-wired-up person reacts to colours in "normal" ways (soothed by green, hot and bothered by red, uses the same noise red to point at the same flowers, fire-engines, blood, and so on that the rest of us do). Note that colour-blind persons don't count: they behave palpably differently (e.g. on tests). So what is there left? If their pulse goes up in a bright red room, and mine does too, if their threshold of perception is the same (say it's easier to detect a very faint red than a very faint yellow)... concede all the physiological inwardness, all the brain states, and there's nothing left. (Technically this line I'm rubbishing is called the inverse spectrum argument.)

Of course, if they see it differently and we can detect a different brain state by external means (ECG or fMRI) then we're just in the realm of experimental physiology. There could well be such non-standard wirings. But they don't involve qualia. There are no such things as qualia.

What on earth could it mean to say someone "saw" something differently when everything detectable is the same? What would you be trying to say about them? Nothing.