And if you are talking about oil Paints, for example The primaries are red/yellow/blue though ideally in practice you have two versions of each primary - one “warm” - one “cool”
If you’re asking how you could do this from the audience with available materials: a prism might not be at hand, but a diffraction grating (common in novelties and some optics). a hologram on a credit card, or it you are really skillful, create an infinitesmal slit between two edges. If you wear standard bifocals, or have an SLR camera with split right focusing, you may be able to orient them to generate a fringe rainbow at the interface between lenses.
But why, dammit? 
I understand the subtractive CMYK printing process - but if you’re talking oil paints, and worried about the tones, I’m sure that this must be accurate (although, of course, a mixed green for a painting wouldn’t rival the vibrance of a genuine pigment green.) Why is it seemingly different for oil paints, and what’s the process that explains this?
Anyone with any insight on how the myth started? I’m gonna keep on with my hypothesis that we get taught this in schools but it’s really a bluish red and a greenish blue that just sorta approximate magenta and cyan.
(The reds were always a little bluish.)
Actually, if you’re wondering why there are three primary colors, the answer can’t be found in physics. It’s physiological. Our eyes have two types of receptos: rods for brightness information and cones for color information. The cones, in turn come in three 'flavors’each with a color receptor pigment with rather broad color response.
One pigment detects blue most strongly with a peak response at ca. 420nm, and then falls off rapidly. The other two actually have minipeaks that make them as responsive to the blue peak as the blue pigment, but they respond more strongly to other frequencies. All three allow us to see blue light (useful for vision) but we rely on the blue receptor to help us distinguish the blue element of “color” as we see it.
Our ‘red’ receptor has a much broader response than our blue pigment, and responds well to basically the entire visual range of colors. In fact the true peak of this gently curving response curve is 558 nm - which is more like a orange or yellow. Why? Probably Because we live under a yellow sun, so perceiving reflected sunlight is probably as useful to us as perceiving bright reds.
Our green receptor pigment also has a broad response, but not as broad as our ‘red’ receptor. It covers “the blue of the sky” and “the green of plants” very well, and the yellow of the sun pretty well. In fact, it matches general natural illumination on Earth - and the response curve of plant chlorophyll- pretty well in general
Calculating color from these three basic receptors (and brightness information from our retinal rods") is complex, but fortunately we’re hardwired to do it, so we don’t have to strain our brains. Some rare women (it is believed they are all women, for genetic reasons), called “tetrachromats” have a fourth color pigment covering the normal visual range, and therefore glean more information about the ‘real’ color of an object. The field is fairly new - there may evenbe different kinds of tetrachromats.
Other species have different or additional pigments. For example, many birds (and yes, bees) have four or even five pigments, which let them see more colors into the UV range, and also have a superior ability to distinguish colors in our visible range.
Yes, they misled us about color perception in grade school, but they misled us in physics (and most college courses) as well.
What about the color mixing systems in intelligent lights such as this? It is a CMY color mixing system, but is it additive, or subtractive? Or is it different because it uses five different filters?
Andrew
Hmm, that’s interesting. My wife and I sometimes get into arguments over what color something is, most notably our cat Zen. I say she’s brown, a.l. says she’s gray. Zen isn’t very forthcoming on the matter; she merely claims that we mere humans cannot fathom the mysteries of catdom.
I remember the red, blue, and yellow my teachers claimed to be primary colors. Blue and yellow made a decent green, and yellow and red made an ok orange, but red and blue made a purplish brown that was more brown than purple. My teacher did not like questions about this and blamed the way we mixed paint. She implied that if we practiced we could get better at it. More telling than the muddy purple was that the purple and yellow made a brown or a fushia depending on whether you mixed the paints or painted one over the other. There are oil paints that do a better job of getting the full range of colors from red, blue, and yellow, and that may be why we think of red blue and yellow as pigment primaries.
[WAG]Personally, I think it may just be that we see red, blue, and yellow as pure colors psychologically, so we assume they must be the primaries. Red and blue strike the eye and the heart as pure so we call them so. Magenta is not one of the first named colors and neither is cyan. Red of blood, blue of sky. What in nature is magenta? To me it looks like it has red and blue in it. I would not have thought it possible that you could add yellow to magenta to get red. [/WAG]
BTW: you can’t make all the colors with CMYK pigments or with RGB leds. There are colors that just can’t be done with. Try getting a good Kelly green on a monitor. http://www.pyrzqxgl.org/~kmartin/color/colorexplorer-p.html shows up such gaps in RGB nicely.
http://webexhibits.org/causesofcolor/index.html is a good discussion of color.
It says “5-wheel color mixing,” so the light from a single bulb goes through 5 filter wheels. Each wheel absorbs a certain color, so it’s a subtractive system. Also the use of CMY filters is a dead giveaway that it’s a subtractive system.
An additive multi-color light would consist of three light sources with R, G and B filters aimed at the same spot. But since the three colors are coming from different directions, there will be strange fringing effects and multicolored shadows. (The shadows from the three light sources don’t overlap.)
This (and my suspicion that the responses may not be linear) is why I think my test involving LEDs and colour fatigue would enable the viewer to discern a difference