Basically, instead of having surface colors & lights defined using RGB channels, you use something like ROYGCBV (Red Orange Yellow Green Cyan Blue Violet) channels.
Then, to find the color reflected by an object, you do a multiplication operation with the light and surface ROYGCBV channels.
This would then be translated to RGB for display by mapping the OYCV channels to their RGB equivalents, and adding everything up.
That way, a yellow light source that's only in the Y channel (rather than an equivalent combination of multiple channels) would cause red objects to appear black, just as they do under monochromatic non-red light sources IRL!
re: explanation, long
@Felthry Oh absolutely, it'd be *much* more intense
Still a neat area of exploration though, I feel ^w^