We can tell the difference between orange-red and red and orange, but would we be able to tell the difference between the lighter green and the darker green with the same resolution? (i.e. could we still tell colors that were 4nm in wavelength apart?)

I think the different colors are a perceptual artifact of our ability to resolve wavelengths of light that are really close together fairly easily. So I think if we could still do that fairly well (and consistently across the spectrum), we would HAVE to be seeing new colors, is what I'm saying.