DougI think the underlying dilemma here is that the "color" we measure of an instance of light by objective measurements of the spectral distribution of the light and subsequent mathematical processing, and the "color" that is a sensation of a human viewing that light, are not the same thing.
The "color" we measure, objectively, is a predictor of the "color" that the "average" human viewer would perceive upon viewing that light.
And in fact the process I described earlier really uses a model of the response of the "average" human visual system to light components of different wavelengths. The three "matching curves" I describe are derived from the curves that describe the response of the three kinds of cone receptors in the "average" human visual system.
Well, to be more honest, it works in the other direction. We cannot (very well) measure the response curves of the eye. In fact we infer those curves from the related "matching curves", which are arrived at by empirical testing.Now would it be more "honest" and "clear" if we did not use the term "color" to refer to the result of this measurement process, reserving that word for the human perception? Sure.
Perhaps we would refer to the result of that measurement as the "protocolor" of the light. Then one could say, "You know, we cannot measure the color of light. All we can measure is its protocolor."
But I did not get the chance to suggest that to the CIE committee that devised the objective definition of "color" and the scheme for determining it. And I would not at all doubt that discussions of that possibility were held. But clearly the decision was to call the defined and measurable metric "color".
And perhaps in my contributions to the discussion here I did not clearly enough illuminate and indicate my recognition that indeed the "color" seen by a human and the "color" measured with an objective process are not the same thing (but we hope they are closely relatable, for the "average" viewer).
This is not the only place in science where we have this duality. Another case is in the "loudness" of sound. That too is a matter of human perception. Yet we use an instrument to measure the acoustic pressure wave arriving at a viewer's location and deliver its finding as the "loudness" of the sound at that point. And, just as in the case of "color", the working of that measurement has "baked in" a model of a certain (somewhat subtle) aspect of human hearing, of course for the "average" listener.
So to answer James' recent question, color (in sense 1) is a fact, and color (in sense 2) is a perception.
Radiation can be measured but light and color is a subjective human sensation and therefore can not be measured.