Such an assertion is unprovable. That fact can never change and it is--once again--the fatal flaw in ruby's internalist model.
That's basically in error. Something can't be a fatal flaw in any claim if it's not in the claim. Rationalism 101.
Let's look at the alternative, saying that colour is not only in the brain (is an objective property of objects or light). Can that be similarly proved either? No. So is that a fatal flaw in that claim too? Wow. Now we see that so-called 'fatal flaws' are all over the shop. What to do? Forget about (that sort of) proof until such time as it is available, that's what. It's a straw man and completely irrelevant. We should not even be discussing it.
--------------------------------------------------------------------------------
Now, the model where it is deemed acceptable to have (or which uses) different definitions of what colour is so that there is (at least)
one definition that covers properties of objects or light (and elsewhere) and also
another definition which covers the phenomenological experience, has already been accepted and its existence acknowledged. It is one way to talk about colour (not an uncommon one). It is not my preferred way. I prefer, at least for the purposes of this discussion, to define colour in brains as the only actual, real colour and consider the external properties of light or objects to be so different that they do not fit the definition of colour and so it would be improper and a misnomer to call them colour. In other words, the properties of objects and light are something else. If you want to call (at least) two different phenomena colour, go ahead. I have already said that many times. All that could be proved is that there are different models using different definitions. I like mine. For one thing, I think it's more precise, including definitionally. Yours has colour in at least four different forms, and associated locations (namely in objects, in light, presumably in the optical nerve, pre-brain processing, since that must surely follow, and in the brain, after processing).
Now if, say, Palmer has since his 1999 book shifted his position (or nuanced it or made progress or whatever) to now accept or use (at least) two definitions of colour, then fine. Even if that had been his original position, that would also have been fine. It was certainly the view of some others before 1999. The alternative model, that actual, real colour is only in brains, is still valid. In other words, one is as valid as the other, and proof that one is right and the other wrong is not only an unavailable option, it crucially
depends on what definitions one uses. If all you have been doing is to try to convince me that there exists another model which rests on a different set of definitions you have been wasting your time since I already accepted that long before starting this thread.
Also, colour relationalism is quite opposed to the sort of colour objectivism you have been expressing throughout. Relationalism is closer to my preferred model than yours.
There is a short section on it in the Stanford entry by Maund which was the source of the Palmer quote in the OP and which you later cited also. From there you can then source more detailed texts by those proposing the various types of colour relationalism. Jonathan Cohen has written much about it in recent times for example (eg see below, from 2004), but you can go back as far as John Locke if you want. Obviously, it can't be proved by anyone that colour is either relational or not. Other properties can at least be subject to potential verification by other measurements, but not colour (which is possibly one reason it was called a secondary not a primary property by those using that ontology, and why the model that has it on a par with other psychological phenomena such as pain is still valid).
Color Properties and Color Ascriptions: A Relationalist Manifesto
http://aardvark.ucsd.edu/color/relational.html#tthFtNtAAD
Note that Cohen's relationalism is not identical with other relationalist models, some of which are more like dispositionalism (in which the properties of objects and light are merely disposed to cause colour experiences, similar to how other properties are disposed to cause pain experiences) though the two are closely related, since in both, colour is essentially subjectivist, perceiver-dependent to at least some degree.