That defends on how broad or narrow bands of color similarity are across different cultures. Those do vary and mutate, and three are also zones of indeterminacy - remember the arguments over 'what color is this dress, blue or gold'?

Linguistic drift can occur due to contact with other cultures, changing environmental conditions, availability of new dyes, or political factors. As an example of the latter, consider an absolute monarch with a color vision deviation from the biological norm. Their opinions about colors could become the 'official' one, and over a long reign this could become institutionalized. Chinese culture, for example, identifies 5 fundamental colors and they're invested with far more symbolic specificity than in western societies: https://en.wikipedia.org/wiki/Color_in_Chinese_culture

And I'm not trying to be dismissive of the paper. I just have some doubts because it needs to pile methodological inferences on top of each other over a wide variation in both space and time. If you throw lots of different circles on a Venn diagram and they all have a pretty consistent overlap, then yes, that probably means something. It's a worthwhile approach.

But the fuzzier the boundaries, the harder that task becomes. I'm happy to buy into the idea that Betelgeuse changed color relatively recently and people noticed: my point was to list some confounding factors that make pinning that down very challenging.