I'm even more fun at parties.

I think we have a little more freedom around where we label colours, as all existing colours are just an active bounding relative to our usually-semi-aligned receptors. Someone who can't see blue could call something blue because others see it as blue, but it's technically still accurate to bound and define strictly from their style of receptors, we just appreciate cohesion in our heuristics due to the nature of our social way of being.

Perceiving something temporarily as a different colour due to the changing environment also altering our internal representation does not make it inaccurate to temporarily alter which made up words appropriately describe the made up boundings around an assumed relationship our internal model representation of wavelengths share with each other, regardless of whether surface contact changing the reflected wavelengths, or a doppler shift giving a temporary relative change in experience or classification.

Also see: why language is more suggestions than law, due to the valid growth of language happening simultaneously in myriad eco-niches with different salience or needs which may alter the 'proper' way of communication specifically within that eco-niche, or in relative shape to other locales. Especially when dealing with preference or style over function, but even bias for 'function' can have a proper local-niche optimization over 'improper' models despite 'proper' local models having functional conflict with broader models. meaning we need better ways to align while diversifying without enforcing the most large and salient local collective preference as law.

Also a lot of our experience is relative or confabulatory in ways that we should recognize before rejecting other localization of experience. recognizing how we handle language at multiple simultaneous scales crafted around-and-for myriad locally/generally salient settings. Unfortunately this becomes difficult to communicate when a politician tells all the people seeing red flowers that the person in a different context seeing the blue flower is just 'wrong,' while the rest are 'correct,' even though the same people with the same logic would argue differently if positions were reversed. This silly hypothetical is the same kind of issue as conveyed around 'the double empathy problem' as a broader systems mechanic, for different boundings on building logic vs model growth purely through passive socialized affirmation, or the same vulnerability to sycophantic delusion you see with "AI" psychosis, but due to a sycophantic group rather than an LLM.

Reality is just weird and dumb, and our existing systems kind of a gauge around what is actively useful as we grow ourselves into new shapes going forward. We definitely need to be wary of socializing strict binaries around local relative confabulatory structures or preferences. Ergo: the rose can be blue, even if that's not permanently useful outside of a strict communication around a personal position and experience. We can recognize the paradox from observer states with both confabulated classifications simultaneously, with both being true and neither being wrong, dependent on the acceptance of both states as valid experiences.