But words have no utility aside from being understood, a word is good as long as there is a consensus as to what it means, and you can always create other words for things it doesn't describe.

Light acting like a wave in some regards and like a particle in others is something we can see experimentally, not just a matter of semantics. It's a conclusion that experience lead us towards. Calling light "continuous" wouldn't work because there is already meaning assigned to that word, and that meaning clashes with observations. Unless we changed the whole word and redefined the continuity of everything, which would be absurd.

That "operating with unclean tools would be fine" clashes with the observation that it can lead to infections, coupled with the axiom that inflicting bodily harm to someone is bad. The axiom could still be changed, but the problem with observations is that they're imposed by reality, they would still be true even if we didn't believe they were. You could also change the word "fine" tho. If you make a language almost identical to English safe that the word "fine" means "an unreasonably dangerous practice" the sentence "Practicing surgery without disinfecting your tools is fine!" is true in that language.

Ceres not being a planet depends only on our definition of planet. It was considered a planet for a while, but what led people to reconsider that isn't just that it was smaller than believed, but also that there were many similar objects in the asteroid belt. Referring to all these objects as "planets" could've been an acceptable truth, but since that would've meant most planets known at the time are small and in the asteroid belt (the Kuiper belt and Port cloud weren't known yet, but now it's just mean most planets are in a belt), and if the likes of Pallas and Juno were included (as was the case once) it also would've meant that most planets weren't round.

Since the previously known planets would've been outliers in several ways, a new word should've been coined for them. It seemed more simple to let them be the only planets and coin the word "asteroid" for the rest (and much later the intermediate category "dwarf planet").

If a different choice had been made, asteroids could be planets, what we now call planets could be called "big planets" and dwarf planets would be called "intermediate planets". This would be an acceptable truth, it wouldn't contradict itself or observations. If it was the consensus, it would be true, but it isn't so it's false, it's as simple as that.

If in the future we find a different definition of life more useful, that definition will be true then. But that won't change what definition is true now. Ceres was a planet. Now it isn't. Something can change category either because it itself changed or because the category changed, like how substances can go from being legal to illegal or vice-versa.