They are 2 different types of balance
I do not see anything resembling an argument for delineating balance against the environment from balance between players here. What makes another player distinct from an environment if both are bound by the rules of the game re: game theory?
Even if we argue case and not type, you are aware that hostile NPCs will have levels in character classes, correct?
your argument fails to show that you cannot balance a game against the environment by simply allowing the environment to also use the "broken toys."
I think maybe you've gotten confused. And the double negative is confusing me a bit here. But if I understand you correctly, and that's a pretty big if: This was not the intended purpose of the argument and I'm not sure how it's relevant here. I was demonstrating that balance - here meaning complexity or obfuscation of player means and goals such that all means approach equal purpose towards goals - is critical to the design of cooperative experiences.
There are a different set of meta rules which govern the environment vs those which govern players. In a 1v1 PVP game, it is normally the case that everyone starts on equal footing and the differentiating factor is player skill, or, in the case where balance is bad, the skewed balance. BG3 is not analogous to a 1v1 PVP game, because you control multiple characters. In BG 3, not only is it not the case that everyone starts off on equal footing (NPCs have items already, the player only has what they find and the player almost always has an information advantage) but there are also other very clear differences, for example, the players are limited to a party size of 4, where as the AI is limited by how many enemies the developers decide to throw at you. The Goblin camp is a nice example of some 20 something enemies that illustrates this.
Even if we ignored those differences however and there was actually a PVP mode, played between 1 group of 4 players and another, it would still not be analogous to a 1v1 PVP game and then, the closest comparison would be team based PVP games. Are classes in team based PVP games balanced 1:1 against each other? Not really, no. Those types of games use the Rock:Paper:Scissors mode of balance, where class A beats Class B, Class B beats Class C and Class C beats class A. In these cases, the system as a whole is seen as in balance even if individual parts are not. This type of balance is much closer to something which could be fairly applied to 5e and even then, you are obviously going to have a situation where some classes will beat others in a matchup, but across the system as a whole are balanced.
If we ask the question, "is balance good," the answer is, in my opinion, it depends what you sacrifice to achieve balance. Imo it is fine to have a slightly out of balanced system, provided it makes for more enjoyable gameplay. Obviously, fun is subjective, so that comes down to player preference, but there are some things which surfaces very clearly do add to the gameplay. For example.
• Opportunity cost of deciding whether you want to remove the surface or leave it there.
• Restricting movement.
• Adding interesting elemental interactions (the whole cloud thing div 2 had).
• Rewarding clever gameplay (being creative with the surfaces).
It also obviously has the following downsides.
• When its always easier to apply a surface than it is to remove it, the opportunity cost is pretty much gone because the best solution is always, "live with it." To avoid this, the relative ease of applying a surface needs to match the ease of removing it (for example, a free action "douse" which would put out a flame on the character and a small space surrounding them would be roughly equal to the ease of the cantrip).
• It messes with balance.
I am a fan of surfaces because of those first 4 points there and I can live with the 2nd negative, the 1st negative is the one which is imo more important because it counters the 4 points that I mentioned above. This was a problem in DOS 2 because of Necrofire (an ooze just needed to move around to apply it, but it took 2 spells to remove it).