If a house is on fire, amazingly enough, the owner, people around, the paint on the walls, the air above the house, and so on, agree on a spectacular number of details about what it means for a house to be on fire.

We form this spectacular congruence into a model: the house is on fire.

Even if your entire thesis were not based on a confusion between the linguistic description of a phenomenon and the phenomenon itself (as if you were a large language model), why wouldn't "objective reality" be retained as a very very very very very very very very very good approximation in the classical limit?