Ah. I wouldn't call that claim the Pythagorean theorem. To me, the Pythagorean theorem is a mathematical statement about mathematical objects called Euclidean triangles (or if we want to get really fancy, it's a statement about vectors in inner product spaces), and there is a separate claim, which is not mathematical, which asserts that a certain model which includes things like Euclidean triangles describes some part of the real world in some way.

In other words, I think it's sensible to enforce a strong separation between talking about the mathematical details of a mathematical model and the relation of that mathematical model to reality. To me this dissolves what I think your original question is (although I am not sure I have correctly understood what your original question is).

Maybe your question is secretly a question about the unreasonable effectiveness of mathematics in the natural sciences?