> Note that this uses two notions of "true" - provable (can be derived from axioms), and actually true.
Even the second notion of truth actually just semantic consequence of the second-order theory of the naturals, which is a mathematical formal concept, not quite the same as actual (ontological) truth.
No, that's not what he did.
He constructed a statement that is true, but that can't be proved in a certain logical system. The statement was basically "This statement is unprovable with these axioms", which can either be proved - meaning you've proved something false, or can't be proved, meaning the statement is true, but is unprovable.
Note that this uses two notions of "true" - provable (can be derived from axioms), and actually true.