Suppose someone proves the polynomial-time hierarchy collapses. Then it didn't collapse because it was always collapsed and in fact was never a true hierarchy. The only reason we call it a hierarchy today is because we don't know that it isn't.
In mathematical reasoning we can't know whether a statement is true unless we have a proof of that statement. However once we have a proof of a theorem like P≠NP then not only do we have P≠NP now but in fact P≠NP was always true even before we had the proof. Despite this we can't help thinking that theorems become true as we see a proof. A year ago undirected connectivity wasn't in log space and now it is. However the theorems that were proven before I started graduate school were all true since the dawn of time.