> There is no logical reason why the second method should be possible at all, but one has found in practice that it does work and meets with reasonable success.

Might be semantics, but there is a lot of logical reason why mathematics should work to describe nature, because mathematical proofs are made based on logics. There is no empirical reason, though.