The fact that "soft" sciences are too often junk sciences is not proof that "empirical proofs" aren't required. The opposite is true, in fact.