"Do not in any way 'prove the rule,' but on the contrary are completely valid disproofs, even though they are rare, indeed, so long as one single exception is demonstrable."