On Evidence in the Metaphysical Sciences