My point is that you are assuming that if knowledge is more remote than second-hand, it will be inaccurate. This is based, in my view, on a certain perception of premodern people that, while popular, is not one I share.
Obviously I'm not making myself clear so I'll give it another shot. Forgive...