can their be such thing as a sort of a reverse "ladder of inference", if that's possible

Yes, I think so. Notice the circular diagram here.

http://www.systems-thinking.org/loi/loi.htm