I think one of the problems with the chain of causation argument, is that it assumes that the chain is not a loop. If it is a loop, then the end result of the chain, could indeed be caused by the previous end result of the chain, although defining what the end of the chain is, is problematical, for the purposes of this argument, consider the end of the chain to be the point just before the loop is rejoined.