Abstract
It is a contradictory argument to say on the one hand that social systems are always open and that there is nothing between closed and open systems, and on the other that there are pseudo-closed systems. Further, Petter Næss has shown that multivariate regression analysis can be used to help uncover mechanisms, something that should be impossible if social systems were always open. He has in addition found that the meaningful activity of urban planning requires for its existence the possibility to make crude, qualitative predictions, which should also be impossible for the same reason. Critical realists therefore need to rethink the whole question of closed and open social systems