Strong Emergence Arising from Weak Emergence

Figure 11

Schematic representation of this article’s main claims. From left to right: hard emergence may be given for a system property X if no a priori macroscopic model can be derived from the known microscopic properties that would predict X sufficiently correct, where bottom-up models fail to ever “push through” to the macroscopic system layer, on which property X manifests, and when also direct micro-to-macro predictions fail. The fact that property X can be post hoc explained or even predicted by macroscopic only prediction methods (regression models, neural network models, or any other machine learning methods) is irrelevant to identify candidates for strong emergence because they operate on the macroscopic system layer only.