This is a crucial point to make. Even if every natural phenomenon can be reduced to very early-level processes (e.g., physics), earlier levels of analysis have little predictive value at later levels of analysis.
Does anyone have a term for this from philosophy of science literature?