'Hyper-Rationalization' Revisited