When time preferences in an intertemporal optimization problem are represented by non–exponential discounting, a control plan that is optimal today may no longer be optimal from the perspective of a future self. We examine this issue by revisiting four notions...