In human communication we only feel the need to give sufficient reason for our xdecisions or actions, not unpack the firing of neorones in our brains. This principle can be applied in explainable AI, when it is tempting to try to explain the minutae of, say, every node and oink in a deep neural network. Instead, we should seek explanations at an appropriate level for the task at hand.
Defined on page 518
Used on pages 441, 518