Abstract
Many challenges we face as individuals and collectives are ill-defined: What are the probabilities for different courses of action? How to compare distinct outcomes? Under such radical uncertainty, conventional optimization solutions fail, prompting calls for wisdom. Yet, the science of wisdom remains fragmented across disciplines. To fill this void, we reconstruct practical wisdom through the lens of cognitive science, asking what wisdom is for, how it works, and how it develops and evolves. Guided by criteria that make a cognitive science concept empirically tractable, formally precise, and continuous with broader theories of mind, we develop a two-level functional cognitive architecture of practical wisdom, in which evoked, higher-order metacognitive processes regulate a culturally-transmitted repertoire of object-level decision strategies. We outline formal elements of this architecture and examine its implications across persons, groups, societies, and natural and artificial intelligences. Finally, we derive testable predictions and close by asking what such a science offers back to the disciplines it draws on.