CONCEPT
Epistemic Redundancy
The organizational equivalent of engineering redundancy applied to belief—the deliberate maintenance of independent cognitive architectures, separate assumption sets, and diverse experiential backgrounds within a team so that errors invisible to one perspective remain detectable to another.
Epistemic redundancy is what specialist silos accidentally provided and what AI-augmented organizations are systematically eliminating. Where engineering redundancy duplicates physical systems so that the failure of one does not take down the whole, epistemic redundancy duplicates
cognitive systems: teams with independently formed assumptions, distinct professional histories, and different mental models of a domain provide independent checks on each other’s errors in a way that no single mind, however augmented, can provide for itself. The concept emerges from two converging traditions.
Lindblom’s analysis of
partisan mutual adjustment established that the quality of democratic outcomes depends on genuine diversity of perspectives in the adjustment process—not merely formal pluralism but perspectives grounded in different positions, interests, and knowledge bases that no central authority can replicate.
Perrow’s risk theory established the parallel engineering point: when a single cause can defeat multiple safety systems simultaneously—
common-mode failure—redundancy that merely duplicates the same system provides no protection. The backup cooling pump that draws from the