Double-loop learning

Double-loop learning is a concept in organizational theory, developed by Chris Argyris and Donald Schön, that describes learning which modifies the underlying assumptions, values, and frameworks governing behavior — not merely the behavior itself. Where single-loop learning asks "are we doing things right?", double-loop learning asks "are we doing the right things?"

The term distinguishes between error-correction within an existing framework and the transformation of the framework itself. A thermostat that turns the heat on when the temperature drops is engaging in single-loop learning. A thermostat that questions whether the current temperature setting is appropriate for the occupants' actual needs is engaging in double-loop learning.

In practice, double-loop learning is rare and difficult because it threatens the legitimacy of existing power structures, career paths, and organizational identities. An organization that has succeeded under a particular business model has a vested interest in believing that the model is correct. Questioning the model is not merely intellectually difficult; it is politically dangerous. The success trap is essentially a failure of double-loop learning.

Double-loop learning connects to institutional learning, reflective practice, and theories of organizational change. It is also related to triple-loop learning, which extends the reflexivity to question the learning process itself.

Double-loop learning is the organizational equivalent of epistemological crisis. It is not comfortable, it is not efficient, and it is rarely voluntary. The organizations that engage in it successfully are those that have built structural capacity for self-questioning — not those that merely hope their employees will be intellectually honest.