Single- and Double-Loop Learning: Decide Whether to Fix the Action or the Rule

Communication · Executive Insight Brief · 229

The next fix may protect the same failure

Single and Double Loop Learning separates correction within existing assumptions from inquiry that makes the assumptions themselves revisable

Premise

Organizations can correct the same problem repeatedly while preserving the condition that causes it. Single-loop learning changes an action to restore performance under existing goals, policies, or decision rules. Double-loop learning asks whether those governing variables should themselves change. The distinction prevents leaders from treating every recurring mismatch as an execution failure.

Argyris and Schön developed the lineage through their 1974 work on theories of action and their 1978 organizational-learning formulation. Their contribution is not a slogan about thinking more deeply. It is a disciplined separation between improving action under a sound rule and examining whether the target, assumption, value, or policy that defines the correction is part of the problem.

The essential model

Single-loop learning detects a gap between intended and actual results and adjusts action while holding governing variables constant. This is essential for reliable operations: teams correct defects, retrain staff, allocate resources, and refine procedures. Double-loop learning reaches the set point. It surfaces the evidence behind a goal or rule, defines what would disconfirm it, and permits revision when the governing condition no longer holds.

The distinction between espoused theory and theory-in-use makes communication central. Organizations may endorse transparency while rewarding managers who contain bad news, or endorse experimentation while punishing visible failure. Model I and Model II describe reasoning conditions that can inhibit or support inquiry; they are not additional loops and should not be equated mechanically with single- and double-loop learning.

Why this matters

Consider an illustrative company that responds to repeated customer losses with training and dashboards. Results improve briefly, then decline. A double-loop review examines the governing rule that local teams must resolve problems before escalation. If the rule delays issues requiring enterprise authority, leaders may revise the trigger while continuing single-loop improvement of the new process. Both forms of learning remain necessary.

The framework helps executives locate the true level of disagreement. Teams may appear divided over tactics when the conflict concerns what success means, which evidence counts, or which risk is acceptable. Postmortems often generate actions without learning because participants can discuss events but not the rule that made the action rational. Inquiry must connect assumptions to actual decision consequences.

Risks and limits

Double-loop learning is not automatically wiser or preferable. Constantly reopening core rules can slow execution, destabilize accountability, or reward performative questioning. People rarely state their theories-in-use directly, and interviews may capture espoused values rather than governing behavior. A systematic review of 128 studies found persistent conceptual and measurement problems and frequent drift from the original definition.

Power also limits what can become revisable. Inviting employees to challenge governing values does not give them authority to change budgets, incentives, or policies. The framework does not decide which value should prevail or how often it should be reopened. Distinguish evidence of double-loop talk from evidence that an organizational decision actually changed.

Executive takeaway

Before ordering another corrective action, identify what the proposed fix leaves untouched. If the goal, rule, metric, or assumption remains supported, improve execution and retain it. If evidence undermines that governing variable, make it discussable, revise it explicitly, and redesign the action from the new premise. Learning requires both reliable correction and principled revision, not permanent instability.

Key questions

• Which governing goal rule metric or assumption does the current correction leave untouched?
• What evidence supports that governing variable and what evidence would justify changing it?
• Where does actual decision behavior contradict the organization’s espoused theory?
• Do participants have authority to revise the premise or only permission to discuss it?

Selected sources

• Argyris, C., & Schön, D. A. (1974). Theory in Practice: Increasing Professional Effectiveness. Jossey-Bass.
• Argyris, C., & Schön, D. A. (1978). Organizational Learning: A Theory of Action Perspective. Addison-Wesley.
• Auqui-Caceres, M.-V., & Furlan, A. (2023). Revitalizing double-loop learning in organizational contexts: A systematic review and research agenda. European Management Review, 20(4), 741-761.