Espoused Theory vs. Theory-in-Use

In the AI Story

The espoused/in-use distinction is the methodological heart of Argyris's research program. It forces the analyst to treat behavior as the primary data and stated beliefs as secondary — a reversal of the usual hierarchy, in which stated beliefs are treated as sincere and behavior as requiring explanation when it deviates from them.

The AI corporate discourse is almost entirely espoused theory: transformation, empowerment, human-machine collaboration, ethical deployment, workforce development. The theory-in-use — inferable from budget allocations, metric choices, governance structures, and the specific content of what the organization will not discuss — is typically different, often contradictory.

The gap is not hypocrisy in the ordinary sense. Hypocrisy implies the actor knows the statement is false. The espoused/in-use gap is typically invisible to the person deploying it; she genuinely holds the espoused theory and genuinely does not see that her behavior operates from a different one. This is what makes the gap durable: it cannot be corrected by pointing it out, because the pointing-out is interpreted within the theory-in-use that produces the distortion.

Closing the gap is what Argyris meant by Model II learning. It requires structured feedback environments in which behavior can be observed, compared with espoused theory, and discussed without triggering the defensive routines that normally protect the gap from examination.

Origin

Argyris developed the distinction through careful transcript analysis of management conversations, comparing what participants said they were doing with what the transcripts revealed they had actually done. The consistent pattern of divergence generated the theoretical framework.

The innovation was methodological before it was conceptual: Argyris's insistence on treating behavior as primary data forced a reconsideration of how organizational learning could be assessed, and the assessment revealed that most organizations were learning far less than they believed..

Key Ideas

Behavior as primary data. What people say they believe is secondary evidence; what their behavior reveals is primary. The inversion of the usual hierarchy is the methodological move that enables everything else.

Invisibility of theory-in-use. The theory actually guiding behavior is typically unavailable to the actor's introspection. This is not a failure of honesty but a structural feature of how cognition operates under pressure.

AI rhetoric as espoused theory. Most corporate AI statements are pure espoused theory. The theory-in-use must be inferred from resource allocation, metric design, incentive structure, and the specific shape of organizational silences.

Closing the gap. The gap does not close through better intention or clearer communication. It closes only through structured feedback environments that can hold the observation of behavior against the declaration of belief without collapsing into defensive activity.

Debates & Critiques

The framework has been criticized for positioning the analyst as having epistemic access to theories-in-use that the actor lacks — a privileged position that itself may be subject to the same distortions. Argyris's response was that the analyst is not privileged; the method is privileged, because it examines behavior against behavior rather than relying on the actor's self-report.