Untested Attributions and Causal Claims

Definition

People routinely make attributions—inferences about why others said or did something, what they are thinking or feeling, what they are capable of. Model I theory-in-use discourages testing these attributions because testing would require either admitting uncertainty or explicitly questioning the other person, both of which threaten the appearance of rational control. Instead, people act on untested attributions and treat them as facts. When these untested inferences become the basis for organizational decisions and policies, they create self-fulfilling prophecies and self-sealing systems that prevent course correction.

In the Book

Argyris documents this repeatedly in his critique of management advice. When a change consultant believes that a line manager "lacks true commitment" and acts on that belief by pushing harder, the manager experiences the pushiness as disrespect and becomes defensive, which the consultant interprets as confirming lack of commitment. The original attribution was never tested with the line manager. In another case, a consultant attributes helplessness to team members and counsels them to work harder, when the helplessness may actually stem from the consultant's own dismissive behavior—an attribution the consultant never tests.

The critique extends to how advice givers present case studies. They describe an executive's success ("Mary gave an hour-and-a-half speech that caused everyone to realize potential and pull together, resulting in excellent final products") but provide no specification of what Mary actually said, what employees actually said, what they actually did, or what other factors contributed to the outcomes. The causal claims are presented without testable evidence. When Argyris asks "What did Mary say? How did she craft it? What did employees say that led to the realization? What were the other causes?" he is asking for the kind of detailed, testable reasoning that would allow others to learn from the case or to diagnose whether they're making the same mistakes.

Why It Matters

Untested attributions undermine both diagnosis and learning. If organizational problems are attributed to "resistant employees" without testing whether employees are actually resisting something specific or resisting the way change is being imposed, interventions will fail. If advice givers make causal claims ("do X and you will get Y") without testing the causal reasoning, practitioners cannot distinguish sound advice from wishful thinking, nor can they learn from failure. The book's methodology—asking for concrete data, specific language, and testable assumptions—is a remedy for this pervasive organizational practice. It is also the basis for actionable advice: you cannot implement something you do not understand, and you cannot understand it without testing the reasoning behind it.