Immunity to change

Robert Kegan and Lisa Lahey's immunity to change (ITC) explains why people fail to change behaviors they genuinely want to change. Their claim: the failure is not weak willpower or insufficient knowledge. It is a competing commitment, held with equal sincerity, protected by a big assumption that is usually untested.

The metaphor is immunological: the system is not broken; it is actively and skillfully protecting itself against something it perceives as a threat.

Evidence: Weak. Limited independent empirical validation; the supporting evidence is largely case material produced by the developers and their practitioners. The adult development theory underneath it is contested. The method's practical defense is different and does not depend on the theory: it converts vague complaints into specific, falsifiable claims about what actually happens to people here, which can then be tested directly. Requires skilled facilitation and genuine safety, and is intrusive when done badly.

Use when: an individual or group repeatedly fails at a change they sincerely endorse; when "we know what to do and we don't do it".

Do not use when: the barrier is structural, capability-based, or a lack of resources. ITC is for genuine self-contradiction, not for obstacles.

The four-column map

Column 1 — Improvement goal. One important commitment, stated positively and in the person's own words. "I am committed to delegating more."

Column 2 — Doing / not doing. Behaviors that work against column 1. Be concrete and non-judgmental. "I take back tasks I've assigned. I rewrite my team's drafts. I attend meetings I've delegated."

Column 3 — Hidden competing commitments. Ask: what would be the worst thing about doing the opposite of column 2? Sit with the fear, then translate it into a commitment. "I am also committed to never having my name on substandard work." Both commitments are real. The person is not weak; they are perfectly designed to produce the current result.

Column 4 — Big assumptions. The beliefs that make column 3 feel necessary. "I assume that if I let something imperfect through, my reputation is permanently damaged and I will lose my standing."

The big assumption is stated as fact and treated as fact, but it has usually never been tested.

Testing big assumptions

Not confrontation — experiment. Design a small, safe, reversible test:

• Modest in scope, low cost if the assumption turns out true.
• Data-generating: decide in advance what you will observe.
• Run it, record what actually happened, and compare with the prediction.

"Let one piece of my team's work go out unedited, and note what actually happens." Most big assumptions do not survive three honest tests.

Collective immunity

The map works at group level, and this is where OD uses it. An executive team's column 1 might be "we are committed to faster decisions"; column 2, "we defer every decision for more analysis"; column 3, "we are committed to never being the one who was wrong"; column 4, "we assume that a visible mistake ends a career here."

Column 4 is then a testable claim about the organization — and frequently a true one, in which case the intervention is not psychological. It is changing what actually happens to people who are wrong. See just-culture, westrum-organizational-culture-typology.

That is the highest-value use of ITC in OD: it converts vague cultural complaints into specific, falsifiable claims about consequences.

Examples by audience

• Health care: a consultant committed to developing juniors who never lets them lead. Competing commitment: never being responsible for a preventable harm. Big assumption: supervision at a distance equals negligence. Testable — in graduated, defined steps.

• Software: a team committed to continuous delivery that keeps a manual release gate. Competing commitment: never being the team that broke production. Big assumption: automated release means unrecoverable failure. Testable, with a feature flag and a rollback drill.

• Executive: a CEO committed to empowering the executive team who reviews every decision. Competing commitment: never being surprised in a board meeting. Big assumption: any surprise reads as loss of control. Testable by agreeing a class of decisions the CEO explicitly will not see.

Limitations

• Time-intensive; a proper map takes hours, and group work takes days.
• Requires real psychological safety and skilled facilitation. Poorly done, it is intrusive and can be coercive — participation must be genuinely optional. See od-ethics.
• Limited independent empirical validation.
• Risks psychologizing structural problems. Always rule out obstacles first.

See also

• double-loop-learning
• force-field-analysis — the organizational analogue.
• Kegan & Lahey, Immunity to Change.