Immunity to Change Kegan Lahey

1. 1Locate a specific improvement goal you have tried and failed to accomplish, and test whether it is adaptive rather than technical: adaptive goals require that you 'get bigger,' not just add a skill or plan.
2. 2Treat the goal as a Trojan horse — a riveting object of attention that, once engaged, can break open and transform the whole mental field rather than merely resolving one behavior.
3. 3Make the invisible visible by mapping four columns: your stated commitment, the doing/not-doing that undermines it, the hidden competing commitments, and the big assumptions that hold the immune system in place.
4. 4Identify what currently 'has you' versus the plan you consciously hold; the meaning-making system operating on you as subject is the barrier, so name it.
5. 5Shift big assumptions from status 'the truth' to 'only possibly true' — this reclassification from subject to object is the mechanism the research shows drives increasing mental complexity.
6. 6Design tests of your big assumptions whose purpose is to 'get information,' not primarily to 'get better,' so the point is checking the assumption's validity, not proving self-improvement.
7. 7Reassess over time in the same sequence: each new mental plateau overcomes the systematic limitations of the prior one, so expect gradual, ordered movement rather than a single leap.
8. 8Judge success by the second, tougher question — did this promote advances in mental complexity permitting a whole range of new abilities — not merely 'did I hit the single goal.'

A medical school faculty (led by Bowe and colleagues) mapped their collective immunity to change around the goal of consistently using the best methods to achieve learning objectives. Column 1 held laudable commitments; column 2 exposed behaviors like overemphasizing factual information and not monitoring students' professional attitudes; column 3 surfaced competing commitments such as 'not having students be unprepared,' 'not threatening residents,' and 'maintaining departmental control of curriculum.' Column 4 named big assumptions — for instance, that students learn essential facts only when formally taught in current time allotments, and that assessment of professional attitudes is too subjective to count for advancement. The group grouped these into assumed limitations of students, faculty, departments, and the institution, and recognized many assumptions were unquestioned truisms 'not recently questioned.' Rather than stop at diagnosis, the reformers designed a series of tests in each category, keeping in mind the purpose was to get information about validity, not primarily to get better. Reclassifying these truisms from 'the truth' to 'only possibly true' opened a way forward to overturn the immunity.

• If a change goal has resisted thoughtful plans and heartfelt resolutions, then treat it as adaptive — requiring you to get bigger — not as a technical fix.
• If your development effort stays directed at the plan or agenda you consciously hold, then your capacity for change is inherently limited; redirect attention to the agenda that 'has you.'
• If a big assumption still reads as 'the truth,' then do not attempt to overturn the immunity yet — first move it to 'only possibly true.'
• If the goal of a test is to 'get better,' then reframe it: the test exists to get information about an assumption's validity.
• If you only ask 'did I meet the single improvement goal,' then apply the tougher second test — did mental complexity advance, enabling a whole range of new abilities.

• Believing the mind, like the body, stops growing after adolescence — the tell is treating current mindset as fixed and pursuing only skill-addition.
• Attending to leadership or the plan while underattending to development — the tell is a change effort that only serves the agenda you already hold, never the one that has you.
• Stopping at diagnosis because the four-column X-ray feels powerful — the tell is confidence that the diagnostic alone will produce the 'treatment,' skipping tests of assumptions.
• Holding unquestioned truisms as timeless truth — the tell is 'can't remember a time when these mindsets weren't accepted' plus resistance when proposals seem to fly in the face of them.
• Treating the immunity work as therapy insight — the tell is raving about a new insight ('I never saw that coming!') without a whole new meaning-making system actually evolving.

• ▸Improvement Goal: the specific change you have repeatedly tried and failed to make
• ▸Adaptive vs Technical: state whether accomplishing this requires you to get bigger
• ▸Doing/Not-Doing: behaviors that work against the goal
• ▸Hidden Competing Commitment: what you are also committed to (the fear if you did otherwise)
• ▸Big Assumption: the belief you currently hold as 'the truth'
• ▸Reclassification: rewrite the big assumption as 'only possibly true'
• ▸Test Design: an action whose purpose is to get information about the assumption's validity
• ▸Second-Question Check: what new range of abilities would advancing mental complexity permit beyond this one goal

1. 1Name one improvement goal in column 1 — a change you genuinely want and know you should make but have failed to make (e.g., 'be a better delegator,' 'be more courageous in my communications'). State it as a commitment, not a complaint.
2. 2Fill column 2 with the concrete behaviors you do or fail to do that work against the column 1 goal. Keep these as observable actions (things a camera would catch), not attitudes or explanations.
3. 3Build column 3 by asking what worrisome, scary outcome you imagine if you did the opposite of each column 2 behavior. Convert each fear into a hidden competing commitment — a commitment to preventing that outcome. This is where the immunity becomes visible.
4. 4Extract the big assumptions underlying column 4: the beliefs you hold as truth (not as hypothesis) that make the competing commitments feel necessary. State each as 'I assume that if X, then Y (disaster).' These are held as fact until surfaced.
5. 5Recognize you cannot diet your way out — working only on column 2 behaviors inside the existing mindset produces no new learning. The point is to transform the immune system, not eliminate anxiety-management altogether; you build a bigger system that can hold both the column 1 and column 3 commitments.
6. 6Design a safe, modest test of one big assumption: change one behavior, observe what actually happens against what the assumption predicts, and collect data. Run it, interpret results, and use disconfirming evidence to loosen the assumption's grip. Repeat with progressively bolder tests.

Take the medical-school faculty's collective immunity to integrating professional competencies. The column 1 commitment is active learning and integrating professional competencies into the curriculum. Column 2 shows the counter-behaviors: setting arbitrary limits with overly detailed syllabi, maintaining departmental control of curriculum, limiting faculty development in new teaching methods. Column 3 surfaces the hidden competing commitments: getting good teaching evaluations, maintaining departmental control, avoiding faculty rebellion, and protecting the school's accreditation. Column 4 exposes the big assumptions treated as fact — 'students require a highly structured setting in which to learn,' 'student evaluations are based on how much is taught rather than what is learned,' and 'faculty will rebel if expected to broaden their teaching and assessment styles.' Seen through this X-ray, the faculty's obstructive behaviors are not laziness but a brilliantly working system devoted to preserving the current way of operating. The path forward is to test one assumption — for instance, whether broadening assessment actually triggers rebellion — rather than merely exhorting faculty to change.

• If the change resists thoughtful plans and heartfelt resolutions despite genuine desire, then treat it as an adaptive challenge requiring you to 'get bigger,' not a technical fix.
• If a column 2 entry is an explanation or feeling rather than an observable behavior, then rewrite it as something a camera could capture before proceeding.
• If working only on behaviors (going on a diet) leaves the mindset intact, then stop and surface the column 3 competing commitment — no new learning occurs inside the old immune system.
• If a test of a big assumption produces no anxiety, then it is too safe to be informative; design one that touches the feeling of being without defenses in the presence of perceived danger.
• If insight arrives fast and feels exhilarating ('I never saw that coming'), then treat it as the start, not the finish — cross the gulf from insight to action through actual tests.

• Stopping at insight. The tell: people rave about the clarity of the new understanding but nothing changes in behavior — the plane is airborne with no idea where to fly or land.
• Working in too constricted a psychological space. The tell: you attack column 2 behaviors directly (dieting, resolving harder) while blind to the bigger dynamics, and effort produces no durable change.
• Trying to eliminate anxiety rather than transform the system. The tell: you frame the goal as removing all discomfort, forgetting you will always need some anxiety-management system.
• Mistaking 'change makes us uncomfortable' for the real mechanism. The tell: you assume the difficulty is the change itself, when the actual driver is feeling defenseless against perceived danger.
• Letting column 3 stay abstract or noble. The tell: your competing commitments sound like virtues rather than self-protective fears, so the immunity never becomes concrete enough to test.

• ▸Column 1 — Improvement Goal: the specific change you want and repeatedly fail to make, stated as a commitment.
• ▸Column 2 — Doing/Not Doing: observable behaviors working against the goal (camera-check each one).
• ▸Column 3 — Hidden Competing Commitments: the self-protective commitments implied by your fears if you did the opposite of column 2.
• ▸Column 4 — Big Assumptions: beliefs held as truth that make the competing commitments feel necessary ('I assume if X then Y').
• ▸Worrisome Outcome: for each column 2 behavior, the scary result you imagine from doing its opposite.
• ▸Test Design: one safe, modest behavior change to check a single big assumption.
• ▸Predicted vs. Actual: what the assumption predicts will happen versus what the test data show.
• ▸Reinterpretation: how the results loosen or revise the assumption, and the next bolder test.

1. 1Locate the Big Assumption in column 4 of the immunity map and restate it in the form 'I assume X is true' — for example, 'I assume that if I criticize the doctors they will get angry, avoid me, criticize me in return.' Naming it exposes it as a claim rather than a fact.
2. 2Reframe the assumption from 'the truth' to 'only possibly true.' This single shift — moving the belief from absolute to testable — is what lowers perceived validity and opens a way to overturn the immunity.
3. 3Do NOT change the column 2 behaviors yet. Instruct people to leave behaviors alone and instead contemplate the big assumptions over several months, as the facilitators asked the doctors and nurses to do.
4. 4Write the biography of the assumption: how long you've held it, where it came from, and whether its validity has been recently questioned. Many turn out to be inherited truisms nobody remembers examining.
5. 5Design tests whose purpose is 'to get information,' not 'to get better' — specifically information about whether the assumption holds. Keep them safe, small pilot projects.
6. 6Run the test, collect data that could either support or debunk the assumption, and review it non-defensively (as Chet did with the survey data about his own leadership).
7. 7Compare before/after data — a baseline witness survey on the column 1 commitment against a follow-up survey — so revised perceptions are grounded in evidence, then move to hooks and releases.

The medical school reformers grouped their institutional Big Assumptions into four areas: assumed limitations of students, faculty, departments, and the institution. One assumption held that students 'are unable to cope with and optimally learn in a less structured, multidisciplinary learning environment.' Rather than argue about it, they designed pilot projects — modestly reducing traditional contact hours to make room for a case-discussion course and independent research — expressly to collect data to support or debunk the assumption's legitimacy. The result surprised the naysayers: over eighty interdisciplinary faculty from seventeen departments volunteered, student reception was enthusiastic, and both faculty and students proved able to apply performance criteria. The supported pilots 'seriously challenged the validity of several institutional assumptions.' Because the assumptions had been treated as testable hypotheses rather than absolute truths, the evidence — not persuasion — dissolved the resistance.

• If the belief automatically dictates a response and feels like plain reality, then treat perceived validity as high and prioritize surfacing it before any test.
• If people admit an assumption's validity 'had not been recently questioned,' then it is a truism ripe for a validity test rather than defense.
• If you are about to run a test, then define its purpose as getting information about the assumption, not getting better — otherwise you contaminate the data.
• If the change target hasn't yet moved a belief from 'the truth' to 'only possibly true,' then don't touch column 2 behaviors; contemplation must precede behavior change.
• If data comes back challenging a cherished assumption, then review it non-defensively and as a learner, or the group reverts to protecting the old immunity.

• Trying to change column 2 behaviors immediately instead of contemplating column 4 assumptions — the tell: people feel forced and revert, because the underlying belief still reads as absolute truth.
• Running tests to 'get better' rather than 'get information' — the tell: the design has no way to debunk the assumption, only to confirm it.
• Arguing against the assumption intellectually instead of testing it with data — the tell: resistance hardens, since the belief is still held as truth, not hypothesis.
• Leaving assumptions unexamined because they are shared truisms 'generally accepted' — the tell: nobody can remember questioning them, and reform proposals that contradict them trigger deep-rooted resistance.
• Reviewing disconfirming data defensively — the tell: the champion (a Chet figure) gets defensive about how their own role is perceived, and the group loses the trust needed to keep testing.

• ▸The Big Assumption (state as 'I/we assume ___ is true')
• ▸Current status (mark: held as 'the truth' or reframed as 'only possibly true')
• ▸Origin/biography (where it came from; last time its validity was questioned)
• ▸What the assumption protects us from (the presumed inadequacy or anxiety)
• ▸Test purpose (write: to get information about validity — NOT to get better)
• ▸The safe test (a small pilot; what you will do differently, once)
• ▸Data that would support it vs. data that would debunk it
• ▸Non-defensive review plan (who reviews results; how you stay a learner)

1. 1Treat every Column 2 obstructive behavior as a symptom, not the thing itself. Ask Perry's question: what will you do to keep from getting what you say you want? List each self-defeating action without trying to fix it yet.
2. 2For each Column 2 behavior, ask what self-protective goal it serves. Complete the sentence 'Doing this protects me from...' — this surfaces the hidden Column 3 commitment (e.g., Peter's 'To preserve my sense of myself as the super problem solver, the one who is in control').
3. 3Rank the competing commitments by felt influence. Peter listed four, but one 'came to feel most influential.' The stronger the commitment feels, the more persistent the obstructive behavior it drives.
4. 4Draw the arrows: connect Column 1 goal, Column 2 behaviors, and Column 3 commitments into a single diagram. You are looking for a system in equilibrium — a foot on the gas and a foot on the brake — not three separate lists.
5. 5Test the strength directly against the improvement goal. If the Column 3 commitment is stronger, the person keeps the obstructive behavior even at real cost (heart patients: only one in seven change even facing death). Do not attribute failure to weak motivation.
6. 6Reframe the commitment as an asset, not a defect. The immune system is 'making a mistake,' not being bad. Peter's stubbornness 'has had a lot to do with the success of this company.' Naming its protective value reduces resistance to examining it.
7. 7Refuse technical fixes. Do not write action plans to eliminate Column 2 behaviors (that is solving an adaptive challenge by technical means). The behaviors are 'brilliantly effective' on behalf of Column 3, so they will regenerate until you address the assumption sustaining the commitment.

Ron Halpern, Peter's president and COO, sets the improvement goals: be more forceful and direct, be a more effective decision maker in the executive staff, stop being overly conciliatory. His Column 2 behaviors work directly against these: overchecking-in, overconsulting, insuring nonegative reactions, trying to please everyone especially the CEO, being overly tuned in to the CEO's point of view. When Ron completes Column 3, the hidden commitments explain the persistence: to be well-liked throughout the company, to maintain the high regard of the CEO, to experience the pleasure of a tight pilot/copilot relationship with the CEO. These commitments are strong enough that Ron's lack of directness is not a skill gap — it is 'brilliant and effective behavior on behalf of his own third-column' goals. A New Year's resolution to 'be more direct' cannot win against a commitment to preserve the CEO's regard, because the resolution treats directness as good and consulting as bad without seeing that consulting is doing exactly what Ron most wants. The two forces cycle enormous energy in opposing directions, so the car does not move.

• If the obstructive behavior returns despite sincere commitment and repeated effort, then a Column 3 commitment is stronger than the stated goal — stop adding motivation and map the competing commitment.
• If a behavior looks purely like a deficit (can't push back, won't delegate), then ask what it protects; if it reliably protects something valued, treat it as 'brilliantly effective,' not broken.
• If one competing commitment 'comes to feel most influential' when named, then prioritize it — that is the load-bearing brake on the system.
• If your proposed intervention only strategizes to remove Column 2 behaviors, then you are applying a technical fix to an adaptive challenge and it will fail.
• If the person shows real emotional stakes when the commitment is named (as staff cry in year-end reviews), then you have located the live self-protective goal, not a rationalization.

• Treating Column 2 behaviors as the problem to extinguish. Tell: the plan is all elimination tactics ('ask more open-ended questions') and no inquiry into what the behavior protects.
• Assuming desire equals capacity. Tell: you keep raising the stakes ('it's life or death') expecting compliance — yet, like the heart patients, six in seven still don't change.
• Naming Column 3 as a character flaw rather than an asset. Tell: the person gets defensive and shuts down instead of recognizing the commitment's protective, even successful, function.
• Listing competing commitments without weighting them. Tell: you have a flat list and can't say which one drives persistence, so interventions scatter.
• Mistaking the internal status quo for stasis. Tell: you assume 'no energy' and try to inject motivation, missing that there is 'an enormous amount of energy' already cycling in opposing directions — foot on gas and foot on brake.

• ▸Improvement goal (Column 1): state the change you sincerely want.
• ▸Obstructive behaviors (Column 2): list what you actually do/don't do that works against the goal.
• ▸This behavior protects me from: for each behavior, name the feared outcome it prevents.
• ▸Hidden competing commitment (Column 3): rephrase each protection as 'I am committed to...'.
• ▸Most influential commitment: mark the one that feels most powerful when named.
• ▸Asset value: state how this commitment has served you or your organization's success.
• ▸Strength test: does the commitment override the goal even at real cost? Yes/No + evidence.
• ▸System arrows: describe the gas/brake loop connecting Columns 1, 2, and 3.

1. 1Name the old obstructive pattern precisely — the specific 'doing/not doing' behavior your immunity produces (e.g., 'I don't delegate; I do the tasks myself'). You cannot replace what you haven't located.
2. 2Tie each new behavior to your improvement goal, not to your comfort. Ask: what action would someone who had already accomplished this goal be doing that I currently avoid? That gap defines the target behavior.
3. 3Redefine the immediate purpose of the new behavior as data collection, not self-improvement. You are running a test to see whether the feared bad outcome actually materializes when you suspend a self-imposed limit — not trying to 'get better' on the first try.
4. 4Design a specific, low-cost test before scaling. Pick a bounded situation, use knowledge of others' strengths to raise the odds of success, and predict in advance what your big assumption says will happen.
5. 5Enact the new behavior, collect the actual result, and interpret it against your prediction. If the expected disaster doesn't occur, you've begun replacing safety-through-limitation with safety-informed-by-learning.
6. 6Pair every behavioral move with mindset work. A new behavior alone is a technical fix to an adaptive problem; without testing the underlying big assumption, the old pattern returns. Keep both behavior and mindset on the agenda simultaneously.
7. 7Widen the repertoire incrementally — add options in new domains (David moving from work delegation to household chores), so the capacity generalizes rather than staying a single trick.
8. 8Recruit signals and feedback loops. Ask people who know your commitment to flag when you slip into the old pattern, and seek explicit feedback so the new behavior gets reinforced externally, not just self-monitored.

David wanted to delegate better so he could focus on select, critical issues instead of day-to-day execution. His old obstructive pattern was doing tasks himself; the new behavior was handing them off. A few weeks into the follow-up process, he chose specific tasks to stop doing and, before asking for help, used his knowledge of each person's strengths and preferences to decide whom to ask — raising the likelihood they'd succeed. He clarified his top three priorities and started working on them. The immediate purpose was not to 'improve' but to generate data testing his big assumption about what would happen if he let go. Notably, the same widening of repertoire surfaced elsewhere: he realized he'd never hired anyone for household chores either, admitting 'I have trouble delegating mowing the lawn!' — showing that the adaptive behavior, once genuinely developed, presses on old limits across domains rather than staying a narrow workaround.

• If a new behavior can be performed inside your current mindset without discomfort, then it's a technical add-on, not adaptive repertoire — expect it to erode under pressure.
• If you're enacting new behavior to 'get better' rather than to 'get information,' then reframe it as a test first; the self-improvement pressure triggers the old anxiety and self-protection.
• If the feared bad outcome does not materialize during a test, then treat that as evidence to expand the behavior; if it does, examine whether the test was too large or the design ignored others' strengths.
• If a change persists only on the behavioral OR only the mindset dimension, then you haven't crossed the threshold — both must stay on the agenda for the repertoire to hold.
• If energy, clarity, or 'seeing more' shows up (Cathy: 'I belong here, and I can see things that should happen that I never noticed before'), then you're accessing released capacity, a sign the immunity is genuinely loosening.

• Treating insight as sufficient. Readers of the earlier work found greater insight 'exhilarating' but insufficient for lasting change — the tell is feeling clarity but no altered behavior weeks later.
• Coping instead of developing — 'adding some new resources' while remaining 'the same person we were before.' The tell: you've widened your response menu but the same agenda still 'has you' out of awareness.
• Favoring only the action dimension — acting for a host of reasons while losing clarity on purpose. The tell: you're performing new behaviors but can't say what big assumption you're testing.
• Skipping test design and going straight to 'do better,' so the first stumble reactivates chronic self-protection. The tell: intelligence diverting to self-preservation, that 'fog' of low-grade anxiety.
• Confining the new behavior to one safe domain and never generalizing. The tell: you delegate at work but still can't ask anyone to do your ironing — the immunity holds everywhere you didn't explicitly test.

• ▸Improvement goal: the specific change you've tried and failed at before
• ▸Old obstructive behavior: the 'doing/not doing' pattern you want to replace
• ▸New behavior to test: one concrete action a person who'd already succeeded would take
• ▸Test boundary: the specific, low-cost situation where you'll try it
• ▸Big assumption predicted outcome: what you fear will happen when you act differently
• ▸Setup for success: whose strengths/preferences you'll use to raise the odds
• ▸Actual result observed: data collected, feared outcome yes/no
• ▸Signal & feedback plan: who flags the old pattern and how you'll invite feedback

1. 1Define the improvement goal in Column 1 as a specific, focused capability statement, not a general aspiration. Convert vague goals like 'close the achievement gap' or 'improve communication' into operational versions such as 'improve how we teach' or 'handle conflict more effectively.'
2. 2For a collective goal, distribute the blank map and have each person independently write their candidate goal in the left-most brainstorming column, then reconcile everyone's Column 1 entries into a single shared commitment. Test the wording by asking five people separately what it means; if answers diverge, it is not yet specific enough.
3. 3Complete Columns 2, 3, and 4 so the goal is anchored to a real immunity system — the doing/not-doing, the hidden competing commitment, and the big assumption — because insight alone is insufficient to bring about lasting change.
4. 4Diagnose whether the goal is technical or adaptive (per Heifetz's distinction). If achieving it requires you to 'get bigger' — to change your meaning-making system, not just add a skill — treat it as adaptive and expect the work to take longer.
5. 5Surface and test the Column 4 big assumption directly, since it is the thing that, once made object rather than subject, permits the goal to be reached. In the school district case, this meant asking whether pushing ELL students 'will always lead to failure and suffering.'
6. 6Sustain the change by running tests that revise the assumption rather than declaring victory on first insight; the achievement is the successful AND sustained attainment, not a one-time behavior shift.
7. 7Assess achievement against two questions: did you accomplish the specific improvement goal, and did the process also advance your mental complexity so a whole range of new abilities becomes available, not just the single goal.

The leadership team of a Southern California school district (superintendent, assistant superintendents, principals; student body over 80% Latino, staff over 80% white) set a Column 1 collective commitment: 'accelerating the rate of academic achievement of our English Language Learners.' Column 2 revealed they 'do not hold high expectations for our ELLs.' Their first Column 3 draft named only the workload — needing new curricula and teaching methods for an already-stretched group — which technically created an immunity picture but did not yet reach the real thing. On a better draft they surfaced the Column 4 big assumption: 'if we really did push our students they would not succeed; they would be crushed; we would feel terrible for them and personally defeated.' Naming this was 'a big eye-opener' — a previously unilluminated side of the group's operation was now in plain view. For the first time the group could ask the crucial question of whether pushing these students always leads to failure, which created the possibility of revising the assumption and overturning a protracted immunity. The achievement of the goal depends on testing and revising that assumption, not merely on seeing it.

• If a stated goal is 'too general or too operational' (e.g., 'improve communication'), then rewrite it as a focused capability ('handle conflict more effectively') before proceeding.
• If five people define the Column 1 goal differently, then the group has not yet reached a genuine collective commitment — reconcile before moving on.
• If reaching the goal requires you to change how you make meaning (not just add a technique), then treat it as adaptive and expect the third threshold to take substantially longer.
• If the team settles quickly on Column 3, then push further — the first answer (often workload or logistics) usually masks the real hidden commitment and the deeper learning opportunity.
• If insight has occurred but behavior has not sustainably changed, then the goal is not yet achieved — greater insight, however exhilarating, is insufficient to bring about lasting change.

• Mistaking insight for achievement. Tell: people report a powerful 'aha' or eye-opener but the old behavioral and conversational routines return; the plane is airborne but nobody knows how to land it.
• Setting a goal too general or too operational. Tell: the group keeps feeling 'we aren't getting anywhere' or 'we're going in circles' because the aspiration is unfocused.
• Stopping at a shallow Column 3. Tell: the team names only the extra work or logistics (the district's first draft) and the map technically shows an immunity but produces no real movement.
• Assuming desire and motivation are enough. Tell: even with life-or-death stakes the change fails — like the heart patients where only one in seven follows through.
• Declaring the goal met after one behavioral change without revising the underlying big assumption. Tell: the improvement is not sustained and the immunity quietly reasserts itself.

• ▸Focused improvement goal (Column 1): state the specific capability, not a general aspiration
• ▸Five-person meaning test: record whether independent definitions converge or diverge
• ▸Doing/not-doing instead (Column 2): behaviors that currently work against the goal
• ▸Hidden competing commitment (Column 3): push past the first, surface-level answer
• ▸Big assumption (Column 4): the belief that, if true, makes the immunity sensible
• ▸Technical vs. adaptive: mark whether reaching this requires you to 'get bigger'
• ▸Assumption-testing action + result: the experiment run to revise the assumption
• ▸Sustained-attainment check: evidence the change held over time, not a one-off
• ▸Complexity payoff: did new abilities beyond this single goal become available?