Adam Grant calls this “confident humility” — the combination of genuine conviction in the current best understanding with genuine openness to revising that understanding when evidence demands it. It’s the intellectual posture of the scientist, applied to beliefs about the world, about people, and about the self. His 2021 book Think Again: The Power of Knowing What You Don’t Know is an extended argument that this posture is both rarer and more valuable than usually assumed, and that the default human mode of thought — the mode that protects beliefs from scrutiny and treats doubt as weakness — produces worse outcomes in professional, personal, and civic life than most people are willing to acknowledge.
The Four Modes of Thought: Which One Are You Using?
Grant’s organizing framework describes four modes that humans shift between when they encounter challenges to their existing beliefs. Three of them are versions of the same basic impulse — defend what you already think — and they’re described using occupational metaphors: the preacher, the prosecutor, and the politician.
The Preacher treats beliefs as sacred commitments. When beliefs are challenged, the preacher doesn’t investigate the challenge — evidence gets marshaled to defend against it instead. The goal is not understanding but confirmation. Sacred beliefs are not to be examined; they are to be protected. Preacher mode tends to kick in when the beliefs in question are identity-connected — political, religious, professional, or personal identities so embedded in self-conception that threatening the belief feels like threatening the self.
The Prosecutor treats a challenge to a belief as an opportunity to defeat the challenger. The goal is not to evaluate the argument but to find its weakness, to win, to invalidate the opposing view. Prosecutorial thinking is most common in adversarial contexts — debate, litigation, political argument — but it bleeds into private thinking too. Spending mental energy identifying the flaws in an argument that’s challenging a belief, rather than genuinely evaluating its merits, is prosecutor mode.
The Politician treats beliefs as positions to be held or abandoned based on audience response rather than evidence. Beliefs change, but not in the direction of truth — in the direction of social reward. The politician believes what the room believes, endorses what the influential endorse, and revises positions when the social calculus changes rather than when the evidence does. This produces the appearance of open-mindedness while achieving the same epistemic result as preaching: no actual updating based on truth-seeking.
Against these three, Grant proposes the Scientist mode — treating beliefs as hypotheses to be tested against evidence, updating them when the evidence requires it, being as willing to disprove existing beliefs as to confirm them. This is the mode that produces learning, that allows adaptation, and — not coincidentally — the mode that’s most uncomfortable, because it requires genuine openness to being wrong about things that matter.
Confident Humility: The Core Intellectual Virtue
The intellectual virtue that enables scientist mode is what Grant calls confident humility — a specific combination of genuine conviction and genuine openness that is easily confused with, but importantly different from, two failure modes that look similar from the outside.
Confident humility is not imposter syndrome. Imposter syndrome is a deficit of confidence — believing you know less than you actually do, accomplishments feeling unearned, operating on the assumption that competent people will eventually see through you. This produces a kind of useful humility, but it also produces genuine epistemic distortion: imposter syndrome makes you doubt the things you actually know well, which is as costly as arrogantly believing things you don’t know at all.
Confident humility is not armchair doubt. The opposite failure mode from arrogance isn’t useful either. The person so uncertain about everything that nothing can be committed to — who treats every belief as provisional to the point of paralysis — has taken the intellectual virtue of humility to a dysfunctional extreme. Confident humility requires actual conviction in the current best understanding, conviction strong enough to act on, combined with genuine openness to revising that conviction when the evidence is sufficient.
What confident humility looks like in practice is the willingness to distinguish between what you know and how you know it — to hold beliefs at the confidence level the underlying evidence actually supports rather than at the level ego or identity demands. The scientist who says “I believe this strongly, but I’d update if I saw X, Y, or Z evidence” is exhibiting confident humility. The one who says “I believe this and no evidence could change my mind” is in preacher mode, regardless of how sophisticated the reasoning sounds.
Identity and Belief: The Core Problem
The deepest insight in Think Again is about the relationship between identity and belief — specifically, the way beliefs fused with identity become immune to evidence. When a belief is just something a person thinks, it can be updated with new information. When a belief is part of who someone is — connected to professional identity, political tribe, religious community, a family’s cultural inheritance — then challenging the belief triggers the same threat-response as an attack on the self.
Grant argues that the solution is not to detach from all identity-connected beliefs — an unworkable prescription — but to make learning and updating central to the identity rather than peripheral to it. Instead of “I am a conservative” or “I am a free-trade advocate” or “I am a vegan” — identities that make updating the associated beliefs feel like betrayal — the prescription is to organize identity around the learning process rather than the conclusions: “I am someone who cares about evidence and updates my views when it changes.”
This identity architecture makes revisiting beliefs feel like an expression of identity rather than a threat to it. Someone whose core identity is “intellectual honesty and evidence-tracking” can update views on climate policy, immigration, or nutrition without a crisis of self. They’re being who they are by updating. Someone whose identity is “I am a true believer in X” is threatened by the very evidence that should be informing them.
Motivational Interviewing: The Art of Changing Other People’s Minds
The core insight from motivational interviewing research is that telling people they’re wrong — or presenting evidence that proves they’re wrong — reliably makes them more committed to the wrong position, not less. The boomerang effect: direct challenge triggers identity defense, which strengthens the position rather than weakening it. The preacher goes louder. The prosecutor finds more flaws in the counter-evidence. The politician looks for the social cues that tell them the room is still on their side.
What works instead is asking questions that help a person find their own contradictions and their own reasons for reconsideration. Not “here’s why you’re wrong” but “help me understand your position” — followed by genuinely curious questions about the basis of the belief, the evidence underlying it, the confidence level it warrants, the circumstances under which it might change. The goal is not to win the argument but to create the psychological conditions under which someone can reconsider a position without feeling that the reconsideration is a defeat.
This technique applies across personal and professional contexts — parenting, coaching, management, negotiation, healthcare. It requires patience and genuine curiosity, which makes it harder than arguing. But it produces actual belief change far more reliably than the confrontational approach most people default to when they want to change someone’s mind.
The Resilient Wisdom Framework: The Belief Audit Protocol
Drawing on Grant’s framework, a practical tool worth building is the Belief Audit Protocol — a structured process for identifying beliefs that may be overdue for revision and moving them from preacher-protected to scientist-evaluated status.
Step 1: Inventory sacred beliefs. In each major domain of life — professional, relational, health, political, philosophical — identify the beliefs held with highest confidence and that would be most uncomfortable to revise. These are the preacher-protected beliefs. They may be correct; high confidence doesn’t automatically indicate error. But they’re the beliefs most in need of scientist-mode evaluation precisely because they’re least likely to receive it.
Step 2: Apply the “what would change my mind?” test. For each identified belief, ask: what specific evidence or argument would, in principle, cause this belief to be revised? If no such evidence can be identified — if the belief is structured such that no possible evidence could revise it — then the belief isn’t actually functioning as a hypothesis. It’s functioning as an identity commitment. Worth knowing about yourself.
Step 3: Seek disconfirming evidence actively. For the highest-confidence beliefs, actively seek out the strongest available counter-arguments. Not strawman versions of opposing views — the best available articulation of the case against the position. If strong counter-arguments can’t be found, the evidence might genuinely be on your side. If you find you’ve been avoiding them, that’s information too.
Step 4: Separate the belief from the identity. For beliefs with significant identity connections, practice articulating the belief in a form that separates it from the identity commitment. Instead of “I believe X because I am the kind of person who believes X,” reconstruct it as “I believe X because of evidence Y and Z, and I would update if I encountered evidence A, B, or C.” Not just semantic — it changes the psychological relationship with the belief and makes updating feel like intellectual integrity rather than identity collapse.
Think Again in the Organizational Context
Grant extends the individual-level framework to organizational and educational contexts, arguing that the same scientist-versus-preacher/prosecutor/politician dynamic plays out at the institutional level. Organizations with “psychological safety” — where employees can express disagreement, raise concerns, and challenge existing approaches without fear of social or professional penalty — update their practices more rapidly and perform better than those where hierarchy and conformity pressure function as organizational preacher modes.
The educational implications are the most practically urgent. Grant argues that the conventional education model — students receive information and demonstrate recall of it — builds knowledge but doesn’t build the thinking infrastructure that allows that knowledge to be updated when the evidence changes. The scientist mode requires specific cultivation: practice at constructing arguments on both sides of questions, exposure to genuine intellectual uncertainty, experience of updating beliefs in response to evidence as something to be practiced rather than avoided.
This is a more radical critique of conventional education than it might initially appear. Not just an argument for teaching more critical thinking. It’s saying that the entire relational model of education — expert transmits knowledge to student, student demonstrates recall — is optimized for filling preacher-protected information repositories rather than for building scientist-mode epistemological infrastructure. The difference in downstream outcomes between these two educational goals is enormous.
Where the Book Excels and Where It Struggles
Think Again is Grant’s most philosophically ambitious work, and the ambition is largely realized. The four-modes framework is clear and useful, the confident humility concept is precisely defined, and the motivational interviewing sections are among the most practically applicable pages in Grant’s bibliography.
The book’s weakness is primarily in its avoidance of the most difficult implications of its own argument. Grant argues for rethinking across political and cultural lines, for opening up to evidence that challenges tribal commitments, for applying scientist mode to beliefs that are identity-connected. But his own examples of beliefs worth rethinking tend to cluster in relatively safe intellectual territory — management practices, team dynamics, educational methodology. The most charged applications — where the rethinking-your-beliefs prescription runs into the strongest identity resistance — are handled more gingerly than the framework’s internal logic would seem to require.
There’s also a tension in the book between two different prescriptions Grant doesn’t fully resolve: the instruction to revise beliefs when evidence demands it, and the acknowledgment that not all questions have clear evidential resolutions. On empirical questions with substantial scientific consensus, the scientist mode has a clear prescriptive direction. On questions that are genuinely contested, where evidence underdetermines belief and values play a significant role in determining which evidence counts and how much, the prescription to “think again” can become either vacuous or politically loaded, depending on which direction of revision is implied. Grant acknowledges this but doesn’t fully work through it.
“The mark of intelligence is not the ability to maintain your position under pressure. It’s the ability to update your position when you’re shown sufficient evidence that updating is warranted. These two things look the same from the outside, which is why we’ve confused them for so long.” — The core epistemic challenge of Think Again
Practical Applications
- Maintain a “beliefs update log.” Keep a record of significant beliefs revised and what evidence or argument prompted the revision. Two purposes: it makes updating feel like an accomplishment rather than a defeat — a track record of intellectual growth — and it helps identify patterns in which categories of belief get revised most and least readily. The patterns are informative.
- Use “I think” instead of “I know.” Small language shift, significant effects on how beliefs get held. “I think climate change is primarily anthropogenic” is a different epistemic claim than “I know climate change is primarily anthropogenic” — the first invites engagement, the second forecloses it. Using “I think” habitually doesn’t undermine conviction; it keeps the belief in hypothesis-mode where it can continue to respond to evidence.
- Apply motivational interviewing to high-stakes disagreements. Need to change someone’s mind — about a treatment recommendation, a business decision, a life choice? Try the motivational interviewing approach before the argumentative one. Ask questions that draw out their reasoning. Express genuine curiosity about the basis of their position. Find the points of agreement and build from those. The goal is to create the conditions for them to find their own reason to reconsider, far more durable than a conclusion reached because someone else won an argument.
- Seek out the best version of the opposing view. In whatever domain the strongest convictions live, identify the most credible, most sophisticated advocate of the contrary position — not a strawman, not the weakest version of the opposition. Read their best argument. If it changes nothing, that’s evidence the position has survived scrutiny. If it changes something, that’s evidence the exercise was needed.
- Build rethinking moments into recurring review cycles. Schedule periodic reviews of major decisions and beliefs — quarterly, annually — with the explicit question: what’s been learned since this decision was made that would change it if it were being made today? This prevents the sunk-cost fallacy from masquerading as consistency and creates a structured opportunity for the kind of updating scientist mode requires but daily life rarely prompts.
Key Lessons from Think Again
- The willingness to rethink is a skill, not a personality trait. People who update their beliefs well have built specific practices and habits that make updating psychologically easier than it is by default. These practices can be built by anyone regardless of starting point.
- Identity fusion with beliefs is the primary barrier to rethinking. The solution is not to hold fewer beliefs strongly but to make “intellectual honesty” the core identity commitment — which turns belief-updating into an expression of identity rather than a threat to it.
- Arguing does not change minds — asking does. The motivational interviewing research is clear: direct challenge to beliefs triggers defensive responses that entrench them. Genuine, curious inquiry about the basis and limits of beliefs creates the conditions in which a believer can find their own reasons to reconsider.
- Confident humility is a performance advantage, not a weakness. Teams and leaders who model confident humility — who hold their positions strongly while remaining genuinely open to revision — make better decisions, adapt faster, and produce more innovative outcomes than those who treat conviction and openness as incompatible.
- The most dangerous beliefs are the ones that feel most obviously true. The beliefs most resistant to evidence are the ones most integrated into identity and community — usually the ones that feel like common sense rather than positions. Common sense is often just a belief that’s been around long enough to stop looking like a belief.
Think Again Summary Q&A
Explicitly not. The “confident” part of confident humility is as important as the “humble” part. Grant is not arguing for epistemic timidity or the inability to hold strong positions. He’s arguing for a specific relationship with those positions: holding them with the confidence the evidence actually supports while remaining genuinely open to updating when the evidence changes. Weak beliefs are not scientist mode — they’re the politician mode masquerading as humility.
How does Think Again apply to situations where the stakes of being wrong are immediate and high?
Grant addresses this with the concept of “process confidence” — confidence in the decision-making process even when the outcome can’t be confidently predicted. In high-stakes situations requiring fast decisions, the prescription isn’t to delay action for additional evidence-gathering — it’s to have built the scientist-mode habits in lower-stakes situations so high-stakes decisions are already emerging from a better epistemological foundation. The time to build confident humility is not in the crisis; it’s in the calm periods before it.
What’s the relationship between Think Again and psychological safety?
Tight. Psychological safety is the organizational infrastructure that makes the scientist mode possible at the team level — the condition under which people feel safe enough to express genuine uncertainty, acknowledge errors, and challenge existing positions without fear of social penalty. Grant’s individual-level confident humility and Amy Edmondson’s organizational-level psychological safety are complementary frameworks pointing at the same fundamental requirement: environments that treat intellectual honesty as a virtue rather than a vulnerability.
Does the book address how to rethink beliefs in relationships?
Yes, with some of the book’s more practically useful content. Grant’s discussion of “argument cultures” versus “inquiry cultures” in intimate relationships — where argument culture treats disagreements as battles to be won and inquiry culture treats them as questions to be explored together — is directly applicable to the relational context. The motivational interviewing approach is particularly useful for parental relationships, where the desire to change someone’s mind (a teenager’s, for example) is high and the effectiveness of direct argumentation is demonstrably low. For more on resilient relationships, see our content on building relationship resilience.
How does Grant address the problem of motivated reasoning?
With appropriate honesty: he acknowledges that everyone is a motivated reasoner and that no amount of intellectual virtue completely eliminates the tendency to process confirming evidence more favorably than disconfirming evidence. What the scientist mode provides is not immunity to motivated reasoning but structural practices — seeking disconfirming evidence, maintaining beliefs-update logs, using “I think” language — that make motivated reasoning more visible and more correctable. Awareness of the problem doesn’t solve it, but it’s the necessary first step.
Is this book politically biased?
It has a liberal intellectual orientation that shows in the examples chosen and the beliefs selected for rethinking. The framework itself is politically neutral — the scientist-versus-preacher distinction applies equally well to rigid left-wing and right-wing beliefs. But readers who come to the book with strong conservative commitments will notice that the examples of beliefs worth rethinking often cluster on the right-leaning end of the political spectrum. Worth noting, not because it undermines the framework, but because applying the book’s own prescription to itself would require Grant to rethink some of his examples.
What’s the book’s view on expert consensus?
Complex and somewhat underresolved. Grant generally argues in favor of updating toward scientific consensus on empirical questions — climate change, vaccine safety, evolution — while advocating rethinking on questions where the evidence is genuinely contested or where conventional wisdom has been overturned by research. The implicit norm is that scientific consensus deserves significant epistemic weight but not immunity from scrutiny — a reasonable position that could be developed with more precision than the book provides.
How does rethinking apply to career and life decisions?
Grant has a specific prescription: “career checkups” — periodic reviews of whether the career currently being pursued is still the one that would be chosen given what’s now known. This is an application of the broader rethinking framework to identity-connected decisions. Major life choices tend to get treated as permanent commitments rather than as hypotheses about what will produce a life worth living — and evaluation of the evidence tends to stop once the choice is made. Building periodic rethinking moments into major life decisions is one of the most practically impactful applications of the book’s framework. See our content on building identity resilience for more on how to make these reviews productive rather than destabilizing.
Books Similar to Think Again
The Intelligence Trap by David Robson — An examination of how higher intelligence can actually increase susceptibility to certain forms of motivated reasoning and belief entrenchment. Complements Grant’s framework with a more psychologically focused analysis of why smart people hold wrong beliefs with such confidence.
Being Wrong by Kathryn Schulz — A deep philosophical and historical examination of error and how humans relate to it. More literary and less practical than Think Again, but illuminating for readers who want to understand the deeper cultural and cognitive roots of the beliefs-as-sacred problem.
Superforecasting by Philip Tetlock — The empirical foundation for the forecasting practices Grant describes. Tetlock’s research on what distinguishes accurate from inaccurate forecasters is the best available evidence for the real-world performance value of the scientist mode epistemic posture.
Black Box Thinking by Matthew Syed — Strong overlap on the failure-to-learn problem and the cognitive dissonance mechanisms that prevent honest self-assessment. The organizational failure culture analysis is the complement to Grant’s individual epistemological framework.
Influence by Robert Cialdini — The dark side of the same psychological mechanisms Grant describes. Where Grant shows how to build better updating habits, Cialdini shows how commitment, consistency, and identity are systematically exploited to prevent updating. Reading both together produces a more complete picture of how human belief systems work.
Straight Talk on Think Again
Think Again is Grant’s most important book and one of the most useful intellectual tools available for anyone making decisions under uncertainty — which is everyone. The scientist-mode framework is precise enough to be actionable, the confident humility concept is carefully enough defined to be useful rather than vacuous, and the motivational interviewing application is among the most practically valuable sections of any book in the contemporary business-and-psychology genre.
Its limitations — the safe-territory example selection, the underresolved tensions between rethinking and expertise, the political blind spots — are real but don’t undermine the central contribution. This is a book that earns its place in the “actually change how you think” category rather than the “confirm what you already believe about being open-minded” category.
Apply the Belief Audit Protocol to something that actually matters. Use the motivational interviewing approach next time changing someone’s mind about something important becomes necessary. Build the beliefs-update log. These practices, applied consistently, produce the kind of adaptive intelligence the scientist mode describes — and that preacher/prosecutor/politician thinking systematically prevents.
Rethinking in High-Stakes Environments: When Updating Is Hardest
The scientist mode is easiest to practice when the stakes are low — when the beliefs being examined are interesting but not identity-connected, when being wrong costs nothing important. The most valuable applications of Think Again’s framework are also the most difficult: the high-stakes contexts where identity protection is strongest, where social costs of updating are highest, and where the belief systems that most need examination are most strongly defended.
Medical contexts are Grant’s most instructive case. A physician who has told a patient that a particular treatment is the right one — who has staked professional judgment on a diagnostic or therapeutic choice — faces exactly the identity-threat problem when evidence emerges that a different approach would have produced better outcomes. The same mechanism that drives an individual’s belief-protection in the medical context also drives the resistance of medical institutions to systematically examining their own failures. The individual cognitive mechanism and the institutional failure culture Syed describes in Black Box Thinking are the same phenomenon operating at different scales.
The prescription is structural in both cases. For individuals, the Belief Audit Protocol applied to professional judgments — with the specific requirement of identifying, before committing to a position, the evidence that would change it — builds the habit of holding professional beliefs in hypothesis-mode rather than identity-mode. This doesn’t eliminate the identity threat when evidence challenges the position. But it pre-commits to a process of genuine evaluation, which makes the updating feel less like defeat and more like intellectual integrity. Someone who said “I believe X, and I would update if I saw Y” can update when Y appears without feeling that professional judgment has been impugned. Someone who just said “I believe X” has no such structure to fall back on.
The organizational prescription is the same as Syed’s: build the structural insulation between belief-examination and blame-assignment. For more on how organizational learning culture and belief systems interact with performance outcomes, see our content on discipline, systems, and institutional learning. The organization that can evaluate whether its professional judgments were optimal without treating the evaluation as an indictment of the people who made them is the organization that learns fastest. This requires leadership that demonstrates confident humility — leaders who model updating their own positions in response to evidence, who treat their own past judgments as hypotheses rather than commitments, and who create the cultural expectation that changing your mind in response to evidence is strength, not weakness.
Think Again and the Resilience Connection
The connection between the scientist mode and psychological resilience is direct and underexplored. Resilience in the face of adversity requires the same epistemological posture Grant prescribes for belief management: the ability to hold a model of reality with sufficient commitment to act on it, while remaining genuinely open to updating that model when the evidence demands it.
The brittle form of resilience — “I am strong and I don’t let things affect me” — is a kind of preacher mode applied to self-perception. It protects a self-concept at the cost of honest assessment of how things are actually affecting a person, which produces the denial and rigidity that characterize brittle coping rather than genuine resilience. The flexible form of resilience — “this is hard, and the approach may need to change while learning more about what works” — is the scientist mode applied to the self in adversity.
The most resilient people in the longitudinal research are not those most certain everything will be fine. They’re the ones who can acknowledge difficulty honestly while maintaining forward orientation and genuine openness to changing approach when the current one isn’t working. Confident humility in adversity: knowing the situation is hard, knowing the current strategy may not be optimal, and continuing to act on the best current understanding while remaining open to revision. This is exactly the cognitive posture that both Grant’s framework and the resilience research commend. For more on the epistemological foundations of resilience, see our content on building adaptive resilience.
Applying Scientist Mode to Your Own Performance Beliefs
The most personally uncomfortable application of Think Again‘s framework for most readers of a performance-oriented publication is the application of scientist mode to beliefs about what produces high performance. This is an identity-connected domain for anyone who has invested heavily in a particular approach — a training methodology, a productivity system, a dietary protocol, a management philosophy. These beliefs aren’t just beliefs. They’re the foundation of a significant amount of time and effort already invested, and challenging them feels like challenging the value of that investment.
The sunk cost fallacy is the specific cognitive error that makes performance belief updating so difficult. Someone who has spent three years building their life around a specific training approach has an enormous psychological investment in the hypothesis that the approach is correct. Evidence that it’s suboptimal doesn’t just challenge the belief — it challenges the value of three years of effort. The identity protection response is to reinterpret the evidence rather than update the belief, to find alternative explanations for disappointing results, to seek out confirming evidence for the original approach and discount disconfirming evidence.
The scientist mode applied to performance beliefs requires the specific practice of identifying, in advance, what results would be expected if the current approach were optimal, and what results would indicate it isn’t. Training that produces injury rather than fitness, or fatigue rather than adaptation, or no progress on the metrics that matter despite consistent application — these are results any honest scientist would take as evidence the approach needs revision. The preacher’s response is to diagnose insufficient commitment and prescribe more effort. The scientist’s response is to diagnose a potentially suboptimal methodology and investigate alternatives. Over long time horizons, the two approaches produce dramatically different results. For more on applying evidence-based thinking to personal performance, see our content on evidence-based health and performance optimization.
Rating: 4/5 stars. Essential reading for anyone making important decisions in uncertain environments, managing people, or trying to change minds in high-stakes contexts. Grant’s most philosophically substantial work, with the practical applications to match.
Related: Hold Me Tight Summary
Related: Courage Is Calling Summary
Related: John Adams Summary
FROM THE LIBRARY ›
References
Editorial StandardsCorrectionsMedical DisclaimerAbout Our ContentAffiliate DisclosureSite Map