There is a particular difficulty in changing one's mind when the old belief is written in one's own handwriting.
The sentence is familiar. It has been defended before. Other people know that you hold it.
Then evidence arrives that does not fit.
Intellectual humility concerns the capacity or disposition to recognize the limits and fallibility of one's knowledge and beliefs, and to respond to those limits without collapsing into either arrogance or intellectual self-erasure.
The idea has philosophical roots and a rapidly growing psychological literature. It is also unusually vulnerable to moral inflation. Because humility sounds admirable, research claims can easily become endorsements before the construct has been defined or measured carefully.
There is no single uncontested definition
A 2022 systematic review led by Tenelle Porter identified 18 definitions and 20 measures, including 16 unique questionnaires, across empirical intellectual-humility research.
That heterogeneity is not a footnote.
Different measures emphasize different facets: owning intellectual limitations, openness to revising beliefs, non-defensiveness, respect for others' viewpoints, independence of intellect and ego, or awareness of the value of other people's knowledge.
These facets overlap without being identical.
A person can be willing to say "I may be wrong" while still dismissing other people. Another can be socially respectful while remaining highly certain of their own beliefs. A third can be genuinely uncertain but unable to revise because identity or group belonging is at stake.
Any claim about intellectual humility therefore needs to specify what version was measured.
Humility is not low confidence
Intellectual humility is sometimes mistaken for underconfidence.
That would make accurate expertise look like a defect.
A scientist with strong evidence can be highly confident and intellectually humble if the confidence remains proportionate to evidence and revisable under relevant counterevidence. A person with chronically low confidence may be poorly calibrated rather than humble.
The distinction matters because humility is not a demand to flatten differences in expertise.
The target is fallibility awareness and appropriate ownership of limits, not compulsory self-doubt.
Metacognitive calibration and intellectual humility may relate in some contexts, but the constructs should not be merged. Calibration is an empirical relation between confidence and performance. Intellectual humility includes broader dispositional, epistemic and interpersonal dimensions.
Self-report creates a paradox
How should a researcher measure a trait whose content includes awareness of one's own limitations?
Self-report is convenient, but it asks people to assess how good they are at recognizing where their own assessments may fail.
Benjamin Meagher's 2022 study compared self- and informant ratings across four intellectual-humility scales and found weak self-other agreement, with different patterns of association with broader personality traits. The result does not make self-report useless. Some internal facets may be difficult for observers to see.
It does show that measurement source matters.
A questionnaire score can reveal a person's self-conception of intellectual humility without serving as a perfect audit of actual belief revision under pressure.
Behavior needs context too
Researchers also study intellectual humility through information seeking, response to disagreement, willingness to revise, attention to evidence quality and interpersonal behavior.
Behavioral measures can escape some self-report problems but introduce others.
Changing one's mind is not always humble. A person can switch views because of social pressure or weak evidence. Refusing to change is not always arrogant. The original belief may be well supported.
The important variable is not change itself but responsiveness to epistemically relevant reasons.
That is harder to measure than counting reversals.
Intellectual humility has cognitive and interpersonal faces
Leary and colleagues' 2017 work connected intellectual humility with openness, curiosity, tolerance of ambiguity and sensitivity to argument quality, while also examining interpersonal consequences.
This dual character is important.
Beliefs are held inside social worlds. Admitting uncertainty can alter status. Disagreement can threaten identity. People can use certainty to protect belonging, competence or moral standing.
An epistemic virtue is therefore expressed through interpersonal conditions.
That makes intellectual humility relevant to teams, science, education and public reasoning, but it also means context and power cannot be ignored. Asking a junior employee to display humility to a senior expert is not the same social act as asking the senior expert to remain corrigible.
The construct should not become a weapon for deference
Humility language can be misused.
People with less institutional power are often already required to soften certainty, hedge claims or show deference. Calling further deference "intellectual humility" can reinforce hierarchy rather than improve reasoning.
Likewise, demands for humility can be asymmetrical: critics are asked to acknowledge uncertainty while authorities speak with unearned certainty.
The research construct does not justify those norms.
A credible institutional application should make corrigibility reciprocal. People with more authority should face at least as much obligation to revise beliefs when evidence changes.
This is an Institute normative synthesis, not a direct empirical conclusion, but it follows from taking the construct's epistemic content seriously rather than using humility as etiquette.
Learning requires enough humility to expose error
Learning often begins when a person discovers that their current model is insufficient.
Intellectual humility can support that transition by reducing the need to defend a belief simply because it is one's own. But the relationship between humility and learning should not be romanticized.
Learning also requires knowledge, cognitive resources, feedback and effective strategies. A person can be willing to be wrong and still lack access to better information.
The Institute should therefore avoid presenting intellectual humility as a master trait that automatically creates wisdom.
It is better understood as one condition that can make revision possible.
Institutions can design for corrigibility without scoring humility
Organisations may want the benefits associated with intellectual humility: error correction, dissent, learning and evidence responsiveness.
The naive response is to build an employee humility score.
The stronger response may be procedural.
Decision logs can record assumptions. Pre-mortems can expose failure modes. Red teams can formalize disagreement. Forecasts can be compared with outcomes. Senior leaders can be required to document belief changes. These structures make revision observable without pretending a questionnaire captures a person's moral worth.
Whether any particular practice works depends on implementation and evidence. The larger principle is that institutions can reduce the social cost of being corrected.
AI makes false intellectual ease cheap
Generative systems can supply arguments, counterarguments and confident summaries on demand. That may support intellectual humility by making alternative perspectives easier to access.
It may also do the opposite.
A user can ask for validation of a preferred view, generate persuasive rationalisations, or mistake breadth of generated argument for actual evidence. The system can produce the performance of perspective taking without the user's belief becoming more corrigible.
Research directly linking generative AI use to intellectual humility is still immature. Claims should remain hypotheses.
A useful Institute question is whether human-AI systems make it easier to encounter disconfirming evidence and to let that evidence matter.
Humility can be local rather than global
A person may be intellectually humble in one domain and defensive in another.
Beliefs tied to identity, morality or group membership may be harder to revise than technical beliefs with little social meaning. This suggests that intellectual humility can have state-like and domain-specific components alongside broader dispositional tendencies.
Research designs that average across topics can miss this variation.
For applied work, it also means an organization should not assume that a general humility workshop will transfer to the precise decisions where status and identity are most threatened.
Disagreement is an especially revealing context
Humility becomes consequential when another person supplies information one would prefer not to accept.
The relevant behavior is not automatic agreement. Healthy disagreement may require maintaining a well-supported position. The metacognitive demand is to represent the possibility that one's own evidence or reasoning is incomplete and to evaluate the other position proportionately.
This is why intellectual humility connects to argument quality rather than mere niceness.
A group that avoids disagreement can appear humble while being epistemically fragile.
Science institutionalizes some forms of humility
Peer review, replication, error correction, preregistration and open data are not personality traits. They are procedures created partly because individual researchers cannot rely on introspection to detect every error.
These practices illustrate a broader principle: institutions can externalize corrigibility.
A scientist may be personally defensive and still work inside a system that exposes claims to criticism. A gracious person may work inside a system that hides negative results.
The distinction is important for the Institute because it moves the conversation from virtue language to infrastructure.
Cultivating intellectual humility can matter. Designing systems that do not depend on perfect humility may matter more.
Humility without epistemic standards can become relativism
If every belief is treated as equally uncertain, evidence loses its role.
Intellectual humility does not require saying that all perspectives are equally supported. It requires allowing confidence to track reasons while preserving revisability.
This is especially important in public discourse. "I could be wrong" should not be used to create false equivalence between a large evidence base and an unsupported claim.
Humility is compatible with strong conclusions when evidence is strong.
Its opposite is not certainty. Its opposite is certainty that has become insulated from correction.
The desirable-outcome literature needs causal caution
Reviews report associations between intellectual humility and outcomes such as openness, learning behavior and less defensive reasoning. Those relationships are theoretically interesting.
They are not all causal demonstrations.
People who score higher on humility measures may differ in education, personality, social environment or other variables that also predict the outcome. Experimental manipulations can help, but short-term shifts in a laboratory are not the same as durable dispositional change.
A responsible article therefore separates three questions: what intellectual humility correlates with, what can temporarily increase humility-related behavior, and what produces sustained change outside the laboratory.
There may be costs to visible humility in some environments
Institutions often claim to value uncertainty while rewarding confident speech. In such environments, openly revising a position can carry reputational cost.
That cost is not evidence against intellectual humility. It is evidence that the social environment can punish the behavior the institution says it wants.
This has methodological implications too. Observed certainty may reflect strategic presentation rather than private dogmatism.
For leaders, the relevant design question is whether correction is rewarded in practice, especially when it comes from people with less status.
Humility and expertise should be mutually reinforcing
The strongest epistemic posture is not "everyone knows equally little." It is differentiated confidence with open correction.
Experts can speak firmly where evidence is strong and remain explicit about limits. Novices can defer appropriately without surrendering the right to ask for reasons.
That balance protects the construct from becoming either arrogance dressed as expertise or deference dressed as humility.
What the evidence does not show
Intellectual humility is not the same as low confidence, low self-esteem, submissiveness or indecision.
A self-report score does not prove that a person revises beliefs appropriately in real conflict.
Changing one's mind is not inherently evidence of humility; the quality of reasons matters.
Associations between intellectual humility and desirable outcomes do not by themselves establish that increasing humility will cause those outcomes.
The construct is not measured uniformly across the literature.
Evidence does not justify using intellectual humility as a moral ranking of people or groups.
And there is not yet strong evidence that generative AI reliably increases or decreases intellectual humility across contexts.
Open questions
Which facets of intellectual humility are most stable, and which are situation-specific? How should researchers integrate self-report, informant report and behavior? What interventions produce durable change rather than socially desirable responding? When does humility improve group reasoning, and when do status dynamics suppress the expression of justified confidence?
Cross-cultural questions are especially important. Norms for modesty, disagreement and public certainty differ. A scale developed in one cultural setting may not carry the same meaning elsewhere.
For institutions, the open question is whether it is more effective to cultivate a trait or to build processes that make correction safe and consequential.
The sentence with a line through it
Intellectual humility becomes visible when a belief is allowed to lose status because evidence has changed.
That may look dramatic in science, where a theory is revised. More often it is small: a teacher updates an explanation, a manager says "my forecast was wrong," a student notices that a preferred study method did not work, a citizen reads the strongest version of an opposing argument before responding.
At the individual level, humility protects the possibility of learning.
At the institutional level, it protects the possibility of correction.
At the civilizational level, it matters because information systems can now generate support for almost any position with unprecedented fluency.
The discipline is not to doubt everything.
It is to remain able to discover that one was wrong without making the discovery psychologically impossible.
Research references
- Porter T, Elnakouri A, Meyers EA, Shibayama T, Jayawickreme E, Grossmann I. Predictors and consequences of intellectual humility. Nature Reviews Psychology. 2022;1:524-536. https://doi.org/10.1038/s44159-022-00081-9
- Porter T, Baldwin CR, Warren MT, Murray ED, Bronk KC, Forgeard MJC, Snow NE, Jayawickreme E. Clarifying the Content of Intellectual Humility: A Systematic Review and Integrative Framework. Journal of Personality Assessment. 2022;104(5):573-585. https://doi.org/10.1080/00223891.2021.1975725
- Leary MR, Diebels KJ, Davisson EK, Jongman-Sereno KP, Isherwood JC, Raimi KT, Deffler SA, Hoyle RH. Cognitive and Interpersonal Features of Intellectual Humility. Personality and Social Psychology Bulletin. 2017;43(6):793-813. https://doi.org/10.1177/0146167217697695
- Meagher BR. An assessment of self and informant data for measuring intellectual humility. Personality and Individual Differences. 2022;184:111218. https://doi.org/10.1016/j.paid.2021.111218
- Fleming SM. Metacognition and Confidence: A Review and Synthesis. Annual Review of Psychology. 2024;75:241-268. https://doi.org/10.1146/annurev-psych-022423-032425
- Rahnev D. A comprehensive assessment of current methods for measuring metacognition. Nature Communications. 2025;16:701. https://doi.org/10.1038/s41467-025-56117-0