Why people believe what they believe
People bring experiences, values, and group ties to a disagreement. Good evidence matters, but it may not answer the concern that makes a belief important to someone. Understanding that concern helps you ask better questions.
Identity-protective cognition
Some beliefs connect to values, community, or personal experience. Ask what a position means to someone before assuming it is only a factual claim.
On a charged question, criticism can sound personal even when you mean to challenge an argument. Separate the person from the claim and ask which concern your evidence is meant to address.
- Identity-Protective Cognition
- The tendency to process information in ways that protect one's membership in valued social groups, even at the cost of accuracy.
A correction is more useful when it is specific and leaves room for the other person to explain what they still find unconvincing.
Motivated reasoning
People can examine evidence carefully and still favor information that fits a prior view. That tendency is worth watching in yourself as well as in others.
If you assume the other side is reasoning in bad faith, you may miss a valid objection. Ask them to explain which part of your case they reject and why.
Test your own preferred argument too. What evidence would change your mind? If you cannot name any, reconsider how you are evaluating the claim.
When a correction does not land
A correction does not guarantee agreement. Someone may question your source, interpret the evidence differently, or hear the correction as a personal attack. Find out which of those is happening before repeating the same fact more loudly.
A correction that needs a conversation
“You present overwhelming evidence that a conspiracy theory is false. You cite expert sources, show the logical impossibilities, demonstrate the lack of evidence.”
If the person treats every source as part of the cover-up, ask what kind of evidence they would trust. You may not resolve the disagreement in one exchange.
Researchers use “backfire effect” for cases where a correction appears to strengthen the mistaken belief. Reviews have not found a robust general effect. Correct a false claim when you can support the correction, then listen to the actual response instead of predicting one. See the research review.
Cognitive dissonance
Conflicting ideas can feel uncomfortable. That discomfort does not tell you which idea is true, but it can signal a question worth examining.
- Cognitive Dissonance
- The mental discomfort experienced when holding contradictory beliefs, values, or behaviors simultaneously.
Someone can respond to conflicting information by questioning a source, revising a belief, or holding uncertainty while they learn more.
Your job is to make the disagreement clear enough that both sides can test it.
The belief stack
The “belief stack” is a prompt for asking whether a stated position connects to a deeper value or concern. It is a model for inquiry, not a map of another person's mind.
Imagine three possible layers: a specific claim, the values used to judge it, and the experiences or identity that may make those values important. Ask which layer is relevant instead of assuming every disagreement runs that deep.
The belief stack in practice
“Someone argues that the city should lower taxes on small businesses. You do not yet know whether their main concern is jobs, fairness, or the cost of operating locally.”
Ask which outcome matters most to them and why. Then test the relevant claim together: which businesses would qualify, what would the city give up, and how would you measure the result? Do not infer a hidden identity or motive from the position alone.
The layers are questions, not diagnoses. A disagreement may be about facts, priorities, or both. Ask which point the other person disputes before choosing a response. Do not assume evidence is irrelevant because a topic feels personal.
Find the right level
The conditions for mind-change
People can reconsider a claim for different reasons. You can make that easier by stating the disagreement clearly, treating objections seriously, and leaving room for either side to update its view.
1. Safety
Avoid turning a disagreement into a judgment about someone's intelligence or character. A respectful tone leaves more room to examine the claim, although it does not guarantee agreement.
Restate the concern in terms the other person recognizes. Explain what evidence would change your own mind, and ask what would change theirs. This models the same openness you are asking of them. (See Chapter 13: Ethos & Credibilityfor more on credibility.)
2. Face-saving
People can revise a position without being mocked for holding it. Acknowledge when new information changes the question, including when that information challenges your own side.
Face-saving language
“Your opponent has held Position A, and you want them to adopt Position B.”
Instead of: 'Position A is obviously wrong.' Try: 'I used to think something similar, until I encountered this new information. It made me reconsider.'
3. Self-discovery
A question can uncover the assumption behind a claim. Ask one because you want to understand the answer, not because you expect it to trap the other person.
After the answer, state where you agree and where the evidence or reasoning still differs. Sometimes a direct correction is more useful than another question.
Ask what is disputed, test it fairly, and stay open to revising your own position. Agreement is possible, but clarity is a useful result too.
Creating conditions for mind-change
Model exchangeAn exchange that starts by agreeing on what evidence would matter
Opponent
“I've been in this industry for 30 years. I've seen trends come and go. This 'AI revolution' is just another fad that consultants are selling to executives who don't know any better.”
Debater
“Thirty years of experience—that's valuable. You've seen a lot of hype cycles that didn't pan out. What were the tells that those were hype versus real change?”
Creating safety through validationOpponent
“Usually it's all talk and no results. Big promises, expensive consultants, and then nothing actually changes at the ground level. The work still gets done the same way.”
Debater
“That's a good framework. So the test would be: is the work actually changing at the ground level? Not the PowerPoints—the actual work?”
Adopting their criteriaOpponent
“Exactly. Show me the work changing, not the slide decks.”
Debater
“Fair. In my team, a research task now takes less time. I still need to compare the quality of the finished work and check whether that holds across projects. Would those measures help us judge whether this is more than a demo?”
Proposing a shared testDebater
“Does that fit your 'hype vs real' pattern? Or does that feel different?”
Inviting self-discoveryAnalysis: The debater asks what would distinguish a lasting change from hype, then proposes a way to test a workplace observation. The observation is not yet proof. Both sides can examine the same measures before drawing a conclusion.
Reading your opponent
You do not need to read someone's mind to debate well. Listen to the claims they make, ask about their priorities, and notice when the conversation needs clarification.
Emotional investment
Pace, volume, and repeated points can mean many things. Do not treat them as reliable clues to someone's feelings or motives. If a point keeps returning, ask whether your answer addressed the concern they meant to raise.
A personal stake may change which examples matter, but it does not make the person unable to consider evidence. Give the claim and the person the same care. (See Chapter 14: Pathos & Emotion for using emotion responsibly.)
Identity vs. surface
A position can reflect both a factual belief and a value judgment. Ask which part of the case the other person questions. If they accept the facts but weigh the tradeoff differently, more statistics alone will not settle the disagreement. Compare the priorities openly.
What are they really defending?
A stated position may leave its supporting concern unclear. Ask what outcome the person wants to protect before supplying an answer.
The hidden concern
“Someone argues strongly against remote work policies. Their stated position: 'Remote work hurts productivity.'”
Ask what they mean by productivity and which work they worry about. They may point to coordination costs, a specific team's results, or an assumption that needs testing. Do not substitute a guess about their motives for their answer.
Once the concern is clear, you can test relevant evidence and compare the tradeoffs instead of talking past each other.
A question to try
Map your belief stack
Choose one of your strongly-held beliefs—political, personal, or professional. Map its belief stack: (1) What is the surface belief you'd state publicly? (2) What deeper values support this belief? (3) What fears or anxieties does this belief address? (4) How is this belief connected to your identity—who you see yourself as? (5) What would have to change for you to abandon this belief?
Show practice hints
- Choose something you actually care about, not something theoretical—the exercise works better when it's personal
- Be honest about the identity connection. Most important beliefs are tied to how we see ourselves.
- The question of what would change your mind is the most valuable. If you can't answer it, that tells you something about the belief.
Checking saved progress…