Most of what you know, you learned from others. You have not personally checked that the earth is about 4.5 billion years old or that Napoleon existed. You rely on teachers, books, scientists, journalists and friends. This is testimony, and it raises hard questions: when should we trust others, whom should we trust, and what should we do when trustworthy people disagree?

Two views of testimony

Reductionists, often traced to David Hume, hold that testimony is justified only by our own evidence that a source is reliable, built up from experience of how often people tell the truth. Anti-reductionists, following the Scottish philosopher Thomas Reid (1710 to 1796), hold that we may accept what others tell us by default, unless we have reason for doubt. Reid argued that humans are naturally disposed both to tell the truth and to believe what they are told, and that a child could never learn anything if it had to check every source first. Both views agree that trust can be defeated by evidence that a source is lying, mistaken or biased.

Whom to trust: experts

Modern life requires trusting experts on matters we cannot judge directly. How can a non-expert tell a real expert from a confident fake? Useful signs include:

• Track record: have their past claims and predictions held up?
• Relevant expertise: is the question actually in their field? A Nobel prize in chemistry does not make someone an expert on economics.
• Agreement among experts: what do most independent specialists in the field conclude? A consensus built from many lines of evidence is far more reliable than any individual.
• Interests: does the source gain from your believing them?
• Openness: do they explain their reasoning, acknowledge uncertainty and correct errors?

Trusting experts is not blind deference. It is a reasoned judgment that someone is more likely to be right than you are on a question where you lack their training and evidence.

Disagreement among peers

Suppose you and a friend, equally informed and equally careful, reach different conclusions. Philosophers call such a person an epistemic peer. What should you do?

• Conciliationists argue you should lower your confidence and move toward the other view, since you have no special reason to think you are the one who erred.
• Steadfast views hold that you may sometimes keep your belief, especially if you assessed the evidence correctly, since the evidence, not the headcount, is what matters.

A common example: you and a friend split a restaurant bill and calculate different amounts. Most people agree you should both become less sure and recheck. Harder cases arise in politics, religion and ethics, where reasonable people disagree persistently.

Intellectual humility

Intellectual humility means being aware of the limits of your knowledge and open to being wrong. It is not having no views, or treating every opinion as equally good. It means holding beliefs with the confidence the evidence supports, seeking out the best arguments against them, and being willing to change your mind. John Stuart Mill wrote in On Liberty (1859): “He who knows only his own side of the case, knows little of that.”

Information environments

Our testimony comes through communities and media we choose. Where people hear only like-minded voices, mistakes go unchallenged; where sources teach audiences to distrust all outsiders, even accurate correction is dismissed. Choosing varied, reliable sources, and noticing which never admit error, is part of epistemic responsibility.

Bringing it together

Epistemology does not hand us certainty. It gives us better habits: aim at truth, proportion belief to evidence, check reasons as well as results, value science's self-correction, and trust others wisely. A democratic community needs these habits in its citizens.

Worked example: the cracked foundation

Sam notices a long crack in the concrete foundation of his house. Three people offer opinions:

• A neighbor who has owned several houses says cracks like that are normal and harmless.
• A foundation repair company offers a free inspection, says the crack is serious, and quotes $18,000 for repairs.
• A licensed structural engineer, paid a flat fee and not selling repairs, says the crack is mostly normal shrinkage but recommends monitoring one section and fixing drainage nearby.

How should Sam weigh this testimony? Relevant expertise: the neighbor has experience but not training; the engineer's field is exactly this question. Interests: the repair company profits if Sam believes the crack is serious, which does not make them wrong but is a reason to want an independent view. The engineer has no stake in the answer. Openness: the engineer explains her reasoning, says what would change her mind (the crack widening), and suggests a way to check.

The reasonable course is to give the engineer's view the most weight while following her advice to monitor the crack, which turns uncertainty into a test. If the crack grows, Sam should update.

Notice what Sam did not do. He did not try to become a structural engineer himself, and he did not simply pick the answer he liked best. He judged sources by expertise, incentives and openness. That is epistemology in ordinary life: not certainty, but reasonable trust, held with humility and open to new evidence.