The More Certain Someone Sounds, the More Questions I Have

I've noticed that the people who sound the most certain about something aren't necessarily the people who understand it the best. Sometimes they're incredibly knowledgeable, and their confidence is justified with a piece of knowledge from schooling, residencies or experience. Other times, they're repeating something they've heard enough times that they've stopped questioning whether it's actually true, or what scientific studies have been concluded, and if it proves to justify the results. What interests me is how difficult it can be to tell the difference, especially when someone speaks with authority, uses the right terminology, and seems completely convinced of what they're saying. I've become increasingly suspicious of explanations that leave absolutely no room for uncertainty.

Recently, I had a conversation with a professional who made a sweeping statement about human behavior. I questioned it, not because I believed I knew more than the person sitting across from me, but because the claim seemed considerably broader than the evidence I understood. What bothered me wasn't necessarily the disagreement. It was the apparent certainty that accompanied an explanation of something as complicated as human behavior. I went looking for research afterward, and the experience reminded me how often we confuse sounding knowledgeable with actually understanding a subject.

There's a psychological phenomenon that explains part of this, and it's fascinating. In 2002, researchers Leonid Rozenblit and Frank Keil published a paper titled The Misunderstood Limits of Folk Science: An Illusion of Explanatory Depth. Their research examined something most of us probably do without realizing it: we believe we understand how things work in considerably greater detail than we actually do. They called this the illusion of explanatory depth, and once you understand the concept, you start recognizing it everywhere.

The researchers asked people to evaluate how well they understood familiar objects and processes, including things like toilets, bicycles, and helicopters. These weren't obscure subjects requiring specialized scientific training. They were ordinary things people had encountered, used, or heard about throughout their lives. Participants initially rated their understanding relatively highly, but when asked to explain the mechanisms step by step, they often discovered that their knowledge was much shallower than they'd assumed. Their confidence declined after they attempted the explanation.

Think about how ridiculous and revealing that is. Someone can use a toilet thousands of times, understand what it's supposed to accomplish, and confidently claim to know how it works. But ask them to explain the precise mechanical process that allows it to flush, refill, and stop filling, and suddenly the explanation becomes considerably less impressive. Familiarity creates the feeling of understanding, even when the actual details are missing. We mistake knowing what something does for knowing how it does it.

I find that distinction particularly interesting because I've spent a lot of time building complicated systems recently. It's remarkably easy to look at something that appears functional and assume you understand what's happening beneath the surface. Then you begin testing unusual situations, examining how different components interact, and discovering that the system behaves differently from what you expected. Sometimes the most important learning happens when something breaks, because failure forces you to confront the difference between your mental picture of how something works and the way it actually operates.

I've experienced enough of that to know how misleading a successful demonstration can be. Something can perform beautifully under ordinary circumstances while containing problems that only appear when conditions change. You can understand the overall purpose of a system without understanding every mechanism responsible for its behavior. That's not necessarily a problem until you start making decisions based on knowledge you haven't actually verified. The dangerous moment is when familiarity starts masquerading as expertise.

And this phenomenon extends well beyond mechanical objects. We encounter people every day who have strong opinions about education, economics, technology, medicine, parenting, and government. Many of those opinions are based on legitimate experience or careful study. Others are built from fragments of information collected over years of conversations, headlines, and assumptions. Yet both can be delivered with exactly the same level of confidence, and the person listening may have very little information available to distinguish between them.

There's a related body of research that makes this even more interesting. In a 2013 paper, Philip Fernbach, Todd Rogers, Craig Fox, and Steven Sloman examined how asking people to explain the mechanisms behind political policies affected their understanding and attitudes. Participants who were asked to explain how policies worked tended to recognize gaps in their knowledge, and their positions became less extreme on average. Simply asking people to provide reasons for their opinions did not produce the same pattern. Explaining why you support something is apparently a different cognitive exercise from explaining how it actually works.

That distinction is enormous.

I can tell you why I believe a particular policy, technology, or approach is beneficial. I can describe the problem it is supposed to solve, the outcomes I hope it produces, and why the alternative seems worse. But if you ask me to explain the precise mechanisms that connect the proposed solution to the expected outcome, I may discover that my understanding has some fairly substantial holes. That's uncomfortable because it forces me to separate what I know from what I assume. It also seems like exactly the kind of discomfort we should be willing to experience before making decisions that affect other people.

I've started thinking about how rarely we ask people to explain the mechanisms behind their certainty. We ask whether they support an idea, whether they believe something works, or whether they agree with a particular position. Those questions invite opinions, and most people have no shortage of them. But asking someone to walk through how something works, what evidence supports it, and what circumstances might cause it to fail requires a different kind of understanding. It turns a declaration into something that can actually be examined.

I think this is especially important when dealing with professionals. Expertise is real, and I have tremendous respect for people who dedicate years to understanding complicated subjects. A physician, engineer, therapist, attorney, or scientist can possess knowledge that someone outside the field simply doesn't have. But expertise isn't the same as infallibility, and a professional's authority shouldn't make their explanations immune to reasonable questions. In fact, I'd expect someone with deep knowledge to be particularly comfortable discussing limitations, exceptions, and the conditions under which their conclusions might change.

There is a difference between a professional saying, "Based on the available evidence, this is the most likely explanation," and saying, "This is simply how it works." Sometimes the second statement is entirely appropriate because the evidence is overwhelming. Other times, it hides uncertainty that deserves to be acknowledged. The problem isn't confidence itself. It's confidence that exceeds the strength of the underlying knowledge.

Of course, there's an uncomfortable possibility here that I can't ignore. I could be doing exactly the same thing.

I have strong opinions about plenty of subjects, and I'm not exactly known for quietly accepting an explanation that doesn't make sense to me. I enjoy challenging assumptions, questioning inefficient processes, and looking for ways to improve things. Those tendencies can be useful, particularly when they lead me to investigate something rather than accepting it at face value. But they can also create their own version of overconfidence. Being good at identifying weaknesses in someone else's explanation doesn't automatically mean my alternative explanation is correct.

That's probably the most useful thing I've taken from this research. The illusion of explanatory depth isn't a defect limited to arrogant people, incompetent professionals, or strangers arguing on the internet. It's a normal feature of human cognition. We rely on simplified mental models because understanding every detail of everything we encounter would be impossible. Most of the time, those simplified models work well enough. Problems arise when we forget they're simplified and start treating them as complete explanations.

I also think this helps explain why some disagreements become so unproductive. Two people can argue passionately about something while neither has a particularly detailed understanding of the mechanism they're debating. They may possess enough information to defend their positions rhetorically, but not enough to evaluate whether those positions are actually supported. The argument becomes a competition over confidence, language, and persuasion rather than an investigation of what is true. Nobody wants to admit they don't know, because uncertainty is often treated as weakness.

That's an unfortunate habit we've developed. I would much rather have a conversation with someone who can explain what they know, identify what remains uncertain, and acknowledge when new information might change their understanding. That person may sound less impressive than someone delivering absolute declarations, but they're giving me considerably more useful information. They're also demonstrating an ability to distinguish evidence from interpretation, which is something I increasingly value.

The internet has made this problem both better and worse. We have access to an extraordinary amount of research, technical documentation, and specialized knowledge that would once have been difficult to obtain. I can encounter a claim, investigate its origins, compare explanations, and sometimes read the underlying research without leaving my chair. That's incredible. But the same technology also allows confident misinformation to spread rapidly, and the appearance of having access to information can create the illusion that we've actually understood it.

Artificial intelligence introduces another complication. I can ask a system to explain almost anything, and it can produce an answer that sounds coherent, detailed, and authoritative. Sometimes that answer is excellent. Sometimes it contains mistakes, unsupported assumptions, or fabricated details presented with exactly the same confidence. The ability to generate a convincing explanation is not proof that the explanation is correct, and I think that's going to become one of the most important distinctions people need to understand.

It also means that asking better questions is becoming increasingly valuable. When I encounter an explanation, I want to know what evidence supports it, whether there are competing explanations, and what would have to be true for the conclusion to hold. I want to understand whether we're talking about an established finding, a reasonable interpretation, or somebody's personal opinion. And when I find myself feeling unusually certain about something, I probably ought to ask whether I could actually explain it in enough detail to justify that confidence.

None of this means we should distrust everyone or treat every expert opinion as equally questionable. That would be an absurd conclusion to draw from research showing that people sometimes overestimate their understanding. Expertise should carry weight when it's supported by relevant knowledge, sound methods, and evidence. The point is that authority and confidence are not substitutes for those things, and reasonable questions are not automatically acts of disrespect.

I'm nearly forty, and the more I learn, the more subjects I discover that I barely understand. That used to bother me more than it does now. These days, I'm increasingly interested in finding the limits of what I know because those limits tell me where there's something worth investigating. I'd rather discover that I was wrong about something than spend another ten years confidently repeating an explanation that never made sense in the first place.

So when someone tells me something with absolute certainty, I tend to become curious about what sits underneath that confidence. Maybe they have years of evidence and experience supporting their conclusion. Maybe they've considered the exceptions, tested their assumptions, and arrived at an answer that deserves to be trusted. Or maybe they're doing what the participants in Rozenblit and Keil's research did, mistaking the feeling of understanding for the real thing.

I can't always tell which one it is, and I know I'm capable of making the same mistake. But I've become considerably more interested in asking people to explain how something works than in listening to them announce that they know.

And I think the next time I catch myself sounding absolutely certain about something, I should probably be the first person to start asking questions.