Why You Don't Actually Know Your Own Reasons

Ask someone why they took their current job, and watch them answer immediately. Confident, detailed, articulate. Three reasons, maybe four.

Now ask them the same question again in six months, after things have gone sideways.

The reasons will be different. Just as confident. Just as detailed. And if you point out the inconsistency, they'll explain why the second set is the real account — the earlier version having been, obviously, a simplification. This is not a flaw in those particular people. It's a feature of how human cognition works — one psychologists call confabulation — and a paper published in Psychological Review in 1977 documented the mechanism so precisely that neuroscientists are still citing it today.

The uncomfortable part isn't that we sometimes get our reasons wrong. It's that we can't tell when we do.

The study that unsettled a generation of psychologists

Richard Nisbett and Timothy Wilson published "Telling More Than We Can Know: Verbal Reports on Mental Processes" in Psychological Review in 1977. It reviewed a wide body of experiments built around one elegant design: create a situation where the actual cause of a choice is experimentally established in advance, then simply ask the participants to explain their choice, and see whether the stated reason matches the known cause.

In one of the most-cited versions, shoppers were asked to evaluate several items displayed in a row and select the best one. There's a well-documented position effect in these designs — people reliably prefer items on the right side of a display, for reasons that have nothing to do with the items themselves. The researchers knew which item would be chosen before participants made their choice.

When asked to explain their selection, participants gave detailed, sincere accounts. Texture. A subtle quality advantage. Some indefinable sense of superiority. Nobody mentioned position. When researchers asked directly whether the item's location in the display had influenced the choice, participants denied it — some with something approaching mild offense at the suggestion.

In a second study, prior exposure to a word subtly shaped which word participants generated in a word-association task. When asked whether that prior exposure had any effect, they confidently said no. The denial was genuine. They had no awareness that the influence had occurred at all, which meant there was nothing to lie about — they simply couldn't see the mechanism that had shaped their output.

Nisbett and Wilson named this confabulation, borrowing the term from neurology, where it describes patients who generate sincere, detailed false memories without any intent to deceive. The neurological version involves measurable damage. What Nisbett and Wilson found is that a functional version of it appears to be the default operating mode for healthy people explaining their own mental processes.

What is confabulation? Confabulation is the sincere production of inaccurate reasons for one's own choices. Unlike lying, it involves no intent to deceive — the person fully believes the story they're telling. Nisbett and Wilson's key finding was that this isn't rare: it is the brain's standard response when asked to explain decisions that were actually driven by processes below conscious awareness.

Timothy Wilson later developed this framework into a full account of the human mind in Strangers to Ourselves

— arguing that a vast "adaptive unconscious" actually governs most of your behavior, and that you have no direct introspective access to it. The explanations you give for your choices aren't retrieved memories of a process you observed. They're constructed stories assembled from plausible cultural material, then experienced as recollection.

They feel like eyewitness accounts. They're closer to informed guesses.

Why confidence proves absolutely nothing

Here's the finding that should make you sit with it for a moment: across Nisbett and Wilson's studies, confidence in the stated reason correlated almost zero with its actual accuracy.

Not a weak correlation. Near zero.

Which means the feeling of certainty about why you did something is statistically meaningless as evidence that you actually know why. The certainty is a property of the story. It's not a measurement of the story's accuracy.

Daniel Kahneman's work on System 1 and System 2 thinking — unpacked in Thinking, Fast and Slow — maps directly onto this structure. System 1 is the fast, automatic, pattern-matching system that generates most of your judgments before conscious thought has even fully framed the question. System 2 arrives slightly later and does what it was built to do: it constructs a coherent narrative about what just happened and presents that narrative as "your reason."

These two processes are sequential, not simultaneous. By the time you're aware of your conclusion, the actual mechanism that produced it has already finished running. You're not observing the process in real time. You're observing the output — and then reverse-engineering a story about a process you never actually witnessed.

This isn't a design failure. The brain has to make decisions quickly, using pattern recognition that runs faster than deliberate reasoning. The problem isn't that this system exists. The problem is what happens when you treat the story it produces afterward as reliable evidence about the underlying cause — and then use that story to make decisions about what to change.

What your brain is actually building

Here's where this gets interesting rather than just discouraging.

The brain's story-building machinery isn't a bug. It's core infrastructure, and it's genuinely impressive. You are, at a very deep level, a narrative animal. You make sense of your life through story, you remember in story, you predict in story, you persuade in story. The same mechanism that confabulates reasons for past choices is the mechanism that lets you envision future ones.

The problem is specific: the brain appears to have access to the outputs of cognitive processes — the choices, the feelings, the judgments that surface into awareness — but not to the processes themselves. You see the end product of a complicated calculation. The calculation itself runs below the level of conscious access. When you're asked to explain the end product, your narrative machinery does what it always does: it finds the most plausible, socially coherent account and presents it as an eyewitness report.

What you experience is not "I'm constructing a story here." What you experience is remembering. This is why it's so difficult to catch in yourself. Confabulation, from the inside, feels exactly like genuine introspection. The phenomenology is identical.

Jim Rohn used to say that working harder on yourself than you do on your job was the most important investment you could make. That's right. But it assumes your self-knowledge is worth working from. If the raw material of that self-knowledge is largely post-hoc story, you need better data before you can work usefully on the right things. Kahneman's Thinking, Fast and Slow

is the clearest map of this gap — specifically the chapters on System 1 and narrative construction, which show why the story feels so much like observation even when it isn't.

What this is actually costing you

If this were a laboratory curiosity that stayed neatly in academic journals, it would be merely interesting. But Nisbett and Wilson, and the researchers who built on their work, found this pattern operating everywhere that matters.

In relationships, confabulation shows up as confident explanations for conflict that almost never capture the real mechanism. "I'm upset because you didn't text back." That's a real feeling. It's not a fabrication. But that same late text last Tuesday barely registered. Something specific changed this time — and the stated reason, however sincerely offered, is probably not a reliable map to what it was.

In career decisions, it shows up as articulate, detailed rationales for choices that were, in the moment, shaped by factors you wouldn't readily credit. The office's location. The fact that the hiring manager reminded you vaguely of someone you trust. The implied status of a title. These aren't shameful influences — they're human ones. But they don't tend to appear in the version of the story you construct afterward.

Seth Godin has written extensively about this pattern looking outward: when someone offers a reason for a rejection or a decision, that stated reason is almost always a socially safe construction rather than a reliable account of the actual cause. Trying to fix the stated reason is often trying to fix the wrong thing entirely. What's true looking outward is equally true looking inward.

Lars Hall and Petter Johansson's choice blindness research, published in Science in 2005, extended the Nisbett and Wilson findings into even sharper territory: when subjects' selections were secretly swapped for a different option immediately after they chose, most failed to notice — and then generated confident explanations for why they'd preferred the substituted item they had never actually selected. The same self-stabilizing story pattern appeared across political opinions, hiring judgments, and consumer choices in the replications that followed.

The self-improvement industry's quiet blind spot

Most self-improvement frameworks are built, whether explicitly or implicitly, on the assumption that understanding your reasons clearly enough will let you act on them. If you know why you procrastinate, you can stop. If you know why you're stuck, you can get unstuck. If you understand yourself well enough, the path forward becomes clear.

The Nisbett and Wilson research doesn't say that understanding yourself is useless. But it does say that the specific tool most people use to build self-understanding — asking "why did I do that?" and trusting the confident story that emerges — is far less reliable than it feels.

The good news is that Timothy Wilson's later work, and research building on this tradition, found that some forms of introspective questioning are meaningfully less contaminated by confabulation than others. The problem is concentrated in why questions, which almost automatically trigger story-construction. Asking "what specifically changed right before that?" tracks observable behavior rather than narrative interpretation. Asking "under what conditions do I actually do this consistently?" looks for pattern rather than cause. The data tends to be more accurate than the story — because the data doesn't need to be plausible, only true.

This is worth building around deliberately.

How to start today

1. Switch from "why" to "what changed." The single most practical shift from this research. When you want to understand a behavior — why you avoided a conversation, why you kept delaying a project, why a decision surprised you once you made it — don't ask why. Ask what specifically was different this time. That question points at observable conditions rather than constructed stories.

2. Keep a decision journal, not a feelings journal. After significant choices, write out the conditions: what the competing options were, what specifically shifted right before you decided, what condition would have made you choose differently. A dedicated notebook

kept physically separate from your regular notes creates a clear material anchor for this habit — and lets you track behavioral patterns across months rather than relying on memory, which is itself subject to the same confabulation problem.

3. Come back and review. Three months after a decision entry, read what you wrote. Notice where your stated reasons held and where the prediction drifted. The review process is where real self-knowledge accumulates — not in the initial explanation, but in the gap between what you said you'd do and what you actually did.

4. Read the research directly. Dan Ariely's Predictably Irrational

documents specific, real-world cases of hidden drivers operating far below the level of stated reasons — in purchasing decisions, in relationships, in how you respond to pricing and framing. It's one of the most eye-opening accounts of the gap between what we think is causing our behavior and what's actually driving it. Uncomfortable in the best possible way.

5. Build in a skepticism pause. When you notice yourself offering a confident, instantly-available explanation for something you did — or when someone offers you one for theirs — that confidence is not itself evidence. It's just what confabulation feels like from the inside. A one-second pause to ask "is this the story, or is this the cause?" costs almost nothing and occasionally changes everything.

Related read: Why Your Brain Invents Reasons for Your Feelings

Designing your evolution on accurate ground

Designing your evolution — genuinely designing it, rather than crafting a more flattering narrative about who you already are — requires working with your real psychology, not the story you've assembled about it.

Nisbett and Wilson published that paper not to demoralize anyone. They published it because more accurate self-knowledge, even partial, imperfect, incrementally improved self-knowledge, creates a real and compounding advantage. The person who has tested their stated reasons against behavioral evidence over time has something fundamentally more useful than the person with a polished, internally consistent, unverified account of themselves.

Most of the real leverage for change — in habits, in the decisions you make under pressure, in the patterns you keep repeating in relationships — is sitting in the gap between the story you tell about why you do things and what's actually running the behavior.

That gap doesn't close because you're intelligent. It closes because you design practices that produce better evidence than story does.

Related read: Why Your Brain Trusts Stories, Not Facts

What's one belief you hold about why you do something that matters — that you've never actually tested against what you observably did?