On Christmas Eve, 1954, Dorothy Martin sat in her Chicago living room surrounded by followers who believed aliens would land on her sidewalk. When the aliens didn’t arrive, Martin’s followers — according to the psychologist Leon Festinger, who had infiltrated the group — doubled down, believing harder than before. That’s the story told in When Prophecy Fails, a foundational text of cognitive dissonance theory. It became the canonical example of what people do when reality contradicts belief: they dig in.

But recent archival research suggests the story was wrong — that the researchers shaped the data to fit the theory. And the real story of what happened to Martin and her followers reveals something more interesting and more useful about human psychology: the people who actually grow from contradictions aren’t the ones who double down, and they aren’t the ones who quietly rewrite their histories to erase the discomfort. They’re the ones who can hold two contradictory truths about themselves at the same time, without collapsing into either one. In systems engineering, where I’ve spent my career, we have a principle for this: a discrepancy between predicted and actual telemetry from a Mars rover isn’t a problem to be explained away. It’s information. And the worst thing you can do is resolve it prematurely by picking the explanation that makes you most comfortable.

The Theory That Shaped How We Think About Contradiction

Leon Festinger’s theory of cognitive dissonance, formalized in 1957, proposed that people experience genuine psychological discomfort when they hold conflicting beliefs or when their beliefs clash with reality. The theory predicted that people would go to remarkable lengths to reduce this discomfort. They’d change their beliefs, reinterpret the evidence, or seek out social support to prop up the version of reality they preferred.

The idea became one of the most influential concepts in social psychology. It shaped how we understand everything from buyer’s remorse to political polarization. A Psychology Today analysis of political hypocrisy traces a direct line from Festinger’s original framework to how people process contradictory information about leaders and causes they support.

But here’s what the theory, as commonly understood, gets subtly wrong about personal growth: it treats dissonance primarily as something to be resolved. Reduced. Eliminated. The whole vocabulary of the field reflects this. Researchers study “dissonance reduction strategies.” A Frontiers in Psychology paper framing dissonance through the lens of emotion regulation treats the experience as a form of negative arousal requiring management. The assumption running through decades of research is that dissonance is a problem to solve, not a state to inhabit.

What if that assumption is backwards for the people who actually grow?

What the Seekers Actually Did

Thomas Kelly’s recent paper, based on Festinger’s unsealed archives at the University of Michigan, argues that the researchers manipulated both the group and the data. Kelly found that the Seekers didn’t actually double down as Festinger described, but instead showed signs of abandoning their beliefs when predictions failed. The group dissolved. Dorothy Martin herself eventually rewrote her personal history, distancing herself from the failed predictions entirely.

Even Martin’s most devoted follower, Charles Laughead, made his famous declaration of faith not spontaneously but in response to prompting by Henry Riecken, an undercover researcher who had deliberately provoked a crisis of faith and then begged Laughead to reassure him. According to Kelly, the key moments in the narrative — including Martin’s despair and Laughead’s affirmation — were influenced by Riecken’s interventions.

I should note: I’m reporting Kelly’s findings here, not evaluating them as someone with expertise in social psychology methodology. But the pattern he describes — researchers smoothing out messy data to fit a clean theory — is something any engineer recognizes. We have a name for it in systems work: confirmation bias in anomaly resolution. You see what your model predicts, not what the telemetry actually shows.

This matters because the original narrative has shaped seventy years of how we think about people confronting contradictions. The received wisdom said: when reality contradicts belief, people dig in. Kelly’s reading of the evidence suggests something more ordinary and more human. People confronting major contradictions often just… let go. But letting go is not the same as growing.

The Flattering Version Problem

What Martin actually did, according to Kelly, is instructive. She didn’t hold the contradiction between her predictions and reality. She didn’t sit with the discomfort of being someone who genuinely believed aliens were coming and being someone who was wrong. Instead, she rewrote the story. She told an interviewer in the 1980s that her psychic powers came not from a tingling sensation one morning but from a car accident, cancer, and a miraculous healing by Jesus Christ. The failed prophecy and Christmas message were omitted entirely from her later narrative.

Martin picked the more flattering version.

This is the pattern I see everywhere, not just in the psychology literature but in ordinary self-understanding — and very much in engineering organizations. People confronted with two contradictory truths about themselves almost always resolve the tension by choosing the story that makes them look better, or at least makes them feel more coherent. The engineer who failed at a critical design review becomes, in her own retelling, the person who was ahead of her time. The manager who missed a fatal flaw in a test protocol becomes the person who was let down by the process.

These aren’t lies, exactly. They’re the natural result of a mind that treats contradictions as intolerable. And for most people, they work well enough.

The people who actually grow are doing something harder. They’re holding both versions simultaneously, without collapsing into either one.

What It Looks Like to Hold Two Truths

In systems engineering, you learn early that every component has multiple failure modes, and many of them are contradictory. A thermal protection system can fail because it gets too hot or because it gets too cold. A structural member can fail from too much stress or from fatigue caused by repeated loading well below its rated capacity. The system works not because you eliminate all failure modes but because you understand that contradictory failure modes coexist, and you design around all of them at once.

People who grow psychologically are doing something similar. They’re holding the reality that they are both competent and sometimes deeply wrong. Both caring and sometimes negligent. Both brave in one context and avoidant in another. These aren’t contradictions to be resolved. They’re the actual topology of a human personality.

Consider someone who can’t accept a compliment because praise contradicts the identity story they’ve built. That person has committed to a single version of themselves — usually the self-critical one — and experiences praise as dissonance. The resolution isn’t to switch to a purely positive self-image. It’s to hold both: I struggle with this, and I also did something well. Both are true at the same time.

That sounds simple. It is almost unbearably difficult in practice.

Why Coherence Is the Enemy of Growth

The deep problem with dissonance reduction, as a life strategy, is that it prioritizes internal coherence over accuracy. When you smooth out contradictions in your self-narrative, you feel better. Your story makes sense. But you’ve also lost information.

This is something I understand viscerally from my years in mission operations. Unresolved anomalies keep engineers awake at night — and they should. The engineers I most respected at JPL were the ones who could carry an unresolved discrepancy in their minds for days or weeks without forcing a premature conclusion. They treated the discomfort not as a problem to be solved but as a signal to keep paying attention.

Festinger’s original framework assumed that dissonance is aversive and that people are motivated to reduce it. That’s mostly true. But research on dissonance as an emotion regulation challenge has begun to recognize that people differ substantially in their tolerance for the experience. The people who can sit with contradictory self-knowledge for extended periods tend to make better decisions over time. They have access to more information about themselves, because they haven’t pruned away the unflattering data. The people who need to resolve it immediately make faster decisions that feel more confident but are built on an incomplete picture.

The Replication Problem and What It Reveals

Kelly’s critique of When Prophecy Fails is part of a broader reckoning in social psychology. The New Yorker has examined whether cognitive dissonance is even “a thing” in the way it’s been popularly understood, noting that the concept has become so elastic it can explain almost any behavior, which means it may explain nothing at all.

Subsequent studies of religious movements failed to replicate Festinger’s core finding that failed prophecy leads to increased proselytization. Kelly’s review suggests the expected outcome is more mundane: the cult typically dissolves. People don’t double down. They quietly walk away and reconstruct their histories.

But this replication failure is itself informative. What it suggests is that the popular understanding of cognitive dissonance overstates how committed people are to their existing beliefs and understates how flexible (and how revisionist) human memory and self-narrative actually are. We don’t cling to contradictions. We dissolve them, usually by editing the past.

The people who grow are the ones who refuse to make that edit.

The Engineering of Self-Knowledge

When I sat in mission control at JPL, watching telemetry come in from a rover on Mars, the most important skill wasn’t understanding what was happening. It was understanding what wasn’t matching your model of what should be happening. A discrepancy between expected and actual telemetry is information. The worst thing you can do is explain it away prematurely.

The same principle applies to self-knowledge. When your behavior contradicts your self-image, the discrepancy is information. You can explain it away with excuses — you were tired, it wasn’t the real you, everyone does that sometimes. Or you can hold it, examine it, and let it update your model.

I am someone who values patience, and I lost my temper yesterday. Both true. I am someone who works hard, and I avoided a difficult task for three weeks. Both true. I am a good friend, and I failed to show up when someone needed me. Both true.

The growth happens not when you resolve these into a single coherent story but when you can hold them without panic. Without the urgent need to explain, justify, or rewrite.

Why the Flattering Version Is So Seductive

The pull toward the flattering version isn’t just vanity. It serves a real psychological function. Research on information bubbles and well-being suggests that exposure to contradictory information about oneself or one’s beliefs creates genuine discomfort that can affect decision-making and mental health. There’s a reason people curate their information environments. Contradiction hurts.

And the flattering version isn’t always wrong. Sometimes you really were ahead of your time. Sometimes the absence really was in service of something important. The problem isn’t that positive self-narratives are false. The problem is that they’re incomplete.

Dorothy Martin wasn’t wrong that she had psychic experiences (at least from her subjective perspective). She was wrong about what those experiences meant, and when reality demonstrated that, she didn’t expand her self-understanding. She narrowed it. She cut away the parts that didn’t fit the new, more respectable story.

This is what most people do, most of the time. And it works, in the sense that it reduces discomfort and allows them to function. But it doesn’t produce growth. It produces a smaller, tidier version of the self.

The Contradiction as Signal

There’s a connection here to something I think about often on my morning hikes in the San Gabriel Mountains. You look up at the sky and feel both the immensity of what’s out there and the specificity of the ground under your feet. Both are real. Neither cancels the other. That coexistence of vastness and particularity isn’t a bug in human cognition. It’s precisely where curiosity, understanding, and genuine growth originate.

The same is true of self-contradictions. The person who recognizes they are both generous and selfish, both courageous and afraid, both disciplined and lazy (depending on the domain, depending on the day) has more raw material to work with than the person who has edited themselves down to one clean adjective.

Growth requires inputs. Contradictions are inputs. Resolving them prematurely is like throwing away data you haven’t analyzed yet.

What This Looks Like in Practice

Holding contradictory self-knowledge without panic doesn’t mean passivity. It doesn’t mean shrugging and saying, “I contain multitudes.” That’s just another form of resolution, a poetic frame that makes contradiction comfortable rather than productive.

What it actually looks like is more uncomfortable than that. It looks like knowing you made a selfish choice and not immediately constructing a justification for why it was actually the right thing to do. It looks like recognizing that you’re good at your job and also that a particular failure was genuinely your fault, not the result of circumstances beyond your control.

It looks like what Laughead couldn’t do. Laughead told Riecken he couldn’t afford to doubt and had to believe. That’s the voice of someone for whom contradiction is existentially threatening. The stakes are too high for ambiguity.

For most of us, the stakes are lower. We can afford to doubt. We can afford to hold two versions of ourselves in mind simultaneously and let the tension between them teach us something. The question is whether we’re willing to tolerate the discomfort.

The Cost of Always Resolving

Festinger’s researchers, ironically, demonstrated the very phenomenon they were studying. Faced with evidence that their subjects were abandoning their beliefs rather than doubling down, the researchers edited the narrative. They emphasized the data that fit their theory and downplayed or omitted what didn’t. They picked the more flattering version of their own findings.

Kelly’s archival research suggests that paid observer Riecken actively manipulated group dynamics to produce the behavior the researchers expected to see. When reality didn’t match the theory, they didn’t update the theory. They massaged the reality. This is dissonance reduction in its purest form, performed by the very people who named the phenomenon.

The cost was seventy years of partially wrong assumptions about how humans handle contradictions. Not catastrophically wrong, since people do experience dissonance and do try to reduce it. But wrong in the specifics, wrong about how rigid people really are, and wrong about what happens when beliefs collide with reality in a group setting.

The cost of always resolving your own contradictions is similar. Not catastrophic. But persistently distorting. Over time, the gap between who you think you are and who you actually are widens. The edits compound. The flattering version drifts further from the real one.

And the person who could have grown from holding the tension becomes the person who is comfortable, coherent, and stuck.

Tolerance as a Skill

The good news, if you can call it that, is that tolerance for self-contradiction appears to be something people can develop. It isn’t a fixed trait. Like any form of discomfort tolerance, it responds to practice.

The practice is deceptively simple. Notice when you’re constructing a justification. Notice when you’re rewriting a memory to make yourself look better. Notice when you’re reaching for “but” to connect two truths about yourself that don’t need connecting. “I failed at that, but I learned from it” is already a resolution. Sometimes the more honest version is just: “I failed at that.” Full stop. And I’m also someone who does good work. Also full stop. No conjunction required.

The two statements can coexist without a bridge between them. Building the bridge is what flattens the contradiction into a growth narrative that may or may not be earned. Sometimes you did learn from the failure. Sometimes you didn’t. Holding both possibilities open is harder than it sounds.

But the people who can do it are the ones who actually change. Everyone else just becomes a more polished version of whoever they decided to be when the contradictions first became uncomfortable.

In mission control, we never closed an anomaly by choosing the most reassuring explanation. We closed it when we understood what had actually happened — or we left it open, flagged and tracked, an honest acknowledgment that the system was more complex than our model of it. The best engineers carried those open items not as failures of understanding but as invitations to keep looking. The discrepancy between the model and the telemetry was never the enemy. The enemy was the premature comfort of a clean resolution.

The same is true of you. The contradiction between who you believe yourself to be and how you actually behaved last Tuesday isn’t a threat to your identity. It’s the most useful data you have. The question has never been whether you contain contradictions — you do, everyone does, and Kelly’s archival work suggests we’ve overestimated for seventy years how much people fight to deny that fact. The real question is what you do with the tension. You can resolve it, quickly and cleanly, the way Dorothy Martin resolved hers: rewrite the story, omit the parts that don’t fit, and move on with a tidier but smaller self. Or you can hold it. Sit with the discomfort. Let both versions stay on the screen at the same time, the way a good engineer lets contradictory telemetry sit until it teaches you something you didn’t know to ask about. That’s where the growth is. Not in the resolution, but in the willingness to leave the anomaly open.

Photo by MART PRODUCTION on Pexels