Social media is full of the same bugs: “I am the exception,” “I will do the opposite just because,” and “I am righteous, so my attack is allowed”
The easiest way to explain Dostoevsky is this: he wrote down the extremely inconvenient committee meeting inside the human mind and gave every member a character.
There is a part of us that wants profit. A part that wants freedom. A part that wants to be seen as good. A part that hates losing. A part that wants approval from the group. A part that is simply angry.
The problem is that this committee does not publish honest minutes.
The real motive may be “I am furious,” while the public document says, “I am defending justice.” The real motive may be “I do not want to lose,” while the statement says, “I am merely correcting misinformation.” The real motive may be “I want special treatment,” while the explanation says, “My case is objectively exceptional.”
That is why a nineteenth-century Russian novelist can feel strangely like a guide to twenty-first-century social media.
Dostoevsky did not predict social media. Social media simply turned old human tendencies—self-justification, status anxiety, spite, moralization, group loyalty—into a giant searchable public log.
This article uses Crime and Punishment, Notes from Underground, Demons, The Brothers Karamazov, The Idiot, and The Gambler as thought experiments, then checks the parallels against modern psychology.
The goal is not to claim that Dostoevsky secretly invented psychology. Fiction is not a randomized trial. Characters are not a representative sample. But his novels are unusually good stress tests: give a person an idea, push that idea to the limit, and watch what breaks.
0. The thirty-second version: your mind has a defense attorney
Imagine forgetting your homework.
You could simply say, “I forgot.”
But your internal attorney can do better.
“I was tired.” “The teacher did not explain it well.” “There was too much homework.” “Other people did not do it either.” “Does this assignment even have educational value?”
Some of those statements may be true.
The problem begins when “some of this is true” becomes “therefore I have zero responsibility.”
Dostoevsky’s characters make the same leap on a much larger scale.
That is why the books can be funny and uncomfortable at the same time. You laugh at an extreme character, then notice that you perform a miniature version of the same maneuver every week.
1. Crime and Punishment: printing yourself an “exception coupon”
Raskolnikov develops, in simplified form, an idea that ordinary people must live under ordinary moral and legal rules, while extraordinary people may sometimes cross those rules in pursuit of a larger historical purpose.
The important part is not “smart people may commit crimes.”
It is the loophole.
The general rule remains intact. Murder is wrong. Law matters. But perhaps my case is different.
That is the psychological shape worth watching.
Modern psychology does not have a single theory called “the Raskolnikov effect,” but motivated reasoning is close to part of it. Kunda’s classic review argued that motivation can influence how people search for and evaluate reasons, while still being constrained by the need to construct justifications that feel plausible to themselves.
The mind is often less like a journalist asking, “What happened?” and more like a skilled attorney asking, “What is the strongest case for the conclusion my client wants?”
On social media, this can look like:
“I dislike this person” becoming “I am attacking this person for the public good.”
Sometimes the public-good claim is genuinely true. That is exactly why this is difficult.
A complete lie is often easier to detect than a claim made from 60 percent fact, 20 percent emotion, and 20 percent self-protection.
Why does the heart not simply obey the argument?
Because human beings do not optimize a single objective.
We want money. We want freedom. We want to see ourselves as decent. We want social acceptance. We want status. We fear punishment. We care about reputation.
An action can serve one goal and violate another.
That is where cognitive dissonance becomes relevant.
But the popular summary—“if you violate your values, guilt will automatically punish you”—is too simple.
People can change their behavior. They can change their belief. They can label the case an exception. They can shift responsibility to someone else. They can decide the violation was not important.
The mind is not only a guilt machine. It is also a contradiction-management machine.
2. Is there really a moral points card?
Moral licensing is the idea that doing something good can sometimes make people feel more permitted to do something less good later.
Older meta-analytic work suggested a modest average effect.
A 2026 meta-analysis of 115 experiments involving 21,770 participants added a striking qualification: licensing was much stronger when people were observed, while bias-adjusted evidence did not support a clear licensing effect when behavior was unobserved.
So the mechanism may not be as simple as:
good deed → moral points → one free sin.
Reputation may matter.
“I have already demonstrated that I am a good person” can create interpersonal room to behave less consistently.
That possibility is highly relevant to public platforms, where moral statements are visible, archived, and rewarded by an audience.
But this is not a universal law. Moral behavior can also strengthen consistency and lead to more moral behavior later.
The practical lesson is narrower: a positive self-image is not an automatic quality-control system.
3. Notes from Underground: “Even if it is rational, do not tell me I must”
Notes from Underground attacks the dream that human behavior can be perfected by calculating rational self-interest.
The Underground Man’s extreme claim is roughly:
Even if you prove exactly what will make me happiest, I may choose the worse option just to prove that the choice is mine.
Ridiculous? Yes.
Also recognizable.
Psychological reactance describes the aversive response that can arise when people feel their behavioral freedom is threatened.
A 2025 meta-analysis found that highly freedom-threatening language increased anger, negative cognition, and reactance relative to less threatening language, and reactance was associated with weaker persuasion.
“A is probably better for you” and “You must choose A” do not carry the same psychological cost.
The second message includes a threat to agency.
Opposition is not autonomy
This is the most useful correction to the Underground Man.
Someone tells you to choose A. You hate the pressure. So you choose B.
You may feel independent.
But if the other person’s choice of A mechanically caused your choice of B, they are still controlling your behavior.
Real autonomy includes the ability to say:
“I dislike your pressure. Your argument is still correct, so I will choose A.”
Decision rights and factual accuracy are separate questions.
The fact that someone has no right to command you does not prove that their advice is wrong.
4. Demons: when an idea becomes your identity
Demons explores revolutionary ideology, secret groups, vanity, idealism, youth, manipulation, and political violence. The Nechaev affair strongly influenced the novel.
The crude reading is “rational intellectuals become murderers.”
That misses the more interesting problem.
An opinion can become part of the self.
Identity fusion research examines a state stronger than simple group identification: a visceral sense that “me” and “the group” are one.
A meta-analysis of 90 studies and 36,880 participants found a strong positive association between identity fusion and extreme pro-group orientations, though the effects were highly heterogeneous across studies.
Once identity is fused with a group, criticism can change meaning.
“Your policy is wrong” can feel like “You are attacking us.”
A factual disagreement becomes a defensive battle.
Moral disengagement can add another layer.
Instead of abandoning the rule “aggression is wrong,” people can redescribe the current aggression as punishment, necessity, defense, or service to a higher cause.
A meta-analysis of 27 samples involving 17,776 children and adolescents found a positive association between moral disengagement and aggression. That evidence should not be copied directly onto all adults or political groups, but the mechanism is useful to understand.
Good goals do not automatically make every method good.
5. The Brothers Karamazov: freedom is delicious, but it comes with dishes to wash
The Grand Inquisitor episode pushes a brutal question:
Do people really want freedom, or would many rather accept bread, certainty, and authority if someone else carries the burden of deciding?
Freedom feels good.
But freedom also means choosing, risking error, tolerating uncertainty, and accepting responsibility.
That is why “Do whatever you want” is not always supportive.
Modern self-determination theory does not support the idea that humans fundamentally dislike autonomy. A 2024 meta-analysis combining 881 independent samples and more than 443,000 people found that support for autonomy, competence, and relatedness was strongly associated with need satisfaction and subjective well-being, with smaller positive links to performance.
The better lesson is not “people hate freedom.”
It is:
People often want agency and support at the same time.
A strong arrangement is:
You keep the decision right. You also get information, advice, skills, and help.
Freedom and support are not opposites.
“I did not do it directly” does not settle responsibility
The Brothers Karamazov also keeps returning to responsibility.
You did not perform the final act. But you encouraged it. You created the climate. You could foresee the risk. You had power to intervene.
Where does responsibility begin?
The opposite mistake is also possible: “You influenced the situation, therefore everything that happened is your fault.”
Real-world responsibility should be decomposed: foreseeability, degree of involvement, decision authority, avoidability, and causal distance.
Online, both “I only posted it; whatever happened later is not my problem” and “your post caused everything, so you own every consequence” are often too simple.
6. The Idiot: being good does not make you an omnipotent rescuer
Prince Myshkin is unusually kind, nonjudgmental, and sensitive to suffering.
Yet kindness does not magically organize the people around him. Jealousy, shame, romantic obsession, dependence, and self-destruction remain.
The useful distinction is:
good intention is not the same thing as good outcome.
A strong desire to help can accidentally override the other person’s agency. It can solve the helper’s discomfort rather than the other person’s problem. It can ignore third parties. It can turn into “I must save this person.”
The quality check for compassion is therefore not merely, “Am I kind?”
It is:
Does this protect the person’s interests, agency, and long-term options?
Goodness is not a license to become someone else’s administrator.
7. The Gambler: “I know this is bad” does not equal “I can stop”
You lose. You want to recover the loss. You bet more. You lose more. Now stopping feels like making the loss permanent.
This is loss-chasing.
A 2023 systematic scoping review distinguishes within-session chasing—continuing or intensifying play after losses—from returning later to recover earlier losses.
The key is that the objective changes.
At first: make money. After a loss: erase the loss. After humiliation: do not end as the loser.
The same shift can occur in an online argument.
The original goal is to clarify a fact.
Then someone mocks you.
The goal becomes: “Do not let this thread end with me looking defeated.”
Seventeen replies later, the information problem ended hours ago. What remains is emotional loss-chasing.
Knowing that a behavior is irrational and being able to stop it are different skills.
8. Social media is not a Dostoevsky factory; it is a magnifying glass for old bugs
Calling social media pure evil would be another 100–0 claim.
A 2024 Annual Review of Psychology article describes evidence that social media can amplify moral outrage, status seeking, punishment, and intergroup conflict—but can also amplify support, prosocial behavior, and collective action.
The platform is better understood as an accelerator.
It gives large speakers to behaviors that capture attention.
Social media is also a comparison machine
A 2023 meta-analysis of 48 studies and 7,679 participants found that exposure to upward-comparison targets on social media had a small average negative effect on self-evaluations and emotions relative to controls or downward comparison, with an overall effect of g=-0.24.
That does not mean “seeing successful people always harms you.”
Some people learn. Some feel inspired. Context matters.
Meta-analysis describes an average tendency, not an individual destiny.
9. Do smart people escape self-justification?
Not automatically.
Reasoning ability can help you discover why you are wrong.
It can also help you manufacture a better brief for why you are right.
That is the uncomfortable implication of motivated reasoning.
The relevant question is not, “How long did I think about this?”
It is:
Did I search as hard for evidence against my preferred conclusion? Would I use the same rule if an enemy did the same thing? What evidence would make me change my mind?
Thinking harder is not the same as debugging your thinking.
10. What changes when other people are watching?
Public reasoning has extra costs.
Reputation. Group loyalty. Status. Consistency with old posts. Fear of giving opponents a screenshot.
Changing your mind may improve accuracy while damaging face.
That makes public correction expensive.
The 2026 moral-licensing meta-analysis is relevant here because observation strongly moderated the effect. Moral behavior is partly embedded in reputation management.
A useful response is to make revision cheap from the beginning.
Say:
“With the evidence available now…” “This part is strong; this part is uncertain.” “This is my current best guess.” “I would change my view if X appeared.”
Those are not cowardly escape hatches. They are update channels.
11. Eight practical checks for living with human bugs
1. Separate facts, values, and feelings
“He said X” is a factual claim. “X was disrespectful” is an evaluation. “I am angry” is a feeling.
All three matter. They are not interchangeable.
2. Look for the exception coupon
“If anyone else did this it would be wrong, but my case is different.”
Sometimes cases really are different.
If so, the exception should be explainable in a rule that could also apply to another person.
3. Swap the actor
If someone you dislike did exactly the same thing, would you judge it the same way?
“Courage” for allies and “recklessness” for opponents may reveal that identity is doing some of the work.
4. Predefine disconfirmation
What observation would change your mind?
If the answer is “nothing,” the claim has become immune to evidence.
5. Consider the opposite once
Classic experiments found that explicitly generating a plausible opposite explanation could reduce some judgment biases.
It is not magic.
But it is more concrete than chanting, “I should be objective.”
6. Separate autonomy from accuracy
“I do not want you deciding for me” concerns authority.
“Your proposal is wrong” concerns content.
Do not merge them.
7. Treat sudden 100–0 language as a warning light
Always. Everyone. Completely. Obviously. Only one possible answer.
Sometimes reality really is one-sided.
But when those words suddenly multiply during anger, check whether resolution has dropped.
8. Restore the original objective
In an argument, meeting, relationship conflict, or investment decision, ask:
“What was I trying to improve before winning became the goal?”
That question catches many loss-chasing loops.
12. Do not turn psychology into a superior-sounding weapon
Once you learn these concepts, a new temptation appears.
“That person has cognitive dissonance.” “This is reactance.” “They only want approval.” “They are morally disengaged.”
Now psychology becomes a stick.
Worse, the speaker can feel scientific while performing exactly the self-justification they are describing.
Instead of diagnosing strangers, analyze structure.
What factual claim is being made? What value judgment is being added? What does the person stand to lose? Which freedom feels threatened? Is group identity involved? How costly would it be for them to admit error?
Understanding a mechanism is not the same as agreeing with the conclusion.
13. What the research does not prove
Dostoevsky was not a psychologist, and novels cannot prove population-level laws.
Social-media users are not all trapped by self-justification; platforms also support learning, cooperation, and mutual aid.
Identity-fusion effects are strong on average but highly heterogeneous.
The moral-disengagement meta-analysis cited here studied ages 8–18, so it should not be generalized mechanically to all adults.
The upward-comparison effect is an average effect, not a prediction for every person.
And research on dishonesty itself has replication problems. The famous 2008 Mazar, Amir, and Ariely paper on self-concept maintenance received an Expression of Concern in 2024; a large preregistered multi-lab replication did not reproduce the central effect from Experiment 1. This article therefore does not use that study as a foundation.
Good research hygiene requires keeping the limitations attached to the exciting result.
14. Conclusion: the goal is not to become a perfectly rational person; it is to remain correctable
Dostoevsky’s characters are not interesting because they are simply evil.
They are intelligent. They suffer. They want freedom. They want love. They want justice. They want dignity.
Those motives collide.
So the practical lesson is not:
remove emotion, be purely rational, be morally perfect.
It is more modest and more useful.
Question your personal exception coupon. Separate autonomy from opposition. Separate justice from a permission slip to attack. Separate good intention from good outcome. Separate recovering a loss from achieving the original goal. Separate loyalty to a group from the truth of a claim.
And replace:
“I have no bias”
with:
“I know some of the conditions under which I become easier to bias.”
For a child, the whole article can be reduced to five lines:
- Separate what you feel from what is true.
- Check whether you invented a special rule just for yourself.
- Ask whether you would judge an enemy the same way.
- Doing the opposite just because someone told you what to do is not necessarily freedom.
- A person who can correct a mistake is stronger than a person who must pretend never to be wrong.
Open social media and you will probably meet a few Dostoevsky characters today.
The inconvenient part is that, on some days, one of them is you.
Sources / research basis
- Dostoevsky, Notes from Underground, Project Gutenberg: https://www.gutenberg.org/ebooks/600
- Van Bavel et al. (2024), Social Media and Morality. DOI: 10.1146/annurev-psych-022123-110258
- Kunda (1990), The Case for Motivated Reasoning. DOI: 10.1037/0033-2909.108.3.480
- Psychological reactance meta-analysis (2025): https://academic.oup.com/hcr/article/52/1/38/8178818
- Slemp et al. (2024), basic psychological needs support meta-analysis. DOI: 10.1037/pspi0000459
- Varmann et al., identity fusion meta-analysis. DOI: 10.1080/10463283.2023.2190267
- Gini et al. (2014), moral disengagement and aggression. DOI: 10.1002/ab.21502
- Rotella et al. (2026), observation and moral licensing. DOI: 10.1177/01461672251345512
- McComb et al. (2023), upward social comparison meta-analysis. DOI: 10.1080/15213269.2023.2180647
- Lord et al. (1984), Considering the opposite. DOI: 10.1037/0022-3514.47.6.1231
- Banerjee et al. (2023), loss-chasing review. DOI: 10.1016/j.neubiorev.2023.105377
- Expression of Concern on Mazar et al. (2008), 2024. DOI: 10.1177/00222437241285882
