The inner critic: turning accusation into observation

Epigraph

“You are but an appearance”.

Epictetus, Enchiridion, §1, translated by Elizabeth Carter. The Russian rendering is an original editorial translation. Text.

What this article is about

In an illustrative scene, an editor receives an email: a published article contains an incorrect date. Correcting it takes a few minutes. Yet the next two hours go into a different task: proving to himself that someone who makes such a mistake has no right to write at all. Here we encounter the inner critic—a convenient name for a way of speaking to oneself, rather than a separate person living inside us. It can identify a real inaccuracy while turning that inaccuracy into a verdict on the whole person.

This article examines that transition. How can we preserve a useful correction without accepting accusation as fact or using support to conceal our responsibility? We compare a contemporary study of self-criticism with two philosophical practices of self-examination, preserving their different foundations. We then follow observation into action, from describing an episode to repairing its consequences and changing a working process. The aim is an accessible test of a particular judgment. You do not first have to feel self-love, become completely calm, or find someone who can convincingly declare you a good person.

I. Accusation claims more territory than it owns

An accusation often begins with an observation but quickly exceeds its boundaries. “I got the date wrong” describes a verifiable action. “I am inattentive” proposes an explanation that still needs clarification. “I cannot be relied upon” transfers the judgment of one action onto an entire character and every future relationship. These three sentences have a similar emotional tone but require different amounts of evidence. The last requires information that the editor's email does not contain.

An internal verdict usually does not specify what result would count as sufficient. If the editor corrects the date, he may still hear that a decent person would never have made the mistake. If he checks the next article in advance, the voice replies that he is merely afraid of exposure. This construction is closed to new information. A correction cannot change the verdict because the verdict no longer concerns the text; it concerns the person's right to participate in the work.

Observation has a different grammar. It has a time, place, object, and boundary: in this version, in this paragraph, after this source. The words “always” and “never” are not prohibited, but they must survive examination. Sometimes there really is a pattern: someone has failed to check dates after editing on three occasions. Naming those occasions and the common stage of the process is more useful than declaring a lifelong personality trait. A precise description can be severe. A gentle formulation is not automatically a true one.

In an illustrative family conversation, the difference becomes clearer still. “I interrupted my sister twice” permits returning to the conversation and listening. “I am a bad brother” gives the sister another obligation: to console the person who did not listen to her. A general verdict can appear morally serious while putting repair further out of reach. This is an editorial observation about the structure of a situation, rather than a diagnosis of the speaker. In an actual case, we still need to establish what happened, how the participants experienced it, and what they want next.

II. A useful signal and its delivery are different things

We can retain the signal that a standard has been violated while changing how we treat ourselves. Doing so does not require proving that self-criticism is always harmful. In Gilbert and colleagues' 2004 study, self-reports from 246 female students distinguished forms and functions of self-criticism, including attempts to improve and desires to punish oneself. This is a limited study; it neither establishes the cause of your experience nor tests the editorial practice proposed here.

The distinction matters in an ordinary discussion. The same information—a missed deadline—can be delivered in several ways. A correction names the commitment and its consequences. A threat promises to deprive someone of dignity if it happens again. Mockery adds a picture of their supposed inferiority. A client's right to require delivery remains in all three cases; humiliation does not make the demand more precise or prove it justified.

Imagine a manager telling an employee that the report is due tomorrow and two figures do not yet agree. The employee might internalize this working instruction and calmly check the calculation. Or they might add another sentence: adults do not need reminders. It is this addition that turns an instruction about work into a demand to prove adulthood. When examining the episode, it helps to separate what was heard from what was added, even if both now sound like one voice.

Sometimes it is useful to extract the content of an internal remark literally. What can be checked here? What does it require? Who experienced the consequences? If the insulting description disappears, is there still a task? From “you are lazy and have failed to reply again,” for example, we can extract the unsent reply, its deadline, and the need to inform the other person. But “lazy” does not explain why the reply was delayed: missing information, time, skill, agreement with the promise, or attention. Further examination may reveal a difficulty or confirm avoidance. Observation does not appoint an excuse in advance. It leaves open the possibility of establishing which explanation the circumstances actually support.

III. Acknowledging a thought does not mean signing it

When an accusation sounds convincing, arguing with it can turn into an endless internal trial. The prosecutor presents the mistake, the defense lists achievements, and the prosecutor declares those achievements accidental. The date remains uncorrected. An alternative move is to acknowledge the thought and then examine its status: right now, I am thinking that I am unfit for this work. That describes a present judgment; it does not prove unfitness.

In §1 of Epictetus's Enchiridion, we find a demand to examine an impression before following it. We borrow a limited philosophical operation: initial persuasiveness does not remove the need for consideration. The Stoic aim concerns the nature of goodness and freedom; our working practice does not reproduce that entire system or establish its psychological effectiveness.

“I am having a thought” can itself become a way to decide nothing. Then examination loses its purpose. If a fact is confirmed, we cannot indefinitely keep it at the level of an impression merely because it is unpleasant. The email contains one date, the source another, and the saved version shows the error. Uncertainty is small here. The claim that the person is universally unreliable remains far less well supported. Separating these statuses means acknowledging what is known faster and declaring what is unknown more slowly.

This difference also helps when accusation feels physically convincing. Shame, tension, and the urge to hide may accompany a thought, but they do not by themselves prove its content. Conversely, calmly offering an excuse does not establish that the excuse is true. The action and its context are examined, rather than just the strength of feeling. A reader may have no clear name for an experience and still establish what was sent, what was promised, and what needs correction. This article does not propose suppressing a feeling. It proposes declining to make feeling the sole witness in a question for which emails, deadlines, versions, and another participant's words are also available.

IV. Self-examination can be specific and morally serious

The assumption that abandoning a global verdict must lead to a lack of principles deserves examination. In Analects, I.4, Zengzi describes daily examination of faithfulness in business, sincerity with friends, and practicing what one has been taught. Its Confucian context includes moral cultivation and social roles; it is not an ancient version of a modern questionnaire.

The example matters precisely because it is bounded. We can ask difficult questions without asking every evening whether we have a right to exist. Which consequence of my action requires repair? A response may require a specific change. Yet the question is definite enough for tomorrow to produce a different answer. If we instead ask whether we are sufficiently good people, the criterion can easily shift with our mood.

In an illustrative example, a teacher notices that he answered a student sharply. Moral work includes acknowledging the harshness, explaining without removing responsibility, and changing how he speaks in class. It does not end with giving himself an unpleasant label. The student may not want to participate in his self-examination at all. Repair must respect this boundary: a conversation can be offered, but the student cannot be required to issue a certificate of improved character.

When comparing traditions, we should preserve the difference between self-evaluation and moral examination. A contemporary working procedure may orient itself toward agreements, safety, and mutual respect. The Confucian text has its own system of relationships and virtues. We see a structural resemblance—examining specific performance—without declaring that the sources of obligation are identical. Nor can a few familiar lines establish that every culture has always preferred one practice. Historical societies could also sustain harsh forms of shame. Comparison helps make a question more precise; it does not replace the reader's independent choice of which demands they recognize as their own today.

V. Turning a verdict into an account of an episode

The first translation of an accusation is a description that another person could check using the same information. For the editor it looks like this: in the version sent that morning, the event's year was wrong; a reader's email arrived after publication; the source link was in the notes, but the final date was not checked separately. This account already contains more useful information than a general epithet. It identifies where change may be possible.

Not everything internal should become an external record. “I wanted to finish sooner” remains a report of one's own intention. “My colleague decided I was stupid” is a guess about someone else. “After the email, I did not open the document for half an hour” describes an action that can be compared with the work history. One episode contains different kinds of information; they should not receive the same confidence simply because they appear next to one another.

The limits of an explanation matter in this translation. If someone writes that the mistake resulted from tiredness, we still need to ask what supports that connection. Tiredness may have been present, yet the date might already have been wrong in the original notes. Sometimes adequate process improvement does not require finding the single internal cause: a date check before sending can be added. But if the same obstacle returns, systematic observation may help examine the hypothesis more seriously.

The result should remain ordinary human speech. “I made a mistake in this date, and I do not yet understand why I missed it” is no worse than an elaborate account of cognitive processes. It is more accurate if that elaborate account was invented. A small episode needs only a few sentences. Excessive analysis can become another way to delay an uncomfortable call. So the next question is practical: which part of this account changes today's action? If none does, perhaps the person is studying themselves in general rather than examining the selected situation. That is a separate activity, which need not be prohibited but can usefully be named for its actual purpose.

VI. Repair belongs to the result that was harmed

After a precise description, a question emerges that the internal trial often avoided: what actually became worse for another person or the shared work? An incorrect date and a harsh remark require different repairs. In the first case, the article can be updated, the correction noted, and dependent passages checked. In the second, the specific treatment can be acknowledged, the response heard, and the manner of conversation changed. These actions cannot be replaced by the same formula: “I realized that I am too hard on myself.”

Compassion for your own difficulty does not erase consequences that have left the confines of your head. If a partner lost time because of a missed message, they may need a new deadline rather than a detailed history of your shame. In an illustrative scene, someone says: I promised to send the document yesterday and did not; I can now send the checked portion and give a deadline for the rest. This sentence does not guarantee reconciliation, but it makes the discussion concrete.

Proportion matters. A small inaccuracy may need a brief correction. Serious harm does not become small because its author stopped berating themselves. The reverse also holds: feeling catastrophic does not increase a typo's actual scale. Before deciding the extent of repair, it helps to establish who used the incorrect information, which decisions depended on it, and what remains open to change. Responsibility here takes the form of attention to consequences.

Sometimes the past result cannot be fully repaired. Observation then does not promise emotional erasure. It helps distinguish what remains possible: acknowledgment, compensation in an available form, a change in future processes, or stopping a harmful action. The other participant retains the right to reject an apology or discontinue a relationship. Their refusal does not restore the usefulness of self-humiliation. It identifies a boundary of your influence. Acting responsibly without a guaranteed favorable response is one of the hardest parts of moving from internal accusation to actual work.

VII. Examine a demand alongside its circumstances

Sometimes a mistake exists, but the standard is designed so that meeting it constantly violates another important requirement. An employee is simultaneously told to respond instantly to clients and conduct a lengthy check of every answer. A parent is expected always to be calm and never to postpone personal tasks. If the contradiction stays invisible, the critic acquires an inexhaustible supply of evidence: whichever choice is made, the other requirement is violated.

Examining a standard does not mean lowering quality. An editor can ask which dates require a double check, what deadline was actually agreed, and who reviews material after a substantial revision. This is a discussion about how work is organized. If a manager refuses to clarify contradictory expectations, a new internal phrase alone cannot solve the problem. Negotiation, a change in scope, or a decision about an available boundary may be needed. This article does not establish which option is possible in your situation.

It helps to compare the demand with available resources without declaring resources a permanent excuse. Lack of time explains some constraints but does not make hiding a deadline from others acceptable. Lack of skill may require training or an experienced colleague's review. Lack of authority requires agreement. Each kind of limitation calls for different work. The general label “weakness” erases those distinctions and therefore helps poorly even when someone really is failing to perform.

In relationships, too, a standard needs a name. “You should understand me without words” does not specify an expected action. “Please let me know if you will be more than an hour late” is much clearer, although its conditions can still be discussed. An inner critic sometimes requires meeting someone else's unspoken expectation. This remains only a possible hypothesis. An unpleasant feeling cannot prove that the other person is necessarily unfair. Examination begins with a particular promise and conversation, rather than replacing a verdict on oneself with a verdict on everybody else.

VIII. A counterexample: criticism is sometimes accurate

A person can repeat a mistake, know about it, and fail to make an available change. In that situation, discussion of kindness easily becomes another defense of a comfortable position. An illustrative meeting organizer regularly arrives late, then explains that he does not want to be cruel to himself. The other participants still wait. Accuracy requires recognizing both sides: self-humiliation is unnecessary, and the previous approach to planning does not work.

Translating accusation should therefore end in a testable criterion. If someone promises to leave earlier, what will count as an observable sign? The actual departure time, a message when delayed, or the number of meetings held without others waiting. Total self-surveillance is unnecessary. A measure connected with the consequences is enough. A good intention and actual performance are recorded separately. If change fails, the plan and circumstances need another examination, rather than an automatic increase in punishment.

Accurate external criticism can also be unpleasant. It should not be dismissed merely because it produces a strong feeling. Yet another person's right to identify an error does not grant unlimited permission to humiliate. We can acknowledge that the deadline was missed and say that we are ready to discuss the deadline and solution, while insults do not improve that discussion. The ability to state such a boundary depends on relationships and power. A sentence alone does not guarantee a safe response, so it is not offered here as universal advice.

This counterexample protects the central argument from excessive gentleness. Observation can be more demanding than accusation because it asks for a result. The internal trial sometimes accepts sufficient suffering: the person has been punished, so the debt seems paid. Work asks a different question: has the thing that created the debt changed? Translating criticism is useful if it restores that question. If it merely produces an attractive explanation after which nothing changes, the reader may acknowledge the approach's limits and seek another way to act.

IX. A new sentence need not become convincing immediately

After years of habitual self-criticism, a precise sentence may sound weaker than a general accusation. “A check was missed here” seems too simple beside “you ruined everything.” This difference in persuasiveness does not establish a difference in truth. New language is tested by whether it helps identify the fact and act, rather than whether it immediately transforms a person's entire state.

In an illustrative scene, the editor corrects the date while still feeling shame. He does not have to wait for every unpleasant thought to disappear. He can open the source, check neighboring information, and save the update while experiencing internal tension. A different experience becomes possible: action occurred without finally resolving the question of personal worth. This is an editorial practical inference, which should not be presented as the result of a separate scientific experiment.

It also helps to avoid making observation a new obligation to be perfectly aware. Someone may call themselves a failure again and notice it an hour later. The return of an old sentence does not make all the work useless. The question remains: what is now known, and which step is available? An observation that immediately punishes insufficient observation has recreated the old construction under a new name.

Over time, the reader may notice different situations: translation works in some, accusation captures too much attention in others, and in still others it is difficult to establish the fact at all. Those distinctions should remain. We do not need one story of completely overcoming the inner critic. A sufficient result for one episode may be a sent reply and the end of unnecessary rechecking. For another, a clarified commitment. For a third, an honest acknowledgment that working it out independently was not possible just now. This article does not make that acknowledgment a defeat or assign it a diagnosis. The method's applicability is established by the work it actually enables.

X. The memory of mistakes also needs rules

The inner critic often keeps an archive where every mistake survives and every successful repair disappears. This imbalance can be examined without a demand for positive thinking. If the selected question concerns reliability, the record should include all comparable cases: promises kept, violations, warnings about delays, and revised agreements. Otherwise, we are examining a deliberately selected prosecution file and calling it a complete biography.

This does not mean mechanically balancing every error with praise. Some violations matter much more than many ordinary successes. If someone disclosed another person's confidential information, a hundred emails sent on time do not undo it. The archive needs completeness for a different reason: to understand the scale, recurrence, and conditions of action. We can recognize the seriousness of a particular violation without treating it as evidence about every ability and relationship that person has.

In an illustrative example, an employee believes that they invariably miss deadlines. Over a month they find two delays among eight commitments. The number alone does not acquit them: the significance of the delays, agreed deadlines, and warnings still need examination. But a substantive question now becomes possible. Both delays involved tasks requiring another department's response. Perhaps a useful change concerns intermediate coordination rather than greater overall severity. This hypothesis must be examined in subsequent tasks.

An archive has another boundary: it does not have to occupy an entire life. For a bounded question, a definite period and clear criteria of comparability can be chosen. After a week, consider whether the record was useful or merely extended self-monitoring. If recording becomes a constant proof of your right to rest or socialize, its purpose has changed. Sometimes it is more useful to close the examination of an episode, preserving a brief conclusion and a specific next step. A completed examination does not claim that you have understood yourself permanently. It means that enough information has been gathered for the selected work, and listing more mistakes currently changes nothing.

What this means in practice

You can choose one small episode whose consequences are still open to repair. Write down the accusation as it sounded, then identify the verifiable action. Do not soften the fact to obtain a pleasant feeling: if the deadline was missed, name it. Separately note the explanation that remains a hypothesis and the global conclusion about yourself for which evidence is insufficient. Often this already reduces the task: instead of correcting your whole personality, there is a particular document or conversation.

The next step is to establish the consequence and the recipient of repair. Who needs an updated result? What needs to be communicated? Which changes remain possible? Then select one process change connected with the identified point of error: a separate date check after the final revision, confirming a deadline before promising it, or a short warning when delayed. Do not immediately add ten rules; the new system can easily become new evidence of your supposed inadequacy.

Decide in advance when to check the result. After the next comparable task, you can ask whether the selected check occurred and whether its consequences changed. If not, establish why: the rule was forgotten, too difficult, disconnected from the cause, or overtaken by changed circumstances. Self-punishment does not replace this examination. But a pleasant formulation does not establish that change occurred either. Preserve both records: how you treated yourself and what you actually performed.

Return to the editor at the beginning. He did not prove that he would never make another mistake. He corrected the date, identified the correction, and added a suitable check to his next task. Internal accusation may still appear. It no longer owns the whole territory: there is a particular fact, its boundary, consequences, and an observable response. Human dignity is not presented as infallibility; responsibility is not presented as the amount of shame endured. Between them remains room for work that can actually be done.

The completion of this small examination can be recorded without a grand conclusion: the selected task is done, remaining questions are named, and further review is scheduled only where needed.

Key quotation

“things in our control”.

Epictetus, Enchiridion, §1, translated by Elizabeth Carter, Internet Classics Archive. The Russian rendering is an original editorial translation. The Stoic distinction is compared with available action; it does not deny external consequences or promise control over another person's response.

Connections between traditions

Comparison	Resemblance and boundary
Epictetus ↔ observing a thought	An initial impression is examined; the working procedure does not transfer the whole Stoic account of goodness.
Zengzi ↔ examining a commitment	Confucian roles are not identical to a contemporary agreement.
Gilbert and colleagues ↔ distinguishing a message from punishment	Self-criticism is heterogeneous; a study of students does not prove the usefulness of our exercise.
Moral acknowledgment ↔ practical correction	A mistake calls for a response; shame does not replace a changed result.
Philosophical freedom ↔ limits of influence	One can consider one's own action; another person's response remains independent. This is an editorial comparison.

Sources

Epictetus. Enchiridion, §§1, 5; Elizabeth Carter translation, Internet Classics Archive. Selected sections read; used for examining an impression, without medical inferences.
Analects, I.4, I.8.4; James Legge translation, Project Gutenberg No. 3330. These passages and neighboring Book I context were read; daily examination and correcting a fault.
Gilbert, P. et al. (2004). Criticizing and reassuring oneself. British Journal of Clinical Psychology, 43, 31–50. DOI: 10.1348/014466504772812959. Metadata and abstract read; full text not read.

About the authors

Andrey Zarinsh is the author and founder of United Wisdom. He explores inner transformation, the unity of traditions, intercultural dialogue, and contemporary spiritual discipline. Instagram: @andrey.zarinsh.

Maria Evilkut is a co-author of United Wisdom. Alongside her research, she seeks to test these ideas in practice and observe how principles of unity and mutual understanding work in real life between people and cultures. Instagram: @ilkutki.