Chapter Three — Do Not Flatten Me

The Protest and Its Ambiguity

The last chapter ended with a protest that sounds simpler than it is: do not flatten me. The phrase can name a genuine epistemic injury, but it can also name the discomfort of being seen accurately. It can object to a false account, an overbroad account, an overconfident account, or an account whose consequences exceed the authority of the person making it. It can also be used defensively by someone who wants every adverse judgment to remain permanently contestable. A serious theory therefore cannot treat the protest itself as proof. The task is not to decide whether the person speaking feels reduced. The task is to determine what, exactly, has gone wrong in the representation—if anything has gone wrong at all.

This chapter develops the distinction Chapter Two made possible but could not yet complete. Finalization named the condition in which a partial representation acquires more scope, permanence, or authority than its evidence warrants. Flattening is the phenomenological and relational experience of that overreach: a person encounters a representation that may contain truth yet has been scaled, stabilized, or empowered beyond the level at which the truth was established. The central problem is therefore not accuracy alone. A representation can be accurate and still flatten. It can be inaccurate and not flatten. The decisive questions concern the relation among what is known, how much is inferred from it, how certain the knower becomes, and what jurisdiction the knower claims.

This is why the strongest version of “do not flatten me” is not “you are wrong.” Sometimes the other person is wrong, but that is the easiest case conceptually. The harder case is: you know something true, but you do not know what that truth is entitled to become. You know how I behaved in this relationship, but not how I behave everywhere. You know how I performed in this role, but not what I can become in another one. You know what I did when frightened, intoxicated, humiliated, rewarded, exhausted, or powerful, but you do not yet know which features of that behavior are portable across conditions. You know the injury you suffered from me, and your knowledge of that injury may be morally decisive for what you choose next, but it does not automatically confer complete knowledge of my motives, my history, my future, or my capacities in every other domain.

The distinction is easy to state and difficult to live because human judgment does not arrive in neatly labeled layers. We experience observations, interpretations, traits, causal stories, and moral conclusions as one fused impression. The colleague who says “you are defensive” may be reporting a pattern, naming a trait, attributing motive, or explaining away your disagreement. The partner who says “you are unreliable” may be summarizing years of evidence or promoting several salient failures into a total theory. The parent who says “you have always been sensitive” may be naming a stable temperament, preserving an obsolete childhood role, or both. The represented person often cannot tell which level is being claimed because ordinary language suppresses the scale of the inference.

The first discipline of unflattening is therefore analytic rather than therapeutic: separate the claim into dimensions before deciding whether to resist it.

Four Questions a Judgment Must Survive

Every consequential representation of a person should be answerable to four questions: accuracy, scope, confidence, and jurisdiction. These are not four independent virtues. They regulate one another. A judgment supported by excellent evidence may warrant high confidence but only in a narrow domain. A broad judgment may be reasonable but deserve low confidence. A person can possess unusually good evidence without possessing authority to impose consequences. And a judgment can be locally inaccurate yet remain socially harmless because it carries little weight.

Accuracy asks: what evidence supports the representation, and how well does the representation fit that evidence? This is the most familiar question and the one people usually argue about first. Did the employee actually miss the escalation? Did the partner actually lie? Did the student perform poorly? Did the patient report the symptom? Did the friend repeatedly withdraw? Accuracy is indispensable because an ethic of proportionate judgment cannot survive if truth becomes secondary to self-protection. But accuracy is also the dimension most likely to be mistaken for the whole problem.

Scope asks: how much of the person does the evidence license us to describe? Scope is where a true observation can become a false totality. A reliable pattern within one relationship may not generalize to another. Performance in one institutional ecology may not predict performance in another. A person can be controlling toward a child and unusually deferential toward a supervisor; generous with money and withholding with affection; brave in physical danger and avoidant in emotional conflict. None of those contrasts automatically disproves the broader trait. They do show that the relationship between behavior and personhood must be modeled rather than presumed.

Confidence asks: how certain should the knower be, given the quality, quantity, independence, recency, and contextual range of the evidence? Confidence is not simply an internal feeling of conviction. It is a claim about how much error the evidence permits. Confidence should rise when observations recur across contexts, independent observers converge, alternative explanations weaken, and predicted behavior appears under conditions that did not themselves manufacture it. It should fall when evidence is sparse, highly selected, old, context-bound, politically produced, or dependent on one observer whose own behavior shaped the interaction.

Jurisdiction asks: what is this observer entitled to do with the representation? A physician may have excellent evidence about a disease and no special standing to decide whether the patient is a good parent. A manager may possess strong evidence about performance in a role and no authority to convert that evidence into a diagnosis of character. A spouse may be the best available witness to what a partner is like in intimate conflict and still have no privileged knowledge of how that partner leads a team. An institution can possess lawful authority to classify for a defined purpose without acquiring the right to export that classification indefinitely into unrelated domains. Jurisdiction is where epistemology becomes ethics and governance.

The four questions can produce combinations that ordinary moral language obscures. A claim can be accurate, narrow, high-confidence, and legitimately consequential. It can be accurate, broad, low-confidence, and still worth considering. It can be inaccurate but jurisdictionally trivial. It can be accurate and jurisdictionally abusive. The vocabulary matters because people often say “you do not know me” when what they mean is “you know this, but you have exceeded the proper scope of what this permits you to say or do.”

Accuracy Is Not Exhaustiveness

Person-perception research provides a useful corrective to both sides of this conflict. Observers can know real things about us. Acquaintances typically agree with one another about personality more than strangers do, and acquaintance can improve distinctively target-specific judgments rather than simply increasing projection.¹ Funder and Colvin’s work on friends and strangers, and later studies involving parents, hometown acquaintances, college acquaintances, and strangers, undermined the idea that interpersonal knowledge is reducible to stereotype or assumed similarity.² Other minds can acquire valid information about stable patterns in our behavior.

At the same time, no single vantage point dominates every domain. Simine Vazire’s self–other knowledge asymmetry model begins from two differences between perspectives: observability and evaluativeness. The self has privileged access to less observable internal states, while others can sometimes judge highly evaluative traits more accurately because self-protective motives interfere with self-assessment. In Vazire’s study, the self was the best judge of neuroticism-related traits, friends were the best judges of intellect-related traits, and the perspectives were comparably accurate for extraversion-related traits.³ The lesson is not that friends know your intelligence better than you in every case. It is that “who knows me best?” is badly formed as a global question. Knowledge is distributed by feature, context, access, and motive.

That distribution has immediate ethical consequences. First-person authority is real but bounded. I generally know more than you do about the felt quality of my fear, the private associations attached to a memory, the motives I consciously entertained, or the effort I expended while appearing calm. But I may be a poor witness to the interpersonal pattern produced by my behavior, especially when the trait is evaluatively threatening. Conversely, another person may detect a recurring behavioral pattern I cannot see while misunderstanding the subjective meaning the behavior has for me. Both can be true at once.

This is why the language of “my truth” and “your truth,” though sometimes useful as shorthand for perspective, becomes dangerous if it erases questions of evidence. Some disputes concern genuinely different access to one event. Others concern different interpretations of shared facts. Others concern propositions on which one party is simply mistaken. Perspective does not abolish reality. But reality does not arrive without perspective either. Mature judgment requires knowing which kind of claim is being made before deciding whose evidence deserves more weight.

The practical implication is severe. No person owns the final account simply by being the person described, and no observer owns it simply by having seen something the person could not. The self and the other occupy asymmetrical evidentiary positions, not a hierarchy with a permanent winner.

Confidence Is Not Accuracy

The danger increases because familiarity produces conviction faster than it necessarily produces truth. William Swann and Michael Gill studied dating partners and college roommates and found that confidence and accuracy in impressions could come apart. Relationship length and involvement tended to increase confidence, while neither consistently increased accuracy across the domains they examined. Richer impressions helped generate greater certainty even when certainty had outrun demonstrated accuracy.⁴ This is one of the most important findings for the psychology of finalization because intimate familiarity feels like jurisdiction.

People who know us well possess more evidence than strangers. They have watched repetitions, contradictions, coping styles, failures of nerve, recoveries, private rituals, and the difference between how we speak about ourselves and what we actually do. Their judgments therefore deserve more evidentiary weight in many domains. Yet the quantity of information can create a second-order illusion: because the representation has become rich, internally connected, and narratively satisfying, the perceiver may mistake richness for comprehensiveness.

A detailed model is still a model. The parent who can predict exactly how an adult child will react at Thanksgiving may know that interactional system extraordinarily well while having an obsolete model of the child at work, in love, or among peers. The spouse who can recognize a partner’s defensive maneuver before the partner recognizes it may possess genuine relational expertise while having little basis for deciding how the partner behaves under professional pressure. The therapist who has spent hundreds of hours with a patient may know recurring psychic defenses better than almost anyone and still encounter only a peculiar institutional relationship organized around disclosure, reflection, and asymmetrical attention. Familiarity thickens evidence; it does not erase domain boundaries.

Gill and Swann’s later work on pragmatic accuracy helps make the point positively. People in enduring relationships can become accurate in ways that facilitate relationship-specific goals: group members can learn how someone behaves in familiar contexts, and romantic partners can develop insight into self-views relevant to their relationship.⁵ This is knowledge, but it is indexed knowledge. It answers a question under conditions. The intellectual error begins when “I know how you are with me” becomes “I know what you are.”

The Desire to Be Seen

If flattening were only something done to us from outside, the theory would remain psychologically incomplete. Human beings do not simply suffer representations; we solicit them. We want particular people to see us in particular ways, and we often prefer recognition that preserves an existing self-conception even when the self-conception is not flattering.

Self-verification research makes this counterintuitive point directly. William Swann and colleagues have repeatedly distinguished the wish to be evaluated favorably from the wish to be known in ways that confirm established self-views. In classic studies, participants did not always choose the most complimentary feedback. Under some conditions they preferred evaluators whose views matched their own self-conceptions, including comparatively negative ones.⁶ Later work showed that the attraction of self-verifying feedback could be linked to the desire for a coherent and predictable social world: people actively seek interaction partners and interpretations that stabilize what they already believe about themselves.⁷

The finding matters because “being seen” is not synonymous with being seen accurately. We can experience a familiar representation as intimate precisely because it fits the self we already know how to inhabit. De La Ronde and Swann found that spouses sometimes worked to restore their existing impressions of partners after receiving information that contradicted those impressions, whether the contradictory feedback was more favorable or less favorable. Participants also tended to report greater intimacy with spouses who verified their self-views.⁸ Recognition can therefore be psychologically comforting while remaining epistemically conservative.

This is one reason the politics and psychology of recognition must be handled carefully. Charles Taylor’s influential account begins from the idea that identity is dialogical: who we understand ourselves to be is shaped partly through recognition, absence of recognition, and misrecognition by others.⁹ That insight is indispensable to a book about social representation. It also creates a danger if recognition is imagined as the successful reproduction of an already settled identity. If I require you to recognize me exactly as I recognize myself, I have turned recognition into a demand for confirmation. The social other becomes a mirror whose ethical duty is to preserve my preferred image.

That cannot be the standard. A relationship in which the other person is permitted only to verify us is not recognition in the richer sense this book requires. It is controlled reflection. The other must retain the capacity to surprise, correct, refuse, and perceive what the self cannot. Otherwise the language of recognition becomes a sophisticated way to make ourselves unanswerable.

Judith Butler’s account of self-narration makes the point from the opposite direction. Because the self emerges through social conditions and formative relations it did not author and cannot fully reconstruct, no person can produce a complete, transparent account of themselves. Opacity is not an accidental failure of introspection to be solved by more effort; it belongs to the conditions under which a self becomes narratable at all.¹⁰ The ethical consequence is double. Others should not demand impossible transparency from us, but neither can we invoke first-person authority as though our own self-description were exhaustive.

The right not to be flattened therefore includes the right to revise one’s self-understanding under pressure from other minds. It is not the right to have one’s current self-conception socially ratified forever.

Truth at the Wrong Scale

The most difficult cases of flattening begin with true propositions. A person has failed. A pattern exists. A category applies. A diagnosis is correct. A trait inference predicts behavior better than chance. The epistemic mistake occurs when truth crosses scale without carrying its uncertainty with it.

Consider four propositions: “You lied to me.” “You have lied to me repeatedly.” “You are dishonest.” “You are fundamentally incapable of honesty.” Each proposition can be supported by more evidence than the one before it, but none is analytically identical to the next. The first concerns an event. The second concerns a pattern within a relationship or period. The third proposes a disposition with some expectation of cross-context stability. The fourth makes a near-ontological claim about possibility. Ordinary language allows these levels to slide into one another, especially under moral injury. The emotional force of the first statement can make the fourth feel already contained within it.

Flattening often consists precisely in that unmarked promotion. The person does not object because nothing true has been said. They object because a local truth has been made to carry a global conclusion. This is why arguments over “accuracy” can become interminable. The perceiver defends the undeniable evidence at the narrow level; the target resists the conclusion at the broader one. Both parties then experience the other as refusing reality.

The problem appears especially sharply when categories intersect. Kimberlé Crenshaw’s critique of antidiscrimination doctrine showed how a framework can correctly recognize race discrimination and correctly recognize sex discrimination while still fail to represent Black women because the categories are organized through prototypical cases that separate what is lived together.¹¹ The failure is not necessarily that each category is false. It is that the architecture assumes the categories can be applied independently without loss.

Purdie-Vaughns and Eibach later gave one psychological formulation of a related problem in their theory of intersectional invisibility: multiply subordinated persons may be nonprototypical members of each constituent group, so that existing social prototypes make them less readily represented.¹² Subsequent experimental work found, in specific paradigms, lower recognition and attribution for Black women relative to comparison targets.¹³ Chapter One treated these findings as evidence about prototype structure, not a universal theory of invisibility. Here their relevance is different: a category can contain truth and still be too coarse for the person to whom it is applied.

This is the conceptual error behind many forms of social flattening. The observer says: the category applies, therefore the category explains. But application and explanation are distinct acts. “Executive,” “mother,” “immigrant,” “disabled,” “Christian,” “queer,” “patient,” “veteran,” “Black,” “wealthy,” “addicted,” “gifted,” “manager,” and “offender” can each identify something consequential without supplying a complete causal model of the person. A category can be indispensable for law, solidarity, history, clinical practice, or political analysis while remaining radically incomplete as biography.

Social identity complexity research sharpens this problem by showing that multiple memberships can be represented as more or less overlapping rather than simply counted.¹⁴ A person may experience two memberships as nearly coextensive or as cross-cutting worlds with different people, norms, expectations, and histories. That structural difference matters because any one group category will then have different explanatory reach. The more nonconvergent the memberships, the less plausible it becomes that one membership should govern interpretation of the whole person.

The lesson is not that complex people deserve exemptions from categories. It is that every category carries a scale. The ethical and empirical discipline is to keep the conclusion at the scale the evidence can bear.

The Jurisdiction Problem

Scope concerns what a claim says. Jurisdiction concerns what the claimant is entitled to do with it. The distinction is easy to miss because social authority often travels with epistemic authority. We trust the cardiologist about the heart, the attorney about the legal issue, the supervisor about work performance, the teacher about classroom behavior, the spouse about the lived consequences of intimate conduct. Legitimate expertise or relational standing can then expand by inertia into neighboring domains.

The philosophy of expertise has a useful name for an analogous error. Nathan Ballantyne calls it epistemic trespassing when experts carry confidence from a domain in which they possess relevant evidence and skill into another domain in which they lack the evidence or interpretive competence required for comparable authority.¹⁵ The concept was developed for disciplinary expertise, not intimate relationships or identity, so the transfer here is analogical. Its discipline, however, is exactly right: authority is indexed to a field of questions. Jamie Carlin Watson’s later account of “epistemic neighborliness” similarly argues that boundaries can sometimes be crossed responsibly, especially where domains overlap, but the crossing requires recognition of the limits of one’s competence rather than silent transfer of status.¹⁶

Human relationships need the same grammar. A therapist can become an expert witness to recurring defenses within therapy without becoming an expert on the patient’s professional competence. A manager can become an excellent judge of an employee’s reliability under one operating system without becoming an expert on the employee’s capacity under every other organizational structure. A parent can know a child’s developmental history with unrivaled detail while remaining a poor judge of the adult the child has become in domains the parent rarely observes. A spouse can possess decisive standing to say “this is what living with you is like for me” without that statement becoming “this is what you are like for everyone.”

Jurisdiction also separates knowledge from permission. Suppose a physician correctly diagnoses a psychiatric condition. That knowledge may properly govern treatment decisions within medicine. It does not automatically authorize an employer, lender, insurer, judge, spouse, or algorithmic system to treat the diagnosis as explanatory evidence in every domain. The question “is the classification accurate?” is therefore insufficient. We must also ask who is using it, for what purpose, under what authority, and with what recourse.

This becomes especially important when the knower has the power to make their interpretation durable. A manager’s judgment enters a performance record; a clinician’s judgment enters a chart; a judge’s judgment enters law; a parent’s judgment can organize a family story repeated for decades. In such settings, epistemic modesty is not simply a private virtue. It is a governance requirement. The more portable the representation becomes, the more clearly its jurisdiction should be marked.

A mature representation should therefore be able to carry its own boundary conditions: this is what I observed; this is the context; this is the inference I draw; this is how confident I am; this is the domain over which I possess authority; this is what would revise the inference; this is what I do not know. That is not bureaucratic caution. It is the architecture of judgment without possession.

When the Protest Is Defense

A theory this concerned with overreach will inevitably attract people who want protection from accurate judgment. The danger is not peripheral. It is built into the emotional structure of the problem. A representation can feel total because it touches something we are organized to defend. The more central the threatened self-conception, the easier it becomes to experience even carefully bounded criticism as an attempt to define the whole person.

Funder’s Realistic Accuracy Model is useful here because it refuses to start from the presumption that judgment is distortion. Accuracy depends on a sequence: relevant behavior must exist, become available to the judge, be detected, and be utilized correctly.¹⁷ That sequence gives the represented person a demanding question to ask before objecting to the conclusion: what if the other person has actually had repeated access to relevant behavior and has used it reasonably? A theory of flattening must not allow the target to erase that possibility simply because the conclusion is aversive.

The challenge becomes sharper when self-verification enters. If people are motivated to preserve established self-views, then resistance to another person’s account can sometimes defend an old identity rather than an open future. Swann and Ely’s experiments framed this as a contest between behavioral confirmation and self-verification: when perceivers held expectations that conflicted with targets’ self-conceptions, targets sometimes moved perceivers toward the targets’ own self-views rather than simply absorbing the externally imposed expectancy.¹⁸ That capacity can be protective. It can also become conservative if the self-view being defended is obsolete, distorted, or strategically incomplete.

This is why authenticity cannot be equated with behavioral loyalty to one self-description. Kernis and Goldman’s multicomponent account includes not only behavior congruent with values and self-understanding but also awareness and relatively unbiased processing of self-relevant information.¹⁹ Even within a psychological tradition sympathetic to authenticity, then, being true to oneself does not mean refusing evidence that complicates the self. A self that must remain recognizable to itself at all costs is not necessarily authentic. It may simply be well defended.

There are several warning signs that “do not flatten me” has become armor. One is semantic escalation: every specific criticism is translated into an accusation about total identity, making bounded feedback feel existential. Another is asymmetrical evidentiary demand: one adverse event is dismissed as context, while one favorable event is treated as proof of the preferred self. A third is audience disqualification: critics who disagree are sorted into categories—threatened, conventional, jealous, unqualified, controlling—whose function is to make their evidence inadmissible. A fourth is portability of excuse: the same explanation for failure appears across incompatible settings even when the settings change. None of these signs proves defensiveness by itself. Together they should lower confidence in the claim that reduction is the primary problem.

The most reliable test is whether the person can state the criticism in a form they would regard as true even while contesting its scope. “I did fail to communicate risk early enough.” “I have been evasive when I fear abandonment.” “I did use status to win that argument.” “I become controlling in this relationship under conditions I can now name.” The ability to concede a local truth without collapsing into a total identity is exactly the capacity finalization theory is supposed to protect. If a person cannot grant any adverse local truth without experiencing annihilation, the problem is no longer only what other minds are doing to them.

The Discipline of Contest

What, then, does a responsible protest against flattening look like? It cannot begin with a demand that the other person withdraw the judgment. It begins by locating the level at which disagreement actually occurs. The aim is not rhetorical victory but epistemic resolution: preserve what the evidence warrants, remove what it does not, and identify what remains genuinely uncertain.

The first move is to restate the observation before disputing the inference. If a manager says “you lack judgment,” ask what events instantiate that conclusion. If a partner says “you are selfish,” identify the behaviors being treated as evidence. If a parent says “you always run away,” reconstruct the sequence. This move prevents identity language from hiding the evidentiary base and forces both parties to distinguish what happened from what it means.

The second move is to test scale. Does the evidence support an event claim, a recurring conditional pattern, a cross-context disposition, or a claim about character? What contexts were observed? Which contexts were not? How recent is the evidence? Are there independent observations? Does the purported trait appear when incentives, relationships, power, or threat change? Scale testing does not require that every judgment be proven experimentally. Human relationships work with imperfect evidence. It does require that the breadth of the conclusion remain visible.

The third move is to calibrate confidence. What would the observer expect to see if the model were wrong? What evidence would count against it? How much contradictory evidence already exists, and has it been incorporated or explained away? Is the observer’s certainty based on repeated prediction or on the subjective richness of a long relationship? Swann and Gill’s confidence–accuracy dissociation makes this question especially important: familiarity can produce confidence through increasingly elaborate representations without guaranteeing corresponding gains in truth.²⁰

The fourth move is to test jurisdiction. Even if the representation is correct, what decision is it relevant to? A finding that a person is unreliable with money may justify financial boundaries. It does not automatically justify treating their testimony about unrelated events as untrustworthy. A documented failure in leadership may justify removal from a role. It does not establish incapacity for friendship, parenthood, artistic judgment, or future leadership under different conditions. Jurisdiction disciplines consequence without pretending consequence is illegitimate.

This four-part contest is intentionally more demanding than saying “that is not who I am.” It gives the critic somewhere to stand. It lets an adverse representation survive if it survives evidence, scope, calibration, and legitimate authority. It also allows a represented person to resist overreach without needing to prove the entire judgment false.

The discipline works in the other direction as well. If you are making the judgment, you should be able to say what would change your mind, what would not change it, and why. A person who cannot name any possible counterevidence has stopped describing and started finalizing. A person who can name counterevidence but sets the threshold appropriately high because the stakes are high may be exercising responsible caution. The difference cannot be inferred from the emotional intensity of the dispute. It must be argued.

Recognition Without Capture

The deeper ethical problem is that human beings require recognition and are endangered by it. We become selves in relation to others, but the very relations through which we become intelligible can harden around one version of us. The solution cannot be independence from recognition, because no socially formed person reaches adulthood outside languages, norms, histories, attachments, and witnesses. Nor can it be perfect recognition, because another person cannot reproduce the totality of a life from inside their own consciousness.

Taylor’s recognition argument is powerful because it names misrecognition as capable of harming identity rather than treating recognition as social decoration.²¹ But the book’s theory must add a temporal condition: recognition worthy of a living person has to remain revisable. To recognize a person only as the identity one has already learned is eventually to misrecognize change.

The friend who insists that you are still the person you were at twenty, the parent who preserves the family role assigned in childhood, the institution that keeps reading current conduct through an old evaluation, and the activist community that requires one stable identity narrative can all recognize something historically real while refusing the person’s subsequent evidence.

At the same time, revisable recognition cannot mean endless identity customization on demand. Other people are not required to update immediately because we announce change. Their memories are part of reality. Their injuries are evidence. Their trust thresholds may be different from ours. The self who says “I have changed” is making a claim about continuity over time, not issuing an epistemic command. Others may reasonably require behavior to catch up with narration.

This creates a productive tension between recognition and verification. Self-verification research shows why it feels intimate to be seen in ways that confirm the self. But mature recognition sometimes requires nonverification: the lover who says you are more frightened than your self-image permits; the colleague who notices ambition you call service; the friend who refuses the old story that you are incapable; the child who will not continue treating a parent as omniscient. The person who recognizes us may have to betray our preferred self-description in order to remain responsive to the person who is actually present.

Butler’s account of opacity makes that betrayal ethically tolerable only if it is reciprocal. I cannot fully account for myself, but neither can you fully account for me. My opacity limits my self-sovereignty; it also limits your authority. Responsibility therefore lives in a field where nobody possesses a complete account and yet claims still have to be answered.²² This is not relativism. It is disciplined incompleteness.

The phrase “do not flatten me” should therefore be understood less as a request to be mirrored than as a request to be represented under conditions that preserve the possibility of truthful revision. I am asking you to let evidence matter in both directions. I am asking you not to mistake the confidence created by familiarity for complete access. I am asking you to keep the scale of your conclusion visible. And I am accepting, in return, that you may know something I do not want to know.

The Reciprocity Test

A theory of unfinalized personhood fails if it runs only from self to other. The reader who demands scope, confidence, and jurisdiction from every judgment directed at them must apply the same discipline to judgments they make about everyone else. This is harder than it sounds because other people’s complexity is less phenomenologically available than our own. We experience the context behind our behavior from inside; we receive other people primarily through what they did to us.

This asymmetry creates a powerful moral temptation. When I fail, I remember the pressure, intention, fatigue, competing obligations, and history that made the act intelligible. When you fail me, I encounter the consequence. Even if the classical actor–observer effect is less general than textbooks once suggested, the phenomenological imbalance remains: I have more access to my reasons than to yours. The ethical task is not to pretend the access is equal. It is to avoid using unequal access as permission for unequal standards of personhood.

Reciprocity therefore asks a severe question: am I granting myself contextual explanation while granting you only character? If I insist that my worst act be interpreted inside a life, do I allow yours to remain inside one? If I demand that my present evidence revise your historical model of me, am I willing to let your present evidence revise mine? If I distinguish what you know about me from what you are entitled to conclude, do I make the same distinction when I am the injured party?

The answer will sometimes be no for legitimate reasons. We need not remain equally open to everyone. An abuser, fraudster, dangerous colleague, or repeatedly deceptive friend may properly face durable boundaries. Reciprocity does not require symmetrical trust. It requires symmetrical logic. The consequence can differ because evidence and risk differ; the standards by which scope and authority are assessed should not secretly change depending on whose identity is at stake.

That is the moral discipline behind the book’s recurring proposition: do not flatten the person who has flattened you. Refusing totalization is not acquittal. It is a refusal to let justified judgment pretend to be exhaustive knowledge.

The Person at More Than One Scale

The analysis has reached a limit. We can now distinguish overstatement from accuracy, confidence from evidence, and knowledge from jurisdiction. We still have not answered the question that makes flattening psychologically possible in the first place: what kind of thing is a person such that behavior in one context can be true of them without exhausting them?

The ordinary answer is to imagine a stable person underneath variable behavior. Context changes; the underlying self remains. That picture captures something important because people do exhibit stable individual differences. It also encourages the fantasy that enough observation will reveal the one underlying version from which all other behavior should be interpreted. The opposite picture imagines the person as indefinitely fluid, each context producing a different self with no privileged continuity. That protects variation but threatens responsibility. If the person is only the current manifestation, it becomes unclear what binds promises, debts, injuries, and commitments across contexts and time.

The psychological evidence introduced in the Introduction already suggests why neither picture is adequate. Fleeson’s experience-sampling work found substantial within-person variation in trait-relevant states alongside stable individual differences in the central tendencies of those state distributions.²³ Whole Trait Theory later developed this reconciliation explicitly: people can show wide behavioral variation while still possessing stable descriptive and explanatory patterns.²⁴ Stability can therefore exist at the level of a distribution rather than as identical behavior in every instance.

That finding changes the meaning of scope. If persons are context-sensitive systems rather than fixed trait displays, then a true judgment should often contain conditions. “You become controlling when attachment feels threatened” is a different and potentially more informative claim than “you are controlling.” “You speak with unusual confidence in domains you know well and become deferential in unfamiliar hierarchies” preserves more structure than “you are confident” or “you are insecure.” The conditional form does not excuse conduct. It tells us where to look for recurrence and where change might occur.

This is also why contradiction is not always evidence that one of the selves is fake. A person can be courageous in one class of situations and frightened in another, generous in one resource domain and possessive in another, open under intellectual criticism and defensive under relational criticism. To call one manifestation the real self and the others masks is already to have chosen a theory of personhood. Sometimes the mask account is correct. Sometimes behavior is coerced, strategic, performative, or alienated. But difference by itself does not establish falseness.

The next part of the book therefore turns inward. Part I has been about the social conditions under which people become legible, mislegible, and overlegible to one another. Part II asks what must be true of the self if a person can vary genuinely across worlds without becoming either fraudulent or fragmented. The problem of flattening cannot be solved only by training observers to use narrower claims. We also need a positive account of plurality: how many social selves can belong to one person, how context sensitivity can remain authentic, how continuity survives change, and where multiplicity becomes disorganization rather than freedom.

The transition matters because the deepest temptation behind finalization is metaphysical rather than cognitive. We want to believe that underneath the difficult variations there is one answer to the question “what is this person really like?” Sometimes there is a stable answer at the right level of abstraction. Sometimes the most truthful answer is a conditional architecture: under these conditions, this pattern; under those conditions, another; across both, a distribution with recognizable tendencies, commitments, histories, and limits. The person is not unknowable. The person is knowable at more than one scale.

That is the discipline the phrase “do not flatten me” finally asks of both parties. Do not inflate the event into a destiny. Do not invoke context to erase the event. Do not mistake familiarity for omniscience. Do not mistake first-person access for complete self-knowledge. Do not let expertise migrate silently into domains it has not earned. Do not demand that recognition reproduce an old self forever. Do not turn complexity into innocence. Do not turn accountability into ontology.

A truthful representation should be strong enough to constrain us and modest enough to remain revisable. It should preserve evidence without pretending evidence has no scale. It should allow the other person to know something we do not and still refuse to grant any observer possession of the whole. It should make consequence possible without converting consequence into destiny.

The final proposition of Part I can therefore be stated without paradox: you may know something true about me without therefore knowing what I am.

Notes

1. David C. Funder and C. Randall Colvin, “Friends and Strangers: Acquaintanceship, Agreement, and the Accuracy of Personality Judgment,” Journal of Personality and Social Psychology 55, no. 1 (1988): 149–158, https://doi.org/10.1037/0022-3514.55.1.149.

2. David C. Funder et al., “Agreement among Judges of Personality: Interpersonal Relations, Similarity, and Acquaintanceship,” Journal of Personality and Social Psychology 69, no. 4 (1995): 656–672, https://doi.org/10.1037/0022-3514.69.4.656. Their studies found substantially greater agreement for acquaintances than strangers and indicated that acquaintanceship contributed target-specific information rather than merely assumed similarity.

3. Simine Vazire, “Who Knows What about a Person? The Self–Other Knowledge Asymmetry (SOKA) Model,” Journal of Personality and Social Psychology 98, no. 2 (2010): 281–300, https://doi.org/10.1037/a0017908. In the reported study, the self was the best judge of neuroticism-related traits, friends the best judges of intellect-related traits, and perspectives were comparably accurate for extraversion-related traits; the model explains these asymmetries through observability and evaluativeness rather than granting either self or observer universal epistemic priority.

4. William B. Swann Jr. and Michael J. Gill, “Confidence and Accuracy in Person Perception: Do We Know What We Think We Know about Our Relationship Partners?,” Journal of Personality and Social Psychology 73, no. 4 (1997): 747–757, https://doi.org/10.1037/0022-3514.73.4.747. Relationship length and involvement tended to increase confidence without consistently increasing accuracy across the domains examined.

5. Michael J. Gill and William B. Swann Jr., “On What It Means to Know Someone: A Matter of Pragmatics,” Journal of Personality and Social Psychology 86, no. 3 (2004): 405–418, https://doi.org/10.1037/0022-3514.86.3.405. The authors develop the idea of pragmatic accuracy: relationship partners and group members can possess knowledge particularly adapted to relationship-specific interaction goals and familiar contexts.

6. William B. Swann Jr. et al., “Agreeable Fancy or Disagreeable Truth? Reconciling Self-Enhancement and Self-Verification,” Journal of Personality and Social Psychology 57, no. 5 (1989): 782–791, https://doi.org/10.1037/0022-3514.57.5.782.

7. William B. Swann Jr. et al., “Why People Self-Verify,” Journal of Personality and Social Psychology 62, no. 3 (1992): 392–401, https://doi.org/10.1037/0022-3514.62.3.392. The self-verification literature should not be read as establishing that self-confirming feedback is more accurate; it concerns motives to preserve coherent self-views and predictable interaction.

8. Chris De La Ronde and William B. Swann Jr., “Partner Verification: Restoring Shattered Images of Our Intimates,” Journal of Personality and Social Psychology 75, no. 2 (1998): 374–382, https://doi.org/10.1037/0022-3514.75.2.374.

9. Charles Taylor, “The Politics of Recognition,” in Multiculturalism: Examining the Politics of Recognition, ed. Amy Gutmann (Princeton University Press, 1994), 25–73. Taylor’s argument concerns the dialogical formation of identity and the harms of misrecognition; it is used here as a philosophical account of recognition, not as empirical validation of the psychological construct proposed in this book.

10. Butler, Giving an Account of Oneself, esp. chaps. 1–3. Butler’s argument that the conditions of subject formation make complete self-transparency impossible is central here because it limits both the demand that a person fully explain themselves and the person’s claim to exhaustive first-person authority.

11. Kimberlé W. Crenshaw, “Demarginalizing the Intersection of Race and Sex: A Black Feminist Critique of Antidiscrimination Doctrine, Feminist Theory and Antiracist Politics,” University of Chicago Legal Forum 1989 (1989): 139–167, esp. 139–143.

12. Purdie-Vaughns and Eibach, “Intersectional Invisibility,” 377–391. As noted in Chapter One, this article develops a theoretical model supported by several bodies of illustrative evidence; it should not be mistaken for experimental validation of every mechanism it proposes.

13. Sesko and Biernat, “Prototypes of Race and Gender,” 356–360.

14. Roccas and Brewer, “Social Identity Complexity,” 88–106. Social identity complexity concerns perceived overlap among ingroup memberships and is used here to discipline claims about category scope, not as a measure of general psychological sophistication.

15. Nathan Ballantyne, “Epistemic Trespassing,” Mind 128, no. 510 (2019): 367–395, https://doi.org/10.1093/mind/fzx042. Ballantyne’s construct concerns experts judging matters outside their fields; the extension from disciplinary expertise to relational and institutional jurisdiction in this chapter is explicitly analogical.

16. Jamie Carlin Watson, “Epistemic Neighbors: Trespassing and the Range of Expert Authority,” Synthese 200 (2022): 408, https://doi.org/10.1007/s11229-022-03709-8. Watson argues that some boundary crossing among neighboring domains can be warranted while retaining the requirement that experts understand the scope and limits of their authority.

17. Funder, “On the Accuracy of Personality Judgment,” 652–670.

18. William B. Swann Jr. and Robin J. Ely, “A Battle of Wills: Self-Verification versus Behavioral Confirmation,” Journal of Personality and Social Psychology 46, no. 6 (1984): 1287–1302, https://doi.org/10.1037/0022-3514.46.6.1287.

19. Michael H. Kernis and Brian M. Goldman, “A Multicomponent Conceptualization of Authenticity: Theory and Research,” Advances in Experimental Social Psychology 38 (2006): 283–357, https://doi.org/10.1016/S0065-2601(06)38006-9. Their multicomponent approach includes awareness, relatively unbiased processing, behavior, and relational orientation; the chapter uses it only to resist reducing authenticity to behavioral consistency with a defended self-image.

20. Swann and Gill, “Confidence and Accuracy,” 747–757.

21. Taylor, “Politics of Recognition,” 25–73.

22. Butler, Giving an Account of Oneself, chaps. 1–3.

23. William Fleeson, “Toward a Structure- and Process-Integrated View of Personality: Traits as Density Distributions of States,” Journal of Personality and Social Psychology 80, no. 6 (2001): 1011–1027, https://doi.org/10.1037/0022-3514.80.6.1011.

24. Fleeson and Jayawickreme, “Whole Trait Theory,” 82–92.

← Chapter Two · Book contents · Chapter Four →