Explanation connects scattered facts into relations on which one can act. Without explanation, a person can only store fragments and cannot judge causes, responsibilities, and next steps. A stable theory also lets many people use a common language instead of starting from zero each time.
An explanation can also acquire a peculiar protection: supporting material proves it correct, opposing material proves that the object is resisting, and the absence of material proves that control has penetrated deep. However reality changes, the theory only accumulates certainty.
Continue the fictional residency project. As the attribution dispute enters independent review, Ye and the organization each hold a narrative. The open version says that the record conflict may arise from process, communication, and interests, and that new version documents would change the judgment. The closed version says that all of Ye's objections stem from "not yet accepting his true self"; his agreement demonstrates the theory's effectiveness, his opposition is defense, and his silence is exposure.
This is not a psychological diagnosis, nor does it presuppose that any theory of reality is closed. This chapter compares the position of the same material within two explanations, asking what result would lower certainty, whether revision leaves a version trail, and whether a theory can acknowledge that an object temporarily lies outside its scope.
How Explanations Accommodate or Exclude Counterexamples
Encountering one counterexample does not necessarily require abandoning the entire theory. One must first check the observation, the scope of the proposition, and the auxiliary conditions; but if a reliable counterexample genuinely falls within the scope of a universal proposition, that proposition cannot be retained unchanged. Retaining the parts that still have support is one thing; protecting a core that may never change under any circumstances is another. The question is whether the core is stated, and whether its boundaries can be altered by material. If the core shifts with every failure, the theory bears no risk.
When Ye submits new version records that affect only the attribution date, the organization need not thereby repudiate all of its management experience. It should adjust the relevant propositions and consequences. A counterexample has scope. One instance of effective support does not prove that all relations are safe, and one error does not prove that all staff act in bad faith.
A theory can lower certainty, narrow its object, suspend application, or replace a causal claim, rather than jumping only between truth and falsity. Corrigibility consists precisely in this plurality of responses. Stating in advance which materials bear on the core prevents selecting only the convenient explanation afterward.
The Material Is Genuine; What It Proves Still Has to Be Examined
Ye finds an original document with a timestamp. The open explanation compares it with the old records and may acknowledge a process error; the closed explanation says that his obsessive search for documents proves he is still controlled by the past. The latter sentence may describe some psychological possibility, yet it does not answer the content of the document. The subject's state and the facts about the object are fused into a single layer.
The reviewers can confirm the document's authenticity, its version relations, and its authorization, and only then discuss whether Ye's communication caused other effects. The two lines of inquiry do not cancel each other.
If the document is forged, it cannot support Ye's claim; if it is genuine, one must still ask whether it concerns the disputed object and whether it conflicts with existing conclusions. Only if this further step also holds does it constitute grounds for requiring revision of the relevant conclusions. Whatever the outcome, it does not represent his entire personality.
A closed explanation often turns the very act of presenting evidence into a symptom, so that the object can never reach the theory. The order of restoration requires that the entry point of fact precede any overall assessment of personality.
A critic may also speak in a closed manner: "when the institution questions any material, that proves oppression." Verification of authenticity remains necessary; a position of power cannot substitute for evidence.
Whether Definition Revisions Leave a Version Trail
The coordination group initially says that "high concern" refers only to a need for additional communication; after Ye provides records of timely replies and completed scheduled conversations, the group no longer specifies which communications remain incomplete, and instead redefines the term as "possibly concealing difficulties." The label moves from visible behavior to invisible possibility.
Adjusting concepts in the light of experience can be legitimate; both science and institutions require revision. The question is whether the revision acknowledges the failure of the old definition.
If the organization writes down the old proposition, the triggering material, the new definition, and future tests, the revision takes on new risk. If it says only "this is what we always meant," past correctness is manufactured after the fact.
The new definition must also state its use. "Possibly concealing difficulties" can support voluntary counseling, yet it does not suffice to revoke a qualification. The more the concept expands, the fewer counterexamples remain, and actual explanatory power may decline. When any behavior can be contained under "hidden possibility," the theory no longer distinguishes its objects.
Language is not a static contract, but public consequences require traceable meaning. Private reflection may remain open; administrative decisions cannot redefine indefinitely.
How Auxiliary Conditions Undergo Examination
When the theory fails, the organization explains it as insufficient experience of the mentor, Ye's temporary defensiveness, lack of time, or environmental interference. Each auxiliary condition may be true. Multiple conditions do not automatically constitute an excuse. Complex relations are in any case multiply caused.
The question is whether the auxiliary conditions have independent evidence and an upper bound. If "lack of time" matters, it should state when the review will occur; if there is environmental interference, compare outcomes after the environment changes. A condition must bring new observations.
A delay without a deadline makes the theory never fail. The subject waits continuously, while the consequences for resources and relations take effect immediately.
The same name can be applied to different behaviors, and this in itself is no contradiction. The problem is that if speaking, silence, and any subsequent material can all be counted only as defense, and the interpreter cannot say what would weaken the judgment, then no checkable distinction has been left.
The open explanation permits not knowing for the time being. The unknown is not a humiliation of the theory, nor can it be occupied by a closed party as evidence of deeper control.
Motive Does Not Answer in Place of Material
Behavior can be recorded; motive must be inferred from context, history, and the subject's own account. Ye may submit the appeal for attribution, dignity, relations, or several reasons at once. When the organization says he "just wants attention," a single motive explanation can cancel all material. When Ye says the organization "just wants control," he may commit the same error. Even if Ye desires attention, the document conflict still requires verification; even if the managers hold good intentions, the harmful consequences still require repair.
Motive can affect the degree of responsibility — intentional fabrication differs from negligence, for example. Attribution requires words and deeds, interests, and alternative explanations; it cannot rest on outcomes alone.
The subject's self-report is important material, and it may also be incomplete. Respecting the subject does not require treating every self-explanation as final fact. When motive cannot be determined, one may first address visible behavior, authority, and consequences. Responsibility need not wait for complete mind-reading.
Terminology Cannot Substitute for Grounds of Application
The coordination group uses terms such as "defensive," "projecting," and "lacking self-knowledge," which Ye can hardly answer in ordinary language. The terminology appears already to contain professional conclusions.
Clinical and professional concepts have strict contexts and can help in understanding complex experience. When ordinary managers borrow the names, they do not automatically acquire the professional evidence. Professionals may also err — qualification is not an exemption from inspection; those without professional standing may still describe behavior, but should not overstep into diagnosis.
Where terminology is used for consequences to resources and status, it requires sources, definitions, scope of application, and review. Where these cannot be stated, one should fall back to the language of concrete behavior. When Ye learns to strike back at the organization with the same terms, equality is not thereby restored. Both sides may merely contest who is entitled to name, while the facts continue to disappear.
Translate the terms into observable objects, and let the concept resume its explanatory burden. The depth of a theory is not proven by others' inability to rebut it.
Counterexample Relevance and the Self-Confirming Circle
That Ye performs well in another project need not prove there was no conflict at the residency. The situations differ; the counterexample may limit only the personological conclusion. A closed critique that demands any success overturn the risk would also misuse the counterexample.
The scope of the proposition must be fixed first. If the organization says only that Ye leaked material at one meeting, his success in other projects is weakly relevant; if it says he is generally unable to cooperate, performance across situations is directly relevant. The more total the theory, the wider the range of counterexamples it must bear. An interpreter cannot use total language when imposing consequences and then suddenly narrow it when facing counterevidence. Relevance is also affected by time. Past behavior can illuminate history; it does not automatically cover a changed present. Reviewers should state why they admit or exclude a counterexample, so that the boundary is arguable rather than merely declared irrelevant.
The Different Meanings of Missing Records
The project finds no written instruction requiring members to ostracize Ye. The open judgment says that an intentional command is not yet confirmed, while the relational consequences can still be investigated. The closed judgment says that the absence of records proves the instruction was carried out in secret.
Secret actions may indeed leave no trace; absence cannot prove that nothing happened. Nor can it prove, in the reverse direction, that something happened.
One should ask whether the institution normally keeps records, who holds access, whether material was deleted, and whether an independent pattern of consequences exists. The weight of absence differs under different conditions.
Protection can be carried out before a final verdict, especially where consequences are grave and measures reversible. Protection does not require filling the unknown with guilt.
The affected person should not be treated as delusional for lack of a complete record, and the organization should not be perpetually condemned for being unable to prove the absence of secrets.
An honest conclusion can be that evidence is insufficient, that structural risk exists, and that specific intent is undetermined. A layered judgment approximates the material more closely than a single truth.
After Ye is publicly questioned he becomes angry, and the organization says this proves emotional dyscontrol; when he remains calm, he is said to be strategically feigning. Every response enters the same direction. Chapter 4 already distinguished emotion from fact. Here we see further that responses are jointly produced by the environment of evaluation.
If no behavior whatsoever would lower the judgment, the theory is unfit to bear public consequences. It may serve temporarily as private conjecture, but not as a basis for qualification.
Some risks indeed cannot be overturned by short-term good behavior — a long-term assignment, for instance, remains unfinished. The interpreter must state the time and the object of observation, rather than waiting indefinitely.
The subject may also perform counterexamples, and procedures can still examine sustained behavior. The fear of performance does not permit voiding all successes in advance. One of the most severe structures of control is one in which a person's compliance, dissent, silence, and departure all prove the authority correct. Review should require the interpreter to state what obtainable material could change the judgment.
A Predictive Hit Does Not Automatically Confer Authorization
The organization predicts that Ye will dispute again, and he does appeal. The prediction hits, yet it does not explain how the appeal arose or whether it was justified.
Broad predictions hit easily. If any opposition counts as "conflict," it cannot distinguish a flawed procedure from unreasonable conduct. Only by specifying time, behavior, and conditions can comparison occur. Selecting similar events afterward writes contingency into capability. Even where the prediction is accurate, the use of consequences still requires justification: knowing that Ye may appeal, for instance, does not authorize cancelling his appeal in advance.
A failed prediction likewise does not automatically overturn the entire explanation; it may show only that time or variables were insufficient. Revision must leave a version trail.
Reflexive effects require recording: after the organization restricts him in advance, Ye's behavior is no longer the natural outcome under the original conditions.
The Interpretive Authority of Groups and Helpers
Most members accept the organization's version, and Ye becomes the sole dissenter. Numbers indicate group opinion; they cannot directly settle documentary relations.
Members may reasonably trust the organization on the basis of long experience, and may also curtail expression because of relational cost. The meaning of silence remains unreadable.
Who sees the raw material, whether discussion permits dissent, and whether expression is tied to resources — these affect epistemic weight. A minority is not right merely by being isolated. Ye must still provide verifiable material; he cannot obtain an exemption from truth through his position under pressure.
Closed theories often treat consensus as proof of reality and dissent as a defect of the subject. Numbers and psychological explanation then jointly lock the exits.
A procedural outcome can continue to be executed while the factual dispute is retained for review. Joint action need not wait for complete unanimity of thought.
External supporters tell Ye that all his confusion proves the project is under systemic control. He gains language and relations, and may also lose the space in which mixed facts are acknowledged.
Support can increase the opportunities for expression and verification, especially where the original institution monopolizes interpretation; it is not a re-conferral of subjecthood on the person concerned. The risk lies in whether the helpers allow Ye to change his judgment. If Ye acknowledges that the mentor was helpful, he should not be told he is still controlled; if he chooses limited cooperation, that is not betrayal.
Outsiders can provide materials, accompaniment, and alternative points of entry; they do not decide Ye's complete experience on his behalf. Where risks are serious, they may still put forward explicit recommendations for protection.
An anti-control theory that absorbs all dissent as internalized oppression has the same form as the original closed structure; a different ethical stance does not remove the epistemic problem. The subject needs multiple exitable supports. No helping relation may acquire permanent authority through "only we understand you."
An explanation can be logically consistent and still be unfit for use because of false premises, insufficient evidence, or improper consequences. Philosophical structure cannot substitute for investigation of reality.
RC emphasizes relations of observation and process, yet this does not guarantee that its users are naturally open. An interpreter, too, can invoke "all observation participates in reality" to demand that the affected person bear responsibility for harm.
Ye's responses enter into relational outcomes; this does not mean he holds the same options as those who control resources. To discuss differential responsibility, one must further weigh each person's actual conduct, commitments, knowledge, capacity for control, and available actions; responsibility cannot be distributed evenly merely because all participate in the same process. Nor does "all explanation constructs" entail arbitrariness. Objects, records, bodies, and the responses of others continuously constrain the narrative.
One inverse use of the theory cancels temporary facts in the name of process, cancels differences of authority in the name of the observer relation, and then interprets any counterexample as a deeper process.
A critical theory must specify, from within itself, its stopping conditions, its counterevidence, and its responsibilities — not merely say that bad actors misuse it. Self-consistency is a starting point, not a certificate of exemption.
The Tests Revision Must Bear
The residency requires that every high-consequence explanation state its propositions, evidence, deadlines, counterexamples, and scope of decision. Global psychological judgments that cannot fill in this form may not be used for qualification. This set of forms can itself become formalistic. A staff member who writes down a counterexample that will never occur remains closed.
Ye needs the opportunity, under proportionate protection, to complete tasks, submit material, and receive independent review. If all opportunities have been cancelled, the theory cannot demand that performance prove it wrong. Reviewers must be able to change labels, resources, and dissemination, not merely append opinions. Where counterevidence has no consequences, it is not an institutional exit.
An explanation may be maintained under new evidence, but it must still answer why. Openness does not mean that every objection changes the conclusion. Regular sampling of success and failure cases checks whether successes are always credited to the theory and failures to the object. Patterns reveal closure more than slogans do.
If the organization replaces the theory at every criticism, the rules lose predictability. Openness requires a stable core and explicit updates, not continual accommodation. Evidence strength, magnitude of consequence, and reversibility jointly determine the scale of change. Where consequences are high and evidence weak, restrictions may be suspended first; where consequences are low, observation may continue.
Version records state why one believed at the time and why one later changed, so that those affected by old decisions can seek remedy. The old theory may remain valid within its original scope and need not be entirely deleted on account of the revision. The new boundary of applicability should be visible.
The subject also needs stable expectations. Having been judged restored today, one cannot be made a problem again tomorrow by a redefined concept without new material. The results of revision must enter training, permissions, and resources; changing only the text and not the consequences is narrative update, not institutional learning.
Evidence Thresholds and Proposition Scope
Ye discovers that material of the same strength bears different thresholds in the two directions. One smooth exchange supporting the program's effectiveness is written into achievements, while one instance of discomfort questioning the program is required to exclude all other causes. A single success suffices to increase certainty, while many failures are still treated as accidental.
Evidence thresholds need not be mechanically symmetric. A proposition about "at least one improvement exists" can indeed be supported by one reliable case; a proposition about "safe for everyone" is damaged by a single definite counterexample. Symmetry should hold in the relation between proposition and evidence, not in a demand for equal quantities of supporting and opposing material.
"Ye participated in the conversation today" states, first, that he took part in one activity; whether it was voluntary, whether it can be sustained, and what burden it imposed require separate questions. It cannot by itself prove that the conditions for judgment have been restored, that the explanatory framework is generally correct, or that the program does no harm. Each expansion of the conclusion requires additional material.
Conversely, one instance of silence cannot prove the program ineffective. It may relate to fatigue, language, relations, or the events of that day. The open explanation first determines to which proposition this silence is relevant, and then decides whether to adjust certainty; the closed explanation permits material to accumulate in one direction only. Review can decompose key claims into five kinds: existence, frequency, cause, scope, and value. The decomposition is not a language game; it prevents a local fact from endorsing an entire theory.
Shared Conditions and Causal Contribution
The project team says that without this explanation Ye would not have recovered. In reality one cannot have the same person undergo both paths at the same moment, so the counterfactual cannot be directly observed.
This does not mean that no causal claim can be evaluated. One can compare different times, similar cases, alternative supports, and known mechanisms, and can also state honestly which causes still cannot be excluded. Real-world assessment requires professional methods; this chapter treats only the discipline of inference.
Ye's increased communication may derive simultaneously from quiet space, a stable routine, a respected right of refusal, and new relationships. The theoretical language may have helped, yet it is not necessarily the sole or principal factor. When an explanation absorbs all shared conditions into itself, it acquires unlimited scope.
The better practice is to record supporting conditions, times, and outcomes separately, and to compare which explanations fit the material. Separate recording does not amount to identifying the independent causal contribution of each factor, still less to arbitrarily withdrawing necessary support in order to prove a theory.
"If he does not accept the explanation, he will fare worse" may also be used to block alternatives. Such a claim bears a heightened burden of proof, because it directly affects the right of exit. Where it cannot be verified, low-risk alternatives and a way back should be retained, rather than letting uncertainty support only the incumbent power.
Omissions of Classification, Scale, and Translation
The team classifies positive reactions as "genuine improvement" and negative reactions as "a phase of resistance." The classification may derive from clinical or practical experience, but if the category is decided only after the outcome appears, the theory becomes immune to failure. The classification itself requires observable conditions. How long is the resistance phase expected to last, what are its independent manifestations, and what outcome would lead the team to reclassify? Without answers, the name merely packages the unknown as the known.
Ye sometimes participates actively and sometimes leaves early. The team can first record the time, the task, and Ye's reported feelings at each participation and departure, rather than inferring capacity or burden from attendance alone. Concrete records may resist sorting into a single success or failure category, yet they preserve the basis for later comparison. Not every case suits prolonged suspension; where immediate safety is at stake, proportionate protection comes first, and the explanation is reviewed afterward.
The authority of the classifiers must also be visible. Who can modify the file, whether Ye can see, supplement, or appeal it, and whether other professionals can propose different explanations — these determine whether the category becomes a single reality.
Conditions of Applicability After Scaling
An explanation that is useful within an individual support relationship may produce different consequences when extended to the whole institution. A small team corrects misunderstandings through familiarity and trust; a large-scale program depends on forms, training, and performance metrics, and fine-grained conditions are easily lost.
Early cases are therefore not spurious, yet they cannot directly represent the results of scaling. If the expansion simultaneously changes the executors, the subjects, the incentives, or the speed of feedback, the conditions on which the original results depended must be re-examined; the direction of effect cannot be predetermined merely by the increase in numbers. If the original project's success depended on low numbers, long-term relationships, and the director's constant correction, then replicating only the terminology and the procedures replicates none of the actual mechanism.
Managers may select the most presentable success cases for promotion, leaving care that resists standardization to be filled in unpaid at the local level. Whether this is intentional requires evidence, but the transfer of cost is itself observable.
An evaluation of scale should record average outcomes, boundary losses, the handling of complaints, and differences in execution. If a theory holds only in carefully selected environments, it should contract its claims rather than blame reality for being insufficiently ideal.
Translation Should Remain Traceable
Ye's words must be written up as professional records by staff members. Translation is unavoidable: the language of daily life cannot enter every institutional field unchanged. Good translation lets different roles collaborate; the danger is that once the translation replaces the original words, it can no longer be traced.
The same sentence — "I do not want to continue" — can be translated as fatigue, resistance, a risk signal, or an explicit withdrawal of consent. Different categories carry different consequences for authority.
Records can separately preserve Ye's expressions, the staff interpretation, the reasons for adopting that interpretation, and the subsequent decisions. Verbatim recording is not required every time; at key junctures the differences should be made visible.
Translators also require protection. If every judgment can be held to infinite after-the-fact account, frontline personnel will write only the most conservative templates. Clear duties, review support, and honest expressions of uncertainty are more realistic than demands for perfect neutrality. When an institution recognizes only professional language, the subject must first learn to describe himself in terms the system endorses before he can receive help. This structure can improve communication, and it can also strip reality-status from those unable to master the terminology.
The Preservation of Counterexamples and the Cost of Explanation
Ye reports discomfort, and the team says experts must judge whether it constitutes a valid counterexample. Professional judgment can rule out misunderstanding and short-term fluctuation, yet it cannot take over the entire interpretive authority over experience.
The subject can provide important material about feelings, aims, and what is bearable; professionals can provide risk and research judgments in their domains; the institution holds resources and rules. Each party's knowledge may still be incomplete or mistaken. The sources and limits of each party's material therefore need to be stated.
In support that one may freely exit, Ye holds the core decision over whether to continue. Where the direct safety of others is at stake, protective authority may temporarily expand, but its scope, duration, and avenue of appeal must be stated. Joint decision must not turn into a demand that the weaker party accept the professional conclusion as the price of cooperation. Nor can all professional advice be disparaged as power language, leaving necessary risks unborne.
An open system acknowledges that material is dispersed and designs multiple paths of entry. A closed system often concentrates ownership of material in the interpreter, and then proves the interpreter indispensable by the subject's inability to provide qualified evidence.
A theory exists not only in texts but also in meetings, training, evaluations, and files. If each anomaly requires new forms, new personnel, and stronger supervision to maintain the original conclusion, the maintenance cost is itself material.
Rising cost does not automatically prove the theory wrong. Complex objects may indeed require more resources. The question is whether the added resources improve the situation of the object, or chiefly protect the explanation from challenge.
Replacing the framework entirely would also require retraining, interrupt relationships, and produce new blind spots. A critic cannot tally only the maintenance costs of the current system while imagining the alternative as costless.
One can compare on a small scale: retain basic support, change the explanatory language or the decision authority, and observe burden and outcomes. High-risk practices in reality require ethical and professional review; this chapter provides no experimental procedures. When a system requires the subject to prove himself continually while the interpreter scarcely needs to account for failure, its epistemic and ethical costs are both concentrated on one side. This asymmetry deserves examination more than the elegance of the theory.
What is decisive about a closed explanation is not whether it uses religious, psychological, or managerial vocabulary, but its manner of handling material: support, opposition, silence, departure, and failure can all only increase the same conclusion; definitions can move after the fact; and the subject's qualification is bound to acceptance of the explanation. Such an explanation cancels the very conditions under which it could fail.
A theory needs a core, complex causality needs auxiliary conditions, and professional terminology and group experience have genuine value. The line of critique lies in whether reality-reachable counterexamples exist, whether revisions acknowledge their versions, whether the unknown can remain unknown, and whether explanations take effect only within the scope of their qualification. Only when an explanation can say how it might be wrong does it remain connected to reality.
A closed explanation may also affect what the person concerned prepares to express. If Ye anticipates that every objection will be reinterpreted as a symptom, he may delete the non-conforming feelings before opening his mouth. Without any explicit external prohibition, this material may never enter the discussion at hand.
Yet the deletion of dissent may be Ye's temporary strategy for avoiding consequences, and it may also accompany a change in his own judgment; it cannot be determined directly from the outward act. The next chapter compares these situations, asking what material supports the interpretation of "self-censorship," and what material remains missing.