At the end of Chapter 18, Ye's assistance relationships had already contracted into limited services: Su retained only the role of archive and deadline consultant, Cen's residence rules concerned shared living alone, and the peer team evaluated him on his work materials. Lin no longer issued any demands to him, and the project gave no new instructions for his daily life. By intuition, control ought to end when the controller exits the scene. Yet when Ye wrote messages to new colleagues he would still first delete the sentences that carried doubts, still pause at the sound of keys at the door, and still rehearse the worst outcomes before filling out an ordinary form. The external executor had left, but the execution continued. Taking these fictional reactions as its setting, this chapter discusses which relational conditions supporters can change: how agreements are honored, how feedback is verified, and whether the person concerned can pause or raise objections. The "action-feedback unit" is a scale for analyzing agreements, not a unit of measurement for psychological recovery. Chapters 19 through 21 discuss respectively the supporting conditions of relationships, of body and time, and of records; they prescribe no order of precedence for recovery.
After Leaving the Old Environment
In the story, Ye carried the evaluative language of the project period into new settings for expression: first imagining how others would interpret him, then revising and deleting until nothing could be faulted. We call this cross-contextual pre-emptive self-review "internalized surveillance" here only to describe the behavioral relation within the setting; it is not a clinical category. It makes Ye's doubts harder to bring into new relationships, but it does not prove that his new colleagues intend to profit from his silence. Old habits, current risks, and personal choices of expression must be examined separately.
Internalized surveillance cannot be inferred directly from cautiousness. Caution may arise from real risks in the new environment, from professional habit, or from temperament. What supports a judgment of internalization is material of this kind: the content of the review copies the rule-version of the old environment; the review runs equally on new addressees, in new places, even in private notes; and the subject himself cannot say who his present audience actually is.
Ye expected that raising doubts would cost him opportunities, yet the new team had so far produced no corresponding punishment. This poses a question for examining the old expectation, but it is not yet sufficient evidence to refute it: non-occurrence may be due to different circumstances, or it may be due to no one daring to raise doubts. In this setting, non-expression reduces the new material available for comparison. Judgment about reality requires knowing, at the same time, the actual distribution of opportunities, the consequences of similar expressions, and the conditions for appeal; one cannot interpret all caution as residue of the old relationship.
Regarding post-trauma reactions, NIMH materials note that cues related to the event may trigger distress reactions, and that some people continue to feel fear after the danger has ended; individual courses vary, and diagnosis and treatment require professional assessment. This supports the limited judgment that "leaving does not necessarily make the reactions cease immediately"; it cannot prove that Ye's situation falls under some diagnosis, still less attribute every pause to control. That the sound of keys triggers a pause is this chapter's narrative setting, used to show that supporters cannot demand immediate adaptation from the person concerned merely because the old relationship has ended.
The Account of the Rupture Between Action and Result Remains Unsettled
Ye still saved screenshots and prepared fallback plans for ordinary agreements; in the story these behaviors continued the wariness of the old environment. They may consume time, and they may still serve practical purposes in the new environment; they cannot be called unprofitable residue without comparison. Chapter 7's "predictability debt" serves here as a cost metaphor: one should check which additional work remains necessary and which can be reduced through explicit commitments and reliable records, rather than demanding that the person concerned first cease all defensive measures.
"Trusting the world" is difficult to convert directly into commitments that a supporter can honor. By contrast, returning tools, delivering materials as agreed, and explaining residence rules all have definite addressees and verifiable consequences. This chapter calls them action-feedback units, used to compare whether relationships are reliable. Small units make it easy to limit certain costs of failure, yet they are not the sole entry point to recovery; a single honored agreement provides material about that agreement alone and does not prove that the world is safe overall.
Encouraging Ye to trust himself may express support, but it cannot substitute for showing whether the new environment is reliable. For supporters, the more bearable responsibility is to honor specific agreements, explain exceptions, and accept corrections. How Ye comes to form an overall trust may also involve bodily condition, past experience, and professional support, and this book cannot prescribe a uniform sequence for it. Repeatedly checking small matters also has different causes; one should not infer from it that a person has regressed, or is recovering along the correct path.
Locally Predictable Arrangements and Bodily Feedback
Chapter 7 already distinguished predictability from trust, and this chapter pushes that distinction to the recovery side. The world itself promises no stability: equipment fails, people change, organizations rewrite their rules. What Ye needs to recover was never an overall expectation of the world, but expectations of certain specific relationships: within this relationship, what feedback will I receive if I act this way; when feedback differs, is the difference explained with reasons; can errors be brought back for correction. Local predictability is therefore a relational property, not an inner state. Its criterion lies not in how Ye feels but in whether the relationship actually delivers: whether the line between commitment and consequence is observable, whether exceptions come with stated reasons, whether versions are visible. To write recovery as inner work misses the object itself — what needs repairing is the feedback structure.
For an action-feedback unit to accumulate material, three connections must be visible at once: the correspondence of action and feedback (what I did, what came back), the correspondence of feedback and reason (why this result), and the correspondence of feedback and review (whether there is a place to raise the matter when the result is wrong). If any one of the three breaks, the unit degenerates into the decoupled state described in Chapter 7: there is action, there is result, and between them no learnable relation. This constitution shows that local predictability is not a psychological exercise but a miniature institutional fact — even with only two people, it takes a little record-keeping, a little willingness to give reasons, a little space that permits counter-questions.
In this story, after several agreements were honored, Ye tried to extend his expectations of Cen to colleagues and service institutions. This change is a setting of the case, not a stage every recovery must pass through. One honored agreement speaks only to that agreement; a new domain still requires new material. Extending too quickly may let a single disappointment drag down the whole judgment, while refusing any extension may forfeit reliable support. What is compared is the correspondence between expectation and the relevant context, not the demand that Ye learn to trust at a fixed speed.
Ye may still feel anxious after a run of honored agreements, and may feel at ease in the absence of any record of honored agreements. These two fictional situations serve to distinguish feeling from relational evidence; they do not establish which of the two is the norm of recovery. Feelings themselves are material to be taken seriously; whether an agreement is reliable still requires records, reasons, and consequences. Neither kind of material can answer for the other alone.
Bodily Feeling as a Source of Feedback
In the story, Ye began to be able to describe his sleep, diet, and daily rhythm without first submitting to the project's evaluation. His statements of discomfort thereby regained a place, but bodily feeling is not a truth requiring no interpretation: the same manifestation may have multiple causes, and health conditions require appropriate assessment. What is protected here is the standing to report feelings and to take part in arrangements; specific physiological questions are discussed separately, alongside empirical materials, in the next chapter.
The bodily dimension of recovery carries a reverse risk: turning the rebuilding of rhythm into a performance that must meet a standard. Supporters, or Ye himself, might set "regular hours" as the new standard and record every deviation as regression, whereupon local predictability flips over into a new project of self-scrutiny — pursuing the freedom of a new environment with the methods of the old. The beneficiary of this mechanism is the supporter's sense of visible accomplishment; the bearer of its costs remains Ye. The counterexamples must also be kept: external structure can genuinely help rhythm recover, and shared rising times or quiet agreements in a residence have real effects; they cannot all be treated as discipline. The distinction lies in whether the structure is negotiable and modifiable, and whether deviation is interpreted as a defect of personality. The condition for recovery is that rhythm serves the body, not that the body serves a report card for rhythm.
Whether fear diminishes, what treatment to adopt, and when to encounter the relevant situations cannot be derived from the philosophical analysis of the action-feedback unit. Methods within professional treatment require professional judgment, and this chapter does not equate everyday agreements with therapy. Supporters can take responsibility for the safety boundaries and commitments of the current environment, respect pauses, and refrain from interpreting the person's continued fear as a refusal to change; whether Ye's reactions abate remains a matter of narrative setting and cannot be used to demonstrate efficacy.
Changes in Surveillance and Punishment Expectations
The sentence-patterns Ye recalled while pre-censoring had specific origins: "put that way you sound uncooperative"; "admit you might be wrong first, then speak." Identifying the source of the old demands lets him ask anew whether those demands apply to the occasion at hand. In this case, he thereby kept a question he had originally intended to delete. This setting shows how source-comparison enters judgment; it does not prove that identifying origins necessarily lightens self-scrutiny; even knowing where a demand comes from, a person may still continue to follow it out of current risk, habit, or other reasons.
The evidential threshold for a bystander to judge that Ye "remains under internalized control" is the same as Chapter 12's requirements for internalization: only when similar scrutiny recurs across addressees, places, and time — even entering private domains unrelated to the old environment — does it support the stronger judgment. A single rewording, occasion-specific politeness, or strategic softening does not count. This threshold works in both directions: it prevents supporters from prematurely declaring Ye incapable of self-judgment, and it prevents anyone from re-pathologizing every one of Ye's cautious acts. The material for the judgment must come from observable differences in expression, not from conjectures about his interior.
Monitoring one's own self-scrutiny may itself become new self-scrutiny: after each deletion and revision Ye would reproach himself — "revising again" — so that scrutiny is layered upon scrutiny, with a standard that can never be met. This is the structure the old environment was best at manufacturing, now run by Ye alone. The condition for recovery is to change observation from grading into recording: notice the revision, register it, and allow it to be followed by no immediate change. The output of observation should be comparable material, not another failing report card.
How Anticipated Punishment Is Replaced Segment by Segment
In the old project, Ye had lost opportunities because of specific doubts. The story sets his later expectations as extrapolated from those experiences: he began to fear that expressing himself to different people would also be punished. Examining this extension requires comparing addressees, occasions, and consequences; one cannot begin by classifying all defensive measures as over-generalization. Explanation, reliable new experience, bodily condition, and professional support may all affect expectation, and each of their contributions requires its own material; on this basis this book asserts no single path of change.
The new team accepted Ye's questions and stated its reasons for handling them, providing one experience different from the old project. Whether it changes Ye's overall expectation, and how long such change takes, cannot be derived from the number of honored agreements. Supporters can be responsible for truthful feedback, but they should not turn every ordinary interaction into a test of recovery progress. Ye may tell those with whom he works continuously what kind of response he needs, and he may also keep experiences he does not wish to make public; the scope of disclosure remains decided by the task at hand and the will of the person concerned.
One genuinely bad experience during recovery — an expression that did in fact bring punishment — should not be used to prove the old expectations entirely right. The key of the test remains Chapter 7's standard: whether the consequence is relevant to the content, proportionate, and appealable. If the new environment's response to ordinary dissent is a status-based denigration, then this is not a failure of recovery but a problem of the new environment itself, and what the subject learns is a valid prediction about this new environment. The distinction between "residual expectation" and "genuine recurrence" rests on the actual feedback structure of the current environment, not on the emotional intensity with which the expectation is confirmed — the intense sense of familiarity, the "I knew it all along," that accompanies confirmation of an old expectation is precisely the least reliable.
Support Networks and the Labor of Recovery
A single unit has limited recovery value; value comes from multiple units that do not lock one another in: the residence arrangement is one, work delivery is one, the archive consultation with Su is one, bodily rhythm is one. The meaning of multiple paths lies at the system level: when one fails, the rest remain, and the whole does not collapse with any single relationship. This is also what it means to translate corrigibility out of institutional language back into the language of life: for a person to possess corrigible conditions comes close to saying that in his world there exist multiple feedback circuits that can be tested and retained separately. The connections need not all succeed at once — demanding simultaneous improvement across every domain is itself the totalizing demand reprised at a new stage.
A small amount of record-keeping can help the units support one another: whether an agreement was honored resists the rewriting of memory better on paper than in the head — especially for Ye, whose old environment systematically rewrote what had happened. But this chapter makes only a limited claim for records: their function is to preserve comparable material, not to manufacture a duty to produce evidence. Requiring a person in recovery to record everything presses the labor of verification back onto the individual once more. How shared records, versioning, and the preservation of counter-evidence more systematically restore trust in one's own judgment is the dedicated topic of Chapter 21, and we do not pursue it here.
After several successful direct expressions, Ye might explain them this way: it was only luck, only that the other had not yet seen him clearly, only an exception this time. Success at the level of experience is confiscated by the level of interpretation, and the old interpretation sustains itself by absorbing counterexamples — this is the closed structure analyzed in Chapter 11, now run by Ye alone. The condition for recovery is not to force him to rewrite his self-narrative at once, but to allow interpretation to lag behind experience for a while: he can complete direct expressions while still not believing they will last. Interpretation usually updates only after a large body of material; the inconsistency in between is not hypocrisy but the normal shape of recovery.
Cen's residence rules are clear, stable, and unconnected to public positions; this predictability is itself recovery material — within it Ye re-experienced that agreements can be honored. But another person's predictability cannot be demanded as an obligation: Cen has the right to change, to travel, to alter arrangements, and to stake recovery on some particular person remaining forever stable is to replace an old single dependency with a new one. Su, Cen, and the peers each provide one source of feedback, which is precisely the polycentric structure described in Chapter 18; recovery therefore requires no one to become a perfect stabilizer — only that sources of stability not be monopolized.
The Distribution of Recovery Labor
Rebuilding local predictability requires available margin of time, space for low-cost failure, and sources of feedback that do not depend on a single interpreter. None of these three is a purely inner resource: a person without time cannot bear the failure of small units, and a person whose housing and income are suspended has no safe testing ground. The principal labor of recovery falls to Ye irreplaceably — no one can reconnect action and result on his behalf — but the conditions for that labor are supplied mostly by the environment. The beneficiaries are all those who will work with Ye in the future, receiving a collaborator whose judgment is corrigible; if the costs are borne entirely by the individual in recovery, that distribution itself deserves scrutiny.
Here there is a harsh possibility that must be stated squarely: the structure that caused the harm may cite "external control has already disappeared" as its reason to declare everything that remains a private affair. The old project settles its account by paying off the contract, the industry does not repair the stigma it spread, public institutions provide no transition resources — and then everyone says in earnest: he is free now; what remains is his own homework. This is isomorphic with the "it was all your choice" responsibility-eraser criticized in Chapter 17, merely displaced to the recovery end: using a formal endpoint to cancel the structure's obligation of repair. The mechanism is clear — maintaining the status quo favors the party that no longer pays; the evidential threshold is to ask who created the unusable conditions and who has long refused to repair them while able to; the counterexample is equally real — personal labor does have an irreplaceable part, and structural repair cannot cancel it, just as it cannot exempt structural responsibility. How the archive arrives ahead of the person himself, and where the authority of judgment falls on returning to work, will be tested concretely in Chapter 23 by following Ye's first week; here we set down only the principle: recovery is personal, its conditions are public.
If Cen finds the support overly dependent on him, he can discuss with Ye whether other sources are needed, while checking that any new relationship is safe, credible, and adopted by Ye voluntarily. Contact cannot be forced in the name of diversification, nor can other help be blocked on the grounds of being needed. This reminder connects with Chapter 18's analysis of dependency structures; here we add only one point: in Ye's setting, stable responses provide material for comparing new and old situations. To judge whether dependency constitutes a restriction, one must further examine permissions, the cost of refusal, and actual consequences; one cannot distinguish new from old relationships by the number of sources alone.
Re-exploitation and Current Constraints
This chapter also retains a possibility requiring examination: Ye may, out of craving for stable responses, overlook the other costs of a new relationship. One cannot conclude from this that everyone in recovery suffers from "feedback hunger," nor judge a person to be captured because he likes immediate responses. The specific conditions of the new relationship should be compared: whose goals the feedback serves, whether refusal carries proportionate consequences, and whether external contact and exit remain usable.
The more structured form makes certainty itself into a product: clear hierarchies, instant responses, a system in which every step has a definite consequence. On its surface this is exactly what Ye lost and craved; the cost is that its consequences point toward the system's goals rather than the subject's capacity for revision — as Chapter 7 said, a harsh institution can be highly predictable, and predictability alone never proves its consequences just; the Ye of the story still needs to check this clause. The mechanism of dependency's reproduction is isomorphic with the old environment: feedback exists stably, but the content of the feedback systematically trains the subject to invest for the sake of feedback itself. The counterexample remains real: time-limited, reviewable, purpose-defined high-structure environments — training, treatment, phased joint projects — whose short-term demands for obedience do not constitute hijacking. The discriminant is the structure's interpretive authority and the position of exit, not the strictness of the structure.
Taken together, the two kinds of stability can be separated by three questions. First, can the feedback be interrogated: when Ye asks "why this result," does he receive reasons or an evaluation of his status? Second, is failure permitted: after a small unit fails, does the relationship continue, or does standing shrink? Third, does the stability point toward more sources: does the system tolerate and support Ye's building feedback circuits elsewhere, or does it redefine external contact as disloyalty? These questions examine different aspects of a stable arrangement; one cannot characterize a whole relationship on the basis of any one of them. This set of distinctions provides no portable instrument of detection; it only directs attention to the structure, away from the surfaces of tone and warmth.
Current Constraints Cannot Be Classified as Psychological Residue
Difficulties after leaving may involve, at the same time, current income, archive restrictions, unresolved disputes, and health status. Comparison before and after the change of environment can supply clues, but it cannot by itself distinguish so-called "internalized residue" from "structural constraint"; the two may also coexist. External arrangements still in force should be checked one by one, appropriate assessment sought for health questions, and one explanation should not be allowed to cancel other conditions that need addressing.
What this chapter has discussed is how everyday agreements provide verifiable conditions of support; it has not demonstrated one path of psychological recovery. Recovery in reality may require many kinds of support, and neither sequence nor outcome can be decided by the story. Nor should the action-feedback unit become an examination the person concerned must complete item by item: it serves to check whether supporters have honored their commitments, not to certify who has recovered adequately.
The rebuilding of small units depends on available margin of time and space for low-cost failure. When both are zero — housing expiring next week, income covering one month, every mistake bearing directly on survival — local strategy cannot accumulate, not because the subject tries too little, but because the conditions for testing do not exist. The condition of failure is itself a clue to responsibility: whoever makes the cost of failure too high for testing is continuously manufacturing positions that cannot be corrected. At that point the direction of repair lies not at the individual level but at the level of conditions; Chapter 20 will develop this head-on in treating the domination of body and time.
The Boundary Between Supporting Conditions and Psychological Recovery
Gathering the scattered conditions of the whole chapter: basic margin of time and body, so that failure can be borne; space for low-cost testing, so that small units can be attempted repeatedly; feedback not monopolized by a single interpreter, so that experience is not confiscated at any single point; the possibility of counter-evidence entering, so that expectations can differentiate rather than totalize. Not one of the four is a purely inner condition; they are distributed across housing, income, relationships, and institutions. This accords exactly with this book's stance of the unity of subject and object: recovery was never an operation a person completes inside himself, but the subject's regaining, within relationships, of a position from which he can test the world through action and revise action through results.
The persons and processes of this chapter are fictional, and the relational checks proposed here cannot substitute for clinical judgment. The next chapter will distinguish bodily condition, task performance, and arrangements of power: what physiological research can support, what further evidence particular relationships require, and how the person concerned can retain participation in the arrangements of body and time while receiving help.