QUESTION-045: Is a mental disorder that outlasts its trigger a protective lock-in at the individual scale, a regime kept stable by behaviour that blocks its own disconfirmation while the capacity to leave it degrades, and does cutting the maintaining link, with the regime's content held fixed, produce lasting change?

The question

TERM-020 defines protective lock-in: a protective regime reorganizes activity so that it “generates the conditions that justify its own continuation and reduces the capacity to leave it”. CLAIM-079 says a regime can be stable without being viable. CLAIM-025 says the operative regime and the capacity to transform it can fail separately. The record’s cases are firms. This question asks whether a maintained disorder is the same mechanism in a person, not an analogy: feedback through the conditions of efficacy, and not the contents of a fearful belief. It also asks what evidence would show that.

Why the record needs it

An individual case would be a second realizer of CLAIM-099’s result, and one with a treatment literature that intervenes on the maintaining links. CLAIM-018 reads the result through Canguilhem as an “illuminating analogue”. This would test it on disorder itself.

What the record holds

• THEORY-058: disorders are “maintained by self-reinforcing causal loops” and “outlast their triggers”: a stable state kept up by hysteresis. Safety behaviour “keeps the meaning from being disconfirmed”. Limit: no individual time series yet.
• THEORY-050: experiential avoidance becomes “a general strategy” and “behaviour narrows”. That is a loss of the capacity to leave the regime. Limit: correlational, mostly self-report.
• THEORY-052: thwarting yields “rigid behavior patterns” that “persist into new situations in which they are not needed”. Limit: direction unknown.
• THEORY-053 says what evidence counts: change in the theorised process that is “specific and comes before the outcome”. Cutting a link while holding exposure fixed is the will line’s intervention criterion (QUESTION-015) in clinical form.
• THEORY-285 (LIT-983, NOTE-766): in an active-inference model, an overly precise prior that one is in a harmful context makes exploration look uninformative, so the agent withholds the evidence that would correct it. Restoring exploratory sampling rebalances the prior; lowering precision only removes symptoms. It answers the question in part, at the model level only. Limit: illustrative simulation with no individual data, and the two intervention arms differ in environment, so exposure is not held fixed while the link is cut.

What would answer it

An experiment that cuts one maintaining link, such as safety behaviour, while holding the regime’s content fixed, and produces lasting change that the same exposure without the cut does not. Or individual data showing hysteresis.

Note of 2026-10-18: overcontrol, and an organizational transfer

The exchange of 2026-10-17 brought a treatment literature this question did not list. A42, at the owner’s U42 prompt on ACT and DBT, added Lynch’s radically open DBT, which “explicitly addresses maladaptive overcontrol, rather than treating deficient self-control as the primary problem”. In this question’s terms overcontrol is a candidate individual-scale protective lock-in whose maintaining link is inhibition itself: a regime of control that restricts the openness and feedback that would show it is no longer needed. The evidence is thin. A42: “promising but not definitive”, with one randomized trial (Lynch et al. 2020) and one systematic review (Hatoum and Burton 2024). Neither is held yet.

A42 then ran the transfer the other way, from the individual to the organization: tighter safeguards that yield “Less disclosure of near misses” and “Fewer challenges to mistaken decisions”. That is filed as CLAIM-349, whose defeater includes this question: the organizational form must predict something the individual-scale lock-in account does not.

A29 §8 offered a dynamical form for “individual data showing hysteresis”, ż = μz − z³ + h, which TERM-020’s note sets out with its caveat.

Note of 2026-10-23: a candidate answer, governance-maintained anxiety

The owner read the exchange’s “a governance prior can become self-protecting by reducing the access of the processes most likely to provide evidence against it” as “a formalization of anxiety” (workbench entry of 2026-10-24, U14). Filed as a candidate answer: CLAIM-446. Its distinctive part is the allocation of authority over inquiry and intervention, not the familiar point that avoidance maintains anxiety. Under the avoidance a defensive process selects, the hypotheses of danger and of safety predict the same outcome, so safe episodes do not discriminate them, and the inference from one’s own precaution to danger lets the regime supply evidence for itself. It joins CLAIM-297’s note on a threat monitor whose interpretation raises its own authority and THEORY-285’s simulation of an over-precise harmful-context prior; neither is patient data. The contrast this question asks for is the claim’s test: hold the represented threat fixed and change the defensive process’s authority or evidence access, and conversely. A proposed simulation is CASE-153, and the cross-scale form, from emergency powers to the security dilemma, is CLAIM-445.

Note of 2026-10-27: the evidential part split off

The status review of this date split this question under ADR-049. What evidence would show that a maintaining link, not the regime’s content, keeps a disorder in place is QUESTION-292, Answered by THEORY-053: not a treatment’s efficacy, but change in the link that is specific and comes before the outcome. This question keeps what is still open: whether any disorder is such a lock-in, and an experiment that cuts the link with the content held fixed.

Note of 2026-10-29: RefraMED, read, gives nothing on mechanism

The randomized trial the note of 2026-10-18 names, Lynch et al. 2020, LIT-1223, is now read (NOTE-1358), so that note’s “Neither is held yet” no longer holds for it. It gives nothing on whether overcontrol is a protective lock-in. Its title promises mechanisms, and the mediators were measured, but no mediation analysis is reported, only contrasts between arms. It selected on refractory depression with undercontrolled presentations excluded, and it did not measure overcontrol. The question is unchanged.

Note of 2026-10-29: the trial’s monograph, read

The earlier note of this date says the BJP paper gives nothing on mechanism. The trial’s NIHR monograph, LIT-1838, now read (NOTE-1421), gives an attempt and no result, so that note’s conclusion stands with a refinement: the mediators were analysed, and the analysis did not identify them. They were sessions attended, alliance and DBT skills, not the flexibility and emotional expression RO DBT names; mediator and outcome were both at 7 months; and the instrumented analysis failed on weak instruments.

Overcontrol was not measured as such. The trial’s proxy is a cluster A or C personality-disorder diagnosis (78% had at least one Axis-II disorder), used as a stratifier, and its interaction with treatment is reported only as not significant. That is the nearest the trial comes to this question, and it is no evidence either way. The question is unchanged.