Lever Freeze-or-Cap¶

Control rule — instantiates Stock-First Control Restoration

Halts or ceilings a lever that has stopped transmitting, so a flat response can no longer be answered with more force — capped where its side-effects overtake its shrinking benefit.

Once a lever is judged non-transmitting, something has to stop hands from reaching for more of it. Lever Freeze-or-Cap is the control rule that does so: it locks the lever at its current setting (a freeze) or forbids it above a ceiling (a cap), converting the instinct "no response, push harder" into an enforced stop. Its defining move is where it sets the ceiling. It does not place the cap by the benefit the lever still delivers — that benefit has gone to nearly zero — but by the side-effect burden the lever keeps piling up as it is pushed. You stop at the point where the harms of leaning harder outrun the vanishing good, and you hold there until the stock underneath is repaired.

Example¶

A patient's chronic pain has stopped responding to a stable opioid dose. The reflex — the patient's and often the clinician's — is to titrate upward. But the receptors have downregulated: this is tolerance, and each increase now buys ≈no added analgesia while respiratory-depression and constipation risk keep climbing. A Lever Freeze-or-Cap intervenes on the dose itself. It freezes the dose at its current level — or caps it at a stated morphine-milligram-equivalent ceiling — and bars further escalation until the underlying stock, receptor sensitivity, has been restored (via rotation to a different agent or a supervised drug holiday). Crucially, the ceiling is chosen where the side-effect burden has overtaken the relief, not where the pain would ideally be. The rule is deliberately blunt precisely so it can resist the in-the-moment "just a little more" that tolerance keeps inviting.

How it works¶

It neither diagnoses nor repairs — it interdicts, and it is built to be hard to argue with in the moment:

Freeze or cap. Either lock the lever at its present value the instant collapse is confirmed, or set a ceiling it may not cross.
Set the ceiling off the harm curve, not the benefit curve. The cap tracks the side-effect-burden meter: keep raising and the meter's reading — toxicity, distortion, collateral cost — crosses the point where marginal harm exceeds marginal good.
Make it dumb on purpose. A hard-edged rule with a narrow override resists the escalation reflex better than a judgment re-litigated every time the outcome looks bad.

The stop is a holding action: it protects the stock from being drained further and buys the room in which repair can happen.

Tuning parameters¶

Freeze vs. cap — a hard lock at the current setting (maximally conservative) versus a ceiling that still allows movement below it. Freeze is safer against depletion; a cap preserves some operating room.
Ceiling height — where the cap sits on the side-effect-burden curve. Lower ceilings protect the stock and limit harm but forfeit any residual benefit; higher ones gamble that a little more force still helps.
Override path — whether and how the cap can be lifted, and on whose authority. No override risks brittleness; an easy override quietly defeats the rule.
Trigger source — what fires the freeze: a Gain-Collapse Test verdict, a side-effect threshold, or a manual call. Automatic triggers act faster; manual ones fold in judgment.

When it helps, and when it misleads¶

Its strength is that it stops the archetype's signature harm — forcing more flow through a dead lever — and, by anchoring the ceiling to side-effects, it also stops the quieter harm of a lever that still "works a little" but poisons the system as you lean on it. It converts escalation pressure into a defended pause in which the stock can recover.

Its failure mode is mis-timing and mis-anchoring: freeze too early and you cap a lever that was merely slow, forgoing real benefit; set the ceiling by the benefit you wish you had rather than the harm you are taking, and you climb straight into the toxic zone. The classic misuse is a cap imposed to look disciplined while the damaged stock goes unrepaired, so the "pause" silently becomes permanent paralysis. The discipline is to pair every freeze with an active, owned repair plan and an exit criterion, so the cap is always a bridge and never the whole strategy. The clinical version of the trap it guards is the closing therapeutic window — as the substrate degrades, the gap between too-little-to-help and enough-to-harm narrows until more dose means only more harm.^[1]

How it implements the components¶

Lever Freeze-or-Cap fills the governing-constraint side of the archetype — the components that stop action rather than measure or mend it:

lever_escalation_stop_rule — it is the enforced stop or ceiling that ends escalation of a non-transmitting lever.
side_effect_burden_meter — it reads the accruing collateral cost and places the cap where that burden overtakes the shrinking benefit.

It does not decide the lever is dead in the first place (normal_flow_lever, responsive_window_model — Gain-Collapse Test) nor repair the stock it is protecting (stock_repair_plan — [Substrate Repair Protocol]); it only interdicts.

Instantiates: Stock-First Control Restoration — this rule is the halt that keeps a diagnosed-dead lever from being escalated.
Consumes: Gain-Collapse Test supplies the collapse verdict that legitimately triggers the freeze.
Sibling mechanisms: Gain-Collapse Test · Deleverage-Before-Stimulus Rule · Fallback Control Mode · Low-Amplitude Reactivation Probe · Hysteresis-Aware Exit Criterion · Minimum Stock-Floor Alert · Replenishment-or-Refill Protocol · Substrate Repair Protocol · Staged Lever Ramp · Stock-Flow Diagnostic Map · Trust-Repair-Before-Persuasion Rule

Notes¶

Freeze-or-Cap governs the active, escalating lever right now; the Deleverage-Before-Stimulus Rule governs not restarting a lever until the stock is rebuilt. They share the "don't pull" instinct but sit at opposite ends of the timeline — halting an over-pushed lever versus withholding a not-yet-earned one — and a full restoration usually needs both.

References¶

[1] Therapeutic window — the dose range above the minimum effective level and below the toxic level. As tolerance downregulates the receptor stock, the window narrows and finally closes, so more dose buys harm rather than help; the fix is to restore the receptor, not to chase the dose. ↩