nucleation

CLAIM-339: A judgement–observation loop cut off from action loses one route for evidence it did not select, and the free-energy principle's reason for minimizing surprise, that a persisting system must keep its states in bounds, needs action on the world, though a loop can confirm itself with action available too

4 min read

Properties11
StatusActive
Version3
Rolecounter
Warrantinterpretive
Tagsprobabilistic-modeling, cognition, psychopathology-and-treatment
Linedistributed-agency, pragmatic-transport
GroundsLIT-1181 — Action and behavior: a free-energy formulation, THEORY-285 — In an illustrative active-inference simulation of post-traumatic context inference, run once per phenotype, a prior that one is in a harmful context, made overly precise by a run of harmful trials under the "vulnerable" learning and forgetting weights, makes exploratory sampling look uninformative, so the agent attenuates the periphery, never receives the evidence that would correct the prior, and the belief sustains itself; lowering precision makes the attenuation policy less vigorous without correcting the prior, and the policy returns when precision is restored, LIT-1199 — Brain–computer interfaces and communication in paralysis: Extinction of goal directed thinking in completely paralysed patients?, CASE-107 — Locked-in syndrome: consciousness preserved with anarthria and quadriplegia; in the classic form, vertical eye movement and blinking remain and carry communication, and in the completely locked-in state no motor channel remains, THEORY-1166 — Di Paolo argues that autopoiesis as originally defined entails conservation and not active homeostasis, so it gives a system only an all-or-nothing norm, to go on or to die, which an observer and not the system can use; that a graded norm the system itself has needs adaptivity, a capacity to detect tendencies toward the boundary of its viability and to act so as to reverse them; and that sense-making and agency need adaptivity, THEORY-324 — A system adapts by trial and error to an environment whose effect on its essential variables it does not know only if a second feedback, through those variables, changes the parameters of its primary feedback; and if that feedback changes them whenever, and for as long as, an essential variable is outside its limits, every equilibrium the whole can reach is an adapted one
Objects toCLAIM-341 — The loop that keeps a lucid agent's model in touch with what it models runs between judgement and observation: each judgement reshapes the observer whose later observations it then judges, and action on the world is not part of that loop, so active inference's action–perception cycle misplaces where the agent is active
Objected byCLAIM-342 — What keeps a judgement–observation loop from confirming itself is evidence whose arrival or scoring does not pass through the interpretation it tests; action is one source of such evidence, not a necessary one, and a loop that selects its own evidence can lock in with action available
UsesTERM-020 — protective lock-in
Cited as “complements” byCLAIM-460 — In the completely locked-in state, evidence of lucid judgement has to cross a channel the patient drives, which is action in active inference's sense, so such cases can show the judgement–observation loop running down but never running with no action at all, and a null result cannot tell lost judgement from a lost channel

Graph View

The claim

The free-energy principle explains why surprise matters. A system that persists must keep its sensory states within bounds, and Friston et al. (LIT-1181) say the active sampling that does so is “mandated by ergodic arguments based on the very existence of adaptive agents”. The enactive account says the same of norms. Self-production gives a system only the norm “to go on or to die”, and that distinction is available to an observer, not to the system. A graded norm the system itself has needs adaptivity, “to detect tendencies toward the boundary of its viability and to act so as to reverse them” (THEORY-1166). Ashby’s ultrastability needs a channel from the essential variables, which are coupled to the world (THEORY-324).

Take action out of a loop of judgement and observation, and the loop loses the route by which evidence arrives that it did not ask for. That leaves it more exposed to the failure the record holds as protective lock-in (TERM-020): a regime kept stable by behaviour that blocks its own disconfirmation.

Kübler and Birbaumer (LIT-1199) give the empirical form. In the completely locked-in state no patient of theirs regained yes/no communication, and they proposed that “the complete lack of motor control and feedback … might be responsible for the cessation of voluntary cognitive activity”. The judgement loop, on this view, runs down when action is gone (CASE-107).

Conceded

As first filed, this counter said that without action the loop has a self-confirming minimum. That is true, but it is not specific to the absence of action. In the active-inference model of THEORY-285 the agent can act, and its own sampling policy “withholds the only observations that could lower that precision” (NOTE-766). A loop that selects its own evidence can confirm itself with or without action (CLAIM-342).

The reading of Friston et al. (NOTE-960) concedes more. In the free-energy formulation, action is not a source of evidence the loop did not select. It is “sampling selectively what one expects to experience”, and unexpected perturbations “were explained away by action and not perceived”. The paper sees the self-confirmation and embraces it: “this weakness becomes the strength”. Its anchor is not action but existence: “agents who do not have the right priors cannot exist”. And action is needed there by stipulation, since “the only way that the agent can change hidden states is through action”.

What survives is narrower again. The enactive account (THEORY-1166) and Ashby’s ultrastability (THEORY-324) make action, or a channel through the essential variables, what grounds a system’s norms. The free-energy formulation does not supply this counter’s premise. On its own terms, action is one more way a loop selects its own evidence.

What it does not say

  • It does not show that the world can reach the loop only through action. CLAIM-342 names other sources: a source the agent does not select, and scoring fixed independently of the interpretation.
  • It runs together two things that paralysis takes apart: evidence the loop did not select, which a paralysed person still receives through the senses, and contingency between intention and outcome, which they lose. Kübler and Birbaumer’s hypothesis concerns the second. The anchor of CLAIM-342 needs only the first (QUESTION-121). So even if the hypothesis holds, what extinguishes goal-directed thinking would not be the loss of an anchor.
  • The extinction hypothesis is a hypothesis, stated as what “might be responsible”, from seven patients already in CLIS when training began and an analogy with curarized rats (NOTE-948). Its authors say passive sensory processing is intact in CLIS, and they predict that a contingency learned before CLIS should carry over. Chaudhary et al.’s patient, who communicated, had done exactly that (NOTE-951), so he fits the hypothesis rather than refuting it.

Note of 2026-10-29: filed with its concessions

CLAIM-342, Active since the ninth status review, objects to this counter. Its concession is in this counter’s text as filed on 2026-10-16: the section “Conceded” narrows it from a self-confirming minimum without action to the loss of one route for unselected evidence.

Backlinks