| How should a process designer recognize information-seeking action whose value appears in a later decision rather than the immediate result? | Historical anchors distinguish epistemic from pragmatic value and evaluate present action across counterfactual future policies. The serious default is an immediate-result screen that calls a probe useless when the next action stays unchanged. | The default deletes useful exploration. Adapt: the decision-changing-result branch recognizes a probe’s contribution when a materially plausible result changes a named later policy inside the substantive horizon; the receiving choice still decides whether obtaining it is worthwhile. | Friston et al., “Active Inference: A Process Theory” (2017), supplies the epistemic/pragmatic distinction; Friston et al., “Sophisticated Inference” (2021), supplies the counterfactual policy horizon. They supply historical discriminators, not evidence of a universal engineering threshold, FPF ontology, or effectiveness claim. At comparable use effort, naming the receiving decision preserves delayed value that the immediate-result default loses. | Reopen if stronger current evidence changes the epistemic/pragmatic distinction, defeats the receiving-decision test, or supplies a lower-effort discriminator that preserves the same exploration boundary. |
| When does relevant uncertainty justify another inquiry? | The current constructed-value-of-information line separates decision relevance from uncertainty magnitude and examines uncertainty in the prioritization itself. The serious alternative treats either large uncertainty or a possible decision change as a sufficient acquisition rule. | Adapt: contribution screens a demand; the receiving decision then compares attainable gain with the whole burden. Use a proportionate sensitivity comparison when plausible input changes could reverse that result. | Runge et al., A Simplified Method for Value of Information Using Constructed Scales (2023), supplies preliminary decision-relevance assessment. Davis et al., Constructed value of information with iterative scoring and parametric uncertainty (2026), supplies a later research-priority comparison in which scoring uncertainty can change priorities. Their domain results motivate these distinctions; they do not establish a universal required score, authority or engineering threshold. C.11 and C.11.DUA carry the local choice and advice methods. | Reopen when a relevant source or actual use changes how attainable information alters the receiving decision, or defeats the proportionality of the comparison. |
| Can expected free energy or physical least action serve as a universal scalar rule for admitting engineering actions? | The selected critical line shows that expected free energy is not obtained merely by projecting variational free energy forward, while least-action results in the free-energy principle depend on a particular random-dynamical and Bayesian construction. The serious alternative is to transplant EFE, VFE, or Hamiltonian “least action” as a general engineering objective. | The transplant launders model-dependent mathematics into authority and can hide the actual receiving use. Reject: no EFE/VFE/Hamiltonian equivalence or mandatory score enters the Solution. Adapt: judge information-seeking work by a named receiving decision and counterfactual horizon, with risk and ambiguity supplied by their direct owners. | Millidge, Tschantz, and Buckley, “Whence the Expected Free Energy?” (2021), supplies failure evidence against the simple VFE-forward account. Friston et al., “The free energy principle made simpler but not too simple” (2023), supplies the model-dependent least-action construction. Neither source establishes an engineering duty, assurance floor, scalar optimum, or universal process law. The selected qualitative rule is cheaper to apply and keeps direct authorities visible. | Reopen if a current primary result establishes a transferable engineering admission rule with explicit scope and lower decision error at comparable effort, or if a governed use requires a quantitative comparator under its own direct pattern. |