Justification and interdependence of epistemic resources

Investigate why, for the manifest variables considered in nomic toy theory, the epistemic horizon for repeatable measurements is determined entirely by the available ready states, and determine how manifest variables and ready states can be justified and whether those justifications make the two learning resources interdependent.

Background

The paper treats a subject's manifest variable and the ready-state variability of its input system as independent modelling choices. Its classification shows that, for repeatable measurements, trivial, Poisson, and complete ready states respectively produce total, knowledge-balance, and absent epistemic horizons for the non-trivial manifest variables considered.

The authors explicitly identify as unresolved both the explanation for this dependence on ready states and the broader question of whether physical motivations—such as thermodynamic low-entropy conditions, evolutionary stability, or the reliability and copyability of records—would impose dependencies between the two kinds of epistemic resource.

References

One open question is why, for the manifest variables considered here, the resulting epistemic horizon for repeatable measurements is determined entirely by the available ready states. More generally, manifest variables and ready states are treated here as independent assumptions, and we derive the epistemic horizons that follow from different choices. It remains to be explored how these choices can be justified, and whether such justifications render the learning resources interdependent.

The View from Within: What Can Embedded Observers (Not) Learn?  (2608.25800 - Gonda et al., 26 Aug 2026) in Section 'Discussion and Outlook', paragraph 'Justifying epistemic resources'

We do not know whether other variables, one whose kernel is not a symplectic subspace, can also be retrodicted for Poisson ready states. The case of complete ready states is covered by \cref{thm:repeatable_complete_MV_complete_RS}. However, answering this question is irrelevant as far as the epistemic horizon is concerned, because identity maps can be retrodicted according to \cref{thm:complete_MV_retrodictive}, and so every variable is retrodictively fixed by \cref{prop:fixed_from_measured}.

The View from Within: What Can Embedded Observers (Not) Learn?  (2608.25800 - Gonda et al., 26 Aug 2026) in Theorem 'EH for retrodiction; complete manifest variable', footnote

If we do not use the previous justification for the constraint on initial ontic states, the question then arises how should one think of ready states here. We leave this question to a large part unanswered.

The View from Within: What Can Embedded Observers (Not) Learn?  (2608.25800 - Gonda et al., 26 Aug 2026) in Appendix, Remark 'New interpretation of ready states'