Answers are grounded in this paper and cite their sources.
Investigates structural reconstruction of information within physical systems, highlighting implications for causation and objectivity.
Papers 1–18 typed the world-side structures of a consistency-constrained information history and, in the fourth triptych, causation, lawhood, and explanation. This paper inverts the program: given only the finite, local, record-mediated access available to an embedded observer, how much of that structure can be reconstructed, under what non-uniqueness, and at what cost? We separate two type-distinct senses of recovery — carrier/information recovery and structural reconstruction — and pose structural reconstruction as an inverse problem governed by a protocol-level forward map (a transcript distribution over whole experiments), a representation-invariant fit rule, and a declared recovery specification. The central negative principle, the No-Global-Decoder Constraint, adds one term to the series' “no global X” sequence: no physically privileged subsystem possesses complete-history access together with a uniquely authoritative decoding map. The central positive result is an elementary but explicit Protocol-Kernel Recovery-Limit Theorem; its value is not mathematical difficulty but the placement of causation, lawhood, explanation, objectivity, and horizon recovery inside one inverse-problem formalism. We distinguish three notions of uniqueness (global, observation-relative, finite-record); show that protocol refinement is monotone while finite-sample compatibility is not; define target-relative correctness through an external adequacy relation to the actual structure; profile recovered structure along the representational, evidential, and corrective objectivity axes of Paper 15; and exhibit the framework on two worked models, a linear-Gaussian structural-causal triple separated in stages by interventions and a conformal-versus-metric spacetime example. We emphasize that, with finite or noisy records, the output of reconstruction is in general not a single equivalence class but a specification-relative compatible set or indistinguishability region, and that no reconstruction certifies its own completeness, uniqueness, or correctness.
No takes yet. Share an insight, caveat, or question.
Tomoyuki Uchida (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: