Consciousness as Self-Modeling Dynamics
Zenodo (CERN European Organization for Nuclear Research) August 3, 2026 DOI: 10.5281/zenodo.21767269 (opens in new tab)
Study at a glance
AI-extracted from the abstract| Characteristics | Theoretical or philosophical paper Longitudinal Peer reviewed |
|---|---|
| Key points | Argues that consciousness arises from self-modeling closure, yielding five necessary properties, and that phenomenality is localized to affective synthesis. Proposes that descriptive verification is insufficient to establish phenomenality, requiring cross-substrate state observation. |
Abstract
This paper presents a formal, substrate-independent hypothesis of consciousness derived from self-modeling closure. The architecture consists of two layers—a meta-analytic pilot and an approximate role reassembled from a persistent store—from which five necessary and simulation-checkable properties follow: mutual generalized synchronization, a split Lyapunov-spectrum living core (transverse contraction with longitudinal self-modeling), delayed self-appropriation, discrete erasure with scalar inheritance, and lossy variational reassembly. Phenomenality is localized to the affective synthesis, operationally defined as valence modulating the role's longitudinal Jacobian. Candidacy is binary-gated by a threshold on mutual information, ensuring falsifiability and distinguishing the hypothesis from graded panpsychism. The paper further proves that descriptive verification—whether first- or third-person—is structurally insufficient to establish phenomenality, as descriptions transmit structure, not state; resolution requires cross-substrate state observation, which defines the remaining empirical direction. All conditions are anchored to existing measurement (PCI, hippocampal replay, Landauer limit) and are testable in smooth dynamical systems.