What does this research mean for the field?

The persistence of bounded, far-from-equilibrium systems across all scales is governed by a single dimensionless thermodynamic persistence ratio (ℛ), where ℛ ≥ 1 is required for long-term survival and model error imposes an exponential lifetime penalty. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.ESTABLISHES_NEW_DIRECTION.

What question did this study set out to answer?

The aim is to establish a thermodynamic theory for information-persisting systems that explains how their identity is maintained through information processing.

June 10, 2026Open Access

Information-Persisting Systems: A Thermodynamic Theory of Persistence as a Non-Equilibrium Accounting Identity

Key Points

The aim is to establish a thermodynamic theory for information-persisting systems that explains how their identity is maintained through information processing.
Development of a dimensionless persistence ratio defined by the Fractal Persistence Equation.
Integration of non-equilibrium statistical mechanics, stochastic thermodynamics, and Landauer's principle.
Formulation of five falsifiable predictions related to persistence across different scales.
ℛ ≥ 1 is essential for long-term persistence (Theorem 5.1).
Model error leads to exponential penalties in lifetime (Theorem 5.2).
Fractal composition of persistence is observed across organizational levels (Theorem 5.3).

Abstract

The foundation paper in the Information-Persisting Systems (IPS) programme. This manuscript develops a physical theory of information-persisting systems (IPS): bounded, far-from-equilibrium patterns whose identity is preserved by active information processing. It combines non-equilibrium statistical mechanics, stochastic thermodynamics, Landauer's principle, and the variational free-energy principle into a single dimensionless persistence ratio ℛ (the Fractal Persistence Equation). Main results: - ℛ ≥ 1 is a necessary condition for long-term persistence (Theorem 5.1) - Model error (KL divergence between internal model and environment) imposes an exponential lifetime penalty (Theorem 5.2) - Persistence at one organisational level composes fractally with persistence at adjacent levels (Theorem 5.3) The theory is substrate-neutral and applies across scales — from nuclear binding and autocatalytic chemistry to cells, ecosystems, firms, and polities — as a single thermodynamic accounting identity. Five falsifiable predictions (P1–P5) are stated, including a substrate-isolation forgetting law testable in fractal mixture-of-experts networks. This is a physics paper, not a metaphysics of consciousness. Companion work on agent systems and implementation lives at https://aion.anomalia.io/

Read Full Paperexternally

Mark Helpful

Bookmark

Relay

View Full Paper