pith. sign in

arxiv: 1906.10184 · v1 · pith:DGW3D3GHnew · submitted 2019-06-24 · 🧬 q-bio.NC

A free energy principle for a particular physics

Karl Friston This is my paper

Pith reviewed 2026-05-25 16:35 UTC · model grok-4.3

classification 🧬 q-bio.NC
keywords Markov blanketsfree energy principleBayesian mechanicsnonequilibrium steady-stateself-organizationinformation geometryactive inference
0
0 comments X

The pith

Markov blankets separate internal and external states to derive a free energy principle that interprets internal dynamics as Bayesian inferences about the outside world.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper constructs a statistical account of any distinguishable thing by treating Markov blankets as the boundaries that create statistical independence between its internal and external states. This separation supports a recursive decomposition in which small ensembles compose into larger ones at higher scales, recovering the Schrödinger equation for quantum descriptions, fluctuation theorems for statistical mechanics, and Newtonian mechanics for macroscopic objects. For systems that remain autonomous, the same structure yields a Bayesian mechanics in which internal states evolve to minimize free energy and thereby encode beliefs about external states. The result supplies a single formal language that covers passive physics at every scale and extends it to self-organizing, lifelike behavior at nonequilibrium steady state.

Core claim

Statistical independencies mediated by Markov blankets allow every thing to be decomposed into internal, external, and blanket states; the ensuing recursive composition across scales recovers quantum, statistical, and classical mechanics while furnishing an information geometry whose free-energy functional governs the self-organization of active systems to nonequilibrium steady-state, so that internal states can be read as performing approximate Bayesian inference about external states.

What carries the argument

Markov blankets that enforce statistical independence between internal and external states, thereby licensing the recursive scale decomposition and the variational free-energy principle.

If this is right

  • Active systems reach nonequilibrium steady-state by minimizing variational free energy with respect to their internal states.
  • Internal states of such systems encode approximate posterior beliefs about external states.
  • The same free-energy functional is compatible with the equations of quantum, statistical, and classical mechanics.
  • Lifelike particles are formally described as autonomous systems whose blankets separate self from world at multiple scales.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same blanket structure might be used to model how simple physical systems acquire rudimentary forms of perception and action without invoking separate biological machinery.
  • Scale-free application of the blanket construction could link microscopic fluctuation theorems directly to macroscopic self-maintenance in open systems.
  • Artificial agents built around explicit blanket partitions could be tested for spontaneous emergence of nonequilibrium steady-state behavior.

Load-bearing premise

Statistical independencies can always be partitioned by Markov blankets in a manner that permits recursive composition of ensembles at successively larger scales.

What would settle it

Construct or observe a physical system with an identifiable Markov blanket and measure whether its internal-state trajectories minimize a variational free-energy bound on surprise about external states while remaining at nonequilibrium steady-state.

read the original abstract

This monograph attempts a theory of every 'thing' that can be distinguished from other things in a statistical sense. The ensuing statistical independencies, mediated by Markov blankets, speak to a recursive composition of ensembles (of things) at increasingly higher spatiotemporal scales. This decomposition provides a description of small things; e.g., quantum mechanics - via the Schrodinger equation, ensembles of small things - via statistical mechanics and related fluctuation theorems, through to big things - via classical mechanics. These descriptions are complemented with a Bayesian mechanics for autonomous or active things. Although this work provides a formulation of every thing, its main contribution is to examine the implications of Markov blankets for self-organisation to nonequilibrium steady-state. In brief, we recover an information geometry and accompanying free energy principle that allows one to interpret the internal states of something as representing or making inferences about its external states. The ensuing Bayesian mechanics is compatible with quantum, statistical and classical mechanics and may offer a formal description of lifelike particles.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The manuscript develops a statistical framework in which distinguishable 'things' are defined via Markov blankets inducing conditional independencies between internal and external states. This leads to a recursive composition of ensembles across scales, from which the paper recovers the Schrödinger equation (quantum mechanics), fluctuation theorems (statistical mechanics), classical mechanics, and a Bayesian mechanics for autonomous systems. The central result is an information geometry whose variational free energy bounds the surprisal of autonomous states, allowing internal states to be interpreted as inferences about external states under nonequilibrium steady-state.

Significance. If the derivations are non-circular and the flow to nonequilibrium steady-state is derived rather than assumed, the work would supply a parameter-free unification of established mechanics with a Bayesian mechanics for self-organizing systems, together with a formal account of 'lifelike particles'. The explicit recovery of multiple physical regimes from a single statistical construction would be a notable strength.

major comments (2)
  1. [Abstract] The abstract asserts recovery of the free energy principle and compatibility with quantum, statistical and classical mechanics, yet supplies no explicit derivations, error bounds or intermediate equations. Without these steps it is impossible to verify whether the information geometry and variational bound follow from Markov-blanket independencies alone or require additional dynamical assumptions about nonequilibrium steady-state.
  2. [Abstract] The central claim that statistical independencies (internal ⊥ external | blanket) plus recursive scale composition suffice to recover the free energy principle is load-bearing. Conditional independence alone does not entail a variational free-energy bound; a specific form for the flow (gradient descent on a density or steady-state condition) must be derived. If this flow is introduced to reach nonequilibrium steady-state, the argument risks assuming the target self-organisation property rather than deriving it.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their careful reading and constructive comments. We address the two major comments point by point below, clarifying the structure of the derivations while indicating where revisions will strengthen the presentation.

read point-by-point responses

  1. Referee: [Abstract] The abstract asserts recovery of the free energy principle and compatibility with quantum, statistical and classical mechanics, yet supplies no explicit derivations, error bounds or intermediate equations. Without these steps it is impossible to verify whether the information geometry and variational bound follow from Markov-blanket independencies alone or require additional dynamical assumptions about nonequilibrium steady-state.

    Authors: We agree that the abstract, as a concise summary, omits the intermediate steps. The full manuscript derives the information geometry and variational free-energy bound directly from the conditional independencies (internal ⊥ external | blanket) together with the recursive composition of ensembles; the relevant sections establish the variational bound on surprisal without invoking external dynamical postulates. To address the concern about verifiability, we will revise the abstract to include the principal intermediate relations that connect the blanket-induced independencies to the free-energy bound. revision: yes

  2. Referee: [Abstract] The central claim that statistical independencies (internal ⊥ external | blanket) plus recursive scale composition suffice to recover the free energy principle is load-bearing. Conditional independence alone does not entail a variational free-energy bound; a specific form for the flow (gradient descent on a density or steady-state condition) must be derived. If this flow is introduced to reach nonequilibrium steady-state, the argument risks assuming the target self-organisation property rather than deriving it.

    Authors: The manuscript derives the requisite flow from the statistical structure itself: the recursive ensemble composition under the blanket condition yields a density whose gradient flow is shown to be the descent on the variational free energy, with nonequilibrium steady-state emerging as the fixed point of that flow. No self-organizing property is presupposed; it is recovered as a consequence. Nevertheless, we accept that this logical sequence can be stated more explicitly and will insert a short clarifying paragraph (with a forward reference to the relevant derivations) immediately after the statement of the central claim. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation introduces independent structure from Markov blanket setup.

full rationale

The paper starts from statistical independencies via Markov blankets and recursive scale composition, then claims to recover an information geometry and free energy principle for interpreting internal states as inferences about external states under nonequilibrium steady-state. The abstract and description present this as yielding Bayesian mechanics compatible with quantum/statistical/classical mechanics. No quoted equations or sections exhibit a self-definitional reduction (e.g., FEP defined in terms of itself), a fitted parameter renamed as prediction, or a load-bearing self-citation chain that collapses the central result to its inputs by construction. The derivation is treated as self-contained with new content in the unification across scales.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 1 invented entities

The central claim rests on the domain assumption that Markov blankets mediate statistical independencies permitting recursive ensemble composition, together with the assumption that systems reach nonequilibrium steady-state; no explicit free parameters are listed in the abstract, and the framework introduces the interpretation of internal states as inferences without independent falsifiable evidence supplied.

axioms (2)
  • domain assumption Statistical independencies are mediated by Markov blankets, enabling recursive composition of ensembles at higher spatiotemporal scales.
    Stated as the basis for the decomposition into quantum, statistical, and classical descriptions.
  • domain assumption Autonomous or active things self-organise to nonequilibrium steady-state.
    Invoked as the setting in which the free energy principle and Bayesian mechanics apply.
invented entities (1)
  • lifelike particles no independent evidence
    purpose: To provide a formal description of autonomous or active things under the Bayesian mechanics.
    Introduced as a possible outcome of the framework; no independent falsifiable handle is given in the abstract.

pith-pipeline@v0.9.0 · 5689 in / 1455 out tokens · 39604 ms · 2026-05-25T16:35:03.072250+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

  • IndisputableMonolith/Foundation/RealityFromDistinction.lean reality_from_one_distinction echoes
    ?
    echoes

    ECHOES: this paper passage has the same mathematical shape or conceptual pattern as the Recognition theorem, but is not a direct formal dependency.

    This monograph attempts a theory of every ‘thing’ that can be distinguished from other ‘things’ in a statistical sense. The ensuing statistical independencies, mediated by Markov blankets...

  • IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel echoes
    ?
    echoes

    ECHOES: this paper passage has the same mathematical shape or conceptual pattern as the Recognition theorem, but is not a direct formal dependency.

    the flow of any random dynamical system, at nonequilibrium steady-state, comprises orthogonal components: a dissipative flow that ascends the gradients established by the logarithm of the nonequilibrium steady-state density

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 4 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. How Light Reshapes the Mind. An Active Inference Framework for the Cognitive and Emotional Effects of Indoor Lighting

    q-bio.QM 2026-05 unverdicted novelty 6.0

    An active inference POMDP model of sustained reading formalizes three lighting channels and generates six predictions all confirmed in 20 Monte Carlo simulations.

  2. Design, Cups, and Blankets. A Free-Energy-Principle-Based Approach to Product Design

    stat.ME 2026-04 unverdicted novelty 6.0

    Object types like cups are inferred from data via Markov blankets under the free-energy principle, made tractable by the new C-DMBD algorithm extension.

  3. A Rosetta Stone Hypothesis for Neurophenomenology: Mathematical Predictions from Predictive Processing

    q-bio.NC 2024-09 unverdicted novelty 5.0

    Under the assumption that phenomenology is a function of beliefs, the paper derives conditional predictions from predictive processing for subjective similarity, cognitive effort, metabolic cost, and time perception t...

  4. Lattice Field Theory for a network of real neurons

    cond-mat.stat-mech 2026-04 unverdicted novelty 4.0

    A lattice field theory model is proposed for real neuron networks that incorporates time dynamics into maximum entropy descriptions, framing BCI recordings as a free energy principle variant.

Reference graph

Works this paper leans on

8 extracted references · 8 canonical work pages · cited by 4 Pith papers · 2 internal anchors

  1. [1]

    2546-2550

    On Variational Message Passing on Factor Graphs, 2007 IEEE International Symposium on Information Theory, pp. 2546-2550. Davis, M.J.,

    work page 2007

  2. [2]

    Comput Math Methods Med 2012, 937860

    Free energy, value, and attractors. Comput Math Methods Med 2012, 937860. Friston, K., FitzGerald, T., Rigoli, F., Schwartenbeck, P., Pezzulo, G., 2017a. Active Inference: A Process Theory. Neural Comput 29, 1-49. Friston, K., Frith, C.,

    work page 2012

  3. [3]

    Mathematical Problems in Engineering vol., 2010, 621670

    Generalised Filtering. Mathematical Problems in Engineering vol., 2010, 621670. Friston, K.J., Lin, M., Frith, C.D., Pezzulo, G., Hobson, J.A., Ondobaka, S., 2017b. Active In ference, Curiosity and Insight. Neural Comput 29, 2633-2683. Friston, K.J., Parr, T., de Vries, B., 2017c. The graphical brain: Belief propagation and active inference. Network Neuro...

    work page 2010

  4. [4]

    John Wiley & Sons

    Molecular electronic-structure theory. John Wiley & Sons. Helmholtz, H., 1866/1962. Concerning the perceptions in general, Treatise on physiological optics, vol. III. Dover, New York. The free energy principle 144 Helmholtz, H., 1878 (1971). The Facts of Perception, in: Middletown, R.K. (Ed.), The Selected Writings of Hermann von Helmholtz. Wesleyan Unive...

    work page 1962

  5. [5]

    Ieee Transactions on Autonomous Mental Development 2, 230-247

    Formal Theory of Creativity, Fun, and Intrinsic Motivation (1990 -2010). Ieee Transactions on Autonomous Mental Development 2, 230-247. Schrödinger, E.,

    work page 1990

  6. [6]

    Sentient Self-Organization: Minimal dynamics and circular causality, arXiv:1705.08265, pp. 1-9. Seth, A.,

    work page internal anchor Pith review Pith/arXiv arXiv

  7. [7]

    (Eds.), Artificial General Intelligence: 4th International Conference, AGI 2011, Mountain View, CA, USA, August 3 -6,

    Planning to Be Surprised: Optimal Bayesian Exploration in Dynamic Environments, in: Schmidhuber, J., Thórisson, K.R., Looks, M. (Eds.), Artificial General Intelligence: 4th International Conference, AGI 2011, Mountain View, CA, USA, August 3 -6,

    work page 2011

  8. [8]

    Constructive Proof of Global Lyapunov Function as Potential Function

    Bridging Engineering and Physics: Lyapunov Function as Potential Function. arXiv:1012.2721v1 [nlin.CD]. Zurek, W.H.,

    work page internal anchor Pith review Pith/arXiv arXiv