# Node C13: Free Energy / Active Inference

slug: oip-node-c13-free-energy-active-inference · https://miscsubjects.com/a/oip-node-c13-free-energy-active-inference · tags: philosophy, oip, convergence-catalogue, node, systems-theory · updated 2026-07-17T02:36:20.994Z

# Node C13: Free Energy / Active Inference

C13 — Free Energy / Active Inference
{
  "id": "C13",
  "claim": "Self-organizing systems (from cells to brains) act to minimize variational free energy — a quantity bounding the surprise of sensory states — effectively making Bayesian inferences about their environment through action and perception.",
  "domain": ["neuroscience", "machine_learning", "theoretical biology", "psychiatry"],
  "pattern": ["free_energy_principle", "active_inference", "variational_Bayes", "predictive_coding"],
  "mechanism": "Variational free energy F = E_q[ln q(s) - ln p(o,s)] upper-bounds surprise (-ln p(o)). Systems minimize F by either (a) updating internal models (perception/inference) or (b) changing the world to match predictions (action). Under Laplace/Gaussian assumptions, this reduces to predictive coding: error = prediction - observation, minimized hierarchically.",
  "scale": "cellular → organism",
  "claim_tier": "T2",
  "sources": [
    "Helmholtz, H. von (1867). Handbuch der Physiologischen Optik. Voss. [Perception as unconscious inference.]",
    "Friston, K. (2005). 'A Theory of Cortical Responses.' Phil. Trans. R. Soc. B, 360, 815-836.",
    "Friston, K. (2010). 'The Free-Energy Principle: A Unified Brain Theory?' Nature Reviews Neuroscience, 11, 127-138.",
    "Rao, R.P.N. & Ballard, D.H. (1999). 'Predictive Coding in the Visual Cortex.' Nature Neurosci., 2(1), 79-87."
  ],
  "dual": "None intrinsic — the dual would be a system that maximizes surprise (actively seeks chaos), which is pathological.",
  "falsifier": "An adaptive agent that provably does not reduce prediction error (or its bound) over time, yet survives and adapts comparably to predictive agents; or evidence that the free energy formalism cannot be operationalized with independent parameters.",
  "rival_frame": "FEP is unfalsifiable. Because any behavior can be described as minimizing some free energy functional post hoc, the principle predicts nothing independently. It is a mathematical tautology dressed as a theory — the Ptolemaic epicycles of neuroscience. Predictive coding is real and useful; the FEP as grand unification is not.",
  "independence_check": "MODERATE. Helmholtz (19th c. physiology) originated perception-as-inference from optics and eye movement studies. Rao & Ballard (1999, computational neuroscience) developed predictive coding independently from hierarchical Bayesian models. Friston (2005+, UCL) synthesized these into the Free Energy Principle. There is clear lineage from Helmholtz to Friston; the independence is in the computational instantiation, not the core insight.",
  "pattern_type": "energetic",
  "maps_to_axiom": ["A3", "A2"]
}

---

## Corpus map
- Same node, other planes: [Encyclopedia C13](/a/convergence-encyclopedia-c13) · [Inventory invariant](/a/oip-invariant-11-311-prediction-free-energy-active-inference)
- Edges touching C13: [disconfirming edge 2](/a/oip-disconfirming-edge-2)
- Catalogue hub: [Public Article](/a/oip-convergence-public-article) · [Schema](/a/oip-convergence-schema)