Evolution of Predictive Processing

The Evolutionary and Physical Foundations of Predictive Processing: A Research Program

The Problem

Predictive processing (PP) — the framework proposing that brains are prediction engines — has become one of the most influential paradigms in cognitive science. Yet it faces persistent criticism: unfalsifiability, the dark room problem, and above all a shallow evolutionary grounding. The standard argument — that predictive architectures were selected because they conferred survival advantage — attributes creative power to natural selection, explaining the persistence of a trait but not its origin. This is precisely the explanatory error identified in Espesset, D. (2021), “Le darwinisme tient-il debout ? — Revue critique,” and developed systematically in Espesset, D. (2023), “Changement de paradigme de l’évolution biologique: du darwinisme à une nouvelle synthèse conceptuelle élargie,” Eikasía Revista de Filosofía.

The Thesis

Predictive processing is not an adaptation selected for by cumulative natural selection. It is a structural inevitability — a necessary consequence of physical constraints on any information-managing system operating under temporal latency and environmental complexity. Selection merely conserves what physics necessitates.

Kriger’s formal results provide the mathematical framework. The Dual-Pressure Convergence Theorem, proven in Kriger, B. (2026), “The evolutionary inevitability of predictive processing: A physical constraint argument” (https://doi.org/10.5281/zenodo.18444910) and extended in Kriger, B. (2026), “Latency and Compressibility: Two Independent Routes to Predictive Architecture in Adaptive Systems” (https://doi.org/10.5281/zenodo.18752102), demonstrates that temporal latency and environmental complexity each independently force adaptive systems toward predictive architecture. Adaptive Observer Theory, in Kriger, B. (2026), “Adaptive Observer Theory: Navigation of Incompleteness and Incompatibility” (https://doi.org/10.5281/zenodo.18812040), proves that viable agents must maintain multiple models and switch between them.

Espesset’s evolutionary biology provides the biological grounding. His critique of the neo-Darwinian paradigm — across Espesset, D. (2018), “L’évolution biologique: vers une nouvelle synthèse conceptuelle étendue pour le 21e siècle”; Espesset, D. (2025), “Schizophrénie darwiniste — Partie 1”; and Espesset, D. (2025), “L’organisme vivant et la construction de l’individu biologique” — establishes that natural selection operates as a conserving and eliminating force, not a creative one. His framework of internal organizational principles — self-organization, information management, autopoietic dynamics, and niche construction (Espesset, D. (2021), “Les maîtres de la construction de niche”) — provides the biological substance for a non-selectionist account of cognitive architecture.

Complexity and the Informational Hierarchy

The question of whether an organism requires predictive architecture cannot be answered without a quantitative account of biological complexity. Espesset’s program on this question — from Espesset, D. (2020), “The concept of complexity in evolutive biology: A synthetic review,” through its formal operationalization in Kriger, B. & Espesset, D. (2026), “The GEKS Index: A Composite Measure of Biological Complexity Across Informational, Structural, Functional, and Evolutionary Dimensions” (https://doi.org/10.5281/zenodo.18775491) — provides the quantitative bridge.

Espesset further proposes that biological organization follows a hierarchy of informational structures of increasing dimensionality: from zero-dimensional units (bits), through one-dimensional sequences (DNA), through two- and three-dimensional molecular associations, to n-dimensional informational matrices at each level of biological integration — organelles, cells, tissues, organs, organisms, populations, ecosystems, the biosphere. The self-organization of matter is fundamentally grounded in the self-organization of information: as informational structure gains dimensionality, material organization follows. This framework connects directly to the Predictive Navigator: at each level of informational dimensionality, the system faces physical constraints that may or may not cross the phase boundary into the predictive regime. Lower-dimensional structures may operate reactively; higher-dimensional structures are forced into predictive architecture. The GEKS Index provides the tool for mapping this hierarchy onto the predictive threshold — making it possible to ask, for the first time, at which level of informational dimensionality predictive processing becomes structurally inevitable.

The question of whether such outcomes are historically contingent or structurally inevitable is addressed in Espesset, D. (2024), “Contingency versus inevitability: a review and reinterpretation of Stephen Jay Gould’s book ‘Wonderful Life’,” Trends in Genetics and Evolution: structural and organizational constraints channel evolutionary outcomes far more narrowly than the contingency thesis allows. Kriger, B. (2025), “The law of imperative uncertainty: Why any complex world requires uncertainty” (https://doi.org/10.5281/zenodo.18101601), establishes the complementary mathematical result: complexity itself generates the conditions under which predictive processing becomes necessary.

The Program: Seven Papers to a Monograph

Paper 1. The Structural Inevitability of Predictive Processing

Integrates Kriger’s Dual-Pressure Convergence Theorem with Espesset’s critique of selectionist explanation. Demonstrates that predictive architecture emerges from physical constraints, not from cumulative selection. Establishes the phase boundary between reactive and predictive regimes.

Builds on: Kriger (2026), “The evolutionary inevitability of predictive processing” (https://doi.org/10.5281/zenodo.18444910); Kriger (2026), “Latency and Compressibility” (https://doi.org/10.5281/zenodo.18752102); Kriger (2026), “Evolutionary and information-theoretic argument for the necessity of representational isolation” (https://doi.org/10.5281/zenodo.18331202); Espesset (2023), “Changement de paradigme,” Eikasía Revista de Filosofía; Espesset (2025), “L’organisme vivant et la construction de l’individu biologique”; Espesset (2025), “Extended book review: How Life Increases Biodiversity.”

Paper 2. The Phase Boundary: Predicting Which Systems Must Be Predictive

Defines the formal threshold for transition from reactive to predictive architecture. Calculates parameters across taxa using the GEKS Index. Generates specific, falsifiable predictions about which taxa must exhibit predictive architecture — the first formally falsifiable formulation of the PP hypothesis.

Builds on: Kriger & Espesset (2026), “The GEKS Index” (https://doi.org/10.5281/zenodo.18775491); Kriger (2026), “The Viability Mismatch Law” (https://doi.org/10.5281/zenodo.18433777); Kriger (2024), “The Transformational Basis of Persistence” (https://doi.org/10.5281/zenodo.18435982); Espesset (2020), “The concept of complexity in evolutive biology.”

Paper 3. Against the Selectionist Theory of Cognition

Argues that the selectionist explanation for PP is explanatorily redundant. Engages with Godfrey-Smith, Sterelny, and Pezzulo. Demonstrates that invoking cumulative selection as creator of cognitive architecture is both unnecessary (Kriger’s theorems) and misleading (Espesset’s critique).

Builds on: Kriger (2022), “Evolutionary Theory of Credence” (https://doi.org/10.5281/zenodo.18379476); Kriger (2025), “The law of imperative uncertainty” (https://doi.org/10.5281/zenodo.18101601); Espesset (2021), “Le darwinisme tient-il debout?”; Espesset (2025), “Schizophrénie darwiniste — Partie 1”; Espesset (2018), “L’évolution biologique: vers une nouvelle synthèse conceptuelle étendue”; Espesset (2018), “Critical scientific reflexions about the Red Queen hypothesis.”

Paper 4. The Dark Room Dissolved

Formal resolution via the Viability Mismatch Law: an agent that ceases to generate prediction error ceases to update its models; environmental drift expands model–reality mismatch until viability is lost. The dark room is not a counterexample to PP but a predictable failure mode. Espesset’s niche-construction framework (Espesset (2021), “Les maîtres de la construction de niche”) provides the biological parallel: organisms actively shape their informational environments.

Builds on: Kriger (2026), “The Viability Mismatch Law” (https://doi.org/10.5281/zenodo.18433777); Kriger (2026), “The Structural Distortion Principle” (https://doi.org/10.5281/zenodo.18452700); Kriger (2026), “The Eruptive Manifestation of Model–Reality Mismatch” (https://doi.org/10.5281/zenodo.18474532); Espesset (2021), “Les maîtres de la construction de niche”; Espesset (2025), “L’organisme vivant et la construction de l’individu biologique.”

Paper 5. Adaptive Observer Theory and the Necessity of Multiple Models

Proves that a predictive system cannot maintain viability with a single generative model. Connects to Espesset’s observation that biological systems maintain redundant and contradictory regulatory mechanisms, and to his analysis of the limits of parsimony (Espesset (2019), “De l’utilisation parcimonieuse du principe de… parcimonie”). The “Bayesian brain” must be a navigator managing a portfolio of models under incompleteness and incompatibility constraints.

Builds on: Kriger (2026), “Adaptive Observer Theory” (https://doi.org/10.5281/zenodo.18812040); Kriger (2026), “The Theorem of Incompatible Truths” (https://doi.org/10.5281/zenodo.18811305); Kriger (2026), “Must Any Consistent Physics Contain Structure Inaccessible to All Internal Observers?” (https://doi.org/10.5281/zenodo.18810467); Kriger (2026), “The Pragmatic Synthesis” (https://doi.org/10.5281/zenodo.18812001); Espesset (2025), “L’organisme vivant et la construction de l’individu biologique”; Espesset (2019), “De l’utilisation parcimonieuse du principe de… parcimonie.”

Paper 6. Coherence, Complexity, and the Predictive Threshold Across Taxa

Applies the GEKS Index to map biological complexity across taxa onto the phase boundary. For organisms above the threshold, predictive architecture is inevitable regardless of phylogenetic path — a quantitative answer to the contingency question (Espesset (2024), “Contingency versus inevitability,” Trends in Genetics and Evolution). Draws on Espesset’s analysis of genome-level complexity (Espesset (2022), “Autour du livre L’entropie génétique et le mystère du génome”) and Kriger’s coherence framework (Kriger (2026), “Coherence Epistemology for AI-Mediated Inter-Species Communication” (https://doi.org/10.5281/zenodo.18778705)).

Builds on: Kriger & Espesset (2026), “The GEKS Index” (https://doi.org/10.5281/zenodo.18775491); Kriger (2026), “Coherence Epistemology” (https://doi.org/10.5281/zenodo.18778705); Kriger (2024), “No final theory” (https://doi.org/10.5281/zenodo.18099738); Espesset (2024), “Contingency versus inevitability,” Trends in Genetics and Evolution; Espesset (2022), “Autour du livre L’entropie génétique et le mystère du génome”; Espesset (2020), “The concept of complexity in evolutive biology.”

Paper 7. The Predictive Navigator: Synthesis

Integrates all results. Addresses the five major PP criticisms — unfalsifiability, the dark room, the evolutionary gap, the implementation gap, the cross-species gap — with formal proofs and evolutionary argument. Proposes the Navigator: not a Bayesian brain, but a viability-maintaining system managing multiple generative models under physical constraints, incompleteness, and incompatibility.

Builds on: All preceding papers, plus: Kriger (2026), “The Predictive Mind and Its Myths” (https://doi.org/10.5281/zenodo.18490146); Kriger (2026), “Against Causation” (https://doi.org/10.5281/zenodo.18851848); Kriger (2026), “Why Mathematics Works” (https://doi.org/10.5281/zenodo.18793228); Kriger (2017), “A Unified Theory of Self-Organizing Systems” (https://doi.org/10.5281/zenodo.18363729); Espesset (2023), “Changement de paradigme,” Eikasía Revista de Filosofía.

The seven papers combine into a co-authored monograph: The Predictive Navigator: From Physical Constraint to Evolutionary Proof.

Publication Plan

Phase I — Foundations (parallel)

Paper 1. The Structural Inevitability of Predictive Processing First choice: Neuroscience & Biobehavioral Reviews — high-impact theory journal, strong PP readership. Second choice: Physics of Life Reviews — interdisciplinary, open peer commentary format.

Paper 3. Against the Selectionist Theory of Cognition First choice: Biology & Philosophy — premier journal at the intersection of evolutionary theory and philosophy of biology. Second choice: Behavioral and Brain Sciences — target article with open peer commentary for maximum engagement.

Phase II — Formal consequences (parallel, after Paper 1)

Paper 2. The Phase Boundary First choice: Journal of Theoretical Biology — formal biological models with quantitative predictions. Second choice: Biosystems — information processing and organization in biological systems.

Paper 4. The Dark Room Dissolved First choice: Philosophical Transactions of the Royal Society B — regularly publishes PP theme issues. Second choice: Adaptive Behavior — interface between formal models and biological behavior.

Paper 5. Adaptive Observer Theory and the Necessity of Multiple Models First choice: Entropy (MDPI) — information-theoretic approaches to cognition and biological systems. Second choice: Synthese — philosophy of science with strong formal theory coverage.

Phase III — Comparative evidence and synthesis (sequential)

Paper 6. Coherence, Complexity, and the Predictive Threshold Across Taxa First choice: Proceedings of the Royal Society B — premier venue for comparative biology with quantitative framework. Second choice: Journal of Evolutionary Biology.

Paper 7. The Predictive Navigator: Synthesis Monograph: The Predictive Navigator: From Physical Constraint to Evolutionary Proof. First choice: MIT Press. Second choice: Oxford University Press.

Existing Publications

Kriger

Predictive processing & dual-pressure framework: Kriger, B. (2026). The evolutionary inevitability of predictive processing: A physical constraint argument. https://doi.org/10.5281/zenodo.18444910 Kriger, B. (2026). Latency and Compressibility: Two Independent Routes to Predictive Architecture in Adaptive Systems. https://doi.org/10.5281/zenodo.18752102 Kriger, B. (2026). Evolutionary and information-theoretic argument for the necessity of representational isolation. https://doi.org/10.5281/zenodo.18331202 Kriger, B. (2026). The Structural Distortion Principle. https://doi.org/10.5281/zenodo.18452700 Kriger, B. (2026). The Predictive Mind and Its Myths. https://doi.org/10.5281/zenodo.18490146

Adaptive Observer Theory: Kriger, B. (2026). Adaptive Observer Theory: Navigation of Incompleteness and Incompatibility. https://doi.org/10.5281/zenodo.18812040 Kriger, B. (2026). Must Any Consistent Physics Contain Structure Inaccessible to All Internal Observers? https://doi.org/10.5281/zenodo.18810467 Kriger, B. (2026). The Theorem of Incompatible Truths. https://doi.org/10.5281/zenodo.18811305 Kriger, B. (2026). The Pragmatic Synthesis. https://doi.org/10.5281/zenodo.18812001

Coherence & complexity: Kriger, B. (2026). Coherence Epistemology for AI-Mediated Inter-Species Communication. https://doi.org/10.5281/zenodo.18778705 Kriger, B. (2026). The principle of optimal coherence. https://doi.org/10.5281/zenodo.18341030

Viability & dynamical systems: Kriger, B. (2026). The Viability Mismatch Law. https://doi.org/10.5281/zenodo.18433777 Kriger, B. (2026). The Eruptive Manifestation of Model–Reality Mismatch. https://doi.org/10.5281/zenodo.18474532 Kriger, B. (2024). The Transformational Basis of Persistence. https://doi.org/10.5281/zenodo.18435982

Evolutionary dynamics: Kriger, B. (2022). Evolutionary Theory of Credence. https://doi.org/10.5281/zenodo.18379476 Kriger, B. (2019). Evolutionary Selection for Atemporal Memory Storage. https://doi.org/10.5281/zenodo.18381880 Kriger, B. (2025). The law of imperative uncertainty. https://doi.org/10.5281/zenodo.18101601 Kriger, B. (2024). No final theory: Law of scale-specific principles. https://doi.org/10.5281/zenodo.18099738

Methodology & epistemology: Kriger, B. (2026). Against Causation. https://doi.org/10.5281/zenodo.18851848 Kriger, B. (2026). Why Mathematics Works. https://doi.org/10.5281/zenodo.18793228 Kriger, B. (2017). A Unified Theory of Self-Organizing Systems. https://doi.org/10.5281/zenodo.18363729

Kriger & Espesset

Kriger, B. & Espesset, D. (2026). The GEKS Index: A Composite Measure of Biological Complexity Across Informational, Structural, Functional, and Evolutionary Dimensions. https://doi.org/10.5281/zenodo.18775491

Espesset

Evolutionary theory & paradigm critique: Espesset, D. (2025). L’organisme vivant et la construction de l’individu biologique: Origine et évolution. Espesset, D. (2025). Schizophrénie darwiniste — Partie 1. Espesset, D. (2025). Extended book review: How Life Increases Biodiversity — An autocatalytic hypothesis by David Seaborg. Espesset, D. (2024). Contingency versus inevitability: a review and reinterpretation of Stephen Jay Gould’s book “Wonderful Life.” Trends in Genetics and Evolution. Espesset, D. (2023). Changement de paradigme de l’évolution biologique: du darwinisme à une nouvelle synthèse conceptuelle élargie. Eikasía Revista de Filosofía. Espesset, D. (2022). Autour du livre L’entropie génétique et le mystère du génome. Espesset, D. (2021). Le darwinisme tient-il debout ? — Revue critique. Espesset, D. (2021). La notion de complexité en biologie évolutive: une revue synthétique. Espesset, D. (2021). Les maîtres de la construction de niche. Espesset, D. (2020). The concept of complexity in evolutive biology: A synthetic review. Espesset, D. (2019). De l’utilisation parcimonieuse du principe de… parcimonie. Espesset, D. (2018). L’évolution biologique: vers une nouvelle synthèse conceptuelle étendue pour le 21e siècle. Espesset, D. (2018). Critical scientific reflexions about the Red Queen hypothesis.

Molecular biology: Espesset, D., Duché, D., Baty, D. & Géli, V. (1996). The channel domain of colicin A is inhibited by its immunity protein through direct interaction in the Escherichia coli inner membrane. The EMBO Journal. Espesset, D., Corda, Y., Cunningham, K. & Géli, V. (1994). The colicin A pore-forming domain fused to mitochondrial intermembrane space sorting signals. Molecular Microbiology. Espesset, D., Piet, P., Lazdunski, C. & Géli, V. (1994). Immunity proteins to pore-forming colicins: Structure-function relationships. Molecular Microbiology.

Community Engagement

The program builds on existing scholarly relationships with Laurent Perrinet (Aix-Marseille / CNRS), whose work on heterogeneous delays in spiking neural networks connects to Paper 1; Giovanni Pezzulo, whose evolutionary PP framework provides context for Paper 3; and Pedro Mediano (Imperial College London), whose work on phase transitions connects to Papers 2 and 6. We welcome inquiries from researchers interested in the program.

What Makes This Program Unique

A formally rigorous theoretical framework with proven theorems and simulation code; a principled evolutionary critique grounded in two decades of published work on biological complexity, self-organization, and the limits of selectionist explanation — from molecular biology through evolutionary theory; a quantitative complexity index (GEKS) already published as a joint tool for cross-taxa prediction; and established relationships with the PP community’s key researchers. This combination exists nowhere else.

The predictive processing community has the theory. We have the proof that it could not have been otherwise.