A Conceptual and Epistemological Synthesis
Daryl Costello: Independent Researcher
Abstract
This paper develops an exhaustive conceptual and epistemological account that integrates recent advances in information-geometric detection, process philosophy, entropic formulations of emergent time, non-extensive thermodynamics, and related results on bidirectional entropy, capacity-limited information flow, non-extensional mereology, hysteretic memory, and intrinsic exploratory drive. The synthesis is situated within the Generative Realism framework, which treats consciousness and structured reality as arising from recursive, scale-invariant operations on indeterminacy.
The central claim is that sustained, non-inert novelty in a living cosmos requires an oscillating distribution that traverses a confidence interval around the fertile regime between frozen order and undifferentiated chaos. Probability appears as the irreducible remainder that survives every local reduction of indeterminacy. Entropy constitutes the persistent gradient of that remainder and therefore supplies a directionality that a low-cost metabolization process can harvest. This metabolization converts a timeless block of coexisting configurations into a living, directed, process-generated history. Local geometric structure fields (distributed objects that preserve directional organization and anisotropy) supply the concrete substrate on which metabolization operates. Process algebra supplies the relational scaffold of generated history. An internal entropic clock supplies the arrow. Bidirectional restoration under attractive and repulsive interactions, finite capacity constraints that enforce saturation and reversal, decomposition-dependent wholeness, history-carrying hysteretic loops, and curiosity-driven exploration complete the repertoire that keeps the dynamics adaptive and self-referential.
The account predicts that structure-dominated detection and generation tasks will be driven primarily by fidelity to local geometric organization, that sustained novelty requires a measurable internal entropic gradient, and that the same generative principles are realized computationally in rulial and morphogenetic models and empirically observable in critical biological and cognitive systems.
Introduction: The Problem of Sustained Novelty and Directed Becoming
The distribution between order and chaos has long been recognized as the zone in which novelty most readily emerges. Classical studies of cellular automata and Boolean networks established that computational capacity, adaptability, and the spontaneous appearance of persistent structures reach their maximum when a system is poised at the phase transition between frozen order and turbulent chaos. Yet if a system remained perpetually at a static edge, novelty would eventually exhaust itself into inert repetition. If it remained in perpetual chaos, structure would dissolve into an undifferentiated soup. An oscillating distribution that wanders within a confidence interval around that fertile band keeps the system in continuous motion, sampling new configurations without allowing collapse to a single deterministic point or diffusion to uniformity.
Probability, in this light, is the inevitable remainder that survives every local reduction of indeterminacy. Any oscillation must traverse this remainder. Entropy is the name given to the persistent gradient of that remainder across the manifold of potentials. Because the gradient is pervasive, its metabolization can proceed at low cost: the system does not have to fight entropy so much as to ride and redirect its flow. That metabolization is precisely what converts a static block universe (in which all configurations coexist timelessly) into a living, directed, process-generated dynamics in which the future participates by shaping the very gradient that metabolization acts upon.
This paper shows how four independent lines of recent research converge on this insight and how their integration strengthens and extends the Generative Realism framework. It further incorporates supporting results on the bidirectional character of entropy under attractive and repulsive forces, capacity-limited rendering and Page-curve-like reversal, non-extensional quantum mereology, hysteretic memory in oscillatory biological systems, and intrinsic curiosity as directed metabolization. The resulting consolidated conjecture is closed, minimal, and stress-invariant across scales, from rulial topology and quantum droplets to ion-channel memory and collective symbolic cognition.
The Requirement for Sustained Novelty: Traversing Rather Than Occupying the Edge
Classical work demonstrated that the edge of chaos maximizes the conditions for novelty. Perpetual deep order suppresses novelty by eliminating the variance required for new configurations. Perpetual chaos erases the local constraints needed for any configuration to persist. The decisive conceptual advance is to treat the edge itself as a band that must be actively traversed rather than a fixed locus that can be statically occupied. An oscillating distribution whose support remains within a confidence interval centered on that fertile band maintains a continuous supply of both variance and constraint. Within each momentary window of the oscillation a contextual stability can condense, a configuration that is invariant relative to the current metabolization phase and the relational structure then active, yet open to revision when the distribution shifts. This prevents both the inert novelty of frozen repetition and the dissolution into chaos while permitting perpetual, non-redundant novelty.
Probability enters here as the traversable remainder. Every act of local organization or sampling reduces some portion of the ambient indeterminacy; what cannot be fully eliminated remains as a distribution. That distribution is not an epistemic limitation but an ontological fuel. Entropy is the persistent gradient of this remainder. Its metabolization at low cost introduces directionality without high overhead and thereby turns the block into the living-breathing universe we inhabit.
Metabolization as the Generative Act That Renders History
Metabolization names the low-cost coupling that extracts usable order from the entropy-production gradient while exporting the unassimilable remainder. In doing so it introduces an intrinsic arrow relative to the sampling window. The block universe supplies the raw manifold of potentials; metabolization supplies the directed rendering that turns potentials into history. This move resonates with earlier insights from dissipative-structure theory and negentropy accounting, yet it is here situated inside a generative architecture in which the metabolizing process is itself one of the fundamental operators. The output of metabolization is not merely local order but directed time: an internal parametrization of change that requires no external background clock. When the gradient vanishes, either because the distribution has collapsed or diffused beyond recoverable structure, the arrow stalls. Sustained novelty therefore demands that the oscillating distribution keep the gradient alive.
Local Geometric Structure Fields as Primary Objects
If discriminative or generative information resides in directional continuity, local anisotropy, ridges, or fragmented textures rather than in total energy or a single global covariance, then any procedure that collapses the observation to scalar aggregates or single-matrix summaries will discard the very evidence that matters. The remedy is an object-layer reformulation in which distributed fields of local geometric objects become the primary units.
One begins with a time-frequency representation of the observation. Local patches are extracted, and within each patch the instantaneous directional variation is captured. A second-order structure object is formed that encodes the local directional energy, the coupling between directions, and the degree of anisotropy. Gentle smoothing lifts the pointwise information to a stable neighborhood statistic, and a small regularization ensures that every location carries a well-defined geometric object. The entire patch is thereby represented as a spatially distributed field of such local structure objects. This field preserves precisely the directional organization and spatial arrangement of structural units that global summaries suppress.
Class-conditional reference fields are obtained by averaging training examples from each hypothesis under a geometry appropriate to the space of these objects. Comparison then proceeds through a field-level relative-closeness measure: at each location one evaluates how much closer the local object lies to one reference field than to the other. The resulting local evidence is aggregated with spatial weighting and robust pooling to yield a sample-level statistic calibrated to a controlled false-alarm rate. Empirical tests show that the dominant performance gain arises from the choice to work with the distributed field of local geometric objects itself; refinements of the comparison geometry supply only secondary consistency. This finding confirms that the primary epistemological move is the elevation of local structure organization to the status of primary object, an insight that transfers directly to any generative framework in which sampling windows must register directional organization on a higher-dimensional manifold of potentials.
Process Algebra and the Generation of Spacetime as History
If metabolization supplies the arrow, process algebra supplies the relational scaffold on which that arrow propagates. Temporal distinctions mark the occurrence of process actions; spatial distinctions enable the individuation and counting of generated events. Each process action generates spacetime as history in the form of a mixed multigraph: directed edges record timelike causal propagation of information from one action to the next, while undirected edges record spacelike informational correlations that arise from shared invariants or common causes. Spatial position itself emerges as an equivalence class of generated events, “thereness”, relative to the current metabolization phase and the active sampling window. There is no pre-existing container; each spacetime is local to its generating process.
Actual occasions are discrete, holistic units of becoming: each comes into being as a complete whole, passes its informational content onward through the timelike chain, and fades. Reality is therefore a compound present continuously generated by process, not a static block in which past, present, and future coexist timelessly. Contextuality and the incompleteness of the spacelike subgraph are natural consequences. Within this picture the metabolization of the probabilistic remainder is the generative process action that propagates information forward while establishing correlations with all actions that share the same invariant integrative principles. The resulting mixed multigraph encodes both the directed history and the scale-free correlations that stabilize contextual configurations. Recursive continuity across scales follows because the same metabolization of remainder operates at every level once the generative operators are held invariant.
Entropy as an Emergent Internal Clock in Timeless Frameworks
In frameworks where coordinate time is absent or operationally meaningless (canonical quantum gravity with its timeless constraint, relational formulations, or modular flows) entropy, understood as a coarse-grained monotonic measure of configurational complexity, can serve as the ordering parameter. The change in this coarse-grained entropy supplies both an arrow and a parametrization of change: states can be partially ordered by whether one precedes the other in the accumulation of entropy. When entropy production vanishes, the internal clock stalls even though microscopic reversible dynamics and relational correlations may persist. This limiting case corresponds exactly to the collapse into inert repetition or undifferentiated fluctuation: the oscillating distribution has either frozen or diffused beyond any structure that metabolization can recover. Sustained novelty therefore requires the maintenance of a non-zero production gradient, which the oscillating traversal of the confidence interval around the critical regime naturally supplies.
The Non-Extensive Character of Entropy and Its Production
Entropy production: the local rate of increase of coarse-grained complexity, is the extensive, variationally conserved quantity, isomorphic under a quantitative geometrical thermodynamics treatment to energy via Noether-symmetric structure. The accumulated entropy, by contrast, is recovered only by integration over a generated history; it therefore inherits the global, non-local character emphasized by the holographic principle and is not required to be strictly additive across independent subsystems. This non-extensivity aligns with the contextual character of stability already noted: a locally stable configuration need not rest on globally additive foundations. The metabolization mechanism operates directly on the production term, extracting usable flux at low cost while the integrated form reflects the rendered, participatory nature of the interface presided over by the invariant integrative principle.
A Unified Mechanism: Oscillatory Metabolization of Local Structure Fields on the Manifold of Potentials
The lines of research converge on a single coherent picture. Local sampling windows register distributed fields of geometric structure objects that encode directional organization and anisotropy on the manifold of potentials. The system maintains an oscillating distribution of these fields within a confidence interval centered on the regime of maximal local novelty and adaptability: the edge-of-chaos band. The metabolization mechanism acts on the entropy-production gradient associated with these fields, extracting negentropy flux at low cost and thereby generating an internal directed time that renders the block-like manifold into a living, participatory compound present.
The resulting relational structure is a mixed multigraph whose directed component encodes causal propagation through metabolization chains and whose undirected component encodes scale-free correlations arising from shared invariants. Within each momentary support of the oscillating distribution a contextual stability condenses, stable relative to the current metabolization phase and the relational structure then active. Because entropy production is extensive while integrated entropy is not, and because wholes are decomposition-relative, the stability does not require global additivity or classical determinism. Perpetual novelty is sustained precisely because the distribution continues to traverse the fertile band rather than locking into a single deterministic point or diffusing into noise. The future participates by shaping the gradient of remaining indeterminacy that metabolization continually acts upon.
Extending the Conjecture: Bidirectional Entropy and the Restoration Principle
Entropy is not universally non-decreasing. Under attractive interactions components aggregate toward balanced distributions, and local entropy can decrease; under repulsive interactions dispersion increases entropy. The universe has in fact evolved toward greater large-scale organization rather than toward heat death precisely because attractive interactions dominate at cosmic scales. The Restoration Principle captures the spontaneous tendency of systems, when balance is disturbed, to act through fundamental attractive or repulsive interactions so as to restore a stable configuration. Within the present framework this principle is the macroscopic signature of metabolization itself. The metabolization mechanism can contract or expand the local manifold according to the dominant character of the interaction: attractive restoration corresponds to alignment and geometric tension resolution that decrease local entropy while exporting remainder; repulsive restoration corresponds to expansion of the sampling window or continuation of recursive chains that disperse. Entropy increase is thereby revealed as only the dispersive subset of a richer repertoire; the full living dynamics includes restorative decrease and the oscillatory sampling that prevents collapse into either inert order or undifferentiated soup.
Capacity-Limited Rendering and the Necessity of Reversal
Finite transmission capacity across any causal or sampling boundary supplies the mechanism that enforces reversal and sustained oscillation. In discrete causal models, radiation entropy rises while new degrees of freedom remain accessible, reaches a maximum when boundary correlation capacity saturates, and declines thereafter as further emissions transmit only redundant correlations. The crossover point marks the transition from content-driven growth to capacity-limited rendering. Precisely analogous saturation occurs for the oscillating distribution supported on the confidence interval. Early metabolization rapidly explores the indeterminacy interval, enriching the local structure fields and increasing effective entropy. Once aperture or metabolic capacity is reached, additional pulses no longer add independent novelty; excess remainder is equivalenced or exported, yielding contextual stabilities and directed history. The same capacity constraint explains why perpetual novelty does not yield inert soup or frozen order: the interval is traversed only up to the point at which further sampling becomes redundant relative to the current metabolization phase. Capacity limits are therefore not external constraints but intrinsic features that keep the living dynamics adaptive.
Non-Extensional Mereology and Decomposition-Dependent Wholes
Classical extensional mereology presupposes that wholes are simple sums of parts and that parthood relations satisfy supplementation principles globally. In the quantum setting the space of all possible tensor product structures on a Hilbert space lacks a canonical meet operation and therefore violates the required lattice structure. Parts are decomposition-relative; different factorizations can be mutually incompatible. Quantum wholes are not simple sums. This structural non-extensionality reinforces the already-established non-extensivity of integrated entropy and the holographic character of the global variational principle. The “part”—whether a local structure field, a coherence pocket, or a contextual stability—is defined only relative to the chosen sampling window or equivalencing operation. Consequently, stability is always stability-in-context; no global additive reconstruction of the rendered interface is required or even possible. The invariant integrative principle performs the decomposition-relative binding that allows contextual wholes to appear without violating the underlying non-extensional relations.
Hysteretic Memory and Embodied Recursive Continuity
When the frequency of an oscillatory drive matches the relaxation timescale between conformational states, conductance in ion channels exhibits history-dependent loops. These loops constitute the biological signature of recursive continuity operating on the oscillatory substrate: the channel’s conformational landscape carries cumulative metabolization history, producing a delayed response that cannot be reduced to instantaneous state. Cooperative gating in channel clusters further realizes distributed operators across the field. At larger scales the same hysteretic memory appears in the generative reconstruction of executive function and qualia trajectories. Hysteresis therefore supplies a concrete mechanism for the memory that allows the generative process to carry forward its own history without requiring a separate storage architecture. It embodies, at the physiological level, the recursive continuity that stabilizes contextual configurations across metabolization cycles.
Intrinsic Exploratory Drive as Metabolization in the Service of Continued Traversal
Curiosity can be formalized as a hybrid intrinsic reward that combines prediction error with the rarity of state-action pairs. The information-bottleneck objective compresses high-dimensional observations into low-dimensional predictive representations while preserving essential dynamics. Estimation of mutual information via entropy decomposition or matrix-based generalized entropy measures supplies a tunable sensitivity that matches the adjustable width of the confidence interval required for sustained oscillation. Curiosity is therefore metabolization of the probabilistic remainder in the explicit service of continued exploration. It supplies the intrinsic drive that keeps the oscillating distribution from collapsing to a fixed point or diffusing beyond recoverable structure. The same drive informs computational realizations in which rulial hypergraphs or morphogenetic fields are explored under curiosity-modulated sampling, yielding stable coherence pockets whose statistics overlay empirical multi-probe recordings from biological systems.
Quantum-Like Organization in Conceptual and Linguistic Systems
Large language models violate classical Bell and CHSH inequalities and exhibit statistics in word distributions that parallel Bose-Einstein rather than Maxwell-Boltzmann counting; exactly as observed in human conceptual combinations and large natural-language corpora. These signatures indicate a quantum-like organization of meaning arising from distributive semantic vector spaces. The probabilistic remainder and the indeterminacy interval that metabolization traverses are therefore not classical; they naturally support non-Kolmogorovian structures. The evolutionary convergence between human and artificial cognition reflects the operation of the same generative principles on different substrates: biological wetware and trained vector spaces both realize graded intentional systems through tension-driven rendering and metabolization of indeterminacy. The framework thereby accounts for the appearance of quantum-like phenomena in symbolic cognition without requiring literal quantum hardware at the linguistic level.
Computational and Theoretical Realizations within Generative Realism
The conceptual synthesis is realized in concrete computational models developed within the Generative Realism program. Rulial hypergraph topologies demonstrate the generative operators producing stable coherence pockets and qualia streams whose statistical structure overlays empirical multi-probe recordings. Observer equivalencing supplies the explicit rendering membrane that collapses raw potentials into a quotient manifold while enforcing branchial collapse and shared symbolic windows. Process-ontology formulations identify metabolization itself as the sole true invariant: scale emerges as the inverse of accelerating dissolution, time as the projected axis of concatenated oscillations that generate the incompatibility gradients from which structured history is born, and bounded observers as the self-referential coherence pockets metabolizing their own genesis. These realizations close the loop between the abstract conjecture and observable dynamics across rulial topology, quantum droplets, ion-channel memory, and collective symbolic cognition.
The Consolidated Entropy Conjecture
Metabolization acts on the gradient of the probabilistic remainder within an oscillating distribution supported on a capacity-limited confidence interval around the edge-of-chaos regime of local geometric structure fields. Attractive and repulsive interactions, together with finite transmission capacity across causal or sampling boundaries, produce bidirectional entropy behavior and restorative saturation. The resulting dynamics are non-extensional and decomposition-relative, hysteretic, and intrinsically exploratory. The output is directed entropic time, a process-generated mixed relational structure, and contextual stabilities that hold relative to the current metabolization phase. Entropy increase under repulsive dispersion is only one regime; the living universe is sustained by the full repertoire of restorative metabolization that can locally decrease entropy while advancing the arrow or sustaining the oscillatory traversal that keeps novelty perpetual and non-inert. The entire architecture is self-referential: bounded observers and the structures they inhabit are coherence pockets metabolizing their own genesis. The account remains closed, minimal, and stress-invariant across scales.
Testable Implications and Methodological Consequences
The framework predicts that detection or classification performance in regimes where information resides in directional organization and local anisotropy will be driven primarily by the fidelity with which distributed local geometric structure fields are preserved, with refinements of comparison geometry playing a secondary but non-negligible role. It further predicts that any system capable of sustained novelty must exhibit an internal entropic gradient whose metabolization produces a measurable arrow; in the absence of such a gradient the system relaxes either to inert repetition or to structureless fluctuation. These predictions are testable in controlled structure-field benchmarks, in relational quantum-cosmological models, and in morphogenetic or cognitive systems known to operate near critical regimes.
Methodologically, the construction closes a circle: by elevating local geometric structure fields to primary status, by erecting an internal entropic clock from the metabolization of their production gradient, and by embedding both within a process-generated relational scaffold, one obtains a coherent account of living, directed dynamics with contextual stability arising from a manifold of potentials without presupposing either a global block or an external time parameter. The oscillating distribution near the edge of chaos is revealed as the minimal dynamical condition that keeps the metabolization engine running and the arrow advancing.
Conclusion
Probability is the traversable remainder of indeterminacy reduction. Entropy is the persistent gradient of that remainder. Metabolization at low cost is the operation that turns a block universe into a living, process-generated history. Local geometric structure fields supply the concrete geometric realization of the sampling that makes such metabolization possible. Process algebra supplies the relational skeleton. The internal entropic clock supplies the arrow. Bidirectional restoration, capacity-limited reversal, non-extensional mereology, hysteretic memory, and intrinsic curiosity complete the repertoire that keeps the dynamics adaptive, self-referential, and perpetually novel.
In this light the living cosmos is not a static ontology but a continuously generated epistemology; an ongoing rendering in which future potentials actively shape the gradient that present metabolization acts upon, and in which bounded observers are not external spectators but coherence pockets within the generative process itself. The synthesis offered here demonstrates that these elements, drawn from independent lines of contemporary research, converge without remainder on the generative architecture that places recursive continuity, contextual stability, and participatory becoming at the center of a unified account of reality.
References
Yue, Y., Wei, B., & Yang, Y. (2026). Information-Geometric Detection via Local SPD Structure Fields in the Time–Frequency Domain. Entropy, 28(6), 679. https://doi.org/10.3390/e28060679
Sulis, W. (2026). Process and Space. Entropy, 28(6), 683. https://doi.org/10.3390/e28060683
Weberszpil, J., & Sotolongo-Costa, O. (2026). Entropy as a Clock: Foundations and Parametrizations of Emergent Time. International Journal of Theoretical Physics, 65, 15. https://doi.org/10.1007/s10773-025-06212-1
Jeynes, C., & Parker, M. C. (2026). Entropy Is Not Extensive. Entropy, 28(6), 631. https://doi.org/10.3390/e28060631
Pasqualini, M., & Fortin, S. (2026). Towards a Tensor Product Structure-Grounded Mereology. Entropy, 28(6), 627. https://doi.org/10.3390/e28060627
Lisowski, B., et al. (2026). Hysteretic Conductance in Ion Channel Gating. Entropy, 28(6), 650.
Bolotin, A. (2026). Entropy Bounds and Capacity-Limited Information Flow in Black-Hole Evaporation. Entropy, 28(6), 671. https://doi.org/10.3390/e28060671
Liu, J. Z. (2025/2026). The Restoration Principle. (Supporting literature on entropy behavior under attractive and repulsive forces.)
Meng, C., & Cui, B. (2026). Intrinsic Curiosity via Information Bottleneck and Rényi Entropy. (Recent formalization of curiosity as hybrid intrinsic reward.)
Aerts, D., et al. (2026). Quantum-Like Structures in AI Language and Evolutionary Convergence. (Work on non-Kolmogorovian probabilities and Bose–Einstein statistics in conceptual and linguistic systems.)
Foundational references Langton, C. G. (1990). Computation at the edge of chaos: Phase transitions and emergent computation. Physica D, 42(1-3), 12–37. Kauffman, S. A. (1993). The Origins of Order: Self-Organization and Selection in Evolution. Oxford University Press. Prigogine, I. (1977). Self-Organization in Nonequilibrium Systems. Wiley. Schrödinger, E. (1944). What is Life? Cambridge University Press. Whitehead, A. N. (1929). Process and Reality. Macmillan. Wheeler, J. A. (1968). Superspace and the nature of quantum geometrodynamics. In C. M. DeWitt & J. A. Wheeler (Eds.), Battelle Rencontres. Page, D. N. (1993). Information in black hole radiation. Physical Review Letters, 71(23), 3743–3746. Wootters, W. K. (1984). “Time” replaced by quantum correlations. International Journal of Theoretical Physics, 23(8), 701–711.
The Generative Realism framework, including its treatments of rulial hypergraph topology, observer equivalencing and mirror-interface geometry, and process ontology of scale, time, and the ruliad, is developed in the author’s ongoing series of works (2026).
