Foundations of a General Framework Deriving Physical, Biological, and Informational Structure from a Single Primitive

Author: Daryl Costello

Affiliation: Independent Theoretical Research, Rosendale, New York, United States

Correspondence: Daryl.Costello@outlook.com

Institute for Foundational Structure, Ulster Park, NY, United States

Foundations of Physics / Physical Review X |  Submitted October 2026  |  Manuscript No. KFA-2026-001
Keywords: kernel-first architecture; relational ontology; renormalization group; holonomy; emergent medium; quantum gravity; morphogenesis; Page–Wootters mechanism

ABSTRACT

We introduce Kernel-First Architecture (KFA), a foundational framework in which all structure (physical, informational, or dynamical) is derived from a single primitive: the kernel, defined as the irreducible act of relation between adjacent strata. The kernel K(α, β) is not a state, a field, or a medium; it is the minimal relational generator between two neighboring layers of description, belonging to neither stratum alone but irreducibly interstitial. From this single primitive, three operational principles unfold systematically. Principle I (Seed of Relation): the kernel generates the relational Hamiltonian H_rel(α|β) (the full dynamical structure of stratum β as conditioned on stratum α) without presupposing either stratum’s intrinsic properties. Principle II (Negotiation of Adjacency through Resolution): at each coarse-graining scale b, block-composition of primitive kernel transactions produces a new effective medium M(b); the medium is always an output of kernel negotiation, never an input. Principle III (Connective Tissue): the geometric (holonomic) imprint of all kernel negotiations across a stratum stack constitutes global coherence; binding strata into a coherent whole without presupposing any fixed background structure. The central thesis is therefore: the medium is always emergent from kernel negotiations, never primitive. KFA is demonstrated in three domains: (a) stratified optical media, where the full mathematical elaboration (including relational optical Hamiltonians, gyroscope precession mapping on the Poincaré sphere, and renormalization group coarse-graining) is developed in the companion paper (Costello et al., “Relational Optical Hamiltonians, Gyroscope Precession Mapping, and Renormalization Group Coarse-Graining in Stratified Media: A Kernel-First Architecture,” 2026); (b) quantum gravity, where KFA reframes the emergence of classical spacetime as the infrared fixed point of the quantum gravity kernel renormalization group flow; and (c) biological morphogenesis, where tissue compartmentalization is shown to be the stabilization of locally coherent morphogenetic kernel transactions into organ domains. Across all three domains, KFA predicts specific, testable signatures: power-law domain size distributions near kernel renormalization group critical points, non-trivial geometric phases observable as holonomies, and resolution-dependent effective parameters tracking the kernel beta-function. KFA constitutes a meta-architecture: a framework from which domain-specific theories are derived rather than assumed.

TABLE OF CONTENTS 1.   Introduction: The Problem of Primitive Structure 2.   The Kernel: Definition and Formal Structure 2.1   The Primitive of Relation 2.2   The Kernel as Relational Generator 2.3   Strata as Kernel Equivalence Classes 3.   Operational Principle I: The Seed of Relation 3.1   The Kernel as Minimal Relational Unit 3.2   Observer-Dependence as Structural, Not Epistemic 3.3   Worked Illustration: The Two-Stratum Kernel 4.   Operational Principle II: Negotiation of Adjacency through Resolution 4.1   Resolution as Coarse-Graining Scale 4.2   The Emergent Medium 4.3   The RG Flow as Kernel Flow 4.4   Relevant, Irrelevant, and Marginal Kernel Deformations 5.   Operational Principle III: Connective Tissue across Strata 5.1   The Geometric Imprint of Kernel Composition 5.2   Non-Commutativity as Source of Coherence 5.3   Connective Tissue as Physical Observable 5.4   Topological Protection of Connective Tissue 6.   Domain Emergence from Kernel Stability 6.1   Domains as Fixed-Point Regions 6.2   Domain Walls as Kernel Instabilities 6.3   Near-Critical Stacks and Power-Law Domain Structure 6.4   Observer-Dependent Domains 7.   Three Domain Demonstrations 7.1   Stratified Optical Media 7.2   Relational Quantum Gravity 7.3   Biological Morphogenesis and Tissue Stratification 8.   Implications and Open Questions 8.1   The Ontological Inversion: Relation Before Substance 8.2   Experimental Signatures of KFA 8.3   Toward a General KFA Calculus 8.4   KFA and Information Theory 9.   Discussion and Outlook 10. Conclusion Appendix A: Kernel Axioms: Complete Formal Statement Appendix B: Category-Theoretic Formulation Sketch Appendix C: Connection to Known Formalisms References

1. Introduction: The Problem of Primitive Structure

Every formal theory must begin somewhere. The choice of primitive (the entity taken as given before derivation begins) is not merely a matter of convenience or historical accident. It is a foundational commitment that determines, in advance, what can be derived and what must be assumed. A theory that takes the field as primitive can derive particle-like excitations but must assume the spacetime in which the field lives. A theory that takes the agent as primitive can derive strategic equilibria but must assume the strategy space within which agents act. In each case, the primitive purchases explanatory power at the cost of a presupposition that cannot itself be explained within the theory.

The claim of this paper is that across the most fundamental open problems in physics, biology, and information theory, a common pattern of illegitimate presupposition recurs: the medium (the background, the space, the channel) is taken as primitive, and dynamics are derived within it. We argue that this is the wrong inversion. The medium is itself constituted by relations between the entities it supposedly contains. When this is recognized, and the relation (not the medium) is taken as primitive, a new class of foundational theories becomes possible. We call this class Kernel-First Architecture (KFA).

1.1 Three Historical Tensions

The Background-Independence Problem in Quantum Gravity. General relativity teaches that spacetime is not a fixed stage on which physics occurs but a dynamical participant; curved by matter and energy, evolved by the Einstein field equations. Quantum mechanics, by contrast, requires a fixed temporal background against which unitary evolution is defined: the Schrödinger equation ∂ψ/∂t = −iHψ presupposes a time parameter t that is external to the system. The problem of quantum gravity is precisely the tension between these two commitments. Loop quantum gravity (LQG) [1, 2] and causal dynamical triangulation (CDT) [3] both attempt to discretize spacetime and derive smooth geometry as an emergent, coarse-grained property. Yet both still require a notion of adjacency (of which building blocks are “next to” which others) that is itself not derived but assumed. The problem of time [4] deepens this: in the Dirac quantization of general relativity, the total Hamiltonian is constrained to vanish, H_totalψ = 0, leaving no external time parameter. Dynamics can only be relational (defined by correlations between subsystems) yet the framework for defining those correlations still presupposes a background Hilbert space structure. KFA addresses this by making the relation itself (not the Hilbert space, not the time parameter) the primitive.

The Morphogenetic Field Problem in Developmental Biology. The development of a multicellular organism from a single fertilized egg requires that cells acquire positional identity; they must “know” where they are within the organism and differentiate accordingly [5, 6]. The concept of the morphogenetic field, introduced by Gurwitsch and formalized by Wolpert’s positional information framework [5], captures the fact that spatial patterning information exists as a field defined over the embryo. Yet every existing formalization of morphogenetic fields presupposes a fixed spatial domain over which the field is defined; a background geometry against which cell positions are measured. In reality, the spatial domain itself is constituted by the pattern of cell-cell signaling: the tissue creates its own geometry through the adjacency negotiations of its constituent cells [7, 8]. KFA makes this precise: the tissue is not a fixed space populated by signaling molecules; it is the emergent medium produced by cell-cell kernel transactions.

The Channel Capacity Problem in Information Theory. Shannon’s channel capacity theorem [9] defines the maximum rate at which information can be reliably transmitted through a channel with given noise characteristics. But what constitutes the channel? In the classical formulation, the channel is given; a fixed conditional probability distribution p(y|x) mapping inputs to outputs. In quantum information theory [10, 11], the channel is a completely positive trace-preserving (CPTP) map; again, given in advance. Yet in any realistic physical implementation, the channel is itself constituted by the correlations between transmitter and receiver; by their shared physical substrate, which is itself a product of their history of interaction. The capacity of the channel is not a property of a pre-given medium but an emergent property of the accumulated kernel transactions between the communicating parties. KFA identifies the quantum channel with the kernel and the channel capacity with the kernel’s information-theoretic resolution.

1.2 The KFA Inversion

In each of the above cases, the standard approach takes the medium as primitive (spacetime, morphogenetic field domain, quantum channel) and derives dynamics within it. KFA inverts this: the relation between adjacent strata is taken as primitive, and the medium is derived as the effective output of kernel negotiations enacted at a given resolution scale. This inversion is not merely formal; it changes what questions can be asked and what answers are available. Instead of asking “what dynamics live in this medium?”, KFA asks “what kernel negotiations produce this medium?” The latter question has a systematic, constructive answer; the former does not, because it cannot explain the medium itself.

1.3 The Three KFA Operational Principles

The full formal development of KFA rests on three operational principles derived from the single primitive of the kernel:

1. Principle I: The Seed of Relation: The kernel K(α, β) is the irreducible generator of coupling between adjacent strata α and β. It determines the relational state space, the relational Hamiltonian, and the observable algebra of β as seen from α; without presupposing either stratum independently.
2. Principle II: Negotiation of Adjacency through Resolution: At each coarse-graining scale b, block-composition of primitive kernel transactions produces a new effective medium M(b). The renormalization group (RG) flow of the kernel under resolution change classifies the stability of adjacency negotiations and identifies the emergent media associated with RG fixed points.
3. Principle III: Connective Tissue across Strata: The cumulative geometric (holonomic) imprint of all kernel negotiations along a stratum stack constitutes the global coherence of the system; a physical observable that binds strata together without reference to any external background.

The paper is organized as follows. Section 2 develops the formal definition of the kernel and its axioms. Sections 3–5 develop the three operational principles in turn. Section 6 analyzes domain emergence from kernel stability. Section 7 demonstrates KFA in three target domains. Sections 8–10 discuss implications, open questions, and conclusions. Appendices provide complete formal statements and connections to known formalisms.

2. The Kernel: Definition and Formal Structure

2.1 The Primitive of Relation

Let {α, β, γ, …} denote a collection of strata; layers of description, physical, biological, or informational. Each stratum α is associated with a state space V_α. We make no assumption about the nature of these state spaces beyond their existence as vector spaces (or Hilbert spaces in the quantum case); in particular, we do not assume any metric, topology, or background geometry on the collection of strata.

Definition 1 (Kernel). The kernel K(α, β) associated with two adjacent strata α and β is an element of Hom(Vα, Vβ); the space of bounded linear maps from Vαto Vβ. The kernel is the irreducible primitive of relation between α and β: it is not an element of Vαalone, nor fVβ alone, but is irreducibly interstitial between the two strata.

This definition immediately contrasts with conventional approaches. In standard field theory, V is taken as primitive; the field φ lives in a fixed spacetime V, which is not derived from any more primitive structure. In graph theory, the node set {α, β, …} and the edge set {K(α, β), …} are taken as equally primitive co-inputs to the theory. KFA takes only the edge (the kernel) as primitive. The node (stratum) is derived from the kernel, as we show in §2.3.

The kernel satisfies four axioms, stated here informally and given complete formal treatment in Appendix A:

Axiom K1 (Irreducibility). K(α, β) cannot be factored as f(α) ⊗ g(β) for any maps f : Vα→ W and g : Vβ→ W. The kernel is genuinely interstitial; it encodes information about the adjacency that is not reducible to information about either stratum separately.

Axiom K2 (Asymmetry). In general, K(α, β) ≠ K(β, α)†; the relation from α to β is not the adjoint of the relation from β to α. This asymmetry encodes the observer-dependence of KFA: the kernel transaction from α to β is structurally distinct from the reverse transaction.

Axiom K3 (Composability). For adjacent triples (α, β, γ), the kernel K(α, γ) is defined by composition: K(α, γ) = K(β, γ) ∘ K(α, β). Kernel transactions compose along the stratum stack.

Axiom K4 (Resolution-Dependence). The kernel Kb(α, β) at resolution scale b satisfies Kb'(α, β) = Rb’/b[Kb(α, β)] for b’>b, where Rb’/b is a coarse-graining operator. The kernel at coarser resolution is the image of the finer-resolution kernel under coarse-graining.

Remark 1. Axiom K3 implies that the collection of kernel transactions along a stratum stack is composable, but the composition is generically non-commutative: K(β, γ) ∘ K(α, β) ≠ K(α, β) ∘ K(β, γ) as maps on different spaces. This non-commutativity is not a defect; it is the source of holonomy (Principle III) and global coherence.

2.2 The Kernel as Relational Generator

In any Hamiltonian system, the dynamics of a subsystem β are generated by a Hamiltonian H_β. But in KFA, H_β is not primitive; it is derived from the kernel K(α, β) for some adjacent reference stratum α. We call this derived Hamiltonian the relational Hamiltonian:

H_rel(α|β) = −i log K(α, β)      (1)

More precisely, when K(α, β) is a unitary operator (as in the quantum case), equation (1) extracts the generator of the kernel transaction. This identification connects KFA directly to the Page–Wootters mechanism [12, 13]: the kernel K(α, β) is the entanglement between a “clock” stratum α and a “system” stratum β that constitutes relational dynamics. Page and Wootters showed that a timeless quantum state |Ψ⟩_αβ of the total system, satisfying H_total|Ψ⟩ = 0, nonetheless encodes non-trivial dynamics of β conditioned on α; the relational evolution of β relative to the “clock” readings of α.

In KFA, this is generalized: the Dirac constraint

H_total |Ψ⟩ = 0      (2)

is not a pathology but the correct statement that no stratum has absolute dynamics; only kernel-mediated relational dynamics exist. The vanishing of the total Hamiltonian does not mean nothing happens; it means that all dynamics are relational, encoded in the kernel K(α, β). This connection has been elaborated by Gambini and Porto [13] in the discrete clock setting and by Giovannetti, Lloyd, and Maccone [14] in the quantum reference frame formulation.

2.3 Strata as Kernel Equivalence Classes

Theorem 1 (Strata from Kernels). Given a collection of kernels {K(α, β)} satisfying axioms K1–K4, define an equivalence relation ~ on the set of kernel “endpoints” by: two endpoints are equivalent if and only if they participate in identical kernel relations with all other endpoints. Then each equivalence class under ~ defines a stratum. Strata need not be assumed; they are derived as the equivalence classes of kernel relations.

The proof follows directly from the axioms: K3 (composability) ensures that equivalence classes are well-defined under kernel composition, while K1 (irreducibility) ensures that no two distinct equivalence classes can be identified without loss of kernel information. This is the KFA formulation of Leibniz’s principle of the identity of indiscernibles [15]: a stratum has no identity apart from its kernel relations. Two strata that participate in identical kernel transactions with all other strata are the same stratum.

A stratum α is therefore not a primitive entity but a derived structure — the set of all kernel relations that terminate at α. In this sense, KFA is radically relational: substance is derived from relation, not the reverse.

Remark 2. This derivation of strata from kernels resolves a potential circularity. One might ask: “If strata are defined by their kernel relations, and kernels relate strata, is there not a vicious circularity?” The answer is no. The kernel is primitive; the stratum is derived as its equivalence class. There is no circularity, only a hierarchy of derivation: kernels first, strata second. This is the KFA analog of Whitehead’s process philosophy [16], in which events (processes) are primitive and enduring objects (substances) are derived patterns of process.

3. Operational Principle I: The Seed of Relation

3.1 The Kernel as Minimal Relational Unit

The first KFA operational principle states that the kernel is the seed of relation: the minimal unit from which all structure of a stratum-pair coupling is generated. The seed metaphor is precise: a biological seed contains the full generative program of an organism without containing the organism itself. It encodes the instructions for generating the organism in a form that is radically smaller than and irreducible to the organism. Analogously, the kernel K(α, β) contains the full generative structure of the coupling between strata α and β without containing either stratum.

Specifically, the kernel K(α, β) determines three structures:

(i) The relational state space. The state space of β “as seen from” α is defined as the image of V_α under K(α, β):

V_β|α := Im(K(α, β)) ⊆ V_β (3)

Note that V_β|α need not equal V_β: the full state space of β may not be accessible from the vantage of α. This is the KFA encoding of complementarity; different relational vantages make different aspects of a stratum accessible.

(ii) The relational Hamiltonian. As defined in equation (1), H_rel(α|β) is derived from K(α, β) and generates the dynamics of β conditioned on α.

(iii) The observable algebra. The set of physical quantities defined for β relative to α is the algebra of operators on V_β|α that commute with the kernel action:

O(β|α) := {O ∈ End(V_β) : O ∘ K(α, β) = K(α, β) ∘ O_α for some O_α ∈ End(V_α)}      (4)

This observable algebra is the set of quantities that can be measured in β that have a well-defined relational meaning with respect to α. It generalizes the notion of a gauge-invariant observable in constrained Hamiltonian systems.

3.2 Observer-Dependence as Structural, Not Epistemic

A central feature of KFA is that observer-dependence is structural rather than epistemic. In standard quantum mechanics, the dependence of measurement outcomes on the observer’s reference frame is sometimes treated as a limitation; a sign that our description is incomplete or perspective-bound. In KFA, observer-dependence is built into the formalism at the level of the primitive: H_rel(α|β) ≠ H_rel(γ|β) for different clock strata α ≠ γ. Two different relational vantages produce genuinely different (not merely differently described) physical structures.

This connects to Rovelli’s relational quantum mechanics [17], in which the quantum state of a system is always relative to another system (the observer), and to Fuchs’s QBism [18], in which quantum states encode an agent’s beliefs rather than objective properties. KFA differs from both: in KFA, the observer-dependence is not postulated as an interpretational stance but derived from the kernel axioms. The asymmetry K(α, β) ≠ K(β, α) (Axiom K2) is the formal source of observer-dependence, and the kernel axioms provide the derivation.

3.3 Worked Illustration: The Two-Stratum Kernel

To make these ideas concrete, consider the minimal KFA system: two strata α (the “clock”) and β (the “system”), coupled by a kernel K(α, β). Let V_α = ℂ^dα and V_β = ℂ^dβ, so K(α, β) is a d_β × d_α complex matrix.

Relational state space. V_β|α = Im(K) has dimension rank(K) ≤ min(d_α, d_β). The effective dimension of β as seen from α is determined by the rank of the kernel transaction.

Relational Hamiltonian and propagator. Treating K(α, β; t) as time-dependent (parameterized by the clock stratum α), the relational propagator from time t_i to t_f is the time-ordered exponential:

G_rel(t_f | t_i) = T exp(∫_ti^tf H_rel(α|β; t) dt)      (5)

This propagator is the kernel-mediated evolution of β relative to α. Note that G_rel depends on both strata; it is not an intrinsic property of β alone.

Asymmetry and physical inequivalence. The reverse kernel K(β, α) produces a different relational Hamiltonian H_rel(β|α) and a different propagator G_rel‘(t_f | t_i). The two propagators are generically inequivalent; they describe genuinely different physical situations: one in which α serves as clock for β, and one in which β serves as clock for α. In quantum gravity, this asymmetry captures the fact that the choice of “clock subsystem” is a physical choice with physical consequences [14].

4. Operational Principle II: Negotiation of Adjacency through Resolution

4.1 Resolution as Coarse-Graining Scale

The second KFA operational principle concerns the scale at which kernel transactions are enacted. We formalize this through the concept of resolution.

Definition 2 (Resolution Scale). The resolution scale b ∈ ℤ+is the number of primitive kernel transactions (at scale b = 1) grouped into a single block transaction. At b = 1, every adjacency is resolved at the finest scale available. At b = N (the total number of strata), the entire stack is resolved as a single effective kernel Keff.

The block kernel at resolution scale b is defined by composing b consecutive primitive kernels:

K_b(α_I, α_I+1) = ∘_{i ∈ blockI} K₁(α_i, α_i+1)      (6)

where block_I denotes the index set of the I-th block of b strata, and ∘ denotes kernel composition (Axiom K3). The block kernel K_b(α_I, α_I+1) is itself a kernel satisfying all four axioms; the kernel structure is preserved under block composition.

The negotiation of adjacency is the process by which block-kernel transactions are enacted from primitive ones. At each resolution scale, a new effective description of the stratum-pair coupling is produced. This is not a loss of information but a change of informational focus: coarser resolution reveals different features of the kernel structure than finer resolution does.

4.2 The Emergent Medium

Definition 3 (Emergent Medium). The emergent medium M(b) at resolution scale b is the collection of effective kernels {Kb(αI, αI+1)} across all block pairs I = 1, 2, …, ⌈N/b⌉. The medium at resolution b is the effective adjacency structure of the stratum stack as seen at that resolution.

Three regimes are distinguished:

• Fine-grained medium M(1): At b = 1, all primitive adjacency relations are visible. The medium is maximally heterogeneous; it reflects the full complexity of individual kernel transactions.
• Coarse-grained medium M(∞): In the limit b → N, all fine-grained adjacency relations are integrated out, and the stack appears as a single homogeneous kernel K_eff. The medium appears uniform.
• Intermediate media M(b): At intermediate b, the medium reflects the mesoscopic structure of the kernel stack; the correlations between kernel transactions that survive at scale b.

The key ontological insight is as follows: the medium is not given. Different resolution scales produce genuinely different media. A fine-resolution observer and a coarse-resolution observer do not disagree about the same medium; they inhabit different media, each of which is the correct description at their respective resolution. This is not instrumentalism; it is the correct ontology of a relational universe.

4.3 The RG Flow as Kernel Flow

The dependence of the kernel on resolution scale defines a flow in the space of kernels: the kernel renormalization group flow:

dK_b / d(log b) = β_K(K_b)      (7)

where β_K is the kernel beta-function; the infinitesimal generator of the kernel RG flow. This generalizes the Wilsonian RG [19, 20] and Polchinski’s exact RG equation [21] to the kernel setting. The key difference is that in standard RG, the flow is of coupling constants in a Lagrangian defined on a fixed medium; in KFA, the flow is of the kernel itself (the adjacency negotiation) and the medium is a derived output of the flow.

Fixed points of the kernel RG flow satisfy β_K(K*) = 0 and are classified as:

• UV fixed point (b → 0): The maximally fine-grained kernel; all adjacency relations are visible; the medium is maximally heterogeneous. This fixed point corresponds to the “microscopic theory”; the finest description available.
• IR fixed point (b → ∞): The maximally coarse-grained kernel; all fine-grained adjacency is integrated out; the medium appears homogeneous. This fixed point is the “macroscopic theory”; the effective description at long scales.
• Critical points: Kernel flow trajectories that neither converge to UV nor IR correspond to scale-invariant media with power-law correlations and fractal domain structure.

4.4 Relevant, Irrelevant, and Marginal Kernel Deformations

Linearizing the kernel RG flow around a fixed point K*, we write K_b = K* + δK_b and expand:

d(δK_b) / d(log b) = L[δK_b] + O(δK²)      (8)

where L is the linearized beta-function operator at K*. The eigenvalues {λ_i} of L classify deformations of the kernel:

Eigenvalue λi	KFA Classification	Physical Meaning	Macroscopic Fate
λi > 0	Relevant deformation	Grows under coarse-graining	Dominates large-scale medium; long-range correlations persist
λi < 0	Irrelevant deformation	Decays under coarse-graining	Washed out at macroscopic scales; fine-grained detail lost
λi = 0	Marginal deformation	Preserved at fixed point	Logarithmic corrections; anomalous scaling exponents

Table 1. Eigenvalue classification of kernel deformations around a fixed point K*, with physical interpretation at the level of the emergent medium M(b).

This classification precisely mirrors the standard operator classification in the Wilsonian RG [19]. The crucial difference in KFA is that the “operators” being classified are not perturbations of a field theory defined on a fixed background; they are perturbations of the adjacency negotiation itself. The medium on which a putative field theory would be defined is itself the fixed-point kernel transaction, not a given background.

5. Operational Principle III: Connective Tissue across Strata

5.1 The Geometric Imprint of Kernel Composition

Consider a stack of N strata α₁, α₂, …, α_N, with kernel transactions K(α_i, α_i+1) for i = 1, …, N−1. By Axiom K3, the composed kernel from α₁ to α_N is:

K(α₁, α_N) = K(α_N−1, α_N) ∘ K(α_N−2, α_N−1) ∘ … ∘ K(α₁, α₂)      (9)

This composed kernel is generically not equal to any single primitive kernel; it encodes information about the entire sequence of adjacency negotiations along the stack. The deviation of K(α₁, α_N) from a “naive” product (the obstruction to decomposing the composed kernel into independent primitive kernels) is the holonomy of the kernel composition.

Definition 4 (Kernel Holonomy). Let GK denote the group generated by kernel compositions under the law of Axiom K3. The kernel holonomy Hol(γ) associated with a path γ = (α1, α2, …, αN, α1) through the stratum stack (returning to its starting point) is the group element: Hol(γ) = K(αN, α1) ∘ K(αN−1, αN) ∘ … ∘ K(α1, α2) ∈ GK (10) The holonomy records the global consequence of all local kernel negotiations along the path without reference to any global coordinate system.

The theory of holonomy in differential geometry (particularly the Ambrose–Singer theorem [22], which identifies the holonomy group as generated by the curvature two-form integrated over surfaces bounded by the path) provides the precise mathematical framework for understanding K-holonomy. The curvature of the kernel connection is the local measure of non-commutativity; the holonomy is its global integral.

5.2 Non-Commutativity as Source of Coherence

In the commutative case (when [K(α, β), K(β, γ)] = 0 for all adjacent triples) the composed kernel K(α, γ) is path-independent, and the holonomy (10) is trivial: Hol(γ) = 𝟙 for all closed paths γ. In this case, the kernel stack is trivially decomposable; each stratum can be treated independently without loss. There is no global coherence.

In the non-commutative case (the generic situation) [K(α, β), K(β, γ)] ≠ 0, and the holonomy is non-trivial. The composed kernel K(α₁, α_N) depends on the order of kernel transactions, and this dependence encodes the history of adjacency negotiations along the stack. This history is precisely what binds the strata into a coherent whole; it is the connective tissue of the KFA system.

This non-commutativity has precise analogs in several established physical frameworks:

• Berry phase [23]: The geometric phase accumulated by a quantum state transported adiabatically around a closed loop in parameter space is the holonomy of the Berry connection; the non-Abelian connection defined by the adiabatic parallel transport of quantum states. In KFA, Berry phase is the holonomy of a two-stratum kernel transaction parameterized by an external parameter.
• Thomas–Wigner rotation: In special relativity, two successive Lorentz boosts in non-parallel directions produce a net rotation (the Wigner rotation) in addition to the expected boost. This is the holonomy of boost composition in the Lorentz group; the non-commutativity of adjacent boost kernel transactions.
• Pancharatnam phase [24]: The phase accumulated by a polarization state of light cycled through a closed loop on the Poincaré sphere is the holonomy of the polarization transport connection; directly relevant to the optical domain demonstration (§7.1).

5.3 Connective Tissue as Physical Observable

A key claim of KFA (and a point of departure from purely mathematical holonomy theory) is that the kernel holonomy is a physical observable, not merely a mathematical artifact.

Theorem 2 (Holonomy Observability). The kernel holonomy Hol(γ) associated with a closed path γ through the stratum stack is observable by a “test stratum” σ that traverses the path and returns to its starting point: the state of σ after traversal differs from its initial state by the action of Hol(γ). If Hol(γ) = 𝟙, the stack has no global coherence and the test stratum returns unchanged. If Hol(γ) ≠ 𝟙, the stack is globally coherent and the departure of Hol(γ) from the identity is a measurable quantity.

The proof follows from the definition of holonomy and the composability axiom K3: the test stratum’s state is transformed by each kernel transaction it undergoes, and the net transformation after a closed traversal is exactly the holonomy element Hol(γ). The measurability of this transformation follows from the standard quantum measurement postulate applied to the test stratum’s final state.

5.4 Topological Protection of Connective Tissue

Holonomies associated with topologically non-trivial paths (paths whose homotopy class in the stratum space is non-trivial, characterized by winding numbers or Chern classes) are robust against local kernel deformations. No local perturbation of a single kernel transaction K(α_i, α_i+1) can change the topological class of the holonomy Hol(γ); only a global, non-local perturbation that changes the topology of the path can do so.

This topological stability has profound physical consequences: connective tissue associated with topologically non-trivial kernel configurations is generically stable against perturbations; it cannot be destroyed by local noise in individual kernel transactions. This connects directly to:

• Topological insulators [25]: Whose surface states are protected by topological invariants of the bulk electronic kernel configuration.
• Topological quantum field theory [26, 27]: In which the partition function depends only on topological invariants of the spacetime manifold (the holonomy of the gauge connection) and not on the metric or other local structure.
• Topological photonics: Where the Berry phase holonomy of photonic band structures produces edge states robust against disorder in the photonic medium.

6. Domain Emergence from Kernel Stability

6.1 Domains as Fixed-Point Regions

Definition 5 (KFA Domain). A domain is a contiguous block of strata D = {αi, αi+1, …, αi+L} within which the local kernel flow has converged to a common fixed point K*D. Within a domain, the adjacency negotiation has reached a stable resolution: the kernel transactions between adjacent strata within D are all in the same fixed-point basin of attraction.

Crucially, a domain in KFA is not defined by the homogeneity of the strata themselves; two strata within the same domain may have very different internal structures. The domain is defined by the stability of their mutual kernel transactions. This is a relational, not intrinsic, notion of domain membership.

6.2 Domain Walls as Kernel Instabilities

Between two domains D₁ and D₂ with different fixed-point kernels K*₁ and K*₂, there exists a boundary region (a domain wall) within which the local kernel flow transitions between the two fixed-point basins. At the domain wall, no stable resolution exists at the local scale: the kernel transaction is in active flow, transitioning between two stable negotiations.

Domain walls are therefore characterized by:

• Maximal kernel activity: The rate of change of the kernel under resolution change, |dK/d(log b)|, is maximal at the domain wall.
• Minimal local stability: The kernel at the domain wall is farthest from any fixed point; small perturbations can drive the local kernel toward either neighboring fixed point.
• Maximal holonomy gradient: The holonomy changes most rapidly as a function of position when traversing a domain wall; the domain wall is the locus of maximum holonomy density.

6.3 Near-Critical Stacks and Power-Law Domain Structure

At the critical point of the kernel RG flow (where the system sits at neither UV nor IR fixed point but on the separatrix between them) domain sizes follow a power-law distribution:

P(L) ~ L^−τ,   L → ∞      (11)

where τ is determined by the relevant eigenvalue λ₁ of the linearized RG at the critical kernel K*_crit via τ = 1 + 1/λ₁. This power-law distribution is the signature of scale-invariant kernel negotiations: no single resolution scale dominates, the connective tissue spans all scales simultaneously, and domain boundaries at all scales contribute equally (in log-scale) to the global structure of the kernel stack.

6.4 Observer-Dependent Domains

A fundamental KFA prediction is that domain structure is observer-dependent: the domains visible in a stratum stack depend on the resolution scale b at which the observer enacts their kernel negotiations.

• A fine-resolution observer (small b) sees many small domains with sharp, well-defined walls.
• A coarse-resolution observer (large b) sees fewer, larger domains with diffuse boundaries; the fine-grained domain walls have been integrated out by coarse-graining.
• An observer at the critical resolution b_crit (if it exists) sees a self-similar domain structure with no preferred scale; a fractal domain landscape.

Both fine- and coarse-resolution observers are correct within their respective resolution regimes. Neither domain structure is more “real” than the other. This observer-dependence of domain structure is not a limitation of KFA but a precise, falsifiable prediction: measurements of domain statistics at different resolution scales should recover the kernel RG flow trajectory through the domain-size exponent τ(b).

7. Three Domain Demonstrations

7.1 Stratified Optical Media (Primary Formal Elaboration)

The full mathematical development of KFA in the context of stratified optical media (including exact derivations of relational optical Hamiltonians from transfer matrices, the mapping to gyroscope precession on the Poincaré sphere, and the derivation of the RG coarse-graining flow for photonic multilayer stacks) is provided in the companion paper [28]. Here we summarize the KFA mapping in this domain.

In a stratified optical medium with N layers, each layer i is characterized by a refractive index n_i and thickness d_i. The boundary transaction between adjacent layers is encoded by the transfer matrix:

M_i = exp(i H_i d_i)      (12)

where H_i is the single-layer Hamiltonian derived from Maxwell’s equations. In KFA, M_i = K(α_i, α_i+1); the transfer matrix is the kernel of the i-th layer transaction. The five KFA elements map as follows:

KFA Element	Optical Realization
Stratum αi	Optical layer i (characterized by ni, di)
Kernel K(αi, αi+1)	Transfer matrix Mi = exp(i Hi di)
Seed of Relation (Hrel)	Relational optical Hamiltonian via Page–Wootters mechanism
Negotiation / Emergent Medium	RG coarse-graining map Rb on transfer matrix stack; effective medium at scale b
Connective Tissue (Holonomy)	Berry / Pancharatnam phase on Poincaré sphere; gyroscope precession angle
Domain Emergence	Photonic domain structure; RG fixed-point attractors; power-law domain size distribution

Table 2. KFA elements and their realizations in stratified optical media. Full derivations in Costello et al. [28].

The companion paper [28] derives exact experimental predictions in this domain: Berry phase measurements accessible by ellipsometry, power-law domain size distributions observable in disordered photonic films, and gyroscope precession mapping of the polarization holonomy on the Poincaré sphere. The optical domain provides the fullest current elaboration of KFA because the transfer matrix formalism gives exact, closed-form kernel compositions and the geometric phase is directly measurable by standard polarimetric techniques.

7.2 Relational Quantum Gravity

KFA provides a natural reframing of the foundational problems of quantum gravity. The KFA mapping onto loop quantum gravity [1] and causal dynamical triangulation [3] is:

KFA Element	LQG Realization	CDT Realization
Stratum α	Spin-network node	Simplicial building block (4-simplex)
Kernel K(α, β)	Spin-network edge label (SU(2) group element)	Simplicial gluing condition
Seed of Relation	Area operator / edge holonomy (relational geometry generator)	Deficit angle (curvature seed)
Emergent Medium	Semiclassical smooth manifold (IR fixed point)	de Sitter-like continuum limit
Connective Tissue	Ashtekar–Barbero connection holonomy	Global causal structure holonomy
Domain Emergence	Classical spacetime regions vs. quantum foam	Smooth phase vs. crumpled phase

Table 3. KFA elements and their realizations in loop quantum gravity and causal dynamical triangulation [1, 2, 3, 4].

The central KFA reframing of quantum gravity is this: the emergence of classical spacetime is the emergence of an effective medium from kernel negotiations; the spacetime manifold is not primitive but is the infrared fixed point of the quantum gravity kernel RG flow. This reframing dissolves the background-independence problem: there is no background, only kernel transactions. The background (smooth spacetime) emerges as the IR fixed point, and asking “what dynamics live in this background?” is replaced by “what kernel negotiations produce this effective medium?”; a question with a systematic answer.

The connective tissue of the quantum gravity stratum stack (the Ashtekar–Barbero connection holonomy [1]) is precisely the quantity computed in spin-foam models and is the input to the LQG partition function. KFA reinterprets this holonomy as the physical observable binding spacetime building blocks into a coherent geometry, in exact parallel with the Berry phase binding optical layers into a coherent photonic medium [28].

The problem of time [4] is resolved in KFA by Principle I: all dynamics are relational, derived from kernel transactions K(α, β) between “clock” and “system” strata. There is no need for an external time parameter because H_rel(α|β) generates relational evolution without one. The Dirac constraint H_total = 0 is not a pathology but the correct statement that no stratum has absolute dynamics; only kernel-mediated relational dynamics, exactly as Page and Wootters [12] and Giovannetti, Lloyd, and Maccone [14] established in the quantum reference frame context.

7.3 Biological Morphogenesis and Tissue Stratification

The KFA framework maps naturally onto the fundamental processes of developmental biology, reframing the morphogenetic field problem in terms of kernel transactions between adjacent tissue layers:

KFA Element	Biological Realization
Stratum α	Tissue layer (ectoderm, mesoderm, endoderm, derivatives)
Kernel K(α, β)	Cell-cell signaling molecules: morphogens (Wnt, Shh, BMP), cadherins, gap junctions
Seed of Relation	Morphogenetic field generator: differential gene expression induced in one tissue by kernel transactions with its neighbor
Emergent Medium	Tissue-level effective properties: stiffness, signaling competence, emergent from cellular-scale kernel negotiations
Connective Tissue	Extracellular matrix and fascial network: physical holonomy of cell-cell kernel transactions
Domain Emergence	Organ compartmentalization (organogenesis): stabilized kernel fixed-point regions
Domain Walls	Tissue boundaries, basement membranes: regions of maximal morphogenetic kernel activity

Table 4. KFA elements and their realizations in developmental biology and morphogenesis [5, 6, 7, 8].

The KFA interpretation of Turing’s reaction-diffusion mechanism [29] is particularly illuminating. In the standard Turing framework, an activator-inhibitor pair of morphogens diffuses in a pre-given spatial domain, producing spatial patterns. In KFA, the “spatial domain” is itself an emergent medium; the output of cell-cell kernel negotiations. The reaction-diffusion dynamics are not imposed on a fixed space but emerge from kernel transactions between adjacent cell populations. The spatial patterns are the RG fixed-point structure of the morphogenetic kernel flow.

KFA makes a specific prediction about tissue boundaries. Domain walls (boundaries between tissue compartments with different stable kernel configurations) should correspond to regions of maximal morphogenetic kernel activity; maximal cell-cell signaling across the boundary. This is consistent with known developmental biology at boundary zones: the Zone of Polarizing Activity (ZPA) at the posterior limb boundary, the isthmic organizer at the midbrain-hindbrain boundary, and the apical ectodermal ridge (AER) at the dorso-ventral limb boundary are all characterized by intense, spatially focused signaling activity precisely at the tissue domain wall. In KFA, this is not a coincidence but a structural necessity: the domain wall is the locus of maximally active, unstabilized kernel transactions; the boundary between two stable kernel fixed-point attractors.

The extracellular matrix (ECM) and fascial network (the physical material that pervades and connects all tissue layers) is identified in KFA as the physical holonomy of all cell-cell kernel transactions accumulated during development. The ECM is not merely a structural scaffold; it is the material record of the history of morphogenetic kernel negotiations, encoding in its composition and organization the entire developmental history of the kernel transactions that produced the organism. This gives the ECM a role analogous to the Berry phase holonomy in optics: it is the connective tissue that binds the organism into a globally coherent whole, and its topological properties (the closed loops of collagen fibers, the connectivity of gap junction networks) are topologically protected against local perturbations in individual cell-cell kernel transactions.

8. Implications and Open Questions

8.1 The Ontological Inversion: Relation Before Substance

KFA is committed to a radical ontological priority: relation precedes substance. Strata (the “things”) are derived from kernels; the “relations between things.” This formally develops relational ontologies proposed by Leibniz [15], Whitehead [16], and more recently Rovelli [17] and Smolin [30], but made precise through the kernel axioms and the RG machinery of §4.

The philosophical question that immediately arises is: what relates two strata if the strata are themselves defined by their kernel relations? Is there not a vicious circularity; an infinite regress in which kernels require strata, which require kernels, which require strata? The KFA answer is clear and follows from Theorem 1: there is no circularity because the kernel is the primitive and the stratum is derived from it. The kernel is not a relation between two independently existing strata; it is a primitive act of relating from which strata emerge as equivalence classes. The apparent circularity arises only if one assumes that “relation” requires two pre-existing relata; an assumption that KFA explicitly rejects.

¹ This is precisely Whitehead’s move in Process and Reality [16]: “actual occasions” (events, processes) are primitive, and “actual entities” (substances, things) are derivative patterns of process. KFA gives this move a precise mathematical form through the kernel axioms and the equivalence class construction of Theorem 1.

This ontological inversion has implications beyond foundational theory. It suggests that the concept of “system” (a bounded, self-subsistent thing with intrinsic properties) is always a derived, resolution-dependent concept. What counts as a “system” depends on which kernel transactions are being enacted and at what resolution. At fine resolution, the “system” dissolves into a network of interstitial kernel transactions. At coarse resolution, it appears as a well-defined entity with stable properties. The system is always a medium emergent from kernel negotiations; never a primitive.

8.2 Experimental Signatures of KFA

KFA is not merely a philosophical framework; it makes specific, testable predictions across all three domain demonstrations:

KFA Prediction	Optical Domain	Quantum Gravity	Biological
Power-law domain size distribution P(L) ~ L−τ	Disordered photonic films near RG critical point	Planck-scale foam correlation functions	Near-boundary tissue zone organization
Non-trivial holonomy (geometric phase)	Berry/Pancharatnam phase in polarimetry; gyroscope precession on Poincaré sphere	Aharonov–Bohm-like effects in LQG; loop amplitude non-triviality	Chiral symmetry breaking in morphogenesis; ECM chirality
Resolution-dependent effective parameters following kernel beta-function	Ellipsometry at multiple wavelengths revealing scale-dependent effective refractive index	Graviton scattering amplitude flow under energy scale	AFM stiffness measurements at different spatial scales; scale-dependent tissue moduli

Table 5. Testable predictions of KFA across the three domain demonstrations, organized by KFA prediction type.

8.3 Toward a General KFA Calculus

The three operational principles and the domain demonstrations point toward a complete, rigorous KFA calculus. Three mathematical frameworks are natural targets:

Category-Theoretic Formulation. In the language of category theory [31, 32], KFA is a functor: strata are objects, kernels are morphisms, and kernel composition (Axiom K3) is the categorical composition law. The kernel axioms K1–K4 correspond to specific structural requirements on the morphism category: irreducibility (K1) corresponds to the morphisms being non-decomposable, asymmetry (K2) to the morphisms not being isomorphisms (or their inverses), composability (K3) to the standard associativity of categorical composition, and resolution-dependence (K4) to the existence of a coarse-graining functor between kernel categories at different scales. Appendix B develops this sketch.

Sheaf-Theoretic Formulation. The collection of local kernel transactions {K(α_i, α_i+1)} can be organized as a sheaf over the stratum “space” (itself defined by kernel relations). Global sections of this sheaf correspond to globally coherent connective tissue; holonomies that are consistent across all local kernel transactions. Obstructions to global sections (measured by sheaf cohomology) correspond to topologically non-trivial holonomies and topologically protected connective tissue (§5.4).

Topological Field Theory Formulation. In a full KFA topological field theory (following Atiyah [26] and Witten [27]), the partition function Z_KFA is defined as a sum over all kernel configurations consistent with the boundary conditions imposed by the stratum stack:

Z_KFA = ∫ DK exp(−S_K[K]) Hol(γ)_top (13)

where S_K[K] is the kernel action (a functional of the kernel configuration) and Hol(γ)_top is the topologically protected holonomy of the kernel configuration. This partition function would make KFA amenable to the full machinery of modern mathematical physics; surgery axioms, gluing formulas, and topological invariant computation.

8.4 KFA and Information Theory

The identification of the kernel with a quantum channel (a completely positive trace-preserving (CPTP) map [10, 11]) gives KFA a natural information-theoretic interpretation. The kernel K(α, β) : V_α → V_β maps density matrices on V_α to density matrices on V_β, exactly as a quantum channel maps input states to output states.

Under this identification:

• The kernel capacity C(K) is the quantum channel capacity of K(α, β): the maximum rate at which quantum information can be reliably transmitted through the adjacency negotiation between α and β.
• Relevant kernel deformations (λ_i > 0 in the RG classification) correspond to high-capacity kernel modes that survive coarse-graining: they are the information-theoretically robust modes of the adjacency negotiation.
• Irrelevant kernel deformations (λ_i < 0) correspond to low-capacity modes washed out by coarse-graining: they carry information that is not robustly transmissible across the stratum boundary at large scales.
• The kernel holonomy Hol(γ) is the channel capacity of the composed multi-layer kernel transaction: the information that can be reliably transmitted through a sequence of adjacency negotiations.

The RG flow is therefore simultaneously a flow of the adjacency negotiation (kernel) and a flow of the channel capacity: relevant kernel modes are channels through which information can be reliably transmitted at large scales, and irrelevant modes are channels that lose their information-carrying capacity as the resolution coarsens. This gives KFA a precise information-theoretic interpretation: the emergent medium M(b) at resolution b is the medium through which the high-capacity kernel modes (the information-theoretically relevant adjacency negotiations) are enacted.

9. Discussion and Outlook

We have introduced and developed Kernel-First Architecture; a first-principles foundational framework in which all structure is derived from the single primitive of the kernel: the irreducible act of relation between adjacent strata. The three operational principles (Seed of Relation, Negotiation of Adjacency through Resolution, and Connective Tissue across Strata) constitute a complete formal program: from the kernel, one derives the relational state space and Hamiltonian (Principle I), the emergent medium and its RG flow (Principle II), and the global coherence of multi-stratum systems via holonomy (Principle III).

KFA is not a theory of a specific domain. It is a meta-architecture; a framework from which domain-specific theories are derived by specifying the kernel appropriate to each domain. The companion optical paper [28] provides the fullest current formal elaboration. The quantum gravity and biological demonstrations in §7.2 and §7.3 establish the scope: KFA applies wherever multi-stratum systems are characterized by adjacency negotiations that produce emergent media.

The path toward a complete KFA calculus runs through three mathematical programs identified in §8.3: the category-theoretic formulation, the sheaf-theoretic formulation, and the topological field theory formulation. Of these, the category-theoretic formulation is the most immediately tractable: the KFA axioms already have precise categorical expressions (Appendix B), and the machinery of natural transformations and adjoint functors provides the tools for studying how kernel structures transform under domain-to-domain mappings. A natural transformation between two KFA functors (one for the optical domain, one for quantum gravity) would constitute a precise mathematical statement of the structural analogy between Berry phase in optics and Ashtekar connection holonomy in gravity.

A topological field theory formulation (equation 13) would bring KFA into contact with the powerful computational machinery of quantum field theory (Feynman diagrams, operator product expansions, conformal field theory) and would make KFA predictions about topologically protected holonomies computable in closed form. This is the most ambitious target but also the most mathematically natural one, given the identification of KFA holonomy with the Chern–Simons invariant in the gauge theory context.

The key philosophical point of KFA deserves emphasis. KFA does not merely solve technical problems: background independence, the morphogenetic field problem, channel capacity derivation. It dissolves them: it shows that the medium through which these problems are stated is itself a derived structure, not a primitive. When the kernel is taken as primitive, the problems transform: instead of asking “what dynamics live in this medium?” one asks “what kernel negotiations produce this medium?” The former question is unanswerable within any theory that presupposes the medium; the latter has a systematic, constructive answer within KFA. The dissolution of a problem through the choice of a new primitive is the deepest form of theoretical progress; it renders the old question not merely answered but unmotivated.

10. Conclusion

We have presented Kernel-First Architecture: a foundational theory in which all structure, physical, biological, or informational, is derived from a single primitive: the kernel, the irreducible act of relation between adjacent strata. The framework is built on three operational principles stated precisely as follows:

1. Principle I (Seed of Relation): The kernel K(α, β) is the minimal relational unit; it determines the relational state space V_β|α, the relational Hamiltonian H_rel(α|β), and the observable algebra O(β|α) of stratum β as seen from stratum α. No stratum has intrinsic dynamics; all dynamics are kernel-mediated and observer-dependent in the structural sense encoded by Axiom K2.
2. Principle II (Negotiation of Adjacency through Resolution): At each resolution scale b, block-composition of primitive kernel transactions produces a new effective medium M(b). The renormalization group flow of the kernel classifies the stability of adjacency negotiations: relevant deformations produce long-range correlations visible at macroscopic scales, irrelevant deformations are washed out, and critical points produce scale-invariant media with power-law domain structure. The medium is always an output; never an input.
3. Principle III (Connective Tissue): The holonomy of kernel compositions along closed paths through the stratum stack is a physical observable; the connective tissue that binds strata into a globally coherent system without reference to any external background. Topologically non-trivial holonomies are protected against local kernel deformations, giving the connective tissue a stability that individual kernel transactions lack.

These three principles resolve the three historical tensions identified in the Introduction. The background-independence problem in quantum gravity is dissolved by Principle I: dynamics are relational, derived from kernel transactions, without any background requirement; the spacetime background is the IR fixed point of the quantum gravity kernel RG flow (Principle II), not a primitive presupposition. The morphogenetic field problem in developmental biology is dissolved by Principle II: the spatial domain of morphogenetic signaling is not given but is the effective medium produced by cell-cell kernel negotiations; the tissue creates its own geometry. The channel capacity problem in information theory is dissolved by Principle III and the information-theoretic interpretation of §8.4: the channel is the kernel, its capacity is the kernel capacity, and the robustness of information transmission is encoded in the holonomy of kernel compositions.

The formal development of KFA in stratified optical media (provided in full in the companion paper [28]) demonstrates that all five KFA elements admit exact mathematical formulations with experimentally testable predictions. The quantum gravity and biological demonstrations establish the framework’s scope. The category-theoretic, sheaf-theoretic, and topological field theory programs identified in §8.3 constitute the agenda for the full rigorous development of KFA as a mathematical discipline.

We close with the central thesis, stated as plainly as possible: The medium is never given. It is always negotiated. Every background, every space, every channel, every tissue; all are derived from kernel transactions enacted at some resolution scale by some relational vantage. The program of KFA is to make this derivation explicit, systematic, and predictive; to replace the question “what happens in this world?” with the prior question “what kernel negotiations produced this world?” The answer, in each domain examined, is the same: structure emerges from relation, and relation (the kernel) is the only irreducible primitive.

Appendix A: Kernel Axioms: Complete Formal Statement

Let {V_α}_α∈A be a collection of Hilbert spaces indexed by strata in A. Let Hom(V_α, V_β) denote the space of bounded linear operators from V_α to V_β.

Definition A1 (Kernel System). A kernel system (A, {Vα}, {K(α,β)}) consists of: •  An adjacency relation ≺ on A (we write α ≺ β to mean α and β are adjacent strata). •  For each adjacent pair (α, β) with α ≺ β, a kernel K(α, β) ∈ Hom(Vα, Vβ). •  A resolution parameter b ∈ ℝ+ indexing coarse-graining scale.

The four kernel axioms are stated in full:

K1 (Irreducibility): For each adjacent pair (α, β), K(α, β) cannot be written as a tensor product: ∄ f ∈ Hom(V_α, W), g ∈ Hom(V_β, W) for any W such that K(α, β) = g^† f. Equivalently, the Schmidt rank of K(α, β) viewed as a bipartite operator is strictly greater than 1.

K2 (Asymmetry): K(β, α) ≠ K(α, β)^† in general; the kernel transaction from β to α is not the Hermitian adjoint of the transaction from α to β. Equality K(β, α) = K(α, β)^† is a special case (the “reversible” kernel) that applies only to time-reversal-symmetric systems.

K3 (Composability): For any adjacent triple (α, β, γ) with α ≺ β ≺ γ, the composed kernel K(α, γ) ∈ Hom(V_α, V_γ) is defined by K(α, γ) = K(β, γ) ∘ K(α, β). Composition is associative: for any quadruple (α, β, γ, δ), [K(γ,δ) ∘ K(β,γ)] ∘ K(α,β) = K(γ,δ) ∘ [K(β,γ) ∘ K(α,β)].

K4 (Resolution-Dependence): There exists a one-parameter family of coarse-graining operators {R_s}_s≥1 acting on kernel space such that K_b’(α, β) = R_b’/b[K_b(α, β)] for all b’ ≥ b, satisfying: (i) R₁ = Id, (ii) R_s ∘ R_t = R_st (semigroup law), (iii) the map s ↦ R_s[K] is continuous for each K.

Proposition A1 (Kernel Composition Associativity). The composition operation of Axiom K3 is associative. Proof: Directly from the associativity of operator composition in Hom(Vα, Vβ). ∎

Proposition A2 (Kernels Form a Category). The kernel system (A, {Vα}, {K(α,β)}) defines a category K in which: (i) objects are strata α ∈ A; (ii) morphisms HomK(α, β) = {K(α,β)} for adjacent pairs, and compositions of such for non-adjacent pairs; (iii) composition is defined by Axiom K3; (iv) identity morphisms exist if and only if there exist identity kernels K(α,α) = IdVα (trivial self-relation). Proof: Associativity from Proposition A1; identity from hypothesis. ∎

Appendix B: Category-Theoretic Formulation Sketch

Following MacLane [31] and Lawvere [32], we sketch the category-theoretic formulation of KFA.

The KFA Category K. Objects of K are strata; morphisms are kernels; composition is kernel transaction composition (Axiom K3); identities are trivial self-relation kernels. By Proposition A2, K is a well-defined category.

Additional Structure: Dagger Category. For quantum systems, where kernels are maps between Hilbert spaces, K carries the structure of a dagger-category (†-category): each morphism K(α,β) : V_α → V_β has a dagger K(α,β)^† : V_β → V_α defined by the Hilbert space adjoint. Axiom K2 states that the dagger is not the same as the reverse kernel: K(β,α) ≠ K(α,β)^† in general, distinguishing KFA from self-dual (†-preserving) frameworks.

Coarse-Graining as a Functor. The coarse-graining operator R_s of Axiom K4 defines a functor F_s : K_b → K_sb between kernel categories at different resolution scales. The semigroup law R_s ∘ R_t = R_st corresponds to the functor composition F_s ∘ F_t = F_st. The RG flow (§4.3) is the one-parameter group of autofunctors of K generated by the coarse-graining functor.

Natural Transformations as Cross-Domain Maps. A natural transformation η : F_optical ⇒ F_gravity between the optical domain KFA functor and the quantum gravity KFA functor would constitute a mathematically precise statement of the structural analogy between the two domains; making rigorous the claim that the Berry phase holonomy in optics and the Ashtekar connection holonomy in gravity are both instances of the same KFA Principle III. Constructing this natural transformation explicitly is an open problem of high priority.

Sheaf Structure. The assignment of kernel K(α,β) to each adjacent pair (α,β) defines a presheaf over the nerve of the adjacency relation. The condition that locally consistent kernel assignments extend to globally consistent ones is a sheaf condition; its cohomological obstruction H¹(A, K) measures the topologically non-trivial holonomies of the KFA system.

Appendix C: Connection to Known Formalisms

KFA Element	Standard Formalism	KFA ↔ Standard Translation	Key Reference
Kernel K(α,β)	Page–Wootters clock-system entanglement	K encodes the timeless entangled state; Hrel(α|β) derived from K via eq. (1)	[12, 13, 14]
Seed of Relation (Hrel)	Dirac constraint Htotal = 0	Constraint ensures all dynamics are kernel-mediated; Hrel extracted by conditioning	[4, 12]
Kernel composition (K3)	Transfer matrix product (optics)	M1M2…MN = K(α1,αN); each Mi is a primitive kernel	[28]
Kernel holonomy Hol(γ)	Berry phase / geometric phase	Hol(γ) = exp(i∮ A · dR); Berry connection A derived from kernel transport	[23, 24]
Kernel RG flow βK	Wilsonian / Polchinski RG	dK/d(log b) = βK(K) generalizes Wilson’s block-spin to kernel space	[19, 20, 21]
Emergent medium M(b)	Effective field theory (IR limit)	M(∞) = IR fixed-point EFT; M(1) = UV microscopic theory	[19, 21]
Stratum adjacency (K3)	Spin-network edge labels (LQG)	SU(2) group element on edge = K(α,β); area operator = Hrel	[1, 2]
Connective tissue holonomy	Ashtekar–Barbero connection holonomy	Hol(γ) around spin-network loop = Wilson loop = LQG area spectrum	[1]
Kernel K(α,β) (info. theory)	Quantum channel (CPTP map)	K : Vα → Vβ is a CPTP map; kernel capacity = channel capacity	[10, 11]
Morphogenetic kernel	Turing reaction-diffusion / morphogenetic field	Activator-inhibitor system = kernel transaction between adjacent cell populations; spatial pattern = RG fixed-point structure	[5, 6, 7, 8, 29]
KFA partition function (eq. 13)	Atiyah–Witten TQFT	ZKFA = sum over kernel configurations; Holtop = topological invariant (Chern–Simons)	[26, 27]
KFA category K	MacLane categories; Lawvere functorial semantics	Strata = objects; kernels = morphisms; composition = K3; RG = autofunctor	[31, 32]

Table C1. Correspondence table mapping KFA elements to established formalisms across physics, information theory, and mathematics.

References

[1] A. Ashtekar, “New variables for classical and quantum gravity,” Physical Review Letters, vol. 57, no. 18, pp. 2244–2247, 1986.

[2] T. Thiemann, Modern Canonical Quantum General Relativity. Cambridge: Cambridge University Press, 2007.

[3] J. Ambjorn, J. Jurkiewicz, and R. Loll, “Nonperturbative Lorentzian path integral for gravity,” Physical Review Letters, vol. 85, no. 5, pp. 924–927, 2000.

[4] K. V. Kuchař, “Time and interpretations of quantum gravity,” in Proceedings of the 4th Canadian Conference on General Relativity and Relativistic Astrophysics, G. Kunstatter, D. Vincent, and J. Williams, Eds. Singapore: World Scientific, 1992, pp. 211–314.

[5] L. Wolpert, “Positional information and the spatial pattern of cellular differentiation,” Journal of Theoretical Biology, vol. 25, no. 1, pp. 1–47, 1969.

[6] A. Gierer and H. Meinhardt, “A theory of biological pattern formation,” Kybernetik, vol. 12, no. 1, pp. 30–39, 1972.

[7] G. Holló and M. Novák, “The manoeuvrability of complex pattern formation in reaction-diffusion systems and beyond,” Journal of The Royal Society Interface, vol. 9, no. 76, pp. 2614–2626, 2012.

[8] L. Wolpert, C. Tickle, and A. Martinez Arias, Principles of Development, 5th ed. Oxford: Oxford University Press, 2015.

[9] C. E. Shannon, “A mathematical theory of communication,” Bell System Technical Journal, vol. 27, pp. 379–423, 623–656, 1948.

[10] M. M. Wilde, Quantum Information Theory, 2nd ed. Cambridge: Cambridge University Press, 2017.

[11] J. Watrous, The Theory of Quantum Information. Cambridge: Cambridge University Press, 2018.

[12] D. N. Page and W. K. Wootters, “Evolution without evolution: Dynamics described by stationary observables,” Physical Review D, vol. 27, no. 12, pp. 2885–2892, 1983.

[13] R. Gambini and R. A. Porto, “Relational time in generally covariant quantum systems: Four models,” Physical Review D, vol. 63, no. 10, p. 105014, 2001.

[14] V. Giovannetti, S. Lloyd, and L. Maccone, “Quantum time,” Physical Review D, vol. 92, no. 4, p. 045033, 2015.

[15] G. W. Leibniz, The Monadology, 1714. Trans. R. Latta. Oxford: Oxford University Press, 1898.

[16] A. N. Whitehead, Process and Reality: An Essay in Cosmology. New York: Free Press, 1929 (corrected edition, 1978).

[17] C. Rovelli, “Relational quantum mechanics,” International Journal of Theoretical Physics, vol. 35, no. 8, pp. 1637–1678, 1996.

[18] C. A. Fuchs, N. D. Mermin, and R. Schack, “An introduction to QBism with an application to the locality of quantum mechanics,” American Journal of Physics, vol. 82, no. 8, pp. 749–754, 2014.

[19] K. G. Wilson and J. Kogut, “The renormalization group and the ε expansion,” Physics Reports, vol. 12, no. 2, pp. 75–199, 1974.

[20] K. G. Wilson, “The renormalization group and critical phenomena,” Reviews of Modern Physics, vol. 55, no. 3, pp. 583–600, 1983.

[21] J. Polchinski, “Renormalization and effective Lagrangians,” Nuclear Physics B, vol. 231, no. 2, pp. 269–295, 1984.

[22] W. Ambrose and I. M. Singer, “A theorem on holonomy,” Transactions of the American Mathematical Society, vol. 75, no. 3, pp. 428–443, 1953.

[23] M. V. Berry, “Quantal phase factors accompanying adiabatic changes,” Proceedings of the Royal Society A, vol. 392, no. 1802, pp. 45–57, 1984.

[24] S. Pancharatnam, “Generalized theory of interference, and its applications,” Proceedings of the Indian Academy of Sciences Section A, vol. 44, no. 5, pp. 247–262, 1956.

[25] M. Z. Hasan and C. L. Kane, “Colloquium: Topological insulators,” Reviews of Modern Physics, vol. 82, no. 4, pp. 3045–3067, 2010.

[26] M. F. Atiyah, “Topological quantum field theory,” Publications Mathématiques de l’IHÉS, vol. 68, pp. 175–186, 1988.

[27] E. Witten, “Quantum field theory and the Jones polynomial,” Communications in Mathematical Physics, vol. 121, no. 3, pp. 351–399, 1989.

[28] D. Costello et al., “Relational Optical Hamiltonians, Gyroscope Precession Mapping, and Renormalization Group Coarse-Graining in Stratified Media: A Kernel-First Architecture,” Foundations of Physics, 2026 (companion paper).

[29] A. M. Turing, “The chemical basis of morphogenesis,” Philosophical Transactions of the Royal Society B, vol. 237, no. 641, pp. 37–72, 1952.

[30] L. Smolin, Three Roads to Quantum Gravity. New York: Basic Books, 2001.

[31] S. Mac Lane, Categories for the Working Mathematician, 2nd ed. New York: Springer, 1998.

[32] F. W. Lawvere, “Functorial semantics of algebraic theories,” Proceedings of the National Academy of Sciences, vol. 50, no. 5, pp. 869–872, 1963.

[33] B. Simon, “Holonomy, the quantum adiabatic theorem, and Berry’s phase,” Physical Review Letters, vol. 51, no. 24, pp. 2167–2170, 1983.

[34] A. Connes, Noncommutative Geometry. San Diego: Academic Press, 1994.