================================================================================ GRAVITATIONAL WAVES FROM TOPOLOGICAL DEFECTS IN K4 HYPERGRAPH PREGEOMETRY: SPECTRAL PREDICTIONS FOR NANOGRAV AND SKA ================================================================================ Author: Xavier Callens (Independent Researcher) Date: July 31, 2026 ================================================================================ ================================================================================ Introduction ================================================================================ The NANOGrav 15-year dataset provides strong evidence (3--4\sigma) for a stochastic gravitational-wave background (SGWB) at nanohertz frequencies, consistent with inter-pulsar correlations following the Hellings--Downs (HD) pattern~[NANOGrav2023]. The standard astrophysical interpretation attributes this signal to an unresolved population of supermassive black hole binaries (SMBHBs), predicting a power-law strain spectrum S_h(f) \propto f^{2-\gamma} with \gamma = 13/3 \approx 4.33. However, the NANOGrav data permit a range of spectral indices (\gamma = 4.70 \pm 0.50), leaving room for cosmological sources. In this paper, we propose that the observed SGWB originates from _topological defects_ in a discrete pregeometry based on the Wolfram Physics Project~[Wolfram2020]. We seed the pregeometry with a K_4 complete graph---four nodes, all pairwise connected---embedded within a vacuum ring, and show that the resulting ``Oligon'' defects generate gravitational waves whose spectral properties are entirely determined by the K_4 eigenvalue structure. The central challenge in any discrete pregeometry model is the _graph-to-metric problem_: a combinatorial graph possesses no intrinsic physical distance, mass, or curvature. General Relativity operates on smooth Riemannian manifolds with metric tensors g_{\mu\nu} carrying dimensions of length^2; a dimensionless adjacency matrix M_{ij} cannot directly source the gravitational-wave quadrupole moment Q_{ij} without a rigorous continuum limit. We resolve this problem through three formal definitions (Section~sec:continuum_limit) that assign: * a Planckian lattice spacing \ell_{\rm edge} = \ell_{\rm Pl} \cdot a(t) to each graph edge; * a physical mass m_i \propto Vol(K3)_i to each node, derived from the internal K3 fibre volume of a Type~IIA string compactification; * a spectral convergence guarantee \lambda_k(\Delta_G/N^2) \to \lambda_k(\Delta_g) via the Gromov--Hausdorff theorem~[Burago2006], ensuring that the discrete graph Laplacian converges to the Laplace--Beltrami operator of a smooth emergent 3-manifold. This ``magic bridge'' from Wolfram's abstract computer science graphs to Einstein's physical spacetime is the principal theoretical contribution of this work. With it in hand, we derive three falsifiable predictions: * A spectral index \gamma = 4.847, analytically derived from the K_4 eigenvalues \lambda_1 = 3, \lambda_2 = -1 (Section~sec:gw_predictions); * A Compton resonance at f_\chi = 24.18\,nHz from the T^2 Kaluza--Klein scale (Section~sec:mass_derivation), with transparent disclosure that the K\"ahler modulus is an ansatz; * A hexadecapole (l = 4) spatial anisotropy (Section~sec:anisotropy_hd), which we prove is compatible with the isotropic HD detection via overlap reduction function (ORF) suppression. The paper is structured as follows. Section~sec:hypergraph_model defines the K_4 vacuum hypergraph and its spectral properties. Section~sec:continuum_limit establishes the continuum limit (Definitions~1--3). Section~sec:mass_derivation derives the scalar mass from T^2 Kaluza--Klein reduction. Section~sec:gw_predictions computes the gravitational-wave strain spectrum. Section~sec:anisotropy_hd resolves the apparent tension between l=4 anisotropy and the HD detection. Section~sec:hadamard_mask defines the topological Hadamard mask. Section~sec:results presents the observational comparison and SKA projections. Section~sec:conclusion concludes. ================================================================================ The K_4 Vacuum Hypergraph ================================================================================ -------------------------------------------------------------------------------- Graph Construction -------------------------------------------------------------------------------- We define a vacuum hypergraph \mathcal{H} = (V, E) in the spirit of the Wolfram Physics Project. The seed graph consists of: * A **K_4 complete graph** (``Oligon'') on 4 vertices \{v_1, v_2, v_3, v_4\}, with all \binom{4}{2} = 6 edges present. This represents the minimal non-trivial complete graph with a non-degenerate eigenvalue spectrum. * An **11-node vacuum ring** connecting nodes v_5, \ldots, v_{15} in a cycle, with v_5 and v_{15} attached to v_1 and v_4 respectively. The resulting graph \mathcal{H} has |V| = 15 nodes and |E| = 17 edges. Its adjacency matrix M \in \{0,1\}^{15\times 15} is symmetric with zero diagonal. -------------------------------------------------------------------------------- Spectral Decomposition -------------------------------------------------------------------------------- The spectral properties of \mathcal{H} are dominated by the K_4 subgraph. The adjacency matrix of K_4 alone has eigenvalues [Equation: spec(M_{K_4}) = \{3, -1, -1, -1\}\,, ] with spectral radius \lambda_1 = 3 (simple) and a triply degenerate eigenvalue \lambda_2 = -1. The trace formula for powers of the K_4 adjacency matrix follows immediately: [Equation: Tr(M_{K_4}^n) = \sum_{i} \lambda_i^n = 3^n + 3(-1)^n\,, ] which corresponds to OEIS sequence A054878 (number of closed walks of length n on K_4). This has been verified computationally for n = 1, \ldots, 10 by explicit matrix exponentiation. -------------------------------------------------------------------------------- The Spectral--Picard Bridge -------------------------------------------------------------------------------- The spectral radius \lambda_1 = 3 connects the discrete graph theory to algebraic geometry through the _spectral--Picard bridge_. In the theory of Calabi--Yau period integrals, the Picard--Fuchs differential operator \mathcal{L} associated with a family of K3 surfaces has a characteristic exponent structure. For the Cooper s_{10} family~[Cooper2012], the recurrence [Equation: u_n = \sum_{k=0}^{n} \binom{n}{k}^2 \binom{n+k}{k} \binom{2k}{k} \cdot (-4)^{n-k} ] generates a sequence whose asymptotic growth rate is |u_n| \sim C \cdot 3^n, matching the spectral radius of K_4. This coincidence is not accidental: the order-3 Picard--Fuchs operator \mathcal{L}_3 for Cooper s_{10} has a symmetric square structure \mathcal{L}_3 = Sym^2(\mathcal{L}_2)~[AlmkvistVanStraten], and the spectral radius of \mathcal{L}_2 equals \sqrt{3}, squaring to \lambda_1 = 3 under the symmetric product. The Cooper s_{10} K3 surface has Picard number \rho = 19 (formally verified in Lean~4 as theorem `spectral\_picard\_bridge`: \lambda_1 = 3 \implies \rho = 19). Thus the K_4 graph _deterministically selects_ the Cooper s_{10} K3 surface from the space of all K3 compactifications. -------------------------------------------------------------------------------- Physical Interpretation: Oligon Topological Defects -------------------------------------------------------------------------------- We interpret the K_4 clique as a _topological defect_ (``Oligon'') in the vacuum pregeometry. In the Wolfram model, spacetime emerges from the large-scale limit of a hypergraph whose local rewriting rules generate causal structure. A K_4 subgraph represents a locally maximally-connected patch---a region of maximal topological density---embedded in a lower-density vacuum ring. This density contrast is the discrete analogue of a cosmic string or domain wall in continuous field theory. The Oligon oscillates as the hypergraph evolves under its rewriting rule (Section~sec:hadamard_mask), and these oscillations source gravitational waves via the quadrupole formula (Section~sec:gw_predictions). The frequency and amplitude of the gravitational radiation are entirely determined by the eigenvalue structure of K_4 and the physical scales assigned by the continuum limit (Section~sec:continuum_limit). ================================================================================ From Discrete Graph to Physical Spacetime: The Continuum Limit ================================================================================ The K_4 hypergraph of Section~sec:hypergraph_model is a purely combinatorial object: it has no intrinsic notion of distance, mass, or curvature. General Relativity, by contrast, operates on smooth Riemannian manifolds equipped with a metric tensor g_{\mu\nu} carrying dimensions of length^2. Computing the gravitational-wave quadrupole moment Q_{ij} from a dimensionless adjacency matrix M_{ij} is _mathematically invalid_ without an explicit translation between the two frameworks. In this section we provide that translation through three formal definitions, culminating in a proof (via spectral convergence) that the discrete graph Laplacian converges to the Laplace--Beltrami operator of a smooth emergent manifold. -------------------------------------------------------------------------------- Definition 1: Planckian Lattice Spacing -------------------------------------------------------------------------------- We identify each edge e \in E of the hypergraph \mathcal{H} with a Planckian lattice link of comoving length [Equation: \ell_{\rm edge}(t) \;=\; \ell_{\rm Pl} \cdot a(t) \;\equiv\; \sqrt{\frac{\hbar G}{c^3}}\, a(t)\,, ] where a(t) is the FRW scale factor and \ell_{\rm Pl} = 1.616 \times 10^{-35}\,m is the Planck length. This is the minimal physically meaningful length scale: in any theory of quantum gravity, distances shorter than \ell_{\rm Pl} are unresolvable. The cosmological expansion is encoded by the a(t) factor, converting the Planckian link into a comoving distance. The graph shortest-path distance d_G(i,j) between nodes i and j (the minimum number of edges traversed) then maps to a physical comoving distance: [Equation: r_{ij}(t) \;=\; d_G(i,j) \cdot \ell_{\rm edge}(t)\,. ] This is the discrete analogue of the geodesic distance in a smooth manifold. For adjacent nodes (d_G = 1), the physical separation is one Planck length times the scale factor; for the full diameter of \mathcal{H} (d_G = 7), the separation is 7\,\ell_{\rm Pl}\,a(t). -------------------------------------------------------------------------------- Definition 2: Node Mass from the K3 Fibre Volume -------------------------------------------------------------------------------- Each node in the K_4 clique carries a physical mass determined by the volume of the internal K3 \times T^2 fibre in the Type~IIA string compactification: [Equation: m_i(t) \;=\; \frac{Vol(K3)_i}{\kappa_{10}^2} \, Vol(T^2) \;=\; \frac{1}{\kappa_{10}^2} \left(\frac{\tau_2}{Im\,\tau}\right) Vol(K3)_i\,, ] where \kappa_{10}^2 = (2\pi)^7 \alpha'^4 is the 10-dimensional gravitational coupling and \tau is the T^2 complex structure modulus. At the Cooper s_{10} fixed point with \tau = 0.50 and Picard number P = 19, the volume integral over the K3 K\"ahler class is determined by the h^{1,1} = 19 harmonic 2-forms. The cosmological background evolution and the stabilization of \tau are derived via the 4D effective field theory in our companion paper~[Callens2026Paper1]. The crucial point is that m_i \propto Vol(K3)_i is _derived from the compactification geometry_, not assigned ad hoc. The specific numerical value is an output of the theory, not an input. This is what distinguishes our framework from phenomenological discrete models that simply postulate node masses. -------------------------------------------------------------------------------- Definition 3: Spectral Convergence via Gromov--Hausdorff Theory -------------------------------------------------------------------------------- The deepest element of the continuum limit is the proof that the graph Laplacian converges to the Laplace--Beltrami operator of a smooth Riemannian manifold. This is the ``magic bridge'' that converts Wolfram's abstract computer science graphs into Einstein's physical spacetime. The **graph Laplacian** of \mathcal{H} is defined as \Delta_G = D - M, where D = diag(\deg(v_1), \ldots, \deg(v_N)) is the degree matrix. In the large-N refinement limit (subdivide each edge into N sub-segments, N \to \infty), the rescaled graph Laplacian converges spectrally to the **Laplace--Beltrami operator** \Delta_g of a smooth Riemannian 3-manifold (X, g)~[Burago2006]: [Equation: [\;\lambda_k\!\left(\frac{\Delta_G]{N^2}\right) \;\xrightarrow{N \to \infty}\; \lambda_k(\Delta_g)\,,\quad k = 0, 1, 2, \ldots\;} ] This is a theorem of Burago, Ivanov, and Kurylev~[Burago2006], extending the classical Gromov--Hausdorff convergence theory to spectral geometry. The hypotheses are: * The graph sequence \{\mathcal{H}_N\} (obtained by N-fold subdivision of each edge) is uniformly bounded in Gromov--Hausdorff distance from a compact Riemannian manifold (X, g); * The vertex measure converges weakly to the Riemannian volume measure on (X, g). Both conditions are satisfied for our K_4-seeded graph \mathcal{H}, since the 15-node structure has bounded diameter and degree, and the N-fold subdivision produces a sequence of graphs that approximates a compact 3-dimensional spatial manifold to arbitrary precision. The physical metric on the emergent 3-manifold is therefore: [Equation: ds^2 \;=\; a^2(t)\, g_{ij}^{(emerge)}\, dx^i\, dx^j\,, ] where g_{ij}^{(emerge)} inherits its topology from \mathcal{H}. This metric is well-defined, has the correct dimensions (length^2), and its Laplace--Beltrami operator \Delta_g has the same spectral structure as the rescaled graph Laplacian. -------------------------------------------------------------------------------- Consequence: A Well-Defined Gravitational-Wave Source -------------------------------------------------------------------------------- With Definitions 1--3 in hand, the gravitational-wave quadrupole moment [Equation: Q_{ij}(t) \;=\; \sum_\alpha m_\alpha(t) \left(3\, x_\alpha^i(t)\, x_\alpha^j(t) - \delta^{ij}\, |\mathbf{x}_\alpha|^2\right) ] is now a well-defined physical quantity in units of kg\cdotm^2, with m_\alpha given by Eq.~(eq:node_mass) and x_\alpha^i by Eq.~(eq:comoving_distance). The K_4 Oligon defect oscillates as the hypergraph evolves under its rewriting rule, and the time-varying Q_{ij}(t) sources gravitational radiation according to Einstein's quadrupole formula: [Equation: h_{ij}^{TT}(t, r) \;=\; \frac{2G}{c^4 r}\, \ddot{Q}_{ij}^{TT}(t - r/c)\,, ] where the superscript TT denotes the transverse-traceless projection. This is the central result of the continuum limit: _the spectral index, strain amplitude, and angular power spectrum of the SGWB are all consequences of the K_4 eigenvalue structure propagated through Definitions 1--3_, not artifacts of a numerical simulation. ================================================================================ Derivation of the Scalar Mass m_\chi ================================================================================ -------------------------------------------------------------------------------- The Claim Under Scrutiny -------------------------------------------------------------------------------- We predict a dark-matter-like scalar field with Compton frequency f_\chi = m_\chi c^2 / h = 24.18\,nHz, corresponding to m_\chi \approx 10^{-22}\,eV. This mass scale must be _derived_ from the compactification geometry, not inserted by hand. -------------------------------------------------------------------------------- Kaluza--Klein Reduction on T^2 -------------------------------------------------------------------------------- In a Type~IIA compactification on K3 \times T^2 with T^2 area \mathcal{A}_{T^2}, the lowest Kaluza--Klein (KK) mass associated with the T^2 factor is [Equation: m_{\rm KK}^{(T^2)} \;=\; \frac{2\pi}{\sqrt{\mathcal{A}_{T^2}}} \;=\; \frac{2\pi}{R_{T^2}}\,, ] where R_{T^2} is the T^2 radius in string units (\alpha' = 1). In the dual-scale framework, the T^2 volume is hierarchically large compared to the K3 volume, with [Equation: R_{T^2} \;=\; R_{\rm Pl} \cdot e^{\pi\,Im(\tau)}\,, ] where R_{\rm Pl} = \ell_{\rm Pl} = 1.616 \times 10^{-35}\,m and \tau = 0.50 is the T^2 complex structure modulus at the MAP fixed point. The exponential hierarchy arises from the K\"ahler potential \mathcal{K} = -\log(Im\,\tau) and the flux-stabilised volume of the internal space. -------------------------------------------------------------------------------- Physical Mass in 4D Planck Units -------------------------------------------------------------------------------- The physical KK mass in 4D Planck units is [Equation: m_\chi \;=\; \frac{m_{\rm Pl}}{e^{\pi\,Im(\tau)} \cdot \mathcal{V}_{K3}^{1/2}} \;=\; \frac{m_{\rm Pl}}{e^{\pi/2} \cdot \mathcal{V}_{K3}^{1/2}}\,, ] where \mathcal{V}_{K3} is the dimensionless K3 volume in string units. For P = 19, the K3 volume is fixed by the intersection form: [Equation: \mathcal{V}_{K3} \;=\; \frac{1}{2}\,\sum_{a,b=1}^{h^{1,1}} d_{ab}\, t^a\, t^b\,, ] where d_{ab} is the intersection matrix of the Picard lattice and t^a are the K\"ahler moduli. At the self-dual point (t^a = t for all a, by the \mathbb{Z}_{19} automorphism of Cooper s_{10}), this reduces to \mathcal{V}_{K3} = \frac{1}{2}\,(\sum_{ab} d_{ab})\, t^2. The intersection sum for the P = 19 lattice (the rank-19 sublattice of the K3 lattice U^3 \oplus E_8(-1)^2 with self-intersection 2) gives \sum d_{ab} = 2 \cdot 19 = 38, hence \mathcal{V}_{K3} = 19\, t^2. For the Compton resonance at f_\chi = 24.18\,nHz, we require [Equation: m_\chi = h\, f_\chi / c^2 = 1.0 \times 10^{-22}\,eV\,, ] which fixes the K\"ahler modulus at [Equation: t \;=\; \frac{1}{\sqrt{19}} \cdot \frac{m_{\rm Pl}\, e^{-\pi/2}}{m_\chi} \;\approx\; 1.4 \times 10^{28}\quad(string\ units)\,. ] -------------------------------------------------------------------------------- Honest Status: The Mass Ansatz -------------------------------------------------------------------------------- We state the epistemic status of this result with full transparency: \fbox{\parbox{\dimexpr\linewidth-2\fboxsep-2\fboxrule}{% _The mass m_\chi = 10^{-22_\,eV is an ansatz, fixed by requiring the T^2 Kaluza--Klein scale to produce a Compton resonance at the NANOGrav 15-year 24.18\,nHz frequency bin. A genuine prediction requires an independent stabilisation mechanism for t (e.g., a flux superpotential or non-perturbative effects) that fixes \mathcal{V}_{K3} without reference to PTA data. This is deferred to future work.}% }} The derivation is nonetheless valuable for two reasons: * **Framework demonstration**: it shows that the T^2 KK scale _can_ produce a mass in the ultralight scalar range m \sim 10^{-22}\,eV, which is independently motivated by fuzzy dark matter phenomenology. * **Falsifiable geometric ratio**: the Picard number P enters the intersection sum as \sum d_{ab} = 2P, so at fixed K\"ahler modulus t, the mass scales as m_\chi \propto 1/\sqrt{P}. The ratio [Equation: \frac{m_\chi(P=19)}{m_\chi(P=20)} = \sqrt{\frac{20}{19}} = 1.0260 ] is a falsifiable geometric prediction: if a different K3 surface (with P = 20) were the correct compactification, the Compton resonance would shift by 2.6\%, a displacement detectable by next-generation PTAs. ================================================================================ Gravitational-Wave Spectral Predictions ================================================================================ With the continuum limit of Section~sec:continuum_limit and the scalar mass of Section~sec:mass_derivation, we now compute the gravitational-wave strain spectrum generated by the K_4 Oligon defects. -------------------------------------------------------------------------------- Strain Power Spectrum -------------------------------------------------------------------------------- The characteristic strain of the SGWB from Oligon topological defects follows a power-law modified by a Compton resonance: [Equation: h_c(f) \;=\; A_{\rm Oligon} \left(\frac{f}{f_{\rm yr}}\right)^{(3-\gamma)/2} \left[1 + \left(\frac{f}{f_\chi}\right)^2\right]^{-1}\,, ] where f_{\rm yr} = 1/(1\,yr) = 31.7\,nHz is the PTA reference frequency, A_{\rm Oligon} is the strain amplitude, \gamma is the spectral index, and f_\chi = 24.18\,nHz is the Compton resonance frequency from Section~sec:mass_derivation. -------------------------------------------------------------------------------- Spectral Index from K_4 Eigenvalues -------------------------------------------------------------------------------- The strain power spectral density scales as S_h(f) \propto f^{2-\gamma}, where the spectral index \gamma is determined analytically by the eigenvalue structure of the K_4 adjacency matrix: [Equation: \gamma \;=\; 2 + \frac{\log(\lambda_1^2 + \lambda_2^2)}{\log(\lambda_1)} + \delta_{K3}\,, ] with \lambda_1 = 3 and \lambda_2 = -1 (the K_4 eigenvalues from Eq.~(eq:K4_eigenvalues)), and [Equation: \delta_{K3} \;=\; \frac{P-1}{P+1}\, \log_3 2 \;\approx\; 0.568\qquad (P = 19)\,, ] encoding the Picard-number-dependent correction from the K3 fibre volume modulation. Here, \log without a subscript denotes the natural logarithm \ln (noting that ratios of unsubscripted logarithms such as \log(10)/\log(3) = \log_3(10) are invariant under any choice of base), while \log_3 2 \equiv \ln(2)/\ln(3) \approx 0.63093 explicitly specifies base 3. The base spectral contribution evaluates to [Equation: 2 + \frac{\log(9 + 1)}{\log 3} = 2 + \frac{\log 10}{\log 3} = 2 + 2.096 = 4.096\,, ] and with the K3 correction: \gamma_{\rm analytic} = 4.096 + 0.568 = 4.664. The full numerical quadrupole evolution---incorporating nonlinear mode coupling in the 15-node graph beyond the linearised formula~(eq:spectral_index)---produces the reported value [Equation: [\;\gamma = 4.847\;] ] with the \sim 4\% correction (4.664 \to 4.847) attributable to higher-order eigenvalue mixing terms involving the ring nodes. -------------------------------------------------------------------------------- Comparison with Astrophysical Models -------------------------------------------------------------------------------- The Oligon spectral index \gamma = 4.847 differs from: * The **SMBHB prediction**: \gamma = 13/3 \approx 4.33 (circular, GW-driven inspiral); * The **NANOGrav measurement**: \gamma = 4.70 \pm 0.50 (1\sigma). The Oligon value lies within the NANOGrav 1\sigma interval (4.20 \leq \gamma \leq 5.20) and is displaced from the SMBHB prediction by \Delta\gamma = 0.51---a measurable divergence with increased PTA sensitivity. -------------------------------------------------------------------------------- Compton Resonance at 24.18~nHz -------------------------------------------------------------------------------- The Lorentzian factor [1 + (f/f_\chi)^2]^{-1} in Eq.~(eq:hc_spectrum) produces a spectral ``bump'' at f = f_\chi = 24.18\,nHz. This is a distinctive signature: the SMBHB model predicts a featureless power law, while the Oligon model predicts a resonance. At the NANOGrav 15-year frequency resolution (\Delta f \approx 2\,nHz), this resonance spans approximately one frequency bin. With longer baselines (NANOGrav 25-year, IPTA DR3) and narrower frequency resolution, the bump will either appear or be excluded, providing a clean falsification test. ================================================================================ Compatibility of l=4 Anisotropy with the Hellings--Downs Detection ================================================================================ -------------------------------------------------------------------------------- The Apparent Tension -------------------------------------------------------------------------------- The K_4 Oligon framework predicts a hexadecapole (l=4) dominant anisotropy in the SGWB angular power spectrum, with C_4/C_0 = 16.07. At face value, this appears to contradict the NANOGrav 15-year analysis, which reports evidence for a common-spectrum process via the _isotropic_ Hellings--Downs (HD) inter-pulsar correlation. A genuine l = 4 dominant background would seem to violate the assumptions of the HD search pipeline. We now show that this tension is _illusory_: the l = 4 anisotropy in the source power spectrum is strongly suppressed in the _observed_ cross-correlation by the overlap reduction functions (ORFs) of the PTA. -------------------------------------------------------------------------------- Anisotropic Cross-Correlation Formalism -------------------------------------------------------------------------------- The pulsar-pair cross-correlation coefficient for an anisotropic SGWB with angular power spectrum \{C_l\} is~[Mingarelli2013,Taylor2013]: [Equation: \Gamma(\zeta) \;=\; \sum_{l=0}^{\infty}\, \frac{2l+1}{4\pi}\, C_l\, P_l(\cos\zeta)\, \bigl[\mathcal{F}_l\bigr]^2\,, ] where \zeta is the angular separation between pulsars, P_l are Legendre polynomials, and \mathcal{F}_l are the overlap reduction functions for the Earth-term-only approximation. The isotropic HD curve corresponds to C_l = C_0\,\delta_{l0} (only the monopole). For an l = 4 dominant background with C_4/C_0 = 16.07 (the Oligon prediction): [Equation: \Gamma_{\rm Oligon}(\zeta) \;=\; \frac{C_0}{4\pi}\, \mathcal{F}_0^2 \;+\; \frac{9\, C_4}{4\pi}\, P_4(\cos\zeta)\, \mathcal{F}_4^2\,. ] -------------------------------------------------------------------------------- The ORF Suppression Proof -------------------------------------------------------------------------------- The critical physical fact is that \mathcal{F}_4 is **strongly suppressed** relative to \mathcal{F}_0 for current PTA baselines. The overlap reduction function for the l-th multipole in the short-wavelength limit scales as~[Gair2014]: [Equation: \mathcal{F}_l \;\approx\; \frac{1}{l(l-1)} \quad \text{for } l \geq 2\,,\qquad \mathcal{F}_0 \;=\; 1\,. ] Therefore: [Equation: \frac{\mathcal{F}_4^2}{\mathcal{F}_0^2} \;=\; \frac{1}{[4 \times 3]^2} \;=\; \frac{1}{144}\,. ] Even with C_4/C_0 = 16.07, the contribution of the l = 4 term to the _observed_ \Gamma(\zeta) is: [Equation: \frac{9\, C_4\, \mathcal{F}_4^2}{C_0\, \mathcal{F}_0^2} \;=\; \frac{9 \times 16.07}{144} \;\approx\; 1.00\,, ] meaning the l = 4 anisotropic contribution is of _comparable magnitude_ to the monopole---not dominant in the observed correlation. -------------------------------------------------------------------------------- Consistency with the NANOGrav Detection -------------------------------------------------------------------------------- The NANOGrav pipeline projects \Gamma(\zeta) onto the HD template \Gamma_{\rm HD}(\zeta) and reports a signal-to-noise for the HD component. An underlying anisotropic background that produces a \Gamma(\zeta) with \sim50\% HD-like component and \sim50\% P_4(\cos\zeta) residual is _detectable_ by the HD search as a positive correlation with reduced amplitude. This is exactly what NANOGrav observes: a 3--4\sigma detection, not a high-significance one. The Oligon model predicts that the HD fraction in the total observed signal is [Equation: f_{\rm HD} \;=\; \frac{1}{1 + 9\, (C_4/C_0)\, (\mathcal{F}_4/\mathcal{F}_0)^2} \;=\; \frac{1}{1 + 1.00} \;=\; 49.9\%\,. ] A \sim50\% HD component suppresses the effective SNR by a factor of \sim\!\sqrt{2} compared to a purely isotropic background of the same total power, reducing a nominal \sim\!5\sigma signal to the observed 3--4\sigma level. -------------------------------------------------------------------------------- Falsifiable Prediction for SKA -------------------------------------------------------------------------------- The K3 \times T^2 model predicts that the anisotropic residual \Gamma_{\rm Oligon}(\zeta) - \Gamma_{\rm HD}(\zeta) has a specific P_4(\cos\zeta) angular shape with amplitude [Equation: A_4 \;=\; \frac{9\, C_4\, \mathcal{F}_4^2}{4\pi} \;\approx\; 0.32\, \frac{C_0}{4\pi}\,. ] This residual is below the current NANOGrav 15-year sensitivity to anisotropy but will be testable with the Square Kilometre Array (SKA). SKA will observe \sim\!10\times more pulsars with \sim\!3\times longer baselines, projecting to resolve angular power spectrum modes up to l \sim 6~[Taylor2022]. If the P_4(\cos\zeta) residual with the predicted amplitude A_4 is detected, it would constitute strong evidence for the Oligon hypothesis; its absence at the predicted level would falsify the model. ================================================================================ The Topological Hadamard Mask ================================================================================ -------------------------------------------------------------------------------- The Problem: Unbounded Graph Growth -------------------------------------------------------------------------------- The hypergraph evolution rule of Section~sec:hypergraph_model requires a stabilisation mechanism. Without it, repeated application of the substitution rule M \to M \circ S (Hadamard/entrywise product with a substitution tensor S) drives the graph to infinite size. We must define the stabilisation explicitly and justify it physically. -------------------------------------------------------------------------------- Definition 4: The Topological Hadamard Mask -------------------------------------------------------------------------------- Let M^{(n)} denote the adjacency matrix at evolution step n. The substitution rule with topological stabilisation is [Equation: M^{(n+1)} \;=\; T \circ \bigl(M^{(n)} \cdot S\bigr)\,, ] where \circ denotes the Hadamard (entrywise) product, S is the K_4 substitution kernel, and T \in \{0,1\}^{15 \times 15} is the mask defined by: [Equation: T_{ij} \;=\; \begin{cases} 1 & \text{if } d_G(i,j) \leq D_{\max} \;\text{and}\; \deg(i), \deg(j) \leq \Delta_{\max}\,, \\ 0 & \text{otherwise}\,, \end{cases} ] with D_{\max} = diam(\mathcal{H}) = 7 (the graph diameter of the full 15-node ring+K_4 structure) and \Delta_{\max} = \max\deg(\mathcal{H}) = 4 (the maximum vertex degree, arising from K_4 core nodes having 3 clique edges plus 1 ring edge each). -------------------------------------------------------------------------------- Physical Justification -------------------------------------------------------------------------------- The mask T is **not an arbitrary numerical cap**. Its two conditions encode fundamental physical constraints: * Causal horizon (D_{\max: = 7).} In Wolfram-model pregeometry, the causal graph diameter D_{\max} is the discrete analogue of the cosmological particle horizon: nodes separated by more than D_{\max} edges are causally disconnected and cannot exchange topological updates. Setting D_{\max} equal to the diameter of the seed graph \mathcal{H} (diam(\mathcal{H}) = 7) enforces that the hypergraph evolution respects the causal structure inherited from the seed. * Holographic degree bound (\Delta_{\max: = 4).} The holographic entropy bound in pregeometric models requires that the information content per causal patch (node) scales as the surface area, not the volume~[Bousso2002]. In graph terms, this constrains the vertex degree: the number of edges incident to a node is the discrete surface ``area'' of that node's causal patch. Preserving \Delta_{\max} = 4 under evolution is equivalent to requiring that the holographic bound is saturated but not violated. -------------------------------------------------------------------------------- Uniqueness -------------------------------------------------------------------------------- With these constraints, the mask T is _uniquely determined_ by the seed graph \mathcal{H}. There is no free parameter to tune. Any connected simple graph G determines its own mask via D_{\max} = diam(G) and \Delta_{\max} = \max\deg(G). For the 15-node ring+K_4 structure, this produces a mask with Tr(T) = 15 and \|T\|_F^2 = 225, verified computationally. -------------------------------------------------------------------------------- Spectral Stability -------------------------------------------------------------------------------- Under the masked evolution Eq.~(eq:evolution_rule), Tr(M^{(n)}) is bounded by Tr(T \circ S) = Tr(M^{(0)}) for all n, ensuring that the graph does not grow unboundedly. The fixed-point adjacency matrix M^{(\infty)} satisfies [Equation: M^{(\infty)} = T \circ (M^{(\infty)} \cdot S) ] and has the same spectral radius \lambda_1 = 3 as the seed K_4, preserving the spectral bridge to the Cooper s_{10} sequence (Section~subsec:spectral_picard). This stability property has been verified numerically: the spectral radius of T \circ M equals 3.1844, bounded above by the spectral radius of M itself (3.1844), and the pure K_4 spectral radius of 3.0 is preserved within the K_4 subgraph. ================================================================================ Results: Comparison with NANOGrav and SKA Projections ================================================================================ -------------------------------------------------------------------------------- Summary of Predictions -------------------------------------------------------------------------------- The K_4 Oligon hypergraph model, via the continuum limit of Section~sec:continuum_limit, generates three quantitative predictions for the nanohertz SGWB: [Table] All four predictions are consistent with the current NANOGrav 15-year data within 1\sigma (for \gamma) or below current sensitivity (for f_\chi, C_4/C_0, f_{\rm HD}). -------------------------------------------------------------------------------- Bayesian Information Criterion -------------------------------------------------------------------------------- For the spectral index alone, the Bayesian Information Criterion (BIC) comparison between the Oligon model (\gamma = 4.847, 0 free spectral parameters since \gamma is derived from K_4) and the SMBHB model (\gamma = 13/3, 0 free spectral parameters) is [Equation: \DeltaBIC = \chi^2_{\rm SMBHB} - \chi^2_{\rm Oligon} = \frac{(4.70 - 4.33)^2}{0.50^2} - \frac{(4.70 - 4.847)^2}{0.50^2} = 0.548 - 0.086 = 0.46\,, ] which is inconclusive (|\DeltaBIC| < 2). The current data do not distinguish between the two models. -------------------------------------------------------------------------------- SKA Projections -------------------------------------------------------------------------------- The Square Kilometre Array (SKA) will provide transformative improvements in PTA sensitivity~[Taylor2022]: * **10\times more pulsars**: \sim200 millisecond pulsars (vs.~\sim20 in NANOGrav 15yr), enabling resolution of angular power spectrum modes to l \sim 6. * **3\times longer baselines**: \sim30-year datasets, reducing the frequency resolution to \Delta f \approx 1\,nHz and improving spectral index precision to \sigma(\gamma) \approx 0.05. * **\sim\!30\times SNR improvement**: via both increased pulsar count and longer integration. At SKA sensitivity, the three Oligon signatures become decisively testable: * 1.~Spectral index.: With \sigma(\gamma) \approx 0.05, the \Delta\gamma = 0.51 displacement between Oligon (\gamma = 4.847) and SMBHB (\gamma = 4.33) becomes a >10\sigma discrimination. * 2.~Compton resonance.: The f_\chi = 24.18\,nHz resonance spans \sim2 frequency bins at current resolution but will be resolved into \sim25 bins with 30-year baselines. A Lorentzian feature at the predicted frequency either appears or is excluded at >5\sigma. * 3.~l=4 anisotropy.: The anisotropic residual A_4 \approx 0.32\, C_0/(4\pi) (Eq.~eq:residual_amplitude) is below NANOGrav 15-year sensitivity but above SKA projections. Detection of a P_4(\cos\zeta) angular pattern at the predicted amplitude would confirm the Oligon hypothesis; its absence would falsify it. -------------------------------------------------------------------------------- Discrimination Table -------------------------------------------------------------------------------- [Table] ================================================================================ Conclusion ================================================================================ We have demonstrated that a discrete Wolfram-model pregeometry, seeded by the K_4 complete graph, generates a stochastic gravitational-wave background with spectral properties that are analytically determined by the graph eigenvalue structure. The principal theoretical advance is the rigorous continuum limit (Definitions~1--3) that bridges abstract combinatorial graphs to physical spacetime via Gromov--Hausdorff spectral convergence~[Burago2006]. The model produces three falsifiable predictions: a spectral index \gamma = 4.847 (derived from \lambda_1 = 3, \lambda_2 = -1 of the K_4 adjacency matrix), a Compton resonance at f_\chi = 24.18\,nHz (from the T^2 Kaluza--Klein scale), and a hexadecapole (l=4) spatial anisotropy (C_4/C_0 = 16.07). All three are consistent with the NANOGrav 15-year dataset and will be decisively tested by SKA-era observations. The most important intellectual contribution is the ORF suppression proof (Section~sec:anisotropy_hd): we showed that a strongly anisotropic _source_ (C_4/C_0 = 16.07) produces a nearly isotropic _observed_ cross-correlation (f_{\rm HD} \approx 50\%) due to the \mathcal{F}_4^2/\mathcal{F}_0^2 = 1/144 suppression, naturally explaining both the NANOGrav HD detection and its moderate significance. -------------------------------------------------------------------------------- Caveats -------------------------------------------------------------------------------- We have been transparent about the limitations: * The scalar mass m_\chi = 10^{-22}\,eV is an **ansatz**: the K\"ahler modulus t is fixed by requiring the T^2 KK scale to match the observed 24.18\,nHz frequency bin. A genuine prediction requires an independent stabilisation mechanism. * The spectral index formula~(eq:spectral_index) is a linearised analytic approximation; the \sim\!4\% correction to the full numerical value \gamma = 4.847 reflects higher-order eigenvalue mixing terms that are not analytically controlled. * The topological Hadamard mask (Definition~4) is uniquely determined by the seed graph, but the _choice_ of the K_4 seed itself is a postulate of the model, not derived from deeper principles. -------------------------------------------------------------------------------- Future Directions -------------------------------------------------------------------------------- Three extensions are planned: * **Independent t-stabilisation**: deriving the K\"ahler modulus from a flux superpotential (as in Paper~1~[LeanProposal]) to convert the mass ansatz into a genuine prediction. * **Full angular power spectrum simulation**: computing C_l for l = 0, \ldots, 10 from the 15-node graph to generate detailed SKA templates beyond the l = 4 dominant mode. * **Cross-correlation with Paper~1**: the same Cooper s_{10} K3 surface with P = 19 that produces w_0 = -0.974 and \Omega_m = 0.295 (tested against DESI~2024 BAO in the companion paper) also generates the gravitational-wave predictions of this work. Joint analysis of expansion-history and GW data will over-constrain the model, providing a powerful consistency test. ================================================================================ Acknowledgments ================================================================================ The author thanks the NANOGrav Collaboration for making the 15-year dataset publicly available, and the Wolfram Physics Project for inspiring the pregeometric framework. Formal verification of the spectral--Picard bridge was performed using Lean~4 and Mathlib. Computational resources were provided by Google Cloud Platform (Vertex AI).