emergent-gravity.com · the audited vacuum · july 2026
A theory of vacuum monitoring. Two gravities buried. A third on trial.
The vacuum maintains a locally conserved budget and monitors only violations of its own conservation law — never states. What that principle demonstrably yields: dark states derived (zero-at-rest exact at the marginal point); a falsifiable decoherence knee; a Coulomb-class static force owned by the constraint's own zero; Gauss's law from the left kernel. What it could not yield on any perfect lattice: Einstein's radiation — two crystalline gravity programs died by their own frozen gates (one ledger cannot carry both forces; no crystal can carry both polarizations — causes of death on the status page). The successor lives on the wreckage's lesson: the amorphous two-ledger vacuum — a glass, because the vacuum froze in finite time. Its first foundation evidence did not survive its own later construction (the 2026-07-12 adversarial review: wrong graph, wrong probes — nothing from that era is citable), but what replaced it is sharper: an exact theorem — the audit reads everything that lives on sites; dark is what lives irreducibly in relations (dimension formula exact to the unit, 12/12) — whose corollaries name the candidates: the photon is a circulation, the graviton a bond-metric modulation, and every displacement is silent. The candidate now stands trial before the reviewer's consolidated gate (content → dynamics → cone → coupling → the 1919 test, in that order, no backward inference), with every fail clause pre-written. And the program's oldest debt is now its declared purpose: Le Sage was right — gravity is a push from an omnipresent flux, and the masses shade each other; he only had the wrong currency. Corpuscles carry momentum and die four classical deaths (Maxwell, Laplace, Poincaré, saturation); an audit carries information and dies none of them — the shadow is derived, linear, and Newtonian (n270). Read it with the legend: every claim carries its class — [theorem] proven · [toy] numerically observed in finite models, by the authors · [reading] interpretation · [wager] conjecture with skin in the game. "Toy" never means laboratory.
(A·δk)ᵢ = −cᵢ + ρ̇ = ℒviolations[ρ]
the ledger — matter drains a conserved budget: a Gauss law by definitionthe audit — the vacuum measures its books, never the worldeverything else — is derived, measured, or honestly declared absent
sector I toy 5/5 reading
The audited sector is Coulomb-like — read as electrostatics.
The Coulomb form's origin, reattributed by the trunk audit: the long-range pole is forced by the staggered zero of the unsigned constraint operator itself — any positive local cost delivers the exact Green form (verified with the identity cost, residual 10⁻¹⁵); the information metric is the declared price that fixes the coefficient and short-range structure, and attraction within the chosen source sector is an implication of constraint + positivity, not a discovery of the metric. With that said honestly: the conserved sector delivers exact Coulomb interactions in one dimension (fit residual 3×10⁻⁶) and Coulomb-consistent quadratic infrared dispersion in two and three (0.24% / 0.36% on pre-registered gates — dispersion checks, not direct force-law measurements), two charge classes with the like-repels-like sign rule (sublattice parity — with the transport debt open: a one-site translation flips the label), emergent neutrality, a potential provably invisible to the constructed clocks — a clock-invisibility theorem; the gauge-invariance reading is retracted (audit §49: gauge symmetry needs local redundancy, charged-matter transformations, and Ward identities, none built — Wilson loops would be physical here). The charge-superselection claim is also retracted (audit §50: it contradicted the dark-state trunk — two dressed branches satisfying the constraint are both dark, so the monitor cannot distinguish their charges; n171's charge-blind verdict had already killed the mechanism internally), a gapless collective mode protected by the conservation law (the Gauss zero mode is exact at every regulator) — read as the photon, an interpretation the model cannot yet cash out (no vector field, no polarizations, no Maxwell dynamics, no Lorentz invariance: the debts are listed), and a short-range neutral residual measured at r⁻³·⁵. Continuum couplings pinned: S∞ = 202.588 (d=2), 493.83 (d=3).
sector II toy reading
Gravity is what the audit costs — the static reading stands; the radiative program is dead. †
The monitor's back-action induces a second force — and three measured facts fix its character. Within the RWA linear-mediator family (audit §16 scope: passivity = Δ > 0 in that family; broader passive susceptibilities are outside the proof), the induced force is repulsive; attraction requires the negative-detuning branch — whose reading as a pumped active vacuum is a hypothesis: the computed model contains no pump (audit §9), and the demon-needs-a-battery line is the motivation for the unbuilt pumped-reservoir construction, not a result. The active medium's stability, recomputed without any regulator under a frozen adversarial gate (n183, repairing the n74 flaw): zero unstable modes; the only two zero modes are the classified conserved pair (global time origin, global energy offset), exactly decoupled; physical margin −1.3×10⁻³. The complete symmetrized matter-coupled noise matrix is now computed: under Δ → −Δ the branches are identical to machine precision in every matter-coupled quadrature, while the coherent response flips sign — response odd, noise even, now prediction 33 with a parameter-free curve and an analog protocol. Scope carried openly (the reviewer's corrections adopted): the Δ < 0 branch is a rotating-frame detuning reversal, not a constructed inverted reservoir — the pump that would realize a true inversion is unmodeled, the device-class label awaits the quadrature-gain computation, and the SI normalization is underived, so no laboratory bound currently tests this (the LPF leg returned NO PHYSICAL VERDICT) — and its idle energy bill is exactly zero at Gaussian order, where naive vacuum energy overshoots by 120 orders. In the attractive branch, clocks near mass run slow — and cannot do otherwise: the rate sector is sign-rigid. Gravity is weak because the audit is overdamped — as a reading of the shared-parameter identity γknee/Ωgrav = κ/|Δ| ≡ Q (audit §12: the ratio is an identity of the single-mode model; its promotion to the observed hierarchy assumes five named transfers, none yet established).
the hinge theorem
Redshift is quantum — ħ is the hinge.
A theorem, exact to machine precision: classical drift rates couple only to tides (∇²Φ), never the potential. Quantum phases read Φ linearly, universally, at long range, through one identity: ω = E/ħ. In this model class, energy-derived phases read the potential while the classical observables we constructed read only its gradients; every energy-derived rate (pulsar rotation included) then redshifts identically, and the residual converges to exact as 1/N. (General relativity describes classical redshift as proper time; the model's sharper claim — that its clock sector needs ħ — holds for this class, not yet for nature.) Consequence: the equivalence principle is asymmetric — force-EP could break someday; clock-EP structurally cannot. Within the model class: the price is state geometry, the clocks are phases, the currency is ħ — an inversion of the usual quantization program, offered as a hypothesis about nature, not a result.
the closure toy: signs scale open
One anharmonicity could feed both Mercury and dark energy.
The linear theory has β = ½ exactly (measured: the F² response is zero to machine) — the missing half of Mercury's perihelion is the field's self-coupling. With γPPN pinned by Cassini, the sky measures the substrate's first anharmonic coefficient to four digits: Δβ = +0.5000 ± 0.0001, demanding a stiffening anharmonicity (ξ > 0, measured). The layer's form is no longer a choice: the Nordtvedt gate (n112) killed the cubic vertex (IR-divergent far-field charge) and selected the derivative coupling — under which self energy and binding energy source gravity with one coefficient as an exact lattice identity (ηN = 0 as a lattice identity — audit §23: not the relativistic Nordtvedt parameter; LLR and MICROSCOPE test a larger dynamical statement, not this identity). On the surviving layer the chain re-closes: the chain is now dismantled by the audit (§§25–27): the fitted ξ = 0.2312 lies outside controlled perturbation theory (the second-order term is 3× the first; the full Newton solver diverges; no stable nonlinear branch is demonstrated), the Mercury script as published is broken (calls the divergent solver, prints a sign after non-convergence — gate defect recorded), and the sign link is vacuous: ΔP ∝ +ξ² is positive for either sign of ξ, so Mercury contributes nothing to Λ's sign — the only shared fact is the symbol ξ. What survives: the derivative integration-by-parts identity (exact) and a positive regulated auxiliary-power correction (finite regulator limit verified by the auditor). Neither is a cosmological constant. Reading that cost as dark energy's sign is [reading]: a cosmological constant is a covariant stress tensor with p = −ρ, and the map pump-power → Tμν (equation of state, homogeneity, conservation, magnitude — the scale wound is up to 10⁸⁹, carried openly) is a registered open derivation. Mercury calibrates ξ; the sign linkage is a model result and a direction, not an established identity.
the gate everything rides here
One experiment carries the falsifiability.
A mass superposition at rest decoheres not at all — constraint-satisfying states are dark states of the audit. Modulated at frequency ω, it decoheres along the Lorentzian (dT₀₀/dt)²/(γ²+ω²) (shape honesty per the trunk audit: the Lorentzian is the transfer function of the single-pole monitor class — a consequence of the Markovian-relaxation postulate, not of the substrate Hamiltonian; n48's registered gates P1/P2/knee failed as written — unit-convention calibration, recorded in the prereg itself — and the 0.3% Lorentzian match is a flagged post-hoc comparison against the assumed single-pole transfer function; the n128 mediator model holds it to within 11% at the knee). Quantized general relativity carries no intrinsic knee-shaped signature (graviton-decoherence master equations exist and predict their own parameter-dependent spectra — the honest comparison is model-by-model); white-noise collapse models predict a rest floor (colored-noise CSL variants carry structure of their own); this theory predicts zero at rest, then a knee, then more-than-collapse above it — a shape none of the named rivals shares. The knee's predicted location collapsed to a clean formula when G cancelled out: γknee = α·mc²/ħ — the fine-structure rate of the mass the vacuum counts in; the window between the two possible anchors is exactly mp/me, and the natural bet is the extremal dressing: the electron anchor, ħγ ≈ 3.7 keV — soft X-ray. The framework predicts the shape (zero at rest, ω² onset, Lorentzian knee), and now a derived functional with no post-hoc tuning within its declared model choices (n128 → n180 → n181; ζ is geometry-dependent per the family sweep, and the 0.4834 was a partial-cycle averaging artifact (the drive window covered 0.038 cycles near maximum — trunk audit §11); the exact coefficient under the registered sinusoidal protocol is ζ = 1/4 (DC convention: 1/2)). The functional bills per coherence cell (ℓc = c/γ — a named assumption, which at the electron anchor rewrites to the Bohr radius) — and executed as its own frozen gate, it retracted both of the theory's absolute targets: the milliwatt beam (form factor 10⁻¹³) and then the 3 pN single-ion figure (benchmark geometry factors 3.5×10⁻³ and 1.9×10⁻², outside the frozen [0.1, 10]). No absolute laboratory target is armed; rates are computed per design. The gate's benchmark returned a shape ratio (dumbbell/sphere = 5.27 in the selected geometry) that is geometry-dependent, not yet a universal prediction — the shape discriminator's magnitude awaits a registered geometry sweep. The experiment →
geometrogenesis — the point was never a point; it was a complete graph (n117, live)
worked out toy: dH=dU readings
Two questions the framework answers in full: de Broglie, and antigravity.
De Broglie's internal clock, given a mechanism [reading]: the quantum phase as the vacuum's receipt counter — ω = E/ħ is the billing rate of existence, the predicted knee is exactly α × the anchor's de Broglie frequency (one audit per 137 turns), and universality of redshift is derived rather than postulated. In the physics page →Antigravity, fully accounted: the only real one is dark energy (the pump's exhaust, undirectable); the theoretical one is a de-pumped vacuum — priced at P = mc²|Φ|γ ≈ one solar luminosity per levitated kilogram (0.92 L☉ at Earth's surface) [a CONDITIONAL energetic estimate, anchor-dependent — the microscopic pump power and the de-pumping "button" itself are underived], against a vacuum that heals in 0.18 attosecond; the folkloric ones are dead (antimatter falls — measured; negative mass forbidden by c ≥ 0; shielding impossible — the back-action is unmediated); and the energy books close exactly: |ΔH| = |ΔU| to four digits [toy] — antigravity cycles net zero, the machine is a battery, never a source. Plus the four-line checklist that rejects every antigravity claim in history. The machine, designed honestly → · the four faces · the energy no-go
the treatise
Read it in full.
The physics — six master equations, five mechanisms, the boundaries, the calibration. Results — every theorem, measurement, lock, and inversion. Parameters — every constant interrogated, every bracket with its source. Predictions — thirty-seven standing, each with its kill condition or exposure. The experiment — the knee, for the laboratory that will run it. Status — what is not explained, the wounds, the debts, the provenance, and the archive of how this was found (including the thirty-four ideas buried on the way). The paper — the two theorems and the constructive result, seven pages. The second paper — the topological dark sector: born from a falsification, keV cold defects, collisionless by its own arithmetic, three kill conditions, four pages. The third paper — the junction closure: GR's radiation content by constraint counting, three pages.