Gardi-Lori: mathematician's review (PhD mathematician role)

Scope: the seven central claims in claims.md, checked against the source PDFs in source folder: Gardi-Lori (read only). Calculations were run in Python (numpy) and are shown below. Page images of the Yee-Gardi "Geometry of Spacetime" v2.3 (pp. 5-9, 27-28) were rendered and read [read: page image p.5, p.6, p.7, p.8, p.27, p.28].

Source gap (repeated from Step 1, with additions)

• Equations are images in the Yee-Gardi papers, so the text layer is empty there. I read pp. 5-9 and 27-28 of the Spacetime v2.3 paper as images. I did not read its sections on the strong force, Planck constant (5.5), magneton or particles (pp. 10-26) as images, so any statement about those is [memory] or from the Step 1 summary only.
• Not in the folder and not read: "The Relationship of the Mole and Charge", "The Relationship of the Fine Structure Constant and Pi", Distinti's papers, Gardi's "Planck's Constant and the Law of Capacitance" and "Calibrating the universe". Nothing below says anything about them beyond what the folder's papers state about them.
• Electrigen paper (Dinu-Gardi) read to p.8 of 10 by Step 1; I did not re-read it. It contains no equations, so there is nothing to check mathematically; its claims are tested only against known experiments.
• Where a conclusion rests on [memory] alone it is marked.

Reader-level definitions used below

• c = speed of light, 299 792 458 m/s. h = Planck's constant 6.626e-34 J s; hbar = h/(2 pi). alpha = fine-structure constant, a pure number about 1/137.036, which measures the strength of the electromagnetic force. e = elementary charge. eps0, mu0 = electric and magnetic constants of free space. Z0 = sqrt(mu0/eps0) = 376.73 ohm, the ratio of electric to magnetic field strength in a light wave in vacuum.
• Planck units: l_P = sqrt(hbar G / c^3) (length), m_P = sqrt(hbar c / G) (mass), q_P = sqrt(4 pi eps0 hbar c) (charge). r_e = e^2/(4 pi eps0 m_e c^2) = 2.818e-15 m is the "classical electron radius". It is defined by that equation, which says electrostatic energy of charge e at distance r_e equals the electron rest energy.
• "Identity" below means true for all values by algebra, so it checks nothing about nature.

Calculations (run, not recalled)

Verified on the web (retagged from Step 1 [memory])

• [web] https://arxiv.org/abs/1602.03920 (Connaughton et al. 2016): Fermi GBM weak transient above 50 keV, about 0.4 s after GW150914, lasted about 1 s, 2.9 sigma (false-alarm probability 0.0022), not seen by other instruments; localisation ill-constrained.
• [web] https://arxiv.org/abs/1710.05834: GRB 170817A followed GW170817 by (+1.74 +/- 0.05) s; difference between speed of gravity and light constrained to between -3e-15 and +7e-16 of c.
• [web] https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.127.100401 (Deppner et al.): lowest temperature 38 pK (-7/+6 pK), about 100,000 atoms.
• [web] https://physics.nist.gov/cgi-bin/cuu/Value?mu0: mu0 = 1.25663706127e-6 N/A^2, relative uncertainty 1.6e-10; it is no longer defined as exactly 4 pi x 1e-7.
• [web] https://arxiv.org/abs/0908.1832 (Fermi GRB 090510): limit on linear Lorentz-invariance violation requires a quantum-gravity mass scale significantly above the Planck mass.
• Not verified on the web (budget): INTEGRAL non-detection of the GW150914 transient (still [memory], medium); EHT ring circularity (still [memory], high); cavity tests of Lorentz invariance (still [memory], high).

---

Claim 1. Light propagates in a physical medium (aether); a lattice of bound e+e- pairs or Planck-scale granules

a. Mathematics. The argument has three steps. (i) Rewrite c = 1/sqrt(eps0 mu0) as sqrt((1/mu0)/eps0) to look like v = sqrt(stiffness/density). This is only a rewriting; "1/mu0 is a stiffness" is an assumption, not a derivation [read: Medium paper, Sect. 3]. (ii) NaCl check: v = sqrt(A/(B+C)) with A the binding energy of one ion pair and B+C its mass. It gives about 3700 m/s (C2: 3667). Dimensionally fine (J/kg = m^2/s^2), but a real sound speed is sqrt(K/rho), with K the bulk modulus, which depends on the curvature of the potential energy curve, not on the binding energy itself. The cohesive energy per unit mass gives the right order of magnitude (a few km/s) for any solid. So the NaCl "success" is an order-of-magnitude coincidence of dimensional analysis [memory, high]. (iii) Vacuum check: A = 2 m_e c^2 and B = 2 m_e, so sqrt(A/B) = c by algebra (C1). The same cancellation works for proton-antiproton, quark pairs, or any mass, which the paper states as a virtue. It is the reason the check carries no information: whatever you choose, you get c. Also "binding energy equal to the entire rest energy of the pair" means the pair has zero total energy; it is simply E = mc^2 renamed. Later papers (Spacetime v2.3): the "granule" is a Planck mass of radius l_P. By C14 its Schwarzschild radius is 2 l_P, so it lies inside its own horizon on the paper's own numbers. The paper's lattice spacing a = 2 l_P e and wavelength 2 q_P e^2 are stated, not derived; "they will be proven later" (p.5), and the later "proof" is the N_A fit (C8). OM paper (2023): circular wave of circumference lambda travelling at tangential speed c: frequency f = c/lambda; with lambda = Compton wavelength h/(m c) this gives hf = m c^2, which is consistent but is also the definition of the Compton wavelength. Minor inconsistency: the text says the peak and valley are separated by the radius lambda/(2 pi); two opposite phases of a rotating wave would sit a diameter apart. The figure (Yin-Yang symbol) is what sets the radius, so I flag it as unclear, not wrong.

b. Experimental consistency. The Michelson-Morley "circularity" argument (light is used to detect the medium for light) is not valid. Round-trip light-speed tests with different setups (Kennedy-Thorndike, Ives-Stilwell, modern cryogenic and optical-cavity resonator comparisons to about 1e-17) and matter-only Hughes-Drever tests of rotation invariance give null results [memory, high; not individually web-verified]. Fermi GRB 090510 constrains linear energy-dependent photon speed at beyond the Planck scale [web: arxiv.org/abs/0908.1832], which is hard for a Planck-scale lattice to meet. A real e+e- sea at Compton spacing would have density of order 1e5 kg/m^3 (2 m_e per lambda_C^3 = 1.3e5 kg/m^3, my calculation, which assumes a spacing the paper does not state); the paper's use of "per unit volume" lets that number cancel out of the formula and never confronts it.

c. Tested? Aether-drift experiments have been done for 140 years with null results [memory, high]. The specific e+e- medium and Planck lattice have not been confronted with any measurement because they make no quantitative prediction beyond c itself. The paper cites an "unofficial acoustic MM" in air, which tests nothing about light. The 2015 LIGO claim is under claim 5.

d. What would change. Special relativity would become an emergent or approximate theory with a preferred frame; the quantum vacuum (a Lorentz-invariant state in quantum field theory) would be replaced; and a lattice would need an explanation for the absence of any preferred direction at the 1e-17 level.

Verdict (split). - Medium with a detectable preferred frame: Tested, not supported. - Lorentz-ether versions designed to be undetectable, and the "e+e- pair medium gives c" derivation: Not testable as stated (the derivation is an identity, C1). - Granule of radius l_P and mass m_P: Internally inconsistent with the Schwarzschild radius of the stated mass (C14), with the caveat that this uses general relativity at Planck scale where its validity is itself an open question.

---

Claim 2. Charge is a length; mass is standing-wave energy; Coulomb's law, electron mass and gravity are geometries

a. Mathematics. Everything checks algebraically and nothing is derived. - CNPS 2018: F = m c^2 r_e/r^2 equals k e^2/r^2 because r_e is defined by m c^2 r_e = k e^2. Numerically 29.0535 N at r = r_e on both sides (checked). A definition restated. - Mass-and-Charge paper: m_e = mu0 e^2/(4 pi r_e) is that same definition solved for m_e (identity; the paper's remark that it is "without the c^2 factor" is only because mu0 = 1/(eps0 c^2)). Coulomb's constant k = mu0 c^2/(4 pi) is the SI relation k = 1/(4 pi eps0). [read: summary and Eq. 5.8.2 in Spacetime p.28.] - Re-dimensioning: coulomb to metre, ampere to m/s, mu0 to kg/m. This can be made algebraically consistent (every equation just has the numbers relabelled), and the paper's units table does that. But it is a change of notation. Dimensions are a choice (Gaussian units already have e^2 = force x length^2; natural units set hbar = c = 1 and make e dimensionless, sqrt(alpha)). Equating charge to a length does not predict a new number. The number "Planck charge as a length" in the lattice (page 5: a granule displaced by q_P, wavelength 2 q_P e^2) uses q_P = 1.876e-18 (its SI numerical value) as metres only if the SI number is lifted without its unit conversion; I could not see the value in the extracted text, so I treat that as unconfirmed. If taken literally, the displacement would be 1e17 times larger than the granule radius and cell spacing (8.8e-35 m), which would not be a small harmonic oscillation of a unit cell. - The 2023 OM paper (Sect. 5.1) says that equating charge with wave amplitude and cancelling m^2/C^2 "should not be done" and criticises both Macken and Yee. Yee-Gardi (2019) does exactly that. The two positions by the same author contradict each other; the 2023 paper does not say it has abandoned the earlier work.

b. Experiment. Charge quantisation, charge conservation, the measured e/m of electrons (Thomson, Millikan), and Gauss's law are all consistent with charge as an independent conserved quantity. A length-valued "charge" is a relabelling, so nothing here conflicts with experiment, but it also does nothing a coulomb does not do [memory, high].

c. Tested? No quantitative prediction was made, so nothing to test. Identities are verified as identities.

d. What would change. If charge really were a length with no additional constant, then e^2/(hbar c) would be a ratio of lengths (alpha) and the SI base unit set would lose a dimension. Since the SI of 2019 fixes e, that is only a convention change.

Verdict. Re-dimensioning: Not testable as stated (a unit convention). "Derivations" of Coulomb's law and the electron mass: reduce to definitions (see claim 3). Cross-paper treatment of charge-as-length: Internally inconsistent (2019 papers vs 2023 OM paper).

---

Claim 3. Derivations of constants (c, k, alpha, Bohr radius, G, anomalous magnetic moment) show the model reproduces measurements

Step 1 said each checked case "reduces to a definition or Planck-unit identity". That is right for most, but two cases are not identities, and one case is wrong in a different way. See Corrections to Step 1.

a. Mathematics. Case by case. 1. G (Spacetime Eq. 5.8.1-5.8.7): C6. The starting point 5.8.1, F_g = mu0 c^2 (e^2/(4 pi r^2)) (alpha l_P^2/r_e^2), is the Coulomb force times a factor alpha (l_P/r_e)^2 chosen to be the "shading" geometry. Substituting the definition of r_e and l_P gives G m_e^2/r^2 identically. The factor is exactly the known ratio of gravity to electric force for two electrons (2.4006e-43), so the model's content is the postulate of that ratio, not its derivation. The result (l_P c^2/m_P) = G is a Planck-unit identity. And the equation is then applied to arbitrary masses, but it was derived only for electrons. 2. Compton wavelength (Eq. 5.7.2): C7, identity given r_e and alpha. 3. Anomalous moment (OM Eq. 8): C3 and C4. It is alpha/(2 pi) by identity and the measured value is 0.15% different. The paper's statement that the fine-structure constant "is anomalous and not required by this model" is not credible, since alpha enters through e^2/Q_p^2 and also through r_e. 4. Planck units "using h" (OM Eq. 4-7, 9-13): by construction Mp Lp^2/Tp = h and Mp Lp/Tp^2 = c-type relations return the inputs (checked: reproduces h and G exactly). The paper's own Section 3.1 admits the model "does not seem to work" with hbar; this is a choice of convention. 5. Alpha as geometry (Spacetime 2.2.2-2.2.4, read from page image p.8): alpha_2 = x y/(4 pi l^2 + pi d l + pi d^2), with x = y = d = r_e and l = pi r_e. The result is 1/(4 pi^3 + pi^2 + pi) = 1/137.0363. The measured 1/alpha is 137.035999. The agreement holds to 6 digits and stops there (C5; difference 2.2 ppm, alpha known to 0.15 ppb). The formula 4 pi^3 + pi^2 + pi is a long-known numerological near-miss [memory, medium]. The construction has free choices: the sphere radius and cone slant length are both called l and l is set to pi r_e "by hand", the cone radius d and rectangle sides are set equal to r_e, and the formula is a sum of three terms. With about five free choices of this kind, hitting a target to 6 digits is not surprising. Also, alpha is not constant: the effective alpha at the Z-boson mass is about 1/128 [memory, high]. A geometric ratio of fixed lengths cannot explain that running. The paper gives 0.00729734 and compares "to CODATA" at three digits, which hides the 2 ppm miss. 6. Avogadro's number (Eq. 5.4.1): a0/(2 l_P e) = 6.02236e23 (C8), 3.7e-5 off. N_A is the number of atoms in a mole, where the mole is tied to the gram. A different mass unit (say, pound) would change N_A while leaving a0/l_P untouched. A geometric ratio of lengths cannot equal a unit-dependent count. Also a0/l_P = m_P/(alpha m_e) by algebra, and the paper needs a free factor 2e to land near N_A. This is a coincidence of the choice of unit; also the result is not within error of the SI value (3.7e-5 versus a0 known to 1e-9, l_P (through G) to 1e-5). 7. Bohr radius, Rydberg energy, Bohr magneton: not checked beyond Step 1 (these sit in pp. 10-26, which I did not render). The Bohr radius is a0 = r_e/alpha^2, so given r_e and alpha it is an identity [memory, high]. 8. "Five constants" input (Spacetime p.2-3): Planck mass, length, time, charge and r_e. But r_e = alpha lambda_bar_C already contains alpha and m_e. So alpha and the electron mass are inputs, not outputs.

b. Experiment. All of the outputs agree with CODATA because the inputs are CODATA values rearranged. The two non-identity outputs (alpha at 2 ppm, N_A at 37 ppm) disagree with measurement at more than the experimental error.

c. Tested? Only by comparison with tabulated numbers. No new, independent prediction (no value that was not an input) has been identified in any paper I read.

d. What would change. Nothing, as stated. A genuine derivation of alpha from geometry would be a major discovery (no accepted theory derives it) and would need to explain its measured energy dependence.

Verdict (split). - "Derivations" of G, k, Compton wavelength, Bohr radius, anomalous moment, electron mass as independent reproductions: Refuted by calculation (they are identities of the inputs, C1, C3, C6, C7). - Alpha = 1/(4 pi^3 + pi^2 + pi): Refuted by calculation (C5; off by 2.2 ppm, roughly 10^4 standard errors). - N_A as a count of unit cells: Not testable as stated (unit-dependent; numerology at 3.7e-5). - OM "a_m = 0.0011614 is the anomalous moment": Tested, not supported (measured 0.00115965).

---

Claim 4. Planck's constant is the energy of one cycle; photon energy is an arbitrary 1 s accumulation; Modified Unit Analysis (MUA) removes quantum uncertainty

a. Mathematics. - The unit argument. E = h f with h in J s and f in 1/s gives J. Writing the "1" in 1/s as "one cycle" does not unbalance anything: a cycle is a count (an angle of 2 pi radians), the same status as the radian, which the SI treats as dimensionless [memory, high]. Adding a base unit "cycle" does not change a single numerical prediction. This is a notational proposal, and the claim "the units do not balance" is false in ordinary dimensional analysis. [read: V3 Sect. 2.] - "Hidden measure time". She writes h = h_Delta t_m with t_m = 1 s hidden. This introduces a number (1 s) into a constant of nature. Then h_Delta = 6.626e-34 J per cycle depends on the choice of second: change the unit of time to minutes and the "energy per cycle" changes by a factor of 60. Physical energy per photon of 500 THz light does not change with the unit of time. The relation E = h f with f in cycles/s is invariant, the per-cycle number is not. - Consistent reading (a test on its own terms). If each cycle carried h_Delta, a pulse of N cycles carries N h_Delta, so its energy depends on duration: a 10-cycle and a 10,000-cycle pulse of the same colour would differ by 1000. Standard photoelectric data say the energy transferred per photon depends only on f (Millikan 1916) [memory, high]. - MUA example (C12). A = P/(hf) and F_s = hf/c multiply to P/c. It reproduces the textbook number because it is algebraically the same expression. The paper's labelling "perfect reflector" with F = P/c is itself off by a factor 2 (reflector: 2P/c). This is also in Step 1's notes [memory, high]. - The uncertainty argument (V3 Sect. 6, Eqs. 13-17). Starting from Delta E Delta t >= h, she replaces h by h_Delta Delta t (with t_m := Delta t) and cancels Delta t to get Delta E >= h_Delta = 6.626e-34 J. Two errors. (1) She set t_m = 1 s earlier and now t_m = Delta t, an arbitrary variable. (2) The correct relation is Delta E >= h/Delta t (up to a factor 1/(4 pi)); the energy resolution improves with longer measurement time, so there is no floor. The conclusion "no energy change smaller than 6.626e-34 J can be detected" would put a floor 2000 times above what optical clocks resolve (C13) [memory, medium]. - The further inference of a minimum temperature 4.8e-11 K (V3 Sect. 7). Her own text notes MIT reached 5e-10 K "one order of magnitude higher". A 2021 microgravity experiment reached 38 pK = 3.8e-11 K, below her floor [web: PRL 127, 100401]. - LC paper "quantum of momentum h/c = 3.5177e-43" is hbar/c (C9). So within MUA, where h is the per-cycle quantum, the paper uses the reduced quantity under the label h. It also uses "reduced Planck charge", while the OM paper states the model "does not work with hbar". - The "light as in-phase vs quadrature" argument (V3 Sect. 10-11, "Figure 4 is plane wrong"). She argues that in every oscillator potential and kinetic parts are 90 degrees out of phase, so light's E and B must be 90 degrees out of phase, and Maxwell et al "made a mistake". Check: for a plane wave E_x = E0 cos(kz - omega t) travelling along z, Faraday's law dB/dt = -curl E gives B_y = (E0/c) cos(kz - omega t), in phase with E. If E and B were in quadrature the time-averaged Poynting flux (E x B/mu0) would be zero: no net energy transport. Her own power-line analogy (P = V I) has the same property: V and I in phase for a travelling wave on a line (V/I = Z), in quadrature only for a standing or reactive exchange (average power zero). So her correction contradicts the property of a travelling wave she wants to model. [This section is in V3; I read it as text; the claim about Maxwell's mistake is in her Appendix B, which I did not read.]

b. Experiment. The photoelectric effect (threshold depends on f, not on intensity), Compton scattering, and single-photon counting each show quantised energy hf per photon. A constant energy per cycle fails the pulse-duration test above [memory, high].

c. Tested? The standard statement has been tested for over a century [memory, high]. The MUA claim has made no independent measurement or prediction: its one worked example agrees by construction. Not replicated by anyone I can cite; no source found for any independent test.

d. What would change. The photon concept, the SI definition of h in J s, and the uncertainty principle (derived from the commutator of position and momentum, not from units) would all fail. The proposal would also have to reproduce blackbody radiation, which is derived from quantised hf per mode.

Verdict (split). - "Units do not balance, add a base unit cycle": Not testable as stated (notation; no measurable change). - Constant energy per cycle, photon as 1 s accumulation: Tested, not supported (photoelectric effect and duration argument). - Removal of uncertainty / minimum detectable energy 6.626e-34 J / minimum temperature: Refuted by calculation (error in the cancelling of Delta t; counterexamples C13 and the 38 pK result). - LC and "out of phase" correction of Maxwell: Refuted by calculation (Faraday's law; zero average Poynting flux).

---

Claim 5. Michelson-Morley does not exclude an aether; GW150914 plus the Fermi transient is an aether P-wave then S-wave; future LIGO events should show the same follow-up

a. Mathematics. In a solid, P (pressure) waves travel at a speed v_P fixed by the elastic constants, and S (shear) waves at a slower v_S, with ratio fixed by the material. Given a source distance d, the gap between them is d (1/v_S - 1/v_P), proportional to d. For GW150914 (about 1.3e9 light years) a 0.4 s gap means a fractional speed difference of 1e-17 (C16). Applied to another source at 1.3e8 light years, the same medium gives about 0.04 s. For GW170817, the observed gamma-ray follow-up was 1.74 s after the gravitational wave [web: arxiv.org/abs/1710.05834], 40 times the scaled value, and the gravitational wave came first. The paper's visual comparison between the LIGO strain trace and a seismic P-wave is a similarity of pictures, not a fit (no model waveform, no statistics). The "P-wave" is inferred only from a chirp that LIGO matched to a numerical-relativity template of a merger [memory, high]. Also, Step 1's wording needs correcting: the paper says that if all future signals are followed by gamma rays "the aether idea should be revisited", a conditional, not a prediction that they will be.

b. Experiment. The Fermi transient had 2.9 sigma significance, was not seen by other instruments, and had poor localisation [web: arxiv.org/abs/1602.03920]; the authors themselves treat chance coincidence as possible. Later BBH events (dozens in the O1-O3 runs) without reported gamma counterparts are [memory, medium], not verified here. GW170817 shows light and gravity speeds equal to within a few 1e-15 [web]. The circular-reasoning argument about MM is dealt with in claim 1b.

c. Tested? Yes, for the MM part (many independent tests, all null; see claim 1). The P/S claim is testable and has made a testable statement (conditional). Not supported by anything cited, and I found no sourced report of its confirmation. I cannot say how many follow-ups were searched for without a source.

d. What would change. The nature of LIGO detections as tensor spacetime ripples (GW polarisation, matched to general relativity templates) and special relativity's single invariant speed.

Verdict (split). - MM null "explained by circularity": Tested, not supported (independent tests of isotropy). - GW150914 and Fermi as P/S aether waves: Refuted by calculation (distance scaling vs GW170817; C16), and the underlying coincidence is weak (2.9 sigma) in its own paper. - "All future LIGO events should show the gamma follow-up": Tested, not supported [memory, medium; unverified count].

---

Claim 6. Particles are standing-wave clusters of Planck wave centres (neutrino fundamental); decay is collapse; Mandelbrot set models black holes (pear-shaped photon ring)

a. Mathematics. Particles (Geometry of Particles, sampled text only; equations not extracted): - Wave-centre counts K = 8, 20, 28, 50 are said to "match nuclear magic numbers". The magic numbers (2, 8, 20, 28, 50, 82, 126) come from nuclear shell structure with spin-orbit coupling [memory, high]. Leptons are not nuclei and no mechanism connects them. A match of 4 numbers from a short list of small integers carries almost no weight. - No lifetime, branching ratio or mass ratio is computed in what I read. I have no calculation to check. - A "fundamental" neutrino conflicts with masses: the electron is 0.511 MeV, neutrinos below about 1 eV (direct bound from tritium beta decay is below about 2 eV [memory, high], KATRIN lower still [memory, medium, unverified]), so composites of neutrinos would be lighter, not heavier, unless binding energy is hugely positive; no calculation supplies it. Mandelbrot paper (read abstract, intro, sections 4-6, 15-16): - The set M is the set of complex numbers c for which z -> z^2 + c, started at z = 0, stays bounded. The paper says points in the black region "collapse past the digits of precision of the computer", that periodic orbits in the set are floating-point artefacts, and "aside from (0,0) and (-1,0) there are no stable periodic orbits". That is wrong. In the main cardioid every c has an attracting fixed point (exact: z = (1 - sqrt(1-4c))/2 with multiplier 2z, |2z*| < 1), and in every "bulb" an attracting cycle. C15 reproduces this with 60-digit arithmetic: convergence is to a nonzero limit and exact. The core "analogy" (collapse to infinitely small singular regions) therefore rests on a wrong fact about M. - The "photon sphere" grey region and "undecided" halo are set by the user's choices of maximum iteration count and escape radius (any radius >= 2 gives the same set M but different grey bands). They are properties of the picture, not of M. A GR photon sphere at r = 3GM/c^2 is an unstable circular photon orbit surface, outside the horizon at r = 2GM/c^2; nothing in M has that meaning. - The M87 overlay: translation, scale, rotation and transparency are adjustable, and the number of iterations (2) is chosen. "The black hole overlays exactly the (0,0) point" has no error estimate. A hand alignment that can be adjusted is not a test.

b. Experiment. Quarks inside protons are supported by deep inelastic scattering (SLAC, 1968 onward) [memory, high]; no quantitative contradiction can be computed because the cluster model gives no cross-section. The EHT image of M87 (2019, ring about 42 microarcseconds) and Sgr A (2022) are close to circular with brightness asymmetry explained by Doppler beaming [memory, high; not web-verified here]. The Ra-224 pear-shaped nuclei exist (Gaffney et al. 2013) [memory, high], but that is a nuclear deformation of order 1e-15 m and has no relation to black holes. Atoms are not black holes: the Schwarzschild radius of an atom's mass is about 1e-52 m, vastly less than its size [memory, high; 2 G m/c^2 with m about 1e-26 kg].

c. Tested? Particle clusters: never confronted with a quantitative test; no output to test. Mandelbrot prediction ("oblate or slightly pear-shaped, depending on how deep we see"): stated so loosely that nearly any image fits; EHT results exist after the paper was written (it says the data were not yet released), and showed near-circular rings; the paper did not revisit. I could not find a sourced follow-up from the author [unknown].

d. What would change. The Standard Model (quarks, leptons as separate fundamental fields, lepton number), and general relativity's black hole shadow shape.

Verdict (split). - Particle-cluster model with neutrino as fundamental unit: Never properly tested (no quantitative output; it was not dismissed by a test either) with a note that the magic-number match is Not testable as stated (numerology). - Mandelbrot-set "analogy" premises (no periodic orbits; iteration artefacts as photon sphere): Refuted by calculation (C15). - Pear-shaped photon ring and M87 position: Not testable as stated (loose prediction; hand overlay), with the EHT near-circular result pointing against it [memory].

---

Claim 7. Side positions: electric charge obsolete and current is a diffusion wave of "electrigen"; dielectric and magnetic fields are exact duals; free space is an LC circuit with impedance Z0/(2 pi)

a. Mathematics. - Steinmetz duality (read in full): flux = L i, Psi = C e, stored energy L i^2/2 and C e^2/2. These are standard circuit theory and correct. A real duality exists: in source-free regions Maxwell's equations are unchanged by (E, cB) -> (cB, -E). "Exact" duality fails where charge exists: div D = rho_free but div B = 0 (no magnetic monopole observed) [memory, high]. The statement "dielectric lines are tension, magnetic lines pressure": in Maxwell's stress tensor both kinds of field have tension along the lines and pressure sideways [memory, high], so that dichotomy is wrong. The paper presents Steinmetz's words as paraphrase, so it is his claim as written in 1914. - LC paper: C10 and C11. The impedance 59.96 ohm is Z0/(2 pi) because the paper multiplies the definition eps1 = 4 pi eps0 by an extra factor 2 in the capacitance and inductance. The statement "only one 2 pi can appear, so Z0/(2 pi) is the correct impedance of free space" does not follow: Z0 is the measurable ratio of electric to magnetic field strength in a plane wave (E/H); redefining eps and mu by 4 pi changes the label, not E/H (still 376.73 ohm in SI). Cross-paper inconsistency: the same author gets Z0/(2 pi) here and Z0/(4 pi) in the OM paper (C10). "Resonance" omega^2 = 1/LC is an identity (C11). The "quantum of momentum h/c" is hbar/c (C9). - Complex numbers paper: complex numbers as 2x2 real matrices a I + b J with J = [[0,1],[-1,0]], J^2 = -I, is the standard isomorphism (checked: matches the paper's Eqs. 6-7). Euler's formula exp(theta J) = cos(theta) I + sin(theta) J is correct. The claim that this removes the need for an imaginary unit is not right: J^2 = -I is the same equation, and J is a 90-degree rotation, which was already the geometric meaning of i. As pedagogy it is fine. - Electrigen paper (Dinu): no equations, no numbers; not checkable mathematically.

b. Experiment. Electrigen: the Tolman-Stewart experiment (1916) measured the charge-to-mass ratio of conduction carriers in metals, consistent with electrons [memory, high]. Thomson's e/m, Millikan's oil drop and the Hall effect show negatively charged light carriers; a Group 0 neutral element has no such measured signature [memory, high]. Mendeleev's 1902-3 "x, y" elements are a documented speculation [memory, medium]. A liquid aether supporting only longitudinal waves cannot support polarisation of light (light has two polarisations, as measured), a conflict stated in Step 1 and one I accept. Steinmetz duality: consistent. Z0/(2 pi): the value of Z0 is used in antenna and optics measurements as 376.73 ohm [memory, high].

c. Tested? Electrigen: the carriers measurements above, not as a test of electrigen specifically; I found no test designed around it. Duality and Steinmetz: textbook. LC impedance: no measurement predicted.

d. What would change. Electrigen would need the periodic table, Maxwell's equations (charge density source term), and the electron as charge carrier to be revised. Z0/(2 pi) changes nothing measurable.

Verdict (split). - Duality of electric and magnetic fields (circuit/source-free sense): Established. "Exact" and tension/pressure statements: Internally inconsistent with Maxwell's equations as stated (monopole absence; stress tensor). - Free space impedance Z0/(2 pi) as the "correct" value: Internally inconsistent (Z0/(2 pi) vs Z0/(4 pi) across papers; tautological resonance) and otherwise Not testable as stated (unit convention). - Charge obsolete, electrigen, diffusion wave: Tested, not supported (carrier e/m measurements, Hall effect; the electrigen idea itself was never given its own test, and I found none). - Complex numbers as matrices: standard mathematics, Established (the "no imaginary unit needed" gloss is not accurate).

---

Summary table

What would settle the main open points

• Claim 3: any one output of the Spacetime model that is not an input (a number nobody has fed in), computed in advance and compared with measurement. Alpha to 9 digits from geometry would do.
• Claim 4: a single-photon-counting experiment on the same light with two pulse durations (the model says energy scales with the number of cycles; the standard view says it does not).
• Claim 5: the distance dependence of any gamma-ray lag across many gravitational-wave events (a P/S medium predicts lag proportional to distance).

---

Corrections to Step 1

1. LC paper, quantum of momentum. facts.json and claims.md carry "h/c = q_pl^2 mu1 = 3.5177e-43" as stated by the paper. The number is hbar/c; h/c is 2.2102e-42 (C9). The paper mislabels it, and the LC paper uses the reduced Planck charge while the 2023 OM paper says the model only works with h. Add to the LC and OM entries.
2. Claim 7 impedance. claims.md gives free-space impedance Z0/(2 pi) only. The 2023 OM paper gives Zi = Z0/(4 pi) = 29.979 (its Eq. 3 and the text beside it). The author's two impedances differ by a factor 2 (C10).
3. Claim 3 review note ("each checked case reduces to a definition or Planck-unit identity"). Too broad. The alpha "geometric ratio" and the Avogadro result are not identities: 1/(4 pi^3 + pi^2 + pi) = 1/137.0363, which differs from the measured value by 2.2 ppm (C5), and a0/(2 l_P e) is 3.7e-5 off N_A (C8). These are numerological fits with free choices, not definitions. The facts.json Spacetime entry also says the fine-structure constant "as a geometric ratio ... equal to alpha when all lengths equal r_e". Replace "equal to" with "numerically close to (2 ppm), not equal". Step 1 also writes 5.8 "gravity derivation ... cancels to F = ... = G identically", which I confirm; the factor alpha l_P^2/r_e^2 is itself the known gravity-to-electric ratio, so the postulate carries the content (C6).
4. Claim 5 wording. claims.md says "all future LIGO events should show the same gamma-ray follow-up". The paper says "if all future gravitational waves are followed by gamma-ray signals the aether idea should be revisited". A conditional, not a prediction that they will.
5. Retagged [memory] items. "Fermi GBM transient 0.4 s later, claimed by Connaughton et al. 2016" is now [web] (arxiv.org/abs/1602.03920): 2.9 sigma, about 1 s, not seen by others. Fermi GRB 090510 LIV limit now [web] (arxiv.org/abs/0908.1832); Step 1's own note that it was unverified is resolved. INTEGRAL non-detection remains [memory], medium, unverified.
6. NaCl figure. Step 1 says the arithmetic gives "about 3699 m/s". With standard atomic masses I get 3667 m/s (C2). The paper's own value is 3700; the difference is 1% and does not change the conclusion.
7. Minor, new items Step 1 did not mention. The 2023 paper's statement that Km = mu0/(4 pi) is "exactly 1e-7" was true under the pre-2019 SI; since the 2019 redefinition mu0 is measured (relative uncertainty 1.6e-10) [web: NIST]. The Spacetime paper's "Planck-mass granule of radius l_P" lies inside its own Schwarzschild radius (C14). The V3 section on E and B being in quadrature contradicts Faraday's law (claim 4). The Mandelbrot paper's statement that periodic orbits are floating-point artefacts is false (C15). None of these changes a Step 1 verdict-relevant fact, but they bear on claims 4 and 6.