Dunning-Davies (with R. L. Norman): Science Contrarian review

Reviewer role: Science Contrarian (judges by experiment; separates "tested and failed", "never properly tested" and "dismissed without testing"; holds the researchers and the critics to the same standard).

Source gap (repeated from Step 1, partly narrowed). The 2024 Hadronic Journal paper (NormanDunning-Daviescopy.pdf) has no text layer and Step 1 viewed only about 10 of 29 page images; I did not read it further. "Gravity v.2" appendices (pp. ~20-69) were not read in detail by Step 1 or by me. I read myself: the full 2020 quaternion preprint (text plus page images p.10-11), the Gravity v.2 front section (pp. ~1-20), the "Consequences of an Aether" tunnelling/hadronic section, and all three pages of the CIA memo. Where a verdict below rests on [memory] alone, it says so. The web budget was used on five checks; a sixth (Santilli neutron synthesis) was refused because the search quota was exhausted, so that item stays [memory].

What the evidence base actually is

Nothing in this folder reports a measurement made by the authors. Dunning-Davies and Norman offer derivations, quotations from other people's reports (Sweet and Bearden, Laithwaite, Schauberger via Callum Coats, T. Townsend Brown's own 1952 proposal, Santilli), and an unreviewed-by-mainstream publication channel (Hadronic Journal; ResearchGate preprints) [read]. Every experimental claim is therefore second-hand. The strongest honest statement for the project is therefore "claims relayed, not tested", and the question for each claim is whether anyone neutral ever tested the original.

Claim 1: A seventh scalar field T exists (quaternion Maxwell, "Temporal Field")

a. Mathematics. The paper defines T = -(1/c) dphi/dt + div A (eq. 1) with E = -grad phi - (1/c) dA/dt and B = curl A [read: page image p.10]. I checked eqs. (3) and (5) symbolically (SymPy, [calculation]). They are not identities of the definitions. Substituting the potentials into curl B - (1/c) dE/dt - grad T leaves -laplacian A + (1/c^2) d2A/dt2 + (2/c) grad(dphi/dt), which is a constraint on the potentials, not an identity. The standard Lorenz combination is L = div A + (1/c) dphi/dt (plus sign); with L the residual is just the ordinary wave operator on A, as expected. The authors' T has the opposite sign, so it does not vanish in the Lorenz gauge (their eq. 7, which is the standard Lorenz condition [read]); that is why they find T = 2 div A != 0. Two further problems [calculation]: (i) the plus-sign combination div A + (1/c) dphi/dt is the Lorentz-invariant four-divergence; the authors' minus-sign T is not, so it is not a relativistic scalar; (ii) under a gauge change A -> A + grad(lambda), phi -> phi - (1/c) dlambda/dt, T shifts by laplacian(lambda) + (1/c^2) d2lambda/dt2, so T is gauge-dependent: it is bookkeeping about the choice of potentials, not a measurable field. Step 1 flagged the sign difference; the gauge and invariance points are new. b. Experiment. Vector Maxwell theory with Lorenz or Coulomb gauge has passed every EM test; there is no reported measurement that needs a seventh component [memory: high; no web check made]. Because T is gauge-dependent, no experiment can distinguish it from a gauge choice. c. Tested? No test by anyone. The paper cites the CIA memo and Bearden books as if they established it; neither contains data [read: memo p.1-3]. Jack's 2003 arXiv note is a notation paper [memory: medium; cited in refs]. Dismissed without testing? Not really: it was never put in a testable form (no observable stated). d. If true. Gauge invariance of EM would fail for a physical field, and the Lorentz-invariant Lorenz condition would not be the natural scalar. Charge conservation (which makes the potentials gauge-redundant) would have to be modified. Strongest honest case. The authors are right that Poynting's vector is ambiguous up to the curl of an arbitrary vector (the memo says so too [read: memo p.1]; standard result [memory: high]), and that Maxwell's Treatise used quaternion expressions in places (claimed in the paper; [memory: medium]). Quaternion rewriting is legitimate mathematics. But those facts do not produce a new field. Verdict: Internally inconsistent (as derived, eqs. 3 and 5 are constraints, not consequences, and T is gauge-dependent and non-invariant). Weak form (a scalar combination of potentials can be written down): Established but trivial/gauge, not a new physical field.

Claim 2: Scalar waves exist, carry energy, and travel at c*sqrt(3)

a. Mathematics. The pure-gauge solution A = grad S, phi = -(1/c) dS/dt gives E = grad(S_t/c) - grad(S_t/c) = 0 and B = 0 [read: memo p.2; calculation]. That is a gauge transformation of the zero potential, so nothing physical propagates. Standard EM energy density (E^2 + B^2)/(8 pi) or equivalent is then identically zero, so "carries energy" is refuted by calculation within the theory the authors invoke. The csqrt(3) derivation [read: p.10-11 images]: with (7) and T = 2 div A = 2 laplacian S, grad(laplacian S - S_tt/c^2 + T) = 0 gives laplacian S = S_tt/(3 c^2), speed csqrt(3). Three self-contradictions [calculation, SymPy and hand algebra]: (1) eq. 7 with A = grad S and phi = -S_t/c already gives laplacian S = S_tt/c^2; combined with the csqrt(3) equation this forces laplacian S = 0 and S_tt = 0 (no wave). (2) The paper's own phi equation, laplacian phi - phi_tt/c^2 = -(1/c) T_t with T = -(2/c) phi_t, gives laplacian phi = 3 phi_tt/c^2, i.e. speed c/sqrt(3), slower than light, from the same equations. Since phi = -S_t/c is the paper's own relation, phi and S cannot both obey their equations unless S_ttt = 0. (3) The step "if the constant of integration is zero" is an admitted assumption [read]. This is new relative to Step 1 (Step 1 found contradiction (1) only). b. Experiment. No detection of superluminal EM or scalar signalling [memory: high; no web check]. Superluminal signalling conflicts with relativity tests; and for gravity the direct constraint is [web] https://arxiv.org/pdf/1710.06394 (GW170817 versus GRB 170817A: speed difference between -3e-15 and +7e-16 of c). c. Tested? The memo itself says: "Their existence is not proven, but if they exist their energy source is not clear" [read: memo p.3]. The authors and Bearden-circle sources present no detector. Tested and failed? No dedicated test found. Never properly tested for the pure-gauge object because it has no observable. d. If true. Relativistic causality, the Poynting theorem and energy conservation. Strongest honest case. The CIA analyst was cautious and fair (author unknown, redacted); "extension to Maxwell's understanding" is a respectable question. Longitudinal modes are real in plasmas, which the memo notes [read: memo p.3], and that is not what the authors claim. Verdict: Superluminal speed csqrt(3): Internally inconsistent (contradicts eq. 7 and its own phi equation, which gives c/sqrt(3)). Energy-carrying pure-gauge scalar wave: Refuted by calculation (E = B = 0 gives zero field energy; the Poynting-ambiguity loophole is real but gives no energy source). Existence of a physical longitudinal EM scalar mode in vacuum: Never properly tested (no neutral experiment on record; not tested because no observable is specified).

Claim 3: A gas-like aether exists; use total derivative D/Dt; relativity valid only in vacuum

a. Mathematics. Thornhill's statement (quoted [read: p.13]) is that Ampere's and Faraday's laws must use D/Dt = d/dt + u.grad in a gas-like ether. This breaks the Lorentz covariance of Maxwell's equations by design: it makes the equations depend on a medium velocity u. Whether Maxwell with D/Dt is self-consistent for a moving medium is a legitimate classical question (it is the Hertz-type convective derivative). But the authors then say that "away from mass, the medium is stationary and the partial derivative returns, so longitudinal waves travel near-instantaneously" [read: p.13-14]. No equation is given that produces near-infinite speed. A statement from Wal Thornhill (a popular writer, not a peer-reviewed source) is used as the evidence. Internal consistency can't be checked because no derivation exists. b. Experiment. Michelson-Morley (1887) and later Kennedy-Thorndike, Ives-Stilwell and modern optical-cavity tests found no preferred frame [memory: high; no web check]. These were designed to catch exactly a gas-like ether with velocity varying from point to point. The "aether is dragged near masses and stationary far away" move is the usual way to save the idea, and the authors do not give a quantitative drag model that can be compared with them. c. Tested? Ether-drift tests have been repeated by many groups; the null result is not in dispute, including by Dayton Miller's critics (Miller's positive claim was reanalysed [memory: medium]). This is not a case of dismissal without testing. d. If true. A preferred frame; relativity valid only approximately; the constancy of c would need to be only apparent. Strongest honest case. "Vacuum as a structured medium" (Casimir effect, zero-point energy, QED vacuum) is mainstream language [memory: high]; the authors cite it. That is not a material gas with Maxwellian statistics. Relativity's restriction to vacuum is also a fair reminder: the measured speed of light in a medium differs from c, but SR does not claim otherwise. Verdict: Gas-like aether with D/Dt and near-infinite longitudinal waves: Tested, not supported (ether-drift and Lorentz-invariance experiments, [memory: high], unchecked on web). The rhetorical claim "SR only holds in vacuum": Not testable as stated in the sense used (SR already is stated for vacuum). Historically, the aether was a historically superseded but instructive position.

Claim 4: Gravity is a superluminal longitudinal scalar wave coupling to EM; antigravity, variable mass and over-unity power have been achieved

Split into five sub-claims.

4A. Gravity propagates much faster than c. a. Whittaker's 1903 paper is correctly summarised in the preprint: the potential 1/r can be written as a superposition of spherical waves sin(mu(Vt - r))/r with any V; the sum is static [read: p.4-5]. The mathematics is sound, but V is arbitrary, so it proves nothing about the actual speed: it shows that a static field can be represented as waves, and Whittaker says V "need not be the same as that of light" [read]. Using this as support for superluminal gravity is a non sequitur. b. GW170817 and GRB 170817A [web: https://arxiv.org/pdf/1710.06394]: gravity-light speed difference within about 1e-15 c. That is a direct measurement of the speed of gravitational waves; it rules out "far above c" for waves. Van Flandern's argument (instantaneous gravity for orbit stability) has a mainstream answer: in GR the velocity-dependent terms cancel the aberration expected from a retarded static-looking force [memory: medium; to my knowledge Carlip 2000; not checked on web]. Binary pulsar decay matches GR's quadrupole emission [memory: high; Hulse-Taylor; not web checked]. c. Tested by LIGO/Virgo/Fermi, many groups, with multi-messenger replication. Not dismissed without testing. d. If true: GR, and the observed gravitational-wave signal timing, would have to be reinterpreted. Verdict: Tested, not supported.

4B. Lensing and redshift are refraction; dark matter is unnecessary. The paper's argument is that "Einstein rings are blue, therefore diffraction" [read: p.2, quoted from a secondary source]. Refraction is chromatic; observed lensing is achromatic to good precision [memory: medium]. The authors do not give an equation for a refractive index profile that reproduces the deflection angle or time delays. Not supported / never quantitatively tested by the authors; mainstream lensing is quantitatively tested. Verdict: Tested, not supported (weak support only; the authors' own version is not specified enough to test, so also Not testable as stated in its quantitative form).

4C. Laithwaite's gyroscope loses weight. Laithwaite did give a demonstration at the Royal Institution in 1974 lifting an 18 kg spinning disc, and claimed it violated Newtonian mechanics; most of the scientific community did not accept this and the lectures were not published at the time [web: https://en.wikipedia.org/wiki/Eric_Laithwaite via search summary; https://www.rigb.org/explore-science/explore/blog/gyroscope-engineer-and-christmas-lectures]. Ordinary gyroscopic torque and precession account for the demonstration [memory: medium; not verified on web in this review]. Here the contrarian points out a real process issue: the lectures were withheld from publication. But the claim is a mass reduction, and no balance measurement of weight change is shown in the sources I read. Verdict: Tested, not supported by balance-type tests found in neither the sources nor memory with high confidence; the demonstration as described is explained by standard mechanics [memory: medium]. I could not confirm a controlled mass-balance test; I label this less certain than other verdicts.

4D. Sweet/Bearden: open-loop gain 5e4 to 1.5e6 and 90% weight reduction with 175 microwatts in, 1 kW out. First, a numbers check [calculation]: 1000 W / 175e-6 W = 5.7e6, which is 3.8 times larger than the top of the quoted range of gain (1.5e6). The quoted abstract is therefore not internally consistent about its own gain [read: p.4, quoted from Sweet and Bearden]. Second, controls: Sweet never had the device certified by independent testing, and the one replication reported (Michael Watson, England) was unsuccessful; Sweet kept his magnet-conditioning process secret [web: search summaries from https://www.rexresearch.com/sweet/1nothing.htm and https://groups.google.com/g/sci.energy/c/CdeKzCe451U; these are fringe-community pages, not neutral sources]. So it is neither "tested and failed" by a neutral lab nor accepted: it is never properly tested, and the reason is largely the inventor's secrecy, not dismissal by mainstream science. No data (traces, calibrations) appear in the folder. Energy conservation: if the claim were real, the output energy must come from somewhere; the authors posit the vacuum, but no energy budget is offered. Verdict: Never properly tested (not dismissed without testing by mainstream, since mainstream never received a testable device; the inventor withheld it). Internally, the quoted gain figures are inconsistent.

4E. Brown electrogravitics; Schauberger self-sustaining turbine (15,000-20,000 rpm, takes off). Evidence is a quotation from Callum Coats's book, a 1936 statement by Schauberger himself ("we can create anything out of nothing") and Brown's 1952 proposal [read: Gravity v.2 pp.~10-12]. Brown's text says classified work "precluded publication"; he wrote this proposal to obtain funding, so it is a self-interested source. The NATO AGARD 1990 reference is described by the authors; I did not obtain the report [memory: low; unverified]. No neutral test of Schauberger's turbine is on record in these documents. Verdict: Never properly tested (no controlled neutral test in the record; unverified against external literature).

Overall verdict for claim 4 (split): superluminal gravity Tested, not supported; refraction lensing Tested, not supported; Laithwaite Tested, not supported (medium confidence); Sweet, Brown, Schauberger devices Never properly tested; "over-unity" as a general energy claim Never properly tested as an experimental claim, while the underlying principle conflicts with energy conservation, which has no known exception [memory: high]. Not Refuted by calculation because no measured power budget exists to apply the law to. What would have to change if true: GR (wave speed, lensing), conservation of energy, and the equivalence principle.

Claim 5: Quantum and nuclear phenomena have a causal aether basis; neutron is a proton-electron bound state; no quarks or neutrinos

a. Mathematics. The tunnelling "explanation" is verbal: "scalar waves may be understood as voltage over time expressions", with no equation, no transmission coefficient and no predicted number [read: Consequences of an Aether, section "Some Further Thoughts"]. The existing WKB/Schrodinger treatment predicts tunnelling currents and alpha-decay half-lives quantitatively (Gamow) [memory: high]. The claim that the Schrodinger equation "does not admit positive binding energies" for an electron in a proton [read] is not shown; the relevant calculation is a constraint, not an obstruction. A confinement calculation [calculation]: for an electron localised to nuclear scale r ~ 1 fm, momentum p ~ hbar/r, so pc ~ 197 MeV fm / 1 fm ~ 200 MeV (about 100 MeV even using hbar/2r). That is over 100 times the 0.782 MeV claimed binding scale (m_n - m_p - m_e = 0.782 MeV [memory: high]). Under standard quantum mechanics a neutron made of a proton and an electron cannot be bound by 0.78 MeV. Santilli's answer is that standard QM does not apply inside the hyperdense proton and replaced it with hadronic mechanics; that is a different theory, not a derivation within accepted law. b. Experiment. Loophole-free Bell tests (Delft, 2015) violated the inequality with electron spins 1.3 km apart [web: https://www.nature.com/articles/nature15759]; further loophole-free tests with photons (NIST, Vienna) in 2015 and superconducting circuits in 2023 [web: https://www.nature.com/articles/s41586-023-05885-0]. These rule out local hidden-variable models, which includes the Mayants "abstract versus concrete" reading, if that reading is local realism. If the authors accept non-locality via superluminal scalar waves, Bell is not a refutation of their picture, which is a fair point; but then they must say how a superluminal channel avoids signalling. Neutrinos: detected directly (Cowan-Reines 1956, and later Super-Kamiokande and SNO flavour oscillation) [memory: high; no web check]. Quarks: deep inelastic scattering [memory: high]. c. Tested? Standard QM, neutrino physics and Bell tests: thoroughly tested and replicated by independent groups. Santilli's neutron synthesis from hydrogen: experiments claimed in the Hadronic Journal and by his own institute; independent confirmation by mainstream nuclear groups is unknown to me [memory: low; web check was refused for quota]. Mind-matter double-slit and random-number-generator claims (Radin, Jahn, Tressoldi [read: Enigma docx]): I did not verify the literature; there is no accepted replication I can cite [memory: low]. I won't assert either way. d. If true. Quantum mechanics would need to be replaced by a deterministic aether theory with new nuclear mechanics, neutrino physics and the quark model. Strongest honest case. Quantum foundations are not closed: hidden-variable theories with non-locality (de Broglie-Bohm) survive Bell [memory: high]; the mass-deficit/positive-binding oddity of the neutron decay is a good exam question; the authors correctly note mainstream theory gives no intuitive mechanism for tunnelling. These are interpretation questions; the authors' specific scalar-wave voltage account makes no predictions. Verdict (split): tunnelling/strong force/Coulomb barrier via scalar-wave interference: Not testable as stated (no equations or predictions; it competes with a quantitatively successful theory). Bell non-locality as misreading: Tested, not supported for local models (loophole-free, replicated [web]). Proton-electron neutron under accepted QM: Refuted by calculation (confinement energy about 100-200 MeV vs 0.78 MeV; caveat: hadronic mechanics is a different theory). Absence of neutrinos and quarks: Tested, not supported [memory: high]. Mind-matter on interference: Never properly tested to a neutral standard in the record [memory: low; unverified].

Claim 6: Hadronic mechanics yields practical clean energy (waste neutralisation, transmutation, PlasmaArcFlow)

a. Not checkable: the 2024 paper sections were only partly viewed by Step 1, and I did not read the remaining pages. The quoted statements (resonant photons stimulating electron capture; PlasmaArcFlow reactors "already in industrial operation", atmospheric nitrogen synthesis, earth-core heat from transmutation) are assertions with no energy balance in what I saw. b. Electron capture is a real weak process [memory: high], but resonant-photon stimulation of it at useful rates is not part of accepted nuclear physics [memory: medium; unverified]. c. Tested by Santilli's own organisations; independent replication by neutral national labs or university groups: none known to me [memory: low]. The web check failed (quota exhausted), so I do not state that independent tests did or did not occur. Dismissed without testing? Unknown. d. If true: the nuclear decay rates would be controllable by photons; the neutron structure would change. Verdict: Never properly tested by a neutral party on the record available to me (claim rests on [memory]/unverified; I could not complete the web check).

Claim 7: This science is used and concealed by government/military programmes

The documents cited do not show that. The CIA memo is a three-page summary memo on request, ending "Their existence is not proven, but if they exist their energy source is not clear", and suggesting an "implied Tesla connection" with part of paragraph 2 redacted [read: memo p.1-3]. It is evidence that an intelligence analyst in 1987 looked at the scalar-wave literature and Bearden's name, and nothing more. The Brown 1952 statement is from a funding proposal. The 1954 quote ("the ideas are secret") is from an edited collection by Thomas Valone [read: p.3, quoted]. The paper itself says that "the working quaternion mathematics in question is omitted, presumably for purposes of weapons' design secrecy" [read: Axioms p.2], yet then derives its mathematics from an open 2003 arXiv paper (Jack) and gets self-contradictory results; that sits badly with the claim that working mathematics is secret and valid. Unfalsifiable form: "concealed" explains the absence of evidence. Applying the same rigour to critics: the contrarian notes that dismissal of Laithwaite (lectures not published) and Sweet (no certified test, but he did not allow one) are documented sociological facts, and mainstream bodies rarely run controlled tests of fringe claims; "dismissed without testing" is a fair description for the electrogravitics line generally. But the existence of a working concealed technology is not supported by anything in the folder. Verdict: working concealed technology: Not testable as stated (unfalsifiable) and not supported by the cited documents. Narrow factual form (an analyst wrote a 1987 memo about scalar waves): Established [read].

Summary table

What the contrarian would still want tested

1. A bench test of "scalar wave" energy: transmit a pure-gauge E = B = 0 configuration and look for any received power. By the authors' own account it should be zero Poynting flux, so a null is the expected result; a non-null would be anomalous.
2. A neutral, instrumented replication of Sweet's device (calibrated input and output power, independent magnets) if it can be obtained; none exists on record.
3. A balance and torque-controlled test of spinning-mass weight change (Laithwaite), with mass measured on a force plate over a range of rpm.
4. A speed-of-gravity test is already done (GW170817); no further test needed.

Corrections to Step 1

• Step 1 recorded the GW170817 speed bound as [memory: medium]; it is now [web] https://arxiv.org/pdf/1710.06394 (difference between -3e-15 and +7e-16 of c). The Bell test claim (2015 loophole-free) is now [web] https://www.nature.com/articles/nature15759.
• Step 1's own check found the contradiction between eq. 7 and the c*sqrt(3) result (verified correct). Additional contradiction Step 1 missed: the paper's own phi equation gives a wave speed of c/sqrt(3) (slower), so the two derivations disagree with each other. Also T is gauge-dependent and not Lorentz-invariant.
• facts.json says "Sweet and Bearden devices achieved gains of 5e4 to 1.5e6 and 90% weight reduction" as a claim; the quoted abstract's own numbers (175 microwatts in, 1 kW out) imply a gain of 5.7e6, outside the stated range. That is an inconsistency inside the source, not Step 1's error.
• facts.json calls the 2020 preprint "Axioms Deductions..." by file name; the real title is as given in the title field. No change needed. Step 1's description of the memo (CIA memo p.1-3, redactions, Bearden named, "not proven") checks out against the page images I viewed.
• claims.md item 2 states "with T = 2 div A the authors derive c*sqrt(3)"; accurate, but note the derivation also needs the assumption "constant of integration is zero" (admitted in the paper).
• Step 1's key_claim "reduce to standard Maxwell when T is constant" is stated in the paper only for grad T = 0; with the authors' sign convention that is itself a constraint on the potentials, so the "reduction" is not automatic.
• Step 1's statement that the Gravity v.2 bundle includes a "negative-mass-density energy term" and "comments section" was not checked by me (appendices unread).