Physicist review: Gift-Stephan (role: PhD physicist)

Source tags: [read] from the documents, [web] URL, [memory] with confidence. Web-search budget was exhausted in this session, so checks were done with WebFetch of Wikipedia pages only. Items I could not check are marked "check failed".

Source gap (repeated from Step 1, now partly closed)

• 2001 gravity paper (Physics Essays 14(4)): the text layer is garbled. Step 1 read pdf pp. 1-5 and 8. I additionally rendered and read pp. 3, 5, 6, 7, 9, 10 [read: page images]. The earlier gap on pp. 6, 7, 9, 10 is closed, and reading it changed the verdict (see Corrections).
• 2009 magnetism paper: text layer is usable but equations are garbled; I read sections 1-4.2 and the Greenspan passage from text. Sections on chemical bond examples were only skimmed.
• Not in the folder, so second-hand through Gift only: Shtyrkov's aberration data, Marmet/Kelly/Sato GPS claims, Greenspan's simulations, Aspden's 1980 force law, Hatch 2004.
• Where a conclusion below rests on [memory] alone, it says so.

Corrections to Step 1

1. claims.md claim 4 and facts.json say the scalar-graviton theory "gives ... orbit precession" and "passes the three classical tests" via the field equation. That is not what the paper does [read: p.6-7 images]. The field equation (eqs 7-9) gives only Newton's law. Perihelion precession (eqs 12-15) comes from importing Aspden's modified force law F = (1 + 3v^2/c^2) G M m / r^2 and then solving the resulting orbit equation. The paper itself says light bending and redshift "involve the interaction of gravity and electromagnetism and therefore the new theory by itself cannot predict the outcome". It then uses a heuristic photon "mass" E/c^2 in Newtonian-style arguments. Only the redshift (eqs 18-24) is a clean result, and it is first-order Newtonian energy bookkeeping. Radar delay, orbit decay and DI Herculis are explicitly deferred ("upon completion of our investigation"). So "has passed the three classical tests" overstates it.
2. Step 1 summary says the paper explains binary-pulsar decay "qualitatively by wave potentials". The paper gives no calculation for it [read: pp.6-9]. Treat it as an assertion.
3. Step 1 omitted Table II (p.9), where the paper claims the same time-local gauge trick yields pions as the strong-force mediators and "unifies" gravity with the strong force, and it quotes Nordenson calling relativity "one of the most serious logical incoherencies in the history of science". Context only; no claim there is derived.
4. Confirmed correct: the 414 ns figure (I reproduce 414.8 ns from 4omegapi*R^2/c^2 with R = 6378 km), the eq 1-8 chain in the 2018 paper, the Greenspan numbers (2.20e14 vs 1.38e14 Hz), and the statement that the 2006 paper's own end results (v/c) equal the special-relativity ones.
5. One refinement, not an error: in the 2018 paper the MM arm result (eq 17, 2 l w/c^2 east-west) is a difference of one-way times, so it needs synchronised clocks. A single clock timing a round trip gives 2l/c with no first-order term. The paper does not say this.

---

Claim 1: Light speed is c - v eastward and c + v westward relative to the Earth's surface (GPS clocks, range equation, time transfer, GPS-clock Michelson-Morley)

a. Is the mathematics internally consistent? The algebra is correct. The CCIR/ITU clock-sync rule gives dt = l/c + l v/c^2 east and l/c - l v/c^2 west. I checked eq 2-4: 2omegaA/c^2 with A = (l/2 pi r) * pi r^2 gives omegarl/c^2 = l v/c^2. [read] The inference is the problem. The correction term in the rule is the Sagnac term, which exists precisely because the Earth-fixed clocks have been synchronised to inertial ECI (Earth-centred inertial) time, in which light is assumed to travel at c. Then "speed relative to the surface = l / dt" gives c - v by construction. Gift's own text says the GPS specification takes light at constant c in ECI [read: 2018 p.4], and Hatch's "clock bias" reading is cited in the conclusion [read]. So the result is an input, not a measurement. The "Michelson-Morley with GPS clocks" arm times (eqs 13-22) have the same circularity: eq 13 (the GPS time) is set equal to eq 14 (ether theory), but both are the same convention. Worked check: for l = 1000 km at 40 deg N, v = 356 m/s, l v/c^2 = 3.96 ns.

b. Consistent with experiment? The measured content, a closed-loop or synchronised-clock delay of 414 ns around the equator and the Michelson-Gale fringe shift, is real and is accepted. It is predicted by special relativity too: von Laue showed this in 1911 [web: https://en.wikipedia.org/wiki/Sagnac_effect]; Michelson-Gale measured 230 +/- 5 vs 237 predicted (parts in 1000) [web, same URL]. GPS operators apply the Sagnac correction routinely; that is why it is in the specification (the same page confirms GNSS must account for Earth rotation). The one-way speed of light "cannot be measured independently of a convention as to how to synchronize the clocks"; experiments claiming to measure it were reanalysed as two-way [web: https://en.wikipedia.org/wiki/One-way_speed_of_light]. Direct physical constraint: any real anisotropy would show in resonator experiments, which Gift cites at dc/c < 1e-17 [read: 2018 p.2], yet his "effect" is 1e-6 (v/c ~ 1.5e-6). He reconciles this by saying such tests are second-order or not relevant, but a true c +/- v difference would be first order.

c. Tested? By whom? Replicated? Marmet (GPS data) and Kelly are cited [read]; their papers were not read. No neutral test is on record because no test can discriminate: ether and special relativity give identical readings. Hatch (cited by Gift) agrees the numbers are right and offers a synchronisation interpretation [read]. Verdict on testing: the measurement was done and replicated daily by GPS operations; the interpretation was never distinguishable.

d. If true, what would change? One-way speed would be physically anisotropic, so the Einstein synchronisation convention would be a physical choice, and the isotropy of cavity resonators (1e-17) would need an explanation. The constant-c postulate for local clocks would fail. But then GPS itself, which assumes constant c in ECI and not in the surface frame, would be working by a preferred-frame rule rather than special relativity. This is a large change and is not needed.

Verdict (split): - Measured part (east-west clock-sync difference, Sagnac delay): Established, and equally predicted by special relativity. - Claim that this shows c - v / c + v physically: Not testable as stated (convention-dependent; no measurable difference from special relativity). Part of the argument is circular, because the assumed c in ECI is used to derive the result, but I would not call it internally inconsistent in its equations.

---

Claim 2: Earth's orbital motion is detectable by apparatus at rest on Earth (planetary radar, Shtyrkov's aberration, Roemer, stellar Doppler)

a. Math. Radar section (eqs 41-55): the barycentric-frame round-trip model assumes light at c in that frame, then divides L by leg time to get c +/- v at the antenna. The same circularity as Claim 1. Roemer/Doppler (2006): the final results are v/c, the textbook ones, and the paper admits special relativity gives them too [read]. The intermediate speeds c - v / c + v are the GGT (generalised Galilean transformation) assumption restated. Aberration: tan(alpha) = v/c gives 20.49" for 29.8 km/s; I confirm 20.489" [calc]. Internally consistent.

b. Consistent with experiment? Roemer, Doppler, stellar aberration and radar ranging are all established and are all reproduced by special relativity. Shtyrkov's claim is different: a source and receiver both fixed to the Earth showing a 20" aberration. In special relativity a source and receiver at rest relative to one another in one inertial frame show no aberration between them, so his result would contradict relativity if real. It contradicts also the very large body of satellite tracking and VLBI work; I am not able to say more without his data [memory: low]. His 1998-99 data have apparent seasonal variation but I did not examine them.

c. Tested/replicated? Gift himself writes "We await confirmation by others of this very interesting experiment" [read: 2018 p.18]. I found no replication. Check failed: web budget exhausted, so I could not look for a published replication or critique. State plainly: no replication known to me; not verified.

d. If true? The detection of 30 km/s would mean a preferred frame accessible from the ground. It would also be inconsistent with the claimed rotation-only effects: if an ether-frame existed with the Sun moving at ~370 km/s relative to the CMB [memory: high], the c +/- v from that motion should be 10-1000 times bigger than the effects Gift sees. The paper's answer is "multiple preferred frames carried by Earth and Sun" [read], which is ad hoc.

Verdict (split): - Roemer, Doppler, aberration, radar ranging as physics: Established. - Reading them as ether detection: Not testable as stated (identical predictions). - Shtyrkov's co-moving-aberration result: Never properly tested (single-author, unreplicated, source not read; no controlled independent test on record; not dismissed without testing as far as I can tell, simply unreplicated).

---

Claim 3: A cosmic ether exists, light-speed invariance and hence special relativity are invalid; Michelson-Morley was the wrong kind of test; CMB is the ether's radiation

a. Math. The non sequitur is internal: the 2006 paper obtains exactly the special-relativistic v/c results for Roemer and Doppler [read] and then says they "invalidate" special relativity because of the intermediate speeds. The FLL (Fitzgerald-Lorentz-Larmor) contraction is chosen so that the second-order results agree with special relativity (Lorentz ether theory is mathematically equivalent to special relativity for all outcomes; this is Mansouri-Sexl territory [memory: high]). Gift cannot both say contraction hides the ether and say first-order tests can see it, when the first-order results are the same as special relativity's. Also: a first-order effect of c +/- v would appear in any one-way synchronisation-independent test; he gives none.

b. Consistent with experiment? Ives-Stilwell, Kennedy-Thorndike and modern cavity experiments (his cited 1e-17 [read]) constrain preferred-frame effects; the combination of Michelson-Morley and Kennedy-Thorndike plus Ives-Stilwell fixes the Lorentz transformation [memory: high; not web-checked]. In 2006 he lists some first-order null results (maser, Mossbauer) himself [read]. Dayton Miller's 1925 claim was attributed by Shankland et al. to temperature drift [read: 2018 p.2 and 2006 p.1, via Gift]; I did not check that analysis.

c. Tested? Yes, extensively: second-order tests and Kennedy-Thorndike/Ives-Stilwell type tests have been run many times with improving precision, by independent groups [memory: high]. The preferred-frame models were tested and constrained. Not a "never tested" claim. Gift's new first-order claims are not new tests; they re-read old data.

d. If true? The Lorentz group would be only an approximate symmetry; the GPS specification itself would reflect a preferred frame; the CMB dipole (Earth moves ~370 km/s relative to the CMB [memory: high]) would define it. But cosmology already has a preferred frame (comoving matter / CMB) without any ether, and special relativity is not violated by it.

Verdict (split): - "SR is invalid because the experiments show light-speed variation": Tested, not supported (second-order, Kennedy-Thorndike, Ives-Stilwell and cavity tests all agree with SR; the cited "detections" have SR's own predictions), with an additional Internally inconsistent step: the claimed "invalidating" results equal SR's results. - "A preferred frame exists but is undetectable (Lorentz ether theory)": Not testable as stated / Historically superseded but instructive. - CMB as ether radiation: Not testable as stated; the CMB gives a cosmological rest frame, which is fine with SR.

---

Claim 4: The equivalence principle is false, so GR is invalid; replace it with a Newtonian-type scalar-graviton gauge theory with no black holes, dark matter or gravity waves

a. Math. (i) The "non-equivalence" argument (pp.2-3) compares rocket A (gravity F plus floor thrust R) with rocket B (only R) and says different forces act, so they differ. The equivalence principle (EP) is a statement about local experiments in a small region, in which gravity is not a force acting on the body but is indistinguishable from the acceleration. Calling gravity "a force F" assumes the Newtonian framing the EP denies, so the argument is circular [read: p.3 image]. (ii) The Larmor test: a charge at rest in accelerating rocket B radiates, one in hovering rocket A does not. This is the old Peierls/Born-type puzzle. The accepted answer (Born 1909, Fulton-Rohrlich 1960, Boulware 1980) is that "radiation" is not an observer-invariant, locally measurable quantity for a uniformly accelerated charge [memory: medium; not web-checked, check failed]. A local test in a small box cannot detect the radiation, which is spread over a region where tidal differences matter, so the EP (local) is not violated. Gift calls Peierls's resolution "arbitrary" without argument. (iii) The replacement theory. eq 6 couples a "scalar" A to psi-bar gamma^0 psi, the time component of a current, which Gift admits is not Lorentz-invariant and "allowed by assumption" [read: p.5]. A gauge shift A -> A + d(alpha)/dt with alpha = alpha(t) is a spatially uniform shift of potential energy, the trivial freedom in Newtonian mechanics (the zero of potential energy is arbitrary). It does not fix the dynamics of A; the kinetic term for A (eq 7 source) is added by hand and its invariance under the shift is not shown [read, p.5-6]. Static limit: eq 9 is the Poisson equation, and the sign of G is simply chosen. (iv) Precession and bending are not derived (see Corrections 1): eq 12 is inserted from Aspden; the orbit equation eq 14 has the same 3GM u^2/c^2 term as GR; with that, Delta = 6 pi G m/(a c^2 (1-e^2)) [eq 15, the paper writes m where M is meant, a typo] is just the GR formula. Using eq 12 with v the tangential speed only is an approximation that ignores radial velocity [calc]. The light-bending derivation (eqs 16-17) uses the same equation with a "photon mass" E/c^2 and ends with "the exact formula given by GTR"; it is not derived from the field equation. (v) Source coupling: the field couples to m * psi-bar gamma^0 psi, i.e. to the constituents' rest masses times number density, not to total energy. Binding energy would not gravitate: iron-56 binding is 8.8 MeV per nucleon, about 0.94% of its mass [calc; 8.79/931.5; binding energy per nucleon is [memory] high], so test bodies of different composition would fall differently at about the 1e-3 level.

b. Consistent with experiment? Weak equivalence principle is tested to roughly 1e-13 in torsion balances and lunar laser ranging [memory: high for 1e-13; the satellite MICROSCOPE result is about 1e-15, medium; check failed, web budget exhausted]. The composition estimate in (v) is far above that. For scalar theories in general: Nordstrom's scalar theory predicts zero light bending and a perihelion lag of about 7"/century against 43" advance observed [web: https://en.wikipedia.org/wiki/Nordström's_theory_of_gravitation]. Gift's theory is not Nordstrom's (his coupling is non-covariant and he imports the force law), so this is an analogy, not a refutation of his equations. The observed light deflection 1.75" at the Sun's limb and the 43" Mercury precession are accepted GR results [memory: high; not checked here]. Binary pulsar decay: Hulse-Taylor shows "energy loss in close agreement with the predicted energy radiated by gravitational waves" [web: https://en.wikipedia.org/wiki/Tests_of_general_relativity]. LIGO detection of gravitational waves in 2015 [memory: high; not checked here]. Gift's explanation, "wave potentials", has no calculation.

c. Tested? The EP and GR: yes, extensively, by many independent groups with replications (torsion balance, lunar laser ranging, Mercury, light deflection, binary pulsars, LIGO). Gift's theory: it has never been put to a test by anyone; the paper itself defers radar delay, orbit decay and DI Herculis [read: p.6]. The DI Herculis anomaly he cites (apsidal motion) is a case for which he gives no calculation [read]; I did not check the astronomy literature [check failed].

d. If true? Everything on the GR side would need rewriting: the EP, light bending, gravitational waves (LIGO's detections would need another source), black-hole observations, cosmology. The physics would need a covariant replacement because his theory breaks Lorentz invariance in gravity and in the (strong) form he proposes also in the magnetic sector (Claim 5).

Verdict (split): - "EP is false": Tested, not supported (WEP tested to ~1e-13; the Larmor argument rests on a confusion about local vs global radiation). The rocket-force argument is Internally inconsistent in framing (it assumes a Newtonian gravitational force to refute a theory that denies one). - "The scalar-graviton gauge theory gives Newton, redshift, precession, bending": Newton's law and first-order redshift follow trivially; precession and bending are Internally inconsistent with the claim of derivation (imported from Aspden/GR forms, paper admits the theory cannot predict bending). Composition-dependence of fall rates: Refuted by calculation (about 1e-3 vs 1e-13 bounds; [memory] caveat above). - "No black holes, dark matter, gravitational waves": Tested, not supported for gravitational waves (Hulse-Taylor [web], LIGO [memory: high]); black holes and dark matter are not Gift-specific, no test offered by him. Not independently checked here.

---

Claim 5: Magnetism is a separate interaction mediated by massless spin-1 "magnatons", explaining the Pauli principle and chemical bonding

a. Math. The 3-vector field A_k is Maxwell's magnetic vector potential. Eq 3.12 is the magnetostatic Poisson equation, which is a theorem of Maxwell/QED already (in Coulomb gauge); the coupling psi-bar gamma_k psi A_k is the spatial half of the QED coupling j^mu A_mu. So the "new" field equation is reproduced by the existing photon field; the paper identifies A_k with the Aharonov-Bohm potential [read] which is the photon's potential. A space-local, time-independent phase change is a subset of ordinary QED gauge transformations, so no new field is forced. The claimed non-covariance (Postulate 1) is introduced to allow the separate 3-vector. Step 3.13: the dipole force 3 mu1 mu2 /(4 pi r^4) for aligned dipoles is classical [read: p.14, Griffiths-level]. Note the paper describes the force from magnatons as "consistent with classical magnetism" while saying QED gives something opposite; the puzzle statement mixes spin direction and moment direction. For electrons the magnetic moment is opposite to spin, and the positronium (e+ e-) hyperfine splitting also involves an annihilation term, so using the orthopositronium example to say QED's magnetic part is attractive for side-by-side parallel electrons is not established in the paper [memory: medium; not web-checked].

b. Consistent with experiment? Bar magnets interacting does not need exchange of photons by charge: the force is from currents and spin magnetisation, and there is no net-charge obstruction (the "problem" of zero net charge is not a problem; the amplitude depends on the magnetisation current) [memory: high]. Strength: the dipole-dipole energy between two electron spins at one Bohr radius is mu0 mu_B^2/(4 pi a0^3) = 5.8e-23 J = 3.6e-4 eV, while the Coulomb energy at the same distance is 27 eV (and hydrogen molecule binding is about 4.5 eV [memory: high]) [calc]. So a 1/r^3 magnetic force cannot be the Pauli force, the solidity of matter, or the bond; it is 4-5 orders too small at atomic distances. Pauli exclusion is a consequence of antisymmetry of identical fermions (spin-statistics) and needs no force [memory: high]. The paper's Greenspan "experimental confirmation": a classical simulation in which assuming bond electrons attract gives 1.38e14 Hz rather than 2.20e14 Hz [read]. That shows only that a model with an attractive effective force fits one number; it is not an experiment, it is not tied to magnetic dipole strength, and a quantum calculation of H2 gives the right vibrational frequency without any new force [memory: high]. I did not read Greenspan; check failed.

c. Tested? The new particle: no direct test is proposed or on record. Searches for new massless spin-1 bosons exist as dark-photon/fifth-force work but not for this object [memory: medium]. Greenspan's simulations are the only support, and they are not independent tests of magnatons. No replication.

d. If true? QED would have two separate massless vectors coupled to the same current, with the magnetic interaction counted twice (once by photons, once by magnatons), contradicting the precise agreement of QED magnetic-moment calculations (electron g-2 [memory: high]). Lorentz covariance would be broken in the magnetic sector. Quantum chemistry (Pauli antisymmetry plus Coulomb) would be misattributed.

Verdict (split): - Magnatons as a new, separate force: Refuted by calculation (strength 1e-4 eV vs eV scale at atomic distances; the field equation is already Maxwell's) and Internally inconsistent in being offered as new while reproducing exactly the existing equations. - Pauli exclusion / bonding from the magnetic force: Refuted by calculation (same numbers). - Greenspan "confirmation": Never properly tested as a magnaton test (a fit, not a controlled experiment; not read).

---

Summary of verdicts