Builder-G: Science Contrarian review

Reviewer role: Science Contrarian (PhD-standard; judges by experiment, not consensus; applies the same rigour to the researcher as to the critics).

Source gap (repeated from Step 1): the only document is G. Builder, "Ether and Relativity", Aust. J. Phys. 11, 279-297 (1958; manuscript received 20 Nov 1957). The PDF has no text layer. I rendered and read all 19 pages as images [read: page images pp.1-19] (journal pages 279-297; PDF page = journal page minus 278). Not available and not read: Dingle's Nature letter (1957) and Builder's earlier papers (Aust. J. Phys. 10:246, 10:424). What Dingle said is known here only from Builder's quotation on p.1 [read] and a web check that the Nature letters exist. Anything about replications and modern tests below is [web] or flagged.

What the paper is, in one paragraph

Builder answers H. Dingle's argument that clock retardation (the twin effect) depends on velocity, so velocity must be absolute, so special relativity (SR) is wrong. Builder keeps Dingle's premise and reverses the conclusion: SR itself "compels" a causal role for absolute velocity (velocity relative to the universe), which is harmless because dynamics and electrodynamics cannot measure it. He then advocates the Lorentz-Poincare ether presentation of SR as pedagogically superior. There is no experiment, no data and no numerical example [read]. Nothing here is a claim of anomalous data, free energy or a failed prediction. It is an interpretive argument, so the contrarian's standard question ("was it tested, or dismissed untested?") has an unusual answer: the interpretation cannot be tested against SR, and it was not suppressed. It was published in a mainstream journal.

Key equations I checked myself

• Eq. (1)-(3): proper time of a clock, dtau/dt = (1 - u^2/c^2)^(1/2). Standard. The retardation integral is the same number in every frame [read p.2, standard SR].
• Eq. (6a), (6b), (7), (7a), (7b), (8): force between a charge and a magnetic pole, and between two point charges, to order v^2/c^2, evaluated at simultaneous positions in frame S.
• Exact field of a uniformly moving point charge at its present position: E = q r_vec (1 - b^2) / [r^3 (1 - b^2 sin^2 th)^(3/2)], with b = v/c and th the angle between v and r. Expanding: E ~ (q/r^2) r_hat [1 + b^2/2 - (3/2)(b.r_hat)^2]. This reproduces Builder's bracket in (7a) and (6b) exactly. Adding the magnetic term (q1 v1 x B)/c with B = v2 x E / c gives the second bracket in (7) and (8). I find the equations correct to O(v^2/c^2). (Step 1 did not check them.)
• (7b): q2 at rest, q1 moving: pure Coulomb force. Correct.
• My own transformation check (hand calculation, not in the paper): take (7b) (q2 at rest, separation r is proper length). Boost to q1's rest frame, where q2 moves at v. The simultaneous separation there is contracted, R_par = r_par/gamma, R_perp = r_perp. Then R^2(1 - b^2 sin^2 th) = r^2/gamma^2, so E' = q (r_par + gamma r_perp)/r^3. Transforming the force back (F_par unchanged, F_perp divided by gamma) gives q r_vec/r^3, exactly (7b). So the "same relative velocity, different force" of (7a) vs (7b) is entirely accounted for by Lorentz transformation of the separation and of the force between frames. It is not evidence of a privileged frame. [read for equations; calculation mine, analytic]

Claim 1. Clock retardation depends on individual speeds in an inertial frame, so it demands a causal role for velocity relative to a unique absolute system (pp.2-3, 7-9; journal 280-282, 287-289)

a. Mathematics. The formulas are right. The inference is a non sequitur. Eq. (3) is the proper time along each worldline, an invariant "path length" in spacetime. Builder writes it with S-speeds u, v and argues that a function of speeds must reflect an interaction with "something universal" (p.3). But the S-speeds appear only because S is a coordinate choice; he himself notes (p.3) the result is the same in every S. Replacing them by speeds relative to an unidentifiable S0 (eq. 4-5) adds a quantity that never appears in any observable, which Builder concedes ("we cannot measure these speeds", p.4). A quantity that is explanatory only by stipulation is not a derivation. Also internally: he says it is "obvious" that there is no alternative to the interaction ("the only hypothesis that is tenable") while the next sentences use the geometric form of eq. (1)-(3), which is the alternative.

b. Experiments. Time dilation itself is well supported: Hafele-Keating 1971 (flown caesium clocks: eastward -59 +/- 10 ns, westward +273 +/- 7 ns vs predictions -40 +/- 23 and +275 +/- 21 ns) [web: https://en.wikipedia.org/wiki/Hafele%E2%80%93Keating_experiment via search summary; hyperphysics.phy-astr.gsu.edu/hbase/Relativ/airtim.html]. Ives-Stilwell-type Doppler tests (modern storage-ring versions bound deviations at ~1e-8) [web: https://en.wikipedia.org/wiki/Test_theories_of_special_relativity]. Nothing experimental distinguishes Builder's causal reading from the geometric one.

c. Tested? The effect: yes, repeatedly, by many groups. The "causal absolute velocity" reading: no test is possible, because by construction (and by Builder's own admission) S0 is not identifiable. Dingle's opposing view (SR is self-contradictory): not supported; the consensus is that his logical objection fails [web: https://en.wikipedia.org/wiki/Herbert_Dingle]. Builder, notably, argues against Dingle, and this was a live 1957 debate (Nature 179:865-866 and 180:1275 exist [web: https://www.nature.com/articles/179865a0, https://www.nature.com/articles/1801275b0]); I did not read these.

d. What would have to change if true. Nothing in SR's predictions (identical by construction). What changes is interpretation: the metric/spacetime picture becomes a derived summary of motion through a medium, and the relativity principle becomes a dynamical conspiracy (all forces and matter must contract and slow in exactly compensating ways). The cost is an undetectable ingredient.

Strongest honest case. The geometric account really is "unexplained" in a causal sense; a clock's rate being set by its path through spacetime says nothing about why. Lorentz and Poincare did build a working causal story, and Bell later argued the same approach is pedagogically legitimate [memory: medium; Bell, "How to teach special relativity", not checked]. Builder is also right that Dingle's reasoning (velocity-dependent effect therefore absolute velocity therefore SR false) is itself invalid: the dependence is on the relative worldline, not on a preferred frame. Weakness of the researcher. "Compels us" (p.4) overstates; in his own Section V he later says the ether hypothesis does not prove itself correct (p.15), conceding the point.

Verdict: split. (i) Clock retardation as an invariant of the worldline: Established. (ii) That it "demands" or "compels" causal absolute velocity: Not testable as stated (no measurable difference from SR; also the inference is not logically forced). Historical note: Historically superseded but instructive as a 1950s rebuttal to Dingle.

Claim 2. An absolute system (ether) is compatible with and sufficient for SR, because absolute velocity is unmeasurable dynamically (pp.4, 11, 13-16; journal 282, 289-295)

a. Math. Consistent. Lorentz-Poincare theory reproduces the Lorentz transformations, and Builder's observation that Einstein's derivation is unaffected if the second postulate is reworded as "light behaves as if propagated in a medium" (p.15) is a fair formal point. Footnote p.16: the light-clock argument checks. A rod of rest length L0 moving at v through the ether has length L0 sqrt(1-b^2); round trip along the motion is L/(c-v)+L/(c+v) = 2L0/(c sqrt(1-b^2)), so the clock slows by sqrt(1-b^2). Correct (I recomputed). Limit: this proves the slowing only for a light-clock. That all other clocks slow identically is not derived but assumed, via the principle of relativity (which Builder says he adds).

b. Experiments. Consistent with Michelson-Morley, Kennedy-Thorndike, Ives-Stilwell, because it is built to be. Wikipedia's summary: Lorentz ether theory (with exact contraction and dilation) gives no experimentally verifiable distinction from SR; SR is preferred for dropping an undetectable aether and for making the relativity principle non-ad-hoc [web: https://en.wikipedia.org/wiki/Lorentz_ether_theory].

c. Tested? No test can separate them. Modern work (Robertson-Mansouri-Sexl framework, sapphire resonators, etc.) bounds any Lorentz violation tightly, with Michelson-Morley/Kennedy-Thorndike-type bounds far below the 1e-8 level [web: https://arxiv.org/pdf/1006.1376, https://en.wikipedia.org/wiki/Test_theories_of_special_relativity]. These null results are not failures of Builder's position (he predicts nulls), but they close the door on an ether that is detectable.

d. What would have to change. Nothing empirical. Philosophy of physics: SR taken as a statement about a medium. Conflicts with the mainstream preference (Occam), not with data.

Strongest honest case. A real, still-respected point: empirical equivalence is genuine, and a critic who says "the ether is refuted" is wrong; it was never refuted by experiment, only made redundant. Quantum gravity searches for Lorentz violation exist precisely because a preferred frame has not been excluded in principle (details [memory: medium], not checked). Rigour on the researcher. "The tenability and sufficiency ... are beyond question" (p.14) is overstated: sufficiency in prediction is shown, but the ether adds no prediction, explains no new fact, and requires an exact coincidence (all matter and fields obey the same compensations). His claim of "no conceptual difficulties" is rhetorical: the unobservability of the rest frame is itself the difficulty.

Verdict: split. Empirical equivalence to SR: Established. Superiority ("sufficient", "pedagogically and heuristically better"): Not testable as stated; as a 1950s position it is Historically superseded but instructive.

Claim 3. Relativistic electrodynamics itself depends on individual (absolute) velocities: same relative velocity gives different forces; Newton's third law fails (eq. 6-8; pp.5-8; journal 283-286)

a. Math. Equations (6a-8) are correct (checked above). The statement that Newton's third law fails for two moving point charges (eq. 7 vs 8) is also correct and standard (momentum is carried by the field; the field momentum compensates). The magnetic-pole case uses a magnetic monopole, a hypothetical object not observed in nature (Builder uses it as a convenient dual; the same effect appears for charges in (7) so no harm). The inference "so phenomena depend on absolute velocity" is the problem. What differs between (a) and (b) are two different configurations in S: relative velocity is equal but the pair's proper separation is not equal, once translated into the pair's rest frame. I showed that (7a) and (7b) are the same physics related by a Lorentz boost; the difference in S-force is the frame dependence of forces and simultaneous separations. In SR, every inertial frame yields the same physics, so no frame-dependent number is an "absolute velocity" effect.

b. Experiments. The O(v^2/c^2) forces here are tiny (b ~ 1e-3 gives 1e-6 relative corrections for lab motion) and are the content of standard electrodynamics, verified indirectly in every particle accelerator and in magnetic-induction devices. Any test that measures a difference attributable to an absolute frame would be a Kennedy-Thorndike / preferred-frame test; those find nothing [web: https://en.wikipedia.org/wiki/Test_theories_of_special_relativity].

c. Tested? The equations: yes, via standard Maxwell-Lorentz theory. The claim of "absolute-velocity dependence": no test distinguishes it, and Builder concedes it cannot be measured (p.11, 13).

d. What would have to change. The equivalence of inertial frames in electrodynamics, i.e. SR's first postulate, would have to be false at observable level, which experiment excludes.

Strongest honest case. Builder's observation is real and often under-taught: velocity-dependent, non-reciprocal forces between isolated charges are genuine in a given frame, and the Einstein 1905 opening about the magnet-conductor asymmetry is a statement about relative motion that holds only at O(v/c) (p.7). A "relativity is just relative motion" slogan (Dingle-style) is wrong, and Builder correctly refutes it. Weakness. He sets up S as "any" frame and then says all S-observers "must agree" the phenomena depend on individual velocities (p.8). This is true only of S-dependent descriptions. There is no S-independent observable in his paper that shows it. The magnet-moving-through-a-loop example (p.6) claims an observable dependence on individual velocities, but he gives no number and the induced current (the invariant observable) is the same in all frames.

Verdict: split. The equations and Newton's-third-law failure: Established. The conclusion that this shows dependence on absolute velocity: Refuted by calculation (the two cases are Lorentz-boost related; my transformation above reproduces (7b) exactly from (7a)'s structure, with no preferred frame). Nobody needs an experiment for that.

Claim 4. Absolute acceleration implies an absolute inertial system and hence absolute velocity; Mach's hypothesis must be rejected, leaving the ether; absolute velocity may be kinematically measurable, e.g. by Doppler asymmetry of starlight (pp.9-13; journal 287-291)

a. Math. There is no new mathematics. The logic fails at one step: acceleration is absolute in Newtonian mechanics too, yet Newtonian mechanics has Galilean invariance and no absolute velocity. Absolute acceleration only needs a notion of "inertial frame" (an affine structure, equivalently "straight worldlines"), not a rest frame. Builder's step (p.10): "the existence of absolute acceleration entails the existence of absolute velocity" is therefore unsupported; he builds it in by defining acceleration as a change in "absolute velocity". The Mach step is a false dichotomy: he rejects only Mach's instantaneous action at a distance and then says only the ether remains; the middle option (field theory with inertia determined by matter, i.e. Einstein's "Mach's principle" in GR) is mentioned in his own footnote p.13 and then left unused.

b. Experiments. * Kinematic measurability: he predicted that if the solar system's velocity "were an appreciable fraction of c" it would show as Doppler asymmetry of starlight (p.12). The cosmic microwave background (CMB) dipole later gave exactly such a measurement: Sun's velocity relative to the CMB frame = 369.82 +/- 0.11 km/s, b = 1.234e-3, towards (l,b)=(264.02 deg, 48.25 deg) [web: https://arxiv.org/pdf/1807.06205 via Planck 2018 results; summary https://arxiv.org/html/2106.07666v1]. This is "velocity relative to the universe as a whole" in a precise sense, which is his own definition of "absolute". It was discovered in the late 1960s-70s, after his paper [memory: medium, not checked]. This is the fairest point in his favour. * But the CMB frame is the standard cosmological comoving frame; it is a preferred observer-class without any physics (clock rates, force laws, light speed) depending on it. Kennedy-Thorndike and Michelson-Morley bounds cap any dependence on motion through it far below the first-order levels (see claim 2). * Mach/frame dragging: Gravity Probe B measured frame dragging of -37.2 +/- 7.2 mas/yr vs GR's -39.2 mas/yr (about 20% precision) [web: https://einstein.stanford.edu/content/sci_papers/papers/PhysRevLett.106.221101.pdf]. That mass-energy motion locally drags inertial frames supports a (partial) Machian effect through a field, not by instantaneous action at a distance.

c. Tested? The kinematic claim: tested and, as a bare velocity, supported (CMB dipole). The causal claim that the "absolute system" acts on bodies: not tested. Not suppressed or dismissed without testing; no one ran a controlled experiment for it because it predicts nothing beyond SR.

d. What would have to change. If an absolute system truly acted on matter, SR's frame-equivalence of physical laws would hold only as an approximate/hidden symmetry. GR's treatment of inertia (locally inertial frames from the metric) would need an extra, non-gravitational medium.

Strongest honest case. Defining "absolute" as "relative to the universe" is a sober choice (p.11), and the CMB dipole shows such a reference frame is real and measurable; Builder correctly anticipated measurability by kinematics alone and correctly separated that from any dynamical detection. He also correctly notes (p.14) that Einstein's 1920 lecture restored an ether of a sort in GR [read]; this is a straightforward reading, though Einstein's ether carries no velocity (a point Builder's own footnote p.13 concedes). Weakness. Dismissing Mach's idea as "untenable" (p.13) is stated, not argued, and it hid the field-theory middle ground. He also says "we cannot ascribe any physical difference" to the CMB-like frame beyond kinematics, so "absolute velocity" becomes a label for the CMB frame, which makes it much weaker than the "causal significance" asserted in the Summary.

Verdict: split. (i) Absolute acceleration implies absolute velocity: Internally inconsistent / non sequitur (his own Newtonian precedent has the former without the latter); in effect Refuted by calculation only in the sense of logic, so I use Internally inconsistent. (ii) Mach's instantaneous-action hypothesis is untenable, leaving only the ether: Historically superseded but instructive (GR and frame-dragging give a third option; GP-B ~20% confirmation). (iii) A cosmological rest frame is kinematically measurable: Established (CMB dipole 369.82 km/s), but with no dynamical effect.

Claim 5. A clock moving through the ether at v0 slows by (1 - v0^2/c^2)^(1/2) as a consequence of Lorentz-FitzGerald contraction, so one hypothesis, not two, accounts for Michelson-Morley; the ether presentation is pedagogically superior (pp.16-19; journal 294-297)

a. Math. The light-clock footnote is correct (see claim 2). Where he is imprecise: Broad's "two hypotheses" are contraction (for Michelson-Morley) and time dilation (for Kennedy-Thorndike, which tests the combination); Builder's footnote that contraction "entails" the clock slowing holds only for the light-clock. For a general clock, equivalence to relativity must be assumed. Also (p.17) the claim that the "second clock is observationally distinguishable from the first because it accelerates" is correct and standard: acceleration is detectable by an accelerometer, a point where the contrarian credit goes to him (not Dingle).

b. Experiments. Michelson-Morley (1887) and Kennedy-Thorndike (1932) nulls are explained as in Lorentz theory; they are equally explained by SR. No experimental difference [web: https://en.wikipedia.org/wiki/Lorentz_ether_theory].

c. Tested? Pedagogical superiority: never tested; it is an opinion; no education-research evidence is cited. Physics content: same as SR.

d. What would have to change. Nothing in prediction; the textbook picture of "what is real" (length contraction as ether-induced physical shrinking vs. spacetime geometry) would flip.

Strongest honest case. The pedagogical claim has serious support in the literature (Bell, Builder's own stated aim to explain "reciprocity" physically: p.19 on why A sees B slow and B sees A slow); the reciprocity explained by one-way light speed unmeasurability is neat and mathematically right. Weakness. The "explanation" is not simpler: one must still postulate that every force law and clock obeys Lorentz covariance, i.e. the relativity principle, as an unexplained fact.

Verdict: split. One-hypothesis derivation of the Michelson-Morley null and clock slowing for the light-clock case: Established (correct mathematics, same predictions as SR). Pedagogical/heuristic superiority: Not testable as stated (opinion; no controlled teaching evidence). As history, Historically superseded but instructive.

How this claim set was treated (the contrarian's questions about the rejection)

• Was it dismissed without testing? There is nothing to test: the paper predicts the same results as SR. It was published in the Australian Journal of Physics (CSIRO) with no sign of blocking [read: p.1, footer]. I found no record of suppression and I will not invent one.
• The real intellectual story is Dingle's: the later Dingle campaign against SR (he was President of the Royal Astronomical Society 1951-53) ended with a broad consensus that his logical objection was unsound [web: https://en.wikipedia.org/wiki/Herbert_Dingle]. Builder's paper is an informed reply from the ether side that accepts SR's predictions. Whether Builder's ether side was "marginalised" is outside what these documents show.
• Researcher rigour: no unit errors, no bad data (there is none). Overreach: "compels", "beyond question", "no tenable alternative" in a context where the paper concedes (p.15) that the ether is not proved and gives no discriminating observable.

Summary table

Corrections to Step 1

1. facts.json says the validity of Builder's equations (6b), (7), (8) "was not checked". I checked them: they are correct to O(v^2/c^2) (exact moving-charge field expanded), and their apparent frame asymmetry is accounted for by Lorentz transformation of separation and force. Step 1's mainstream_conflict line "His claim ... is an observable asymmetry" slightly mischaracterises him: Builder says the asymmetry is observed in S and cannot be turned into a measurement of absolute velocity (pp.8, 11, 13).
2. claims.md page numbers for Claim 5 ("pp.14-16") are too early: Section VI and the clock-slowing argument are on PDF pp.16-17 (journal 294-295); the concluding pedagogical section is PDF p.18 (journal 296). Claim 2's "pp.4, 11, 13-14" should read pp.4, 11, 13-16. (PDF page = journal page minus 278.)
3. claims.md does not record that Claim 4's Mach step omits Einstein's "Mach's principle" field-theory option, which Builder mentions in his own footnote on p.13. Step 1 did not mention that Builder's "absolute velocity" is defined as velocity relative to the universe or fixed stars (p.11-12), which is the CMB frame in modern terms. This is the single most testable, and most defensible, part of the paper.
4. Facts record the author only as "G. Builder"; the paper gives no first name [read: p.1], so none should be assumed.
5. The remaining facts.json summary matches my reading of the page images.