Kelly-AG: Quality Engineering (Step 4b)

QE agent: Opus. Date: 2026-10-04. Inputs checked: _text/manifest.json, _textprep.log, the one digest (digest/Kelly-2-1a1021.01.md), deep/Kelly-2-1a1021.md, facts.json, claims.md, sites.md, all 11 notes in research/, the saved files in research/files/ (Wang 2003, Kelly 2000 Hafele-Keating page and Monograph 2 checked directly), physicist.md, mathematician.md, contrarian.md, arbiter.md, primer.md, example.py (run). I read the whole extracted text, _text/Kelly-2-1a1021.txt, pp.1-13, myself.

Page numbers are the === page N === markers in the extracted text. They match the PDF pages used by the deep reader and the reviewers.

Verdict: MINOR. The six central verdicts hold up, and the main argument is sound: the arbiter's ruling on claim 1 (weak form supported by Wang et al.; strong form refuted by Kelly's own line integral) is correct and is checked numerically in example.py. The defects are these: - Two pieces of experimental support that Kelly offers for "light travels relative to the laboratory" are missing from claims.md, the arbiter and the primer: Michelson's 1913 rotating-mirror test and the ring-laser standing-wave argument. - Pages 6-11 never had a deep read. - Every reviewer and the arbiter use an off-scale verdict label ("Tested, supported"), and the arbiter's verdict for claim 2 is inconsistent between its body and its summary. - There are a few small accuracy slips, the main one being how Wang's ride-along gyroscope is described.

All of these can be fixed by updating facts.json, claims.md, arbiter.md (a "QE revision" section) and primer.md.

---

1. Coverage

Coverage is complete. The text layer is clean prose throughout. The only garbling is in symbols (omega and pi are dropped, e.g. "dt = 4A/c2" on p.5 and "4nA" on p.12).

Defect C1 (digest gap, propagated). The digest leaves out four items on pp.6-11: - Michelson's 1913 rotating-mirror test (pp.6-7); - Kelly's ring-laser standing-wave argument and its sources, Martin 1986 and Anderson 1986 (p.7); - Einstein's 1916 and 1919 quotations, with Kelly's charge that authors "selectively quote Einstein's writings" (p.8); - the Saburi (0.9 m) test (p.10).

It also leaves out the p.11 statement "The disc was 50 cm diameter, and the enclosed area of the light path was 666 cm2". The reviewers found that one by reading the text layer. The other four were never picked up (section 3).

---

2. Shortcuts

Defect S1: pages 6-11 had no deep read. facts.json says so honestly: "Pages 6-11 and 13 ... rest on the digest only". These pages carry four things: - the ring-laser evidence for claim 4; - the evidence for claims 1 and 5 (the Harress and Pogany polygon paths, p.8; Wang et al., p.10); - the inertial-frame ("orbital angle") argument, pp.9-10; - the "SR has no role" section (claim 3), p.10.

In practice the shortcut was partly made good, so this is minor: - the arbiter and the primer re-read pp.9-13 in the text layer; - the mathematician cites pp.5-7 and 11 directly; - claim 1 was checked against Wang's own paper, which is saved and was read.

But pp.6-8 were read by nobody in full. That is how the omissions in section 3 got through. Remediation: deep reads 1 and 2.

S2: the equations on p.12 and p.7 were checked by algebra, not by page image. The arbiter discloses this. The deep reader did render p.12. Its reading ("2 Delta t = 2D/c^2 = [8 pi n A]/[c^2]") settles the "4nA versus 4 pi n A" question that the arbiter leaves open: it is an extraction garble, not Kelly's error. The arbiter should cite the deep read for this. The ring-laser formula on p.7 ("df = [4A]/[P]" in the text layer; it should read 4AΩ/(λP)) was never seen as an image. It does not affect any verdict, but deep read 1 should confirm it.

S3: Investigator time. sites.md records about 3.5 minutes of web time against a 10-minute cap. The run was cut short "by time/choice", and several high-value targets were left as not reached: - the full text of Vigier 1997; - the Irish Times archive for 1996; - the Hafele 1971 report in the Precise Time and Time Interval (PTTI) meeting proceedings; - Infinite Energy 2001.

This is acceptable for a first run, because sites.md lists them for the next run. But the arbiter's open questions 4 and 5 depend on exactly these sources.

---

3. Missed material (from my own reading of the digest and the full text)

1. Michelson 1913, rotating mirrors (pp.6-7; ref. [9], Astrophys. J. 37, 190-3). Kelly presents this as "a clear confirmation that the conclusions in this paper ... are correct" [read: p.7]. Light was reflected from mirrors turning at 1800 rpm. From the fringe shift, Kelly says, Michelson showed three things: - (1) light does not bounce like a tennis ball off a moving racquet; - (2) the reflected light does not take up the mirror's velocity; - (3) "the light actually ignores the whole rotating apparatus. It moves solely with respect to the laboratory" [read: p.7].

He gives the accuracy as 1:50. He also says Michelson made a factor-of-2 error that was cancelled by a second omission, an error "uncovered by a colleague who has been corresponding with this author" [read: p.7].

This is a second experiment offered in support of claim 2, and of the source-independence half of it in particular. Only the mathematician mentions it, as "an interpretation I could not check" [memory, low]. It is absent from claims.md, the physicist's and contrarian's reviews, the arbiter and the primer. Retention defect (section 5).

Mainstream context the reviewers should weigh: Michelson 1913 is usually cited as a test against the emission (ballistic) theory of light, i.e. as evidence that reflected light does not take the speed of a moving mirror [memory, medium]. Read that way, it supports source-independence, which SR also assumes, and does not discriminate SR from Kelly's picture. The same pattern applies to claim 2.

2. Ring-laser standing waves (p.7). Kelly argues as follows: - in a ring laser, the standing waves "stay stationary in relation to the laboratory" (Martin 1986, ref. [13]); - when the ring rotates, the nodes pass an observation post at a rate proportional to the rotation rate; - for a square or a triangle that rate differs from the rate for a circle; - "This phenomenon is further confirmation of the proposal in this paper, that light, in small scale experiments, travels relative to the laboratory" (Anderson 1986, ref. [14]) [read: p.7].

This is a third line of evidence for claim 2. No reviewer, nor the arbiter or the primer, addresses it. The physicist should say whether the standing-wave picture is the standard account of the ring-laser beat [memory, medium: the standard textbook picture of a ring-laser gyroscope does describe the standing wave as fixed in inertial space, with the beat seen as nodes sweeping past the detector]. If it is, it is the same non-discriminating situation as claim 2. Note the phrase "in small scale experiments": it is a qualification of claim 2 that claims.md leaves out.

3. Kelly's argument about how Einstein's writings are used (p.8). Kelly quotes: - Einstein 1905, on clocks moving along polygonal and curved paths; - Einstein 1919, "The rotating observer notes very well that of his two equivalent clocks, that placed on the circumference runs slower"; - Einstein 1916, on "special" meaning uniform translation.

He charges that "authors can selectively quote Einstein's writings, whenever there is a seemingly insurmountable difficulty with SR theory" [read: p.8]. This is the core of his case that SR's restriction to straight-line motion is applied selectively. The arbiter treats the 1905 polygon argument (claim 5) but not the 1916/1919 quotations or the charge itself. It belongs in claims.md as part of claim 5, or as a note under claim 3. The reviewers' standard answer (SR handles rotation in an inertial frame; the 1916 restriction is about which frames are equivalent) should then be stated beside it.

4. The Saburi test (p.10). Kelly cites "the M & G (600m) or Saburi (0.9m) tests" as paths whose curvature cannot be distinguished from a straight line [read: p.10]. Saburi is not in Kelly's reference list [read: p.13]. [memory, medium] This is probably Saburi et al. 1976, a Sagnac-correction test with clocks carried round, or with signal round trips, in Japan; it was not checked. Record it as an unreferenced citation.

5. Kelly's GPS-consistency charge (p.9). Kelly writes: "Pretending that SR does not apply to rotation, while at the same time applying it daily to experiments like the Global Positioning System, that has a far greater amount of rotation than the Sagnac experiment, is indefensible ... a very biased selection of what is termed an 'Inertial Frame'" [read: p.9]. The arbiter rules the orbital-angle numbers a category error. But Kelly's underlying accusation of inconsistency in how frames are chosen is a distinct argument, and it should be stated in his words beside that ruling. The primer has only "orbital angle ... Internally inconsistent" in its evidence table and never explains the argument to the reader.

6. Kelly's Doppler point (p.12, item 3). Kelly notes that Sagnac's own observations had no Doppler effect, because the camera rode on the disc, and that a lab observer would see a negligible one [read: p.12]. It is also relevant that Post's quoted figures compare the SR correction with the Doppler correction (p.10). The point is minor; facts.json could hold it.

7. Investigator lead dropped: Kelly's own Faraday experiments. Monographs 5 and 6 (1998, unipolar induction, with his own apparatus) are saved in research/files and listed under "The inventor's own work" in sites.md. They appear in no review, nor in the arbiter or the primer. They are peripheral to Sagnac, but they are the inventor's own experimental work. The primer's "Kelly's own works" list should name them.

8. Investigator lead dropped: Essen (1977). Essen's Hafele-Keating critique (Creation Research Society Quarterly 14, 46), cited by Kelly 2000, is in the contrarian's review only. The arbiter and the primer keep the Hafele-Keating numbers but not this earlier critique, which Kelly built on. Minor.

---

4. Accuracy spot checks (against the extracted text and the saved files)

Accuracy note A1 (Wang's ride-along gyroscope). Wang et al. write: "The FOG travelled with the mechanical conveyor and its uniform motion did not cause any phase shift because the FOG is only sensitive to the rotational movement" [read: Wang p.2]. They also write that the spare 50 m coil "moved uniformly with the FOG so as not to produce any phase shift". Two corrections follow.

• What the statement is. It is part of the design, stated with a reason. It is not reported as a separate controlled null measurement. The arbiter's "the one relevant experiment is a null" and the primer's "what Wang saw for the gyroscope riding along the conveyor" (section 5, Motion 3; table row 1b) state it a little too strongly. Suggested wording: "Wang et al. state that the FOG's own uniform motion produced no phase shift, which is what the line integral predicts."
• How the FOG moves. The primer (section 4, "exact point of departure") says the source and detector "ride round with it". In fact the FOG translates with the conveyor; it does not go round the loop. The physics argument is unaffected.

Accuracy note A2. The arbiter's absence-of-evidence item 3 says Kelly published, per Vigier, "in Phys. Lett. A through a citing author". That wording could suggest Kelly published in Phys. Lett. A. He did not: Vigier did, and cited Kelly's 1995 monograph. The sentence should be reworded.

---

5. Retention

Check: was pertinent material from the source, or from the Investigator's findings, left out of the arbiter or the primer because a reviewer judged it negatively?

• Kept well. All of the following are in both the arbiter and the primer, with sources and without dismissive labels:
• Kelly's Hafele-Keating per-clock numbers;
• the Irish Times 1996 dispute, described as documented in one sympathetic published source;
• Cowan's memoriam (with his stake noted);
• Vigier's citation;
• Marett's Michelson-Gale recomputation;
• Wang's zero-area and figure-8 results;
• the Einstein 1914 Harress notes;
• the open hollow-core question.

The arbiter's absence-of-evidence analysis follows SOURCE HONESTY well: it sets out competing explanations and says what record would decide between them. The strongest case for Kelly under claim 1 is stated explicitly. - Defect R1 (Michelson 1913 rotating-mirror test). Kelly offers it as a "clear confirmation", and it is his second experimental pillar for claim 2. Only the mathematician mentions it, and only as unchecked. It is not in the arbiter or the primer. This is not a deliberate negative exclusion; it was missed because pp.6-8 had no deep read. The effect is the same, though: the reader never learns that Kelly had a second experiment. - Defect R2 (ring-laser standing-wave argument). Kelly's third line of evidence for claim 2, missing from every file. - Defect R3 (Kelly's "selective quotation of Einstein" and GPS-frame inconsistency charges). These appear in the arbiter only as the reviewers' rebuttal of the orbital-angle numbers. Kelly's own argument, in his words, is absent (section 3, items 3 and 5). - Minor R4. The Faraday monographs and Essen 1977 are dropped (section 3, items 7 and 8).

---

6. Process

---

7. Remediation actions

Deep reads (most important first)

1. Kelly-2-1a1021, pp.6-8, Sonnet. Questions: - The Michelson 1913 rotating-mirror test as Kelly describes it: apparatus, the 1800 rpm figure, the three conclusions, the 1:50 accuracy, the factor-of-2 errors and the unnamed correspondent. - The ring-laser section: render p.7 to confirm the beat formula df = 4AΩ/(λP) and the numbers given for Macek & Davis 1963 and Bilger et al. - The standing-wave "nodes pass an observation post" argument, with what Kelly attributes to Martin 1986 and Anderson 1986, and the "in small scale experiments" qualifier. - The lock-in "luck" passage. - The Einstein 1905, 1916 and 1919 quotations, and Kelly's "selective quotation" charge. - The Harress and Pogany description.
2. Kelly-2-1a1021, pp.9-10, Sonnet. Questions: - The inertial-frame argument: every number (1.6 deg, 2.9e-6, 2.3e-10, 1.66e-13, 2.8e-13, the 1e7 and 5e6 ratios; the Bilger square of side 0.866 m on an 8,750,000 m disc), with a check of each. - The GPS-consistency charge, in Kelly's words. - The full Post (1967) quotation. - The Saburi (0.9 m) reference. - Kelly's exact description of Wang et al. 2003. - "What have we measured" (p.10).
3. Kelly-2-1a1021, pp.10-12, Sonnet. Questions: - The "SR versus Sagnac" section: Kelly (2001) and the Post statements about the dilation factor and the Doppler correction. - The p.11 description of Sagnac's layout (50 cm disc, 666 cm^2). - The claim of a "first general derivation without an ether". - p.12 item 3, the Doppler point. - Confirm from the p.12 page image that "4nA" is a garble of "4 pi n A".

Other actions

• facts.json: add key claims for Michelson 1913, for the ring-laser standing waves, and for the "selective quotation"/GPS-frame charge. Add Saburi as an unreferenced citation. Update how_read after the deep reads.
• claims.md: under claim 2, list Kelly's three lines of evidence (Sagnac-type derivations; Michelson 1913 rotating mirrors; ring-laser standing waves) and the qualifier "in small scale experiments". Under claim 5 (or claim 3), add Kelly's Einstein-quotation and GPS-frame inconsistency arguments in his words.
• arbiter.md, "QE revision" section:
• Weigh Michelson 1913 and the ring-laser standing-wave argument under claim 2. Expected: both support source-independence and "c in a non-rotating frame", and neither discriminates against SR. Expect no change to the claim 2 verdict, but state it.
• Replace "Tested, supported" with a scale verdict (P1).
• Fix the claim 2 label inconsistency (P2).
• Reconsider the claim 4 orbital-angle label (P3).
• Reword the "Wang null" (A1).
• Note that the Stedman 1995 paper is a companion to Kelly's cited Bilger IEEE papers (spot check 7).
• Reword the Vigier/Phys. Lett. A sentence (A2).
• Cite the deep read's p.12 image for the 4 pi n A garble (S2).
• primer.md:
• Add Michelson 1913 and the ring-laser standing-wave argument to section 4, and add a row or note to the section 6 evidence table.
• Add one paragraph explaining Kelly's orbital-angle/GPS argument and the reply (rotation rate, not angle swept) for the reader.
• Fix "ride round with it" and the Wang ride-along wording (A1).
• Bring the claim 1 and claim 2 verdict labels into line with the revised arbiter.
• Add Monographs 5 and 6 (Faraday experiments, 1998) to "Kelly's own works" and Essen 1977 to the related reading.
• Investigator, next run (not part of this remediation): the targets not reached in sites.md, prioritised as follows: the 1996 letters in the Irish Times archive; the Hafele 1971 PTTI report (to check Kelly's Table 3); the full text of Vigier 1997; Michelson 1913 (Astrophys. J. 37, 190), to check Kelly's account of it; Saburi 1976.

---

Recheck

Date: 2026-10-04. Re-read claims.md, facts.json, arbiter.md (including its "QE revision" section) and primer.md, checked by search against each defect above. example.py was not re-run (not touched by the fixes).

Remaining defect (1): 1. primer.md line 205 (and, for clarity, line 65): wording says the source and detector go round the loop with the conveyor. Wang's FOG translates with the conveyor and does not circulate. Suggested text: "the fibre circulates: its two straight runs move in opposite directions while the source and detector move in a straight line with the conveyor."

Final: OPEN (1 issues)