QE (Step 4b): Hieronymus-TG

QE run 2026-10-04 (Opus). Inputs read: _text/manifest.json, digest/US2482773-f82c2d.01.md, deep/US2482773-f82c2d.md, facts.json, claims.md, sites.md, all seven research/ notes, physicist.md, mathematician.md, contrarian.md, arbiter.md, primer.md, example.py. Checked against the extracted text (_text/US2482773-f82c2d.txt) and a fresh render of patent page 1 (Figs. 1-4) at 200 dpi in /tmp/qe-hier [read: page image p.1].

Verdict: MINOR

The work is sound. The project has one document, it was fully read, and the verdicts that matter rest on calculations that I reproduced. The defects are: findings by the Investigator that the arbiter and primer left out; two statements in the primer whose sources are wrong; facts.json and claims.md not updated with corrections the reviewers agreed on; and one arbiter sub-verdict (the isobar extension of claim 3) that rests on a keyword skim of a book nobody has read in full. None of these changes a headline verdict, so it is not necessary to redo Steps 2-4.

1. Coverage

• No documents were skipped. The manifest has one entry, and the textprep log matches it (1 document, 1 chunk).
• The project folder holds no secondary or earlier AI-written analyses.
• research/files/ is empty. The Investigator found four key items but downloaded none, because each download needs the owner's approval: the 1988 autobiography (59 MB), JBR May-June 1988, Hite's PDF, and Astounding June and Aug 1956. Not downloading them follows the process. The consequence is that the only reading of these sources is the Investigator's keyword-search summary. See 2 and 6.

2. Shortcuts

• Patent: acceptable. It was read in full, and the digest's mistakes were caught by the deep read (for example the garbled numerals "detector 12" and "unit 10", which are really 72 and 70, as the GB 663,978 abstract confirms).
• Autobiography: a defect, minor. The Investigator read the autobiography's OCR text by keyword search only, and says so (research/2026-10-04-archive-org.md: "read by keyword search, not in full; a deep read is recommended"). Three verdict elements rest on that skim:
• (a) the arbiter's claim 3 (ii), that the isobar extension is "Refuted by calculation";
• (b) the Be ~186 / Bi ~1097 consistency point in primer Step 6;
• (c) the "test plan, not stated whether done" wording.

Item (a) carries a hidden assumption. The 0.0007 deg vs 0.028 deg argument assumes that Analyzer No. 508's "prism dial setting" uses the same scale and resolution as the 1600-division quadrant from pp.237-238. The two numbers come from different passages (pp.266-267 vs p.233/237), and possibly from different instruments. The Investigator also reports only that he "computed" the factor and "plotted factor against prism dial setting". The skim does not show whether he claimed that the instrument separates Ti-50 from Cr-50 by angle. The arbiter's second point holds whatever the book says: the free-nucleon factor puts the pair in the opposite order to their real masses, and I verified those numbers. The resolution half is conditional until the book is read. - Campbell 1956, Hite and Tompkins & Bird: acceptable. All are summarised from the Investigator's notes only, and every reviewer and the arbiter say so. No headline verdict depends on them, apart from the "works unplugged / operator-dependent" observations. Those are correctly treated as reported observations.

3. Missed material

Material in the patent that is missing from claims.md, facts.json and the primer (from the digest and deep read):

1. The converse claim [read: p.3]: "Conversely, the flow of electrons along a conductor produces a radiation having characteristics of the radiation from each respective element." Only the mathematician touches on it. This is the patent's seed of the later "broadcast"/treatment mode, which the 1972 notes say needs power. It belongs in claims.md as part of claim 2, or as a separate claim.
2. The lens variant (Fig. 4; patent claims 10-12, "a lens having a focal point for each known element"; the scale is read along the axis) [read: p.4-6; page image p.1]. Three of the 13 patent claims depend on it. claims.md and the primer do not mention it. The dispersion argument applies to it too: focal length varies with n only through (n-1), a change of about 15% over 400-200 nm. One sentence in the primer would cover it.
3. "Radiations of the elements from the light" [read: p.3]. The plant paragraph says sunlight carries element radiations. This is the patent's own link between claims 2 and 6. The primer's claim 6 row and the departure table could quote it.
4. "Which zone has as yet not been fully explored" [read: p.2]. Only the contrarian notes this (and rebuts it). It is a short historical point worth one line in the primer: UV spectroscopy was mature by 1946.
5. Attorney. Signed on p.1; the signature reads "[C.] Earl Hovey" on my render, which matches the autobiography's "C. Earl Hovey" (Investigator). The deep read called it "illegible". This is minor and goes in facts.json.

4. Accuracy spot checks (against _text and page image p.1)

Hidden assumption in the worked example. Fig. 3 itself stops at A ≈ 120. The primer's 52-58 deg range (H to Bi) extrapolates the line to A = 209. The text supports the extrapolation ("continue up to include all known elements when the full size chart is employed", deep read §3), and so does the inventor's later Bi ≈ 61.7 deg. Without extrapolating, the drawn range H to A = 115 is still about 29-32 deg against 3-6 deg available. The conclusion stands. The primer should state the extrapolation and the robust lower figure.

5. Retention

Investigator findings missing from both arbiter.md and primer.md, or present in the arbiter but dropped from the primer. Most of these look like condensation rather than removal because of a negative judgement, but the rule makes their absence a defect either way:

6. Process

• Source tags. Present throughout. Small gaps:
• Primer §1 states biography (KCP&L engineer, 1895-1988) untagged. It should carry [web] Hite / Wikipedia / autobiography.
• Primer §6, claim 2 row attributes the ink-drawn "symbolic" prism to "Campbell (1956)". The Investigator's notes on the June and Aug 1956 issues do not mention the symbolic machine. That comes from Campbell's letters reproduced in the autobiography (around p.123) and from Wikipedia. Its date, and whether Campbell reported that it worked, are not established in the research notes. The mathematician's "Campbell's reported success with a prism 'drawn in ink'" has the same problem. Retag as "[web] autobiography ch. on Campbell letters; Wikipedia; date and reported result unverified".
• Primer §7 "Solid" calls "the patent's detector (a lacquered metal plate on a tube amplifier fed from 60 Hz mains)" a textbook electrovibration setup. The patent does not specify the power supply or AC on the plates. That detail is from Hite's Pathoclast tracing and the 1972 notes. Reword so it does not attribute this to the patent.
• Web checks.
• Physicist: used 2 of about 6. Its medium-confidence items (skin depth, flash-greening, ultra-weak emission, patent examination practice) are declared not load-bearing.
• Contrarian: its one check failed. Its glass-dispersion statement was [memory: medium], which the arbiter replaced with a recomputation.
• Arbiter: 5 of 8 used, with useful upgrades (electrovibration 1954 parameters, LPOR/DPOR, Ti/Cr masses).
• Gap: the Malitson Sellmeier coefficients, which every claim 1/3 verdict uses, are still [memory: high] and unfetched. The values in example.py are the standard published set (0.6961663, 0.4079426, 0.8974794; 0.0684043, 0.1162414, 9.896161 µm) [memory: high]. Acceptable, but a single fetch would close it.
• Verdict scale. Physicist claim 1 heads its verdict "Tested, not supported in the strong form". No test is on record, and its own split says "Refuted by calculation", so the label is misapplied. The arbiter corrected it. Elsewhere the scale is applied consistently, and the split verdicts are well defined.
• Primer structure. All 9 STEP 4 sections are present in order, with answers at the end. Symbols are defined (λ, f, c, h, k, T, n, θ, α, δ, L, C, Z, A, p). The statistics refresher meets the reader profile. Textbooks are named without guessed chapter numbers.
• example.py. Runs with the project venv and prints "All checks passed". Its numbers match primer §5 exactly.
• Interested parties and absence of evidence. Both are handled well. The arbiter's analysis of the record going dark (five explanations, evidence for and against each, and the records that would decide) follows SOURCE HONESTY. The Hite "unsourced admission" is kept as his assertion, as it should be.
• Investigator. Time-capped run, queries recorded, and the sites not reached are queued with "Recheck? yes". The GB 1951/1952 and corporation 1967/1975 date conflicts are correctly left open.

7. Remediation (MINOR)

Deep read (manifest document): 1. US2482773-f82c2d, pp.2-6 (Sonnet). Extract and quote the converse electron-flow statement (p.3), the "radiations of the elements from the light" interpretation (p.3), the "not been fully explored" sentence (p.2), and the lens variant (Fig. 4; claims 10-12: focal point per element, axial scale). Say how each extends claims 2, 3 and 6.

Sources outside the manifest (owner approval is needed to download them into research/files; then deep read): 2. Autobiography The Story of Eloptic Energy (1988), pp.88-90, 123-124, 173, 227-242, 266-267, 308, 315 (Opus). Questions: - Was the inter-operator / inter-instrument test plan (p.233) carried out, by whom, and was it blind? - Which instrument and scale produced the Be 186 / Bi 1097 readings, and which produced the No. 508 dial settings? - Does he claim that the instrument separates Ti-50 from Cr-50 by angle, or only that the computed factor does? - What does the Campbell correspondence say about the symbolic machine's results and date? - Are there data tables? 3. Astounding June and Aug 1956, plus the 1957 issue (vol. 58 p.63) (Sonnet). Confirm the operator, unplugged and defective-tube statements verbatim. Find where and when the symbolic machine was first reported. 4. Hite PDF, pp.4-14 and 33-41 (Sonnet). Confirm the production numbers, the 1972 notes verbatim, and the source (if any) for the "admitted wording" claim.

File updates: - facts.json: - Hall US 2,115,578 is a thermal-pneumatic "Radiant energy receiver" [web]. - Attorney: C. Earl Hovey (p.1 signature; autobiography p.89). - Add the lens variant and the converse claim to key_claims. - Add patent family GB 663,978 / CA 486,008. - Note that the Fig. 3 drawing stops at A ≈ 120. - claims.md: - Claim 4: add patent claim 6's "unaffected" end point beside the specification's wording. - Claim 6: "the control plant" (singular), "relatively healthy", no species or counts; add the patent's own interpretation (radiations "from the light"). - Claim 2: add the converse (electron flow produces element radiation). - Claim 3: mention the lens variant. - arbiter.md, new "QE revision" section: - Mark claim 3 (ii), the isobar extension, as conditional. The resolution argument holds only if the No. 508 dial shares the quadrant's scale and he claims angular separation. The factor-ordering argument stands regardless. - Add the retention items from §5 to "Investigator findings retained". - Note that physicist claim 1's "Tested, not supported" label was superseded. - primer.md: - Fix the claim 2 row's attribution of the symbolic machine. - Reword the §7 "patent's detector ... fed from 60 Hz mains" sentence. - Tag the §1 biography. - Add Campbell's age/training pattern and the defective-tube observation to the claim 5 row and to the open question. - Add Hieronymus's objection to the symbolic machine. - Add the production mix (about 10 prism analyzers vs about 200 medical units) to "Other material kept". - Add the mineral-prospecting / moon-cycle claim, the atomic-test and isotope "radiation distance" claims, and the other Apollo claims, with sources. - Add one sentence on the lens variant and the "not fully explored" remark. - In §5 Step 4, state the extrapolation beyond the drawn chart, and the robust figure (H to A = 115 needs about 30 deg, still 5-10 times the dispersion). - Mark the isobar sub-verdict as conditional, matching the arbiter's QE revision. - sites.md (next Investigator run): add Astounding 1957 (vol. 58) and the Campbell symbolic-machine source to the queue; keep the queued USPTO, Espacenet, Missouri/Georgia, newspaper, IAPSOP and WorldCat items.

Recheck

Rechecked 2026-10-04 against the current facts.json, claims.md, arbiter.md, primer.md and sites.md.

Fixed: - Section 3 missed material (converse, lens variant, "from the light", "not fully explored", attorney): now in facts.json, claims.md (claims 2, 3, 6), arbiter.md "QE revision" and primer (rows for converse, lens, Step 5b). - facts.json updates (Hall 2,115,578 described, Hovey, GB 663,978 / CA 486,008, Fig. 3 stops at A ~120): done. - claims.md updates (claim 4 "unaffected" end point, claim 6 singular control plant and "relatively healthy", claim 2 converse, claim 3 lens): done. - Isobar sub-verdict marked conditional in arbiter.md (3 (ii) split) and primer claim 3 row; physicist "Tested, not supported" label noted as superseded. - Section 5 retention items (Campbell age/training and defective-tube, Hieronymus objection, production mix, mineral/moon claim, isotope and atomic-test claims, Apollo extras): added to arbiter.md and primer.md. - Primer source and wording fixes: symbolic-machine attribution retagged and flagged unverified; Solid paragraph no longer attributes 60 Hz mains to the patent; Step 4 extrapolation and the robust 28.5-31.9 deg figure stated; section 1 biography tagged.

Remaining: - sites.md: the queue still lacks the Astounding 1957 (vol. 58) item and a dedicated entry for the source of Campbell's symbolic-machine report (only a lineage note exists at line 66). Minor.

Not defects, still open by process: remediation items 2-4 (autobiography, Astounding 1956, Hite PDF deep reads) need the owner's approval to download. The primer and arbiter carry the "keyword search only" caveats and conditional wording meanwhile.

Final: OPEN (1 issues)