Liber Ruth

Translated 2026-08-18, one translator, 3 chunks (2,457 Latin words), on Opus at Wilson's call. verify-english.mjs clean: columns, notes and sections aligned, no duplicate paragraphs. Apparatus: 1 [sic:] · 3 [var:] · 0 [cj:] · 0 [d:] · 0 [ed:] · 0 [nt:] · 23 [n:] (exact parity with the Latin twins). Structure: CAPUT PRIMUM, CAPUT II, CAPUT III, CAPUT IV. Zero guillemets in either language.

This work is the conventions pilot for the whole Glossa block (62 works, ~783K words, PL 113–114). What §1 settles binds the books that follow; what §2 records is this book's own trouble.

§1 · THE GLOSSA CONVENTIONS — SETTLED HERE

1. The verse mark stays; the lemma is Englished. This was precedent, not invention.

Migne sets each gloss as VERS. 1.-- Facta est. — a verse address, then the Vulgate words being glossed, then the comment. The queue note asked whether to keep the Latin lemma beside the English or go English-only. The corpus already answered: work 21413 (Alcuin, In Cantica canticorum, PL 100) is shipped and uses exactly this form, and its English keeps VERS. n.-- verbatim while translating the lemma — VERS. 1.-- Veni in hortum meum → VERS. 1.-- Come into my garden.

So: VERS. n.-- is an address and passes through untouched. The lemma is translated. The Latin lemma is never lost to a reader — it stands opposite, on the Latin side of the same page. This is also the shape of Wilson's 2026-08-18 ruling on Migne's conjecture notes: "the cj: is for scholars where the english is for all."

⚑ Look for the precedent before ruling. The convention question was live for a day and was answered in one grep of the shipped English. The corpus is its own control here as much as it is inside a single work.

2. A fragmentary lemma is rendered as the phrase it stands for

Most lemmas survive word-for-word — Quando hora → "When the hour", De vestris → "Of your own". A few are fragments that are not English at all: VERS. 3.-- Non te. (Ruth 3:3, non te videat homo). Rendered "Let no man see thee" — the phrase the fragment opens, not a word-for-word stump. The alternative ("Not thee") would be faithful to the letter and useless to a reader, and the Latin fragment is opposite in the parallel column.

3. The attribution sigla need no new convention — they are already [n: …]

Migne credits his sources inline: (RAB.) for Rabanus, (ISID.) for Isidore, and in other books (BEDA, ubi supra.), (GREG., lib. 24 cap. 13.), (HIER.). These reach us as ordinary [n: …] notes, and the locked register already governs them: markers pass through verbatim, Latin abbreviations unanglicized. Ruth carries three (RAB.) and one (ISID.), all untouched. No sigla convention is needed, and none should be invented.

⚠ The sigla are Migne's attributions, and the whole recension's attributions are known to be unreliable (the "Walafrid Strabo" ascription is a myth; the honesty label on every Glossa work page says so). Passing a siglum through verbatim is not endorsing it.

4. The standard work-page template is sufficient

No lemma-aware layout is needed. The glosses are ordinary paragraphs opening with a verse mark; the parallel columns, the column anchors and the note styling all behave. Revisit only if a long book (Psalms, 106 chunks) reads badly at length — not on this evidence.

### 5. Migne prints no conjecture apparatus in this volume — and that is recorded as a ### CHECKED ZERO, not as a skip

All five of Ruth's plate pages were read at 300 dpi before translating (archive.org patrologiaecurs04migngoog, pp. 271–275; page = (column + 11) / 2, verified against the running head). p. 271 carries the Judges/Ruth division mid-page — the case that lost three notes on PL 202 — and was read there. No foot-of-page notes anywhere. Five further pages across two other runs of PL 113 gave the same result. data/plate-notes/coverage.json records Ruth as tier 3, coverage full, notes_recovered: 0, with the method: a later reader must be able to tell a book that was checked from a book that was skipped.

§2 · THIS BOOK'S CRUCES

0534A · significaRuth — the digitization joined two lines and ate a letter. PATCHED.

The TEI reads quos significaRuth. The plate reads quos significat Ruth., broken across a line at the head of the right-hand column (p. 272, col. 534A, line 1; read at 300 dpi and again magnified). This is Corpus Corporum's transcription, not Migne — so it is a data/tei-patches/8968.json fix, applied before chunking, exactly as the rule provides. The plate is untouched and unemended.

0537D · VERS.9 — the same class, punctuation only. PATCHED.

The TEI drops the space after VERS. at this one verse mark out of twenty-three. The plate sets it like every other (p. 274). Patched, for the same reason: uncorrected, the Latin side of the page shows VERS.9 and the defect reads as ours.

0533D · effoetandi — Migne's own, and it does not construe. RENDERED LITERALLY.

et maxime illud, ut superata a veritate infirmam se cognoscit effoetandi: « Revertimini, filiae meae. » The genitive gerund has nothing to govern it — infirmam se cognoscit wants an infinitive or ad with the gerund. Checked on the plate: Migne prints it so (p. 272, left column). Rendered "she acknowledges herself too weak for child-bearing" — the gerund taken with infirmam, which is the one construction the words allow. ⛔ The first draft read "too weak, and spent with child-bearing", inventing a coordinating conjunction that is nowhere on the page; both blind readers caught it independently (§3). No conjecture is offered: the fault is legible, it is Migne's, and Pattern 7 governs. A reader who wants to see the difficulty has the Latin opposite.

0535B · sul tectum meum — broken type. [sic: sul].

Domine, non sum dignus ut intres sul tectum meum (Matt. 8:8), for sub. Legible and plainly a broken b; carried through under Pattern 12 so the oddity reads as the plate's and not as ours.

0534A · percepta gratia baptismi et societate fieri — strange, and Migne's

"having received the grace of baptism and its fellowship" renders it; fieri sits loose in the clause and has no counterpart at all in our English, because it will not construe where it stands. Verified on the plate (p. 272, left column, VERS. 14) — it is printed thus. societate fidei, the phrase used at 0533C, is the obvious suspect and is deliberately NOT adopted. Not marked, and the reason is worth stating: nothing here is broken type (Pattern 12), nothing is missing from the digitization (Pattern 13), and no marker in this corpus means "Migne prints a word that will not construe". Saying so here beats inventing one for a single site — but a reader comparing the columns will see a Latin word with no English, so it must be said somewhere, and this is where.

0534C, 0535C · Migne's scripture citations are printed as he printed them

[n: (Marc. VI, XVI; Joan. XVI)] on Elegi vos de mundo and [n: (Matth. XVI)] on the Great Commission (Matt. 28:19) are not our errors and are not corrected in the text. Citation repair belongs to data/citation-corrections.json at index time, never to the running text — the standing rule, and this book gives two clean instances of it.

0534C · Et alibi- — a hyphen where a colon belongs

Punctuation only, in the TEI; not checked against the plate, and it changes nothing in the sense. Recorded so a later reader knows it was seen and let alone, not missed.

0534B · quando lex incarnationis suae ordinem … convertit — an ambiguity kept

suae may be taken with incarnationis (the order of his incarnation) or read with lex. Rendered "when the law turned the order of his incarnation to the mystery of the passion of Christ", which the following sentence supports (the barley harvest = the time of the passion). The other reading is not excluded; nothing in the argument turns on it.

§3 · THE BLIND POLARITY READ (2026-08-18) — 17 sites, 13 of them OURS

Two fenced readers on Opus, neither permitted cruces.md, this file, the style rulebook or any grep of the English directory; each read all three chunks. Different lenses on purpose — polarity and unexpressed subjects for the first, completeness/construction/scripture- conformation/etymology-pairing for the second. Both reported clean chunks and rejected candidates with reasons, as required.

⭐⭐ The work's characteristic fault was NOT polarity. It was SILENT REPAIR of Migne's own readings — four scripture words quietly conformed to the familiar English Bible. Not one negation was dropped or added anywhere in the work; both readers checked the clusters at 0533C–0533D, 0536B and 0537B–0537C and cleared them. The defect class that actually cost us was the one Pattern 7 exists to forbid, committed by a translator who was marking [sic: sul] correctly two columns away. Consistency in using the apparatus is not evidence that the apparatus was used where it was needed.

The readers DISAGREED on the most serious site, and the Vulgate settled it

0538C — Quem scilicet per legem gignere non potuit … et nutricis fungebatur officio, quia matris carebat privilegio. Reader A called the unexpressed subject wrong (HIGH); reader B read the same 46-word chain and reported it "rendered complete and correctly parsed."

A is right, and the evidence is not a judgement call: the gloss lifts its words from Ruth 4:16 — susceptumque Noemi puerum … et nutricis ac gerulae fungebatur officio — where the subject is NOEMI, who is the synagogue on every page of this book. Both susceptum and nutricis fungebatur officio are that verse. The English had four bare "she"s whose nearest antecedent was the Gentile Church, which made the Gentile Church into Christ's nurse rather than his mother — the exact inversion of the gloss's argument. Fixed by naming the subject.

⚑ Two readers agreeing proves nothing and two readers disagreeing proves nothing; the source text decides. B's lens (parsing) could not see this, because the sentence parses perfectly either way. It is settled only by knowing which verse the words come from.

Fixed — OURS (13)

Three [var: …] markers now stand where a Vulgate divergence was being erased (0533B, 0535C, 0538A). That is what Pattern 14 is for, and it is the right apparatus here: the printed reading keeps its English, and the received text stands beside it for the reader who wants to know why the words look strange.

Migne's own, flagged and left standing (4)

• 0533D effoetandi — a genitive gerund governing nothing; see §2.
• 0534A societate fieri — fieri is a whole word on the plate with no counterpart in our English, because it cannot be construed where it stands (societate fidei, the phrase used at 0533C, is the obvious suspect and is NOT adopted). Recorded here rather than marked: no marker in this corpus means "Migne prints a word that will not construe", and inventing one for a single site is worse than saying so plainly.
• 0535B sul tectum meum — broken type, carries [sic: sul]. Correctly handled.
• 0537A cum praecursoris sui adventum quem … conspexit — one accusative too many and quem with no antecedent it can take. Our English resolves it onto the forerunner and supplies the object of convertit; the resolution is defensible and is the only one that construes. Nothing is emended.

Rejected, with the reason (both readers proposed these; both were declined)

• Aethiopia praeveniet manus ejus Deo → "shall soon stretch out her hands" — Douay wording, but Migne's Latin is verbatim Vulgate Ps 67:32, so the traditional echo IS the literal rendering. The register permits it exactly there.
• Lege, scilicet quae… read as a nominative heading — the ablative is recovered two clauses later at "In this the Church dips her morsel"; no reader is misled.
• Imprecatur prospera → "He prays" — the singular verb has a singular governing noun (Decem majorum natu benedictio), and this gloss routinely speaks of a collective in the singular (Agnoscit synagoga, Cognoscit mater Ecclesia).
• incarnationis suae → "his incarnation" where lex is the grammatical subject — reflexive suus for the logical subject is ordinary here, and "the law's own incarnation" is nonsense.
• [n: (Matth. XVI)] on Ite, docete omnes gentes (Matt. 28) — Migne's citation misprint, rendered as printed. Citation repair belongs to citation-corrections.json at index time.
• Non te → "Let no man see thee" — the expansion supplies Migne's own printed verse, not a foreign one; noted as an inconsistency with the shorter lemmas, kept as §1.2 rules.