Liber Paralipomenon II

Anselm of Laon and his school, on II Paralipomenon. Translated 2026-08-25 by two translators working in parallel — chunks 0000–0005 (cols 0663D–0678A, 5,241 Latin words) and chunks 0006–0010 (cols 0678B–0692C, 5,592 words) — and merged here. 11 chunks, 10,833 Latin words (by chunk frontmatter) → 16,327 English, 1.51×. verify-english.mjs clean over the whole work: column anchors, note parity, section alignment, no duplicate paragraphs.

Apparatus fired, work-wide: 15 [var:] · 2 [cj:] · 0 [sic:] · 0 [d:] · 0 [ed:] · 0 [nt:] · 0 [cn:]. 122 [n:] notes at exact parity. Zero guillemets in either language; zero em-dashes in the English; question marks 3 Latin / 3 English in the body.

⚑ A raw count over the English files returns a fourth ? and it is not a defect: it stands inside the [var:] at 0669A, where the marker quotes the Clementine's Nonne omnes sunt … qui haereditatem capient salutis? — a question Migne's own lemma prints as a flat assertion (capiunt). The translator carried the assertion, declined to supply the mark, and put the Vulgate's interrogative where it belongs, in the marker. That is Pattern 8 working exactly as written, and it is why parity is counted on the body and not on the file.

Ruth (8968) cruces.md §1 governs and was not reopened: VERS. n.-- passes through verbatim, the lemma is Englished, a fragmentary lemma is rendered as the phrase it stands for, and the attribution sigla ride as ordinary [n: …] notes with their contents unanglicized. Tobit (8970) §1(b) governs the abridgment formula: usque ad → as far as, with no verb ever supplied to complete a resumption tail. I Regum (8952) is the model, never the authority.

§M · THE MERGE — what was reconciled between the two stints, and what needed nothing

Two translators cannot see each other's work, and they diverge on terminology without knowing it (11551 shipped one half calling a column's members stylus/epistylium and the other "epistyle"/"shaft", which left the author's own etymological argument explaining nothing). So the merge is a check, not a concatenation. Both stints stated their vocabulary before it could be compared, and the comparison was then made against the English files themselves, not against the tables — a table records what a translator believed it did.

Checked across the seam, and consistent without intervention:

• The number-names. senarius · centenarius · denarius · binarius · quadragenarius · duodenarius → the senary · the centenary · the denary · the binary · the quadragenary · the duodenary, never flattened to "six" or "a hundred", because in CAPUT IX–X the number-name is the argument. All 12 occurrences fall inside chunk 0006; the first stint's range contains none, so there is no seam here to fail at.
• Ecclesia / ecclesia. Capitalization mirrors the plate, word for word. The second stint stated the rule; the first had followed it silently — its single lowercase "the church" (0003, in the atrium sacerdotum gloss) stands opposite the single lowercase ecclesia its Latin twin prints, against three capitalized forms in the same chunk. Verified at the site, not taken on trust.
• doctores → teachers, never "doctors", in both stints and across nine of the eleven chunks (18 occurrences). praedicatores stays preachers, held apart.
• pascha → pasch, in both (0001, 0009, 0010). No "Passover" anywhere in the work — which matters here, because 0010 is where Migne prints phase and Phase beside it and the second stint kept those form for form. Flattening either word to "Passover" would have destroyed a distinction the plate itself makes.
• The abridgment formula, rendered as far as throughout, every occurrence inside its own italic span answering an italic usque ad in the Latin. Both stints decided usque per occurrence rather than by rule, which is what the brief requires: 50 tokens work-wide, of which 10 sit inside another word (utriusque, cujusque) and the remainder split 28 formula / 13 ordinary prepositional uses. The trap at 0680D — where usque ad is not the formula but 9:26's own words — was caught.
• Pattern 17. thou = singular, you = plural — 21 second-person singular forms in the English (thou 7, thee 4, thy 10), each named in its stint's table with the Latin that fixes the number.
• Proper names kept form for form as Migne prints them, in both — Philitshiim, Neopolis, Eroge, Amulcus, Ononimus, Zareas, Benennom, and Jesus where Migne writes it for Josue. None normalized toward a familiar English Bible spelling.

Held apart deliberately, and stated here because a later sweep would otherwise churn it: charitas is charity and dilectio is love, throughout. The pair is not decorative — the [var:] at 0668C exists because Migne's Rom. 13:10 lemma prints charitas where the Clementine prints dilectio, and an English that used one word for both would make that marker unreadable. The same applies to typice / figuraliter (typically / figuratively), scientia / sapientia (knowledge / wisdom), and significat / figurabat / designatur.

Nothing was rewritten in either stint's English at the merge. No drift was found that required it, and a merge that silently edits a translator's chunk destroys the only record of what that translator actually read.

One thing the merge deliberately does NOT do: reconcile the two lemma collations into one list. They cover disjoint column ranges, each was checked against clementine-flat.txt independently, and each names the lemmata it found to agree as well as those that diverge — roughly 200 spans between them. Flattening them would lose which reader checked what, which is exactly what a later blind read needs to know.

§7a · PLATE CANDIDATES, WORK-WIDE — one list, both stints

Consolidated because a plate read is bought a page at a time and both stints' candidates fall in the same tome. None fired a marker, so the work ships asserting nothing about the plate (ship test B): a [sic:] is a public claim, and seven false ones were withdrawn from the previous batch for exactly this reason.

Non-words and split runs — our digitization's error 19 times out of 19, so raised, never patched: first stint, latitudis · obediende · omma · subbatismum and the split runs holo causti · de cem · coena culum; second stint, eight more, with for es (0679B) and tranferunt/decjiciunt (0692B) the sharpest.

⚠ The split runs and the clipped-looking forms sit at or near a left-column head — which is exactly where our archive.org copy of PL 113 (patrologiaecurs04migngoog) is known to clip its left margin (p. 320 and p. 354 confirmed). That is a fact about our photograph, not about Migne.

⭐ And this work needs no download to settle any of them. data/volumes.json lists a second witness for PL 113 — Gallica ark:/12148/bpt6k5505319w, IIIF, server-side crops, one HTTP request per read, map f = (column + 1) / 2 (NOT the archive map, which is (column + 11) / 2). That is how both of the previous batch's "unreadable" candidates were settled in a single fetch each, and it went opposite ways on them: one suspicion withdrawn, one proved to be Migne's own broken sort. The candidates below are the cheapest open items in this work, and none of them blocks it.

⭐ ONE CANDIDATE WAS SETTLED AT THE PLATE, AND IT WENT MIGNE'S WAY — (Rabac. III) @ 0680C

Not raised by either translator, because it is inside an [n:] note and notes pass through verbatim: the citation index caught it instead, as the work's single unparsed reference. Rabac. is no book of the Bible, and an R for an H in Habac. is exactly the shape of an OCR fault — the non-word class that has been ours 19 times out of 19.

It is Migne's. Read on the second witness (Gallica bpt6k5505319w, f340 = cols 679/680, band C, one HTTP request, map f = (column + 1) / 2 confirmed against the printed corner numbers before the read): the plate prints sanitas (Rabac. III) in that form. So the non-word rule has met its first counter-instance from inside a citation, and the shape of the exception is worth naming — the rule was measured on words of the TEXT, and an abbreviated book-name in a note is a different class of thing, short enough that a single wrong sort produces a plausible-looking abbreviation rather than visible nonsense.

What was done: the printed form stays in the text verbatim, as the sacred-marker rule requires, and the index resolves to Hab.3.8 through data/citation-corrections.json, which keeps refDisplay verbatim and the printed key auditable. The work's unparsed count went 1 → 0. No Ibid. depends on the reference.

⚑ The same crop independently confirmed two of the second stint's calls, which is why it is recorded here and not merely in the corrections file: Migne really does print equitatus tuus sanitas against the Clementine's et quadrigae tuae salvatio (so the [var:] at 0680C stands on the plate, not only on the file), and he really does print usque ab conversionem, which that stint declined to touch with a [cj:]. Both declines were right, and one page settled both.

Translated 2026-08-25, one translator, 6 chunks, 5,241 Latin words (frontmatter counts: 638 + 1,588 + 147 + 1,511 + 930 + 427). verify-english.mjs clean over this range (the only failures it reports are the peer's still-missing 0009 and 0010).

Apparatus fired: 12 [var:] · 2 [cj:] · 0 [sic:] · 0 [d:] · 0 [ed:] · 0 [nt:] · 57 [n:] (exact parity with the Latin twins). Zero guillemets in either language, zero [cn:] (this volume's plate spot check is a checked zero). Question marks: 1 in the Latin, 1 in the English, same band. Em-dashes in the English: zero.

Ruth (8968) cruces.md §1 governs and was not re-opened: VERS. n.-- passes through verbatim, the lemma is Englished, a fragmentary lemma is rendered as the phrase it stands for, the attribution sigla ride as ordinary [n: …] with contents unanglicized. Tobit (8970) §1(b) governs the abridgment formula: usque ad → as far as, no verb ever supplied to complete a resumption tail. I Regum (8952) is the model, never the authority.

§0 · WORK-WIDE VOCABULARY — for the merge with the peer's 0006–0010

Fixed at first occurrence and never varied inside this range. Named here so drift between the two stints can be reconciled without re-reading either.

§1 · THE LEMMA COLLATION — every italic span, checked against sources/vulgate/clementine-flat.txt

Migne's book is 2Par in that file, numbering identical. A lemma named here with no comment AGREES with the Clementine at its own verse, verbatim, or is a clean clip of it — that is why the agreements are named. Everything that diverges carries its divergence, and the twelve that earned a public claim are argued in §3.

CAPUT PRIMUM (0663D–0665C)

Confortatus est ergo Salomon, (1:1, verbatim) · Data est mihi omnis potestas in coelo et in terra (Mt 28:18, verbatim; Clem. caelo, ligature orthography only, not a divergence) · Gabaon, (1:3, verbatim single word) · Da mihi sapientiam, (1:10, verbatim) · Ab excelso, (1:13, verbatim clip) · Tabernaculum autem Domini … eadem tempestate erant in excelsis Gabaon, (1 Par 21:29 — Clem. ea tempestate erat in excelso Gabaon. [var:] FIRED, §3a) · Et duodecim millia equitum. (1:14, verbatim) · Sexcentis argenteis. (1:17, verbatim) · Consummatum est. Et inclinato capite, emisit spiritum (Jo 19:30 — Clem. tradidit spiritum. [var:] FIRED, §3b) · Centum. (1:17 — a clip of centum quinquaginta; see §2a for why it is expanded) · Necesse est omnia impleri quae scripta sunt in lege Moysi et prophetis et psalmis de me (Lc 24:44 — Clem. necesse est impleri omnia; word order only, no marker, §4).

CAPUT II (0665C–0666B)

Decrevit Salomon aedificare domum, (2:1 — Clem. Decrevit autem Salomon aedificare domum nomini Domini; a clip, and the dropped autem, §4) · Salomon. (verbatim) · Et erunt servi. (2:8, verbatim clip) · Filium mulieris. (2:14, verbatim) · Qui novit operari in auro et argento. (2:14, verbatim; Migne's clip stops where the Clementine goes on to aere, et ferro) · Nos autem, (2:16, verbatim) · Super quem videritis Spiritum descendentem et manentem super eum, hic est qui baptizat (Jo 1:33, cited Matth. III — Clem. videris, SINGULAR. [var:] FIRED, §3c; the clip before in Spiritu Sancto is a clip and nothing else) · Numeravit igitur, (2:17, verbatim) · Tria millia et sexcenti. (2:17 — Clem. tria millia sexcenti, without et, §4).

CAPUT III (0666B–0670D)

In area Ornan. (3:1, verbatim) · Mense secundo, (3:2) · Anno quarto regni sui. (3:2) · Porticum vero, (3:4) · Et cedro omnis domus vestiebatur, (3 Rg 6:18 — Clem. Et cedro omnis domus intrinsecus vestiebatur. [var:] FIRED, §3d) · Stravit quoque pavimentum, (3:6) · Postes et parietes, (3:7) · Ego sum ostium, (Jo 10:9, verbatim) · Oraculum habebat viginti cubitos longitudinis, et viginti latitudis et viginti altitudinis (3 Rg 6:20 — Clem. prints cubitos three times and latitudinis; the dropped cubitos is compression, §4, and latitudis is a NON-WORD and a plate candidate, §2c) · Oraculum in abditis (no such phrase anywhere in the Clementine; the gloss's own interiori domo points at 3 Rg 6:19 Oraculum autem in medio domus, in interiori parte fecerat. No [var:] — the base verse cannot be fixed with enough confidence for a public claim, §4) · Coelavit cherubim. (3:7, verbatim; coe-/cae- orthography only) · Fecit quoque domum, (3:8) · Laminis aureis texit eam (2 Par 3:8, verbatim, though the note cites III Reg. VI) · Talentis sexcentis, (3:8 — Clem. quasi talentis sexcentis, clip) · Sed et clavos. (3:9) · Mihi autem adhaerere Deo, bonum est (Ps 72:28, verbatim; Migne's comma for Clem.'s none) · Coenacula quoque. (3:9) · Abscondes eos in abscondito faciei tuae, a conturbatione hominum (Ps 30:21, verbatim) · Quae sursum sunt quaerite (Col 3:1, verbatim — and note that the gloss's continuation ubi Christus est in dextera Dei sedens is the rest of the same verse, printed as roman prose, and is rendered as prose) · Fecit etiam. (3:10) · Ita ut ala. (3:11 — Clem. ita ut una ala haberet; the clip drops una, §4) · Plenitudo autem legis est charitas (Rom 13:10 — Clem. Plenitudo ergo legis est dilectio. [var:] FIRED, §3e) · Tangebant parietem. (3:12 — Clem. tangebat parietem, SINGULAR. [var:] FIRED, §3f) · Omnes enim sunt administratorii spiritus … qui haereditatem capiunt salutis (Heb 1:14 — Clem. Nonne omnes sunt … qui haereditatem capient salutis? — a question in the Vulgate, a statement in Migne. [var:] FIRED, §3g) · Ala ejus. (3:12) · Et facies eorum versae. (3:13 — Clem. et facies eorum erant versae; compression, §4) · Fecit quoque. (3:14) · Et intexuit. (3:14) · Ante fores etiam templi duas columnas. (3:15, verbatim) · Nemo venit ad Patrem, nisi per me, (Jo 14:6, verbatim) · Qui cum venisset ad regem Salomonem, fecit omne opus ejus, et finxit duas columnas aereas, (3 Rg 7:14–15, verbatim across the verse join, joined with a comma, §4) · Triginta et quinque. (3:15, verbatim) · Vobis datum est nosse mysterium regni Dei (Mc 4:11, verbatim, correctly cited) · Aperuit illis sensum, ut intelligerent Scripturas (Lc 24:45 — Clem. opens Tunc aperuit; clip) · Quasi catenulas. (3:16 — Clem. Necnon et quasi catenulas in oraculo, clip) · ( sic ) — Migne's own, reproduced verbatim; no [sic:] of ours stacked on it (runbook false positive 5) · Capitibus columnarum. (3:16) · Malogranata. (3:16) · Ipsas quoque. (3:17).

CAPUT IV (0670D–0674C)

Fecit quoque, (4:1) · Viginti cubitorum. (4:1) · Et decem cubitis. (under VERS. 1 — Clem. 4:1 has et decem cubitorum altitudinis; the identical lemma at VERS. 3 matches 4:3 et decem cubitis verbatim. Case only, and no part of the gloss turns on it, §4) · Mare etiam. (4:2) · Quinque cubitos habebat. (4:2) · Funiculus triginta cubitorum. (4:2) · Funiculus triplex difficile rumpitur (Ecl 4:12, verbatim) · Et decem cubitis. (4:3, verbatim) · Posteriora. (4:4 — a complete one-word lemma, rendered as it stands; contrast Centum. and Ducenti., §2a) · Grossitudo autem luteris trium unciarum (3 Rg 7:26 — Clem. trium unciarum erat, §4) · Trium unciarum, (re-quotation of the same) · Porro vastitas. (4:5) · Tria millia metretas. (4:5) · Fecit quoque conchas. (4:6) · Fecit quoque decem luteres aeneos. Quadraginta batos capiebat luter unus. Eratque quatuor cubitorum. Singulos quoque luteres per singulas decem bases posuit, (3 Rg 7:38 — Clem. per singulas, id est, decem bases, posuit; the dropped id est and full stops for Clem.'s colons, §4) · Et posuit quinque. (4:6) · Candelabra aurea. (4:7, verbatim) · Lucerna pedibus meis verbum tuum, et lumen, (Ps 118:105, verbatim clip) · Quia mandatum lucerna est, et lex lux (Pr 6:23, verbatim) · Necnon et mensas. (4:8) · Fecit etiam atrium. (4:9) · Et basilicam grandem. (4:9) · Non potui vobis loqui quasi spiritualibus, sed quasi carnalibus, (1 Cor 3:1, verbatim) · Et basilicam. · Et ostia in basilica. (4:9) · Fecit autem Hiram. (4:11) · Quorum intuentes exitum conversationis intueamini et fidem. (Heb 13:7 — Clem. quorum intuentes exitum conversationis, imitamini fidem. [var:] FIRED, §3h) · Et creagras. (4:11) · Et phialas. (4:11) · In argillosa. (4:17) · In regione Jordanis. (4:17) · Gaudebunt campi et omnia quae in eis sunt (Ps 95:12, verbatim) · Audivimus eam in Ephrata, invenimus eam in campis silvae (Ps 131:6 — Clem. opens Ecce audivimus; clip, §4) · Fecitque Salomon. (4:19) · Nihil sine me potestis facere (Jo 15:5 — Clem. sine me nihil potestis facere; word order only*, no marker, §4) · Omne enim datum optimum, et omne donum perfectum desursum, (Jas 1:17 — Clem. has no enim, §4) · Altare aureum. (4:19) · Et forcipes, (4:21) · Et ostia coelavit. (4:22, verbatim; coe-/cae-) · Sicque completum est omne.* (4:22).

CAPUT V (0675C–0676B)

Intulit igitur Salomon omma, (5:1 — Clem. omnia; omma is a NON-WORD and a plate candidate, §2d) · benedicit Dominus pusillis cum majoribus. (Ps 113:21 — Clem. Benedixit omnibus qui timent Dominum, pusillis cum majoribus. [var:] FIRED, §3i) · Vectium autem. (5:9) · Nihilque aliud erat. (5:10 — Clem. Nihilque erat aliud in arca; word order only, §4) · Cupio dissolvi et esse cum Christo (Phlp 1:23 — Clem. desiderium habens dissolvi, et esse cum Christo. [var:] FIRED, §3j) · Vestiti byssinis. (5:12) · Sacerdotes. (5:12, from et cum eis sacerdotes centum viginti) · Igitur cunctis. (5:13) · Compleverat enim. (5:14) · Vobis datum est nosse mysterium regni Dei, (verbatim Mc 4:11, here cited Matth. XIII, where the Clementine reads mysteria regni caelorum. No [var:]: the printed words agree with the received text exactly — it is the reference that points elsewhere, which is the refDisplay/refKey layer's business, not Pattern 14's, §4).

CAPUT VI (0676B–0676C)

In caligine. (6:1, verbatim) · Nubes latibulum ejus (Job 22:14, verbatim) · Caligo sub pedibus ejus: et posuit tenebras latibulum (Ps 17:10 + 17:12 conflated — Clem. 17:12 Et posuit tenebras latibulum suum; the dropped possessive and the join, §4) · Confitebor tibi, Domine, pater coeli et terrae, quoniam abscondisti haec… (Mt 11:25 — Clem. Confiteor tibi, Pater, Domine caeli et terrae, quia abscondisti. [var:] FIRED, §3k).

CAPUT VII (0676C–0677A)

Dedicavit. (7:5) · Sanctificavit. (7:7) · Sacrificium laudis honorificabit me, et illic iter, (Ps 49:23, verbatim clip; and note Migne prints no colon after Unde here, where he prints one everywhere else — preserved, Pattern 8) · Fecit ergo. (7:8) · Igitur in die, (7:10).

CAPUT VIII (0677B–0678A)

Expletis autem viginti, (8:1) · dederat Hiram, (8:2) · Bethoron. (8:5) · Et omnes. (8:6 — Clem. et omnes urbes firmissimas; clip) · sicut palma florens ascendit de virtute in virtutem, ut videat Deum deorum in Sion (Ps 83:8, with the opening woven from Ps 91:13 Justus ut palma florebit — Clem. Ps 83:8 ibunt de virtute in virtutem: videbitur Deus deorum in Sion. [var:] FIRED, §3l) · In Libano. (8:6) · Aperi, Libane, portas tuas (Zach 11:1, verbatim) · Ducenti. (8:10 — a clip of ducenti quinquaginta; see §2a) · Qui erudiebant. (8:10) · Holocausta. (8:12) · In Asiongaber. (8:17) · In Ophir. (8:18).

§2 · THIS RANGE'S CRUCES

2a · Centum. (0665C) and Ducenti. (0677C) are CLIPS and are expanded; Posteriora. is not

Both are one-word lemmata and both are clips whose own gloss proves the missing word. Centum. (1:17 equum centum quinquaginta) is glossed Qui sunt ter quinquageni — thrice fifty, which is 150, not 100. Ducenti. (8:10 ducenti quinquaginta) is glossed by taking the centenarius and then the quinquagenarius in turn — 2×100 + 50. Rendered "A hundred and fifty." and "Two hundred and fifty.", per Ruth §1.2: the fragment is rendered as the phrase it stands for. Rendered as bare "A hundred" and "Two hundred" each gloss would state a falsehood about its own lemma.

⚑ And the control's class, per the brief. Posteriora. (0671C) and Sacerdotes. (0676A) are complete one-word lemmata with nothing elided, and are rendered as they stand. Contrasting them with Centum. would manufacture an inconsistency that is not there.

2b · 0665B · doctorum testamentum praedicatione — will not construe. RENDERED AS PRINTED, NO MARKER.

Quia doctorum testamentum praedicatione ad cognoscendam fidei veritatem per praedicatores in usum veri Salomonis homines adducuntur. The clause is passive (adducuntur), so the nominative/accusative testamentum has nothing to govern it and nothing to be governed by; testamenti, with praedicatione, is the construction the words want, and the same paragraph's per utrumque testamentum two lines below shows the phrase is live in his mind. The conjecture is deliberately NOT adopted: nothing here is broken type (Pattern 12), nothing is lost from the digitization (Pattern 13), and this corpus has no marker meaning "Migne prints a word that will not construe" — the Ruth 0534A precedent (societate fieri) is exactly this and took none. Rendered "Because by preaching, the testament of the teachers, men are brought…", which leaves the stranded word visible where the plate leaves it.

2c · Four NON-WORDS, all rendered for their evident sense, NO [sic:] FIRED

The 2026-08-23 addendum measured non-word suspicions at ten for ten as our digitization's error rather than Migne's, and a [sic:] is a public claim that his plate is defective. All four are logged here as plate candidates and none carries a marker (the ship test: an unread candidate is safe iff it fires nothing).

2d · Three SPLIT runs — same class, same treatment. NO [sic:] FIRED.

⚠ All three sit at or near the head of a left-hand column, which is where our archive.org copy of PL 113 clips the margin. That is a reason to want the plate, not a reason to patch: nothing here was conjectured into the Latin, and if a plate read confirms the joined forms these are data/tei-patches/8965.json fixes and nothing else.

2e · 0674D–0675D · TWO [cj:], and why each is Pattern 18 and not a corpus-frequency patch

Both are real, well-formed Latin words, so Pattern 7 renders them and Pattern 12 has nothing to wrap; the addendum's rule (a real word is Migne's until a plate says otherwise) is obeyed — the printed word keeps its English and the conjecture stands beside it, additively.

1. 0675C · in area nihil fuit contrarium legi Domini → in the floor [cj: area; read arca, "the ark"]. The lemma is 5:10 Nihilque aliud erat in arca, and Migne's own next sentence opens Erat autem in arca secundum Paulum urna aurea. The rival reading (that area is meant, the threshing-floor of 3:1) loses because the whole paragraph inventories what was inside the ark — urn, rod, tables, bars — and a threshing-floor holds none of them.
2. 0675D · Erat et virgo Aaron, quae excisa denuo floruerat → the virgin [cj: virgo; read virga, "rod"]. virga Aaron quae fronduerat is Heb 9:4, which is the very list Migne is walking (secundum Paulum), and excisa … floruerat is said of a cut rod, not of a person. The rival (virgo as a Marian figure) loses because nothing in the sentence or its neighbours makes the turn to Mary, and the clause's three companions are all objects in the ark.

⛔ Neither was allowed to replace Migne's word in the running text; the reader meets "the virgin" and "the floor" on the page, with the conjecture beside them, which is what Pattern 18 requires and what a [sic:] would have destroyed.

2f · 0674D · Erat et tabulae testamenti — singular verb, plural subject. NO MARKER.

Rendered "There was also the tables of the testament", disagreement and all. English has an exponent for verb number here, so normalizing it would be a silent repair (Pattern 12's ⭐(2) runs the other way only where English has no exponent). Migne prints Erat three times in the list and only the third has a plural subject.

2g · 0666A · Significat quod gentes … transmittit — plural subject, singular verb; and TWO construals

Rendered "It signifies that the Gentiles, the report of the Gospel having been heard, sends diverse persons through faith unto baptism", disagreement carried, for the reason in 2f. The rival construal was weighed and lost: audita fama Evangelii could be read as a nominative subject ("the heard report of the Gospel sends…"), which would remove the disagreement altogether — but audita fama + genitive is the standard ablative-absolute shape, and that reading leaves gentes and diversas personas stacked as two unexplained accusatives. A construal that exists only because it tidies the grammar is the thing this edition is for refusing.

2h · 0667C · latitudo … significant — the same class, in the sicut … ita frame

Sicut latitudo parietis … provectus virtutum … vel sanctorum choros … significant. Rendered "As the breadth of the wall … signify …", plural as printed. The ita half has no verb at all in the Latin, and none was supplied: the English gaps it too ("so the evenness of the pavement, their concordant humility, wherein…").

2i · 0672B · In utrique parte templi … ut utraque populo — the two forms are exchanged

utrique stands where utraque belongs (with parte) and utraque where utrique belongs (with populo), in consecutive clauses. English carries no exponent for either case, so there is nothing to show and nothing to emend (Pattern 9's shape). Rendered "On either side of the temple … that to either people…". Logged so a later reader can tell a seen difficulty from a missed one; ⚑ the transposition is the sort of thing a plate settles in one look.

2j · 0670D and 0672A and 0678A · THREE sentences Migne leaves without a terminal stop

Pattern 8: a sentence that simply stops, stops in the English too. None was closed.

• 0670D …et hilaritatem dilectionis etiam in eos qui persequuntur — no verb, no stop.
• 0672A …in se ablutis confert omnium remissionem peccatorum — no stop.
• 0678A …ut unusquisque ordo officio suo functus digne ministret — no stop; and the same paragraph opens VERS. 14.--Secundum, with no space after the dashes and no italic lemma at all, alone among the seventy-odd verse marks in this range. Reproduced verbatim.
• 0678A …et eum sapientiae et consilio divina ascribentes tribuunt — no stop, and see 2l.

2k · 0671B · Unde audit: — a colon introducing nothing

The paragraph ends on Migne's colon and the quotation never arrives; the next line is the following lemma. The colon is preserved and nothing is supplied. This is the abridgment showing through (the recension cuts a Bede/Rabanus quotation and leaves its cue), but it is not the usque ad formula and was not treated as one.

2l · 0678A · et eum sapientiae et consilio divina ascribentes tribuunt — DECLINED, with the reason

tribuunt has neither object nor dative, and eum has no role that the clause can give it. The obvious repair is ei for eum — "ascribing to him divine gifts of wisdom and counsel" — and it is deliberately not adopted, and no [cj:] was fired, because the difficulty is larger than the one word: even with ei, tribuunt is left bare. Firing [cj:] on eum would tell a reader the rest of the clause construes, which it does not. Rendered as closely as the printed words allow: "and, ascribing divine things to wisdom and to counsel, they attribute him." ⚑ This sits in a resumption tail. The tail was parsed as strictly as ordinary prose — the abridgment removes words, it does not suspend grammar — so its oddity is reported, not excused.

2m · 0677A · non placuerunt. illi. — Migne's full stop inside the clause, PRESERVED

quia cum Salomon dedisset non placuerunt. illi. The dative illi is stranded on the far side of a stop. Rendered "…they did not please. him." Pattern 8's "never delete" governs (the 11542 @1041D precedent). ⚑ Also a plate candidate, as punctuation of this shape is exactly what a clipped scan manufactures.

2n · 0677A · ad quam usque Jesus persecutus est inimicos reges — rendered "Jesus"

The man is Joshua, and Migne prints Jesus, the Greek form. Rendered as printed, per the names rule; the crux is here because the reader meets a Jesus pursuing kings in Ephraim. I Regum's twin site (8952 @0556A) prints Josue, which is a hint about the recension and not an authority over this plate.

2o · 0677B · VERS.--6.-- — a malformed verse address, REPRODUCED VERBATIM

Alone among this range's verse marks, this one prints the dashes before the number. Convention 1 makes the address pass through untouched, so it passes through untouched. Not normalized, and not a Pattern 9 section-label case: Pattern 9 normalizes a translated structural head, and this address is not translated at all.

2p · 0664D · Interpretatur enim Gabaon sublimitas, vel collis patruelis — read clean

patruelis is a real word ("a cousin's", of a father's-brother's son) and construes; rendered "a cousin's hill". No conjecture offered and none needed.

2q · Sections read clean, and named so a blind reader can tell them from unread ones

CAPUT PRIMUM entire (0663D–0665C) beyond 2b · CAPUT II entire beyond 2g · CAPUT III's 0666B–0667B, 0668A–0669D and 0670C–0670D · CAPUT IV's 0671A–0672A, 0672C–0673D and 0674A–0674C · CAPUT VI entire · CAPUT VII entire beyond the Unde colon noted in §1 · CAPUT VIII's 0677C–0677D. Candidates rejected in reading, with reasons: nuto at 0665B (pro nuto suo for nutu — a 4th→2nd declension shift attested across medieval Latin; Pattern 9, no marker, rendered "at his own beck") · psalmatio at 0666A (a well-formed -tio noun, and the Hebrew z-m-r root behind Zocri is exactly "to sing", so it is not a slip at all — rendered "a singing of psalms") · scenophegiae at 0676D (attested for scenopegia) · Chyros/Zor at 0666A (transliterations, not Latin) · coelavit for caelavit (orthography, corpus-wide).

2r · Suspiciously TIDY forms — the normalization class (brief §4), and a checked negative

Reading with brief addendum C in view (read the whole line, not the target word), the one form in this range that reads too smoothly for its neighbourhood is 0668D Similiter ergo alae decem cubitos complent, sitting four words from the disagreement class of 2f/2h. It construes perfectly, which is exactly what a normalization does. ⚑ Raised, not marked, and recorded as what it is: a candidate, not a finding. No other form in these six chunks drew the suspicion, and that negative is worth having.

§3 · THE TWELVE [var:], AND WHY EACH EARNED A PUBLIC CLAIM

Every one rests on sources/vulgate/clementine-flat.txt, cited by book chapter:verse. None rests on recollection, and in none of them was Migne's wording replaced — the marker stands beside the faithful rendering.

a · 0665A · 1 Par 21:29. Migne eadem tempestate erant in excelsis Gabaon; Clem. ea tempestate erat in excelso Gabaon. Three changes at once, two of them number. It earns the marker because the gloss introduces it as Supra enim scriptum est — a claim about what stands written.

b · 0665C · Jo 19:30. Migne emisit spiritum; Clem. tradidit spiritum. The word Migne prints is Matthew's (Mt 27:50 emisit spiritum) under a note reading Joan. XIX. A different content verb in a proof-text adduced for the sixth day's completion.

c · 0666A · Jo 1:33. Migne videritis, plural; Clem. videris, singular. Under Pattern 17 the English must move with it, and does: "Upon whom you shall see", not the familiar "thou shalt see". ⚑ This is the 7a″ trap in its purest form — the remembered English is singular, and nothing but the printed letters decides it.

d · 0667B · 3 Rg 6:18. Migne drops intrinsecus, and the gloss it introduces is about walls clothed on their inner face with cedar and gold — so the dropped word is the one the exposition needs.

e · 0668C · Rom 13:10. Migne Plenitudo autem legis est charitas; Clem. Plenitudo ergo legis est dilectio. Fired for charitas, not for autem: this range holds charitas and dilectio apart everywhere (see §0), and the verse is adduced to prove that the angels' twofold love is charity.

f · 0668D · 2 Par 3:12. Migne Tangebant parietem, plural; Clem. tangebat, singular. The gloss is built on the plural — Duos aeque parietes alis suis tangunt, Extendunt … alas, all the way down the paragraph. Marking it is what lets a reader see that the exposition and the lemma agree with each other and not with the Vulgate.

g · 0669A · Heb 1:14. Migne Omnes enim sunt … qui haereditatem capiunt salutis; Clem. Nonne omnes sunt … qui haereditatem capient salutis? Migne turns a rhetorical question into a flat assertion and the future into a present. ⚑ No question mark was supplied — Pattern 8 forbids it, and the whole point of the marker is that his sentence is not a question.

h · 0673B · Heb 13:7. Migne quorum intuentes exitum conversationis intueamini et fidem; Clem. … imitamini fidem. "Behold ye also their faith" for "imitate their faith", in a gloss whose next clause is ut … et ipsi magni efficiantur — becoming, not merely beholding.

i · 0675C · Ps 113:21. Migne benedicit Dominus pusillis cum majoribus; Clem. Benedixit omnibus qui timent Dominum, pusillis cum majoribus. Present for perfect, and Dominus promoted to subject because the glossator has taken omnes qui timent up into his own sentence (omnes quos timentes se et diligentes). The tense is what earns the marker.

j · 0676A · Phlp 1:23. Migne Cupio dissolvi et esse cum Christo; Clem. desiderium habens dissolvi, et esse cum Christo. A different verb, and the form every patristic citation of this verse uses. Stated, not explained: Pattern 14 forbids speculating about why two witnesses differ.

k · 0676C · Mt 11:25. Migne Confitebor tibi, Domine, pater coeli et terrae, quoniam abscondisti; Clem. Confiteor tibi, Pater, Domine caeli et terrae, quia abscondisti. Future for present, and the two vocatives exchanged.

l · 0677C · Ps 83:8. Migne sicut palma florens ascendit de virtute in virtutem, ut videat Deum deorum in Sion; Clem. ibunt de virtute in virtutem: videbitur Deus deorum in Sion. Active for passive, singular for plural, purpose for statement — three changes that together make the Church the subject, which is exactly what the gloss needs her to be.

§4 · DIVERGENCES CHECKED AND DELIBERATELY LEFT UNMARKED

Recorded because a list of findings alone cannot be told from a report by a reader who never looked, and because Pattern 14 fires on divergence that changes the sense, not on every difference.

• Word order only: Lc 24:44 omnia impleri / Clem. impleri omnia (0665C) · Jo 15:5 Nihil sine me / Clem. sine me nihil (0674C) · 2 Par 5:10 Nihilque aliud erat / Clem. Nihilque erat aliud (0675C).
• A dropped or added particle: 2 Par 2:1 without autem (0665C) · 2 Par 2:17 tria millia et sexcenti (0666B) · Jas 1:17 with an added enim (0674C) · Ps 131:6 without Ecce (0674B) · Lc 24:45 without Tunc (0670A) · 3 Rg 7:38 without id est (0672A) · 3 Rg 7:26 without erat (0671C) · 2 Par 3:11 without una (0668C) · 2 Par 3:13 without erant (0669B).
• A dropped possessive: Ps 17:12 latibulum for latibulum suum (0676B), inside a conflation of 17:10 with 17:12 that is itself Migne's weave and not a variant.
• A clip that stops early: Jo 1:33 before in Spiritu Sancto · 2 Par 2:14 before aere, et ferro · Col 3:1, whose remainder Migne prints as his own prose · Ps 49:23 before quo ostendam.
• Case or orthography with nothing turning on it: Et decem cubitis under VERS. 1 for Clem. decem cubitorum (0671A) · coelavit/caelavit · coelo/caelo · Coelavit cherubim · the quae/quæ ligature class throughout.
• A citation pointing elsewhere while the WORDS agree: Vobis datum est nosse mysterium regni Dei at 0676B, cited Matth. XIII and printed exactly as Mc 4:11. No [var:] — Pattern 14 is about the text, not the reference, and Migne prints the identical words correctly cited to Marc. IV at 0670A in this same range. The internal agreement is itself the evidence.
• A citation whose base verse cannot be fixed: Oraculum in abditis (0667D–0668A). Not in the Clementine at all; the nearest is 3 Rg 6:19. Declined — a [var:] naming a verse I cannot establish would be a public claim resting on a guess.
• Migne's own attributions, e.g. Laminis aureis texit eam cited III Reg. VI where the words are 2 Par 3:8 (0668B). Passing a siglum or a reference through is not endorsing it.

§5 · THE TESTS, RUN, AND WHAT THEY RETURNED

Test 1 (a [sic:] must not swallow the word it marks). Vacuously clean: zero [sic:] fired in this range. Run anyway on the two [cj:], which are the only markers sitting inside a clause: stripped of the bracket, "There was also the virgin of Aaron" and "that in the floor there was nothing contrary to the law" both read complete, with subject, verb and object intact (the fourth question — does the clause still have the thing it acts on? — asked and answered). All six files were then stripped of every marker mechanically and swept for a stranded article or preposition sitting on punctuation: no hits beyond the whitespace the stripping itself leaves.

Test 2 + 2a (name the host of every printed negative). All 53 printed negation tokens in the six Latin chunks were extracted mechanically and each was matched to the single Latin word it attaches to, then found on that word in the English. One defect, found by this test and repaired: at 0675A Non enim in hoc altari sanguis hostiarum sed thymiama tantum incenditur had been rendered "For not on this altar is the blood of victims but only incense burned" — the negative floated off its host onto the prepositional phrase, and the English read well, which is what made it invisible. The sed thymiama tantum correlative proves the host is sanguis hostiarum, not in hoc altari and not the verb. Now: "For on this altar not the blood of victims but only incense is burned." ⚑ Exactly the class 2a predicts, with a grammatically plausible rival adjacent to the real host.

The trigger population 2a names (a negative on a verb with a quantifier or correlative nearby) was walked one by one: Nec enim semper … arctari (0674A) and non semper vasa … tenebantur (0674A) both take the negative on semper and are proved by their sed / ubi pivots · non tantum in regione Jordanis, sed and non solum in Judaea, sed (0674B) both take it on the phrase, proved by sed · non ad tectum templi, sed usque ad coenaculum (0666D), phrase, proved by sed · non in oculto agant, sed exemplum … praeferunt (0668A), phrase, proved by sed · nulli tamen absque peccato esse donat (0675B), on nulli · tanto minus interna gaudia vident (0675C), a comparative left uninverted ("by so much the less") · nihil habeant roboris … vel industriae (0670D), on the two genitives, not on habeant — see 2s below.

Test 2b (each anchor against the WORD it stands beside). All 55 column anchors in the range were checked against their Latin neighbour word, not merely for presence and order. Five sit mid-construction and were split to match: habendum → "to be [0670C] held" · dissolvi → "be [0671C] dissolved" · texit eam → "[0668B] he covered it" (inside the italic lemma, as Migne sets it) · est in id [0668C] est · construunt → "they [0666B] build up". Four sit with no space in the Latin ([0665A]VERS. 10, [0667B]Notandum, [0671D]VERS. 5, [0676B]VERS. 13, [0678A]VERS. 14) and are reproduced without a space.

Test 3 (punctuation by column band, never by total). Colons, semicolons, question marks and exclamation marks were compared band by band across all 55 anchor-partitioned bands in the six chunks. Two mismatches were found and repaired — a supplied semicolon at 0666A (Migne has a comma before de quo dicitur) and a supplied semicolon at 0677D (Migne has a comma after Holocaustum totum incensum dicitur). Final state: zero mismatched bands. The single question mark (0667C, nisi aliquid mysticum ostendat?) is in its own band in both twins, and the lower-case sed Migne prints after it is preserved.

Test 4 (does it ASSERT what he asserts?). Every paragraph was read back against the Latin with the counts already known clean. It returned 2j (three sentences that stop without a verb or a stop, none of them closed), 2k (a colon introducing nothing), 2l (the resumption tail that will not construe — caught precisely by refusing to let "the abridgment made it odd" excuse it), 2b (testamentum), and the two [cj:] sites of 2e. None of these is visible to any count-based check: every one of them is present, in order, at parity.

The usque ad count, since the brief named it. The Latin of these six chunks carries 28 tokens containing usque. 18 are the abridgment formula, each in its own italic span, and all 18 are as far as in the English with no variation. Five are ordinary and were decided one at a time from what stands on either side: usquequo aedificaretur templum (0665A, "until") · sed usque ad coenaculum inferius pervenerunt (0666D, "but as far as") · deinde usque ad supremum templi tectum (0666D, "then as far as") · ad quam usque Jesus persecutus est (0677B, "as far as which") · a fluvio Cophene usque ad regionem Indiae (0678A, "as far as the region of India"). Five more are look-alikes inside another word and are not usque at all: utriusque populi (0669A), cujusque justi (0670B), utriusque Testamenti (0671A), cujusque electi (0677C), cujusque temporis (0678A). ⚑ No lemma's own words were swallowed by the formula reading — the two that could have been, 0666D's twin usque ad pair, sit in Migne's own prose about the thirty cubits and not in any citation.

2s · 0670D · the nihil with three genitives — how the host was fixed

cum in se firmae fidei ad contemnendas saeculi pompas, ad desideranda bona invisibilia, nihil habeant roboris ad corrigendos, vel industriae ad intelligendos eorum quibus praesunt errores. The single nihil governs roboris and industriae, and the two ad-phrase pairs divide between the genitives; the negative is not on habeant. Rendered "…they have nothing of strength for correcting, nor of diligence for understanding, the errors of those over whom they are set", with the English gapping the second genitive exactly as the Latin does.

§6 · UNEXPRESSED SUBJECTS — every one supplied, named, and what named it

Per the brief's dominant-defect clause. Where two readings both construe it is said so, not chosen silently.

§7 · PLATE CANDIDATES — what a plate read of PL 113 would settle, in one list

No plate access from this stint, and nothing was patched or conjectured into the Latin.

1. 0667B latitudis · 2. 0670A obediende · 3. 0675C omma · 4. 0677A subbatismum — four non-words (§2c).
2. 0670D holo causti · 0672A de cem · 0677B coena culum — three split runs (§2d), all at or near a left-column head, where our archive.org copy of PL 113 clips the margin. ⚠ Raised, not patched: a conjecture entered as a patch would wear Migne's voice.
3. 0677A non placuerunt. illi. — a stop inside a clause (§2m).
4. 0672B utrique/utraque exchanged (§2i).
5. 0668D complent — the normalization candidate of §2r, raised as a checked negative would be worth having too.

If a plate read confirms any of 1–2, each is a data/tei-patches/8965.json fix and not a [sic:]; if it shows Migne printing the broken forms himself, the [sic:] markers can be added then, on evidence.

Translated 2026-08-25, one translator, 5 chunks, 5,592 Latin words (frontmatter sum) → 8,147 English, 1.46×. Chunks 0000–0005 are a peer's and go to cruces-0000.md; the merge is the orchestrator's.

verify-english.mjs clean on the whole work. In this batch: column anchors 57/57 verbatim and in order, [n:] notes 65/65 at parity, italic spans 152/152, paragraph counts 1:1 in every chunk. Apparatus fired here: 3 [var:] · 0 [sic:] · 0 [cj:] · 0 [d:] · 0 [ed:] · 0 [nt:]. Zero guillemets in either language, zero em-dashes in the English, question marks 2 Latin / 2 English (both at 0684A, in the Augustine sentence).

Ruth (8968) cruces.md §1 governs throughout — VERS. n.-- passes verbatim, the lemma is Englished, a fragmentary lemma is rendered as the phrase it stands for, the attribution sigla are ordinary [n:] notes with their contents unanglicized. Tobit (8970) §1(b) governs the abridgment formula: usque ad → as far as, no verb ever supplied to complete a resumption tail. Nothing in either was reopened.

§A · WORK-WIDE VOCABULARY — the choices this half made, for the merge

Stated so the peer's half can be reconciled against it. Nothing here is a new convention; it is a record of which English word was fixed to which recurring Latin one.

Proper names are kept form for form as Migne prints them, never normalized: Philitshiim (0680C), Neopolis (0682C), Eroge (0688A), Amulcus (0688B), Ononimus (0691C), Zareas (0682D), Jesus for Josue (0682B–C), Benennom (0689A). The Ruth Lia ruling.

How a bare lemma was decided — the rule applied, since the brief warns against manufacturing an inconsistency

Two classes, and they are different (brief, "name the control's class"):

• A lemma that yields a readable English phrase naming the thing glossed is rendered as it stands — Sexcenta. "Six hundred." · Quadraginta. "Forty." · Trecenta quoque. "Three hundred also." · Fecit igitur. "He made therefore." · Regem vero. "But the king." · Et cedrorum. "And of cedars." An English cardinal numeral stands as a substantive, so nothing is elided in the English.
• A lemma that leaves a dangling English word is expanded to the phrase it opens (Ruth §1.2), four times in this batch: Quod postquam, → "Which after she had seen," (9:3) · Tantamque. → "And so great an abundance." (9:27 Tantamque copiam praebuit argenti) · Obturaverunt cunctos. → "They stopped up all the fountains." (32:4 obturaverunt cunctos fontes) · Carduus qui, → "The thistle which is in Libanus," (25:18) · Ad quam Deus. → "To which God commanded me to go in haste." (35:21 ad quam me Deus festinato ire praecepit). Each keeps Migne's own comma or etc.

§B · THE LEMMA COLLATION — every italic span in 0678B–0692C, checked in sources/vulgate/clementine-flat.txt

A lemma named here with no comment AGREES with the Clementine at its own verse, verbatim or as a clean clip of it — that is the point of naming it. Everything that diverges carries its divergence. Migne's book is 2Par; chapter and verse numbering is identical, so no conversion arises.

CAPUT IX (0678B–0681A)

Regina quoque. (9:1) · Quod postquam, (9:3 — Clem. Quae postquam; §C1) · Beati viri tui, (9:7) · De lignis. (9:11, addressed VERS. 12 — §C2) · Sexcenta. (9:13) · Omnesque reges Arabiae. (9:14) · Fecit igitur. (9:15) · Hastas aureas. (9:15) · Trecenta quoque. (9:16) · Sex quoque brachiola. (Clem. 9:18 Sex quoque gradus … et brachiola duo altrinsecus. [var:] FIRED, §C-fire-1) · Scabellum aureum. (9:18) · Stantes juxta brachiola: sed et alios duodecim leunculos, (9:18–19, clean across the verse join; Clem. comma for Migne's colon) · Brachiola, (9:18) · Sex gradus. (9:18–19) · Non fuit tale solium, (9:19) · Omnia quoque. (9:20) · Argentum. (9:20) · Siquidem naves regis ibant in Tharsis, (9:21) · Quadraginta. (9:25) · In Jerusalem. (9:25) · Philisthinorum. (9:26) · Tantamque. (9:27) · Et cedrorum. (9:27) · Reliqua vero. (9:29 — Clem. Reliqua autem; a particle, no marker).

Scripture the gloss quotes inside CAPUT IX: Surget regina austri in judicium cum generatione ista (Mt 12:42 — Clem. Regina austri surget in judicio: word order, and an accusative for an ablative. Case, not sense — no marker, the 8952 @0557D precedent) · Estote perfecti sicut et Pater vester coelestis perfectus est (Mt 5:48 — Clem. Estote ergo vos perfecti; two words clipped) · Si vis ad vitam ingredi serva mandata (Mt 19:17 — Clem. Si autem vis) · Si vis perfectus esse, vade et vende omnia quae habes, (Mt 19:21 — Clem. vade, vende quae habes, et da pauperibus: Migne's omnia is not in Mt 19:21; it is Lc 18:22's omnia quaecumque habes, so the citation is harmonized. No marker: "sell all that thou hast" asserts nothing Mt 19:21 does not) · omnis qui reliquerit domum, and centuplum accipiet et vitam aeteram possidebit. (Mt 19:29, verbatim apart from the non-word aeteram, §E1) · Reges Tharsis et insulae munus offerent, (Ps 71:10 — Clem. munera, plural; §C3) · Et adorabunt eum omnes reges, omnes gentes servient ei (Ps 71:11 — Clem. omnes reges terrae; one word dropped) · Lectulum Salomonis ambiunt sexaginta for es (Ct 3:7 — Clem. En lectulum Salomonis sexaginta fortes ambiunt ex fortissimis Israel: reordered and clipped; Migne's own [n:] says (Prov. III) where the verse is Canticles — his citation, left standing, §C4) · Non nobis, Domine, non nobis, sed nomini tuo da gloriam (Ps 113:9, verbatim) · Et duae manus hinc et inde tenentes sedile (3Rg 10:19 — Clem. hinc atque inde) · Sine me nihil potestis facere (Jo 15:5, verbatim; the [n:] reads (Joan. V) for XV — Migne's, left standing) · Fortitudo et decor indumentum ejus, (Pr 31:25, verbatim) · Multae filiae congregaverunt divitias: tu supergressa es universas (Pr 31:29, verbatim; Clem. semicolon for Migne's colon) · quae fit in eis fons aquae salientis in vitam aeternam (Jo 4:14 — Clem. sed aqua quam ego dabo ei, fiet in eo: the glossator has bent the verse into a relative clause on animae sanctorum, so the plural and the present are his grammar, not a divergent text; no marker, and the English follows his) · Currus Dei decem millibus, (Ps 67:18, clipped; the [n:] reads (Psal. LXXVI) — Migne's) · Ascendens super equos tuos, et equitatus tuus sanitas (Hab 3:8 — Clem. Qui ascendes super equos tuos, et quadrigae tuae salvatio. [var:] FIRED*, §C-fire-2) · Coram illo procident Aethiopes, (Ps 71:9, verbatim) · Similes illis fiunt qui faciunt ea, (Ps 113:16 — Clem. fiant, jussive; §C5) · Potens est Deus de lapidibus istis suscitare filios Abrahae (Mt 3:9, verbatim from potens*).

CAPUT X–XIX (0681B–0684A)

Reliquit consilium senum, (10:8) · Responditque rex, (10:13) · Amavit, (11:21) · Anno autem quinto regni Roboam ascendit. (12:2) · Sesac rex. (12:9) · Clypeos. (12:9) · Semeiae. (12:15) · Addo. (12:15) · Semeron. (13:4) · Bethel. (13:19) · Ephron. (13:19) · In diebus. (13:20) · Egressus est, (14:9, addressed VERS. 2 — §C2) · Azarias autem, (15:1, addressed VERS. 15 — §C2) · Aromatibus, (16:14) · Combusserunt super eum, (16:14) · Ambulavit in viis David, (17:3) · Dedit omnis Juda munera Josaphat. (17:5, verbatim) · Itaque factus est pavor, (17:10) · Descenditque, (18:2) · Omnem causam quae venerit ad vos, (19:10).

Scripture quoted: Vae tibi, terra cujus rex puer est, (Ecl 10:16 — Clem. sets a comma after terra; punctuation only) · Ubi plurima verba, ibi frequenter egestas (Pr 14:23 — Clem. ubi autem verba sunt plurima; reordered, sense identical) · Juxta combustiones majorum facient tibi. (Jr 34:5 — Clem. et secundum combustiones patrum tuorum, regum priorum qui fuerunt ante te, sic comburent te. Loose throughout rather than divergent at a point, and the sense is identical; no marker — the 8952 @0555B precedent. Migne prints no [n:] for it, only Unde alibi) · Vos estis genus electum, regale sacerdotium, (1Ptr 2:9 — Clem. Vos autem genus electum: Migne supplies the copula and drops the particle) · Haereticum hominem post primam et secundam correctionem devita (Tit 3:10 — Clem. post unam et secundam correptionem; §C6).

CAPUT XX–XXIV (0684B–0687A)

De Ammonitis. (20:1) · De his locis, (20:2) · Engaddi. (20:2) · Ascensuri enim per clivum nomine Sis, (20:16 — Clem. ascensuri enim sunt per clivum; the copula clipped) · Venit ergo Josaphat, (20:25) · Vallis benedictionis. (20:26) · Omnes hi, filii Josaphat, (21:2; Clem. sets no comma) · Ecce Dominus percutiet te, (21:14) · Et octo annis regnavit, (21:20) · Azarias filius Joram, (Clem. 22:6 Ochozias filius Joram. [var:] FIRED, §C-fire-3) · Ochozias, (the gloss's own second lemma, 22:6/22:2) · Adversum Jehu, (22:7) · Si quidem Athalia, (22:10 — Clem. siquidem, one word; word division only) · Super gradum. (23:13) · Pepigit autem Joiada foedus. (23:16; Clem. spells Jojada, i/j orthography) · Post quae placuit Joas, (24:4) · Et filii ejus, (24:7) · At illi conducebant, (24:12) · Reliquam partem pecuniae, (24:14) · Et facta est ira. (24:18) · Zachariae filii Joiadae, (addressed VERS. 25, where Clem. 24:25 has only filii Jojadae sacerdotis; the name Zacharias comes from 24:20 — a compression across five verses, not a divergent reading, no marker) · Insidiati vero sunt, (24:26).

Scripture quoted: Hic est Filius meus dilectus, in quo mihi complacui (Mt 17:5 — Clem. in quo mihi bene complacui; one word) · Unicuique nostrum data est gratia secundum mensuram donationis Christi (Eph 4:7 — Clem. Unicuique autem nostrum) · Verumtamen non fiebant ex eadem pecunia hydriae templi et fuscinulae (4Rg 12:13 — Clem. hydriae templi Domini, et fuscinulae, et thuribula, et tubae…; clipped, and the negation stands in both).

CAPUT XXV–XXXIV (0687B–0690A)

In vallem Salinarum, (25:11) · Scio, inquit, quod cogitaverit Dominus, (25:16 — Clem. Deus; a divine name, §C7) · Carduus qui, (25:18) · Bethsames. (25:21) · Et apud Obededom, (25:24) · Qui postquam recessit. (25:27) · Ozias cum regnare coepisset. (26:3) · Super portam vallis. (26:9) · Sed cum roboratus esset, (26:16) · Ingressusque templum Domini, (26:16) · In agro regalium sepulcrorum. (26:23; Clem. sepulchrorum) · Viginti quinque annorum erat Joatham, (27:1) · Ipse aedificavit portam, (27:3) · Ophel. (27:3) · In valle Benennom. (28:3) · In ipso anno, (Clem. 29:3 Ipse, anno et mense primo regni sui; §C8) · Peccaverunt patres nostri, (29:6) · Misit quoque, (30:1) · Fregerunt simulacra, (31:1) · Obturaverunt cunctos. (32:4) · Aedificavit. (32:5) · Et egit poenitentiam valde coram Deo patrum suorum. (33:12, verbatim) · Ossa praeterea. (34:5) · Ad Oldam prophetidem, (34:22).

Scripture quoted: Et turris gregis nebulosa filiae Sion (Mch 4:8 — Clem. Et tu, turris gregis: the apostrophe becomes a bare nominative; one word, and nothing in the exposition turns on it) · turris Ophel. (labelled In Hebraeo habetur by Migne — a report of another text, not collated as a Vulgate reading; the 8952 @0541D precedent) · Posuit tenebras latibulum suum (Ps 17:12, clipped from Et posuit) · Magnus Dominus et laudabilis nimis (Ps 47:2, verbatim) · Et consummatus, factus est omnibus obtemperantibus sibi causa salutis aeternae, (Hbr 5:9, verbatim) · Ego sum ostium. Per me si quis introierit, salvabitur (Jo 10:9, verbatim) · Excelsus super omnes gentes Dominus (Ps 112:4, verbatim) · Etenim pascha nostrum immolatus est Christus (1Cor 5:7, verbatim) · Nolite sanctum dare canibus, neque mittatis margaritas ante porcos. (Mt 7:6 — Clem. Nolite dare sanctum canibus … margaritas vestras; word order and one possessive).

CAPUT XXXV–XXXVI (0690B–0692C)

Fecit autem Josias in Jerusalem, (35:1) · Ponite arcam in sanctuario templi, quod aedificavit Salomon filius David rex Israel. (35:3, verbatim) · Non fuit Phase simile huic, (35:18) · Ad quam Deus. (35:21 — Clem. ad quam me Deus festinato ire praecepit; the pronoun clipped) · In campo Mageddo. (35:22) · Constituitque. (36:4) · Et vertit. (36:4) · Regem vero. (36:10).

Scripture quoted: non in fermento veteri, neque in fermento malitiae et nequitiae: sed in azymis sinceritatis et veritatis (1Cor 5:8, verbatim, both negations intact) · Dominus fortis et potens: Dominus potens in praelio (Ps 23:8, verbatim; Migne's colon for Clem.'s comma) · Christus salvum facit populum suum a peccatis eorum (Mt 1:21 — Clem. ipse enim salvum faciet; §C9) · Zelus domus tuae comedit me (Ps 68:10, clipped from Quoniam zelus) · Cujus ventilabrum in manu sua, et permundabit aream suam (Mt 3:12, verbatim; comma for Clem.'s colon) · Dominus jussit me ascendere ad terram (not a Psalm: it is Rabsaces at 4Rg 18:25 / Is 36:10, Dominus dixit mihi: Ascende ad terram hanc. A paraphrase inside the glossator's reported speech, and his own frame says so — Unde et Rabsaces dixit. No marker; the [n: (Psal. LXXXIII)] is his, left standing) · Dormivit Joakim cum patribus suis, regnavitque Joachin filius ejus pro eo (4Rg 24:5–6 — Clem. Et dormivit … et regnavit).

§C · THE THREE [var:] FIRED, AND EVERY DIVERGENCE DELIBERATELY LEFT UNMARKED

FIRED

C-fire-1 · 0679A/B · Sex quoque brachiola. → [var: Vulg. 2Par 9:18 Sex quoque gradus … et brachiola duo altrinsecus, "six steps … and two little arms on either side"] Migne's lemma joins Sex quoque — which in 9:18 belongs to the steps — directly to brachiola, of which the verse says there were two. The English therefore prints "Six little arms also", which reads exactly like a translator's slip; that is the 8952 @0543A test and it is why the marker fires even though the gloss expounds only the six (In senario bonorum operum perfectio), which is a real six in the verse. ⚑ The rival, and why it lost: this could be read as an ordinary Glossa clip joining two non-adjacent words, the class declined at 8952 @0562C (Cumque Amalecitae). It lost because that clip only pulls a subject forward and asserts nothing new, whereas this one transfers a number from one noun to another and makes a false statement about Solomon's throne. Migne prints Sex gradus. as its own lemma eleven lines later, so he is not treating the two as one clip.

C-fire-2 · 0680C · Ascendens super equos tuos, et equitatus tuus sanitas → [var: Vulg. Hab 3:8 et quadrigae tuae salvatio, "and thy chariots salvation"] A different noun and a different predicate: cavalry for chariots, health for salvation. The 8952 @0543A shape (two texts asserting different things), and the plainest divergence in this batch. Migne's [n: (Rabac. III)] — his spelling of Habacuc — rides verbatim.

C-fire-3 · 0685B/C · Azarias filius Joram, → [var: Vulg. 2Par 22:6 Ochozias filius Joram, "Ochozias the son of Joram"] The most material of the three: the whole gloss is about this name. Migne addresses VERS. 6 with Azarias where the Clementine reads Ochozias, and then expounds both — Qui superius Ochozias, id est, apprehendens Dominum: Azarias vero adjutorium Domini — arguing that the name is changed for the better. Without the marker a reader has no way to know which of the two names is the received text at that verse, which is the one fact the exposition depends on. The Hebrew of 2Chr 22:6 does read Azariah; that is a research question and, per Pattern 14, the marker states the two readings and says nothing about why.

DECLINED — every one is a real difference from the Clementine, and none changes the sense

§C1 · 0678C · Quod postquam — the Clementine's Quae is feminine and makes the queen the subject of vidit; Migne's Quod is neuter. The lemma is a fragment (etc. follows), rendered per Ruth §1.2 as the phrase it opens, "Which after she had seen," — and "she" is the queen of Saba of 9:1–3, which is also who the gloss's provocata est and venit are about. No marker: nothing in the exposition turns on the gender of the relative, and English has no exponent for it either way.

§C3 · 0679A · munus for munera — worth naming even though it takes no marker, because the gloss is built on the singular: the sentence immediately before it is offert rectae fidei munus. A number difference in a noun is the 8952 @0545B class (one word) and Pattern 14 fires on a change of sense, not of number; but a reader who wants the reason Migne's citation reads as it does now has it.

§C4 · 0679B · the Canticles verse cited (Prov. III) — Migne's own citation, wrong, and not corrected in the running text: citation repair belongs to data/citation-corrections.json at index time. Two more in this batch: Sine me nihil potestis facere cited (Joan. V) for XV at 0679D, and Currus Dei cited (Psal. LXXVI) for LXVII at 0680C, and Dominus jussit me ascendere cited (Psal. LXXXIII) at 0690D for a verse that is in Kings and Isaias.

§C5 · 0681A · fiunt for fiant — indicative for jussive, one letter, and the indicative is what the gloss needs (corda gentilium insensibilia … quos lapides pro Deo coluerunt). Declined on the sources/vulgate/README.md caution that a call turning on a single letter should say so rather than trust this copy to that precision, and on the 8952 @0558B precedent (mood recast, received sense unchanged). Rendered as printed: "They that make them become like unto them."

§C6 · 0684A · primam/correctionem for unam/correptionem — declined, and for the reason that governed the charitas/dilectio pair at 8952 §4a: this is the ordinary patristic citation form of Tit 3:10, and a marker that fires here would have to fire on every Latin father who quotes the verse. correptio and correctio are the standard Latin pair; primam for unam is idiom, not doctrine.

§C7 · 0687B · Dominus for Deus — a divine name substituted in a quoted verse. Declined on the 8952 @0542C precedent (Miserere mei, Deus for Domine, "vocative"). The gloss immediately expounds Qui neminem vult perire, which is indifferent to which name stands.

§C8 · 0689A · In ipso anno for Ipse, anno et mense primo regni sui — the pronoun's case is changed (Ipse → ipso) so that it modifies the year instead of being the subject, and primo is gone. Declined: ipso is a real, well-formed word, so the printed reading is Migne's until a plate says otherwise (2026-08-23 addendum A), and the gloss's point — in principio regni sui cultum Dei magnificavit — is the beginning of the reign either way. Rendered "In that same year,".

§C9 · 0690B · Christus salvum facit for ipse enim salvum faciet — the pronoun is replaced by its antecedent (an adaptation to the glossator's sentence, which has just given the etymology salus Domini) and the future becomes a present. Declined as tense recast with the sense unchanged, and the present is rendered as printed — the 7914 @1071D rule runs this way, not the other: it is the Vulgate's tense that must not be restored over Migne's.

§D · usque ad — DECIDED PER OCCURRENCE, as the brief requires

The five chunks print 22 tokens containing the letters usque, of which 4 are inside another word and are not the preposition at all (utriusque 0679A · minusque 0682A · Ingressusque 0688A · cujusque 0690A) — they are named because a mechanical count of the formula would swallow them.

10 are the abridgment formula, every one set in its own italic span, and all 10 are as far as in the English with no verb supplied to complete the resumption: 0684A (GREG. on the throne of the Lord) · 0684B (RAB. on the ascent of Sis) · 0686B (Joas and the priests) · 0687A (Zacharias) · 0687D (the gate of Ephraim) · 0689A (the valley of Benennom) · 0689B (HIERON. on 29:6) · 0689D (Christ building the Church) · 0692A (HIERON. on the name Joachin) · 0692C (RAB. ex Beda).

The other 8 are not the formula and were each decided from what stands on either side:

⚑ The dangerous one is 0680D, where the words are 2Par 9:26's own. It was caught by looking the verse up, exactly as 8952 @0543C and @0548C were.

§E · THIS BATCH'S CRUCES

E1 · Five NON-WORDS — all rendered for their evident sense, NO [sic:] FIRED, all plate candidates

The 2026-08-23 addendum's finding is that a suspicion aimed at a non-word has been our digitization's corruption every time it was tested, and a [sic:] is a public claim that Migne's plate is defective. None was fired. Each is logged here so the plate can settle it; if the plate reads the reconstructed form, each is a data/tei-patches/8965.json fix and nothing else.

E1a · 0684A · reperetur — a non-word with TWO candidate reconstructions, and the choice is recorded

et tanto durius reperetur ad poenam, quanto securius viveret in culpa (the tail of the Gregory abridgment). reperetur is a form of no Latin verb; measured against the whole corpus of src/latin/, it occurs exactly once — here — while reperiretur's family runs to hundreds of well-formed tokens and reservare to twenty-six. Two reconstructions were weighed:

1. reperiretur — "would be found". Two letters dropped (reperiretur), and the imperfect subjunctive matches viveret in the answering clause.
2. reservaretur — "would be reserved". This is Gregory's own idiom (ad poenam reservare), and idiomatically the stronger reading.

Reading 1 was taken and rendered "would be found unto punishment." Why reading 2 lost: a non-word is a transcription accident, and the criterion for a transcription accident is orthographic distance, not idiom — 1 is two dropped letters, 2 is five substituted ones. A plate read settles it in one look, and if the plate has reservaretur the English needs one word changed. ⚑ Plate candidate — 0684A.

E2 · 0678D · the arithmetic does not work, and both words are REAL. Rendered as printed.

Senarius per senarium multiplicatus, sexaginta facit. Centenarius quoque sexies ductus centum et sex reddit. Six by six is thirty-six, not sixty; a hundred taken six times is six hundred, not a hundred and six. The lemma above is Sexcenta. (9:13, sexcenta sexaginta sex talenta), so the passage is deriving 600 + 60 + 6, and the two conjectures that yield it are visible: per denarium in the first sentence (6 × 10 = 60) — and Migne's very next sentence is Denarius mandatorum significat perfectionem, a word with nothing else in the paragraph to attach to — and sexcenta reddit in the second. Neither is adopted and no marker fires: every word printed is a real, well-formed Latin word, which the 2026-08-23 addendum refuted five times out of five as a ground for patching. Rendered exactly: "The senary multiplied by the senary makes sixty. The centenary also, drawn out six times, yields a hundred and six." The conjectures are here by decision, not by oversight.

E3 · 0680B–C · one sentence printed twice, with a different first word. NO [d:], and why.

Senarius vero praesens tempus significat, in quo apostolica doctrina Novi Testamenti populum nutrit … [two sentences later] … Duodenarius vero praesens tempus significat, in quo apostolica auctoritas populum gentium nutrit Novi Testamenti doctrina. The verse is 9:25, whose second number is twelve (curruum equitumque duodecim millia), so Duodenarius is the word the argument needs; and the first copy interrupts the quadragenary exposition, which resumes immediately after it with Bene ergo per quadragenarium numerantur equi Salomonis.

Pattern 11a was weighed and NOT applied. Its case (11613 @1230D) is an eye-skip: the same clause re-set, adjacent, one copy ending in a word copied from the line before. Here the two copies are two sentences apart, and their tails differ in wording as well as in the first word, so they are not one run set twice, and a [d:] would be a claim about compositor behaviour I cannot support from the file. ⛔ And the thing 11a exists to prevent is not at risk: nothing in either copy reverses the argument, and senarius as a figure of perfection is used in this very chunk (In senario bonorum operum perfectio). Both are rendered as printed. ⚑ Plate candidate — 0680B, on the narrow question whether Migne prints Senarius there.

E4 · 0687D · priusquam where the sense wants postquam. Rendered as printed, NO [cj:].

Sed priusquam per superbiam illicita tentavit, lepra percussus nominis et honoris offuscavit dignitatem. Ozias was struck with leprosy after the presumption, not before, and Migne's own gloss on 26:16 four lines later narrates it in that order (Sed postquam Zacharias obiit … unde lepra in fronte percussus est). Declined: priusquam is a real, well-formed word occupying its slot, the sentence construes as printed if the clause is read as bounding the preceding quandiu recte egit, and Pattern 18's bar is a negation or a correlative, which this is not. Rendered "But before he attempted unlawful things through pride, being struck with leprosy he darkened the dignity of his name and of his honour."

E5 · 0680C–D · usque ab conversionem — a preposition that cannot govern its case. NO [cj:].

ab inchoatione sui in Jordane baptismatis usque ab conversionem omnium gentium ad fidem Christi redundat gratia potestatis. The frame is ab inchoatione … usque ad, "from the beginning … even to"; ab takes no accusative, so ab conversionem is not Latin, and the English cannot show the difference between the two prepositions except by rendering the one printed. Rendered "even from the conversion of all nations to the faith of Christ", which contradicts the ab inchoatione it answers — that contradiction is Migne's and is what the edition exists to show. [cj:] weighed and declined, on the nearest control there is: 11535 @1360C, ab for ad subdiaconatum, direction reversed, declined and closed by the 18a census. The conjecture is ad; it lives here.

E6 · 0681A · quos lapides pro Deo coluerunt — the relative's antecedent, named

cum corda gentilium insensibilia, ad veri Dei intellectum quos lapides pro Deo coluerunt, ad meditationem legis … convertit. quos is masculine and cannot look back to corda; its antecedent is the tapidum (= lapidum) of the paragraph's first line (§E1), picked up by the appositional lapides. Rendered "the senseless hearts of the gentiles, which stones they worshipped in place of God" — the clause kept where the sense binds it, with the English clauses reordered for readability, nothing added and nothing dropped.

E7 · 0692A–B · et se subditos convertunt ad impietatem — two construals, and which was taken

1. "they turn themselves as subjects to impiety" — se … subditos as reflexive object plus predicate;
2. "they turn those subject to them to impiety" — reading se subditos as sibi subditos.

Reading 1 was taken. Why 2 lost: the whole run of clauses on either side is about what these rulers do to themselves (degenerantes relabuntur, nec fidem … custodiunt, nec … foedera servant, semel indurati), and sibi subditos would require a dative Migne does not print. The rival is not excluded and is recorded here because a reader will see it.

E8 · 0689D · Contra paganos et haereticos pugnantes — an ambiguity kept

pugnantes is accusative plural and agrees with paganos et haereticos, which is how it was rendered: "Against pagans and heretics who fight, diligent caution must be used." The rival reads it as the unexpressed plural agent of the gerundive ("for those fighting against pagans and heretics") and is not excluded; the counsel is the same on either reading, and nothing was supplied to settle it.

E9 · 0690D · a paragraph that ends with no point. NOT supplied.

…quia prophetae praedixerant eum esse ascensurum, quod eum minime latuit — the paragraph simply stops. Pattern 8 forbids supplying a terminal stop, so the English stops too: "…which was by no means hidden from him". The Latin column shows the same.

E10 · 0682C · Ephron glossed with the twentieth milestone, and Idumoeorum — spellings left alone

• 0687A Idumoeorum against 0684B Idumaeos and 0684C Idumaei, three columns apart. An oe/ae difference in a proper name has no English exponent (all three are "Idumaeans"), so nothing is emended and nothing is marked; recorded so a later reader knows it was seen. Pattern 9.
• 0682C Neopolim for Neapolim, 0688A Eroge for En-rogel, 0691C Ononimus, 0682D Zareas for Zara, 0682B–C Jesus for Josue: Migne's forms, all kept, none normalized.

E11 · 0681C · Ecclesiae qua … provocata est — a relative in the wrong case, no English exponent

Conveniunt haec Ecclesiae qua auditis miraculis Christi provocata est. provocata est is feminine nominative, so the relative must be quae; qua is ablative. English "who" carries no case, so there is nothing to render either way and nothing to mark (Pattern 9, the 8952 @2p/@2q class). Rendered "These things agree with the Church, who, when the miracles of Christ had been heard, was stirred up to seek him."

E12 · 0692C · scilicet et civitatis Dei de qua ejecti fuerant, recipiunt — a genitive with nothing to govern it

The tail of the last abridgment in the book. recipiunt takes an accusative and civitatis Dei is genitive; the noun it depended on was removed with the abridged matter. No verb and no noun supplied (Tobit §1(b)), and the English keeps the genitive: "namely, and of the city of God from which they had been cast out, they receive." The brief's warning was heeded — the tail was parsed as strictly as ordinary prose, and it is the abridgment that took the word, not the grammar that changed.

E13 · Migne's verse addresses that do not match his own words — reproduced, never reordered

Four in this batch, on the Jude precedent: 0678C VERS. 12 on 9:11's De lignis · 0682C VERS. 2 (CAPUT XIV) on 14:9's Egressus est · 0682D VERS. 15 (CAPUT XV) on 15:1's Azarias autem · 0686C VERS. 25 on a lemma whose name comes from 24:20. Also 0681A prints VERS. 27 twice, for Tantamque. and Et cedrorum., both of which really are 9:27. Nothing renumbered.

E14 · 0681C · the numbers in the gloss on 11:21 do not match the verse. Gloss prose, so no [var:].

Migne's gloss gives Roboam eighteen wives, thirty concubines, eighteen sons, sixty daughters; 2Par 11:21 has eighteen wives, sixty concubines, twenty-eight sons, sixty daughters. Two of the four numbers differ. No marker, and the reason is the convention's own boundary (18a): [var:] is for scripture quoted as scripture, and this is the glossator's own prose reporting the verse, not an italic lemma. It is logged here because it is the largest factual divergence in the batch and a reader comparing the columns will meet it. The same class, smaller, at 0682D: the gloss gives Zareas ninety thousand foot and ten thousand horse where 14:9 has decies centena millia — Migne is following Josephus there, as his neighbouring sentences show.

E15 · 0691B–C · qui vixit annos decem et triginta — Joakim's age, real words, left as printed

Ten and thirty is forty; 4Rg 23:36 makes Joakim twenty-five at his accession with eleven years' reign, so thirty-six. sex et triginta is the obvious suspect and is NOT adopted — every word printed is a real word (addendum A), and the passage is reporting Josephus, not the Vulgate. Rendered "who lived ten and thirty years".

§F · THE TESTS, RUN, AND WHAT THEY RETURNED

Test 1 — a [sic:] must not swallow the word it marks. No [sic:] was fired in this batch, so the test was run on the three [var:] instead, by deleting each marker and reading what remained: all three sit beside a complete rendering (the lemma is Englished in full at each site) and none stands in for one. The two 0006 markers follow their italic span; the 0008 marker sits between the lemma and its [n:].

Test 2 + 2a — name the host of every printed negative. 44 printed negatives across the five chunks, each traced to the single Latin word it attaches to. The five sites that fired 2a's trigger condition (a negative with a quantifier or correlative nearby) and how each was proved from structure, not from what reads well:

Test 2b — every anchor checked against the WORD it stands beside in the Latin, not merely for presence. 57 anchors, all verified in place; four were deliberately positioned against English word order to keep the Latin adjacency — 0679B inside a lemma (sexaginta [0679B] for es → "sixty [0679B] valiant men"), 0680C (insolentia prioris [0680C] populi → "the former [0680C] people"), 0684A (communionem [0684A] non devitant → "the fellowship and communion [0684A] of them that err") and 0688D (scientiae [0688D] magnitudinem → "the greatness of knowledge [0688D] from the eyes of men"). 0690C splits an italic span exactly as Migne does (comedit [0690C] me → "hath eaten [0690C] me").

Test 3 — punctuation compared BY COLUMN BAND, never by total. Colons, semicolons, question marks and exclamation marks were counted band by band against the Latin twin, with our English-only markers stripped first: 62 bands, zero mismatched. It caught three defects that a total would have hidden — a supplied semicolon at 0682B (Qui et Abdo, hic est), a supplied semicolon at 0688B, and a colon that had landed on the wrong side of the 0692C anchor. All three repaired.

Test 4 — does it ASSERT what he asserts? Every paragraph re-read against the Latin with the counts already clean. It is what produced §E3 (the doubled sentence), §E6 (the antecedent of quos), §E7 and §E8 (the two ambiguities recorded rather than silently chosen), and the four fragment expansions in §A. ⚑ The one place a smooth English sentence was a suspect and survived: 0688A's quam sacerdos auri lamina protegebat reads easily as "which the priest covered with a plate of gold", and the plate-of-gold is the high priest's own lamina of Ex 28:36, which is why it reads well; checked and left.

Unexpressed subjects supplied in this batch — every one named, and what named it

§G · THE PLATE, AND WHAT A READ WOULD SETTLE

Migne prints no numbered Forte apparatus in this work — a checked zero, carried from the volume's spot check; no [cn:] is in the Latin and none belongs in the English. No plate access was available to this translator, so nothing here was patched and nothing was marked on the strength of a reconstruction.

Candidates for a plate read, in priority order:

1. 0684A reperetur (§E1a) — the only site where the reading affects an English word that is not otherwise recoverable; two candidate reconstructions, and the plate chooses between them.
2. The eight non-words of §E1 (0678D · 0679B · 0680A · 0680D · 0682C · 0684C · 0692B ×2) — each is a data/tei-patches/8965.json fix if the plate has the sound form, and not one is a [sic:] until a plate says the defect is Migne's.
3. 0680B Senarius (§E3) — whether the plate prints the sentence twice, and with which first word.
4. 0678D centum et sex (§E2) — real words, so the printed reading stands either way; a plate read would only tell us whether the arithmetic failure is Migne's or older than him.

⚠ Our archive.org copy of PL 113 clips the left margin on some pages. Two of the candidates above sit at a column head where that clipping would show — for es at the head of 0679B and tranferunt / decjiciunt in the 0692B column — so a truncated left margin is a live alternative explanation for them and is raised here rather than patched. Nothing was conjectured into the text on that account.

Sections read clean, and named so a blind reader knows they were looked at: CAPUT XIII (0682B–C, the three place-name glosses) · CAPUT XXIII (0686A, both entries) · CAPUT XXVII's Ophel exposition (0688C–D, seven scripture citations, all seven verbatim in the Clementine) · CAPUT XXXI and XXXIII (0689C–D) · CAPUT XXXV VERS. 1 and VERS. 3 (0690B–C).

§Z · THE BLIND POLARITY READ (step 4a), AND WHAT IT REPAIRED

Run 2026-08-25 by a reader fenced from this file, per the runbook: given the cruces first, every agent in the 2026-07-28 sweep returned "all already logged" and found nothing new, while blind readers on the same works found four live defects. All 11 chunks read against their Latin twins.

⭐⭐ EVERY PRINTED NEGATION IN THE WORK WAS CORRECT AND CORRECTLY CARRIED. Not one polarity defect. That is now the fifth consecutive book of this series where the name of the class has proved to be something else. The yield was the REFERENT, and it is worth stating as the finding rather than as a list: what English cannot leave implicit and Latin can.

Repaired, seven sites:

1. 0672B — the grace of confirmation had become the grace of baptism. cum post ablutionem baptismi gratia implentur spirituali: the genitive baptismi belongs to ablutionem, and gratia … spirituali is the ablative with implentur. Our English lifted the genitive off its noun and fused the ablative into it — "filled with the grace of baptism that is spiritual". ⛔ It collapsed the exact distinction the paragraph exists to draw, which its own next sentences spell out: Lavatur … aqua baptismi against per impositionem manus episcopi, accipit donum Spiritus sancti. Now "after the washing of baptism they are filled with spiritual grace."
2. 0686B — a supplied pronoun made Athalia a man, in a sentence that names her in its own middle clause: "Not that he then had sons … Athalia endeavoured … he calls his sons". molita est is feminine, and the lemma's ejus is hers (2 Par 24:7). The lemma itself carried it too, "And his sons". Now And her sons / she / her.
3. 0689D — a participle bound to the wrong noun. Contra paganos et haereticos pugnantes became "Against pagans and heretics who fight", making the caution owed to a subset of heretics and evaporating the reason for veiling the mysteries — which is why the proof-text is Nolite sanctum dare canibus. Now "For those fighting against pagans and heretics".
4. 0692A–B — agency inverted. se subditos convertunt ad impietatem read as "turn themselves as subjects", which removes the subjects from a sentence about rulers corrupting them. Migne's own Hieronymus excerpt six lines later is the control: praepositorum depravati exemplis cedunt cum ipsis in praedam diaboli. Rendered "turn those subject to them to impiety" with a [cj: se; read sibi] — the printed word construes, so the reading is shown, not smuggled.
5. 0688A — the wrong man pleased God. placuit Deo sat as a bare "he" immediately after "Zacharias the priest … lived", and its subject is Ozias (2 Par 26:5 is the control: Ozias sought the Lord in diebus Zachariae). Named. 6–7. Two numeral lemmata contracted to what Migne prints — see §Z2.

§Z2 · THE LEMMA-EXPANSION FINDING — RAISED, ACTED ON, AND THEN WITHDRAWN

⛔⛔ THE CORRECTION STANDS FIRST, because a false contraction shipped for part of this session and the reasoning that produced it was inverted. The reader flagged five lemmata printing longer in English than in the Latin twin, each addition exactly the Vulgate continuation. Two of them — Centum. and Ducenti. — were contracted to "A hundred." and "Two hundred." That was wrong, and both have been reverted.

What the plate did and did not settle. Gallica bpt6k5505319w (map f = (column + 1) / 2, calibrated against the printed corner numbers) shows f333 a bare Centum. and f339 a bare Ducenti.. That answered the reader's honest caveat — the words are not dropped from our Latin, our Latin is right — and it answered nothing else. Whether a bare lemma should be expanded in English is a question about convention 3, and no plate can decide it.

⛔ And convention 3 decides it the other way. Both lemmata are clips of two-word numerals: 2 Par 1:17 reads equum centum quinquaginta and 2 Par 8:10 ducenti quinquaginta. A lemma whose governed word is missing is a fragment, and the Ruth pilot's convention 3 renders a fragment as the phrase it stands for, not as a word-for-word stump. The settled ruling on a work that treats the same bare lemma two ways is explicit: resolve it by EXPANDING the stump, not by contracting its twin.

⛔⛔ The gloss argument was backwards, and it is the part worth remembering. The contraction was justified on the ground that Migne's gloss supplies the figure itself — Qui sunt ter quinquageni under Centum., Centenarius … Quinquagenarius vero under Ducenti. — so that an expanded English "pre-empts the gloss's own arithmetic." It does the opposite. Ter quinquageni is thrice fifty, which is 150, not 100: the gloss is not withholding the figure, it is confirming it, and it is the proof that Migne's Centum. means centum quinquaginta. Contracting the English left the lemma saying one hundred on a page whose gloss immediately says one hundred and fifty.

So all five were licensed expansions and none was a defect. Carduus qui, etc. (carrying Migne's own etc.), Obturaverunt cunctos., Ad quam Deus. — the second stint's deliberate, stated applications of convention 3 — and the two numerals with them.

⚑ What survives as a real observation is smaller and duller: the two stints of this work state different rules for a bare cardinal. The second stint ruled that an English cardinal stands as a substantive so nothing is elided (Sexcenta. "Six hundred.", Quadraginta. "Forty."), which is right for a complete numeral and does not reach a clipped one. Migne's Sexcenta. is the whole of 2 Par 1:17's six hundred; his Centum. is half of its hundred and fifty. Same-looking lemma, different class — which is the brief's own warning that a control must be the same class as the thing it controls. No further change was made on either account.

⚑ The general lesson, and the reason this is written out rather than quietly reverted: a blind reader's finding can be right about the FACT and wrong about the RULE. The five English lemmata really did print words the Latin twin does not — that observation was correct and worth having. What it could not supply is whether the convention licenses them, and acting on the fact without checking the rule turned three sound renderings into candidates and two into errors.

⚑ One flag from the reader that is NOT a gap — recorded so it is not re-raised

It reported that 0687D's priusquam (for postquam) and 0680D's usque ab are carried literally with "no crux or marker", leaving a reader in front of a self-contradicting sentence with nothing to explain it. Both are in fact argued in this file (§E4 and §D), and the work page links its cruces. The reader could not know that, because it was fenced from this file on purpose — which is the fencing working, not failing. ⚠ A blind reader's claim that something is undocumented is always about what it could see, and must be checked here before it is believed.

Rejected candidates are listed in the reader's own terms and were not acted on: four lower-confidence parses (0674B, 0670D, 0673C) where a genitive floats but the sense survives, and eight rejected outright with reasons — among them 0672B's subjectless deliquerat (Migne's, English must supply something and supplied the least committal thing) and both comparatives at 0682A, checked for the 11088 reversal and running the right way.

Chunks read and found CLEAN: 0001, 0004, 0005, 0007. Named, because a report that lists only findings cannot be told apart from a report by a checker that never read.