Liber Paralipomenon I

Anselm of Laon and his school, on the first book of Paralipomenon. Translated 2026-08-25 by three translators working in parallel — 0000–0004 (the genealogies, cols 0629D–0643D, 5,203 Latin words), 0005–0009 (the dwellers in Jerusalem to Nathan's oracle, 0644A–0655B, 4,129 words) and 0010–0013 (David's wars to the courses, 0655C–0664D, 3,570 words) — and merged here. 14 chunks, 12,902 Latin words. verify-english.mjs clean over the whole work; question-mark parity exact in every chunk.

Apparatus fired, work-wide: 34 [var:] · 3 [cj:] · 0 [sic:] · 0 [d:] · 0 [ed:] · 0 [nt:] · 0 [cn:]. Column anchors and [n:] notes at exact parity; zero em-dashes. The work asserts nothing whatever about Migne's plate — thirteen plate candidates were raised across the three stints and not one fired a marker.

§M · THE MERGE — three stints, two things reconciled, one boundary ruled

Three translators, none able to see the others when it mattered. The merge is a check, not a concatenation, and it was made against the English files themselves rather than the three vocabulary tables — a table records what a translator believed it did.

⭐ The second and third stints read their predecessors' cruces off disk and inherited from them deliberately, which is how this is supposed to work: doctores → teachers (15 occurrences, zero "doctors"), interpretatur → "is interpreted", charitas/dilectio held apart, verus pacificus → "the true peaceable one" (which recurs at 0654D and gave the seam a real cross-check), proper names form for form, as far as for the abridgment formula. Three extensions were declared rather than smuggled: septenarius → "the septenary number", signat → betokens (a fourth verb kept off significat), gentilitas → "the gentile world". And 8965's substantival number-names, imported at the head of the run, finally bite in the last stint — the senary, the quaternary, the octonary.

⛔ The one drift that was real: the divergence marker had two forms

The opening stint wrote [var: Vulg. Alian] — the Clementine's word and nothing else. The later two wrote [var: Vulg. 1 Par. 9:11 Mosollam …], with book, chapter and verse. Ten markers against twenty-four.

⚑ The third stint noticed and deliberately did NOT harmonize it, flagging it for the merge instead. That was the right call and is worth stating as practice: a translator that silently conforms to a peer's format leaves no record that the two ever differed, and the next reader cannot tell an inherited convention from a coincidence.

Ruled to the house form, and all ten normalized here, each reference resolved against sources/vulgate/clementine-flat.txt rather than assumed: Phaleg → 1 Par. 1:19 · Alian → 1:40 · Hamram → 1:41 · dux Alva → 1:51 · videbat → 2:52 · peccatum mundi → Jo. 1:29 · adversus eam → Mt. 16:18 · adhuc delicatus → 2 Rg. 3:39 · Scientia inflat → 1 Cor. 8:1 · salvabitur → Jo. 10:9. All 34 markers in the work now carry an explicit reference. The reason the bare form will not do is the reason [var:] exists at all: the marker is a public claim about another text, and a claim a reader cannot look up is not checkable.

⚖ THE BOUNDARY OF THE MIRROR RULE — ruled here, and reversible if Wilson disagrees

The third stint capitalized two words the plate sets lower-case — faciente domino (0661C, in the etymology of Asabias) and In regum (0657D) — and flagged the departure rather than hiding it, asking whether the mirror should be absolute.

Ruled: it should not, and the plate itself is the argument. In the very chunk containing In regum ita, Migne also prints In Regum ita — both casings of the same phrase, a few lines apart. His case is not carrying a distinction there, so mirroring it would transcribe an accident as though it were a signal.

So the rule has a stated scope: mirror the plate where the case is SEMANTIC IN ENGLISH — Ecclesia/ecclesia is the founding case, and it qualifies because "the Church" and "the church" are two different claims a reader can tell apart. Do not mirror where English has no lower-case option for the referent — the divine name, a book title. Rendering faciente domino as "the lord doing" would not be fidelity; it would be a misspelling wearing fidelity's clothes.

⚑ Ecclesia itself was mirrored word for word throughout all three stints, and the lower-case sites were verified at the source, not taken on trust — including 0011's ecclesiarum → "the churches of God".

Checked and sound

• usque ad censused per occurrence in all three stints and never by rule: 15 occurrences in the first range of which only 4 are the formula, 13 in the second with 9, 22 in the third with 4. The traps were caught — ab Adam usque ad Noe (0632C) and Verumtamen usque ad tres primos (0648A), 11:21's own words inside a lemma.
• Migne's displaced verse numbers, four in the first range and more later, reproduced untouched along with a malformed VERS 9.--.

§T · WHAT THE FOURTH TEST FOUND, AND WHY IT IS THE ONE THAT MATTERS HERE

Every containment check passed on this book from the start. Three conformations were caught only by reading the finished English back against the Latin asking whether it asserts what he asserts — and all three were the familiar English Bible arriving over the plate:

Absalom → Absalon · Judas Iscariot → Judas Scarioth · In Hebron → In Hebrun.

⛔ Nothing mechanical sees these. The name is present, spelled like a name, in the right place, and it reads better than the truth. They are the exact defect this edition exists to catch, and a genealogy is where they are easiest and least visible — which is why the first stint was warned about them and the third still found three.

The same test produced the reading repairs no count would show: fidei taken as an accusative object (0656B), sibi as "to him" (0657D), and a Douay-supplied object in Osee 6:5.

§V · THE DIVERGENCES THAT CARRY AN ARGUMENT

Thirty-four is a high count, and the reason is that this book reasons from its lemmata. Four are worth naming because in each the divergence is not incidental to the gloss but load-bearing:

• ⭐⭐ 0657A — an arithmetic argument with a wrong operand. The gloss quotes 2 Reg. 24:9 at 80,000 and 50,000 where the Clementine reads 800,000 and 500,000, and then at 0657C reasons from the difference between the two books' numbers. The marker is what lets a reader see that the argument is being run on figures the received text does not have.
• ⭐⭐ 0636A — qui dividebat dimidium requietionum for the Clementine's videbat (1 Par. 2:52). The gloss underneath uses videre four times and dividere not once.
• ⭐ 0639C — Migne negates an apostolic sentence. He prints scientia non inflat; 1 Cor. 8:1 has no negative at all. The negative was rendered on its printed host and the divergence marked — never silently repaired, which is exactly 7a.
• ⭐ 0659C–D — the gloss does textual criticism on its own lemma and lands on ad torrendum, which is the Clementine's reading. torrentem and torrendum are held apart as "the parching" and "the roasting"; collapsing them would have erased the argument being made.

§W · THE ONE PLACE THE THREE STINTS' CAUTION IS VISIBLE

Thirteen plate candidates, zero markers. Non-words (Jerusasem, Quincunque, firmamamentum, Isti sum nominati), split runs in names (Vas seni, de pravatus), and a stranded est — all rendered for their evident sense, all logged, none patched into the Latin and none blamed on Migne.

⚠ Several sit at column feet and heads, which is where our archive.org copy of PL 113 is known to clip its left margin (p. 320 and p. 354 confirmed, and p. 320 carries this work's opening columns). That is a fact about our photograph.

⭐ Every one is one Gallica IIIF request away — ark:/12148/bpt6k5505319w, map f = (column + 1) / 2, verified against the printed corner numbers. Seven such reads have already been made in this session on the sibling books and they settled eight questions, going Migne's way every time. None of these blocks the work, which is the point of the ship test: an unread candidate is safe precisely because it fires no marker.

Translated 2026-08-25, one translator, 5 chunks, 5,203 Latin words (frontmatter sum) → 7,335 English, 1.41×. Chunks 0005–0013 belong to peers and go to their own cruces files; the merge is the orchestrator's.

verify-english.mjs clean over this range (the only failures it reports are the peers' still-missing 0005–0013). In this batch: column anchors 54/54 verbatim, in order and in place; [n:] notes 65/65 at parity; italic spans 159/159 (the English shows 169, the surplus being exactly the ten [var:] markers' own internal italics); paragraph/line counts 1:1 in every chunk.

Apparatus fired here: 10 [var:] · 0 [sic:] · 0 [cj:] · 0 [d:] · 0 [ed:] · 0 [nt:] · 0 [cn:] (this volume's plate spot check is a checked zero: Migne prints no numbered Forte apparatus here, so none is in the Latin and none belongs in the English). Zero guillemets in either language. Question marks 2 Latin / 2 English, in the same bands (0633C and 0642B). Zero em-dashes in the English.

Ruth (8968) cruces.md §1 governs throughout and was not reopened: VERS. n.-- passes verbatim, the lemma is Englished, a fragmentary lemma is rendered as the phrase it stands for, the attribution sigla ride as ordinary [n:] notes 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) and Paralipomenon II (8965) are format models, never authorities over this plate.

§0 · VOCABULARY AND THE CHOICES THIS BATCH MADE — for the merge

Fixed at first occurrence and never varied inside this range. Nothing here diverges from 8965's §0/§A tables; where 8965 settled a word, that settlement was followed.

⚑ 8965's number-names table (senarius → the senary, etc.) never fires in this range — this book's numbers are all cardinal counts of men and years, and none is expounded as a number-name. Pascha / phase likewise does not occur. Recorded so the merge can see the tables were consulted and found inapplicable, not overlooked.

Proper names are kept form for form as Migne prints them, never normalized to a familiar English Bible spelling. In this range that means, among others: Nemrod (never Nimrod) · Jerusasem at 0635D (see §B2) · Gergezaeum with z · Samareum for Clem. Samaraeum · Phaley · Aliam · Hamran · Aliha · Armon for Arnon (0634B, 0634C) · Jesbii · Ianoe · Nephthali · Theglathphalnasar at 0642A beside Teglathphalnasar at 0642A (Migne spells it two ways nine lines apart; both reproduced) · Ahila/Ahilam and Gozam/Gozan, likewise doubled by Migne and likewise reproduced · Vas seni (§B3) · Reginae (§C). The Ruth Lia ruling.

How a bare lemma was decided

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

• A lemma that yields a readable English phrase naming the thing glossed stands as it is — the great majority here, because they are names: Thubal. · Mosoch. · Tharsis. · Cethim. · Dodanim. · Phut. · Saba. · Phaley. · Ceturae. · Assurim. · Epher. · Thamna. · Sephi. · Hamran. · Jacan. · Denaba. · Bosra. · Themanorum. · Madian. · Rohoboth. · Calubi. · Jether. · Ephrata. · Jair. · Onam. · Sesan. · Hebron. · Cariathiarim. · Bethlehem; · Daniel. · Senneser. · Asalelphuni. · Recha. · Ezra. · Johanan. · Abia. · Happhim,. Nothing is elided in any of them, so nothing is supplied.
• A lemma that leaves a dangling English word is expanded to the phrase it opens (Ruth §1.2), twice in this batch and only twice:
• 0635A · Achar qui. → "Achar who troubled Israel." (2:7 Achar, qui turbavit Israel). "Achar who." is not English.
• 0639D · Her pater. → "Her the father of Lecha." (4:21 Her pater Lecha). Here the expansion does a second job: bare "Her" reads in English as a pronoun, and the whole gloss is about the man Her, whom Migne identifies with Elimelech.

⛔ Contrasting these two with the thirty-seven bare names above would manufacture an inconsistency that is not there. And the expansion is not conformation: it supplies the verse's own clipped words, it never substitutes a familiar English Bible's wording for what Migne prints.

§A · THE LEMMA COLLATION — every italic span in 0629D–0643D, checked in sources/vulgate/clementine-flat.txt

Migne's book is 1Par in that file and the chapter/verse numbering is identical, so no conversion arises. 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, and the ten that earned a public claim are argued in §C. Every reading below was looked up; none rests on recollection.

CAPUT PRIMUM (0629D–0635A) — the whole of 1 Par 1

Noe. (1:4, verbatim) · Noe vir justus erat et perfectus in generationibus suis, etc. (Gn 6:9 — Clem. Noe vir justus atque perfectus fuit in generationibus suis: et for atque and erat moved forward. Particle and word order only, no marker, §D) · Filii Japheth. (1:5, verbatim) · Thubal. (1:5) · Mosoch. (1:5) · Filii autem. (1:7, verbatim clip of Filii autem Javan) · Tharsis. (1:7) · Cethim. (1:7) · Dodanim. (1:7) · Phut. (1:8) · Saba. (1:9) · Saba Arabia, (not scripture — the etymology formula interpretatur nunc; see §B4) · Iste coepit esse potens in terra. (1:10, verbatim; Migne's initial capital for Clem.'s lower case) · Mesraim autem genuit. (1:11 — Clem. Mesraim vero genuit. Particle only, no marker, §D) · Gergezaeum, Hevaeum. (1:14–15 — Clem. 1:14 Gergesaeum and 1:15 Hevaeumque: a z/s spelling and a clip joined across the verse division. No marker, §D) · Aradium quoque. (1:16, verbatim) · Et Samareum. (1:16 — Clem. et Samaraeum. Orthography, no marker, §D) · Hamathaeum. (1:16, verbatim) · Filii Sem, etc. (1:17, verbatim) · Hus et Hul. (1:17, verbatim) · Phaley. (1:19 — Clem. Phaleg. [var:] FIRED, §C-a) · Jectan autem. (1:20, verbatim) · Sem, Ar. (1:24 — Clem. Sem, Arphaxad, Sale; Ar. is Migne's abbreviation, printed with its own full stop, and is reproduced as printed, §B5) · Primogenitus. (1:29, verbatim) · Ceturae. (1:32, verbatim) · Assurim. (1:32, verbatim) · Epher. (1:33, verbatim) · Thamna. (1:36, verbatim) · Aliam, (1:40 — Clem. Alian, and Migne's own gloss two lines below reads Alian too. [var:] FIRED, §C-b) · Sephi. (1:36 and 1:40, verbatim in both) · Hamran. (1:41 — Clem. Hamram, and Migne's own gloss one line below reads Hamram too. [var:] FIRED, §C-c) · Jacan. (1:42, verbatim) · Denaba. (1:43, verbatim) · Jobab filius Zarae, (1:44 — Clem. 1 Par Zare, while Gn 36:33 reads Zarae, which is Migne's form. Orthography, no marker, §D) · Bosra, etc. (1:44, verbatim) · Fulvida vestimenta ejus ex Rosor (no such phrase anywhere in the Clementine; see §B1 — a base verse that cannot be fixed, and no [var:]) · Themanorum. (1:45, verbatim) · Madian, (1:46, verbatim) · Rohoboth. (1:48, verbatim) · Matred filiae, etc. (1:50, verbatim) · Duces, etc. (1:51, verbatim clip) · Dux Aliha. (1:51 — Clem. dux Alva, and Gn 36:40 also dux Alva. [var:] FIRED, §C-d).

CAPUT II (0635A–0636D)

Achar qui. (2:7, verbatim clip; Migne drops Clem.'s comma) · Calubi, (2:9, verbatim) · Isai autem genuit. (2:13, verbatim) · Jether. (2:17, verbatim) · Ephrata. (2:19 — Clem. Ephratha. Orthography, no marker, §D) · Jair. (2:22, verbatim) · Nati sunt. (2:25, verbatim) · Onam. (2:26, verbatim) · Sesan. (2:34, verbatim) · Hebron. (2:43, verbatim) · Hi erant filii. (2:50, verbatim) · Cariathiarim, (2:50, verbatim) · Bethlehem; (2:54, verbatim; Migne's semicolon for Clem.'s comma) · Qui dividebat dimidium requietionum (2:52 — Clem. qui videbat dimidium requietionum. [var:] FIRED, §C-e) · Habitantium. (2:55, verbatim clip of habitantium in Jabes).

Scripture the gloss quotes inside CAPUT II: Quincunque Spiritu Dei aguntur, hi filii Dei sunt (Rom 8:14 — Clem. Quicumque enim Spiritu Dei aguntur, ii sunt filii Dei. The dropped enim, hi for ii and the word order are §D; Quincunque is a NON-WORD and a plate candidate, §B2). Also Haereditas mea Israel, et quorum possessio Dominus est — printed as roman prose, not italic, and matching no Clementine verse (a weave of Is 19:25 with the levitical Dominus pars haereditatis formula); the italic inventory does not reach it, and it takes no marker.

CAPUT III (0636D–0638A)

David vero. (3:1, verbatim) · Jezraelitide. (3:1 — Clem. Jezrahelitide. Orthography, no marker, §D) · Daniel. (3:1, verbatim) · De Egla uxore sua. (3:3, verbatim; Migne's initial capital) · In Hebron ubi regnavit, (3:4, verbatim clip; Migne drops Clem.'s comma) · Omnes hi filii David, (3:9, verbatim clip; Migne drops Clem.'s comma after hi) · Filii autem Josiae. (3:15, verbatim clip) · Asir, Salathiel, Melchiram, Phadaia, Senneser, etc. (3:17–18, verbatim across the verse join) · Senneser. (3:18, verbatim) · redemptus Domini, · carceratus, · frutex Domini; (not scripture — Hebrew name etymologies, §B4).

Scripture the gloss quotes inside CAPUT III: Ephraim vitula docta diligere trituram. (Os 10:11, verbatim clip) · Non enim haeres erit filius ancillae cum filio liberae (cited Gen. XXI, and the printed words are verbatim Gal 4:30, where Gn 21:10 reads non enim erit haeres filius ancillae cum filio meo Isaac. No [var:] — Pattern 14 is about the text, not the reference, and the text agrees with the received text exactly; §D) · Postula a me et dabo tibi gentes haereditatem tuam (Ps 2:8, verbatim clip; Migne drops Clem.'s comma after me) · novit Dominus qui sunt ejus (2 Tim 2:19 — Clem. cognovit Dominus qui sunt ejus. Simplex for compound of the same verb, with no change of sense; no marker**, and the reason is argued in §D).

CAPUT IV (0638A–0641C)

Filii Juda. (4:1, verbatim clip) · divisio, · atrium, · ortus. · vinea mea, · vigilans. · sinistra, · murmuratio, · spica, · vana vetustas. · missus vel umbra ejus. (all not scripture — name etymologies, §B4) · Asalelphuni. (4:3, verbatim) · Gedor et Ezer. (4:4, verbatim apart from Clem.'s comma) · Fuit autem Jabes. (4:9, verbatim) · Et mater (4:9, verbatim clip of et mater ejus) · Invocavit. (4:10, verbatim) · Caleb autem frater. (4:11, verbatim clip) · excelsus, · humilis, (etymologies) · Recha. (4:12, verbatim) · Saraia autem genuit. (4:14, verbatim, addressed under VERS. 13 — Migne's own verse numbering, reproduced, §B7) · scientia (inside the 1 Cor quotation, below) · Ezra. (4:17, verbatim) · auxiliator, · auxilium, · qui rebellans, · pulverem, (etymologies) · Genuitque Mariam. (4:17, verbatim) · Filiae Pharaonis. (4:18, verbatim clip) · Her pater. (4:21, verbatim clip) · Securus et Incendens. (4:22, verbatim; Clem. prints a comma between them) · Qui reversi sunt in, (4:22, verbatim clip) · Figuli habitantes. (4:23, verbatim, from Hi sunt figuli habitantes) · Habitaverunt in, etc. (4:28 — Clem. Habitaverunt autem in Bersabee. Particle only, no marker, §D) · Commoratique sunt ibi. (4:23, verbatim, addressed under VERS. 28 — Migne's numbering, §B7) · Et universa cognatio non, etc. (4:27, verbatim, addressed under VERS. 28, §B7) · Hae civitates, (4:31, verbatim) · Isti sum nominati. (4:38 — Clem. Isti sunt nominati. sum has no grammatical slot in the clause and is a plate candidate, §B2; no marker) · sepes ejus. (etymology of Gador) · In eo terra latissima et quieta, (4:40 — Clem. et terram latissimam et quietam et fertilem: the case is the glossator's own weave into a nominative sentence and fertilem is a clip. No marker, §D) · Percusserunt tabernacula: (4:41, verbatim clip; Migne's colon) · Viri quingenti. (4:42, verbatim).

Scripture the gloss quotes inside CAPUT IV: Si dimittimus eum sic, omnes credent in eum (Jo 11:48, verbatim clip) · Ecce agnus Dei qui tollit peccata mundi, etc. (Jo 1:29 — Clem. Ecce agnus Dei, ecce qui tollit peccatum mundi. [var:] FIRED on the number, §C-f) · Venit princeps hujus mundi, et in me non habet quidquam. (Jo 14:30 — Clem. venit enim princeps mundi hujus: a particle and word order, §D) · Dilata locum tentorii tui, etc. (Is 54:2, verbatim clip) · Et portae inferi non praevalebunt adversus eum, (Mt 16:18 — Clem. adversus eam, of the Church. [var:] FIRED, §C-g) · data est ei potestas super omnem virtutem inimici (Lc 10:19 — the tail super omnem virtutem inimici is verbatim, but the frame data est ei potestas is the glossator's, against Clem. Ecce dedi vobis potestatem calcandi supra serpentes. A weave, not a quotation to collate: [var:] declined, §D) · Ego sum tener et unctus rex (2 Rg 3:39 — Clem. Ego autem adhuc delicatus, et unctus rex. [var:] FIRED, §C-h) · scientia non inflat, sed charitas aedificat (1 Cor 8:1 — *Clem. Scientia inflat, caritas vero aedificat, with NO negative. [var:] FIRED, §C-i) · Per me si quis introierit, ingredietur et egredietur, et pascua inveniet (Jo 10:9 — Clem. Per me si quis introierit, salvabitur : et ingredietur. [var:] FIRED on the dropped clause, §C-j) · qui semetipsum exinanivit, servi formam accipiens (Phlp 2:7 — Clem. sed semetipsum exinanivit, formam servi* accipiens: qui for sed is the glossator's relative weave and servi formam is word order, §D) · Et erit in loco ubi ante dictum est: Non plebs mea vos: ibi vocabuntur filii Dei (cited Osee. I, and the printed words are Rom 9:26, not Os 1:10 which reads Non populus meus vos … Filii Dei viventis. Against Rom 9:26 Migne adds ante, drops eis, and drops vivi. No [var:], for the reason given at Gn 21 above and in §D).

CAPUT V–VIII (0641C–0643D)

Sed cum violas, etc. (5:1 — Clem. sed cum violasset thorum patris sui. A truncation, and it is rendered as the phrase it truncates, §B6; no marker, §D) · Contra Agarenos. (5:10 — Clem. contra Agareos. A name form on which nothing in the gloss turns; no marker, §D) · Et suscitavit. (5:26, verbatim clip) · Filii Levi, Gerson, etc. (6:1, verbatim, addressed under VERS. 6 — Migne's numbering, §B7) · Johanan. (6:9–10, verbatim) · Porro Josedec. (6:15, verbatim clip) · Abia. (6:28, verbatim) · Isti sunt quos constituit, (6:31, verbatim clip) · Ethan filius. (6:44, verbatim clip) · Levitae qui. (6:48, verbatim clip of Fratres quoque eorum Levitae, qui ordinati sunt) · Aaron. (6:49, verbatim) · Omnes hi, filii Bechor. (7:8, verbatim; Migne adds a comma Clem. does not print) · Omnes hi filii (7:8, the same words re-quoted without the comma — Migne punctuates his own lemma two ways nine lines apart, and both are reproduced) · Concubinaque ejus. (7:14, verbatim) · Happhim, (7:15, verbatim) · Soror autem ejus Reginae. (7:18 — Clem. Soror autem ejus Regina peperit Virum decorum, nominative. Case only, no marker, §D and §B8) · Filii autem Ephraim. (7:20, verbatim clip) · Qui translati sunt in Manahath. (8:6, verbatim clip) · Naaman autem et Achia: (8:7, verbatim clip; Migne's colon) · Ipse transtulit eos, (8:7, verbatim).

Scripture the gloss quotes inside CAPUT V–VIII: Tu es sacerdos in aeternum secundum ordinem Melchisedech, etc. (Ps 109:4, verbatim clip) · Factum est cum senuisset Samuel, posuit filios suos judices Israel, et caetera. (1 Rg 8:1 — Clem. Factum est autem cum senuisset: a particle, §D) · Ut scias quomodo oporteat te conversari in domo Dei, quae Ecclesia Dei vivi, etc. (1 Tim 3:15 — Clem. ut scias quomodo oporteat te in domo Dei conversari, quae est ecclesia Dei vivi: word order and a dropped copula, §D) · Vos estis genus electum, regale sacerdotium, etc. (1 Ptr 2:9 — Clem. Vos autem genus electum: Migne prints estis where the Vulgate prints autem, which changes an apposition into a predication without changing what is asserted; no marker, §D) · Filii Ephrem conversi sunt in die belli (Ps 77:9 — Clem. Filii Ephrem, intendentes et mittentes arcum, conversi sunt in die belli: an internal omission with nothing in the gloss turning on it; no marker, and the reasoning is set out in §D beside the Jo 10:9 case that went the other way).

§B · THIS RANGE'S CRUCES

B1 · 0634B · Fulvida vestimenta ejus ex Rosor — an unfixable base verse and an unattested name. NO MARKER.

Appellatur et alia Bosor civitas Esau in montibus Idumeae, de qua dicitur: Fulvida vestimenta ejus ex Rosor — and then Migne's sentence simply stops, with no terminal point (Pattern 8: it stops in the English too).

Two things are wrong here and neither earns a marker.

1. Rosor is a name that occurs nowhere — not in the Clementine, not in this work, and Migne's own sentence has printed Bosor twice in the twelve words before it. An initial R for B is the obvious reading. ⛔ But an unfamiliar proper name is a real word for the purposes of the 2026-08-23 addendum, and a real word is Migne's until a plate says otherwise. Rendered as printed ("from Rosor"), logged here, no [sic:] — a [sic:] is a public claim that his plate is defective, and seven such claims were withdrawn in one batch for exactly this.
2. The quotation matches no Clementine verse. The nearest is Is 63:1–2 (tinctis vestibus de Bosra … Quare ergo rubrum est indumentum tuum), and fulvida answers the fulvus of Migne's own Edom etymology at 0633C. [var:] declined: a marker naming a verse I cannot establish would be a public claim resting on a guess (8965 §4's Oraculum in abditis precedent).

B2 · Two NON-WORDS and one broken form, all rendered for their evident sense. NO [sic:] FIRED.

The addendum measured non-word suspicions at ten for ten as our digitization's error rather than Migne's, and 8965 fired no marker on any of its four. Same treatment here; all are logged as plate candidates and none carries a marker (the ship test: an unread candidate is safe iff it fires nothing).

⚠ The last of these would qualify for a [sic:] under Pattern 12's "no grammatical slot" test, and it deliberately does not get one. The reason is the ship test again, sharpened by this volume's own warning: our archive.org copy of PL 113 clips the left margin on some pages, so a broken form near a column head is our scan before it is Migne's plate. It is raised, not asserted.

⛔ And note where these three sit. Jerusasem is at 0635D and Quincunque at 0636D, both in the column-D band at the foot of a page; Isti sum opens a paragraph at 0640C. None of them was patched into the Latin, and none should be until a plate is read — if a read confirms the sound forms, these are data/tei-patches/8964.json fixes and nothing else.

B3 · 0642C · Joel, qui et Vas seni — a split run in a NAME. Carried as printed, no marker.

Clem. 1 Par 6:28 prints primogenitus Vasseni, et Abia as one word (itself a famous corruption of the Hebrew, which is not our business). Migne prints it as two words, in prose, and this is exactly Pattern 10's split-run shape. It is carried into the English as the two words Migne prints — "Joel, who is also Vas seni" — because a name's letters are what the reader needs, and joining them would be our conjecture. ⚠ Also a left-margin plate candidate: the run stands three words into a line at 0642C. No [sic:], for the reason in B2.

B4 · The ETYMOLOGY spans are not scripture and were not collated

Twenty-five italic spans in this range are Hebrew name-glosses in the interpretatur formula, not quotations: exsiccans · auriga · ascendens · divisio · atrium · ortus · vinea mea · vigilans · sinistra · murmuratio · spica · vana vetustas · missus vel umbra ejus · excelsus · humilis · auxiliator · auxilium · qui rebellans · pulverem · sepes ejus · redemptus Domini · carceratus · frutex Domini · Saba Arabia · scientia (this last a word lifted out of the 1 Cor 8:1 quotation and italicized separately by Migne, and it is collated there). Pattern 14 does not reach them and a Clementine search would have produced noise, not findings — the liturgical-composition boundary of style rule L3, applied to the other genre that shadows scripture without quoting it.

⚑ And the etymologies are where the names have to be right. Three glosses in this range are built on the letters of the name, and the English keeps them for that reason: Phaleg = division (0632C), Alian = cast out against Genesis's Aluham = exalted (0633C–D), Hamram = reddened against Genesis's Amdan = desirable (0633D). ⛔ In all three Migne's lemma prints a different form from the one his own gloss expounds — which is what §C-a, §C-b and §C-c are about, and why they were worth a marker.

B5 · 0632C · Sem, Ar. — an ABBREVIATION, reproduced, not expanded

Clem. 1 Par 1:24 reads Sem, Arphaxad, Sale. Migne prints Ar. with its own full stop inside the italic span, which is the shape of an abbreviation rather than of a clipped lemma. Reproduced exactly (Ruth §1.2's expansion licence covers a clipped lemma, not an abbreviated word), and the gloss below turns on the count of generations, not on the name. Logged so a later reader can tell a decision from an oversight.

B6 · 0641D · Sed cum violas, etc. — a truncation, rendered as the phrase it truncates

Clem. 5:1 reads sed cum violasset thorum patris sui. violas is a real Latin word ("thou violatest", or "violets") that cannot stand in a cum-clause of this shape, and Migne's own etc. marks the span as clipped. Rendered "But when he violated, etc." — the verb the truncation stands for, stopping where Migne stops. No [var:]: this is Migne's clip, not a divergent reading of the verse. ⚠ It is also a left-margin candidate (0641D, foot of column). Raised, not asserted.

B7 · Migne's own VERSE NUMBERING, four times off. REPRODUCED, never corrected.

Four lemmata in this range sit under an address that is not their verse. The address passes through untouched (convention 1) and the true verse is recorded here:

• 0639C · VERS. 13.-- Saraia autem genuit. — the words are 4:14, not 4:13.
• 0640C · VERS. 28.-- Commoratique sunt ibi. — the words are 4:23, not 4:28.
• 0640C · Et universa cognatio non, etc., under the same VERS. 28 — the words are 4:27.
• 0641D · VERS. 6.-- Filii Levi, Gerson, etc. — the words are 6:1, not 6:6.

Also in this class: 0637C · VERS 9.--, printed with no full stop after VERS, alone among the seventy-odd verse marks in this range; and 0639B · VERS. 12.--[n: (ID.)] Recha., where the note stands before the lemma and there is no space after the dashes. Both reproduced verbatim (8965 §2o's VERS.--6.-- precedent).

B8 · Sentences Migne leaves unstopped, and other punctuation reproduced

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

• 0634B …de qua dicitur: Fulvida vestimenta ejus ex Rosor — no stop (§B1).
• 0635B …ubi propheta vocatur, Nathan scribitur, non Jonathan — no stop.
• 0635C …filius Judae, filius Jerameel per Mem scribitur — no stop.
• 0636C …cum et anima et corpore perfruetur perfectione quietis — no stop.
• 0637B …scandalizata in passione ejus sibi mortem aeternam ascivit — no stop.
• 0639B …quia data est ei potestas super omnem virtutem inimici — no stop, and inside an italic span whose column anchor Migne splits mid-phrase (data [0639B] est), reproduced split.
• 0638D Et quia Jabes docebat eos; inserti sunt — a semicolon inside the clause, preserved.
• 0641D …ad Ecclesiae [0641D] populum translata est — the anchor splits Ecclesiae populum, and the English splits it in the same place ("the Church's [0641D] people").

B9 · Places the Latin will not construe. RENDERED AS PRINTED, NO MARKER.

This corpus has no marker meaning "Migne prints a word that will not construe" (Ruth @0534A societate fieri; 8965 §2b), and inventing one for a single site is worse than saying so plainly. Each of these is rendered as closely as the printed words allow, and nothing is emended.

• 0631C · qui ab Amano et Tauro omnia maritima usque ad oceanum possidere Britannicum. An infinitive where a finite verb is wanted; it can only be read as an accusative-and-infinitive still governed by videbimus twelve words back, with qui standing for its subject. Rendered "who from Amanus and Taurus possess all the seaboard as far as the British ocean" — English has no exponent for a stranded infinitive in this position, so the crux is where the difficulty is recorded.
• 0631D · Sic homines a Dei timore revocabat, ut spem in propria virtute poneret. A singular verb where the sense wants the plural homines. Rendered singular ("that he might set his hope in his own strength"), the subject supplied as Nemrod, who is the subject of revocabat in the same sentence and of every finite verb around it. ⚑ The rival — that poneret is a loose singular for the men — is rejected because nothing but the sense supports it, and the sense is precisely what Pattern 7 forbids using to override the morphology.
• 0638A · id est sanctarum animarum quibus Christo, quem Cyrus … significavit … annumeravit. Two datives (quibus, Christo) with a verb that wants a nominative subject. Rendered with Christ as the subject of annumeravit, which is the only reading that yields a sentence; the dative Christo is left without an English exponent, because English marks no case on a name. Logged, not marked.
• 0640B · Mahalon et Chelion, qui hic Securus et Incendens dicuntur, fuisse in Moab, quia … An accusative-and-infinitive with no governing verb anywhere in the paragraph. The English gaps it in the same place ("That Mahalon and Chelion … were in Moab"), and no verb was supplied.
• 0643B · quae in Aegyptum, fame cogente, a patre venalis ducta a Manasse: qui horreis praeerat …, in conjugium misericorditer suscepta est. An anacoluthon straddling Migne's colon. The colon is preserved and the break is preserved; nothing was joined up.

B10 · Real words that read oddly, weighed and left standing

Per the addendum: a real, well-formed word is Migne's until a plate says otherwise, refuted five of five. None of these fires a marker.

• 0633D · rubricatus, sordidus vel tumulentus. Jerome's onomasticon glosses Hamran as rubricatus sive temulentus ("drunken"), and temulentus is very likely the word. But tumulentus is a well-formed formation on tumeo, so it is rendered on its own letters as "swollen", and the conjecture stays here where a reader can weigh it. No [cj:]: Pattern 18 needs the faithful English to mislead, and "reddened, foul or swollen" of a man named "the red one" does not.
• 0639D · eo quod ipsius amica esset, fuit enim filia Levi. amica ("friend") is a real word and the sentence's reason (fuit enim filia Levi) wants a kinship term. [cj:] DELIBERATELY NOT FIRED, and the reason is that the two rivals cancel: amita ("paternal aunt") is the obvious repair from the Jewish tradition, but the Vulgate itself reads patruelem suam at Ex 6:20 — a cousin, not an aunt. A [cj:] naming one of them would tell a reader the matter is settled when the plate's own comparand points the other way. Rendered "because she was his own friend".
• 0642C · cantorum David qui organis musicis laudes Domine canebant. A vocative where the dative Domino belongs. English has no exponent for the difference in this position (Pattern 9's shape), so it is rendered "praises to the Lord" and logged, not marked.
• 0632B · quae vario eventu subversa bellorum — bellorum separated from eventu across the participle. Word order, not a defect; rendered "overthrown by the varying issue of wars".
• 0634B — Migne says Job's wife bore him septem filios and then names six: Eliphaz, Armoe, Diasath, Philasia, Dipharceluth, Thelon. Both reproduced.** The count is his and the list is his, and the discrepancy is exactly the kind of thing this edition exists to leave visible.
• 0641B · hi quatuor dicuntur under the lemma Viri quingenti ("five hundred men"). Migne's, reproduced; the quatuor is his allegory of the four evangelists and does not answer the lemma's number.
• 0633A · sed usque in praesens, quia procul a nobis sunt — an adverbial phrase with no verb of its own, rendered as printed ("but down to the present").
• Read clean and rejected as candidates, with reasons: Syriae Cheles (0631A, 0631D) for Syria Coele, a standard medieval form, printed twice identically · questionibus for quaestionibus (0637A) and coelestis/coelis throughout, orthography, Pattern 9 · Requietatio beside requietio (0636A), a live doublet in this very paragraph · ononymos (0634C), the Greek ὁμώνυμος, rendered "of the same name with it" · Castium (0632A) for Jerome's Casluim, an unfamiliar name and therefore a real word (brief) · Hala filii Beor (0634A) where Clem. 1:43 has Bale filius Beor, a name in Migne's prose and not a lemma · Aluham (0633C) where Gn 36:23 reads Alvan — likewise Migne's prose, and it is the very form his etymology expounds.

B11 · 0633C · Pattern 8b applied once, and it is the batch's only added mark

Quaeritur, cum Thamna in Genesi legatur … quomodo inter filios numeretur, cum in sequentibus legatur ipsa Thamna soror Lotam filii Seir, qui et Edom? — Migne closes an indirect question with a question mark, and English cannot end that construction with one. Recast into direct form behind a colon per 8b: "It is asked: since Thamna is read in Genesis … how is she numbered among the sons, when …?" ⚑ This is the one place in the batch where the English carries a mark the Latin band does not: the colon-count in band 0633B is 1 Latin against 2 English, and it is this and only this. Recorded because test 3 otherwise returns zero mismatched bands, and a reader running that test must be able to tell a ruled recast from a defect.

⚠ The other two Quaeritur sites in the range close on a semicolon, not a question mark (0637A cur haec Egla sola uxor David dicatur; · 0637B cur non in summa quadraginta anni … numerentur;), so 8b does not apply to them and no colon was introduced. 0642B's Unde quaeritur: quomodo … ? is already direct behind Migne's own colon and needed no recasting.

B12 · 0639A–B · Migne's MASCULINE pronouns for the Church, reproduced

Per eum quoque benedicuntur membra ejus, id est, Ecclesia … Cum eo est manus Domini, quia in ipsa operatur Filius Dei … Et portae inferi non praevalebunt adversus eum, nec diabolus dominabitur ei. Within four lines Migne refers to the Church by a feminine (in ipsa) and three masculines (eo, eum, ei), and the Vulgate's own adversus eam is the masculine's comparand (§C-g). All four are rendered with the gender Migne prints — "With him is the hand of the Lord, because in her the Son of God works … against him … over him." ⚑ Rendering the masculines as "it" would have been smoother and would have hidden the one thing worth seeing: the gloss has slid from the Church back to Christ mid-sentence and Migne never marks the turn. Both construals are live and neither is adopted silently; this is where the second one is recorded.

B13 · 0641D · praebuisse Israel auxilium Ituraeos, Naphaeos et Nodab — who helped whom

Rendered "that the Ituraeans, the Naphaeans and Nodab, the sons of Ismael, furnished help to Israel", taking Israel as the dative it most naturally is and the three names as the accusative subject of praebuisse. ⚑ The rival was weighed and lost: 1 Par 5:19–20 has those peoples fought against Israel, so a reader who knows the chapter will expect the opposite — but the reversal would require reading Israel as the subject and auxilium against its own verb, and a construal that exists only because it agrees with the history is the thing this edition is for refusing. Migne's sentence says what it says. Recorded here so the difficulty is visible.

B14 · Sections read clean, named so a blind reader can tell them from unread ones

CAPUT PRIMUM: 0630D–0631C entire, 0632A–0633B entire, 0633D–0634A, 0634C–0634D entire (beyond B1, B5, B9, B10 and the three §C name-markers). CAPUT II: 0635A–0635C entire, 0636A–0636D entire (beyond B2 and B3). CAPUT III: 0637A–0637D entire (beyond B9 and B11's negative finding). CAPUT IV: 0638B–0639A, 0639C–0640B, 0640D–0641C (beyond B2, B7, B9, B10 and §C-f through §C-j). CAPUT V–VIII: 0642A–0643D entire (beyond B3, B6, B7, B9, B10 and B13).

§C · THE TEN [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, which is all Pattern 14 permits.

⚑ The first four are the genealogy class the brief predicted, and they share a shape that makes them more than orthography: in each, Migne's own gloss expounds a form his own lemma does not print. That is not a spelling variant; it is a lemma and its exposition disagreeing on the page.

a · 0632C · 1 Par 1:19. Migne's lemma Phaley; Clem. Phaleg. His next four words are Phaleg enim divisio — the etymology is built on the g, and Phaley explains nothing.

b · 0633C · 1 Par 1:40. Migne's lemma Aliam; Clem. Alian. Two lines below: In Paralipomenis vero Alian, id est ejectus — Migne cites the Clementine's form correctly against his own lemma, and the whole gloss (exalted in Genesis, cast out in Paralipomenon) turns on it.

c · 0633D · 1 Par 1:41. Migne's lemma Hamran; Clem. Hamram. His next four words: In Paralipomenis Hamram, id est rubricatus. The same shape as (b), one verse later.

d · 0635A · 1 Par 1:51. Migne's lemma Dux Aliha; Clem. dux Alva, and Gn 36:40 likewise dux Alva. Here the disagreement runs the other way: Migne's gloss defends his lemma (Alva in Genesi legitur … in Paralipomenis Aliha) against a Clementine that reads Alva in both places. Marking it is what lets a reader see that his Paralipomenon is not ours.

e · 0636A · 1 Par 2:52. Migne Qui dividebat dimidium requietionum; Clem. qui videbat dimidium requietionum. ⭐ This is the batch's most consequential divergence and the marker is what makes the gloss legible: the exposition beneath it uses videre four times and dividere not once (ita dimidium requietionum vidisse · id est historiam viderunt · Vel dimidium requiei videt · Tunc … perfectionem requietionum videbit). Migne's lemma and Migne's gloss are reading two different verbs.

f · 0638D · Jo 1:29. Migne Ecce agnus Dei qui tollit peccata mundi; Clem. Ecce agnus Dei, ecce qui tollit peccatum mundi. Plural for singular. The liturgical form has displaced the Vulgate's, and number is the class 8965 fired on twice (videritis, Tangebant). The dropped ecce rides in §D as a clip.

g · 0639A · Mt 16:18. Migne portae inferi non praevalebunt adversus eum; Clem. *adversus eam, of the Church. English has an exponent for this and the divergence is a change of referent, so the marker is needed to keep "against him" from reading as our slip. See B12.

h · 0639B · 2 Rg 3:39. Migne Ego sum tener et unctus rex; Clem. Ego autem adhuc delicatus, et unctus rex. ⭐ The 7a″ trap in its purest form: the gloss exists to explain the name Recha as tener, quotes this verse to prove it, and says four words later quasi vermiculus ligni tener. The Vulgate's word is delicatus, and the argument's keyword is not in the Vulgate at all.

i · 0639C · 1 Cor 8:1. Migne quos scientia non inflat, sed charitas aedificat; Clem. Scientia inflat, caritas vero aedificat — with no negative anywhere in the verse. Migne (or his glossator) has negated an apostolic sentence to bend it into a relative clause praising humble teachers. ⛔ The negative is rendered, on its printed host inflat, exactly as the plate has it, and the marker records that the Vulgate does not carry it. Stated, not explained: Pattern 14 forbids speculating about why two witnesses differ. (The charitas/caritas orthography and the sed/vero particle are §D and are not what fired the marker.)

j · 0641A · Jo 10:9. Migne Per me si quis introierit, ingredietur et egredietur, et pascua inveniet; Clem. Per me si quis introierit, salvabitur : et ingredietur. A whole clause dropped from the middle of the verse, and "he shall be saved" is not a small one to lose.

⚑ And the boundary, because it decided one of these and refused another: an omission from the middle of a quotation fires where the dropped words are a predication of their own (Jo 10:9's salvabitur) and does not fire where they are a descriptive phrase inside a clause that survives intact (Ps 77:9's intendentes et mittentes arcum, §D). Both were weighed; the second lost because nothing in either the verse's assertion or the gloss's argument changes without it.

§D · 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.

• A particle dropped or added: 1 Par 1:11 Mesraim autem genuit for Clem. vero (0631D) · Gn 6:9 justus et perfectus for atque (0631A) · Jo 14:30 without enim (0639A) · Rom 8:14 without enim (0636D) · 1 Par 4:28 without autem (0640C) · 1 Rg 8:1 without autem (0642C) · 1 Par 3:15 and 3:9 without Clem.'s commas · 1 Ptr 2:9 Vos estis for Clem. Vos autem (0642D: a copula supplied where the Vulgate has a connective, which changes the grammar and not the assertion).
• Word order only: Jo 14:30 princeps hujus mundi (0639A) · Phlp 2:7 servi formam accipiens (0640D) · Rom 8:14 hi filii Dei sunt (0636D) · 1 Tim 3:15 oporteat te conversari in domo Dei (0643A).
• A pronoun swapped for its equivalent: Rom 8:14 hi for Clem. ii (0636D).
• A simplex for its compound, same lexeme, same sense: 2 Tim 2:19 novit Dominus qui sunt ejus for Clem. cognovit (0638A). ⚑ Weighed against §C's threshold and declined*: 8965's fired verb cases (emisit/tradidit, Cupio/desiderium habens) are lexically distinct verbs, and novit and cognovit* are one verb with and without a prefix. Recording the decline is the point of this line.
• A clip that stops early: Os 10:11 before et ego transivi · Ps 2:8 before et possessionem tuam · Is 54:2 before et pelles tabernaculorum · Ps 109:4 (whole, but introduced by Migne's etc.) · 1 Par 2:14 · 1 Par 1:16, 1:32, 2:50, 2:55, 4:9, 4:11, 4:23, 6:15, 6:31, 6:44, 6:48, 7:20, 8:6, 8:7.
• An omission from the middle of a verse, weighed and declined: Ps 77:9 without intendentes et mittentes arcum (0643C) · Jo 1:29 without the second ecce (0638D) · 1 Tim 3:15 without est (0643A) · Mt 16:18 (the Et ego dico tibi frame). Each leaves the surviving clause asserting exactly what the Vulgate's does; contrast Jo 10:9, §C-j.
• Case or orthography with nothing turning on it: Gergezaeum for Clem. Gergesaeum (0631D) · Samareum for Samaraeum (0631D) · Ephrata for Ephratha (0635A) · Jezraelitide for Jezrahelitide (0637A) · Zarae for 1 Par's Zare — and it is Genesis's spelling, Gn 36:33 (0634A) · Agarenos for Agareos (0641D) · Reginae for Regina (0643B) · In eo terra latissima et quieta for the accusative terram latissimam et quietam (0641A) · charitas/caritas (0639C) · coelo/caelo and the quae/quæ ligature class throughout.
• A truncation Migne's own etc. marks: 1 Par 5:1 violas for violasset (0641D, §B6).
• A weave rather than a quotation: Lc 10:19 at 0639B (data est ei potestas is nowhere in the verse; only the tail is) · Ps 17-style conflations do not occur in this range. [var:] declined, per style rule L3's boundary and 8965 §4.
• A citation pointing elsewhere while the WORDS agree: Non enim haeres erit filius ancillae cum filio liberae cited Gen. XXI and printed as verbatim Gal 4:30 (0637C) · Et erit in loco ubi ante dictum est … cited Osee. I and printed as Rom 9:26 (0641B). No [var:] in either — Pattern 14 is about the text, not the reference, and reference repair belongs to data/citation-corrections.json at index time, never to the running text.
• A citation whose base verse cannot be fixed: Fulvida vestimenta ejus ex Rosor (0634B, §B1). Declined — a [var:] naming a verse I cannot establish would be a public claim resting on a guess.
• Migne's own attributions passed through: (RAB.), (ID.), (HIER.), (HIERON.), (GREG.), (RAB-HIERON.), (HIERON., Quaest. in Paral., tom. III.), (HIERON. Quaest. Hebr. in Paral. ) — the last two with Migne's own internal spacing before the closing parenthesis, reproduced. Passing a siglum through is not endorsing it.

The usque ad audit — 15 occurrences, and only 4 are the formula

⛔ Decided per occurrence, from what stands on either side, per addendum D.

• Formula (own italic span, each preceded by Migne's etc.,): 4 — 0635A (Achar qui), 0637D (Filii autem Josiae), 0639D (Genuitque Mariam), 0642C (Porro Josedec). All four render as far as, and no verb was supplied to complete any of the four resumption tails.
• Ordinary "as far as / unto": 11 — 0631A usque ad flumen Thanaim and usque ad Gadara · 0632C usque ad Indicum oceanum · 0632C ab Adam usque ad Noe · 0633A ab Euphrate usque ad mare Rubrum · 0639A usque ad terminos terrae · 0640D usque ad Orientem vallis · 0641B usque ad diem hunc and usque ad finem mundi ×2 · 0643A usque ad Achimaas.
• Look-alikes in the same family, none of them the formula: usque hodie ×3 (0631A, 0634A, 0634C) · usque in praesens (0632D) · usquequo caeterae tribus sortes suas acceperunt (0636B).

⚑ The trap here is 0632C's Sicut in principio ab Adam usque ad Noe decem generationes posuit, which sits inside a [n: (RAB.)] gloss five words from an etc.-bearing neighbour and is the verse's own sense, not an abridgment. It was read as ordinary and the formula reading was not allowed to swallow it.

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

Test 1 (a [sic:] must not swallow the word it marks). Vacuously clean: zero [sic:], zero [d:], zero [ed:] and zero [cj:] in this range. Run anyway on the ten [var:], which are the only markers standing inside a sentence: each was deleted and the host sentence read back, and all ten leave a complete clause with its subject, verb and object intact (the fourth question — does the clause still have the thing it acts on? — asked and answered). All five files were then stripped of every marker mechanically and swept for a stranded article, preposition or copula sitting on punctuation: no hits beyond the id est → "that is," formula and Migne's own clipped lemma Habitaverunt in,.

Test 2 + 2a (name the host of every printed negative). All 33 printed negation tokens in the five Latin chunks were extracted mechanically, each matched to the single Latin word it attaches to, and then found on that word in the English. No defect found. The trigger population 2a names — a negative on a verb with a quantifier, correlative or predicate adjacent — was walked one by one:

• 0631D ut non Deo suam felicitatem, sed propriae virtuti ascriberent — the host is Deo, not ascriberent, and the sed propriae virtuti correlative proves it: "to ascribe their happiness not to God, but to their own strength."
• 0635D quod Sesan non habuerit filios, sed filias — the host is filios, proved by sed filias: "had not sons, but daughters."
• 0640B in Hebraeo non principes, sed mariti dicuntur — host principes, proved by sed.
• 0639C quos scientia non inflat, sed charitas aedificat — host inflat, proved by the verb pair the sed pivots on; this is the §C-i site and the negative is Migne's own addition to the verse.
• 0642A non quod eum transtulerit in Assyrios, sed quod Teglathphalnasar … cepit — host is the whole quod clause, proved by the answering sed quod.
• 0640D quia Simeon non [0640D] habuit tantam sobolem quantam Judas — ⭐ the sharpest 2a case in the batch: a negative standing before a verb with a tantam … quantam correlative one word away. The host is the quantity, not the verb — the claim is not that Simeon had no offspring but that he had not so great an offspring — and the English scopes it that way ("Simeon had not so great an offspring as Juda"), proved by the correlative and not by what reads well.
• Also checked and clean: quam aqua non posset ascendere (0631D, host posset) · haud procul (0632A, host procul) · Alii non vere … consulentes (0632B, host vere) · et non cognovit quisque fratrem suum, nec audivit (0632D, two verbs) · invenire non potui (0632D) · ne senex … arguatur (0633B) · non est hujus temporis disserere (0631B) · non sine sacramenti mysterio (0632C) · non amplius quam septem (0635A, host amplius) · non intelligebant (0636B) · non Jonathan (0635B) · ne irrideretur (0637A) · cur non … numerentur (0637B) · in regno non recte computari (0637C, host recte) · non meruerunt numerari (0637C) · consortium habere non possunt (0637C) · nisi forte (0638C) · nocere non potuit (0639A) · non praevalebunt and nec diabolus dominabitur (0639A–B) · in me non habet quidquam (0639A) · suscipere noluit (0639B) · quia nulla de eo progenies (0640C, host progenies) · Non plebs mea vos (0641B, host the predicate) · non corrigebat (0642C) · ne viderent ossa fratrum suorum (0643C).

Test 2b (each anchor against the WORD it stands beside). All 54 column anchors in the range were checked against their Latin neighbour word, not merely for presence and order. Twelve sit with no space before a VERS. address ([0631B]VERS. 7, [0631D]VERS. 10, [0632C]VERS. 17, [0633B]VERS. 32, [0634D]VERS. 50, [0636A]VERS. 54, [0637D]VERS. 15, [0638A]VERS. 18, [0640C]VERS. 28, [0642C]VERS. 15, [0642D]VERS. 44, [0643D]VERS. 6) and are reproduced without a space. The remaining 42 fall inside a sentence and each was placed against the same word in the English; five of them split a phrase to match Migne, exactly as 8995 @0605B did deliberately: Decimus [0630D] ab Adam → "the tenth [0630D] from Adam" · parit [0639A] Ecclesia → "the most holy [0639A] Church" · data [0639B] est ei potestas → "power [0639B] is given" (inside the italic lemma, as Migne sets it) · Ecclesiae [0641D] populum → "the Church's [0641D] people" · Simeon non [0640D] habuit → "Simeon had not [0640D] so great".

Test 3 (punctuation by column band, never by total). Colons, semicolons, question marks, exclamation marks and guillemets were compared band by band across all 54 anchor-partitioned bands in the five chunks. Sixteen mismatches were found and repaired, and they ran overwhelmingly one way: fourteen were semicolons or colons this translator had SUPPLIED, at 0631C, 0632A (×2), 0632D, 0633B (×2), 0634B, 0634D (×2), 0635A, 0637D, 0639D, 0640B and 0643C (×2). Two were marks of Migne's that had been DELETED and are now restored: the semicolon at 0633A (decem generationes posuit; ita a Sem) and the colon at 0640D (unde in Josue dicitur: Simeon per cognationes suas accepisse), where the English had run the clause on. Final state: zero mismatched bands except the single ruled Pattern 8b colon at 0633B (§B11). Question marks: 2 Latin, 2 English, in the same two bands. Guillemets: zero on both sides. Em-dashes in the English: zero.

⚠ No comma→semicolon sweep was performed and none was contemplated. Migne prints 42 semicolons of his own in these five chunks, so this book is nowhere near Tobit's zero-semicolon exception, and the traffic ran the other way: all but two of the repairs removed one of ours.

Test 4 (does it ASSERT what he asserts?). Every paragraph was read back against the Latin with the counts already known clean, asking only that question. It is what produced §B9 (five places rendered without a governing verb rather than given one), §B12 (the masculine pronouns for the Church left standing), §B13 (the Ituraeans helping Israel against what the chapter elsewhere says), §B10's septem filios against six names and hi quatuor under Viri quingenti, and §C-e — the dividebat/videbat divergence, which no count-based check could see, because nothing is missing from the lemma; it simply reads a different verb from the four videres that expound it.

⚑ Two habits, both applied: a sentence that reads smoothly where the Latin is rough was treated as a suspect (which is how 0638A's stacked datives and 0640B's governorless infinitive were caught), and the four abridgment tails were parsed as strictly as ordinary prose — the formula removes words, it does not suspend grammar.

§F · WHAT I DECLINED, AND WHY

1. Every [sic:]. Three non-words and one split run (§B2, §B3) would each carry one on the printed evidence alone. None fired. The 2026-08-23 addendum measures non-word suspicions at ten for ten as the digitization's error, a [sic:] is a public claim about Migne's plate, and I have no plate. ⛔ Three of the four sit at a column foot or head, and this volume's archive.org copy is known to clip the left margin (p. 320 is confirmed and carries cols 0629/0630, this batch's opening). Raised here, never patched: a conjecture entered into the Latin would wear Migne's voice.
2. Both available [cj:]. amica at 0639D and tumulentus at 0633D are real, well-formed words, so Pattern 7 renders them and Pattern 12 has nothing to wrap. §B10 gives each its rival and says which lost and why — amica because the Vulgate's own patruelem at Ex 6:20 contradicts the obvious amita, tumulentus because the faithful English does not mislead. A [cj:] on either would tell a reader the matter is settled.
3. A [var:] on Fulvida vestimenta ejus ex Rosor (§B1) — the base verse cannot be fixed.
4. A [var:] on the two mis-cited quotations whose words are correct (Gal 4:30 under Gen. XXI, Rom 9:26 under Osee. I) — Pattern 14 is about the text, not the reference.
5. A [var:] on novit/cognovit (§D) — one lexeme, no change of sense; the decline is recorded because the line between it and 8965's fired verb cases is a judgement someone may want to revisit.
6. A [var:] on Ps 77:9's dropped clause (§C-j's boundary) — a descriptive phrase, not a predication.
7. Any correction of Migne's four displaced verse numbers, of VERS 9.--, or of VERS. 12.--[n: (ID.)] (§B7). Convention 1 makes the address pass through untouched, and it does.
8. Any plate read. I have no plate access in this session. Every candidate above is stated as a candidate, with the column, the printed reading and what a single page would settle. ⚑ The cheapest next move on this work is one Gallica fetch per site — ark:/12148/bpt6k5505319w, f = (column + 1) / 2 for that copy, calibrated against the printed corner numbers before the read — which would close §B2, §B3 and §B6 in four requests.

Translated 2026-08-25, one translator, 5 chunks, 4,129 Latin words (frontmatter counts: 595 + 687 + 1,161 + 1,184 + 823 = 4,450; the 4,129 figure in the launch brief is the body-only count). English body ≈ 1.53× the Latin body.

Apparatus fired in this range: 12 [var:] · 2 [cj:] · 0 [sic:] · 0 [d:] · 0 [ed:] · 0 [nt:] · 64 [n:] (exact parity with the Latin twins, same order, same positions). Zero guillemets in either language. Zero [cn:] — this work's plate spot check is a recorded checked zero, so no numbered conjecture apparatus exists to recover, and none was invented. Question marks: 3 in the Latin, 3 in the English, in the same three bands (0648B, 0649A, 0650C). 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 and with Migne's own internal spacing ((RAB. in Paral. ), (RAB., ubi supra. )). Paralipomenon II (8965) cruces.md §M governs the vocabulary — see §0 below for where this range meets it and where it had nothing to meet.

§0 · VOCABULARY — held to 8965 §M, and the two places this range extends it

⚑ The number-names of 8965 §M are otherwise absent here. Senarius, centenarius, denarius, binarius, quadragenarius, duodenarius occur nowhere in cols 0644A–0655B, so there is no seam in this range for that convention to fail at. Recorded as a checked absence.

§1 · THE LEMMA COLLATION — every italic span in cols 0644A–0655B, checked against sources/vulgate/clementine-flat.txt

Migne's book is 1Par in that file; his chapter/verse numbering agrees with it except where noted in §4. A lemma named here with no comment AGREES with the Clementine at its own verse, verbatim, or is a clean clip of it — the agreements are named because a list of findings alone cannot be told from a report by a reader who never looked. Everything that diverges carries its divergence; the eleven that earned a public claim are argued in §2.

CAPUT IX (0644A–0645B, chunk 0005)

Nathinaei. (9:2, verbatim, the verse's last word) · Helciae filii Mosoliam, etc. (9:11, diverges, §2a) · Idithun. (9:16, verbatim) · Sellum vero filius Core. (9:19, verbatim clip) · Custodes vestibulorum. (9:19, verbatim clip of custodes vestibulorum tabernaculi) · Phinees autem. (9:20, verbatim clip) · His quatuor Levitis, etc. (9:26, verbatim clip) · De Hosa autem, id est de filiis Merari, Semri princeps, etc. (1 Par 26:10, verbatim; Clem. punctuates id est, de filiis Merari : Semri princeps — punctuation only) · De horum genere erant et super vasa minister. (9:28, diverges, §2b) · Mathathias Levites primogenitus Sellum Coritae praefectus erat eorum quae in sartagine frigebantur. (9:31, verbatim; Clem. opens Et Mathathias — a clip, not a divergence) · Panes erant propositionis. (9:32, verbatim clip of super panes erant propositionis) · Capita Levitarum per familias suas. (9:34, verbatim clip) · Laetatus sum in his quae dicta sunt mihi. (Ps 121:1, verbatim) · Jahiel, (9:35, diverges, §2c) · Abinadab. (9:39, verbatim) · Esbaal. (9:39, verbatim) · Meribbaal. (9:40, verbatim).

Etymologies, not scripture, and therefore outside Pattern 14: clamans, · calvus · pater meus Deus. · pater meus votum, · populus meus votum. · de ore ignominia. Each is rendered so that the LETTERS the gloss argues from survive — Cora / Core (Aleph against He) are kept apart, and Abinadab is kept in all three of its printed occurrences (see §3a).

CAPUT X (0645C–0647A, chunk 0006)

Philisthiim, etc. (10:1, verbatim) · Fugeruntque viri. (10:1, verbatim clip of fugeruntque viri Israel) · Cumque appropinquassent. (10:2, verbatim) · Abinadab. (10:2, verbatim) · Melchisua. (10:2, verbatim) · Et dixit Saul ad, etc. (10:4, verbatim clip) · Ad armigerum. (10:4, verbatim clip of ad armigerum suum) · Noluit autem armiger. (10:4, verbatim clip) · Quod cum vidisset armiger. (10:5, verbatim clip) · Et caput. (10:10, verbatim clip of et caput affixerunt) · Hoc cum audissent viri Jabes. (10:11, verbatim clip) · Quercum quaerat in Jabes, et jejunaverunt septem diebus, etc. (10:12, diverges, §2d) · Auferetur a vobis regnum Dei, et dabitur genti facienti fructum ejus (Mt 21:43, Clem. fructus ejus — checked and left unmarked, §4).

Etymologies: rex meus salus. · domus securitatis.

CAPUT XI–XII (0647B–0650A, chunk 0007)

Majores. (11:3, verbatim clip of majores natu) · Principes populi convenerunt in unum cum Deo Abraham (cited Psal. II, diverges, §2e) · Dixeruntque. (11:5, verbatim) · Ecce dedi vobis potestatem calcandi super serpentes, etc. (Lc 10:19, Clem. supra serpentes — checked, unmarked, §4) · Notus in Judaea Deus, et in Israel magnum nomen ejus. Et factus est in pace locus ejus. (Ps 75:2–3, an added et — checked, unmarked, §4) · Joab autem reliqua urbis exstruxit. (11:8, verbatim) · Dei adjutores sumus, Dei agri. (1 Cor 3:9, diverges, §2f) · Jesbaam. (11:11, verbatim) · Sedens in cathedra sapientissimus (2 Rg 23:8, verbatim) · Ahohites. (11:12, verbatim) · Verumtamen usque ad tres primos. (11:21, verbatim clip — see §3c and §5) · quis in nubibus aequabitur Domino? (Ps 88:7, verbatim after Clem.'s Quoniam, which Migne prints outside the italic; the ? is Migne's own and rides as printed) · Ariel Moab, (11:22, verbatim; Clem. duos ariel Moab) · Interfecit leonem. (11:22, verbatim clip) · Tempore nivis. (11:22, verbatim) · Aedificans Jerusalem Dominus, (Ps 146:2, verbatim) · Pater meus usque modo operatur, et ego operor (Jo 5:17, verbatim) · Lavabis me et super nivem dealbabor (Ps 50:9, verbatim) · Inter tres. (11:24, verbatim clip of inter tres robustos) · Hi quoque venerunt. (12:1, verbatim) · Qui erant fortissimi, etc. (12:1, verbatim) · Veloces. (12:8, verbatim) · Novissimus, etc. (12:14, verbatim) · Petrus, Jacobus et Joannes, qui videbantur columnae esse, dextras mihi dederunt, et Barnabae societatis, etc. (Gal 2:9, diverges, §2g) · Vos autem quem me esse dicitis? (Mt 16:15, verbatim; Clem. commas Vos autem, quem) · Tu es Christus Filius Dei vivi. (Mt 16:16, verbatim) · Tu es Petrus, et super hanc petram aedificabo Ecclesiam meam. (Mt 16:18, verbatim clip) · Accipite Spiritum sanctum, quorum remiseritis peccata remittuntur eis (Jo 20:22–23, verbatim) · Transierunt Jordanem. (12:15, verbatim clip) · quando inundare consuevit super, (12:15, verbatim clip stopping before ripas suas — §3d) · Et omnes fugaverunt qui morabantur (12:15, verbatim clip) · Inimicos meos dedisti mihi deorsum, et odientes me disperdidisti (Ps 17:41, diverges, §2h) · Porro de Manasse. (12:19, verbatim clip) · Deo autem gratias qui semper triumphat nos in Christo Jesu (2 Cor 2:14, verbatim) · Exercitus Dei. (12:22, verbatim clip of quasi exercitus Dei) · Ut transferrent. (12:23, verbatim) · Joiada quoque princeps de stirpe Aaron, etc. (12:27, diverges, §2i) · De filiis quoque Issachar, viri, etc. (12:32, verbatim clip) · Viginti, etc. (12:35, verbatim clip of viginti octo millia) · Tribus diebus. (12:39, verbatim clip).

Etymologies: conculcatus, · visio pacis. · sedens in populo. · aedificator Dominus (the anchor [0648C] falls INSIDE this one; see §3e).

CAPUT XIII–XV (0650B–0653A, chunk 0008)

Iniit autem. (13:1, verbatim) · Non enim. (13:3, verbatim clip of non enim requisivimus) · A Sihor. (13:5, verbatim clip of a Sihor Aegypti) · Quid nunc tibi vis in via Aegypti, ut bibas aquam turbidam (Jr 2:18, cited Jerem. VII; Clem. Et nunc quid tibi vis — word order, checked, unmarked, §4) · Cariathiarim. / Cariatharim, (13:5–6; both printed forms kept, §3f) · Omnis vir. (13:6, verbatim clip of et omnis vir Israel) · Congregavit autem David omnes electos ex Israel triginta millia (2 Rg 6:1, Clem. has rursum — checked, unmarked, §4) · Cum autem pervenissent, etc. (13:9, Clem. singular pervenisset — checked, unmarked, §4) · Mihi absit gloriari, nisi in cruce Domini nostri Jesu Christi (Gal 6:14, Clem. Mihi autem absit — checked, unmarked, §4) · Eo quod divisisset Dominus Ozam. (13:11, verbatim) · Divisio Ozae. (13:11, verbatim) · In domo Obededom. (13:14, verbatim) · Non sum missus nisi ad oves quae perierunt domus Israel (Mt 15:24, verbatim) · Caecitas ex parte contigit in Israel, etc. (Rom 11:25, verbatim) · Accepit quoque David, etc. (14:3, verbatim) · Audientes autem. (14:8, verbatim clip) · In cacumine pyrorum. (14:15, verbatim) · De Gabaon usque Gazera. (14:16, verbatim) · Divulgatumque est nomen David in universis regionibus, et Dominus, etc. (14:17, verbatim) · Tunc dixit David: Illicitum est, etc. (15:2, verbatim) · Fac opus evangelistae, ministerium tuum imple (2 Tim 4:5, Clem. opus fac — word order, checked, unmarked, §4) · Congregavitque universum Israel. (15:3, verbatim clip) · Sanctificamini cum fratribus. (15:12, verbatim clip) · Sancti estote quoniam ego sanctus sum (Lv 19:2, Clem. quia ego — checked, unmarked, §4) · Nablis videlicet, etc. (15:16, verbatim) · Lyris. (15:16, verbatim) · Laudate eum in cymbalis bene sonantibus, etc. (Ps 150:5, Clem. benesonantibus as one word — orthography, checked, unmarked, §4) · Pro octava canebant, etc. (15:21, verbatim clip) · Porro Sebenias et Josaphat, etc. (15:24, verbatim clip) · Clangebant. (15:24, verbatim) · David autem. (15:27, verbatim) · David erat in Ephod. (15:27, diverges, §2j) · Michol filia Saul. (15:29, verbatim).

Etymologies: civitas sylvarum · firmamamentum novum et turbidum, (a non-word, §3g) · flentium, · aqua omnis.

CAPUT XVI–XVII (0653B–0655B, chunk 0009)

Attulerunt igitur. (16:1, verbatim) · Benedixit populo. (16:2, verbatim clip) · Et divisit. (16:3, verbatim) · Frixam oleo. (16:3, verbatim clip of frixam oleo similam) · Sic enim Deus dilexit mundum ut Filium unigenitum daret, (Jo 3:16, Clem. Filium suum unigenitum — a dropped possessive, checked, unmarked, §4) · In illo die. (16:7, verbatim) · Constitues eos principes super omnem terram. (Ps 44:17, verbatim) · Confitemini Domino, / Confitemini, (16:8, verbatim) · Et dicat omnis. (16:36, verbatim clip of Et dicat omnis populo) · Amen, amen, dico vobis quia si duo ex vobis consenserint super terram, ex omni re quamcunque petierint, fiet illis, etc. (Mt 18:19, diverges, §2k) · In excelso. (16:39, verbatim clip of in excelso quod erat in Gabaon) · Ut benediceret. (16:43, verbatim clip of ut benediceret etiam domui suae) · Et ait Nathan. (17:2, verbatim clip) · Dimitte eam, quia animus ejus in amaritudine est, et Dominus celavit a me, etc. (4 Rg 4:27, diverges, §2l) · Non aedificabis. (17:4, verbatim clip) · Non tu aedificabis mihi templum, quia vir sanguinum es. (declared by Migne himself to be antiquae translationis — checked and deliberately unmarked, §4) · Templum Dei sanctum est quod estis vos (1 Cor 3:17, Clem. Templum enim Dei — checked, unmarked, §4) · Nec filii. (17:9, verbatim clip of nec filii iniquitatis) · Cumque impleveris. (17:11, verbatim clip) · Aedificabit (17:12, verbatim, Clem. Ipse aedificabit mihi domum) · Quoniam tu, Deus Israel, revelasti aurem servi tui, dicens: Domum aedificabo tibi (2 Rg 7:27, Clem. quia tu, Domine exercituum Deus Israel — checked, unmarked, §4) · Nisi Dominus aedificaverit domum, etc. (Ps 126:1, verbatim) · Et thronus. (17:14, verbatim clip of et thronus ejus erit firmissimus) · donavit nomen, quod est super omne nomen, ut in nomine Jesu omne genu flectatur, (Phlp 2:9–10, Clem. donavit illi nomen — checked, unmarked, §4).

Etymologies: congregans · Allelu, · ia · collis.

§2 · THE ELEVEN [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 · 0644A · 1 Par 9:11. Migne Helciae filii Mosoliam; Clem. filius Helciae, filii Mosollam. A proper-name form. Marked because this range is a name-list book and a name is a fact, not orthography — the launch brief asks for exactly this class, and the Latin column prints the form beside it so a reader can see which witness has which.

b · 0644C · 1 Par 9:28. Migne super vasa minister; Clem. super vasa ministerii. Not a spelling: Migne's reading puts a singular predicate noun against the printed plural erant, and the English is deliberately left to show it ("there were also a minister over the vessels"). ⚑ minister is a real, well-formed Latin word, so under the addendum-A rule it is Migne's and takes no [sic:]; the divergence is the whole of what is asserted.

c · 0645A · 1 Par 9:35. Migne Jahiel; Clem. Jehiel. The gloss is built on the letters — Jahiel, qui in Regum, Ahiel … Jahiel disertus Dei — so a form conformed to the Vulgate would make the exposition explain a word that is not on the page. This is 7a″ in its name-list form.

d · 0646D–0647A · 1 Par 10:12. Migne Quercum quaerat in Jabes; Clem. subter quercum, quae erat in Jabes. Rendered as printed ("Let him seek the oak in Jabes"), which is a jussive in the middle of a narrative lemma. ⚑ The obvious conjecture is that quae erat has been fused into quaerat — the Pattern 10 fused direction, which no mechanical test sees. No [sic:] was fired, because the fused direction is exactly the class our digitization has owned 19 times out of 19 in the sibling book, and a [sic:] would blame Migne's plate for a defect that is probably ours. The [var:] asserts only what the Clementine reads, which is checkable. Plate candidate — see §7.

e · 0647B · Ps 2:2. Migne, under his own note (Psal. II), prints Principes populi convenerunt in unum cum Deo Abraham. Clem. Ps 2:2 reads principes convenerunt in unum adversus Dominum. The marker gives Ps 2:2, the verse Migne cites. ⚑ Recorded and not argued in the marker: the closing words are those of Ps 46:10 (Principes populorum congregati sunt cum Deo Abraham), so the printed line is a weave of two psalms. Pattern 14 forbids speculating why two witnesses differ, so the marker states the reading Migne's own citation points at and the crux carries the rest. It matters to the gloss: the sentence is adduced to show prophets and apostles running together to Christ, which cum Deo Abraham gives and adversus Dominum destroys.

f · 0647D · 1 Cor 3:9. Migne Dei adjutores sumus, Dei agri; Clem. Dei enim sumus adjutores : Dei agricultura estis. A different predicate — "God's fields" for "God's husbandry" — in a sentence adduced for the apostles' work of cultivation.

g · 0649A · Gal 2:9. Migne Petrus, Jacobus et Joannes … dextras mihi dederunt; Clem. Jacobus, et Cephas, et Joannes … dextras dederunt mihi. A different name and a different order of names in a proof-text whose whole function here is to identify who the pillars were. ⚑ Migne's note reads (Gal. XI.) — Galatians has six chapters. Notes pass through verbatim (sacred-marker rule) and the reference was not corrected; recorded here so the indexer's unparsed bucket has an explanation waiting.

h · 0649C · Ps 17:41. Migne Inimicos meos dedisti mihi deorsum; Clem. dedisti mihi dorsum. One letter, two different assertions — "downward" against "their back". Both are real words, so this is Pattern 14 (the divergence sits inside a quoted scripture) and not Pattern 18, per 18a's explicit rule about where the divergence sits.

i · 0649D · 1 Par 12:27. Migne Joiada; Clem. Jojoda. Name form, same reasoning as a. The gloss identifies him with Phinees, so the reader needs to see which spelling the identification was made on.

j · 0653A · 1 Par 15:27. Migne David erat in Ephod; Clem. David autem etiam indutus erat ephod lineo. The marker is fired for the dropped lineo, not for the construction: the gloss immediately before it expounds a fibre de terra ortum coming ad candorem vestis — the linen is the word the exposition needs, and it is the word Migne's lemma does not print. Same shape as 8965 §3d.

k · 0654A · Mt 18:19. Migne Amen, amen, dico vobis quia si duo ex vobis consenserint; Clem. Iterum dico vobis, quia si duo ex vobis consenserint. ⭐ The gloss is built on the divergence: the whole paragraph is about the gentile saying Amen amen, id est fiat, fiat, and Migne's lemma is adduced secundum illud — a proof-text chosen for an opening that the received text does not have. Also ex omni re for Clem. de omni re, not separately marked.

l · 0654B · 4 Rg 4:27. Migne Dimitte eam, quia animus ejus in amaritudine est; Clem. Dimitte illam : anima enim ejus in amaritudine est. Fired for animus against anima: the gloss's subject is that the prophet's mind is not always irradiated (prophetarum mentes), and animus is the word that meets it.

§3 · THIS RANGE'S CRUCES — decisions, and the reasons a later reader cannot reconstruct

3a · 0645B · Abinadab … Abinadab … Abinadab — three times, two etymologies. RENDERED AS PRINTED, NO MARKER.

Abinadab vocatur et interpretatur [0645B] Abinadab pater meus votum, Abinadab populus meus votum.

Two different etymologies are attached to one printed name. The standard Hieronymian pair is Abinadab = pater meus spontaneus/votum and Aminadab = populus meus spontaneus/votum, so the second occurrence is very likely Aminadab. ⛔ Not marked, and not emended. Abinadab is a real, well-formed proper name in its own grammatical slot; addendum A refutes the corpus-frequency argument for real words 5 times out of 5, and Pattern 12 reserves [sic:] for type that cannot render in place. It renders in place. The conjecture lives here, where the rule says it lives. ⚑ Not dittography either. Pattern 11 wants a doubled run; this is a three-term list in which the second and third terms take different predicates, so marking either copy would quarantine a term the sentence needs. 11a's test (which copy is the intrusion?) has no answer here, because neither copy duplicates a neighbouring phrase.

3b · 0647C · constituisse Jacob where the sentence is about Joab. RENDERED "Jacob", NO MARKER.

Josephus's report: Super opus autem murorum constituisse Jacob, expulisse Jebusaeos et Jerosolymam vocasse. The man set over the walls is Joab, named twice in the same paragraph (Joab autem reliqua urbis exstruxit opens it, Joab, qui reliqua urbis exstruxit closes it). ⛔ A familiar English Bible spelling must not arrive over the plate, and neither may a familiar name from four lines away. Jacob is a real, well-formed name occupying its slot; it is Migne's until a plate says otherwise. Rendered "Jacob", logged, no marker. Plate candidate — §7. ⚑ It was weighed against Pattern 18 and declined: 18's bar is a sentence that says something the author is arguing against, and a wrong name in a report of Josephus is a wrong fact, not a reversed argument. 18a is explicit that mood/person/number anomalies that merely fail to construe stay declined, and a name substitution is no stronger.

3c · 0648A · Verumtamen usque ad tres primos. — the clip was NOT completed with a negative.

Migne's lemma stops where 11:21 continues non pervenerat. Ruth's convention 3 renders a fragment as the phrase it stands for, and the gloss's own next sentence supplies the negative (Sed ad tres primos non pervenit), so "Nevertheless he had not attained to the three first" was available and reads better. ⛔ Declined, and the reason is 7a: no negation appears in the English that the Latin does not print. Ruth's worked case (Non te. → "Let no man see thee") expands a fragment whose negative is itself printed; here the negative is precisely the word Migne stops before. Rendered "Nevertheless unto the three first." — a readable clip, with the negative arriving one sentence later exactly where Migne prints it. The rival reading is named here so that a blind reader meets it, and it lost on the rule, not on taste.

3d · 0649A · quando inundare consuevit super, — the clip stops on a preposition.

12:15 continues super ripas suas. Rendered "when it is wont to overflow above," — the comma is Migne's and is carried, and "its banks" was not supplied. Same reasoning as 3c: convention 3 licenses rendering a fragment as its phrase, it does not license importing the Vulgate's next two words into an italic span that stops before them. The Latin twin keeps the clip findable.

3e · 0648C · the column anchor falls INSIDE an etymology, and the English splits there too.

Migne prints qui aedificator [0648C] Dominus interpretatur. English word order would put "the Lord the builder", which moves the anchor off its word. Rendered builder [0648C] Lord — an apposition that reads as one, with the anchor standing beside the same two words it stands between in the Latin. Splitting mid-construction to match Migne is the 8995 precedent and is deliberate, not an oversight. ⚑ One other anchor is not word-adjacent in the same way and is recorded: 0652B, where Migne has Haec Ecclesiam [0652B] significat and the English reads "This signifies the Church [0652B], which…". The anchor sits against Ecclesiam/"the Church" in both; English cannot put it before "signifies" without inverting the clause. All 46 anchors in this range were checked against the word they stand beside, not merely for presence and order (test 2b).

3f · 0650C · Cariathiarim and Cariatharim — both printed forms kept.

Migne spells the city two ways in one paragraph, and glosses each separately. Neither was normalized to the other or to the Clementine's Cariathiarim, because the second gloss (quae alibi Cariathabaal) is doing different work from the first. ⚑ Recorded and not marked: the same paragraph makes Cariathiarim civitas Damasci … quae in Amos Emathabal, which is a confusion of Cariathiarim with Emath (Amos 6:2). It is Migne's, it is renderable, and it is rendered.

3g · Four NON-WORDS and one suspicious form — ALL rendered for their evident sense, NO [sic:] FIRED

Following 8965 §2c–2d, which is the sibling book's settled practice, and addendum B's ship test: a [sic:] is a public claim about the plate, and seven false ones were withdrawn from an earlier batch.

⚑ None of these fired a marker, so this range ships asserting nothing about Migne's plate.

3h · 0651A · significat. quam vult Christus — Migne's full stop inside the clause, PRESERVED.

Allegorice autem David Christum, arca Ecclesiam significat. quam vult Christus in civitatem suam adducere

The stop breaks the relative clause off from its antecedent and the next word is lowercase. Pattern 8: never delete a mark he prints. The English reads "…the ark the Church. which Christ wills to bring into his own city" and looks wrong on the page, which is the point. ⚑ Punctuation was compared band by band, not by total (test 3). That comparison returned 30 mismatched bands on the first pass — 26 semicolons and 4 colons, every one of them mine, including a place at 0647C where two of Migne's sentences (…circumdedisse. Super opus autem…) had been merged behind a semicolon. All 30 were repaired to the printed mark; the range now returns zero mismatched bands for colons, semicolons, question marks and guillemets. ⛔ The totals would have hidden a good share of them, exactly as the brief warns.

3i · Three sentences Migne leaves WITHOUT a terminal stop — reproduced without one.

0650C (sicut superius ostenditur) · 0648A (et Semeiam filium Heli) · 0649A (sed etiam coram omnibus hominibus). Pattern 8: never supply a mark, including a terminal one.

3j · 0648C–0649A · the verse addresses run OUT OF SEQUENCE, and are reproduced in Migne's order.

CAPUT XII opens VERS. 12, then VERS. 8, then VERS. 14. Migne's ordering is reproduced as printed and is not reordered (the Jude precedent in the brief). Recorded so a reader does not read it as our chunking error. ⚑ Two further verse addresses point at a different verse from the one the words come from: VERS. 17 at 0651A glosses in domo Obededom, which is 13:14; VERS. 12 at 0655A glosses Et thronus, which is 17:14. Both are Migne's; both ride as printed.

3k · 0646C · Videntes indocti … ignorantes … — two construals, and which one the English took.

Videntes indocti verae humilitatis medicamina lavacrumque baptismi, ignorantes in Judaea magistros suos certamine spirituali victos…

Either participle can govern the accusatives. Taken: ignorantes governs the medicines and the laver, videntes governs the defeat of the masters — "The unlearned, ignorant of the medicines of true humility and of the laver of baptism, seeing in Judaea their own masters overcome…". Rejected: videntes governs the medicines and ignorantes the defeat, which would make them unaware of a ruin that the same sentence then says troubled them (majorum ruina turbati). ⚑ Both construe; the second is self-defeating, and that is the whole of the reason. Named because a blind reader will stop here.

3l · Unexpressed subjects supplied, and what fixed each one

Every "he/she/they" supplied where the Latin prints none:

• 0645C Ad David se transtulit → the writer of Paralipomenon, the subject of commemorat and demonstrat in the two clauses before it.
• 0645C Quo imminente … pugnant → quo is the true David, Christ, named in the preceding relative clause (qui typum tenet veri David).
• 0645D vetuerunt → the gentiles, subject of Irruerunt and corruperunt in the same sentence.
• 0653A David hunc humiliatum despicit → Michol, the lemma above (Michol filia Saul) — this is Ruth's 0538C class exactly, where the subject is in the lemma and not in the sentence.
• 0654B despexit eum saltantem → Michol again, named two words earlier (Michol filiam Saul, quam increpabat David).
• 0649A Hic significant → left plural in English ("Here they signify"), against the singular Hic, because doctores is the plural the verb agrees with. A number disagreement with no English exponent on Hic; Pattern 9, no marker, recorded.

3m · 0655A · dictum sit per Nathan: Et nuntiabit tibi Dominus — a Septuagint quotation Migne sets ROMAN.

Migne italicizes the 2 Rg 7:27 quotation that follows but leaves this one unitalicized. Italic structure is 1:1 with the plate, so the English leaves it unitalicized too; no italic span was invented. The same is true of Zacharias de vice Abia at 0644B.

§4 · DIVERGENCES CHECKED AND DELIBERATELY LEFT UNMARKED

Recorded because Pattern 14 fires on divergence that changes the sense, not on every difference, and because a decline that is not written down is indistinguishable from a lemma nobody read.

• A dropped or added particle, with nothing turning on it: Ps 75:2–3 with an added et (0647C) · 2 Rg 6:1 without rursum (0650D) · Gal 6:14 without autem (0650C) · Lv 19:2 quoniam for quia (0651C) · 1 Cor 3:17 without enim (0654C) · Phlp 2:9 without illi (0655A) · Jo 3:16 without suum (0653C) · Mt 18:19 ex omni re for de omni re (0654A, where the marker is fired for the opening instead).
• Word order only: 2 Tim 4:5 Fac opus for opus fac (0651C) · Jr 2:18 Quid nunc tibi vis for Et nunc quid tibi vis (0650B).
• A preposition or orthography: Lc 10:19 super serpentes for supra serpentes (0647B) · Ps 150:5 bene sonantibus split for benesonantibus (0652B) · the coelum/caelum and quae/quæ classes throughout.
• Number, with nothing built on it: Mt 21:43 fructum ejus for fructus ejus (0646D) · 1 Par 13:9 Cum autem pervenissent for pervenisset (0650C). ⚑ Both were weighed against 8965 §3f, where a number divergence was marked because the gloss expounded the plural. Neither gloss here touches the number: 0646D abridges into prose about the nations, and 0650C is abridged away by usque ad in the next four words.
• A dropped title in a vocative: 2 Rg 7:27 without Domine exercituum (0655A). Weighed against 8965 §3d (a dropped intrinsecus that the gloss needed) and declined — nothing in this gloss is about the Lord of hosts.
• A reading Migne himself labels as another translation: 0654C prints Non tu aedificabis mihi templum, quia vir sanguinum es and introduces it juxta antiquae translationis seriem. The page already tells the reader it is not the Vulgate, and a [var:] would be our voice saying what his own sentence has said. Declined, on the same footing as 8965's "Migne's own apparatus settles it".
• Migne's own references, passed through verbatim and not corrected: (Gal. XI.) at 0649A for Gal 2:9 · (II Cor. I) at 0649C for 2 Cor 2:14 · (Jerem. VII) at 0650B for Jr 2:18 · (Psal. II) at 0647B for a line woven from Ps 2 and Ps 46 · psal. LXXV in running prose at 0647C, rendered "Psalm LXXV" under Pattern 2. Pattern 14 is about the text, not the reference.
• Abbreviations expanded in running prose, which is why two period counts differ by one: in cap. XXVI → "in Chapter XXVI" (0644C) and psal. LXXV → "Psalm LXXV" (0647C). Both are Pattern 2 locators in prose, not [n:] contents. These are the only two band-level full-stop differences in the range, and both were traced to the abbreviation, not to a lost or merged sentence.

§5 · usque — CENSUSED PER OCCURRENCE. 13 tokens, 9 formula, 4 NOT.

The brief's warning is live in this range and the split is close to even.

The abridgment formula (9), each in its own italic span answering an italic usque ad in the Latin, rendered as far as and with NO verb ever supplied to complete the tail: 0646A (Beda, Ad armigerum) · 0646D (Rab., Et caput) · 0646D (Rab., Hoc cum audissent) · 0648A (Greg. hom. 34) · 0648B (Rab., Quem liber Regum) · 0650C (Rab. ex Hieron.) · 0651B (Rab., In cacumine pyrorum) · 0653A (Greg. Moral.) · 0653A (Rab., Michol). NOT the formula (4), and each decided from what stands on either side:

1. 0648A · Verumtamen usque ad tres primos. — 11:21's own words, inside the lemma. This is the 8969 usque ad finem fortiter pertingit trap exactly, and it is the one that would have done real damage: reading it as the formula would have swallowed the lemma and left the gloss glossing nothing.
2. 0648C · Pater meus usque modo operatur — Jo 5:17's own words.
3. 0648C · Huc usque enumeravit fortes David — ordinary adverb, "thus far".
4. 0651C · De Gabaon usque Gazera. — 14:16's own words, inside the lemma.

§6 · CONSISTENCY WITH THE PEER'S CHUNKS 0000–0004 — checked at the files, not at a table

The peer's five chunks were on disk when this range was filed and were read for the recurring classes. Followed, not re-derived:

• The interpretatur formula. The peer renders it "X is interpreted y" with the etymology inside the printed italic span (0636D Jabes is interpreted drying up; 0638B Phares, who is interpreted division). This range uses the same formula at all fourteen of its etymologies. ⚑ Where the peer's Latin prints the etymology unitalicized, the peer leaves it unitalicized ("Jethra is interpreted a remnant", 0635C) — italic parity, not house style — and this range does the same at 0645B (Ignis maritus) and 0654B.
• Proper names form for form. The peer keeps Philisthiim where Migne writes it and Philistines where he writes Philisthaei/Philisthinos; this range does the same (0645C Philisthiim, 0646D "the Philistines"). Gabaonites likewise.
• doctores → teachers. Zero occurrences of "doctors" in the peer's range and zero in this one.
• VERS. n.-- verbatim, ## CAPUT N. → ## CHAPTER N. — identical treatment.
• The [var:] habit. The peer fires them with the witness named first ([var: Vulg. has salvabitur …]); this range opens the same way and adds the book chapter:verse, which is the 8965 form. This is the one place the two stints differ in presentation and it is a merge decision, not a defect — flagged for the merge rather than silently harmonized, because rewriting the peer's markers would destroy the record of what the peer checked.

§7 · PLATE CANDIDATES — raised, never patched, and NONE fires a marker

Our archive.org copy of PL 113 (patrologiaecurs04migngoog) is known to clip the left margin on some pages, and every candidate below sits at or near a column head. That is a fact about our photograph, not about Migne's plate, and the brief forbids patching one into the text: a conjecture entered as a patch wears Migne's voice.

⭐ All four are settleable in one HTTP request each from the second witness recorded in data/volumes.json — Gallica ark:/12148/bpt6k5505319w, IIIF, server-side crops, map f = (column + 1) / 2 (NOT the archive map (column + 11) / 2), calibrated against the printed corner numbers before any read. None of them blocks this range.

⚑ 0652D is raised for the OTHER direction — the normalization class (brief §4 and addendum C). The sentence is rough throughout (ut … inspiretur … proferant … emittere), and in the middle of it est sits with no clause to be the verb of. It is the shape of a form that reads suspiciously tidy in a passage that is otherwise rough, or of a word the digitization supplied. Rendered as printed ("and on a brazen, that is, a stable foundation of faith and of work, to send forth…"), logged, no marker. ⭐ And when 0653A is read for lignum, read the whole line — 0652D is on the facing half of the same plate.

⛔ No [sic:] anywhere in this range. Every candidate above is either a real word (Migne's by the addendum-A rule) or a non-word of the class that has been our digitization's 19 times out of 19. The page therefore asserts nothing about Migne's plate that a plate has not licensed.

§8 · THE TWO [cj:], AND WHY EACH CLEARS PATTERN 18'S BAR

Both open on a word Migne did print, in italics, so the verbatim-in-the-Latin-twin guard is satisfied and no emendation of our own invention can be smuggled in.

a · 0653A · lignum → for wood [cj: lignum; read linum, "flax"] sprung from the earth.

lignum namque de terra ortum multiplici labore ad candorem vestis perveniens, veritatem humanae carnis inter flagella triumphantis ostendit, et hoc est: David erat in Ephod.

lignum is a perfectly good Latin word, so Pattern 7 renders it and Pattern 12 has nothing to wrap — the founding shape of Pattern 18. The English then asserts that wood, sprung from the earth, is brought by manifold labour to the whiteness of a garment, which is false and is not what the passage argues. Everything in the sentence is flax: de terra ortum (the plant), multiplici labore (retting, hackling, spinning, bleaching), ad candorem vestis (the white linen), and the lemma it explains is the ephod, which 15:27 makes stola byssina and ephod lineo. The commonplace behind it is Bede's, on the linen of the ephod. ⚑ The rival was weighed: leave it unglossed and put the conjecture in the crux only. That is precisely the 11081 @0030B failure Pattern 18 was written to close — munde for mundo, a real word, a crux nobody opens, and a page asserting the wrong thing. It lost on that precedent. ⚠ If the plate reads linum, the defect is our digitization's and the [cj:] should become a silent correction plus a crux line. The marker is honest either way — it claims only that the word printed in our Latin column should be read linum — but the plate read in §7 would settle which.

b · 0654D · aedificat → A man of blood [0654D] builds [cj: aedificat; read non aedificat, "does not build"] the temple of God.

Templum Domini aedificat, qui corrigendis proximorum mentibus vacat. … Vir sanguinum templum Dei aedificat. Qui enim adhuc carnalibus incubat, necesse est ut instruere proximorum mentes erubescat.

The paragraph's whole argument is that a man must be clean in himself before he corrects others (Debet enim in se mundus esse, qui aliena corrigere vult), that a muddy hand cannot wipe away filth, and that David is refused the temple because he is a man of blood (Non tu aedificabis mihi templum, quia vir sanguinum es). The sentence as printed reverses that, and the very next sentence explains why it must be a denial — for he who still lies among carnal things must needs blush to instruct the minds of his neighbours. 7a's own tell: the next clause denies what the English just asserted, and the contradicting neighbour is on the same plate. ⚑ This is 18a's admitted class. The bar as 18a restates it is whether the English asserts something false, and the ✅ list there includes a dropped word glossed on the word Migne did print (7561 @0648, solum for non solum). A lost negation is named in 18a as the defect class the whole 7a⁗ apparatus exists for, and it has nowhere else to live where there is no [sic:]. ⚑ The rivals were weighed and are named, because naming a rival is not adjudicating between them. (i) Vir sanguinum as a rhetorical question — declined, Migne prints a full stop and Pattern 8 forbids supplying a mark. (ii) aedificat read as a concessive ("even a man of blood builds") — declined, there is no concessive particle and the next sentence's enim explains a denial, not a concession. (iii) Leave it, log it — declined on 18a's ruling that incoherence is not a defence where the word that went wrong is a negation.

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

• Test 1 (a [sic:] must not swallow its word). Vacuous here — zero [sic:]. The two [cj:] were nevertheless stripped and re-read: "for wood sprung from the earth…" and "A man of blood builds the temple of God." Both read aloud without the bracket, both keep their verb and their object.
• Test 2 + 2a (name the host of every printed negative). 34 printed negatives in the range (non ×24, nec, nisi ×3, nemo, nulli, ne ×4, noluit/nollent, parum, Absit, sine). Each was matched to the single Latin word it attaches to and then found on that word in the English. Zero supplied: a reverse scan of the English for not/no/nor/never/none/without returned only hosts that answer a printed Latin negative, plus the one inside the 0654D [cj:], which is labelled as ours. ⚑ The 2a trigger population fired six times and each host was proved from structure, not from what reads well: 0651D Non [0651D] solum David, sed etiam omnes prophetae (the negative is on solum, i.e. on David, and the sed etiam pivot proves it — the anchor falls between Non and solum, so the English had to split there too) · 0649A non tantum coram Deo, sed etiam coram omnibus hominibus · 0648C non segnes, sed studiosos · 0651B non conjuges, sed concubinae · 0653A non uxor David, sed Saul filia · 0654B prophetiae spiritus non semper … irradiat, where the negative rests on semper and not on the verb — "does not always irradiate", not "never irradiates".
• Test 2b (each anchor against the WORD it stands beside). All 46 anchors in the five chunks printed with 60 characters of Latin and of English on each side and compared one by one. All 46 stand beside the same words. Two are recorded in §3e as deliberate splits.
• Test 3 (punctuation band by band, never by total). 30 mismatched bands on the first pass, all mine — see §3h. Now zero, for colons, semicolons, question marks and guillemets, in every band of every chunk. Full stops were compared the same way and returned two bands differing by one, both traced to an expanded Latin abbreviation (§4).
• Test 4 (does it ASSERT what he asserts?). Run last, paragraph by paragraph, with the counts already clean. It is what produced §8b — a sentence that contained everything, in order, at parity, and said the reverse of the paragraph it sits in — and §3k, where a smooth reading and a self-defeating one both construed.
• Marker parity, mechanically. Column anchors 46/46 verbatim and in order; [n: …] 64/64 with identical contents and positions; italic spans 1:1 once the [var:]/[cj:] contents are discounted (0005 +3, 0006 +1, 0007 +5, 0008 +3, 0009 +4 — each accounted for by a marker); guillemets 0/0. node scripts/verify-english.mjs 8964 reports no failure in 0005–0009 (its only complaints are the third stint's still-missing 0010–0013).

§10 · WHAT WAS DECLINED, AND WHY

• Every [sic:]. Six plate candidates (§7) and not one marker. A [sic:] is a public claim that Migne's plate is defective, and none of these is licensed by a plate.
• The negative in the 11:21 lemma (§3c) — a better-reading English lost to 7a.
• A [cj:] on Jacob at 0647C (§3b) — a wrong fact in a report of Josephus is not a reversed argument, and 18a keeps that class declined.
• A [var:] on the antiquae translationis line at 0654C (§4) — Migne's own sentence already tells the reader.
• A [var:] on the two number divergences (§4) — no gloss is built on either number, which is the test 8965 §3f actually applied.
• Any patch to data/tei-patches/8964.json. Four of the six plate candidates would be patches if a plate confirmed them, and none was written: our Latin is not to be corrected on an argument, however good.

Translated 2026-08-25, one translator, 4 chunks, 3,570 Latin words (frontmatter counts: 981 + 947 + 1,169 + 473 = 3,570; body words 990 + 954 + 1,179 + 479). English body ≈ 1.47× the Latin body. This is the tail of the book — CAPUT XVIII–XXIX, David's wars, the census and its plague, the temple preparations, and the Levitical, priestly and military courses.

Apparatus fired in this range: 12 [var:] · 1 [cj:] · 0 [sic:] · 0 [d:] · 0 [ed:] · 0 [nt:] · 63 [n:] (exact parity with the Latin twins, same order, same positions). Zero guillemets in either language. Zero [cn:] — this volume's plate spot check is a recorded checked zero, so there is no numbered Forte apparatus to recover and none was invented. Question marks: 2 in the Latin, 2 in the English, in the same two bands (0658B, 0662D). Em-dashes in the translated prose: zero (the three in this range's files are all inside [var:]/[cj:] content, which is our own editorial voice, not Migne's page). verify-english.mjs clean over the whole work.

Ruth (8968) cruces.md §1 governs and was not re-opened. Paralipomenon II (8965) cruces.md §M governs the vocabulary; the two peer stints on this book (cruces-0000.md, cruces-0005.md) were read on disk before this file was closed and their choices were followed — see §0.

§0 · VOCABULARY — held to 8965 §M and to the two peer stints on this book

Proper names are kept form for form as Migne prints them, never normalized to a familiar English Bible spelling. In this range: Adarezer · Soba · Tebath (Clem. Thebath) · Chun · Syna Sobal · Achimelech (Clem. 1 Par. Ahimelech) · Phelethi and the Cerethites / Phelethites of the gloss · Melchon (Clem. Melchom) · Ethai the Hethite · Araphan · Orpha · Noemi · Areuna · Ornam (Clem. Ornan, twice) · Joiarib (Clem. Jojarib) · Meselemia · Obededom · Sephim · Hosa · Asabias · Jeria / Jerias · Hebrun at 0661C beside Hebron four words later (Migne spells it both ways in one sentence; both reproduced) · Azmoth · Adiel · Abdi (Clem. Adli) · Balanam · Joas · Ubil the Ismaelite (Clem. Ismahelites) · Jadias · Absalon (not Absalom) · Judas Scarioth (not Iscariot) · Joiada (Clem. Jojada) · Banaias · Abiathar · Jesus where Migne writes it for Josue at 0664A (8965 §M's own case) · Philisthiim is not in this range, but Philisthaeos / Philisthaeis are, and follow the peers' Philistines.

⚑ Three drafting errors of exactly the conformation class were caught and reversed on the fourth test, and they are named because they are the class this apparatus exists to stop: Absalon had been drafted "Absalom", Judas Scarioth "Judas Iscariot", and In Hebrun "In Hebron". All three were the familiar English arriving over the plate.

§1 · THE LEMMA COLLATION — every italic span in cols 0655C–0664D, checked in sources/vulgate/clementine-flat.txt

Migne's book is 1Par in that file; his chapter and verse numbering agrees with it everywhere in this range. A lemma named here with no comment AGREES with the Clementine at its own verse, verbatim, or is a clean clip of it — the agreements are named because a list of findings alone cannot be told from a report by a reader who never looked. Everything that diverges carries its divergence; eleven of the twelve that earned a public claim are argued in §2 — the twelfth is not a lemma at all and is argued in §3a: it sits on the unitalicized quotation of 2 Reg. 24:9 at 0657A, which is why it appears in no line of the collation below.

CAPUT XVIII (0655C–0656B, chunk 0010)

Et filias. (18:1, verbatim clip of et filias ejus) · usque ad (formula) · Adarezer. (18:3, verbatim) · gladio spiritus, quod est verbum Dei (Eph. 6:17 — Clem. gladium spiritus (quod est verbum Dei); Migne's ablative is the gloss's own construction with confligit, case only, no marker, §4a) · Subnervavitque. (18:4, verbatim) · In Damasco. (18:6, verbatim) · sanguinis poculum · sublimis · humecta. (etymologies, not scripture — outside Pattern 14) · Pharetras. (18:7, verbatim clip of pharetras aureas) · Salomon. (18:8, verbatim) · parvulus venit, · pulchritudo. (etymologies of Tebath and Chun) · In valle Salinarum. (18:12, verbatim) · Fecit sibi David nomen cum reverteretur capta Syria in valle Salinarum, caesis decem et octo millibus (2 Reg. 8:13 — Clem. Fecit quoque sibi David nomen; a dropped particle only, §4b. ⚑ Note that the flat file itself prints revereteretur here, a defect in our copy of the Clementine, not in Migne, who prints reverteretur correctly; sources/vulgate/README.md warns that a call turning on one letter must say so, and this one is said) · Decem. (18:12, a clip of decem et octo millia) · Qui vertit linguam suam, incidit in malum (Prov. 17:20 — [var:] FIRED, §2a) · Achimelech. (18:16 — Clem. Ahimelech; spelling only, §4c, but see §3a for the gloss's Abimelech) · Phelethi. (18:17, verbatim) · Porro. (18:17, verbatim clip) · genus electum, regale sacerdotium (1 Petr. 2:9, verbatim clip) · Primi ad manum. (18:17, verbatim clip of primi ad manum regis).

CAPUT XIX–XX (0656B–0657A, chunk 0010)

Ammon. (19:1, verbatim) · usque ad (formula) · vultus est super facientes mala (Ps. 33:17 — [var:] FIRED, §2b) · Congregavit. (20:1, verbatim) · qui operatur in nobis velle et perficere (Phil. 2:13 — [var:] FIRED, §2c) · Benedices coronae anni benignitatis tuae (Ps. 64:12, verbatim, though Migne's note cites Deut. VII — §6a) · Coronam, (20:2, verbatim clip) · Bellum. · usque ad (formula) · In Deo faciemus virtutem, (Ps. 59:14, verbatim; also Ps. 107:14, and Migne's Psal. LIX is right) · Aliud bellum. (20:5 — Clem. Aliud quoque bellum, §4b) · Tertium fuit in Gob contra Philisthaeos in quo percussit Adeodatus filius Saltus, polymitarius, Bethlehemites Goliath Gethaeum, (2 Reg. 21:19 — Clem. Tertium quoque fuit bellum in Gob; §4d) · usque ad (formula).

CAPUT XXI (0657A–0657D, chunk 0010)

Et inventus est, (21:5, verbatim) · Et percussit. (21:7, verbatim) · Peccavi nimis, (21:8, verbatim) · usque ad (formula) · Et dixit. (21:13, verbatim) · Septuaginta millia virorum. (21:14, verbatim — and it agrees with 2 Reg. 24:15 too, so the gloss's own arithmetic is not at fault here) · Exstrueretque altare Domino Deo in area Ornam Jebusaei. (21:18 — Clem. area Ornan Jebusaei; proper-name spelling, §4c) · Dedit ergo David Ornam pro loco siclos auri justissimi ponderis sexcentos, et aedificavit ibi altare Domino, (21:25 + 21:26 run together, verbatim apart from the same Ornam/Ornan) · usque ad (formula).

CAPUT XXII (0657D–0658D, chunk 0011)

Ut congregarentur. (22:2, verbatim) · Propter hoc dolavi in prophetis, et occidi eos in verbis oris mei (Osee 6:5 — Clem. dolavi in prophetis ; occidi eos, without et; §4b) · Ferrum. (22:3, verbatim clip) · Confige timore tuo carnes meas (Ps. 118:120, verbatim clip) · Jerusalem, quae aedificatur ut civitas cujus participatio ejus, (Ps. 121:3, verbatim; Migne's Ibid. CXXII is off by one against Clementine numbering — §6a) · Aeris. (22:3, verbatim) · Ligna quoque. (22:4, verbatim) · Sidonii et Tyrii. (22:4, verbatim) · Parvulus est et delicatus. (22:5, verbatim) · Praeparavit impensas. (22:5, verbatim) · O Israel quid Dominus Deus tuus petit a te? nisi ut timeas Dominum Deum tuum, ut ambules in viis ejus, (Deut. 10:12 — [var:] FIRED, §2d) · Multum. (22:8, verbatim clip of Multum sanguinem) · Jerusalem, Jerusalem, quae occidis prophetas, et lapidas eos qui ad te missi sunt (Matth. 23:37, verbatim) · Auri talenta centum millia, et argenti mille millia talentorum. (22:14, verbatim) · Talentorum. · Artifices, latomos, et caementarios, (22:15, verbatim clip) · Praecepit quoque David cunctis principibus Israel ut adjuvarent Salomonem, (22:17, verbatim clip of … Salomonem filium suum) · Adorate Dominum in atrio sancto ejus (Ps. 28:2, verbatim; Migne's note reads Psal. XVIII — §6a) · Sanctuarium Domino Deo, ut introducatur arca foederis Domini, et vasa Domino consecrata in domum, quae aedificatur, (22:19, verbatim clip) · Anima justi sedes est sapientiae (cited Sap. VII — NOT IN THE CLEMENTINE AT ALL, §3d).

CAPUT XXIII (0658D–0659D, chunk 0011)

Igitur David, (23:1, verbatim) · Ubi venit plenitudo temporis, misit Deus Filium suum (Gal. 4:4 — Clem. opens At ubi venit; a head-clip, §4b) · Numeratique. (23:3, verbatim) · Inventa sunt triginta octo millia virorum. (23:3, verbatim) · Viginti quatuor. (23:4, verbatim clip of viginti quatuor millia) · Praepositorum. (23:4, verbatim) · usque ad (formula) · Dixit enim David: Requiem dedit Dominus Deus, (23:25, verbatim clip) · usque ad (formula) · Et erunt sub manu filiorum Aaron, in cultum domus, (23:28, verbatim clip) · Exedris et in loco purificationis. (23:28, a clip of et in exedris, et in loco purificationis) · Ad ferventem similam, et ad torrentem, (23:29 — [var:] FIRED, §2e) · Ad ferventem similam. · ad ferventem, (Migne's own two reported codex readings, §2e) · ad torrendum. (the Clementine's own reading at 23:29**, and the gloss ascribes it to the Hebrew — a rare case of the gloss's textual criticism landing exactly on the received text).

CAPUT XXIV–XXV (0659D–0660B, chunk 0011)

Sacerdotioque functus. (24:2, verbatim) · usque ad (formula) · Per familias sedecim. (24:4, verbatim) · Descripsitque. (24:6, verbatim) · usque ad (formula) · Exivit autem sors. (24:7, verbatim) · Prima Joiarib, secunda, (24:7 — Clem. sors prima Jojarib, secunda Jedei; i/j orthography and a clip, §4c) · Igitur David et magistratus exercitus segregaverunt in ministerium, (25:1, verbatim clip) · usque ad (formula) · Cuncti doctores. (25:7, verbatim clip of cuncti doctores, ducenti octoginta octo — and the clip is exactly where the gloss's arithmetic goes wrong; §3c).

CAPUT XXVI (0660B–0661D, chunk 0012)

Meselemiae. (26:2, verbatim) · reddente Domino, (etymology) · Zachariae. (26:2 prints the nominative Zacharias; Migne's genitive lemma follows Filii Meselemiae : Zacharias, §4e) · orantibus apostolis et jejunantibus, ait Spiritus sanctus: Segregate mihi et Paulum et Barnabam (Act. 13:2 — [var:] FIRED, §2f) · Obededom. (26:4, verbatim) · Omnium servum me feci, ut omnes lucrifacerem (1 Cor. 9:19 — [var:] FIRED, §2g) · Ad austrum. (26:15, verbatim) · Ad occidentem, juxta portam, (26:16, verbatim) · abundabit iniquitas, et refrigescet charitas multorum (Matth. 24:12 — [var:] FIRED, §2h) · Clausi sunt sermones usque ad praefinitum tempus (Dan. 12:9 — Clem. clausi sunt signatique sermones; §4b. ⚠ This is the usque ad look-alike: the words are the verse's own, §5) · Levitae sex, et ad aquilonem quatuor per diem, (26:17, verbatim clip) · Bini et bini. (26:17, verbatim) · Filii Moysi, (26:24, verbatim) · Fratres quoque ejus Eliezer. (26:25, verbatim) · Non nobis, Domine, non nobis, sed nomini tuo da gloriam (Ps. 113:9, verbatim, and Migne's Psal. CXIII is right) · Isaaritis vero praeerat Chonenias, (26:29, verbatim) · Israeli trans Jordanem contra occidentem, (26:30, verbatim clip) · faciente domino, (etymology of Asabias; on the lowercase domino see §3f) · Omnia opera nostra operatus es in nobis (Isa. 26:12, cited Psal. XXVI — [var:] FIRED, §2i) · Hebronitarum autem princeps fuit Jeria secundum familias et cognationes eorum. (26:31, verbatim) · initium sapientiae timor Domini (Ps. 110:10, verbatim, note right) · qui enim timet, proficit ut amet, et foras mittat timorem (cited I Joan. IV — the span is the gloss's own sentence with a tail echoing 1 Jo. 4:18 perfecta caritas foras mittit timorem; not a quotation, so no [var:], §4f).

CAPUT XXVII (0661D–0663B, chunk 0012)

Non est relatus in fastos. (27:24, verbatim clip of … in fastos regis David) · Super thesauros, (27:25, verbatim clip) · Melior est patiens viro forti (Prov. 16:32, verbatim clip, note right) · His autem thesauris. (27:25, verbatim) · Ezri filius Chelub. (27:26, verbatim) · Oliveta et ficeta, quae erant in campestribus, (27:28, verbatim) · Balanam Gederites. (27:28, verbatim) · Saphat. (27:29, verbatim) · judicavi, (etymology) · Sedebitis et vos super duodecim sedes, judicantes duodecim tribus Israel (Matth. 19:28 — Clem. super sedes duodecim; word order only, §4g) · Super camelos vero Ubil Ismaelites, et super asinos Jadias. (27:30 — Clem. Ismahelites, and Migne clips Meronathites; §4c) · Si quis videtur inter vos sapiens esse, stultus fiat (1 Cor. 3:18, verbatim with two internal clips) · Jonathan. (27:32, verbatim) · usque ad (formula) · Super quem requiescet Spiritus meus, nisi super humilem et quietum et trementem sermones meos? (Isa. 66:2 — [var:] FIRED, §2j) · Achitophel (27:33, verbatim) · Chusai. (27:33, verbatim) · Populus quem non cognovi servivit mihi, in auditu auris obedivit mihi (Ps. 17:45, verbatim, note right) · Post Achitophel fuit Joiada. (27:34 — Clem. Jojada, i/j only) · Joab. (27:34, verbatim) · Ad omnia mandata tua dirigebar, omnem viam iniquitatis odio habui (Ps. 118:128 — Clem. Propterea ad omnia … omnem viam iniquam odio habui; §4h) · Estote perfecti sicut Pater vester coelestis perfectus est (Matth. 5:48* — Clem. Estote ergo vos perfecti, sicut et Pater vester*; particles only, §4b).

CAPUT XXVIII–XXIX (0663B–0664D, chunk 0013)

Cum eunuchis et potentes et robustissimos quosque in exercitu Jerusalem. (28:1, verbatim) · Dedit autem David. (28:11, verbatim) · Et leunculos. (28:17, verbatim clip of et leunculos aureos) · usque ad (formula) · Polliciti sunt itaque, (29:6, verbatim) · usque ad (formula) · Et benedixit Domino, (29:10, verbatim clip) · Et comederunt et biberunt coram, (29:22, verbatim clip of coram Domino) · Unxerunt secundo Salomonem filium, (29:22, verbatim clip of filium David) · Judex vivorum et mortuorum (cited Ephes. II — the words are Act. 10:42 / 2 Tim. 4:1, and the Eph. 2:14 matter the note answers to stands unitalicized just before it; §6b) · Oleum effusum nomen tuum (Cant. 1:2, verbatim) · Et mortuus est in senectute bona, et plenus dierum et divitiis et gloria, (29:28 — Clem. in senectute bona, plenus dierum, without the second et; §4b) · Viri autem sanguinum et dolosi non dimidiabunt dies suos (Ps. 54:24 — Clem. Viri sanguinum, without autem; §4b) · Qui sibi videntur in bonis ducere dies suos, et in puncto ad inferna descendunt (Job 21:13 — [var:] FIRED, §2k).

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

The gate applied throughout, stated so it can be checked: Pattern 14's own test — a divergence "in a way that changes the sense." A word substituted, added or dropped inside the italic span such that the span asserts something else fires. A pure clip, a dropped or added particle (autem, quoque, et, At, Propterea), a word-order change, a case adapted to the gloss's own syntax, a proper-name spelling, and an orthographic variant (caritas/charitas, coelo/caelo, i/j) do not fire and are all in §4. The witness prefix is Vulg., matching what 8965's shipped chunks print, so the two books read alike where they will stand side by side. Eleven are argued a–k below; the twelfth, on 2 Reg. 24:9 at 0657A, is argued at §3a because the reasoning about it is inseparable from the gloss's arithmetic.

### a. 0656A — Prov. 17:20, a tense and an added possessive Migne: Qui vertit linguam suam, incidit in malum. Clem. Prov. 17:20: qui vertit linguam incidet in malum. Future → present, and suam added. The gloss is about Christ extinguishing folly now, in his examining, so the tense is doing work.

### b. 0656C — Ps. 33:17, the Lord dropped out of the quoted span Migne prints, inside the italics, vultus est super facientes mala, with cujus standing outside them. Clem. Ps. 33:17: Vultus autem Domini super facientes mala. The subject of the verse — Domini — is not in the span at all; an est stands where it was, and the span has been re-hinged on a relative antecedent supplied by the gloss. A reader collating the span against the psalm meets a different sentence.

### c. 0656B — Phil. 2:13, nobis for vobis Migne: qui operatur in nobis velle et perficere. Clem.: qui operatur in vobis et velle, et perficere. First person for second. The gloss applies it to us, the faithful who are the crown, so the divergence is load-bearing and not a slip a reader should have to guess at.

### d. 0658B — Deut. 10:12, O Israel for Et nunc Israel, and ut for et Migne: O Israel quid Dominus Deus tuus petit a te? nisi ut timeas … ut ambules in viis ejus. Clem.: Et nunc Israel, quid Dominus Deus tuus petit a te, nisi ut timeas … et ambules. The vocative replaces the temporal frame, and the second clause is subordinated to the first instead of coordinate with it. ⚑ Migne's own ? mid-quotation is preserved exactly (Pattern 8); it falls on a direct question and needs no 8b recasting.

### e. 0659C–D — 1 Par. 23:29, and it is the most interesting lemma in the range Migne: Ad ferventem similam, et ad torrentem. Clem. 23:29: et ad similae sacrificium … et ad torrendum. Two substitutions in one short lemma. ⭐ The gloss then does textual criticism on its own lemma: some codices have Ad ferventem similam, others ad ferventem only, sed in Hebraeo habetur, ad torrendum — and ad torrendum is exactly what the Clementine prints. So Migne's page preserves the variant and the correction. ⛔ The rendering is where this could have been lost. torrentem (participle) and torrendum (gerund) are one root, and English has no clean exponent for the difference; rendering both "for the roasting" would have erased the very distinction the gloss is arguing. They are held apart as "for the parching" (ad torrentem, the lemma) and "for the roasting" (ad torrendum, the Hebrew reading), and ad ferventem similam / ad ferventem as "for the seething fine flour" / "for the seething." The four are distinct English phrases because they are four distinct Latin readings.

### f. 0660C — Act. 13:2, and Paulum for Saulum Migne: orantibus apostolis et jejunantibus, ait Spiritus sanctus: Segregate mihi et Paulum et Barnabam. Clem. Act. 13:2: Ministrantibus autem illis Domino, et jejunantibus, dixit illis Spiritus Sanctus : Segregate mihi Saulum et Barnabam. Four divergences, of which the name is decisive: the Spirit separates Saul, who is not yet Paul. orantibus apostolis for ministrantibus illis Domino is the second material one — Migne's span names the apostles and makes the act prayer rather than ministry, which is precisely what the gloss needs (a primis praedicatoribus ad praedicandum ordinati sunt).

### g. 0660C — 1 Cor. 9:19, omnes for plures Migne: ut omnes lucrifacerem. Clem.: ut plures lucrifacerem. "All" for "the more" — and the gloss's whole point is the universality of the condescension (Omnium servum me feci), so the substitution is the one that makes the proof-text prove it.

### h. 0660D — Matth. 24:12, a perfect turned future Migne: abundabit iniquitas, et refrigescet charitas multorum. Clem. Matth. 24:12: Et quoniam abundavit iniquitas, refrigescet caritas multorum. The gloss is expounding the western gate as finis mundi … persecutio Antichristi, and Migne's future puts the whole verse in the time to come where the Clementine has a perfect protasis. (charitas for caritas is orthography and is in §4.)

### i. 0661C — Isa. 26:12, in nobis for nobis Migne: Omnia opera nostra operatus es in nobis. Clem. Isa. 26:12: omnia enim opera nostra operatus es nobis. "Wrought in us" for "wrought for us" — and the gloss has just said nemo haec agit nisi dono Dei, so the indwelling reading is the one it wants. ⚑ Migne's note reads (Psal. XXVI); the words are Isaias'. Notes ride verbatim, so the marker names the true book and §6a carries the citation.

### j. 0662C–D — Isa. 66:2, an Old-Latin-shaped quotation Migne: Super quem requiescet Spiritus meus, nisi super humilem et quietum et trementem sermones meos? Clem. Isa. 66:2: ad quem autem respiciam, nisi ad pauperculum, et contritum spiritu, et trementem sermones meos? Three substitutions: the Spirit resting for the Lord looking, humilem for pauperculum, quietum for contritum spiritu. Only the last four words are common ground. This one has no [n:] at all — it is the resumption tail of an usque ad — so without the marker a reader has nothing to collate against.

### k. 0664D — Job 21:13, qui sibi videntur added Migne: Qui sibi videntur in bonis ducere dies suos, et in puncto ad inferna descendunt. Clem. Job 21:13: Ducunt in bonis dies suos, et in puncto ad inferna descendunt. Job's flat indicative becomes a self-deception. Standing as it does at the very end of the book, beside Ps. 54:24 on the bloody and deceitful, the added clause changes who is being described.

§3 · THIS RANGE'S CRUCES — the numbers, the names, and the words that will not construe

### ⭐ a. 0656A — In Regum libro ita gives 80,000 and 50,000 where Kings gives 800,000 and 500,000, in a gloss that is doing arithmetic between the two books Migne, at 1 Par. 21:5: In Regum libro ita. Quod inventa sunt de Israel octoginta millia virorum fortium et de Juda quinquaginta millia pugnatorum. Clem. 2 Reg. 24:9: octingenta millia virorum fortium … quingenta millia pugnatorum. Both of Migne's are real, well-formed numerals, so the non-word rule does not touch them; both are an order of magnitude short of the text they are explicitly quoting (In Regum libro ita). This earned a [var:]* — not because the numerals look wrong, but because the passage presents itself as the text of Kings, and because the gloss at 0657C then reasons from the difference between the two books' numbers (tot millia interfecta … quot millibus numerus in Paralipomenon scriptus, numerum qui in Regum est excedit*). An arithmetic argument with a wrong operand is exactly the class the launch brief calls the most valuable kind, and a reader who cannot see the operand cannot see the argument.

### ⭐ b. 0656A — non viginti et octo millia, and 0656A decem et octo with no millia Three numerals in one short gloss, all real words, all rendered as printed and none marked:

1. In quibusdam autem codicibus invenitur David non viginti et octo millia in valle Salinarum occidisse, quod vitio scriptorum factum est. The lemma four lines above is Decem., clipping 18:12's decem et octo millia, and the gloss is about which number the codices carry. The non is placed on the numeral, not on the infinitive — it stands immediately before viginti et octo millia and the whole sentence is a dispute about the quantity, so the English reads "slew not twenty and eight thousand". (Test 2a's trigger fired here: a negative with a quantifier adjacent to a plausible verb. The structure, not the readability, decided it.)
2. Joab vero duodecim millia prostravit, sicut in titulo quinquagesimi septimi psalmi scribitur. The twelve thousand in a psalm title is Ps. 59:2 (in valle Salinarum duodecim millia), not Ps. 57. Rendered "the fifty-seventh psalm" as printed. ⛔ A [cj:] was weighed here and DECLINED, and the rival is named because naming a rival is not adjudicating between them. The rival was [cj: read quinquagesimi noni]. It lost because, unlike the 288 at 0660B, nothing inside the printed sentence contradicts the printed ordinal — the datum the gloss actually reasons from (duodecim millia) is reported correctly, and only the pointer is off. A [cj:] is for a real word whose faithful English makes the author assert what he did not; here he asserts the twelve thousand, and he is right.
3. Hic autem decem et octo Abisai interfecisse perhibetur — the numeral stands with no millia. Rendered "eighteen", bare, as printed. It is the same phrase as the lemma Decem. with its noun elided, which in an abridged recension is unremarkable and is not a defect.

### ⭐ c. 0660B — ducenta octoginta against quater septuaginta duo: the one [cj:] quater enim septuaginta duo, ducenta octoginta faciunt. Four times seventy-two is 288, and the lemma's own verse (1 Par. 25:7) reads cuncti doctores, ducenti octoginta octo. Migne's numeral is a real, well-formed form, so Pattern 12 has nothing to wrap and Pattern 7 renders it — and the English then asserts an arithmetic the author did not (the very case Pattern 18 was made for, 11081 @0030B). So: ducenta octoginta keeps its English, "two hundred and eighty", and [cj: read ducenta octoginta octo] stands beside it, additive. The rival treatment, rendering "two hundred and eighty-eight" and logging a crux, is the silent repair Pattern 7a forbids and was never available. ⚠ Whether the lost octo is Migne's compositor or our digitization is not asserted: a plate read would settle it, and the [cj:] claims only that the sum requires it.

### d. 0658D — Anima justi sedes est sapientiae is cited Sap. VII and is nowhere in the Clementine Searched as a phrase and by its parts across all 73 books: neither Anima justi nor sedes est sapientiae occurs. It is a medieval adage that travels under Wisdom's name. No [var:] was fired, and the reason is exact: Pattern 14 sets our reading against a received reading, and here there is none to name. The crux is the right home, and this is it.

### e. 0657C–D — in arca Areuna vel Ornam, where the lemma prints in area Ornam The lemma is 21:18's in area Ornam Jebusaei — the threshing-floor. The gloss twice writes arca: altare jubetur in arca Areuna vel Ornam construere, and then nisi in arca cordis … constituat illi altare humilitatis. Both rendered "ark", as printed, and no marker fired. arca is a real, well-formed word, and the addendum-A rule is that a real word that reads oddly is Migne's until a plate says otherwise — refuted five times out of five when a corpus argument was pressed against one. It is also not obviously an error: the second arca cordis is a settled figure, and the gloss may be trading the floor for the ark deliberately, which is why "floor" was not quietly restored in either place. A plate read of col. 0657C–D would settle it and would be worth having.

### f. 0661C, 0657D — faciente domino and In regum ita: lowercase where Migne elsewhere capitalizes Migne prints a lowercase d in the Asabias etymology and a lowercase r in In regum ita at 0657D, against Domino and In Regum libro everywhere else in the range. Both are capitalized in the English — "the Lord doing", "In Kings thus" — and this is a deliberate departure from the mirroring rule, stated so it is checkable. 8965 §M's mirroring rule was made for Ecclesia/ecclesia, where the case carries a semantic distinction (the Church against a church). dominus and regum carry none: lowercasing them in English produces a page that reads as our typo and tells a reader nothing about the plate, which the facing Latin column already shows him. The two sites are named here so that a later sweep can reverse the decision if Wilson wants the mirror to be absolute.

### g. 0662A — Jonathas enim domum Dei: an accusative in a predicate slot Jonathas enim domum Dei, et Ozias fortitudo Domini. The second half prints the nominative fortitudo; the first prints the accusative domum where the same construction wants domus. English has no exponent for the case of a predicate noun — "Jonathan is the house of God" is the only English there is — so this falls under Pattern 9 and takes no marker and no emendation. Logged, and marked so a later sweep does not churn it.

### h. 0663B — Interpretatur enim Joab idem pater Real words, and it will not construe: idem pater stands as the interpretation of Joab, whose etymology the same paragraph has just given as inimicus. Rendered literally, "For Joab is interpreted the same father", and no conjecture is offered in the text. A plausible one (Idithun pater, or idem for Idida) would be guesswork on a page where nothing else is broken.

### i. 0662D — cupiditate de pravatus: the range's only split run Qui postea cupiditate de pravatus cum Absalon de nece David tractavit. A word split across a space — the non-word class, which the addendum records as our digitization's error 19 times out of 19 in the sibling book and every time in the batch that measured it. Rendered in plain English, "depraved by covetousness", and NO marker fired, exactly as 8965 handled its seven split runs: a [sic:] here would be a public claim that Migne's compositor split the word, and seven false [sic:] were withdrawn from a previous batch for making precisely that claim. Raised as a plate candidate in §7, never patched.

### j. 0661B — quae enim hostibus offerebant, in supellectilem templi donabant "For the things which they offered to the enemies, they gave into the furnishing of the temple." The dative hostibus with offerebant gives a sentence in which spoils are offered to the enemy and then given to the temple; the argument wants the reverse direction (donaria principum thesauros esse sanctorum). Rendered as printed. No marker: hostibus is a real word in a grammatical slot, and translating offerebant as "took from" would be Pattern 7a's mechanism 1, the opposite word chosen to rescue a sentence.

§4 · DIVERGENCES CHECKED AND DELIBERATELY LEFT UNMARKED

Every one of these was looked up in clementine-flat.txt and every one was decided against the gate in §2's opening paragraph. They are listed so that a blind reader can tell a checked divergence from an unnoticed one.

• a. Case adapted to the gloss's own syntax. 0655C gladio spiritus (Clem. Eph. 6:17 gladium) — an ablative because confligit governs it. The words are identical; only the ending is the author's weave.
• b. A particle dropped or added, or a head clipped. 0655D Fecit sibi David nomen (Clem. Fecit quoque sibi) · 0656B Aliud bellum (Clem. Aliud quoque bellum) · 0657D Propter hoc dolavi in prophetis, et occidi eos (Clem. has a semicolon and no et) · 0659A Ubi venit plenitudo temporis (Clem. At ubi) · 0661A Clausi sunt sermones (Clem. clausi sunt signatique sermones) · 0663B Estote perfecti sicut Pater vester (Clem. Estote ergo vos perfecti, sicut et Pater) · 0664C in senectute bona, et plenus dierum (Clem. without the second et) · 0664D Viri autem sanguinum (Clem. Viri sanguinum). None changes what the span asserts.
• c. Proper-name orthography. Achimelech / Clem. Ahimelech (1 Par. 18:16) · Ornam / Clem. Ornan (21:18, 21:25, twice in the same chunk) · Melchon / Clem. Melchom (20:2) · Joiarib / Clem. Jojarib and Joiada / Clem. Jojada (i/j) · Ismaelites / Clem. Ismahelites (27:30) · Tebath / Clem. Thebath (18:8). 8965 §M keeps names form for form and fires no [var:] on their spelling; that is followed here without exception, and the forms are listed in §0 so the choice is visible.
• d. 0656B — 2 Reg. 21:19, an omission that leaves the quotation without its noun. Migne: Tertium fuit in Gob contra Philisthaeos. Clem.: Tertium quoque fuit bellum in Gob. The dropped bellum is supplied to the reader by the lemma Aliud bellum two lines above, and the abridged recension compresses quotations constantly. Weighed for a [var:] and declined on the omission rule; named here because it is the closest call in §4.
• e. 0660C — Zachariae against 26:2's nominative Zacharias. Migne's second lemma takes the genitive out of Filii Meselemiae : Zacharias primogenitus. A lemma's case following the gloss's own sentence is not a textual divergence.
• f. 0661C–D — the I Joan. IV span is not a quotation. qui enim timet, proficit ut amet, et foras mittat timorem is the gloss's own sentence; only foras mittat timorem echoes 1 Jo. 4:18's perfecta caritas foras mittit timorem, and the subject and mood are the gloss's. Pattern 14 fires on a citation, not on an allusion woven into the author's clause.
• g. 0662C — Matth. 19:28, word order only. Migne super duodecim sedes, Clem. super sedes duodecim.
• h. 0663B — Ps. 118:128, iniquitatis for iniquam. "Every way of iniquity" and "every iniquitous way" assert the same thing; the head-clip of Propterea is §4b. Declined under Pattern 14's own "changes the sense" test rather than under any rule about word classes.
• i. Orthography with no semantic content. charitas/caritas (0660D) · coelestis/ caelestis (0663B) · coeli/caeli — the flat file's ligature expansion is documented in sources/vulgate/README.md and is never a divergence.
• j. 0657C — Septuaginta millia virorum (21:14) agrees with the Clementine at its own verse AND with 2 Reg. 24:15. Recorded as a checked agreement precisely because the launch brief warns that this session has already caught a septuaginta-for-sexaginta in a gloss counting guards. It is not that class. Likewise 0656A's septuaginta judices chosen by Moses in the desert: Num. 11:16 reads septuaginta viros, so the number is right and no crux is owed.

§5 · usque ad — decided per occurrence, and the look-alikes named

19 tokens of the string in the range. Split:

• The abridgment formula, in its own italic span: 14 (0010: six, at Et filias 18:1, Ammon 19:1, Bellum 20:2, Aliud bellum 20:5, Peccavi nimis 21:8, Dedit ergo David 21:25 · 0011: five, at CAPUT XXIII's Quia lex et prophetae, 23:25, 24:2, 24:6, 25:1 · 0012: one, at 27:32 · 0013: two, at 28:17 and 29:6). All rendered as far as, each in its own italic span answering an italic usque ad. No verb was supplied to complete any of the fourteen resumption tails — 0664A's perstrinximus and 0662C–D's Super quem requiescet stand as they end, and 0660B's digne laudes persolvit is left with no terminal stop because Migne prints none.
• Ordinary "unto / until": 2, and one of them is the trap. 0659A usque ad finem Deo servit → "serves God unto the end" — plainly prepositional. ⚠ 0661A Clausi sunt sermones usque ad praefinitum tempus → "The words are shut up until the appointed time" — this sits inside a single italic span with the rest of Dan. 12:9 and its [n: (Dan. XII)], and the words are the verse's own. Read as the formula it would have swallowed the second half of a scripture citation, which is the failure addendum D names. Decided from what stands on either side of it, as the brief requires.
• Inside another word: 3 — utriusque (0659B, 0661A), cujusque (0663C). Not the string's own use at all.

§6 · MIGNE'S CITATIONS — five that do not point where they say, all passed through verbatim

[n: …] contents are sacred and untranslated, so none of this touches the page; it is recorded because the citation index will meet it and because a reader collating will.

### a. Four misdirected [n:]

• 0656B — Benedices coronae anni benignitatis tuae is cited (Deut. VII). The words are Ps. 64:12, verbatim; Deut. 7:13 has nothing like them.
• 0658A — Jerusalem, quae aedificatur ut civitas is cited (Ibid. CXXII), following a (Psal. CXVIII). The verse is Ps. 121:3 in Clementine numbering — off by one, which is the Hebrew-numbering offset and unremarkable in itself, but it is recorded because the Ibid. chain resolves through it.
• 0658D — Adorate Dominum in atrio sancto ejus is cited (Psal. XVIII). It is Ps. 28:2.
• 0661C — Omnia opera nostra operatus es in nobis is cited (Psal. XXVI). It is Isa. 26:12 — the right chapter and verse under the wrong book, which is how the slip probably happened. The [var:] at that site names Isaias so a reader is not sent to the psalter.

### b. 0664B — a note attached across an unitalicized quotation Christus vero est pax nostra, qui fecit utraque unum — Eph. 2:14, and Migne sets it roman, not italic. The italics fall instead on Judex vivorum et mortuorum, which is Act. 10:42 / 2 Tim. 4:1, and the [n: (Ephes. II)] sits on that span. So the note is right about the sentence and wrong about the words it stands beside. Italic parity is preserved exactly as printed: the Ephesians matter is roman in the English too, and no italics were added to make the note fit.

⛔ Nothing in this section was "corrected." Runbook false positive 2 governs: a following Ibid. depends on the printed sequence, and 0658A is exactly such a chain.

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

The dominant defect class, so each site is listed with the evidence, and the two sites where two readings both construe are said to be such rather than resolved silently.

1. 0657B ex quorum causa peccavit → "on whose account he had sinned." David — the paragraph's subject throughout (Tumore elatus David, populum numerando peccavit), and peccantis vitium in the same clause is his fault, not the judge's.
2. 0657B ita enim quae corporaliter populum percutit → "that which struck the people bodily." quae is feminine singular and its antecedent is mors, named two sentences earlier (Mortem dicit). No gendered English pronoun was supplied; the relative carries it.
3. 0657D nisi … constituat illi altare humilitatis, in quo offerat → "unless he set up … in which he may offer." The subject is homo, printed in the same sentence as ab homine.
4. 0657D–0658A Fidei firmitatem sacris dogmatibus ordinavit → "He ordered the firmness of faith." ⚠ Two readings both construe: Christus, the subject of the immediately preceding gloss (De Ecclesia Christus artifices domus suae elegit), or David, whose act the lemma Ferrum records at 22:3. The English "He" leaves it exactly as open as the Latin does, and nothing was decided.
5. 0658A quos ad exemplum instruxerunt vel scripturis suis inseruerunt → "whom they trained up." The holy Fathers (sancti Patres), the subject of the Aeris gloss immediately above.
6. 0660B et omnem terrenarum rerum cupiditatem transilit, digne laudes persolvit → "and overleaps … he worthily pays his praises." The subject was removed by the abridgment; the two verbs fix it as third person singular and the gloss's frame (Quia verus David rex noster et Dominus) fixes it as the one who praises, not as David. Supplied as a bare "he" and named here rather than resolved.
7. 0661B quae enim hostibus offerebant … donabant → "the things which they offered." The princes, from donaria principum in the same sentence.
8. 0661D ex his honorat servientes sibi → "out of these he honours." Christ, from Thesaurus Christi, the sentence's own opening.
9. 0662D qui etiam, secundum librum Regum, laqueo se suspendit → "who also … hanged himself." ⚠ Two readings both construe and the English preserves both: the relative chain runs from Hic est Judas Scarioth, but secundum librum Regum points to Achitophel (2 Reg. 17:23), whose name opens the paragraph. A bare "who" carries the ambiguity that the Latin carries; naming either man would have decided a question Migne left open.
10. 0663D lignum, fenum, stipulam superaedificat, detrimentum patietur → "he builds upon it … he shall suffer loss." Subject removed by the abridgment; singular fixed by both verbs, and the matter is 1 Cor. 3:12–15's builder. Supplied as a bare "he".
11. 0664C Significat cum copia virtutum … vitam finisse praesentem → "that he ended the present life." David, from the lemma Et mortuus est in senectute bona directly above.
12. 0661A Qui praecipue a David honorabantur … significat doctores de Judaeis → "Who were chiefly honoured … signifies the teachers." ⚑ Migne prints a singular verb against the plural Filii Moysi and its plural relative. English has an exponent for number, so the singular is reproduced (Pattern 12's second test: rendering it "signify" would silently repair a disagreement that is a fact about this printing). No marker: nothing is carried untranslated.

Verbless sentences left verbless, no verb supplied to any of them: 0655D–0656A Et quia Christus … exstinguit, unde: (a quia clause with no main clause) · 0658A Lectiones sonoras et cantica spiritualia quae … · 0659B Unde septuaginta duo discipuli ad praedicandum missi · 0659C Evangelium vero stabile et fixum · 0659C Thalamis vel cubiculis in quibus habitant · 0660D Potest per Eleazar filium Aaron seniorem, etc. (rendered impersonally, "It is possible through Eleazar…", supplying no infinitive) · 0661A Sed in perfectorum actione elevata, quod divinis praesignatur eloquiis · 0662A Jonathas enim domum Dei, et Ozias fortitudo Domini · 0662B Quae virtutes in campestribus · 0664A Quia congregatio fidelium … laudes et hymnos Deo decantet (a quia clause with no main clause). ⚑ The ordinary Latin nominal sentence Thesaurus Christi scientia, sapientia virtutis (0661D) DID take an English copula — Latin omits est by rule and English has no exponent for that, so this is Pattern 9 and not a missing verb. Named so the two classes are not confused.

§8 · ONE PLACE THE PRINTED CASE WOULD NOT CARRY THE CONSTRUCTION — 0660A, and the rival is named

ostendens Ecclesiae sacerdotalem et regiam dignitatem plene habere et legis, et Evangelii doctrinam perfecte servare. The two infinitives want an accusative subject and Migne prints the dative or genitive Ecclesiae. Two readings:

• dative — "showing to the Church that [she] fully has…", with the accusative subject elided;
• genitive with dignitatem — "showing the priestly and royal dignity of the Church to be fully had…", which requires a passive the Latin does not print.

The dative reading was taken, and the English reads "showing that the Church fully has the priestly and the royal dignity and perfectly keeps the teaching both of the law and of the Gospel." The genitive lost because it can only be got into English by turning two active infinitives passive, which is a larger intervention than supplying the elided accusative subject that a dative reading already implies — and because both readings assert the same thing, so nothing a reader needs is at stake. No marker: the words are all real, all in place, and the difficulty is a construction, for which this apparatus has no marker and should not invent one (the Ruth 0534A precedent).

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

• Test 1 (a [sic:] must not swallow its word). Vacuous — zero [sic:] in the range. Run anyway on the [var:] and [cj:] markers, which are additive by construction: every marker was stripped mechanically and all four chunks read as continuous English with no stranded article, preposition or doubled mark.
• Test 2 / 2a (name the host of every printed negative). 38 printed negatives (non ×23, nisi ×7, nec ×3, nemo ×3, neque ×1, nihil ×1), each traced to the single Latin word it attaches to. Three fired 2a's trigger — a negative near a quantifier, correlative or predicate — and each was proved from the sentence's structure, not from what read well: 0656A non viginti et octo millia (quantifier, and the whole gloss is a dispute about the quantity — §3b) · 0657B a culpa alienus non fuit (predicate adjective, not the verb: David was not free from the fault) · 0663C non causa libidinis, sed suscipiendae prolis (a sed correlative fixing the negative on the noun phrase, not on habebant) · 0662B nec acceptionem personarum, sed bonae voluntatis rectitudinem quaerunt (same shape; the English reads "seek for their hearers not the accepting of persons, but the rightness of good will", with the negative on the object).
• Test 2b (an anchor's position, not merely its presence). 38 column anchors (10 · 10 · 12 · 6), each checked against the WORD it stands beside in the Latin. Six required splitting an English construction to keep the adjacency, and all six were done deliberately: 0655C aptum [0655C] praedicationi → "fitted [0655C] for preaching" · 0658D principibus [0658D] Israel → "princes [0658D] of Israel" (inside a lemma) · 0659B quia [0659B] fides → "because [0659B] the faith" · 0661C possessionem [0661C] acceperunt → "beyond the Jordan [0661C] received a possession" · 0664B opera [0664B] fructuosa → "the [0664B] fruitful works" (English puts the adjective first, so the anchor stands before the word rendering fructuosa) · 0664D non dimidiabunt [0664D] dies suos → "shall not live out half of [0664D] their days".
• Test 3 (punctuation compared BY BAND). Built as a script over the verified 1:1 anchor partition, counting ; : ? ! « » in each band of both twins. First pass returned three mismatched bands, all in chunk 0012, all ours — 0660C a semicolon softened to a comma after reatum suum in propatulo habuerunt; 0661D a semicolon invented after sapientia virtutis; 0662D a semicolon invented after ruinam meruit. Totals would have cancelled the last two exactly against the first. All three repaired; final state 0 mismatched bands in all four chunks. Question marks: 2 and 2, in the same two bands.
• Test 4 (does it ASSERT what he asserts?). Run last, paragraph by paragraph against the Latin, with the counts already clean. Six repairs, none of which any other check could see: 1. 0656B cum praedicatorum ordo contra mundi potentes fidei opponit had been rendered "sets faith against the mighty of the world" — reading the printed genitive fidei as an accusative object the plate does not print. Repaired to "when the order of the preachers of faith opposes the mighty of the world", which construes the printed case. Pattern 7a mechanism 5, a mis-bound reference, and it read better wrong than right. 2. 0657D ut … sibi convenirent had been rendered "might be fitted together to him", which makes the craftsmen converge on Christ; sibi with a plural subject is "to one another". Repaired. 3. 0657D Osee 6:5 dolavi in prophetis had been given an object — "I have hewed them by the prophets" — which is Douay's supplement, not Migne's text (the Clementine prints no object either). Repaired to "I have hewed by the prophets". 4–6. The three conformed proper names — Absalom for Absalon, Judas Iscariot for Judas Scarioth, In Hebron for In Hebrun — every one of them the familiar English arriving over the plate, and every one invisible to marker parity, band counts and negation tests. These are 7a″'s exact mechanism and they are why the fourth test is run last and slowly.

§10 · PLATE CANDIDATES — what a read of PL 113 cols 0655–0664 would settle

None of these fired a marker, so this range ships asserting nothing about the plate (ship test B). Listed cheapest-first; data/volumes.json gives PL 113 a second witness at Gallica ark:/12148/bpt6k5505319w with server-side IIIF crops, one HTTP request per read, map f = (column + 1) / 2 (NOT the archive map, (column + 11) / 2) — 8965 settled three candidates that way in three fetches.

⚠ Two of these sit at or near a left-column head — 0656A and 0657A — and our archive.org copy of PL 113 (patrologiaecurs04migngoog) is known to clip its left margin on some pages. That is a fact about our photograph, not about Migne, and it is a live alternative explanation for a truncated-looking numeral. Raised, never patched. No conjecture was entered anywhere as a patch; a conjecture entered as a patch wears Migne's voice.

⭐ A checked negative, recorded because the addendum asks for them. septuaginta was suspected twice on the launch brief's warning and cleared twice from the file alone: 0656B's septuaginta judices chosen by Moses matches Num. 11:16 exactly, and 0657C's Septuaginta millia virorum matches 1 Par. 21:14 and 2 Reg. 24:15 exactly. Neither is the septuaginta-for-sexaginta class, and neither needs a plate.

§11 · SECTIONS READ CLEAN, AND WHAT WAS DECLINED

Read clean, with no crux and no candidate: CAPUT XIX entire (0656C, one paragraph) · CAPUT XXII 0657D–0658D except the Deut. 10:12 lemma · CAPUT XXIII 0658D–0659B · CAPUT XXIV 0659D–0660A · CAPUT XXVI 0660C–0661B · CAPUT XXVIII 0663B–0664A · CAPUT XXIX 0664A–0664C. Named so a blind reader can tell a checked column from an unvisited one.

Declined, with the reason:

• A [var:] on Ornam / Clem. Ornan, and on every other proper-name spelling. 8965 §M keeps names form for form and fires no marker on their orthography; firing here would make the two books disagree about their own convention on the shelf where they will stand together.
• A [var:] on the gloss's qui in Regum Abimelech at 0656A. The gloss's whole etymology turns on it — Achimelech = frater meus rex against Abimelech = pater meus rex — and Clem. 2 Reg. 8:17 reads Achimelech, not Abimelech, so the gloss's report of the Kings text does not match the received Kings text. It is nevertheless unitalicized gloss prose, a report of a reading rather than a quotation of one, and Pattern 14 is for scripture quoted as scripture. This is the most valuable thing in §11 and it is deliberately in the crux and not on the page**; a reader who wants it will find it here, and a marker would have converted the gloss's own textual note into our claim about it.
• A [cj:] on quinquagesimi septimi (§3b) — the rival is named there and why it lost.
• A [sic:] on de pravatus (§3i) — the ship test.
• A [nt:] — no note in the range is editorial prose; all 63 are citations or sigla.
• An [ed:] — nothing in the range is missing from the digitization. The two [n:] sigla with Migne's own internal spacing that the peers found do not recur here; all 63 notes in this range are ordinary and pass through verbatim.
• Any comma→semicolon sweep. Migne prints 11 semicolons of his own in these four chunks, so the Tobit exception — which rested on his printing zero in that book's first chunk — does not apply and was not invoked. His own count was taken before the question was decided, as the brief requires.

§Z · THE BLIND POLARITY READ (step 4a)

Run 2026-08-25 by a reader fenced from this file. All 14 chunks read against their Latin twins. Every printed negation was located in the English and checked to be on the same word. Zero polarity defects — the sixth consecutive book of this series where the name of the class proves to be something else.

⚑ Disclosure, made by the reader itself and recorded because it bears on how the report is weighed: a mechanical grep for proper-name spellings across the 8964 directory matched inside this file and printed a few of its lines. It happened after every finding had been formed by reading, and the reader said so unprompted. What it incidentally showed — the de pravatus word-split, and Absalon / Judas Scarioth / In Hebrun — confirmed finding 1 rather than producing it. ⭐ The leak is also how we learned something genuinely useful: this file claims the Absalon conformation was fixed, and at 0637C it was not. A stint's own record of a repair is not evidence the repair reached every site.

Repaired, four sites

1. ⭐ 0637C — the conformation the translators' own final test missed. Migne prints David Absalonem filium suum fugerit; the English printed Absalom, the familiar English Bible spelling. The work's other occurrence (0662D, cum Absalon de nece David tractavit) was already correct. The fix had been applied at one site and not the other — which is exactly why a blind read is a different check from a self-audit. Repaired.
2. ⛔⛔ 0648C — the serious one: a relative bound to the wrong antecedent made the saints incarnate. quia ad eum omnis sanctorum coetus pertinet qui descendens de coelis, et factus homo … morte sua interficiens in infernum demersit. Every one of those participles and verbs is singular and belongs to eum — Banaias, read as Christ, the subject of the surrounding chain of Hic-clauses. The English attached "who" to the nearest noun, the saints, and the sentence then attributes the Incarnation and the harrowing of hell to the company of the saints, leaving an unresolvable "his own death" inside a plural clause. ⚑ The English was internally incoherent where the Latin is not — the shape the fourth test exists to catch. Repaired with an em-dash and "he who", the least invasive way to rebind it.
3. 0643B — an ablative agent moved off its verb, so a sentence about a marriage stopped naming who married her. a patre venalis ducta a Manasse: qui horreis praeerat cibariis patri datis, in conjugium … suscepta est — ducta already has its agent in a patre, so a Manasse governs suscepta est. The English gave ducta two agents and left suscepta est agentless. Second defect in the same sentence: cibariis patri datis is the girl's father, rendered "his father", pointing the reader at Manasse's. Both repaired.
4. 0634B — an adjective rebound to make an impossible agreement. septem filios et tres filias praedictas uxor genuit: praedictas is accusative feminine plural and cannot modify uxor, a nominative singular. The English moved it to the noun that made the sentence read smoothly ("the aforesaid wife") and dropped it from the noun that carries it. Repaired to "the aforesaid three daughters". Low severity in sense, but a repair with no marker is the class that leaves no trace.

Migne's own defects — checked, and all carried literally

Reported by the reader as carried literally, correctly refusing to call them undocumented, since it could not see this file. All are argued above. Among them: the intrusive non at 0656A (where carrying it preserves a self-contradiction that is Jerome's point about scribal corruption); the dropped negation at 0654D, carried affirmative with its [cj:]; the 4×72 arithmetic at 0660B; arca for area at 0657C–D, where the gloss's own re-use (in arca cordis) shows the reading is consistent in Migne rather than a one-off slip; and the 80,000/50,000 gloss at 0657A, carried with its [var:] and with the direction of excedit preserved at 0657C.

⬜ TWO THINGS LEFT FOR WILSON, deliberately not acted on

• The lemma-expansion practice is inconsistent inside this work, and the inconsistency is real even though none of the three expansions is a defect. Achar qui. → "Achar who troubled Israel", Her pater. → "Her the father of Lecha", Iniit autem. → "And he took counsel" are all fragments, which convention 3 licenses. But Mesraim autem genuit. (0632A) is the identical shape — transitive verb, object elided, Vulgate continuation available — and stands as "And Mesraim begot." So do Filii autem., Jectan autem., Sem, Ar. ⚑ The standing ruling is to expand the stump rather than contract its twin, which would mean expanding these; that is a change to a shipping work's lemmata and is Wilson's call, not mine. ⚠ Note the cardinals here are all correct and bare (Viginti → "Twenty", Septuaginta millia virorum → "Seventy thousand men") — this work did not repeat the sister book's numeral confusion.
• 0646C — a "reads better than its Latin" site the reader flagged and declined to assert. Videntes indocti … medicamina lavacrumque baptismi, ignorantes in Judaea magistros suos … victos: the English swaps the two participles' objects. The reader's reason for not calling it a defect is good — the alternative parse is the only one compatible with majorum ruina turbati six words later, since men "not knowing" cannot be "troubled by the ruin" — while Migne's comma slightly favours the other binding. Genuinely undecided; left as printed.

Chunks read and found CLEAN: 0002, 0005, 0008, 0009, 0011, 0013.