Epistola I ad Corinthios

PL 114, cols 0519C–0550D · 14 chunks · 12,614 Latin words. Attributed by Migne to Walafrid Strabo; in fact Anselmus Laudunensis et schola. Translated 2026-08-30 by three Opus stints, none of which could see the others, then read by three blind polarity readers and gated at the plate.

This file is the MERGE LAYER, and it is not a copy of the range files. cruces-0000.md (chunks 0000–0004, ARGUMENTUM + CAPUT I–VI), cruces-0005.md (0005–0009, CAPUT VII–XII) and cruces-0010.md (0010–0013, CAPUT XIII–XVI) stand beside it and carry the per-site detail — the full lemma collations, the negation tables, the unexpressed-subject rulings, the plate candidates. This file records only what was settled across the three ranges, or after them by the plate.

⚠ Kept to the merge layer alone, per the 8986 finding: build-cruces.mjs concatenates every cruces*.md in the directory, so a merge that embeds its range files publishes the apparatus twice.

§0 · Apparatus, counted from the English after the plate gate

0 [sic:] · 14 [var:] · 6 [cj:] · 0 [ed:] · 0 [d:] · 0 [nt:] — counted from src/english/8981/0*.md after the gate and the polarity repairs, not summed from the reports.

The three stints filed 14 [var:] · 1 [cj:] · 0 [sic:]. The plate gate added five [cj:] that no stint and no blind reader could have fired, because firing them required the page.

verify-english.mjs 8981 → OK, 14 chunks. scan-raw-markers.mjs → clean. ✅ plate-gate.mjs 8981 PASSES. polarity-record → chunks: 14 · sites: 12 · ours: 8.

⭐ All three stints independently declined to fire a single [sic:], each giving the same reason unprompted: a [sic:] accuses Migne's type, our Latin is Corpus Corporum's transcription and not the plate, and no column had been read. That is now the fourth work running where the stated marker policy produced zero false accusations — and the third where the plate then proved the instinct right in detail.

§1 · THE PLATE GATE — 12 pages, and the result

Pages 266–268, 272, 274–280 read at 300 dpi with 600 dpi crops for every letter-level call, plus p.269 for the polarity read's open items. Read by the session directly, not by fenced agents — deliberately, because briefing several readers on one suspicion produces convergence, not corroboration. Calibration: p.266 prints 521/522, checked before the first read; the map page = (col + 11) / 2 held at every subsequent page.

ALL 14 [var:] STAND · 0 REVISED · 0 FALSE.

⭐⭐ The marker policy predicts the false rate, not the volume. This is PL 114, the volume that returned 65% false on a letter-level sweep in August. Four consecutive works that fired no letter-level type accusation have now returned 36/37, 14/14, 7/7 and 14/14.

§2 · ⭐⭐⭐ THE GATE CREATED FIVE MARKERS RATHER THAN DESTROYING ANY

Every one is Pattern 18: Migne prints a real word, the faithful English asserts what the author did not, and the conjecture stands beside it additively. Every one had been correctly WITHHELD by a stint or a blind reader for want of the column — the marker policy working in both directions, which is the thing this run demonstrates that the previous three did not.

⭐ 0522D is the one no plate read could have found and no reader could have settled alone. The session confirmed at p.266 that the full stop after spiritui is Migne's own, so Pattern 8 keeps it; the reader showed that our English had then given the genitive Dei a dative's job to rescue the fragment. Text untouched, [cj:] names the elision (the settled 8993 @0655C treatment).

§3 · WHAT THE PLATE REFUSED

• ⛔ 0546C — a blind reader's proposed [cj:] is DECLINED. Our file reads min re; the reader proposed [cj: min re; read minore]. The plate prints minore whole and clean at 600 dpi, so the defect is Corpus Corporum's and a conjecture would have accused Migne of a word he never set. TEI patch, no marker. This is exactly the class 8993 had to WITHDRAW two [cj:] for; caught here before shipping. ⭐⭐ Stint C had already reasoned it out from the files alone: min re and ves tra eleven words apart is a scanning signature, not a compositor's, and it withheld the marker on that ground. The plate confirmed it.
• ⛔ 0538C — a blind reader's plate request REFUTED. It proposed disserentis for deserentis at medium-high confidence, arguing from the contrast with praecipientis. The plate prints deserentis. The English stands; the crux is Migne's.
• ✅ 0528C–D Spiritus and 0527C veritatis — two more reader suspicions that the plate answered in Migne's favour. Both were correctly raised without a proposed emendation.

⚑ The standing rule held again: blind readers are strongest on what the English SAYS and weakest on what the page PRINTS. Of six reader claims about the printed page, one was confirmed, three were refuted, and two were declined — while every reader claim about the English stood.

§4 · SIX HIDDEN DEFECTS, and two were MIGNE'S

data/tei-patches/8981.json. Our Latin read correctly at both normalized sites, so no check we own could have found either.

• 0522A qni for qui — a turned u, MIGNE'S OWN, silently corrected by CC. The controls are in the same breath: quibus on the next line and Filii qui dederunt two lines below, both a clean u in his own type.
• 0522C a spurious et — CC inserted a word the plate does not print (a spirituali homine, personaliter legitur). It had reached the English as "and it is read personally".
• Four ordinary CC corruptions: min re→minore, ves tra→vestra (the pair eleven words apart), nomines→homines, spituale→spirituale.

⚠ A methodological warning worth more than the patch it belongs to. The first 600 dpi crop taken for nomines landed on a different occurrence eight lines away (amplius non dissolventur homines) which our file transcribes correctly. Had that been accepted as the control, a correct line would have "verified" a defect on another. The control has to be the line the defect is on, not the nearest line containing the same word.

§5 · CONVENTION DRIFT, resolved across the three ranges

• ⭐ Quasi dicat — three stints, three renderings. A "as if he should say" (×4) · B "As if he said" (×3) · C "as if to say" (×9). Settled by the corpus, not by preference: "as if he should say" runs 239 occurrences corpus-wide and is the form used in every sibling Pauline epistle in this volume (8986, 8993, 8987, 8990) and in 9006. Normalized work-wide, 16 sites, each checked 1:1 against the Latin first. Stint A's ac si diceretur (passive) and bare Quasi: were correctly left alone as different constructions.
• ✅ Head form — the launch brief was WRONG and all three stints were right to ignore it. The brief's example said CHAPTER SEVEN.; the corpus has zero spelled-out forms, CHAPTER XIII. in 26 works, and CHAPTER ONE. in 61. The rule is to mirror the plate: Migne sets CAPUT PRIMUM spelled and CAPUT XIII in numerals, so the English does the same. All three stints converged on this independently. The brief's example wants correcting.
• ✅ Register: ye = 0, you = 103. thou 44 / thee 4 / thy 15, singular address only. The first work in four where the merge layer had nothing to convert — 8993 needed twelve.

§6 · WHAT IS KNOWN TO BE UNRESOLVED

• ⚠ 0533 rationē. The plate sets sed ratione praeditarum animarum with a clear MACRON over the final e; our file has plain ratione. The macron expands to rationem, which does not construe (praeditus governs the ablative), while ratione does — which is what makes the tidying invisible. Recorded, NOT patched, NOT marked: the defensible claim is "the plate prints a macron and our file does not", not "Migne intended rationem". The second witness (raw/scans/pl202/copyB_patrologiaecurs114unkngoog.pdf) closes it in one page.
• ⚠⚠ FOUR PAGES IN THIS WORK CLIP THE LEFT MARGIN — p.274, p.276, p.278, p.280, each losing the left column's opening characters and its left-hand corner number (537, 541, 545, 549). This is new information about the volume: PL 114's clipping was on record for p.265/p.281 (RIGHT) and p.290/p.330/p.340 (LEFT). It is per-page, not per-volume, and now confirmed on both margins inside a single work. No marker here depends on a clipped column; any future crux in one needs the second witness.
• ⬜ workStatus is null, so the page publishes the weaker "New English translation". Whether a prior English of the Glossa on I Corinthians exists is a first-English determination and is Wilson's call — the same one made for 8947 against 8946.

Range: cols 0519C–0529B (ARGUMENTUM, CAPUT PRIMUM–CAPUT VI). Latin source is Corpus Corporum's transcription; of these columns only p.265 (cols 0519/0520) has ever been read at the plate, and that was a checked zero. No [sic:] was fired anywhere in this range — see §4 for why, and for the list of forms that would have taken one under a plate-blind policy.

Four [var:] were fired, all resting on sources/vulgate/clementine-flat.txt and cited by book chapter:verse.

1. Conventions taken in this range

• VERS. n.-- passes verbatim; the lemma is Englished; a fragmentary lemma is rendered as the phrase it stands for (brief convention 3). Expanded fragments in this range: Paulus vocatus. → "Paul, called"; quia non multi. → "for there are not many wise"; Ut non glorietur. → "That no flesh may glory"; Non enim judicavi. → "For I judged not myself to know anything"; Fundamentum enim. → "For other foundation no man can lay"; Sic nos. → "So let a man account of us" (both occurrences, identically); Ab humano. / Humano. → "By man's day" / "Man's day" (same fragment twice — expanded both, per the ⚑ ruling that the stump is expanded, not its twin contracted); Saecularia igitur judicia. → "If therefore you have secular judgments".
• usque ad → as far as, following the settled sibling practice in this volume (8986 ×7, 8987 ×3, 8990 ×1, 8993 ×9, 8995 ×6, 8998 ×19). All five occurrences in this range (0523A, 0523B, 0523C, 0524B, 0524D, 0527C, 0528A) are the abridgment formula — checked one by one against what stands on either side (addendum D). None is a lemma's own words: there is no usque ad inside any italic span in this range.
• pseudo (indeclinable, 0524D "de gloria pseudo", 0525C "de magisterio pseudo", 0528B "sequentes pseudo") is rendered "the false apostles" throughout, matching the work's own spelled-out pseudoapostolorum at 0525D. Not italicized: it is an abbreviation of a phrase the work itself writes out, not a technical term with no English equivalent.
• baptistae → "baptizers" throughout (0519D, 0524B, 0525C), not "Baptists".
• Pattern 17. The gloss's own voice is plural "you" for the Corinthians. "thou/thee" is used only where the Latin is grammatically singular: 0522A (te facis, tu facis, tu non faceres), 0525C–0525D (te discernit, baptizatus es, gloriaris, accepisti, acceperis), 0529B (parcis tibi, te sibi fecit, evertis, peccas). Two "ye" occur in this range: none — no archaic plural was used anywhere.
• Question-mark parity: 0000 = 1, 0001 = 0, 0002 = 1, 0003 = 5, 0004 = 0, matching the Latin twin band by band. No 8b recast was needed: every ? in this range closes a direct question in Latin already.
• Punctuation was compared band by band (test 3), colons, semicolons, question marks and guillemets, and every band now matches its Latin twin exactly. Two marks had been destroyed and were restored: a colon at 0524C (sanctos homines et angelos accipe:) and a semicolon at 0527A (diabolo eum vexandum traderet;). Twenty supplied semicolons/colons — all of them Migne's commas quietly promoted — were reverted to commas.

2. VULGATE COLLATION — every lemma in 0519C–0529B

Comparand: sources/vulgate/clementine-flat.txt, I Cor 1–7 read in full, plus Lc 8, Mt 13, Mt 23, Mt 25, Mc 4, Sir 27, Jo 8.

2a. Divergences MARKED [var:] (4)

1. 0522B, chunk 0001, gloss on I Cor 2:12. Migne prints Qui habet, dabitur ei … Qui ergo non habet, et quod habet, auferetur ab eo, and attributes it [n: (Luc. VIII)]. Lc 8:18 reads Qui enim habet, dabitur illi : et quicumque non habet, etiam quod putat se habere, auferetur ab illo. Migne's wording is instead Mt 13:12 almost verbatim (Qui enim habet, dabitur ei … qui autem non habet, et quod habet auferetur ab eo). Material, because the gloss's whole point is plene habere est scire unde habeat — Luke's quod putat se habere is precisely the clause that would have carried it, and it is absent. ⚠ The [n: (Luc. VIII)] was passed through verbatim and NOT corrected (runbook false positive 2: an anaphora-bearing citation is never repaired in place). The mismatch is recorded here and in the marker, not in the note.
2. 0522D, chunk 0001, lemma of I Cor 2:15. Migne prints Spiritualis autem homo. I Cor 2:15 reads Spiritualis autem judicat omnia — with no homo. Material: homo is the word the gloss then expounds (Est vel vita vel scientia, and the whole distinction of the spiritual man in life from the spiritual man in knowledge), and it builds the parallel with Animalis autem homo of 2:14, which is the previous lemma and is the Vulgate's own wording there. The added word is doing the work of the paragraph.
3. 0523B, chunk 0002, Augustine's proof-text. Migne: Vasa figuli probat fornax, et homines tentatio tribulationis probat [n: (Eccli. XXVII)]. Sir 27:6 reads Vasa figuli probat fornax, et homines justos tentatio tribulationis — with justos and with no second verb. Material: the Vulgate restricts the trial to the just; Migne's line, adduced to prove that the fire tries every man's work, does not.
4. 0526C, chunk 0003, gloss on I Cor 4:15. Migne: Ne vobis dicatis Patrem in terra, unus est enim Pater Deus [n: (Matth. XXIII)]. Mt 23:9 reads Et patrem nolite vocare vobis super terram : unus est enim pater vester qui in caelis est. Material: Migne's Pater Deus is the identification the gloss then argues from (ne gratia Dei … sanctitati cujusquam hominis tribuatur), and the Vulgate does not print it.

2b. Divergences FOUND, examined, and DELIBERATELY NOT MARKED

Each is a real divergence from the Clementine; none changes the sense, and Pattern 14's bar is "in a way that changes the sense." Listing them is the point of this section.

2c. Lemmata CHECKED AND FOUND TO AGREE with the Clementine

Verbatim (allowing for the fragment cut and for tanquam/tamquam, nunquam/numquam):

I Cor 1 — Paulus vocatus · Apostolus Jesu Christi · Et Sosthenes · Frater (1:1) · Cum omnibus (1:2) · Gratia · Et Domino Jesu Christo (1:3) · In omnibus (1:5) · Obsecro (1:10) · Chloes (1:11) · Hoc autem dico · Ego quidem (1:12) · Divisus est · Christus (1:13) · Gratias ago Deo (1:14) · Sed evangelizare · Non in sapientia · Non in sapientia verbi · Verbi (1:17) · Pereuntibus · Dei virtus (1:18) · Scriptum est enim (1:19) · Placuit Deo per stultitiam praedicationis salvos facere credentes (1:21, verbatim, whole) · Videte enim · quia non multi (1:26) · Sed quae stulta (1:27) · Ut non glorietur (1:29) · Ex ipso autem (1:30). ⭐ nulli at 0521B is not a lemma but the gloss's own contrast-word (Non autem ait, nulli); it is not in the verse and correctly claims no divergence.

I Cor 2 — Non enim judicavi (2:2) · Perfectos (2:6) · In mysterio · Quae abscondita (2:7) · Si enim cognovissent · nunquam Dominum gloriae crucifixissent (2:8) · Nec in cor hominis · Ascendit (2:9) · Scrutatur (2:10) · Sed Spiritum qui ex Deo est · Ut sciamus (2:12) · Spiritualia (2:13) · Animalis autem homo · Quia spiritualiter examinatur (2:14) · Omnia (2:15).

I Cor 3 — Adhuc enim carnales (3:2) · Ministri ejus (3:5) · Unum sunt (3:8) · Adjutores · Agricultura · Aedificatio (3:9) · Fundamentum enim (3:11) · Si quis autem (3:12) · Dies enim · Ignis probabit (3:13) · Mercedem (3:14) · Nescitis (3:16) · Si quis autem · Templum (3:17) · Sapientia enim hujus mundi (3:19) · Nemo ×2 · In hominibus (3:21) · Omnia enim · Sive vita (3:22) · Vos autem Christi · Christus autem Dei (3:23).

I Cor 4 — Sic nos ×2 · Et dispensatores (4:1) · Hic jam quaeritur · Jam (4:2) · Meipsum · Ab humano · Humano (4:3) · Quoadusque veniat Dominus · Illuminabit · Tunc laus erit unicuique (4:5) · Me et Apollo · Ne · unus (4:6) · Quis enim te discernit · Quid autem · Si autem accepisti (4:7) · Jam saturati estis · sine nobis · Regnetis (4:8) · Puto enim · tanquam morti destinatos · Et angelis (4:9) · Propter Christum (4:10) · Tanquam purgamenta (4:13) · Non ut confundam vos (4:14) · Paedagogorum · Per Evangelium ×2 (4:15) · Si Dominus voluerit (4:19) · In virtute (4:20).

I Cor 5 — Omnino auditur (5:1) · Inflati estis (5:2) · In nomine Domini (5:4) · Tradere · Ut spiritus salvus sit (5:5) · Totam massam (5:6) · Conspersio (5:7) · Sinceritatis (5:8) · Alioquin (5:10) · Nunc autem · Nominatur ×2 (5:11).

I Cor 6 — Audet aliquis (6:1) · Quoniam (6:2) · Saecularia igitur judicia (6:4) · Ad verecundiam (6:5) · Fratribus (6:8) · An nescitis · Nescitis (6:9) · Sed sanctificati estis · sed (6:11) · Omnia mihi licent (6:12) · Esca ventri (6:13) · Corpora · Membra sunt · Christi (6:15) · Unum corpus (6:16) · Qui autem adhaeret (6:17) · Fugite fornicationem · Qui autem fornicatur (6:18) · Templum (6:19) · Et portate Deum · In corpore (6:20).

Unattributed scripture inside the gloss, also checked: qui gloriatur, in Domino glorietur (0521C = I Cor 1:31, verbatim); quia in igne revelabitur implied at 0523C (I Cor 3:13, paraphrase not quotation, unmarked).

3. THE UNEXPRESSED SUBJECT — every supplied "he/she/it/they"

Sites where English had to name a subject the Latin leaves unexpressed, with what decides it. Where two readings construe, both are stated.

1. 0519C, Frater. Per hoc removet, quod non malo animo ei notificaverit. Rendered "By this he removes it, that he had made it known to him…". The subject of removet is Paul (he is the one writing); the subject of notificaverit is Sosthenes (the frater just named, and the man qui inter eos versabatur in the preceding lemma); ei is Paul. ⚠ A second reading construes: notificaverit = Paul, ei = an unnamed recipient, i.e. Paul removing the suspicion that he reported the Corinthians' faults out of ill will. The rendering leaves both open by keeping "he … him" and adds nothing.
2. 0520C, Dei virtus. Dum occisus diabolum vicit, et hominem liberavit… vel virtutem dat credentibus. Supplied "he" = Christ, from crux, id est mors Christi eight lines above and from Dei virtus itself (I Cor 1:18, Christum … Dei virtutem, 1:24). Not in doubt.
3. 0521A, VERS. 19. Ideo non misit me in sapientia, quia praedixerat quod eam non reciperet… Supplied "he" = God (the Scriptum est enim of Isa 29:14 quoted at I Cor 1:19 is God's own speech, and misit answers Non enim misit me Christus of 1:17). ⚠ The sender at 1:17 is Christ, so "he" could be Christ throughout. Both construe; the English says only "he", which is what the Latin says.
4. 0521B–C, VERS. 27. Qui etsi primum elegerit pauperes … non tamen relinquit — "he" = God, from Deus humilitatem proponat and Venit enim humilis Deus in the same sentence chain. Not in doubt.
5. 0524A–B, the two AUG. notes on Templum. Habitat itaque in singulis Deus names its own subject; the following Capitur autem habitans ab aliis amplius and in eis occulte agit inherit God. Not in doubt.
6. 0524D, VERS. 23. Ut sicut haec nobis, ita nos Christo subjiciamur — the first clause has no verb; "are subject" is carried over from subjiciamur, within the same sentence. Flagged as an ellipsis resolution, not a supplied subject.
7. 0523A, VERS. 8. tamen Deus magis per hunc quam per illum — no verb; "gives" carried over from dant in the same sentence. Same class as 6.
8. 0523D. Qui autem aurum, argentum, lapides pretiosos de utroque igne securi sunt — no verb in the relative clause; "have built" carried over from superaedificaverunt in the immediately preceding clause. Same class as 6.
9. 0527B. vel tradit, ut Spiritus sanctus … salvus sit eis — subject of tradit is Paul (the Apostolus of Sic Apostolus … Satanae tradit four lines above); the object supplied as "him" is the fornicator of I Cor 5:5. Both from the lemma above, not the sentence.
10. 0527D. tollerentur potius, ne perverse malos evitando quisque … mittat ad gehennam — subject of tollerentur is mali (from auferendos esse malos). Not in doubt.

4. PLATE REQUESTS — column, printed reading, why, and what would settle it

No [sic:] was fired on any of these. Each is raised here as a plate candidate and nothing more. The reasoning is the volume's own: PL 114's marker sweep returned 65% false, and both recent works in this volume that refused letter-level [sic:] came back 36/37 and 14/14. Three of the six below are non-words, which addendum A says are worth raising and have been right every time; three are real words, which addendum A says are Migne's until a plate says otherwise, and they are rendered as printed.

Non-words (hapax in the whole corpus — a grep -r over src/latin/ returns one hit each, this work's)

1. 0519D · ahos (VERS. 10, invitat ahos ad horum similitudinem). Not a Latin word. Certainly alios: the sentence contrasts hos (the praised) with the rest, and horum stands three words later. Englished for its evident sense ("the others") and left unmarked. Settled by: reading p.265 col. 0519D. If Migne prints alios, this is our transcription's l→h; if he prints ahos, it is his type.
2. 0520C · devidebant (per hominum se nomina devidebant). Not a Latin word; dividebant is certain (divisus est is the lemma being glossed, twice, on the same column). Englished as "were dividing" and left unmarked. Same plate request, col. 0520C.
3. 0519C · suffraganeiis (VERS. 2, omnibus scribit suffraganeiis Corinthiis). Not a Latin form; suffraganeis is certain. Englished as "suffragan" and left unmarked. ⚠ All three of these are on p.265, the one page in this range with a plate read on record, and that read was a checked zero. It is worth knowing whether these three forms were within that read's scope — if they were, all three are Migne's and the English above is a silent repair that should be reverted to carried type. ⚠ Note also that p.265's right column clips the right margin in our scan (measured on per ipsum Ch, praedic, hominibu, conscient, diliger). No truncated word occurs in our chunk 0000 — I looked for the class and found none; the clipping falls in the Romans work that shares the page, whose text ends mid-page at 0519C.

Real words that read oddly — rendered as printed, no marker (addendum A)

1. 0524A · voluerit for expected violaverit (sed si quis voluerit templum quod vos estis, disperdet illum). I Cor 3:17 reads si quis autem templum Dei violaverit, disperdet illum Deus, and the gloss is expounding that verse word for word: with voluerit the clause has no sense at all ("if any man shall have wished the temple…"). This is the strongest candidate in the range. It is a real Latin form, so under addendum A it is Migne's until the plate says otherwise, and I declined the [cj:] as well as the [sic:]: an [cj:] asserts that Migne printed voluerit, which is exactly what has not been checked, and the loss of -ia- is precisely the shape our digitization produces. Rendered literally. Settled by: col. 0524A. ⛔ If Migne prints violaverit, this needs a tei-patch, not a marker, and the English sentence must be re-rendered ("shall have violated").
2. 0528C · suo potestate judicis (nolo redigi suo potestate judicis). suo cannot agree with potestate. I Cor 6:12, the verse being glossed, reads ego sub nullis redigar potestate — so sub is what the construction wants, and suo/sub is a plausible misreading. Rendered "under a judge's own power", carrying suo as "own" rather than dropping it. Settled by: col. 0528C.
3. 0528D · sanctificatus … Spiritus (Non potest non esse sanctum corpus, quo sanctificatus utitur Spiritus). sanctificatus agrees with Spiritus, which yields "the sanctified Spirit" — theologically impossible in a gloss arguing that the Spirit is God (0529A: si enim non esset Deus, templum utique nosipsos non haberet). sanctificator or sanctificans would give the required sense. Rendered as printed ("which the sanctified Spirit uses"), because it is a real, well-formed word. Settled by: col. 0528D.

Punctuation / structure candidates

1. 0522D · the full stop in sed spiritui. Dei subjectus certissime ac fideliter judicat. The stop splits what wants to be sed spiritui Dei subjectus … judicat ("but, subject to the Spirit of God, judges most certainly and faithfully"). Pattern 8 obeyed: the stop is preserved and the English is broken in the same place ("…but to the spirit. Subject to God, he judges…"). Settled by: col. 0522D, whether the plate prints the point.
2. 0524A–B · Habitare autem ideo in animalibus hominibus et parvulis … quia in eis occulte agit. An infinitive with no governing verb (dicitur? est?). No verb was supplied; the infinitive is rendered as a verbal noun ("And his dwelling in sensual men … is for this reason, that…"), which adds only a copula that the Latin ellipsis licenses. Settled by: col. 0524B — a word may have been lost.

⭐ Suspiciously TIDY forms — the §4 normalization class, recorded as negatives

Per brief §4 and addendum C, a form that reads too well in a rough passage is the only handle on the digitization's silent corrections. I read whole lines rather than target words throughout. Nothing in this range struck me as tidied. Two forms I examined and cleared: querebantur (0522D, for quaerebantur) and tollerentur (0527D, for tolerentur) are both attested medieval orthography (Pattern 9), not tidying and not defects — no marker, no crux beyond this line. Recording the negative because the class is only ever found this way.

5. Candidates RAISED AND REJECTED, with reasons

• personaliter legitur: examinatur (0522C). The sense the passage seems to want is impersonaliter — that examinatur is read as an impersonal. ⛔ Rejected as a conjecture and rendered "it is read personally". Migne prints a real word; the reading is arguable either way (the examinatur of I Cor 2:14 does have a personal subject in the second exposition Migne then gives, quod animalis est); and nothing licenses reversing a word to its opposite under Pattern 7a. Logged, not repaired.
• veritatis for veritas (0527C). ut sinceritas mundam vitam faciat, et veritatis omnem fraudem excludat — the pairing with I Cor 5:8's sinceritatis et veritatis makes veritas as a second subject very attractive. ⛔ Rejected: the genitive construes as printed (omnem fraudem veritatis, "every deceit of truth", with sinceritas still the subject), so there is no defect to mark and no need for a conjecture. Rendered on the genitive reading. ⚠ The rival — veritas as subject, "and that truth may exclude every deceit" — is named here and lost because it requires changing a printed letter, while the printed text construes.
• potentia rationalis (0522B), nominative where the ablative potentia rationali is expected after pro. ⛔ Rejected: a case mismatch with no English exponent (Pattern 9). Rendered "for the rational power of the soul itself"; no marker, no repair.
• Deo … judicat (0522C), where de Deo is expected. ⛔ Rejected as a marker candidate for the same reason: rendered as printed ("who judges God according to the phantasy of bodies"), logged here. A plate read at 0522C would settle whether the de is Migne's or ours; it is a two-letter loss, so it belongs to the same class as items 4–5 above and takes no marker.
• Column-band gaps. 0519C → 0519D → 0520B (no 0520A), and 0522D → 0523A across the chunk boundary. ⛔ Not reported as defects: runbook false positive 1. No text in this range is broken mid-word or mid-construction across an anchor. I checked every anchor in the range against the word it stands beside in the Latin (test 2b); three sit mid-construction and were placed mid-construction in the English on purpose: privilegio [0519C] nominis, necessitate [0520D] instante, hospitium [0527B] non praebuerunt.
• Migne's verse ordering. VERS. 5 is glossed at 0519D and VERS. 4 is not glossed at all; I Cor 4:8's sine nobis is glossed under VERS. 9, not VERS. 8. ⛔ Not reordered and not reported as an error — this is his arrangement (brief: "Migne's own printing is not yours to tidy"). Recorded so a blind reader does not raise it.
• Migne's own ( sic ) — none occurs in this range. Checked.

6. Sections read CLEAN, by name

Every paragraph below was read back against its Latin twin asking only test 5's question — does the English assert what he asserts — with the counts already known clean.

• 0000 — ARGUMENTUM (whole); CAPUT PRIMUM VERS. 1, 2, 3, 5, 10, 11, 12, 13, 14, 17, 18, 19, 21, 26, 27, 29, 30. Clean apart from ahos (0519D), devidebant (0520C), suffraganeiis (0519C), all §4 above.
• 0001 — CAPUT II VERS. 2, 6, 7, 8, 9, 10, 12, 13, 14, 15. Clean apart from the Lc VIII divergence (§2a.1), the homo addition at 2:15 (§2a.2), the personaliter crux and the 0522D full stop. The four [var:] of this range stand at 0522B, 0522D, 0523B and 0526C; two of them are in this chunk.
• 0002 — CAPUT III VERS. 2, 8, 9, 11, 12, 13, 14, 16, 17, 21, 22, 23, and the four Augustine notes at 0523C, 0524A, 0524A, 0524B. Clean apart from voluerit (0524A) and the infinitive at 0524A–B.
• 0003 — CAPUT IV VERS. 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 13, 14, 15, 19, 20. Clean apart from the Mt XXIII divergence (§2a.3) and the two immaterial I Cor 4:4 divergences (§2b). ⚠ The lemma Ne etc., unus, etc. at 0525C is two one-word lemmata with a gloss (Ab illo, id est meliori quam ille) whose antecedent is I Cor 4:6's alterum/alio, not anything in the gloss's own sentence. Rendered "Than that man, that is, better than he"; flagged because the construction is obscure and a blind reader will stop there.
• 0004 — CAPUT V VERS. 1, 2, 4, 5, 6, 7, 8, 10, 11 and the AUG. note at 0527C; CAPUT VI VERS. 1, 2, 4, 5, 8, 9, 11, 12, 13, 15, 16, 17, 18, 19, 20 and the AUG. note at 0528A. Clean apart from suo potestate (0528C) and sanctificatus … Spiritus (0528D).

Negation audit (test 2 / 2a), run on the trigger population

Every printed negative was matched to the single Latin word it attaches to. The sites where a negative sits on a verb with a quantifier, correlative or predicate nearby — 2a's exact trigger — are these, and each was resolved from the sentence's own structure:

• 0519C non malo animo … sed ex charitate: host is malo animo, not notificaverit. The sed clause (ex charitate) is a phrase, so the negative is on the phrase.
• 0520B si in nomine majorum non est gloriandum, nec in nomine illorum: host of the second negative is in nomine illorum, carried by the parallel.
• 0520C ita nec illius, sed Christi esse cognoscitur: host is illius, not cognoscitur — the point is whose the baptism is, not that it is unknown.
• 0521B non tamen relinquit sapientes, divites, nobiles: host is the verb; the correlative is etsi primum elegerit … non tamen relinquit, so the scope is the verb.
• 0524B nec ubique habitans, sed in templo suo: ⭐ host is ubique, not habitans — God is everywhere present but does not dwell everywhere. The English reads "yet not dwelling everywhere, but in his temple", with the negative on the quantifier. (Rendering it "not dwelling, but in his temple" would be the 8990 @0614B defect exactly.)
• 0526D Non solum ergo illum reum mortis, sed etiam inhaerentes illi, non esse immunes: the second negative is on esse immunes inside the sed etiam limb, not on demonstravit. Getting this wrong makes Paul deny the accessories' guilt.
• 0527D ostendit satis non temere, et quolibet modo, sed per judicium auferendos esse: host is the adverbial pair temere, et quolibet modo, shown by sed per judicium.
• 0528B nequaquam … funditus deserant, sed vel aliis … committant: host is the verb, and the sed limb confirms it.
• 0529A Non potest non esse sanctum corpus: two negatives, both carried ("It cannot be that a body is not holy"). Neither was collapsed into an affirmative.
• 0526B nec illa elevant, nec ista frangunt: both carried, each on its own verb.
• 0521B nec usus est eis · 0525C Nihil ille habet a baptista, nihil tu · 0526C non ut hoc nomen caeteris tollatur, sed ne … tribuatur: all carried, hosts named.

Test 1 (a marker must not swallow its word)

Not applicable in the ordinary way — no [sic:] exists in this range. The four [var:] are additive notes standing beside a complete rendering; stripping all four from the English leaves four grammatical sentences with nothing stranded. Checked mechanically.

Range: CAPUT VII–CAPUT XII, chunks 0005–0009. This is the middle stint of a three-agent split; the merge layer must reconcile this file with cruces-0000.md / the other range files.

No column in this range has been read at the plate. Per the ship test (brief §B and CLAUDE.md hard rule 8a) no [sic:] was fired anywhere in the five chunks — not one. Every suspicion about the type is logged below as a plate request instead. Three of the candidates below are real, well-formed Latin words that read oddly, which is exactly the population the 2026-08-23 batch measured as Migne's own, 4 times out of 4; a corpus-frequency argument was deliberately not used to convict any of them.

[var:] markers were fired, five of them, each resting on sources/vulgate/clementine-flat.txt and cited by book chapter:verse.

1. Conventions applied (and one that needed a ruling)

1a. ⚠ THE HEAD FORM — a conflict between the launch brief and this work's own shipped English

My launch message gave CAPUT VII. → CHAPTER SEVEN. I did not follow it. This work's already-shipped src/english/8981/0001.md prints ## CHAPTER II. against Latin ## CAPUT II., and Ruth (8968), the pilot named as the format model, does the same. The consistent rule visible in both is: a spelled-out Latin ordinal becomes a spelled-out English one (CAPUT PRIMUM. → CHAPTER ONE., which 8981/0000 also prints), a Roman numeral stays a Roman numeral. So this range prints CHAPTER VII. … CHAPTER XII.

✅ Checked after all three stints landed: the work is consistent. Chunks 0000–0013 now print CHAPTER ONE. (for CAPUT PRIMUM.) and CHAPTER II. … CHAPTER XVI. for the Roman forms, including CHAPTER XV. (cont.) at a chunk boundary. All three agents converged on the same rule independently. Nothing for the merge layer to fix here; recorded because the launch brief's example would have produced a second convention had it been followed literally.

2. Vulgate collation — THE DELIVERABLE

Every italic span in cols 0529C–0542D was checked in place against sources/vulgate/clementine-flat.txt. Below: first the divergences, then, in §2c, the lemmata checked and found to agree, because a findings-only list cannot be told apart from a report by an agent that never looked.

2a. Divergences that took a [var:] (5)

(Six rows, five markers — the last two rows are the single Matthew marker plus the 12:28 one.)

2b. Divergences found, logged, and deliberately given NO marker

Pattern 14 fires on divergence that changes the sense. These do not, and a marker on each would be spam on every page:

• 0530A Nam caeteris vs Vulg 1Cor 7:12 Nam ceteris — ae/e, orthographic.
• 0531A Tanquam non habentes sint vs 1Cor 7:29 tamquam — orthographic; same for Tanquam Judaeus (0534A / 9:20), Nunquid for Numquid (0533A, 0533C ×2 / 9:4, 9:8, 9:9), Quotiescunque for Quotiescumque (0539B / 11:26), Si quidem for siquidem (0532C / 8:5).
• 0532A De his vs 1Cor 8:1 De iis; 0534A His qui sub lege vs 9:21 iis qui sub lege — the his/iis pair, the same variant twice.
• 0532C Idolum nihil est in mundo vs 8:4 nihil est idolum in mundo — word order only.
• 0533B Sororem mulierem vs 9:5 mulierem sororem — word order only.
• 0534A Nam et si vs 9:16 Nam si — an added et, no sense change.
• 0534A Omnis enim vs 9:25 Omnis autem — connective only.
• 0531A Ego sum Pauli, ego Apollo, etc. vs 1Cor 1:12 Ego quidem sum Pauli : ego autem Apollo — compression of a verse the work already glossed in chunk 0000. Same at 0538B (Ego sum Pauli).
• 0538B Ut exhiberet sibi Ecclesiam gloriosam vs Eph 5:27 ut exhiberet ipse sibi gloriosam Ecclesiam — a dropped ipse and an inverted adjective; English has no exponent for the order and the sense is untouched.
• 0537B Si hominibus placerem, servus Christi non essem vs Gal 1:10 si adhuc hominibus placerem, Christi servus non essem — a dropped adhuc and inverted order. ⚠ 7a″ was live here: the familiar English is "if I yet pleased men". The English on the page says "If I pleased men", because that is what Migne prints.
• 0539C Qui manducat me, ipse vivet propter me vs Jo 6:58 et qui manducat me, et ipse vivet propter me — dropped connectives.
• 0540A Non inferetur cadaver tuum in sepulcro patrum tuorum vs 3Rg 13:22 in sepulchrum patrum tuorum — accusative in the Clementine, ablative in Migne, plus the ch/c spelling. ⚠ Logged rather than marked because the brief warns that a call turning on a single letter should not trust this copy to that precision (sources/vulgate/README.md).
• 0540A Invicem exspectare vs 1Cor 11:33 invicem exspectate (imperative → infinitive). One letter; same caution; and the gloss (Ut multorum oblatio simul celebretur) is indifferent to it.
• 0540A Quia Deus operatur. Omnia in omnibus. vs 1Cor 12:6 idem vero Deus qui operatur omnia in omnibus — Migne splits one relative clause into two lemmata and recasts qui as Quia Deus. This is the glossator's own segmentation, not a text divergence.
• 0541B Ut vult vs 1Cor 12:11 prout vult — a clipped lemma.
• 0539B Laudo vos in hoc? — not a divergence at all. Migne prints Vel sic distingue, i.e. he is proposing a re-punctuation of 11:22 (laudo vos ? in hoc non laudo). Rendered as the glossator's proposal, both question marks kept.

2c. Lemmata CHECKED AND FOUND TO AGREE with the Clementine

Recorded because a checked negative is worth having. All verified in place, not from the pre-scan's ✓.

• 1Cor 7 (chunk 0005): De quibus 7:1 · Hoc autem 7:6 · Unusquisque 7:7 · Uri / melius 7:9 · Quod si discesserit + Et vir uxorem 7:11 · Si quis frater 7:12 · Sanctificatus est + Alioquin 7:14 · Non enim + In pace 7:15 · Unde enim 7:16 · Divisit Dominus 7:17 · Circumcisus aliquis 7:18 · Circumcisio nihil est 7:19 · Servus + Magis utere 7:21 · Libertus 7:22 · Empti estis + Servi hominum 7:23 · De virginibus 7:25 · Necessitatem 7:26 · Hoc itaque dico + Tanquam non habentes sint 7:29 · Cogitat + Ut sit sancta 7:34 · Non ut laqueum 7:35 · Si quis autem 7:36 · Melius facit 7:38 · Mulier alligata + Quod si dormierit 7:39 · Beatior + Puto autem 7:40 · neque nubent, neque nubentur = Mt 22:30 / Mc 12:25 verbatim · Si me sciretis, forsitan et Patrem meum sciretis = Jo 8:19 verbatim · Putas inveniet fidem in terra? = Lc 18:8 from putas on.
• 1Cor 8–9 (chunk 0006): Scientia inflat 8:1 · Si quis autem diligit Deum 8:3 · De escis + Nisi unus 8:4 · Si quidem sunt dii multi 8:5 · Deus / Ex quo / Et unus 8:6 · Polluitur 8:7 · In Christum 8:12 · Non sum liber 9:1 · Operandi 9:6 · Quis militat ×2 + Quis plantat vineam, et de fructu ejus non edit? + Quis pascit + Et de lacte 9:7 (the Chrysostom lemma is verbatim 9:7) · Ne quod offendiculum 9:12 · Nescitis 9:13 · Ego autem + Bonum est + Gloriam 9:15 · non est mihi gloria + vae + Necessitas 9:16 · Si enim volens + Si autem invitus + Dispensatio mihi 9:17 · Ex omnibus + Omnium + me servum feci 9:19 · Omnibus omnia 9:22 · Nescitis + Bravium 9:24 · Non quasi 9:26 · Castigo corpus meum + Ne forte 9:27.
• 1Cor 10 (chunk 0007): Nolo enim + Quoniam patres + Sub nube 10:1 · In Moyse baptizati + Baptizati sunt 10:2 · Eamdem escam + Spiritalem 10:3 · Petra erat Christus / Consequente eos petra / Petra autem 10:4 · Neque idololatrae + Sedit populus + Ludere 10:7 · Neque tentemus Christum 10:9 · In figura + Scripta sunt + Fines saeculorum 10:11 · Videat ne cadat 10:12 · Tentatio vos + Possitis sustinere 10:13 · Propter quod 10:14 · Calix benedictionis + Panis quem frangimus 10:16 · Quoniam unus 10:17 · Videte + secundum carnem 10:18 · Quid ergo? + idolis immolatum sit aliquid 10:19 · Socios fieri daemoniorum 10:20 · An aemulamur Dominum + Omnia mihi licent 10:22–23 · Omne quod in macello 10:25 · Propter illum 10:28 · Ut quid enim 10:29 · Sive ergo manducatis 10:31 · Ecclesiae Dei 10:32 · Omnibus placeo 10:33. Plus the non-italic quotation at 0536C, Deus deorum Dominus locutus est, et vocavit terram a solis ortu usque ad occasum = Ps 49:1 verbatim (see §4 — this is an usque ad look-alike).
• 1Cor 11 (chunk 0008): Imitatores 11:1 · Laudo 11:2 · Viri + Caput Christus est 11:3 · Orans 11:4 · Quoniam imago 11:7 · Verumtamen + In Domino 11:11 · Omnia autem + omnia ex Deo 11:12 · Vos ipsi 11:13 · Si comam 11:14–15 · Hoc autem praecipio + Non laudans 11:17 · Oportet et haereses 11:19 · Dominicam coenam 11:20 · Unusquisque enim 11:21 · Ecclesiam + Aut Ecclesiam Dei + Laudo vos? 11:22 · Ego enim accepi + In qua nocte 11:23 · Fregit 11:24 · Quotiescunque + Donec veniat 11:26 · Indigne 11:27 · Judicium sibi manducat 11:29 · Ideo inter vos 11:30 · Dum judicamur + Corripimur 11:32.
• 1Cor 12 (chunk 0009): De spiritualibus 12:1 · Scitis 12:2 · Ideo notum + Nemo in spiritu + Dicit + Nisi in Spiritu sancto 12:3 · Divisiones 12:4 · Idem vero Deus 12:6 · Unicuique 12:7 · Sapientiae 12:8 · Alteri fides 12:9 · Sermonum 12:10 · Unus atque idem spiritus + Dividens singulis 12:11 · Ita et Christus 12:12 · Etenim + in uno spiritu 12:13 · Nam et corpus 12:14 · Si totum 12:17 · Quod si essent omnia unum 12:19 (clipped before membrum, otherwise verbatim) · Aut iterum caput pedibus 12:21 · Inhonesta sunt 12:23 · Honesta autem + Sed Deus temperavit 12:24 · Et si quid patitur 12:26 · De membro 12:27 · Et quosdam quidem + Primum + apostolos + Secundo prophetas + Genera linguarum 12:28 · Excellentiorem viam 12:31.

3. PLATE REQUESTS — every suspicion, none of them marked

Nothing here fired a marker. Each names the column, the printed reading, why it is suspect, and what would settle it. A [cj:] was considered for the first three and withheld: it too asserts what stands on the plate, and our Latin is Corpus Corporum's transcription, not the plate.

1. 0541D — Memora, unumquodque eorum sicut voluit (lemma at VERS. 18, glossing 1Cor 12:18). The Clementine reads posuit Deus membra, unumquodque eorum in corpore sicut voluit, and the gloss beneath it is about the members of the body (Ut nihil desit illi: ita et in Ecclesia diversas personas). Memora is a real Latin form (imperative of memoro) and construes with nothing here. Carried into the English as it stands, inside the already-italic lemma, no second marker (Pattern 10's nested-italics rule). Settled by: one plate read of col. 0541D. If Migne prints Membra, this is a transcription corruption and the English should be re-cut to "The members, every one of them as he would."
2. 0542A — qui cum honesti sunt egestate et habitu, aliquando tamen vitae mundioris sunt. The tamen demands a privative in the first limb, and the whole paragraph glosses inhonesta sunt — the expected reading is inhonesti. honesti is a perfectly real word, so rendered as printed ("though they are comely in poverty and in habit, are nevertheless sometimes of a cleaner life"), which is the sentence Migne's words make. Settled by: one plate read of col. 0542A.
3. 0535B — Solet res, quae significatur nomine rei quam significat nominari. The Augustinian commonplace runs solet res quae significat … nomine rei quam significat nuncupari; the printed passive significatur reverses which term is named after which, in a sentence whose next clause (Petra. Id est Christus, sequebatur) needs the active. Rendered as printed. Settled by: one plate read of col. 0535B.
4. 0530A — quia potior quod a falsariis dicitur esse appositum (inside the (AMBR.) note). Verbless: potior has no copula and the relative quod has no clear antecedent. Rendered "because he is the mightier, which is said to have been inserted by falsifiers", i.e. the copula is supplied because English cannot omit it, and nothing else. Very possibly Migne prints potior est with a full stop before Quod. Settled by: col. 0530A.
5. 0542B/C — Nunquid omnes pro, etc. (lemma at VERS. 29). pro, is a word cut in the middle. Rendered by convention 3 as the phrase the fragment stands for, Are all prophets, etc., from 1Cor 12:29 Numquid omnes prophetae?. It is not at a column edge, so the margin-clipping caution does not obviously apply. Settled by: col. 0542C — is it a compositor's break, Migne's own compendium, or ours?
6. 0531D — nec eis verecundiam numerositas inferre. numerositas is nominative and cannot be the object of debeo; the sentence is an anacoluthon as printed. Rendered "nor ought their multiplicity to bring shame upon them", supplying no word. Settled by: col. 0531D — Migne may print numerositate or numerositatem.
7. 0538C — Non deserentis solertia requiratur, sed praecipientis auctoritas attendatur. disserentis ("of one arguing") would give the expected contrast with praecipientis. deserentis is a real word, so rendered as printed ("of one who forsakes"), no conjecture in the text. Settled by: col. 0538C.
8. 0541D — the printed opus est?. — a question mark followed immediately by a full stop. Both marks reproduced (Pattern 8: never delete). Settled by: col. 0541D.
9. 0537C — quasi. Etsi de aliis culpo, laudo tamen de hoc: ironia. A bare quasi closed by a full stop, mid-sentence in sense. Reproduced as printed ("as if. Though I blame concerning other things…"). Settled by: col. 0537C.
10. 0536A — Vel. modo loquor infirmis, ut supra prudentibus. Vel closed by a full stop and followed by a lower-case modo, where the work's other seventeen Vels take a comma. Reproduced as printed ("Or. now I speak to the weak…"). Settled by: col. 0536A.
11. 0542A — VERS 23.--, with no full stop after VERS, unique in this range against the 124 other VERS. n.-- addresses of this range. Reproduced verbatim. Settled by: col. 0542A.
12. 0535B–0536D — verse addresses out of step with the Clementine's versification. Migne addresses In figura as VERS. 10 and Scripta sunt as VERS. 11, where the Clementine has both inside 10:11 (10:10 being Neque murmuraveritis). Likewise VERS. 22 at 0536D carries An aemulamur Dominum, which the Clementine numbers 10:22 together with Omnia mihi licent — glossed here as VERS. 23. Order and numbers reproduced exactly (brief: Migne's own printing is not ours to tidy). This is a numbering question, not a defect, and needs no plate read.

Candidates examined and REJECTED, with reasons

• caeteris · Tanquam · Nunquid · Quotiescunque · coenam · charitas · Moyse — Migne's ordinary orthography, Pattern 9. Not defects, not marked, not raised.
• spiritalem (0535A) beside spiritualibus (0540A) — both attested; the spirital-/ spiritual- pair is the Vulgate's own at 1Cor 10:3–4. Rejected as a Pattern 9 item.
• septemviram (0531D) — a rare word, but well formed and exactly right for the Sadducees' woman of Mt 22:25–28. Rendered "the woman of seven husbands". No plate request.
• theosebia (0540D) — a transliterated Greek noun, printed in italics by Migne and kept in italics untranslated, glossed by the sentence itself. Not a defect.
• propensiorem (0529D) / propensiorem poenam (0535C) — the same real word twice in the range, in two senses; no anomaly.
• de pseudo gloriantes (0535D) — a bare pseudo used substantivally. Rendered "the false ones", not expanded to "the false apostles", although pseudoapostolorum / pseudoapostolis stand twice within two columns. Expanding would be a supply, and the substantival pseudo is attested medieval usage. Raised here so a reader is not surprised; not a plate request.
• A comma→semicolon sweep — explicitly not run (brief, Test 3's caveat): Migne prints plenty of semicolons in this range himself.

4. usque ad — decided one occurrence at a time (Addendum D)

Six occurrences in the range. Five are the abridgment formula (each has an etc., before it and a resumption tail after) and were rendered as far as with no verb supplied: 0533A (Bede, on 9:4) · 0534A (Jerome, on 9:20) · 0536B (Bede, on 10:16) · 0536B (Augustine, De doctr. Christ., on 10:20) · 0538A (Augustine, De Trin., on 11:7) · 0539A (Ambrose, on 11:21). That is six formulae.

The look-alike: 0536C vocavit terram a solis ortu usque ad occasum — the words are Ps 49:1's own and mean "unto"; rendered "unto the going down thereof". It sits four lines from a genuine formula and is exactly the trap Addendum D names.

Each resumption tail was parsed as strict prose. The tails carry their own connectives unchanged — 0533A's potestas igitur est, non jussio Domini keeps both the igitur and the non; 0538A's ut male viventes bene loquendo turpitudinem suam contegant keeps its purpose clause.

5. Supplied pronouns and subjects — every one, named

Sites where English cannot leave the subject or object unexpressed and the Latin does not print it. Where two readings both construe, both are stated.

1. 0530B, quo forsan possit fieri fidelis — supplied "she". The antecedent is the infidelem of the preceding clause, and 1Cor 7:12–13, the verse being glossed, makes that the unbelieving wife (Si quis frater uxorem habet infidelem). fidelis is common in gender, so the Latin does not decide it; the lemma above does. ⚠ If the plate's Nam caeteris gloss is read as covering 7:13 as well (the unbelieving husband), "he" would construe equally. Flagged, not hidden.
2. 0530C, sic et tu exspecta — "thou", singular, per Pattern 17. Not a supply, a number call: tu is printed.
3. 0532A, nec sufficienter scire dicit — rendered "and says that they do not know sufficiently." The rival is "nor does he say that they know sufficiently" (taking nec on dicit rather than on the infinitive). Both construe. I took the first because the sentence's other limb is Quos primum arguit — he is reproving them, which is an assertion about them, not a withholding of one. The rival is not obviously worse.
4. 0533C, Curae quidem sunt; verum non ita, ut de iis legem instituat — rendered "They are indeed a care; but not so, that he should establish a law concerning them", with no "to God" supplied, although 1Cor 9:9 prints cura est Deo. The dative is not on the page here.
5. 0538A, Quia in aliis ubi est, potest velare — supplied "he" = the vir of the lemma Orans (1Cor 11:4 Omnis vir orans). Secure.
6. 0541C, Non illius persona contemnenda vel praeferenda — supplied "his". illius has no printed antecedent in the sentence. Two readings construe: the individual believer who received a gift (which is what the paragraph is about — quia unus et idem in omnibus operatur), or the Spirit himself. I took the first, because the parallel clause forbids giving God's glory to men, so the person at issue is a man's.
7. 0541D, Ut nihil desit illi — supplied "it" = the body, from corporis in the VERS. 17 gloss immediately above.
8. 0542B, ut non minus pro alio quam pro se — supplied "it" (the member), not "each one"; the subject of the whole period is membra.
9. 0530A, quia potior [est] — the copula supplied, see plate request 4.
10. 0532D, Noster vero homines deos facit — rendered "But ours makes men gods", keeping Migne's bare Noster; the antecedent is Deus noster six words earlier.

6. Punctuation, and the one deliberate departure

Marks were compared band by band against the Latin twin, colons, semicolons and question marks separately, across all 54 column bands in the five chunks. Question marks are at exact parity in every band (0005: 2 · 0006: 4 · 0007: 4 · 0008: 3 · 0009: 2). Colons and semicolons are at parity in 53 of 54 bands.

⚠ The single deviation is deliberate and is Pattern 8b. At 0539C Migne prints Quaeritur quomodo hoc, cum Dominus dicat: Qui manducat me, ipse vivet propter me? — an indirect question closed with ?, which English syntax cannot take. Per 8b it was recast into direct form behind a colon: "It is asked: how is this, when the Lord says: He that eateth me, he himself shall live by me?" That adds one colon to that band. Nothing was supplied or deleted; the sentence type and the mark both survive.

Marks preserved against the ear: opus est?. (0541D) · quasi. (0537C) · Vel. (0536A) · the space before the paren in Ego sum Pauli ) (0538C band, VERS. 19; the paren-space is Migne's) · the .-- after etc. at 0529B (Hoc autem, etc.--Dicet aliquis).

7. Column anchors

All 54 anchors present, verbatim, in order, and checked against the word they stand beside in the Latin, not merely for presence (brief 2b). Where English word order made an exact seat impossible the anchor was kept beside its own Latin word rather than moved to a tidier place:

• 0533C — Latin pastores [0533C] erant apostoli; the English was written "soldiers, and husbandmen, and shepherds [0533C] were the apostles" rather than the smoother "the apostles were … shepherds", precisely so the anchor keeps its seat.
• 0530C — Latin confirmatur [0530C] conjugium; English "the marriage is confirmed [0530C] through the knowledge of God", the anchor kept immediately after confirmatur's English.
• 0542C — Latin omnium [0542C] judices; English "and judges of all [0542C]", the anchor seated after omnium.
• 0538C — Latin in revelatione [0538C] Evangelii; English "in the revelation [0538C] of the Gospel".
• 0540A — Latin qui vivus [0540A] potuit terreri; English "who while alive [0540A] could be terrified".

8. Read clean — named

So that this report can be told apart from one by an agent that did not look, the following were read in full against their Latin twins and returned nothing to raise:

• 0005 — the whole of CAPUT VII except the four sites in §3 (nos. 4, 6) and §5 (nos. 1–2). In particular the long Augustinian marriage argument at 0529C (Non enim quia incontinentiae ibi malum est …) was read three times for polarity: five printed negatives, all present, each on its printed host, and the non propter illud malum clause was re-cut so the negative sits on the phrase and not on the verb, per 2a's trigger condition (non … sed correlative with a plausible verb adjacent).
• 0006 — CAPUT VIII entire and CAPUT IX to 9:27. The Trinitarian Ex quo passage at 0532D (alioquin jam non omnia per Patrem, nec omnia per Filium…) was checked negative by negative and clause by clause: clean. The five sigla all carry through in place.
• 0007 — CAPUT X entire. The Ps 49:1 quotation and the usque ad look-alike beside it (§4) were the only things worth recording; the long Augustinian passage on singing and eating at 0537A–B is a sustained singular address and was rendered thou throughout, consistently with the rest of the range.
• 0008 — CAPUT XI entire, including the head-covering material at 0537C–0538C and the eucharistic material at 0538D–0540A. Read specifically for 7a″ conformation pressure, which is at its worst here (Hoc est corpus meum is not on this page but its neighbours are); no lemma was conformed. The one real find is the 11:14 [var:], which is a conformation in the source, not in us.
• 0009 — CAPUT XII entire. The body-and-members similitude at 0541C–0542B was read for supplied subjects (see §5, four of the ten sites are here).
• Marker hygiene: verify-english.mjs 8981 passes; scan-raw-markers.mjs reports no marker rendering as literal brackets; frontmatter is byte-identical to each Latin twin; [n:] counts are 1 / 5 / 3 / 3 / 2 as the frontmatter declares, counted before and after writing; no [nt:] was needed, since all fourteen notes in the range are attribution sigla or citations, not editorial prose.
• What the fourth test asked (does the English assert what he asserts, not merely contain it): each paragraph was read back against its Latin with the counts already known clean. The three paragraphs where a smooth English reading was actively suspicious were §3's nos. 2, 3 and 7 — all three read better under the conjecture than under the print, which is why all three are logged and none is repaired.

Third and final stint of this work: CAPUT XIII–XVI. Latin twins src/latin/8981/0010.md–0013.md. To be merged with cruces-0000.md and the middle stint's file.

Counts, checked before and after writing each chunk:

Colon/semicolon bands were compared against the Latin twin per Brief test 3 and brought to exact parity (markers excluded): five semicolons of mine were converted back to Migne's commas in 0012, two in 0010, two in 0011, and one semicolon of his that I had softened was restored in 0012 (@0549A per charitatem fiet;) and one in 0013 (@0549D absorpta es;).

§0 · ⛔ THE MARKER POLICY AS I APPLIED IT — read this before reading anything below

No column in 0543A–0550D has been read at Migne's plate. Our Latin is Corpus Corporum's transcription. I therefore fired ZERO [sic:], and every non-word and every split run I found is logged below as a plate request, not as an accusation against Migne's type. Three of the four candidates in my range are word-division or letter-loss defects of exactly the shape Corpus Corporum manufactures, and two of them stand eleven words apart in the same paragraph (@0546C ves tra, and min re eight lines above at the same column band), which is the clearest internal evidence I have that the fault is the transcription's rather than the compositor's — one plate can set a broken sort, but two in one short paragraph is a scanning signature.

How the four are handled in the English. Pattern 13a clause 1 governs — words lost mid-sentence, reading certain from context → restore silently, log the crux — because the alternative (carrying the broken run in bare italics) would either invent an italic span the Latin does not have, breaking span parity, or require the [sic:] this range may not fire. Each restoration is named below with the printed run, so nothing is concealed and a plate read can undo any of them in one edit.

[var:] — four fired, and each stands on an unread column. plate-gate.mjs gates [var:] as well as [sic:], so this work owes plate reads on 0543A, 0546A, 0548B and 0550C before it can be marked ours. I fired only where the divergence is multi-word, i.e. where no single-character transcription slip could have manufactured it — the class the gate's own rationale (judicabit→judicavit, 9003) exists to catch. Every single-letter or single-word divergence I found is in §2 with no marker, however tempting.

§1 · THE VULGATE COLLATION — every lemma in 0543A–0550D, walked

Source: sources/vulgate/clementine-flat.txt. This is the deliverable. The launch brief warned that CAPUT XIII and XV are the two stretches whose English every reader knows by heart, so each lemma below was rendered from Migne's Latin first and only then compared.

1a · DIVERGENT, marked [var:] (4)

1b · DIVERGENT, and I took NO marker — with the reason in each case (13)

These are the ones a [var:] would have over-claimed. Every one is a real divergence and every one is here so a later reader can overrule me.

• 0543B · Tunc cognoscam. — Vulg 1Cor 13:12 tunc autem cognoscam. A dropped autem in a clipped lemma; no sense is at stake. No marker.
• 0543B · sicut ipse est, — this is not the 1 Cor 13:12 lemma (sicut et cognitus sum); it is the gloss's own italicised echo of I Jo 3:2 videbimus eum sicuti est, with no siglum printed. Marking it would assert a divergence in a citation Migne never made.
• 0543A · non novi vos. Neither Mt 25:12 (nescio vos) nor Mt 7:23 (numquam novi vos). ⚠ Declined because Migne prints no siglum here and the words sit inside the glossator's own sentence (quibus non dicetur … sed non novi vos) as a paraphrastic allusion, not a quoted proof-text. Note that at 0545D the same work quotes Nescio vos exactly and then glosses it id est non novi vos esse meos — which is positive evidence that non novi vos is the glossator's idiom, not his text.
• 0544A · Et sine voce. — Vulg 1Cor 14:10 et nihil sine voce est. ⛔ The most dangerous fragment in my range, and the reason it is here: convention 3 would expand it to the phrase it stands for, and that expansion supplies a nihil Migne does not print. I rendered the printed words only — "And without a voice." — per 7a. No marker.
• 0547D · Si omnino, etc. — same trap. Vulg 15:29 si omnino mortui non resurgunt. Rendered "If at all, etc."; no negative supplied.
• 0547D · Propter vestram. — Vulg 15:31 per vestram gloriam. A genuine preposition divergence, and I declined the marker because Migne adjudicates it himself on the same line: Vel: Per vestram gloriam. The reader meets both readings without our apparatus. Rendered so the divergence survives in English too: "On account of your glory. Or: By your glory."
• 0546D · Deinde qui sunt. and the Alia littera beside it — Vulg 15:23 deinde ii qui sunt Christi, qui in adventu ejus crediderunt. The lemma drops ii; Migne then prints his own variant notice, Alia littera: Deinde hi qui sunt Christi in adventu ejus. This is Migne's apparatus, not a defect — reproduced, unmarked.
• 0547B · Novissime autem inimica destruetur mors. — Vulg 15:26 Novissima autem. ⚠ The richest declined candidate in the range and the one I most nearly marked. Migne prints the Vulgate form as the verse lemma (VERS. 26.-- Novissima autem.) and the adverbial form in the re-quotation, then expounds the adverb twice more (sed novissime, id est post omnia destruetur mors; Novissime. Quia non erit aliquid quod destruat). Three internal attestations. Declined anyway, on my multi-word rule: novissima→novissime is one letter, and one letter is precisely what the digitization manufactures. The Vulgate form stands four words away on the same line, so the reader is not misled. ⭐ Plate request 0547B is the way to settle it; if the plate reads Novissime, this becomes a [var:].
• 0548C · Quem visurus sum ego ipse, et non alius. (Job 19:27) — the Clementine has quem visurus sum ego ipse, et oculi mei conspecturi sunt, et non alius. An elision inside a quotation, not a variant reading; the negation is intact. No marker.
• 0548B · Corrumpunt bonos mores, etc. — Vulg 15:33 corrumpunt mores bonos colloquia mala. Word order only (bonos mores / mores bonos). No marker.
• 0549C · Media nocte clamor factus est: ecce sponsus venit. — Mt 25:6 Media autem nocte … Ecce sponsus venit, exite obviam ei. A dropped autem and a truncation at the point Migne stops. No marker.
• 0548D · Et eadem caro. — Vulg 15:39 Non omnis caro, eadem caro. ⛔ Declined deliberately, and this is the one where declining costs most, because what Migne's lemma lacks is the Vulgate's negation. I did not fire a [var:] because the shape of the divergence — an opening Et where the verse has Non omnis caro, — is exactly what a transcription drops or supplies at a lemma head, and a [var:] would publish a claim about Migne's page on that evidence. I also did not supply the negative in the English: the lemma reads "And the same flesh.", flatly. ⭐ Plate request 0548D.
• 0550D · Spiritum meum et vestrum, etc. — Vulg 16:18 refecerunt enim et meum spiritum, et vestrum. Same words re-ordered into a citation head. No marker.

1c · CHECKED AND AGREEING WITH THE CLEMENTINE — named, not summarised

Verified verbatim against clementine-flat.txt at the verse glossed, allowing only Migne's standing orthography (charitas/caritas, coelestis/caelestis, nunquam/numquam, spirituale/spiritale, Hierosolymis/Jerusalem), which is Pattern 9 and not a divergence:

1 Cor 13 — Si linguis (13:1) · Charitatem autem (13:1) · Velut aes sonans (13:1) · Et si habuero prophetiam (13:2) · Noverim mysteria (13:2) · Scientiam (13:2) · Charitas patiens est (13:4) · Non aemulatur (13:4) · Non inflatur (13:4) · Non quaerit (13:5) · Omnia credit · Omnia sperat · Omnia sustinet (13:7) · Charitas · Nunquam excidit (13:8) · Evacuabitur quod (13:10) · Nunc · Tunc autem facie ad faciem · Speculum · Aenigma (13:12) · Fides, spes · Major autem horum est charitas (13:13).

1 Cor 14 — Qui enim loquitur lingua (14:2) · Nunc autem (×2) · linguis loquens · Aut in revelatione · Aut in scientia · Aut in prophetia · Aut in doctrina (14:6) · Tamen quae sine (14:7) · Etenim si incertam (×2, 14:8) · Si ergo nesciero (14:11) · Sic et vos (14:12) · Spiritus meus orat · Sine fructu (14:14) · Orabo spiritu · Orabo (14:15) · Amen (14:16) · Gratias ago Deo (14:18) · Sed malitia parvuli estote (14:20) · In lege · Aliis linguis · Et nec sic exaudient (14:21) · Itaque linguae in signum · linguae sunt in signum (14:22) · Si ergo conveniat (14:23) · Quid ergo · Unusquisque · Psalmum habet · Doctrinam (14:26) · Aut ut multum tres (14:27) · Sibi autem loquatur (14:28) · Prophetae duo (14:29, autem dropped — clipping) · Quod si alii (14:30) · Non enim est dissensionis Deus · Sicut in omnibus (14:33) · Turpe est (14:35) · Si quis videtur propheta · Cognoscat (14:37) · Si quis autem ignorat (14:38) · Nescio vos (Mt 25:12, exact).

1 Cor 15 — Per quod et salvamini · Nisi frustra credidistis (15:2) · Resurrexit tertia die (15:4) · Visus est Cephae (15:5) · Abortivo (15:8) · Non sum dignus (15:9) · Gratia autem Dei · Abundantius · Non ego, sed gratia (15:10, autem dropped) · Et sic credidistis (15:11) · Si autem Christus (15:14) · Invenimur autem · Adversus Deum (15:15) · Quod si Christus non (15:17) · Quoniam quidem per hominem (15:21) · Et sicut · sicut (15:22) · Unusquisque · Qui in adventu ejus crediderunt (15:23) · Deinde finis · Cum evacuaverit (15:24) · Oportet autem illum regnare · donec (15:25) · Oculi nostri ad Dominum Deum nostrum, donec misereatur (Ps 122:2, exact) · Novissima autem · Inimica (15:26) · Praeter eum qui subjecit ei omnia (15:27) · Ut sit Deus omnia · Ut sit Deus · Omnia · In omnibus (15:28) · Pro mortuis (15:29) · Per vestram gloriam (15:31) · Si secundum hominem · pugnavi · Manducemus et bibamus, cras enim moriemur (15:32, exact) · Nolite seduci (15:33) · Evigilate · evigilate · justi · Nolite peccare (15:34) · Sed dicet aliquis · Insipiens (15:35–36) · Insipiens tu: quod seminas (15:36, Migne's colon for the Clementine's comma) · Deus autem dat · Unicuique seminum (15:38) · Coelestia (15:40) · Stella enim a stella differt (15:41) · In domo Patris mei mansiones multae sunt (Jo 14:2, exact) · Seminatur (15:42) · Corpus animale · Surget spirituale (15:44, corpus elided) · Si est corpus · si · animale est (15:44) · Homo (15:45) · Sed non · Sed non prius · quod est spirituale, sed quod est animale, deinde spirituale (15:46, word order + supplied est) · Primus homo · Secundus homo · de coelo · coelestis · Coelestis (15:47) · Hoc autem dico · Quia caro · Neque corruptio incorruptelam possidebit (15:50, exact) · Omnes quidem (15:51) · Immutabimur (15:52) · Absorpta est (15:54) · Ubi est, mors, victoria tua? (15:55 — same words, Migne merely setting mors between commas; the ? is his and the Clementine's alike).

1 Cor 16 — De collectis · Ita et vos (16:1) · Per (16:2) · Manebo vel etiam hiemabo (16:6) · Obsecro autem vos, fratres, nostis domum Stephanae et Fortunati (16:15, truncated at etc. before et Achaici) · ut et vos subditi sitis (16:16).

Two lemmata declined expansion on the 7a″ ground, and only those two: Et sine voce and Si omnino (§1b). Every other clipped lemma was expanded per convention 3 to the phrase it stands for (Si linguis → "If I speak with tongues"; Evacuabitur quod → "That which is in part shall be done away"; Tamen quae sine → "Yet the things which are without life"; Prophetae duo → "Let two prophets speak"; Sicut in omnibus → "As in all the churches"). ⚠ Complete one-word lemmata were NOT expanded — Scientiam → "Knowledge", Speculum → "A glass", Aenigma → "An enigma", Homo → "Man", Omnia → "All things", Inimica → "The enemy", Coelestia → "Heavenly" — because a complete one-word lemma and a clipped fragment are different classes (Brief, "NAME THE CONTROL'S CLASS"), and contrasting them manufactures an inconsistency that is not there.

⭐ The one place this work does treat the same bare lemma twice: Etenim si incertam @0544A, once bare and once with etc. Resolved per the standing ruling by expanding the stump, not contracting its twin: both read "For if the trumpet give an uncertain sound", the second carrying Migne's etc.

§2 · PLATE REQUESTS — column, printed reading, why, and what would settle it

None of these fires a marker. All four in 2a are restored silently in the English per Pattern 13a clause 1 and named here so the restoration is reversible.

2a · Non-word / split-type candidates (4) — corpus frequency says "an error exists", never whose

What would settle all four: one page. All are on PL 114 pp. 278–280 by the volume map in CLAUDE.md (patrologiaecurs06saingoog, PDF page = (column + 11) / 2 — 0546 → p. 278·5, so pp. 278–279; 0547 → p. 279; 0549 → p. 280; calibrate against the printed corner numbers before reading, per the standing rule). A second witness for PL 114 is not recorded in data/volumes.json the way PL 113's Gallica copy is; if our copy clips, look for one.

2b · Suspiciously TIDY forms — brief §4, the normalization class (recorded negatives)

The digitization also silently corrects. I looked for forms that read too well in rough company and found none I would raise in 0543A–0550D. Two I examined and cleared:

• 0549C · septemvira muliere ("the woman of seven husbands", of the Sadducees' riddle). A rare coinage, not a normalization; septemvir- is well formed and the sense is exact.
• 0546A · Incassum est fides — masculine/neuter adverbial incassum with feminine fides reads oddly, but incassum is an indeclinable adverb here ("in vain"), not an adjective. Nothing tidied. No crux.

2c · Punctuation Migne prints that I carried against the grain of English

Reproduced exactly, per Pattern 8; each is a candidate for a plate read only if someone doubts the transcription.

• 0542D/0543A · si adsit, habentur omnia? — a ? closing a plain statement ("if it be present, all things are had?"). Carried. This is his mark, and recording it is the point.
• 0546C · quod est causa mortis Et si in peccato tenetur — no stop before a capitalised Et. Carried: "which is the cause of death And if he is held in sin". Not repaired.
• 0543C · quod modo est secretum, Speculum. Est anima — a comma where a full stop belongs, immediately before a new lemma. Carried.
• 0544A · …prolationem et non significationem sciret — the paragraph ends with no terminal mark at all. Carried; my English ends unpunctuated too.
• 0546B · Gratia autem Dei Primum sola gratia and 0546B · Non ego, sed gratia, etc — a missing stop after an italic lemma, and an etc with no point. Both carried.
• 0544C · ut videret imagines mens Joseph illuminata est — no mark between imagines and mens, so two clauses run together. Carried: "that he might see the images the mind of Joseph was enlightened, that he might understand." ⚠ Reads as a defect in English because it is one in Latin.
• 0544D · VERS 20.-- and 0544B · VERS. 10-- — Migne's own irregular verse addresses (missing point after VERS, missing point after 10). Passed through verbatim, per the settled convention that VERS. n.-- is an address, not prose.

2d · Column-anchor placement (Brief 2b) — every anchor checked against the WORD it stands beside

All 31 anchors are verbatim and in order; I additionally checked position, not presence. The mid-construction ones and where they landed in English:

• 0543B before portat → "Patience, with an even mind, [0543B] bears the evils…"
• 0543C before sperando → "If by believing and [0543C] hoping we love…"
• 0543D before quamvis → "he understands a likeness, [0543D] albeit an obscure one"
• 0544A before prophetat → "he who [0544A] prophesies is greater"
• 0544B before linguam → "By myself I have shown [0544B] that a tongue avails not"
• 0544C before similitudines → "where the [0544C] likenesses of bodily things"
• 0544D before interpres → "Neither the Greek nor the Latin [0544D] translator"
• 0545A before qua → "not in revelation, [0545A] of which they are unworthy"
• 0545B before quo modo → "he begins to determine [0545B] in what manner"
• 0545C before verbum → "stands every [0545C] word"
• 0545D before tamen → "to glory [0545D] nevertheless attain" (English inverted so the anchor keeps its Latin neighbour)
• 0546A before non tenetis → "Which is, if [0546A] you do not hold"
• 0546B before labore → "not in dignity, [0546B] in labour and in preaching"
• 0546C before pseudoapostoli → "The [0546C] false apostles were denying" (split inside the noun phrase, the 8995 @0605B precedent, because Latin puts the verb first)
• 0547B before regni → "not a longer [0547B] continuance of the kingdom"
• 0547C before esse subjectus → "so by no reason can he [0547C] be subject to any"
• 0547D before Si omnino → placed before the lemma
• 0548A inside an italic span, before et bibamus → "Let us eat [0548A] and drink"
• 0549A before quia → "God is all in all, [0549A] because God is charity"
• 0549B before esse → "it is established [0549B] to be so"
• 0549C before factus est → "as if to say, [0549C] he was made"
• 0549D before de resurrectione → "the seven husbands, [0549D] for these men were thinking carnally about the resurrection"
• 0550A before quia corruptio → "shall not possess the kingdom of God: [0550A] because"

The remaining eight (0543A, 0546D, 0547A, 0548B, 0548C, 0548D, 0550C, 0550D) stand at paragraph or VERS. heads and are reproduced there.

⚠ Band gaps are not defects (runbook false positive 1): 0550 prints A, C, D in our text with no B, and 0546–0549 run A–D cleanly. Not reported as a hole, and no arithmetic was done over band letters.

## §3 · SUPPLIED SUBJECTS AND PRONOUNS — every "he/she/it/they" the Latin does not print, ## where two readings construe

The Brief calls this the dominant defect class. Named sites, with what decided each:

1. 0543B · Quia non eam aliena felicitas contristat — eam is charitas, from the lemma Non aemulatur above. Rendered "it". Decided by the lemma, not the sentence.
2. 0543B · Non dicit omnibus credit — "He does not say, it believes all men". The unexpressed subject of dicit is the Apostle (the work's standing subject for dicit / ait / subdit); of credit, charity, from the lemma Omnia credit. Both fixed by the lemma above.
3. 0544A · quando lingua loquens se non intelligebat — "when one speaking with a tongue did not understand himself". Indefinite; no person supplied beyond the participle's own.
4. 0544C · Ratio quare non debet quaerere linguas (0545A) — impersonal debet, rendered "the reason why one ought not to seek after tongues". ⚠ A second reading is available ("why he ought not", i.e. the Corinthian): both construe, and the choice is invisible in the English's force. Named here rather than chosen silently.
5. 0546B · Vel similis abortivo, quia sum minimus — ⭐ first person, and I kept it. The gloss slips out of dicitur / accepit into Paul's own voice mid-sentence. The easy English would have been "because he is the least", and it would have concealed the shift. Rendered "because I am the least".
6. 0547B · Quia non erit aliquid quod destruat — "there will not be anything which it may destroy". The subject of destruat is death, from the lemma Novissime and from postquam hoc mortale induat immortalitatem. ⚠ quod destruat also construes as "which may destroy [death]" — i.e. nothing left to do the destroying. Both readings work grammatically; I took the first because the paragraph's whole point is that death is destroyed last, which requires that it still had work to do until then. Named, not chosen silently.
7. 0547B · Inimica. Cui inimicamur: vel ipsa nobis — ipsa is feminine agreeing with mors. English has no exponent for that (Pattern 9), and rendering "or she to us" would personify Death on our authority. Rendered "or it to us".
8. 0547A · Sed hominibus collectis in angelis et hominibus omnis praelatio cessabit — parsed as hominibus collectis (ablative absolute) ∥ in angelis et hominibus … praelatio cessabit: "when men have been gathered in, among angels and men every prelacy will cease." ⚠ The rival parse takes collectis in angelis together ("men gathered among the angels"), which then strands et hominibus. The rival loses because it leaves a dative with nothing to govern it; recorded so the reader can weigh it.
9. 0548A · et non credis jam immortalis — the parallel Crederes patri iterum morituro requires a dative, and immortalis is not one. Rendered "and thou dost not believe one who is now immortal", supplying the pronoun the case cannot supply. Not marked: this is a case defect with no English exponent worth a [cj:], and Pattern 9 governs.
10. 0548B · a pseudo — Migne uses the Greek stem absolutely, where the work elsewhere writes pseudoapostoli (0546C, 0546B). Rendered "by the false one", not expanded to "false apostles": the plural would supply a number the Latin does not print.
11. 0549C · unde sadducaeis quidam non poterant respondere — quidam nominative ("certain men"), sadducaeis dative of the persons answered. Rendered "certain men could not answer the Sadducees". The rival ("certain of the Sadducees could not answer") needs sadducaeorum and is refused by the case.

§4 · NEGATION — test 2 and 2a, run site by site

Every printed non · nec · neque · nisi · nihil · nullus · nunquam · ne in 0543A–0550D was located, its host word named, and the host found in the English with the negative on it. The sites where a quantifier, correlative or predicate stood next to the verb — 2a's trigger population — are these, and each was resolved from the sentence's structure, not from what read well:

1. 0543D · Est autem aenigma non omnis, sed obscura allegoria. Host = omnis, not est. The sed correlative pivots on it. English: "an enigma is not every allegory, but an obscure one." (The wrong host gives "an enigma is not an allegory", which the next clause denies.)
2. 0546B · quia sum minimus, vel tempore et vocatione, non dignitate, labore et praedicatione. Host = the ablative group dignitate, labore et praedicatione, not a verb. English keeps the negative on the nouns: "whether in time and calling, not in dignity, in labour and in preaching."
3. 0546B · cum non praecederent nisi mala merita. non … nisi is a single exceptive. English: "since no merits went before save evil ones" — both negatives accounted for, none doubled.
4. 0547A · nihil horum aliquos terrenos vel coelestes habuisse ex se, sed ab illo. ⛔ The hardest site in the range. Host = nihil, the object of habuisse; aliquos terrenos vel coelestes is the accusative subject. The sed ab illo correlative proves the scope. English: "that none of these things did any earthly or heavenly beings possess of themselves, but from him from whom are all things." ⚠ The plausible-and-wrong reading floats the negative onto the subject — "that no earthly or heavenly beings had any of these" — which reverses what the sentence asserts (the point is that they had them, from another).
5. 0547B · excludit majorem manifestationem, non ampliorem regni permanentiam. Host = ampliorem … permanentiam, correlative with majorem manifestationem. Kept on the noun phrase.
6. 0547C · non tantum in omnibus … sed et omnis creatura fatebitur. Host = tantum; the sed et correlative fixes it. English: "not only in all … but every creature also."
7. 0545D · quia non perit unus de pusillis. Host = perit, with the quantifier unus adjacent — 2a's exact trigger. Checked: the negative belongs on the verb here (not one perishes), because there is no correlative pivoting on unus. English: "because not one of the little ones perishes."
8. 0548C · Sicut non omnis caro est ejusdem dignitatis. Host = omnis. English: "As not all flesh is of the same dignity" — deliberately not "all flesh is not", which English hearers read as universal denial.
9. 0549D · non omnes vivi reperti immutabuntur, sed soli sancti. Host = omnes, with sed soli pivoting. English: "not all who are found alive will be changed, but only the saints."
10. 0550C · ut in operibus suis non regnum Dei, sed istorum acquisitionem intueatur. Host = regnum Dei, correlative sed. English: "so that in his works he looks not to the kingdom of God, but to the getting of them."
11. 0545A · quibus … non patet vel placet veritas. One non over two verbs joined by vel. English: "the truth neither lies open nor is pleasing" — the negative distributed, not doubled.

No negative in this range was relocated, dropped, or supplied. In particular, the two places where expanding a clipped lemma would have added one (§1b: Et sine voce, Si omnino) were left unexpanded for that reason.

§5 · MARKERS FIRED, AND THE ONE [cj:]

• [sic: …] — none. See §0.
• [var: …] — four, §1a. All four owe plate reads (0543A, 0546A, 0548B, 0550C).
• [ed: …] — none. Nothing in this range is a demonstrated hole in the digitization rather than a corruption inside it; the four corruptions are §2a and are restored under Pattern 13a clause 1.
• [nt: …] — none. All five [n:] notes in my range ((Matth. XXIII), (ORIG.), (Ose. VI), (THEODOR.), (AMBROS.) — five in total) are citations or attribution sigla, not editorial prose. Passed through verbatim, untranslated, Latin abbreviations unanglicized. ⚠ Two of them sit inside an italic lemma (0543A, 0550C) and are reproduced inside the span, with no asterisks inside the marker.
• [cn: …] — none present. This range carries no recovered foot-of-page note. ⚠ The plate spot check for 8981 is owed or already recorded elsewhere; I could not verify it from my range and did not assume a sibling book's zero.

The one [cj:] — @0546B, Inveni

he says: I have found, etc. [cj: Inveni; read Invenimur, "we are found"]

Why a [cj:] and not silence, an expansion, or a [sic:]. Inveni is a perfectly good Latin form, so Pattern 12 has nothing broken to wrap and the [sic:] question does not arise. But the faithful English — "I have found" — asserts something Paul does not: the lemma being resumed is Invenimur autem ("but we are found"), printed eight words earlier at the head of the same VERS. 15, and the whole paragraph is about the Corinthians being found false witnesses. That is Pattern 18's founding case exactly: real word, misleading English, gloss additive.

The rival I rejected: treating Inveni, etc. as a clipped resumption cue and rendering it by convention 3 as "We are found, etc." That is very likely what it is — but convention 3 governs a phrase cut short, not a word cut short, and taking it would have put a first-person-plural verb on the page where Migne prints a first-person singular, with nothing to show for it. The [cj:] keeps Migne's word and puts the conjecture beside it, which is what the pattern exists to do. If a plate read shows Invenimur in full, the marker comes out and the reading is simply restored.

§6 · REGISTER, AND WHAT THE MERGE LAYER SHOULD CHECK

• "you", never "ye", in the gloss's own voice. ⛔ Zero occurrences of "ye" in chunks 0010–0013 — checked by grep, not by memory. This is the site where 8993's third stint drifted and twelve sites had to be converted; it did not happen here.
• "thou" is used, and only where the Latin is grammatically singular (Pattern 17), in exactly five places, all of them singular address: 0548A si pater tuus surgeret, crederes? … et non credis … Crederes patri (the interlocutor); 0548B Insipiens … qui non attendis quod quotidie vides and 0548C Insipiens tu: quod seminas (Paul's singular fool); 0550A Ne putares (the reader); 0549C–0550B Ubi est, mors, victoria tua? … Vicisti … victa es … Victoria tua (Death, addressed singular). Every plural vos / estis / vestra is "you" — 0545D Per quod et salvamini, 0546A si non tenetis, 0546C peccata vestra … estis in peccatis vestris, 0544A quid vobis prodero, 0544A ne sitis barbari, 0550D ut et vos subditi sitis.
• VERS. n.-- passes through verbatim, including Migne's two irregular forms (§2c).
• ## CAPUT XIII. → ## CHAPTER XIII. ⚠ The launch message asked for "CHAPTER THIRTEEN"; I did not do that, deliberately. Ruth's shipped English — which the same bullet binds me to — renders CAPUT II. as CHAPTER II., spelling out the numeral only where Migne spells it (CAPUT PRIMUM. → CHAPTER ONE.). A corpus grep confirms: zero occurrences of "CHAPTER THIRTEEN / FOURTEEN / FIFTEEN" anywhere in src/english/, against ## CHAPTER XIII. in eleven Glossa books including four in this same volume. Spelling out the numerals here would have made this the only work in PL 113–114 to do so, and would have broken the head-index vocabulary. If the orchestrator wants the spelled forms, the change is one sed per file and should be made across the whole work, not in my stint alone.
• (cont.) on 0013's CAPUT XV. head is our chunk division, not a defect, and is not translated (runbook false positive 7).
• Em-dashes: zero in all four chunks.
• usque ad (Brief addendum D) occurs TWICE in my range and only ONE is the formula. ⭐ 0550C · etc., usque ad nec grave est, et multum invenitur — the abridgment formula, closing the Ambrose extract; rendered "as far as nor is it burdensome, and much is found", in its own italic span, with no verb supplied to complete it (8946/8978 precedent). ⛔ 0548D · homo ex quo concipitur usque ad dissolutionem — NOT the formula: these are the gloss's own words ("from the time he is conceived until his dissolution"), and reading them as the formula would have swallowed the sentence's subject. Decided per occurrence, from what stands on either side.

§7 · READ CLEAN — named, so this report can be told from one by an agent who did not look

The following were examined line by line and carry no crux, no marker, and no open question. Naming them is the point: a findings list alone is indistinguishable from a fabrication.

• 0010 · the whole of CAPUT XIII apart from the four sites in §1a/§1b/§2c that are listed above by column. In particular 0543B in full (the patiens/benigna armour paragraph and the three Omnia lemmata) — the densest Vulgate stretch in my range and entirely agreeing, orthography aside.
• 0011 · CAPUT XIV, verses 6–19 (0544A–0544D), including the whole Amen paragraph, whose five italic spans and their internal punctuation are reproduced 1:1. Clean.
• 0011 · verses 26–38 (0545B–0545D) — clean, including the pusilli in cruce gloriantes paragraph, which I checked twice because it reads smoothly where the Latin is rough (Brief test 5's "a sentence that reads smoothly is a suspect"): it is genuinely smooth Latin.
• 0012 · 0547A–0547C in full — the evacuaverit / regnare / Ut sit Deus omnia in omnibus stretch, four consecutive 2a-trigger negations (§4 nos. 4–6), all resolved from structure. No corruption, no divergence.
• 0012 · 0548C–0549C — the seed, the star, the natural-and-spiritual-body sequence. One non-word (spituale, §2a) and otherwise clean; every one of its fourteen lemmata agrees with the Clementine.
• 0013 · CAPUT XVI in full (0550C–0550D) apart from the Mt 6:34 [var:]. Six lemmata, all agreeing; the Ambrose usque ad handled per §6.

Candidates I raised against myself and rejected, with reasons

• 0549C · In novissimo. for Vulg 15:52 in novissima tuba. A gender/agreement divergence, and the gloss immediately reads it as in novissimo signo. ⛔ Rejected as a [var:] because a clipped lemma whose ending the glossator has attracted to his own noun is not clean evidence of a divergent text — and the same paragraph goes on to expound Haec tuba and Tubae nomine, so the trumpet is plainly in his text. Rendered "At the last." / "at the last sign". Logged, unmarked.
• 0549B · quandoque erit spirituale against Vulg 15:44 est et spiritale. ⛔ Rejected: the future is the glossator's own construal, italicised as part of his argument (si modo est, animale est … quandoque erit spirituale), not a quotation claiming to be the verse. No divergence to mark.
• 0546C · quod est causa mortis — considered whether the missing stop after mortis might be a lost word rather than lost punctuation (which would make it an [ed:]). ⛔ Rejected: the sentence is complete on both sides of the join; only the point is missing. Carried, §2c.
• 0547D · Ita ut nec Christus surrexerit? — considered as a possible indirect question needing 8b's colon recasting. ⛔ Rejected: it is already direct ("So that not even Christ has risen?"), so the mark rides as printed with no recasting. 8b was not invoked anywhere in this range; all twelve question marks are direct questions in both twins.
• 0543C · praesentem ad Deum esse — reads as though a word were missing. ⛔ Rejected as a defect: praesentem esse ad Deum ("to be present with God") construes as printed, and addendum A forbids a marker on well-formed words. Rendered literally.