Moralitates in Cantica canticorum

Translated 2026-08-08 by six Opus agents over 35 chunks (40,570 Latin words), each range logging to its own cruces-NNNN.md. Those six files are the record and are unabridged; this file is the merge. It exists because a work split across agents will drift on its load-bearing vocabulary and no agent can see the others — reconciling that is a required pipeline step (translation-runbook.md step 4), not a tidy-up.

Work totals after the blind read's repairs (§8): 138 [var:] · 18 [sic:] · 1 [ed:] · 0 [d:] · 0 [nt:] · 0 [f:]. As first translated: 139 / 17 / 1 — one [var:] was deleted as spurious and one [sic:] added that had been missing. verify-english.mjs clean at 35/35. EN/LA ratios 1.35–1.63.

§1 · THE STRUCTURE OF THE BOOK, AND THE TYPOGRAPHY SEAM

⭐ This work changes how it marks scripture partway through, and the change is a BOOK BOUNDARY, not a printer's whim. Books I–II mark quotations with italics; Book III onward marks them with guillemets « ». The switch falls between chunks 0013 and 0014, at cols 0523/0523B — exactly at Explicit secundus tomus / INCIPIT TERTIUS TOMUS. Corpus-wide counts: 316 italic runs, 286 guillemet runs.

Do not normalize the seam in either direction. Under Pattern 6 a « » is a mark Migne PRINTS and is reproduced 1:1; a " " is ours. Two things look like exceptions and are not:

• 0013's single guillemet pair is the Sub tuam protectionem antiphon — a liturgical class Migne guillemets even inside the italic books.
• 0017's single italic run is a printed locator (Exod. [0533D] XXVII), not scripture.

After the seam italics survive by function: words discussed as words (Eleos, Oleos, misericordia), Migne's marginal source-edition page numbers (303–313), and a lemma nested inside a guillemet quotation.

Guillemet asymmetries are Migne's and are mirrored, never balanced (Pattern 5 corollary): unclosed « at 0532C–D, 0565A, 0579D; stray » at 0568D; unopened » at 0572D and 0557C. A 270-word run at 0569A–C swallows Philip's own commentary inside Migne's quotation marks; reproduced as printed.

§2 · WHAT THE MERGE CHANGED

1. tomus = "Tome"; liber = "book". Ruled here, swept 16 places. The seven divisions of the work are tomi; the Song itself is the liber. Five agents collapsed both to "book"; one used "Tome"; that one was right, and was itself inconsistent (Tome at 0019, tome at 0020). ⚑ The decisive text is 0019 @0541C: Ad quartum tomum exponendum hujus libri accedentes. Rendered with one English word that becomes "the fourth book of this book". Migne's own gloss at 0005 settles the sense — Explicit primus tomus, id est primum volumen — a division of a work, not the work. Capitalized as a locator per Pattern 2. Also normalized in the same sweep: INCIPIT = "HERE BEGINS", EXPLICIT = "Here ends" (one agent had "BEGINNETH"/"endeth").

2. Pattern 3, "Holy Scripture" — 11 instances swept across FOUR agents' ranges. The English carried 6 × "divine Scripture", 3 × "divine Scriptures", 2 × "holy Scripture" against 17 Latin instances of divina/sacrae/divinae Scriptura(e). Pattern 3 fixes the standing name as Holy Scripture, both words capitalized, and says explicitly that a capital S alone is not enough. Chunks touched: 0000, 0001, 0002, 0003, 0004, 0014, 0023, 0025, 0028.

3. candidatio — translated, not carried. Libanus candidatio interpretatur occurs at 0005 and 0009 in the same etymological construction; one agent gave "a whitening", the other kept "candidatio". Translated, because both passages argue from whiteness → purity → virginity, and the Latin left standing makes the inference unavailable to an English reader. The register rule reserves italic Latin for technical words with no clean English equivalent; candidatio is an ordinary noun and has one.

§3 · WHAT THE MERGE RATIFIED — no edit needed, do not churn

Migne's lower-case after a misplaced full stop STANDS. Eight instances (0003, 0006, 0011 ×2, 0012 ×2, 0014, 0015), and all three agents who met the construction independently preserved it — there was no drift, only a request to have it ruled. Reasoning, since a later sweep will be tempted: Pattern 9 normalizes only where English has no exponent for the Latin defect (grammatical gender, a typo'd section label). Lowercase-after-a-stop is expressible in English and expresses exactly what the plate shows, so Pattern 8 governs and both the mark and its consequence stand. Capitalizing would make the passage read as a deliberate sentence fragment — a worse misrepresentation than the lowercase.

Terminology checked across all 35 chunks and found already consistent:

• The Song's colour triple, spanning two agents and 60 instances: nigra = black (28/28) · fusca = swarthy (9/9) · formosa = comely (23/24). Kept strictly distinct; fusca is never "black".
• charitas = charity, never "love" — verified: no chunk under-renders it. 0023 prints the doublet charitas, sive dilectio, which forces the split.
• clementia = clemency kept distinct from misericordia = mercy (28/25 and ~58/54).
• moralitas = the morality · cellaria = storerooms (vs cella vinaria = wine cellar) · recti = the upright · adolescentulae = maidens · torcular = winepress · anaglypta and scalio kept italic Latin.

Pattern 17 (thou = singular, you = plural; a number marker, not a reverence marker) was ruled the same day as this translation and carried in every agent's brief. 0504A is the worked case — one sentence carrying peccata nostra and inter vos together.

§4 · FINDINGS NO SINGLE AGENT COULD HAVE MADE

These are the merge's own results: a divergence attested in more than one agent's range is evidence about Philip's copy of the Bible, not a local slip.

⭐ Zech. 4:14 — duo filii olei SPLENDORIS — four chunks, three agents. 0008, 0015 (×2), 0016, 0018. The Vulgate reads duo filii olei; Migne adds splendoris every single time. And Philip knows it: at 0015 @0528C he says so outright — in Isaia simpliciter oleum appellatur, in Zacharia vero cum additamento, splendoris oleum nominatur — and builds a Trinitarian exposition on the addition (Father as oil, Son as splendour, non tamen tres splendores sunt, sed splendor unus). A lemma conformed to the Vulgate here would void a whole chapter, and the consistency across three agents shows the reading is his text, not a compositor's slip on one page.

⭐ Gen. 3:19 — Terra es et in terram ibis — two chunks, two agents. 0020 and 0033, guillemeted both times, against the Vulgate's pulvis es et in pulverem reverteris. The second half of the work expounds terra (we are earth; the earth given the serpent for food; the earth that yields her fruit). On the Vulgate text those chapters would have no proof-text at all.

Cant. 1:4 — sicut pellis Salomonis, singular (Vulg. pelles). Philip raises the singular himself as a quaestio magna, says it stands nowhere else in either Testament, and resolves it a parte totum. Kept singular in all five recapitulations.

The running lemma is re-quoted against itself and both readings stand: the head of Tome VII prints ne vagari incipiam post greges, and 0575B re-quotes it as per greges. Neither was conformed to the other — the within-work merge rule reconciles our vocabulary, never Migne's printed text.

§5 · POLARITY — rendered as printed, NOT repaired

The class this edition exists to expose. Each was rendered literally and its conjecture kept out of the running text.

• 0498A, Prov. 16:5 — Migne prints Qui exaltat cor suum, mundus est coram Deo; the received text has immundus. Philip's paragraph argues the proud virgin is unclean (Eve corpore virgo… sed mente immunda). Rendered "is clean before God". The 11536/8625 class exactly.
• 0560B, abs re — the idiom is non abs re; the printed text asserts the opposite of what the next paragraph proves. Negation not supplied.
• 0507A, munitum for minutum (×2) — rendered "fortified", which contradicts the argument it sits in. That contradiction is the plate's.
• 0566B, Rom. 5:19 — per obedientiam unius justi constituentur multi moves the Vulgate's predicate into the genitive and leaves multi with nothing to be made. Rendered awkwardly and deliberately.
• 0545A–B — the esse passage asserts the creature must be per se and the next sentence denies it. Corruption not localizable to one word, so no marker fired. ⚠ Flagged by its agent as the passage in the work most likely to be wrongly "fixed" later.

⭐ One silent repair was caught in draft and reverted, which is what the rule is for: at 0513C the non-word abur had first been rendered "the ivory" — invisible, because the sentence is about ivory. It is now carried as [sic: abur].

§6 · OWED — plate reads, and nothing else blocking

1. 0494D — text lost in our SOURCE, not (demonstrably) the plate. The TEI itself reads …aeternaliter constit</p> and starts a new paragraph, so the loss is upstream of our chunker — verified in sources/pl/tei/11638.xml, not inferred. Carried as a fragment with the work's only [ed:] (Pattern 13 cl. 3). If the plate at PL 203 col. 494D carries more, this becomes a data/tei-patches/11638.json entry and the [ed:] comes out in the same commit.
2. 0523C–D — a clause is lost from a list running one oil-nature per sentence: « Qui peccatum non fecit » is welded to an orphan homines reficit. Rendered as printed with no [ed:], because plate-vs- digitization cannot be told apart here without the leaf. Second plate check.
3. 0530C prints Oleos where 0525D prints Eleos for ἔλεος. Rendered as printed both times and deliberately unmarked — ⚠ a reviewer who "corrects" one to the other destroys the pun the argument runs on.
4. ⛔ The blind polarity read has NOT been done and cannot be done by any of the six translating agents or by the merger (runbook 4a: findings are formed blind and reconciled against this file afterwards, never before). This work has no polarity-record.mjs entry and must not be deployed until it has one.

§7 · METHOD DATUM — naming the words worked

Every agent was handed a verbatim list of every quotation printed in its own range, extracted mechanically from the Latin (602 in all), with the instruction that the divergence list was the most valuable thing it would produce. This is the 8930 finding applied deliberately rather than hoped for.

Result: 139 [var:] markers across 35 chunks, plus roughly 120 further quotations collated and explicitly logged as unmarked in the range files — so the absence of a marker is recorded evidence rather than silence. Compare 8930, where the range given a strong but generic "read every lemma carefully" produced 1 marker and a later blind read found seven defects in it.

The blind read will test that. Until it runs, this number is an input, not a result.

§8 · BLIND POLARITY READ — 2026-08-08, three readers, reconciled

Runbook step 4a. One reading pass over every chunk, split across three Opus readers (0000–0011, 0012–0023, 0024–0034), each carrying BOTH lenses — negation/ scope and quotation-fidelity/markers — and each fenced off from this file, the six range files, next-session-resume.md and git history. Findings were formed blind and reconciled here afterwards, never before.

46 sites, 37 ours. Final markers after repair: 138 [var:] · 18 [sic:] · 1 [ed:].

⭐ The result that matters: the prose is sound; the defects were in the apparatus

Two of the three readers found NO dropped negation and NO substituted-opposite anywhere in their ranges — 23 chunks between them — having checked every printed negation particle, including at the places where carrying one makes the sentence contradict its own neighbour. The third found one, below.

That is the 7a″ profile the genre predicts, and it is the opposite of a homily collection's. The failures clustered in three places nobody had looked for them: our own [var:] markers, cross-agent convention drift, and head punctuation — none of which a negation-hunting read would have surfaced.

The one polarity finding — a negative particle whose force was dropped

0023 @0559C, 1 Reg. 12:23 — mechanism 3, which leaves no lexical trace. Migne prints « Absit hoc peccatum a me in Domino, quo minus cessem orare pro vobis… »; the Vulgate has ut cessem. The English read "that I should cease to pray for you" — the Vulgate's clause, not Migne's, silently dropping the negative element quominus carries. Nothing looked wrong: the sentence was smooth, the citation right, the verifier green. Aggravating it, the accompanying [var:] elided precisely that clause behind an ellipsis, so the divergence was neither rendered nor marked anywhere on the page. Fixed — the English now reads "that I should not cease", and the [var:] names ut cessem and states that Migne's quo minus reverses the clause.

Repaired — ours (the rest)

Marker defects — a class no previous work in this corpus has surfaced, because no previous blind read was told to audit the MARKERS rather than the prose:

• A [var:] that manufactured a divergence that does not exist (0015 @0526A, Luc. 1:68). It asserted Migne diverges by printing Dominus — but the Clementine, this edition's base witness, prints Dominus too. Deleted.
• A [var:] quoting a "Vulgate" that exists in no edition (0025, Prov. 17:3) — a hybrid of the Clementine's Sicut igne probatur and the Stuttgart's quomodo probatur in conflatorio. Replaced with the Clementine verbatim.
• A [var:] asserting a divergence in the wrong direction (0026, Joan. 17:5): it claimed an "accusative" divergence that is a Clementine/Stuttgart split, not Migne's, while missing the real one — Migne drops apud te. Rewritten.
• A [var:] citing the wrong chapter in our own note (0032): nolo mortem impii is Ezech. 33:11; 18:23 reads numquid voluntatis meae est mors impii. Corrected — this is OUR note, so refDisplay's "keep what Migne prints" does not shelter it.
• A [var:] annotating nothing a reader can see (0016, Cant. 2:16): in liliis against inter lilia has no English exponent, and the marker's gloss repeated the running English word for word. Demoted to this crux — the page was asserting a divergence it could not show.
• Two [var:] markers stranded 25 and 60 words downstream of the quotations they annotate (0018, 0034). Both moved to the close of their quotation.
• A [sic:] on a word that had ALSO been translated (0023 @0558A, repacicavit): the English rendered "made peace again" and appended a detached [sic:] at sentence end. Pattern 12's operational test forbids both — "rendered as printed" means no marker, "carried untranslated" means the marker wraps the run. ⚠ The [sic:] content was a verbatim Latin substring, so verify-english passed it: this defect is invisible to the verifier. Now carried inside the marker.
• A carried non-word with no marker at all (0001 @0494D, constit) and two [sic:] markers missing their italics (0008 victuram, 0009 vestitvit) — the verifier strips asterisks, so these too passed silently.
• One unmarked material divergence added (0028 @0568C, Psal. 119:5): Migne prints cum habitantibus Cedar, the Vulgate habitationibus — dwellings, not inhabitants — and Philip's gloss immediately personalises it, id est cum his qui tenebras diligunt. The exposition rests on Migne's word.

Cross-agent drift the §2 merge had missed — all found by readers checking against ranges no translator could see:

• charismata rendered THREE ways: "charisms" (0000, 0004), "graces" (0017), "free gifts" (0019 ×2, 0022). Normalized to charisms, which keeps gratia = grace and dona = gifts distinct.
• ⭐ oleum splendoris rendered two ways — "oil of splendour" (0015 ×5, 0016) against "oil of brightness" (0008, 0018). This one is load-bearing and is the §4 finding's own vocabulary: Philip's argument at 0528C is that Zechariah adds splendoris, and it runs "Splendour therefore is the Father, splendour is the Son, splendour none the less is the Holy Spirit". A second English word for the same Latin breaks the chain the argument is made of. Normalized to splendour.
• Joan. 16:33 Confidite — "Have confidence" (0004) vs "Be of good heart" (0008). Normalized to the first, which is also the received English.
• sacra Scriptura → "sacred Scripture" at 0016 @0529D. ⚠ The §2 Pattern 3 sweep matched divina and holy and missed sacra entirely — a sweep is only as wide as its pattern list.
• RECAPITULATIO A SUPERIORIBUS rendered FOUR ways across 23 heads: "FROM WHAT WENT BEFORE" · "FROM WHAT PRECEDES" · "FROM WHAT IS ABOVE" · "FROM THE FOREGOING" — one per agent, and the most visible drift in the work, since heads land in the TOC and in heads[] indexing. Normalized to FROM THE FOREGOING (15 heads). ⚠ 0023/0024 are NOT part of it: their Latin reads RECAPITULATIO SECUNDUM ANAGOGEN, with no a superioribus to render.

Two mis-renderings of ordinary words:

• 0017 @0534D repetitur → "found", which is reperitur. Repetitur construes perfectly ("is repeated"); Pattern 12 is explicit that a real word merely wrong is rendered as printed. Now "is repeated".
• 0028 @0569C Ista ergo nos oportet omnia … possidere → "it behooves us all". Omnia is neuter accusative and modifies ista, the things possessed, not nos. Now "All these things therefore it behooves us…".

Head punctuation, Pattern 8 — seven supplied marks removed. Colons supplied where Migne prints commas (0012, 0013, 0032, 0033, 0034), a comma supplied where he prints none (0017, 0027).

⚠ A reader recommendation DECLINED, after checking

One reader reported QUALITER as rendered inconsistently ("HOW" vs "IN WHAT MANNER") and recommended picking one form work-wide. Declined — the Latin uses two different words, and pairing every head against its source shows the English was tracking them correctly: QUOMODO → "HOW" (0009, 0010, 0011, 0012, 0027), QUALITER → "IN WHAT MANNER" (0013, 0016, 0032, 0033, 0034). Collapsing them would have destroyed a distinction the plate makes. Only 0026 was actually wrong — QUALITER rendered "HOW" — and that one instance is fixed. A majority-rule "fix" here would have been a silent normalization of Migne.

Confirmed correct, no action — recorded so a later sweep does not churn them

• The colour triple holds across two agents and 60 instances: nigra = black (32/32), fusca = swarthy (9/9), formosa = comely. Including 0025's si inter tribulationes fusca, id est nigra appareo — the one sentence where collapsing them would be fatal.
• Every guillemet asymmetry is Migne's and is mirrored, not balanced: unclosed « at 0532C, 0552B, 0565A, 0579D; stray/orphan » at 0552C, 0557C, 0568D, 0572D. All three readers checked open/close counts against the Latin, not against English expectation.
• The typography seam is handled faithfully on both sides — and the hardest case is the one a translator would most naturally have tidied: at 0015 @0525A Migne marks the lemma not at all, and the English supplies straight quotes (ours) rather than guillemets. It wasn't tidied.
• Migne's lower-case after a misplaced full stop stands in all eight places (see §3).
• 0020's Gen. 1:1 printed three ways in one chunk — fecit / creavit / Fecit — rendered "made / created / made" respectively, unconformed.
• 0014's Rom. 9:5 printed once without Deus and once with it, rendered accordingly, unconformed in either direction.

⬜ OWED — measured, not done

[var:] reference format is inconsistent work-wide: 71 markers use the roman Vulg. Zach. IV, 14 form, 59 use arabic Vulg. Zach. 4:14, and 8 carry no reference at all (which Pattern 14's own worked example also does, so it is permitted). The same verse appears as Isa. LIII, 5 at 0014 and Isa. 5:3 at

1. Deliberately not swept: both forms name the verse correctly, so this is presentation in our own apparatus rather than a claim about the text, and regex-rewriting 130 markers at speed risks damage out of proportion to the gain. Do it deliberately, with a tested script, and prefer the arabic form — it distinguishes our note from Migne's roman [n:] by eye, which is the Pattern 6 principle applied to the apparatus.

Smaller, left for the same pass: three unbracketed English glosses on italic Latin (0015 @0525D, @0528C; 0016 @0530C — Eleos … misericordia, mercy), where our voice sits unmarked in running text; and a handful of commas promoted to semicolons inside quotations (0014 head, 0016 @0528D, @0529A, 0019 @0543A).

⚠ One attribution could not be settled from our files and is NOT recorded as either: the deleted semicolon in ideoque; audacter aiebat (0018 @0542B) is either Migne's plate or the TEI transcription. If the plate carries it, Pattern 8 forbids the deletion; if the TEI invented it, it is a tei-patches question. It needs the leaf — the same plate check already owed at §6.

# 11638 — Philip of Harvengt, Moralitates in Cantica canticorum (PL 203, 489–584) ## Cruces for chunks 0000–0005 (cols 0489C–0504B)

Agent 1 of six. This file must be merged with the other agents' cruces-*.md before the work ships (runbook step 4). Head-of-file terminology decisions are binding on the whole work at merge time — where a later agent has rendered the same term otherwise, one voice must win (Song-of-Songs ruling: across works each stands, WITHIN one work reconciliation is required).

A. Terminology fixed in this range (merge candidates)

B. Second person (Pattern 17) — the singular/plural map for this range

The work slides between a singular reader and a plural congregation, and the slide is on the page:

• Singular → thou: 0002 @0497C Qui ergo istum librum legis vel audis… perfectus esse satage and the whole poteris / potes / credis et retines / adimples / repraesentabis / custodis chain that follows; 0000 @0493B invenies; 0004 @0500A Intuere, o homo, temetipsum… si bene vixeris; Ps 36:4 dabit tibi petitiones cordis tui; Ps 30:20 dulcedinis tuae… abscondisti; Ps 44:8 tu dilexisti… unxit te… sanctis tuis; John 17:3 cognoscant te… misisti; John 17:21 sicut ego et tu unum sumus.
• Plural → you: 0001 @0495C non dubitatis… videte ut… custodiatis; Matt 5:48 Perfecti estote… Pater vester; Ezek 36:26 Tollam a vobis… dabo vobis (0003); 1 John 2:27 vos… docet vos (0004); Isa 59:2 inter vos et Deum (0005).
• ⚠ 0005 @0504A carries the number clash inside one sentence — peccata nostra fecerunt murum inter vos et Deum: "our sins have made a wall between you and God." Carried as printed; see §C.

C. Vulgate divergences — the deliverable list

Every printed quotation in the range was read word by word against the Vulgate. Twenty-three material divergences fired [var: …] (2 · 4 · 8 · 2 · 6 · 1 across chunks 0000–0005); the rest are logged here without a marker. Migne's reading is what the English renders in every case.

Marked [var: …]

1. 0000 @0490C, Ps 30:20 — Migne Magna, inquit, multitudo dulcedinis tuae; Vulg. Quam magna multitudo…. The loss of Quam turns an exclamation into a statement, and Philip's gloss (Vere magna et admirabilis est multitudo) follows Migne's declarative, not the Vulgate's exclamatory. Note that the [n:] falls inside the guillemets on the plate and is reproduced there.
2. 0000 @0491D, 1 Cor 10:11 — Migne Quoniam quidem omnia quae contigerunt patribus nostris, in figura contigerunt illis; Vulg. Haec autem omnia in figura contingebant illis. A full re-cast, and patribus nostris is Philip's own hinge for the next sentence (Si ergo apud illos figura fuit…). Not italicized on the plate; rendered as running text, as printed.
3. 0001 @0494B, John 17:3 — Migne drops solum Deum verum. Material and load-bearing: Philip immediately supplies the missing predicate himself (subauditur: et Spiritum sanctum, unum et verum Deum esse), which only makes sense against a text that lacks it. A conformed lemma here would have made Philip's gloss redundant.
4. 0001 @0494B, Luke 2:14 — Migne Gloria in excelsis Deo; Vulg. in altissimis. Familiar-liturgical form against the Vulgate's.
5. ⛔ 0001 @0495C, Luke 12:35 (cited as Luc. XXII) — Migne prints Sint lumbi nostri praecincti. The Vulgate has lumbi vestri. This is the range's sharpest 7a″ trap: the remembered "let your loins be girded" arrives before the printed word is read, and it also silently repairs Migne's chapter number. Rendered "Let our loins be girded." The reference is also wrong (Luke 12, not 22) — printed as-is per refDisplay policy, not corrected.
6. 0001 @0496A, Ps 93:19 — Migne Secundum multitudinem tentationum mearum; Vulg. dolorum meorum. The whole paragraph is about temptation (ab omnigena tentationum infestatione vulneratis), so the divergent word is the one the argument runs on. Unmarked as a quotation on the plate apart from the italics; no [n:].
7. 0002 @0496D, Ps 35:7 — Migne Judicia Dei abyssus multa; Vulg. judicia tua. Third person for second; Philip's frame requires the third.
8. 0002 @0496D, Eccl 9:2 — Migne Omnia reservantur homini post mortem incerta; Vulg. sed omnia in futurum servantur incerta. post mortem is not in the Vulgate at all and is exactly what Philip's argument (the sealed well of God's counsel until judgment) needs.
9. 0002 @0497A, Matt 25:41 — Migne Ite, maledicti…; Vulg. Discedite a me maledicti… qui paratus est diabolo. praeparatus for paratus is an orthographic variant only (Pattern 9, no marker of its own).
10. 0002 @0497A, Matt 25:34 — Migne percipite regnum; Vulg. possidete paratum vobis regnum.
11. 0002 @0497B, John 17:21–22 (cited as Joan. XVIII) — Migne Pater, volo ut sicut ego et tu unum sumus, ita et ipsi unum sint in nobis; the Vulgate's clause order and wording are wholly different. Reference likewise off by one chapter; printed as-is.
12. 0002 @0497B, 1 John 3:2 (cited as Joan. III) — Migne nondum apparet (present); Vulg. nondum apparuit (perfect). The 7914 @1071D class (tense quietly restored to the Vulgate). Rendered "it doth not yet appear."
13. 0002 @0497D, Matt 5:48 — Migne Perfecti estote, sicut et Pater vester perfectus est qui in coelis est; Vulg. Estote ergo vos perfecti, sicut et Pater vester caelestis perfectus est.
14. ⛔ 0002 @0498A, Prov 16:5 — A DROPPED NEGATIVE, and it inverts the argument. Migne prints Qui exaltat cor suum, mundus est coram Deo. The maxim as commonly cited in the schools is immundus est coram Deo omnis qui exaltat cor suum (a loose form of Prov 16:5, abominatio Domini est omnis arrogans). Philip's whole paragraph argues that a virgin who is proud of her virginity is unclean — Eve was corpore virgo et casta, sed mente immunda — and the very next clause says nec prodest quidquam virginem et castum esse corpore, si fuerit incestus in mente. As printed, the Scripture proof asserts the exact contrary of the sentence it is adduced for. Rendered literally ("He that exalteth his heart is clean before God") per Pattern 7a; not repaired. The lost im- is a plate defect of exactly the 11536 @1085D / 8625 @0940B class. This is the single most important entry in the file.
15. 0003 @0499A, Ezek 36:26–27 (cited as cap. XI et XXXVI) — Migne Tollam a vobis cor lapideum; Vulg. auferam cor lapideum de carne vestra.
16. 0003 @0499D, 1 Cor 6:17 — Migne Qui adhaeret Deo, unus spiritus efficitur cum ipso; Vulg. Qui autem adhaeret Domino, unus spiritus est. Philip's own chunk 0000 close (unus spiritus effici merear cum ipso, @0493A) is built on Migne's expanded form, not the Vulgate's.
17. 0004 @0501A, 1 Pet 2:5 — Migne superaedificamini in habitaculo Spiritus sancti; Vulg. superaedificamini domus spiritualis. Migne's reading is what the surrounding argument (the spirit as habitaculum Dei) requires.
18. 0004 @0501C, Ps 44:8 — Migne unxit te Deus Pater tuus oleo jubilationis prae omnibus sanctis tuis; Vulg. unxit te Deus, Deus tuus, oleo laetitiae prae consortibus tuis. Three divergences in one verse, and the third is load-bearing: the sentence immediately before reads significavit Christum prae omnibus sanctis impletum Spiritus sancti charismatibus — the exposition quotes Migne's phrase, not the Vulgate's.
19. ⛔ 0004 @0502B, Ps 83:8 — Migne prints deorsum in Sion where the Vulgate has Deus deorum in Sion. deorum → deorsum is a one-letter plate corruption that turns "the God of gods in Sion" into "downward in Sion," and Philip has already supplied Deum outside the quotation (donec videat Deum deorsum in Sion regnantem), which is what makes the corrupt reading construe at all. Rendered "until it see God, downward in Sion, reigning." Not repaired.
20. 0004 @0502C, Phil 2:8 — Migne adds Patri (factus est obediens Patri usque ad mortem), which the Vulgate lacks; the addition is Philip's point (obedience reconciling the race to the Father).
21. 0004 @0502D, 1 John 2:27 (cited as II Joan. II) — Migne Et vos unctionem a Spiritu sancto accepistis, et, ut aliquis doceat vos, necesse non habetis; the Vulgate's clause is a jussive (maneat in vobis) and names no Spirit.
22. 0004 @0502D, John 7:37–38 — Migne compresses, dropping Qui credit in me, sicut dicit Scriptura and reading de ventre ejus fluent aquae vivae for flumina… aquae vivae; also dicebat de Spiritu sancto for dixit de Spiritu.
23. 0005 @0504A, Isa 59:1–2 — Migne Non est invalida manus Domini ad salvandum, sed peccata nostra fecerunt murum inter vos et Deum; Vulg. Ecce non est abbreviata manus Domini ut salvare nequeat… sed iniquitates vestrae diviserunt inter vos et Deum vestrum. Note the internal number clash Migne prints — peccata nostra but inter vos — carried as printed (§B).

Divergences logged, NOT marked (sense unaffected, no gloss turns on them)

• 0000 @0491D, Cant 1:1 — exact Vulgate; no divergence. Ditto Zach 6:12 (@0492B), Ps 21:23 (@0492C), Isa 7:14 (@0493A), Ps 84:11 (@0493A).
• 0000 @0492C, Matt 28:10 — Ite, nuntiate fratribus meis etc.; Migne's own etc. covers the truncation.
• 0000 @0492B, Gen 29:6–11 — narrative paraphrase in Philip's own voice (et ecce, inquit Scriptura…); not a quotation, not marked.
• 0001 @0493B, Ps 118:130 + 1 Cor 14:20 conflated — declaratio sermonum cujus illuminat et spirituale intellectum parvulis malitia manifestat, cited to 1 Cor 14. Two verses welded; construes as printed (parvulis dat. + malitia abl.), so rendered straight. spirituale intellectum is a gender mismatch with no English exponent (Pattern 9).
• 0001 @0493D, Rom 10:4 — Migne drops enim. Immaterial.
• 0001 @0495B, Ps 91:13 — exact; the sentence that follows it (Justi florebunt sicut lilium ante Dominum) is not Vulgate and is not marked as a quotation on the plate — it is the antiphon form (cf. Hos 14:6 germinabit sicut lilium). Rendered as running text, as printed.
• 0001 @0494D, Job 5:6 + Wisd 7:16 conflated (nihil fit in terra sine causa + in manu Dei sumus et nos et sermones nostri), cited to Job V only.
• 0002 @0496C, Heb 4:11 — Migne drops ergo. Immaterial.
• 0002 @0497A, 1 Cor 13:12 — Migne quotes three fragments in an order the Vulgate does not have (in aenigmate before per speculum); no gloss turns on the order.
• 0004 @0500A, Gen 26:19–22 — Migne's puteum illum Abundantiam appellaverunt; the Vulgate names the wells Calumnia, Inimicitias, Latitudo. Narrative paraphrase, not a marked quotation; rendered "Abundance."
• 0004 @0501D, 1 John 3:2 — Migne prints sicuti est videbimus (reversed from the Vulgate's videbimus eum sicuti est, and without the object). Rendered in Migne's order — "as he is we shall see" — with the object left where Philip put it (Deum… exspectamus). No marker: nothing in the gloss turns on the order, unlike 7914 @1079B.
• 0004 @0502A, Jas 2:26 — paraphrase in Philip's own voice, not quoted.
• 0004 @0502B — Per obedientiam perveniunt ad occultorum Dei notitiam, attributed to divina Scriptura and cited [n: (II Cor. II)]. The sentence is in no book of the Vulgate. Reference printed as-is; no correction proposed.
• 0004 @0500C, Gal 6:14 / @0500C John 16:33 (Confidite quia ego vici mundum, Vulg. without quia) / @0500C–D John 14:6 split into two quotations — all immaterial.
• 0004 @0501A, Dan 2:45 — abcissus for abscissus; attested medieval spelling, Pattern 9, no marker.
• 0005 @0503A, Cant 4:15 — exact. @0503A Wisd 1:4 — fragment, present tense habitat for the Vulgate's future habitabit; Philip supplies his own nullatenus, so the negation is his, not a dropped one. @0504A Wisd 1:5 — Migne Quia for enim; immaterial.

D. Text cruces — printed readings rendered literally, not repaired

1. ⛔ 0001 @0494D — the sentence breaks off mid-word. Our TEI ends the paragraph …in unam personam Dei et hominis aeternaliter constit. The verb is almost certainly constituat, but the truncation is at the level of our digitized source, not demonstrably Migne's plate, so it takes neither [sic:] (which would blame the plate) nor a silent repair. The fragment is carried in italics and an [ed: …] marks the hole (Pattern 13, clause 3): a reader-visible break that would otherwise read as our own bug. Recovery: read PL 203 col. 494D on the plate; if the plate carries more, the fix belongs in data/tei-patches/11638.json, and the [ed:] comes out in the same commit.
2. ⛔ 0004 @0500B–C — a clause with no verb. Migne prints …et pugnare contra se ipsum, ut dominans animo suo, fortior ut expugnatore urbium. The allusion is Prov 16:32 (qui dominatur animo suo, expugnatore urbium), where the missing word is melior est / sit. The second ut stands where a finite verb belongs and cannot be construed with the ablative expugnatore. Nothing is supplied. Rendered "so that, ruling over its own mind, stronger as the stormer of cities." — the English stops where the Latin stops. Every ut Migne prints is on the page (Pattern 7a: the printed word is carried even when the sentence will not construe).
3. 0004 @0500C — ut vineat seipsum et mundum. vineat is not a Latin word; vincat is certain from the context (ut vincat seipsum et mundum, answering ego vici mundum two lines later). Non-word class, Pattern 7 as ruled 2026-07-28: carried untranslated in italics inside [sic: vineat], not read as vincat.
4. 0004 @0501B — a missing terminal stop. O quam magna est clementia nostri Redemptoris Quid enim innixio Dei… — the plate prints no punctuation between the exclamation and the following question. Not supplied (Pattern 8); the English runs on identically.
5. 0003 @0499B — a full stop inside a relative clause. fides, spes, charitas sunt. quae nostras simplices enutriunt cogitationes. The stop before the lowercase quae is the plate's. Not deleted (Pattern 8, the 11542 @1041D class): the English reads "…are faith, hope, charity. which nourish our simple thoughts."
6. 0000 @0493A — David prophetae dicit. Dative where the parallel at 0004 @0501C reads David propheta loquitur ad Filium Dei. Rendered literally as a dative ("David says to the prophet"); this is a real word merely mis-inflected, so under Pattern 12 it takes no [sic:] — the English word on the page is ours. Conjecture: propheta, near-certain.
7. 0000 @0493A — Junge ergo, si tali placet, lilium violae vel rosae. Bare dative tali with no antecedent; rendered "if it please such a one." Conjecture: si talia placent or si ita placet. Nothing supplied.
8. 0001 @0494C — Duo ejusdem e diverso dentes accedunt, in a sentence whose next clause is simul conjunctis labiis. dentes (teeth) where labia (lips) is what a kiss requires and what Philip himself writes three words later. Rendered "Two teeth of the same one draw near from opposite sides"; a real word, so Pattern 7 not Pattern 12 — no marker. Conjecture: labia, but the dentes/labia pair may be Philip's own (two teeth = the two natures meeting), so this is flagged, not resolved.
9. 0002 @0497D–0498A — two finite verbs, no relative. ideo bene quidem in nostra recapitulatione sequantur lilia virginitatem, castitatemque, et munditiam significant. Conjecture: lilia, quae virginitatem… significant. Rendered as two clauses exactly as printed: "…let the lilies follow, they signify virginity and chastity and cleanness."
10. 0000 @0491B — a promise of seven, a list of two. omnes et singulae sententiae… septem in se ad minus continere videntur, scilicet… followed by only the divinity/humanity and the Christ/Church conjunction before the sentence turns to rhetoric and the arts. Rendered as printed; the count is made good at 0000 @0493B (In ista sententia invenies septem sacramenta…), which lists seven. No defect asserted.
11. 0001 @0494A — per omnen textum sui for omnem: orthographic, Pattern 9, translated normally.
12. 0003 @0499B — ne nos dimersos obruant for demersos: attested medieval spelling, Pattern 9, no marker.
13. 0004 @0500A — terra obice obstrusum manebat: obstrusus is attested (from obtrudo); Pattern 9, translated normally ("shut up by a barrier of earth").

E. Structural notes

• Marginal numbers 287 … 294 (the earlier edition's pagination, which Migne sets inline in italics) are reproduced verbatim, in place, including where one splits an italic run — 0002 @0498B prints Beati mundo corde, quoniam ipsi Deum 291 videbunt, and the English preserves all three runs. 292 opens the ## head at 0004 and again at 0005 and is kept inside the head line.
• Guillemets: one pair in the whole range (0000 @0490C, the Ps 30 quotation), reproduced 1:1 with the [n: (Psal. XXX)] inside them, as the plate prints it, and with inquit left inside the quotation per Pattern 5. Everything else in chunks 0000–0005 is marked by italics only, so no " " were needed either — this range supplies no quotation marks of our own at all. ⚠ Later agents: the work switches to guillemet-dominant marking from chunk 0014.
• Column anchors glued to text — the plate/TEI prints [0497C]Ad moralitatem (0002) and [0498C]In veteri lege (0003) with no space. Reproduced identically.
• Anchor order at 0005 runs 0503A · 0503B · 0504A · 0504B · 0503. The final bare 0503 (no band letter) carries the Explicit. Left exactly as printed; column-band arithmetic is a known non-signal.
• (cont.) on the 0005 head is our chunker's marker, not Migne's; carried untranslated, per the standing rule.
• Chunk 0000 carries two ## heads (INCIPIT PRIMUS TOMUS. and PROOEMIUM) where the frontmatter heads[] lists one. Body count governs; the English mirrors both.

F. Verification run

node scripts/verify-english.mjs 11638 reports no findings for 0000–0005 (remaining failures are the unwritten chunks of other agents). Anchor sequences match the Latin twin exactly in all six chunks (14 · 11 · 9 · 6 · 12 · 5); [n:] counts match the twin exactly (13 · 12 · 14 · 4 · 24 · 4; a raw grep reports 5 for 0003 only because that chunk's incipit: frontmatter field ends with a truncated [n: — body counts govern); guillemets 1‖1; ## heads 2 · 1 · 2 · 1 · 1 · 1.

PL 203, coll. 0503B–0518B. Agent 2 of 6. Merge with the other agents' cruces files at assembly (runbook step 4); the terminology block below is offered for that reconciliation.

Terminology decisions (for the merge — these are the ones a peer agent could differ on)

The work's whole argument is a fixed pair of virtues, and elegant variation on it would be a doctrinal error of the same class as 7a″'s lemma-conformation. Fixed here:

Second person (Pattern 17). Nothing in this range addresses a plural except quoted scripture. Philip's own singular addressee — o auditor qui audis et legis … stude et tu (0506A), Sic enim habes in Evangelio scriptum (0514D), si … adjunxeris … adimplebis (0515B) — takes thou throughout. Quoted plurals take you: Lac vobis potum dedi, Vos unctionem habetis … docet vos, Patres vestri … moriemini, Ut fides vestra, Confidite, Poenitentiam agite, vos palmites estis. Quoted singulars take thou: ubera tua, Scio, Domine … es tu, in te projectus sum, in manus tuas, Si tu es Christus, Consilium tuum, Semen tuum, In semine tuo, Venter tuus. The range therefore shifts number inside one paragraph at 0506A (thou → Lac vobis) and that is the plate's own shift, not drift.

Quotation typography. Migne marks every quotation in 0006–0011 with italics; the plate prints no guillemets at all in this range (verified: 0 « in all six Latin chunks). Per Pattern 6 none were supplied, and no " " were supplied either — the italics already do the work, and adding straight quotes on top would double-mark. Guillemet parity with the Latin twin is therefore 0 = 0 in every chunk. This is the italic-dominant end of the work; the switch to « » begins at 0014, outside this range.

Inline edition apparatus. The plate carries italic page-numbers of an earlier edition inside the running text — 295 (0505A), 296 (0506D), 297 (0508C), 298 (0510B), 299 (0512A), 300 (0513D), 301 (0011 head), 302 (0517B). All are reproduced verbatim, untranslated, in place, per the Joel precedent. Three of them fall inside an italic scripture lemma or a head; they are kept there.

A. Vulgate-divergence list — the [var: …] decisions

Every printed quotation in the range was read word by word against the Vulgate. Twenty markers fired. The ones that fired are listed first; the ones examined and deliberately not marked follow, because that list is the evidence that the marked ones were chosen, not merely noticed.

Fired

1. 0504D · Jon. 4:2 — Migne: Scio, Domine, quia clemens et misericors es tu. Vulg.: scio enim quia tu Deus clemens et misericors es. The plate replaces the predicate Deus with a vocative Domine. Material, and Philip's gloss hangs on exactly this word: Scio pro certo quia tu, qui nunc Deus solummodo omnibus innotescis… — he is expounding a tu that the Vulgate makes explicitly tu Deus. [var: Vulg. Jon. 4:2 …].

1. 0505A · Rom. 4:15 + Hebr. 7:19 — Migne prints Lex iram, (id est mortem) operatur, et nihil ad perfectum adduxit lex under a single [n: (Rom. IV)]. Only the first clause is Romans; nihil ad perfectum adduxit lex is Hebr. 7:19. Migne's own note is reproduced verbatim and uncorrected (a reference is never emended here); the [var:] states the two loci so a reader is not left thinking Rom. IV contains both.

1. 0505B · Ps. 21:10–11 — Migne: Ab uberibus matris meae in te projectus sum. The words are Vulgate words but drawn from either side of a verse division: v. 10 ends spes mea ab uberibus matris meae, v. 11 begins In te projectus sum ex utero. Marked because the stitched clause is what Philip glosses (Occidentem litteram … deserui), i.e. the argument rests on the join.

1. 0505D · I Joan. 2:20 — Migne: Vos unctionem habetis a Spiritus sancto. Vulg.: vos unctionem habetis a Sancto ("from the Holy One"). Material: the whole paragraph is about the gifts of the Holy Spirit descending from Christ, and the plate has already supplied the identification the Vulgate leaves open. (For the grammatical defect in a Spiritus sancto see § B.1.)

1. 0506A · II Cor. 2:15 (+14) — Migne: Bonus odor sumus Christi in omni loco. Vulg. v. 15: Christi bonus odor sumus Deo. In omni loco belongs to v. 14. The plate has swapped the addressee-dative for the locative phrase of the previous verse.

1. 0506A · II Cor. 2:14 — Migne: Gratias agimus Deo qui odorem notitiae suae per nos manifestavit in omni loco. Vulg.: Deo autem gratias qui … odorem notitiae suae manifestat per nos in omni loco — present, not perfect, and the opening is restructured. Marked chiefly for the tense: Philip's point is a standing apostolic fragrance, and the perfect closes it into a past act.

1. 0508B · I Cor. 2:8 — Migne: Si enim Jesum Dominum cognovissent, nunquam Regem gloriae crucifixissent. Vulg.: si enim cognovissent, nunquam Dominum gloriae crucifixissent. King for Lord of glory, plus the added object Jesum Dominum. The commonest sort of 7a″ trap in this work: the familiar English ("the Lord of glory") is exactly what the plate does not say.

1. 0508C · Ezech. 18:31–32 / 33:11 — Migne: Et quare moriemini filii Israel? quia nolo mortem peccatoris, dicit Dominus, sed ut convertatur et vivat. Vulg. 18:31–32: quare moriemini domus Israel … quia nolo mortem morientis dicit Dominus Deus, revertimini et vivite; the tail sed ut convertatur et vivat answers to 33:11. Two substantive word-changes plus a conflated close.

1. 0509B · Zach. 4:14 — Migne: Isti sunt duo filii olei splendoris. Vulg.: isti sunt duo filii olei. The added genitive is not in the received text.

1. 0509D · Ps. 102:10 — Migne: Domine, non secundum peccata nostra facias nobis, neque secundum iniquitates nostras retribuas nobis. Vulg. is indicative: non secundum peccata nostra fecit nobis, neque … retribuit nobis. The plate gives the liturgical petition-form. Material: a statement about what God has done becomes a prayer for what he may do, and Philip's frame (si Deus ad mala … vellet respicere) reads either way, which is precisely why the substitution is invisible.

1. 0511B · I Petr. 1:7 — Migne: Ut fides vestra multo pretiosior sit auro quod per ignem probatur. Vulg.: ut probatio fidei vestrae multo pretiosior auro… The Vulgate prices the trial of the faith; the plate prices the faith. Philip's argument (gold = the Catholic faith) is built on the plate's shorter form.

1. 0512C · Isa. 25:1 — Migne: Consilium tuum, Domine, antiquum verum fiat. Vulg.: quoniam fecisti mirabilia, cogitationes antiquas fideles, amen. An entirely different Latin of the verse. Marked with the Vulgate reading only; no conjecture is offered about its origin (Pattern 14 forbids speculating why witnesses differ).

1. 0512D · Ezech. 17:3–4 (+22–23) — Migne's eagle sentence is a free conflation: Vulg. 17:3–4 has aquila grandis magnarum alarum, longo membrorum ductu, plena plumis et varietate, venit ad Libanum et tulit medullam cedri, and the transport ends in terram Chanaan, in urbe negotiatorum; the placing in monte sancto et excelso belongs to 17:22–23. The whole Marian exposition that follows depends on the mountain, so the join is load-bearing.

1. 0513D · Gen. 3:15 — Migne: Semen tuum conteret caput serpentis. Vulg.: ipsa conteret caput tuum. The plate's reading is the semen-reading, not the ipsa-reading, and Philip's point (the promise of the coming of the Son) requires the seed. This is the single most consequential divergence in the range.

1. 0515B · Isa. 11:2 — Migne omits spiritus Domini before spiritus sapientiae et intellectus. Marked because Philip is counting: he needs seven gifts to add to five virtues and reach twelve, and the plate's shorter list gives exactly seven.

1. 0516C · Apoc. 12:15 — Migne: Et misit draco flumen post mulierem. Vulg.: et misit serpens ex ore suo post mulierem aquam tamquam flumen. The whole surrounding exemplum is elephant-and-dragon, so the substituted subject is the one the argument needs.

1. 0517A · Dan. 4:8 — Migne: cujus proceritas pertingebat usque ad coelum, et rami illius dilatabantur usque ad terminos terrae. Vulg. 4:8: proceritas eius contingens caelum, aspectus illius erat usque ad terminos universae terrae — the branches enter only at v. 9. Philip then identifies the branches with the Apostles and quotes Ps. 18 on their sound going out to the ends of the earth: the exposition rests entirely on a word the Vulgate does not have at this verse. Textbook 7a″.

1. 0517A · Hebr. 4:14 — Migne: Habemus pontificem magnum Christum Jesum qui penetravit coelos. Vulg.: habentes ergo pontificem magnum qui penetravit caelos, Iesum Filium Dei. The christological title is replaced.

1. 0517C · Eccli. 41:17 (Migne's note reads Eccli. XL., kept verbatim) — Migne: Sapientia occulta et thesaurus absconditus. Vulg.: sapientia absconsa et thesaurus invisus. Both nouns carry a different adjective.

1. 0517D · Cant. 7:2 — Migne: Venter tuus sicut crater tornatilis. Vulg.: umbilicus tuus crater tornatilis, numquam indigens poculis; venter tuus sicut acervus tritici. The plate has moved the turned bowl from the navel to the belly. Philip adduces it for roundness* (thronus ille rotundus … de virtute in virtutem), which the Vulgate's acervus tritici* would not give him. He is quoting his own base text, the Song — which makes this the most surprising divergence in the range.

Examined and NOT marked (sub-material; recorded so the absence of a marker is evidence)

• 0503C · Joan. 6:56 — Caro mea vere est cibus for Vulg. caro enim mea vere est cibus: dropped connective at the head of a quotation.
• 0503B/C · Cant. 1:1 — fragrantia for Vulg. fraglantia: spelling (Pattern 9).
• 0505A · Isa. 28:9 (Migne's note Isa. XXIX, kept) — avulsos for apulsos: the Clementine itself reads avulsos; no divergence.
• 0505C · Luc. 23:21 / Joan. 19:6 — Crucifige illum for the doubled Crucifige, crucifige eum.
• 0505C · Exod. XXI — Si quis hoc, aut illud peccatum fecerit, moriatur has no Vulgate exponent at all: it is Philip's own summary of the law's capital sanctions (cf. Exod. 21:12ff), printed in italic like a quotation. No received reading exists to place in a [var:], so it is logged here only. Flagged for the merge: if a peer agent has met the same summary-as-lemma elsewhere in the work, the two should agree about not marking it.
• 0506A · I Cor. 3:2 — Lac vobis potum dedi, truncated before non escam.
• 0506A · I Petr. 2:2 — Quasi for Vulg. sicut: the received liturgical incipit.
• 0506B · Jac. 4:4 — constituetur (future) for constituitur; hujus saeculi for saeculi huius.
• 0506C · Exod. 16:14 — tusum for tunsum (Pattern 9 spelling).
• 0507C · Joan. 10:18 — ponendi animam meam for ponendi eam; sense unchanged.
• 0508B · Joan. 6:49 — word order only (manducaverunt manna in deserto).
• 0508C · Joan. 8:21/24 — in peccato vestro (sing.) against 8:24 in peccatis vestris; 8:21 has the singular, so the plate is quoting 8:21 with 8:24's dixi vobis.
• 0509A · III Reg. 10:18–20 and 0509B · Zach. 4:2–3 — collated in full, verbatim Vulgate apart from illius/eius.
• 0510C · Joan. 16:33 — Confidite, quia ego vici mundum, added quia.
• 0512A · Apoc. 4:5 — procedunt (present) for procedebant.
• 0512D · Deut. 32:11 — provocat for provocans, eos for eum.
• 0513A · Eccli. 24:17 — verbatim. NB the citation is printed inside the italics as bare parentheses with a column anchor inside it: (Eccli. [0513B] XXIV). Reproduced exactly, anchor in place; it is not an [n:] and was not converted into one.
• 0513D · I Petr. 2:22 — ore ejus for ore ipsius.
• 0514A · Gen. 8:17/9:1 — Crescite et multiplicamini, truncated.
• 0514A · Gen. 22:18 — reordered, terrae dropped.
• 0514B — Perfectus Deus, perfectus homo, ex anima rationali et humana carne subsistens is the Athanasian Creed, verbatim; not scripture, no marker.
• 0514D–0515C — III Reg. 10:19, Marc. 16:16, Coloss. 2:9 (Habitat for inhabitat), Joan. 1:16, III Reg. 10:20, Joan. 1:14: all collated, all sub-material.
• 0515A · Matth. 4:17 — enim dropped.
• 0516A · Eph. 2:14 — enim dropped. 0516A · Prov. 9:1 truncated before excidit columnas septem — and Philip supplies the missing half himself in prose at 0516D (domus … super septem columnas subsistentem), which is why the truncation is not a loss.
• 0516C · Prov. 31:10 — verbatim, Migne adding the question mark.
• 0517B · Joan. 15:5 — estis supplied.
• 0517B · Ps. 18:5 — Migne replaces the Vulgate's eorum with a non-italic gloss apostolorum set inside the italic run: in omnem terram exivit sonus apostolorum, et in fines orbis terrae verba eorum. The italic break already shows the reader that the word is Philip's, so no [var:] was added; the English preserves the break at the same point.
• 0517C · Ps. 73:12 — operatus est Deus salutem, subject supplied against Vulg. operatus est salutem. Borderline; not marked because nothing in the exposition turns on the explicit subject.
• 0517D · Ps. 44:14 — eius dropped. The note prints the psalm number as (Psal. XXXXIV) — the four-X numeral is kept verbatim inside the [n:] per the verbatim rule; it is not a defect we may normalize.

B. Carried type, defects and unconstruable readings

Marked [sic: …] — carried untranslated (Patterns 7, 10, 12)

1. 0507B · perficis — tot convicia a perficis hominibus pertulit. Not a word; conjecture perfidis ("faithless"), near-certain: the whole chapter is about the perfidia Judaeorum and 0508B prints apud perfidos Judaeos correctly. Carried.
2. 0509A · victuram — sculpsit in eis victuram cherubim (III Reg. 6:32). Conjecture picturam, the Vulgate reading. Carried; sits inside an italic lemma, so it takes no second italic marker (Pattern 10) — [sic: victuram], asterisks deliberately omitted.
3. 0512A · vestitvit — Et vestitvit eum auro fulvo nimis. Conjecture vestivit, which the same lemma prints correctly nine columns earlier at 0509A. Inside an italic lemma; no second italic marker.
4. 0513C · abur — si pannus lineus super ipsum abur mittatur. Conjecture ebur, printed correctly three words later (sua virtute ebur pannum … custodit). In roman prose, so italics are kept inside the marker. ⚠ The trap here was ours, not Migne's: the first draft of this chunk rendered abur as "the ivory" — a silent repair of exactly the class Pattern 7a describes, invisible afterwards because the sentence is about ivory. It was caught on the marker sweep and reverted.
5. *0516B · draeo — draeo* quippe, ut ait beatus Augustinus, serpens vel animal est. Conjecture draco, printed correctly in the next clause and throughout.
6. 0518B · quodidie — quae filiis suis … quodidie patefacit regna coelestia. Conjecture quotidie; the work prints quotidiano at 0513A and quotidie at 0518A, so this is a one-off and not an attested variant (Pattern 9 does not cover it).

### Real words that are merely wrong — rendered as ordinary English, no marker Pattern 12's boundary: these can be rendered, so the English word on the page is ours even though the reading is Migne's.

1. 0507A (×2) · munitum for minutum — Multum, inquam, manna istud munitum erat and Valde, inquam, apparuit munitum istud manna. Rendered "fortified" both times. The argument requires minutum: the sentence before it reads manna fuisse minutum describitur, the sentence after glosses se minoravit, and the entire excursus is the etymology minuo → minutum. The literal rendering therefore says the opposite of the argument, and it stands. Translating it "small" would have been the 11547/8625 silent repair with no trace left on the page.
2. 0506D · per eos for per os — quare Spiritus sanctus per eos Moysi prophetae dicat apparuisse minutum. Rendered literally, "says through them to Moses the prophet". Conjecture per os Moysi prophetae, "through the mouth of Moses the prophet" — the very phrase the work prints at 0512A (per os historiographi). The printed genitive Moysi prophetae is left dangling in the English as it dangles in the Latin; no connective was supplied to make it construe.
3. 0511D · pius for prius — Verumtamen pius considerandum nobis est. A nominative adjective where an adverb is required. Rendered "piously"; conjecture prius, "first" (which the paragraph's Superest igitur … Tunc demum … Verumtamen sequence wants).
4. 0513C · Qui for Cui — Qui ergo manifeste non clarescit vel tenuiter scienti… Nominative against the dative scienti. Rendered literally ("Who therefore … does not manifestly grow clear, even to one who knows but slightly"); conjecture Cui. This is Pattern 7a mechanism 5 (reference mis-bound), and the negation non is carried.
5. 0517B · sub terra for subter eam — quia sub terra habitant bestiae agri, where the quotation five lines above reads Subter eam habitabant bestiae agri. Rendered "under the earth". The plate contradicts its own quotation of Dan. 4:9 within one paragraph, and the English shows it.
6. 0504D · efficieris — futurum est, ut … pro nobis homo efficieris: future indicative where futurum est ut requires the subjunctive. Rendered "thou shalt be made"; no English exponent for the mood (Pattern 9 class).
7. 0512A / 0514B · nominative-for-genitive slips with no English exponent — divinitatis hominem illum quem assumpsit … provexit (for divinitas), and summitatis throni rotunda sit (for summitas). Rendered by sense; there is no emendation to make in English and none to undo (Pattern 9).

Unconstruable — rendered as closely as the printed words allow

1. 0518A — Et nisi ipse Deus homo fieret verus Deus, nullatenus cum Patre et Spiritu sancto pari honore adoraretur. The sentence does not construe: fieret wants one predicate and has two (homo, verus Deus), and the preceding sentence has already used Nisi enim Deus homo fieret in the other direction. Rendered word-for-word — "And unless he himself, God, became man true God, he would in no wise be adored…" — with no connective supplied to smooth it. Probable intent: nisi ipse Deus homo esset verus Deus, "unless this same God-made-man were true God." Escalated rather than fixed.
2. 0514A–B · floruit — In quinta autem aetate quae erat a David usque ad nativitatem beatae Mariae floruit, quia… Two finite verbs (erat, floruit) in one relative clause. Rendered by taking both with aetate and supplying the connective in brackets — "which was from David unto the nativity of blessed Mary [and] flourished" — so that the supplement is visible as ours.

Punctuation carried against the sense (Pattern 8 — preserved, never repaired)

1. 0504C — Ideo optavi adventum tuum. id est osculum tuum — full stop mid-clause, lowercase continuation. Kept: "…thy coming. that is, thy kiss".
2. 0506C, 0508A — Manhu, quod significat. Quid est hoc? The Vulgate's quod significat quid est hoc is one clause ("which means, What is this?"); Migne's full stop cuts significat off from its object. It is printed twice at 0506C and 0508A, both times the same way, so it is the plate's settled punctuation, not a slip in one line. Kept both times: "Manhu, which signifieth. What is this?"
3. 0517B — infimum in sancta Ecclesia obtinent locum isti vero. veluti spirituales… — no mark where the clause turns, a full stop three words later. Kept exactly: "…obtain the lowest place in holy Church the latter however. as being spiritual…"
4. 0518A — Duae autem manus quae tenent sedile. divinitas et humanitas Christi sunt — full stop for a comma. Kept.
5. 0512C — (si dici fas est!) — exclamation mark inside the parenthesis, kept.
6. 0507D — the citation (ibid) is printed without a closing period and inside the italic run, not as an [n:]. Reproduced exactly as printed matter, untranslated.

Chunk-boundary defect

1. 0511B, end of chunk 0008 — the file ends …et ipse est propitiatio pro peccatis nostris (I Joan. II. , The citation's parenthesis is never closed, the italic run swallows it, and a bare comma stands after the closing asterisk. Reproduced verbatim, including the orphan comma; no »-style supply and no [ed:], because nothing here is demonstrably lost from the digitization — the text simply breaks at our chunk division and chunk 0009 opens a new head. Per the standing rule on (cont.) boundaries this is ours to merge, not to mend. Flagged for the merge: whoever assembles the work should confirm against 0009 that no words fall between them.

C. Things deliberately NOT reported

Checked and settled, per the known-false-positive list: the missing A-band at 0504 (0503D → 0504B) and the missing 0508-band gaps are positional quarter-guides, not lost text — no anchor in this range breaks a word or a construction. noteCount matches the body exactly in all six chunks (17 / 11 / 10 / 14 / 8 / 14). Every Ibid.-style note ((ibid.) at 0506A, 0507D ×2, 0509A, 0509B, 0512D) resolves to the obvious preceding antecedent and none was touched. No [f: …] tags: this is a lemma-and-gloss commentary, not a florilegium, and Pattern 4's trigger (citation tails as running apparatus) is absent — Philip's citations are all [n:] notes.

Agent 3 of six. Range 0012–0017 straddles the work's quotation-typography seam; see §1, which is the section the merge must read first.

0. Terminology decisions (for the required cross-agent merge)

Set at first occurrence in this range and held throughout it. Where another agent's range fixed a different English for the same Latin, the merge (runbook step 4) must pick one — within a work there is one voice.

Second person (Pattern 17). The whole range is singular except inside plural quotations, and is rendered thou/thee/thy accordingly: the reader addressed in 0012 and 0016 (Quicunque … legis vel audis, Recordari debes, qui haec audis), the Virgin in 0013 (Domina mea), the Bridegroom in 0014 and 0015, Sobna and Pilate in 0016. Plural vos takes you even mid-page: Estote misericordes … Pater vester (0015), Ergo vos elegi de mundo … fructus vester (0015), Dei agricultura estis (0015), Vos qui secuti estis me (0015), Nescitis … ? (0015), Si semen Abrahae estis … vultis me interficere (0016), Vos autem virum homicidam … petiistis (0016), filii vestri in quo ejiciunt (0016), Non regnet peccatum in vestro mortali corpore … exhibuistis membra vestra (0017). Note 0016's Pilate quotation: Si hunc dimittis, non es amicus Caesaris is singular ("If thou release this man, thou art no friend of Caesar") sitting two lines from the plural Crucifige eum crowd — the number distinction is visible with no apparatus, which is the point of the rule.

1. ⚠ THE TYPOGRAPHY SEAM — where this work changes how it marks quotations

The seam falls between chunk 0013 and chunk 0014, i.e. at the end of the second book and the opening of the third: col. 0523 (the Explicit secundus tomus) / col. 0523B (the INCIPIT TERTIUS TOMUS).

Measured on the Latin twins in this range (body only):

Two further facts the merge needs:

• The changeover is a BOOK boundary, not a column boundary. Books I–II (chunks 0000–0013) mark scripture with italics; Book III (0014 on) marks it with guillemets. The single « » in 0013 is not a precursor: it is the Sub tuam protectionem antiphon at col. 0520B — a liturgical text, not scripture, which is exactly the class Migne guillemets even in the italic books. The single italic run left in 0017 is likewise not scripture but a printed locator, (Exod. [0533D] XXVII).
• After the seam the two systems coexist by FUNCTION, not by drift. Guillemets carry the quotation; italics inside the range now carry (a) Greek and Latin words discussed as words — Eleos, Oleos, misericordia, misericors — and (b) Migne's marginal page numbers 303–313, and (c) one scripture lemma nested inside a guillemet quotation (0016, Sap. III inside the Augustine extract). Nothing in this range was normalized in either direction; guillemet parity against the Latin twin is 1:1 in all six chunks (verify-english.mjs clean).

Do not harmonize across the seam at merge time. Chunks 0000–0013 must keep their italics and 0014–0034 their guillemets; the difference is Migne's plate, and Pattern 6 makes it the reader's only cue to what the plate prints.

1a. Unclosed guillemet — 0016, col. 0532C–0532D

The Augustine extract « Sciamus pro certo, quia post ascensionem Domini… opens at 0532C and is never closed; the chunk (and, in the Latin, the paragraph) ends mid-quotation at ad poenam aeternam. Per Pattern 5's corollary the English mirrors the plate: no » supplied. Latin twin: 21 open / 20 close; English: 21 / 20. This is the work's own defect and must not be "fixed" at merge.

1b. Guillemets that swallow the author's own prose — 0015, 0016

Three times in this range Migne opens « on a scripture lemma and closes it many lines later, after Philip's own commentary has run on inside the quotation:

• 0015 @0525A–0525B — the Jer. XXXI lemma opens the quote, and De hac nimirum ineffabili charitate Christi… Joannes apostolus dicit: plus the whole I Joan. IV citation are inside it before the » falls.
• 0016 @0530A — « Vinea, inquit Isaias, facta est dilecto meo, in cornu filio olei. then four sentences of Philip's own argument (Quoniam quidem de vinea aliquantulum… restat nunc ut…) and the whole Luc. IX voice-from-heaven quotation, all before the ».
• 0015 @0527B — the bracketed lemma [Et plantavit illam electam] is Migne's own square brackets, not ours; see §3.

All three reproduced 1:1. They are Pattern 5 cases, and any "quotation count differs" warning downstream is this, not an omission.

1c. [n:] inside the guillemets

In 0015–0017 the note marker almost always falls inside the closing » (e.g. … misericors est [n: (Luc. VI)]. »). Reproduced where it stands, every time.

2. ⛔ VULGATE DIVERGENCES — the deliverable list

Every printed quotation in the range was read word by word against the Vulgate. 34 [var: …] markers fired (0013 ×5, 0014 ×1, 0015 ×10, 0016 ×11, 0017 ×7). Below: first the ones that carry the argument, then the marked remainder, then the divergences judged immaterial and left unmarked (logged so a later sweep is not re-doing the work).

2a. Load-bearing — the exposition rests on the divergent word

1. 0015 @0525C and again @0528B, and the whole of 0015's closing argument — Zech. IV, 14. Migne prints « Isti sunt duo filii olei splendoris, qui assistunt dominatori universae terrae ». The Vulgate has duo filii olei with no splendoris. This is the single most consequential divergence in the range: at 0528B–0528D Philip builds an explicit argument on it — Sciendum praeterea quod in Zacharia propheta amplius aliquid dicitur quam in Isaia… in Zacharia vero cum additamento, splendoris oleum nominatur — and from that "addition" he derives the Father as the oil, the Son as the splendour of the oil, and a Trinitarian splendor unus. The whole peroration of chunk 0015 is exegesis of a word the received text does not print. Marked at the first occurrence only (0015 @0525C); the later occurrences are the same lemma re-quoted and are unmarked by design.
2. 0013 @0520A — Ps. XLIV, 7. Migne: « Thronus tuus, Deus, in saeculum saeculi »; Vulg. Sedes tua. The entire chapter is an allegory of Solomon's throne applied to the Virgin, and it is introduced as Tu es ille thronus de quo Psalmista dicit — the proof-text supplies the word only because the plate reads Thronus. Conforming it to Sedes would have voided the proof.
3. 0013 @0519C — III Reg. X, 18. Migne: auro clarissimo nimis; Vulg. auro fulvo nimis. The gold is glossed as the divinity, and the Virgin is then called omni claritate et splendore solis clarior — clarissimo is picked up, fulvo is not.
4. 0015 @0527C — Prov. XVIII, 10. Migne: ad eam confugiet justus, et salvabitur; Vulg. ad ipsum currit justus et exaltabitur. The gloss that follows is Homines namque ideo ad turrim confugiunt… in qua homines salvantur — both of Philip's verbs come from Migne's reading, and neither from the Vulgate's.
5. 0017 @0536B — Ezech. XV, 2. Migne: quod pretiosum est prae caeteris lignis silvarum; Vulg. ex omnibus lignis nemorum quae sunt inter ligna silvarum, with no pretiosum. The vine-simile turns on the comparative quanto prius exstitit charior, pretiosior… two lines earlier.
6. 0016 @0532B — Prov. XXXI, 23. Migne: Nobilis in portis vir sanctae Ecclesiae, cum sederit cum senatoribus populi sui; Vulg. vir ejus, quando sederit cum senatoribus terrae. The ecclesial reading is written into the lemma itself, and Philip then observes that Solomon "names them the senators of the people" — a claim true only of this text.
7. 0015 @0526B — Luc. I, 69. Migne: Erexit cornu salutis in domo David; Vulg. erexit cornu salutis nobis in domo David. Philip's own next sentence reads hoc est cornu salutis quod erectum est nobis — he knows the fuller text; the plate's lemma has lost it.

2b. Marked, material but not load-bearing

• 0013 @0521A, Deut. XXI, 23 — Maledictus … qui in ligno pendebit vs Vulg. maledictus a Deo est qui pendet in ligno (future for present; a Deo absent).
• 0013 @0521B, Prov. XXV, 11 — verbum prolatum in tempore suo vs Vulg. qui loquitur verbum in tempore suo. NB Migne prints prolatum outside the italic run; the three-run italic structure is reproduced exactly, and the gloss Malum quippe prolatum depends on it.
• 0013 @0522C, Luc. I, 30 — apud Dominum vs Vulg. apud Deum (and enim dropped).
• 0014 @0523D, Isa. LIII — « Vulnera nostra sanavit » is not Vulgate wording at all; Vulg. 53:5 livore ejus sanati sumus.
• 0015 @0525B, I Joan. IV — non quia nos dilexerimus eum vs Vulg. 4:10 non quasi nos dilexerimus Deum, conflated with 4:12 caritas ejus in nobis perfecta est.
• 0015 @0526A, Luc. I, 68 — Benedictus Dominus Deus Israel vs Vulg. Benedictus Deus Israel; also plebis suae for plebi suae.
• 0015 @0526D, "Col. III" — the text quoted is Ephes. V, 25, and diverges from it: uxores vestras, sanctam Ecclesiam, tradidit semetipsum propter illam vs Vulg. sicut et Christus dilexit ecclesiam et se ipsum tradidit pro ea. Marked for the wording only. The misattribution to Col. III is NOT corrected — printed references stand (and a following Ibid. chain must never be disturbed).
• 0015 @0526D, Cant. I, 14 — Quam pulchra es … quam pulchra es vs Vulg. ecce tu pulchra es … ecce tu pulchra. Philip's gloss (repetendo duplicat) survives either reading, so this is marked but not listed in 2a.
• 0015 @0527B, Joan. XV — Ergo vos elegi de mundo ut eatis conflates 15:16 with 15:19.
• 0015 @0527C, Act. IV, 12 — Migne adds a closing clause the Vulgate lacks: nisi in nomine Domini nostri Jesu Christi; and in quo oporteat salvari for in quo oporteat nos salvos fieri. The nisi is carried as "save" (Pattern 7a).
• 0015 @0527D, Isa. V, 3 — order reversed (viri Juda et habitatores Jerusalem) and the second inter dropped; the gloss id est de vinea mea is Philip's own and sits inside the guillemets.
• 0016 @0529A, "Luc. XX" — the parable is quoted largely in Matthew's words (Homo quidam paterfamilias… locavit eam agricolis) against a note reading Luc. XX; and the conspiracy speech (Ecce hic est haeres: Iste igitur occidatur, et sic ad nos haereditas transferetur) matches neither evangelist verbatim. Two markers.
• 0016 @0530A, Luc. IX, 35 — Filius meus dilectus in quo mihi complacuit vs Vulg. Luke's Filius meus electus.
• 0016 @0530A, Cant. II, 16 — qui pascitur in liliis vs Vulg. inter lilia.
• 0016 @0530D, Isa. XXII — 17 and 19 conflated into one sentence (expulsion de statione tua belongs to v. 19).
• 0016 @0530D–0531A, Jer. XXII, 24 — Jeconias rex Juda sicut annulus aureus in manu mea vs Vulg. Jechonias filius Joachim rex Juda anulus in manu dextera mea; and evellam eum et projiciam, an added verb.
• 0016 @0531C, Joan. VIII — Si semen Abrahae estis (from v. 37) welded to v. 39, and utquid vultis me interficere for nunc autem quaeritis me interficere.
• 0016 @0532A, Act. III, 15 — auctorem vitae morti tradidistis vs Vulg. interfecistis.
• 0016 @0532A, Matth. XII, 27 — in principe daemoniorum for in Beelzebub, plus ut dicitis imported from the Lucan parallel.
• 0016 @0532B, Isa. III, 14 — cum senioribus populi sui for cum senibus populi sui et principibus ejus; matters because the next sentence aligns seniores with Solomon's senatores.
• 0016 @0532C, Joan. XII, 26 — Pater volo ut ubi ego sum, illic sit et minister meus fuses Joan. XVII, 24's Pater, volo ut onto XII, 26's ubi sum ego, illic et minister meus erit.
• 0017 @0534A, "Luc. VI" — the sentence quoted is Matth. VII, 18's (Non potest arbor bona fructus malos facere); Luke 6:43 reads non est arbor bona quae facit fructus malos.
• 0017 @0534A, Matth. XXV, 1 — Simile est for Vulg. simile erit.
• 0017 @0534B, Isa. IX, 5 — erit in combustionem ignis, Vulg. erit in combustionem et cibus ignis.
• 0017 @0534B, Rom. VI, 19 — Migne prints the infinitive ita nunc exhibere membra vestra where the Vulgate has the imperative exhibete. Rendered as an infinitive ("so now to present your members"), which leaves the sentence without a main verb; that is the plate's state and is not repaired.
• 0017 @0535A, II Petr. I, 14 — Scio, fratres, quod velox sit… dignatus est revelare mihi vs Vulg. certus quod velox est… significavit mihi.
• 0017 @0535B, Rom. V, 5 — per inhabitantem Spiritum vs Vulg. per Spiritum Sanctum.

2c. Divergences examined and left UNMARKED (immaterial, or not scripture)

Recorded so that a later pass does not have to re-derive them.

• 0012 @0519A Sap. VII — Anima justi sedes est sapientiae is not in the Vulgate at all, at Wis. VII or elsewhere; it is a proverbial tag. No [var:] fired, because Pattern 14 compares wordings of a cited verse and there is no verse here to compare with; and Migne's reference is not ours to correct. Flagged for the indexer: this [n:] will not resolve to a real locus.
• 0013 @0519D Exaltata es super choros angelorum ad coelestia regna — introduced as sicut scriptum est but it is the Assumption antiphon, not scripture. Same treatment as above: translated, unmarked.
• 0013 @0520B Sub tuam protectionem confugimus… — the Sub tuum praesidium antiphon in an expanded form. Not scripture; the only « » in 0013 (see §1).
• 0013 @0521A maledictus, sive maledicta quae non reliquerit semen in Israel — no such Vulgate verse; Philip's own summary of the Mosaic law. The masculine / feminine doublet has no English exponent (Pattern 9) and is rendered "cursed is he, or cursed is she" so that both printed forms survive.
• 0013 @0521A Eccli. XXIV, 12 — Creator omnium is fronted out of the earlier clause of the same verse; compression, not divergence.
• 0013 @0521B Psal. XVIII, 6 — procedens dropped, but the host sentence supplies processit. Immaterial.
• 0013 @0521B Cant. I, 12 — Cum for Vulg. Dum esset rex.
• 0014 @0523C Philip. II, 9 — Dedit Deus ei nomen for donavit illi nomen; sense unchanged.
• 0014 @0523C Rom. IX, 5 — first citation truncated (est super omnia benedictus, no Deus); the second, at 0524B, prints the full qui est super omnia Deus benedictus. Both rendered as printed.
• 0015 @0525B Exod. XXVII, 20–21 — verbatim but for word order (ardeat semper lucerna).
• 0015 @0525C Isa. V, 1–2 — verbatim, including the Vulgate's own difficult in cornu filio olei, which is rendered literally ("on a horn, the son of oil") because Philip's entire allegory parses it word by word.
• 0015 @0525D Luc. VI, 36 — ergo dropped.
• 0015 @0527B Cant. II, 15 — vineas nostras picks up vinea nostra from the following clause of the same verse.
• 0015 @0528A Sap. III, 8 — Judicabunt sancti nationes, subject supplied from the context of the sentence introducing it.
• 0015 @0528C Hebr. I, 3 — gloriae dropped from splendor gloriae.
• 0016 @0529A Isa. V, 7 — viri Juda plural for Vulg. vir Juda.
• 0016 @0530D Isa. V, 2 — Elegit Dominus lapides ex illa, subject supplied and the order changed.
• 0016 @0532A Joan. XIX, 6 — Crucifige eum for the doubled crucifige crucifige eum.
• 0017 @0533D Isti sunt viri misericordiae, quorum justitiae oblivionem ante Deum non acceperunt — attributed to beatus Paulus, but it is Eccli. XLIV, 10 (viri misericordiae sunt quorum pietates non defuerunt), and even then not verbatim. Migne prints no quotation mark of any kind here, so per Pattern 6 the English supplies " " — ours, not his. Attribution left as printed.
• 0017 @0535A Marc. VI noted against lucerna corporis tui… est oculus tuus, which is Matth. VI, 22 / Luc. XI, 34. Reference untouched.
• 0017 @0536D Deut. XXXII, 32/35 — both clauses verbatim; the connective usque ad illum locum, ubi ex persona Dei ait and the closing subauditur, in die judicii quod merentur are Philip's own and sit inside the guillemets.
• 0017 @0537A Osee. VII — vv. 11 and 9 conflated; both verbatim in themselves.
• 0017 @0537B Psal. LIII, 5 — quidem added. Prov. V, 9 — verbatim, in two pieces (Ne des alienis honorem tuum … et annos tuos crudeli), the second quoted separately with its own guillemets and its own Ibid. note; both reproduced as two quotations, not merged. Joan. VIII, 44 — Quia ipse for ille.

3. Carried type, [sic: …] — five instances

Wrapped per Pattern 12; each is verbatim in the Latin twin (verifier check 9).

Deliberately NOT marked (Pattern 12's boundary — a real word merely wrong is rendered, not carried):

• 0016 @0530C — Oleos for Eleos. At 0525D the plate prints Eleos siquidem Graece; at 0530C the same etymology is reintroduced as Oleos enim Graece, misericordia Latine dicitur. This is the most interesting reading in the range and it is not a [sic:]: the transliterated Greek is already carried in italics untranslated on both occasions, so a marker on one and not the other would blame the plate for what is, in form, the same operation — and Oleos is precisely the contamination the argument invites, since Philip is punning oleum against ἔλεος across four columns. Rendered as printed, both times, and logged here. A reviewer who "fixes" 0530C to Eleos has destroyed the evidence.
• 0015 @0528B — qui assistant Dominatori, subjunctive, against assistunt in the same lemma at 0525C. A mood difference with no English exponent in a relative clause (Pattern 9); rendered "who stand before" and logged, not marked, not emended.
• 0013 @0523A — liber Regnum for liber Regnorum. A real (if wrong) form; rendered "the book of Kings."
• 0017 @0534D — offeret (future) where offert is expected; rendered as a future ("which it will offer").
• 0013 @0519D — omnia esse mea possent loqui. esse is almost certainly for essent/entia; rendered "all that is mine could speak", logged.

4. Punctuation and capitalization carried from the plate (Pattern 8)

Migne's stops are reproduced, including where they break a sentence. Where the plate prints a full stop followed by a lower-case word, the English keeps the lower case, because that is what makes the misplaced stop visible on our page; this is a per-range convention and the merge should extend or reverse it work-wide, not leave it split.

• 0012 @0519A–0519B — the twelve sacraments are listed with two internal full stops (…catholicae fidei. susceptio sacri baptismi, communicatio corporis et sanguinis Christi. confessio…). Kept.
• 0013 @0519D — quia sicut scriptum est. followed by the antiphon. Kept.
• 0013 @0520D–0521A — Sicut enim Filius ejus maledictum legis … incurrit. Maledictus nempe, ait Moyses… The stop severs the sicut … sic correlation; kept, and the English reads as the plate does.
• 0014 @0523D — Adhibitum igni sedentibus in tenebris lumen donat. et Christus… Kept, lower-case and.
• 0015 @0526D–0527A — sanctam Ecclesiam sepivit? vinea nempe vocatur… Kept.
• 0015 @0528A — inside the Matth. XIX quotation: sedebitis super sedes duodecim. judicantes duodecim tribus Israel. Kept, lower-case "judging".
• 0016 @0529A — ejecisti gentes, et plantasti, eam: the stray comma before eam kept.
• 0017 @0533A — Enim vero si tandiu volumus exspectare loqui de bono, donec simus boni; vel si tandiu de perfectione tacemus donec perfecti simus. A protasis with no apodosis; the sentence simply stops. Rendered as printed, nothing supplied.

Migne's own square brackets at 0015 @0527B — [Et plantavit illam electam] — are his, not ours. Reproduced as brackets; the content is translated (unlike [ al., …] variant readings, which stay Latin), because it is a scripture lemma the next clause expounds, not apparatus. Flagged for the merge in case another agent met the same construction.

Marginal page numbers 303 (0012), 304 305 (0013), 306 (0014), 307 308 (0015), 309 310 (0016), 311 312 313 (0017) are printed in italics mid-sentence and are reproduced verbatim in position, untranslated. They are the reason a naive italic-extraction of this work over-counts italic quotations.

5. Passages that do not construe — rendered literally, conjecture here only

1. 0012 @0518D–0519A. Et quoniam thronus rotundus dicitur fuisse, manifeste indicat quia thronus Dei desiderat esse, omni perfectioni studeat, mentem et linguam corpusque ab omni peccato custodiat. There is no subject for desiderat and no connective before the two jussives: something like quisquis thronus Dei desiderat esse has dropped out. Rendered as printed ("it plainly shows that he desires to be the throne of God, let him strive…"). No word was supplied to make it construe.
2. 0014 @0523C–0523D — a lost clause, and the most likely real hole in the range. …sicut Petrus dicit: « Qui peccatum non fecit [n: (I Petr. II)] » homines reficit, et Christus nos sanguine et carne sua potat et satiat. The passage is running through the seven natures of oil, one per sentence; homines reficit is the predicate of a new nature ("oil refreshes men") whose subject oleum is missing, so the Petrine quotation and the refreshing clause have been welded together. Rendered exactly as printed. No [ed: …] was fired: Pattern 13 requires demonstrated loss in our digitization, and this cannot be distinguished from a defect in Migne's own plate without the leaf. If a plate check is ever done in this work, this is the column to look at.
3. 0013 @0522C. Modestiam ipsius, id est religionem quibus praeconiis possumus, ut dignum est attollere, quam scimus… — an anacoluthon (no quibus praeconiis … attollere possumus main clause, and it closes with a full stop, not a question mark, unlike the parallel sentence that follows it). Rendered close and left declarative, per Pattern 8.
4. 0017 @0537A. Nonnullos enim magnae religionis et satis ardui propositi novi viros, quos utinam non novissem, qui, quandiu licet eis… — the accusative nonnullos … viros never receives its verb; the sentence runs on into a tandiu clause. Rendered as one long sentence with the anacoluthon intact.

6. Negation audit (Pattern 7a)

Every printed non / nec / neque / nisi / nullus / nullatenus / nunquam / nihil / minime / haud in the six chunks appears in the English, and none was added. Two places where carrying it makes the sentence read oddly, and where it was carried anyway:

• 0017 @0533A — nusquam Deum laudabimus, quoniam sicut necesse est, nullatenus in hac vita unquam boni et perfecti … erimus digni. Both the nullatenus and the unquam are rendered ("we shall in no wise ever in this life be…"), producing the flat pessimism the sentence actually states.
• 0015 @0527C — Act. IV's Non enim est aliud nomen … nisi in nomine Domini nostri Jesu Christi: the nisi-clause is Migne's addition to the verse (§2b) and is rendered "save in the name of…", keeping both the negation and the exception.

No silent repair was made anywhere in this range; where a sentence would not construe it is listed in §5 with the conjecture kept out of the text.

7. Notes

All 101 [n: …] markers in the range (0012 ×2, 0013 ×12, 0014 ×8, 0015 ×30, 0016 ×22, 0017 ×27) are citations and stay verbatim Latin. No [nt: …] fired — Pattern 15's prose-note class does not occur in this range; every note is a book/chapter locator, several of them bare (ibid.) whose antecedents were left untouched.

8. Verification state

node scripts/verify-english.mjs 11638 → OK. For this range specifically: column anchors 1:1 and in order in all six chunks (including 0013's plate-order oddity [0523A] … [0524A] … [0523], which is reproduced as printed and is not a defect); note counts 2 / 12 / 8 / 30 / 22 / 27 matching; guillemets 0-0 / 1-1 / 6-6 / 29-29 / 21-20 / 26-26 matching; one ## head per chunk.

Agent 4 stint (cols 0537C–0562B). Filed under the first chunk of the range per the runbook; this file is one input to the required merge-time reconciliation across the six agent stints (translation-style.md, Song-of-Songs ruling, "⚠ THE BOUNDARY" paragraph — within a work, one voice).

1. Terminology decisions made in this range — read these first at merge time

These are the choices a neighbouring stint most likely made differently. Where a peer chose otherwise, reconcile to ONE form across the work; my reasons are given so the merge can adjudicate rather than coin-flip.

2. ⛔ Vulgate divergence list — the deliverable

Every printed scripture quotation in 0018–0023 was read word by word against the Vulgate. Marked with [var: …] in the text (material divergence):

Divergences observed and deliberately NOT marked (immaterial to sense; logged so a later sweep does not re-open them): 2 Petr. 2:22 word order ad vomitum suum / ad suum vomitum · Cant. 5:8 Nuntiate dilecto meo, quia amore ejus langueo for ut nuntietis ei quia amore langueo · Cant. 7:12 fructus claritatis et bonae operationis parturiunt (the author's own gloss taken up into the quotation) · Jer. 15:19 word order · Joan. 15:4 non valet fructificare for non potest ferre fructum a semet ipso · Cant. 5:6 ut Dilectus meus mihi locutus est for ut locutus est · Amos 1:3 non convertam without eum · Jer. 31:3 miserans tui · Joan. 12:32 traham ad me without ipsum · Joan. 8:25 Ego principium qui loquor vobis (the patristic form, universally so cited) · Gen. 1:2 tenebrae erant · Gen. 1:2 Tenebrae erant super abyssum without faciem (the author's own next sentence restores faciem) · Gen. 2:7 Fecit for formavit, Dominus dropped · Gen. 3:10 Audivi, Domine, vocem tuam · Cant. 6:1 heavy compression · Cant. 2:4 Introduxit me rex · II Reg. 12:13 Peccavi Domino, ora pro me and the expanded Nathan reply · I Reg. 24 / 26 conflation and rewording throughout David's two speeches · II Reg. 2:6 pro opere quod fecistis in eo for quoniam fecistis verbum istud · II Reg. 1:17–27 re-ordered: Migne runs vv. 24, 23, 21, 22, 19/27 in that sequence, and prints abiit for rediit and robusti et nobiles et egregii viri for fortes · Apoc. 1:5 lavit sanguine suo without nos a peccatis nostris · II Reg. 18:33 quis mihi det for quis mihi tribuat · Exod. 32:31 peccatum maximum for magnum, Domine added, educ for duc · II Reg. 24:17 oro for obsecro · I Reg. 16:1 abjecerim for proiecerim · Cant. 8:6 Fortis est ut mors charitas, sive dilectio (author's doublet) · Ps. 118:131 attraxi Spiritum sanctum at 0023 where 0019 prints Spiritum and the author himself supplies subauditur sanctum — the plate is inconsistent with itself across four columns, and both readings are rendered as printed.

3. Plate defects, carried type, and punctuation carried under Patterns 7–8, 10–12

Carried type — [sic: …], two instances only:

• 0019 @0543D armaliolum — infra armaliolum cordis nostri. Not a word; conjecture armariolum, "little cupboard/casket", near-certain (the sentence needs a container to receive the charismata Spiritus sancti). Carried untranslated in italics per Pattern 7's non-word class.
• 0023 @0558A repacicavit — reconciliavit et repacicavit. Not a word; conjecture repacificavit, near-certain from the paired verb and from pacificando conjungeret three columns earlier (0546B). Carried; the marker follows the sentence because the English clause "made peace again" renders the sense the sentence must bear, and the broken run is exhibited beside it rather than inside it — nested italics inside the clause were not expressible without splitting the verb pair.

Misplaced quotation marks — reproduced 1:1 in place, per Pattern 5's corollary. All three are Migne's, not ours:

• 0018 @0542A–B — « opens at Isti sunt filii olei splendoris… (Zach. 4:14) and the closing » does not arrive until after the author's own prose (Quia vero persaepe jam superius praefati sumus quod oleum significet misericordiam…) and a second quotation (Ps. 58:18). One printed quotation therefore swallows two verses and a sentence of commentary. Reproduced exactly; the English reads as strangely as the Latin does, which is the point.
• 0019 @0541C — « opens at Trahe me post te. and closes only after the Amos citation, swallowing Sicut in octonario opere Dei in libro Genesis, et in octo sententiis Amos prophetae, ubi dicitur: — the author's own comparison. Same treatment.
• 0022 @0552B–C — a genuine unclosed/orphaned pair. The Nathan speech (Noli, inquit, timere…) opens with « and never closes; three sentences later an orphan » appears after Mea, inquit, manus auxiliabitur ei [n: (ibid.)], with a space before the comma that follows it. Chunk-level counts balance (23/23) only because the two errors cancel. Both marks are reproduced where they stand.
• 0023 @0557C — a bare orphan » inside Caput quippe Christi secundum apostolum Paulum » Deus est [n: (I Cor. XI)], with no opening mark anywhere near it. This is why 0023's counts are 26 « against 27 », in the Latin and in the English alike. Reproduced; not balanced by supplying an opening mark.

Punctuation carried under Pattern 8 (preserve, never supply, never delete):

• 0018 @0538A — full stop mid-sentence: interiit amor deceptoris hujus mundi. et ad nihilum redacta est…, the plate continuing in lower case. The stop is kept; English capitalization after a stop is our orthography, not a supplied mark.
• 0018 @0539C — Illud quoque quod Isaias propheta tantopere exprimit, ut asserat quod Deus ex vinea sua lapides elegerit. closes with a full stop although the clause has no main verb; the next sentence begins De nostra anima dictum non dubitemus. Kept as two sentences.
• 0021 @0548C, 0022 @0551C, 0023 @0560D — a printed ? closing a statement: …in atria Domini? (Ps. 83:3), …in cathedra pestilentiae non sedit? (Ps. 1:1, and the same verse again at 0551D as etc.), …et implebo illud? (Ps. 80:11). All three are declaratives in their sources. Kept as printed.
• 0021 @0550B — the chunk ends with no terminal stop: « Gaudium vestrum nemo tollet a vobis [n: (Joan. XVI)] ». Not supplied.

4. Readings that will not construe — rendered as printed, conjecture here only

Pattern 7a: none of these is repaired in the running text.

• 0019 @0541C inferre — quatenus velamen litterae Veteris Testamenti dignetur nobis inferre, ut … sensum spiritualem … valeamus reperire. Printed inferre ("bring in upon") makes the prayer ask God to impose the veil so that the spiritual sense may be found — self-contradictory. Conjecture auferre ("take away"), which is what the purpose clause requires and what the same author asks in the parallel at 0555B. Rendered "bring in upon us" as printed. This is a real word wrongly used, not broken type, so no [sic:] (Pattern 12's boundary).
• 0019 @0541C, no main verb — Et enim velamen magnum super istam sententiam positum, scilicet: «…» has no finite verb (est wanting). Rendered as the verbless clause it is.
• 0020 @0545A–B, the esse passage — sed quod ab eo creatum est, ut per ipsum consistat qui creavit, necesse est per se esse, et sine alio esse, unum dignoscitur esse, quod ad Deum proprie pertinet. As printed this asserts that the creature must be per se — the exact opposite of the sentence that follows (Per alium esse et sine alio esse non posse, creaturae proprium comprobatur esse). Something has dropped, most likely a non or a per ipsum for per se; the corruption is not localizable to a single word, so no marker fires. Rendered as printed, including the self-contradiction. ⚠ This is the passage in the range most likely to be "fixed" by a later reader; it should not be.
• 0023 @0560B abs re — tunc profecto nunc abs re factum esse creditur, quoniam quidem quadraginta duae virtutes in serie hujus catalogi positae inveniuntur. The idiom is non abs re ("not without purpose"), and the whole following paragraph exists to show the forty-two was purposeful (42 mansions, 42 generations). The printed text therefore says the opposite of what the author immediately proves. A dropped negation of exactly the Pattern 7a class. Rendered "beside the purpose" as printed; the non is NOT supplied.
• 0023 @0559C–D Eam majorem — Eam majorem quippe charitatem nemo habet, quam ut pro persecutoribus suis exoret… Expected Ea major / majorem hac (Joan. 15:13). Eam is a real word in the wrong case; rendered literally ("no man has that greater charity, than that…"). No [sic:] per Pattern 12's boundary.
• 0022 @0553D debuisses — audita morte ejusdem inimici sui Saulis de cujus utique interitu debuisses gaudere — a second-person singular pluperfect subjunctive in a clause whose subject elsewhere is David (debuisset expected). Rendered as printed with thou (Pattern 17: the morphology decides, not the sense).
• 0022 @0554C, number shift mid-sentence — o tu amator vel auditor divini verbi … te oportet … considerare et diligenti intuitu … perpendite. Singular address then a plural imperative in one sentence. Both kept: "it behoves thee to consider … and do ye weigh". Not smoothed to one number.
• 0023 @0555D fratribus vestris — et cum fratribus vestris nullatenus charitativi apparemus — "your brethren" governed by a first-person plural verb; nostris expected. Rendered as printed.
• 0022 @0554A in montem Gelboe — accusative after in where the locative ablative is required (in monte, as the same chunk prints at 0553C). Rendered "on Mount Gelboe"; English has no exponent for the case error, so Pattern 9 applies — logged, not marked.
• 0023 @0560A fortitudo in adversariis — "fortitude among adversaries" where the catalogue's parallel member is temperantia in prosperis and in adversis is expected. Construable as printed; rendered literally.
• 0023, Ramasse / Ramesse — the same station is spelled two ways nine lines apart (@0561A Ramasse, @0561B Ramesse, the latter inside the Exod. 12:37 quotation). Both kept as printed.
• 0020 @0546B Sathaelis — the fallen angel named Sathael, not Satanas. Rendered "Sathael", not normalized.

5. Reference observations — recorded, no correction proposed

Per the standing rule (refDisplay keeps what Migne prints), none of these is altered and no citation-corrections.json entry is implied.

• 0018 @0540A — the quotation is introduced in libro Sapientiae scriptum reperitur while the note reads [n: (Eccle. VII)]; the verse is Eccle. (Qoheleth) 7:30. The prose and the note disagree about the book. Both stand.
• 0018 @0542B — [n: (Joan. III)] for I Joan. 3:2, twice in the range. Migne's normal abbreviation for the epistle in this work; not an error.
• 0020 @0546D — [n: (II Par. II)] for Solomon's Tu solus nosti corda filiorum hominum, which is II Par. 6:30.
• 0022 @0552A — [n: (II Reg. II)] on the Bethsabee adultery, which is II Reg. 11. The same note content (II Reg. II) recurs correctly at 0553D for the Jabes Galaad message (II Reg. 2:5–6), so the two are not a single systematic slip.
• 0023 @0562A — [n: (Psal. CVII)] on vox exsultationis et salutis est in tabernaculis justorum, which is Ps. 117:15.
• 0023 @0556B — [n: (Isa. LXIII)] sits before the clause it supports rather than after; reproduced in position.

6. Marker counts, verified against the Latin twin

Column anchors match the Latin twin sequence exactly, in order, including the out-of-sequence runs that are Migne's own: 0019 opens [0541] … [0541C] … [0542C] (the third-tome colophon sits at 0541 while the fourth tome's body begins at 0541C), and 0020 closes [0547A] … [0548A] … [0547] (the Explicit quartus tomus colophon is printed at 0547 after text at 0548A). Both are the known false-positive class 1 — band arithmetic, not a break.

0023's guillemet asymmetry (26 open / 27 close) is Migne's and is present identically in the Latin twin; see §3.

7. Pattern 17 (second person) — how it was applied here

Singular addressees, all rendered thou/thee/thy: the auditor (0019 @0542C, 0020 @0544D, 0022 @0551D and @0554C, 0023 @0558B), Moses apostrophized (0020 @0544C, 0023 @0558C), the individual conversus (0018 @0541C), God/Christ addressed by the Bride and the maidens throughout 0021, David's interlocutors in 0022, and every scriptural tu (Osee 2, Jer. 15, Gen. 3:19, Ps. 80:11, I Reg. 24 and 26, Exod. 32, II Reg. 24, I Reg. 16, Luc. 19:42).

Plural addressees, rendered you/your: Zach. 2:8 Qui tangit vos, Joan. 15:4 nisi in me manseritis, Jac. 5:8 Patientes et vos estote, Luc. 21:19 In patientia vestra, Matth. 25:34 quod vobis paratum est, Joan. 16:22 Gaudium vestrum, Matth. 5:44 Diligite inimicos vestros (both occurrences), II Reg. 2:5–6 Benedicti vos, II Reg. 1:24 Filiae Israel … cultui vestro, I Reg. 12:23 orare pro vobis, Matth. 10:20 Spiritus Patris vestri, and 0023 @0555D's cum fratribus vestris.

The two are adjacent within single sentences more than once (0022 @0554C, above); in every such case both were carried rather than levelled.

PL 203, coll. 0561–0574. Agent 5 of 6. Range: the close of Book V (0024) and the whole of Book VI (0025–0029), which expounds Cant. 1:4–5 (Nigra sum, sed formosa) four times over — allegorically of the Church, of the flesh of Christ, of our nature, then secundum moralitatem of the soul, and finally of the blessed Virgin.

1. Terminology decisions (for the merge — reconcile against the other five agents)

The lemma of Cant. 1:4–5 recurs some forty times in this range, in fragments. One English form was fixed for each fragment at first occurrence and never varied:

Other fixed equivalents in this range:

• tomus (the work's own division: Explicit quintus tomus, INCIPIT SEXTUS TOMUS, Explicit sextus tomus) → "book". Not "volume" and not "tome": Philip is numbering the parts of one work.
• moralitas → "the morality", in the heads (SECUNDUM MORALITATEM → "ACCORDING TO THE MORALITY") and in the prose (moralitatem quae in ea latet → "the morality which lies hidden in it"). Chosen over "the moral sense" because it is the work's own title-word (Moralitates) and because Philip pairs it with allegoria, which is not glossed either. Flag for merge: if another agent has "moral sense", one of us must move.
• anagoge → "the anagoge" (0024 head).
• recapitulatio → "recapitulation"; a superioribus → "from what went before".
• Sponsa → "the Bride"; adolescentulae → "the maidens".
• perfidus / perfidia of the Jews → "faithless / faithlessness" throughout (never "perfidious"), matching infidelitas → "unbelief" as a distinct word.
• sol justitiae → "the sun of justice"; formositas → "comeliness", to keep it visibly cognate with formosa.
• fusca → "swarthy", kept distinct from nigra "black", because Philip glosses one by the other (fusca, id est nigra appareo, 0564B) and the gloss dies if the two words collapse.
• clarificare → "to make bright" / claritas → "brightness", kept cognate through the resurrection passage at 0565D–0566A, where the same root carries the argument from decoloravit to clarificabit.
• Jerusalem (0025) and Hierusalem (0029) are the same name in two spellings. English has one exponent; both render "Jerusalem". Pattern 9, no marker.

2. Vulgate divergences — the deliverable list

Every quotation printed in the range was read word by word against the Vulgate. Fifteen divergences were judged material and carry a [var: …]; the rest are recorded here without a marker.

2.1 Marked [var: …]

1. 0563A — Cant. 1:4, sicut pellis Salomonis (Vulg. sicut pelles Salomonis, plural). The most load-bearing divergence in the range, and the reason this chunk exists: from 0563B to 0564A Philip builds an argument out of the singular. He says he can find pellis Salomonis written nowhere else in either Testament, raises it as a quaestio magna, and resolves it by the figure a parte totum — one skin standing for the whole skin-built tabernacle at Gabaon. A plural "curtains/skins" would leave nothing to explain. The English is singular everywhere the Latin is singular, including in the recapitulations at 0565D, 0569A, 0570A, 0572B.
2. 0562C — Prov. XVII, « Sicut igne probatur aurum, sic justos probat tribulatio tentationum. » Not the Vulgate's Prov. 17:3 (quomodo igne probatur argentum et aurum camino, ita corda probat Dominus) at any point: the subject, the object and the verb all differ. Cited as Prov. XVII notwithstanding; the reference is left as printed.
3. 0563C — Exod. XXV, « Vide, inquit, ut tu facias tabernaculum secundum formam illius tabernaculi quod in monte vidisti. » Vulg. Exod. 25:40 is inspice et fac secundum exemplar quod tibi in monte monstratum est; Migne's wording is nearer Hebr. 8:5 but is not that either (tabernaculum is supplied twice).
4. 0563D — I Par. XVII, arca Dei est in medio pellium. Vulg. 1 Par. 17:1 has arca foederis Domini sub pellibus est. Material because the whole inference — that Moses' tabernacle was made of skins — is drawn from this clause.
5. 0563D — Psal. LXVII, « Ascendens Christus in altum, captivam duxit captivitatem. » Attributed to David, but Ps. 67:19 reads ascendisti in altum, cepisti captivitatem; the printed wording is Ephes. 4:8, with the name Christus added, which stands in neither.
6. 0564D — Act. XX, lupi graves, qui non parcent vobis. Vulg. Act. 20:29: lupi rapaces … non parcentes gregi. Two changes, and the second moves the object of the sparing from the flock to the hearers.
7. 0566B — Rom. V, per obedientiam unius justi constituentur multi. Vulg. Rom. 5:19: per unius oboeditionem iusti constituentur multi, where iusti is the predicate of multi ("many shall be made just"). In Migne the word has migrated into the genitive phrase and become "of one just man", leaving multi with no predicate at all. Rendered as printed ("many shall be constituted"), which is awkward English and is meant to be: the printed clause is incomplete. This is the range's purest 7a″ trap — the familiar English ("many shall be made just") arrives instantly and would silently restore a word the plate does not carry.
8. 0566D — Gen. III, « Si de fructu vetito comederitis… » The Vulgate's serpent says in quocumque die comederitis ex eo; "of the forbidden fruit" is not in Gen. 3:5 in any form.
9. 0567B — Luc. VIII, « Sole orto aruerunt. » Luc. 8:6 reads natum aruit; sole orto belongs to Matt. 13:6, where the verb is aestuaverunt. Migne's clause is a fusion of the two and stands as such in neither.

Also marked, in the same class:

1. 0566A — Joan. XVII, claritate quam habui antequam mundus fieret. Vulg. 17:5 has the accusative claritatem … priusquam mundus esset apud te. The ablative changes the construction ("with the brightness") and apud te is gone.
2. 0565C — Joan. XII: « Videtis quia nihil proficimus, ecce totus mundus post ipsum vadit. Si dimittimus eum sic, omnes credent in eum. » Sentence 1 is Joan. 12:19 but with totus mundus … vadit for mundus totus … abiit; sentence 2 is Joan. 11:48, under the single citation (Joan. XII).
3. 0566B — Luc. XIII, relinqetur (future) for the Vulgate's present relinquitur.
4. 0570D — Thren. III, plural Oculi mei depraedati sunt for the Vulgate's singular oculus meus depraedatus est. Marked because the exposition immediately following is plural (infelices oculi … depraedantur et oppugnant) and so rests on the printed number.
5. 0571A — Matth. XII, Ex ore tuo justificaberis, et ex ore tuo condemnaberis. Vulg. 12:37: ex verbis tuis. Material and marked because the entire paragraph is about the mouth (os nostrum … os nostrum nos justificat), and the Vulgate's "words" would leave the proof-text arguing something adjacent to the point.
6. 0571B — Eccli. XIII, « Sicut qui tangit picem ab ipsa inquinatur, sic qui tangit uxorem proximi sui, immundus prorsus efficitur. » Only the first limb is Sir. 13:1; the Vulgate's second limb is qui communicaverit superbo induet superbiam. The clause about the neighbour's wife is Prov. 6:29 in substance and stands nowhere in Ecclesiasticus.

(Count as filed: 0025 ×6, 0026 ×3, 0027 ×3, 0028 ×3 = 15 markers.)

2.2 Divergences logged but NOT marked (minor: no change of sense)

• 0563A Hebr. XIII Non habemus hic — Vulg. non enim habemus hic.
• 0563B Psal. XXIV; Joan. XV odio habuerunt me gratis — agrees with Joan. 15:25; the Psalm reference (Ps. 24 for the more usual 34:19 / 68:5) is Migne's and is left as printed per the standing rule on references.
• 0564C I Joan. II Ex nobis exierunt … Si enim essent ex nobis — Vulg. prodierunt … nam si fuissent. Tense/lexis only.
• 0564D Act. XIII repulistis … judicastis (perfect) for the Vulgate's present repellitis … iudicatis.
• 0565A Matth. XXVIII Ite, docete omnes gentes — Vulg. euntes ergo docete.
• 0565B Luc. XXIII; Joan. XIX Crucifige illum — Vulg. crucifige crucifige eum / crucifige crucifige. Pronoun and repetition only.
• 0565B Jer. XII dedit contra me vocem suam — suam is not in Jer. 12:8. Not marked: it is the word the exposition needs and the sense is unchanged.
• 0567D Gen. II Plantaverat autem Dominus — Vulg. Dominus Deus.
• 0567D Gen. II noli comedere — Vulg. ne comedas.
• 0568B Ezech. II inter scorpiones habitas — Vulg. cum scorpionibus habitas.
• 0568D Ephes. V Fuistis aliquando tenebrae — Vulg. eratis enim aliquando.
• 0572A Luc. II pertransibit gladius — Vulg. pertransiet.
• 0573A Psal. CXLVII Tu ergo Hierusalem, lauda Dominum — Vulg. Ps. 147:12 lauda Hierusalem Dominum. Migne supplies the pronoun and the connective; the sense of the verse is untouched, and Philip's id est Ecclesia gentium attaches to the vocative either way.
• Throughout: the lemma prints quod fusca sum (indicative) where the Vulgate has quod fusca sim (subjunctive). No English exponent; not marked.
• 0567D Gen. II custodiret illud for the Vulgate's illum (paradisum is masculine). Pattern 9 gender mismatch — English "it" either way, no marker.

3. Defect markers fired

• [sic: incrudelitatem] — 0566A (quanto magis illum a se per incrudelitatem ejecerunt). incrudelitas is a hapax in this corpus (grepped: one occurrence, this one) and is not an attested medieval form as far as can be shown here. Its prefix is privative, so any rendering of it as "cruelty" would be the silent repair Pattern 7a forbids — translating a word as something other than what it means in order to rescue the sentence. Conjecture: crudelitatem, which the parallelism (quanto magis … tanto magis) and the sense both want, and which is one minim-cluster away. Carried untranslated in italics per Pattern 7's non-word class.
• [sic: neam] — 0571B (vineam neam non custodivi). Plainly for meam. The printed lemma appears twice within eleven lines, correctly as vineam meam at 0571C, which is itself the proof that the first is compositor's type and not the author's. Only the broken word is inside the marker; vineam is a real word and is translated ("the vineyard"), so the English reads …in the vineyards, the vineyard [sic: neam] I have not kept. — deliberately odd, because the plate is.

No [d: …] (no dittography found) and no [ed: …] (nothing lost in the digitization) in this range.

4. Guillemet defects in the plate — reproduced, not repaired (Pattern 6 / Pattern 5 corollary)

Three unbalanced runs. All are Migne's, all are mirrored exactly, and each is why this range's guillemet counts are odd in the verifier's output.

1. 0565A (chunk 0025) — an opening « never closed. « Ite, docete omnes gentes [n: (Matth. XXVIII)]. Sed tunc vineam suam non custodierunt… runs on into Philip's own narrative and to the end of the chunk without a closing mark. Latin «11 »10; English «11 »10. Per the Pattern 5 corollary the missing » is not supplied.
2. 0568D (chunk 0028) — a closing » with nothing to close. « Fuistis aliquando tenebrae, » id est peccatores; « nunc autem lux, » id est justi, in Domino [n: (Ephes. V)]. » — two openings, three closings. The stray third mark is reproduced where it stands. Latin «12 »13; English «12 »13.
3. 0572D (chunk 0029) — a closing » with no opening. Exsulta et lauda, habitatio Sion [n: (Isa. XII)], » id est Ecclesia de gentibus congregata… The Isaias quotation is printed unopened. Latin «3 »4; English «3 »4.

Also worth recording, because it looks like an error and is not:

• 0569A–0569C (chunk 0028): a guillemet run that swallows the commentary. Migne opens at « Tabernaculum fuit constructum de pellibus rubricatis… and does not close until …virga Aaron quae fronduit [n: (III Reg. VIII; Hebr. IX)]. » — some 270 words later. Everything between (the exposition of the red skins as martyrs, the hyacinth skins as confessors, the haircloths as sinners, the ark, the four Gospels, the two bars) is Philip's own prose, inside Migne's quotation marks. The run is reproduced 1:1; nothing is re-punctuated.

Pattern 5 instances (an attribution verb inside the quotation, translated in place and left inside): 0563C « Vide, inquit, ut tu facias… »; 0566B « Vinea autem Domini exercituum, sicut dicit Isaias, domus Israel est. »; 0571A « Ex ore tuo, inquit, justificaberis… ». None was broken around an English speech tag.

5. Other cruces

• 0566A, quia et ipsa … originem duxit. ipsa is feminine, and in the sentence Mater Christi, secundum carnem, Synagoga fuit, quia et ipsa per patriarchas et reges, et per beatam Mariam originem duxit the natural antecedent by sense is Christ (masculine) — but the whole recapitulation is spoken by the caro Christi, which is feminine and is the speaker of every verb around it. Rendered "because it too drew its origin", carrying the feminine referent as the flesh. No emendation; flagged here because a reader may expect "he".
• 0570C, sicut materia filiis suis sustentatur et adjuvatur. Rendered literally, "as matter is sustained and helped by its own sons." Conjecture: mater — the comparison is of the flesh governed by its five senses to a mother sustained by her sons, and materia makes the simile say nothing. materia is a real word and construes grammatically, so under the Pattern 12 boundary it takes no [sic: …]: the English word on the page is ours, and the conjecture lives here.
• 0570D, Dumque eum exterioribus delectationibus inficit … illam debilitat et occidit. The masculine eum has no clean antecedent (the nearest nouns are odoratus, masculine, which is the subject, and animam, feminine, which is illam). Rendered as printed — "while it infects him … it weakens and kills her" — with the shift of gender left visible. Possibly eam in the plate's copy-text; not emended.
• 0562C, aurum clarissimum, id est obrysum.* obrysum is an attested spelling of obryzum (refined gold; the corpus has obryzo* twice at 11703). Pattern 9: translated normally, "the brightest gold, that is, refined gold." No italics, no marker.
• 0563C, excelsum majus. Rendered "the greater high place", kept identical at its second occurrence (Quid vero est illud excelsum majus) because the second passage glosses the first.
• 0567A, Dicat ergo natura nostra; — the plate prints a semicolon where a colon is expected, before direct speech. Preserved (Pattern 8); the English reads "Let our nature then say; Through the first Adam…".
• 0565D (chunk 0026), Per Cedar ergo qui tenebrae interpretatur, tenebras mortis — the first limb has no finite verb; it borrows debemus intelligere from the second limb. Supplied in brackets, "[we ought to understand]", per the corpus's existing convention for supplied English words.
• 0569C (chunk 0028), quae per duos vectes figuraliter, spiritualiter intellexerit — two adverbs and one verb, and figuraliter has no verb of its own. Rendered "which [are set forth] figuratively by the two bars", with the supplied verb bracketed.
• Column sequence at the end of 0029: [0573A] → [0574A] → [0573]. Reproduced verbatim and in order. The jump to 0574A and the bare [0573] on the Explicit line are what the source carries; per the standing false-positive rule on column-band arithmetic, no correction is proposed.
• 0567 (chunk 0027) ends with two heads and no text: RECAPITULATIO A SUPERIORIBUS followed by SECUNDUM MORALITATEM., then the bare anchor [0568A]. This is the chunk boundary — the body of the section is chunk 0028, whose head repeats as SECUNDUM MORALITATEM. (cont.). Both heads are translated and both blank lines kept; not a defect.

6. Second person (Pattern 17)

Singular addressees take thou/thee/thy, plural you/your, decided by the Latin morphology alone:

• thou — Exod. 25:40 to Moses (ut tu facias … vidisti, 0563C); David to Nathan (Videsne, 0563D); Pilate (Si hunc dimittis, non es amicus Caesaris, 0565C); Ps. 120:6 (sol non uret te, 0567B) and Philip's gloss on it (tentatio nulla decipiet te); the reader addressed directly at 0568C–D (Sed et tu quicunque … scito te habitare); Matt. 12:37 (Ex ore tuo, 0571A); Simeon to Mary (tuam ipsius animam, 0572A) and the quasi diceret gloss expanding it; the Church of the Gentiles addressed from 0572D to 0574A (custodiris … Tu es enim … tibi … hostes tuos), including Ps. 147:12 inside it.
• you — the daughters of Jerusalem throughout (Nolite me considerare, O vos filiae Jerusalem); the serpent to Adam and Eve (Gen. 3:5, comederitis … oculi vestri … eritis, 0566D); Paul to the Ephesians (0568D) and to the Jews at Antioch (Act. 13:46, 0564D–0565A); Act. 20:29 (in vos … parcent vobis); the Jews addressed by Mary at 0572C (o vos Judaei … quem vos interfecistis); Luc. 13:35 (relinquetur vobis domus vestra).

No case in the range required a judgment call: every second person is morphologically unambiguous.

7. Marker counts as shipped

verify-english.mjs 11638 reports no error and no warning against any of these six.

Agent 6 of 6 (the end of the work: 0034 is the final chunk). PL 203, coll. 0573B–0584D. This file is one of six to be merged; it opens with the terminology decisions so the merge can reconcile them.

1. Terminology decisions made in this range (for the merge)

Second person throughout follows Pattern 17: tu/tuus → thou/thy; vos/vester → you/your. The two places where the number changes inside a page are deliberate and are the rule's payload:

• 0031, John 14:9–10 — Migne prints singular non credis where the Vulgate has plural Non creditis; the English reads "thou believest not." ([var:] fired.)
• 0033, John 8:44 — Vos ex patre diabolo estis is plural and reads "You are…", inside a chunk otherwise addressing one man ("Render therefore, O man… thou canst answer").

2. Markers fired

[d:], [ed:], [nt:]: none fired. No dittography in the range; nothing in the digitization is lost against what the plate carries; every [n:] in the range is a bare citation, none is editorial prose.

3. ⛔ The Vulgate divergence list — the deliverable

Every printed quotation in 0030–0034 was read word by word against the Vulgate. 17 [var:] markers fired; the divergences that were logged but left unmarked are listed after.

Marked [var:] — material divergences

0030 (2)

1. 0573C, Coloss. II, 9 — Migne: in ipso omnis plenitudo divinitatis inhabitat. Vulg.: in ipso inhabitat omnis plenitudo divinitatis corporaliter. The word order is inverted and corporaliter is absent. Marked because the missing word is the one Col 2:9 is normally adduced for.
2. 0573D, Joan. I, 14 — Migne: Vidimus eum quasi unigenitum a Patre. Vulg.: et vidimus gloriam ejus, gloriam quasi unigeniti a Patre. Migne's object is Christ himself, not his glory, and unigenitum is accusative to match. Philip's argument (the topaz "shines with every colour," i.e. Christ is adorned with every virtue) rests on seeing him, so the divergence is load-bearing.
3. 0575A, Psal. VIII, 6 — Migne adds a vocative Domine the Vulgate does not print. Marked because the added vocative turns a description into an address.

0031 (6)

1. 0575A–B, the lemma itself — the chapter head and the first re-quotation print Cant. 1:6, ne vagari incipiam post greges sodalium tuorum; the second re-quotation, twenty words later, prints Ne forte vagari incipiam per greges sodalium tuorum. ⚠ This is the range's most important reading: the author re-quotes his own lemma with a different preposition and an added forte, and then expounds vagari. Both forms are rendered as printed and neither has been conformed to the other.
2. 0575B, Matth. XVI, 13–14 — Migne: Quem dicunt homines esse hominis Filium? Responderunt discipuli: Alii Eliam, alii Jeremiam…. Vulg.: …esse Filium hominis? At illi dixerunt: Alii Joannem Baptistam, alii autem Eliam, alii vero Jeremiam…. John the Baptist is dropped from the list of mistaken identifications — and the list is precisely what the topaz/chrysolite exposition is counting.
3. 0576A, Joan. XVII, 21 — Migne: Pater, volo ut sicut tu es in me…. Vulg.: ut omnes unum sint, sicut tu, Pater, in me, et ego in te…. Recast as a first-person volition.
4. 0576A, Joan. XIV, 9–10 — Migne: videt et Patrem meum, non credis quia ego in Patre et Pater est in me [full stop, not a question]. Vulg.: videt et Patrem… Non creditis quia ego in Patre, et Pater in me est? Three divergences at once: meum added, plural → singular, question → statement. Punctuation follows the plate (Pattern 8): the English is a statement.
5. 0576B, Isa. XVIII, 4 — Migne: Quiescam in loco meo, et considerabo quasi meridiana lux. Vulg.: Quiescam, et considerabo in loco meo, sicut meridiana lux clara est. Migne's word order attaches in loco meo to quiescam, and the next four sentences of commentary interrogate exactly that phrase (Quid est ergo quod dicit in loco meo quiescam?). Conforming to the Vulgate here would have destroyed the exposition.
6. 0576B, Isa. XXVIII, 12 — Migne: Haec est requies mea, refice lapsum. Vulg.: Haec est requies mea… reficite lassum. Singular imperative for plural, and lapsum "the fallen" for lassum "the weary." The very next sentence glosses peccatorem lapsum … exhortamur — the commentary is built on lapsum, and the familiar "refresh the weary" would have left it unsupported. This is the range's clearest 7a″ trap.
7. 0576C, Isa. LXVI, 2 — Migne: Super quem requiescam, nisi super humilem et quietum, et trementem verba mea? Vulg.: ad quem autem respiciam, nisi ad pauperculum et contritum spiritu, et trementem sermones meos? An Old-Latin/LXX-shaped form (LXX ἐπὶ τὸν ταπεινὸν καὶ ἡσύχιον). It is adduced under the heading Quae est ergo requies Dei — the divergent verb requiescam is the whole reason the verse is here.
8. 0576D–0577A, II Cor. V, 10 — Migne: ut recipiat unusquisque prout gessit. Vulg.: ut referat unusquisque propria corporis, prout gessit.

0032 (5)

1. 0578D, Isa. IX, 2 — Migne: Populus gentium qui ambulabat in tenebris. Vulg. has no gentium; the commentary immediately reads it as populis gentium.
2. 0579A, Ephes. III, 17 — Migne: in cordibus nostris. Vulg.: in cordibus vestris. A person change, and the gloss that follows continues in the first person plural.
3. 0579C, Ezech. XVIII, 23 — Migne: Nolo mortem peccatoris. Vulg.: nolo mortem impii; and the whole passage is about the peccator.
4. 0579D, "Ezech." (ibid.) — Peccator quacunque hora ingemuerit, salvus erit is not in the Vulgate at all, at Ezekiel or elsewhere; it is the medieval adage drawn from Ezech. XXXIII, 12. Marked, because a reader who checks the citation will find nothing there.
5. 0580A, Heb. VI, 7 — Migne: Terra bibens desuper venientem imbrem, et fructificans opportunam, ab his, a quibus colitur, accipit benedictionem. Vulg.: Terra enim saepe venientem super se bibens imbrem, et generans herbam opportunam illis a quibus colitur, accipit benedictionem a Deo. Note opportunam is left with no noun to agree with — rendered "what is seasonable," not silently supplied with "herb."
6. 0580C, I Joan. II, 15 — Migne prefixes Filioli (from I Joan. II, 12/18) and prints si quis dilexerit mundum for Vulg. si quis diligit mundum.

0033 (5)

1. 0580D, Gen. III, 19 — Migne: « Terra es, et in terram ibis. » Vulg.: pulvis es, et in pulverem reverteris. ⚠ The single most important divergence in the range. The entire second half of the seventh book is an exposition of terra — "we ourselves are earth," the earth given to the serpent for food, the earth that yields her fruit. On the received Vulgate text the argument has no proof-text at all. A conformed "dust thou art" would have been smooth, correctly cited, verifier-clean, and would have destroyed the chapter.
2. 0581A, Eccle. VII, 30 — Migne: Deus fecit hominem simplicem et rectum. Vulg.: quod fecerit Deus hominem rectum. The gloss immediately splits the two words (simplicem, ut malum nunquam faceret, rectum autem, ut…), so simplicem cannot be dropped.
3. 0581C, Joan. XVII, 3 — Migne: ut cognoscant te, et quem misisti Filium tuum unum et verum Deum esse. Vulg.: ut cognoscant te solum Deum verum, et quem misisti Jesum Christum. A doctrinally loaded recasting: Migne's form predicates "the one and true God" of the Son. Rendered as printed.
4. 0582A–B, Psal. CI, 26 — Migne: Et tu in principio, Domine, terram fundasti. Vulg.: Initio tu, Domine, terram fundasti. The next sentence turns on it: Filium autem Dei Propheta hic principium appellat.
5. 0582C, Isa. XXVIII, 21 — Migne: Opus ejus alienum fuit ab eo, ut faceret opus suum, prius fecit opus alienum. Vulg.: ut faciat opus suum, alienum opus ejus: ut operetur opus suum, peregrinum est opus ejus ab eo. Philip himself calls the words valde obscura and says they cannot easily be found in the other prophets — which is a fair description of a text that is not the Vulgate's.

0034 (1)

1. 0584B–C, Psal. XLIV, 11 — Migne: Intellige, filia, et vide, et inclina aurem tuam. Vulg.: Audi, filia. Woven, unguilleted, into the Father's speech; rendered "Understand, O daughter."

Logged, NOT marked (immaterial or word-order only)

• 0573C Job XXVI — the quotation itself agrees with Vulg. Job 28:19; the note points to Job XXVI. Reference crux only, not corrected (refDisplay rule).
• 0574D Psal. VIII, 6 Minuisti eum paulo minus ab angelis — agrees.
• 0575C Matth. XVI, 16; 0576B Psal. IV, 9; 0576C Psal. LXXV, 3 (In pace factus est for factus est in pace, order only); 0577C Psal. XXXII, 5; 0577D Psal. LXXXVIII, 2; 0579A Psal. CIII, 18, I Cor. X, 4; 0579D Deut. XXXII, 2 (adds et); 0580A Isa. XXVI, 19, Psal. XLV, 5, Psal. XXXV, 9; 0581B Joan. VIII, 44 (perficere vultis / homicida ab initio erat, order); 0581B Psal. LXXXIV, 13 (drops Etenim); 0581D Psal. XXX, 6; 0582A Joan. XIV, 2 (multae mansiones for mansiones multae); 0582A Eccle. XVIII, 1; 0582B Psal. CIII, 24; 0582C Gen. I, 1 — all agree or differ only in word order or a connective.
• 0577A Psal. LXI, 12 — Migne Duo haec quae audivi for Vulg. duo haec audivi; the added relative leaves the clause without a main verb, which is why the commentary answers it with Quae sunt? Immaterial to the sense, logged only.
• 0577B Joan. V, 22 — Pater non judicat quemquam for Vulg. Neque enim Pater judicat quemquam; same sense.
• 0577B Eccle. V, 4 — Deus altissimus patiens redditor est for Vulg. Altissimus enim est patiens redditor; order only, and the commentary glosses both words.
• 0576D « Per obedientiam pervenitur usque ad occultorum Dei notitiam. » — introduced as divina Scriptura but it is not scripture; it is a Gregorian sententia. Guillemets reproduced as printed; no [var:], because there is no received text to compare against.
• 0577D [n: (I Cor. II)] attached to neminem vult perire, sed omnes ad agnitionem sui vult venire — the sentence is I Tim. II, 4, not I Cor. II. Not marked (it is unguilleted, the author's own prose); reference crux.
• 0578B [n: (Psal. X)] attached to Qui aeternos montes de superioribus rigat suis… — the allusion is Psal. CIII, 13. Reference crux; unguilleted prose, no marker.

4. Textual cruces — printed readings rendered literally

1. 0575D (0031), quamvis. Tu me deprecaris quamvis remota omni ambiguitate te certificem ubi… — quamvis ("although") will not construe with the ut-sense the sentence needs; ut is the obvious conjecture. Rendered "that, although with all ambiguity removed I should assure thee where…", i.e. as printed. No emendation.
2. 0575D (0031), pascor for pascam. Christ's answer prints the passive in meridie pascor et accubo where the lemma, the question, and the parallel answer in 0032 (in meridie pasco et accubo, col. 0578A–B) are all active. Rendered "at midday I am fed and lie down" in 0031 and "at midday I feed and lie down" in 0032 — the two are deliberately different because the plate is. Do not harmonize them at merge.
3. 0576B (0031), aurae. quare se comparat meridianae luci plus quam aurae vel alteri horae diei? — in a list of hours of the day, aurorae ("dawn") is the near-certain intention; aura ("breeze") is a real word and is rendered as printed ("the breeze"), per Pattern 7 and the requiescere faciam boundary. No marker (a real word merely wrong is not [sic:] matter).
4. 0578A (0032), Respondeat ergo justum est Christus. An unconstruable run: an impersonal justum est dropped inside the jussive. Rendered "Let Christ therefore answer — it is just — to holy Church"; conjecture Respondeat ergo (justum est) Christus or Respondeat ergo ut justum est.
5. 0578D–0579A (0032), et sicut coeperunt. …in cordibus eorum est plantata et radicata, et sicut coeperunt in se Christum habere habitatorem — sic is expected. Rendered "and as they began to have Christ as an indweller."
6. 0575C (0031), eam for cultum. cultum solius divinitatis ab omnibus sanctis removeat, et eam soli Deo … attribuat — feminine pronoun for a masculine antecedent (adorationem understood). Pattern 9: gender has no English exponent, "attributes it"; no marker, no churn at merge.
7. 0581A (0033), postquam for per quam. rectitudinem mentis quam a Deo acceperat, et postquam ipsi soli placere debuerat … intorsit — per quam is the sense required. Rendered "and after which he ought to have pleased him alone," as printed. The whole period is anacoluthic (the quia clause never reaches its verb before juxta Dei dispensationem actum est takes over) and is rendered as printed, including the shift to possit at the end.
8. 0583C (0033), insensibile spiraculum vitae. Rendered "the insensible breath of life." Odd of Gen. 2:7; possibly insensibiliter or insensibili. Rendered as printed, no marker.
9. 0583D–0584A (0033), redemit without an object. Filius ejus Dominus Jesus Christus sanguine suo redemit — the object is supplied in square brackets, "redeemed [thee]," on the ~95-instance corpus precedent for supplied English words.
10. 0584D (0034), refertur. Daniel propheta … memoria abundantiae suavitatis Spiritus sancti refertur … Christum montem interpretatur — refertur cannot be "is reported" here, because the sentence's finite verb is interpretatur; read as refertus (est) / refercitur, "is filled." Rendered "is filled with the memory of the abundance of the sweetness of the Holy Spirit." Conjecture logged, no emendation and no marker (a real word merely wrong).

5. ⚠ The unclosed guillemet at 0579D (0032) — mirrored, not repaired

Migne prints:

Moyses namque dicit: « Concrescat ut pluvia doctrina mea, et fluat ut ros eloquium meum [n: (Deut. XXXII)]. Isaias quoque ait ad Dominum: « Quia ros lucis, ros tuus [n: (Isa. XXVI)]. »

The Moses quotation is opened and never closed, and the sentence attributing the next quotation to Isaiah is swallowed inside it. Per Pattern 5's corollary the missing » is not supplied: the English opens where Migne opens and closes where Migne closes, so chunk 0032 carries 15 « against 14 » in both twins, exactly as the plate does. verify-english.mjs check 7 passes on this because it counts against the Latin twin, not against balance. Do not "fix" this at merge.

6. (cont.) heads

0031, 0033 and 0034 carry (cont.) on a repeated ## head. It is our chunk-boundary artifact, left untranslated and unmarked; the builder merges these. The head of 0031 also prints a stray space before the closing italic asterisk (sodalium tuorum *.) — reproduced.

7. Verification

node scripts/verify-english.mjs 11638 passes on 0030–0034 (only 0016, 0017, 0023 were still missing at the time of writing, from other agents' ranges). Per-chunk marker parity against the Latin twin, body only:

The italic marginal page numbers of Migne's own setting (334 … 339) are reproduced verbatim in place.