Epistola ad Philippenses

Translated 2026-08-18, one translator, 4 chunks (3,220 Latin words → 4,817 English, 1.50×; per chunk 1.49 / 1.47 / 1.49 / 1.52). verify-english.mjs clean.

Apparatus: 7 [var:] · 1 [sic:] · 0 [cj:] · 0 [d:] · 0 [ed:] · 0 [nt:] · 8 [n:] (exact parity with the Latin twin: 1 + 2 + 2 + 3). 32 column anchors (9 + 6 + 10 + 7), 5 heads, 0 guillemets, 5 question marks EN vs 5 LA, 0 em-dashes.

Conventions are the Ruth pilot's (src/english/8968/cruces.md §1) and the NT agent brief's, and are not re-opened: VERS. n.-- passes verbatim, the lemma is Englished, a fragmentary lemma is rendered as the phrase it stands for, usque ad → as far as with no verb supplied to complete the sentence it breaks, the attribution sigla ride as ordinary [n: …] with contents untranslated. CAPUT PRIMUM. → CHAPTER ONE., CAPUT II./III./IV. → CHAPTER II./III./IV., matching 8982 and 8988 as shipped.

§0 · PLATE STATUS — foot-of-page apparatus is a CHECKED ZERO

Read this session by the orchestrator, merged here. Method: archive.org patrologiaecurs06saingoog, leaf n = (column + 9) / 2, verified against the running head (n307 = GLOSSA ORDINARIA. — EPIST. AD PHILIPP.). Leaves n305, n306, n307, n308 read at full page width across the whole foot, plus n309, where this work ends and Colossians begins mid-page below the rule — the case that lost three notes on PL 202, and clean here.

No foot rule, no numbered notes: the columns simply end. notes_recovered: 0, tier 3, coverage full. Consequence: this work takes no [cj:], and had it taken one it would have been ours, not a conjecture Migne had already answered at the foot of his own page.

Plate corroborations passed to me, used but not substituted for my own collation:

• n306 (cols 603/604), VERS. 7 — Factus. Secundum formam servi, qui erat infectus secundum formam Dei et etiam suo. Habitu. Id est conversatione. ⚑ The plate prints a FULL STOP after suo; our TEI twin runs it on as et etiam suo Habitu. The English translates the committed twin (which is what a reader meets in the facing column), so it reads "…and also by his own Habit. That is, by his manner of life." On the plate the gloss ends at suo with its noun unexpressed, and Habitu. opens a fresh lemma — which is the better sense. Recorded, not patched: this is punctuation, not lost words, and Pattern 8 governs the twin as it stands. A future TEI patch here would be defensible.
• n306, VERS. 8 — Humiliavit, etc. (CHRYS.) Sponte Filius Patri obedivit, non quasi in servilem, etc., usque ad quod omnium maxime Patrem honoravit. Confirms (a) the (CHRYS.) siglum printed inside the gloss run, (b) non quasi in servilem carrying its negation, (c) the usque ad leaving the sentence unfinished. No verb was supplied.
• 0602D VERS. 29. — Non solum, etc. Utrumque ergo ad Dei gratiam pertinet ✓
• 0605–0606 VERS. 19. — Quorum deus venter est. ✓
• 0607–0608 VERS. 20. — Nostra conversatio and Impleat omne desiderium vestrum. Non in terrenis, sed in gloria. ✓ — all as the Latin twin has them.

§1 · THE LEMMA DIVERGENCE LIST — ALL 136 ITALIC SPANS CHECKED (7a″)

The dispatched list carries 136 spans; the Latin prints 140 italic runs (34 + 29 + 42 + 35). The four the list does not carry are usque ad ×4 (0603D, 0603D, 0605C, 0605C) — Migne's abridgment formula, not lemmata. Recorded so the next reader knows the gap was found and accounted for, not inherited.

Every span was collated word by word against the Clementine Vulgate. Agreements are listed too, so this cannot be told apart from a report by an agent that never looked. Where the span is a fragment (Migne's usual practice — an incipit of two or three words, often closing on a comma or etc.) the verdict "agrees (fragment)" means: the printed words are the Vulgate's, in the Vulgate's order, truncated.

Chapter I — chunk 0 (34 spans)

Chapter II — chunk 1 (29 spans) · ⚑ THE CARMEN CHRISTI IS §3, BELOW

(Two of the twenty-nine runs in this chunk are the usque ad formula and are not lemmata; they are counted in the 140 and excluded from the 136.)

Chapter III — chunk 2 (42 spans)

(Two of the forty-two runs are usque ad.)

Chapter IV — chunk 3 (35 spans)

§2 · THE DIVERGENCES DECLINED, each with its reason

Twelve spans diverge from the Clementine and took no [var:]. A blind reader will look here, so each is argued rather than listed.

1. 0601C sicut justum est sentire, — Vulg. 7 sicut est mihi justum hoc sentire. The glossator has just quoted the same clause's opening as its own lemma (Sicut est mihi) and is now re-quoting the remainder inside his own sentence, dropping the pronoun he has already used. Word order and one dropped demonstrative, no sense at stake. Pattern 14's bar is a divergence that changes the sense.
2. 0602A Pro vobis gratias ago: Confido. Vos autem volo scire, — not a citation at all but the glossator's précis of the epistle's movement so far, stitching vv. 3, 6 and 12. Marking a précis as a variant would misdescribe what is on the page. ⚑ But note what it proves: his own précis prints Vos autem volo scire, WITH the pronoun that his lemma at the head of the same paragraph drops. See §4.
3. 0602C Christus mihi vivere, — the glossator's re-quotation of v. 21, reordered so that it can sit inside quia …. Same class as (1).
4. 0603B Superiores, — agrees; the note is about the ENGLISH. The Vulgate clause is superiores sibi invicem arbitrantes and the fragment-expansion convention would give "esteeming others better than themselves." It is rendered "Superior," because the gloss beneath it expounds the WORD, twice (quo superior sit, quo videmur superiores esse), and an expansion into the familiar clause would leave the gloss with no hook — the Song-of-Songs ruling's case 2 (stylus/epistylium, murenulae/ferculum): where the gloss is about the word, the word stands.
5. 0604B Deus enim. — Vulg. 13 Deus est enim. A two-word incipit that has lost the copula. Nothing in the gloss (an Augustinian excerpt on perseverance) turns on it, and the English "For it is God." supplies no word the Latin lacks: it renders Deus enim as the clause it opens.
6. 0604D Qui Spiritu Dei, vel, Deo, — Vulg. 3 qui spiritu servimus Deo. This is a real divergence and it took no marker because Migne prints the received reading himself, in the same breath, as his own alternative (vel, Deo,). The [cj:] decline-class "anything the edition already glosses" is written for conjectures, but its argument is general: a reading the reader already meets does not need ours beside it.** Logged here so the decline is visible; a later editor who disagrees will find it named.
7. 0605A Quanquam et ego, — Vulg. 4 quamquam ego habeam confidentiam et in carne. The et has moved from "also in the flesh" to "I also." A real shift of emphasis — but the gloss (Ne videatur Apostolus contemnere in aliis quod non habet) rests on neither position, and Migne re-quotes the lemma two lines later as the bare Quanquam. Declined as positional; named because it is the closest call in this section.
8. 0607B modestia nota sit omnibus — vestra dropped and the quotation cut before hominibus so that the glossator can finish it his own way (fidelibus et infidelibus). The truncation is the gloss's method, not a variant text.
9. 0607D superat omnem sensum: — the simplex for the Vulgate's compound exsuperat, in the glossator's own second, alternative construal. See §4 for how the English keeps the two verbs apart without importing the familiar English of either.
10. 0607D quaecunque sunt amabilia, — sunt supplied, which the Vulgate carries only on the first member of the eight-fold series. A compositor's regularization at worst; no sense at stake.
11. 0608C Qui recipit justum vel prophetam, and in nomine justi vel prophetae. — Matt. X, 41 is a doubled saying (prophet … prophet, just man … just man) and the glossator explicitly collapses it into one in order to make his point about datum versus fructus: non ait simpliciter … sed addidit. He is analysing the verse, not quoting it. A [var:] here would report as a textual variant what is an act of exegesis.
12. 0608D In odorem. — Vulg. 18 odorem suavitatis, in apposition, without in. The preposition changes nothing the gloss uses (placent Deo ut odor suavissimus).

Also declined, and it is not a lemma: the Augustine excerpt at 0603C–D opens Exinanivit autem se formam servi accipiens against the Vulgate's sed semetipsum exinanivit, formam servi accipiens. Pattern 14 is for scripture quoted as scripture; this is Augustine's own sentence in the Enchiridion, reported by Migne as an excerpt. Rendered as printed, no marker.

§3 · ⭐ THE CARMEN CHRISTI (Phil. II, 5–11) — THE FULL COLLATION

This was the flagged high-risk ground, and the result is a clean negative, which is worth stating exactly because a negative that nobody can distinguish from an unread passage is worthless. Every word of every lemma in 0603B–0604B was collated against the Clementine letter by letter.

Zero divergences. [var:] count for the hymn: 0. Migne's Philippians lemmata for the Carmen Christi are the Clementine text, truncated. (Contrast chapters I, III and IV, which between them produced all seven markers — so the zero here is not the collation failing to bite.)

The two places where the hymn's ENGLISH could have gone wrong, and what was done

(a) Non rapinam. — where the negation attaches. non governs rapinam, not arbitratus est: he did not judge it robbery. The English is "He thought it not robbery." — negative sitting on "robbery," which is where Migne prints it. The fragment is expanded to its clause (convention 3) and the expansion coincides with the traditional English because the printed Latin coincides with the Vulgate, which the register expressly permits; it was not reached for.

⚠ The gloss then offers a second construal in which the negation moves: Vel, non est arbitratus se aequalem Deo esse secundum humanitatem — here non is on the verb ("he did not judge himself equal to God according to his humanity"). Migne prints two different negation placements two lines apart, and the English carries each where he puts it. That contrast is the single most easily flattened thing in the chunk.

(b) Factus. … qui erat infectus … — infectus is in-factus, "unmade", played against the lemma factus. It is a privative and it is carried: "he who was unmade according to the form of God." Rendering it "uncreated" would be the theologically expected word and would lose the play on the lemma; rendering it "tainted" (the other infectus) would be a different word altogether. Plate-confirmed at n306. See also §0 on the plate's full stop after suo.

§4 · THE CRUCES PROPER

⭐⭐⭐ 0607A (VERS. 1) · Gaudium, et consolatio — THE LEMMA IS DEFECTIVE AND ITS OWN GLOSS PROVES IT

Migne prints VERS. 1.-- Gaudium, et consolatio. Id est per quos in praesenti laetificor et in futuro coronabor. Discipulis enim in agone victoribus, dignus erit corona magister.

Phil. IV, 1 reads gaudium meum, et corona mea. Migne's lemma has consolatio where the Vulgate has corona — and the two sentences of gloss beneath it expound corona and nothing else: the master will be crowned, the master is worthy of the crown. The lemma word has no purchase on its own gloss.

Rendered as printed — "My joy, and my consolation." — with [var: Vulg. gaudium meum, et corona mea, "my joy and my crown"]. This is the single most valuable finding in the work and it is exactly the shape 7a″ describes in reverse: not a translator conforming a lemma to the familiar English, but a plate whose lemma has drifted off the verse its own gloss is written for.

⚠ No [cj:]. The bar is that the printed word makes the sentence assert something the author argues against, and consolatio asserts nothing false — the gloss simply answers a word that is not there. Pattern 14 is the right home, and Pattern 18's own boundary rule settles it: [var:] inside a quoted scripture, [cj:] in the author's own prose.

⭐⭐ 0602A (VERS. 12) · Scire autem volo — A DROPPED vos, CONTRADICTED SIX WORDS LATER

Vulg. I, 12 is Scire autem vos volo fratres. Without vos the lemma says "I wish to know," not "I wish you to know" — the subject of the infinitive changes, which is as material as a divergence gets in three words.

And the same paragraph carries the correction. The glossator's own précis two lines below prints Vos autem volo scire, with the pronoun. So the plate contains both readings, and the lemma is the defective one.

Rendered as printed ("But I wish to know, etc.") with [var: Vulg. Scire autem vos volo, "I wish you to know"]. The précis is rendered with its own pronoun ("But I wish you to know,"), because that is what Migne prints there. ⚠ The two English readings differ on the page and that is not an inconsistency — it is the plate.

⭐⭐ 0602B (VERS. 17) · Non puro animo for non sincere

Vulg. I, 17 is Christum annuntiant non sincere. Migne's lemma is Non puro animo, "not with a pure mind" — a different phrase, not a different form of the same word, and quite possibly an Old Latin reading. The gloss is built on it (Et ne viderentur correcti, addit). Rendered as printed, [var: Vulg. non sincere, "not sincerely"].

⭐⭐ 0607A (VERS. 20) · etiam Dominum exspectamus for Salvatorem … Dominum nostrum Jesum Christum

Vulg. III, 20 is unde etiam Salvatorem exspectamus Dominum nostrum Jesum Christum. Migne's lemma keeps etiam … exspectamus and replaces the whole double object with a bare Dominum. Rendered as printed, [var:] carrying the Vulgate in full. ⚑ Note this is the one divergence where the gloss is indifferent (praeter commoda quae jam habemus), so it would have been the easiest of the seven to leave unmarked.

0601D (VERS. 10) · sine offensione for sine offensa

A different noun of the same root, in the same slot, with the same sense. Marked because it is a lexical difference rather than a matter of order, and because Pattern 14's own worked instance (10083) turns on precisely such a substitution; the sense is not at stake and the crux says so, so a later editor who wants to withdraw this marker has the argument here.

0602C (VERS. 21) · Mihi autem vivere Christus est for Mihi enim

enim → autem: an explanatory connective becomes an adversative one, at the head of the most-quoted sentence in the chapter. Marked. The gloss (Quasi diceret: Merito Christum magnifico sive vita, sive morte) reads it as resumptive, which suits either particle.

0605B (VERS. 8) · praeeminentem for eminentem

Vulg. III, 8 propter eminentem scientiam. Migne prints the compound, and prints it after its noun, split across his gloss (propter scientiam quam de Jesu habemus, praeeminentem, caeteris scientiis). Rendered "pre-eminent," with the marker. The word order is the glossator's construction and took no separate note.

0606D (VERS. 19) · circumcicis — [sic:], WITH THE SENSE KEPT OUTSIDE

et gloriantur in pudendis circumcicis. There is no such Latin word. It is circumcisis with a c for an s — the whole corpus of Latin here prints 172 circumcis- forms and not one other circumcic-, which is Pattern 9's own test ("reserve carry-through for forms unique to this plate") answered by count rather than by feel.

English: they glory in the parts of shame, circumcised [sic: circumcicis], which is to be wise in earthly things.

⚑ 7a⁗ test run, all four questions. Strip the marker and the clause reads "they glory in the parts of shame, circumcised, which is to be wise in earthly things" — it keeps its verb (gloriantur), it carries no negation to lose, it keeps its conjunction, and it keeps its object (in pudendis). The marked run is an attributive adjective, and its sense stands in plain English beside it. Nothing is stranded.

✅ THE PLATE WAS READ AT THIS WORD, AND IT AGREES WITH THE TWIN. Raised here as an open question (is this Migne's defect or our transcription's?), it was answered the same session by the orchestrator: leaf n307 (cols 605/606, right column), read at full resolution, prints circumcicis — the second c unmistakably a c and not an s, matched against the letterforms of faciunt and sapere on the neighbouring lines. Our twin agrees with the plate. So the defect is Migne's own type, the [sic: circumcicis] stands, and there is no TEI patch to make.

⛑ The negative is worth recording because two of its siblings came back positive. The same batch found two places where the twin had silently departed from the plate — Esther 0739D (convivium for the plate's convivum) and Colossians 0611A (consilio for the plate's concilio). This was the third candidate of the three and the only one to come back clean. "Is this Migne's defect or our transcription's?" is now a live question in this corpus, not a formality, and the corpus-frequency evidence that made the plate read worth doing (172 circumcis- forms and no other circumcic-) is the way to raise it.

0605D · resurrectionis Christi Et in hac vita — A FULL STOP MIGNE DOES NOT PRINT

quae nostra resurrectio est virtus et efficacia resurrectionis Christi Et in hac vita habere societatem in patiendo. The capital Et announces a new sentence and no stop precedes it. Pattern 8 forbids supplying one, so the English runs on identically: "…the power and efficacy of the resurrection of Christ And in this life to have fellowship in suffering." The reading is odd on the page and it is odd on the plate's terms, not ours.

The clause it sits in is itself unconstruable as printed — quae nostra resurrectio est virtus et efficacia resurrectionis Christi ("which our resurrection is the power and efficacy of Christ's resurrection"), with quae and nostra resurrectio competing for the same slot. Rendered as closely as the words allow; no emendation, no marker — nothing here is a non-word, and the defect is syntactic, which Pattern 18a explicitly leaves declined.

0606B · Et si ita de me. quid me qui de aliis? — RENDERED LITERALLY, [cj:] DECLINED

Printed exactly so: a full stop mid-clause and an interrogative that will not construe. English: "And if it is thus of me. what of me who of others?"

• The stop is Migne's and stays (Pattern 8).
• The question mark is Migne's and stays; it is one of the five in the work.
• The likely intention is quid de aliis? ("what of others?"), with me qui an eye-skip from the preceding de me. Declined for [cj:] because Pattern 18's bar is an English that asserts something false, and this English asserts nothing at all — 18a's own ruling puts anomalies that merely fail to construe outside the class, and this is not a negation or a correlative. Recorded here, which is where 18a says an uncertain reading lives.

0607D · Exsuperat vs superat — TWO VERBS, TWO ENGLISH WORDS, NEITHER OF THEM THE FAMILIAR ONE

The lemma is Exsuperat omnem sensum (Vulgate); the glossator's alternative construal re-quotes it as the simplex superat omnem sensum. English has one obvious word for both, and the one obvious word ("passeth all understanding") is the Authorized Version's, which would conform the lemma AND erase the difference in a single stroke.

Taken: "It surpasseth all understanding." for the compound, "surmounts all understanding:" for the simplex. Neither is the received English phrase; the pair keeps Migne's two verbs distinguishable in the English as they are in the Latin. Declined: rendering both "surpasses" (the difference would vanish), and rendering the second "passeth" (the AV import, on the weaker of the two verbs).

0604D · de caetero pseudoapostolos, ne legalia ab eis recipiatis moneo — ANACOLUTHON, KEPT

pseudoapostolos is accusative with no verb to govern it: moneo takes the ne clause, not the noun. English: "for the rest, the false apostles, I admonish you not to receive the observances of the law from them" — the noun left hanging as it hangs in the Latin, with the comma Migne prints rather than a dash or a supplied preposition. No marker: a real word in a real case, merely unattached.

0606A · jam perfectus sum … viator. Perfectus enim erat viator — MIGNE'S OWN PERSON SHIFT

First person, then third, of the same man, one sentence apart. Reproduced. The same shift runs through 0601C (Et exponit causam → bonum vobis desidero) and 0605B (ut in futuro Christum habeant praemium: et hic sim in illo membrum). This gloss moves between reporting Paul and speaking as Paul without warning, and it does so at least four times. It is a feature of the book, not a defect list; no marker anywhere, and the English follows the person printed.

0602D–0603A (VERS. 30) · a qua spiritum malignum exclusi

Acts XVI, 16–23. excludo is "shut out, expel," not "shut in": "out of whom I shut out the malignant spirit." Recorded because the first draft read "shut up," which reverses it, and because 8982 @0616B glosses the same episode with a different verb (ejecerat) — a neighbouring book is a hint, never an authority.

0605A · latria and dulia stay Latin AND stay roman

Migne prints them in roman. The register would ordinarily italicize a technical term left in Latin — but italic parity with the twin is sacred and outranks the register's typography preference. They are carried as English-alphabet words in roman, unexplained, exactly as he prints them.

0605A (VERS. 5) · quando facta est separatio in servo Salomonis

Jeroboam, Solomon's servant (III Kings XI–XII). Migne does not name him and neither does the English: "when the separation was made under the servant of Solomon." in + ablative of attendant circumstance; "in the servant of Solomon" is not English and "in the time of Jeroboam" would supply a name the plate withholds.

§5 · THE UNEXPRESSED SUBJECT — EVERY SUPPLIED PRONOUN, NAMED

The dominant defect class. Each supplied English pronoun that the Latin does not print, with what fixes it. Christ, the Father, Paul, Timothy, Epaphroditus, the Philippians, the hirelings and the false apostles all move through these paragraphs.

§6 · NEGATION AUDIT — every particle, and the word it attaches to

Test 2 says the question is not "are all the negatives present" but "is each negative on the word it attaches to." Counted mechanically over the four Latin chunk bodies, the work prints 115 negation-bearing tokens: non 75 (12 + 23 + 20 + 20) · ne 13 · nec 7 · nisi 7 · nondum 3 · nihil 2 · nulla/nullo/nulli/nullum 1 each · nescio 1 · nequit 1 · nunquid 1 · nolite 1. neque, haud, nunquam and nullus do not occur. All 115 are in the English; none is in the English that Migne does not print. The ones where placement, not presence, was the risk:

⚑ No sentence was rescued. The two places where carrying the printed word leaves awkward English are 0605D (Christi Et) and 0606B (quid me qui de aliis?), and both are logged in §4 rather than smoothed.

§7 · MIGNE'S OWN PRINTING — reproduced, not tidied

• Verse addresses. VERS. 1, 3, 5, 6, 7, 8, 10, 11, 12, 13, 14, 17, 18, 21, 22, 24, 25, 28, 29, 30 (ch. I) · 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 14, 15, 20, 21 (ch. II) · 1, 3, 4, 5, 7, 8, 9, 10, 11, 13, 14, 15, 18, 19, 20 (ch. III) · 1, 2, 3, 4, 5, 6, 7, 8, 10, 11, 12, 17, 18, 19 (ch. IV). Every chapter's sequence is monotonic — no out-of-order address in this work, and none was reordered. Ch. II jumps 11 → 13 (v. 12 unglossed); ch. III jumps 11 → 13 and 15 → 18; ch. IV jumps 12 → 17. All Migne's, all skips rather than inversions.
• VERS. 1, 2.-- at the head of chapter II is a double address for one gloss. Reproduced verbatim, comma and all.
• The four [n:] sigla — (Aug., Ench. c. 35 et 36), (CHRYS.) ×2, (AUG.) ×2 — plus the four scripture locators (Psal. LXVII), (Joan. XIV), (Luc. XI). All eight ride verbatim, contents untranslated. None is editorial prose, so no [nt:]: the longest, (Aug., Ench. c. 35 et 36), is a book-and-chapter locator, which Pattern 15 keeps as [n:] however long. All three scripture locators were checked and all three are right — a clean sweep that 2 Timothy did not get.
• (AUG.) at 0608A sits INSIDE the italic run (Bonae famae. [n: (AUG.)]), as (Psal. LXVII), (Joan. XIV) and (Luc. XI) do inside theirs. The italics are reproduced around the marker, not moved off it.
• Lower-case deus at 0606D (Quorum deus venter est, ut deus sit venter) against capital Deo elsewhere in the same paragraph. His typography, and it is pointed: the belly as god gets the small letter. English follows with "God"/"a god" as the syntax requires and the crux records the plate's practice.
• Quasi: / Quasi dicat: / Quasi diceret: are three different formulas in this work and are kept distinct — "As if:" / "As if to say:" / "As if he would say:". The bare Quasi: occurs five times and is never expanded into Quasi dicat.
• The ? after an indirect question does not occur here. All five question marks fall on direct questions (0602B Quid enim? · 0604A Quomodo potest fieri…? · 0605C quomodo tua? and Nunquid tu tibi imposuisti legem? · 0606B quid me qui de aliis?), so Pattern 8b was not needed and no colon-recast was performed anywhere.
• 0604A: Migne continues after the question mark with a lower-case sed. Kept lower-case.

§8 · FIXED RENDERINGS HELD ACROSS ALL FOUR CHUNKS (one voice)

§9 · READ CLEAN — named, so a blind reader can tell checked ground from unexamined

• The whole Argumentum (0601A, 3 sentences): nec receperunt carried; the subject of collaudat named rather than pronouned.
• 0601A–0602A, chapter I vv. 1–12: 18 lemmata, 1 [var:], no defect found. The Perficiet gloss (Ipse ut velimus operatur incipiens, qui volentibus cooperatur perficiens) is a compressed Augustinian tag and construes as printed.
• 0603C–0604B, the Carmen Christi and its gloss: §3. Zero divergences; two negation placements checked individually; infectus carried.
• 0604B, the De dono perseverantiae excerpt: the longest single sentence in the work (60 words), five negation particles, one non solum … sed etiam correlative and one non … nisi. Read three times against the Latin. This is the passage most like the one 2 Thessalonians got wrong and it is the passage a blind reader should re-read first.
• 0605C, the four-sentence diatribe on meam justitiam: two question marks, both direct, both kept; the singular thou fixed here and nowhere else.
• 0606C, the two-righteousnesses paragraph: four negations, a licet non immerito litotes kept as a litotes ("although not undeservedly").
• 0608B–0608C, the four scit clauses (scit humiliari · scit abundare · scit satiari · scit esurire): a deliberate four-part figure, kept four-part with one English verb each. ⚠ The third has no verb of its own (Qui acceptis alimentis non ad ingurgitationem ventris, nec plus carni tribuit…) — tribuit serves both limbs. English supplies "[gives it]" in square brackets, our standing convention for a supplied word, rather than silently inventing a verb.
• All 32 column anchors counted in the Latin before writing and in the English after; all eight [n:] likewise; all five ? likewise. ⚠ Counting is not placing — one of the 32 was in the wrong place and no count could see it. See §11, repair 3.

The three tests from the brief, run on these files

1. [sic:] must not swallow the word it marks. One marker in the work. Stripped, its clause keeps verb, object, conjunction and needs no negation. §4.
2. A negation can be RELOCATED. §6 lists every particle with the single Latin word it attaches to and the English word it sits on. The 0604B si velint site — the 2 Thessalonians shape exactly — is affirmative in both languages.
3. Punctuation compared BY COLUMN BAND, not by total. Both twins partitioned by the 32 verified anchors and compared band by band for : ; ? « » !. The first run returned 21 mismatched bands — 14 bands carrying a supplied semicolon, 3 a supplied colon, and 4 bands where a printed mark had been DESTROYED: a semicolon at 0604A and colons at 0604B, 0605B and 0606D. ⚑ The destroyed marks are the ones a total would have hidden — chunk 0's surplus of supplied marks and chunk 1's and 2's deficits sit in different chunks, and within chunk 2 the supplied semicolons at 0606B–0606D would have masked the lost colon at 0606D exactly. All 21 corrected against the Latin. Final: 0 mismatched bands, all four chunks. ⚠ Migne prints 8 semicolons of his own in this work, so the Tobit zero-semicolon exception does not apply here and was not invoked.

§10 · WHAT A BLIND READER SHOULD HIT HARDEST

1. 0604B, the perseverance excerpt. Longest sentence, densest negation, one correlative, and the si velint clause that has already broken one book in this batch.
2. 0603C, the two placements of non in the Non rapinam gloss. They are two lines apart and they are not the same claim. If one of them has drifted onto the other's word, that is the defect this work is most likely to be carrying.
3. 0602C, Pro hac vita mortale corpus tradidit. Two subjects construe. I took Paul and said why in §5; Christ is the theologically expected reading and I may be wrong.
4. 0607A, Gaudium, et consolatio. If a later reader thinks consolatio should instead have taken a [cj:] rather than a [var:], the argument is in §4 and I would not resist it strongly; the decisive point for me was that the divergence sits inside a quoted lemma.
5. 0606D, circumcicis. The one [sic:], and the one place where a plate read would settle whether the defect is Migne's or Corpus Corporum's. §4 says how to check it.
6. 0603B, Superiores, rendered as the bare word rather than expanded. The fragment-expansion convention arguably wanted "esteeming others better than themselves"; §2 (4) says why the gloss forbids it. This is the one place I depart from the default reading of convention 3, and I depart from it deliberately.
7. 0607D, surpasseth / surmounts. A pair of English words chosen to keep two Latin verbs apart. If the edition would rather both read "surpasses", the change is trivial and the divergence is recorded here either way.

§11 · THE BLIND POLARITY READ — 2026-08-18, and the adjudication

A blind reader that never saw this file read all four chunks. Eight sites, two substantive. All eight repaired. verify-english.mjs re-run clean; band test re-run to 0 mismatched bands; italic-span parity re-checked and now exact (68 / 58 / 84 / 70, markers stripped).

Upheld

• The Carmen Christi zero (§3) stands, and for the right reason. The reader collated all twelve lemmata independently and confirms Migne's printed text there genuinely agrees with the Clementine — so there was nothing to conform. Both documented hazards verified: the two placements of non at 0603C are distinct in the English and each sits where Migne puts it, and infectus → "unmade" preserves the play against Factus where "uncreated" would have read better and killed it.
• 0604B, the De dono perseverantiae excerpt — clean on all five negatives. The densest site in the work, named in §10 as the one to hit hardest, held.
• 0603B Superiores, left bare — endorsed. An expanded lemma would pre-empt its own exposition.
• The single [sic:] at 0606D passes the swallow test. Rendering stands beside the marker.
• 0602C Pro hac vita … tradidit — "Paul" is correct, but my argument was worthless and is replaced in §5. See there.

⭐⭐⭐ Repair 1 · 0605D (VERS. 11) · Occurram — A VULGATE CONFORMATION THAT SEVERED THE GLOSS'S OWN HINGE

Was: "I may attain." — Douay's rendering of Phil. III, 11. Now: "I may meet."

The whole gloss exists to expound occurram in its literal sense: occurram Christo in judicio · occurrent Domino in aera deportati ab angelis — the running-out-to-meet of 1 Thess. IV, 17. I rendered that same verb "meet" twice within four lines and the lemma itself "attain", so an English reader could not see that the "meet" of the exposition was the lemma's own word.

⚠ Aggravated by a collision I had not noticed: "attain" was also doing duty for assequi at 0606C (illam non assequitur). Two Latin verbs on one English word — and the one that lost its identity was the one being glossed. Fixed by giving each verb its own English (§8): occurrere = meet · assequi = attain · pervenire = come. That third move required changing ad quod pervenimus from "whereunto we have attained" to "whereunto we have come" in all three places.

⛔ This is 7a″ exactly, and it landed where I was not watching. §1 records the lemma as "agrees" with the Vulgate — which is true, and which is precisely why the guard was down: a lemma that agrees with the Vulgate still has to be TRANSLATED from the Latin. 7a″ is usually described as conforming a divergent lemma to the familiar English; this is the other half of the class, where the Latin agrees and the familiar English is still the wrong word because the gloss is built on the Latin verb's own sense. The next translator should read the divergence list as not covering this.

⚠ [var: Douay "I may attain"] was offered and is DECLINED. Pattern 14 marks a divergence between Migne's printed scripture and the received text; here Migne prints the Vulgate's own word and there is no textual divergence at all — only a difference between our literal rendering and Douay's. A [var:] there would tell a reader that the Latin diverges when it does not. The received English rendering is recorded here instead, which is where a note about a translation belongs.

⭐⭐⭐ Repair 2 · 0607A (VERS. 1) · THE LEMMA HALF-CONFORMED TO THE READING ITS OWN MARKER FLAGS

Migne prints Gaudium, et consolatio. — no possessives. I wrote "My joy, and my consolation.", supplying two, unbracketed — and the words supplied were exactly meum and mea, which the [var:] on the same line declares to be the Vulgate's text and not Migne's. The lemma and its own apparatus contradicted each other on one line.

Now: "Joy, and consolation." — the marker's claim and the text's rendering agree.

⚑ Worth stating plainly, because it is a new shape: the conformation hid inside the very entry that was catching the conformation. §4 argued consolatio against corona correctly and at length, and while doing it imported the two possessives from the same verse. Marking a divergence in one word is no protection at all for the words beside it.

⭐⭐⭐ Repair 8 · the 0607A [var:] reworded to carry its own inference

The marker recorded that the Vulgate reads corona mea — but not the thing that matters. It now reads: [var: Vulg. gaudium meum, et corona mea, "my joy and my crown". The gloss printed beneath this lemma expounds the crown and nothing else, and leaves Migne's own word unglossed] so that a reader meets the inference (the Glossa's exemplar read corona; Migne's lemma is corrupt at that word) rather than having to assemble it from the facing column.

⭐⭐ Repair 3 · 0601D · A COLUMN ANCHOR ~55 ENGLISH WORDS DOWNSTREAM

Migne breaks the column at Id est intimo | [0601D] amore Christi. My anchor had migrated to the second occurrence of amore Christi, three sentences later — so the locator pointed a reader at the wrong fifth of the column. Moved to "with the inmost [0601D] love of Christ".

⛔ The limitation this exposes, and it is the line I most want the next translator to have: AN ANCHOR CAN BE PRESENT, VERBATIM, AND IN THE RIGHT ORDER, AND STILL BE IN THE WRONG PLACE. verify-english.mjs checks anchors for presence and order, not position, so a displaced anchor passes silently. ⚑ And the band test does not catch it either, which I can now state as a measured fact rather than a worry: the ~55 words that moved between bands 0601C and 0601D contain no colon, semicolon or question mark, so the band comparison returned 0 mismatches both before and after the repair. Two independent mechanical checks, and the site is invisible to both. Column anchors are the citation-resolving unit of this edition; the only test for their position is reading the twin at the anchor, and it costs one pass. The other 31 are exact to the word, including the ones deliberately split mid-construction to match Migne (non [0605B] recedentes → "not [0605B] departing"), which is what made this one anomalous against my own standard rather than against a rule.

⭐⭐ Repair 4 · 0608B · A CONNECTIVE ADDED INSIDE LEMMA ITALICS

Migne's italicised lemma is modo refloruistis, two words. My English lemma carried three — "but now you have flourished again" — while his actual sed, which stands outside the lemma, was separately rendered "whereas." The connective had been duplicated and half of it pushed inside quoted Scripture. Now: "now you have flourished again, but before you were occupied…" ⚑ Italic-span parity could not see this: the span count was right, only its contents were wrong.

⭐ Repair 5 · 0606C · AN et … et CORRELATIVE FLATTENED

et propter illam quae jam inerat, perfectum, et propter istam quae adhuc deerat imperfectum se esse dicebat. The first et was unrendered. The pair is "both … and", and the sentence's whole point is the simultaneity of the two self-descriptions — the Titus-ita class, where the correlative is what marks a paradox as deliberate rather than as an inconsistency. Now: "said that he was both perfect, on account of that which was already within, and imperfect, on account of this which was still wanting."

⭐ Repair 6 · 0605D · A SUPPLIED POSSESSIVE MOVING TOWARD THE VULGATE, AND A BRACKETING INCONSISTENCY

et configurationem mortis → I had "the conformation of his death", unbracketed; the "his" is the Vulgate's morti ejus (Phil. III, 10), which Migne does not print. Now "[his]".

And the bracketing practice is now consistent throughout, which it was not: I bracketed [gives it] at 0608C and then left two other gapping recoveries bare within eight lines of it. All three gapped verbs are now bracketed — [gives it] (0608C, tribuit serving two limbs), [guards] (0608A, custodit gapped in quisquis autem famam), [did it] (0608C, the raven clause) — as is [loved] at 0601D. The rule I am now holding: any English word supplied to fill a Latin ellipsis is bracketed, every time.

⭐ Repair 7 · 0604A · ejus LOSING ITS PRINTED SIGNAL

nunc de claritate ejus … ostendit → was "now he shows concerning his glory", where English "his" reads by default as the subject's. ejus, not sua, is Migne's printed instruction that the possessor is not the subject: Paul shows, the glory is Christ's. Now "concerning his [Christ's] glory."

⚑ §5 had already identified this site and named the mechanism correctly — and left the English alone. That is the 7a′ failure exactly ("a logged column is not a checked column"), and it is worth recording that it happened here in the very table built to prevent it. The other ejus at 0602B (Constantia enim ejus) reads "his constancy" with Paul the possessor and Constantia the subject, so there is no pronoun there to mis-bind; it needed no repair and got none.

⭐ THE TWO-READING SITE, DECIDED · 0606D · ut omnes idem de illo sapiamus

de illo is masculine or neuter — the form does not decide it, and a flag is not a decision. Decided: NEUTER. The English now reads "that we all be of the same mind concerning that same thing whereunto we have come by faith and knowledge" (was: "concerning him").

For the neuter:

1. The relative that follows is neuter and immediately adjacent — de illo, ad quod pervenimus. Read as one referent the sentence is seamless; read as two it has a masculine antecedent and a neuter relative side by side with nothing between them.
2. The subject of the whole gloss is a thing, not a person — the regula disciplinae, and quod commune sit et modestum in veritate evangelica, three lines above; and the lemma it closes on is et in eadem permaneamus regula.
3. No person has been named for several lines, and the nearest masculine noun (ad cultum Dei) is a genitive inside another clause, reached across an intervening subdit.
4. The paragraph before it has live neuter-ish deixis in istam … illam for the two righteousnesses, so ille is already working as a pointer rather than as a person.

Against it, honestly: the paragraph's frame is ad cultum Dei excogitaveritis, so "of one mind concerning him" construes and is not strained; and if the Glossa meant the neuter, de eo or de eodem would have been the ordinary choice. The reading was taken on (1) — the adjacency of ad quod — which is textual rather than thematic, and a later editor who prefers the masculine has the case for it above, in full, rather than a bare flag.

Sites the reader raised that needed no change

None. All eight were taken, and the two-reading site was decided rather than re-flagged.

What this read changes about §10

§10's list was wrong about where the risk was. It named 0604B (clean), 0603C (clean), 0602C (conclusion right), 0603B (endorsed) and 0606D-circumcicis (confirmed on the plate). The two substantive defects were at 0605D and 0607A — one lemma I had recorded as agreeing with the Vulgate, and one lemma I had spent a starred crux entry arguing about. ⭐ The lesson generalises: the passages a translator flags are the passages a translator has already read. The defects were in the two places I had written down as settled.