Liber Esther

Translated 2026-08-18, one translator, 4 chunks (3,252 Latin words), on Opus. verify-english.mjs: clean except one expected, pre-announced failure — see §0.

Apparatus: 10 [var:] · 1 [sic:] · 2 [cj:] · 0 [d:] · 0 [ed:] · 1 [nt:] · 47 [n:] (18 / 12 / 11 / 6, exact parity with the Latin twins). Structure: PROLOGUS + CAPUT PRIMUM → CAPUT X, eleven heads, 1:1. Guillemets: 107 « / 106 », 1:1 with the Latin. Three of Migne's own defects account for the imbalance and stay uncorrected (§3). Italics: 3 spans (1 usque ad, 2 guilleted-elsewhere scripture quotations), 1:1. Question marks: ONE in the Latin (0740D), one in the English. None supplied. Column anchors: 34, verbatim and in order.

§0 · ⛔ ONE OPEN DEPENDENCY — READ BEFORE SHIPPING

verify-english.mjs reports exactly one error and it is expected:

0000.md: [sic: convivum] is not verbatim in the Latin twin

This is the orchestrator's plate finding at 0739D (§2.2). Migne's plate prints convivum; our TEI reads convivium. The [sic:] is correct against the plate and cannot pass the substring check until the pending data/tei-patches/8948.json restores convivum to the Latin side. The orchestrator holds that patch and will apply it after these files close. Do not "fix" the marker by deleting it — that would put back the silent upstream repair this edition exists to expose.

Every other check passes: columns, notes, sections, guillemets, italics, question marks, frontmatter verbatim, no duplicate paragraphs.

§1 · CONVENTIONS — INHERITED, NOT RE-DERIVED

Ruth (8968) §1 and Tobit (8970) §1 govern this book unchanged and nothing here re-opened them: VERS. n.-- passes through verbatim; the lemma is Englished; a fragmentary lemma is rendered as the phrase it stands for; sigla ride as ordinary [n: …] notes with Migne's own internal spacing; guillemet marks are sacred and 1:1, and notes and column anchors legitimately sit inside guillemet spans (this book does it constantly — e.g. 0742B, where [0742B] falls inside the I Cor. 15:9 quotation, and 0740C, where [0740C] falls inside the Matt. 22:2 quotation).

What Tobit did NOT cover, decided here, for the OT guillemet books still to come:

(a) usque ad in this book is USUALLY A LEMMA, not the abridgment formula

Tobit ran 17 usque ad abridgments. Esther runs one — at 0739C, in the opening Rabanus extract, and it is the only italic in chunk 0. But Esther also prints « Usque ad fores. » at 0744A — Esther 4:2 — which is a guilleted lemma, not an abridgment, and is Englished as one ("« As far as the doors. »"). ⚑ The typography is the test and it is reliable in this recension: the formula is always italic, a lemma is always guilleted. An agent working from a word-search rather than the markup will conflate them.

No verb was supplied to the one genuine abridgment; it ends mid-clause and stays there.

(b) The fragment rule YIELDS to the negation rule where they collide

Ruth's settled rule renders a stump as the phrase it stands for. Twice here that would have imported a word from the Vulgate, and the two cases were decided opposite ways on purpose:

• 0741C · « Nolentes cogeret ad bibendum, » etc. The Vulgate clause is nec erat qui nolentes cogeret ad bibendum. Migne prints only the tail, and prints NO negative. Completing the fragment would have added a nec the plate does not carry — the exact thing 7a forbids — so the fragment stands as a stump: « Should compel the unwilling to drink, » etc. It is awkward English and it is deliberate. (Nolentes → "the unwilling" carries the only negation Migne actually prints here.)
• 0744A · « Non enim. » (Esth 4:2, non enim erat licitum … aulam regis intrare). Completed to « For it was not lawful. », because the supplied word is licitum and supplying it changes no polarity. The negative Migne prints is rendered on the word he prints it on.

⚑ The rule this book adds: complete a fragment freely, EXCEPT where the completion would supply or move a negative. There the stump wins.

(c) ⭐ A BRACKETED SUPPLY BELONGS IN GLOSS PROSE, NEVER INSIDE A GUILLEMETED LEMMA

The rule, and the remaining OT books inherit it: bracketed supplies belong in gloss prose. Inside a guilleted lemma a bracket of ours reads as an interpolation into Migne's quoted scripture, not as a supplied word in our own sentence, and the supply must therefore be made in the surrounding prose or not made at all.

The case that set it: at 0747A the lemma « Et plurimum posse cognoverant. » has its antecedent genuinely off the page — the quem (Mardochaeus) that the abridgment cut off sits in the previous verse — so bracketing "he" as [Mardochaeus] was offered, at nil risk, because there is no competing candidate. Declined anyway, and the ground is the convention rather than the risk: the supply would sit inside a quoted verse, and a reader cannot tell our bracket from an editorial insertion into scripture.

Checked before ruling: every bracketed supply in this work, and in Tobit (8970) and Judith (8960), is in gloss prose. The practice was already uniform and undeclared; this states it. ⚑ Where a lemma genuinely needs its referent named, name it in the gloss sentence that follows — which is what this book does at 0742C (Hanc… revocavit, §5.3) — or leave it in the cruces.

(d) Migne's own dropped « is not ours to restore

Three lemma-openings in this book lose their opening guillemet and the space after -- with it (§3). The English reproduces both the missing mark and the missing space, e.g. VERS. 8.--Every one what he would, » etc. This looks like our typo and is not.

§2 · PLATE STATUS — READ, AND WHAT IT SETTLED

1. Foot-of-page apparatus: CHECKED ZERO — tier 3, coverage full, notes_recovered: 0

Read by the orchestrator during this session, not by the translator, and merged here. Method: archive.org patrologiaecurs04migngoog (on disk at raw/scans/pl113/), PDF page = (column + 11) / 2, verified against the running head (p. 376 = cols 741/742, head GLOSSA ORDINARIA. — LIB. ESTHER.). Pages 375, 376, 377, 378 and 379 read at full page width across the whole foot. No foot rule and no numbered notes; the columns simply end. p. 375 was read at the mid-page work division, where Judith ends and LIBER ESTHER. begins below the rule — the case that lost three notes on PL 202.

Write this down as a checked zero, not a skip. Third sampling of PL 113; the volume's "no conjecture apparatus" finding now rests on fourteen pages across three books. ⛔ Therefore the two [cj:] markers in this work are MINE, not Migne's (§5.1, §5.2).

⚑ Not apparatus, not reported: p. 378's foot carries the signature line PATROL. CXIII. and the gathering number 24. Bookbinding.

2. ⭐ 0739D · convivum — the digitization SILENTLY CORRECTED Migne. A NEW CLASS.

The plate prints « Grande convivum cunctis principibus. » — convivum, no second i. Read at 900 dpi on p. 375; the letters are unambiguous. Our TEI reads convivium.

⚑⚑ This runs the OPPOSITE way from every digitization defect the corpus has met. Tobit's Nephtlali was Corpus Corporum corrupting what Migne printed; here Corpus Corporum has normalized a defect away, so our Latin column would print a word Migne did not set. Name the three classes apart, because the next translator will meet all of them:

Rendered "« A great banquet [sic: convivum] for all the princes. »" — the sense is untouched, so the English carries a real word and the bracket only records provenance (the Tobit failure mode, §0 of the brief, explicitly avoided).

What makes the defect local and therefore visible: Migne prints convivii correctly at VERS. 5 two lines on (« Dies convivii. »), and convivium/convivio correctly six more times in the work. It is one lemma, not his habit.

⚠ The lesson, and it is uncomfortable: a lemma that looks perfectly correct may look correct because someone fixed it, and nothing in the Latin can tell you. I have flagged the one other word in this book that struck me as suspiciously tidy — Nochum at 0739C (§5.5) — for a plate read, and can offer no systematic detection beyond that.

3. The Jerome prologue is a CROSS-REFERENCE, and nothing is missing

p. 375 prints, in order: the rule · LIBER ESTHER. · B. HIERONYMI IN LIBRUM ESTHER PROLOGUS. · beneath it, in parentheses, (Vide inter opera B. Hieronymi, tom. VIII.) · a short rule · CAPUT PRIMUM. — and only then the text. Migne prints the head and sends the reader to Jerome's own works instead of reprinting the prologue. Our twin has this exactly right. No lacuna is logged, none exists, and the head is translated like any other.

4. Corroborated by the plate, not merely inferred

cols 739/740: the opening lemmata and Susan metropolis est in Perside… · cols 741/742: VERS. 11. — « Ut ostenderet, » etc. and VERS. 15. — « Evoluto autem. » · cols 745/746: CAPUT VIII and CAPUT IX's VERS. 4. · cols 747/748: the close.

⚑ One item in the orchestrator's corroboration list — « Si negavi quod volebant, pauperibus… » (Job 31:16) — is not in this work at any column, and is not in our twin. It belongs to the band of another work sharing cols 747/748. Reported so it is not later read as text we lost.

§3 · THREE UNBALANCED GUILLEMET SITES — ALL MIGNE'S, ALL REPRODUCED

Guillemets run 107 « / 106 ». A naive left-to-right counter reports the wrong sites, because one unclosed open at 0741C silently swallows the next close; these are the real four, found by locating each defect individually:

⚑ The 0741D and 0742A pair is one defect twice: both are lemma-openings that lost « — the mark and the following space — so the address reads VERS. 8.--Unusquisque where every sound address reads VERS. 7.-- « Vinum. ». The missing space is the tell, and it is preserved. Whether the loss is Migne's compositor or our transcription cannot be settled from the text; it was not on the orchestrator's plate-read list and is offered as a candidate for a future read.

⚑ 0747A is the one to watch for a span-extractor, exactly as Tobit's 0728A was: the unclosed « makes the two Acts quotations plus the intervening Item: read as a single 40-word "lemma". They are two quotations, Acts 2:43 and Acts 5:13.

§4 · THE LEMMA LIST — ALL 103 SPANS, CHECKED AGAINST THE VULGATE

Esther's Latin is among the least stable in the Vulgate — Jerome translated the Hebrew book and appended the Greek additions separately — so this section is the deliverable. Every span in the work appears below, including the ones that agree exactly.

4a · THE DIVERGENCES — eight [var:] fired

(Ten rows, ten markers, one per divergent span: 1 in chunk 0, 5 in chunk 1, 3 in chunk 2, 1 in chunk 3. Counted in the files.)

4b · DIVERGENCES FOUND, MARKER DECLINED — with reasons, because a blind reader will look

• 0740C · Simile est regnum coelorum (Matth. XXII). Matt 22:2 reads Simile factum est. A dropped auxiliary; the comparison is unchanged. No marker.
• 0741B · Columnas fecit argenteas (Cant. III). Cant 3:10 has Columnas ejus fecit. A dropped possessive in a lemma the gloss does not build on.
• 0741B · Beati enim pauperes spiritu (Matth. V). Matt 5:3 has no enim. A connective added to fit the sentence frame — the citing author's, not a textual variant. Same class: Fiebat enim omni animae timor (0747A, Acts 2:43 autem), Nemo audebat… et magnificabat (0747A, Acts 5:13 sed), Cumque et secundo (0743A, Esth 2:19).
• 0741D · Ubi venit plenitudo temporis, misit Deus Filium (Gal. IV). Gal 4:4 has At ubi… and Filium suum. Elision, no sense change.
• 0742B · Ego sum minimus apostolorum (I Cor. XV). 1 Cor 15:9 has Ego enim sum.
• 0742B · Quia vos indignos judicatis vitae aeternae (Act. XIII). Acts 13:46 has quoniam indignos vos judicatis aeternae vitae. Connective and word order only.
• 0743B · Historiis et annalibus traditum coram rege (Esth 2:23). The Vulgate has traditumque est historiis, et annalibus mandatum coram rege. Recast and compressed into a participle; the assertion — that it was entered in the records before the king — is identical, and the gloss reads it as the two Testaments either way. Declined, and this is the closest call in §4b; a reader who thinks it should have fired has the reading here.
• 0743C · Justus angustia liberatur, et traditur impius pro eo (Prov. XI). Prov 11:8 has Justus de angustia liberatus est, et tradetur impius. Perfect and future flattened to a gnomic present, and de dropped. A proverb asserts the same thing in either tense. Declined.
• 0744B · Omnia quaecunque voluit fecit in coelo et omnibus abyssis (Psal. CXIII). A conflation, not a variant: the head is Ps 113:11 (omnia quaecumque voluit fecit), which is what the note cites and cites correctly; the tail et omnibus abyssis comes from Ps 134:6, whose in terra, in mari is dropped and whose in before omnibus is gone (rendered "and in all the depths", the preposition governing both members). No [var:]: the printed citation is right for the verse it names, and a marker would have to assert a received reading for a line that is two lines joined.
• 0744C · Nolite solliciti esse de crastino (Matth. VI). Matt 6:34 has Nolite ergo solliciti esse in crastinum. Same meaning; the semicolon inside the guillemets is Migne's and is kept.
• 0745C · Tunc invocabunt, et non exaudiam eos (Prov. I). Prov 1:28 has invocabunt me… et non invenient me. Objects moved, not changed.
• 0746A · Custoditur justo substantia peccantis (Prov. XIII). Prov 13:22 has peccatoris. Participle for noun, identical sense.
• 0746A · Auferetur a vobis regnum (Matth. XXI). Matt 21:43 has regnum Dei. A dropped genitive; the gloss glosses it as the Jews' dignity, which needs no Dei.
• 0745D · Est qui accusat vos, Moyses, in quo speratis (Joan. V). John 5:45 has in quo vos speratis. Dropped pronoun.
• 0747A · Fiebant prodigia et signa multa (Act. V). Acts 5:12 has signa et prodigia. Order only.
• 0746B · Tradidit Mardochaeo (Esth 8:2). The Vulgate reads deditque eum Mardochaeo. A synonym for the verb and the pronoun clipped; the lemma is a stump either way. Declined and flagged — Esther's transmission is unstable exactly here, and this is the one lemma in the book where I would not be surprised to learn Migne's copy-text simply read tradidit.

4c · CHECKED AND AGREEING WITH THE VULGATE — named, not summarized

Esther lemmata, chunk 0 (0739C–0742B): Susan civitas, (1:2) · Tertio igitur anno. (1:3) · Grande convivum cunctis principibus. (1:3 — the words agree; the spelling does not, §2.2) · Dies convivii. (1:5) · Et pendebant. (1:6) · Et columnis marmoreis fulciebantur. (1:6, word for word) · Super pavimentum. (1:6) · Vinum. (1:7) · Nolentes cogeret ad bibendum, (1:8, tail of the clause — §1b) · Unusquisque quod vellet, (1:8) · In palatio, ubi rex Assuerus, (1:9) · Ut ostenderet, (1:11) · Quae renuit. (1:12) · Responditque Mamuchan audiente rege atque principibus: Non solum regem laesit regina, (1:16, word for word including the colon).

chunk 1 (0742C–0744C): Qui fuit. (2:7) · Esther. (2:8) · Et septem puellas. (2:9) · Quae noluit. (2:10) · Ingredientesque. (2:13) · Atque inde. (2:14) · Evoluto autem. (2:15) · Cumque et secundo. (2:19, §4b) · Duo eunuchi regis, (2:21) · Historiis et annalibus traditum coram rege, (2:23, §4b) · Aman filium. (3:1) · Missa est sors. (3:7) · Tulit ergo. (3:10) · Usque ad fores. (4:2 — a lemma, not the abridgment formula, §1a) · Non enim. (4:2, §1b) · Rursumque. (4:15) · Virgam auream. (5:2) · Si inveni in conspectu. (5:8).

⚑ Tebeeth. at 0742D is not a Vulgate divergence and is not marked. The Vulgate reads Tebeth (Esth 2:16); Migne prints Tebeeth, and the same clause prints Xudimos for the Greek month (Ἀυδυναῖος / Audynaeus). Both are transliterated proper names in a list of three calendars, which is exactly the environment in which spellings vary legitimately, and neither is a non-word. Rendered as printed, no [sic:], no [var:]. Declined, and named so a reader with the plate can close it — if the plate reads Tebeth, this is a digitization corruption of the Tobit class and takes a patch, not a marker.

chunk 2 (0744D–0746D): Noctem illam duxit rex insomnem, (6:1, word for word) · Ventum est. (6:2) · Festina. (6:10) · Tulit itaque. (6:11) · Secunda die, (7:2 — no VERS. address, §5.6) · Intravit in hortum. (7:7) · Et statim. (7:8) · Harbona. (7:9) · In domo Aman. (7:9) · Die illo. (8:1) · Domum Aman. (8:1) · Tulitque rex. (8:2) · Tradidit Mardochaeo. (8:2, §4b) · Procidit ad pedes, (8:3) · Obsecro ut. (8:5) · Accitisque scribis, (8:9) · Qui centum viginti. (8:9) · Interficerent, (8:11) · In Susan, (8:14).

chunk 3 (0747A–0748C): Et plurimum posse cognoverant. (9:4 region) · Itaque percusserunt Judaei. (9:5) · Quorum ista sunt nomina. (9:6) · In tantum. (9:16) · Dies autem. (9:17) · Et misit Phur, (9:24) · Isti sunt dies. (9:28) · Rex vero Assuerus omnem. (10:1).

✅ Et plurimum posse cognoverant. at 0747A — CLOSED AT THE BLIND READ: clean, no defect, no marker. The lemma is the tail of an accusative-and-infinitive whose head quem (Mardochaeus) the abridgment cuts off, and whose subject — "they", the officials of Esth 9:3 — sits in the previous verse. My English is correct as written: "they" for the officials, "he" for the severed quem, plurimum posse literal, cognoverant pluperfect ("had learned"). ⚠ Record the basis honestly: this was settled from the reading agent's own knowledge of Esther 9:4, with no witness on disk — the same evidentiary footing as my own declination, resolved by a second reader rather than by a text. Bracketing "he" as [Mardochaeus] was offered and DECLINED: the antecedent is genuinely off-page, but it has no competing candidate, and the supply would sit inside a guilleted scripture lemma, where a bracket of ours reads as an interpolation into the quoted verse rather than as a translator's supply. Every bracketed supply in this work is in gloss prose. The convention holds.

The open question as I left it, kept because the shape of it is worth a later agent's time: No Vulgate text is on disk in this repo, so every check in this section is against my own knowledge of the Clementine, and for this phrase — which sits at the Esth 9:3/9:4 seam, where the chapter-and-verse division itself varies between editions — that was not enough to be sure whether Migne's cognoverant stands where the Clementine puts it, or whether his VERS. 4. address is one verse high. Declined rather than fired on a half-check, which is the call Tobit's Exsurgens vindicated. A reader with the text can close it in one look. The rendering ("And they had learned that he could do very much") is right for the printed words either way.

4d · NON-ESTHER SCRIPTURE QUOTED INSIDE THE GLOSSES — read word by word

The brief names these as where the reach for a familiar English Bible happens, so each was checked against Migne's Latin and not against the Douay:

Exact: Homo quidam fecit coenam magnam: et vocavit multos (Luke 14:16 — the colon is Migne's and rides inside the guillemets) · Accinxi te et non cognovisti me (Isa 45:5) · Non sum missus nisi ad oves, quae perierunt domus Israel (Matt 15:24) · Ego confirmavi columnas ejus (Ps 74:4) · Discite a me, quia mitis sum (Matt 11:29) · Charitas Dei diffusa est in cordibus nostris (Rom 5:5; Charitas for caritas is orthography, Pattern 9) · Qui potest capere capiat (Matt 19:12) · Qui se humiliat exaltabitur (Luke 14:11) · Alias oves habeo quae non, etc. (John 10:16) · Ibunt hi in supplicium aeternum: justi autem in vitam aeternam (Matt 25:46) · Pater tuus Amorrhaeus, et mater tua Cethaea (Ezek 16:3) · Astitit regina a dextris tuis (Ps 44:10) · attingit a fine usque ad finem fortiter: et disponit omnia suaviter (Wisd 8:1, minus ergo) · Data est mihi omnis potestas in coelo et in terra (Matt 28:18) · Non remanebit ex eo quidquam usque mane (Exod 12:10, Non for nec and reordered — sense identical) · Deposuit potentes de sede, et exaltavit humiles (Luke 1:52) · Convertetur Libanus in Chermel, et Chermel in saltum reputabitur (Isa 29:17) · Quia omnis qui se exaltat, humiliabitur (Luke 14:11) · Cum fortis armatus custodit atrium suum (Luke 11:21) · Confidite, ego vici mundum (John 16:33) · Sortes mittuntur in sinum: sed a Domino temperantur (Prov 16:33) · Cum judicatur, exeat condemnatus (Ps 108:7) · In matutino interficiebam (Ps 100:8) · Si vis ad vitam ingredi, serva mandata (Matt 19:17, minus autem).

Divergent: the eight of §4a and the fourteen of §4b.

Unidentified, and reported as such: 0746B · « Nutrivit corda fidelium. » I cannot place this in the Vulgate. It is guilleted and cited as scripture (unde:) and carries no [n:]. Rendered literally ("He hath nourished the hearts of the faithful"), unmarked — a [var:] requires a received reading to set against it and I have none. Named so that a reader who recognizes it can close it.

4e · ⭐ FIVE WRONG CITATIONS — AND ONE IS AN OFF-BY-ONE CHAIN. NOTHING CORRECTED.

Runbook false positive 2, and Tobit's three wrong locators were all correctly left alone. Nothing here is corrected either.

• 0741C · (Luc. X) on Qui se humiliat exaltabitur — Luke XIV:11 (or XVIII:14).
• 0744A · (Psal. XIV) on Astitit regina a dextris tuis — Ps XLIV:10. An X for an XL.
• 0745B · (Luc. XXIV) on Quia omnis qui se exaltat, humiliabitur — Luke XIV:11. ⚑ The same verse is misattributed twice in this book, to two different wrong chapters (Luc. X at 0741C, Luc. XXIV at 0745B), which is worth recording because it argues the errors are compositorial rather than a bad copy-text.

⭐ The chain at 0744C, and it is the most valuable single finding in the apparatus. Three quotations run in a row and the notes are displaced one place to the right:

Every note is right for the quotation before it and wrong for the one it sits on. The first quotation is left bare and the last is left holding a locator two books away. ⛔ Reproduced exactly as printed, all three, including the bare first quotation. Correcting even one would break the chain's shape and destroy the evidence that it is a chain — and Migne's own text names Jacob and Genesis in the running prose (et Jacob in Genesi ait), which is how the displacement is detectable at all.

⚑ Note also that (Gen. 30) prints an arabic numeral where all 46 other notes in the work print roman. Reproduced. It is consistent with a note set out of sequence in haste.

A near-miss that is NOT an error: (Sap. VII) at 0744A is correct for Wisd 7:30, and (Ibid., VIII) correctly steps to Wisd 8:1 — Migne's Ibid. chain here works perfectly, which is the positive control for §4e.

§5 · THE BOOK'S OWN CRUCES

5.1 ⭐ 0744D · In se mobilis permanens — the sentence denies its own argument. [cj:] FIRED.

In se mobilis permanens cursus temporum et actus omnium contemplatur, et nulla eum latet cogitatio cui omnia praesentia.

Rendered literally: "Remaining movable in himself, he contemplates the course of times and the acts of all." The clause it introduces says nothing escapes him to whom all things are present — the standard Boethian contrast between the unmoved contemplator and the moving times. The printed word asserts the opposite of the sentence it opens.

The conjecture is immobilis, two letters away, and it yields exactly the argument: "Remaining unmoved in himself." Pattern 18's shape precisely — mobilis is a real word in a real case, so Pattern 12 has nothing to wrap and Pattern 7 must render it, and the English then asserts something the author did not. Marked additively:

Remaining movable [cj: mobilis; read immobilis, "unmoved"] in himself

⚑ The rival construal is recorded because it is not silly. mobilis could be taken with cursus — "abiding in himself, he contemplates the movable course of times" — which construes as sense and needs no conjecture. It fails on form: cursus here must be accusative (object of contemplatur), and the accusative plural of mobilis is mobiles, not mobilis. The rival reading requires its own emendation, so it is not the cheaper one. Both are on the page for a reader who disagrees.

⚠ The [cj:] is mine, not Migne's — the foot of p. 376 is a checked zero (§2.1).

5.2 0741B · Ex alibi — a formula this book uses correctly three other times. [cj:] FIRED.

Migne prints Ex alibi: « Columnas fecit argenteas… » where the book prints Et alibi: at 0740D, 0741C and 0744C, and alibi alone at 0747A. Ex is a real preposition and it governs nothing here, so it has no slot (Pattern 12 test 1) — but neither is it a non-word, so it cannot be carried in a [sic:]. Rendered "From elsewhere" and marked [cj: Ex; read Et, "And elsewhere"].

⭐ Declining this was live and was rejected on Tobit's notionum ground: the faithful English is merely odd rather than false, so Pattern 18a's bar is not obviously cleared — but declining would leave the site with no apparatus at all, and a reader would meet "From elsewhere" with nothing to tell him the page is broken. The reportable defect is the absence of a marker, not the choice of reading.

5.3 ⭐ THE UNEXPRESSED SUBJECT — every supplied pronoun, and what decided it

This is the brief's dominant defect class and this book's crowded cast (Assuerus, Vasthi, Esther, Mardochaeus, Aman, the Jews, the Church, the Synagogue, Christ, the devil) is its worst ground. Five sites needed a decision. All five took a bracketed supply rather than a bare pronoun, because a bare pronoun binds silently in a sentence whose referent changes mid-clause — Ruth's 0538C failure.

(i) 0739D · incarnationis suae sacramentum patefecit. → "he [Christ] laid open the mystery of his incarnation." The lemma is Tertio igitur anno and the gloss's own next sentence names the three times, the third sub gratia. The banqueting king is Assuerus in the letter and Christ in the allegory, and incarnationis suae can only be Christ's. Not close.

(ii) 0740C · Judaeos… de manu inimicorum liberat, atque ipsos juste condemnat. → "delivers the Jews… out of the hand of their enemies, and justly condemns those same [enemies]."

⚠ This is the sentence the whole book turns on and both readings parse. ipsos is adjacent to Judaeos and an English reader will bind it there — "and justly condemns them", i.e. the Jews — which is a theologically loaded claim the gloss does make elsewhere (0743C, 0747A). It is the wrong reading here. Two things decide it: the verb liberat has just made the Jews the beneficiaries, so atque cannot swing to condemning them without an adversative; and the Esther story the clause is glossing has Assuerus condemn Aman and the enemies, which is what inimicorum names two words earlier. The bracket is there because the wrong reading is the one that arrives first.

(iii) 0745B · Sed ubi Esther petitionem suam aperuit, damnatus ad poenam secessit. → "he [Aman], being condemned, withdrew to his punishment." damnatus is masculine and Esther is the cum-clause subject; the paragraph's next sentence names Aman by name and casts him out of the banquet. No live alternative.

(iv) 0745C · reginam pro salute sua deprecari contendit. → "he [Aman] strove to beseech the queen for his own safety." The ablative absolute has just made rex the subject of properante, so English reaches for the king — and the king is exactly who it is not.* Esth 7:7 has Aman rise to beg Esther for his life, and the following clause (sed tempus opportunum non habuit, and then the Proverbs 1:28 proof-text about calling too late) is about the doomed man, not the king. sua* referring to a non-subject is ordinary medieval Latin.

(v) 0746B · Laudante et praedicante instantiam magistrorum, a Domino… accipiunt. → "[The king] praising and proclaiming the earnestness of the masters, they receive from the Lord grace for grace." The ablative absolute has no noun at all — a participle pair hanging on nothing. The lemma is Tulitque rex (Esth 8:2), so the king is the one supplied, and magistrorum being genitive rules the masters out as the praiser. The plural accipiunt is the masters. ⚠ The BRACKETS were added at the blind read (§7.3): my first version supplied "The king" bare, so a reader met a well-formed ablative absolute and could not learn that the noun is not on the page.

(vi) ⭐⭐⭐ 0748C · historialiter, enim non omnem terram tributariam fecit, quia nec habuit. → "historically, for he [Assuerus] made not all the earth tributary, because neither did he possess it."

⚠ ADDED AT THE BLIND READ (§7.1), and it is the worst site in the work. My first version read a bare "he", and the only antecedent the English offered was Christ, named three words earlier and subject of both preceding verbs. The subject is Assuerus. historialiter marks the register, and the register is the whole architecture of the sentence: allegorically the verse's "all the earth" is true, of Christ, who filled the whole world; historically it is not, because the king of Esther 10:1 never possessed all the earth to tax it. The tell I had in front of me and did not read: quia nec habuit is false of Christ under this gloss's own premise, and my English therefore had the page saying Christ did not make the earth tributary because he did not possess it — flatly contradicting implevit totum mundum eight words earlier, in the same sentence.

⛔ This is Ruth's 0538C in Esther's clothing, and note what did NOT catch it. The subject is not ambiguous between two women; it is ambiguous between the literal and the allegorical register, and the marker that disambiguates it (historialiter) is an adverb, not a noun. Every antecedent-hunting habit I had was looking for a name.

⚑ And the one place a supplied pronoun is deliberately NOT bracketed: 0742C. Hanc Nabuchodonosor… transtulit, sed pietas divina… revocavit — "called her back". The antecedent is Hanc, printed two clauses earlier and referring to Esther, and no second feminine is in play. Recorded so the absence of a bracket reads as a decision.

⚑⚑ On the Vasthi/Esther "she" the brief warned about: it does not occur in this book. The gloss reads Vasthi's refusal (0742A) and Esther's raising (0742B onward), but Migne never runs the two women through unmarked pronouns in one sentence — each paragraph names its subject or inherits it from its own lemma. Checked paragraph by paragraph and reported as a checked absence, not as a section skipped.

5.4 The negation audit — 61 particles, every one carried, one RELOCATED and caught

Brief test 2 run as specified: for each printed negative, the single Latin word it attaches to was named, then that word found in the English and the negative checked to be on it. Not "are all the negatives present."

⭐ It caught one, and count-based checking would not have: 0742A · Non solum regem laesit regina. My first version read "The queen hath not only injured the king" — the negative floated onto the verb, where the Latin puts it on regem. The verse's whole force is that the injury reached beyond the king (sed et omnes populos), which the misplaced negative blurs. Corrected to "The queen hath injured not only the king." ⚠ Nothing was added and nothing dropped, so every count matched before and after. This is the 2 Thessalonians mechanism exactly, in a lemma rather than in gloss prose.

Every other site, checked and carried (attachment named in each case): Neque on necesse (0740C) · non on in hoc · nec on Aaron · Nec on Salomon · non on cognovisti (0740D) · non on demonstrant — the work's only question mark, preserved · Non… nisi on sum missus (0741A) · Nolentes (0741C, §1b) · Nemo on cogitur · non on servos · non on necessitatem · nulli on noceat (0741D) · non on augeat ×2 · noluit / noluerunt on introire / venire (0742A) · non on sum dignus (0742B) · absque on misericordia (0742C) · noluit (lemma) · non… ultra on sustinuit · nec… ultra on habet (0742D) · nisi on restituatur · Non on sufficit (0743A) · non inside the John 10:16 stump · nolunt on consentire (0743C) · Non enim (0744A, §1b) · Nullus on potest · non on esse · sine on intermissione (0744B) · non on deputatur (0744C) · Non… sed on in praesenti · Nolite on solliciti esse · Non on remanebit · nulla on cogitatio (0744D) · Non on est · nunquam on tradentur · invitus (0745A) · non valentes on occultare · nihilum (0745B) · non on habuit ×2 (0745B, 0745C) · non on exaudiam and on invenient · Ne on pullularent (0746D) · Nemo on audebat (0747A) · Non solum on alios operarios · nec on curant · non on exquirimus (0747B) · nec on cogimus · non on in toto · non on est historia · nec on subsistere · nec on procedere · non on humanum (0748B) · sine on manibus (0748B) · non on fecit (0748C) · nec on habuit · nulla on expositio (the closing [nt:]).

None dropped, none added. ⚑ Two are worth naming because carrying them produces strain:

• 0744A · Sapientia enim vincit malitiam (§4a). Migne's Wisdom quotation has no negative where the Vulgate has one. None was supplied. The [var:] is the only thing on the page that tells the reader.
• 0748C · quia nec habuit — "because neither did he possess it", of Christ and the earth, closing the work on a flat denial that reads oddly beside implevit totum mundum four words earlier. It is the historical/allegorical distinction the sentence makes, and it is Migne's. Rendered as printed.

5.5 Proper names and numbers rendered as printed

• 0739C · post Darium patruum cognomento Nochum — Artaxerxes III is Ochus, and cognomento Ochum → cognomento Nochum is the shape of a word-boundary slip. No [sic:]: it is a proper name in its correct case, not a non-word, and Pattern 7 renders it ("surnamed Nochus"). ⚑ This is the one other word in the book I would ask the plate about (§2.2's closing warning), because the alternative is that our TEI made it.
• 0739C · Susi fluvio and 0746D · Susis equitatio vel evertens interpretatur. The book gives Susan two incompatible etymologies eight columns apart — from a river, and from "riding/overturning". Both rendered as printed; the contradiction is Migne's and is not adjudicated.
• 0742D · Xudimos (§4c). Rendered as printed.
• 0746C · centum et viginti and the arithmetic denarius × duodenarius = 120, plus septenarius → 127. Rendered as printed; the sum is right and the Vulgate's 127 provinces is what it is reaching.

5.6 Structure and layout as printed

• CAPUT VII opens with no VERS. address at all — « Secunda die, » etc. stands bare where every other chapter opens with one. Reproduced; nothing supplied.
• Verse numbers ascend within every chapter. No out-of-sequence verse mark anywhere in this book — checked chapter by chapter and recorded as checked, unlike Jude.
• Column 0747D is absent: the sequence runs 0747C → 0748A. Migne's, or the plate's column-band arithmetic; not renumbered, not reported as loss.
• Columns 0740A and 0740B are absent from this work: the sequence runs 0739D → 0740C. This is the documented interleaving with Judith (8960), which runs to 0740B in the same columns — runbook false positive 1, not reported as a defect.
• The [nt:] at 0748C. The closing note is editorial prose — a statement that the old copies carry no exposition past this point and that chapters 11–16 are therefore omitted — and takes Pattern 15's translation, at the same marker index. The opening note ( Vide inter opera B. Hieronymi, tom. VIII.) stays [n:]: it is a cross-reference locator, not a statement, and §2.3 shows it is doing the work of a see-under on the plate. Judged by content, not by length; the two are within four words of each other.
• Migne's internal spacing in the two parenthetical notes is reproduced character for character, including the space after ( and before ).

5.7 Punctuation the plate prints and the English keeps

Test 3 (band-by-band comparison, never by total) was run against the column anchors, which partition both twins identically. Zero mismatched bands at file close — but it caught seven sites during drafting, and the totals would have hidden at least three of them:

⚠ In chunk 0 the colon totals matched at 23 against 23 and the bands matched too — that is a genuine clean, not the Tobit cancellation. In chunks 1–3 the totals were off by one in each direction and would have looked like noise; the bands named the exact sites.

⛔ No comma→semicolon sweep was performed and none was considered. Migne prints 25 semicolons in this work's four chunks (7 / 6 / 11 / 1), so Tobit's narrow exception — which rested on his printing zero in that book's first chunk — does not apply here and was not invoked. The plate was counted before the question was asked.

Kept as printed, individually:

• 0740C · a colon inside a quoted verse: « Homo quidam fecit coenam magnam: et vocavit multos ». Inside the guillemets, where he sets it.
• 0741B · a full stop inside the guillemets before the note: « Beati enim pauperes spiritu, quoniam ipsorum est regnum coelorum. [n: (Matth. V)]. » — two terminal stops, one before the note and one after. Both kept.
• 0744C · a semicolon inside a quoted verse: « Nolite solliciti esse de crastino; ».
• 0741C · Item. with a full stop where the parallel formulae read Et: with a colon.
• 0742A · Non solum regem laesit regina carries a colon after principibus inside the guillemet span, and the span opens with a whole clause of narrative before the quoted speech. Both inside, as printed.
• 0748C · historialiter, enim non omnem terram — a comma between historialiter and enim, splitting a postpositive from its clause. Rendered with the comma where he sets it.
• 0740C · -- standing alone between two guillemet spans (« Grande convivum… » -- « Simile est regnum coelorum…). Reproduced. Not an em-dash of ours and not to be reported as one.

5.8 Sentences left broken because the plate breaks them

• 0741A · In circumcisionis sacramento Dominicae resurrectionis, et Novi Testamenti et verae circumcisionis: et octo sunt beatitudines… The first limb has no verb, and the colon before et octo sunt leaves it dangling. Rendered as printed, verbless. Nothing supplied. (The argument wanted is the octave — eighth day, eight beatitudes — and the missing predicate is presumably significatur or the like. Not supplied, not conjectured: what is missing is a verb, and [cj:] requires a printed word to gloss.)
• 0741A · Magnatum vel primitivorum fidelium primi refecti sunt. Two genitives before a bare primi. Rendered "The first of the great ones, or of the primitive faithful, were refreshed." The rival binding — primitivorum fidelium as a unit, "of the primitive faithful", against magnatum… primitivorum as parallel genitives — makes no difference to the sense and was not adjudicated.
• 0746C · ita decalogi custodia apostolica traditione per totum orbem diffusa, septiformis Spiritus gratia in cordibus fidelium infunditur vel est consummata. The ita clause opens with a nominative absolute that never gets its verb, and then a second subject (gratia) takes over. Rendered as printed with the anacoluthon intact. Nothing supplied.
• 0747A · Non solum alios operarios… damnandos judicat, sed etiam carnales Judaeos… sensuros. The second limb's esse is gapped, which Latin permits and English does not; rendered by recasting the second limb as a that-clause ("but also the carnal Jews… that they shall feel"), which supplies no word and loses no clause. Logged because the asymmetry is visible in the English and is Migne's.
• 0745A · qui si videbatur gloriosior, infra alios apparet vilior. The si is odd where sense wants qui alone or quondam. Rendered with the si — "he who, if he seemed the more glorious, appears beneath others baser and lower". No [cj:]: the printed sentence construes as it stands and asserts nothing false, which is Pattern 18a's bar. ⚠ REPAIRED AT THE BLIND READ (§7.2). My first version silently dropped the si altogether, and that entry claimed it had been rendered when it had not. See §7.2.

5.9 [sic:] and [cj:] candidates considered and DECLINED, with reasons

• 0742D · Tebeeth / Xudimos — §4c. Proper names in a calendar list. Declined.
• 0739C · Nochum — §5.5. Proper name in its correct case. Declined; plate read wanted.
• 0743B · Justitiae for justitia — took a [var:] (§4a), not a [sic:]. Justitiae is a real word in a real case (nominative plural) with a real slot, so Pattern 12's boundary excludes it; what is wrong is the agreement with the received text, which is Pattern 14's business. The two markers answer different questions and only one applies.
• 0746D · rediviva plantaria — redivivus of plants is attested; not a defect.
• 0742D · competentium — a genuine technical term (catechumens enrolled for baptism). ⚠ Rendered "the competentes" in ROMAN, not italic Latin, contrary to the register rule for untranslatable technical terms, because chunk 1's Latin twin prints zero italics and an italic span here would break italic parity 1:1. The register rule and the parity rule collide and parity wins; recorded because it is the first time in this block they have collided, and a later book will meet it again.

5.10 Read clean, and named so

Chunk 3 carries no [sic:] and no [cj:] — one [var:] (I Thess.) and one [nt:], and its guillemet imbalance is Migne's (§3). CHAPTER III (0743C–0743D), CHAPTER V (0744B–0744C) and CHAPTER X (0748B–0748C) were each read with no defect found. The heavy apparatus is concentrated in chunks 0 and 1 (the banquet allegory) and at 0744D–0745A.

§6 · WHAT A BLIND READER SHOULD OPEN FIRST

Rewritten after the blind read (§7); items 1–5 of the original list have been adjudicated and are now closed. Kept in place so the record shows what was offered and what came back.

1. ✅ 0744A, Sapientia enim vincit malitiam (§4a) — offered because a reader who does not open the marker will not notice a negative is missing. Endorsed; the [var:] wording was sharpened ("reversed in what Migne prints, which also has no negative"), because "the negative is absent" did not say absent from what.
2. ✅ 0740C, atque ipsos juste condemnat (§5.3 ii) — offered as the site where the wrong reading arrives first and parses perfectly. Endorsed, and the bracket explicitly told to stay: ipsos sits nearer Judaeos, which is exactly why the wrong reading arrives first, but Christ cannot in one clause both deliver the confessors and justly condemn them.
3. ✅ 0744D, mobilis (§5.1) — endorsed. The immobile-God-contemplating-mobile-time topos is the sentence's whole point, and the loss is a single-syllable haplography after In se.
4. ✅ 0741B, Ex alibi (§5.2) — endorsed, and on firmer ground than I had: Ex alibi is not Latin at all, and Et alibi is this work's own formula, used verbatim four times in these four chunks.
5. ✅ 0747A, Et plurimum posse cognoverant (§4c) — closed clean, no defect. Basis recorded honestly in §4c: an agent's knowledge, not a witness on disk.

What is still open, and it is short:

• 0746B, Nutrivit corda fidelium (§4d) — an unidentified guilleted quotation cited as scripture with no [n:]. A reader who recognizes it closes it.
• 0739C Nochum and 0742D Tebeeth (§5.5, §4c) — two plate-read candidates, live now that §2.2 has shown the digitization can tidy Migne as well as break him.
• 0741A, the magnatum / primitivorum binding (§5.8) — deliberately not adjudicated because nothing turns on it.
• 0741D / 0742A, the origin of the three dropped « (§3) — Migne's compositor or our transcription; the plate read did not cover it.

§7 · THE BLIND POLARITY READ — a measured ZERO on the class that beat Ruth, and three repairs

A reader that never saw this file read all four chunks.

7.0 ⭐⭐ THE HEADLINE IS A NEGATIVE RESULT: the scripture-conformation hunt came back EMPTY

Every guilleted quotation was checked against the Vulgate, and at all ~15 places Migne diverges the English follows HIM, not the familiar Bible — ten carrying a [var:], the rest silently faithful. Singled out: Hab 3:8 rendered on Migne's equitatus / sanitas rather than the Douay's chariots and salvation, and 1 Thess 4 rendered on his resurgemus / residui sumus with the [var:] naming the Vulgate's rapiemur … in nubibus, called the best-executed non-conformation in the work.

⛔ Record this as a MEASURED ZERO, not as an absence of findings. Ruth — same volume, same recension, same testament — gave up four silent conformations on exactly this ground, and all four were made by a translator using [sic:] correctly two columns away. A zero here is evidence about the method, and it is the only kind of evidence that class ever yields. The reader also confirms no [sic:] swallow anywhere: every marker stripped leaves a complete sentence.

⚑ And the zero did NOT generalize. The three defects it did find are all in gloss prose, none in a lemma — the same distribution as 2 Thessalonians, where all 57 named lemmata were right and both defects sat in the sentences between them. The lemma list does its job; it does not cover the prose.

7.1 ⭐⭐⭐ 0748C · a bare "he" that inverted a theological claim — REPAIRED

Full treatment at §5.3(vi). "he" → "he [Assuerus]". As written, the page had Christ not making the earth tributary because he did not possess it, contradicting implevit totum mundum in the same sentence.

⚑ A second, smaller repair was taken in the same edit: scope negation. non omnem terram tributariam fecit negates the scope, not the act, and "he did not make all the earth tributary" reads to many as "he made none of it tributary." Now "he made not all the earth tributary." ⭐ This is the same class as the Non solum regem relocation I caught myself (§5.4) — a negation sitting on the wrong side of a quantifier — which means my own test-2 pass found one instance of it and missed another two columns apart in the same work. The rule the pair yields, and it is sharper than "check every negative": a negative next to a quantifier (omnis, solus, totus) attaches to the QUANTIFIER, and English word order must put it there.

7.2 ⭐⭐ 0745A · si dropped, and the smoothed English read as an undeclared emendation — REPAIRED

Migne prints qui si videbatur gloriosior. My English rendered no particle at all — "he who seemed the more glorious appears beneath others baser and lower."

⛔ The danger is second-order and worth naming precisely: the smoothed clause is exactly what you would get by translating qui prius videbatur gloriosior — a plausible emendation that nobody made and nothing on the page declares. A reader could not learn that Migne prints a particle there at all. Repaired to the concessive conditional: "he who, if he seemed the more glorious, appears beneath others baser and lower." No [cj:] — the sentence construes as printed, so the conjecture threshold is not met (§5.8).

⚠ And the worse half is the record, not the text. §5.8 already carried an entry on this si, and that entry asserted the particle had been rendered when it had not. A cruces line saying "rendered with the si" is exactly the kind of note a later reader trusts instead of checking. A note that a call was made is not evidence the call reached the sentence — Tobit's §7 lesson (carrying a number faithfully is not carrying a referent) in its most literal form.

7.3 ⭐ 0746B · a supplied subject left unbracketed — REPAIRED

"The king" → "[The king]" over Migne's noun-less ablative absolute. The supply was right; it was bare, in a work that brackets supplied subjects in five other places (§5.3), so the one place the noun is genuinely missing from the plate was the one place the English hid it. Established practice, no new apparatus.

7.4 ⚑ A PROJECT GAP, recorded rather than worked around

Four calls in this batch were declined or hedged for want of a Vulgate on disk, across two books — this work's Et plurimum posse cognoverant (0747A), Tebeeth (0742D), the Ps 113/134 conflation (0744B), and Tobit's Exsurgens (0731A). Two were later closed from an agent's knowledge rather than from a text we hold, which is a weaker footing than this edition's own standards want for a marker decision. Every §4 collation in this file rests on the same footing. The orchestrator is recording this as a project gap; it is not a request for anyone to keep working around it.