MIGNE.APP

Cruces

Liber Ecclesiasticus

Anselm of Laon and his school — read the work →

These are the edition's working notes, published as they were written. They record, work by work, where Migne's plate is defective or the reading uncertain: the printed form, what the English does with it, and why. They are not a groomed apparatus. Where a note argues with itself, withdraws an earlier decision, or leaves a question open, that is the record of how the reading was actually reached, and it is left standing — a conjecture that is visible can be checked, and one that has been tidied away cannot.

The English never silently repairs the page. A construable but wrong reading is rendered literally and the conjecture is recorded here, never applied to the text. How the English is produced and checked →

PL 113, cols 1183C–1232A · 18 chunks · translated 2026-08-29 by four agents (0000–0004, 0005–0009, 0010–0013, 0014–0017). This file is the MERGE of their four cruces files; the per-stint files are preserved below in full, unedited, under their own headings. Nothing has been summarised away.


⭐ WORK-WIDE CONVENTIONS — reconciled at the merge, 2026-08-29

Read this before touching the English. These are the decisions that differed between stints and were settled work-wide. A later sweep that "fixes" one of them will be re-introducing the drift this section exists to remove.

1. Migne's elision formula usque adas far as (settled; 135 occurrences)

The stints split cleanly down the middle: 0000–0009 rendered it "down to" (48×) and 0010–0017 "as far as" (88×) — one work, two renderings of the same editorial formula, which is precisely the failure the merge step exists to catch (cf. 11551's stylus/"epistyle" split, 11063's ALLEGORICE).

Resolved to "as far as", because it is the form already shipped in the sister Glossa books — 8976 (Acts) and 8985 — and this is an editorial formula, not a lemma. The "each commentary stands on its own" ruling (do not harmonise a LEMMA across works) does not reach it: harmonising Migne's own abridgement mark across the Glossa is desirable, and one voice per work is required regardless.

⚠ Two genuine prose uses of "down to" survive and are CORRECT — "despising to be bowed down to earthly things" and "come down to us from above". They carry no asterisk. Any future sweep must match the italic formula (down to / *down to:), never the bare words.

⚠ One site, 0003 @1191D, prints the formula with no closing asterisk — as far as He taketh seven other spirits… — because the Latin does the same (usque ad Assumit septem alios spiritus nequiores se): the italic runs on through the quoted continuation. That is mirrored deliberately. Do not "repair" it.

2. impiusungodly (settled; 10 Latin occurrences)

Three renderings were in use across the work: "impious" (0005, 0006, 0008), "wicked" (0006's Ps. 1 quotation, 0013), "ungodly" (0014, 0015). Normalised to ungodly, the form shipped in 8985 (2/2).

The other occurrences of "wicked" in this work are CORRECT and must not be swept: they render nequam, iniquus, nequior and malitia, not impius — e.g. "wicked servants" @1217C, "the offerings of the wicked" @1218D, "more wicked than himself" @1191D. The normalisation was applied site by site against the Latin, never by find-and-replace on the English.

3. Non-words: carried through, per Pattern 7 — NOT rendered for the evident sense

A live contradiction in the corpus, flagged not swept. The 0005–0009 stint raised that the sister Glossa work 8976 ruled work-locally that non-words are rendered for their evident sense and not carried, firing no marker on six sites — the opposite of translation-style.md Pattern 7, which Wilson ruled on 2026-07-28: "carry it through… Pattern 7 has no exceptions."

8947 follows Pattern 7. The markers stand; nothing was unwound. 8976's departure is recorded here for a ruling, and is NOT retrofitted (the 2026-08-18 no-retrofit rule).

4. Vocabulary held apart deliberately — do not collapse these

kept apartwhy
Libanus the mountain → "Lebanon" vs Libanus the gum → kept Latinthe gloss @1209D is an etymology of the word (in Hebraeo candorem significat: in Graeco, θοῦς); "frankincense" would leave it explaining nothing. Note ad collem thuris IS "the hill of frankincense" — a different word, thus.
stultus → "fool" vs fatuus → "senseless man"CAPUT XXI prints both as lemma-words four verses apart (Fatuus in risu 21:23, Stultus a fenestra 21:26); collapsing them erases a distinction the plate makes.
ungula (24:21) kept Latinthe gloss IS Graece ὀνύξ, Latine ungula — "onycha" makes it circular.
litare / sacrificare (35:3) kept Latinthe whole gloss is the distinction between them.
praecordia → "inward parts" vs cor → "heart"both stand in one gloss @1216D.
scientia "knowledge" vs sapientia "wisdom"; charitas "charity" vs dilectio/amor "love"inherited from 8976 §V.
acharis (20:21) kept Latin inside the italic lemmaMigne's gloss is its etymology.

Agreed across all four stints, no reconciliation needed: doctorteacher (never "doctor" — 0 occurrences work-wide), praedicator → "preacher", CAPUT N. → "CHAPTER N.", VERS. N.-- kept verbatim, virtus decided per occurrence (power of God, virtue where moral), second person per Pattern 17 (tu thou/thee/thy, vos you).

5. Greek

All 8 Greek words are reproduced in Greek script with accents intact, none transliterated: φοινίκον · ἐλαἴον · ἐλαία · πλατός · ἄκανθον · στακτήμ · ὀνύξ · θοῦς. They fall almost entirely in CAPUT XXIV (chunk 0010).

6. Plate status — ⭐ THE GATE HAS BEEN RUN (2026-08-29)

All 22 pages bearing a marker were read at Migne's plate, covering all 69 original markers. Scan raw/scans/pl113/patrologiaecurs04migngoog.pdf, PDF page = (column + 11)/2, corner numbers read BEFORE the text on every page, every site re-verified at 600 dpi. Reads recorded in data/plate-reads.json; node scripts/plate-gate.mjs 8947 passes.

Our scan CLIPS THE LEFT MARGIN inside this work (pp. 598, 604, 620). Left-column sites were settled at the Gallica second witness ark:/12148/bpt6k5505319w, f = (column + 1)/2.

Result — and it ran hard in both directions:

[var:]56 of 56 CONFIRMED on the plate. Not one was withdrawn.
[sic:] standing4magnituditudinem 1184C, obsentiamus 1197A, reconciliatione 1207A, erudiunmultos 1210D
[sic:] WITHDRAWN as false9 — every one was a Corpus Corporum corruption we were about to blame on Migne
[sic:] newly owed4 — defects of Migne's own type that CC had silently repaired, invisible to every check we run

Nine of thirteen [sic:] markers were false accusations against Mignemisericodia (plate: misericordia), ex (plate: et), Possuntt (plate: Possunt), Ad monitio (plate hyphenates Ad-/monitio = Admonitio, the 8946 rejoin class), Domini (plate: Domino), cum (plate: eum), nunde (plate: nunc/de), jurejarando (plate: jurejurando), JesusChristus (plate: two words). All withdrawn; TEI patched.

And four of Migne's own defects had been hidden from us by CC's silent tidying: loquendeum 1190C, descretur 1199D, interrogatia 1200C, orationm 1202C. Restored by TEI patch and now carried with a [sic:] per Pattern 7. ⚑ Note what this means: the translating agents could not have found these, because our Latin already read correctly. Only the plate shows them.

All 13 corrections are in data/tei-patches/8947.json, which runs in both directions.



Agent batch: chunks 0000–0004, cols 1183C–1196C (CAPUT I – CAPUT XI). No foot-of-page conjecture apparatus is available for this work (plate spot check 2026-08-29 records none on the pages read); every reading below is rendered as printed and conjectured only here.


A. Carried type and non-construing readings (Pattern 7 / 12)

### 1184C — magnituditudinem Printed: vel magnituditudinem bonitatis ejus et misericordiae. Non-word (dittographic syllable). Carried as [sic: magnituditudinem]. Conjecture: magnitudinem, near-certain. The clause keeps its governing verb (posse … penetrare) and its negation (neminem) outside the marker.

### 1184D — bracketed supply [non est in charitate] inside the 1 John 4:18 lemma Printed: Timor [non est in charitate] animae in charitate, [n: (Joan. IV)], etc. The square brackets are in the source, not ours. The bracketed words are the Vulgate's (1 Jn 4:18 Timor non est in charitate); what remains outside them, animae in charitate, does not construe with them. Rendered with the bracket preserved and its contents translated (they are scripture, not an al. variant). Not marked [sic:]: the type is legible and the bracket looks editorial. ⚠ Candidate left-margin clip: this sits at the head of a column band. Plate check wanted.

### 1187D–1188A — Cui Deum honorem constituit Printed: Cui Deum honorem constituit, non debet a filiis inhonorari. Rendered literally as a double accusative ("He for whom he appointed God as his honour…"). Conjecture: Cui Deus honorem constituit, "for whom God appointed honour" — which is what the following clause (quia Deus judicio suo exquiret) expects. Not emended in the text.

### 1189C — Per fores autem potes Scripturas et eorum doctores intelligit. Second-person potes and third-person intelligit in one clause; it will not construe. Rendered as closely as the printed words allow ("And by the doors thou canst, he understands the Scriptures and their teachers"). Conjectures: potest … intelligi, or potes … intelligere. Both are real words, so nothing is carried under Pattern 12.

### 1196B — nec immunes sunt a dilecto dilecto is a real word (abl. of dilectus), so it is rendered, not carried: "nor are they free from the beloved." Conjecture: delicto, "from fault" — which is what quia meditatio eorum est in dolo in the same sentence requires. Logged, not repaired.

### 1186A / 1194C — Hypocrita.. and VERS 37.-- Printed punctuation preserved: the double stop after Hypocrita (1186A) and the missing stop after VERS at 1193C. Pattern 8, not defects of ours.


B. Vulgate divergences marked [var: …] in the text

colMigne printsVulgatenote
1184Cin mandatis ejus cupit nimis (Ps 111:1)volet nimisverb changed; the gloss does not turn on it
1184DTimor Domini sanctus permanet in saeculum saeculi (Ps 18:10)permanens⚠ the SAME verse is printed correctly as permanens at 1187C, four columns later — internal evidence that 1184D is the defective setting
1185BNon amat pestilens eum qui se corrigit (Prov 15:12)corripit, "him that rebuketh him"reflexive vs. transitive; changes who is corrected
1185Cqui autem fatuus est aperuit stultitiam (Prov 13:16)aperietperfect for future
1185C–Dostendat conversationem bonam in mansuetudine sapientiae (Jas 3:13)ostendat ex bona conversatione operationem suamthe object of ostendat is different
1186DBeatus qui confidit in Domino, et erit Deus fiducia ejus (Jer 17:7)Benedictus vir qui confidit in Domino, et erit Dominus fiducia ejusopening and divine name both diverge
1186DQui autem dereliquerint Deum, confundentur et in terra scribentur (Jer 17:13)omnes qui te derelinquunt confundentur; recedentes a te in terra scribentursecond-person address of the Vulgate lost
1187APrope est Deus omnibus invocantibus se in veritate (Ps 144:18)Prope est Dominus omnibus invocantibus eumdivine name; cf. the timor/amor risk class
1188Bmaxime qui laborant in doctrina et verbo (1 Tim 5:17)in verbo et doctrinaword order reversed
1188Dthesaurizas tibi iram in die judicii (Rom 2:5)in die irae
1189AVade et revertere, cum cito possis dare (Prov 3:28)cras dabo tibi: cum statim possis dare
1189A–Bclamabit, et non exaudiet Dominus (Prov 21:13)et ipse clamabit, et non exaudieturpassive turned active with a supplied subject; the negation is present in both and is carried
1189Bsubstantiam mundi … fratrem suum necesse habentem (1 Jn 3:17)substantiam hujus mundi … necessitatem habere⚠ the SAME verse is printed with hujus mundi … necesse habere at 1193C
1190ANegotia, dum venio (Lk 19:13)NegotiaminiNegotia is a real (if rare) active imperative, so it is rendered, not carried
1190Cstultitia magis speranda est, quam correctio illius (Prov 29:20)quam illius correptioword and order
1191Alemma Amicus fortis (Sir 6:14)Amicus fidelis protectio fortisthe adjective has migrated onto amicus; the gloss at VERS. 16 then repeats Amicus fidelis
1191Blemma Qui timet Dominum (Sir 6:17)Qui timet Deum⚠ MODERATE confidence — verify on the plate
1191Cut cognoscant te solum verum Deum (Jn 17:3)solum Deum verumword order
1192ANon est homo super terram, qui faciat bonum (Eccl 7:21)Non est enim homo justus in terrajustus absent from Migne
1192Aet magis illo spem habet stultus (Prov 26:12)magis illo spem habebit insipiens
1192Blemma Noli amare mendacium (Sir 7:13)Noli arare mendacium⚠ the strongest single finding in this batch: the Vulgate's odd "plough a lie" has become "love a lie", and the gloss (Omne mendacium malum) is compatible with either, so nothing on the page flags it
1194Ablanditiis labiorum traxit eum (Prov 7:21)protraxit illum
1194DPraecipe divitibus non altum sapere (1 Tim 6:17)Divitibus hujus saeculi praecipe non sublime sapere
1195DHomo enim videt ea quae patent (1 Sam 16:7)quae parent, "the things that appear"both are real words; the sense shift is small but real

C. Divergences observed and NOT marked (logged only)

Truncated lemmas are not listed; only mid-lemma or mid-quotation differences judged too small, or too uncertain against the recension this gloss follows, to carry a [var:].


D. Notes on structure (no action)


Stint: chunks 0005, 0006, 0007, 0008, 0009 = CAPUT XII–XXIII, cols 1196C–1208B. Latin words 4,531. verify-english.mjs clean; column anchors and [n:] notes at exact 1:1 parity with the Latin twins in all five chunks; zero em-dashes, zero guillemets (the Latin twin prints none).

Apparatus fired in this stint: 17 [var:] · 4 [sic:] · 0 [cj:] · 0 [d:] · 0 [ed:] · 0 [nt:] · 0 [f:]. Every one is listed below with its column for the plate check.

No editorial apparatus was available. The 2026-08-29 spot read of this work (pp. 597–598 = cols 1183–1186) found no foot-of-page conjecture notes, and the spot coverage does not license extending that to cols 1196–1208. Nothing below was settled against Migne's own conjecture, because there is none in hand. The left-margin clipping confirmed at p. 598 did not recur in any column of this stint: no lemma or gloss in 0005–0009 has lost word-openings at a column start.


§V · LOAD-BEARING VOCABULARY (for the merge)

Fixed at first occurrence and never varied inside this stint. Where the peer stint (0000–0004) had already committed, I inherited from it rather than re-deciding.

LatinEnglishnote
usque ad (in Migne's abridged Raban excerpts)down toinherited from the peer stint 0000–0004. The sister Glossa work 8976 (Acts) uses as far as; within 8947 the peer's form governs.
CAPUTCHAPTER
VERS. N.--kept verbatim, unanglicized
doctrinateachingnever "doctrine", even in sana doctrina → "sound teaching" (1199A). Peer stint uses "teaching" at 1189C.
doctorteacherofficium doctoris = "the office of a teacher" (1197B); sancti doctores = "the holy teachers" (1206A). Peer uses "teachers" for doctores at 1189C.
stultusfool / foolish
fatuussenseless (man)held apart from stultus deliberately: CAPUT XXI prints both as lemma-words within four verses (Fatuus in risu 21:23, Stultus a fenestra 21:26) and collapsing them would erase the plate's own distinction.
insipienssenseless only in the fixed Ps. 48 tag, rendered "beasts without understanding" (1201A) to avoid colliding with fatuus
sensussense / understanding"understanding" where it means the mind's grasp, "sense" where it means the meaning of a text.
haereticus / catholicusheretic / catholic (lower case, as Migne prints)
Ecclesiathe Church (capital, mirrored from the plate)
Synagogathe Synagogue
eleemosynaalms / almsgiving
timor Dominithe fear of the LordPs. 110:10 at 1202D checked word-for-word: Migne prints timor, not amor. So does every other occurrence in this stint.
Mystice / Allegorice / Historice / Moraliter / Tropologice / ad litteramMystically / Allegorically / Historically / Morally / Tropologically / according to the letteradverbs, following the plate's adverbs (Pattern 9 precedent, 11057).
casula (Ecclus. 14:25)"little dwelling"no gloss depends on the word.
loramentum (Ecclus. 22:19)"binding" / "wooden binding"the gloss expounds ligna and fundamentum, not the binding itself.
acharis (Ecclus. 20:21)kept as Latin inside the italic lemmaMigne's gloss IS the etymology (Sine gratia Spiritus sancti; vel amarus) — case 2 of the "each commentary stands on its own" ruling. No second italic marker inside an italic lemma (Pattern 10).

⛔ §1 · THE VULGATE COLLATION — every lemma and every quotation, collated word by word

Collated against the on-disk Clementine text (sources/vulgate/clementine/), not from memory. Truncation is not divergence and no short lemma is reported below. Roughly 130 lemmata and inline quotations were collated; the ones that agree are not listed.

§1a · MARKED [var:] — 12, with columns for the plate check

colMigne printsVulgatewhy it is material
1197BSi enim sal infatuatum fuerit, ad nihilum valet ultra (Matth. 5:13)Quod si sal evanuerit, in quo salietur? ad nihilum valet ultradifferent verb, and the Vulgate's question is absent entirely. The gloss applies it to the heretic-teacher who claims to be a physician of souls.
1198DOperamini dum dies est: venit enim nox, quando jam non licet operari (Joan. 9:4)Me oportet operari opera ejus qui misit me, donec dies est: venit nox, quando nemo potest operariMigne turns a first-person statement into a plural imperative and the argument rests on the imperative ("he exhorts that we offer").
1199BQui biberit aquam quam ego dabo vobis, fiet in eo fons aquae salientis in vitam aeternam (Joan. 4:14)qui autem biberit ex aqua quam ego dabo ei, non sitiet in aeternum: sed aqua quam ego dabo ei, fiet in eo…conflated, and plural vobis against Vulg. singular ei, with fiet in eo singular left standing after it.
1199Dquoniam justus Deus, et justitias dilexit (Psal. 10:8)quoniam justus Dominus, et justitias dilexit⭐ a divine-name swap, and this work quotes the same verse correctly at 1202C (Justus Dominus, et justitias dilexit: aequitatem vidit vultus ejus). Two printings of one verse, four columns apart, disagreeing.
1199DA tribus impiis deseretur (Ecclus. 16:5, lemma)tribus impiorum deseretur⭐⭐ the strongest lemma finding in the stint. Vulg = "the tribe of the impious shall be made desolate"; Migne = "by three impious men it shall be deserted". The gloss expounds Migne's reading, not the Vulgate's: qui cogitatione, locutione, opere seipsos sequestrant — thought, word, deed, three of them. 7a″ exactly: the exposition rests on a word that is not in the received text.
1200Binterrogatio hominum (Ecclus. 16:22, lemma)interrogatio omniumthe gloss then names the classes of men (idolaters, heretics, false catholics), following hominum.
1201BStatuit terminos gentium juxta numerum angelorum Dei (Deut. 32:8)constituit terminos populorum juxta numerum filiorum Israel⭐⭐ the LXX/Vet. Lat. reading, and the whole gloss depends on it: the lemma is Rectorem and the gloss reads it as Angelos, quibus commisit Deus custodiam singularum gentium. On the Clementine reading the proof-text proves nothing.
1201CCurrite dum lucem habetis, ne tenebrae vos comprehendant (Joan. 12:35)Ambulate dum lucem habetis, ut non vos tenebrae comprehendantdifferent verb; the urgency ("Run") is the point of the gloss (non est tempus operandi, sed praemium recipiendi).
1202BSermo durus suscitat rixas, lenis mitigat suscitatas (cited Prov. XV)Prov. 15:1 sermo durus suscitat furorem; Prov. 15:18 qui patiens est mitigat suscitatasa conflation of two verses four columns apart in Proverbs, with rixas for furorem and lenis for qui patiens est. Cited again at 1207B in the same conflated form, so it is Migne's stable text here, not a one-off slip.
1203DQui enim scit voluntatem domini et non facit, vapulabit paucis (Luc. 12:47)…vapulabit multis⭐⭐⭐ the most consequential finding of the stint. The Vulgate gives multis to the servant who KNEW and paucis to the one who did not; Migne's 1203D gives paucis to the one who knew, inverting the verse, and then glosses it exiguo enim conceditur misericordia — building an argument on the inverted reading. And this same work prints the verse correctly at 1206C: Servus enim qui scit voluntatem domini sui, et non facit: vapulabit multis. Qui autem nescit, paucis. Three columns apart, the plate contradicts itself. Rendered as printed at 1203D per Pattern 7; the [var:] carries the received reading. Plate-check both 1203D and 1206C together.
1204CVoca pauperes et debiles, caecos, claudos, et beatus eris (Luc. 14:13)voca pauperes, debiles, claudos, et caecos: et beatus erisword order of the last pair reversed, plus et inserted after pauperes. Reported because the pair is a fixed list.
1205BBonus homo de bono thesauro cordis sui profert bonum (Matth. 12:35)Bonus homo de bono thesauro profert bonacordis sui imported from Luc. 6:45; singular bonum for plural bona.
1205CAdversarius noster diabolus (I Petr. 5:8)adversarius vesterfirst person for second. A person-of-pronoun swap has a full English exponent, which is why it is marked and the eum/eam case below is not.
1206DSicut paleae, etc. (Ecclus. 22:21, lemma)Sicut pali in excelsis, et caementa sine impensa posita⭐⭐ paleae (chaff) for pali (stakes) — a completely different image, and the gloss follows Migne by supplying the house built on sand (Matth. 7:26) rather than stakes on high ground.
1207DAnima callida (Ecclus. 23:22, lemma)Anima calida quasi ignis ardens⭐⭐ callida (crafty) for calida (hot), one letter. Migne's own gloss contradicts his own lemma: it reads ardentem animam habent — a burning soul — which is calida. So the plate diverges from the Vulgate in the lemma and agrees with it in the gloss, in one entry.
1208AHaec in Ecclesia (Ecclus. 23:34, lemma)Haec in ecclesiam adduceturablative for accusative, and the verb is gone. The gloss reads it locatively (in conventu fidelium… declaratur), so the divergence is what is expounded.

(All sixteen rows carry a [var:] in the text — 1197B, 1198D, 1199B, 1199D ×2, 1200B, 1201B, 1201C, 1202B, 1203D, 1204C, 1205B, 1205C, 1206D, 1207D, 1208A — and a seventeenth sits at 1203B (Ecclus. 19:7 verbum durum et nequam for Vulg. verbum nequam et durum), listed with the unmarked divergences in §1b because the pair's order is all that moved. 17 [var:] in all.)

§1b · COLLATED DIVERGENCES DELIBERATELY LEFT UNMARKED

Real differences from the Clementine text that fire no marker, because English has no exponent for them or the sense does not turn on them. Listed so that a later sweep does not re-derive them, and so that the absence of a marker is itself on the record.

§1c · MIGNE'S OWN [n:] CITATIONS THAT POINT TO THE WRONG PLACE

Reproduced verbatim and untranslated in every case, per the note rule. Recorded here only so the indexer's refKey resolution has the true target, and so nobody proposes a text change. ⚠ None of these is a defect claim.


§2 · MARKED DEFECTS — 3 [sic:], with columns

⭐ 1207A — reconciliatione : a FUSED word-division defect that eats the sentence's negation

Migne prints, at Ecclus. 22:27, in the gloss:

Effugiet amicus. Non rediturus, nec post reconciliationem; unde: Amico reconciliatione credas in aeternum.

The maxim as printed says "trust a reconciled friend for ever" — the exact opposite of the gloss it is adduced to prove, whose own words are Non rediturus, nec post reconciliationem ("not to return, nor after reconciliation"). The received sententia is Amico reconciliato ne credas in aeternum. Migne's reconciliatione is reconciliato + ne run together, the fused direction of Pattern 10 — and the ne it swallows is the clause's only negation.

Treatment, and it is the pair of rules working together:

This is precisely 7a⁗-b's dangerous shape: the quarantine is the size of a lexeme, and the lexeme was carrying the negative. Marking the run and letting the sentence go affirmative would have printed, on a reading page, the reverse of what the plate's own next words argue. Plate-check 1207A.

1197B — obsentiamus : a non-word carrying the clause's verb

Non ergo eis obsentiamus, nec potestatem tribuamus. obsentiamus is not a Latin word; the conjecture is assentiamus ("let us not assent to them"), which the dative eis and the coordinate nec potestatem tribuamus both support. Carried untranslated in italics with [sic: obsentiamus] per Pattern 7's non-word class and Pattern 12. 7a⁗ checked: the negation Non is outside the marker and stands in the English, and the coordinate clause supplies a predicate, so the sentence is not left verbless.

Flagged for the merge as a live disagreement of practice, not as a settled call. The sister Glossa work 8976 (Acts) ruled in its own cruces, work-locally, that non-words are rendered for the evident sense and not carried, and fired no marker on six such sites. I have followed translation-style.md (Pattern 7's non-word ruling, Wilson 2026-07-28) instead, because that is the file my brief names as authoritative. If 8947 wants 8976's practice, this marker and the ones at 1203D and 1205A are the three to unwind — and the unwinding is mechanical, since both cruces entries carry the conjecture. Do not unwind them silently.

1203D — misericodia : a non-word carrying the clause's SUBJECT

Qui enim scit voluntatem domini et non facit, vapulabit paucis: exiguo enim conceditur misericodia. misericodia is misericordia with its r dropped — a single missing letter that leaves a form which is not a word. Carried untranslated in italics with [sic: misericodia] (Pattern 7 / Pattern 12).

7a⁗ checked, and it changed the rendering: the marked run is the clause's subject, so quarantining it alone would have left an English verb with nothing to do it — "for to the small is granted." The sense is therefore supplied beside the marker, in apposition and in plain English: for [sic: misericodia], mercy, is granted in small measure. The verb, the subject and the sense all survive outside the quarantine.

This is the same column as the Luke 12:47 paucis finding (§1a). One column carrying both an inverted scripture quotation and a broken word is worth reading as a whole. Plate-check 1203D.

1205A — ex : a real word with no grammatical slot

Homo acharis… vel amarus, quia amarum ex noxium sensum ingerit auditoribus. ex governs the ablative and cannot stand between two accusative adjectives agreeing with sensum; the conjecture is et. Marked [sic: ex] under Pattern 12's first test — a real word that has no grammatical slot cannot be carried by any English — and both adjectives are rendered around it, so nothing is swallowed. Plate-check 1205A.


§3 · RENDERED LITERALLY, NO MARKER — the Pattern 7 cases

Each of these is a reading that does not construe, or barely construes. None of them is repaired in the English. There is no Migne conjecture to consult (see the plate note above), so each is rendered as closely as the printed words allow and left here.

⭐ 1203C — ictum : the clause's object is the wrong noun, and the English is nonsense

Stultus enim perturbat sensum, ut gemitum doloris exprimat per actum nefandum: sed non conturbat ictum quidquid acciderit.

Rendered as printed: "but whatever shall have happened disturbs not the blow." The sense the passage plainly needs is a person — the wise man of the opening sentence (Non aequaliter stultum et sapientem commovet stultum verbum), and the next clause is about illi qui de firmitate cordis confidit. The likeliest conjecture is justum, "the just man" (ictum / justum, an easy misreading); sapientem would also serve the sense but is not close to the type.

Deliberately NOT marked [cj:]. Pattern 18 fires where the faithful English misleads — where a reader takes away a false claim and never knows. "Disturbs not the blow" misleads nobody; it is visibly broken, and a reader who meets it knows to look. Firing [cj:] here would put our conjecture on the page for a case that does not need protecting, which is the spam direction of that marker. ⚠ Note also that the negation non is printed, is carried, and is not what is wrong with the sentence.

1196D — a verbless fragment after the Amos citation

Dubiam et dolosam doctrinam, quam in expositione divinae Scripturae expendunt.

An accusative noun phrase with a relative clause and no governing verb — rendered as the fragment it is: "A doubtful and deceitful teaching, which they expend in the exposition of Holy Scripture." It reads as a stranded gloss on a lemma that is not printed (Ecclus. 12:4 territory). 7a⁗ question 1 asked and answered: nothing was quarantined here, so nothing was swallowed; the verblessness is the plate's. Rendered, not repaired. Worth a plate read at 1196D — if anything is lost before Dubiam, it is source loss and would take an [ed:], not a [sic:].

1198D — cruciendam for cruciandam

Gerundive of cruciare in -endam rather than -andam. Rendered normally ("to be tormented"), no italics, no marker: first/third-conjugation gerundive confusion is an attested medieval orthographic habit, and Pattern 9 explicitly reserves carry-through for forms unique to this plate. Recorded so a later sweep does not churn it.

1206C — a sentence that simply stops

The Raban excerpt at Ecclus. 22:10 ends …mors enim corporis, finis laboris mundani with no terminal mark at all. Pattern 8: never supply. The English stops where the plate stops, without a period. Not a defect of ours.

1201C — postquam, which does construe

In praesenti vita, postquam non est tempus operandi, sed praemium recipiendi. Read as post quam ("after which"), which construes and matches the parallel argument at 1198D (post mortem enim… non est tempus operandi, sed mercedem recipiendi). Checked and dismissed as a non-defect — recorded because it looks like one at first sight.


§4 · CHECKED NON-DEFECTS — do not spend a plate read on these


§5 · FOR THE MERGE

  1. §V is the vocabulary table. The only term where I could see the peer stint had already committed is usque ad → "down to", and I inherited it. doctrina/doctor were checked against 0002 and agree.
  2. The 8976-vs-Pattern-7 non-word split (§2) needs a ruling, and it is a two-marker change either way.
  3. 1203D and 1206C must be plate-read as a pair — the same Luke 12:47 quoted two ways three columns apart is either the most interesting thing in this work or a transcription artifact, and only the plate can say which.
  4. 1199D and 1202C are the same pair-shape for Ps. 10:8 (justus Deus / justus Dominus).
  5. Nothing in this stint was left untranslated and no chunk was abridged. Five files, five complete chapters' worth of glosses, all lemmata collated.

Agent batch: chunks 0010–0013, cols 1208C–1219C (CAPUT XXIV – CAPUT XXXVI). 1,081 + 1,138 + 1,024 + 1,108 Latin words. Note parity 28 / 27 / 22 / 25 exact; column anchors 11 / 11 / 11 / 12, all verbatim and in order.

Apparatus fired in this batch: 8 [var:] · 1 [sic:] · 0 [d:] · 0 [ed:] · 0 [nt:] · 0 [cj:] · 0 [cn:]. Zero em-dashes. Zero guillemets.

No foot-of-page conjecture apparatus is available for this work (the 2026-08-29 plate spot check, cols 1183–1186, records none on the pages read, and coverage is "spot"). Every reading below is rendered as printed and conjectured only here. ⚠ The same read confirms the scan clips the left margin inside this work; nothing in cols 1208C–1219C looked clipped to me, but §A.7 and §C.3 are the two candidates if a plate read is ever run.


§V · Vocabulary fixed in this batch — for the merge

Inherited from 8976 §V (Glossa on Acts, the sister book) without change: doctorteacher (never "doctors"), praedicatorpreacher, charitascharity vs. dilectio/amorlove, scientiaknowledge vs. sapientiawisdom, usque adas far as, Ecclesia capitalization mirrors the plate, virtus decided per occurrence. VERS. N.-- and ## CHAPTER N. follow 8976's shipped format exactly, including whether the head carries a final stop (CAPUT XXIV does not; XXV–XXXVI do).

Fixed here, at first occurrence, and held across all four chunks:

LatinEnglishwhy
Libanus (mountain, 24:17; sub Libano monte, 24:36)Lebanonit is the mountain, and the gloss says so (mons est Phoenicis).
Libanus (the gum, 24:21)Libanus, kept Latinthe gloss is an etymology of the wordin Hebraeo candorem significat: in Graeco, θοῦς — and "frankincense" would leave that etymology explaining nothing. Held apart from the mountain deliberately: a later sweep collapsing the two destroys the distinction the plate makes. Note the SAME gloss renders ad collem thuris (Cant. 4:6) as "unto the hill of frankincense" — that is thus, a different word.
ungula (24:21)ungula, kept Latinditto: the gloss is Ungula est quae in Exodo onycha dicitur … Graece ὀνύξ, Latine ungula. Rendering the lemma "onycha" or "nail" makes the gloss circular.
virtusper occurrencepower at 1208B (virtutis illius, of God), 1208C (Dei virtute, potentia divinitatis); virtue everywhere moral (1208C altioribus … virtutibus, 1216C quatuor principalium virtutum, 1218A aedificium virtutum). Both stand in this range.
timor Dei / timor Dominithe fear of God / the fear of the Lord, mirroring the plate word for wordsee §B.0.
praecordia (33:5)inward partsheld apart from cor → heart, which stands two clauses later in the same gloss (Cor stultorum).
litare / sacrificare (35:3)kept Latin, italicthe whole gloss is the distinction between them; there is no English pair that carries it.
caprea / ibices / avices (27:22)roe / ibices / avicescaprea is glossed by a Latin pun (quasi capiens ardua) which is reproduced in Latin beside the English; ibices and avices are the names the gloss is explaining, so they stay. avices is not standard Latin but is the form Isidore's etymology needs (ad instar avium) and the gloss states that etymology — so it is rendered as a name, not marked.
eleemosyna (29:16–18, 31:28)alms / almsgivingon the brief's risk list; both occurrences are in glosses, not lemmas.

§B · VULGATE COLLATION — every lemma and inline citation, cols 1208C–1219C

Collated word by word against sources/vulgate/clementine-flat.txt (Clementine). Truncation of a lemma is not divergence and is not listed. Migne's coelum/caelum orthography is not listed. Verse-number mismatches are §C.

### B.0 — the timor risk class: CLEAN in this batch Every occurrence of the phrase in cols 1208C–1219C prints timor, never amor: 1211C Timor Dei (lemma, Sir 25:16, agrees with Vulg) · 1213C Qui Deum timet · 1214A quasi semper praesentem pertimescit · 1216C Catholicus enim Deum timens · 1219B mulier timens Deum (the one divine-name divergence, B.2 below). No Amor Domini.

B.1 — marked [var: …] in the text (8)

chunkcolMigne printsVulgatewhy it is marked
00101209BQuasi platanus dilatata sum in plateis (Sir 24:19)quasi platanus exaltata sum juxta aquam in plateisthe gloss turns on the divergent word. Migne introduces the quotation with Nomen ergo, et formam arboris expressit dicens, and closes the paragraph Christum significat, qui per membra sua in orbe dilatatur. The Vulgate's exaltata would leave both the introduction and the application explaining nothing; juxta aquam is simply absent.
00101210Dlemma Ego quasi fluvius Dorix (Sir 24:41)ego quasi fluvii dioryx⭐ two changes, and the gloss builds on both: nominative fluvius for genitive fluvii, and Dorix for dioryx (Gk διῶρυξ, "channel"). Migne's exposition ends unde tertia lingua Graecorum Dorica dicitur — i.e. it has read the channel-word as the Doric dialect. The whole gloss is downstream of the divergence.
00111211BEt ecce timor Domini ipse est sapientia, et recedere a malo intelligentiam (Job 28:28)Ecce timor Domini, ipsa est sapientia; et recedere a malo, intelligentiaintelligentiam is accusative where the construction needs a nominative predicate; the clause does not construe as printed. Latin case has no English exponent (Pattern 9), so the English reads "understanding" and the marker carries the difference. ipse agrees with timor rather than sapientia — grammatical either way, but it is not what the Vulgate prints.
00111212AStatim eam sequitur quasi bos ad victimam, et quasi agnus lasciviens, et ignorat quod ad victimam pertrahatur (Prov 7:22)quasi bos ductus ad victimam … et ignorans quod ad vincula stultus trahatur⭐ the second half of the verse is a different sentence: Migne repeats ad victimam where the Vulgate has ad vincula stultus, so the printed proof-text has lost the fool and the bonds. Sense changes; marked.
00111213ANon laudes ante mortem quemquam (Sir 11:30)Ante mortem ne laudes hominem quemquamprinted negation swapped (non for ne), hominem absent, order reversed. Marked because the brief names printed negations as the class to surface; the sense is unchanged.
00131217Clemma Si est tibi fidelis (Sir 33:31)Si est tibi servus fidelis⭐ the noun the lemma is about is missing from the lemma, and the gloss immediately supplies it: Boni servi legem tradidit. Not truncation — fidelis follows.
00131219Alemma Extolle adversarium (Sir 36:9)Tolle adversariumthe strongest single finding in this batch. Tolle adversarium, et afflige inimicum is a prayer that the adversary be taken away; Migne prints Extolle, "exalt". The gloss (Significat per adventum Christi … pro statu Ecclesiae et fidelium salute) is compatible with either and flags nothing.
00131219Bmulier timens Deum ipsa laudabitur (Prov 31:30)mulier timens Dominum, ipsa laudabiturdivine name. Marked on the strength of the brief's timor Domini risk class and the peer batch's precedent at 1186D / 1187A / 1191B, where the same substitution recurs in this work.

B.2 — divergences NOT marked in the text (logged only)

Pattern 14's gate is "differs … in a way that changes the sense". These do not clear it. They are recorded so the collation is complete and so a later sweep does not re-find them as discoveries.

chunkcolMigne printsVulgateclass
00101210ADocete omnes gentes … docentes servare omnia (Mt 28:19–20)docentes eos servarepronoun dropped
00101210CNemo novit Filium nisi Pater, nec Patrem quis novit (Mt 11:27)nequeconnective; both negative, both carried
00101210DAqua sapientiae salutaris potavit illum (Sir 15:3)potabitperfect for future
00101210DFlumina de ventre ejus fluent aquae vivae (Jn 7:38)agrees exactlyno divergence — but the note reads [n: (Job. VII)] for John 7. Reference only; carried verbatim per the note rule and the standing "do not correct a reference" ruling.
00101210Dfulgebunt sicut splendor firmamenti (Dan 12:3)quasiand autem absent from the opening. See §A.1 for the fused run in the same verse.
00111211Asicut et Christus Ecclesiam (Eph 5:25)sicut et Christus dilexit Ecclesiamverb of the second clause dropped; ordinary ellipsis in citation
00111211ASi duo consenserint ex vobis, fiet illis (Mt 18:19)si duo ex vobis consenserint super terram, de omni re quamcumque petierint, fiet illisword order (the rest is truncation)
00111211BQui autem meditatur in lege Domini, dabit fructum in tempore suo (Ps 1:2–3)conflates in lege ejus meditabitur (v.2) with quod fructum suum dabit in tempore suo (v.3)loose quotation from two verses
00111211BCorde creditur (Rom 10:10)Corde enim crediturparticle
00111212CQui servat mandata mea, ille est qui diligit me, et diligetur a Patre meo (Jn 14:21)Qui habet mandata mea, et servat ea: ille est qui diligit me. Qui autem diligit me, diligetur…compression
00111213CDimittite et dimittetur vobis (Lk 6:37)Dimitte, et dimittemininumber and voice; sense identical
00111213CNoli esse nimium justus (Eccl 7:17)Noli esse justus multumorder/word; all three negations (noli, neque, ne) present in both and carried
00121214Alemma Multis (Sir 29:10)Multi⚠ one letter, and the gloss reads it as the nominative plural (Illos notat, qui neminem laedere volunt). Deliberately NOT marked — a [var:] asserts something about Migne's plate, and a single-letter difference is exactly what our own transcription could have produced. Plate check wanted.
00121214Blemma Qui perdiderit (Sir 29:23)cum perdiderit⚠ same reservation, compounded by the verse-order anomaly at §C.1. Not marked.
00121215BViri sanguinum non dimidiabunt dies suos (Ps 54:24)Viri sanguinum et dolosi non dimidiabuntphrase dropped mid-verse; negation present and carried
00121215Dlemma Et vome (Sir 31:25)evomeword division. Not carried under Pattern 10: Et and vome are both real words and the run construes, so nothing is quarantined.
00121216BQui major est vestrum (Lk 22:26)qui major est in vobisgenitive for prepositional phrase
00121216Clemma Hora surgenti (Sir 32:15)Et hora surgendi non te tricesdative participle for gerund
00131217Dut possimus consolari eos (2 Cor 1:4)ut possimus et ipsi consolari eosphrase dropped
00131218Blemma Propitiationum (Sir 35:3)propitiationem⚠ single-letter class again; not marked
00131219CInvenit sibi passer domum, et turtur nidum, ubi reponat pullos suos (Ps 83:4)Etenim passer invenit sibi domum, et turtur nidum sibi, ubi ponat pullos suosorder and compound verb

B.3 — lemmas checked and found to AGREE (so the list above is a list, not a sample)

Every remaining lemma in cols 1208C–1219C agrees with the Clementine text within the words it prints: Sir 24:2–3, 5, 6, 7, 12, 13, 14, 15, 17, 18, 19 (Quasi oliva, Quasi platanus), 20, 21 (all six spice lemmas incl. Balsamum non mistum, whose closing quotation Et quasi balsamum non mistum odor meus matches word for word), 22, 29, 34, 35, 36, 37, 38, 39, 41 (Quasi trames), 46, 47 · 25:2, 3, 5, 6, 9, 13, 16, 17, 18, 22, 26, 30 · 26:1, 4, 5, 6, 11, 12, 16, 18, 23, 25, 28 · 27:1, 5, 7, 8, 13, 15, 17, 22, 24, 30 · 28:8, 13, 19, 21, 23, 28 · 29:5, 12, 13, 16–18, 20, 27, 28, 29, 30, 33, 35 · 30:1, 8, 14, 15, 18, 21, 23, 26 · 31:1, 12, 13, 19, 22 (both), 24, 27, 28, 30, 31, 32 (both) · 32:1, 7, 14, 22, 24 · 33:5, 6, 7, 8, 10 (both), 13, 15, 18, 19, 25 · 34:2, 8, 9, 10–11, 12, 23, 26, 28, 30 · 35:1, 4, 10 (both), 13, 14, 21 · 36:1, 7, 16, 17, 19, 20, 24 (both), 25, 26, 28. Inline citations that agree exactly: Prov 10:4, Prov 6:12–13, Prov 15:7, Gen 3:5, Ps 33:9, Ps 44:3, Ps 7:11, Mt 5:19, Lk 22:31, 1 Cor 13:13, 2 Cor 9:6. Every printed negation in every lemma of this range is present in the English, including Sir 25:22 Non est caput, 24:38 Non investigabit, 29:13 Et non abscondas, 35:10 Et non minuas, 35:14 Noli offerre, 36:25 Non est vir, 36:28 Qui non habet nidum.


§A · Carried type and non-construing readings (Patterns 7, 10, 12)

### A.1 — 1210D · erudiunmultos — the only [sic:] in the batch Printed (Dan 12:3, inside the italic lemma): et qui ad justitiam erudiunmultos, quasi stellae in perpetuas aeternitates. A fused run (Pattern 10, the direction no grep finds). The Vulgate has erudiunt multos. Carried as [sic: erudiunmultos]no second italic marker, per Pattern 10, because it sits inside an already-italic scripture lemma. ⚠ 7a⁗ applied: the fused run is the clause's only verb and its object. Deleting the marker would leave "and they who … unto justice", a relative clause with no predicate. The English therefore states the sense in plain words beside the marker — and they that instruct many [sic: erudiunmultos] unto justice — satisfying both Pattern 12 (the marker must not swallow what it marks) and 8976 §M's ruling (render for the evident sense).

### A.2 — 1209A · Platanus ab altitudine dicta Printed: Platanus ab altitudine dicta, vel quia patula et ampla. Nam πλατός Graeci amplum vocant. Rendered literally, "so called from its height". ⚠ The sentence contradicts itself as printed and that is exactly what is being shown: the alternative etymology and the Greek gloss both require breadth, not height (Isidore, Etym. 17.7.40, a latitudine foliorum). Conjecture latitudine, near-certain. Not emended, not marked — both altitudine and latitudine are real words, so there is nothing to carry under Pattern 12, and a [var:] would be wrong because this is Migne's own prose, not a scripture citation.

### A.3 — 1209A · ad, coelestia se sustollit A comma between preposition and object. Preserved (Pattern 8: punctuation follows the plate). The English reads "raises itself up to, heavenly things". Not a defect of ours.

### A.4 — 1215B · Sicut spado. Non facit ad complexum virginis, sic adoratio idoli… Will not construe: an unannounced third-person subject, then sic with no sicut. Rendered as closely as the printed words allow ("He does not serve for the embrace of a virgin, so is the adoration of an idol, unfitting and unprofitable to those who adore"). ⚠ The printed negation is carried; the smooth reading ("As an eunuch is unfit for the embrace of a virgin, so is the adoration of an idol") would require reading Non facit as concessive and is not what is printed. Not emended.

### A.5 — 1213A · loquela ad mala dicendum prona Gerund/gerundive mismatch (ad mala dicendum for ad mala dicenda). Rendered "prone to the saying of evil things". Real words; nothing carried. Latin case has no English exponent.

### A.6 — 1213D · Pejor est interitus animae quam occidunt haeretici, quam quilibet cruciatus corporis. Two quam clauses, the first relative (quam = the soul) and the second comparative, inside one sentence. Rendered literally ("Worse is the destruction of the soul, which heretics slay, than any torment whatever of the body"). Reads as a zeugma; not emended.

### A.7 — 1212D · Sicut in medio. — an orphan Printed: VERS. 1.-- Propter inopiam, etc., [n: (RAB.)] Sicut in medio. Hic comparatione, etc., usque ad…Sicut in medio. stands between the note and the Hic comparatione formula and governs nothing. ⚠ Candidate second lemma whose closing words were lost, or a left-margin clip. Rendered as printed ("As in the midst."), not marked. Plate check wanted.

### A.8 — 1218A · aedificium virtutum praeparere praeparere for praeparare. Rendered normally as "prepare", not carried and not marked: a/e vowel confusion is ordinary medieval orthography (Pattern 9 — "attested medieval spellings are correct Latin"), and carrying it would bury the real defects in noise. Logged so that a sweep does not churn it.

### A.9 — 1216C · bonis [operibus] omnibus replevit The square brackets are in the source, not ours. Following the peer batch's ruling at 1184D, the bracket is preserved and its contents translated: "filled us with all good [works]". Not marked [sic:] — the type is legible and the bracket reads as editorial.

### A.10 — 1217B · quia non aliter potest habere pacem et quietem Singular potest after plural admonet … ut … non omittant. Rendered literally ("because not otherwise can he have peace and quiet"). Number mismatch; no emendation made.

### A.11 — 1216C · the chunk ends without a full stop …per sanctos praedicatores divinis arguetur sententiis — no closing period on the plate. Preserved (Pattern 8). This is a chunk boundary of ours; the sentence continues into 0014, which is another agent's file. Flagged for the merge, not a defect.


§C · Migne's verse numbering — observed, not corrected

### C.1 — 1214A–B · CAPUT XXIX runs 16, 17, 18 → 23 → 20 → 27 VERS. 23 (Qui perdiderit) is printed before VERS. 20 (Gratiam fidejussoris). Reproduced in the printed order. This is the printed sequence and anaphora resolves against it; no reordering, no correction.

### C.2 — lemma-to-verse offsets, all reproduced as printed 1209C VERS. 21 · Balsamum and Quasi myrrha are Vulg 24:20 · 1215B VERS. 1 · Lignum offensionis is Vulg 31:7 · 1216C VERS. 15 · Hora surgenti is Vulg 32:15 (agrees) · 1219A VERS. 15 · Extolle adversarium is Vulg 36:9. Migne's numbering of Ecclus. runs ahead of and behind the Clementine division at several points in this range. Not reported as divergence; the lemma text is what was collated.

### C.3 — 1215C · the anchor is printed [01215C], with a leading zero Every other anchor in cols 1208C–1219C is four digits. Reproduced verbatim in the English, as the marker rule requires. ⚠ Almost certainly a digitization artifact rather than Migne's type — it is a column guide, not text — so it is not marked [sic:]. Flagged for whoever owns the anchor normalizer.


§D · Checked non-defects — do not spend a plate read on these


Cols 1219D–1232A: CAPUT XXXVII–LI (the close of the work). Translated as one batch; chunks 0000–0013 were rendered in parallel by other agents and terminology is to be reconciled at merge.

Load-bearing vocabulary choices (for the merge)

LatinEnglish used here
usque ad (excerpt-elision formula)as far as (matches shipped 8985 Glossa practice)
VERS. N.-- kept verbatim, lemma translated inside the italics
CAPUT N.CHAPTER N. (Roman numeral kept; XXXIX has no stop in the plate and none here)
juxta historiam / ad litteram"according to the history" / "to the letter"
mystice"mystically"
doctor"teacher" (never "doctor")
praedicator / rector"preacher" / "ruler"
conversatio"conversation" (in its older sense: manner of life)
sacramentum / mysterium"sacrament" / "mystery"
impius / iniquus / injustus"ungodly" / "unjust" / "unjust"
disciplina / documentum"discipline" / "lesson"
electi / subjecti"the elect" / "those subject to him"
virtus"virtue" of a moral quality, "power/strength" of God
second personPattern 17: tu → thou/thee/thy; vos → you

⛔ Vulgate collation — divergences found (Clementine text, sources/vulgate/)

Every VERS. N lemma in cols 1219D–1232A was collated word-by-word against the Clementine Sirach; the inline proof-texts were collated against their own books. Truncated incipits are not listed. Seven divergences were judged material enough to carry a [var: …]; the rest are recorded here only.

Marked [var: …] in the English

  1. 1220A · Ecclus. 37:14 — Migne Non attendite (2nd pl. imperative); Vulg. Non attendas his in omni consilio (2nd sg.). Number of the address is load-bearing under Pattern 17, so the divergence has an English exponent ("attend ye not" vs "attend thou not"). Marked.
  2. 1222B · Ecclus. 39:41 — Migne Et nunc in omni tempore suo; Vulg. 39:41 Et nunc in omni corde et ore collaudate, et benedicite nomen Domini. Migne's lemma has taken up in tempore suo from vv. 39–40, and the gloss rests on the divergent word: it expounds tempus ("Say not that former times were better… Virtues make the days good"). A conformed lemma would have destroyed the gloss. Marked.
  3. 1223A · Ecclus. 40:32 — Migne conculcabitur inopia ("want shall be trodden down"); Vulg. condulcabitur inopia ("want shall be sweetened"). One letter, opposite sense. Marked.
  4. 1223C · Ecclus. 41:19 — Migne Verumtamen revertimini ("return ye"); Vulg. Verumtamen reveremini ("revere ye"). The gloss rests on Migne's reading: suadet ut revertantur ad studium docendi — he expounds it as a returning. Marked.
  5. 1223D · Ecclus. 41:20 — Migne omnem irreverentiam observare; Vulg. omnem reverentiam observare. Sense-reversing, and the neighbour of no. 4 — the two sit in the same short passage, which is worth a plate check as a pair. Marked.
  6. 1229C · Ecclus. 50:5 — Migne ingressum domus atrii; Vulg. ingressum domus et atrii. The dropped et fuses two things the gloss keeps apart: in rudibus, quos signat atrium; et in perfectis, qui signantur per domum. Marked.
  7. 1229C · Ecclus. 50:13 — Migne de manu sacerdotis (sg.); Vulg. de manu sacerdotum (pl.). The gloss repeats and builds on the singular (Partes accipit de manu sacerdotis, qui … exigit a ministris). Marked.

Divergences recorded, NOT marked (immaterial to the sense or to the gloss)

Proper-name forms in the Praise of the Fathers — checked, all ATTESTED (Pattern 9)

Cols 1226B–1229B carry the dense name sequence. Each was checked against the Clementine and against ordinary medieval usage; none is a defect and none is marked: Henoch (Clem. Enoch at 44:16, but Henoch at 49:16 — Migne is internally consistent with the second), Noe, Phinees filius Eleazari, Jesus Nave, Josue, Roboam, Jeroboam, Ezechias, Josias, Joram, Achaz, Ezechiel, Nehemias, Josedec, Sichem/Neapolis, Seth, Sem, Onias. The -as/-ias Latin nominatives and the coelum spelling throughout are Migne's house orthography, not variants. ⚠ One name IS worth a note but is not a Vulgate matter: 1228B, 48:1, the gloss says the two captains of fifty were sent ab Ozia rege; IV Reg. 1 names Ochozias. Rendered as printed ("king Ozias"); Ozias is an attested short form of the same name in medieval usage, so no marker.

Carried type — every [sic: …] in these four chunks

Each entry gives column, the run carried, and the conjecture. Per 7a⁗ the English supplies in plain words whatever force the quarantined run carried.

  1. 1219D–1220A (0014, 37:5)Possuntt for Possunt. Doubled t; a non-word, so carried. It was the clause's only finite verb, so the English says "can be understood" beside the marker.
  2. 1220D (0014, 38:21)Ad monitio for Admonitio. Pattern 10 split type; the whole run is carried, including the fragment Ad which is a real word. English supplies "An admonition" beside it.
  3. 1221C (0014, 39:20)Domini for Domino in Ps. 32:2, Confitemini Domini in cithara. A real word form, but the genitive has no slot after confitemini, so it fails Pattern 12's "does it render in place?" test and is carried; the English supplies "to the Lord". This is simultaneously a divergence from the Vulgate (Domino): it is marked once, as type, because a one-letter case ending is far likelier to be the compositor than a variant text. Worth a plate check. No second italic marker is used since the run sits inside an italic lemma.
  4. 1222A (0014, 39:35)cum for eum in bestiae quoque contra cum saeviunt. Real word, no slot after contra; carried, with "against him" supplied beside it.
  5. 1226B (0016, 44:1)nunde for nunc de. Pattern 10 fused run. It swallowed a preposition, so the English supplies "now concerning" beside the marker.
  6. 1226C (0016, 44:22)jurejarando for jurejurando, inside the italic lemma (hence no nested asterisks in the marker content). English supplies "by an oath".
  7. 1230C (0017, 51:1)JesusChristus. Pattern 10 fused run; it was the clause's subject, so the English supplies "Jesus Christ" beside it.

Other cruces — rendered as printed, no marker

Plate

No plate read is available for cols 1219–1232 (the 2026-08-29 spot check covered 1183–1186 only) and Migne prints no conjecture apparatus on the pages that were read. Every [var: …] and [sic: …] above therefore stands on the Corpus Corporum text alone and carries its column here so it can be checked against the leaf before shipping. The left-margin clipping confirmed on p. 598 was not observed in these chunks: no column start in 1219D–1232A shows lost word-openings.

Completeness

Chunk 0017 ends the work at CAPUT LI, VERS. 37, col. 1232A (Laetetur anima), which is the last verse Migne glosses. Nothing is truncated.


⭐ BLIND POLARITY READ — 2026-08-29 · 18 chunks · 4 readers · 5 defects OF OUR OWN, fixed

Four readers, one per stint, each fenced from this file, from cruces-.md, data/tei-patches/8947.json and data/plate-reads.json. That fencing is why the pass worked: three of reader 1's five findings turn out to be already logged in §A below (a dilecto 1196B, potes 1189C, Cui Deum* 1188A) — independent rediscovery, which is the evidence the reader actually read — while the two that were new are both defects of OURS, the class a translator's own cruces cannot catch because the same agent wrote both.

OUR defects — found and FIXED (5)

colwasnowwhy
1186D"he who truly fears bears his mercy patiently""…patiently awaits his mercy"sustinet = awaits. The old rendering made God's mercy a thing to be endured, fighting the lemma it glosses one clause earlier (qui sperant super misericordia ejus).
1192A–B"; and he who is weighed down … should wish to be made a judge""; and lest he who is weighed down … wish to be made a judge"velit is a second subjunctive still governed by ne. The scope failed to cross the semicolon, so the gloss appeared to advise the man crushed by his own faults to seek judicial office — the exact reverse of its thesis. A dropped-negation effect produced by scope, not omission: invisible to any particle count.
1199C"No one can, etc.""No one can [comprehend] the foreknowledge of God, etc."praescientiam Dei was simply absent from the English. Our loss, not Migne's.
1205C"testimonies of Scripture, foreign to their depraved sense""…estranged by a depraved understanding"As it stood the heretic gathered texts foreign to a depraved sense in order to confirm his error — the opposite of the point. a pravo sensu is ablative of means.
1215D"Sleep is lighter on a small fare""Sleep is finer on a small fare"subtilior = finer, less gross. "Lighter" reads in modern English as shallower, i.e. worse — against the lemma (Somnus sanitatis) and the mystical gloss. Register, but it inverted the sense for an English reader.

⛔ ONE FINDING REJECTED ON THE RULE — 1207A, and do not re-raise it

Reader 2 held that the English supplies a negation the Latin does not print: plate reads Amico reconciliatione credas in aeternum, English reads "trust thou not for ever."

Rejected. translation-style.md 7a⁗ rule 2 governs and requires exactly what is there: "Does the clause still carry its negation? If the marked run contained non/nec/ne … the English must be negative. The marker is not a place to park a negative and let the sentence go affirmative." 7a⁗-b then names word division as the dangerous shape, because quarantining a fused run removes a whole lexeme rather than one letter. reconciliatione swallowed the ne of reconciliato ne; the marker isolates the defective TYPE, never the FORCE it carries. The reader was right about what it SAW — no separate ne token — and wrong about which convention licenses it.

Migne's defects newly logged (English faithful — do NOT fix)

Yield by genre — the prediction held

CAP. XXXVII–XXXVIII (dense moral commentary) produced both of stint 4's reversals; the Praise of the Fathers, CAP. XLIV–L, produced none — list-like narrative has no neighbour to contradict. Chunks returned CLEAN and named as such: 0000, 0005, 0011, 0012, 0015, 0016. Roughly sixty candidates were considered and rejected with reasons, itemised in the readers' reports; the densest negation sites in the work (1217A four negations, 1218A's bag-with-holes paragraph with four sed non turns) were read hardest and are intact.

node scripts/polarity-record.mjs 8947 --chunks 18 --sites 5 --ours 5 recorded.