Liber Leviticus

Glossa ordinaria on Leviticus, Anselm of Laon and his school. PL 113 cols 0295D–0380A, 30 chunks, 32,212 Latin words, translated in six stints on 2026-09-07. Merged from six stint files. The concatenation is the easy half; what follows is the reconciliation, which is the required step.

Apparatus as shipped: 90 [var:] · 3 [sic:] · 1 [cj:] · 0 [ed:] · 0 [d:]. Every marker stands on a column read at Migne's plate (plate-gate.mjs 8961 green); the whole book was collated on both PL 113 printings, corners read first on every leaf. 22 TEI patches were filed and applied — all of them digitization losses, none of them Migne's defect, so none carries a marker.

Work-wide conventions — what was reconciled at the merge, and what was left alone

⛔ conversatio — ONE real drift, and it went THREE ways

Six agents who could not see each other produced three renderings of one recurring term:

manner of life 20 sites stints 0000, 0005, 0020, 0025 conversation 6 sites stint 0010 (5) + stint 0020 (1) converse 4 sites stint 0015

Every site was checked against the Latin before anything moved: all thirty render conversatio or its verb, and none is a different word. Swept to "manner of life" at 29 of the 30 sites. The majority reading is also the right one for a reader: conversatio is manner of life, and both "conversation" and "converse" are false friends that a general reader will take as talk. This is the 11551 epistylium class — a term the book uses as a technical one, split across stints so that a reader cannot see two passages using one word.

⭐ The last stint had already called it. Its own vocabulary table (§ chunks 0025–0029, table 1) reads conversatio | manner of life | "~20 occurrences; never 'conduct' or 'converse'" — written before it could see that two peers had done exactly that. A stint that fixes its vocabulary and says so in writing is what makes a merge decidable; without that line this would have been three defensible renderings and no way to choose.

⚑ One site is deliberately NOT "manner of life" and must stay as it is: @0329C (chunk 0012), qui cum peccatoribus conversantur, ad conversationem, non ad consensum vel imitationem. Migne repeats the root, and the sense is association, not manner of life. Rendered "who associate with sinners, for association, not for consent or imitation" — the figure kept, the sense right. Sweeping this one would have produced "who associate with sinners, for the manner of life", which does not construe.

✅ Checked for drift and found CONSISTENT — recorded so nobody re-sweeps them

• mundus / purus do not collide, and Leviticus is the book where they would. mundus is clean (150 sites, all six stints), purus is pure / purely / purity (10 sites). The positive control is @0366A, where Migne sets both in one clause — purissimum oleum, id est mundam vitam — and the English keeps them apart.
• usque ad, the abridgment formula, is "as far as" at all 244 sites. ⚠ A raw count turns up six "up to" / "down to" and they look exactly like drift: every one is an ordinary preposition (comes down to the legs, bent down to the earth, stirred him up to repent). Count the construction, never the word.
• The exposition tags. Mystice, Allegorice, Moraliter, Tropologice are capitalized where they open a sentence and lowercase where the English sentence puts something before them. ⚠ This also looks like drift — stint 0005 has one "Mystically" and one "mystically" — and it is not: @0310D the translator moved autem forward ("But mystically…") and @0315C left quoque in place ("Mystically too…"). Both are ordinary sentence casing, and this is the ALLEGORICALLY./ALLEGORICAL. split of 11063 not happening.
• praeputium is rendered two ways on purpose. @0345B qui in praeputio adhaerent Deo is the Pauline state and is uncircumcision; @0374D vitiorum polluti praeputio is the concrete metaphor and is foreskin. Two senses, correctly distinguished, not a drift.
• oblatio / offering, hostia / victim, concupiscentia / concupiscence, desiderium / desire, libido+luxuria / lust — all consistent across the six ranges. ⚠ "offering" outnumbers Latin oblatio in every stint alike, because it also renders munus and the bare sacrificial nouns; that is a shared convention, not a divergence.
• The leprosy vocabulary of Lv 13–14 was fixed by the stint that owned both chapters and is tabulated below (§ chunks 0010–0014, table 1): volatilis = flying, not "spreading"; effloruerit = flower forth, because the gloss @0333D–0334A is built on the flower; morticinum = carrion against cadaver = carcass. Nothing outside that range competes with it.
• The feast, jubilee and valuation vocabulary of Lv 23–27 was fixed the same way by the last stint (§ chunks 0025–0029, table 1) and follows the earlier stints' sicles / obols. One deliberate exception is recorded there: magisterium takes two English words, because Migne has two senses (mastership @0371B, teaching @0380A).

⚑ What the two printings did to this book — the finding that outlives the stints

Neither copy of PL 113 may be trusted alone, and this run caught both lying, in opposite directions, inside one book.

• Gallica — our TEI's exemplar, and the better witness for non-Latin type — drops letters the archive carries at eight sites: ingessae @0301C, occuli @0301D, conten nat @0317D, tranferente @0324C, ministare @0327D, tan git @0354C, dis cernimus @0358D, and three word-initials in a single paragraph @0367D (egem / imeres / prin-eps). A Gallica-only run would have fired at least seven false [sic:] against Migne.
• The archive is jbig2, a lossy symbol-substitution codec, so its errors are well-formed and they construe. Twenty archive-only readings were refused in the last range alone (§ chunks 0025–0029, table 3); three of them would have inverted a sense (adjectis for abjectis, ad non Deus for ad nos Deus, nomini for nemini). ⭐ Caught in the act @0346D: Gallica prints conjunctionem broken across a line with no hyphen, and the archive has lost the j and supplied a hyphen, giving con-/unctionem.
• ⛔ The numeral 3 in this fount reads as 5 at scan resolution. Vers. 5 @0360A is Migne's Vers. 3; the same glyph explains Vers. 15 = 13 and Vers. 25 = 23. A [var:] or [sic:] fired on a verse number is the exposed class — read every digit at 3× or better, on both copies.
• ⛔ A narrow or left-clipped crop manufactures a dropped letter at every line beginning on this tome. Two candidates (vicearius, Aegyto @0311A) evaporated when re-cropped with the margin in. Always crop the full column width.

⚑ Three "misfired citations of Migne's" were ours

@0352D (I Cor. XL.), @0354A (Psal. XXXVIII) and @0374A VERS. 23. on Gladium are all corruptions in our transcription; both plates print (I Cor. xi.), (Psal. xxxiii.) and VERS. 25, and Migne is right at all three. Read in our file alone each looks exactly like the nine citation misfires in this book that genuinely ARE his (listed in the stint sections below). The Psalm one carries its own control: our file cites the same verse correctly at 0345C. All three were patched in both twins, because verify-english content-checks [n: …].

The blind polarity read — six readers, 2026-09-07

Step 4a of the runbook, run blind: no reader saw this file until its findings were formed, because a reader given it first returns "all already logged" (measured, 2026-07-28). All thirty chunks were read against the Latin twin, paragraph by paragraph.

Result: 7 sites, 6 of them OURS. Every one survived verify-english, the two-printing plate collation, the 90-marker Vulgate check and the cruces merge — because not one of them is a missing negation. Five are mis-bound grammar and two are words supplied or altered where the plate does not print them. translation-style.md 7a is written around dropped negations; this run says the richer seam in a lemma-and-gloss commentary is agency.

Corrected — ours

⭐ 0310B and 0377D are one class, and it is the class to hunt next time. A gerundive with dative of agent, and a postponed relative bound to the object instead of the subject. Both leave every negation in place, so nothing mechanical can see them. 0377D has three controls on its own plate — the lemma it expounds (dabit ille qui voverat eum), the clause after it (huic enim obediet), and the identical construction three sentences earlier at 0377C (Ille enim … novit, qui correxit) which the same translator rendered correctly. That inconsistency inside one paragraph is what proves it a slip rather than a reading.

⚑ 0360D was corrected in the same pass: sed ubi et ipsi fideles persistamus is "where they too were faithful", not "let us ourselves also persist".

Migne's — rendered as printed, logged, no marker

• 0340C — Superius fluxum sanguinis patientem … et hic quoque eum qui patitur fluxum sanguinis. The first must be seminis: the same chunk sets out the scheme four paragraphs earlier (fluxum seminis patientem, obscena et corrumpentia docentem diximus … restant qui Gentilium doctrinam spargunt, quae opportune fluxus sanguinis dicitur), so as printed the sentence reads "above, blood-flux … and here likewise, blood-flux" and destroys the et hic quoque it is built on. ⚑ No [sic:]: it is a single word, and on 9004 eight of twenty-seven [sic:] fell at the plate, every fall a single word. Whose it is has not been established.
• 0315A — delictum cum non facimus quod debemus: peccatum vero cum non facimus quod non debemus. The second non makes sin a second kind of omission and cancels the distinction the sentence exists to draw. Migne's own page refutes it two paragraphs down: Peccatum est perpetratio mali, delictum desertio boni. Rendered literally.
• 0337A — Et dimittet passerem. Quia quodammodo tenetur sub contemplatione humani intellectus. The lemma releases the bird and the gloss gives its being held as the reason. A dropped negation is likely; quodammodo may be doing that work. Flagged, not asserted.
• 0372A — the gloss Et dabo, et superfluum est discusses an et its own lemma (Dabo vobis pluvias) does not print. Settled by one look at whether Migne's lemma reads Et dabo. Unresolved; no English consequence.

⛔ Raised and REFUTED — 0344A ad fortem Domini destinantur

A reader called this its highest-priority finding: an unmarked silent emendation of fortem to sortem, invisible to every gate. It is not an error. fortem is the accusative of fors, "lot" — and the proof is one column earlier in the same exposition of the two goats of Lv 16:8, where Migne prints super quem cadit fors diaboli. The reader's own corroboration ("the work uses the right word twice nearby") pointed the opposite way: fors diaboli is the same word.

⭐ This is CLAUDE.md 8a with another data point: internal corroboration proves an error exists, never whose it is — and here it did not even prove that much. Acting on it would have put a [sic:] on type Migne sets correctly.

What the readers checked and found clean

Four of six ranges came back genuine negatives, each naming the sentences where the class would have shown and did not: the Lv 11–13 clean/unclean and leprosy diagnostics (including the paradoxical toto corpore leprosum Gentilem … non expellit; sed eum mundum judicat, rendered as printed); the Lv 18 prohibition chain, checked one by one rather than as a rhythm; the Lv 22 eat/not-eat pairs; and the Lv 26 blessing/curse hinge, whose confirmed Migne negation at 0372D (quia caro non concupiscet adversus spiritum) is carried with its [var:] and not smoothed.

⚑ Every reader ran a negation-particle count as a cross-check and every one reported it useless — English exceeds Latin in all thirty chunks (periphrasis, un- compounds, marker text), and in the one chunk containing a real Migne defect the count ran four in the wrong direction. It is not a detector.

Stint 0000 — cruces-0000.md

As filed by the stint, unchanged.

Glossa ordinaria on Leviticus, Anselm of Laon and his school. PL 113. Stint file — the orchestrator merges this with the peers' files.

Plate reads: every column in this range was collated word for word against Gallica ark:/12148/bpt6k5505319w, leaves f148–f154, page map calibrated on the printed corner numbers of f148 (295/296), f150 (299), f151 (301), f152 (303), f154 (307/308). Two columns (0299–0300, 0301–0302) were additionally corroborated on the on-disk archive.org PDF, pp. 155–156, whose corners read 299/300. Recorded in data/briefs/8961-PLATE-READS-0000.json.

Lemma count: 179 marked spans in the Latin of chunks 0000–0004 (all italic; this book uses no guillemets in this range), paragraph-bounded count identical to the raw count, and equal to the brief's 179. No mis-split.

Press variants — two printings differ, both clean, NO marker fired

Both are the same shape and both were found by putting the two witnesses side by side. In each the Gallica printing (our TEI's exemplar) carries a defective form and the archive.org printing carries the correct one. There is nobody to accuse, so no [sic:]; the English renders the sense, which is the same under either reading, and the fact is recorded here.

• 0301C — Gen 46:26 lemma. Gallica prints ingessae (Cunctae animae quae ingessae sunt cum Jacob in Aegyptum); the archive prints ingressae, which is also the Clementine. Verified at 3× on Gallica and 3× on the archive; the Gallica form shows no r-stroke at all. Rendered "that went in." ⚑ Note against the jbig2 caution: jbig2 SUBSTITUTES a symbol, it does not INSERT one, so a well-formed r on the archive where Gallica has none is unlikely to be a codec artifact.
• 0301D — Gallica prints occuli (sanantur occuli ex infirmitate lacrymantes); the archive prints oculi. The same sentence's next clause prints oculos correctly on both. Rendered "eyes."

[sic:] — one, confirmed on BOTH printings

• 0300A — trangressus for transgressus (Si enim promissionem suam trangressus fuerit). A non-word. Read on Gallica f150 and on archive p.155; both print it. Marked transgressed [sic: trangressus], the sense rendered beside the marker per the 2026-09-05 amendment to Pattern 7.

[var: …] — Migne's scripture against the Clementine (15 fired)

Every one rests on a collated column, cited book chapter:verse against sources/vulgate/clementine-flat.txt.

Divergences DELIBERATELY not marked — synonym, word order, or truncation only

Listed so that a later reader can see they were checked rather than missed. In each the sense the gloss builds on is unaffected, and marking them would bury the fifteen above in noise.

• 0297C Gal 4:22 Scriptum est quoniam Abraham habuit duos filios — Vulg. duos filios habuit, and Migne drops enim. Word order only.
• 0297D Hos 10:11 — Migne drops et ego before transivi.
• 0299A Ex 1:1 — Migne stops at Aegyptum, dropping cum Jacob. Truncation.
• 0301C Acts 27:37 Eramus autem omnes animae — Vulg. vero universae. Two synonyms.
• 0302A Rom 12:3 — Migne drops the second sapere and adds sed.
• 0302B 1 Cor 13:10 Cum enim venerit — Vulg. autem.
• 0302C Lev 2:4 Cum obtuleris — Vulg. Cum autem obtuleris.
• 0303D Prov 5:3 Favus distillans — Vulg. stillans, and enim dropped.
• 0304D Heb 5:14 Perfectorum autem est cibus solidus — Vulg. solidus cibus. Word order.
• 0305B Mt 10:16 Estote simplices sicut columbae — Migne quotes only the second half.
• 0305C Jas 1:14 Quisque enim tentatur — Vulg. unusquisque vero.
• 0306B 1 Pet 1:2 — Migne drops Jesu before Christi.
• 0306C Heb 10:20 — Migne drops quam and novam et viventem. Compression.
• 0306C Isa 6:6 quem forcipe tulit — Vulg. tulerat, and the opening et dropped.
• 0306D Jas 3:17 — Migne drops est; the span is a fragment anyway.
• 0307A Heb 13:10 — word order (de quo non habent potestatem edere / de quo edere non habent potestatem).
• 0308D Lev 4:32 Si autem de pecoribus — Vulg. Sin autem.

Migne's own defects reproduced as printed (Pattern 8 / Pattern 7)

• 0301A — the plate opens a parenthesis it never closes: ( Isich. in Levit. Admirabilis verborum subtilitas, etc., … ubi Christus est in dextera Dei sedens. Reproduced unclosed; no mark supplied. Confirmed on Gallica f151.
• 0305D–0306A — the same defect again: (ORIG. Observandum est quod in peccato pontificis legislator non addit …, never closed. Confirmed on Gallica f153.
• 0302C — omnes tamen scientias superans et ad perfectionem exaltans ends with no terminal stop before VERS. 4.. Confirmed at the plate; not supplied in the English.
• 0302D — De hac re siquidem locutionem ab apostolis et a prophetis suscipimus, Verba apostolorum sunt panes — a comma followed by a capital where a full stop is wanted. Plate confirms; reproduced.
• 0305D — Ideo aut omnino additur: ut et proximis parcamus et nobis. aut is a real word occupying its slot, so Pattern 7 renders it literally ("Therefore either it is added altogether"). The conjecture is autem ("Therefore, however, it is added altogether"), which the following ut clause wants. No [sic:] (the type is a real word) and no [cj:]: the literal English is odd, not false, which is the Pattern 18 bar. Confirmed on Gallica f153.

Not defects — checked and left alone (Pattern 9)

• 0297B prescrutari for perscrutari; 0305B epatis for hepatis; 0300D spiritalium; 0305D nescibat for nesciebat (an attested late/syncopated imperfect on the pattern of scibat, not broken type). All confirmed at the plate, all translated normally.
• 0301B and elsewhere, (ID) / (ISICH). / (ISICH. in Levit. appear as running text rather than as [n: …] in the TEI. Kept verbatim Latin in the English, on the same footing as note contents — they are Migne's citation abbreviations, not prose.

Constructions rendered literally, logged because they will look wrong

• 0297B — agricolam seminantem semen bonum, justos in the list of Matthew 13's equations. Matthew makes the sower the Son of Man and the good seed the sons of the kingdom; Hesychius's compressed list equates the sower-of-good-seed with the just. Rendered as printed. Not a lemma (no italics), so no [var:].
• 0299C — et Spiritus sancti semen virtutum fruge ditavit: semen is accusative where the construction wants an ablative (semine). Rendered literally, "enriched the seed of the Holy Spirit with the fruit of the virtues." Gender/case with no clean English exponent; no marker.
• 0302A — the italic run Non plus sapere quam oportet, sed sapere ad sobrietatem. Quorum unus, puts Rom 12:3 and the resumed Leviticus 2:2 lemma (quorum unus tollet pugillum) inside one span, with the (Rom. XII) note after both. Reproduced as one span; the reader should know the last two words are Leviticus, not Romans.
• 0305B — Reticulum jecoris: Id est, epatis glosses the Latin word by the Greek-derived one, and English has one word for both. epar is kept in italic Latin so the gloss does not collapse into a tautology.

Stint 0005 — cruces-0005.md

As filed by the stint, unchanged.

Stint file; the orchestrator merges it with the peers'. Every column in the range was read at the plate before these entries were written — Gallica ark:/12148/bpt6k5505319w f155–f162, f = (column + 1) / 2, calibrated against the printed corners on f155 and f162; the archive copy patrologiaecurs04migngoog (PDF page = (column + 11) / 2, calibrated on p.164) was brought in as second witness at cols 0315–0318. Reads recorded in data/briefs/8961-PLATE-READS-0005.json.

Plate vs. our TEI

0315C — Sacrificium salutare, a word our TEI drops. The plate reads Sacrificium salutare jubetur fieri in oleo on both printings. The English renders the plate; patch filed in data/briefs/8961-PENDING-TEI-PATCHES-0005.md.

0320A — ὑπωρίς, not ὑπωμίς. The plate sets ὑπωρίς (rho, full descender, read at 3× on a native crop); our TEI gives ὑπωμίς. Neither is a Greek word — the sense wants ἐπωμίς, the shoulder-piece — but only one of the two is Migne's. English carries the plate's form; patch filed. No [sic:] fired, because the marker's content must be verbatim in the twin and the twin does not yet carry it.

0317D — conten nat / contemnat: the two printings differ, and NO marker is fired. Gallica sets et gulam conten nat with an unmistakable word space; the archive copy sets contemnat clean; our TEI has contemnat. A split run is exactly Pattern 10's shape, but the reliable prior here is the class split, not the shape: a single word seen defective on only one printing does not license a public accusation against Migne's type. Logged as a press variant, English reads "and despise gluttony", nothing marked, nothing patched. (Same class as 8950's Mititte/Mittite and PL 114 @0073A Bethehem/Bethlehem.)

0318D — habeba, and the [sic:] IS fired. Gallica sets habeba (a non-word); the archive sets habebat; our TEI has habeba, i.e. it follows the exemplar printing. Unlike 0317D the defect stands on the printing our text descends from and the twin carries it verbatim, so [sic: habeba] is checkable and true of the page we edit. The sense is rendered beside the marker per the 2026-09-05 amendment: "he had [sic: habeba] therefore two tunics" — the clause keeps its verb, and Convenienter ergo habebat duas tunicas four lines later is Migne's own control. ⚑ A later gate should know the archive printing refutes it; that is recorded here on purpose rather than left for someone to rediscover.

0311A — two candidates REFUTED at the plate, and both were artefacts of my own crop. A first column crop clipped the left margin and appeared to read vicearius and Aegyto. Re-cropped with the margin included, the plate reads vice-narius and Aegy-pto, exactly as our TEI has them. Recorded because the failure shape is general: on this tome a narrow left crop manufactures a dropped letter at every line beginning.

0315C — two archive-only candidates refuted by Gallica. The archive copy reads quas debens offerre and proxima, sunt, necessario offerenti. sacerdotis sunt; Gallica reads debent and offerentis, as our TEI does. jbig2 symbol substitution, and the standing rule that the archive is never the sole witness for a letter held.

Readings that will not construe (rendered as printed, per Pattern 7)

0315B — Non bonum scilicet aut ovium. Printed thus on both printings. The sense the context wants is boum — "not of oxen, namely, or of sheep" — and bonum leaves the abridged lemma with no construction at all. Declined for [cj:] under 18a: this is not a negation, not a correlative, and the clause is an usque ad fragment that asserts nothing false; it merely fails to construe. Rendered literally, "Not good, namely, or of sheep."

0312B — ablatis reliquis iterum renovati, quo ardeant alia quae imponuntur. renovati (nom. pl. participle) has no subject in the sentence; renovetur would construe with ignem. Plate-confirmed as printed. Rendered as closely as the words allow ("the remains being taken away, it be again renewed"), the participle carried as the passive it is; no marker — a real word in a slot English can absorb is Pattern 7's business, not Pattern 12's.

0319D — magaritae for margaritae. Plate-confirmed. Left unmarked and translated normally ("pearls") under Pattern 9: the metathesised form is attested in medieval orthography, and italicising it would bury the real defects of this range in noise.

0312B — Quem nutriet sacer. Migne abbreviates the Lv 6:12 lemma mid-word with a period. Under L5 (contraction ≠ truncation) the sentence's grammar contains the whole word — sacerdos is the subject of nutriet — so it is rendered whole: "Which the priest shall feed."

Scripture as Migne prints it — the divergence list

⛔ Every entry below was read word by word against the plate and then against sources/vulgate/clementine-flat.txt, cited by book chapter:verse. Six entries carry a [var:] in the English; the rest are recorded here because the divergence is real but does not change what the passage asserts, which is the Pattern 14 bar.

Marked [var:]

Recorded, not marked

• 0309A, Lv 5:2. Migne's lemma is Anima qui tetigerit where the Clementine — and Migne himself, nine words earlier on the same column — has Anima quae tetigerit aliquid immundum. A relative-pronoun gender has no English exponent (Pattern 9); the English uses "who" for qui and "that" for quae so the two lemmata do not collapse into one sentence, and nothing is emended.
• 0310A, Lv 5:11. Non mittet in eum for in eam (the oblation, fem.). Same class.
• 0310B, Mt 18:22. Non dico tibi septies, sed usque septuagies septies — the Clementine's first usque is absent. Cited (Matth. XXVIII); the verse is Mt 18:22. refDisplay keeps what he printed.
• 0310B, Jonah 3:9. Quis scit si convertatur et ignoscat Deus, et relinquat post se benedictionem — cited (Joan. III) for Jonas III, and conflated with Joel 2:14.
• 0312A, 2 Cor 1:12 quoted and cited (Eph. ult.).
• 0312A–B, 1 Cor 2:6 and 2:2. Sapientiam loquimur without autem; nihil judicavi me scire inter vos for non enim judicavi me scire aliquid inter vos. Both cited (II Cor. I).
• 0313B, 1 Tim 4:5. Sanctificatur autem for enim.
• 0314C, Is 9:5. erit in combustione for in combustionem.
• 0314D, Heb 12:29 quoted verbatim (Deus enim noster ignis consumens est) but cited (Deut. IV), where Dt 4:24 has Dominus Deus tuus.
• 0315B, Lk 5:32 / Mt 9:13. Non veni vocare justos, sed peccatores, cited (Matth. IX); the wording is Luke's, without Matthew's enim.
• 0315C, Job 2:4. et omnia quae habet homo for cuncta. Plate-confirmed on both printings; a synonym, so no marker.
• 0315D, Lv 7:12. Si pro gratiarum actione fuerit oblatio for oblatio fuerit.
• 0316A, Rom 9:16. Non enim volentis neque currentis, sed Dei miserentis est for Igitur non volentis, neque currentis, sed miserentis est Dei. The same verse is quoted again at 0317D in a third order (Non est volentis… ) — the variance is Migne's, per occurrence.
• 0316B, Mt 13:52. Sint, inquit, nova et vetera — an inquit inside the quotation; guillemets are not printed here and none are supplied.
• 0318C, Sir 45:9 / 2 Sm 13:18. Vestivit eum tunicam talarem, cited (Eccli. XLV), and interiorem, cited (II Reg. XV). Neither is verbatim in the Clementine at the address given.
• 0319A, Gal 5:2. Si circumcidimini for the subjunctive circumcidamini, and nihil vobis for vobis nihil.
• 0319D, Mt 7:6. ne sanctum detur canibus et magaritae ponantur ante porcos, cited (Marc. VII).
• 0320B, Mt 5:19. sed qui fecerit et docuerit hic magnus vocabitur, cited (I Tim. III; Matth. V); the received text adds in regno caelorum.
• 0321A, Ps 4:7. Signatum est super nos lumen vultus, Domine — Clementine lumen vultus tui.
• 0321B, Is 61:1. eo quod unxerit me — Clementine eo quod unxerit Dominus me.
• 0321C, Ps 44:8 quoted a second time (Unxit te, Deus Deus tuus) and cited (Psal. LIV.), where the first quotation at 0321B is correctly cited (Psal. XLIV).
• 0322C, Mk 14:23–24. Et sumens, gratias agens dedit eis, et biberunt ex eo omnes, et dixit eis for Et accepto calice … ex illo … Et ait illis. The cup drops out of the narrative clause; the gloss's argument rests on the pouring, not the cup, so no marker.
• 0322D, Ex 16:15. Hic est panis quem dedit vobis Dominus for quem Dominus dedit vobis ad vescendum, cited (Exod. XXVI).
• 0322D, 1 Cor 12:28. primum apostolos, deinde prophetas for secundo prophetas, and without Et quosdam quidem.
• 0323A, Jn 17:19. Pro eis ego sanctifico meipsum — the Clementine opens Et pro eis.
• 0310C, Ex 30:38. Quicunque talem compositionem fecerit, peribit de populis suis — the tail is verbatim Clementine; the head paraphrases quicumque fecerit simile.

Typography and structure preserved

• 0314C, Pattern 8. …mundatur ab originali ends with no terminal stop on the plate; the English ends unstopped too. Nothing supplied.
• 0323A. Migne sets VERS 39. with no period after VERS, alone in this range; kept.
• 0311D–0312A. Numbering runs VERS. 8, 10, 11, 12, 14, 16 — Migne addresses the verses he glosses, and the gaps are his.
• 0318B. VERS. 28. stands between VERS. 37. and CAPUT VIII and glosses Lv 7:38 (In monte Sinai). The number is Migne's; not corrected.
• 0317B–C. (HIERON., epist. ad Fabiolam, tom. I.) is set as running italic, not as a bracketed note, and carries the column anchor [0317C] inside it. Kept verbatim Latin in the English, on the same footing as an [n:] citation.
• 0313A. Migne prints à masculorum dignitate with a grave accent; not reproduced and not marked — a compositor's accent on a Latin preposition carries no reading.
• No guillemets anywhere in this range, in Latin or English: this book marks its lemmata with italics only, and every quotation mark the English shows is a supplied " " (Pattern 6).

Stint 0010 — cruces-0010.md

As filed by the stint, unchanged.

Glossa ordinaria on Leviticus 9–13, Anselm of Laon and his school, PL 113. Every column in this range was read at the plate (Gallica IIIF ark:/12148/bpt6k5505319w, f = (column + 1) / 2, f162–f168; corner numbers read first on every leaf), and three sites were checked on the on-disk archive.org copy as well. The read record is data/briefs/8961-PLATE-READS-0010.json; TEI divergences are in data/briefs/8961-PENDING-TEI-PATCHES-0010.md.

Lemma-span count: 205 marked spans in the Latin of chunks 0010–0014, counted both as a whole-file total and paragraph-bounded, with no unmatched delimiter in either direction. The brief's figure is 205. They agree, and the two counting methods agree with each other. This work sets its lemmata in italic and carries no guillemets at all in this range.

1. Fixed English equivalents for the leprosy vocabulary (Lv 13, chunk 0014)

Held to one equivalent per Latin word across the whole range, because the same word carries a literal and an allegorical sense in adjacent sentences and elegant variation would hide that. A peer stint owns the neighbouring range and cannot see these.

vitiligo, tinea and livida do not occur in this range.

2. ⚑ Press variants — two printings, and Gallica drops letters the archive carries

Both sites were checked on both copies and both copies are clean at both sites. There is nobody to accuse, so no [sic:] was fired, and each is recorded here instead.

• 0324C — Gallica (our TEI's exemplar, and our TEI) prints tranferente; the archive.org copy prints transferente. Rendered "transferring", which is the sister printing's own reading and not our conjecture.
• 0327D — Gallica and our TEI print ministare; the archive.org copy prints ministrare. Rendered "to minister", same reasoning.

⛔ A stint working PL 113 from Gallica alone will read both of these as Migne's defects and mark them. They are not. Conversely, two archive-only readings in this range (condendum for comedendum @0324C, Moyes for Moyses @0327C) are jbig2 artefacts and are not reportable: the codec substitutes a visually similar, well-formed character, so its errors construe.

3. The one [sic:]

• 0327B — sacricium for sacrificium, in Debet autem comedi hoc sacricium in loco sancto. Confirmed on BOTH printings and present in our twin, so the defect is Migne's. The English renders the sense with the type inside the marker, per Pattern 7 as amended 2026-09-05: "But this sacrifice [sic: sacricium] ought to be eaten in the holy place."

4. Scripture divergences — the [var:] markers, twelve in all

Each was collated word for word against sources/vulgate/clementine-flat.txt and confirmed on the Gallica plate before firing.

1. 0324B Convertimini et agite poenitentiam, put in Peter's mouth at Pentecost. The words are Ez 18:30 verbatim; Acts 2:38 reads Poenitentiam agite, Acts 3:19 Poenitemini igitur et convertimini.
2. 0324C Hic est panis quem dedit vobis Dominus ad comedendum, hic est sermo quem constituit Dominus (Exod. XVI). Vulg. Ex 16:15 Iste est panis quem Dominus dedit vobis ad vescendum, Ex 16:16 hic est sermo quem praecepit Dominus. Four separate departures in one quotation, and the gloss turns on sermo being the thing appointed.
3. 0325C Dominus Deus noster ignis consumens est (Deut. IV). Vulg. Dt 4:24 has tuus; the form with noster is Heb 12:29. Fired because the gloss makes the fire the Spirit given to us.
4. 0326A Deus meus et Dominus meus (Joan. XX). Vulg. Jo 20:28 Dominus meus, et Deus meus — the order reversed, in a gloss whose whole point is the two things Thomas touched and saw, carnem et divinitatem, in that order.
5. 0330A Crura eorum recta, et pennati pedes eorum (Ezech. I). ⭐ The strongest of the twelve: no received verse reads crura eorum recta. Vulg. Ez 1:7 is et pedes eorum, pedes recti, et planta pedis eorum quasi planta pedis vituli, and the whole gloss is built on the contrast of crura with pedes (Habere nos oportet crura sublimiora pedibus).
6. 0330B ut in ore duorum testium stet omne verbum (Matth. XVIII). Vulg. Mt 18:16 reads duorum, vel trium testium. Material: the gloss immediately draws the two witnesses as the two laws (et in utraque lege laudes Deum), which three witnesses would spoil.
7. 0332D ipse loquetur ad populum, et erit os tuum, tu autem eris in his quae ad Deum pertinent (Exod. IV). Vulg. Ex 4:16 has pro te after loquetur and ei after eris; both are dropped, and the passage is precisely about who stands to whom.
8. 0333B capilli capitis vestri omnes numerati sunt, cited as (Matth. X). Mt 10:30 reads Vestri autem capilli capitis omnes numerati sunt; the word order printed here is Lc 12:7's. The pre-scan's bare tick against Luke was the tell.
9. 0334A Date eleemosynam et omnia munda erunt vobis. Vulg. Lc 11:41 et ecce omnia munda sunt vobis — future for present, in a sentence promising that avarice can be cut away.
10. 0334C Ignis adversarios consumet (Isa. LXXIV). No verse reads thus. Is 26:11 has ignis adversarios tuos devoret; Heb 10:27 ignis aemulatio, quae consumptura est adversarios. Isaiah has sixty-six chapters, so the address is impossible as well. Confirmed on the plate, chapter number and all.
11. 0335A Qui peccat in Spiritum sanctum, non remittetur ei. Vulg. Mt 12:32 qui autem dixerit contra Spiritum sanctum, non remittetur ei — sinning against, for speaking against.
12. 0335C Qui hanc doctrinam non affert, non recipiatis eum in domum, et ave ei non dixeritis (II Joan. 10). Vulg. Si quis venit ad vos, et hanc doctrinam non affert, nolite recipere eum in domum, nec ave ei dixeritis. ⭐ Migne quotes this same verse at 0326B in a third form (Si quis venit ad vos … nolite eum recipere in domum), and both forms are on the plate. The 0326B form differs from the Vulgate only in word order and took no marker.

5. Divergences examined and DECLINED, with the reason

• 0326B nolite eum recipere in domum (2 Jn 10) — word order only against the Vulgate's nolite recipere eum; no sense at stake. Recorded at §4.12 as the companion of a marker that did fire.
• 0325C Et egressus est ignis a Domino (Lv 9:24). Vulg. Et ecce egressus ignis a Domino, devoravit holocaustum. An added est, a dropped ecce and an abridgment; the Pentecost parallel the gloss draws is untouched.
• 0328B Vidit Deus cuncta quae fecerat (Gn 1:31). Vulg. Viditque. One enclitic.
• 0327C Voluntarie enim peccantibus non relinquitur hostia pro peccato (Heb 10:26). A heavy abridgment of Voluntarie enim peccantibus nobis post acceptam notitiam veritatis, jam non relinquitur pro peccatis hostia, with the sense entire. Both negations carried.
• 0327C quia Spiritus sanctus disciplinae effugiet fictum (Sap 1:5). Migne's quia stands where the Vulgate has enim; that is the commentator's connective, not a variant reading.
• 0331A Scripsi vobis non commisceri fornicariis, cited (I Cor. III) — the words are 1 Cor 5:9–10 and the reference is wrong by two chapters, but the reference is inside an [n:] and passes verbatim; the wording turns the Vulgate's Ne commisceamini into an infinitive without touching the sense.
• 0331A Moab olla spei meae, cited (Rom. II) — the words are Ps 59:10 verbatim. Nothing diverges but the address, which is Migne's and rides in the note.
• 0332A In loco sancto edent ea. Nearest are Lv 6:16 comedetque illam in loco sancto and Lv 10:13 Comedetis autem eam in loco sancto; the point being made is only that a court can be called a holy place.
• 0333C (Joan. XI) for the Lord asking Peter a third time — the scene is Jo 21:15–17. Migne's, on the plate, inside a note, so it passes verbatim.
• 0335C/0335D Qui praedicas non furandum, furaris; peccatori autem dixit Deus: Quare tu enarras justitias meas? under one (Rom. II.) — a conflation of Rom 2:21 with Ps 49:16, but both halves are quoted accurately. Only the single address is short.
• 0335D Sicut olera herbarum cito decident (Ps 36:2, which has quemadmodum). One connective.

6. Printed readings rendered as they stand (Pattern 7), with the conjecture here

• 0330C Si naturaliter essent immunda, immundus esset viventia etiam qui tangeret. The clause will not construe: viventia is an accusative with nothing to govern it. Rendered as closely as the words allow — "he would be unclean who touched them even living" — and the hole is left visible. Confirmed on the plate.
• 0335C quae cum tentationis fuso in filiorum tenuitatem torquetur — "into the fineness of sons", in a sentence about a spindle twisting the warp. filorum, "of threads", is what the image needs and is one letter away. ⛔ [cj:] DECLINED: Pattern 18's bar is whether the English asserts something false, and "the fineness of sons" is baffling rather than false — it does not put a claim in the author's mouth. Recorded here per 18a. Confirmed on the plate.
• 0335B Calvus ac mundus est. Simplex, non stultis — "Simple, not to the foolish". The dative has no obvious governor. Rendered literally. Confirmed on the plate.
• 0333B sanctas autem deponere sanctum, damnum est — sanctum read as the dative-less masculine ("a loss to a holy man"); it may equally be a neuter predicate ("to lay aside holy ones, holy though it be, is a loss"). Rendered on the first construal; the second is here.
• 0328C quicunque enim pisces pinnati sunt et squamosi. — a sentence with no main verb of its own. Rendered as it stands.
• 0328C qui autem tales non sunt, cum his nec cibum sumere. — bare infinitive, no governing verb. Rendered as it stands, with both negations.
• 0332D Corinthiani for Corinthii. An attested-enough formation and no English exponent distinguishes them; Pattern 9, no marker.
• 0330D Migne prints VERS. 34 twice, the second time for what is Lv 11:38. His, on the plate, carried.
• 0335C VERS. 48 is printed with no period after the numeral, alone in the range. His, carried.
• 0326A Migne glosses Elisaphan twice and with two different words — Dei speculum ("the mirror of God") at 0325D–0326A, then specula Dei mei ("the watchtower of my God") at 0326A. Both are on the plate and both are rendered as they stand rather than harmonized; the following clause's speculari leans on the second.

7. Notes on the digitization, filed as patches rather than marked

• 0325D — Migne prints מישאל before Misael and אלצפן before Elisaphan; our TEI has lost both. MT-verified against Ex 6:22. The loss leaves the sentence sound, so no [ed:] (Pattern 13a clause 1); filed in 8961-PENDING-TEI-PATCHES-0010.md for a data/hebrew-patches/8961.json entry.
• 0330B — the plate reads Falsa dicendo testimonia, dignum dicendo indignum; our TEI prints a full stop before the lowercase dignum. The English follows the plate.

8. Truncated lemmata — stop where Migne stops

0328D Quidquid pinnulas, etc. (Lv 11:10) — the printed lemma breaks off before the negation, so the English breaks off with it: "Whatsoever fins, etc." An earlier draft read "Whatsoever hath not fins", which supplies a non the plate does not print at that point and is exactly the 7a failure the rule names. The full verse is Quidquid autem pinnulas et squamas non habet, and the gloss that follows carries the negation for itself.

Stint 0015 — cruces-0015.md

As filed by the stint, unchanged.

Glossa ordinaria on Leviticus, Anselm of Laon and his school. PL 113. Stint file — the orchestrator merges this with the peers' files.

Plate reads: every column in this range was collated word for word against Gallica ark:/12148/bpt6k5505319w, leaves f168–f174. The page map (f = (col + 1) / 2) was re-calibrated in this stint on the printed corner numbers of all seven leaves: f168 = 335/336, f169 = 337/338, f170 = 339/340, f171 = 341/342, f172 = 343/344, f173 = 345/346, f174 = 347/348. Two sites were additionally corroborated on the on-disk archive.org PDF (page = (col + 11) / 2), pp. 178 and 179, whose printed corners read 345/346 and 347/348. Recorded in data/briefs/8961-PLATE-READS-0015.json.

Lemma count: 230 marked spans in the Latin of chunks 0015–0019 (all italic; this book uses no guillemets in this range), paragraph-bounded count identical to the raw count, and equal to the brief's 230. No mis-split.

Our TEI diverges from the plate — English renders the PLATE, patches pending

Filed in data/briefs/8961-PENDING-TEI-PATCHES-0015.md. None of these takes a [sic:]: the plate is clean at each, so a marker would publicly accuse Migne of the digitization's defect.

• 0343D — our twin prints the non-word vetrorsum (Ego autem non contradico, vetrorsum non abii); Gallica f172 prints retrorsum, which is also Isa 50:5. Read at native resolution and again at 3×: the initial is an italic r, not a v. English reads "I went not backward."
• 0346D — our twin prints con junctionem with a space (Talium con junctionem turpitudinem appellat). Gallica f173 breaks the word across a line with NO hyphen — the line ends Talium con and the next begins junctionem — so Migne's word is conjunctionem, entire. English reads "the conjunction of such persons." ⚑ The archive is the defective witness here, and instructively so: archive p.178 prints con- / unctionem, i.e. it has lost the j and supplied a hyphen. A stint working from the archive alone would have marked a [sic: conunctionem] on a word Migne sets correctly. This is the jbig2 caution in its exact predicted form — a substitution that construes.
• 0339C — our twin's Greek reads έπισαγμα ὅνον (tonos-for-smooth-breathing, rough breathing on ὄνον); the plate reads ἐπίσαγμα ὄνον. English carries the plate's accents.
• 0347A — our twin supplies a terminal stop after …confessione deletum vel detectum. Neither printing has one: Gallica f174 and archive p.179 both end the paragraph bare. Under Pattern 8 ("never supply a mark the plate lacks") the English stops there without a period.

Migne's own defects, reproduced without a marker

• 0344B — Migne opens (ORIG., ubi supra. at VERS. 29 and never closes the parenthesis; the paragraph runs to id est abjectus et refutatus. with no ). Confirmed on the plate. Reproduced unclosed, per the Pattern 5 corollary.
• 0345B — the same shape: (secundum illud: Qui devorant plebem meam sicut escam panis [n: (Psal. XIII)] obfirmat faciem suam Deus… — the parenthesis opened before secundum is never closed. Confirmed on the plate. Reproduced unclosed.
• 0338D — the verse address reads VERS. 20 where the lemma (Ponet super extremum auriculae dextrae illius qui) is Lv 14:25, and it stands after VERS. 21 in the printed sequence. Confirmed on the plate; Migne's own numbering, reproduced.
• 0342C — Migne's note on Ecce vir, oriens reads (Psal. XV.), repeating the citation attached to Non dabis sanctum tuum four lines earlier; the phrase is Zech 6:12. His citation, verbatim, unaltered. Compare the note two clauses later, (Zachar. VI), attached to visitavit nos oriens, which is Lk 1:78. The pair is his, not ours.
• 0338C — the note (Psal. XIV.) covers Factae sunt lacrymae meae panes (Ps 41:4) and Lavabo per singulas noctes lectum meum (Ps 6:7); neither is Ps 14. Verbatim.
• 0345D — the note (Psal. XLIII) is attached to Captabant in animam justi, which is Ps 93:21. Verbatim.
• 0348B — the note (Prov. XIII.) is attached to Mulierem fortem quis inveniet, Prov 31:10. Verbatim.

Sentences that do not construe, rendered as printed

• 0336D — Quomodo enim rasis pilis succedunt novi, si novae cogitationes ad meliora pullulant veteribus abrasis. The Quomodo … si … frame has no apodosis and the sentence closes on a full stop. Confirmed on Gallica f168 exactly as our twin has it. Rendered word for word ("For just as, when the hairs are shaved, new ones follow, if new thoughts sprout up toward better things when the old have been shaved away"); no repair, no marker.
• 0342C — …et non morietur immoderatis, scilicet quaestionibus. The ablative immoderatis is stranded before its noun by the interposed scilicet. Printed thus; rendered thus ("and he shall not die of immoderate questionings, that is").

Readings checked and deliberately NOT marked

• 0339D ffuxum (VERS. 29, col 0340C) is the fl-ligature, not a doubled f: the same word prints cleanly as fluxum twice in the neighbouring lines of the same paragraph. No [sic:].
• 0337D Ut auditum muniat — the n of muniat is ink-damaged on Gallica and reads at first glance as a k. It is muniat, the reading of our twin. Damage, not type.
• 0344C loco non circonscripsit — an attested medieval spelling of circumscripsit (Pattern 9). No marker.
• ZERO [sic:] fired in this range. No candidate survived collation on both printings; the one that looked strongest (con junctionem) proved to be our file's, and the archive's own version of it proved to be the archive's.

[var: …] — Migne's scripture against the Clementine (18 fired)

Every one rests on a collated column, cited book chapter:verse against sources/vulgate/clementine-flat.txt.

Divergences from the Clementine noted and deliberately NOT marked

Each was collated and judged immaterial: an abridgment the etc. already announces, a spelling, or a word order that leaves the gloss's argument untouched.

• 0336B Quicunque in Christo baptizati sumus — Rom 6:3 quicumque baptizati sumus in Christo Iesu; word order, and Iesu dropped.
• 0336D cum hujusmodi nec cibum sumere — 1 Cor 5:11 cum ejusmodi.
• 0341D In terra deserta invia et inaquosa — Ps 62:3 in terra deserta, et invia, et inaquosa.
• 0342A Non veni legem solvere, sed adimplere — Mt 5:17 non veni solvere, sed adimplere; the object is carried down from the verse's own first clause.
• 0342C Non dabis sanctum tuum — Ps 15:10 nec dabis.
• 0343C Malitia ejus ascendit coram me — Jon 1:2 ascendit malitia ejus coram me; word order only.
• 0344A Sciens Jesus quia omnia dedit ei Pater — Jn 13:3 sciens quia….
• 0346A Si invicem comeditis — Gal 5:15 quod si invicem mordetis, et comeditis.
• 0346C quia qui videt Filium, et credit in eum, habet vitam aeternam — Jn 6:40 ut omnis qui videt Filium et credit in eum, habeat vitam aeternam.
• 0347D Si circumcidimini — Gal 5:2 si circumcidamini.
• 0348A Quis enim consensus templo Dei cum idolis? — 2 Cor 6:16 quis autem consensus templo Dei cum idolis; unmarked in Migne (no italics, no note), so it is the gloss's own voice.

Stint 0020 — cruces-0020.md

As filed by the stint, unchanged.

Stint file; the orchestrator merges it with the peers'. Every column in the range was read at the plate before these entries were written — Gallica ark:/12148/bpt6k5505319w f174–f181, f = (column + 1) / 2, calibrated against the printed corners on f174 (347 / 348) before any reading; the archive copy patrologiaecurs04migngoog (PDF page = (column + 11) / 2) was brought in as second witness at cols 0351–0358, calibrated on pp. 181–184 (351/352 … 357/358). Reads recorded in data/briefs/8961-PLATE-READS-0020.json; TEI patches in data/briefs/8961-PENDING-TEI-PATCHES-0020.md.

Plate vs. our TEI — five digitization losses, no [sic:] fired

All five have both printings agreeing against our file, which is the class the standing rule assigns to the digitization and not to Migne. Detail in the patches file; in brief:

⚠ 0354A is the one that would have produced a false public finding. Read only in our file, (Psal. XXXVIII) on Vultus autem Domini super facientes mala looks like one of Migne's misfired citations (Ps 33:17 is the verse). The plate reads xxxiii on both printings: the misfire is ours. Nobody should log it against Migne.

Press variants — the two printings differ, so no marker and no patch

0354C — tan git / tan-git. Gallica breaks the word across the line with no hyphen (hence our TEI's tan git); the archive (p.182, 400 dpi) prints tan- cleanly. A word-division defect is exactly Pattern 10's shape, but a defect on one printing only does not license a public accusation against Migne's type. English reads "openly touches him who does the same as his neighbour"; nothing marked, nothing patched. Re-cropped with the full right margin in before the call, per the standing crop caution.

0358D — dis cernimus / dis-cernimus. Identical mechanism, identical disposition: Gallica sets dis at the line end with no hyphen, the archive sets dis-. English reads "by the nose we distinguish odours and stenches".

0352D — the terminal stop after sanctorum vindicans injuriam. Gallica prints none; our TEI supplies one; the archive appears to carry one. One printing only, so nothing is asserted: the English keeps the stop.

Migne's own misfired citations — reproduced verbatim, plate-confirmed

Each of these was checked at the plate to make sure the misfire is HIS and not our file's:

• 0348D Omne opus servile non facietis in eo [n: (Exod. XXXI)] — the wording is Lv 23:7 (and parallels), not Exodus 31. Confirmed on f174.
• 0351A Ponite filium meum Salomonem in mulam meam [n: (Eccli. XIX)] — the verse is 3 Rg 1:33, and Migne himself attributes it correctly to David in the same sentence. Confirmed on f176 and archive p.181. A [var:] stands beside it for the wording.
• 0352A in paralytico grabatum portante glorificabant Deum [n: (Matth. VIII)] — the crowd glorifying God for the paralytic is Mt 9:8. Confirmed on f176.
• 0358A Anima quae peccaverit ipsa morietur [n: (Ezech. XVIII; Isa. LIII)] — Ezech. XVIII is right; the second member does not answer to Isa 53 in any wording. Confirmed on f179. (The archive's digits here read xliii; jbig2 is never a sole witness for a numeral, and Gallica and our TEI agree on LIII.)
• 0359C Non superbum, non iracundum, non vinolentum [n: (I Tim. II.)] — the catalogue is Tit 1:7; Migne's own next clause, Est enim Dei dispensator, is the continuation of that same Titus verse. Confirmed on f180.

Readings rendered as printed, per Pattern 7

0359A — the sentence that simply stops. Qui viam Dei pergere non potest, et boni operis exsors est, non ut claudus, qui saltem cum infirmitate tenebat — and there the gloss breaks off, with no object, no terminal stop, and VERS. 20 beginning on the next line. This is Migne's own truncation, read at the plate (f180): it is not a digitization loss, so no [ed:] is licensed, and Pattern 8 forbids supplying the stop. The English breaks off in the same place: "not like the lame man, who at least with his weakness was holding".

0359D — Inordinate agendo populum scandalizent. A jussive subjunctive with no governing clause, in a gloss on Et contaminare non debet. Rendered as printed ("By acting disorderly let them scandalize the people"); the sense the passage wants is a ne clause. No marker: the type is sound and nothing is defective, so neither [sic:] nor [cj:] reaches it (18a).

0350B — ne aliena delicta diluere non valeat. Both negatives are Migne's and both are in the English ("lest he be unable to wash away others' offences"). Recorded because a negation stacked on ne is exactly the shape 7a says not to tidy.

Vulgate divergences — what was fired, and what was checked and cleared

244 marked spans in the brief; 245 in my Latin. The single extra is not a lemma: at 0352A our TEI wraps a bare citation in italics, [n: (Matth. VIII)],, and the brief's generator (rightly) skips a span with no word content. Every other span matched the brief in order. Paragraph-bounded count: 65 / 22 / 49 / 62 / 47 = 245, no unmatched delimiter, and this book carries zero guillemets, as the peers found.

Fired — 15 [var: …]

⭐ The two that carry their own gloss on their backs. At 0349A the whole gloss is built on the plural: Nota quia non ait, sabbatum, sed sabbata — and the Vulgate's Ex 16:25 has the singular sabbatum, with a passive non invenietur where Migne has an active second-person non invenietis. Conform the lemma and the gloss beneath it stops being about anything. At 0362B the argument is Legale pascha significat, and Migne's Immolabis pascha is precisely what the Vulgate does not print there (Phase).

Two more worth naming: 0353A In senibus (for In antiquis) sits under the lemma Coram cano capite and supplies the very word — old men — the gloss is expounding; and 0353C / 0355A / 0358C are one recurring habit, the glossator turning a second-person scripture into a first-person plural (judicaverimus, mortificaverimus, consurreximus) so that the precept falls on "us". That is a pattern, not three accidents, and it is why each was marked rather than smoothed.

Checked against the Clementine and deliberately NOT marked

Every ⚠ and ⚑ in the brief was walked word by word. These came back divergent-but-immaterial — abridgment, an etc. cut, a particle, or a word order — and take no marker:

• 0349B 1 Tim 2:4 qui vult omnes homines salvos fieri (Vulg. qui omnes homines vult salvos fieri) — word order only.
• 0350D Lv 19:16 lemma Non ambulabis in domo and 0350A Non levabitis, 0351B custodita homini, 0351C Visitatio erit eis, 0355C In conspectu filiorum generis sui, 0356B Sine filiis morietur, 0360B Omnem immunditiam animae, 0361C linguam incisam, 0361D Non offerent haec Domino, 0360D Super ipsum: all explicitly labelled LXX by Migne. A divergence Migne himself attributes to another version is not a [var:].
• 0350B Non nocebis proximo and Non rapies — Augustine's own summary formulae inside a quotation of Augustine, not lemmata of Leviticus.
• 0350C Jac 2:10 Qui enim in uno offenderit, omnium reus est — a compression of Quicumque autem totam legem servaverit, offendat autem in uno, factus est omnium reus; unitalicised in Migne, i.e. not offered as a quotation.
• 0353A Lv 19:31 lemma Neque ab ariolis (Vulg. nec ab ariolis) — a particle.
• 0353B Act 15:10 super cervices discipulorum (Vulg. cervicem) — number of a noun with no weight in the gloss.
• 0353C Mt 15:3 Quare transgredimini … propter traditiones vestras — drops et vos, plural traditiones; the gloss turns on neither.
• 0354A Lv 20:3 Ego ponam faciem meam (Vulg. et ponam) — a particle.
• 0357D Is 1:8 Relinquetur filia Sion sicut umbraculum vineae (Vulg. derelinquetur … ut umbraculum in vinea) and Os 2:5 Fornicata est mater eorum (drops enim) — worn quotation formulae; the gloss reads the same either way. ⚠ Borderline; recorded so a later reader can disagree cheaply.
• 0358B 2 Cor 11:2, 0359C 1 Tim 6:10, 0360C 1 Cor 11:29, 0361B 1 Cor 15:41 (Stella enim differt a stella, word order), 0361D 1 Cor 6:18 (enim for autem), 0359D Heb 13:10 (comedere for edere) and Heb 10:19–20 (carnem ejus for carnem suam): checked, immaterial.
• 0360D Pr 23:1 Cum sederis … quae posita sunt ante te (Vulg. Quando sederis … quae apposita sunt ante faciem tuam) and 0362A Sir 11:30 Ante mortem non laudes hominem (Vulg. ne laudes hominem quemquam) — abridged proverbs quoted for their drift.
• 0354B Lv 18:7 Non revelabis turpitudinem patris tui — the glossator's back-reference to his own earlier lemma, not a fresh citation.
• 0358A Lv 21:12 lemma De sanctis non exibit (Vulg. Nec egredietur de sanctis) — Augustine's lemma form, and Migne says so with (ID., quaest. 82.).
• 0352D Ps 21:21 Erue a framea, Deus, animam meam, et de manu canis Unicam meam — agrees with the Clementine exactly, capital U and all; the capital is Migne's and is reproduced.

Pattern 9 items declined

• Moses beside Moyses in the same work (0359C vs 0348D) — both attested; no marker.
• grossescunt (0359A), refricant (0355B), cicinnos (0352C), tornicantur-corrected fornicantur: only the last is a defect at all, and it is ours (above).
• pascha for Phase at 0362B is a real-word variant of the received text, so it took a [var:], not a [sic:].

Marker tally

[var:] 15 · [sic:] 0 · [ed:] 0 · [d:] 0 · [cj:] 0.

Zero [sic:] is the honest number for this range: the five candidates that looked like Migne's broken type (edehat, tornicantur, nommetur, tan git, dis cernimus) all dissolved at the plate — three into digitization losses and two into press variants. The prior held: on a collated lemma be generous with [var:], on a single word be stingy with [sic:].

Stint 0025 — cruces-0025.md

As filed by the stint, unchanged.

Glossa ordinaria on Leviticus 23–27, Anselm of Laon and his school, PL 113. This range carries the book to its end: the feasts, the jubilee, vows and valuations, and the closing Haec sunt praecepta at 0380A.

Every column in this range was read at the plate, on BOTH printings, word by word. Gallica IIIF ark:/12148/bpt6k5505319w, leaves f181–f190; the on-disk archive.org copy patrologiaecurs04migngoog, PDF pages 186–195. Both page maps were re-calibrated against printed corner numbers before any collation (Gallica f181 = 361/362; archive p.186 = 361/362; archive p.195 = 379/380). The read record is data/briefs/8961-PLATE-READS-0025.json; the eight TEI divergences are in data/briefs/8961-PENDING-TEI-PATCHES-0025.md.

Lemma-span count: 359 marked spans in the Latin of chunks 0025–0029, counted both as a whole-file total and paragraph-bounded, with no unmatched delimiter in either direction (all 718 asterisks pair inside their own line). The brief's figure is 359. They agree, and the two counting methods agree with each other. This work sets its lemmata in italic and carries no guillemets at all in this range, as the peers found for cols 0295–0336.

No [sic:] was fired anywhere in this range, and no [d:]. Nothing in these nineteen columns was found defective on both printings. Markers fired: 24 [var:] and one [cj:].

1. Fixed English equivalents — the feast, jubilee and valuation vocabulary

Held to one equivalent per Latin word across the whole range. Where the peers' table (cruces-0010.md §1) already fixed a word, that choice is followed — in particular sicle for siclus and obols for oboli, both already in the shipped 0000–0014 English.

2. ⭐ The one that matters most — a printed negation reversing the Apostle

• 0372D — Migne prints securi erimus in terra nostra, quia caro non concupiscet adversus spiritum [n: (Galat. IV)]. The Vulgate at Gal 5:17 reads caro enim concupiscit adversus spiritum — the flesh does lust against the spirit; it is the premise of Paul's whole argument. Confirmed word for word on both printings (Gallica f186 at 3×; archive p.191). The negation is carried into the English exactly as printed, per 7a, and a [var:] stands beside it. Migne's own address, Galat. IV, is also wrong: the verse is in ch. V. ⚠ Note that the gloss is built on the negative — the promise of Lv 26:5-6 is security because the flesh will not fight — so this is not a slip whose sense the context refutes. It is a divergence the exposition depends on, which is exactly why it must be shown and not smoothed.

3. ⛔ Twenty archive-only readings — NOT reportable, listed so nobody re-raises them

The archive.org copy is jbig2, a symbol-substitution codec, so its errors are visually similar, well-formed, and construe. Every one of these was clean on Gallica and agreed with our TEI. None is a press variant and none is Migne's. Recorded because two of them would have been damaging if believed.

⭐ marks the three that construe and would have inverted a sense: adjectis ("the sinners being added") for abjectis ("cast out"); ad non Deus — the same s→n substitution as the 11064 apud non case; and nomini for nemini, which would have swallowed the negation of Jn 8:33 and left "we have served a name."

4. ⚑ Press variants — the two printings differ, so there is nobody to accuse

No [sic:] fired at any of these, and no TEI patch. Rendered by sense, logged here.

4a. ⭐ THE GALLICA COPY DROPS THREE WORD-INITIAL LETTERS IN ONE PARAGRAPH — 0367D

Migne's Et eduxerunt gloss. Gallica f184 prints, on three consecutive lines:

Experimento ipso egem confirmatam ostendit, ut et tu blasphemare imeres, et peccatum fugeres, quod non solum prin-/eps, sed tota multitudo vindicat.

The archive (p.189) prints legem, timeres, princeps. Our TEI, which follows the Gallica exemplar, carries egem and imeres (it silently joined the prin-/ceps hyphenation). Verified with the full left margin in frame at 3×, from x = 0 of the leaf: the affected lines begin at the same left margin as their neighbours, so this is the copy's own left-edge inking failure and not a crop artefact. A stint working from Gallica alone would have fired two false [sic:] against Migne here — the exact hazard the runbook names. English renders "the law", "mightest fear", "the prince".

4b. 0366D — interdura / interdum, the reverse direction

Gallica and our TEI print the non-word interdura; the archive prints interdum. Gallica's type at that line-end is smudged, so it may be a damaged interdum rather than a distinct setting. Either way one printing is clean and the other is not: rendered "sometimes", no marker.

4c. 0376D — ntelligibilem, archive defective

The archive drops the word-initial i of intelligibilem (cognationem tam sensibilem quam ntelligibilem). Re-cropped with the whole column margin in frame to be sure. Gallica and our TEI are whole. This is the mirror of 4a and worth pairing with it: neither printing is the correct one, and the direction of the defect changes from leaf to leaf.

4d. 0368A — an unresolved verse number, VERS. 4 / VERS. 3

At the Seres agrum lemma the archive prints VERS. 4 — the correct locus, Lv 25:4 seres agrum tuum — while Gallica and our TEI print VERS. 3, repeating the number that stands over the preceding Sex annis gloss. Both digits are well-formed at 4× on their own copy, so this is not obviously a jbig2 substitution; a substitution that happened to correct Migne would be a remarkable coincidence, which argues for two genuine press states. The English follows the twin (VERS. 3) rather than silently supplying a number the reader's Latin column does not show. Flagged for the orchestrator: if the patch pass reads a third copy, this is the site.

4e. Two terminal stops the archive carries and Gallica does not

0372B non det escam and 0377C neque ipsius emptoris fuerat. Both sentences end without a stop in Gallica and in our TEI, and with one in the archive. The English carries the stop, on the press-variant rule (one printing is clean); both are recorded here so that a reader meeting our full stop can see it is one printing's and not our invention.

5. Migne's own misfired citations — his, confirmed on both printings, not ours

Reproduced verbatim in the [n: …] markers, as the rule requires. Where the quoted text also diverges, the point is folded into the [var:]; where only the address is wrong, it is recorded here alone.

• 0364A (Matth. IX) on Vade et vende omnia quae habes — Matthew has it at ch. XIX. Folded into the [var:].
• 0364C (Joan. X) on Cumque elevaverit eos sacerdos cum panibus primitiarum — the words are Lv 23:20 verbatim, a Leviticus lemma with a John address. Text clean; no [var:].
• 0364D (Matth. VIII) on Sunt enim quae nunc non possunt nobis revelari — the allusion is to Jn 16:12. Not a quotation, so no [var:].
• 0367A ⭐ (Joan. XXXI) — John has twenty-one chapters. The sister of the peers' Isa. LXXIV find at 0334C. Folded into the [var:] on Jn 3:18/3:36.
• 0367C (Psal. LIX) on Filii hominum, dentes eorum arma et sagittae — the verse is Ps 56:5. Text clean against the Clementine; no [var:].
• 0369C (Matth. VI) placed on Dabo benedictionem meam, which is Lv 25:21; the Matthew address belongs to the solliciti quid comederent allusion three words earlier.
• 0370D (Matth. XVII) for Mt 7:2. Folded into the [var:].
• 0372D (Galat. IV) for Gal 5:17. Folded into the [var:] of §2.
• 0373A (Matth. XVII) on Hoc genus non ejicitur — Mt 17:20, correct.
• 0374A (Joan. III) on Semen Abrahae sumus — the verse is Jn 8:33, quoted verbatim. Address wrong, text clean; no [var:].
• 0375A ⭐ (I Tim. II) on Caecitas ex parte contingit in Israel — the passage is Rom 11:25, and I Timothy has nothing resembling it. Folded into the [var:].
• 0377A (Joan. XIV) on cujus vitis Christus, Pater agricola — Jn 15:1. Confirmed on both printings (the archive's xvi is a jbig2 reading, §3).

6. The [var:] markers — 24, each collated word for word and plate-confirmed

Against sources/vulgate/clementine-flat.txt, then confirmed on the plate before firing.

Chunk 0025 — 0364A Mt 19:21 (omnia imported from Lc 18:22, cited Matth. IX) · 0364B 2 Cor 2:14 scientiae for notitiae, the word the gloss then expounds · 0364D Rom 11:33 without divitiarum · 0364D–0365A Heb 4:10 in illam requiem for in requiem eius, in a gloss whose whole point is that God's rest signifies ours.

Chunk 0026 — 0366A Eccl 9:8 non deficiet (future indicative) for jussive non deficiat, which turns the precept into the promise of abiding mercy the gloss reads out of it · 0366C Prv 3:28 + 27:1 recast, with nescitis pluralizing the second clause · 0366C the peccatum non fecit proof-text cited (Isa. LIII) but worded as 1 Pt 2:22 · 0367A Jn 3:18 with in Filium imported from 3:36, plus the impossible Joan. XXXI.

Chunk 0027 — 0368C 1 Th 4:15 plural descendent and incorrupti (from 1 Cor 15:52) for primi · 0369A "1 Cor 3:22" printed as Omnia vestra sumus sicut et vos nostra, which is the shape of 2 Cor 1:14 · 0369C ⭐ Mt 6:34 as Nolite cogitare de crastino where the Vulgate has Nolite solliciti esse — the gloss's own next clause is lex jubet ne COGITENT de tribus annis, so the contrast the sentence is built on lives in the divergent word · 0370D Mt 7:2 in quo judicas for in quo iudicio iudicaveritis · 0371C 1 Tim 5:8 Qui enim for Si quis autem, turning a condition into a maxim, with abnegavit for negavit.

Chunk 0028 — 0372D ⭐ Gal 5:17 with the printed non (§2) · 0373B Mt 25:40 heavily recast · 0373B "Col. III" Venit ad nos Deus et habitat in nobis, which is in no Vulgate verse · 0373C Bar 3:36,38 reputabitur ... ad eum for aestimabitur ... adversus eum, and in terra for in terris, in the gloss on Ambulabo that makes the point of the incarnation · 0375A ⭐ Rom 11:25 contingit/subintraret cited as I Tim. II · 0375B Mal 3:1 Subito for statim.

Chunk 0029 — 0375C Mt 5:48 without caelestis · 0375D ⭐ 1 Pt 2:2 infantes rationabiles, moving rationabile from the milk to the infants — and the gloss is precisely about the one-month-old newly begotten in the faith, so the transfer is load-bearing · 0377D ⭐ Lv 27:25 Omnis aestimatio siclo sanctuarii pondera for ponderabitur: Migne's lemma has lost its verb, and the gloss then supplies one of its own (ut omnis definitio sanctis fiat ponderibus) · 0378B 1 Cor 4:17 filius meus dilectus for carissimus et fidelis, and ipse for qui · 0379A 2 Tim 4:11 Assumens Marcum, deduc tecum for Marcum assume, et adduc tecum.

7. The one [cj:] — 0368B, Dominum for Dominus

Migne prints Non prohibet Dominum vesci eis quae sponte sunt, sed reponere in usum proprium. The accusative is a real Latin form in a real slot (subject of the infinitive after prohibet), so nothing is broken and Pattern 12 has nothing to wrap — but the faithful English then says that the Lord is the one forbidden to eat, when the sentence is plainly about what the lawgiver forbids the people. That is the agency-inversion class of 7a's fifth mechanism, and Pattern 18 is where it lives. Rendered as printed with the gloss beside it: He does not forbid the Lord [cj: Dominum; read Dominus, "the Lord does not forbid"] to feed upon the things that grow of themselves. Confirmed on both printings.

8. Divergences checked and CLEARED, with reasons

Itemized because a report that lists only findings cannot be told from one by a reader who never read. All were collated against the Clementine and, where the wording was at issue, against the plate.

• 0362C ut in omnem terram exeat sonus eorum (Ps 18:5 exivit) — the subjunctive is the gloss weaving the verse into its own purpose clause, not a divergent copy-text. No marker.
• 0363B Mt 26:28 without enim and novi testamenti — an abridgment in a sentence about the libatio; nothing in the gloss turns on the dropped words.
• 0363D Mt 5:21-22 without quia — same class.
• 0364A Jn 1:29 peccata mundi for peccatum mundi — a number difference with no exponent in the gloss, which is about the seven gifts. Declined.
• 0365B Rom 6:19 abridged mid-verse; 0365C Mt 18:4 with se humiliaverit transposed; 0365C 2 Cor 5:1 compressed inside an usque ad tail — all three are abridgment, not divergence.
• 0366B Ps 109:7 verbatim ✓. 0366C Jn 1:9 truncated at venientem ✓ (no in hunc mundum, but the sentence stops there).
• 0366B Lv 24:3 Usque in mane for usque mane (the brief's one ⚑ in this range, matching only Ex 16:23) — a preposition, no sense change. Declined.
• 0367D Mt 5:39 close enough to the Clementine to be an ordinary memoriter quotation.
• 0368B Ps 127:2 verbatim ✓. 0368D Heb 12:14 verbatim ✓ — worth naming, since this is the verse the 7914 read found silently conformed (castimoniam → "holiness"); here Migne agrees with the Vulgate and there is nothing to mark.
• 0369D Ps 127(128):8 verbatim ✓. 0370B Mt 13:44 verbatim ✓. 0371C Lc 19:9 with huic domui transposed — word order only.
• 0372B Prv 3:18 apprehenderunt for apprehenderint — tense inside an usque ad tail.
• 0372C Ps 103:15 recast as Panis enim cor hominis confirmat, vinum laetificat — I fired no [var:] here and want the reason on record, because it was close: the division of labour the gloss builds on (bread strengthens, wine gladdens) is the Vulgate's own; what Migne changes is the order and the mood, not the assignment.
• 0372C Phil 3:13 abridged. 0373A Mt 17:20 cum oratione for per orationem — a preposition. 0373B Rom 13:9 restauratur for instauratur — near-synonym, confirmed on both printings.
• 0373D Jn 6:61 without et. 0374A Jn 8:33 verbatim ✓ (address wrong, §5). 0374D Rom 2:28 abridged. 0375B Ps 83:8 and Ps 49:2-3 verbatim ✓. 0375A Lv 26:44 Attamen for Et tamen.
• 0378A Lc 2:23 verbatim ✓ — note this citation is printed inside the italic run as (Lucae II), not as a [n: …], and is reproduced in place.
• 0364C Non complebitis missionem vestram — not a Vulgate reading at all, and deliberately so: the gloss's whole point is differentiam innuit ipsa mutatio verborum, Migne setting two forms of Lv 19:9 / 23:22 against each other. Marking one of them as divergent would be marking the thing the sentence is about. No marker.
• The LXX renderings (holocaustomata 0362D, χίδρα νέα 0363B, pauperi et proselyto 0364D, egressus 0365D, Homo homo 0367B, Contritionem pro contritione 0367C, Ad agrum terrae aestimabuntur 0370B, Ligna campi 0372B, In conspectu vestro morte vestra 0372D, Et apponam castigare vos septies 0373D, Peccata sua placita habebunt 0374D, honorem animae suae 0375B) and the Aquila/Symmachus/Theodotion readings at 0375B and 0376B are announced as such by Migne and are not measured against the Clementine.

9. Migne's own reproduced oddities, carried and not marked

• 0365A, 0372C — Migne's own square-bracketed alternatives, Quasi [sicut] non est opus and cum obscura [occulta] explanantur. Confirmed on both printings. They are his editorial brackets, so they pass through the English untranslated, on the same footing as an [ al., …] variant. Neither is ours.
• 0372A Si in praeceptis with no stop before Quid fieri — both printings, both columns of the lemma. Carried.
• 0369D, 0373A Et centum, Usque ad annum — lemmata with no closing stop before the gloss begins. Both printings. Carried.
• 0374B the elliptical Gladium, spiritus, quod est verbum Dei, ... ne Vetus Testamentum habere cum Evangelio contradicant — the final clause will not construe cleanly (an infinitive governed by nothing). Rendered as closely as the printed words allow, "lest they gainsay the having of the Old Testament together with the Gospel," and logged rather than smoothed.
• 0380A frustra conatis hunc subducere, illum inducere — a dative/ablative participle with no head noun, an anacoluthon. Rendered "in vain for those who strive to withdraw this one," not converted into a second-person "in vain do you strive," which would have been a silent repair. Confirmed on both printings.
• 0365C verum enim et sabbatum sabbatorum annuntiavit — verum stands with an elided sabbatum. Rendered "for he proclaimed a true sabbath too, and a sabbath of sabbaths."
• 0378B our TEI's meditatur,nec (no space after the comma) is a whitespace artefact of the digitization; both plates space it normally. Not a defect of Migne's, too trivial for the patch file.

10. Not defects (Pattern 9 and the known false positives) — declined, and why

• The 0379A → 0380A anchor jump in chunk 0029 looks like three missing bands. It is not: f190 / archive p.195 shows Leviticus occupying only band A of col. 0379, the LIBER NUMERI title block and Praefatio filling the rest of the leaf, and the Leviticus text resuming at band A of col. 0380. Read at the plate rather than assumed, because this is the seam where the book ends.
• The recurring VERS. 3 at 0368A (§4d) is a press variant, not a band-arithmetic gap.
• Judaeis / Judaeos, Moses / Moyses alternating within a column: attested variation in Migne's own spelling, not carried type.
• deculpantur (0373A) — an unusual verb, but a well-formed one in its slot; translated normally ("are cleared of blame"), no marker.
• The three misfired citations whose text is clean (§5) take no [var:], because Pattern 14 marks a divergent quotation, and an Ibid.-style address is adjudicated at index time.