Evangelium secundum Marcum

Merged from the five translation stints' files (cruces-0000, -0005, -0010, -0015, -0020) after the whole apparatus had been read at Migne's plate. The stints' own sections are preserved below, verbatim and unabridged, because their reasoning is the record of how each reading was reached. Where a stint's attribution CLASS disagrees with the plate, the plate governs, and §0 says so site by site. Nothing below §1 has been silently altered.

This is a nineteenth-century abridged recension of the marginal gloss: a VERS. n.-- address, the Vulgate words being glossed in italics, then the comment, with attribution sigla (BEDA., HIER., ISID., RAB.) arriving as ordinary notes. Migne's attributions in this recension are known to be unreliable, and passing a siglum through is not endorsing it.

§0. THE PLATE READ — what it settled, and what it overturned

Source, and why it is not our usual one. Our on-disk PL 114 scan (raw/scans/pl114/patrologiaecurs06saingoog.pdf) clips both margins through Mark's range, badly enough to eat whole words — 2–4 characters off the right column's line-ends at p.95, 3–6 off the left column's line-starts at pp.114 and 118. That was recorded as a caveat at chunking, before it could bite. The read therefore used a second witness: archive.org patrologiaecursu0114mign, a different copy of PL 114, colour, 2215×3603, clean at both margins. Same remedy that settled all seven contested markers on PL 196 (2026-08-31).

https://archive.org/download/patrologiaecursu0114mign/page/n<N>.jpg

leaf n = (column + 11) / 2, calibrated against the printed corner numbers at n95 (179/180), n115 (219/220) and n126 (241/242) before any read — never from the digits in the identifier, which name the item and not the volume. ⚠ The endpoint returns HTTP 200 with an empty body to a bare curl; it needs -L and a browser user-agent. A usable witness would otherwise have been discarded as having no page images.

23 leaves, 29 markers, all read. Recorded in data/plate-reads.json under 9001.

The result

The 11535/11550/9004 asymmetry again, and at full strength. Every marker aimed at a quoted clause was Migne's. Seven of eight aimed at a single word were ours. Suspicion of a word is suspicion of our file; suspicion of a clause is evidence about Migne.

⚠ BUT THE RATE IS NOT COMPARABLE TO 9004's, AND MUST NOT BE QUOTED AS IF IT WERE

9004 withdrew 8 of 27 [sic:]; this work withdrew 7 of 8. The stints here were briefed on addendum A and deliberately did not fire on the non-words they were most confident about — two of the five fired zero [sic:], logging their non-words as cruces and asking for the plate instead. The eight that were fired are a residue, not a sample. The direction agrees with the earlier works; the denominator does not.

⭐ Stint 1 predicted its own result in its §0, before any plate was pulled: "I expect some or all to become data/tei-patches/9001.json entries rather than accusations against Migne." All three of its [sic:] fell. Stint 3 named its likeliest withdrawal by name and was right.

⭐⭐ THE ONE THAT SURVIVED IS THE INTERESTING ONE

0203A magnitudino — magnitudino signi notesceret. A genuine non-word, and Migne prints it. Addendum A's measured claim was that non-word suspicions came back 10 for 10 as the digitization's corruption. This is the first counter-example in the series. A non-word can be his, and the rule is therefore a prior, not a law.

The seven withdrawals, and what each cost to find

⭐ 0188A decided the translation, exactly as the stint said it would. It flagged that destruere gives "do away with" while astruere would reverse the sense, and declined to assert which. The plate reads destruere; the rendering stands.

⛔ 0187D: THE OBVIOUS PATCH WAS THE WRONG ONE. Our Nunquia is corrupt, but Migne's lemma reads Nonquid — with an o — while his own Matthew quotation eleven words later reads Nunquid. Patching to Nunquid would have conformed the lemma to his quotation and printed a word he does not set at that spot. ⚑ Note what this does to a familiar argument: "the correct form prints eleven words later" is the reasoning 9004 identified as a failed attribution test, and here the two forms genuinely differ on the page. Migne's own Nonquid/Nunquid inconsistency is his; it is owed a [sic:] and none is fired.

⭐⭐⭐ TWO REAL DEFECTS THAT NO MARKER POINTED AT

Both found by reading the whole line while checking something else four words away — the brief's addendum C, earning its place twice on one work.

• 0203A cursus habentis saeculi, where Migne prints labentis. A real, well-formed word, so no corpus argument could have touched it, and addendum A's rule sent stint 2 to log it as a crux and ask for the plate rather than accuse Migne on the strength of an argument. The rule worked exactly as designed. English repaired: "the course of the passing age", not "the age that now holds".
• 0202D multae. fidelium catervae. Migne sets it hyphenated across a line break — mul- / tae fidelium catervae — with no full stop. The stop is the digitization's, at that same break. Stint 2 carried it faithfully under Pattern 8 and said plainly it could not tell whose it was, which is the right behaviour; the English was otherwise about to publish "many. bands of the faithful".

Both stint behaviours are the ones to keep. Neither fired a marker, both named what they could not settle, and both were right to leave it to the plate.

Two silent normalisations — Migne's reading restored, a [sic:] owed

Corpus Corporum does not only corrupt; it also tidies. Neither of these was flagged by anyone, and nobody could have flagged them: a normalised form reads perfectly.

• 0185D — Migne sets the italic lemma A quator portabatur and his roman gloss on the same line reads Quia quatuor virtutibus. Both forms in his own type, directly comparable: the ideal control, and the same shape as 8993 @0650A (dia/die). Confirmed at 4×. CC supplied the u, so our Latin held a word Migne does not print.
• 0203A — Migne sets Pascha cruxifixum Dominum; CC tidied it to crucifixum. Confirmed at 4×. Surfaced only because the whole line was read for cursus labentis, four words away.

⚑ Each is owed a [sic:], and neither is fired here. Firing one is a fresh public claim per site, and that is Wilson's call, not the merge's.

11 TEI patches in data/tei-patches/9001.json — 9 corrupted, 2 normalised. No chunk boundary moved. Every withdrawn marker's rendering was checked, not merely unmarked: 0210C's lemma needed "And spitting" restored once the [sic: Ex] came out.

§1. WORK-WIDE CONVENTIONS — reconciled at the merge, and now settled

Five stints could not see one another, and they diverged. Each axis below was tested against the Latin before it was called drift, per the brief's rule that a control must be the same class as the thing it controls. Do not re-open these; a later sweep that churns them is undoing measured work.

1. The commentator's own voice is not archaised

Stint 5 rendered the metalanguage verb saith eight times (Luke saith, as Matthew saith, as the Apostle saith, and one lemma). The other four ranges use says 51 times and saith zero, on the same dicit/ait — and stint 5 itself used says and said elsewhere, so it was inconsistent within its own range. translation-style.md forbids archaising the gloss voice. All eight normalised to says/said. This is the 9004 stint-4 defect (168 -eth forms) in milder form.

⚑ Everything else that looks archaic in this work is inside quoted scripture and is left: 47 -eth forms, of which the italic-membership test found only a handful outside a quotation, and those sit in unitalicised scripture (Ezek 33:11, 1 Cor 2:9) where 9004's ruling leaves them.

2. turba → crowd; multitudo → multitude

turba was rendered crowd (~20 sites), throng, and multitude across the ranges. The decisive site is 0211C, where one sentence prints both words — Turba multa triduo Dominum sustinet, cum multitudo fidelium… — and the English rendered them identically. Six sites repaired; the three surviving "multitude"s all render multitudo.

⚠ "Throng" was NOT drift and was left alone. It renders the verb comprimere (ne comprimerent eum), not the noun turba — a different class, and contrasting the two would have manufactured an inconsistency that is not there.

3. daemonium → demon; diabolus/satanas → devil

Both English words were in use for daemon-, twice within a single chunk (0002, 0010), while diabolus was consistently "devil". Eight sites repaired. Every surviving "devil" now sits in a chunk where diabolus or satanas is present. ⚑ "Devils" for daemonia is also the Douay rendering, so this was mild conformation to the familiar English Bible as well as internal drift.

4. gentes → Gentiles in the typology, nations in the generic sense; natio → nation

The 9004 axis. 9004 settled on "gentiles" after checking that no natio occurs in the work, so nothing was flattened. Mark does print natio — twice — and one of them settles the convention mechanically: 0243B reads Omni nationi gentium, both words in a single phrase. They cannot share an English word.

So: natio is always nation; gentes in the Judaei/gentes typological opposition is Gentiles; gentes in a plainly generic sense keeps nations — the peoples descended from Noah's sons (0198B), Judea distinguished from the surrounding lands (0217B), and the scripture phrases omnes gentes at Matt 28:19 and Mark 13:10, which two different stints rendered identically without conferring. Two typological sites were corrected to "Gentiles" (0197D per gentes praedicant, 0207B ad salvandas gentes). Singular gens of the Jewish people ("their own nation", 0224C, 0227A) is a third class again and stands.

⚑ Ethnicus at 0235D → "the heathen" is a fourth Latin word and is not part of this axis.

5. Not drift, checked and left

mysterium / sacramentum map to mystery / sacrament at near-perfect 1:1 in every range — a useful control, because it shows the five stints could agree, and therefore that the four axes above are real divergence rather than noise.

§1a. THE BLIND POLARITY READ — 8 sites, 6 ours, and it paid for itself in Latin

Three blind Opus readers, disjoint eights (0000–0007 / 0008–0015 / 0016–0023), all 24 chunks, each fenced from the stint cruces files until its own findings were written. Recorded in data/polarity/9001.json.

NOT ONE dropped or added negation anywhere in the work. All three ran particle tallies and found no deficit in any chunk, consistent with their close reads. The dominant class was mechanism 5 again — 5 of the 6 defects were a reference silently re-bound to the contextually easier noun, exactly as on 9004 (5 of 8) and 11613 (12 of 25).

⭐ THE WORST SITE IS A CLASS OF ITS OWN: A CRUX THAT RECORDED A REPAIR NOT IN THE TEXT

0199A. cruces-0005.md §6.2 states the decision explicitly — "I take qui gravatur as the Lord … The English follows the first" — and the shipped English carried the rival reading. No re-reading of the cruces could ever have surfaced this, because the crux says the right thing; only reading the English against the Latin does. This is why the blind read is a separate gate and not a review of the apparatus.

⭐⭐ AND IT PAID FOR ITSELF IN LATIN, THREE TIMES

The readers cannot see the plate. Twice their diagnosis was right and their proposed remedy would have been wrong, which is the addendum A lesson arriving from a new direction: a defect correctly SEEN is not a defect correctly ATTRIBUTED.

• 0191B. A reader saw that "in the passion" was neither the Latin's poena nor Bede's sense, and proposed rendering poena literally with a [sic:]. Back to the plate: Migne prints in coena — at the supper. The English was smoothed over a corruption that was ours, and the proposed [sic:] would have accused Migne of a word he never set.
• 0193A. The same reader read si forte intelligant, nulla as an intrusion and proposed a [sic:]. The plate prints it exactly as we have it: the contradiction is Migne's own. No marker — nulla is a well-formed word, not broken type, so Pattern 12 does not apply. Rendered literally, with the subjunctive tightened to "should understand".
• The six unmarked non-words. A reader asked, as a class-level question rather than a finding, whether the shipped English should be silently reading through simpiliciter, juugitur, praegnantibu.,, curcifigentes cum, us que ad and ve peram with no marker — two of them lemmata. Put to the plate: six for six OURS. Every one is now a TEI patch and no [sic:] was ever owed. The stints' restraint was right in every single case.

⚑ So the running count for this work: 14 non-word suspicions have now met the plate, 13 were the digitization's and one (0203A magnitudino) was Migne's. Addendum A's rule survives as a strong prior with a named exception, which is what a rule of that kind should be.

⚑ One reader's finding declined as genuinely two-way: 0197D potentiam tamen ejus stupendo et mirando per gentes praedicant — whether potentiam is the object of praedicant or of the gerunds. Both construe; the reader itself offered it as a query. Left as shipped and recorded here so a later pass does not re-derive it.

⚑ Two bookkeeping gaps in the stint files, passed on rather than repaired: cruces-0010 has no lemma-expansion section though its chunks expand at least five stumps, and cruces-0020 files the Cleophas site at 0237C when it is at 0240B.

§2. THE STINTS' OWN FILES, VERBATIM

Preserved unabridged. Where §0 overturns a stint's attribution, §0 governs — but the reasoning is kept because it is the record of how the reading was reached, and because on this work the stints' conclusions were mostly right where their attributions were not.

<!-- ===== stint file: cruces-0000.md, verbatim ===== -->

Stint 1 of 5. Cols 0179A–0192D, chunks 0000–0004, 5,495 Latin words → 8,042 English words (ratio 1.52–1.56, on the register's expected ~1.5×). verify-english.mjs clean on all five: column anchors, [n:] notes, italic spans, VERS. n.-- addresses and paragraph counts all 1:1 with the Latin twins.

Apparatus fired in this range: 3 [sic:] · 6 [var:] · 1 [nt:] · 0 [cj:] · 0 [d:] · 0 [ed:]. Notes carried: 170 [n:] (+1 converted to [nt:]). Guillemets: zero in either language, in either direction.

§0 · ⛔ PLATE READ OWED — the three [sic:], and I do NOT claim to know whose they are

All three stand on columns nobody has read. Each is a non-word — a form that is no Latin word at all — so Pattern 7 cannot render it and Pattern 12 requires the marker on carried type; there is no third treatment available inside the conventions. But addendum A's own measurement says these are as likely to be OUR corruption as Migne's: every one of the ten non-word suspicions in the 2026-08-23 batch was confirmed as the digitization's error, not the plate's, and eight of 9004's twenty-seven [sic:] were withdrawn on exactly this class, every fall a single word. Read these three first; I expect some or all to become data/tei-patches/9001.json entries rather than accusations against Migne.

⚑ The two 0003 sites share a shape and that is worth a look at the plate in one pass: both are a final -d standing as -a (Nunquid→Nunquia, destruere→aestruere is de-→ae-, but 8718's twin is the same -d→-a class). Both twins occur in other works, which argues — without settling — for a source-side class rather than this plate. I offer no attribution. 9004 measured translating agents' attribution guesses at 5 right, 9 wrong.

7a⁗ compliance on all three — the marker does not swallow the sense:

• 0183C: the marked run was the clause's object; "the woman" stands beside it.
• 0187C: the marked run carried the interrogative; "Can" stands beside it, and the lemma keeps Migne's terminal period rather than a ? he does not print. Per Pattern 10 the marker takes no second italic because it sits inside an already-italic lemma.
• 0188A: the marked run was the clause's only verb; "do away with" stands beside it. The conjecture is destruere and the reason it beats astruere (a real word, and the reading a bare ae/a spelling would give) is the argument itself: the Ebionites are the party that keeps the letter and repudiates Paul, so the apostles' plucking undoes the letter of the sabbath, it does not affirm it — and 8923's twin, at a verse whose received text is destruere, points the same way. If the plate reads astruere the English must be reversed to "build up", so this is the one of the three where the plate read changes the translation and not only the apparatus.

Strip every marker from these five chunks and each sentence still reads aloud without it: no stranded article, no stranded preposition, no doubled mark. (Test 1, run mechanically.)

§1 · THE LEMMA DIVERGENCE LIST — all 256 italic spans walked (7a″ point 4)

Brief count check, run first (brief §7a″ addendum): data/briefs/9001-lemmata-0000.txt declares 256 spans for chunks 0000–0004. Counting the italic spans in my own Latin gives 0 + 80 + 45 + 49 + 82 = 256, and the file's first and last entries are the first and last spans of my range (0179A Initium Evangelii · 0192D In propria venit). The totals agree and the cut is aligned for this range — no span is missing from my file and none belongs to the next stint. (Recorded because the split has been wrong on this very work; on the evidence here stint 1 is the range it did not damage.)

The pre-scan ordered the work; every span was then read word by word against sources/vulgate/clementine-flat.txt at the verse actually being glossed, which is what the pre-scan's ✓ does not do.

1a · The line I drew, stated before the findings so it can be argued with

[var:] fired where a whole word or clause is replaced by different matter that changes what the verse says. [var:] declined, and the divergence logged here instead, where the difference is (i) a clip at either end of the quotation, (ii) an inflection or a single letter, or (iii) an adaptation to the citing sentence (case, person, or a relative pronoun bent to the gloss's own antecedent).

⚑ The reason (ii) is a decline and not a judgement call: CLAUDE.md records 9003's collation finding Corpus Corporum departing from Migne at ~3 sites per page, and finding the digitization manufacturing a Vulgate divergence — judicabit→judicavit, a single letter — that a collation then dutifully "found." A single-letter [var:] on an unread column is that trap exactly. Six such sites are logged below unmarked, and each is a plate candidate if anyone wants them.

1b · [var:] FIRED — six sites

1c · Divergences found and DECLINED — twenty-two sites, with the class

Clips at either end of the quotation (Migne's citing practice; convention 3 governs, and per the ⚑ ruling expansion is not conformation):

• 0179A Ecce mitto. — Mc 1:2/Mal 3:1 read Ecce ego mitto. The same gloss quotes Ecce ego mitto angelum, correctly three lines below, which is what makes this a clip.
• 0179C Rectas facite semitas. — Mc 1:3 adds ejus.
• 0180A Gratia salvati estis — Eph 2:5/8 Gratia enim estis salvati, and the words are in the other order.
• 0181B Hoc sentite in vobis — Phlp 2:5 Hoc enim sentite; Propter quod et Deus illum exaltavit — Phlp 2:9 exaltavit illum.
• 0181B Et baptizatus est in Jordane. — Mc 1:9 baptizatus est a Joanne in Jordane.
• 0181D Jesus plenus Spiritu sancto — Lc 4:1 Jesus autem plenus.
• 0184B Nemo principum hujus saeculi cognovit. — 1 Cor 2:8 opens quam nemo principum; Migne makes the relative clause a sentence.
• 0184C Date eleemosynam: et ecce omnia munda sunt. — Lc 11:41 ends munda sunt vobis.
• 0184D Dixit et facta sunt — Ps 32:9 Quoniam ipse dixit.
• 0186B ne deterius tibi contingat — Jo 5:14 ne deterius tibi aliquid contingat. ⚑ The first draft here read "lest something worse befall thee," which silently restored the dropped aliquid. Caught and removed; the English now stops where Migne stops.
• 0186D Et relictis omnibus secutus est eum — Lc 5:28 relictis omnibus, surgens secutus est.
• 0189A Fili, non sit manus tua porrecta ad capiendum — Sir 4:36 has no Fili, sets Non sit porrecta manus tua, and reads ad accipiendum. A near-synonym plus a vocative Sirach uses habitually; the sense of the precept is untouched.
• 0189A a latere cadent mille — Ps 90:7 Cadent a latere tuo mille; the tuo goes because the gloss has bent the verse into a relative clause on cujus.
• 0192D non habebitis vitam — Jo 6:54 non habebitis vitam in vobis.

Inflection or a single letter (the manufactured-divergence class; see 1a):

• 0179C custodiunt … exquirunt — Mal 2:7 custodient … requirent: present for future, and a different compound of the same verb. Nothing in the gloss (that priests are called angels) turns on either.
• 0180D Testimonium perhibuit — Jo 1:34 perhibui, and Filius Dei for Migne's Dei Filius. Third person for first, because Migne has recast John's testimony into his own reporting frame. English cannot render the Dei Filius order at all.
• 0186D Justus in primis accusator est sui — Pr 18:17 Justus prior est accusator sui. Different word, same sense.
• 0186D Fecit eis convivium magnum Levi — Lc 5:29 Fecit ei. One letter, and it moves the feast from Christ to the company — the gloss's own next words (Qui enim Christum intus recipit) favour ei. Left unmarked deliberately, and it is the single strongest candidate in this list for a [var:] after a plate read.
• 0187C praecedent vos in regno Dei — Mt 21:31 in regnum Dei.
• 0188D Misericordiam volo — Os 6:6 volui; the present is Matthew's form of the same verse (Mt 9:13, 12:7), which is likelier the source of Migne's wording than a variant.
• 0189B magis autem laboret operans — Eph 4:28 operando.
• 0191A imposuit eis nomen Boanerges — Mc 3:17 nomina.
• 0192B Fiat contra Dominum semper — Ps 108:15 Fiant.

Adapted to the citing sentence:

• 0180B In Jordane addressed to VERS. 5, where Mc 1:5 reads in Jordanis flumine (in Jordane is verse 9's wording). The gloss is an etymology of Jordan and works either way.
• 0180D Ego baptizo vos aqua (present) against the same chapter's own lemma Ego baptizavi vos aqua at VERS. 8 — Isidore is quoting Mt 3:11, not Mark. Both forms are rendered as printed, and they differ in the English as they differ on the page.
• 0180D Fortior me — Mc 1:7 prints Venit fortior post me; fortior me is Lc 3:16's and the Greek's. The gloss turns on fortior, not on me.
• 0181C Vidit apertos coelos — Mc 1:10 vidit caelos apertos. Word order only, and English has no exponent for it; recorded because 7a″'s own worked case (@1079B, cinis et pulvis) was an order reversal that did matter. Here nothing in the gloss depends on it.
• 0185C in qua corrigit adolescentior viam suam — Ps 118:9 In quo corrigit … viam suam?. The feminine is bent onto the gloss's sublimitas … Scripturae, and the Psalm's question mark is gone. Pattern 8: we do not supply the mark Migne does not print.
• 0184D corpora sua — Rom 12:1 corpora vestra, because Migne has recast the imperative into an infinitive after offerre.
• 0191B Mandabo nubibus meis ne pluant — Is 5:6 nubibus mandabo ne pluant super eam imbrem; the possessive is added, and the gloss is precisely about him that suspends the waters, so it fits the argument rather than the verse.
• 0191B ut leo et bos simul comedant, et lupus cum agno — Is 11:6–7 compressed and re-worded (leo quasi bos comedet paleas · Habitabit lupus cum agno). This is an allusion woven into the gloss's own ut clause, not a quotation; Pattern 14 fires on quotations.

Checked and found to agree with the Clementine EXACTLY — logged because a findings-only list cannot be told from a report by an agent that never looked (7a″ point 4). Mc 1:1 Initium Evangelii · Jesu Christi · Filii Dei · Mt 1:1 filii David filii Abraham · Mc 1:2 Sicut scriptum est · Mc 1:3 Clamantis · In deserto · Parate · Viam Domini · Mc 1:4 Joannes · Baptizans · Mt 10:8 gratis accepistis, gratis date · Mc 1:5 Confitentes · Ps 95:6 Confessio et pulchritudo in conspectu ejus · Lc 1:28 Benedicta tu in mulieribus · Mc 1:6 Et erat Joannes · Vestitus pilis cameli · Mc 1:7 Solvere corrigiam · Mc 1:8 Ego baptizavi vos aqua · Ille vero baptizabit vos Spiritu sancto · Ps 45:5 Fluminis impetus laetificat civitatem Dei · Apc 2:7 Vincenti dabo edere de ligno vitae, quod est in paradiso Dei mei · Jo 1:14 Plenum gratiae et veritatis · Mc 1:10 Manentem in ipso · Mc 1:11 Tu es Filius meus dilectus · Mc 1:12 Spiritus · Lc 4:1 et agebatur a Spiritu in desertum · Mc 1:13 Quadraginta diebus · Eratque cum bestiis · Mc 1:14 Postquam autem · Praedicans · Jo 18:36 Regnum meum non est de hoc mundo · Mc 1:15 Poenitemini · Mc 1:16 Vidit Simonem · Mc 1:21 Et ingrediuntur · Capharnaum · Mc 1:22 Erat enim · Mt 5:21/22 Audistis quia dictum est antiquis · Ego autem dico vobis · Mc 1:24 Quid nobis · Mc 1:25 Obmutesce · Mc 1:26 Discerpens · Mc 1:27 Quidnam est? · Mc 1:30 Et statim dicunt ei · Ps 18:13 Delicta quis intelligit? ab occultis meis munda me · Mc 1:31 Apprehensa · Ministrabat eis · Mc 1:32 Vespere autem · Afferebant ad eum omnes male habentes · Mc 1:33 Ad januam · Mc 1:34 Quoniam sciebant eum · Mc 1:35 Et diluculo · Mc 1:38 Ad hoc enim veni · Mc 1:39 Et erat praedicans in synagogis · Mt 8:2 Et ecce leprosus veniens adorabat eum, dicens: Domine, si vis, potes me mundare · Mc 1:40 Et genu flexo · Si vis potes · Mc 1:41 Extendit manum · Volo, mundare · Mc 1:42 et cum dixisset, statim discessit ab eo lepra · Ps 113:11 omnia quaecunque voluit fecit · Mc 1:44 Vide · Ostende te · Mc 1:45 At ille egressus · Ita ut jam · Mc 2:2 Ita ut non caperet · Mc 2:3 Paralyticum · A quatuor portabatur · Mc 2:4 Nudaverunt tectum ubi erat · Mc 2:5 Cum autem vidisset · Fili · Mc 2:5 Dimittuntur tibi peccata · Mc 2:7 Quis potest dimittere · Mc 2:9 Quid est · Mc 2:10 Filius hominis · Mc 2:11 Vade in domum tuam · Mc 2:13 Et egressus · Mc 2:14 Levi · Lc 14:33 qui non renuntiat omnibus quae possidet, non potest meus esse discipulus · Secutus est eum · Mc 2:15 Et factum est · Publicani · Mc 2:17 Non necesse · Mc 2:18 Et erant discipuli · Mt 9:15 Nunquid possunt filii sponsi lugere quandiu cum illis est sponsus · Mc 2:21 Nemo assumentum · Mc 2:22 Sed vinum novum · Mc 2:23 Et vellere spicas · Mc 2:25 Nunquam legistis · Mc 2:26 Sub Abiathar · Panes propositionis · Et dedit eis · Mc 2:27 Homo propter sabbatum · Mc 2:28 Itaque Dominus · Mc 3:1 Manum aridam · Mc 3:2 Et observabant · Mc 3:4 Licet sabbatis benefacere · Animam · Perdere · Mc 3:5 Extende manum · Et restituta est manus illi · Lc 3:11 Qui habet duas tunicas, det non habenti: et qui habet escas, similiter faciat · Mc 3:6 Exeuntes autem · Cum Herodianis · Mc 3:7 Secessit · Et multa turba · Mc 3:9 Navicula · Ne comprimerent · Mc 3:11 Et spiritus immundi · Mc 3:12 Ne manifestarent · Mc 3:13 Et ascendens · Mc 3:14 Ut essent duodecim · 1 Cor 14:22 linguae in signum sunt non fidelibus, sed infidelibus · Mc 3:16 Et imposuit · Jo 1:42 Tu vocaberis Cephas, quod interpretatur Petrus · Jo 21:18 Alius te cinget et ducet quo tu non vis · Mc 3:17 Et Jacobum · Mt 17:5 Hic est Filius meus dilectus · Ps 100:1 Misericordiam et judicium cantabo tibi, Domine · Jo 1:1 In principio erat Verbum · Mc 3:18 Et Andream · Et Philippum · Et Bartholomaeum · Mt 5:9 pacifici, quoniam filii Dei vocabuntur · Lc 6:35 Diligite inimicos vestros · Et Matthaeum · Et Thomam · 2 Cor 12:2 scio hominem in Christo · 2 Cor 12:4 quae non licet homini loqui · Et Jacobum Alphaei · Thaddaeum · Et Simonem · Mc 3:18 Et Simonem Cananaeum · Mt 5:5 qui lugent, quoniam ipsi consolabuntur · Ps 105:30 Et cessavit quassatio · Ps 108:14 Et peccatum matris ejus non deleatur · Jr 9:23 Non glorietur sapiens in sapientia sua · 1 Cor 1:31 qui gloriatur, in Domino glorietur · Mc 3:20 Et veniunt · Jo 1:11 In propria venit, et sui eum non receperunt · Mc 3:21 Et cum audissent.

⚑ Every re-quotation inside the gloss was then checked against the rendering fixed for its lemma (7a″ point 3). The two that matter: 0184D Volo, mundare is quoted three times in one paragraph and reads identically each time; 0189A licet sabbato benefacere re-quotes VERS. 4's Licet sabbatis benefacere with sabbato for sabbatis, and the English follows the number Migne prints in each place rather than levelling them.

§2 · MIGNE'S CITATION MISPRINTS — rendered as printed, NOT corrected in the text

Citation repair belongs to data/citation-corrections.json at index time, never to the running text (Ruth's standing rule). Five in this range:

§3 · THIS RANGE'S OWN CRUCES

0179A · The prologue chunk, and why its one note became [nt:]

Chunk 0000 is ten words: a head and [n: ( Vide Opera ejus. )] — Migne does not reprint Jerome's prologue, he points at it. Pattern 15 decides by content, not length, and Pattern 4 already classes a cross-reference (Vid. notas, ibid., nunc tomo V) as editorial comment that is translated, against a locator (work title + number) that is not. This is a sentence with a verb, addressed to the reader; it is a locator only in the loosest sense. It is also the exact shape of the asterisk-layer ruling of 2026-08-24: an editorial cross-reference gets an English twin because leaving it Latin-only strands an English-only reader in front of the only line on his page.

⭐ The corpus settles it independently. 8976 (Actus Apostolorum, PL 114, same recension) opens with the identical device — ## PROLOGUS B. HIERONYMI. then Vide inter opera ejus. — and its shipped English translates it: See among his works. There the words are body italics, here they arrive inside a [n:], so the vehicle differs and the ruling does not. Rendered [nt: ( See his Works. )], keeping Migne's spacing inside the parentheses and his capital on Opera, which in a cross-reference of this kind names the collected edition.

⚠ One deliberate divergence from 8976, flagged rather than swept: its English head reads THE PROLOGUE OF THE BLESSED JEROME. for the same Latin. I have written THE PROLOGUE OF BLESSED JEROME., which is the launch instruction and also the corpus majority for B. (three works: …OF BLESSED JEROME ON ESDRAS AND NEHEMIAS, …ON THE BOOK OF ECCLESIASTES, …ON THE BOOK OF ESTHER; 8976 is the lone "OF THE BLESSED"). Cross-work harmonization is not required and I have not proposed one; this is recorded so the difference is known to be seen.

Heads, and the numerals rule: CAPUT PRIMUM. → CHAPTER ONE. · CAPUT PRIMUM. (cont.) → CHAPTER ONE. (cont.) (22 corpus precedents for that exact string; the marker is chunking's, never translated) · CAPUT II. → CHAPTER II. · CAPUT III. → CHAPTER III. The corpus was checked: CHAPTER ONE. in 61 works, CHAPTER XIII. in 26, zero spelled-out forms above ONE.

0182A · transferamur — the paragraph ENDS WITH NO FULL STOP, and so does ours

…a quibus corpore absoluti in coelum transferamur — Migne closes VERS. 13's second gloss with no terminal mark at all. Pattern 8: never supply. The English stops in the same place, in the same way. A later reader or a proof-pass will read this as our typo; it is not.

0182A · mysterium angelorum — a real word, rendered as printed, NOT emended

mysterium angelorum meremur, a quibus corpore absoluti in coelum transferamur. The relative a quibus wants persons, the verse being glossed is et angeli ministrabant illi (Mc 1:13), and ministerium is the word the sense asks for — "we deserve the ministry of angels, by whom…". Rendered "the mystery of the angels", which is what the page says. mysterium is a perfectly good Latin word occupying its slot, so Pattern 12 has nothing to wrap and addendum A governs: a real, well-formed word that reads oddly is Migne's until a plate says otherwise. The conjecture is ministerium and it is not adopted. Plate candidate; the two words differ by three letters, so it is worth the look.

0184B · Neque enim solo illius divinitatis auctoritas — the case has no host

solo is ablative, and there is no ablative for it to agree with: per hoc is accusative and auctoritas is the nominative subject. solum, the adverb, is the obvious suspect and is not adopted. Rendered by taking solo with the instrument — "For not by this alone is the authority of his divinity shadowed forth, but also the humility of his voluntary emptying" — which is the one construction the printed words allow.

⚑ This site is a test-2a trigger and was worked as one: a negative (Neque) with solo adjacent and a verum etiam correlative. The correlative pivots on what is shadowed forth (auctoritas … verum etiam … humilitas), which puts the negative on solo, not on adumbratur. Rendering it "the authority is not shadowed forth" would have been the 8990 @0614B defect exactly, and it reads better than the true one.

0187C · Publicani — two parses, both construe, neither chosen silently

…nec ut Pharisaei calumniantur in vitiis permanentes, sed poenitentes. permanentes and poenitentes are nominative or accusative plural, so either (a) the converted publicans do not slander as the Pharisees do, who remain in vices, but repent; or (b) they do not slander, as the Pharisees do, those remaining in vices, but those repenting. Rendered "and slander not as the Pharisees do, remaining in vices, but repenting" — English participles dangle exactly as the Latin ones do, so the ambiguity survives into the English rather than being adjudicated in it. (a) fits the paragraph's argument better; nothing decides it.

0187A · An open parenthesis Migne never closes — NOT supplied

Ipse (secundum hoc quod scriptum est: Justus in primis accusator est sui; Matthaeum se et publicanum nominat… The parenthesis opened after Ipse is never closed anywhere in the paragraph, and a semicolon stands where the close belongs. Mirrored: the English opens the parenthesis in the same place and leaves it open, per Pattern 5's corollary on unclosed quotations, which is the same principle. Plate candidate.

0188C · Abrathar against the lemma's Abiathar two lines above — no marker

Achimelech pater, et Abrathar filius — and the verse address five words earlier is VERS. 26.-- Sub Abiathar,. Corpus-wide: Abrathar once (here), Abiathar seven times. ⛔ No [sic:], and the reason is the Pattern 12 test rather than the frequency count: a proper name renders in place. "Abrathar the son" is perfectly good English; nothing has to be carried untranslated, so there is no carried type for the marker to wrap. Rendered as printed, logged, and worth the plate read while §0's are being done.

0189C · in crimen reputant., quod — a period AND a comma, both carried

Our Latin prints reputant., quod at VERS. 5. Pattern 8 forbids deleting either mark, so the English carries both: "they reckon it for a crime., that at his word the hand was healed." It will read as our defect and it is not ours. ⚠ Very likely the digitization rather than the plate — a stray period is the classic transcription artifact — but that is a guess and I make no claim. Plate candidate. The sentence's own parse is also loose: Magnum studium nequitiae is taken as an exclamatory nominative in apposition, with quod… as the object of in crimen reputant; the alternative (Magnum studium nequitiae as the object) makes them reckon the zeal, not the healing, a crime, which the argument denies.

0191A · in pluviam spargerent in terris. quoniam Dominus — lowercase after a full stop

Carried as printed, both the stop and the lowercase opening. Same class as the above and the same caveat: an artifact of either layer, unread.

0183A · Pharao dimissus ab Israel, persequimur eum

"Pharao, let go by Israel, we pursue him." The Exodus logic runs the other way (Pharaoh pursues Israel), and persequitur would give it — but persequimur is a real, well-formed first-person plural in a sentence whose subject shifts to "we" for the moral application (diabolus contemptus surgit in scandala follows). Rendered as printed; persequitur named and not adopted.

0190B · Uterque with plural participles

Uterque et plagas habentes, et immundi spiritus. A singular pronoun over two plural members. Rendered "Each of the two, both they that had plagues, and the unclean spirits" — the number stands in the English because English has an exponent for it, unlike the gender mismatches Pattern 9 exempts. No emendation; nothing to mark.

0180B · commuta — an imperative in a resumption tail

…et sic mundatos Christianam religionem promittentes in Ecclesiae membra commuta. Second person singular imperative closing an abridged Bede extract whose frame has been removed by the usque ad. commutat would give a third-person indicative. Rendered as the imperative Migne prints — "and so change those who are cleansed…" — under the standing rule that the usque ad formula removes words, it does not suspend grammar; a tail that reads oddly is not evidence the abridgment made it odd. Not adopted: commutat.

0180C · qui baptismum, etc. — NO VERB SUPPLIED

Debent enim melioris vitae novitatem promittere, qui baptismum, etc. The relative clause is cut off before its verb by Migne's own etc. The English stops where he stops — "who baptism, etc." — under Tobit's rule that no verb is ever supplied to complete an abridgment. It is abrupt because the plate is abrupt. Same treatment at 0190D, ut virtus ostensa fidem, etc., which is rendered "that the power shown, faith, etc."

⚑ usque ad — 38 formula occurrences and TWO look-alikes, decided per occurrence

Brief addendum D. In this range the plate's own italics do the work and were used as the detector, then each was checked from what stands on either side:

• 38 italic usque ad are the abridgment formula, rendered as far as — the corpus's fixed English for it (1,875 instances; zero alternatives anywhere).
• 0180D, roman: …etc., usque ad et usque ad centum quadraginta quatuor millia pervenit. The second one is not the formula — it is reaching up to the 144,000 of the Apocalypse, and it is set roman while the formula beside it is italic. Rendered as ordinary prose.
• 0181B and 0191B, roman usque: — a bare usque with a colon, not usque ad, introducing the closing words of an elided quotation (Propter quod et Deus illum exaltavit; quae non licet homini loqui). Rendered "even to:" in roman, distinct from the formula's italic.

⚑ Unexpressed subjects supplied, and what fixed each one

Every "he/she/it/they" the Latin does not print, named:

• 0181A saltus dat, culmina pratorum carpit, palmas in vertice portat … summatim gustat — four third-singular verbs whose subject is cervus, from the simile's own opening (Sicut cervus, Ps 41:2); palmas in vertice are a stag's antler-palms, which is the tell. Named as "the hart" so no bare pronoun carries it.
• 0180D …et usque ad centum quadraginta quatuor millia pervenit — the subject fell inside the usque ad elision and is genuinely unrecoverable. Two readings construe: gratia (the grace the question is about) and Ille qui (the last expressed nominative before the cut). Rendered "it reaches", which takes the first. Nothing decides it and the crux is the record, per the brief's instruction where two readings both construe.
• 0184C Leprosus … mundatur qui non vult mortem peccatoris — the relative cannot be the leper: the words are God's, from Ezek 33:11, and the antecedent is Domini four words back. Rendered "is made clean by him who willeth not the death of the sinner," binding it explicitly rather than leaving an English relative that would attach to the leper. This is Ruth @0538C's class (the Noemi site) and it was worked the same way — from the verse the words come from, not from the sentence.
• 0186A Cum grabato ante Jesum deponitur — the paralytic, from Paralyticum at VERS. 3 and grabato itself.
• 0185C Sobrietatem et sapientiam docet — Wisdom, the subject of Sap 8 and of the book Migne names in the citing sentence.
• 0184D–0185A tribuit · dubitat · praesumit · tangit · probat — the leper for the first three, Christ for the last two; the switch is marked in the Latin by Dominus legis est.

⚑ Negation — every particle named to its host, and the four correlative sites

135 negative particles across the four gloss chunks, each carried, none supplied. The four that sat on the test-2a trigger (a negative on a verb with a quantifier or correlative adjacent) were each proved from the sentence's structure, not from what read well:

• 0182C non ex majestate sensuum, sed ex verborum judicant utilitate — host is the prepositional phrase, proved by sed ex … utilitate; the verb judicant stays affirmative. (This is 8960 @0740B's model case, handled the same way.)
• 0182D non subductis ad terram navibus tanquam cura redeundi, sed tanquam vocantem et jubentem secuti sunt — the sed pairs the two tanquam phrases, so the negative governs the first of them and not secuti sunt: "they followed him, not with the ships drawn up to land as though with a care of returning, but as one calling and commanding."
• 0185A Non ideo tangit, quia sine tactu curare non poterat, sed sic probat — host is ideo, i.e. the reason, not the touching. He does touch.
• 0184B Neque enim solo … — see the entry above. Also worked, all straightforward non … sed on a noun or a phrase rather than the adjacent verb: 0179C non societate naturae, sed officii dignitate · 0180B non mundatis non proficit, sed in deterius mutat (two negatives, two different hosts) · 0183B Non est hic voluntatis confessio, sed necessitatis extorsio · 0185B non legis ordine, sed gratia · 0185C Non omnibus manifestatus est Jesus … sed his · 0186D non solum cives, sed et maris habitatores · 0190B non solum timore coacti ad procidendum, verum etiam ad confitendum · 0190B Non solum daemones … sed et apostoli · 0191C non solum remissio peccatorum, sed etiam in apostolorum numero esse · 0190D linguae in signum sunt non fidelibus, sed infidelibus.

⚑ Punctuation — compared BAND BY BAND, not by total (test 3), and it caught nineteen sites

Column anchors partition both twins identically, so : ; ? ! were counted per band across all 41 anchors and every mismatch closed. Nineteen supplied marks were found and removed — sixteen semicolons and three colons that English habit had introduced where Migne prints a comma or nothing (e.g. 0179A Salvator Latine, Christus Graece had become "in Latin; Christus"; 0186A Mira humilitas, despectum had become "humility: him"), plus one colon restored at 0190B (in eis daemones: non solum) that had been softened to a comma. The totals had NOT cancelled here, but a per-chunk total would still have hidden the 0190B restoration behind the removals in the same chunk — which is the reason the test is per band. Final state: zero mismatched band/mark pairs in all five chunks.

Question marks: 6 in the range, all direct questions, all preserved in place (0180D Quis fortior est gratia…? · 0182C Quem non inebriat eloquentia saecularis? · 0183A Quidnam est? · 0183D Delicta quis intelligit? · 0188C An non Davidem … meministis? · 0188D cur in his causam improbatis…?). No Pattern 8b site exists in this range: the one Quaeritur is at 0182D (Quaeritur quomodo primum Petrum … vocaverit) and Migne closes it with a period, so it is rendered as an indirect question with no mark and no recasting. The single ! (0189A quanto magis Filius hominis verus Rex et Sacerdos!) is preserved.

⚑ Pattern 17 — thou/you

thou throughout for a printed singular: 0185A Vides, 0187D agnosceres, 0188D Non te turbet, 0190D te attentum reddit and Jo 21:18's quo tu non vis. you for the printed plural at 0188C (meministis, improbatis — the Pharisees) and inside quoted scripture that is plural (gratis accepistis, gratis date; Diligite inimicos vestros; Nisi manducaveritis).

§4 · SECTIONS READ CLEAN, AND CANDIDATES REJECTED WITH REASONS

Read clean — walked line by line and nothing raised: chunk 0000 entire · 0179B–0180A (VERS. 1–3, the Isidore etymologies) · 0181C–0182A (VERS. 10–13) · 0183B–0183D (VERS. 24–33) · 0186B–0186D (VERS. 6–13, the Arian polemic and the five causes) · 0188A–0188B (VERS. 21–23 after the aestruere site) · 0189D–0190A (VERS. 6–9) · 0191C–0192A (the apostle etymologies Philip through Thaddeus) · 0192C (the Judas etymology).

Rejected, with the reason:

• Potinum (0185A) for Photinum — Migne's spelling of Photinus, an attested Latin form of the name and the sort of orthography Pattern 9 exempts. Rendered "Potinus". No marker, no crux beyond this line.
• calcamentorum (0180D) for calceamentorum — attested medieval orthography (Pattern 9). Rendered normally, no italics.
• convexans (0183B) — flagged and dropped: Migne himself frames it as a variant reading (sicut quidam codices habent), so it is a reported reading, not a defect, and convexare is attested late Latin. Rendered "shaking".
• Scariotem / Scarioth (0192B–C) — Migne prints both forms within four lines. Both are kept ("Scariot" for the accusative lemma, "Scarioth" for the indeclinable), on Ruth's Lia precedent that Migne's name-forms stand. Not an inconsistency of ours.
• Andraeas / Andream (0182B, 0191B) — spelling variation in Migne; "Andrew" throughout, since English has no exponent and no gloss depends on it.
• (cont.) on the repeated CAPUT PRIMUM head (chunk 0002) — runbook false positive 7. Not a defect, marker not translated.
• Column-band arithmetic — the range runs 0179A→0192D with bands appearing as Migne set them; no gap was reported and none should be (runbook false positive 1). No text breaks mid-word or mid-construction across any anchor in this range. ⚑ Two anchors DO split a construction and were placed to match: non [0179C] societate naturae and non [0185A] obsequitur legi, both rendered with the anchor inside the English construction at the same juncture (the 8995 @0605B precedent). Every anchor was checked against the word it stands beside, not merely for presence and order (test 2b).
• Pattern 4 [f:] inline locators — not applicable, this is lemma-and-gloss commentary (runbook false positive 4). None tagged.
• A lemma expanded vs. a one-word lemma left as it stands — not an inconsistency; the two are different classes and contrasting them has been declined three times. Expanded here as fragments: Non necesse → "They have no need" · Quid est → "Which is easier" · Nemo assumentum → "No man seweth a piece" · Apprehensa → "Taking her by the hand" · Cum sabbatis → "When on the sabbath days." Left as they stand, because nothing is elided: Clamantis · Parate · Joannes · Baptizans · Spiritus · Poenitemini · Obmutesce · Discerpens · Paralyticum · Fili · Levi · Publicani · Animam · Perdere · Secessit · Navicula · Thaddaeum · Jacobum · Joannem · Filii.
• Suspiciously tidy forms (brief §4, the normalization class): the neighbours of every line examined closely were read in the same look, per addendum C. Nothing struck me as tidied — but that is a negative recorded, not a clearance: a normalized form reads perfectly and the only handle is a tidy word in a rough passage. The roughest passages here (0189C's reputant.,, 0191A's lowercase quoniam) are rough in their punctuation, not their words.

§5 · WHAT I WANT THE PLATE READ FOR, in priority order

1. 0183C Femmam · 0187C Nunquia · 0188A aestruere — the three [sic:]. Expect withdrawals; expect tei-patches. 0188A also decides the English: destruere gives "do away with", astruere would reverse it.
2. 0182A mysterium for ministerium, and 0188C Abrathar for Abiathar — real words, no markers fired, both plausibly ours.
3. 0186D Fecit eis convivium magnum Levi — one letter against the Clementine's ei, and the strongest declined [var:] in the list.
4. 0189C reputant., and 0191A in terris. quoniam — punctuation artifacts of unknown layer; cheap to settle on the same page pass as (1) and (2), since 0188A–0189C sit together.
5. 0184B solo for solum — three letters, and the case has no host on the page.
6. 0187A — the unclosed parenthesis after Ipse, to see whether Migne closes it and our file lost it.

⚠ The launch brief records that our PL 114 scan clips both margins on p. 95 (cols 179/180), losing 2–4 characters off the right column's line-ends. Nothing I have raised falls in 0179–0180 except the calcamentorum spelling, which is declined anyway — so the clipping should not obstruct any of the six above. Per the two-witness rule, data/volumes.json lists PL 114's other scans if a line will not come clear.

<!-- ===== stint file: cruces-0005.md, verbatim ===== -->

Stint file. Cols 0193A–0205A, CAPUT III (tail) through CAPUT VI, 4,786 Latin words, 187 italic spans (every span the launch brief lists for this range except the fifteen at 0205A that open CAPUT VII — those belong to chunk 0010, not to this stint; see §6.1).

Written to be merged with the other four stints' files. Nothing below has been silently altered after the fact.

§1. THE POLICY THIS STINT APPLIED, stated before the findings so it can be judged

1. Pattern 7 throughout. Every printed reading is rendered as printed. Where a real, well-formed Latin word reads oddly, it is rendered as printed and logged here with no marker, however good the internal argument — per addendum A of the brief and the launch instruction. Ten such sites are listed in §4.
2. Markers fired: one [sic:], three [var:], zero [d:], [ed:], [cj:]. The [sic:] sits on a non-word (§3.1). The three [var:] each rest on sources/vulgate/clementine-flat.txt, cited by book chapter:verse, and each sits where the gloss's own exposition turns on the divergent word — that is the bar this stint used, and it is stated here so a reader can see which divergences were not marked and why.
3. Divergences that are real but do not bear on the gloss take no marker — orthography, word order, a dropped autem, a compressed citation, a wrong [n:] reference. They are all listed in §2 anyway, because a list of findings alone cannot be told apart from a report by an agent that never looked.
4. Attribution is never presented as settled. Where a defect could be Migne's type or Corpus Corporum's transcription, both are named and neither is asserted. (9004: the stints' guesses about whose a defect was were wrong 9 times in 14.)

§2. THE LEMMA DIVERGENCE LIST — all 187 spans, verdict for each

Verdicts: = agrees with the Clementine exactly (allowing coeli/caeli-class orthography and initial capitals) · frag a clipped lemma, agreeing so far as it goes · DIV a real divergence, itemized below the table · ref the wording agrees but Migne's [n:] locator does not resolve to it.

2.1 Chunk 0005 — CAPUT III tail, cols 0193A–0194A (16 spans)

• (a) 0193A Ab Hierosolymis. The Clementine spells Jerosolymis, and Migne's own gloss prose four lines later spells it Jerosolymis twice (a Galilaea et Jerosolymis · Scribae ab Jerosolymis). Latin orthography of a proper name has no English exponent (Pattern 9 family) and rendering the lemma "Hierosolyma" against the prose's "Jerusalem" would manufacture an inconsistency inside one paragraph. Rendered "From Jerusalem." in both places; no marker.
• (b) 0193C Qui enim fecerit for the Clementine's Qui autem fecerit (Mt 5:19). One particle, no change of assertion. English reads "For he that shall do and teach" — note that 8947/0011 renders the same verse "But he that shall do and teach," from its plate; the two Englishes correctly differ. No marker.
• (c) 0193C Qui blasphemaverit in Spiritum sanctum vel verbum, etc., cited to Matth. XII. Mt 12:31–32 reads Spiritus autem blasphemia non remittetur and qui autem dixerit contra Spiritum Sanctum; Migne's cue conflates 12:31 with the verbum contra Filium hominis of 12:32 under one verb. Declined a [var:]: the span is a cue closing in etc., the gloss hangs on the part elided (neque in hoc saeculo neque in futuro), and Pattern 14's "received reading" cannot be stated cleanly for a conflation. Rendered as printed; logged here.
• (d) 0193D Honora patrem et matrem for Ex 20:12 Honora patrem tuum et matrem tuam — the possessives dropped, i.e. the New Testament's own citation form (Mt 15:4, Mc 7:10). Rendered as printed; no marker.

2.2 Chunk 0006 — CAPUT IV, cols 0194B–0196D (42 spans)

• (e) 0194D Aliud cecidit secus viam. Mark reads circa viam; secus viam is Luke's wording (Lc 8:5), and the gloss's very next words quote Luke. A harmonization, not a change of sense (both = "by the way side"). No marker; the English reads "by the way side" either way, which is why this one is here and not in the text.
• (f) 0194D Concupisces sapientiam: serva mandata, [n: (Prov. VII)] — [var:] FIRED. Prv 7:2 reads serva mandata mea, et vives; no clause about coveting wisdom stands anywhere in Proverbs. The nearest received wording is Sir 6:37 concupiscentia sapientiae dabitur tibi — where the desire is given, not commanded. The marker fires because the gloss's whole point is that the desire is a precept (Quod praeceptum est. Praecipitur enim), so a reader must be able to see that the precept is not in the cited book.
• (g) 0194D [n: (Psal. III)] — the words are Ps 111:1 verbatim (Beatus vir qui timet Dominum, in mandatis ejus volet nimis). The locator is wrong (III for CXI). Locators pass through verbatim (brief §4, runbook false positive 2): reproduced as printed, no marker.
• (h) 0195B quando illuminabit Deus abscondita tenebrarum, et manifestabit consilia cordium [n: (Matth. X)]. This is 1 Cor 4:5, not Matthew, and the Clementine reads qui et illuminabit — Migne supplies Deus as the subject where Paul has a relative. No [var:]: supplying the subject changes nothing the gloss uses (the gloss uses only the fact of a disclosure at the judgment). Locator reproduced as printed.
• (i) 0195C Quoniam adest messis for Mc 4:29 quoniam adsit messis — indicative for subjunctive. The gloss expounds messis (= consummatio saeculi), not the mood. No marker.
• (j) 0196A Seorsum autem discipuli. Mc 4:34 reads seorsum autem discipulis suis disserebat omnia — dative, governed by disserebat; Migne's lemma is nominative discipuli. A one-letter difference, and it is exactly the class where the plate is worth a look (see §5.2). It construes as printed — the gloss's own continuation (Illi enim digni erant audire) takes the disciples as subject — so it is rendered "But apart, the disciples." with no marker.
• (k) 0196C Migne: incredulitatem illorum · quia hi qui viderunt eum resurgere. Mc 16:14: incredulitatem eorum · quia iis, qui viderant eum resurrexisse. Free re-quotation, no change of assertion. No marker.
• (l) 0196D Migne drops haec and prints sic for ita (Lc 24:26 Nonne haec oportuit pati Christum, et ita intrare). No marker. The ? Migne closes it with is preserved.

2.3 Chunk 0007 — CAPUT V, cols 0197A–0200B (61 spans)

Every one of the following agrees with the Clementine at its own verse, allowing for clipping: Et venerunt, (Mc 5:1) · Homo in spiritu. (5:2) · Et neque catenis, (5:3) · Torqueas. (5:7) · Quod tibi nomen est? (5:9) · Legio mihi nomen est. (5:9) · Mitte nos. (5:12) · Grex praecipitatus est. (5:13) · Qui autem. (5:14) · Et vident illum. (5:15) · Vestitum. (5:15) · Et timuerunt. (5:15) · Et rogare coeperunt eum. (5:17) · Vade in domum tuam. (5:19) · Et venit quidam, (5:22) · Quoniam filia mea. (5:23) · Et comprimebant. (5:24) · Et mulier, (5:25) · Multa perpessa, (5:26) · Venit in turba. (5:27) · Tetigit. (5:31) · Adhuc, (5:35) · Archisynagogo. (5:36) · Et non admisit, (5:37) · Flentes et ejulantes. (5:38) · Non est mortua. (5:39) · Ejectis omnibus. (5:40) · Talitha cumi. and Tibi dico, surge, (5:41) · Si quis mihi ministrat, me sequatur (Jo 12:26) · Noli me tangere: nondum enim ascendi ad Patrem meum (Jo 20:17) · Justificata est sapientia ab omnibus filiis suis (see (q)) · mortuos sepelire mortuos suos (Mt 8:22) · plus the seventeen usque ad abridgment spans and illuminans / illuminatus: (gloss-Latin, not scripture).

Divergences:

• (m) 0197C Et rogabant eum ne imperaret illis, ut in abyssum irent [Luc. VIII]. Lc 8:31 reads rogabant illum. One pronoun. No marker.
• (n) 0197D Sedentem ad pedes ejus. attributed to Lucas. Lc 8:35 has the words but split: invenerunt hominem sedentem, a quo daemonia exierant, vestitum ac sana mente, ad pedes ejus. Migne closes the gap. The very next lemma (Vestitum.) is the word he jumped over, so the reader has both. No marker.
• (o) 0198A Exi a me, Domine, quia homo peccator sum, etc. [Luc. V]. Lc 5:8 puts Domine last: Exi a me, quia homo peccator sum, Domine. Word order only. No marker.
• (p) 0198C Quia filia unica erat illi fere annorum duodecim, et haec moriebatur [Luc. VIII]. Lc 8:42: quia unica filia erat ei. Order + pronoun. No marker.
• (q) 0199A Post Dominum Deum ambulabis [Ose. XI]. Os 11:10 reads Post Dominum ambulabunt — third plural, and no Deum. Migne's second-singular future is closer to the Deuteronomic formula. [var:] DECLINED: the gloss turns on post ("she does not see God present in the flesh, but afterwards…"), which both readings carry, and neither the added Deum nor the person touches it. Rendered "Thou shalt walk after the Lord God."
• (r) 0199C Super cathedram Moysis sederunt scribae et Pharisaei [n: (Matth. XXVIII)]. Mt 23:2 spells Moysi; and the locator says XXVIII where the verse is XXIII. Orthography + a wrong locator, both reproduced as printed. No marker.
• (s) 0199C Justificata est sapientia ab omnibus filiis suis [n: (Matth. XI)]. The words are Luke's (Lc 7:35, ab omnibus filiis suis); Mt 11:19 has a filiis suis, without omnibus. A cross-gospel harmonization under a Matthew locator. No marker — nothing in the gloss (mysteries revealed to the faithful, not to mockers) turns on omnibus.
• (t) 0199C Laetentur omnes qui sperant in Domino, et in aeternum exsultabunt [Psal. V]. Ps 5:12: Laetentur omnes qui sperant in te; in aeternum exsultabunt. Second person made third, one et added. No marker.
• (u) 0199D Nolumus vos ignorare, fratres, de dormientibus [I Thess. IV]. 1Thes 4:13 has Nolumus autem vos ignorare. A dropped particle in a lemma cut for citation. No marker.
• (v) 0200A Credidit Abraham Deo, et reputatum est ei ad justitiam [Jac. II]. Jac 2:23 (and Rom 4:3, Gal 3:6) all read illi. One pronoun. No marker.

2.4 Chunk 0008 — CAPUT VI, cols 0200C–0204B (56 spans)

Agreeing exactly (allowing clipping): Unde huic, (Mc 6:2) · Faber. (6:3) · Non est propheta. (6:4) · Et non poterat, (6:5) · Et mirabatur. (6:6) · Et vocavit. and Et dabat illis, (6:7) · Ne quid tollerent, (6:8) · Calceatos sandaliis. and Ne induerentur duabus tunicis. (6:9) · Excutite pulverem de pedibus. (6:11) · Et ungebant oleo. (6:13) · Et audivit rex. (6:14) · Philippi fratris sui. (6:17) · Non licet tibi habere, (6:18) · Et juravit. (6:23) · Et contristatus, (6:26) · Caput ejus in disco. (6:27) · Et tulerunt. (6:29) · Et convenientes. and Omnia. (6:30) · Venite seorsum. / Requiescite. / Erant enim, (6:31) · Et pedestres. (6:33) · Et misertus est. (6:34) · Et cum jam hora. (6:35) · Date illis. (6:37) · Quinque. (6:38) · Secundum contubernia. (6:39) · Per centenos et quinquagenos. (6:40) · Et acceptis. / Fregit. (6:41) · Et sustulerunt, (6:43) · Quinque millia. (6:44) · Et statim coegit. / Ad Bethsaidam. (6:45) · Abiit in montem. (6:46) · Infirmatur quis in vobis?…ungentes eum oleo (Jac 5:14, exact to the point where etc. cuts) · plus twelve usque ad spans.

Divergences:

• (w) 0200B Ipse baptizabit vos in igne et spiritu [n: (Joan. I)] — [var:] FIRED. The Clementine has this only at Mt 3:11 / Lc 3:16, and there in the reverse order and with Sancto: ipse vos baptizabit in Spiritu Sancto, et igni. Jo 1:33 (the locator Migne gives) has only qui baptizat in Spiritu Sancto — no fire at all. The marker fires because the gloss is built on the order: Faber is glossed qui operatur igni et spiritu, and Migne adduces the verse to prove that pairing. Under the received order the proof-text stops proving it. ⚑ Rendering note: faber is Englished "craftsman" and fabricatus est "crafted", so that Faber est qui fabricatus est auroram et solem keeps the figura etymologica the gloss is made of. "Carpenter" (the received English of Mc 6:3) would leave "he who works with fire" explaining nothing — the stylus/epistylium rule in translation-style, Song-of-Songs §2.
• (x) 0200C Quis est filius Isai [n: (III Reg. XII)]. 3Rg 12:16 reads vel quae haereditas in filio Isai; Quis est filius Isai is Nabal's, 1Rg 25:10. Migne's words are a real verse under the wrong locator. Locator reproduced as printed, no marker — the gloss needs only "meanness attends a man's origin", which either verse supplies.
• (y) 0200C Humilia Dominus respicit, et alta a longe cognoscit [Psal. CXXXVII]. Ps 137:6: quoniam excelsus Dominus, et humilia respicit, et alta a longe cognoscit. Migne compresses the first half. Same assertion. No marker.
• (z) 0200D In Jesu nomine surge et ambula [Act. III]. Act 3:6: in nomine Jesu Christi Nazareni surge, et ambula. Compression. No marker.
• (aa) 0201B Audivit Herodes tetrarcha omnia quae fiebant, etc. [Luc. IX]. Lc 9:7 has Audivit autem Herodes. A dropped particle at the head of a citation. No marker.
• (bb) 0202A Abierunt in deserto. Mc 6:32 reads abierunt in desertum locum seorsum. Migne's lemma is the ablative in deserto against the Vulgate's accusative of motion, and drops locum seorsum. [var:] DECLINED and this is a judgment call I want recorded: the case difference does have an English exponent ("in" vs "into"), but the span is also a truncation, and the gloss (Petendo solitudinem…) is indifferent to the case. Rendered literally, "They went away in the desert.", and named as a plate candidate in §5.2 — a two-letter ending is exactly the shape Corpus Corporum alters in both directions.
• (cc) 0202C Et misertus est eis, et curavit languores eorum. — [var:] FIRED. Attributed to Matthew; Mt 14:14 reads curavit languidos eorum — the sick, not their infirmities. The marker fires because Migne's own next sentence expounds the Vulgate word, not his own: et languidos curare. This is the 7a″ shape exactly, running the other way: the quotation has drifted off the word the exposition uses. Both are rendered as printed — "healed their infirmities" in the quotation, "the languishing to heal" in the gloss — and the marker records the divergence.

2.5 Chunk 0009 — CAPUT VI cont., cols 0204C–0205A (12 spans)

All twelve agree with the Clementine: Et ipse solus, (Mc 6:47) · Et circa quartam. and Et volebat praeterire eos. (6:48) · Putaverunt. (6:49) · Nolite timere. (6:50) · Non enim intellexerunt. (6:52) · Cognoverunt. (6:54) · Fimbriam vestimenti. (6:56) · Ut quid, Domine, recessisti longe (Ps 9:22 in the Clementine's numbering, which is Migne's Psal. X: a numbering system, not a divergence — the words agree letter for letter) · plus three usque ad spans. The Isaiah tail inside the abridgment at 0204C (Cum ambulaveris in igne, non combureris, et flamma non ardebit in te) is Is 43:2 verbatim and is set roman, not italic, in the Latin; it is carried roman.

§3. MARKERS FIRED — the four public claims this stint makes

3.1 [sic: magnitudino] — 0203A, chunk 0008. The only [sic:] in the range.

Printed: ut illis se non habere testantibus, magnitudino signi notesceret. magnitudino is not a Latin word in any inflection: magnitudo, -inis has no such form, and the clause needs the nominative subject of notesceret. Corpus-checked: the form occurs nowhere else in this work.

Rendered "the greatness [sic: magnitudino] of the sign might become known" — 7a⁗ applied: the marker records the type, and the sense is supplied in plain words beside it, so the clause keeps both its subject and its verb when the bracket is stripped. (Fourth question, per the brief: the clause still has its object of knowledge, signi → "of the sign".)

⛔ Attribution not settled. Addendum A says a non-word suspicion has been right every time that a defect exists, and 9004 says the guess about whose it is was wrong nine times in fourteen. If the plate reads magnitudo, this marker must be withdrawn and the fix is a data/tei-patches/9001.json entry, not an apparatus note. This is the highest-priority line in §5.1.

3.2 [var:] ×3

Each content string was checked not to appear verbatim in the Latin twin (Pattern 14's mirror test), and each cites the Clementine by book chapter:verse, per the brief's [var:] rule.

3.3 Zero [d:], zero [ed:], zero [cj:]

• No dittography anywhere in the range. Checked by eye at every etc. and every abridgment seam, which is where an eye-skip would sit.
• No [ed:]: nothing in these five chunks reads as digitization loss. The frontmatter noteCount matches the body [n:] count in all five chunks (14/25/40/34/6, with the incipit's note not double-counted), and no list, heading or verse address is missing.
• [cj:] considered twice and declined both times, per §4 (b) and §4 (c) below. Neither clears Wilson's bar — the printed word must make the sentence say something the author is arguing against — and neither falls in the narrow negation/correlative class the 2026-08-15 and 2026-08-17 rulings admit.

§4. RENDERED AS PRINTED, LOGGED, NO MARKER — the real-word class

Ten sites. Every one is a real, well-formed Latin word or phrase, so addendum A governs: Migne's until a plate says otherwise, however good the argument. The argument is given anyway, because it is what a plate reader would want to know.

(a) 0193C — Ubi ostendit quaedam revelari in futuro. (chunk 0005) The frame is Mt 12:32, neque in hoc saeculo neque in futuro — about remission, and the gloss's own next sentence is about pardon (quorum etiam haec vita veniam promeruit). One expects remitti. But revelari is a real infinitive in its slot and yields a defensible sense (what is forgiven is disclosed hereafter). Rendered "certain things are revealed in the world to come." Plate candidate, low priority.

(b) 0197A — Ruptis catenis (ut Lucas scribit) subito agebatur a Domino in deserto. (chunk 0007) ⚠ The single most consequential reading in this range. Lc 8:29 reads Et ruptis vinculis agebatur a daemonio in deserta, and Bede's source has the demon. As printed, Migne's sentence says the demoniac was driven into the desert by the Lord, which is the opposite agency, in a paragraph about the Gentiles bursting the restraint of law and being carried off by lust. A [cj:] was considered and declined: Domino is a real word in the right case, the sentence is a paraphrase in the gloss's own prose (so Pattern 14 does not reach it), and the printed reading is not one the author is visibly arguing against — it can be read as the Lord permitting the driving. Rendered "he was suddenly driven by the Lord into the desert." ⭐ Ask the plate. If Migne prints a daemonio, this is a Corpus Corporum substitution of one real word for another — the concilio/consilio class (brief §4) — and belongs in a tei-patch.

(c) 0197A — Catenis et compedibus graves, et duae leges gentium signantur. (chunk 0007) The sentence wants gravibus (ablative, with compedibus) or a different construction: as printed, a nominative plural graves has no subject to agree with and the clause does not construe. Rendered literally, "Heavy with chains and fetters, and the two laws of the Gentiles are signified…" A [cj:] was declined: a case anomaly that merely fails to construe stays declined under the 2026-08-15 scoping (the efferte / suscepit class), and the sentence asserts nothing false.

(d) 0194D — Via mens, metu carnalium cogitationum trita. (chunk 0006) trita ("worn smooth") wants meatu — the traffic of carnal thoughts wearing the road — which is the received figure and the only one that makes trita mean anything. metu ("by fear") is a real noun in the right case. Rendered "worn by the fear of carnal thoughts." Plate candidate: a single dropped a is a shape the digitization produces.

(e) 0196A — quae tanto apertius excrescunt, viliores fiunt. (chunk 0006) A stranded tanto with no quanto. The 11535 ruling admits a stranded comparative whose answering member is printed; here neither member's partner is printed, so the sentence is merely elliptical, not false. Rendered "which grow up so much the more openly, become the more worthless." No marker.

(f) 0198B — quia unus eorum clarioris nominis, est famosioris. (chunk 0007) Reads as though an et has been lost (clarioris nominis et famosioris), or est is intrusive. Rendered as printed: "because one of them was of a more illustrious name, is of a more famous." A [d:] was considered — a doubled comparative could be an eye-skip — and declined: the two comparatives are different words, so nothing is repeated, and Pattern 11 has nothing to wrap.

(g) 0200D — datur eis potestas praecepta docendi. (chunk 0008) Two nominatives/accusatives in apposition with no connective. Rendered "there is given them power, precepts of teaching." No marker; the asyndeton is characteristic of this abridgment.

(h) 0202A — Non solum requisitionis causa. (chunk 0008) The lemma is Venite seorsum … et requiescite, so one expects requietionis / requiei ("for the sake of rest"). requisitio ("a seeking again") is a real noun. Rendered "Not only for the sake of a seeking again." Plate candidate, low priority.

(i) 0203A — cursus habentis saeculi. (chunk 0008) Bede's De tabernaculo tradition reads cursus labentis saeculi, "the course of the gliding age", and labentis is what the list of the five books' contents wants. habentis is a real participle. Rendered "the course of the age that now holds." Plate candidate — an l/h confusion is a scan shape, and this is the strongest of the real-word cases after (b).

(j) 0205A — Rumore, non facie vel magnitudine signorum: etiam vultu plurimis notus erat. (chunk 0009) The sentence contradicts itself: they knew him by report, not by his face, and then he was known to very many by his countenance. Under 7a the contradiction is carried: "By report, not by his face nor by the greatness of his signs: he was also known by his countenance to very many." No word is translated as its opposite to rescue it. (A likely reading is etiam vultu plurimis notus non erat, or paucis; both are conjecture and stay here.)

§5. WHAT I WANT THE PLATE FOR

5.1 Priority order

1. 0203A magnitudino — the range's one [sic:]. Confirm or withdraw.
2. 0197A a Domino (Lc 8:29 = a daemonio) — the reading with the largest consequence. Read the whole line, not the word (addendum C): its neighbours graves and subito are both suspect in the same clause.
3. 0203A cursus habentis saeculi — read habentis / labentis.
4. 0202D multae. fidelium catervae — see 5.2.
5. 0194D metu / meatu · 0196A discipuli / discipulis · 0202A Abierunt in deserto / in desertum locum — three one-or-two-letter endings, cheap to settle if the pages are being pulled anyway.

5.2 The stray full stop at 0202D, chunk 0008 — the one shape I could not resolve

Migne (as our file has him) prints, inside a [n: (BEDA.)] paragraph:

Christum deserta gentium petentem multae. fidelium catervae relictis moenibus priscae conversationis … sequuntur

A full stop stands between the adjective and its noun, and multae … catervae plainly agree. Under Pattern 8 the mark is carried and not deleted, so the English reads "many. bands of the faithful follow" — placed at the same point in the same construction, so that the mark is visible as the source's and not as ours.

⛔ I could not tell whether this is Migne's type or the digitization's. It has the shape of a lost abbreviation point or a line-break artifact (the digitization's), and equally the shape of a compositor's dropped space (Migne's). Both readings are stated; neither is asserted. If the plate has no stop, the repair is a tei-patch and the English loses the period silently. This is the one place in the range where a reader will think the edition has erred, so it is worth a page even though nothing turns on the sense.

5.3 A negative worth recording

Per brief §4 ("record the negatives too"): Matthaeus is spelled Mattheus, once, at 0201A (chunk 0008), against six Matthaeus elsewhere in this range. English has no exponent for it — "Matthew" either way — so nothing is concealed, and Pattern 9 governs. Raised here only so a plate reader glancing at 0201A knows it was seen and not missed.

§6. WHAT WAS READ CLEAN, AND WHAT WAS REJECTED

6.1 Scope

• The brief's last fifteen spans (0205A, CAPUT VII: Et conveniunt. through Nihil est extra,) are NOT in this stint. Chunk 0009 ends at Fimbriam vestimenti (Mc 6:56); src/latin/9001/0010.md opens CAPUT VII with colFirst: 0205B and colContext: 0205A. They belong to whoever holds 0010. Not checked here — do not read their absence from §2 as clearance.
• All 187 spans that are in range are accounted for in §2. None was skipped.

6.2 Read clean, and named so a blind reader knows where I looked

• Verse-address order. All five chunks run strictly ascending (0005: 22–34 · 0006: 1–40 · 0007: 1–41 · 0008: 2–46 · 0009: 47–56). No out-of-sequence address anywhere in the range — the Jude phenomenon does not occur here.
• Question marks. Latin 0 / 4 / 1 / 1 / 0; English identical, band by band. Every mark Migne prints is preserved (Nonne ut super candelabrum ponatur? · Quid timidi estis? · Quid me praesente timetis? · …intrare in gloriam suam? · Quod tibi nomen est? · Infirmatur quis in vobis?) and none is supplied. Pattern 8b was not needed: this work prints no indirect question closed with ? in this range.
• Colons, semicolons, guillemets — compared BY COLUMN BAND, not by total (brief test 3). All 54 bands across the five chunks match at zero mismatched bands. Thirteen mismatches were found and repaired this way before filing, every one of them a colon or semicolon I had supplied where Migne prints a comma; the totals would have hidden at least two of them, which is the reason the test exists. No guillemets occur in this range.
• Column anchors. 49 anchors, verbatim and in order, and each checked against the word it stands beside in the Latin, not merely for presence (test 2b). Nine are deliberately split mid-construction to match Migne: principem [0193B] daemoniorum → "the prince [0193B] of demons" · volet [0195A] nimis → "he shall delight [0195A] exceedingly" · major est [0195D] omnibus oleribus → "greater than [0195D] all herbs" · plures [0197C] sectae → "more [0197C] sects" · annis duodecim [0198D] → "of twelve years [0198D]" · Mittuntur duodecim [0200D] apostoli → "The twelve [0200D] apostles are sent" · ad fractionem [0203A] panis → "to the breaking [0203A] of bread" · ut edant et [0203B] bibant → "that they may eat and [0203B] drink" · praesentiae [0204D] suae → "his own [0204D] presence". Two sit one word downstream of their Latin position because English word order forced it and the alternative was unreadable: 0198A (Latin between legem and dicunt; English before "because" = quia) and 0204C (Latin before dicitur; English before "said").
• usque ad — decided per occurrence (addendum D). 38 occurrences in the range. 36 are the abridgment formula and are rendered as far as, in italics, following 8945/8947 precedent. Two are not, and one of them is the trap: 0201B, quod praedicatio usque ad illos pervenerit — "that the preaching reached even unto them", the words' own sense, inside a gloss on shaking the dust from the feet. Rendered as ordinary prose, not as the formula. The other is 0193A esse persequendum usque ad mortem, rendered ordinary prose, 'even unto death'.
• Negation, word by word, with the host named (tests 2 and 2a). Every non · nec · neque · nisi · nihil · nullus · ne · nunquam in the range appears in the English, and none appears that the Latin does not print. The five sites that fired the 2a trigger condition (a negative with a quantifier or correlative nearby) were each proved from the sentence's own structure: 0193D Nec fratres injuriose contemnit, sed opus spirituale praeficit (the sed pivot puts the negative on the contemning, not on the manner) · 0196A Non quod nullum sermonem … sed quia nullus sermo · 0195C etsi non aeque multum omnes — the negative is on aeque multum, not on omnes: "though not all equally much", not "though not all bear wheat", which the sentence's own first half (omnes enim ferunt frumentum) forbids · 0196C Non quod omnia animata … sed majestate conditoris (the sed shows the denial is of animacy, so English's ambiguity here matches the Latin's) · 0202A/0202B the two Non solum … sed etiam pairs, negative on solum both times.
• Unexpressed subjects — named, per the brief's dominant-defect section. Every supplied "he/she/it/they" was resolved from the lemma above or from the verse it comes from: 0193A gentiles desiderabant → object supplied ("longed for him", i.e. Christ, from illum … perduxit in the same sentence) · 0193C dicebat → the blasphemer of the lemma · 0197A agebatur → the demoniac of Mc 5:2–3, not the Gentile people of the allegory · 0197B dimittit … possidet → the demon (ei in the same clause) · 0197C in abditis agunt → the demons/idolaters left by the freed Gentiles, not the swine · 0199A non videt → the woman with the issue (Mc 5:27) · 0199A emundat → Christ, not the woman (the sacraments are suis vestimentis, his garments) · 0199B quae … substantiam → the woman · 0200A spargit … colligit → anima, feminine, rendered "it" · 0202A venit seorsum, ducit quos eligit → Christ. One site where two readings both construe, recorded rather than chosen: 0199A Quia qui … gravatur, solo fideli Ecclesiae corde tangitur. I take qui gravatur as the Lord (thronged by the crowd of heresies, touched only by the Church's faithful heart), on the strength of the preceding sentence's quem passim comitans turba comprimit. The rival reading makes the burdened man the heretic. The English follows the first; the second is named here because a blind reader will see the ambiguity.

6.3 Candidates raised and rejected, with reasons

• [f:] inline citation tails. Not applicable — runbook false positive 4. This is a lemma-and-gloss commentary; its scripture arrives inside the prose and inside [n:], not as apparatus tails. Nothing tagged.
• Column-band arithmetic. 0194A → 0194B, 0200B → 0200C and the rest are positional quarter-guides. Runbook false positive 1. Not reported.
• (cont.) on the CAPUT III and CAPUT VI heads (chunks 0005 and 0009). Our chunk division, not Migne's. Carried verbatim, untranslated: ## CHAPTER III. (cont.), ## CHAPTER VI. (cont.). Runbook false positive 7.
• Head numerals. CAPUT IV. → CHAPTER IV., CAPUT V. → CHAPTER V., CAPUT VI. → CHAPTER VI. — numerals mirror the plate. No head in this range is spelled out in the Latin, so none is spelled out in the English.
• 0199C — the siglum (ID) set as bare text inside the italics. VERS. 38.-- Flentes et ejulantes. (ID). — no [n: ] wrapper, no internal period, and the sentence's own period falls outside the italics. Every other siglum in the range is [n: (ID.)]. Reproduced exactly as printed, untranslated, italics and stop positions included. Not a defect to mark: it is a transcription-layer difference (this one was never wrapped as a <note>), and the frontmatter noteCount: 40 counts the 40 wrapped notes, so nothing is missing.
• 0199C — ejulantes in the lemma, ululantes in the gloss four lines later (inter flentes et ululantes mortua jacet). Not a defect: Mc 5:38 has ejulantes, the gloss paraphrases. Rendered "wailing" and "howl" respectively — the variation is Migne's and is preserved.
• 0197D — Sedentem ad pedes ejus. followed by lower-case quod signat after a full stop. Migne's own punctuation; carried, lower case and all ("Sitting at his feet. which signifies that…"). Pattern 8: never delete a mark he prints, and no case-repair is ours to make.
• Em-dashes. None used in the five chunks. Not a defect class; recorded only so the merge knows this stint carries none.
• Lemma expansions under convention 3, each named because expansion is a judgment. Four spans were expanded to the phrase they stand for rather than left as stumps: 0197A Homo in spiritu. → "A man with an unclean spirit." (Mc 5:2 in spiritu immundo) · 0197B Torqueas. → "Torment me not." (Mc 5:7 ne me torqueas — ⚠ the expansion imports the ne; it is the Ruth Non te. → "Let no man see thee" precedent exactly, and it is flagged because importing a negation is the one expansion a 7a check should look at twice) · 0197D Qui autem. → "But they that fed them." (Mc 5:14) · 0200C Non est propheta. → "A prophet is not without honour." (Mc 6:4) · 0199A Tetigit. → "Who touched me." (Mc 5:31 Quis me tetigit? — the period Migne prints is kept; no ? is supplied). Bare one-word lemmata with nothing elided were not expanded and stand as they are: Scribae, · Fratres. · Lucerna, · Ascendit, · Vestitum. · Dormiens. · Archisynagogo. · Omnia. · Requiescite. · Quinque. · Fregit. · Putaverunt. · Cognoverunt. ⚑ Per the brief's "name the control's class" note, these two groups are different classes and the difference between them is not an inconsistency.

<!-- ===== stint file: cruces-0010.md, verbatim ===== -->

Stint file. Chunks 0010, 0011, 0012, 0013, 0014, CAPUT VII through CAPUT IX, 4,948 Latin words / 7,795 English. One translator, Opus. To be merged with the other four stints' cruces at assembly.

verify-english.mjs clean on all five chunks (no errors, no warnings): frontmatter verbatim, column anchors 14/16/0/16/2 in order, [n:] 39/34/0/27/4 at parity, question marks 2/5/0/1/2 at parity, ratios 1.54 / 1.59 / 1.43 / 1.59 / 1.59.

Apparatus fired in this range: 4 [sic:] · 3 [var:] · 1 [cj:] · 0 [d:] · 0 [ed:] · 0 [nt:]. Guillemets: zero in the Latin, zero in the English (checked band by band).

⛔ Nothing in this range has been read at Migne's plate by me. Every [sic:] and [var:] below is listed again in §5 with the column it stands on, so plate-gate.mjs has a work list. I have stated for each what I think happened and, per the 9004 measurement, I do not present any attribution as settled — the stints' whose-is-it guesses there were wrong 9 times in 14.

§1 · THE MARKERS, ONE BY ONE

[sic:] — four, all of them a form that has no grammatical slot

Rendered and marked in every case (7a⁗): strip the four markers out of the English and every clause still reads aloud with its verb, its negation and its object.

⚠ The one I am least sure of is Ex, and it is the one that is a real word. Two readings are open and I cannot choose between them from the file: (a) Migne (or the digitization) set Ex for Et; (b) the compositor's eye slipped and re-set the first syllable of exspuens, which would be a partial dittography and would take [d:], not [sic:]. I took [sic:] because it makes the weaker claim — that the type as printed will not construe — and because a [d:] would additionally assert which copy is intrusive. If the plate reads Et exspuens, withdraw the marker and patch the Latin; that outcome would not surprise me.

[var:] — three, and only where the sense turns on it

1. 0217A · Eccli. 12:13. Migne prints Quis medebitur incantatori a serpente percusso? The Clementine (Sir 12:13) reads miserebitur — "who will pity a charmer struck by a serpent". Material, and the gloss is built on the divergent word: its question is quo alio doctore corrigitur — by what other teacher is the apostate corrected. Healing, not pity, is what the sentence is arguing about. [var: Vulg. Eccli. 12:13 reads miserebitur, "shall have pity on"]. ⚑ Migne's own citation (Eccli. XII.) is correct here.
2. 0211D · Matt. 4:10. The gloss's whole point is a contrast of two forms of address — Diabolo dicitur: Vade retro, Satanas. Petro: Vade retro me. The Clementine at Matt 4:10 reads simply Vade Satana (and Lk 4:8 has no vade at all); retro appears in the Vulgate only at Mark 8:33, Vade retro me Satana, which is the second member of the very contrast. So the distinction the gloss draws does not exist in the Vulgate's wording. Rendered with one English word for retro in both members ("Go behind" / "Go behind me"), so the contrast survives; [var:] records the received reading.
3. 0215D · Mark 9:36. Migne re-quotes his own lemma as Quisquis unum ex hujusmodi pueris receperit, me recipit — without in nomine meo, which the Clementine has, and with no etc. to mark the omission. This is not ordinary compression: the very next clause of the gloss reasons from the missing phrase (intelligit Joannes quod in nomine ejus eos non recipiant). Without the [var:] a reader meets a gloss whose hinge is not in its own proof-text.

[cj:] — one

0212D–0213A · nequeant. Migne: in judicio, in forma servi et bonis et malis, ut impii quem sprevere, quem Judaei negavere, milites quem crucifixere, Pilatus Herodesque quem judicavere nequeant judicem agnoscere. Rendered as printed — "may not be able to recognize the judge" — and glossed [cj: nequeant; read queant, "may be able"].

The bar is met on Pattern 18's own terms and on 18a's negation clause: the sentence's stated reason for his appearing in forma servi et bonis et malis is precisely that those who spurned him should know their judge, and the printed negative makes the page assert the opposite of the clause it is the purpose of. The word that went wrong is a negation, which is the class 18a admits even where the result is merely incoherent. The gloss is additive: Migne's word keeps its English and the conjecture stands beside it.

⛔ Rival considered and rejected: reading nequeant … agnoscere as a litotes ("cannot fail to recognize"). Latin does not license that without a second negative, and supplying one would be the silent repair Pattern 7 forbids. The conjecture queant is the minimal change and is the only one that makes the purpose clause serve its stated reason.

§2 · THE LEMMA / SCRIPTURE COLLATION — ALL 186 SPANS WALKED, INCLUDING THE AGREEMENTS

Per 7a″ point 4. Every span in data/briefs/9001-lemmata-0010.txt was checked against sources/vulgate/clementine-flat.txt at its own verse, not merely against the pre-scan's ✓/⚠. Recorded here are (a) the divergences that took a marker, (b) the divergences that did not, with the reason, and (c) the classes that agree.

⚑ A verse-numbering check first, because it could have manufactured findings and did not. Migne's VERS. numbers in this range track the Clementine's Mark chapter-and-verse exactly, not the modern one: his CAPUT VIII ends at VERS. 39 Amen dico and CAPUT IX opens at VERS. 1 Et post dies, which is Clementine Mc 8:39 / Mc 9:1. No offset, nothing to log.

2a · Verse lemmata that agree with the Clementine word for word (no marker, no crux)

Mark 7: Et conveniunt (7:1) · Communibus manibus · Non lotis (7:2) · Pharisaei (7:3) · Baptismata (7:4) · Quare discipuli tui (7:5) · Bene prophetavit (7:6) · Bene irritum (7:9) · Vos autem dicitis · Tibi profuerit (7:11) · Nihil est extra (7:15) · Et vos imprudentes (7:18, fragment of Sic et vos imprudentes estis) · Et in secessum (7:19) · De corde hominum (7:21) · Et inde surgens · Neminem (7:24) · Mulier enim (7:25) · Sine prius saturari filios · Filiorum (7:27) · De micis (7:28) · Propter hunc sermonem (7:29) · Tyri · Sidonem · Decapoleos · Venit ad mare Galilaeae inter medios fines Decapoleos (7:31) · Surdum et mutum · Et deprecabantur (7:32) · Et apprehendens eum · Seorsum · Tetigit linguam (7:33) · Et suscipiens · Ingemuit · Adaperire (7:34) · Et statim apertae · Et loquebatur recte (7:35) · Et praecepit illis ne cui dicerent · Quanto autem eis praecipiebat (7:36).

Mark 8: Misereor super turbam · Triduo sustinent (8:2) · Si dimisero eos · De longe (8:3) · Septem (8:5) · Super terram (8:6) · Pisciculos (8:7) · Manducaverunt · Quod superaverat · Septem sportas (8:8) · Et statim (8:10) · Quaerentes (8:11) · Si dabitur (8:12) · Ascendit navim (8:13, fragment of ascendit iterum navim) · Quid cogitatis (8:17) · Video homines (8:24) · Deinde iterum (8:25) · Et misit · Nemini dixeris (8:26) · Quem me dicunt · Homines · Vos vero (8:27–29) · Tu es Christus (8:29) · Et comminatus est (8:30) · Quoniam oportet (8:31) · Et apprehendens (8:32) · Vade retro · Vade retro me · Quae Dei sunt (8:33) · Si quis vult · Deneget · Tollat crucem (8:34) · Qui enim · Qui enim me (8:38) · Amen dico (8:39) · Regnum Dei (8:39, chunk 0012).

Mark 9: Et post dies · Et post dies sex · In montem excelsum · Transfiguratus (9:1) · Et vestimenta (9:2) · Et apparuit · Elias cum (9:3) · Rabbi · Bonum est (9:4) · Nubes obumbrans · Hic est (9:6) · Et statim (9:7) · Praecepit illis (9:8) · Verbum continuerunt · Conquirentes (9:9) · Quid ergo (9:10) · Restituet · Quomodo scriptum (9:11) · Turbam magnam (9:13) · Et interrogavit (9:15) · Et dixi discipulis (9:17) · O generatio (9:18) · Statim spiritus (9:19) · Ab infantia (9:20) · Et frequenter eum · Adjuva (9:21) · Discerpens · Factus (9:25) · Occident eum (9:30) · Ignorabant (9:31) · Capharnaum · Quid in via (9:32) · Inter se disputaverunt (9:33) · Et residens (9:34) · Complexus (9:35) · Quisquis unum · Et quicunque · Et quicunque me susceperit (9:36) · Respondit illi · Magister, vidimus quemdam · Et prohibuimus eum · Nolite prohibere (9:37–38) · Et quisquis scandalizaverit · Mola asinaria (9:41) · Et si scandalizaverit · Manus tua · Quam duas manus (9:42) · Ubi vermis (9:43) · Omnis enim igne · Omnis victima sale salietur (9:48) · Quod si sal · Habete in vobis sal et pacem (9:49). Habete in vobis sal et pacem at 0217A is the last span in my range; Et inde exsurgens and Pharisaei interrogabant belong to chunk 0015 and are not mine.

Two internal re-quotations agree too and are worth naming because a divergence there would have been invisible: Quanto autem eis and Statim apertae at 0208B, and Omnis enim igne / Omnis victima sale salietur re-quoted twice each at 0216D.

2a-bis · THE REGENERATED BRIEF (coordinator's correction of 2026-09-03)

The first data/briefs/9001-lemmata-0010.txt was split by column BAND and was therefore short at its head; the corrected file is split by chunk membership, 197 spans, bands 0205A–0217A. ⚑ I counted my own Latin before trusting it: 62 + 57 + 1 + 68 + 9 = 197 italic spans in chunks 0010/0011/0012/0013/0014. The totals agree.

Nothing in this stint's collation changes, because the collation above was made by walking the five Latin chunks themselves rather than the brief, so the thirteen head-of-range spans that had slid out of the old file (band 0205A, Et conveniunt through Nihil est extra) were already walked and already stand in §2a. Confirmed span by span against the regenerated file:

Et conveniunt (7:1) · Communibus manibus (7:2) · Non lotis (7:2) · Pharisaei (7:3) · Baptismata (7:4) · Quare discipuli tui, (7:5) · Bene prophetavit (7:6) · Bene irritum (7:9) · Vos autem dicitis (7:11) · usque ad (the abridgment formula) · Tibi profuerit (7:11) · Nihil est extra, (7:15) — all agree with the Clementine, and the two carrying Migne's comma (Quare discipuli tui, · Nihil est extra,) keep it with etc. in the English.

⭐ One span in the head gap is a scripture quotation and not a Mark lemma, and it is the only one of the thirteen that needed a verse look-up rather than a lemma check: 0205A Quia abscondisti haec a sapientibus et prudentibus, et revelasti ea parvulis [n: (Matth. XI)]. The Clementine Mt 11:25 reads quia abscondisti haec a sapientibus, et prudentibus, et revelasti ea parvulis — verbatim agreement, and the citation is right. (Lk 10:21 carries the same words with quod for quia; Migne's Quia is Matthew's, so the locator and the wording point the same way. No marker, nothing to log beyond this line.)

⛔ The old file's tail also carried two spans that are not mine — Et inde exsurgens and Pharisaei interrogabant (Mark 10:1–2, chunk 0015). They are struck from §2a above.

2b · Divergences from the Clementine that took a marker

The three [var:] of §1: Eccli. 12:13 medebitur / miserebitur · Matt. 4:10 Vade retro, Satanas / Vade Satana · Mark 9:36 the dropped in nomine meo.

2c · Divergences from the Clementine that took NO marker, with the reason

Every one of these was checked at the verse and declined; they are listed so a blind reader can see the line I drew (Pattern 14 fires on a divergence that changes the sense).

2d · The usque ad audit (brief addendum D)

The formula occurs 20 times in this range and is rendered as far as throughout, after the shipped precedent in this same work (9001/0015 and 9001/0005). ⚑ Two occurrences in the range are NOT the formula and were not rendered as one:

• 0205B communio apostolorum non tota extendit palmites suos usque ad mare — the words of Ps 79:12 (Extendit palmites suos usque ad mare), rendered "as far as the sea".
• 0207B sed usque ad mare venisse — Bede's own geography of Mark 7:31, rendered "as far as the sea".

Both sit within a line of a real formula occurrence, which is exactly the trap 8969's Sap 8:1 lemma set.

§3 · MIGNE'S OWN, RENDERED AS PRINTED — NO MARKER

Each of these is a real, well-formed Latin word or construction that reads oddly. Per addendum A they are Migne's until a plate says otherwise: rendered as printed, logged here, no marker fired, and no corpus-frequency argument treated as licence.

1. 0205B · communio apostolorum non tota extendit palmites suos usque ad mare. The negation is printed and is carried: "the communion of the apostles does not wholly stretch out its branches as far as the sea." tota is nominative and agrees with communio, so the scope is on the subject, not on the verb (Test 2a). Jerome's antithesis wants the apostolic communion to be the vine that does reach the sea against the Pharisees' barren cleanness, so the printed non pulls against the argument. I have not conjectured: the sentence has no correlative or quantifier whose member is missing, so Pattern 18 does not reach it, and the non may be doing work the abridgment has cut away.
2. 0205C · crebro eleemosynarum, aliorumque justitiae fructuum lamento sua opera purgare. lamento is a real ablative and occupies its slot; rendered "by the frequent lament of alms and of the other fruits of justice". A washing-word would suit "purge" better, but I decline to name one: I have no witness on disk for Bede's text here and a conjecture from memory is exactly what the edition forbids. Worth a look if the plate is being read for something else on the page (addendum C).
3. 0205C–D · reprimit furca rationis, id est, Moysi et Isaiae increpatione*. The nominative instrument and the ablative apposition do not agree; both are rendered as printed. Gender/case slips of this kind are not defects under Pattern 9 and no marker is right.
4. 0205A · et ideo cupita salute patiuntur. patior does not govern an ablative, so cupita salute is taken as an ablative absolute: "and therefore, having desired health, they suffer." ⛔ [cj: patiuntur; read potiuntur] was weighed and DECLINED. potiuntur ("they obtain") takes the ablative, restores the antithesis with the Pharisees, and would match the 10703 @1126B shape (a lexical slip leaving a stranded ablative). It loses because Pattern 18's bar is that the English assert something the author argues against, and "they suffer" asserts nothing false — it is merely limp. 18a is explicit that a merely uncertain reading stays in the crux. This is the strongest declined conjecture in the range.
5. 0206C · Mulier est mater nostra Romana Ecclesiae. Rendered by the only construction the printed cases allow: "The woman is our Roman mother of the Church" (Ecclesiae genitive with mater). Ecclesia nominative — "our mother the Roman Church" — is the obvious reading and is deliberately not adopted; adopting it would be a silent repair of a case.
6. 0206D · non panem infractum litterae cupit. infractus can be read privatively ("unbroken", against the micas of the preceding clause) or as "broken/weakened". Rendered "the unbroken bread of the letter", which is the reading the crumb/loaf antithesis supports. The other reading is not excluded and nothing in the gloss decides it.
7. 0207D · Spuma de carne Domini, divina sapientia. spuma is "foam", not sputum; rendered "the foam from the flesh of the Lord". The context is unmistakably the spittle of Mark 7:33, but spuma is a real word in its slot and is left alone.
8. 0207C · dum odorat emungitur. Rendered "its nose is wiped while it smells" — the literal force of emungo — inside a list of twelve paradoxes where every member is a real word. Nothing is emended.
9. 0208C · cujus auditum … divina gratia referat. referat ("brings back") is rendered as it stands; reparet/reficiat* would be smoother and is not adopted.
10. 0209A · quod viam praesentis saeculi, aliter incolumes transire nequimus, alimento verbi sui Christus nos reficiat. aliter has no answering nisi, so the second clause hangs with no conjunction. Rendered with a colon and nothing supplied: "…we cannot otherwise pass unharmed the way of the present world: that Christ refresh us with the food of his word." ⛔ A [cj:] on the 11535 @1231B model (a correlative whose answering member is not printed) was weighed and DECLINED — aliter … nisi is a weaker pair than quam after a comparative, and 18a says that ruling was not to be read wider. The colon leaves the gap visible.
11. 0210D · et quae vix et quasi per gradus ad lucem redeat — a relative-subjunctive clause with no clean antecedent; taken with caecitatis and rendered so.
12. 0211B · ne confessio eorum vulgi opinione, sed veritatis agnitione videatur infirmata. The sed limb has no predicate of its own and infirmata cannot serve it ("seem weakened by the recognition of the truth" is the opposite of the point). Rendered with both limbs as printed, sharing the printed participle. No conjecture: the word in doubt is a participle, not a negation or a correlative, and 18a keeps that class declined.
13. 0212A · Sed perfectioribus, quanta passurus quod a mortuis resurrecturus esset aperuit. passurus has no esset; rendered "how great things he was to suffer, and that he was to rise again from the dead", with the copula taken from the second limb. The English "and" is doing the work of the missing coordination and is flagged here rather than bracketed, because the two objects of aperuit are unambiguous.
14. 0215A · quia eis in figuris loqui consueverunt. A plural verb where the subject is the Lord (consuevit). Rendered with the plural as printed; a number slip is a real form in a real slot and 18a keeps number anomalies declined.
15. 0214D · Credulitas nostra, ut rostrata lingua infirma est. "like a beaked tongue" — rendered literally. The simile is opaque to me and I offer nothing.
16. 0217A · …eo plus a concordiae virtute desipiunt. subjungit: … Migne prints a full stop and then a lower-case subjungit, leaving the long Quia autem sunt multi… period without an apodosis. The stop is preserved (Pattern 8) and the English capitalizes at the sentence break, since capitalization after a full stop has no English exponent for the defect. Recorded so a later reader knows it was seen.
17. 0208C · quantoque frequentius [ferventius] praedicare debeant. [ferventius] is Migne's own bracketed variant and is carried verbatim and untranslated, like every other bracketed variant in this corpus. No marker of ours is stacked on it.
18. 0216B · Σκάνδαλον is reproduced in Greek script exactly as printed; the three glosses that follow (offendiculum vel ruina vel impactio pedis, and ut alii, scrupulus) are translated.

Sections read clean

Chunk 0012 in its entirety (0212C, 42 words) — nothing to report; the (cont.) in its head is our chunk division and is not translated. Chunk 0014's VERS. 43 paragraph (0216C–D) and chunk 0013's VERS. 33–35 (0215B–D) were read twice for negation scope and are clean. Chunk 0010's VERS. 25–29 (0206C–0207A), the densest scripture cluster in the range, carries five citations and every one agrees with the Clementine.

§4 · MIGNE'S CITATIONS PRINTED AS HE PRINTED THEM

Six [n:] locators in this range do not point where their text comes from. All are reproduced verbatim; none is corrected in the running text. Citation repair belongs to data/citation-corrections.json at index time (the standing rule, and 8968 §2's two worked instances).

⚠ 0208B [n: (Matth. XV)] on Civitas in monte posita is Matt 5:14 — but note this one sits on a non-italic paraphrase, so it is a citation attached to the gloss's own words.

§5 · WHAT I WANT READ AT THE PLATE, IN PRIORITY ORDER

Our PL 114 scan clips the margins on leaves in this range, so these are requests, not assumptions. Columns carrying a marker come first — plate-gate.mjs will refuse to ship without them.

Gate-blocking (marker stands on the column):

1. 0214D — Hoc genus nonullo ([sic:]). A non-word; the question is only whether the plate reads in nullo (then the fault is the digitization's and the Latin needs a tei-patch) or something else.
2. 0210C — Ex exspuens ([sic:]). The likeliest withdrawal in the range. See §1.
3. 0207D — serminobusque ([sic:]). Non-word. ⚑ While that line is up, read Spuma and emungitur on the same column (addendum C: both normalizations found in the last batch were found while reading for a different word).
4. 0212B — vineunt ([sic:]). Non-word.
5. 0217A — medebitur ([var:]). The [var:] rests on the Clementine, which is on disk; the plate read is to confirm Migne prints medebitur and not miserebitur. ⚑ Read subjungit (§3.16) and phylargyria on the same column while it is up.
6. 0211D — Vade retro, Satanas ([var:]). Confirm retro is printed in the first member.
7. 0215D — the Mark 9:36 re-quotation ([var:]). Confirm in nomine meo is absent from the printed quotation and that no etc. stands where our file has none.

Not gate-blocking, but worth the look while the leaf is up:

1. 0205C — lamento (§3.2). A real word; I have fired nothing and offer no conjecture, but this is the reading most likely to be a normalization or a broken sort.
2. 0206C — Romana Ecclesiae (§3.5), and infractum four lines below it (§3.6).
3. 0205B — non tota (§3.1). The negation that pulls against Jerome's antithesis.
4. 0209A — the missing nisi after aliter (§3.10).
5. 0212D — nequeant (§1, [cj:]). Not gate-blocking, since [cj:] makes no claim about the plate, but if Migne prints queant the marker comes out and the English changes.

§6 · UNEXPRESSED SUBJECTS SUPPLIED, AND WHO I TOOK THEM TO BE

Per the brief's dominant-defect section. Every "he/she/it/they" below is one the Latin does not print; where two readings construe I say so rather than choosing silently.

• 0206B, VERS. 21 — ut si pulchram mulierem nos crebro viderit inspicere, intelligit nos amare. Supplied "he". The subject is cut away by the usque ad abridgment; from Bede's argument on Mark 7:21 the observer is the devil, learning our inclinations from what we look at. ⚠ Not certain, and it cannot be settled from what is on the page — the sentence is the tail of a resumption. Rendered with a bare "he", which leaves the ambiguity where Migne leaves it.
• 0207B, VERS. 31 — ut transferat se ad alias gentes. Supplied "he", taken as the Lord, which is what Mark 7:31 narrates and what the whole column is about (the withdrawal to Tyre, Sidon and Decapolis). The rival is Judea/the covering transferring itself, which also construes; the Lord wins because cui Dominus two clauses earlier makes him the sentence's active person and because the gloss's point is the passage to the Gentiles.
• 0206C, VERS. 25 — dedit intelligere quod ambo tacuerunt. "Both" is Migne's word; I take it of the Lord and the woman, who are the pair who cum silentio ingressi sunt. Matthew and Mark, the other candidate pair, do not enter a house.
• 0206D — Micas intellectus quaerit, non panem infractum litterae cupit. Rendered "she"; the nearer antecedent is Nata (the daughter, the western barbarism), the further is Mulier, and it is the mother who asks for crumbs in Mark 7:28. The Latin leaves it open and so does the English — this is deliberate, not an oversight.
• 0209A — ipse ab infimis delectationibus abstractam pane coeli reficit. abstractam is feminine accusative with no noun printed; supplied as [the soul] in brackets, which is the standing convention for a supplied English word. animam is the only candidate the gloss offers.
• 0207C, VERS. 32 — ut eos quos humana fragilitas nequit, gratiae dextera salvet. The verb of the relative clause is elided; supplied as [heal] in brackets.
• 0210D, VERS. 24 — ut David se vocari hominem. The governing verb is elided; supplied as [refused] in brackets, since sed canem mortuum et pulicem unum requires a verb of refusing.
• 0216D, VERS. 48 — qui corpus et animam, etc. The verb falls inside the abridgment; supplied as [offers] in brackets.
• 0213C, VERS. 6 — ipsaque illustrabitur gratia et in perpetuum protegetur. Rendered "it", left as vague in English as in the Latin; the resumption tail gives no subject and I decline to name one.

§7 · PUNCTUATION, BAND BY BAND (Test 3)

Compared band by band across all 14+16+0+16+2 column anchors, for : ; ? ! « », with our markers stripped from both sides first.

• Question marks: exact parity in every band. The two lemma-borne marks (Quem me dicunt? at 0211B, Quis medebitur…? at 0217A) ride as printed; the indirect-question case of Pattern 8b does not arise anywhere in this range.
• ! at 0205C (Mira stultitia!) preserved.
• Colons: zero deficits. Four bands initially showed a printed colon softened into a comma or a semicolon in the English (0206D, 0208C, 0210D, 0214B) and two showed a colon we had supplied where Migne prints a comma (0205B, 0205D); all six were repaired before filing, and the band scan now shows no band in which the English carries fewer of any mark than the Latin.
• Semicolons: the English carries a surplus — 28 in the English against 9 in the Latin, the difference falling where Migne prints commas or full stops in asyndetic pairs (Munditia … sterilis est, communio apostolorum …; Micas intellectus quaerit, non panem … cupit). These are the translator's punctuation of a Latin comma, not a rendering of any mark on the plate. Recorded as a class, not swept: per the 2026-08-14 measurement this is a rate question, and Migne prints semicolons himself in this book (0215B, 0215D).

§8 · CONVENTIONS FOLLOWED (so the merge can check them against the other four stints)

• VERS. n.-- verbatim, untranslated. Lemma Englished. usque ad → as far as (matching 9001/0005 and 9001/0015, already on disk).
• Heads: CAPUT VII. → CHAPTER VII., CAPUT VIII. → CHAPTER VIII., CAPUT IX. → CHAPTER IX. — numerals mirror the plate. (cont.) carried untranslated on the two continuation heads (0012, 0014); it is our chunk division and the builder merges it.
• Sigla (BEDA.) (HIER.) (HIERON.) (ISID.) (AUG.) (ID.) (ISID., RAB.) pass through inside [n: …] verbatim, unanglicized, in place. The work prints both (HIER.) and (HIERON.) and the distinction is preserved as printed — it is not normalized.
• Pattern 17: tu/tibi/tuus → thou/thee/thy (0205D Proderit tibi?, In tuos usus; 0212A tu non vis; 0215D in nomine tuo); vos → you (0205D Vos … contemnitis, 0207A vos gentes, 0211B Vos vero).
• Pattern 2: De concordia evangelistarum at 0210B rendered as a work title, On the Concord of the Evangelists.
• Em-dashes: none used anywhere in the five chunks.
• Repeated ## heads, [f:] tails, [nt:] prose notes, [ed:] holes, [d:] doublings: none in this range — the four runbook false-positive classes (band arithmetic, Ibid., noteCount, Pattern 4 tails) were checked and none applies.

<!-- ===== stint file: cruces-0015.md, verbatim ===== -->

Stint of five chunks, cols 0217B–0232C, CAPUT X – CAPUT XIV (through VERS. 30 of CAP. XIV), 6,329 Latin words → 9,862 English (1.56×). One translator, Opus.

Apparatus fired in this range: [sic:] 0 · [var:] 5 · [cj:] 1 · [d:] 0 · [ed:] 0 · [nt:] 0. [n:] 157, at exact parity with the Latin twins. verify-english.mjs clean for 0015–0019 (columns, notes, sections, no duplicate paragraphs).

Mechanical checks run and clean, band by band (column anchors partition both twins 1:1): question marks (3·1·1·2·3), colons, semicolons (0·0·2·5·3), guillemets (zero in both languages), italic spans, [n:] count and position, paragraph count. Every residual band delta is exactly the content of a marker of ours and is itemised below. Anchor POSITION (brief test 2b) was checked word-against-word for all 62 anchors: every one stands beside the same word as in the Latin. The one loose case is 0224C, where Migne sets δευτερώσεις [0224C] vocant) and English word order puts the verb first; the anchor is one word off inside the same parenthesis. Not moved.

⚠ Range boundary, stated explicitly. This stint is chunks 0015–0019 and nothing past the last line of 0019 — CAPUT XIV, VERS. 30, ending et trinae negationis sordes lavit lacrymis at col. 0232C. Everything from VERS. 32.-- Gethsemani. onward (the Gethsemane glosses, all under band 0232C) belongs to chunk 0020 and another stint. It was not translated, not collated, and not checked here. ⛔ Its absence from this file is not clearance for it.

⚠ Brief correction of 2026-09-03, and I re-did the count rather than trusting it. data/briefs/9001-lemmata-0015.txt was regenerated (split by chunk membership rather than by column band), from 312 spans to 297. Counting … spans in my own five Latin files gives 73 + 48 + 53 + 42 + 81 = 297, which agrees exactly. The correction dropped 15 tail rows that were chunk 0020's and restored two head rows that the band split had pushed into the previous stint's file; both are collated in §2d below.

§0 · THE THINGS I MOST WANT A PLATE READ FOR

Ordered by value. None of these fired a marker; each is a candidate, per addendum A and the ship test (B). ⛔ Attribution below is argued, never settled — 9004's stints guessed whose a defect was wrongly 9 times in 14.

1. Three NON-WORDS, one per chunk, all singletons in the whole work

Addendum A's decisive class. Each is a form that is no Latin word; each occurs once in 9001; in each case the work itself prints the correct form elsewhere, which is the 8995 circumcicis control done in the affirmative direction.

No [sic:] was fired on any of the three, deliberately. Addendum A's measured result is that a non-word suspicion is confirmed as our corruption — which makes [sic:], a public claim against Migne's plate, the wrong marker, and is exactly how eight of 9004's twenty-seven were withdrawn. Each is rendered for its evident sense (simply · is joined · them that are with child) and logged here. If the plate carries the defect, a [sic:] belongs at that word; if the plate is clean, a data/tei-patches/9001.json entry does. I cannot tell which from the file, and the rendering is correct either way.

⚑ None of the three falls on a clipped column. Under PL 114's verified map (PDF page = (column + 11) / 2, patrologiaecurs06saingoog), 0224C is the right column of p. 117, 0227B the left of p. 119, 0230B the right of p. 120 — and the launch's clipping is confined to the LEFT column of pp. 114 and 118. All three should be clean, unobstructed reads.

2. 0225B — the sed with no non. THE ONE [cj:] I FIRED.

Nam in hoc infirmitatem filii, sed quia alter in altero operatur ostendit.

The sed has nothing to answer. Rendered as printed and glossed additively:

For in this he shows the weakness [cj: infirmitatem; read non infirmitatem, "not the weakness"] of the Son, but that the one works in the other.

Why this one and not the others. Pattern 18a's bar is that the English asserts something the author is arguing against, and it is explicitly reopened for a dropped negation answering a printed correlative — 7561 @0648, where Migne's setting drops the non of non solum … sed, is the same shape and took the same treatment (the gloss opens on the word Migne DID print, so the verbatim-in-the-Latin-twin guard is satisfied by infirmitatem). Without the gloss the page says Christ shows the weakness of the Son, which the next clause and the whole Sede a dextris argument deny: Nam et filius subjicit inimicos Patri, et Deum Patrem glorificat super terram. ⛔ Not [sic:] — nothing is broken type. Not [var:] — this is the glossator's own prose, not a quoted scripture (18a's WHERE the divergence sits rule). ⭐ Plate value: if Migne prints non infirmitatem the defect is ours and this marker must be withdrawn and a patch filed. This is the single highest-value plate read in the stint.

3. 0226B — Recedent de templo Domino

Bede's lemma-gloss opens Recedent de templo Domino, etc. Recedent is a real word (future, 3 pl.) in a slot that wants the ablative absolute Recedente, which de templo Domino is otherwise built for. Rendered as printed — "They shall withdraw from the temple, the Lord, etc." — which does not construe, and is meant not to: Pattern 7. No marker: a real, well-formed word is Migne's until a plate says otherwise (addendum A), and this is also the exact shape of a lost final letter in transcription. Not glossable under Pattern 18 either: the anomaly is person/number, which 18a leaves declined.

4. 0225D — in libro Regnum

For in libro Regum, "in the book of Kings" — Regnum is a real word with no genitive-plural slot here. Rendered "in the book of Kings", because Migne's own siglum three words later fixes the reference beyond doubt ([n: (IV Reg. XII)]). No marker; one letter, real word, plate candidate. I record that the rendering is a reading and not a transcription, so a later sweep can see it.

5. 0219B — Et assument. for the Vulgate's Et assumens

Lemma at Mc 13:32 [recte 10:32]; see §2 row for it. A one-letter, OCR-shaped divergence in a lemma whose gloss (Bede, on Christ foreseeing the disciples' trouble) needs the participle. Rendered as printed, no [var:] — see §2's stated policy.

§1 · THE [var:] MARKERS — five, each resting on sources/vulgate/clementine-flat.txt

Policy applied throughout, stated so it can be checked: a [var:] fires where a different lexeme or a different construction changes what the English says. It does not fire on orthography (quicunque · maechatur · coelum · nunquid · Joiada), on word order alone, on a dropped or swapped connective (autem · enim · vero · et), on a synonym (dilectio for caritas, facile for cito), on a lemma clipped short, or on a one-letter divergence of the shape a scanner makes. Every declined case is still listed in §2 — a list of findings alone cannot be told apart from a report by an agent that never looked.

1. 0218A · 1 Cor. 14:20. Migne: Nolite effici parvuli sensibus, sed malitia parvuli estote. Clementine: nolite pueri effici sensibus, sed malitia parvuli estote. Migne collapses the verse's two different words into one, in a gloss whose whole subject is childlikeness (Talium … non istorum; similem habent simplicitatem). Marked [var: Vulg. 1 Cor. 14:20 nolite pueri effici sensibus, "be not children in understanding"].

1. 0222B · Matt. 25:41. Migne: Discedite a me, mali, in ignem aeternum. Clementine: Discedite a me maledicti in ignem aeternum. A different word, in a proof-text quoted for the sentence of the reprobate. Marked [var: Vulg. Matt. 25:41 maledicti, "ye cursed"]. ⚑ Two further divergences in the same quotation are logged in §2 and took no marker: Migne's quia for enim (he conflates vv. 41 and 42) and manducandum for manducare, with a comma before it that the English carries.

1. 0223C · John 16:12. Migne: Adhuc multa habeo vobis dicere, quae non potestis portare modo. Clementine: sed non potestis portare modo. A relative for an adversative — a different construction, not a different word only. Marked [var: Vulg. Jo. 16:12 sed non potestis portare modo, "but you cannot bear them now"].

1. 0228D · 1 Cor. 15:34. Migne: Vigilate, justi, et nolite peccare. Clementine: Evigilate justi. The whole paragraph is an exposition of vigilate (VERS. 36, omnes vigilare praecipimur), so the proof-text has been drawn to the argument's keyword — the 7a″ shape exactly, in the direction that matters. Marked [var: Vulg. 1 Cor. 15:34 Evigilate, "Awake"].

1. ⭐ 0229D · 2 Cor. 5:16 — A DROPPED NEGATION INSIDE A QUOTED SCRIPTURE. Migne prints Etsi noverimus Christum secundum carnem, sed nunc jam novimus. The Clementine reads sed nunc jam non novimus — "but now we know him so no longer." As printed, the proof-text asserts the opposite of the sentence that adduces it: Bede's point is that Christ is no longer joined to us corporali praesentia et familiaritate sicut nunc, and the next clause says Spiritualiter autem semper est nobiscum. Rendered without the negation, as printed — Pattern 7a forbids supplying one — and marked [var: Vulg. 2 Cor. 5:16 sed nunc jam non novimus, "but now we know him so no longer"]. ⛔ Deliberately [var:] and not [cj:]: 18a's decline class is explicit that a divergence inside a quoted scripture is Pattern 14's, wherever it would otherwise meet Pattern 18's bar — the rule that separates them is WHERE the divergence sits, not how badly it reads. ⚑ Also in the same quotation and not marked: noverimus for the Clementine's cognovimus.

Marker-content parity. Every [var:]/[cj:] content string was checked against the guard: the four [var:] contents do not occur in the Latin twin (they are the Clementine's words), and the [cj:]'s leading italic infirmitatem does occur verbatim in the Latin twin. That is the inverted test and both halves hold.

§2 · THE LEMMA DIVERGENCE LIST — 7a″ point 4, agreements included

Every italic span in cols 0217B–0232C, collated against clementine-flat.txt by book chapter:verse. Agreements are recorded, because a list of findings alone cannot be told apart from a report by an agent that never looked.

2a. Checked and AGREEING with the Clementine exactly (orthography aside)

Verse lemmata, in order: Mc 10:8 Et erunt · 10:9 Homo non separet · 10:10 Et in domo · 10:13 Discipuli autem comminabantur offerentibus · 10:14 Talium enim est regnum Dei · 10:15 Quisquis non receperit regnum + Regnum Dei · 10:17 Procurrens quidam genu · 10:18 Quid me dicis · 10:19 Ne adulteres · 10:20 Magister · 10:21 Et veni, sequere me · 10:23 Quam difficile · 10:25 Facilius est camelum · 10:26 Et quis potest · 10:27 Apud Deum · 10:28 Ecce nos dimisimus + et secuti sumus te · 10:30 Centies tantum · 10:31 Multi autem (×2) · 10:32 Et stupebant · 10:35 Et accedunt · 10:38 Aut baptismo · 10:39 Calicem quidem · 10:40 Sedere autem + Vobis · 10:41 Indignati sunt · 10:42 Jesus autem vocans eos · 10:46 Et veniunt Jericho + Et proficiscente eo + Bartimaeus caecus · 10:47 Qui cum audisset + Fili David · 10:49 Et vocant caecum · 10:50 Qui, projecto vestimento suo + Exsiliens · 10:52 Vidit et sequebatur + In via · 11:1 Bethaniae + Duos ex discipulis (Migne omits suis) · 11:2 Et statim · 11:4 Pullum ligatum · 11:6 Et dimiserunt · 11:7 Et sedit · 11:8 Alii autem · 11:9 Et qui praeibant · 11:10 Benedictus qui venit · 11:11 Et introivit + In templum + Et circumspectis · 11:13 Cumque vidisset · 11:14 Jam non amplius · 11:15 Coepit ejicere + Vendentes · 11:16 Et non sinebat · 11:17 Vos autem fecistis · 11:20 Aridam factam · 11:21 Recordatus Petrus + Amen, dico vobis · 11:23 Quicunque dixerit · 11:25 Et cum stabitis · 11:28 In qua potestate · 11:29 Interrogabo · 11:31/32 Si dixerimus (×2) · 11:33 Neque ego dico vobis · 12:1 Lacum + Peregre profectus · 12:6 Quia reverebuntur · 12:7 Coloni autem · 12:8 Et ejecerunt · 12:10 Lapidem quem reprobaverunt · 12:13 Et mittunt ad eum · 12:14 Magister, scimus · 12:15 Qui sciens versutiam + Denarium · 12:17 Reddite ergo quae sunt Caesaris + Quae sunt Dei Deo (×2) · 12:18 Sadducaei (×2) · 12:20 Septem ergo · 12:25 Neque nubent, neque + neque nubent + neque nubentur · 12:26 Ego sum Deus · 12:28 Interrogavit eum · 12:29 Primum omnium · 12:32 Bene magister · 12:34 Et nemo · 12:36 Sede a dextris · 12:38 Cavete · 12:40 Qui devorant domos viduarum · 12:41 Gazophylacium + Aspiciebat quomodo + Divites · 12:42 Quadrans · 12:44 Omnes enim · 13:1 Et cum egrederetur · 13:2 Vides has · 13:3 Et cum · 13:5 Et respondens · 13:6 Multi enim venient · 13:8 Exsurget enim · 13:9 Videte autem + Videte · 13:10 Et in omnes gentes · 13:14 Cum autem videritis · 13:21 Et tunc si quis · 13:24 Sed in illis diebus · 13:25 Decidentes + Et virtutes quae + Movebuntur · 13:27 A quatuor ventis · 13:28 A ficu autem · 13:31 Coelum et terra transibunt · 13:32 Neque filius · 13:33 videte, vigilate · 13:34 Et janitori · 13:35 Vigilate ergo (×2) · 13:37 Quod autem vobis · 14:1 Erat autem pascha + Et quaerebant summi sacerdotes · 14:2 Non in die festo · 14:3 Et cum esset Bethaniae + Simonis leprosi + Et recumberet, venit + Mulier + Alabastrum + Nardi spicati + Et fracto · 14:4 Erant autem quidam + Ut quid perditio ista · 14:5 Poterat enim unguentum · 14:6 Bonum opus operata · 14:7 Me autem non semper · 14:8 Quod habuit · 14:9 Amen dico vobis · 14:10 Uno de duodecim + Abiit ad · 14:11 Et quaerebat quomodo illum · 14:13 Ite in civitatem + Homo lagenam aquae + Sequimini · 14:15 Coenaculum grande · 14:17 Vespere autem · 14:18 Amen dico + Unus ex vobis · 14:19 Contristari · 14:20 Unus ex duodecim + Qui intingit + Unus ex vobis tradet me · 14:21 Vae autem · 14:22 Manducantibus + Accepit · 14:23 Gratias + Biberant · 14:25 Jam non bibam + Vitis · 14:26 Et hymno dicto + In montem Olivarum + Omnes scandalizabimini · 14:27 Percutiam pastorem (×2) · 14:29 Petrus autem.

In-gloss scripture agreeing exactly: Luc. 9:23 Si quis vult post me venire · Psal. 24:16 Unicus et pauper sum ego · Psal. 61:11 Divitiae si affluant, nolite cor apponere · Psal. 33:6 Accedite ad eum et illuminamini · Joan. 12:26 Si quis mihi ministrat, me sequatur · Psal. 131:17–18 Paravi lucernam Christo meo: inimicos ejus induam confusione · Luc. 1:32 Hic erit magnus, et filius Altissimi vocabitur · Psal. 4:7 Signatum est super nos lumen vultus tui (Migne stops before Domine and prints etc.) · Psal. 110:10 initium sapientiae timor Domini · Matt. 22:29 Erratis nescientes Scripturas neque virtutem Dei · Psal. 108:7 Oratio ejus fiat in peccatum · Joan. 8:20 Haec verba locutus est Jesus in gazophylacio, docens in templo · Job 26:11 Columnae coeli contremiscunt, et pavent ad nutum ejus · Act. 1:7 Non est vestrum nosse tempora vel momenta quae Pater posuit in sua potestate · Matt. 25:34 Venite, benedicti Patris mei · 1 Cor. 5:7 Etenim Pascha nostrum immolatus est Christus · Psal. 18:3 nox nocti indicat scientiam · Psal. 40:10 Qui edebat panes meos, magnificavit super me supplantationem · Matt. 26:22/25 Nunquid ego sum, Domine? / Rabbi? · 1 Thess. 5:7 Qui ebrii sunt, nocte ebrii sunt · Marc. 4:11 Vobis datum est nosse mysterium regni Dei · Luc. 10:19 Dedi vobis potestatem · Zach. 13:7 Percute pastorem (the prophet's imperative) against Marc. 14:27 Percutiam pastorem (the Father's answer) — Migne's contrast is exact in both directions and is not a defect.

2b. Checked, DIVERGENT, and NOT marked — with the reason in each case

2c. Fragmentary lemmata expanded under Glossa convention 3

Recorded so nobody reads them as conformation later. ⛔ Per the "name the control's class" note, a clipped lemma and a complete one-word lemma are different classes and are not to be contrasted: Magister. · Denarium. · Alabastrum. · Quadrans. · Contristari. · Gratias. · Exsiliens. · Decidentes. · Movebuntur. · Bethaniae. · Vitis. · Cavete. · Interrogabo. have nothing elided and are rendered as they stand.

Expanded: Si licet viro. → "Whether it is lawful for a man to put away his wife" (Mc 10:2) · Quicunque dimiserit. → "Whosoever shall put away his wife" (Mc 10:11) · Quid me dicis. → "Why callest thou me good" (Mc 10:18) · Et quis potest. → "And who can be saved" (Mc 10:26) · Ecce nos dimisimus. → "Behold, we have left all" (Mc 10:28) · Procurrens quidam genu. → "A certain man running up, on bended knee" (Mc 10:17, genu flexo) · Multi autem straverunt. → "But many spread their garments" (Mc 11:8) · Bis gallus vocem. → "The cock shall twice give voice" (Mc 14:30).

⚑ A distinction kept on purpose: Mc 10:28's lemma is dimisimus and the gloss's re-quotation of the same act is Matthew's reliquimus. They are rendered with two different English verbs ("we have left" / "we have forsaken") so the change of word is visible; collapsing them would be the conformation 7a″ names.

§3 · CRUCES BY COLUMN — Migne's own prose

0217D · Non discipuli interrogaverunt prius — where the negative sits

Rendered "It was not the disciples who asked first": the non stands before the subject, not before the verb, and English's only exponent for subject-focus negation is the cleft. The rival ("The disciples did not ask first") puts the negative on the verb, which is the 2a misattachment class exactly. Both construe; the word order decides, and it is recorded here because it is a judgment and not a reading.

0217D · Deus separat: qui conjungit … sic habent uxores quasi non habentes

Migne's sentence changes number in the middle: singular qui conjungit, then plural habent. Rendered as printed, with the colon and the parenthesis where he sets them; no verb supplied and no number smoothed. The quasi non habentes (1 Cor. 7:29) is carried with its negative.

0220B–C · Populus gentium … contradicebant multi — an anacoluthon, kept

The nominative Populus gentium is left hanging and the sentence continues with contradicebant multi. Kept hanging in the English. Straightening it would put our hand on a break the plate makes, and nec tamen … poterat … prohibere is carried as "nor yet could".

0221C · Pedes sunt extremi quos ad jungendum constituit Apostolus

ad jungendum has no expressed object. Rendered "which the Apostle appointed for the joining together", and the relative qui is taken as the feet, which is the only antecedent available.

0221D–0222A · significat quod adventum Christi … salus est

An accusative where the quod clause wants a nominative. English has no exponent for the case, so the clause is rendered with "the coming of Christ" as its subject and the anomaly is invisible on our side — recorded here for that reason, and left unmarked (Pattern 9's boundary: the defect has no English exponent, and unlike 11632's signa … ardeat nothing about the printed sentence is recoverable by marking one word).

0223C · cum aut hic cui quaerit minus intelligit

cui where the sense wants qui — "he who seeks". Rendered as printed ("he to whom he seeks"), which does not construe. Real word, one letter, plate candidate; no marker. The correct reading is plain from the sentence's own second limb (aut odio aut contemptu veritatis indignus est).

0224C · recte hujusmodi fabulam fingunt

Bede's sentence, as printed, says the Sadducees rightly frame the seven-brothers fable to convict of raving those who assert the resurrection of bodies — which is not what Bede is arguing. Rendered with recte as "rightly", not softened to "fittingly" or "consistently": both are readings, and Pattern 7 forbids choosing one inside the text. Flagged here so a blind reader does not take it for a silent repair. Declined for [cj:]: recte is neither a negation nor a correlative, and 18a's extension is scoped to those.

0226B · usque ad erat luitura — a supplied "it"

The abridgment removes the subject of erat luitura (feminine singular, presumably the city or the nation named inside Bede's elided sentence). English cannot leave it unexpressed; "it" is supplied and is ours. The same applies to three other resumption tails in the stint — 0217C concessa est, 0218C testatur, 0219B fruuntur — where the unexpressed subject is inside the elided text and the supplied pronoun is a translator's necessity, not a reading of the Latin. Brief §"unexpressed subject" requires this to be said, so it is said here in one place.

0227B · uxorem Jonathae … filium sinu suo lapsum recepisse perpetuo claudum

Migne's libro Historiarum anecdote; recepisse takes filium … claudum. Rendered "took back her son, fallen from her bosom, lame for ever". No difficulty beyond the truncated praegnantibu. in the lemma above it (§0.1).

0230A · passuros post paucos dies praescit

A dangling accusative plural in a clause whose subject is Christ; the sense wants passurus ("about to suffer"). Rendered as printed — "those about to suffer after a few days, he foreknows that his Gospel is to be preached in the whole world" — which does not construe, deliberately. Declined for [cj:]: 18a leaves number/person anomalies that merely fail to construe declined (11064 @1154C is the named precedent), and this is one.

0230A · futura impii contumelias tacet

futura (neuter plural) against contumelias (feminine plural). Gender/number disagreement with no English exponent — Pattern 9. Rendered "is silent about the coming reproaches of the wicked man". Marked here so a later sweep does not churn it.

0230A · nomine non numine

A jingle Migne prints and English cannot keep: nomen / numen. Rendered "in name not in divine power", which loses the sound and keeps the sense. Recorded because the loss is ours.

0231C · Judas aliis contristatis … qua Dominum tradit; etiam manum … mittit

The semicolon falls inside what should be one clause and the ablative phrase has no verb until after it. Migne's semicolon is kept exactly where he sets it (Pattern 8), and the English breaks where he breaks.

0231D · exstinguuntur inimici quae sunt mysteria Ecclesiae Christi

quae (neuter plural) after inimici (masculine plural). Rendered as printed, "the enemies are extinguished, which are the mysteries of the Church of Christ" — gender with no English exponent again, and the resumption tail's own referent is inside the elided text.

0232A · VERS. 26 lacks its --

Migne sets VERS. 26. Et hymno dicto. where all 111 other verse addresses in this stint print VERS. n.-- . Reproduced exactly as printed, per "Migne's own printing is not yours to tidy." The verse addresses are otherwise in sequence throughout cols 0217B–0232C; no out-of-order verse number in this range.

0225D–0226A · VERS. 42 for Mc 12:42, and VERS. 36 for Mc 13:37

Two places where Migne's verse address does not match the verse his lemma quotes (VERS. 36.-- Quod autem vobis. is Mc 13:37). His numbering, reproduced.

0224C · nubere / ducere — a lexical point English cannot carry in the same words

Bede's note is that Latin says nubere of women and ducere of men, so neque nubent, neque nubentur has to be taken loosely. Rendered "it is properly of women that marrying is said: and of men, the taking of a wife", then neque nubent / neque nubentur as "neither shall they marry" / "neither shall they be married". No new italic span was opened for the mentioned verbs, because italic spans are 1:1 with the Latin.

0224B, 0226D · Greek carried verbatim

δευτερώσεις (0224C) and Φυλάττειν (0225D) are reproduced exactly as printed, untranslated where Migne leaves them untranslated and glossed where he glosses them (Φυλάττειν Graece, servare Latine).

0223D–0224A · Migne's Regnum / two elliptical verbs

Ficus enim non peccavit … sed sacerdotes (0222B) is elliptical; English supplies "did" to close it ("but the priests did"), which is an English ellipsis and not a supplied Latin verb. Recorded.

§4 · SECTIONS READ CLEAN — named, so a blind reader knows where I looked

The following were read line by line against the Latin and are clean: no defect, nothing declined, nothing left:

• 0217B–0217C, CAPUT X opening through VERS. 5 (the Judaea/Galilee geography and the Cornuto syllogismo tail).
• 0218A–0219A entire, VERS. 13 through VERS. 28 — fifteen consecutive glosses, the densest stretch of straightforward lemma-and-comment in the stint, including all four usque ad resumptions and the two Psalm citations, both of which agree verbatim.
• 0219C–0220A, VERS. 35 through VERS. 42 (the sons of Zebedee), apart from the Act. X siglum noted in §2b.
• 0221A–0221C, CAPUT XI VERS. 1 through VERS. 8.
• 0222C–0223B, VERS. 16 through VERS. 32 — the temple-cleansing and the In qua potestate exchange, entire.
• 0224A–0224B, VERS. 10 through VERS. 17 (the tribute money), entire.
• 0225A, VERS. 28 through VERS. 34 as far as Et nemo.
• 0226C–0227A, CAPUT XIII VERS. 3 through VERS. 10, entire.
• 0227C–0228A, VERS. 21 through VERS. 28, entire.
• 0229A–0229D, CAPUT XIV VERS. 1 through VERS. 6, entire (the Bethany anointing) — the only entry from this stretch is the §1.5 [var:] at its far end.
• 0230C–0231B, VERS. 12 through VERS. 18, entire.
• 0232B–0232C, VERS. 27 through VERS. 30, entire.

Zero guillemets in either language across all five chunks — checked, not assumed. Zero usque ad look-alikes missed: the string usque ad occurs 59 times in this stint and was decided per occurrence (addendum D). One occurrence is not the formula — Matt. 28:20 at 0229D, §2b — and one is the formula standing inside the same italic span as its resumption, usque ad Deus meus misericordia ejus praeveniet me [n: (Psal. LVIII)] at 0225D, which is reproduced as one span.

Foot-of-page apparatus: none in this range per the launch's checked zero for the work; no [cn:] is expected or present in the Latin twins, and none was invented.

§5 · CANDIDATES CONSIDERED AND REJECTED, with reasons

Each of these was a live suspicion. All were declined; the reasoning is recorded so it is not re-run.

1. [sic:] on simpiliciter, juugitur, praegnantibu. — declined. Addendum A: a non-word suspicion has been confirmed as OUR corruption every time it was tested, and [sic:] is a public claim against Migne's plate. Cruces + plate request instead (§0.1). This is the single decision in the stint most likely to be reversed by a plate read, and it should be — in the direction of adding three markers, not removing any.
2. [sic:] on Recedent, illo, assument, cui, Regnum, prima, et videam — declined. All are real, well-formed Latin words. Addendum A is 4-for-4 that a real word reading oddly is Migne's, and two of the four refuted arguments in that batch were excellent ones. Rendered as printed, logged.
3. [var:] on the dilectio/caritas, facile/cito, vero/autem, noverimus/cognovimus pairs — declined as synonym-level. Firing on these would put a marker on most pages and bury §1's five.
4. [var:] on every dropped autem/enim/et — declined; Pattern 14's test is a change of sense, and eight such were found (listed in §2b).
5. [cj:] on recte (0224C), passuros (0230A), futura (0230A), adventum (0221D) — declined. 18a's reopened class is negations and correlatives (plus a stranded comparative); mood, case, number and person anomalies that merely fail to construe stay declined, and 18a is explicit that its census is closed and not to be reopened.
6. A [d:] on Percutiam pastorem printed twice at 0232B — declined, and it is not close. The two occurrences are the two halves of Jerome's own argument (the prophet asks Percute, the Father answers Percutiam), and the second is the answer, not an eye-skip. Marking it would destroy the point.
7. A [d:] on Sadducaei, etc. printed twice at 0224B–C — declined. Two separate glosses on the same lemma, one Jerome's and one Bede's, each with its own comment. The Glossa does this throughout; it is structure, not dittography. Same for Si dixerimus at VERS. 31 and VERS. 32, which are two different verses, and Quae sunt Dei Deo at 0224B, which Migne glosses twice under Aliter.
8. Supplying ? at 0223A (quid amplius rogabit, nisi quod perseveret in eo quod obtinuit?) — not needed: Migne prints the mark and it is kept. Conversely, no ? was supplied at 0217C Nunquid Deus contrarius sibi est, etc., 0227A Quare haec omnia inferantur, or 0219C's Baptismo habeo baptizari … perficiatur., all of which Migne closes without one. Question-mark parity is exact per band in all five chunks.
9. A comma→semicolon sweep in 0015 and 0016, where Migne prints zero semicolons in 2,558 Latin words — declined as an initiative, and instead simply obeyed: the English carries zero semicolons in those two chunks. The Tobit exception was not invoked, because it does not need to be; the plate's own practice was counted first (Latin 0·0·2·5·3 across the five chunks) and matched exactly, band by band. The English shows 0·0·3·5·3 only because the [cj:] marker's own fixed form carries a semicolon inside it at 0225B.
10. A [nt:] conversion of any note — declined. All 157 notes in this range are citations or bare attribution sigla ((BEDA.), (HIER.), (ID.), (BED.), (ISID.), (HILAR.), (Id.), (Cap. I)), which Pattern 15 keeps as [n:]. Note that Migne prints both (BEDA.) and (BED.), and both (ID.) and (Id.), within a few columns; the variance is his and is carried per occurrence, unnormalized.
11. A [f:] tag anywhere — declined per runbook false positive 4: this is lemma-and-gloss commentary, not a florilegium, however much inline scripture it carries.
12. Reporting the column-band gaps (0217B→0217C with no A; 0223C→0223D; 0226A→0226B) — declined per runbook false positive 1. No text breaks mid-word or mid-construction across any anchor in this range; that was checked, and it is the only signal in the class.
13. BEDA. at 0230B, set as an italic span in running text rather than as a note — carried verbatim, untranslated, in its italics. It is a siglum, and sigla pass through (convention 4); it is not a lemma and was not rendered as one. The same treatment was given to the bare HIER. at 0224D, which Migne sets in roman inside the paragraph.

§6 · WHAT THE FOURTH TEST TURNED UP

Run last, with the counts already known clean: each finished paragraph read back against the Latin asking only does it assert what he asserts? Three repairs came out of it, all in the negation-scope family that no count-based check sees, and all made after the chunks otherwise verified:

• 0220D — non falsas divitias quaeramus, sed lucem had been rendered "let us not seek false riches, but the light", floating the negative onto the verb. The sed pivots on falsas divitias. Now "let us seek not false riches, but the light."
• 0223D — Non hoc ignorando dixit … sed had been "He did not say this in ignorance", same class, same fix: "He said this not in ignorance."
• 0231D — sanguis … non omnes emundat had been "does not cleanse all", where English scope is genuinely ambiguous and the wrong reading (that it cleanses nobody) is available. Now "cleanses not all", which is the 2a trigger population exactly: a negative on a verb with omnes adjacent.

⚑ All three read perfectly well before the fix. That is the point of the test.

Two further sentences read smoothly where the Latin is rough and were therefore treated as suspects rather than successes; both were left rough on purpose and are logged above — 0226B Recedent (§0.3) and 0230A passuros (§3).

§2d · THE TWO HEAD SPANS RESTORED BY THE 2026-09-03 BRIEF CORRECTION

Both stand at band 0217A — the last anchor of chunk 0014 — while belonging to chunk 0015, which is why the band split lost them. Both were translated in the normal course (they open CAPUT X in my Latin, and the English carries them); what was missing was their entry in the collation, and this supplies it. Both agree with the Clementine exactly, and neither is a divergence:

⚑ Nothing in either changes a rendering already filed, and no marker follows from them. The third span in that head region, Si licet viro at 0217C, was in the original brief and is adjudicated at §2b — it is the one interesting divergence in the head, and it runs the unexpected way, with the Clementine as the outlier.

§7 · WHAT THE SHIP TEST (CLAUDE.md 8a) STILL OWES ON THIS RANGE

node scripts/plate-gate.mjs 9001 currently fails for the whole work — 27 [sic:]/[var:] markers against 0 recorded plate reads — and five of those 27 are mine, at cols 0218A · 0222B · 0223C · 0228D · 0229D. They cannot ship until those columns are read. Recorded here so the debt is attached to the range that incurred it.

⚑ The measured prior says these five are the safe half of the population: on 9004, 46 of 46 [var:] stood at the plate and 8 of 27 [sic:] fell, every fall a single word and not one a quoted clause. All five of mine are [var:] on quoted clauses of three words or more, which is the class that has never fallen. I fired zero [sic:], which is why the three non-words of §0.1 are cruces and not markers.

⚑ The [cj:] at 0225B (§0.2) is not in the gate's population, and it is the one marker in this stint I would most want a plate to adjudicate before ship — not because it is likely wrong about the sense, but because if Migne prints non infirmitatem the defect is Corpus Corporum's and the marker must be withdrawn in favour of a patch.

Columns I would put in front of a reader, in order: 0224C (simpiliciter, and read the whole line — addendum C) · 0225B (the sed with no non) · 0230B (juugitur) · 0227B (praegnantibu.) · 0226B (Recedent) · then the five [var:] columns for the gate. None of the six candidate columns falls on the two clipped leaves named in the launch (pp. 114 and 118, cols 0217/0218 and 0225/0226) — 0225B is on p. 118 and is in the LEFT column, so that one may be affected; the other five should be clean reads.

<!-- ===== stint file: cruces-0020.md, verbatim ===== -->

Range: cols 0232D–0244C, the tail of CAPUT XIV through CAPUT XVI (the Passion and the Resurrection), 4,702 Latin words, 4 chunks. One of five stints on idno 9001; this file is cruces-0020.md for merging.

§0 · Counts, and what shipped

Apparatus fired in this range: [var:] × 2 · [sic:] 0 · [cj:] 0 · [d:] 0 · [ed:] 0 · [nt:] 0. Anchors, notes, italic spans and paragraph counts are 1:1 with the Latin twins; : ; ? « » ! were compared band by band against the Latin (Brief test 3) and every band matches, with the two sanctioned exceptions named in §6.4. verify-english.mjs 9001 reports no error against any of my four chunks (it fails only on peer chunks not yet on disk).

⭐ The range ends where the manifest says it ends. 0023 stops at [0244C] verbum per totum orbem seminaverunt. — the last words of Mark. Nothing of the Luke commentary that opens below it on the same plate page is present in my Latin, and none was translated.

§1 · ⛔ THE PLATE-READ REQUESTS — three split/non-word runs, and NO marker fired on any of them

These are the three places in my range where the type is not a Latin word. Per the brief's addendum A a non-word is worth raising explicitly as a plate candidate, and per the launch instruction I raise them here and fire no marker, because addendum A's own data says this class has been the digitization's corruption, 10 times out of 10 — a [sic:] would blame Migne publicly for what is most likely Corpus Corporum's error, which is exactly the seven withdrawals of the 2026-08-23 batch. Each is rendered on the page at the reading Migne self-evidently printed, and the crux is the receipt.

⚑ Why all three are one finding. curcifigentes is an r/u metathesis; us que and ve peram are word-division splits. All three are the shape addendum A names (per verse, Qui dam, quorum dam) and all three read as line-break artifacts. us que ad is the strongest of the three as evidence, because the identical formula is set correctly thirty-nine other times in my four chunks — a single split against 39 clean occurrences within 4,700 words is not a compositor's habit.

⚠ curcifigentes cum is a double defect — the metathesis and cum for eum — and it sits in a lemma, where a reader meets it first. If only one line of my range gets a 400 dpi look, make it col. 0237D. If the plate reads crucifigentes eum, the fix is data/tei-patches/9001.json and nothing in my English changes. If the plate really prints curcifigentes cum, then the English needs a [sic:] and I have under-marked it — I say so plainly rather than guess.

⛔ I make no claim about WHOSE these are. 9004 measured stints' attributions wrong 9 times in 14, and my argument above is a probability, not a finding.

§2 · The conventions I applied, stated so a merge can check them

1. VERS. n.-- verbatim, lemma Englished (Glossa convention 1–2). 77 verse addresses in range, none altered.
2. usque ad → as far as, in its own italic span — settled on 8978, reused on Tobit (8970 cruces §1b) and not re-opened. 40 occurrences in my range (0020: 13 · 0021: 9 · 0022: 7 · 0023: 11). ⭐ No verb, subject or object was supplied to any of them, including the four whose resumption lands mid-clause (0233D adversariis dare didicerunt — the object of dare is inside the excised middle and stays missing in the English; 0236D purpura vel cocco vestitur; 0238B Hic sortita, etc.; 0240C lineo terreno celebrat).
3. ⚠ usque ad LOOK-ALIKES (brief addendum D): four in my range, decided one at a time, and NONE of the four is the formula. - 0232D Humiliavit semetipsum factus obediens usque ad mortem — Phil 2:8's own words, inside a quoted scripture. Rendered "unto death". - 0235A inviolata permanet castitas usque ad consummationem saeculi — ordinary prose. "even to the consummation of the world". - 0238D/0239A excitantur vincula inferni usque ad cervicem — ordinary prose. "even to the neck". - 0243A etc., usque ad: Euntes in mundum universum, praedicate Evangelium. — this one IS the formula, and it is the one case where Migne prints the colon inside the italic span and runs the resumption on. Reproduced exactly, colon and all.
4. A lemma is expanded only where the printed words cannot stand as English. This is Ruth's own calibration and not a new rule: Ruth expanded Non te. → "Let no man see thee" but left She went out, etc. unexpanded. So in my range — - expanded (the stump is not English): Summum, → "The high priest," (Mc 14:53); Tristis est, → "My soul is sorrowful," (Mc 14:34); Sed non quod. → "But not what I will." (Mc 14:36); Nam et Galilaeus. → "For thou art also a Galilean." (Mc 14:70); Alios salvos. → "He saved others." (Mc 15:31); Si mortiferum, → "If they shall drink any deadly thing," and Super aegros manus, → "They shall lay hands upon the sick," (Mc 16:18); Et induunt. / Ut crucifigerent. / Et non accepit. take the object English cannot omit. - left as printed (the stump IS English): Vere, → "Truly,"; Currens autem, → "And running,"; Velare faciem, → "To cover the face,"; At ille,, Tunc,, Novissime,, Illi autem,, Maria autem,, Jesus autem,, Et quidam,, Vide,, Signa autem,, Qui omnes,, Et exsurgens, and the rest. ⚑ Per the brief's "name the control's class" section, contrasting these two groups would manufacture an inconsistency that is not there — they are different classes, and where I had to choose I expanded the stump rather than contracting its twin.
5. (cont.) on the repeated heads of 0020 and 0022 is our chunk division and is not translated — carried as printed, per runbook false-positive 7.
6. Heads: numerals mirror the plate. CAPUT XIV. → CHAPTER XIV., CAPUT XV. → CHAPTER XV., CAPUT XVI. → CHAPTER XVI. Checked against the shipped corpus, which has CHAPTER XI./XV./XVI. in numerals throughout and zero spelled-out forms above ONE.
7. Pattern 17: thou wherever the Latin is singular (Peter at 0234B/0235B me negabis; Pilate to Jesus at 0235D Tu dicis; the psalmist's addressee at 0236C tabernaculo tuo; Death at 0241A aculeus tuus; the young man at 0241C in adolescentia tua), you wherever it is plural (0233A ut non intretis; 0233D potestas vestra; 0234C Et videbitis; 0236A Quid ergo vultis; 0237B vobis relinquens exemplum; 0241D Vos autem, quid timetis; 0242A Praecedit vos). No archaism in the commentator's own voice.
8. Greek passes through verbatim, unitalicized where Migne prints it roman: ἀναβόλαιον (0239A) and Παρασκευὴ (0240B). The gloss on the second (Παρασκευὴ Graece, praeparatio Latine) sets praeparatio in roman, so it is translated ("preparation in Latin") rather than kept as italic Latin — adding italics would break span parity.
9. Pattern 4 [f:] does NOT fire. This is a lemma-and-gloss commentary, not a florilegium (runbook false-positive 4). Its inline scripture is text, not apparatus.

§3 · The two [var:] markers, and why the other nineteen divergences did not get one

3.1 ✅ FIRED — 0234B, Mc 14:60, the lemma Non respondens,

Migne prints VERS. 60.-- Non respondens, etc. The Clementine reads Non respondes quidquam ad ea quae tibi objiciuntur ab his ? — a finite second-person verb inside the high priest's question, not a participle. The gloss immediately below turns on the question being asked (Ad responsum provocat, ut ex qualibet occasione sermonis locum accusandi inveniat — "he provokes him to an answer"), so the divergence is material: Migne's participle converts the priest's interrogation into a description of Jesus. Rendered as printed — "Not answering, etc." — with [var: Vulg. Mc 14:60 Non respondes, "Dost thou not answer"] beside it. ⚑ Why this is Pattern 14 and not Pattern 18: the divergence sits inside a quoted scripture, and [var:] is the convention for that boundary (Pattern 18a's last decline class states it). ⚑ The rival I rejected: that respondens is a plate slip for respondes and deserves a plate read. It may be — but it is a real, well-formed Latin word, which addendum A says is Migne's until a plate says otherwise, so no [sic:] and no [cj:]. The [var:] records the divergence without asserting a defect, which is the only claim I can stand behind.

3.2 ✅ FIRED — 0238C, Rom 6:6, quoted as Sicut dicit Apostolus

Migne: Vetus homo noster simul cum illo crucifixus est ut destruatur corpus peccati, ut ultra non dominetur super nos peccatum. Clementine Rom 6:6: vetus homo noster simul crucifixus est, ut destruatur corpus peccati, et ultra non serviamus peccato. Two divergences, the second material: Migne's final clause is Rom 6:14's peccatum … non dominabitur ("sin shall not have dominion") pulled into 6:6's slot, so the proof-text asserts sin's loss of dominion where the Vulgate asserts our ceasing to serve. Rendered as Migne prints it, with the Vulgate reading in the marker. The added cum illo is not separately marked — it is explicative, and marking it would double a note the reader already has.

3.3 ❌ DECLINED, with the reason in each case

Every one of these was checked against sources/vulgate/clementine-flat.txt and cited by book chapter:verse. None is unchecked; all are declined on the merits.

Checked and found to agree with the Clementine exactly (7a″ point 4 — the list is worthless without these): Mc 14:32 Gethsemani · Sedete hic · donec orem — 14:33 Et coepit pavere — 14:34 Tristis est — 14:35 et orabat — 14:36 Sed non quod · Non quod ego volo, sed quod tu — 14:37 Et ait Petro — 14:38 ut non intretis in tentationem — 14:42 Surgite · Surgite, eamus — 14:44 Signum — 14:45 Rabbi — 14:47 Unus — 14:50 Tunc — 14:52 At ille · Ab eis — 14:53 Summum — 14:54 A longe · Atrium · Et calefaciebat se — 14:60 Et exsurgens — 14:61 Tu es Christus — 14:62 Et videbitis · A dextris virtutis — 14:63 Summus autem sacerdos · Vestimenta sua — 14:64 Qui omnes — 14:65 Velare faciem · Alapis eum caedebant — 14:68 At ille · me negabis · Et exiit foras — 14:70 Nam et Galilaeus — 14:72 Et recordatus est — 15:1 Et confestim · Vincientes Jesum — 15:2 Tu es rex Judaeorum? · Tu dicis — 15:4 Vide · Vide in quantis — 15:11 Ut magis Barabbam — 15:12 Quid ergo vultis faciam regi Judaeorum? — 15:13 Crucifigatur (Migne's passive for the Vulgate's Crucifige eum — see below) — 15:14 Pilatus vero · Quid enim mali fecit? · At illi magis — 15:15 Barabbam, et tradidit · Jesum flagellis · Jesum flagellis caesum — 15:17 Et induunt — 15:19 Et percutiebant · Adorabant eum — 15:20 Ut crucifigerent · Et educunt illum — 15:21 Et angariaverunt · Patrem Alexandri — 15:22 Calvariae locus — 15:23 Et dabant · Et non accepit — 15:24 Mittentes sortem — 15:26 Et erat titulus — 15:27 Et cum eo crucifigunt · Duos latrones — 15:29 Moventes capita — 15:31 Alios salvos — 15:32 Descendat nunc · Et qui cum eo — 15:33 Et facta hora sexta · Tenebrae — 15:35 Et quidam — 15:36 Currens autem — 15:37 Jesus autem — 15:38 Et velum templi — 15:39 Videns autem · Vere — 15:40 Erant autem · Jacobi minoris et Joseph — 15:42 Et cum jam sero esset — 15:43 Venit Joseph — 15:46 Joseph autem · Quod erat excisum — 15:47 Maria autem — 16:1 Et cum transisset — 16:2 Et valde mane — 16:3 Et dicebant ad invicem — 16:4 Revolutum lapidem — 16:5 Juvenem — 16:6 Nolite expavescere · Jesum · Crucifixum · Surrexit, non est hic · Ecce locus — 16:7 Et Petro (×2) · Praecedit vos · Praecedit vos in Galilaeam — 16:8 Fugerunt · Et nemini — 16:9 Surgens autem — 16:11 Et illi audientes — 16:12 Duobus ex his · Ambulantibus — 16:14 Novissime · Duritiam cordis — 16:15 Omni creaturae — 16:16 Qui crediderit — 16:17 Signa autem · Daemonia ejicient · Linguis loquentur novis — 16:18 Si mortiferum · Super aegros manus — 16:19 In coelum · Et sedet a dextris — 16:20 Illi autem. Also, verbatim: Jo 18:13 adduxerunt eum ad Annam primum · Mt 27:19 Nihil tibi et justo illi · Is 5:7 Exspectavi ut faceret judicium, et ecce iniquitas: et justitiam, et ecce clamor · Ps 90:10 Flagellum non appropinquabit tabernaculo tuo · Ps 31:10 Multa flagella peccatoris · Jo 19:1 Apprehendit Pilatus Jesum et flagellavit · Jo 19:2 Et milites plectentes coronam de spinis · Mt 27:34 cum felle mistum · Et cum gustasset, noluit bibere · Eph 3:17–18 (the 0238C full form) · Ps 2:6 Ego autem constitutus sum rex ab eo · Ps 115:3 Quid retribuam Domino pro omnibus quae retribuit mihi? · Ps 75:2 Notus in Judaea Deus, in Israel magnum nomen ejus · Mt 10:5 In viam gentium ne abieritis (twice, 0240A and 0243B) · Jo 19:25 Stabant autem juxta crucem Jesu mater ejus · Ps 117:19 Aperite mihi portas justitiae · Ps 29:6 Ad vesperum demorabitur fletus et ad matutinum laetitia · Is 35:10 Fugiet dolor et gemitus.

⚑ Crucifigatur, @0236A is Migne's paraphrase of Mc 15:13 Crucifige eum inside his own narrative sentence (Cumque respondissent: Crucifigatur), not a lemma. Rendered "Let him be crucified." Not marked: he is reporting, not quoting.

§4 · ⛔ THE INVENTORY BRIEF FOR THIS RANGE IS SHORT BY 17 SPANS — please fix upstream

data/briefs/9001-lemmata-0020.txt says my range holds 206 italic spans. My four Latin chunks hold 223. The seventeen missing are all at the head of chunk 0020, cols 0232D–0233B — the brief's listing opens at [0233B] Caro autem infirma., so everything from VERS. 32.-- Gethsemani. through VERS. 38.-- ut non intretis in tentationem, was never pre-scanned. The generator appears to have keyed the slice on a later column than the chunk's own colFirst: 0232D.

The seventeen: Gethsemani. · Sedete hic. · Donec orem. · usque ad · Humiliavit semetipsum … pro amicis suis · Et coepit pavere. · Venit princeps hujus mundi … · Tristis est, · usque ad · Et orabat. · usque ad · Non quod ego volo, sed quod tu. · Sed non quod. · Et ait Petro, · Et si oporteat me mori tecum … · Ut non intretis. · ut non intretis in tentationem,.

I collated all seventeen myself against the Clementine and they are in §3 above — three of the range's more interesting divergences (Jo 15:13's word order, Jo 14:30's, and the Mt 26:35 / Mc 14:29 conflation) are in that unscanned head. ⚠ If the same generator produced the other four stints' briefs, their head columns may be missing too.

§5 · This range's own cruces — Migne's readings, rendered as printed

0233A · somnium for somnum — RENDERED "sleep", and here is why that is not a silent repair

nunc tristitiae magnitudine somnium vincere non potest, of Peter in Gethsemane. somnium is "a dream"; the sense wants somnum, "sleep" (Lc 22:45 invenit eos dormientes prae tristitia). No marker, and no carried Latin: somnium/somnus interchange is ordinary medieval orthography and Pattern 9 governs it — carrying every non-classical form would bury the real defects. Recorded so a later sweep does not churn it.

0233B · tres personas nos habere in orando — rendered literally, sense not adjudicated

Trina Domini vigilia tres personas nos habere in orando docet, et de praeteritis et de praesentibus, et futuris veniam rogare. Rendered "teaches us that we have three persons in praying, and to ask pardon both for things past and for things present and for things to come." personas here is odd — the triad that follows is of times, not persons — but persona is a real word in its slot and the abridgment gives no more of Jerome's sentence. Not marked.

0233B · ut passuri gaudium et confidentiam videant — TWO READINGS, and I say so rather than choose

Both construe. (a) passuri nominative plural attaching loosely to the nos of eamus: "so that, we being about to suffer, they may see joy and confidence" — the subject of videant is then the same party as the subject of inveniant two clauses earlier, the arresting men. (b) passuri as the subject of videant: "that those about to suffer may see joy and confidence" — grammatically tidier but makes the captors, who are not about to suffer, disappear from a sentence that has just named them (nos non inveniant). Shipped as (a), because the nos non inveniant … sed ultro eamus obviam frame keeps one subject running through inveniant and videant, and because (b) leaves ultro eamus obviam pointing at nobody. ⛔ (b) is not thereby refuted, and a reader should know it was weighed.

0234A · sicut Isaac caligantibus oculis, Jacob sub manibus non agnoscebat — the case decides it

Read as: Isaac, eyes dim, did not know Jacob under his hands (Gen 27:22–23), Jacob being the indeclinable object. The tempting alternative — Jacob as a second subject alongside Isaac — leaves sub manibus with nothing to attach to and contradicts the source story, where the hands felt are Jacob's and the man deceived is Isaac. Shipped accordingly. ⚑ This is the brief's unexpressed-subject class exactly: nothing in the sentence names who fails to recognize whom, and only Genesis decides it.

0234A / 0234B / 0233C / 0233B — every supplied pronoun in this range, named

Per the brief, each "he"/"she"/"they" I supplied that the Latin does not print, with what fixed it:

• 0233B Dat signum osculi → "He gives the sign of a kiss" — Judas, from Signum / Mc 14:44 Dederat autem traditor ejus signum eis.
• 0233C Ipse enim pietatis nunquam obliviscitur → "he himself" — Christ, from Dominus tetigit auriculam in the preceding sentence.
• 0234A Quorum et praesentiam detestabatur … servavit in mente → "he" — the young man who fled (Mc 14:51–52), the subject of the lemma At ille … Ab eis. Not Peter, who is two verses away and is the subject of the next paragraph.
• 0234B Quia negationi proximus … quod Dominum non reliquit → "he" — Peter, from the lemma A longe / Mc 14:54.
• 0236D purpura vel cocco vestitur, cum triumpho gloriosorum martyrum gloriatur → "she" — the Church. Feminine singular, and the martyrs-and-purple topos in this gloss has no other subject; the abridgment has removed the sentence's own antecedent, so this is stated rather than assumed.
• 0237C Cleophae videtur eam dicere → "He seems to call her" — Bede, the speaker of the whole [n: (BEDA.)] paragraph.
• 0240B si forte valeant imitari → "they … imitate him" — the women of solae mulieres.
• 0242D ut ipsis evangelistis et apostolis resurrectionem evangelizaret → "she" — Mary Magdalen, from the paragraph's own mulier auctor culpae / In principio mulier.

0234C · Sacerdos interrogat Filium Dei; Jesus respondit Filium hominis — the accusatives are ambiguous in Latin and in English

Both accusatives can be read as the person addressed or as the title spoken. Rendered "The priest asks after the Son of God; Jesus answers the Son of man", which keeps the ambiguity English can keep. The argument (ut intelligamus eumdem Filium Dei esse et Filium hominis) requires the titles reading, and a reader who wants it has the sentence.

0237B · Oriens de virtute fulget — rendered "The east shines from the power"

The quadrata mundi topos assigns the four quarters to the four arms of the cross; de virtute most likely means the cross's top (virtus for the head or summit of a thing). Rendered literally as "from the power" and not repaired to "from the top", because that is Pattern 7's whole business. The conjecture lives here.

0237B · arcton — Migne's Greek accusative in a Latin sentence

arcton dextera tenet. ἄρκτος = the Bear, the north. Rendered "the right hand holds the north". Not italicized (Migne sets it roman) and not marked.

0237D · Id pro quo patitur — a two-word sentence with no verb of its own

Following Et non accepit. Rendered "That for which he suffers." Elliptical in Latin, elliptical in English; nothing supplied.

0238B · Erat autem hora ter. — Migne's contraction, expanded

Mc 15:25 reads Erat autem hora tertia. The lemma prints ter. with a stop, which is a contraction of a word the sentence grammatically contains (Pattern L5's test), not a truncation that stops short — and the gloss immediately reasons about the third hour against the sixth (nam sexta hora tenebrae suffuderunt terram). Expanded to "Now it was the third hour." ⚠ The alternative reading, and it loses: that ter is the adverb "thrice". It cannot be — hora ter is not Latin, and the gloss's own arithmetic is about hours.

0238B · praepositi, separati in the four orders — rendered "the prelates, the set apart"

conjugati, viduati, praepositi, separati — the married, the widowed, the prelates, the set-apart. The last two are unusual as an order-list and I have not smoothed them into the more familiar virgines; they are what Migne prints.

0238C · Quandiu ergo agunt opera nostra — rendered "So long therefore as our works act"

agunt is transitive with no object, because the usque ad has excised the rest of the sentence along with it. No object supplied.

0238C · sinistri latronis mentem imitatur et actus — singular verb, plural relative

Qui autem pro humana gloria … imitatur answers Et qui hoc … gerunt two clauses earlier. Rendered with the number Migne prints on each ("they who do this … but whoever does it").

0239C · Ut arcam testamenti … appareant — accusative subject of an intransitive

arcam is accusative where appareant wants a nominative. English carries no case ending, so there is nothing to render either way — Pattern 9's exemption in its exact form. Logged so a later sweep does not churn it; the Latin column shows it.

0240A · [n: (Matth. X)]*.; — Migne prints a full stop AND a semicolon

et in Evangelio prius: In viam gentium ne abieritis.; et post passionem: … Reproduced verbatim, both marks, in the English. Pattern 8: never delete a mark the plate prints, however odd. ⚠ Worth one glance on the plate if it is being read anyway — a stop-then-semicolon is also the shape a digitization produces at a line end.

0240B · Sicut non excluditur … ita non repellitur — negation hosts named (test 2/2a)

A sicut … ita pivot with a negative on each side is precisely the trigger condition. Both negatives sit on their verbs (excluditur, repellitur) and there is no quantifier or correlative nearby competing for the host: the sentence's structure is a parallel of two exclusions, not of a solus/omnis scope. Rendered "is not shut out … is not driven back." Checked, and clean.

0242C · sequenti nocte … cum sequenti nocte — the phrase stands twice, ten words apart

Sepultus ergo sexta sabbati … circa ve peram sequenti nocte et die sabbati, cum sequenti nocte in monumento positus, die tertia … surrexit. Both are rendered, in place. ⛔ Not marked [d:]. Pattern 11's dittography is the same run set twice, adjacent; these are ten words apart with et die sabbati between, and the sentence's arithmetic (one day, two nights) arguably needs two nights named. A [d:] would assert a compositor's doubling I cannot demonstrate. The crux is the honest home for it, and it is a plate-read candidate of second priority.

0244B · quod inde eis non nocet si mortiferum bibunt — the quod clause hangs loose

Rendered "thereby it does them no hurt if they drink a deadly thing", with both negatives (nec … perducunt, non nocet) on their own verbs. The quod has no governing verb in Migne's sentence; nothing was supplied to give it one.

§6 · Sections read clean, and the tests run

### 6.1 Read clean, nothing to report Chunk 0022 in its entirety (cols 0239C–0240D, VERS. 35–47) apart from the two items above (us que ad @0240C, arcam @0239C) — it is the most regular stretch of the four, with no non-words, no scripture divergence beyond spelling, and no ambiguous subject. Chunk 0023 cols 0241A–0242B (VERS. 1–7) likewise. Chunk 0021 cols 0236D–0237C (VERS. 19–22) likewise.

### 6.2 Test 1 — a [sic:] must not swallow the word it marks Vacuously clean: no [sic:] was fired. The strip test was still run on both [var:] sites — delete the marker and each sentence reads whole, because a [var:] by construction stands beside a full rendering and never inside one. Neither marker is load-bearing for a verb, a negation, a conjunction or an object.

### 6.3 Test 2 / 2a — negation hosts Every printed negative in the range was matched to the single Latin word it attaches to and then found on that word in the English. The four trigger-condition sites (a negative on a verb with a correlative or quantifier adjacent) were worked explicitly and are recorded above or here:

• 0240B sicut non … ita non — §5, clean.
• 0243A non tam est illorum infirmitas, quam nostra firmitas — the non hosts on tam, not on est. Rendered "is not so much their weakness as our strength", which is the only English that puts it there. Had it drifted onto the verb ("it is not their weakness but our strength") the concessive would have become a denial.
• 0234D Non ut scelera eorum non videat, sed … — two negatives, the first on the purpose clause as a whole (Non ut), the second on videat. Both carried: "Not that he might not see their crimes, but that…". The self-contradictory look is Migne's construction, not a defect.
• 0238C nec occidendo potuerunt efficere quod non esset eis rex — nec on occidendo ("not even by killing"), non on esset. Both carried; collapsing them into one negative would reverse the sentence.

### 6.4 Test 3 — punctuation compared BY BAND, not by total Every column band in all four chunks matches its Latin twin for : ; ? « » !, with two deliberate exceptions, both licensed and both here so a checker does not re-find them:

1. 0241B, one extra colon — Pattern 8b. Migne closes an indirect question with ?: Quomodo lapis per angelum sit revolutus, Matthaeus dicit? English cannot end that construction with a mark, so it is recast direct behind a colon: "Matthew says: how was the stone rolled away by the angel?" The mark is kept, nothing supplied, nothing deleted. This is the range's only 8b site.
2. 0234B and 0238C, extra colons inside [var:] content (Mc 14:60, Rom. 6:6) — ours, not Migne's, and stripped from the ratio count.

⚠ No comma→semicolon sweep was run. Migne prints semicolons freely in this range (20 of them), so Tobit's narrow exception does not apply. Where a first draft had reached for a semicolon on English rhythm I put the Latin's own mark back — nine such corrections in 0020 alone.

### 6.5 Test 2b — anchors checked for POSITION, not just presence All 47 anchors were checked against the word they stand beside in the Latin, not merely for presence and order. Six needed the English clause rebuilt to keep them there rather than letting them drift to the end of a sentence: 0232D (dilectionem [ ] nemo), 0234B (saecularis [ ] circumitus), 0236D (martyrum [ ] gloriatur), 0240C (resurrectione [ ] glorificata), 0242B (gloriam [ ] post), 0243C (remittuntur, [ ] peccata). The last is the one worth naming: ea quae in baptismo eis remittuntur, peccata, extraxerunt puts peccata in apposition, so the English reads "…which are remitted to them in baptism, [0243C] their sins" — the anchor lands exactly where Migne has it because the apposition was seen. Read as a direct object instead, the anchor has nowhere to go.

### 6.6 Test 5 — does the English ASSERT what he asserts Every paragraph was read back against the Latin for this and not for containment. Three sites where a smooth English reading was available and was rejected as the wrong assertion:

• 0239D emisit smoothed to match tradidit — rejected; see §3.3. The seam is the argument.
• 0234A Jacob read as a subject rather than the object — rejected; see §5.
• 0241A Ubi est, mors, aculeus tuus reset to "O death, where is thy sting" — rejected. It reads better and it is 7a″'s conformation exactly.

### 6.7 Brief §4 / addendum C — the normalization class (a form that is suspiciously tidy) Read for, and nothing raised. ⚑ Recording the negative because addendum C's own point is that two of its four instances were found while reading a neighbouring word. The candidates I looked at and let go: 0234B obreptionis (rare but classical, and obreptio is standard in this register — rendered "of stealth"); 0236B apompeius (a Latinization of ἀποπομπαῖος, printed roman and left roman, "the emissary"); 0238D ἀναβόλαιον (Gen 49:11's Greek, carried verbatim); 0242C simplae/duplae (real adjectives, deliberately paired — "single"/"twofold" — and the whole paragraph's arithmetic depends on them).

§7 · What I want the plate read for, in priority order

1. 0237D, curcifigentes cum — one line. Decides whether the range needs a [sic:] or a tei-patch, and it is in a lemma.
2. 0240C, us que ad and 0242C, ve peram — same class, same page range, cheap to take in the same look. ⚑ Per addendum C, read the whole line at each of the three, not the target word.
3. *0240A, [n: (Matth. X)].;** — is the stop-plus-semicolon Migne's, or the digitization's?
4. 0242C, the doubled sequenti nocte — if the plate prints it once, that is a [d:] we do not need and a patch we do.

⚠ The launch note says our PL 114 scan clips both margins on leaves in this range. All four requests are mid-column, none at a margin, so a clipped leaf should still answer them.