Evangelium secundum Joannem

Merged from the six translation stints' files (cruces-0000, -0005, -0010, -0015, -0020, -0024) 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 §2 has been silently altered; the two corrections that were made to a stint's own text are named in §0.4.

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 (AUG., BEDA., HIER., ALCUIN., GREG.) arriving as ordinary notes. Migne's attributions in this recension are known to be unreliable, and passing a siglum through is not endorsing it.

John is printed inside the Glossa proper, under the running head GLOSSA ORDINARIA. — EVANG. JOAN. — unlike 9003 and 9004 in the same volume, whose bylines were corrected off the Glossa author. There is no attribution question here.

§0. THE PLATE READ — 23 of 23 [var:] stand, and every correction came from the DECLINED list

Source. archive.org patrologiaecursu0114mign — a second copy of PL 114, colour, 2215×3603, clean at both margins where our on-disk scan (patrologiaecurs06saingoog.pdf) clips them. Established and calibrated on 9001, so this work never had to stop and hunt one.

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

leaf n = (column + 11) / 2, i.e. leaf n carries columns (2n−11) and (2n−10). Calibrated on this work's own range at n183 = cols 355/356, and each of the six readers calibrated independently on the printed corner numbers of its own first leaf before reading anything. ⚠ The endpoint returns HTTP 200 with an empty body to a bare curl; it needs -L and a browser user-agent.

23 leaves, 26 markers, all read. Recorded in data/plate-reads.json under 8999; per-reader files in data/plate-findings/8999-plateA.md … -plateF.md.

§0.1 The result

No stint fired a single [sic:] in 29,728 words, so nothing could fall. That is the launch brief working as designed, not an absence of defects: §2 of the brief told the stints their attribution guess carries no evidential weight, and to log a suspect word unmarked for the plate instead. They logged nineteen.

⭐ With 9004 (46/46) and 9001 (19/19), that is 88 of 88 [var:] standing across three Glossa books. Suspicion of a quoted clause is evidence about Migne. Suspicion of a single word is suspicion of our file.

§0.2 ⭐⭐ EVERY ONE OF THE 17 CORRECTIONS CAME FROM THE DECLINED LIST

The 23 leaf-reads spent on fired markers produced no corrections at all. All seventeen TEI patches came from the nineteen oddities the stints logged and refused to attribute, plus lines read in passing. A stint that had guessed would have spent its credibility on exactly the wrong sites — and it is worth saying that the reverse also held: the markers they did fire were right 26 times out of 26.

§0.3 ⭐⭐⭐ THE NON-WORD PRIOR DOES NOT HOLD WITHIN A SINGLE WORK

Same author, same recension, same volume, ~200 columns apart. So reference_plate-sweep-rate-varies-by-work is too weak: the rate varies by RANGE inside one work, and no policy may be inherited from a sibling book. The only safe procedure is the one the brief already mandates — log, do not attribute, and read the plate.

⚑ 9001 @0203A magnitudino remains the standing counter-example on the other side: a genuine non-word that Migne prints. A non-word is not automatically ours, and manao is explicitly not that case — its plate is clean.

§0.4 ⚑ FIVE WITHDRAWALS ARE ONE MECHANISM: THE HYPHEN-REJOIN CLASS

Corpus Corporum mis-rejoining a word Migne broke at a line or column turn:

Four readers found it independently, on four different leaves, unable to see each other. It manufactures a plausible non-word and no frequency test can distinguish it from Migne's own. It is greppable, and it is a candidate pass over other works. The 0361A case is the sharpest: CC resolved the hyphenation and kept the orphan syllable as well, so our file printed the syllable twice.

⚑ The reader who settled interrogate noted that the same column prints interroges, interrogeris, interrogavit and interrogatus correctly — and explicitly refused to rest its verdict on that, because corroboration proves an error exists and never whose it is. The ground was the printed inter- / rogare.

§0.5 ⚑ AND THE READ RAN IN BOTH DIRECTIONS — FIVE SILENT REPAIRS

Sites where Corpus Corporum quietly fixed Migne's own broken type, i.e. where a [sic:] would have been true and our clean file hid it:

This is the class no Vulgate collation can ever reach, because the digitization removed the evidence before any collation could run — the mirror of 9003's manufactured judicabit → judicavit. Migne's readings are restored in data/tei-patches/8999.json; the [sic:] each one owes is deliberately NOT fired, because firing one is a fresh public claim per site.

⛔ A recorded disagreement. The reader who found the last two argued against restoring them, on the ground that it "would introduce two non-words." That is precisely the reasoning the NORMALIZED convention exists to override, and 9001 settled it by restoring Tlbi (a wrong dotless-l sort) and cruxifixum. Our Latin is not a corrected Latin; it is Migne's page. The objection is recorded rather than silently resolved.

§0.6 ⭐ Panes nostri IS MIGNE'S — the hardest decline in the work, and the stint would have guessed wrong

0373B, Jo 4:20. The gloss four words later is entirely about patres adoraverunt and patriarchae in montibus, so "Our loaves" looks indefensible. Read at 12× on a native crop, the plate sets an unambiguous italic n — no t ascender, no r flag — with the genuine tr of nostri two letters away as an on-page control. Migne's. The stint declined for the right reason and had leaned toward Patres. A [var:] against Jo 4:20 is now fired and the English stands as printed. Fourth measured case in this project of a confident attribution refuted by the page.

§0.7 Two corrections made to the stints' own text

1. cruces-0000 §c.14 reported that super Filium hominis occurs in the Clementine "zero times". That was a case-sensitive search error: it occurs twice, Ps 79:16 and 79:18. The finding itself survives and is the cleaner for it — the ✓ came from a different verse, which is the documented trap in its pure form. The pre-scan had no defect at that site.
2. cruces-0003's 0365B [var:] quoted the normalised constringitur; corrected to name Migne's printed contringitur.

§0.8 ⛔ A TRAP IN THE PLATE ITSELF, AND ONE IN OUR ANCHORS

Leaf n215's printed right-hand corner reads 421, not 420. Verified at 6×. Leaf n216 then reads 421 / 422, so 421 is printed twice in PL 114 and 420 is never printed at all — the compositor mis-numbered it. The running text settles it (n215-right runs VERS. 37 Tu dicis → Sicut agnus coram tondente, and n216-left opens with that sentence's continuation, our [0421A]), so n215-right is column 420. A reader trusting corner numbers alone would misfile every finding on that column by one.

And a <pb> anchor is not a locator for a phrase near it. Four sites sit in a different band from the anchor that precedes them — actionem personarum at 386D not 0387A, the [cj:] nobis at 414D not 0415A, liberatum at 0420A not 0420B, Separatur at 0422B not 0422A. Two of those errors were mine, in the readers' briefs, and both came from reading the TEI <pb> position instead of the plate. This is CLAUDE.md's band-arithmetic false positive seen from the other side, it is not a defect, and nothing was patched.

§1. WORK-WIDE CONVENTIONS — tested at the merge, and only one needed a sweep

⭐ This is the first Glossa book to carry the conventions INTO the translation rather than reconcile them afterwards. The launch brief handed all six stints Mark's four settled axes before they wrote a line. The merge cost on 9001 was 100+ edits and on 9004 was 147. Here it was 15, all on one axis.

Every axis below was measured mechanically against the Latin, not judged by impression.

1. usque ad → as far as — the one real drift, 15 sites swept

The abridgment formula, always italic and always following etc.,. Each stint was internally consistent; the drift was purely between stints, and it split three against three:

Settled on as far as — the within-work majority (26 to 15) and the form 9001 uses throughout, so Mark and John now agree. All 15 normalised; the count is now 41 / 0.

⚑ Only the italic formula was touched. Unitalicised "down to" is ordinary prose ("goes down to Capharnaum", "down to the last times of Domitian") and was left alone — contrasting two different word classes is how the 9001 merge nearly manufactured a throng/turba inconsistency that was not there.

⚑ Stint 6 flagged this axis but misread its shape, reporting a 2:1 split with chunk 0000 on the down to side. 0000 carries no instance of the formula at all; its "down to" is prose.

2–5. The four inherited axes — ALL HELD, NO EXCEPTION FORCED

Measured across the whole work:

natio does not occur in this work at all (0 instances), so nothing is flattened by the gentes convention. All ten English "nation"s fall in the classes convention 4 permits: the generic sense (divisions after the flood, omnes gentes at Matt 28:19, "the wise men of the nations"), or singular gens of a people ("the Jewish nation", "their own nation", pro gente at Jo 11:50, gens sancta at 1 Pet 2:9). Twenty-eight "Gentiles" carry the typology.

⚑ A measurement trap, recorded because it cost time and would cost it again. A naïve \bturba grep matches the verb turbare (turbavit, turbatur, turbata — "to be troubled"), which is a different word from the noun turba, and it made chunks 0017–0020 look like five unrendered "crowds". They contain none. \bdaemoni likewise misses daemones, and BSD grep -o '\bnation' matches inside incarnation, damnation and predestination. Three separate false drift signals in one merge, every one an artifact of the measuring instrument. Check the instrument before reporting the drift.

§1a. THE BLIND POLARITY READ — 26 sites, 19 ours, and it overturned two closed rulings

Three blind Opus readers, disjoint ranges (0000–0009 / 0010–0018 / 0019–0027), all 28 chunks, each fenced from this file and from every cruces-*.md until its own findings were on disk. Per-reader reports: data/polarity-findings/8999-reader1.md … -reader3.md. Recorded in data/polarity/8999.json.

26 sites adjudicated: 19 ours and repaired, 5 confirmed as Migne's, 2 withdrawn by their own reader on the merge's evidence. Mechanism 5 accounts for 11 of the 19 — a reference silently re-bound to the contextually easier noun, the dominant class on 9001 (5 of 6), 9004 (5 of 8) and 11613 (12 of 25), and dominant again here.

§1a.1 ⛔ ALL THREE PARTICLE TALLIES FOUND NOTHING, AND THE WORST SITE SCORED CLEAN

Every reader ran a per-chunk tally of non/nec/neque/nisi/haud/nunquam/nullus/ nihil. Not one finding in this work came from a tally. Chunk 0016 scored exactly 57/57 and carries the flattest self-contradiction in its range; chunk 0022 scored clean and carries the worst defect in the book.

⭐ And the reason 0022 scored clean is worth writing down, because it defeats every mechanical check we have. At 0415A the stint correctly diagnosed Migne's nobis for non and fired a [cj:] — but parked the negation inside the bracket, so the running English read "they to such a degree understand for us". The marker's text contains the word "not", so the paragraph looks negative to any counter, to verify-english, and to the reader's own tally, while the sentence a human reads asserts the opposite of itself and of three neighbours on one line ("proverbs to men who do not understand" before it; "they do not even understand that they do not understand" seven words after). This is translation-style.md 7a⁗ in its purest form: the force must be carried in the running text, not in the apparatus. Repaired.

§1a.2 ⭐⭐ THE READ OVERTURNED TWO RULINGS THIS FILE HAD ALREADY CLOSED

This is the argument for blindness, and neither could have been found by re-reading the apparatus.

1. 0366C — logged here as MIGNE'S defect, and it is OURS. The crux ruled the printed et might be for nec and queued a plate check. But non already governs both perfect subjunctives (si non natus fueris … et [non] susceperis), and Chrysostom's Greek has καί under a single μή. The Latin is sound. The English negated only the first limb, so the page said that a man who has received the certainty of the doctrines is the one far from the kingdom — contradicting its own apodosis. Compounding it, alicubi had been moved out of the apodosis into the protasis, which is exactly what made the un-negated limb read tolerably and so hid the fault. Repaired; the ruling and its plate-check item are struck.
2. 0409B — the crux's justification was not in the text it justified. The [cj:] was declined on the express ground that modo construes as "only" — and the shipped English printed "presently", the construal the crux had rejected. Plate reader D confirmed no negation stands on the page, so the decline was right and the rendering was not. Repaired to "only". ⚑ This is the 9001 @0199A class again: a crux that records a decision the text does not carry. No re-reading of the apparatus can ever surface it, because the crux says the right thing.

§1a.3 The eleven mechanism-5 repairs

The worst four, each of which produced a doctrinally false sentence in fluent English:

Also repaired: 0374D (inducit/affirmat given to Christ where the speaker is the Samaritan woman — and two entries earlier the identical formula Pedetentim … instruit genuinely is Christ, which is what made the wrong binding plausible), 0422B (the Church entering after the Church; the subject of intrabit is the Synagogue, and the Rom 11 eschatology was lost), 0387A, 0395A, 0403B, 0411D, 0397A.

§1a.4 Four repairs that are not mechanism 5

• 0366C — negation scope, above.
• 0386D — the non sits on the infinitive in Latin (non solvere vos sabbatum creditis, "you believe you do not break the sabbath") and had been raised to the main verb ("you do not believe that you break"), which is an absence of belief where the argument needs an assertion.
• 0390C — a silent emendation of the printed Latin, undone. Migne prints ut nos; the English had quietly corrected it to "as you do". The emendation is almost certainly right in sense (Christ is answering Vos secundum carnem judicatis) — but a defect repaired invisibly tells the reader the plate says something it does not. Now rendered literally with a [cj:].
• 0406A — passive cogitur made active, and et ("even") dropped, which is the word carrying the reluctance the whole gloss is about.

§1a.5 Confirmed as Migne's — two given a [cj:] the plate licenses, three logged

• 0385D audit — the sentence says the world does not hear its own false witnesses, when the verse glossed is Jo 7:7 odisse and the whole point is that the world does not hate its own. Plate reader C confirmed the au is unmistakable at 2.3× — the reading is Migne's — so a [cj: audit; read odit] is exactly right: it conjectures on his text and asserts nothing about whose the defect is. Fired. ⚑ Reader 2 dissented from the stint's "odd, not false" decline and was right; this reads as perfectly smooth English and asserts a falsehood.
• 0387D assensionem — plate reader C confirmed the doubled ss. [cj: assensionem; read ascensionem] fired; the same chunk expounds the ascension throughout.
• 0422D Noli peccare — the next sentence says what the angels forbid is her tears, so the expansion wants Noli plorare. [cj:] fired on a column recorded read.
• 0396D, 0384A, 0401A — logged, no marker. 0396D is the clearest contradicts-its-own-neighbour in the book (both halves of an e contra antithesis end sed in fine perseverant, so the page asserts the reprobate persevere) and plate reader D confirmed both clauses really do read that way; no determinate conjecture is available, so none is offered.

§1a.6 What the readers withdrew, and why that matters

Reader 3 withdrew two of its own findings on seeing this file: the four "silent non-word corrections" it flagged are the ruled NORMALIZED policy of §0.5, and the 0410D full stop is Migne's own. Reader 1 withdrew its quodum objection as deliberate Pattern 7 policy — and the merge then read the plate anyway and found the reader's instinct right: leaf n191 prints quodam, so the word was ours and is now patched and translated. Reader 2 conceded 7 of its 15 as already known and "usually better reasoned" there.

⚑ A structural observation from reader 2 worth keeping. This file's mechanism-5 census names three sites, all inside chunks 0010–0011 — and all four of that reader's serious re-bindings sit in 0012–0018, the range that never got a census. A stint's own apparatus is a complete record of what it looked for and silent about what it did not.

§2. THE STINTS' OWN FILES, VERBATIM

Six files, unabridged and in range order. Each states the policy it applied before its findings, so it can be judged; each records its span count against the brief's total. Where §0 overturned or corrected something below, the correction is named there, not here.

<!-- ===== cruces-0000.md ===== -->

Stint 1 of six. Cols 0355B–0369C (Prologue; CAP. I, II, III to VERS. 29). Written against data/briefs/8999-launch-brief.md, translation-style.md, the 9741/0000 anchor pair, src/english/9001/cruces.md §1, and data/briefs/8999-lemmata-0000.txt.

(a) THE POLICY I APPLIED — stated before the findings, so it can be argued with

1. Attribution. I made no attribution claim anywhere, and I fired no [sic:]. Brief §2 is the governing rule and I took it literally. Every defect I met in this range, whether a real word reading oddly or a broken run, is rendered as printed (or, for the one duplicated syllable, read through in the manner 9001's stints used) and logged below with no marker and no guess about whose it is. Where I have an internal argument — the corrected form prints eleven words later, the gloss expounds the other reading, the run breaks exactly at a column anchor — I have written the argument down and then explicitly declined to convert it into a class claim, because 9004 measured that exact reasoning at 5 right and 9 wrong. §0 below is a list of things for the plate, not a list of accusations against Migne.

⚑ Corollary I want on the record: sciebat/bat at the 0361A break (§0.1) is the site I would have marked under any earlier convention. I did not. If the plate confirms it is ours, the right outcome is a data/tei-patches/8999.json entry, not a marker.

2. [var:] is the only marker class I fired — eight of them, every one on a QUOTED CLAUSE, every one cited to sources/vulgate/clementine-flat.txt by book chapter:verse. None rests on the brief's pre-scan or on recollection; I read each verse out of the file. The measured record (9004 46/46, 9001 19/19, 11535+11550 91/92) says a clause-level [var:] is evidence about Migne and a word-level suspicion is evidence about our file, so I let that asymmetry set my threshold: a divergence gets a marker when it is a clause, citable, and changes what the verse says; a divergence of word ORDER or ORTHOGRAPHY alone is logged and not marked, however real.

3. Where the gloss itself adjudicates the divergence, I did not mark it. Two sites (Joanna/Jona at 0362A, and the whole 0356C punctuation dispute) print the rival readings side by side in Migne's own type and expound the difference. The divergence is already on the page; a [var:] there annotates what the reader has just been told. This is the Nicetas calibration (a plate that carries its own apparatus settles its own marker) applied to a gloss.

4. Inherited Mark conventions (§1 of the brief) applied from the first line. dicit/ait → says, never saith, in the commentator's voice; -eth/thou only inside quoted scripture or a printed singular address. mysterium → mystery, sacramentum → sacrament, 1:1. diabolus → devil. No exception was forced. ⚑ The turba/multitudo and natio/gentes axes have no instance at all in cols 0355B–0369C — neither word of either pair occurs — so I have set no precedent on them and later stints should not read my silence as one. Judaei occurs and is the Jews throughout; Pharisaei the Pharisees; Israelita an Israelite.

5. Pattern 7 / 7a / 7a⁗. Every printed negation is carried, including where it makes the sentence fight itself (0357C sed non in resurrectione; 0369A). Nothing is translated as its opposite to rescue a clause. Punctuation follows the plate: the missing full stop at 0368C (diligit mala Et hoc est) is carried as printed and not supplied. Zero em-dashes in all five chunks.

6. Mechanism 5 was the class I checked hardest, it being 5-of-6 on 9001 and 5-of-8 on 9004. Four sites in this range could have bound a reference to the easier noun; §0.3 lists them with the case and gender that decided each, not the sense.

(b) SPAN COUNT — RUN, AND IT AGREES

The brief states 168 italic spans for chunks 0000–0004. I counted the … spans in my own Latin files:

168 = 168. The split is correct for this range, so I am neither short at the head nor carrying a tail that belongs to stint 2. (0000 is the Prologue and carries no italics at all, which is why the count is concentrated in four files.) Marker parity was also run per file: column anchors 15/14/10/14 and body [n:] 8/5/10/3 all reproduce 1:1 against the Latin twins.

(c) THE LEMMA DIVERGENCE LIST — all 168 spans walked

Verdicts: = agrees with the Clementine at the verse glossed (allowing for a lemma clipped mid-phrase, which is this recension's normal habit and is not a divergence) · ≠M a real divergence, marked [var:] · ≠d a real divergence, declined with the reason given · g not a scripture citation at all (gloss-side etymology, re-narration, or an abridgment formula), so 7a″ has nothing to compare.

Chunk 0001 — cols 0356C–0360A, 49 spans

Chunk 0002 — cols 0360B–0363C, 47 spans

Chunk 0003 — cols 0363D–0366A, 31 spans

Chunk 0004 — cols 0366B–0369C, 41 spans

The eight I marked

• §c.5 · 0358A [var:] — Sta in foramine petrae, et cum pertransiero videbis posteriora mea, under an Exod. XXXIII siglum. The Clementine has no such sentence: Ex 33:21 stabis supra petram, 33:22 ponam te in foramine petrae … donec transeam, 33:23 videbis posteriora mea. Migne's is an imperative conflation of three verses, and the gloss expounds its own wording word by word — petra, foramen, pertransiero, posteriora — so the divergence is load-bearing, not decorative. Marked.
• §c.6 · 0358A [var:] — ut cognoscant te unum verum Deum for Jo 17:3 ut cognoscant te, solum Deum verum. A lexical substitution (unum for solum) inside a clause adduced as the definition of eternal life, in a paragraph whose subject is the knowledge of the Trinity. Marked.
• §c.7 · 0358A [var:] — Ostende mihi teipsum under Exod. XXXIII. The Clementine offers ostende mihi faciem tuam (33:13) and Ostende mihi gloriam tuam (33:18); teipsum is in neither, and the gloss's argument is precisely that Moses asked to see God himself and did not. Marked, both candidates named.
• §c.8 · 0358A–C [var:] — nemo novit Patrem nisi Filius under Luc. X. Lc 10:22 prints nemo scit quis sit Filius, nisi Pater: et quis sit Pater, nisi Filius; the Matthean parallel (Mt 11:27 neque Patrem quis novit, nisi Filius) is nearer but is not what the siglum cites. Both given in the marker. ⚑ I state the two readings and no more: whether this is Migne's compositor, his copy-text, an Old Latin form or the glossator quoting from memory is a research question, and Pattern 14 forbids me to guess at it.
• §c.10 · 0360A [var:] — the lemma reads Peccata mundi, plural; Jo 1:29 reads peccatum mundi, singular. The gloss immediately expounds the singular — Peccatum mundi dicitur originale peccatum — so the printed lemma and its own exposition disagree in number on the same line. Marked on the lemma. (I did not also mark the gloss-side qui tollit peccata at 0360B: that is the commentator's own prose, not a quotation.)
• §c.15 · 0365A [var:] — Funiculis peccatorum suorum quisque constringitur under Prov. V. Pr 5:22 prints et funibus peccatorum suorum constringitur, with no quisque. Material because the whole paragraph turns on funiculi, the cords of Jo 2:15, and Migne's proof-text carries the lemma's own word where the received text does not. Marked.
• §c.17 · 0368D and 0369A, two [var:] — Jo 3:22. The lemma prints In terram Judam, et illic demorabatur where the Clementine has in terram Judaeam : et illic demorabatur cum eis; and eight lines later the same clause is re-quoted as et ibi morabatur et baptizabat — a second, different departure from illic demorabatur, inside our own range and contradicting our own lemma. Both marked, and the internal disagreement stated as the fact it is, with no claim about its origin.

The eight I DECLINED, and why

• §c.1 · 0356C, three spans — the punctuation dispute at Jo 1:3–4. Nihil est quod factum est sine ipso (labelled Ordo:), Quod factum est, erat vita in ipso, and Hilary's Sine ipso factum est nihil, quod factum est in ipso. All three transpose or re-divide Jo 1:3–4 rather than disagreeing with it, and re-dividing that sentence is the entire subject of the paragraph: Migne prints Origen's division, Chrysostom's and Hilary's one after another and labels them as such. A [var:] would annotate a divergence the page has just declared. Declined, all three.
• §c.2 · 0357D — Verbum caro factum est: vidimus gloriam ejus plenam gratiae et veritatis, et de plenitudine ejus nos omnes accepimus. Against Jo 1:14 + 1:16 this drops gloriam quasi unigeniti a Patre and reads plenam (agreeing with gloriam) for the received plenum (agreeing with Verbum) — the celebrated grammatical crux of that verse, resolved in Migne's favour. ⚠ This is the closest call in the range and the one I would most like a second opinion on. I declined because the clause is introduced by Quasi dicat — an avowed paraphrase, the evangelist made to summarize himself — and Pattern 14 governs quotations, not the gloss's own reconstructions. If the merge takes the other view, the marker is [var: Vulg. Jo 1:14 gloriam quasi unigeniti a Patre plenum gratiae et veritatis] and I would not argue.
• §c.3 · 0358A — Deum nemo vidit unquam* for Clementine umquam*. Orthography, both forms standard; Pattern 9. Declined.
• §c.4 · 0358A — Non videbit me homo et vivet: Ex 33:20 reads non enim videbit me homo et vivet. A dropped connective at the head of a clipped quotation, no sense change. Declined.
• §c.9 · 0358D — Non sum Elias. Jo 1:21 prints only Non sum; the predicate is supplied from the question two words earlier (Elias es tu?), which the lemma itself quotes. Nothing is added to the sense and nothing is taken from it. Declined.
• §c.11 · 0360C — vidit spiritum. Third-person, inside the gloss's own sentence (ita perhibuit testimonium, scilicet …), against Jo 1:32's first-person vidi Spiritum. This is the glossator re-narrating, not citing; the italics are Migne's way of pointing at the verse, not of quoting it. g, not a divergence. Declined.
• §c.12 · 0360D — In labore et mente sunt differentes, in operatione indifferentes, under unde Apostolus. Not in the Clementine anywhere, in any form; I searched the whole file. ⛔ I decline a [var:] on the ground that there is nothing to cite: Pattern 14 requires a received reading, and I cannot supply one for a sentence I cannot locate in the canon. Logged here instead so the plate read has it. The attribution to Apostolus stands untouched per the brief's rule on this recension's sigla.
• §c.13 · 0362A — Filius Joanna for Jo 1:42 filius Jona. A real divergence. Declined because Migne's very next sentence adjudicates it: Jona dicitur et Joanna indifferenter. Jona columba, Joanna gratia interpretatur. Both forms are on the page, in his type, with their etymologies. A marker here would tell the reader what the gloss has just told him.
• §c.14 · 0363C — Super Filium hominis (twice) for Jo 1:51 supra Filium hominis. ⚑ Worth flagging as a method finding: the brief's pre-scan marks this span ✓ Clementine verbatim, and the ✓ came from somewhere other than the verse glossed, exactly as the brief warned a ✓ can. ⛔ CORRECTED AT THE MERGE: this stint reported that super Filium hominis occurs in the Clementine zero times, which is a case-sensitive search error on its part. It occurs twice — Ps 79:16 and Ps 79:18, et super filium hominis quem confirmasti tibi — and the pre-scan is therefore behaving exactly as lemma-inventory.mjs specifies (it lower-cases both sides). The tool has no defect here and must not be "fixed". The finding itself survives untouched, and is in fact the cleaner for it: this is the documented trap in its pure form — a lemma matching a different verse verbatim while diverging from its own. The divergence is real; I declined it because super/supra are interchangeable prepositions here and no sense turns on the choice. But a stint trusting the ✓ would never have looked.
• §c.16 · 0368A and 0368C — two transpositions in Jo 3. Sic enim dilexit Deus for Sic enim Deus dilexit; and Erant enim opera eorum mala for erant enim eorum mala opera. Word order only, no sense change. ⚑ The second is the more interesting: the same chunk prints Erant enim eorum four lines earlier, in the received order, so Migne has the clause both ways within one paragraph. Logged, declined, and handed to the plate.

(d) THIS RANGE'S OWN CRUCES, BY COLUMN

§0.1 The one site I would have marked, and did not

0360D–0361A — nam sciebat / [0361A] / bat eum Filium Dei esse. The syllable bat stands alone at the head of column 0361A, immediately after a complete sciebat at the foot of 0360D. The shape is a hyphenated line-break (scie- / bat) whose first half has been expanded and whose second half has been kept.

⛔ I do not know whose it is and I have not said. The reasoning that it must be the digitization's — a duplicated fragment, sitting exactly on a column boundary, no lexeme lost — is the reasoning 9004 measured at 5 right and 9 wrong, and 9001's magnitudino proved that a genuine non-word can be Migne's. I have therefore read through it (the English reads "for he knew him to be the Son of God"), exactly as two of 9001's five stints did with their six non-words, and fired nothing. The verb survives, so 7a⁗ is satisfied. Plate: leaf n186 (cols 361/362), left column, band A, first line. If it is ours, this is a 8999.json patch, not a marker.

§0.2 Real, well-formed words that read oddly — logged, unmarked, no attribution claimed

§0.3 Mechanism 5 — the four sites where a reference could have been re-bound

Checked on case and gender, not on which reading reads more easily. Recorded so the blind read can disagree with a stated decision rather than guess at an unstated one.

1. 0358D–0359A, praecessurus … judicaturum … judicandum. Sed ipse Elias praecessurus, in potestate judicaturum, Joannes figuraliter Elias eumdem judicandum. The two accusative participles are active and passive and they are the entire point: Elias goes before the Christ who will judge, John before the same Christ who will be judged, so that the two heralds answer the two advents. Collapsing them into one voice destroys the antithesis. The English keeps them opposed.
2. 0367C, qui omnibus praeest. Non insultat ei, ut superior eo habeatur qui omnibus praeest. The relative attaches to the subject of insultat — Christ, who is set over all — not to Nicodemus, the nearer noun and the easier reading. Nicodemus is the magister in Israel; he is not set over all, and the next clause has Christ willing him to be born of the Spirit. Rendered accordingly.
3. 0360B, reconciliat nos ei. Filius Dei in quo Patri congruit et peccata dimittendo reconciliat nos ei. ei is the Father, named three words earlier in the dative, not Christ. Rendered "reconciles us to him", with the Father as the antecedent established by the clause it stands in.
4. 0361C, illi personam respondent. qui rem, non personam, quid, non quem, ne se videatur ostendere, illi personam respondent. Two different subjects in one sentence — the questioner asks a thing, they answer a person — and the easy reading merges them. Kept distinct.

§0.4 Structure and punctuation carried as printed

• 0359A — an italic span closing over a proper name. Migne's italics run …Ecce agnus Dei. Isaias, and then break to roman for Hoc de Joanne prophetavit. The name is inside the lemma's italics though it belongs to the sentence that follows. Reproduced with the boundary exactly where the plate has it; the English reads …Behold the Lamb of God. Isaias, / prophesied this of John. Not a defect I can attribute, and no marker: an italic boundary is not type.
• 0368C — a full stop that is not there. …odit lucem et diligit mala Et hoc est quod dicit: Pattern 8 forbids supplying a mark the plate lacks, so the English runs on in the same way ("…and loves evil things And this is what he says:"). ⚑ 9001 @0202D is the neighbouring case in the opposite direction — a full stop the digitization added at a hyphenated break — and that stint was right to carry it and say it could not tell whose it was. Same here.
• 0365C — Quia signum quaerunt quare solita commercia de templo ejicit? A question mark closing what is grammatically not a question. Preserved (Pattern 8).
• 0367B — Quae quia incognita sunt carnali, iterum quaerit? Quomodo possunt haec fieri… The question mark falls before the question rather than after it. Preserved.
• 0365C — Anatole … dysme … arctos. Greek words in Latin letters, as Migne sets them. Left in Latin script; §5's single Greek-script word is at chunk 0023 and does not concern this range.
• 0364B, 0368B — triclinium, mundus. Kept in italic Latin where the gloss is expounding the word itself (the triclinium as three ranks of diners; Cosmos/mundus/ornatus), per the Register rule for terms whose English equivalent would make the gloss explain nothing.

§0.5 One note that stays a note

The [n:] markers in this range are all citations or attribution sigla — (AUG.), (ORIG.), (HIER.), (CHRYS.), (ID.), (Exod. XXXIII), (Ibid.), (Luc. X), (Isa. V), (Prov. V). None is editorial prose, so no [nt:] conversion is owed anywhere in chunks 0000–0004. The substantial extracts under [n: (AUG.)] at 0366C, [n: (CHRYS.)] at 0366C and [n: (HIER.)] at 0361A are body text following a siglum, not note contents — the marker is the siglum alone — and they are translated as body text, which is how 9001 handled the same shape.

§0.6 What I am handing to the plate read, in the order I would look

1. 0361A bat (§0.1) — the only site in the range where a marker was genuinely in question.
2. 0359D circa / citra (§0.2) — and read the whole line, per 9001's addendum C, which found two real defects four words away from what it went looking for.
3. 0362B infirmus — decides nothing in the translation, but decides whether the figure is a paradox or a slip.
4. 0357C sed non in resurrectione — the sentence contradicts itself as printed.
5. 0366C et dogmatum susceperis — a connective, and mechanism 3 leaves no trace.
6. 0360D In labore et mente sunt differentes… (§c.12) — an unlocatable "Apostle."

<!-- ===== cruces-0005.md ===== -->

Stint 2 of six. Cols 0369C–0380C (chunk frontmatter bands 0369D–0380C), covering John 3:30–36, 4:1–54 and 5:4–43. Governed by data/briefs/8999-launch-brief.md; the four Mark conventions of §1 were applied from the first line, not reconciled afterwards.

(a) THE POLICY I APPLIED — stated before the findings, so it can be argued with

1. Attribution. I fired no [sic:] and no [cj:]. Every defect I found is logged below with the printed form named and no claim about whose it is. Where our Latin holds a real, well-formed word that reads oddly, and where it holds a non-word, I rendered or carried it as printed and said plainly that I cannot tell Migne's type from Corpus Corporum's transcription. This is the brief's §2 operating rule taken literally, and it cost me two findings I would otherwise have marked with confidence — 0373B Panes and 0379D Sicut autem, both below. On the brief's own measurement, that confidence would have been worth nothing, and both sites are handed to the plate read intact rather than half-decided.

2. [var:]. I fired three, all on quoted clauses, all cited to sources/vulgate/clementine-flat.txt by book chapter:verse, and all where the gloss's own argument rests on the divergent words:

Each is a multi-word, semantically coherent divergence of the kind an OCR letter-swap cannot manufacture. I declined every single-word divergence, without exception — including four I believe are real errors (0372B Dixit, 0379C veniet, 0373B Panes, 0379D autem). That line is deliberate and it is where the brief's measured class split puts it.

3. Negation. Every non, nec, neque, nisi, nullus, nihil, nunquam the Latin prints is in the English, including where it makes the sentence awkward (0377A nec in potestate ideo magnum, rendered "neither in power was it therefore great"). None was added.

4. Mechanism 5. At each site where a pronoun, relative or participle could bind either of two nouns, I checked case and gender and rendered so as to keep the ambiguity the Latin has rather than resolve it to the easier noun. The four sites are listed in §(d): 0370B se distinguit a Christo, 0371D privaverunt, 0375C Judaeis spiritualiter edoctis, 0377B quae … cessare jubet.

5. Register. Gloss voice not archaised; no -eth on a metalanguage verb anywhere in my range. Every saith (3) sits inside an italic lemma rendering Dicit. Pattern 17 applied as a NUMBER rule: thou for singular tu (the Samaritan woman throughout ch. IV, the healed man at 0377A–B), you for plural vos (the Jews at 0378A, 0380A–B).

6. Repairs that would have been invisible, and were not made. Two are recorded in §(d) as written and removed, because 7a′ says a crux describing a repair is not the same as a text that carries the ruling: the supplied ? at 0376D, and the supplied ? I nearly put on 0377C's cur mihi invidetis … operor.

(b) SPAN COUNT — RUN, AND IT AGREES

data/briefs/8999-lemmata-0005.txt declares 148 spans for chunks 0005–0009. Counting the … spans in the body of my own five Latin files (frontmatter excluded):

148 = 148. I went further and diffed the two sequences rather than only the totals: the brief's 148 span texts and my 148, both in document order, are byte-identical, line for line. The head of my range (Illum oportet crescere, 0369C) and its tail (Si alius, 0380B) are both mine, and neither is a neighbour's. scripts/split-lemma-brief.mjs is correct for this range.

Marker parity was checked the same way, per chunk, against the Latin twin: column anchors 4 / 15 / 10 / 14 / 1, [n: …] notes 1 / 5 / 1 / 5 / 0, body lines 7 / 31 / 19 / 31 / 1 — all equal to the Latin, all in the same order and position. verify-english.mjs 8999 reports no warning against any of my five chunks; the only failures it prints are the missing chunks of other stints' ranges.

⚑ One anchor was lost and restored during drafting: [0380C] was dropped from 0009 on the first write and put back before anything else was done. Recorded because the brief asks that markers be counted before writing each file, and this is what that check is for.

(c) THE LEMMA DIVERGENCE LIST — all 148 spans, in order, with verdicts

Verdict key. conforms = the printed Latin agrees with the Clementine at the verse being glossed, allowing for the span's own cut (Migne quotes incipits and tails constantly, and a cut is not a variant). n/a = not a scripture span (an usque ad abridgment formula, or the glossator's own words in italic type). DECLINED = a real divergence from the Clementine on which I fired no marker, with the reason. MARKED = a [var:] fired.

⛔ The pre-scan's ✓ cleared three genuine divergences — nos. 37, 139 and (partly) 20 — exactly as its own header warned it could: the sequence occurs somewhere in the Clementine and so passes, while diverging at the verse in hand. No. 139 is the sharpest instance, and I found it only by reading the line rather than the mark.

Chunk 0005 — cols 0369C–0370C (John 3:30–4:1)

1. [0369C] VERS. 30.-- Illum oportet crescere. — conforms, Jo 3:30.
2. [0369C] Illum oportet crescere. — conforms; re-quotation of no. 1.
3. [0369C] VERS. 31.-- Me autem minui. — conforms to Jo 3:30b. Migne's address (31) is one ahead of the Clementine's versification; a verse-number, not a text divergence — it rides as printed and takes no marker (brief §6.2 family).
4. [0369C] Qui desursum venit, super omnes est, — conforms, Jo 3:31.
5. [0370A] usque ad — n/a, abridgment formula.
6. [0370A] De terra. — conforms, Jo 3:31.
7. [0370A] Et quod vidit. — conforms, Jo 3:32.
8. [0370A] Signavit quia Deus. — conforms, Jo 3:33, cut before verax est.
9. [0370A] Verax est. Quem enim misit Deus. — conforms, Jo 3:33b + 3:34a. The full stop is Migne's splice of two verses, not a divergence.
10. [0370B] Verba Dei loquitur. — conforms, Jo 3:34.
11. [0370B] Non enim ad mensuram dat Deus spiritum. Pater diligit Filium et omnia dedit in manu ejus. — conforms verbatim, Jo 3:34b + 3:35. The longest clean span in the range.
12. [0370C] VERS. 36.-- Qui credit in Filium habet vitam aeternam: qui autem, — conforms, Jo 3:36 (Clementine points with ;, Migne with : — punctuation only, carried).
13. [0370C] Sed ira Dei manet super eum. — conforms, Jo 3:36.

Chunk 0006 — cols 0370C–0374B (John 4:1–24)

1. [0370C] VERS. 1.-- Ut ergo cognovit. — conforms, Jo 4:1.
2. [0370C] Et baptizat. — conforms, Jo 4:1.
3. [0370D] Mundas eam lavacro aquae et verbo vitae [n: (Ephes. V)] — DIVERGENCE, MARKED [var:]. See §(d) 0370D.
4. [0370D] VERS. 2.-- Sed discipuli. — conforms, Jo 4:2.
5. [0370D] VERS. 3.-- Reliquit Judaeam. — conforms, Jo 4:3.
6. [0370D] Et abiit. — conforms, Jo 4:3.
7. [0371A] Et abiit in Galilaeam, — divergence: Clementine Jo 4:3 has et abiit iterum in Galilaeam. DECLINED. The span is a cut (it closes on a comma), the dropped word is an adverb of repetition, and nothing in the gloss — which expounds Galilee as the spiritual sense of the letter — turns on it. A dropped short word is also precisely what our transcription produces. No marker; noted for the plate only because the pre-scan's ✓ conceals it.
8. [0371A] In Galilaeam. — conforms.
9. [0371A] VERS. 4.-- Oportebat. — conforms, Jo 4:4.
10. [0371B] VERS. 5.-- Venit ergo. — conforms, Jo 4:5.
11. [0371B] Quae dicitur Sichar. — conforms, Jo 4:5.
12. [0371B] VERS. 6.-- Erat autem, — conforms, Jo 4:6.
13. [0371B] Jesus ergo fatigatus, — conforms, Jo 4:6.
14. [0371C] Sedebat. — conforms, Jo 4:6.
15. [0371C] Supra. — conforms, Jo 4:6 (supra fontem).
16. [0371C] VERS. 7.-- Venit. — conforms, Jo 4:7.
17. [0371D] usque ad — n/a, abridgment formula.
18. [0371D] Da mihi. — conforms, Jo 4:7, cut before bibere.
19. [0371D] Da mihi. — conforms; re-quotation of no. 31.
20. [0371D] VERS. 8.-- Discipuli enim. — conforms, Jo 4:8.
21. [0371D] VERS. 9.-- Dicit ergo. — conforms, Jo 4:9.
22. [0371D] Non enim, — conforms, Jo 4:9 (non enim coutuntur).
23. [0371D] VERS. 10.-- Respondit Jesus. — conforms, Jo 4:10.
24. [0372B] VERS. 11.-- Dixit ei mulier. — divergence: Clementine Jo 4:11 reads Dicit ei mulier, present. DECLINED, and the pre-scan cleared it wrongly (dixit ei mulier occurs elsewhere in the Clementine, so the phrase search passed). One word, one tense, and Migne prints Dicit ei Jesus correctly at no. 45 six verses later in the same chunk — which is the "the correct form prints a few lines later" argument the brief names as a failed attribution test. It proves an error exists, not whose. No marker; for the plate.**
25. [0372B] VERS. 12.-- Nunquid tu. — divergence orthographic only (Nunquid / Clementine Numquid). DECLINED: an attested medieval spelling with no English exponent, Pattern 9. Not a defect.
26. [0372C] Et ipse ex eo. — conforms, Jo 4:12.
27. [0372C] VERS. 13.-- Qui autem. — conforms, Jo 4:13.
28. [0372C] VERS. 14.-- Aqua. — conforms, Jo 4:14.
29. [0372C] VERS. 16.-- Vade, voca virum tuum, — conforms, Jo 4:16.
30. [0372D] vade, voca virum tuum. — conforms; re-quotation, lower-cased as printed.
31. [0372D] VERS. 17.-- Respondit mulier. — conforms, Jo 4:17.
32. [0372D] Dicit ei Jesus. — conforms, Jo 4:17.
33. [0372D] VERS. 18.-- Quinque enim viros. — conforms, Jo 4:18.
34. [0373B] Quem nunc habes, non est tuus vir. — divergence in order: Clementine Jo 4:18 et nunc, quem habes, non est tuus vir. DECLINED: the span opens mid-verse, so Quem is fronted by the cut and not by a variant reading. Every word, and the negation, is present; the sense is identical.
35. [0373B] VERS. 19.-- Quia propheta tu es. — divergence in order: Clementine Jo 4:19 quia propheta es tu. DECLINED: English has no exponent for Latin word order, and no word of the gloss (which turns on the woman's partial understanding) depends on it.
36. [0373B] VERS. 20.-- Panes nostri, — DIVERGENCE. MARKER DECLINED. The most important crux in the range — Clementine Jo 4:20 reads Patres nostri in monte hoc adoraverunt. Rendered as printed, "Our loaves". Full reasoning at §(d) 0373B.
37. [0373B] VERS. 21.-- Mulier. — conforms, Jo 4:21.
38. [0373D] Vos adoratis Patrem quem nescitis, — DIVERGENCE, MARKED [var:]. See §(d) 0373D.
39. [0373D] Quia salus ex Judaeis est, — conforms, Jo 4:22.
40. [0373D] Crede mihi. — conforms, Jo 4:21.
41. [0373D] Quia nisi credideris non intelliges [n: (Isa. VII)] — DIVERGENCE, MARKED [var:]. See §(d) 0373D.
42. [0373D] Quando neque in monte. — conforms, Jo 4:21.
43. [0374A] quia salus ex Judaeis est, — conforms; re-quotation of no. 52.
44. [0374A] VERS. 23.-- Sed venit hora, et nunc est, quando. — conforms, Jo 4:23.
45. [0374A] In spiritu et veritate. — conforms, Jo 4:23.
46. [0374A] VERS. 24.-- Spiritus est Deus. — conforms, Jo 4:24.

Chunk 0007 — cols 0374B–0376D (John 4:25–54)

1. [0374B] Scio quia Messias venit, qui dicitur, — conforms, Jo 4:25.
2. [0374B] Ille nobis annuntiabit omnia. Dicit ei Jesus. — divergence: Clementine Jo 4:25 cum ergo venerit ille, nobis annuntiabit omnia. DECLINED: Migne drops the temporal clause and ille becomes the subject of the surviving words. That is an abridgment of a lemma, of the kind this recension makes on every page; the words retained are verbatim, and the gloss expounds Messias, not the dropped clause.
3. [0374B] Ego sum qui loquor. — conforms, Jo 4:26, cut before tecum.
4. [0374C] Ego sum Messias, — n/a: not a scripture citation. These are the glossator's own words, introduced by his own unde dicit and set in italic because Migne italicises them; no verse reads so. Nothing is owed against the Clementine, and no marker.
5. [0374C] Reliquit ergo hydriam, — conforms, Jo 4:28.
6. [0374C] VERS. 27.-- Et mirabantur quia. — conforms, Jo 4:27.
7. [0374C] qui venit quaerere quod perierat. — divergence: Lc 19:10 reads Venit enim Filius hominis quaerere, et salvum facere quod perierat. DECLINED: the words are woven into the gloss's own comparative clause (sicut qui venit…), not adduced as a proof-text, and the omission is the author's compression of an allusion. Pattern 14 is for scripture quoted as scripture.
8. [0374C] VERS. 28.-- Reliquit ergo. — conforms.
9. [0374C] ego sum, — conforms; re-quotation of Jo 4:26.
10. [0374C] VERS. 29.-- Venite et videte. — conforms, Jo 4:29.
11. [0374C] VERS. 33.-- Nunquid aliquis attulit. — orthographic only (Nunquid/Numquid), Jo 4:33. DECLINED, Pattern 9, as no. 38.
12. [0374C] VERS. 34.-- Meus cibus est. — conforms, Jo 4:34.
13. [0374C] Meus cibus est ut faciam, — conforms, Jo 4:34.
14. [0374C] Voluntatem ejus. — conforms, Jo 4:34.
15. [0374C] VERS. 35.-- Nonne vos dicitis. — conforms, Jo 4:35.
16. [0374C] Levate oculos vestros. — conforms, Jo 4:35.
17. [0374C] VERS. 36.-- Ut et qui seminat. — conforms, Jo 4:36.
18. [0374C] VERS. 38.-- Ego misi vos. — conforms, Jo 4:38.
19. [0374C] Laboraverunt, et vos in labores eorum introistis. — conforms, Jo 4:38, cut before alii.
20. [0374C] VERS. 40.-- Cum venissent. — conforms, Jo 4:40.
21. [0375C] Et mansit ibi duos. — conforms, Jo 4:40, cut before dies.
22. [0375C] VERS. 42.-- Et mulieri dicebant. — conforms, Jo 4:42.
23. [0375C] VERS. 43.-- Ipsi enim audivimus. — conforms to Jo 4:42. Migne's address (43) is again one ahead of the Clementine's versification; verse-number, not text.
24. [0375C] VERS. 44.-- Quia propheta. — conforms, Jo 4:44.
25. [0375C] VERS. 46.-- Et erat quidam regulus cujus filius. — conforms, Jo 4:46.
26. [0375C] VERS. 47.-- Et rogabat. — conforms, Jo 4:47.
27. [0375C] Nisi signa et prodigia, — conforms, Jo 4:48.
28. [0375C] Dic verbo et sanabitur puer [0376B] meus [n: (Matth. VIII)] — conforms, Mt 8:8, cut before sed tantum. (The column anchor falls inside the italic span; it is reproduced in place, inside the English span.)
29. [0376B] Non inveni tantam fidem in Israel. — conforms verbatim, Mt 8:10. ⚑ Worth saying because it is the one place a conformation would have been easy: the parallel Lc 7:9 reads nec in Israel tantam fidem inveni, and Migne's own siglum names Matthew. He prints Matthew's wording. No divergence.
30. [0376B] Nisi signa et prodigia. — conforms, Jo 4:48.
31. [0376B] VERS. 48.-- Nisi signa. — conforms, Jo 4:48.
32. [0376B] Et prodigia. — conforms, Jo 4:48.
33. [0376C] Descende. — conforms, Jo 4:49.
34. [0376C] Vade, filius tuus. — conforms, Jo 4:50, cut before vivit.
35. [0376C] Credidit ipse. — conforms, Jo 4:53.
36. [0376C] Hoc iterum secundum signum fecit Jesus. — conforms, Jo 4:54.
37. [0376C] VERS. 54.-- Hoc iterum secundum signum fecit [0376D] Jesus cum venisset a Judaea in Galilaeam. — conforms verbatim, Jo 4:54. (Anchor inside the span, reproduced in place.)

Chunk 0008 — cols 0376D–0380B (John 5:4–42)

1. [0376D] VERS. 4.-- Angelus autem. — conforms, Jo 5:4.
2. [0376D] VERS. 5.-- Triginta et octo. — conforms, Jo 5:5.
3. [0376D] VERS. 6.-- Vis sanus fieri. — conforms, Jo 5:6 in wording. ⚑ In punctuation it does not: the Clementine closes with ?, Migne with a full stop. Pattern 8 governs, the full stop is carried, and the ? I had drafted was removed — see §(d) 0376D.
4. [0376D] VERS. 8.-- Tolle grabatum tuum. — conforms, Jo 5:8.
5. [0376D] VERS. 9.-- Et ambulabat. — conforms, Jo 5:9.
6. [0376D] VERS. 10.-- Dicebant ergo. — conforms, Jo 5:10.
7. [0376D] Tolle grabatum, — conforms, Jo 5:10.
8. [0376D] Non licet. — conforms, Jo 5:10.
9. [0376D] VERS. 11.-- Respondit eis. — conforms, Jo 5:11.
10. [0377C] VERS. 13.-- Nesciebat quis esset. — conforms, Jo 5:13.
11. [0377C] VERS. 14.-- In templo. — conforms, Jo 5:14.
12. [0377C] Noli peccare. — conforms, Jo 5:14, cut before jam.
13. [0377C] VERS. 16.-- Quia haec faciebat in sabbato. — conforms, Jo 5:16.
14. [0377C] VERS. 17.-- Jesus autem respondit. — conforms, Jo 5:17.
15. [0377C] VERS. 18.-- Respondit itaque Jesus. — conforms, Jo 5:18 (its closing words).
16. [0377C] VERS. 19.-- Amen, amen dico vobis. — conforms, Jo 5:19.
17. [0378B] Non potest filius. — conforms, Jo 5:19, cut before a se facere quidquam.
18. [0378B] Haec et Filius. — conforms, Jo 5:19.
19. [0378B] VERS. 20.-- Omnia demonstrat. — conforms, Jo 5:20.
20. [0378B] Et majora his demonstrabit ei, — conforms, Jo 5:20.
21. [0378C] Sicut enim Pater suscitat mortuos et vivificat. — conforms, Jo 5:21.
22. [0378C] Majora his. — conforms, Jo 5:20.
23. [0378C] VERS. 21.-- Sicut enim Pater suscitat. — conforms, Jo 5:21.
24. [0378C] Sic et Filius. [n: (AUG.)] — conforms, Jo 5:21. (Note marker inside the italic span; reproduced in place.)
25. [0378D] usque ad — n/a, abridgment formula.
26. [0378D] quia Filius hominis est. — conforms, Jo 5:27.
27. [0378D] VERS. 23.-- Ut omnes honorificent. — conforms, Jo 5:23.
28. [0378D] Qui non honorificat. [n: (AUG.)] — conforms, Jo 5:23. (Note marker inside the span.)
29. [0378D] VERS. 24.-- Quia qui, — conforms, Jo 5:24.
30. [0379A] Audit et credit. — n/a: not a verbatim citation but the glossator's own catchword, splicing Jo 5:24's qui verbum meum audit, et credit ei qui misit me into two words. His very next sentence re-quotes the clause in full and correctly (Audit verbum meum et credit, non mihi, sed ei qui me misit), which is what shows the two-word form to be his shorthand rather than a variant. No marker.
31. [0379A] Sed transiit. — conforms, Jo 5:24.
32. [0379A] Amen, amen dico vobis, quia venit hora et nunc est, — conforms, Jo 5:25 (the Clementine's comma before et nunc is punctuation only).
33. [0379A] VERS. 25.-- Quia venit. — conforms, Jo 5:25. ⚑ Note the present tense here, printed correctly, against no. 137 below.
34. [0379B] Vocem Filii. — conforms, Jo 5:25.
35. [0379B] Sicut enim Pater. [n: (AUG.)] — conforms, Jo 5:26. (Note marker inside the span.)
36. [0379B] usque ad — n/a, abridgment formula.
37. [0379B] In semetipso. — conforms, Jo 5:26.
38. [0379B] VERS. 27.-- Et potestatem. — conforms, Jo 5:27.
39. [0379C] Beati mundo corde, quoniam ipsi Deum videbunt [n: (Matth. V)] — conforms verbatim, Mt 5:8. (Note marker inside the span.)
40. [0379C] usque ad — n/a, abridgment formula.
41. [0379C] VERS. 28.-- Quia veniet hora. — divergence: Clementine Jo 5:28 reads quia venit hora, present. DECLINED. One word, one letter, in the tense — the exact profile an OCR pass manufactures, and the class the brief assigns to our file. ⚠ The tempting argument, and I am naming it because it is the shape the brief says has failed: the exposition here does set the future resurrection of the flesh against the present resurrection of the mind at no. 129, and no. 129 prints venit correctly eight lines earlier — so a future at v. 28 fits the argument perfectly. That is internal corroboration, and it proves only that the two forms differ on the page, never who set them. The gloss states the distinction in its own words (ita etiam hic est auctor de resurrectione carnis) and never quotes the tense. No marker; for the plate.
42. [0379C] VERS. 30.-- Non possum ego a meipso facere. — conforms, Jo 5:30, cut before quidquam.
43. [0379D] Sicut autem. — DIVERGENCE THE PRE-SCAN CLEARED, and the second most important crux in the range. Clementine Jo 5:30 reads Sicut audio, judico. DECLINED. See §(d) 0379D — the reason it matters is that the gloss's first word after the lemma is Audio.
44. [0379D] Non quaero. — conforms, Jo 5:30.
45. [0379D] VERS. 31.-- Si ego testimonium. — conforms, Jo 5:31. ⚑ Migne sets a doubled full stop after the closing italic (Si ego testimonium..); carried into the English exactly, Pattern 8. Not marked: I cannot tell his type from our transcription, and the class is the same as 9001's multae. fidelium catervae.
46. [0379D] VERS. 34.-- Sed haec. — conforms, Jo 5:34.
47. [0379D] VERS. 35.-- Ille erat, — conforms, Jo 5:35.
48. [0380B] quod illuminat omnem hominem venientem in hunc mundum. — divergence: Clementine Jo 1:9 reads lux vera, quae illuminat. DECLINED: the relative has been conformed to the gloss's own host noun — Christus verum lumen, quod illuminat — which is neuter and takes quod. That is scripture woven into the author's sentence, not a variant text, and every other word of the verse is verbatim. English has no exponent for the gender either way.
49. [0380B] VERS. 37.-- Neque vocem. — conforms, Jo 5:37.
50. [0380B] VERS. 41.-- Claritatem. — conforms, Jo 5:41.
51. [0380B] VERS. 42.-- Quia dilectionem. — conforms, Jo 5:42.

Chunk 0009 — col 0380B–0380C (John 5:43)

1. [0380B] VERS. 43.-- Si alius. — conforms, Jo 5:43.

Tally

148 spans walked. 3 [var:] fired. 12 divergences declined — nos. 20, 37, 38, 47, 48, 49, 61, 66, 70, 137, 139 and 144, of which 38 and 70 are a single orthographic class (Nunquid) and 47, 48 and 144 have no English exponent at all. 7 spans are not scripture at all: nos. 63 and 126, the glossator's own words in italic type, and the five usque ad abridgment formulae, nos. 5, 30, 121, 132 and 136. The remaining 129 conform.

(d) THIS RANGE'S OWN CRUCES, BY COLUMN

0370B — In hoc se distinguit a Christo (mechanism 5, ambiguity preserved)

The reflexive has two candidate subjects on the page: the Baptist, who is the speaker of the whole passage, and the Evangelist, named two sentences earlier (Hoc Evangelista dicit). The next clause, sed quia missus et verba Dei loquitur, apertius addit, then makes the sent one Christ while addit stays with the speaker. The English keeps all three pronouns bare — "In this he distinguishes himself from Christ; but because he is sent and speaks the words of God, he adds more openly" — because resolving them would be an assertion the Latin does not make, and the easier resolution is the one the brief says will be wrong.

0370D — [var:] FIRED · Eph 5:26

Migne: Mundas eam lavacro aquae et verbo vitae. Clementine Eph 5:26: ut illam sanctificaret, mundans lavacro aquae in verbo vitae. Three differences at once — participle to finite verb, illam to eam, and in to et.

Fired because the gloss's argument stands on the third of them. Its own preceding roman words are ejus baptismus fit etiam in corpore, qui est in duobus, scilicet in aqua et verbo vitae — the whole point is that baptism consists of two things, which the Vulgate's in verbo does not say and Migne's et verbo does. A three-site divergence carrying a coordinating conjunction the exposition depends on is not a shape OCR produces; it is a clause, and the brief's class split puts clauses on Migne's side of the line.

⚠ I am not claiming the divergence originates with Migne's copy-text rather than his compositor, nor that the author quoted from memory. Pattern 14 forbids that speculation and the marker states only the two readings.

0371A — VERS. 4, a sentence with no verb in its relative clause

inter legem litterae, quae per Judaeam et divinas leges circa Deum revolutas, quarum figura est Galilaea, est lex naturalis in medio constituta. The relative quae has no predicate: the sentence needs something like significatur. I did not emend and did not mark; the English supplies the missing verb in our own square brackets, "which [is signified] by Judaea", which is the corpus's existing convention for a word we add ([it], [them]). The printed words are otherwise untouched. Whether the verb fell out at Migne's press or at Corpus Corporum's transcription I cannot say.

0372A — legis valle muniti (real word, reads oddly)

Judaeos, qui legis valle muniti, et prophetiae sunt telis armati. muniti + telis armati wants a fortification, and vallo (rampart) would give one; valle is the ablative of vallis, a valley, which fortifies nothing. It is a real, well-formed Latin word, so it is rendered as printed — "who are fenced by the valley of the law" — and no marker is fired. I do not know whether Migne set valle or whether an o became an e in transcription, and the brief is explicit that my belief about that is worth nothing. For the plate.

0372B — the gloss's melior against John's major

The lemma (no. 38) is only Nunquid tu, so nothing italic diverges. But the gloss expands it as Nunquid tu melior es, id est potes melior esse Jacob, where Jo 4:12 reads Numquid tu major es patre nostro Jacob. This is the glossator restating the verse in his own roman prose, not quoting it, so Pattern 14 does not reach it and no marker is owed. Rendered "better" twice, as printed. Recorded because a later reader meeting "better" against the Vulgate's "greater" would otherwise suspect us.

0372D–0373A — quodum, a non-word, CARRIED WITH NO MARKER

sed utebatur quodum non legitimo. quodum is not a Latin form; quodam is plainly what the sentence wants. Under Pattern 7 there is nothing to render, so the run is carried into the English untranslated, in italics — "but was making use of a quodum not lawful" — and under the brief's §2 no [sic:] is fired, because a non-word is not automatically ours (9001 @0203A magnitudino is Migne's) and I cannot settle it.

7a⁗ check, performed before closing it: with the carried run deleted the clause still has its verb (utebatur → "was making use of"), still carries its negation (non legitimo → "not lawful"), and loses no conjunction or interrogative. Nothing is hiding inside the quarantine.

0371D — privaverunt (mechanism 5, ambiguity preserved)

Samaritanos Judaei exsecrantur et supplantatores vocant, quia eos haereditate patris sui Jacob privaverunt abstinendo a cibis et vasis eorum. The subject of privaverunt can be the Samaritans (who supplanted the Jews' inheritance, which is why they are called supplanters) or the Jews (whose abstention from Samaritan food and vessels is the participial clause). Both eos and eorum flip with it. Rendered with bare pronouns throughout, so the English is as undecided as the Latin.

0373B — ⭐ Panes nostri FOR Patres nostri — RENDERED AS PRINTED, NO MARKER

VERS. 20's lemma reads Panes nostri, — "Our loaves". Clementine Jo 4:20: Patres nostri in monte hoc adoraverunt. The gloss that follows is about the fathers and nothing else: pro monte, in quo et non in templo patres adoraverunt, qui Deo placuerunt. Et sic patriarchae in montibus, Judaei in templo adoraverunt. There is no doubt whatever that Patres is the word the passage is expounding.

And that is exactly why no marker is fired. Panes is a real, well-formed Latin word; the divergence is one word inside a two-word lemma, which is the class the brief's measurement assigns to our own file; and the corroboration available to me — that the correct sense prints four words later in the gloss — is verbatim the reasoning the brief lists among 9004's nine wrong calls. A [var:] here would be a public assertion that Migne's page reads Panes, and I have not seen his page.

• English: "Our loaves, etc." — Pattern 7, rendered as printed.
• 7a⁗ check: nothing is quarantined, so nothing is swallowed; the gloss states the fathers' worship in its own words and the argument survives intact beside the odd lemma.
• For the plate: if Migne prints Patres, this is a data/tei-patches/8999.json entry and the English becomes "Our fathers". If he prints Panes, it is owed a [var:] against Jo 4:20. Both outcomes are live and I decline to guess between them.

0373D — [var:] FIRED ×2 · Jo 4:22 and Is 7:9

(i) Vos adoratis Patrem quem nescitis against Clementine Jo 4:22 Vos adoratis quod nescitis. Migne adds Patrem and makes the relative masculine to agree with it. Fired because the gloss is built on the added word: quia nullus adorat Patrem nisi Filium adoret, per quem ad notitiam Patris venitur. ⚑ And the control is on the same column in Migne's own roman type: eight lines later the gloss paraphrases the verse without the addition — vos Samaritani adoratis quod nescitis, sed nos Judaei … adoramus quod scimus — so the two readings stand beside each other on one page. Multi-word, semantically coherent, load-bearing: a clause, not a letter-swap.

(ii) nisi credideris non intelliges [n: (Isa. VII)] against Clementine Is 7:9 si non credideritis, non permanebitis. A whole-clause divergence with a large change of sense (understand against continue) and a change of number, on a verse Migne himself cites by siglum. Fired. Pattern 14 forbids me to say why the two differ and the marker says only what each reads; the crux does not argue it either.

Also on this column, both carried and neither marked:

• id est. quia Christus, qui est salus — a full stop set mid-sentence, after id est. Carried into the English as printed ("that is. because Christ…"), Pattern 8. Whose stop it is, I cannot say; 9001 @0202D is the same class and turned out to be the digitization's.
• The paragraph ends et hoc est. with nothing following. Carried, unfinished, per Pattern 8's "a sentence that simply stops stops in the English too". No [ed:]: nothing demonstrates a loss, and the same formula (et hoc est) is used mid-paragraph three lines above.

0375C — VERS. 42, Judaeis spiritualiter edoctis (mechanism 5, ambiguity preserved)

Postquam Dominus carnis praesentia Judaeis spiritualiter edoctis gratiam suae cognitionis infudit mulieri, id est legi, dicebant… The dative-or-ablative Judaeis … edoctis can be an ablative absolute ("the Jews having been taught") or a second dative alongside mulieri ("poured into the Jews… and into the woman"). Rendered as an English absolute, which is the construction that leaves the reader where the Latin leaves him, rather than choosing the second and quietly giving the Jews the grace the sentence gives the woman.

0376B — the prodigium etymology keeps its Latin

Prodigium quasi porro dictum … Prodigium quasi procul a digito. The gloss is the etymology, so under the Song-of-Songs ruling's case 2 an English word sharing no root would make it explain nothing. The Latin terms are kept in italic inside the gloss with the sense given beside them: "Prodigium as it were porro dictum, spoken far off … as it were procul a digito, far from the finger". These four italic runs are ours, not Migne's lemma italics, which is why my English chunk carries 41 italic spans against the Latin's 37; the same accounts for the counts in 0006 (50 vs 46) and 0008 (51 vs 51 — there the [var:] italics replace none).

0376D — ⛔ A ? I SUPPLIED AND REMOVED

The lemma Vis sanus fieri. closes with a full stop in Migne; the Clementine closes Jo 5:6 with ?. I first wrote "Wilt thou be made whole?" — a mark the plate does not print, supplied because the English sounds wrong without it. It is out. The English reads "Wilt thou be made whole." Pattern 8 forbids supplying a mark, and 8b's colon licence does not apply: this is a direct question, and English can take Migne's full stop perfectly well. verify-english caught it as a quotation-mark parity warning (2 EN vs 1 LA), which is the check earning its place. Recorded here per 7a′, because the ruling has to be in the text and not only in the crux — it is: the shipped English carries the full stop.

0377A — aetate (real word, reads oddly)

et non solum aetate de tua erectione, sed etiam de proximorum sollicitus esto salute. The ablative of aetas does not construe with the non solum … sed etiam pair, which wants two parallel de phrases. Rendered as printed — "be solicitous not only in age concerning thine own raising up, but also concerning the salvation of thy neighbours" — with both members of the correlative intact. No [cj:]: Pattern 18's bar is that the English assert something false, and this English asserts nothing false, only something obscure. No [sic:]: it is a real word and I cannot say whose. For the plate.

0377B — quae … cessare jubet (mechanism 5, ambiguity preserved)

nescientes gratiam qua lex impletur, quae ab omni servili opere … cessare jubet. quae is feminine and both gratiam and lex are feminine, so it may bind either. Rendered "which bids us cease", binding nothing.

0377C — a direct question closed with a full stop, carried

Quasi dicat caeci in lege cur mihi invidetis, si hominem in sabbato sano, qui Deus cum Patre semper operor. A direct cur question that Migne closes with a period. Carried as printed — the English ends on a full stop — and the ? I would have supplied was not supplied. This is Pattern 8's "preserve, never supply" in the direction that looks like our error, and 8b's licence does not reach it (8b is for indirect questions English syntax cannot punctuate).

0379A — ut cognoscant te quem misi Jesum Christum esse unum verum Deum

Not italic, so not a lemma and not a [var:] candidate. But as printed, the accusative-infinitive makes quem misi Jesum Christum an apposition to te and the sentence therefore says that thou, Jesus Christ whom I sent, art one true God — where Jo 17:3, which the gloss is paraphrasing (ut post dicet), distinguishes the Father from the Son he sent. Rendered as printed ("that they may know thee, Jesus Christ whom I sent, to be one true God") and logged. The compression may be the glossator's own; it may be a lost et. Not marked.

Also on this column: the direct question Cur ait: Audit verbum meum et credit, non mihi, sed ei qui me misit: closes with a colon, not a question mark. Carried as printed; the answer (quia verbum ejus est in me) follows the colon, so no mark is supplied and none deleted.

0379D — ⭐ Sicut autem FOR Sicut audio — DECLINED, AND THE PRE-SCAN CLEARED IT

The lemma at VERS. 30 reads Sicut autem. Clementine Jo 5:30: Sicut audio, judico. The gloss's very first word after the lemma is Audio — Audio per unitatem substantiae, id est, scio quia non est aliud audire, quam scire vel esse — and the whole comment is an exposition of audire. The lemma as printed quotes a word the gloss never touches and omits the one it expounds.

No marker. One word; the corroboration is internal and four characters away; and the pre-scan marked this span ✓ Clementine verbatim because sicut autem occurs somewhere in the Clementine — which is precisely the false clearance the brief's header describes, and the reason this was found by reading the line rather than by reading the mark. English: "But as. I hear through the unity of substance…", rendering autem as printed and leaving the gloss's Audio to say "I hear" on its own. For the plate.

0380B — Jo 1:9 with quod for quae

Covered at no. 144. Declined: the relative follows the gloss's own lumen, which is how this recension habitually weaves a verse into its sentence.

INHERITED CONVENTIONS — applied from the first line, and where they were TESTED

Per the brief's §1, each is reported with the evidence in my own range rather than merely asserted. No exception was forced, and nothing was silently departed from.

1. Gloss voice not archaised. dicit/ait/dicat → says, never saith, throughout. The italic-membership test was run mechanically over all five chunks: every one of the 3 saith tokens sits inside an italic lemma rendering Dicit, and of 28 other archaic verb forms, 20 are inside italic lemmata. The remaining 8 were each read individually and all 8 stand: five are second-person singulars compelled by Pattern 17 (savourest, receivest, awaitest, goest, and wast), which is a NUMBER marker and not archaism of the gloss voice; three are the gloss re-quoting its own lemma's scripture words unitalicised — 0370C he does not say, cometh, but abideth (Migne: non ait, venit, sed manet), and 0379A He heareth my word and believeth (Migne: Audit verbum meum et credit). Those are quoted scripture, which is where 9001's ruling leaves them. Zero third-person -eth on a metalanguage verb.
2. turba → crowd · multitudo → multitude. ⚑ Both words are live in my range, so the axis is tested, not assumed: turba three times on one column at 0377C (In turba positus · fugiat turbam suorum affectuum · Non in turba, sed in templo) → crowd every time; multitudo at 0376C (multitudo gentium solo verbo convertendarum) → multitude. No English word does duty for both.
3. daemonium → demon · diabolus/satanas → devil. Neither root occurs anywhere in chunks 0005–0009 (grepped). The axis is untested here and I have introduced no word into it. Antichristus at 0380C is rendered Antichrist and belongs to no part of it.
4. natio → nation · gentes → Gentiles / nations. ⚑ No standalone natio occurs in my range — the four apparent hits are incarnatio, incarnatione, sanationis, seminationis — so, as on 9004, nothing is flattened by the choice. gentes in the Judaei/gentes typological opposition → Gentiles: 0371A ad gentes transire, 0371B gentilis populus, 0371C–D and 0374C Ecclesia de gentibus, 0375D a gentibus facile conversis … revertetur ad Judaeos (the opposition is explicit in that clause), 0376B fides gentium, 0376A Ecclesia gentium, 0376C multitudo gentium. gentes generic → nations: 0373A de divisionibus gentium, the fourth of the five laws, which is the post-diluvian division of Genesis 10 and carries no Jew/Gentile contrast at all. Singular gens of the Jewish people → nation: 0376A aliquem gentis Judaicae doctorem → "some teacher of the Jewish nation".
5. mysterium → mystery · sacramentum → sacrament, 1:1. ⚑ Both are live here too: mysterium at 0373A (adhuc mysteria loquitur), 0374C (mysterium Ecclesiae de gentibus futurae) and 0375A (mysterium redemptionis) → mystery/mysteries; sacramentum at 0377D (constat sacramentum esse quod de requie sabbati legitur) → sacrament. One English word each.

Two further consistencies fixed within my range, so a merge can check them: fons → fount and puteus → well throughout, which the gloss at 0371C requires by expounding the difference between them; and grabatum → bed at all three occurrences (0376D ×2, 0377B), matching the Vulgate's own wording, which Migne prints unchanged.

FOR THE PLATE READ — the five sites I could not settle, ranked

Every one is a place where I know an error exists and refused to say whose. Leaf n = (column + 11) / 2 on patrologiaecursu0114mign.

Secondary, same rule, lower stakes: 0372D quodum (non-word, carried in italics); 0379D–0380A the doubled full stop after Si ego testimonium..; 0373D the full stop inside id est.; 0371A the verbless quae per Judaeam.

Not one of these carries a marker, and that is the deliverable. If the plate agrees with me at every site, nothing was lost by waiting; if it agrees with me at five of nine, which is the brief's measured rate, four false public claims against Migne were not made.

<!-- ===== cruces-0010.md ===== -->

Stint 3 of six. PL 114, cols 0380C–0394C (John VI. 1 – VIII. 59), 6,020 words of Latin in five chunks. Written to the governing document, data/briefs/8999-launch-brief.md, and to translation-style.md; the four Mark conventions of src/english/9001/cruces.md §1 were applied from the first line and none of them was forced open by this book (§d.6 below records the controls).

(a) THE POLICY I APPLIED — stated first, so it can be argued with

1. Attribution is not mine to assert, and I asserted none. Per the brief's §2: every reading in this range that is a real, well-formed Latin word which merely reads oddly is rendered as printed and logged with no marker, however good the internal argument for what was meant. I found no non-word and no split or fused run anywhere in the five chunks — the one word-division candidate, a me ipso at 0387B, is a normal orthographic variant of meipso and is not a defect. I fired zero [sic:]. Nine of the fourteen attribution calls made on 9004 were wrong at the plate; I am not adding a fifteenth guess to that record.

2. [var:] fires only on a coherent CLAUSE-level divergence, never on a single short token. The measured asymmetry (46/46 and 19/19 [var:] stood; 7 of 8 single-word [sic:] fell) is mechanical: a transcription slip manufactures broken or missing tokens and cannot invent a semantically coherent divergent clause. So a substituted or recast clause I may cite against sources/vulgate/clementine-flat.txt; a dropped or altered single token — vivus at 0384B, vestrum at 0389C, de at 0391A, et at 0387B, cognovimus/cognoscimus at 0393D — I log and leave for the plate, because a [var:] asserts that Migne's text diverges and our Latin is not Migne's page. Three [var:] fired, all on clauses, all cited book chapter:verse.

3. Below Pattern 14's bar is not a divergence. Word order, a connective swapped for another of the same force (enim/autem), nunquid for numquid, coelo for caelo, a lemma abridged before its last word: these are logged in §(c) with the reason and take no marker. Marking them would bury the three that matter.

4. [cj:] only where the faithful English ASSERTS SOMETHING FALSE — not where it is merely opaque. Pattern 18a closed the conjecture census and left ~90 candidates declined; the stated bar is that the English asserts what the author did not, and I have held to it strictly. Two [cj:] fired (0381A vanitatem, 0383A me esse Patrem), each of which otherwise puts a falsehood on the page in the author's voice. Two obvious conjecture candidates were DECLINED because the English they produce is odd, not false — 0387D assensionem and 0385D audit — and both are logged for the plate in §(d). ⚑ [cj:] is non-attributive by design (Pattern 18: Migne's word stands first and keeps its English; ours is added beside it), so neither marker claims anything about whose the reading is. If the plate shows our file corrupt at either site, they become data/tei-patches/8999.json entries — not withdrawn accusations against Migne.

5. Negation and mechanism 5. Every non, nec, neque, nisi, ne, nemo, nihil, nullus, nondum the Latin prints is in the English and none is in the English that the Latin does not print, including where carrying it leaves the sentence at odds with itself. Where a relative or participle could bind to either of two nouns I read the case and gender, not the sense; the three sites where the easier reading would have been the wrong one are named in §(d).

6. No plate was read for this range and none is claimed. Everything I could not settle is named as unsettled, with the column, so the read that follows this translation has a list rather than a hunt.

(b) SPAN COUNT — RUN, AND IT AGREES

data/briefs/8999-lemmata-0010.txt states 215 spans for chunks 0010–0014. Counting the … spans in my own Latin:

215 = 215, and the brief's first span ([0380C] Trans mare.) and last ([0394C] Et exivit.) are the first and last italic spans in my own files — so the range is aligned at both ends, not merely equal in count. Every asterisk in all five chunks is inside a body span; the frontmatter carries none. This is the check that caught the −0/+13/−11/+15/−17 splitter error on 9001; it is clean here.

Marker-parity counts, checked before each file was written: column anchors 14 / 6 / 15 / 14 / 7, [n: …] notes 2 / 1 / 8 / 4 / 7 — each equal to its Latin twin and to the frontmatter noteCount. node scripts/verify-english.mjs 8999 passes on all five.

(c) THE LEMMA DIVERGENCE LIST — all 215 spans, in order

Every italic span in the range, with its verdict; every divergence I declined to mark carries its reason. Verdicts were formed by reading the printed Latin word by word against clementine-flat.txt at its own verse — the pre-scan's ✓ is not clearance and was not treated as any (the seven usque ad spans it marks ✓ are the clearest demonstration: they are Migne's abridgment formula and not scripture at all).

Summary: 3 [var:] fired · 21 divergences examined and declined · 7 structural spans · 184 conform.

[0380C] VERS. 1.-- Trans mare. — Conforms, read against its own verse. [0380C] VERS. 2.-- Et sequebatur. — Conforms, read against its own verse. [0380C] VERS. 5.-- Cum sublevasset. — Conforms, read against its own verse. [0380C] VERS. 7.-- Ducentorum. — Conforms, read against its own verse. [0380C] VERS. 9.-- Est puer unus. — Conforms, read against its own verse. [0381A] Sed haec, — Conforms, read against its own verse. [0381A] VERS. 10.-- Facite homines. — Conforms, read against its own verse. [0381A] VERS. 14.-- Illi ergo homines cum vidissent quod, — Conforms, read against its own verse. [0381A] VERS. 15.-- Jesus ergo. — Conforms, read against its own verse. [0381A] Percipite regnum quod vobis paratum est. — DIVERGENCE — [var:] FIRED. Matt 25:34 reads possidete paratum vobis regnum a constitutione mundi. Migne substitutes the verb (percipite) and recasts the clause; a coherent clause-level variant, the class that stood 19/19 on 9001 and 46/46 on 9004. [0381A] Ut autem sero factum est. — Conforms, read against its own verse. [0381A] VERS. 19.-- Timuerunt. — Conforms, read against its own verse. [0381A] Nolite timere, ego sum. — Divergence, DECLINED. Jo 6:20 reads Ego sum, nolite timere (so Matt 14:27, Mc 6:50). Migne reverses the two clauses and changes nothing else: word order, no sense at stake, and the gloss turns on neither clause's position. [0381A] VERS. 22.-- Altera die. — Conforms, read against its own verse. [0381D] VERS. 25.-- Et cum invenissent. — Conforms, read against its own verse. [0381D] Amen, amen dico vobis, — Conforms, read against its own verse. [0381D] VERS. 26.-- Amen, amen dico vobis. — Conforms, read against its own verse. [0381D] VERS. 27.-- Signavit Deus. — Conforms, read against its own verse. [0381D] VERS. 29.-- Hoc est opus Dei. — Conforms, read against its own verse. [0382B] VERS. 32.-- Dixit ergo. — Conforms, read against its own verse. [0382B] VERS. 33.-- Panis enim verus. — DIVERGENCE — [var:] FIRED. Jo 6:33 reads Panis enim Dei est, qui de caelo descendit. Migne's lemma puts the content word verus where the received text has Dei est (possibly drawn from panem de caelo verum, Jo 6:32, the lemma directly above). A substitution, not a dropped token. [0382B] VERS. 34.-- Domine, semper, — Conforms, read against its own verse. [0382B] Qui biberit de hac aqua, non sitiet in aeternum, — DIVERGENCE — [var:] FIRED. Jo 4:13-14: Omnis qui bibit ex aqua hac, sitiet iterum : qui autem biberit ex aqua quam ego dabo ei, non sitiet in aeternum. Migne's single clause collapses the two halves, so that drinking of THIS water is what quenches for ever, which is the reverse of the verse's own contrast; and the gloss builds on the quotation (secundum corpus accepit). The received reading belongs beside it. [0382B] Da mihi [0382C] hanc aquam, — Conforms, read against its own verse. [0382C] Ego sum panis vitae. — Conforms, read against its own verse. [0382C] VERS. 36.-- Sed dixi vobis. — Conforms, read against its own verse. [0382C] VERS. 37.-- Omne quod dat mihi Pater, — Conforms, read against its own verse. [0382C] Veniet. — Conforms, read against its own verse. [0382C] VERS. 39.-- Haec est enim voluntas ejus qui misit me Patris, — Divergence, DECLINED. Jo 6:39 reads Haec est AUTEM voluntas ejus qui misit me, Patris. One connective for another of the same force, below Pattern 14's bar ("in a way that changes the sense"). [0382C] Non perdam, — Conforms, read against its own verse. [0382C] VERS. 40.-- Ut omnis, — Conforms, read against its own verse. [0382C] Qui audit verbum meum et credit ei qui me misit; — Divergence, DECLINED. Jo 5:24 reads Qui verbum meum audit, et credit ei qui misit me. Two transpositions; no word added or lost, no sense changed. [0383A] Qui videt Filium et credit in eum. — Conforms, read against its own verse. [0383A] VERS. 41.-- Murmurabant. — Conforms, read against its own verse. [0383A] VERS. 43.-- Nolite murmurare. — Conforms, read against its own verse. [0383A] VERS. 44.-- Nisi Pater. [n: (AUG.)] — Conforms, read against its own verse. [0383B] usque ad — Structural, not scripture: Migne's abridgment formula, rendered as far as. The pre-scan's tick is a phrase-search artefact, the two words being common in the Vulgate. [0383B] Et ego resuscitabo, — Conforms, read against its own verse. [0383B] VERS. 45.-- Est scriptum. — Conforms, read against its own verse. [0383B] erunt docibiles Dei, — Not a divergence. Jo 6:45 reads Et erunt OMNES docibiles Dei; the gloss has taken omnes into its own sentence (Omnes homines regni illius ... erunt docibiles Dei), so the italics open after it. An artefact of where Migne set the span. [0383B] In prophetis. — Conforms, read against its own verse. [0383C] VERS. 46.-- Non quia Patrem, — Conforms, read against its own verse. [0383C] Ego sum panis vitae. Patres vestri, — Conforms, read against its own verse. [0383D] Ego sum panis vitae. — Conforms, read against its own verse. [0383D] Verba vitae aeternae habes. — Conforms, read against its own verse. [0383D] VERS. 52.-- Et panis, — Conforms, read against its own verse. [0383D] VERS. 54.-- Nisi manducaveritis carnem, — Conforms, read against its own verse. [0383D] usque ad — Structural, not scripture: Migne's abridgment formula, rendered as far as. The pre-scan's tick is a phrase-search artefact, the two words being common in the Vulgate. [0384A] VERS. 55.-- Qui manducat. — Conforms, read against its own verse. [0384B] VERS. 56.-- Caro enim. — Conforms, read against its own verse. [0384B] caro mea vere est cibus. — Not a divergence. Jo 6:56 reads Caro ENIM mea vere est cibus; the span is quoted inside the gloss's own quia clause, which is why the connective stands outside it. [0384B] VERS. 57.-- Qui manducat. — Conforms, read against its own verse. [0384B] VERS. 58.-- Sicut misit me. — Conforms, read against its own verse. [0384B] Et ipse vivet. — Conforms, read against its own verse. [0384B] VERS. 59.-- Hic est panis. — Conforms, read against its own verse. [0384B] Ego sum panis qui de coelo descendi. — Divergence, DECLINED. Jo 6:41/51 read Ego sum panis VIVUS, qui de caelo descendi. One token absent from a self-quotation the gloss introduces with Supra dixerat. A single short token is the class the brief's Sect. 2 measures as at least as likely ours as Migne's, and no [var:] can assert whose it is. Logged for the plate. (coelo for caelo is orthography throughout and is never marked.) [0384B] panis qui de coelo descendit: — Conforms, read against its own verse. [0384B] VERS. 61.-- Multi ergo. — Conforms, read against its own verse. [0384D] Durus est, — Conforms, read against its own verse. [0384D] VERS. 63.-- Si ergo, — Conforms, read against its own verse. [0384D] VERS. 64.-- Spiritus est qui vivificat. — Conforms, read against its own verse. [0384D] Verba quae ego, — Conforms, read against its own verse. [0384D] VERS. 65.-- Sed sunt quidam, — Conforms, read against its own verse. [0384D] Et quis traditurus. — Conforms, read against its own verse. [0384D] VERS. 66.-- Quia nemo potest, — Conforms, read against its own verse. [0384D] VERS. 67.-- Ex hoc, — Conforms, read against its own verse. [0384D] VERS. 71.-- Nonne ego vos, — Conforms, read against its own verse. [0385C] VERS. 1.-- Post haec autem. — Conforms, read against its own verse. [0385C] VERS. 2.-- Dies festus. — Conforms, read against its own verse. [0385C] VERS. 3.-- Ut et discipuli. — Conforms, read against its own verse. [0385C] VERS. 4.-- Nemo. — Conforms, read against its own verse. [0385C] VERS. 6.-- Tempus. — Conforms, read against its own verse. [0386A] VERS. 8.-- Ego enim non ascendo. [n: (AUG.)] — Divergence, DECLINED. Jo 7:8 reads ego AUTEM non ascendo. Connective only, same ground as Jo 6:39 above. The negation is intact and is carried in the English. [0386A] VERS. 10.-- Non manifeste, — Conforms, read against its own verse. [0386A] VERS. 16.-- Respondit eis: Mea doctrina, — Not a divergence. Jo 7:16 reads Respondit eis Jesus, et dixit : Mea doctrina; the lemma abridges the speech frame. Note that Migne interleaves this span with roman gloss (Mea doctrina, hominis, non est mea, sed Dei), and the interleaving is reproduced 1:1 in the English. [0386A] non est mea, — Conforms, read against its own verse. [0386A] VERS. 18.-- Qui a semetipso. — Conforms, read against its own verse. [0386A] Qui autem, — Conforms, read against its own verse. [0386A] VERS. 19.-- Nonne Moyses, — Conforms, read against its own verse. [0386C] VERS. 20.-- Daemonium. — Conforms, read against its own verse. [0386C] VERS. 21.-- Unum opus feci. — Conforms, read against its own verse. [0386C] VERS. 22.-- Circumcisionem. [n: (ALC.)] — Conforms, read against its own verse. [0386C] usque ad — Structural, not scripture: Migne's abridgment formula, rendered as far as. The pre-scan's tick is a phrase-search artefact, the two words being common in the Vulgate. [0386D] VERS. 23.-- Si circumcisionem, — Conforms, read against its own verse. [0386D] VERS. 24.-- Nolite judicare, — Conforms, read against its own verse. [0386D] VERS. 25.-- Dicebant ergo. — Conforms, read against its own verse. [0387B] Generationem ejus quis enarrabit [n: (Isa. LIII)] — Conforms, Is 53:8. The sentence carries NO TERMINAL STOP after the note in the Latin; none is supplied in the English (Pattern 8). [0387B] VERS. 28.-- Et me scitis. — Conforms, read against its own verse. [0387B] et unde sim — Conforms, read against its own verse. [0387B] a me ipso non veni. — Not a divergence. Jo 7:28 reads a meipso non veni: word division inside meipso, an orthographic variant attested both ways, not a Pattern 10 broken run. Negation intact. [0387B] Ego scio. — Conforms, read against its own verse. [0387B] nemo novit Patrem nisi Filius, et cui Filius voluerit revelare [n: (Marc. XI)] — Divergence, DECLINED. The text is Matt 11:27 (neque Patrem quis novit, nisi Filius, et cui voluerit Filius revelare) / Lc 10:22; Migne's wording is a synonymous recasting with nothing of substance added or lost. His CITATION, however, points to Mark XI, where no such text stands. Logged in (d) as a citation crux and not marked: [n:] contents pass through verbatim by rule, and the index's refDisplay/refKey split is where that is answered. [0387B] VERS. 29.-- Sed scio eum. — Conforms, read against its own verse. [0387B] quia ab ipso sum ipse, — Divergence, DECLINED. Jo 7:29 reads quia ab ipso sum, ET ipse me misit; Migne's span runs the two clauses together across the missing et and breaks off with etc. A single short token inside an abridgment span: our file or his, and nothing in our hands decides it. Logged for the plate. [0387C] ab ipso sum, — Conforms, read against its own verse. [0387C] VERS. 30.-- Hora ejus. — Conforms, read against its own verse. [0387C] VERS. 32.-- Audierunt Pharisaei. — Conforms, read against its own verse. [0387D] VERS. 33.-- Adhuc modicum tempus. — Conforms, read against its own verse. [0387D] Quid faciemus? Vobiscum sum et vado. — Not a Vulgate divergence. The italic run MERGES TWO TEXTS: Quid faciemus? is Act 2:37 (there Quid faciemus, viri fratres?), and Vobiscum sum et vado is the return to the Jo 7:33 lemma, which the gloss expounds in the next breath (Non ait ero, sed sum). Migne's italics swallow the join; rendered as printed, in one run. [0387D] Filius [0388A] hominis qui est in coelo, — Conforms, read against its own verse. [0387D] VERS. 34.-- Non potestis venire. — Conforms, read against its own verse. [0387D] Non potes me sequi modo, sequeris autem postea [n: (Joan. XIII)] — Divergence, DECLINED. Jo 13:36 reads non potes me modo sequi : sequeris autem postea. Transposition of modo only. [0387D] VERS. 35.-- Ad semetipsos, — Conforms, read against its own verse. [0387D] VERS. 36.-- Quis est hic? — Conforms, read against its own verse. [0387D] VERS. 37.-- In novissimo autem. — Conforms, read against its own verse. [0387D] Si quis sitit. — Conforms, read against its own verse. [0387D] et cum biberit, flumina de ventre ejus fluent aquae vivae. — Divergence, DECLINED. Jo 7:38 reads Qui credit in me, sicut dicit Scriptura, flumina de ventre ejus fluent aquae vivae. The opening words are the glossator's own resumption of his sentence (Bibat spiritum ... et cum biberit), swallowed by the italics; the quoted portion agrees with the Vulgate word for word, so there is no claim about Migne's scripture text to be made here. [0387D] usque ad — Structural, not scripture: Migne's abridgment formula, rendered as far as. The pre-scan's tick is a phrase-search artefact, the two words being common in the Vulgate. [0387D] Nondum enim, — Conforms, read against its own verse. [0388D] Quia Jesus nondum erat glorificatus, — Conforms, read against its own verse. [0388D] Ex illa ergo turba, — Conforms, read against its own verse. [0388D] Oportet obedire magis Deo quam hominibus [n: (Act. V)] — Divergence, DECLINED. Act 5:29 reads Obedire oportet Deo magis quam hominibus. Transposition only. [0388D] VERS. 45.-- Venerunt ergo. — Conforms, read against its own verse. [0388D] VERS. 47.-- Nunquid et vos, — Not a divergence. nunquid for numquid is Migne's settled orthography (four spans in this range) and is Pattern 9 matter, never marked. Jo 7:47 otherwise. [0388D] VERS. 48.-- Nunquid ex principibus aliquis, — Not a divergence. Orthography as above; conforms to Jo 7:48. [0388D] Ut non videntes videant, et videntes caeci fiant [n: (Joan. IX)] — Divergence, DECLINED. Jo 9:39 reads ut qui non vident videant, et qui vident caeci fiant. Participles for relative clauses; the assertion is identical and the antithesis the gloss rests on survives untouched. Both negations carried. [0388D] VERS. 51.-- Nunquid lex. — Not a divergence. Orthography as above; Jo 7:51 otherwise. [0388D] Et tu Galilaeus es, — Conforms as far as it runs: Jo 7:52 opens Numquid et tu Galilaeus es? and the lemma begins one word in. [0388D] Et reversi sunt, — Conforms, read against its own verse. [0389B] VERS. 1.-- Jesus autem. — Conforms, read against its own verse. [0389C] Qui sine peccato est, primus in illam lapidem mittat. — Divergence, DECLINED. Jo 8:7 reads Qui sine peccato est VESTRUM, primus in illam lapidem mittat. One token absent: single-token class, no marker. (in illam lapidem, a feminine pronoun on a masculine noun, is the Clementine's OWN reading and is not a defect here.) [0389C] VERS. 3.-- Adducunt. — Conforms, read against its own verse. [0389C] VERS. 6.-- Jesus autem inclinans se deorsum. — Conforms, read against its own verse. [0389C] VERS. 8.-- Et iterum se inclinans. — Conforms, read against its own verse. [0389C] VERS. 11.-- Nec ego te condemnabo. — Conforms, read against its own verse. [0389C] Nec ego te condemnabo, — Conforms, read against its own verse. [0389C] VERS. 12.-- Iterum ergo. — Conforms, read against its own verse. [0389C] Ego sum lux mundi. — Conforms, read against its own verse. [0389C] VERS. 13.-- Testimonium. — Conforms, read against its own verse. [0389C] VERS. 14.-- Verum est testimonium. — Conforms (Jo 8:14, verum est testimonium meum), abridged. [0389C] VERS. 15.-- Vos secundum carnem judicatis. — Conforms, read against its own verse. [0390C] Ego non judico quemquam. — Conforms, read against its own verse. [0390C] Et si judico ego, — Conforms, read against its own verse. [0390C] VERS. 16.-- Quia solus [n: (AUG.)] — Conforms (Jo 8:16, quia solus non sum), but the lemma STOPS BEFORE THE NEGATION; the gloss restores it in its first words (Ideo verum, quia non solus), and the English follows the gloss rather than supplying a negative to the lemma. [0390C] VERS. 17.-- Et in lege vestra scriptum est. [n: (ID.)] — Conforms, read against its own verse. [0390C] usque ad — Structural, not scripture: Migne's abridgment formula, rendered as far as. The pre-scan's tick is a phrase-search artefact, the two words being common in the Vulgate. [0390C] VERS. 19.-- Dicebant ergo ei. — Conforms, read against its own verse. [0390D] Ubi est Pater tuus. — Conforms, read against its own verse. [0390D] Forsitan et Patrem, — Conforms (Jo 8:19), abridged before meum, so the English reads "Perhaps also the Father," and does not supply the possessive. [0390D] VERS. 20.-- Haec verba, — Conforms, read against its own verse. [0391A] Et nemo apprehendit eum. — Conforms, read against its own verse. [0391A] VERS. 21.-- Ego vado. — Conforms, read against its own verse. [0391A] Quaeretis, — Conforms, read against its own verse. [0391A] In peccato vestro moriemini. — Conforms, read against its own verse. [0391A] VERS. 22.-- Nunquid interficiet semetipsum? — Not a divergence. Orthography only (Jo 8:22). [0391A] VERS. 23.-- Vos deorsum estis. — Divergence, DECLINED. Jo 8:23 reads Vos DE deorsum estis. The dropped preposition shifts "you are from beneath" to "you are beneath", but it is one short token, the class most often our own file, and the gloss's continuation (ideo terram sapitis) reads the same under either. Rendered as printed ("You are below"), logged for the plate. [0391A] VERS. 24.-- Dixi ergo vobis. — Conforms, read against its own verse. [0391A] Si enim. — Conforms (Jo 8:24, si enim non credideritis); the lemma likewise cuts before non, and the gloss's continuation carries it. [0391A] Credideritis. — Conforms, read against its own verse. [0391A] Ego sum, moriemini, — Conforms: one italic run splicing two fragments of Jo 8:24 (credideritis quia ego sum, moriemini in peccato vestro), reproduced as one run. [0391A] VERS. 25.-- Principium. [n: (AUG.)] — Conforms, read against its own verse. [0391C] usque ad — Structural, not scripture: Migne's abridgment formula, rendered as far as. The pre-scan's tick is a phrase-search artefact, the two words being common in the Vulgate. [0391C] Loquor vobis. — Conforms, read against its own verse. [0391C] VERS. 27.-- Et non cognoverunt. — Conforms, read against its own verse. [0391C] qui misit me, verax est, — Divergence, DECLINED. Jo 8:26 reads sed qui me misit, verax est. Transposition only; the gloss quotes it mid-sentence, which is why sed falls outside the span. [0391D] Cum exaltaveritis Filium, — Divergence, DECLINED, AND THE WITNESS IS WHAT IS ODD HERE. clementine-flat.txt reads Jo 8:28 Cum exaltaveriS Filium hominis, a singular. Migne's plural exaltaveritis is the reading the Clementine editions print and the one the context demands (he is addressing the Jews). A [var:] here would rest on what looks like a defect IN OUR COLLATION FILE. Flagged so the file can be checked; no marker. [0391D] VERS. 29.-- Quia ego sum. — Conforms, read against its own verse. [0391D] Et a meipso facio nihil; — Conforms, read against its own verse. [0391D] Quia ego, — Conforms, read against its own verse. [0391D] VERS. 30.-- Multi crediderunt in eum. — Conforms, read against its own verse. [0391D] VERS. 31.-- Vere discipuli. — Conforms, read against its own verse. [0391D] Liberi. — Conforms, read against its own verse. [0391D] Veritas liberabit vos, — Conforms, read against its own verse. [0391D] Respondit eis, — Conforms, read against its own verse. [0391D] liberabit vos, — Conforms, read against its own verse. [0391D] Et nemini servivimus. — Conforms (Jo 8:33), abridged before umquam. [0391D] Amen, amen dico, — Conforms, read against its own verse. [0391D] VERS. 35.-- Servus autem, — Conforms, read against its own verse. [0391D] VERS. 37.-- Scio, — Conforms, read against its own verse. [0391D] VERS. 39.-- Responderunt et dixerunt ei. — Conforms, read against its own verse. [0391D] Si filii Abrahae, — Conforms, read against its own verse. [0391D] VERS. 40.-- Nunc autem quaeritis, — Conforms, read against its own verse. [0391D] VERS. 41.-- Vos facitis opera. — Conforms, read against its own verse. [0391D] Dixerunt, — Conforms, read against its own verse. [0391D] Nos, — Conforms, read against its own verse. [0392D] VERS. 42.-- Dixit ergo. — Conforms, read against its own verse. [0393A] Ego enim. — Conforms, read against its own verse. [0393A] VERS. 43.-- Quare loquelam. — Conforms, read against its own verse. [0393A] VERS. 44.-- Vos ex patre, — Conforms, read against its own verse. [0393A] Pater tuus Amorrhaeus est, mater tua Cethaea [n: (Ezech. XVI)] — Divergence, DECLINED. Ez 16:3 reads pater tuus Amorrhaeus, et mater tua Cethaea: Migne supplies the copula and drops et. Function words only. [0393A] Et desideria. — Conforms, read against its own verse. [0393A] desideria patris [0393B] vestri vultis facere, — Conforms, read against its own verse. [0393B] Ille homicida. — Conforms, read against its own verse. [0393B] Lingua eorum gladius acutus [n: (Psal. LVI)] — Conforms, read against its own verse. [0393B] Et in veritate non stetit. — Conforms, read against its own verse. [0393B] Quia non est in eo veritas. — Divergence, DECLINED. Jo 8:44 reads quia non est veritas in eo. Transposition only, and the gloss's own next sentence prints Migne's order back (quia non est ipsa in eo), which is evidence for the reading rather than against it. Negation intact in both. [0393B] Ex propriis loquitur, quia mendax est. — Conforms, read against its own verse. [0393B] Et pater ipsius. — Divergence, DECLINED. Jo 8:44 reads et pater ejus. One pronoun for another of the same reference; the exposition (Scilicet mendacii quod invenit) is indifferent to which. [0393B] Ex propriis. — Conforms, read against its own verse. [0393B] VERS. 46.-- Quis ex vobis, — Conforms, read against its own verse. [0393B] VERS. 48.-- Nonne, — Conforms, read against its own verse. [0393B] VERS. 49.-- Honorifico. — Conforms, read against its own verse. [0393B] VERS. 50.-- Ego autem. — Conforms, read against its own verse. [0393B] Daemonium habes, — Conforms, read against its own verse. [0393D] Honorifico Patrem, — Conforms (Jo 8:49), abridged before meum; same treatment. [0393D] est qui quaerat et judicet. — Conforms, read against its own verse. [0393D] Tentat vos Deus [n: (Deut. XIII)] — Divergence, DECLINED. Dt 13:3 reads tentat vos Dominus Deus vester. An abridgment of a proof-text quoted for its verb; nothing substituted. [0393D] Deus intentator malorum est [n: (Jac. II)] — Divergence, DECLINED. The text is Jac 1:13, Deus ENIM intentator malorum est: connective only. Migne's citation reads Jac. II where the verse is Jac. I; logged in (d) with the Marc. XI case, not marked. [0393D] Judica me, Deus, et discerne causam meam [n: (Psal. XLII)] — Conforms, read against its own verse. [0393D] VERS. 51.-- Amen, amen dico vobis. — Conforms, read against its own verse. [0393D] Non videbit. — Conforms, read against its own verse. [0393D] VERS. 52.-- Nunc cognoscimus. — Divergence, DECLINED. Jo 8:52 reads Nunc cognovimus. Present for perfect, one token; the English renders the printed present ("Now we know"). Single-token class. [0393D] VERS. 53.-- Quem teipsum facis? — Conforms, read against its own verse. [0394B] Non rapinam arbitratus est esse se aequalem Deo [n: (Philip. II)] — Conforms, read against its own verse. [0394B] VERS. 54.-- Respondit Jesus. — Conforms, read against its own verse. [0394B] Quem te facis? — Not a Vulgate divergence: it is MIGNE AGAINST MIGNE. His lemma at VERS. 53 prints Quem teipsum facis? (= Jo 8:53), his gloss at VERS. 54 re-quotes it as Quem te facis?, and the close of the earlier gloss prints quem teipsum facis? again. All three are rendered as printed. Whether the short form is his or our transcription's, nothing here decides. [0394B] Est Pater meus. — Conforms, read against its own verse. [0394B] Qui glorificat. — Conforms, read against its own verse. [0394B] Et non cognovistis. — Conforms, read against its own verse. [0394B] Et si dixero, — Conforms, read against its own verse. [0394C] et si dixero. — Conforms, read against its own verse. [0394C] VERS. 56.-- Abraham. — Conforms, read against its own verse. [0394C] Diem meum. [n: (AUG.)] — Conforms, read against its own verse. [0394C] usque ad — Structural, not scripture: Migne's abridgment formula, rendered as far as. The pre-scan's tick is a phrase-search artefact, the two words being common in the Vulgate. [0394C] VERS. 59.-- Et exivit. — Conforms, read against its own verse.

(d) THIS RANGE'S OWN CRUCES, BY COLUMN

Two markers fired in the whole range, both [cj: …], both at sites where the faithful English would otherwise state a falsehood in the author's voice. No [sic:], and no non-word or broken run was found to justify one. Everything else below is rendered as printed and carries no marker, with the column named so the plate read can settle it.

FIRED

0381A — [cj: vanitatem; read unitatem, "the unity"] (chunk 0010, VERS. 9). Est puer unus … Unus, quia vanitatem Ecclesiae futuram praefigurabat. The boy is one because he prefigured something future about the Church; vanitatem makes that the Church's coming vanity, which is not a thing this gloss or any near relative of it says, and the argument (oneness prefiguring a future one-ness) has no purchase on it. vanitatem is a real word, so Pattern 12 has nothing to wrap and Pattern 7 requires the English to render it — which it does, with the conjecture added beside it, per Pattern 18. ⚠ Whose reading it is, I do not know. The two forms are not an obvious optical confusion, which argues mildly for the type rather than the transcription, and that argument is exactly the kind measured as unreliable; it is offered as context, not as a finding.

0383A — [cj: me esse Patrem; read me esse a Patre, "that I am from the Father"] (chunk 0010, VERS. 43). …nemo potest venire ad me nisi a Patre tractus, et ideo vos non creditis me esse Patrem, quia non estis tracti a Patre. Rendered flat, the page has Christ complaining that the Jews do not believe him to be the Father — a claim the same chunk denies twice over (0383A: hoc est credere in Filium quod et in Patrem; 0383C: ego novi Patrem, quia ab illo sum), and which the sentence's own two other clauses contradict, both of which turn on being drawn a Patre. One short preposition is missing. ⚠ That is precisely the shape that comes back as our own file more often than not, so no accusation is made: the English renders Patrem, the conjecture stands beside it, and if the plate prints a Patre this becomes a TEI patch and the marker comes out.

DECLINED — the two conjecture candidates I did NOT mark

0387D assensionem (chunk 0012, VERS. 33). quaeretis me post resurrectionem et assensionem. The sense wants ascensionem; assensio ("assent") is a real word, and the same chunk prints post ascensionem correctly at 0388C. ⛔ I decline the conjecture marker anyway, on the bar in (a).4: "after the resurrection and the assent" is opaque, not false, and the internal corroboration four columns later is the exact argument 9004 measured as worthless for attribution ("the correct form prints eleven words later"). A single c/s difference is also the most ordinary transcription slip there is. Rendered as printed; flagged here as the first thing to look at on leaf n199.

0385D audit (chunk 0012, VERS. 6). et ideo non audit vos mundus suos falsos testes. The verse being glossed is Jo 7:7, Non potest mundus odisse vos: me autem odit, which makes odit the expected verb and the whole sentence turn: the world does not hate you, its own. As printed, the world does not hear its own false witnesses, which is odd but asserts nothing false, so it stays under Pattern 7 with no marker. Rendered "does not hear"; flagged for the plate.

RENDERED AS PRINTED, NO MARKER — read at the plate before anyone corrects them

• 0384A (0010) — sed audivit, unde magis moveantur. A third-singular perfect where the clause wants a present, and it does not construe with moveantur. Rendered "but he heard, whence they are the more moved."
• 0387A (0012) — non est judicandum secundum faciem, id est actionem personarum. The standing phrase is acceptio personarum, and the gloss's next sentence prints ne in judicandis causis personas accipiamus. actionem is a real word; rendered "the action of persons."
• 0387C (0012) — Breviter utrumque complectatur. Subjunctive where the sense wants an indicative; rendered "Let him briefly comprise both."
• 0388B (0012) — quibus alii conderentur, alii docerentur. conderentur construes ("might be founded") and is left; noted only because the pairing with docerentur invites a reader to emend it.
• 0388B (0012) — non fuit ille fons. Perfect where the argument is timeless; carried as a past ("that man was not a fountain"), negation intact.
• 0389D (0013) — Vel signat Ecclesiam, quae relinquens Deum, fornicata erat cum daemonibus. The third allegory of the adulteress makes the Church the one who forsook God and fornicated with demons; the reading a commentator would expect here is the Gentile people. A real word, rendered as printed.
• 0390B (0013) — confirmatur. veni missus a Patre. A sentence opening in lower case after a full stop. English orthography capitalises it; the lower case may be his or the transcription's, and nothing is claimed. No mark was added or deleted.
• 0392B (0013) — isti eidem tributis Caesaris. The verb is gapped from the preceding clause; rendered "these men serve the same by the tributes of Caesar."
• 0393A (0014) — ostendit Deum non patrem eis esse eorum. Both eis and eorum stand; both are carried ("that God is not their father to them"). Not a Pattern 11 dittography — the run is not repeated, the pronoun is doubled by case.
• 0394A (0014) — Hinc discat homo Jesus, non malum, sed bonum reddere. A nominative Jesus where the sense wants a Jesu. Rendered literally, "Hence let man learn Jesus, to render not evil but good," rather than supplying the preposition.
• 0383B (0010) — licet diversa prophetarant. Syncopated pluperfect for prophetaverant: Pattern 9, correct Latin for its period, translated normally with no crux owed. Listed only so a later sweep does not churn it.

PUNCTUATION AND LAYOUT

• ⚑ 0393C (0014), VERS. 49 ends with NO TERMINAL STOP — …sicut ego honorifico Patrem, sic vos me honoraretis — and none is supplied (Pattern 8: never supply, including a terminal stop). The English paragraph ends the same way. This is the one place in the range where an English reader will see something that looks like our error and is not.
• 0387B (0012) — the Isaias citation likewise closes without a stop after [n: (Isa. LIII)]; reproduced.
• 0386A (0012) — quod est - Ego non sum a meipso. Migne's spaced hyphen is reproduced as printed rather than converted to a colon or a dash.
• 0384D (0011) — Suspenditur oratio, Quasi dicat: — a capital after a comma, kept.
• No em-dash occurs in any of the five English chunks (checked mechanically: zero).

CITATIONS THAT DO NOT RESOLVE — logged, not marked

Both pass through verbatim inside [n: …] per the sacred-marker rule; neither is corrected, queried or expanded, and neither is a defect class this edition marks in the text.

• 0387B (0012) — [n: (Marc. XI)] on nemo novit Patrem nisi Filius, et cui Filius voluerit revelare. The text is Matt 11:27 / Lc 10:22; Mark has no such verse anywhere, let alone in chapter XI. ⚑ Per the brief's Sect. 6.2 this is NOT the Ibid. class and is not proposed for correction; it is recorded for the index, where refDisplay keeps what Migne prints.
• 0393D (0014) — [n: (Jac. II)] on Deus intentator malorum est, which is Jac 1:13.

CONVENTIONS INHERITED FROM MARK — the controls, none forced open

1. The gloss voice is not archaised. 33 says/said for dicit/ait across the five chunks; zero saith. Every -eth form in the range (13 of them: believeth, eateth, quickeneth, seeth, speaketh, knoweth, seeketh, judgeth, glorifieth, tempteth) sits inside an italic lemma — checked mechanically, not by eye. The archaic second person in the gloss (thou doest, thou seest, thy words) is Pattern 17's number marker on a singular addressee, not archaism, and it alternates with plain "you" wherever the Latin turns plural, in the same paragraph where it does so (0385B, 0390D, 0393B).
2. turba → crowd, multitudo → multitude. Mechanically 1:1 in this range: 6 turba nouns → six "crowd", 2 multitudo/multitudinem → two "multitude", and no third word used for either. 0381A puts multitudinem in te creditorum eleven words from turba*'s territory and the two stay apart.
3. daemonium → demon, diabolus → devil. daemonium at 0386C and Daemonium habes at 0393D → "demon"; daemonibus at 0389D → "demons"; diabolus/diaboli at 0392C, 0393A, 0393B → "devil". No crossing, and the Douay "devils" pull was resisted at 0386C where it is strongest (daemonium habere, qui daemonia pellit).
4. gentes. de Judaea ad gentes (0385C) and iturus est ad gentes credituras (0388A) are the Judaei/gentes typology → Gentiles; templo gentium (0394C), which closes the range against de figuris Judaeorum, likewise → Gentiles. ut omnes gentes docerent (0388C) is the Matt 28:19 phrase → nations, per the rule. ⚑ One judgement call: 0380C gentium magna turba sequitur glosses Jo 6:2's multitudo magna — a crowd following in Galilee, with no Jew/Gentile contrast in sight — so it is rendered "a great crowd of the nations", generic. Flagged because it is the only site in the range where the axis is genuinely ambiguous, and a merge that prefers "Gentiles" there will get no argument from me. No natio occurs in this range, so nothing was flattened.
5. mysterium → mystery, sacramentum → sacrament, 1:1 and both present: mystery at 0381C, 0390D, 0394C; sacrament at 0381D, 0383D, 0384C. This is the control that shows the other four axes are being tracked rather than asserted.

⚑ No inherited convention was broken and no exception was taken.

MECHANISM 5 — the three sites where the easier reading was the wrong one

Checked by case and gender, not by sense, per the brief's Sect. 4.

• 0384C (0011) — ut homo vivat ex me, participatio facit quae me manducat. The relative is feminine and agrees with participatio, not with homo. The English reads "the participation effects which eats me"; the easy reading ("he who eats me") would have rebound the clause to the man.
• 0383C (0010) — quia verbum est vita hominum quod et carni suae … dat vitam in resurrectione. The relative is neuter: the giver of life is verbum, not vita (feminine) and not the flesh. Rendered so.
• 0382D (0010) — Mihi unitur humilis humili. humili is dative and agrees with Mihi: "the humble man is united to me, myself humble" — not "the humble is united to the humble."

One further binding is genuinely ambiguous on the page and is left ambiguous: 0383C, patribus murmuratoribus quorum estis filii, murmurando non fuit panis vitae — murmurando can attach to quorum estis filii ("whose sons you are by murmuring") or to non fuit panis vitae. The comma before it and its position beside the verb decide the English for the second ("was, by their murmuring, not the bread of life"). ⚑ Recording the ruling here because the brief's Sect. 4 warns that a crux can state one reading while the English carries the rival: the English carries the reading just stated, and I checked the shipped sentence against this line after writing it.

THE usque ad ABRIDGMENTS

Seven in the range (0383B, 0383D, 0386C, 0387D, 0390C, 0391C, 0394C), all rendered as far as, following 9001. The text after the formula is Migne's resumed extract and is translated; the etc. before it is left as printed.

<!-- ===== cruces-0015.md ===== -->

Stint 4 of six. Cols 0394C–0407A (chunk colContext bands), covering CAPUT IX–XIII: the man born blind, the good shepherd, the encaenia, the raising of Lazarus, the entry into Jerusalem, and the supper and washing of feet.

(a) THE POLICY I APPLIED — stated before the findings, so it can be argued with

1. Attribution. I treat my own guess about whose a defect is as carrying no evidential weight (brief §2; measured wrong 9 times in 14 on 9004). Concretely:

• A real, well-formed Latin word that merely reads oddly is rendered as printed, logged in §(d), and takes no marker — however good the internal argument. Where the argument is good I say so and say that it settles nothing.
• A non-word or broken run gets the same treatment: carried, logged, not marked. I fired zero [sic:] in this range. Two non-words are logged (0399B impieretur, 0400C Sea).
• A [var:] I can cite from sources/vulgate/clementine-flat.txt by book chapter:verse I fired — three of them, and every one rests on a multi-word or lexical divergence inside a quoted clause, never on a single letter.
• I fired zero [cj:], zero [d:], zero [ed:].

2. The [var:] bar I set, and why it is narrower than "any divergence from the Clementine." The class split says a quoted clause is evidence about Migne and a single word is suspicion of our file. I read that as a statement about what a letter-swap can manufacture, not about span length — so my test was: could an OCR have produced this? A different verb (Redi for Vade), an inserted conjunction that re-shapes a sentence (nisi Unigenitus), a clause pulled in from another verse — no. An inflection (est for es), a vowel (quid for quod), a fusion (Romam for Romani), word order, an ae/e spelling — yes, and those are all declined below with the reason given per span. ⛔ The hardest decline is 0395D Jo 9:34 natus est totus, which is a genuine sense change (third person for second) inside a quoted clause and would have been my fourth marker: it is declined because CLAUDE.md rule 8a records the digitization manufacturing exactly this shape at 9003 (judicabit→judicavit), and a Vulgate collation cannot see that class by construction.

3. Negation and mechanism 5. Every printed non/nec/nisi/nullus/nihil is carried, including where carrying it leaves the sentence at odds with itself (0396D, 0398C). Where a pronoun or relative could attach to two nouns I checked case and gender rather than sense; the sites where I had to choose are logged in §(d) with the reading the English actually carries, and I have re-read each English clause to confirm it carries the reading the crux claims.

4. Inherited conventions (brief §1). Applied from the first line. No exception was forced by John in this range; the axes are exercised as follows:

5. Other standing rules. Pattern 17: tu Juda (0401D) and Tu quid dicis (0394C) are singular → thou; vos / habetis / creditis → you. Pattern 8b fired once, at VERS. 7 (0394D–0395A), where Migne closes an indirect question with ? — rendered "It is asked: how did he heal…?" Pattern 8: Migne's marks stand, including the missing terminal stop on the 0405D lemma and the medial commas at 0396C and 0405D. Guillemets: Migne prints none in this range and none were supplied; three straight-quote pairs are ours (0397D "of mine", 0401C "faithful", 0407A "whither I go"). Em-dashes: zero.

6. usque ad → "down to" (7 occurrences). ⚠ Flag for the merge, not a finding: the work is already inconsistent on this above my range — chunk 0001 renders it "as far as", chunk 0005 "down to". I follow 0005. One of the three should be swept at merge.

(b) SPAN COUNT — RUN, AND HERE IS THE RESULT

data/briefs/8999-lemmata-0015.txt declares 199 spans for chunks 0015–0019.

I counted the … spans in my own Latin: 25 + 47 + 40 + 50 + 37 = 199. The totals agree exactly, chunk by chunk. No head-of-range span is missing and no tail span belongs to another stint.

Two secondary checks, both clean: column anchors 6 / 12 / 9 / 13 / 10 and [n:] notes 1 / 7 / 5 / 4 / 5 reproduce the Latin 1:1, in order and in place. node scripts/verify-english.mjs 8999 reports verify OK: 28 chunks, columns/notes/sections aligned.

⚠ One caveat on the brief's own column labels. The brief attributes most spans to the chunk's colContext band; I attribute each to the last anchor actually preceding it in the Latin, so the band letters in my list below are finer-grained than the brief's and will not match it line for line. The order and the count match exactly, which is what the check is for.

⚠ And one substantive finding about the pre-scan itself: at 0401A the span Sed ut filios Dei qui erant dispersi congregare. is marked ✓ Clementine verbatim, and a literal grep of it returns 0 hits (Jo 11:52 reads congregaret in unum). So a ✓ in this file is not merely "not clearance" — at least one of them is not a verbatim match at all. Walk the ✓ lines as carefully as the ⚠ ones.

(c) THE LEMMA DIVERGENCE LIST — all 199 spans, in order

= means: read word by word against the Clementine at the verse actually being glossed, and it conforms (or is a bare incipit/fragment of it, or an usque ad abridgment marker, which is apparatus and not scripture). Everything else carries its verdict and, where declined, the reason.

Markers fired: 3 [var:] — 0397C ×2, 0399C ×1. Those three columns need the plate read before ship (plate-gate.mjs). No other marker of any kind was fired in this range.

Chunk 0015 — 25 spans

1. [0394C] Et praeteriens. — =
2. [0394C] Caecum. — =
3. [0394D] Neque hic peccavit. — =
4. [0394D] ut manifestentur opera Dei in illo. — =
5. [0394D] Me oportet operari. — =
6. [0394D] Quandiu sum in mundo. — ⚠ declined. Vulg. Jo 9:5 quamdiu. Orthographic only, and quandiu is this work's own spelling elsewhere (0403C, 0403D). Pattern 9, no marker.
7. [0395A] Et fecit lutum. — =
8. [0395A] Unxit oculos. — = — but the words stand at Jo 9:11 (et unxit oculos meos), not at 9:6, under which Migne prints them as a second lemma. The recension's VERS. addresses cover a span of verses; not a divergence.
9. [0395A] Et venit videns, — =
10. [0395B] Non est hic homo a Deo, — =
11. [0395B] Tu quid dicis, — =
12. [0395B] Ille autem dixit, — =
13. [0395B] Donec vocaverunt, — =
14. [0395B] Da gloriam Deo: Nos scimus, — =
15. [0395C] Quid iterum, — ⚠ declined. Vulg. Jo 9:27 quod iterum vultis audire. One word, and quod→quid is an o/i substitution of exactly the class an OCR manufactures. The gloss (Stomachatur contra duritiam eorum) reads equally on either. Rendered as printed, "Why again". No marker.
16. [0395C] Maledixerunt ergo ei. — =
17. [0395C] Hunc autem nescimus. — =
18. [0395C] Hunc scimus: — = at Jo 7:27 (Sed hunc scimus unde sit), not at 9:29; the gloss's Superius dixerant is correct.
19. [0395C] Nisi esset hic a Deo, — =
20. [0395D] In peccatis natus est totus. — ⚠ declined — the range's most tempting [var:]. Vulg. Jo 9:34 In peccatis natus es totus (2nd person); Migne prints est. Reason declined: a one-letter inflectional difference inside a quoted clause is precisely the class CLAUDE.md rule 8a records the digitization manufacturing — 9003's judicabit→judicavit, which a Vulgate collation then dutifully "found" — and no collation can catch that class. A [var:] here would be a public claim against Migne resting on a letter. Rendered as printed ("He was born wholly in sins"); logged for the plate.
21. [0395D] Et tu doces nos? Et ejecerunt eum foras. — =
22. [0395D] Quis est, — =
23. [0395D] Et vidisti. — =
24. [0396A] Ut qui non vident. — =
25. [0396A] Ego non judico quemquam. — = at Jo 8:15, adduced as a cross-reference under VERS. 39.

Chunk 0016 — 47 spans

1. [0396A] Amen, amen dico vobis. — =
2. [0396B] Per ostium. — =
3. [0396B] Fur. — =
4. [0396B] Pastor est ovium. — =
5. [0396C] Huic ostiarius aperit. — =
6. [0396C] Et oves, — =
7. [0396C] usque ad — =
8. [0396D] Illi autem non cognoverunt. — =
9. [0396D] Omnes, — =
10. [0397A] Si quis. — =
11. [0397A] Et ingredietur. — =
12. [0397A] Ut vitam habeant. — =
13. [0397A] Nisi ut furetur, — =
14. [0397A] Mercenarius autem, — =
15. [0397B] Et lupus rapit. — =
16. [0397B] Quia mercenarius est, — =
17. [0397B] Ego sum pastor. — =
18. [0397C] Ego sum pastor bonus et animam pono pro ovibus. — ⚠ [var:] FIRED. Vulg. Jo 10:11 Ego sum pastor bonus. Bonus pastor animam suam dat pro ovibus suis. Migne's span joins 10:11a to a form of 10:15 (et animam meam pono pro ovibus meis), possessives dropped. Multi-word and structural: a letter-swap cannot manufacture it, and the gloss that follows (Quod non ipse unus fecit) expounds the joined clause.
19. [0397C] Et ego. — =
20. [0397C] Deum nemo vidit unquam, nisi Unigenitus qui est in sinu Patris, ipse enarraverit [n: (Matth. XI)] — ⚠ [var:] FIRED. Vulg. Jo 1:18 Deum nemo vidit umquam: unigenitus Filius, qui est in sinu Patris, ipse enarravit. Migne prints nisi Unigenitus … ipse enarraverit: a conjunction inserted, Filius dropped, the verb out of the perfect indicative. That re-shapes the sentence, not a letter of it. (Migne's note reads (Matth. XI) on a John quotation; reproduced verbatim per §2, not queried.)
21. [0397C] nemo novit Filium nisi Pater, neque Patrem quis novit nisi Filius, et cui Filius voluerit revelare [n: (Matth. XI; Luc. X)] — ⚠ declined. Vulg. Mt 11:27 et cui voluerit Filius revelare; Migne et cui Filius voluerit revelare. Word order only. Pattern 14's bar is a change of sense, and there is none.
22. [0397C] Et animam meam pono. — =
23. [0397C] Pro ovibus. — =
24. [0397D] Oportet me. — =
25. [0397D] Et vocem meam. — =
26. [0397D] Quia ego, — =
27. [0397D] usque ad — =
28. [0397D] Facta sunt autem encaenia, — =
29. [0398A] In porticu. — =
30. [0398A] Circumdederunt ergo eum Judaei. — =
31. [0398A] Quousque animam nostram tollis? — = verbatim Jo 10:24 — ⚑ but a 7a″ near-miss worth recording. The familiar English ("hold our soul in suspense") is not what tollis says, and the gloss turns on tollere animas. Rendered "take away".
32. [0398A] Si tu es Christus, dic nobis palam. — =
33. [0398B] Et nemo potest rapere de manu Patris mei. — =
34. [0398B] Ego et Pater unum sumus. — =
35. [0398B] Sustulerunt ergo lapides. — =
36. [0398C] Facis teipsum. — =
37. [0398C] Ego et Pater unum sumus, — =
38. [0398C] Nonne scriptum est, — =
39. [0398C] Quia ego. — =
40. [0398C] Ego dixi: Dii — =
41. [0398C] Si illos dixit deos. — =
42. [0398D] Operibus credite. — =
43. [0398D] Quia in me est Pater. — ⚠ declined. Vulg. Jo 10:38 quia Pater in me est; Migne quia in me est Pater. Word order only.
44. [0398D] Ego et Pater unum sumus. — =
45. [0398D] qui videt me, videt et Patrem; — =
46. [0398D] Et multi venerunt. — =
47. [0398D] Quia Joannes, — =

Chunk 0017 — 40 spans

1. [0399A] Erat autem. — =
2. [0399A] Languens Lazarus. — =
3. [0399A] Ecce quem amas, — =
4. [0399B] Pro gloria Dei, — =
5. [0399B] Diligebat. — =
6. [0399B] Duobus diebus. — =
7. [0399B] Eamus in Judaeam. — =
8. [0399B] Dicunt ei discipuli. — =
9. [0399C] Absit a te, Domine, — = Mt 16:22 (the italic stops at Domine,). ⚑ The roman continuation Migne sets after it, non fiet tibi istud, diverges from the Vulgate's non erit tibi hoc — outside the italic, so outside 7a″. Logged, no marker.
10. [0399C] Beatus es, Simon Barjona, quia Pater revelavit tibi, — ⚠ declined. Vulg. Mt 16:17 quia caro et sanguis non revelavit tibi, sed Pater meus, qui in caelis est. Migne's span is an abridgment, not an altered reading: every word retained agrees with the Vulgate and the retained clause's sense is unchanged. ⚑ The dropped material contains a non, but no negation is dropped from anything the English renders — the span is shorter, not negated differently.
11. [0399C] Redi post me, Satana, quia sapis non quae Dei, sed quae hominis sunt. — ⚠ [var:] FIRED. Vulg. Mt 16:23 Vade post me Satana, scandalum es mihi: quia non sapis ea quae Dei sunt, sed ea quae hominum. Migne prints Redi post me — a different verb, not a letter-swap — and omits scandalum es mihi. (The non is displaced but not lost, and hominis for hominum is a single letter; the marker rests on Redi, which is what makes it a claim about wording rather than about type.)
12. [0399C] Nonne duodecim sunt horae. [n: (HIER.)] — =
13. [0399D] usque ad — =
14. [0399D] Ut credatis. — =
15. [0399D] Et invenit. [n: (AUG.)] — =
16. [0399D] usque ad — =
17. [0399D] Sed et nunc scio. — =
18. [0399D] Etiam si mortuus fuerit. — =
19. [0400A] Vivet. — =
20. [0400A] Silentio. — =
21. [0400A] Et vocat te. — =
22. [0400A] Secuti sunt. — =
23. [0400B] Infremuit spiritu. — =
24. [0400B] Ubi posuistis eum? — =
25. [0400B] Non novi vos [n: (Matth. VII)] — ⚠ declined. Vulg. Mt 7:23 numquam novi vos; cf. Mt 25:12 nescio vos. A loose citation of a proverbially quoted clause; sense unchanged.
26. [0400B] Ubi est Adam? [n: (Gen. III.)] — ⚠ declined. Gen 3:9 reads Ubi es? Migne gives the traditional third-person paraphrase. The gloss's point — God asking as though he did not know — is unaffected either way.
27. [0400C] Jesus ergo, — =
28. [0400C] Erat autem spelunca. — =
29. [0400C] Domine, jam fetet. — ⚠ declined. Vulg. Jo 11:39 foetet. ae/e orthography, Pattern 9.
30. [0400C] Sea propter populum, — ⚠ NON-WORD — logged, NO MARKER. Migne's span opens Sea where Jo 11:42 reads sed propter populum. Carried untranslated inside the italic lemma (nested italics are not expressible, Pattern 10). The conjecture Sed is certain in sense and unattributable in origin**: d→a is an ordinary OCR shape and an ordinary broken sort alike. The span governs no translated clause, so nothing is quarantined (7a⁗ qq. 1–3 all pass).
31. [0400D] Voce magna. — =
32. [0400D] Et statim prodiit. — =
33. [0400D] Solvite eum. — =
34. [0400D] Credent in eum. Et venient Romani. — =
35. [0401A] Et venient Romam. — ⚠ declined, and see §d. Vulg. Jo 11:48 et venient Romani, which Migne's own lemma four lines above prints correctly. Romam is a real Latin word (acc. of Roma), and ni→m is a textbook OCR fusion. ⛔ "The correct form prints a few lines earlier" is the reasoning §2 names as a failed attribution test: it corroborates that an error exists and says nothing about whose. Rendered as printed. No marker.
36. [0401A] Hoc a semetipso. [n: (AUG.)] — = fragment of Jo 11:51 (Hoc autem a semetipso non dixit); the lemma incipit drops autem.
37. [0401A] usque ad — =
38. [0401A] Pro gente. — =
39. [0401A] Sed ut filios Dei qui erant dispersi congregare. — ⚠ declined — and the brief's ✓ on this span does not reproduce. Vulg. Jo 11:52 sed ut filios Dei, qui erant dispersi, congregaret in unum; Migne has the infinitive congregare and drops in unum. A literal grep of the span against clementine-flat.txt returns 0 hits, so this ✓ is a pre-scan artifact — reported per the brief's own warning that ✓ is not clearance. Declined because the span is a truncated lemma and an abridgment cannot be told from a transcription slip.
40. [0401B] Proximum autem. — =

Chunk 0018 — 50 spans

1. [0401B] Venit in Bethaniam. — ⚠ declined. Vulg. Jo 12:1 venit Bethaniam (no preposition). ⚑ Migne's own gloss on the next line prints prius Bethaniam venit, without the preposition — the two forms stand five words apart, which corroborates that they differ and says nothing as to whose the in is. One particle. No marker.
2. [0401B] Maria ergo. — =
3. [0401B] Dixit ergo unus. — =
4. [0401C] Non quia, — =
5. [0401C] Loculos habens, — =
6. [0401C] Sinite illam. — =
7. [0401D] praevenit ungere corpus meum in sepulturam. — =
8. [0401D] Me autem non semper habetis. — =
9. [0401D] Et venerunt. — =
10. [0401D] Cogitaverunt. — =
11. [0402A] In crastinum. — =
12. [0402A] Clamabant. [n: (AUG.)] — =
13. [0402A] Rex Israel. — =
14. [0402A] Noli timere, filia. — =
15. [0402B] Ecce totus. — ⚠ declined. Vulg. Jo 12:19 ecce mundus totus post eum abiit. An incipit taking the first and third words; the recension abridges lemmata this way throughout.
16. [0402B] Hi ergo accesserunt. — =
17. [0402B] Jesus autem. — =
18. [0402B] Benedictus qui venit in nomine Domini; — =
19. [0402B] Venit hora ut clarificetur. — =
20. [0402C] Amen, amen dico vobis, — =
21. [0402C] Qui amat animam suam, — =
22. [0402C] Nisi granum frumenti. [n: (AUG.)] — =
23. [0402C] Qui amat. [n: (AUG.)] — =
24. [0402D] Qui odit animam, — =
25. [0402D] Si quis mihi ministrat. — =
26. [0402D] si quis mihi ministrat. — =
27. [0402D] Nunc anima mea turbata est. Et quid dicam? — =
28. [0403A] Pater, clarifica. — =
29. [0403A] Non propter me vox. — ⚠ declined. Vulg. Jo 12:30 Non propter me haec vox venit. Lemma abridgment.
30. [0403A] Nunc judicium est. — =
31. [0403B] Ejicietur. — =
32. [0403B] Et quomodo tu, — =
33. [0403B] si exaltatus fuerit Filius hominis. Quis est iste Filius hominis? — ⚠ declined. Vulg. Jo 12:34 Oportet exaltari Filium hominis?; Migne's first clause instead echoes Jo 12:32 si exaltatus fuero a terra. The gloss argues that the crowd remembered his habitual self-designation rather than this sentence, and that argument stands on either wording.
34. [0403B] Ambulate. — =
35. [0403C] Nescit quo vadat. — =
36. [0403C] Dum lucem habetis. — =
37. [0403C] Ab eis. — =
38. [0403C] Ut sermo Isaiae impleretur. — ⚠ declined. Vulg. Jo 12:38 ut sermo Isaiae prophetae impleretur. One word dropped from a lemma incipit.
39. [0403D] Non poterant. [n: (AUG.)] — =
40. [0403D] Gloriam Dei. — =
41. [0403D] Jesus. — =
42. [0404A] Qui credit in me. — =
43. [0404A] Et qui me videt, videt et eum qui, — ⚠ declined. Vulg. Jo 12:45 Et qui videt me, videt eum qui misit me; Migne Et qui me videt, videt et eum qui. Word order plus one added et — and videt et Patrem is genuine Vulgate at Jo 14:9, which Migne quotes at 0398D, so contamination in either direction cannot be told apart.
44. [0404A] Qui videt me. — =
45. [0404A] qui credit in me, non credit in me, sed in eum qui misit me, — =
46. [0404B] Ego lux. — =
47. [0404B] Sermo quem locutus sum. — =
48. [0404B] Mandatum. — =
49. [0404C] Et scio. — =
50. [0404C] Quae ergo ego loquor sicut dixit mihi Pater, sic loquor. — =

Chunk 0019 — 37 spans

1. [0404C] Ut transeat. — =
2. [0404D] In finem dilexit. — =
3. [0404D] Et coena facta. — =
4. [0405A] accepit vestimenta, — =
5. [0405A] Surgit a coena. — =
6. [0405A] Et ponit vestimenta. — =
7. [0405A] Deinde mittit. — =
8. [0405B] Dicit ei Petrus, — =
9. [0405B] Respondit. — =
10. [0405B] Dicit ei Simon. — =
11. [0405B] Qui lotus. — =
12. [0405B] Qui lotus. — =
13. [0405C] Et cum recubuisset, iterum dixit eis: Scitis quid fecerim vobis? Vos vocatis me, — ⚠ declined — the divergence is a COMMA. Vulg. Jo 13:12 cum recubuisset iterum, dixit eis; Migne cum recubuisset, iterum dixit eis, which moves iterum from the reclining to the saying. Punctuation follows the plate (Pattern 8): rendered "when he had sat down, he said to them again". Migne's marks are his editors' work, and recording them is the point. No marker.
14. [0405D] Non de omnibus. — = Jo 13:18. ⚑ Migne's roman gloss beside it reads qui manducat meum panem where Jo 13:18 has mecum panem — outside the italic, so outside 7a″. Logged, no marker.
15. [0405D] levabit contra me calcaneum suum, — =
16. [0405D] Amodo dico vobis. — =
17. [0405D] Amen, amen dico vobis. — =
18. [0406A] Cum haec dixisset Jesus — = Jo 13:21 — ⚑ the span carries no terminal stop in Migne, and none was supplied (Pattern 8).
19. [0406A] Et cum intinxisset panem. — =
20. [0406B] Dedit Judae Simonis Scariothis, — ⚠ declined. Vulg. Jo 13:26 Judae Simonis Iscariotae. Scariothis is a widely attested medieval form of the name; Pattern 9 — translate normally, no italics, no marker.
21. [0406B] Quod facis. — =
22. [0406B] Quia loculos habebat. — =
23. [0406C] Erat autem. — =
24. [0406C] Nunc clarificatus est Filius hominis. [n: (AUG.)] — =
25. [0406C] Non estis mundi omnes, — =
26. [0406C] usque ad — =
27. [0406C] Filioli. [n: (ID.)] — =
28. [0406C] usque ad — =
29. [0406C] Quo ego vado vos non potestis venire. — =
30. [0406D] Mandatum novum do. [n: (AUG.)] — =
31. [0406D] Diliges proximum tuum sicut teipsum, — =
32. [0406D] usque ad — =
33. [0406D] Ut diligatis invicem sicut dilexi vos. — =
34. [0406D] Domine, quo, — =
35. [0407A] Sequeris autem postea. — =
36. [0407A] Animam meam. — =
37. [0407A] Non cantabit gallus, — = ---

(d) THIS RANGE'S OWN CRUCES, BY COLUMN

Nothing below is marked in the running text unless it says so. Where I name a conjecture I am naming a conjecture, not an attribution.

0394D — vitium propter naturam inolevit

Real, well-formed words that read oddly: the idiom the sentence wants is in naturam inolevit, "grew ingrained into nature," which is what makes the clause explain hereditary blindness. propter naturam says "by reason of nature," which is close enough that the argument survives and different enough to notice. Rendered as printed. No marker; I do not know whose it is.

0396A — VERS. 39: the antithesis of Jo 9:39 collapses, and the same gloss serves both halves

Migne prints: Ut qui non vident. Id est, qui se videre putant, quaerunt ut videant, et qui vident, id est, qui se videre putant, et medicum non quaerunt…

The verse is an antithesis — ut qui non vident videant, et qui vident caeci fiant — and the gloss glosses both of its halves with the identical phrase. The second occurrence is right (qui vident = those who think they see); the first is glossing qui non vident with the description of its opposite. As printed, the English says the same thing of both members.

⛔ Not marked as dittography. Pattern 11 governs a run the plate sets twice adjacently; here the two copies are separated by a full clause each and each is syntactically at home. So this is either an eye-skip in Migne's copy-text, a compositor's, or a loss in transcription, and I cannot tell which — which is exactly why 11a's "mark the intrusive copy" cannot be applied: under 11a I would have to identify an intrusion, and the first copy is not intruded matter, it is a wrong gloss standing where a right one belongs. Rendered as printed; the broken antithesis stays visible on the page, which is the evidence.

⚑ The same sentence has no apodosis: caeci fiant is never supplied, so the et qui vident… member trails off into et hoc est judicium. Rendered elliptically, as printed. Best plate candidate in chunk 0015.

0396D — the wolves persevere too: sed in fine perseverant twice

Praedestinatae oves … in fine perseverant, e contra lupi ad tempus audiunt, sed in fine perseverant. The sheep's clause and the wolves' clause end in the identical three words, so e contra ("on the contrary") contrasts nothing at all and the sentence asserts that the reprobate persevere to the end. Every word is well formed. Conjecture: the wolves' verb wants a negative, or pereunt / deficiunt. Rendered as printed, no marker — a plate read settles whether Migne's page says this. ⭐ This is the strongest single plate candidate in my whole range, and it is a sense defect with no lexical trace, so nothing mechanical will find it.

0396D–0397A — Nec talis dicendus est mercenarius, ideo lupo veniente non resistit

The ideo has no protasis, and Migne's comma after mercenarius leaves the two clauses merely juxtaposed. It does not construe as printed. Rendered as closely as the printed words allow, with the comma kept (Pattern 8). No conjecture I would put my name to.

0396C — quia si aliud, praedicat, non audiunt oves

Migne sets a comma inside the protasis, between aliud and praedicat. English word order cannot carry a comma at that point without producing nonsense ("because if something else, he preaches, the sheep do not hear"), so the English reads "because if he preaches something else, the sheep do not hear." ⚠ Recorded as a deliberate departure from Pattern 8's letter: a medial comma whose position has no English counterpart. Pattern 8's "render as printed and log a crux where a printed mark makes the English genuinely unreadable" is the clause I am invoking. Nothing is added or removed; only the comma's position is not reproducible.

0397A — Unde post et abundantius habeat vitam et pascua inveniet

Subjunctive and future indicative coordinated under one et … et. No English exponent for the mismatch in a way that shows; rendered as printed ("and may have life more abundantly, and … shall find pastures"). Pattern 9 family. No marker.

0399B — impieretur: a non-word, carried, NOT marked

Ut quatriduum impieretur. There is no such Latin verb; impleretur ("might be fulfilled") is near-certain, and l→i is one of the commonest OCR shapes — which is an argument that an error exists and no argument at all about whose it is (§2). No [sic:] fired, per the brief.

⚠ How it is rendered, and why the handling needed a decision. The carried run was the clause's only predicate, so 7a⁗ question 1 bites: quarantining it (or here, merely carrying it) leaves the English verbless. Since no marker is available to hold the sense beside it, the English reads:

That the four days might impieretur [be fulfilled].

The printed run stands in italics, untranslated, per Pattern 7; the supplied sense sits in square brackets, which is the corpus's existing convention for our own supplied English ([it], [them]), so nothing on the page claims the bracketed words are Migne's. Flagged for the merge in case the plate read prefers to convert this to a [sic:] with the sense beside it in the 7a⁗ form.

0400A — VERS. 28: non sequeretur, a singular where the paragraph needs a plural

Quia si audissent Judaei eam vocatam fuisse a Jesu, non sequeretur, et ita non viderent miraculum. The very next verb is plural and the next sentence is secuti sunt — the paragraph plainly means the Jews would not have followed. The printed verb is singular, and the only singular subject available is Mary (eam), of whom the statement makes no sense.

Rendered with the printed singular ("she would not follow"), and declined for [cj:] under Pattern 18a: a person/number anomaly that merely fails to construe stays declined, and the English here does not assert something false so much as something incoherent, which 18a's own examples (suscepit for suscepi) put outside the bar. No marker. ⚑ Logged as a mechanism-5 site because the contextually easier reading is the plural — which is precisely why I did not take it.

0400C — Sea: a non-word inside a lemma, carried, NOT marked

Detailed in §(c). Sea for sed at Jo 11:42. Carried untranslated inside the italic span (nested italics are not expressible, Pattern 10). Nothing translated stands under the missing conjunction, so no force is quarantined.

0401A — Et venient Romam, with Romani printed correctly four lines above

Migne's VERS. 48 lemma prints Et venient Romani (= Jo 11:48); the second lemma, in the same column, prints Et venient Romam. The gloss beside it is about the loss of place and nation, and turns on neither.

⛔ I decline to attribute this, and the reason is the interesting part. Romam is a real Latin word; ni→m is a textbook OCR fusion; and the correct form standing a few lines earlier is the exact reasoning §2 records as a failed attribution test on 9004 ("the correct form prints eleven words later"). All three observations together establish that the two readings differ and nothing whatever about which of Migne's type or Corpus Corporum's transcription produced the second. Rendered as printed: "And they will come to Rome." No marker.

0401B — hoc tantum Caiphas

Elliptical; the verb is not printed. The English supplies it in square brackets, "this only did Caiphas [say]", per the supplied-English convention. No marker.

0403B — qui olim a paucis etsi extra tentetur, non laedit intus

Does not construe: a paucis has no verb to govern it, and the passive tentetur has no expressed subject in a clause whose sense wants an active ("though he tempt from without"). ⛔ I did not render tentetur actively — that is Pattern 7a's mechanism 1, a word translated as something other than what it means to rescue a sentence. The English reads "who of old by few, even though he be tempted without, does not hurt within," which is as close as the printed words allow and is visibly broken. No marker; no conjecture I would defend.

0404C — ut est nata veritas ex Pater

ex governing a nominative. A real word in the wrong case — and English has no exponent for the case, so "born of the Father" renders it in place and the defect is invisible in the English. That is the shape Pattern 12's second test was written for (a defect that renders only by hiding itself), and under the brief's rule it still takes no marker, because I cannot say whether the e is Migne's sort or the transcription's. ⚑ Named here explicitly as a marker candidate for the plate read, since it is the one site in my range where the English alone will never show a reader that anything is wrong.

0405A — Praetermisit altitudinem, suppositurus ejus humilitatem

Mechanism 5. ejus can attach to Christ or to the altitudinem just named. I take it as Christ's own lowliness, on the strength of the clause that follows in the same sentence — cum tam altus sic se humiliaverit — and the English carries that reading: "being about to place his lowliness beneath it." Recorded so that the choice is visible and reversible.

0399C, 0400B, 0405D — three quotations whose divergence sits OUTSIDE the italics

Migne quotes scripture in roman type within the gloss as well as in italic lemmata. Three of those roman quotations diverge from the Clementine: non fiet tibi istud for Mt 16:22's non erit tibi hoc (0399C); Ubi est Adam? for Gen 3:9's Ubi es? (0400B, though this one is italic); and qui manducat meum panem for Jo 13:18's mecum panem (0405D). None is a lemma, so none falls under 7a″ and none is marked. Logged because a later 7a″ pass that reads only the italic spans will not see them, and meum for mecum is a one-letter change that turns "eateth with me" into "eateth my bread" — a real difference of sense in a verse about betrayal at table.

Not reported, per brief §6

No column-band gap, no Ibid. antecedent, no noteCount mismatch, no Pattern 4 [f:] tail, no Pattern 9 item is raised above as a defect. The <LACUNA> at 0416D is outside this range. No bracketed Forte occurs in these five chunks.

Summary of markers fired

Plate-read debt created by this stint: three columns — 0397C and 0399C. The seven sites named above as candidates (0396A, 0396D, 0399B, 0400A, 0400C, 0401A, 0404C) carry no marker and create no debt; they are handed to the plate read as questions, which is what the brief asks.

<!-- ===== cruces-0020.md ===== -->

PL 114, cols 0407A–0418A — CAPUT XIV–XVII, the Farewell Discourse and the High Priestly Prayer (John 14–17). Stint 5 of six.

(a) THE POLICY I APPLIED — stated before the findings, so it can be argued with

1. Attribution. I made no claim about whose any defect is. Where the printed Latin is odd, broken, or a non-word, I rendered or carried it, logged it by column, and named both candidates (Migne's type / Corpus Corporum's transcription) without asserting either. I fired zero [sic:]. That is deliberate and it is the brief's §2: on 9004 the stints' class calls were right 5 times and wrong 9, and on 9001 seven of eight [sic:] fell at the plate. My internal arguments here read as decisive too, and that is precisely the reason not to trust them. Every one is written out below so the plate read can settle it.

2. Markers actually fired — 4 in 4 chunks.

Every [var:] sits on a quoted clause and cites sources/vulgate/clementine-flat.txt by book chapter:verse — the class the brief's measured prior says stands (46/46 on 9004, 19/19 on 9001). The one [cj:] is on a lost negation, which is Pattern 18a's ratified ✅ class (11064 @1090B apud non for apud nos is its mirror image); it is additive, asserts nothing about whose the defect is, and the printed word keeps its English beside it.

3. Pattern 7 / 7a. Every non, nec, nisi, nullus, nihil the Latin prints is in the English and none is in the English that the Latin does not print — including at 0409B, where carrying the plate faithfully leaves the sentence arguing for a distinction it does not draw (logged below, marker declined). Printed punctuation is carried, including the comma inside de eo quod, videtur (0410B), the full stop inside est vitis. tamen (0410D), and the comma in quare debeant, eum in hoc imitari (0411D).

4. Mechanism 5 was the live risk and I checked case and gender, not sense, at every site where a pronoun, relative or participle could bind two ways. The Trinitarian sites are worked out one by one in §(d); where I made a construal, the English carries the construal I state (the brief's §4 closing warning — 9001 @0199A recorded one ruling and shipped the rival).

5. Inherited conventions (brief §1). turba does not occur; multitudo (0407C) → multitude. daemonium does not occur; diabolus (0410B) → devil. mysterium does not occur; sacramentum (0410C) / sacramentis (0415C) → sacrament. natio does not occur. The commentator's own voice is not archaised — dicit/ait → says; the -eth/-est forms in these chunks all sit inside italic lemmata (saith, believeth, abideth, hath, knowest, speakest), which is the italic-membership test. Pattern 17: thou renders the singular address to Philip at 0408A–B and to the Father throughout CAPUT XVII; you renders the plural apostles. No inherited convention broke on this range.

⚑ One convention judgment worth naming: gens sancta at 0409B → holy nation (singular gens of a people, brief §1.4), and gentibus praedicando at 0416A → Gentiles, because CAPUT XV has just set the Judaei/mundus opposition expressly (Hic expressius Judaeos ostendit, 0411D). Both are the settled axis, not a departure.

6. Sacred markers. Column anchors, [n: …] notes and ## heads reproduced verbatim, in order, in place; counts checked mechanically against each Latin twin before writing. λόγος at 0417A is reproduced in Greek, verbatim. The <LACUNA> at 0416D is the settled false positive of brief §6.6: our Latin reads ut sint unum sicut et nos, haec est pax illa continuously and no marker is owed there — I confirmed the text is whole and left it alone.

(b) SPAN COUNT AGAINST THE BRIEF'S 187 — I RAN IT, AND IT AGREES

data/briefs/8999-lemmata-0020.txt declares 187 italic spans for chunks 0020–0023. I counted the … spans in my own Latin files:

187 = 187. No head- or tail-slippage in my range. (The one span that straddles a column anchor — Animalis [0414C] autem qui non percipit… — is counted once, as the brief counts it.)

⚠ Two spans that the ENGLISH adds, deliberately, and they are the only two. English 0021 has 39 spans and English 0022 has 52, against 38 and 51 in the Latin. Both extras are non-words carried untranslated in italics under Pattern 7 — liguis (0410D) and interrogate (0415A). Neither carries a [sic:]. These are the only italic spans in my four chunks that do not answer 1:1 to the Latin, and a merge that finds the counts unequal should read this line rather than "repair" them. Anchors and notes are 1:1 in all four chunks.

(c) THE LEMMA DIVERGENCE LIST — every span, in order, with its verdict

How to read it. ✓ / ⚠ / 1w are the brief's pre-scan marks, reproduced. The verdict is mine, formed by reading the printed Latin word by word against the Clementine at the verse being glossed — ✓ was treated as no clearance and ⚠ as no finding. Every divergence I DECLINED carries its reason. The 151 entries reading "Conforms" were each walked; where a span is a lemma cut (Migne addresses a verse by its opening words and stops), conformity means the words he does print agree with the verse he is addressing.

(d) THIS RANGE'S OWN CRUCES, BY COLUMN

Sites are grouped by what they are. Nothing in §d.1 or §d.2 carries a marker, and the class claim in every heading is a description of the shape, never of whose the defect is.

d.1 — Non-words and broken forms: LOGGED, CARRIED, NOT MARKED

These are the three sites I would have fired a [sic:] on under the older habit. Brief §2 says log and hand them to the plate, and that is what this does. I do not know whose any of them is. All three are OCR-plausible and compositor-plausible; the plate settles it, I cannot.

• 0411C — Haec manao. Jo 15:17 reads Haec mando vobis, and the gloss immediately after (Hic ostendit quis sit fructus…) is about the commandment, so mando is what the sentence needs. manao is a form of nothing (mano, manare would give mano). Carried untranslated inside the italic lemma — Pattern 10's nested-italics rule, so it takes no second italic marker and adds no span. A reader meets These things manao, etc. and the facing Latin shows the same. ⚑ Note against a familiar argument: the correct form mandatum prints at 0409A and 0411B, and praecipit twice in the same column — which is exactly the "the right form prints eleven words later" reasoning the brief identifies as a failed attribution test. It proves the error exists. It proves nothing about whose it is.
• 0410D — caeteris liguis. ligonibus (mattocks) is the only word liguis is near, and the sentence is about cut vine-shoots being worth less than other woods (lignis) because they serve no use but fire. liguis is a non-word. Carried in italics, no marker; this is one of the two spans the English adds (see §b).
• 0415A — quae ipsi interrogate volebant. interrogate is a real form — a second-person plural imperative — with no grammatical slot: volebant governs an infinitive, and the clause is third-person narrative about the disciples. Under Pattern 12's own test this is the class that cannot render in place; under brief §2 it takes no marker. Carried untranslated in italics. The second added span.

d.2 — Real words that read oddly: RENDERED AS PRINTED, NO MARKER

Each of these makes a sentence that limps. Under Pattern 7 the English limps with it, and under Pattern 18a's ❌ boundaries none of them reaches the [cj:] bar — the printed word does not make the sentence assert what the author is arguing against; it merely makes it awkward.

• 0409B — the distinction with nothing to distinguish. ⭐ The most interesting decline in the range. Migne prints Audito quod mundus modo videbit eum, sed ipsi videbunt … quaerit Judas causam hujus discretionis. Jo 14:22 is Judas asking why he will manifest himself to them et non mundo, and the gloss at 0408D has just said mundus … post modicum non videbit. With no negative here, the two halves of the contrast say the same thing and discretio has no work to do. I declined the [cj:], and the reason is that a construal survives: modo can be read "only" rather than "now" — the world will see him only [for a while], they always — which restores the distinction without emending anything. Where a plausible construal exists, 18a's ❌ ("merely awkward") governs. Rendered "will presently see him". Plate read wanted at 0409B.
• 0414A — quo clausus non videbo. The parallel eight words later is quo resuscitatus videbor ("in which, raised up, I shall be seen"), so the active videbo reads as a slip for passive videbor. Declined: 18a scopes the ruling to negations and correlatives and leaves voice/person anomalies that merely fail to construe declined (its @1154C suscepit precedent). Rendered "I shall not see".
• 0414D — Redit ad patriam. The three neighbouring clauses are Reliquit mundum … Vadit ad Patrem … Reliquit mundum, and Jo 16:28 is vado ad Patrem, so patriam reads as a slip for Patrem. Declined — "He returns to the fatherland" is coherent and asserts nothing the glossator is arguing against. Rendered as printed.
• 0416A — Quod hic est ex Patre, in futuro erit plene. Against plene the expected correlative is ex parte (1Cor 13:9–10), and the whole paragraph is about knowledge partial now and full hereafter. Declined: ex Patre construes on its own terms (what we have here we have from the Father), and 18a's ratified correlative extension is narrow — a correlative whose answering member is printed and whose own member is not. Here a whole different word would have to be supplied. Named as a plate-read candidate, because Patre/parte is a shape a scan mangles and a compositor mis-sorts equally.
• 0411A — similitudo dans eis aeternam aeternitatem. "Eternal eternity." Both words are real and the phrase is pleonastic, not false. Rendered as printed; no marker.
• 0415D — id est populus innotescat. Intransitive innotescat with populus as subject gives "that the people become known", where the argument wants the Father becoming known through the people. Rendered literally; no marker.
• 0411B — de dilectione proximo. Dative where proximi is expected; construable as a dative of reference and rendered "love toward the neighbour". No marker.
• 0409D — singulas personas … quos. quos is masculine against feminine personas. Pattern 9: Latin grammatical gender has no English exponent — "whom" is the same word either way, there is no emendation to make or to undo. Logged so a later sweep does not churn it.
• 0410D — Christus est vitis. tamen hanc gratiam… A full stop mid-sentence, followed by a lower-case tamen. Pattern 8: never delete a printed mark. Carried into the English exactly ("is the vine. yet he would not give"). The 9001 precedent at 0202D is the reason I do not guess: there a period at a line break turned out to be the digitization's, and the stint that carried it faithfully and said it could not tell whose it was was right to.
• 0410B — de eo quod, videtur. A comma between relative and verb. Carried.
• 0411D — subdit quare debeant, eum in hoc imitari. A comma splitting the verb from its infinitive. Carried.
• 0412D — the head address prints VERS 1.--, without the period that every other address in these four chunks prints (VERS. 1.--). Reproduced exactly as printed.
• 0410B — VERS. 28.-- Credatis. The address repeats VERS. 28, already used at 0409C for Jo 14:28, while credatis stands at Jo 14:29. Reproduced as printed; brief §6.2's family (never correct a printed reference).

d.3 — Mechanism 5: the reference that could bind two ways

The brief names this as the dominant defect class (5 of 6 on 9001, 5 of 8 on 9004). These are the sites in my range where it was live. In each case I checked the morphology, and the English carries the construal stated here — not the other one.

• 0412C — Spiritus procedit a Patre et Filio, sed Filius dicitur a Patre procedere. The tempting reading makes Spiritus the subject of dicitur ("but he is said to proceed from the Father"), which is the smoother theology. The morphology forbids it: Filius is nominative and dicitur is passive, so the Son is the subject. The sentence is then Jo 8:42 (ego ex Deo processi), and it sets up the conclusion exactly: A quo ergo habet ut sit, ab illo habet ut de illo procedat Spiritus — the Son has from the Father both his being and the Spirit's proceeding from him. English follows the printed nominative.
• 0410C — licet non de suo, nec per ipsos, sed per meipsum. suo is reflexive and would ordinarily bind the subject (the husbandman, i.e. Christ) — but sed per meipsum then contradicts it. It must bind palmites. Rendered "though not out of what is their own". Logged because the case alone does not settle it; the meipsum contrast does.
• 0411C — eis tollit nomen servi, qui est ex timore foras mittendo, qui nescit… Two relatives, two antecedents. The first qui is nomen servi (the name that comes of fear, cast out by charity — 1Jo 4:18); the second is the servant of Jo 15:15 who does not know what his lord does. English binds them that way.
• 0408A — ostendit ex se noto nosci Patrem. ex se noto is ablative absolute on the Son, not on the Father: "from himself, once known, the Father is known."
• 0409A/0409B — et quia Pater, tunc ego diligam eum. The elided verb of quia Pater is diliget, from the clause immediately before; supplied in brackets, not silently.
• 0416B — Non enim quia Filio dedit Pater amisit. Pater is the subject of amisit, not of dedit alone: "the Father did not lose because he gave to the Son."
• 0417D–0418A — quia nos sumus ille, cum secundum hominem ejus capitis sumus corpus. ejus capitis is genitive depending on corpus; rendered "we are the body of his head", not "his body, the head".
• 0416D — ut sint unum in natura sua, sicut et nos in nostra sumus unum. The whole point is that the disciples' unity is in their own nature, not with God's; the negatives that follow (non ait, ut sint unum nobiscum) are carried in full. This is the site the brief warns about in kind — a mis-binding here would make the creature consubstantial with God.

d.4 — Notes and citations: passed through, never corrected

• 0414C — [n: (I Joan. I)] on a quotation of 1 Cor 2:14. Reproduced verbatim. The [var:] fired there cites the correct verse in our voice; Migne's siglum stays his.
• 0409C — [n: (Isa. XXVII)] on Pacem super pacem, which is at no verse of Isaiah 27 (see span #47). Reproduced verbatim, uncorrected, unmarked.
• 0409C — [n: (I Cor. IV)] on a clause conflating Hbr 4:13 with 1Cor 4:5 (span #48). Same treatment.
• 0408B — the THEOPH. gloss quotes Jo 14:16 as alium Paracletum mittet vobis, where the Clementine reads alium Paraclitum dabit vobis. This is in the roman gloss, not an italic lemma, so it is outside the 7a″ inventory and takes no [var:]; logged because the verb differs, and because the same Paracletus spelling recurs at 0409C and 0412C — Migne's usage across the work, not a local slip.
• 0416A — the gloss paraphrases Jo 17:2 as potestatem omnis creaturae, where the verse reads omnis carnis. Again roman gloss, not lemma; logged, no marker. The lemma of the same verse (span #156, ut omne quod dedisti ei det eis vitam aeternam) conforms exactly, which is what makes the gloss's word visible as the glossator's own.

d.5 — The <LACUNA> at 0416D: confirmed closed, no marker owed

Brief §6.6. Our Latin reads Gaudium quod supra expressit, ut sint unum sicut et nos, haec est pax illa et beatitudo in [0417A] futuro saeculo — continuous, with nos present and the column break falling where the patch says it does. I checked the text at that point rather than taking the brief's word for it, translated it as whole prose, and fired nothing.

PLATE READ — the columns this range hands forward

In priority order, and with what each would settle:

1. 0411C — Haec manao: is the non-word Migne's or Corpus Corporum's? If Migne's, a [sic:] is owed and the English carrying is already correct; if ours, a data/tei-patches/8999.json entry and the lemma reads "These things I command".
2. 0410D — caeteris liguis, same question, and the full stop in est vitis. tamen on the same column (the 9001 @0202D shape: a stop that may belong to the digitization at a line break). Read the whole line, not the word.
3. 0415A — interrogate, and on the same column nobis for non, where I fired the one [cj:]. If the plate prints non, the [cj:] comes out and the sentence is repaired at source; if it prints nobis, the [cj:] stands and the column is Migne's own.
4. 0409B — the missing negative in mundus modo videbit eum. My decline rests on a construal of modo; the plate may print non videbit and end the question.
5. 0416A — ex Patre / ex parte.
6. 0414A — videbo / videbor; 0414D — patriam / Patrem. Both low priority: real words, coherent as printed, declined on the rule rather than on doubt.
7. The three [var:] columns — 0414C and 0416D — required by the ship test (plate-gate.mjs) before this work can be marked ours, whatever the reading turns out to be.

<!-- ===== cruces-0024.md ===== -->

Stint 6 of six, and the last range of the work. Chunk 0027 closes at 0426B with Nec ipsum. Non dicit spatio non posse capi, sed capacitate legentium…, which is the last sentence of the work; the brief records that the plate (leaf n213/n218 of patrologiaecursu0114mign) shows Actus Apostolorum opening below a rule on the same leaf. Nothing is missing at the end and no [ed:] is owed there.

(a) THE POLICY APPLIED — stated before the findings, so it can be argued with

A1. Attribution: no marker on any defect whose owner I cannot name

Per the launch brief §2 and CLAUDE.md 8a. My guess about whose a defect is carries no evidential weight, and I have not offered one. Concretely:

• A real, well-formed Latin word that merely reads oddly is rendered as printed, logged in §(d) below, and takes no marker — however good the internal argument. Sites: liberatum (0420B), magistri (0420B), Separatur (0422A), Magdalenae/vidit (0423A).
• A non-word gets the same treatment: carried into the English untranslated, in italics under Pattern 7/10, logged, and not marked. One site in this range: undet (0423B). I fired zero [sic:] across four chunks. Two of 9001's five stints did the same and the merge recorded it as the right call; that precedent is what I followed.
• Where a defect could be Migne's type or Corpus Corporum's transcription, I say so and assert neither.

A2. [var:]: fired only where the divergence is bigger than one word, or where the gloss itself corroborates the divergent reading

Three [var:] fired, all on quoted clauses cited from sources/vulgate/clementine-flat.txt. The discriminator I used, and it is the whole of it:

1. Fire where the divergence is a phrase present or absent, or a whole clause of different shape — an OCR letter-swap cannot manufacture that. (0420C Gen. 3:18; 0420D Isa. 53:7.)
2. Fire where the divergence is one word but the surrounding gloss independently expounds the divergent reading — the gloss is then a witness, printed on the same page, that the exemplar carried it, so the divergence is the text's and not a transcription slip. (0420C Jo. 1:29 peccata plural, expounded three times over as peccatorum.)
3. Decline every one-word / one-inflection divergence with no such corroboration — autem for ergo, Jesum for illum, negat for negavit, vidit for vidi, Separatur for separatim. This is the class CLAUDE.md names as unfalsifiable from our files: the judicabit → judicavit case, where the digitization manufactured a Vulgate divergence and a collation then dutifully found it. Every one is logged in §(c) with its reason.

I have not claimed that a declined divergence agrees with the Vulgate; I have claimed only that I cannot tell our file from Migne's plate at that site.

A3. Negation and mechanism 5

Every non, nec, neque, nisi, nullus, nihil, ne, minus the Latin prints is in the English, including where carrying it reads awkwardly. Chunk 0027's closing pile-up (Jn 21:23–25) is carried in full: non sit mortuus → "did not die"; Non enim magnum erat dare dilecto non mori → "it was no great thing to grant the beloved not to die"; Non spatio locorum … capere non posse → "Not by the space of places … could not contain"; Non dicit spatio non posse capi → "He does not say that they cannot be contained by space". Mechanism-5 checks were made by case and gender at every relative and participle; the four that could have gone the easy way are recorded in §(d).

A4. Inherited conventions (brief §1) — applied from the first line, one exception-adjacent note

• dicit/ait in the commentator's voice → says throughout. Archaism appears only inside italic lemmata (He saith: I am not, Thou shalt deny me thrice, Why weepest thou?), and in the unitalicised Matt. 22:30 tag at 0425B neque nubent neque nubentur, where 9004's ruling leaves it.
• turba → crowd (0418B sine turbis → "without the crowds"). multitudo does not occur in this range.
• diabolus → devil (0420B, twice). daemonium does not occur in this range.
• natio does not occur in this range. gentes: Gentiles in the Judaei/gentes opposition (0418B converte ad gentes against Mitte … ad Israel; 0419D a Judaeis traditum gentes interficerent; 0421C Judaeos … gentes significat; 0424A fidem gentium against Thomas who saw); nations where plainly generic (0421D gentium sapientes, omnibus gentibus imperantes — three tongues for three peoples, no typology). gentilium at 0420A is a fourth word and takes Gentiles in the Judaei/gentiles pairing; gentis alienae at 0421A, of a single foreign people, takes nation.
• mysterium → mystery, sacramentum → sacrament, 1:1. ⚑ This range settles it mechanically for the merge: both words occur, and close together. 0421C prints grande mysterium piis; 0422B prints sacramenta Divinitatis, 0422C cujus sacramenta, 0424B tanta … sacramenta. They cannot share an English word. I record that "the sacraments of the Godhead" (0422B) reads strainedly where "mysteries" would be idiomatic, and I did not depart — this is not an exception, it is the convention doing its job.

A5. One drift axis the merge must settle, and it is not mine to settle

usque ad as the recension's abridgment marker. The three chunks of 8999 already shipped are split 2 : 1 — 0000 and 0005 render it down to, 0001 renders it as far as — while the sibling book 9001 (Mark) uses as far as throughout. I used down to (within-work majority, brief §1's "one voice within a work"), at four sites: 0421B, 0425B, 0426A ×2. Flag, counts and both candidates are here so the merge can flip it with one sweep either way.

(b) SPAN COUNT AGAINST THE BRIEF'S TOTAL — RUN, AND IT AGREES

data/briefs/8999-lemmata-0024.txt states 113 spans for chunks 0024–0027.

I counted the … spans in the body of my own four Latin chunks (frontmatter excluded):

113 = 113. No head-of-range or tail-of-range slide. The list in §(c) walks all 113 in the brief's order. English italic parity holds 1 : 1 in three chunks; 0026 carries one extra italic run in the English (31 vs 30) and that is the Pattern 7 carry-through of the non-word undet at 0423B, not a lemma.

(c) THE LEMMA DIVERGENCE LIST — every span, its verdict, and the reason for every decline

Verdicts: ✓ conforms = the printed Latin matches the Clementine at the verse being glossed (punctuation, capitalization and lemma truncation ignored — a lemma is a cue and stops where Migne stops). ▲ fired = a [var:] stands in the English. ○ declined = a real divergence, logged, no marker, reason given.

Chunk 0024 (0418A–0420B) — 31 spans

Declines in 0024:

• #8, 0418B, Mt 26:52. Migne: Omnes enim qui gladium acceperint gladio peribunt. Clementine: omnes enim, qui acceperint gladium, gladio peribunt. Word order only, and Migne also drops the Clementine's commas. No word is added, removed or changed, and the gloss that follows (Quo gladio?) turns on the sword, not on the order. Not material; a [var:] here would be apparatus noise on a page that already carries three.
• #9, 0418B, Jo 18:12. Migne: Cohors autem. Clementine: Cohors ergo. A one-word connective inside a two-word lemma cue, of exactly the class an OCR pass can manufacture and no internal evidence can attribute. Declined per A2.3. ⚑ Note the gloss does not touch the connective, so there is no corroboration either way.

Chunk 0025 (0420C–0422A) — 28 spans

The three fired, with their evidence:

• ▲ #34, 0420C, Jo 1:29. Migne: ecce qui tollit peccata mundi (plural). Clementine: ecce qui tollit peccatum mundi (singular). One word, and therefore inside the class A2.3 declines — except that the gloss expounds the plural three times over: the lemma sits inside In spinea corona nostrorum notatur susceptio peccatorum, and is followed by Spinae namque in peccatorum significatione ponuntur. The exposition is built on the plural, on the same page, in Migne's own type, which is independent evidence that the exemplar carried it. Fired: [var: Vulg. Jo. 1:29 has the singular, "the sin of the world"]
• ▲ #35, 0420C, Gn 3:18. Migne: Terra tua spinas et tribulos germinabit. Clementine: Spinas et tribulos germinabit tibi. A phrase present that the Vulgate does not have, and a word absent that it does* — the shape an OCR pass cannot manufacture. And the gloss is an exposition of terra tua specifically: Id est, conscientia tua punctiones … procreare non desistet.* The gloss explains nothing if the Vulgate's wording is restored. Fired: [var: Vulg. Gen. 3:18 has no "thy earth," and reads "shall bring forth thorns and thistles to thee"]
• ▲ #41, 0420D, Is 53:7. Migne cites (Isa. LIII) and prints Sicut agnus coram tondente sine voce, sic non aperuit os suum — which is not Isaiah's Clementine wording (quasi agnus coram tondente se obmutescet, et non aperiet os suum) but, to the word, the form in which Acts 8:32 quotes it. Whole-clause divergence, plus a tense shift from future to perfect; the gloss then argues from the silence as an accomplished fact (in suo silentio non reus sed innocens habeatur). Fired, naming both witnesses: [var: Vulg. Isa. 53:7 reads "and as a lamb before his shearer shall be dumb, and he shall not open his mouth"; the wording printed here is that of Act. 8:32] ⚑ I do not assert why they differ (Pattern 14 forbids it): an Old Latin reading, the author quoting Acts from memory, and Migne's copy-text are all open.

Declines in 0025:

• #44, 0420D, Jo 19:14. Migne: Erat autem parasceve. Clementine: Erat enim parasceve Paschae. Two differences, both innocent: a one-word connective (A2.3) and the ordinary truncation of a lemma cue before Paschae. Truncation is how this recension prints every lemma and is never a divergence.
• #50, 0420D, Jo 19:16. Migne: Tradidit eis Jesum ut crucifigeretur. Clementine: tradidit eis illum ut crucifigeretur. A proper name for a pronoun with the same referent: no change of sense, and the gloss that follows turns on eis and on crucifigerent vs crucifigeretur, never on the object. ⚑ Note that the gloss's own paraphrase eight words earlier prints tradidit eis Jesum ut crucifigatur — present subjunctive against the lemma's imperfect. Both stand as printed; I make no claim about whose the difference is.
• #58, 0420D, 1Tm 5:8. Migne: Si quis suorum curam non habet, fidem negat. Clementine: Si quis autem suorum, et maxime domesticorum, curam non habet, fidem negavit, et est infideli deterior. The omissions are lemma abridgment (the span carries Migne's own etc. immediately after it), and the residue is a one-word tense shift perfect → present. Declined on both counts; the English reads "denieth the faith", the present that Migne prints, and not the Douay perfect.

Chunk 0026 (0422B–0424A) — 30 spans

(#60–#89 are 30 spans, counting the two re-quotations at #83–#84. The carried non-word undet at 0423B is not a lemma and is not in this list; it is logged in §(d), and it is why the English of 0026 carries 31 italic runs against the Latin's 30.)

Declines in 0026:

• #61, 0422A, Jo 20:7. Migne's lemma is Separatur — a well-formed passive verb, "it is separated". The Clementine reads separatim, the adverb. One word; the sense of the gloss (a nostrae infirmitatis cognitione remota sunt, occultata seorsum) is compatible with either, so there is no corroboration in the direction A2.2 requires. Rendered as printed, "It is separated." I do not know whether this is Migne's type or Corpus Corporum's transcription and I do not guess. The plate settles it in one line.
• #77, 0423A, Jo 20:18 — two divergences, both declined.
• Magdalenae for the Clementine's Magdalene. A genitive where the received text has the indeclinable epithet. Real Latin, one letter, no gloss depends on it; rendered literally, "Mary of Magdalene". ⚑ Note that the gloss on the next lemma but one prints Mariae Magdalenae correctly as a dative at 0423C — which proves an inconsistency exists on the page and proves nothing about whose it is (brief §2: "the correct form prints eleven words later" is the reasoning 9004 measured as a failed attribution test).
• vidit for the Clementine's vidi. One letter, and it converts Mary's direct speech ("I have seen the Lord") into an indirect report ("that she had seen the Lord"). Material in sense, and I have rendered the indirect report Migne prints. But it is exactly the one-letter shape A2.3 declines: the gloss (modo mulier vitam nuntiat viris) reads the same under either. Logged for the plate; no marker.
• #76, 0423A. Not a divergence: the span is Jo 20:16 entire, with the opening words of Jo 20:17 (Dicit ei Jesus) run on to it as the next cue. Migne's lowercase magister against the Clementine's Magister is orthography with no English exponent (Pattern 9). No marker.

Chunk 0027 (0424B–0426B) — 24 spans

No declines and no markers in 0027. Every scripture span in the work's last chunk conforms to the Clementine at its own verse. ⚑ Worth recording because it is the negative that makes the three fired markers credible: the same walk over the same brief returned zero here.

Summary: 113 spans walked · 3 [var:] fired · 7 divergences declined and reasoned · 0 [sic:] · 0 [cj:] · 0 [ed:] · 0 [d:].

(d) THIS RANGE'S OWN CRUCES, BY COLUMN

### 0418C — non quin post sint morituri Rendered "not but that they should die afterwards". Migne's non quin is carried whole; the sentence is deliberately double-negative in Augustine's argument (they are let go, not because they will never die, but because they do not yet believe as the saved believe). No defect.

### 0418B — Illo qui igneus vertitur ante paradisum igneus, masculine nominative, agrees with Illo … qui (the sword), not with paradisum. Rendered "By that fiery one which is turned before paradise". Mechanism-5 site checked by gender; the easy reading would have attached the fire to paradise.

### 0420A — sed ad regnum coeleste credendo venire Anacoluthon. impedio governs dominationem vestram; the sed clause then swings to a bare infinitive with no governing verb printed. Rendered as closely as the printed words allow — "but to come by believing to the heavenly kingdom" — with nothing supplied. No emendation and no marker.

### 0420B — sed ita liberatum est, ut neque se regem neget, nec de mundo se fateatur esse liberatum is a real, well-formed participle and is rendered as printed: "but it is so set free that…". ⚑ I record, and do not act on, the possibility that Migne's text stood at libratum ("so poised/balanced"), which would answer exactly to the same gloss's own formula fifteen lines earlier at 0419A: Ita temperat responsionem, ut nec veritatem taceret, nec defendere se videretur — the identical "neither … nor" shape about the identical kind of answer. That parallel proves an oddity exists. It does not prove whose it is, and under brief §2 it earns no marker: not a [sic:] (the word is real and renders in place), and not a [cj:] (the English asserts nothing the author is arguing against — it is merely odd, which 18a's bar explicitly excludes, and the 18a census is closed). Both negatives are carried.

### 0420B — Id est filius magistri, id est diaboli … Vel Barabbas, id est filius patris Migne glosses Barabbas first as filius magistri and only then as filius patris, which is the etymology (bar-abbas). Rendered as printed, both times. ⚑ The first gloss is corroborated by its own sentence — qui latroni in suo scelere Judaeis in perfidia magister fuit — so whatever it is, it is not a bare slip of a single word; and the recension's Vel signals that Migne is offering two glosses, not correcting one. No marker.

### 0421A — (et si minus illis invidis) Parenthesis, elliptical, and the datives have no governing word printed. Rendered literally, "(and if less so to those envious men)". Logged because it does not construe cleanly and a reader will feel it; nothing is supplied and nothing is marked.

### 0421D–0422A — Quadripertita vestis signat Ecclesiam toto orbe, qui quatuor partibus constat diffusam Mechanism-5 site, decided on gender. qui is masculine and can only be orbe, not the feminine vestis or Ecclesiam; it is the world that consists of four parts, not the garment or the Church. English follows: "the Church spread through the whole world, which consists of four parts". The easy reading (the fourfold garment consisting of four parts) is the wrong one and is the shape 9001's worst site had.

### 0422A — quae in sorte Dei gratia commendatur Genuinely ambiguous: Dei gratia may be ablative ("is commended by the grace of God") or nominative ("the grace of God is commended in the lot"). I took the ablative, which keeps Ecclesia as the subject of commendatur across the relative — the construction the sentence has been running since a quo toto. Logged so the choice is visible; the rival reading is defensible.

### 0422B — sed et mortuum credere noluit Rendered "but refused to believe him even dead", taking et adverbially. This is what makes the antithesis work with the next clause (Petrus Ecclesiam, quae cognovit carne mortuum, viventem credidit Deum): the Synagogue would not believe him even dead, the Church knew him dead and believed him living. The alternative (et for eum) is not needed to construe and would be an emendation; not taken, not marked.

### 0422C — quia longe a nostra nostra is feminine with no expressed noun; from the sentence it is passio. Rendered "because it is far from ours". Nothing supplied.

### 0422C — nec suis oculis, nec oculis illorum duorum credentium putavit putavit takes datives that belong to credere. Rendered "she gave no thought either to her own eyes or to the eyes of those two who believed", which carries both necs and does not substitute a verb Migne does not print. Not repaired to credidit.

### 0423B — Ibi culpa abscinditur undet processit — the one non-word in this range undet is not a Latin word. Carried into the English untranslated, in italics under Patterns 7 and 10 — "There the fault is cut off undet it proceeded" — and no [sic:] is fired, per brief §2 and CLAUDE.md 8a.

• The obvious conjecture is unde ("whence"), which construes at once. A second, and to my eye the better shape, is a fusion of unde + et ("whence also"), which is the class the brief warns is invisible to any mechanical test.
• ⚠ 7a⁗ check, run explicitly. The quarantined run is the clause's relative adverb, which is 7a⁗'s question 3. The clause still has both its verbs (abscinditur, processit) and carries no negation, so nothing false is asserted by the carry; what the reader loses is the connective, and the conjecture is one line away here. I have not silently supplied "whence" in the running text.
• ⛔ I do not claim whose it is. A non-word is not automatically our file's — 9001 @0203A magnitudino is Migne's — and this is precisely a plate question.

### 0423B — et quae tunc verba serpentis modo narrat verba vivificatoris Zeugma: one verb (narrat) serving two clauses in different tenses, with nothing printed for the first. English cannot carry that, so the supplied verb rides in square brackets on the existing supplied-word convention: "and she who then [told] the words of the serpent now relates the words of the Life-giver". Flagged so a later reader can see the bracket is ours.

### 0423B — De qua manu illatus est potus mortis, de ipsa suscipite potum vitae Migne prints no quotation marks here, so the English uses none either and the sentence runs on from ac si ipsis rebus dicat Deus: exactly as the plate has it. Pattern 6: guillemets are his and there are none to reproduce; a supplied " " was considered and dropped as needless.

### 0424B — sacramenta / 0421C — mysterium See A4. Both words occur in this range, three columns apart, so the 1 : 1 convention is forced mechanically rather than by preference. "The sacraments of the Godhead" (0422B) and "such great sacraments" (0424B) read strainedly; that is the convention working, not drift.

### 0424C — sed ad dies, refert Migne's comma falls between dies and refert, splitting the phrase from its verb. Carried (Pattern 8, never delete a printed mark): the English reads "but he refers it to the days:" with the sense-break kept as a colon before the enumeration, which is where his own construction goes. Logged because it is a punctuation judgment and not a free one.

### 0425A — Est ideo vel ad demonstrationem … tenet hominem poena Anacoluthon: Est is stranded with no complement, and the sentence resumes with a finite tenet. Rendered "It is therefore either for the showing of the misery that is owed … that punishment holds a man for a time", which construes the printed words without adding any. No marker.

### 0425A — nec parva videretur culpa, si cum illa auferretur poena The nec is carried: "and the fault would seem no small thing if the punishment were taken away along with it". ⚑ This is a sentence that reads backwards on a first pass and is not a defect: Augustine's point is that a fault whose punishment vanished with it would look trivial. Recorded so a later sweep does not "repair" it.

### 0426A — quanta arcana divinitate prae caeteris esset scripturus divinitate is ablative and has no governing preposition printed; quanta arcana is the accusative object of scripturus. Rendered "how great secret things concerning the divinity he was to write beyond the rest", which is the least-supplied reading available. The rival ("with what divinity") is possible and is logged here rather than in the text.

### 0426B — the end of the work Nec ipsum. Non dicit spatio non posse capi, sed capacitate legentium, quamvis salva fide rerum plerumque verba videantur excedere fidem per hyperbolen. Rendered with all three negations intact. This is the last sentence of 8999 and nothing is missing after it (brief, plate leaf n213/n218); no [ed:] is owed and none is fired.

### Punctuation and layout, carried without comment elsewhere

• 0423A Dixit ergo eis iterum. Pax vobis, — Migne's full stop inside the lemma where the Clementine has a colon. Carried (Pattern 8).
• 0424C Prima, quotiescunque sit facta, uno die resurrectionis. — verbless, carried.
• Em-dashes: zero in all four English chunks.
• Guillemets: zero in the Latin, zero in the English; the two places English wanted a quotation mark (0423B, 0424C) take none, because supplying one would misrepresent the plate and " " would have added nothing a reader needs.

What the plate read should look at first, in this range

In descending order of what a single line of reading would settle:

1. 0423B undet — the only non-word. One line.
2. 0422A Separatur vs the Vulgate's separatim. One word, and it is the lemma.
3. 0423A Magdalenae and vidit in Jo 20:18 — two one-letter questions in one line.
4. 0420B liberatum — is Migne's text libratum?
5. The three [var:] columns, 0420C and 0420D, which the ship test requires read before the work can be marked ours (CLAUDE.md 8a): they are two columns and cover all three markers.