Epistola II B. Petri

Translated 2026-08-18, one translator, 2 chunks (2,058 Latin words → 3,128 English, 1.52×), on Opus. Structure: ARGUMENTUM, CAPUT PRIMUM, CAPUT II, CAPUT III. Apparatus: 2 [var:] · 5 [sic:] · 2 [cj:] · 0 [d:] · 0 [ed:] · 0 [nt:] · 32 [n:] (17 + 15, exact parity with the Latin twins). 22 column anchors (11 + 11), 87 italic spans (39 + 48), 1 question mark in chunk 0 and 0 in chunk 1 — all at parity. Zero guillemets in either language. verify-english.mjs clean.

Conventions follow GLOSSA-NT-AGENT-BRIEF.md and src/english/8968/cruces.md §1 (Ruth, the Glossa pilot): VERS. n.-- passes through verbatim; the lemma is Englished; the attribution sigla ride as ordinary [n:] notes, contents untranslated; usque ad → as far as (8978 §205). Heads: CAPUT PRIMUM. → CHAPTER ONE., CAPUT II. → CHAPTER II., CAPUT III. → CHAPTER III. (corpus-shipped forms).

§0 · PLATE STATUS — foot-of-page apparatus for cols 0689–0694 is a CHECKED ZERO

Read by the orchestrator this session, not by this translator (two translators duplicated this work last batch; it was not repeated). Method: archive.org patrologiaecurs06saingoog, leaf n = (column + 9) / 2, verified against the running head (n349 = cols 689/690, head GLOSSA ORDINARIA. — EPIST. II PETRI.). Leaves n349, n350 and n351 read at full page width across the whole foot. No foot rule, no numbered notes — the columns simply end. Two of the three carry a mid-page work division and were read there: n349 (1 Peter ends, EPISTOLA II B. PETRI begins below the rule, where this work opens) and n351 (this work ends, 1 John's CAPUT PRIMUM begins). That mid-page division is the case that lost three notes on PL 202; it is clean here.

Tier 3, coverage full, notes_recovered: 0. Every conjecture recorded below is therefore ours, not one Migne had already answered in his own apparatus.

Two readings were independently corroborated on the plate and both stand as our files have them: 0690D Non enim indoctas fabulas, with its (BEDA.) siglum attached exactly as printed (and the following gloss's deos appellare non timuerunt / de caetero non divinis), and 0694A Non tardat Dominus. with its gloss's non tardat promissis implendis, quia nondum est tempus. All of those negations are carried.

⚑ Boundary: the plate's col. 694 right-hand column runs on into 1 John (VERS. 5. — Et haec est annuntiatio, carrying three question marks). That is the next work. Our text ends at 0694C and our Latin prints one question mark, which is the number the English meets.

§1 · THE LEMMA DIVERGENCE LIST — ALL 72 ITALIC SPANS CHECKED AGAINST THE VULGATE

Every italic span in both chunks was worked through from the launch list before the English was written, and each was checked word by word against the Clementine Vulgate of the verse it comes from. Agreements are listed as well as divergences, so this can be told apart from a report by someone who only listed findings.

1a · The two that diverge — both marked [var: …]

1b · The seventy that agree with the Vulgate (checked, no marker)

Chunk 0 — 2 Peter 1. Simon. (1) · Servus et Apostolus. (1) · Coaequalem. (1) · In justitia. (1) · Gratia vobis. (2) · multiplicetur; (1 Pet 1:2, and correctly so — the gloss is contrasting the two epistles) · adimpleatur: (2) · in cognitione Dei, (2) · Quomodo omnia. (3) · Maxima et pretiosa nobis promissa donavit. (4) · Ministrate in fide. (5) · In scientia autem abstinentiam. (6) · In pietate autem. (7) · Haec enim si vobiscum, (8) · Caecus est et manu tentans. (9) · Oblivionem accipiens. (9) · Non peccabitis aliquando. (10) · Sic enim. (11) · Et quidem scientes. (12) · Justum autem, (13) · Depositio tabernaculi mei. (14) · Dabo autem operam. (15) · Post obitum, (15) · Sed speculatores facti, (16) · Et habemus firmiorem. (19) · Cui benefacitis attendentes. (19) · Hoc primum intelligentes. (20) · Non enim voluntate humana. (21).

Chunk 0 — scripture quoted from other books, the highest-risk class:

• Matt. V at 0689B — Nisi abundaverit justitia vestra plus quam scribarum et Pharisaeorum, non intrabitis in regnum coelorum. Word-for-word Matt 5:20, only coelorum for caelorum (orthography, Pattern 9). Both negations (Nisi, non) carried into the English.
• Psal. LXXXIII at 0690B — et videbitur Deus deorum in Sion. Ps 83:8 verbatim.
• Psal. II at 0691A — Dominus dixit ad me, Filius meus es tu. Ps 2:7 verbatim except that Migne prints a comma where the Vulgate points a colon. Pattern 8: reproduced as a comma, not silently upgraded. Pattern 17: es tu is singular, so "Thou art my Son".
• Luc. XXII at 0690C — Confirma fratres tuos, Luke 22:32 verbatim; tuos singular → "thy brethren".

Chunk 1 — 2 Peter 2 and 3. Perditionem. (2:1) · Via veritatis blasphemabitur. (2:2) · Si enim Deus angelis. (2:4) · Et civitates. (2:6) · In cinerem. (2:6) · Iniquos. (2:9) · In diem judicii. (2:9) · Ubi angeli. (2:11) · Mercedem injustitiae. (2:13) · Hi sunt. (2:17) · Superba. (2:18) · In desideriis carnis, (2:18) · Si enim, (2:20) · cognitione Domini nostri et Salvatoris Jesu Christi, (2:20) · Hanc ecce vobis. (3:1) · Ut memores. (3:2) · Verborum, (3:2) · a sanctis prophetis (3:2) · apostolorum: (3:2) · Domini et Salvatoris (3:2) · Latet. (3:5) · Unum vero hoc. (3:8) · Sicut mille anni. (3:8) · Non tardat Dominus. (3:9) · Coeli magno. Coeli, (3:10) · Elementa. (3:10) · Novos vero coelos. (3:13) · Satagite immaculati et inviolati. (3:14) · Frater noster. (3:15) · Difficilia. (3:16) · Depravant. (3:16).

Chunk 1 — scripture quoted from other books:

• Luc. XII at 0694B — Beati servi illi quos cum venerit Dominus, invenerit vigilantes. Luke 12:37 verbatim. Checked twice because it is the most familiar sentence on the page and therefore the likeliest place to import a remembered English; it agrees, so the traditional cadence is legitimate here.
• Apoc. VI at 0694A — see §2b.

1c · Divergences examined and DECLINED, with the reason

• 0690B VERS. 12 Vos incipiam semper commonere de his. — Vulg. incipiam vos semper commonere de his (2 Pet 1:12). Migne fronts Vos. Transposition only, no change of sense; Pattern 14 fires on a divergence that changes the sense. Declined, following 8978 §2b (Jude 3's four transposed inner lemmata). Rendered "I will begin always to admonish you of these things."
• 0694A Et dictum est illis, ut requiescerent adhuc modicum tempus… — Apoc 6:11 reads adhuc tempus modicum. Transposition of two words, sense identical, and English cannot even display it ("yet a little time" either way). Declined on the same ground; logged because the launch brief named this span as high-risk and a reader will look here. Everything else in the span — donec compleantur conservi eorum — is verbatim, and Migne stops where Apoc 6:11 continues (et fratres eorum…), which is a truncation, not a variant.
• 0692A VERS. 7 Loth oppressum. — Vulg. justum Lot oppressum. Loth is the ordinary medieval spelling of the name, attested throughout the corpus; Pattern 9, not a defect and not a variant. Rendered "Lot oppressed."
• 0693A VERS. 22 Contigit enim illis. — Vulg. 2 Pet 2:22 Contigit enim eis illud veri proverbii. A synonymous pronoun; no English exponent distinguishes illis from eis, and no sense turns on it. Declined, logged.
• 0693B praecepta — the third-position italic in the VERS. 2 chain. Vulg. 3:2 has the genitive praeceptorum; Migne's praecepta is the gloss re-casting the word into the nominative of its own sentence (et quae praecepta sunt etiam Domini et Salvatoris), not a quotation of the verse. Same class as Jude's In viam Cain abeunt (8978 §2b). Declined.
• 0689B, in the gloss (unitalicized): ut noscatis unum verum Deum, et quem misit Jesum Christum — John 17:3 reads ut cognoscant te solum verum Deum, et quem misisti Jesum Christum. Two divergences, and unum for solum is not nothing. [var:] declined, twice over: the run is not in italics, i.e. Migne does not present it as a quotation, and it has been recast from second person into third to fit ut noscatis. Pattern 14 is reserved for scripture quoted as scripture. Rendered literally, "that you may know the one true God, and Jesus Christ whom he sent." ⚠ Flagged for a blind reader: if the house judgment is that a woven allusion this close to its source should carry a marker, this is the site.

§2 · THE ISAIAH CITATION AT 0694B — Migne's [n: (Isa. LXV)] points at a verse that is not there

The quotation Lux lucernae non lucebit ibi amplius is Apocalypse 18:23, not Isaias. No verse of Isa. LXV — nor of Isa. LX, the chapter that carries the nearest thought (non erit tibi amplius sol ad lucendum per diem) — reads anything like it. The chapter number is plausibly there because the gloss's own preceding sentence is about new heavens and a new earth, which is Isa 65:17, and the compositor (or Bede's source) carried the locator across to the wrong sentence.

Nothing was changed. The note rides verbatim, as every [n:] does. The [var:] marker records only the two readings of the text (ibi against Apoc 18:23's in te), because Pattern 14 says to state the readings and let the crux argue the rest. This is the crux.

The textual divergence itself is real and material in a small way: Apocalypse addresses Babylon in the second person ("in thee"), while Migne's ibi makes it a place-adverb ("there") and so lets Bede apply it to the world that is passing. Rendered "shall shine there no more".

§3 · CARRIED TYPE — five [sic: …], each with its conjecture

None of these is inside a marker that swallows a verb, a negation or a conjunction (7a⁗ questions 1–3 applied to each, and answered before the marker was closed).

No [cj: …] sits on any of these five. Pattern 18's bar is that the faithful English asserts something false on a word that is a real Latin form; every defective run in the table above is a non-word, so [sic:] governs and Pattern 18 is expressly excluded ("anything already inside a marker"). The work's two [cj:] markers are elsewhere, on real forms, and are set out at §3a. Of the remaining real-word oddities, eadem at 0689A and reservant at 0692A are treated at §5 and neither makes the English assert anything false.

§3a · TWO [cj: …], BOTH AT 0692C, BOTH ADDED AFTER THE BLIND READ

Pattern 18: the printed word is a real Latin form, so nothing is broken for [sic:] to wrap, and the faithful English then carries a relation or a polarity the Latin does not. Both are additive — neither changed a word of the translation.

non boni — an intrusive negation, and the reader needs both readings

Vel torquebuntur ibi, id est apud Deum: ubi angeli non boni, non sustinent, sed vindicant si quid exsecrabile contra illos dicitur.

Rendered, unchanged: "where the angels, not good ones [cj: non boni; read boni, "the good angels"], endure not, but avenge".

Why a marker was needed where §4 originally took none. My first treatment chose the attributive reading silently, on the strength of angeli mali in the construal above. That was a silent choice between two readings that both construe, which is the one thing the apparatus exists to prevent. The evidence runs both ways:

• For boni (the marker's conjecture): the clause is set apud Deum, and its angels avenge blasphemy — that is the office of the faithful angels, not of the punished apostates. non boni fights its own sentence. And a doubled non … non inside four words (angeli non boni, non sustinent) is the classic signature of an intrusive particle, plausibly a compositor's dittography for ubi angeli boni non sustinent.
• For non boni as printed: Vel introduces an alternative construal, and an alternative may perfectly well keep the evil angels of the first half — who would then avenge insults offered to themselves (contra illos), which the pronoun allows.

⛔ The translation follows the plate and both nons stand. The marker exists so that the roughness reads as Migne's rather than as ours, which is exactly what it did not do before. A [sic:] was considered and rejected: the words are rendered in English on the page, and Pattern 12's own operational test ("does the reader see our translation or his type?") reserves [sic:] for carried, untranslated runs.

praesumptioni — a stranded dative, no longer papered over

et territi magnitudine poenarum, jam ab illo judicio praesumptioni desistunt

Rendered: "they now desist from that judgment, to presumption [cj: praesumptioni; read praesumptionis, "of presumption"]".

desistere ab illo judicio already takes its ablative, so the dative praesumptioni attaches to nothing. The expected form is the genitive praesumptionis, agreeing with the judicium superbiendi three lines above. My earlier "for presumption" was the defect this edition exists to catch: an English preposition supplying a causal relation the Latin case does not carry, which let the sentence read smoothly straight past a broken form. The bare dative exponent "to" now displays the strandedness instead of concealing it.

§4 · THE UNEXPRESSED SUBJECT — every supplied pronoun, named

The dominant defect class. Each place where the English supplies a person the Latin does not print, with what fixes it. Nothing here was left to construe itself.

Subject = Peter (the epistle's author), fixed by the ARGUMENTUM and by the first-person glosses that surround them. 0689A scribit ("He writes this epistle") — the ARGUMENTUM names Simon Petrus three lines above · 0689A cognoscit ("He acknowledges them as equal to himself"; sibi = Peter, who is the nos of coaequalem nobiscum) · 0689B addit ("Whence he adds") · 0689B scribit twice ("he writes that one to beginners, this one to the more perfect") · 0690B vult monere ("Why does he wish to admonish…") · 0690C vocat ("Beautifully he calls it not death") · 0690D tangit ("Here he touches both the pagans and the heretics") · 0693B ostendit ("he shows how effectual it shall be") · 0694B ait ("He saith not, others") · 0694B dicit ("He says that Paul had written to them").

Subject = God, fixed by the lemma above. 0691D–0692A redegit · contexerit · voluit ("he reduced them to ashes… he willed the surrounding region still to preserve…"). The antecedent is not in the sentence: it is Deus in the VERS. 4 lemma Si enim Deus angelis two glosses up, which the Vulgate's own Et civitates… eversione damnavit continues. Also 0693C videt ("he sees it even as that which is done in one day") and 0694A differt · adimpleat · praevidit, where the printed Dominus of Non tardat Dominus is the subject three words earlier.

Subject = Lot, fixed by the lemma Loth oppressum. 0692B cernens · valebat · reservabat · fuscabat ("which, beholding, he was not able to correct; but he himself kept himself in good works. Nor did he darken himself…").

Subject = the Sodomites among whom Lot dwelt — the one that needed the Vulgate. 0692B quia nihil in illo videbant, nihil de eo audiebant. The sentence prints no subject at all, and English cannot leave the verb bare. It is not the proximi of the sentence before (who are the ones whose deeds tormented him, in the accusative), and it is not Lot. It is 2 Pet 2:8's habitans apud eos — the people he lived among. Rendered "they saw nothing in him, they heard nothing of him". If a blind reader disputes one supplied pronoun in this work, it should be this one.

Subject = David, fixed by the same clause. 0691B prophetavit ("because he prophesied of Christ more clearly than the others") — propheticum sermonem David is printed nine words earlier, and firmiorem inter alios prophetas is the comparative that requires a prophet, not the word, as its subject.

Subject = an individual heretic, and Migne shifts number mid-sentence. 0691C–D: associat (singular) then faciunt · concitant (plural). Rendered as printed, "in those whom, seduced, he associates to his own heresy… whom they stir up" — the number shift preserved rather than smoothed to one or the other. This is a real inconcinnity on the plate and it is not repaired.

⚠ But the English reader has no antecedent for that singular, because the only expressed plural in view is Per haereticos and nothing in the sentence licenses a "he". The reading is almost certainly distributive — whom each of them associates to his own heresy — the singular ranging over the plural haereticos, with the plural verbs then returning to the whole body. Recorded here rather than resolved in the text, so that the roughness reads as Migne's alternation and not as a translator's slip.

⭐ The dative goal of associat is suae haeresi, not the heretic. An earlier draft read "whom he joins to himself, seduced into his own heresy", which kept the sense but moved the government: it made the person the goal and demoted the heresy to an adverbial phrase. Repaired after the blind read. The verse's own logic wants it this way — the men are annexed to the heresy, which is what makes them count as a place where the way of truth is blasphemed.

Two sites where the readings were weighed and both construe:

1. 0689A quia eadem quam ipsi acceperant Christi gratiam et fidem. Is ipsi the recipients ("the same faith which they themselves had received") or the apostles ("which we ourselves had received")? The Vulgate's coaequalem nobiscum sortiti sunt fidem argues for the apostles as the standard of comparison; the Latin pronoun is third person and argues for the recipients. English "they themselves" preserves exactly the ambiguity the Latin has, so no choice was forced and none was smuggled. Recorded so nobody "fixes" it into "we ourselves."
2. 0692C ubi angeli non boni, non sustinent, sed vindicant. ⛔ This entry was wrong and has been superseded — see §3a. It set out two construals of non boni, chose the attributive one silently on the strength of angeli mali above, and took no marker, on the reasoning that "neither reading changes a negation." That reasoning was the error: the competing reading is that the non is itself intrusive, so the polarity of the clause is precisely what is at stake. The translation is unchanged and both nons still stand; a [cj: non boni; read boni] now carries the alternative on the page.

One subject that cannot be recovered, and is not guessed. 0693A, the tail of the Bede excerpt: ante Dei oculos sordidas ipsas etiam lacrymas facit. Whatever governs facit was cut away by Migne's own etc. / usque ad. Rendered with a neutral "it makes even those tears themselves foul before the eyes of God" — no antecedent invented.

§5 · MIGNE'S OWN PRINTING, LEFT ALONE

• 0689A quia eadem … Christi gratiam et fidem. eadem where the accusative eandem is wanted. Pattern 9: a case/gender mismatch has no English exponent — "the same" is invariable. No emendation to make and none to undo; no marker, logged so a later sweep does not churn it.
• 0692A dilectio carnalis … nil tamen in invisibilibus nisi incendium sibi reservant. A plural verb under a singular subject. Unlike the number-agreement case at Pattern 12's second test, English here does have an exponent — "reserve" against "reserves" — so rendering it plural displays the disagreement rather than hiding it, and no marker is needed. Rendered "carnal love, though it smile upon fools, yet they reserve to themselves nothing among things invisible but burning". The plural picks up the stultis and the eis of the following clause; the singular dilectio remains its grammatical subject. Deliberately not smoothed to "it reserves".**
• 0692C jam ab illo judicio praesumptioni desistunt. ⛔ Superseded — see §3a. Pattern 12 does exclude it (the form is real), but the conclusion that "the English asserts nothing false" was wrong: the English did assert something, namely a causal relation ("for presumption") that no case in the Latin carries. Now rendered "to presumption" with a [cj:].
• 0692D VERS. 20 …quam mundus immittit; rursus implicati… Migne points a semicolon in the middle of a conditional protasis. Pattern 8: preserved, not merged. The English reads "If, fleeing the defilement both of soul and of body which the world sends in; being again entangled (as before baptism) they are overcome by those defilements, it appears that…" — the awkwardness is his.
• Verse sequence. Chunk 0 runs 1 · 2 · 3 · 4 · 5 · 6 · 7 · 8 · 9 · 10 · 11 · 12 · 13 · 14 · 15 · 16 · 19 · 20 · 21; chunk 1 runs 1 · 2 · 4 · 5 · 6 · 7 · 9 · 11 · 13 · 17 · 18 · 20 · 22, then chapter III 1 · 2 · 5 · 8 · 9 · 10 · 13 · 14 · 15 · 16. All gaps, never reversals — the recension simply does not gloss every verse, unlike Jude, whose plate prints 23 before 21. Nothing to log beyond the fact that it was checked.
• VERS. 9.-- inside an italic run at 0694A. The Bede excerpt's usque ad tail runs …unde et recte subjungitur: and then the next verse address and lemma follow inside the same italic span. Reproduced exactly in place, address, column anchor and all.
• Migne's own ( sic ): none in this work. Nothing stacked.

§6 · THE JUDE PARALLEL — where 8985 and shipped 8978 print DIFFERENT things

2 Peter 2 and Jude cover the same ground and Migne's Glossa on them overlaps, so this is the list a reader of both actually wants, and the one place a rendering could have been imported silently. No English was carried across from src/english/8978/. Every line below was rendered from this plate and then compared.

The material overlap is far thinner than the biblical parallel would predict. Jude's gloss and this one share their subject (false teachers, the fallen angels, Sodom, the waterless springs) but almost none of their sentences. The shared matter, and how the two plates differ:

⭐ The one place the pull was real, and it was resisted: 0690D indoctas. Jude's plate carries no counterpart, but the verse is one of the best known in 2 Peter and the familiar English ("cunningly devised fables") arrives before the Latin is read. That is the same mechanism that produced Jude's Arbores autem trap (8978 §2a), and it is the reason this work's single most important marker exists. See §1a.

§7 · LEMMA EXPANSIONS — the words this translator supplied, named

Convention 3 (a fragmentary lemma is rendered as the phrase it stands for, not a word-for-word stump; Ruth's Non te. → "Let no man see thee"). Every place it was invoked, with what was supplied, so a reader can see the edge of our hand:

• 0691D VERS. 4 Si enim Deus angelis. → "For if God spared not the angels." Supplied: non pepercit, from Vulg. 2 Pet 2:4. "For if God, to the angels." is not English and angelis is a dative that only parcere governs. ⚠ This is the one expansion that supplies a negation, which is why it is named here rather than passed over: the non is the verse's, printed in the Vulgate the fragment opens, and the alternative under convention 3 was a stump that says nothing. In the Beda excerpt two words later the same gap is filled with the corpus's supplied-word brackets instead — "If God [spared not] the apostate angels" — because that is running prose, not a lemma, and brackets are available there.
• 0690B VERS. 13 Justum autem, → "But it is just,". Supplied: the copula only. arbitror was deliberately not supplied, because the gloss's own next clause renders it (quia justum arbitror → "because I judge it just") and supplying it in the lemma would have pre-empted the gloss.
• 0694A VERS. 10 Coeli magno. Coeli, → "The heavens with a great violence. The heavens,". Supplied: impetu, from Vulg. 3:10 caeli magno impetu transient. The bare ablative magno has no English exponent standing alone. The gloss expounds only the second lemma (Coeli, id est sancti), so nothing downstream rests on the supplied word.
• 0694B VERS. 14 Satagite immaculati et inviolati. → "Be diligent to be unspotted and undefiled." Supplied: the infinitive that English needs between satagite and its predicate adjectives (Vulg. satagite immaculati et inviolati ei inveniri in pace).
• 0693A VERS. 1 Hanc ecce vobis. → "Behold, this to you." Nothing supplied. secundam scribo epistolam was left out on purpose: the gloss immediately says it (hanc etiam secundam scribo Epistolam), so the fragment is intelligible as it stands.
• Everything else was left as a short lemma because short English carried it: Simon. · Equal. · In justice. · For so. · Perdition. · The unjust. · Into ashes. · Proud things. · The elements. · Difficult things. · They pervert. · It is hidden. No stump in the work is unreadable as English.

Two English commas inside lemmata, declared. In scientia autem abstinentiam. → "But in knowledge, abstinence." (the comma marks an ellipsis English cannot carry otherwise) and Hanc ecce vobis. → "Behold, this to you." Both are ordinary English prose pointing, not sentence-type marks, so Pattern 8 is not engaged; a comma that was going to be added to Caecus est et manu tentans. was removed on review, since "He is blind and groping with the hand." needs none.

§8 · READ CLEAN — sections with nothing to report

Named so that a blind reader can tell "checked and clean" from "not looked at".

• The ARGUMENTUM (0689A). One sentence, no scripture, no defect. pietas fixed as "piety" here and held to it at 0689C/D (pietatis opera → "the works of piety") — one voice within the work, per the merge-time reconciliation rule. ⚠ One genuinely undecidable construal, raised for the record by the blind read and left as written. Simon Petrus per fidem huic mundo sapientes mortuos esse declarat: per fidem can attach to the accusative-and-infinitive (the wise are dead to this world through faith — what our English gives) or to declarat (Peter declares it through faith). Word order slightly favours the first, the finite verb's proximity slightly favours the second, and there is no evidence either way — nothing in the sentence, and no gloss below picks the phrase up. Recorded because a reader deserves to know the ambiguity is Migne's and not a reading we resolved quietly.
• CAPUT PRIMUM VERS. 2–8 (0689B–0689D). Six glosses, no divergence, no defective type; the multiplicetur / adimpleatur contrast between the two epistles verified against 1 Pet 1:2 and 2 Pet 1:2 respectively and both are correct.
• VERS. 14–15 (0690C–0690D). The depositio gloss and the dabo operam gloss: no scripture, three negations, all carried, no ambiguity of subject beyond the Peter voice.
• VERS. 16's transfiguration gloss (0691A). Elias / Moses / Peter, James and John read as three ages; the doubled Vel, per Moysen mortuum mortuorum, per vivos vivorum is elliptical in the Latin and is rendered elliptically ("by Moses dead, that of the dead; by the living, that of the living"). Not a dittography — the two limbs differ and neither is an eye-skip; Pattern 11 was considered and declined.
• VERS. 20–21 (0691B–0691C). The four nons of the prophecy argument all carried; the chunk's single question mark is at 0690B (Quare vult monere illos qui sciunt, et confirmati sunt?), a direct question that keeps its mark with no recasting — Pattern 8b is not engaged anywhere in this work.
• CAPUT II VERS. 9–13 and 17–18 (0692B–0692D). Read clean.
• CAPUT III VERS. 2 (0693A–0693B). The five-italic chain (Verborum, · a sanctis prophetis · apostolorum: · praecepta · Domini et Salvatoris) reproduced with each span's own printed punctuation inside or outside the asterisks exactly as Migne sets it, including the absence of a comma after a sanctis prophetis, where English would normally want one.
• CAPUT III VERS. 16 and the closing Depravant gloss (0694C). Three negations (Nullus · non perverterint · non habent), all carried; the work ends where the plate ends.
• Em-dashes: zero in both chunks, checked after writing. Two places where a dash was the first reflex (0690B's Ideo monebu… period and 0692C's quod est exsecrabile) were rebuilt on commas.

§9 · WHAT A BLIND READER SHOULD LOOK AT HARDEST

1. 0692C angeli non boni — §3a. Two readings, evidence both ways, now carried on the page as a [cj:] instead of resolved silently. The likeliest remaining defect in the work, and the one place a reader may reasonably think the marker should have been a decision instead.
2. 0692B quia nihil in illo videbant, nihil de eo audiebant — the supplied "they", §4. The subject comes from 2 Pet 2:8, not from the sentence, and it is the exact shape of the class that beat Ruth at 0538C.
3. 0691D VERS. 4's supplied spared not — §7. The only place this translation supplies a negation the run does not print, defensible under convention 3 and named for that reason.
4. 0689B's John 17:3 allusion — §1c. The declined [var:]; unum for solum is the kind of word an argument could rest on, and this one does not, but the judgment is contestable.
5. 0692A reservant — §5. Rendered plural against a singular subject. If the house view is that the sentence should read "it reserves", that is a repair and should be argued, not made.
6. 0694A Quantor — §3. The decision not to write "The four" beside the marker.
7. 0691C–D's singular/plural shift (associat … concitant) — §4. Preserved rather than harmonized; the distributive reading is argued there but not imposed.
8. 0689A's per fidem — §8. Undecidable, and recorded as such.

§10 · THE BLIND POLARITY READ — what it found, and the transferable mechanism

A blind reader went over this work after the English was filed. It never saw this cruces file, and it read Jude (8978) in Latin only, so neither our reasoning nor our shipped English of the sibling epistle could steer it. It returned four repairs, all now applied, and — more usefully — two negatives on the two risks this work was launched to guard.

10a · The two negatives, with the evidence

No silent scripture conformation, anywhere. Every scripture span was checked individually against the Vulgate. Two were singled out as the model for the class: 0690D indoctas/doctas and the (Isa. LXV) locator at 0694B (§1a, §2). The reader also confirmed that Migne is followed against the Vulgate at Apoc. VI, 0694A on both points, the transposed adhuc modicum tempus and the omitted et fratres eorum, and called that the highest-risk span on the page. §1c had declined a marker there on transposition grounds and logged it; the decline is confirmed.

No Jude contamination. The reader compared the two plates directly and reached §6's conclusion from the other side: Migne's two recensions share almost no wording. The clearest demonstration is the one place the biblical parallel is tightest, the fallen angels:

• Jude's plate — Si enim angelis peccantibus non pepercit
• 2 Peter's plate — Si Deus apostatas angelos traditos poenis inferni

Two different sentences about the same event. Likewise the two Hi sunt glosses, which share a lemma and are about different things (Jude: murmurers envying their brethren; 2 Peter: nihil in eis est nisi vanitas et immunditia). There was, in the end, very little available to import even had the pull been given in to.

10b · ⭐ The mechanism behind two of the four repairs — carry this to the next work

Repairs 1 and 3 are the same error twice: an English preposition supplying a relation the Latin case does not carry.

Why it is a class and not two slips. Latin fixes these relations in the morphology; English must choose a preposition, and the reflex reaches for the commonest one, "for", which in English is ambiguous across purpose, benefit, cause, exchange and duration. Nothing looks wrong afterwards: the sentence is smooth, no negation moved, no lemma changed, and every automated check passes. It is Pattern 7a's fifth mechanism (reference or agency mis-bound) wearing a preposition instead of a pronoun, and no particle-counting check will ever see it.

The habit that catches it: when the English needs a preposition the Latin does not print, name the case and name what governs it before choosing the word. pro + ablative after a verb of relinquishing is exchange, not purpose. A dative after a verb that has already taken its ablative is attached to nothing, and the honest English is the bare exponent plus a marker, never the preposition that makes the clause read well.

10c · Confirmed correct, no action

The [spared not] bracket at 0691D (§7); the [sic:] handling of all five carried-type runs, including leaving Quantor wholly untranslated rather than writing "The four" (§3); Coaequalem → "Equal", since the gloss recovers the force with equal to himself; pseudo and pseudoapostoli kept distinct; quia et primo … et cum read temporally; ita videt, sicut folded; and verba doctorum auditum praebere treated as ordinary medieval case looseness (Pattern 9).

Six comma→semicolon promotions (0689B, 0690A, 0690C, 0690D, 0692B, 0694C) were observed and expressly declined for repair: they are real, positional and all in one direction, but a corpus-wide semicolon measurement closed last batch as a rate question, not an absence, with a standing instruction not to sweep it. Nothing was softened in the other direction anywhere in this work — no printed mark deleted, none supplied — so no reading is altered and no parity check is fooled. Left alone rather than churned across six sites.