Epistola I B. Joannis

Translated 2026-08-19, one agent, 5 chunks (4,275 Latin body words → 6,458 English, 1.51×). Conventions are those settled on Liber Ruth (8968) §1 and restated in GLOSSA-NT-AGENT-BRIEF.md; nothing new is ruled here. The two shipped siblings — 2 John (8984) and 3 John (8989) — were read first and their vocabulary reused (Charissimi → "Dearly beloved"; usque ad → as far as; VERS. n.-- verbatim; fragment lemmata expanded to the phrase they stand for).

Apparatus taken: 6 [var:] · 10 [sic:] · 1 [d:] · 1 [cj:] · 0 [ed:] · 0 [nt:]. Marker parity with the Latin twins: 42 column anchors · 19 [n:] notes · 166 italic spans · 0 guillemets · 11 question marks, all 1:1 and in place. verify-english.mjs 8979: clean (5 chunks, columns/notes/sections aligned, no duplicate paragraphs).

Punctuation checked BAND BY BAND, not by total (test 3). The 42 verified anchors partition both twins into 47 bands; colons, semicolons, question marks, guillemets and exclamations were compared inside every band and zero bands mismatch. No mark was supplied and none deleted; in particular no ? was invented, and Migne's five in chunk 0, two in chunk 2, one in chunk 3 and three in chunk 4 all ride as printed.

§1 · THE LEMMA DIVERGENCE LIST (7a″) — all 156 lemmata collated against the Clementine

Every italic lemma in the work was collated against sources/vulgate/clementine-flat.txt before the gloss under it was rendered, and each rendering was then held fixed against every re-quotation of it in the gloss below. Agreements are named individually below, not passed over, because a list of findings alone cannot be told apart from a list by someone who did not check.

1.1 · The six divergences that took a [var:]

1.2 · Divergences examined and DECLINED, with the reason for each

None of these was absorbed silently; all are recorded because a blind reader will look here.

• The clipped-possessive class (three sites, one habit). 0693D et manus contrectaverunt (Vulg. manus nostrae contrectaverunt); 0695B Remittat nobis peccata (Vulg. remittat nobis peccata nostra); 0699A apparuit, ut peccata tolleret (Vulg. ut peccata nostra tolleret). Declined as a class, not one at a time: three lemmata drop exactly one possessive and nothing else, which is a lemmatizer's clipping habit rather than three independent variants. Rendered without the possessive in every case — "and hands have handled", "That he may forgive us sins", "that he might take away sins" — so the English does not quietly restore what Migne does not print. ⚠ These three are the sites where the familiar English ("our hands", "forgive us our sins") pulls hardest, and none of them followed it.
• 0695A quia Deus lux est for Vulg. quoniam Deus lux est (1 Jo 1:5). Pure connective substitution; identical English ("that God is light"). Pattern 14 fires on sense.
• 0695A quae societas lucis ad tenebras for 2 Cor 6:14 quae societas luci ad tenebras. Genitive for dative; identical English. Migne also drops autem and aut and prints v. 15 before v. 14. The re-ordering is his, is reproduced, and is not a textual variant. No marker.
• 0695D Si mandata ejus observamus for Vulg. observemus (1 Jo 2:3). Indicative for subjunctive; English "if we keep his commandments" either way, Pattern 9's "no English exponent" logic. Recorded, not marked.
• 0697D quicunque negat Christum, etc. The Clementine 2:22 reads Quis est mendax, nisi is qui negat quoniam Jesus est Christus? Migne's italic run is the gloss's own compression, introduced by its own scilicet, not a quotation of a variant text. Same call 3 John made on quae forsitan (8989 §1). No marker.
• 0699B Si enim dixerimus quia peccatum non habemus, nos ipsos seducimus for 1 Jo 1:8 Si dixerimus quoniam peccatum non habemus, ipsi nos seducimus. enim is the gloss's connective; quia/quoniam and nos ipsos/ipsi nos yield identical English. No marker.
• 0703B honorificabit illum Pater meus for Jo 12:26 honorificabit eum Pater meus; 0703B quis contra nos? for Rom 8:31 qui contra nos?. Both render as the same English word. No marker.
• 0704A quoniam cui talia agunt (Gal 5:21 quoniam qui talia agunt, regnum Dei non consequentur). This one is not declined: it took a [cj:] instead, see §2.4.
• Orthography, all Pattern 9, none marked: Charissimi for Carissimi (0698C, 0701C); charitas for caritas throughout; unquam for umquam (0702A); exsecrabilis for execrabilis (0700D). These are PL's nineteenth-century house spellings, printed corpus-wide. The precedent is 3 John §1, which declined Charissime on the same ground.
• 0694D Beati qui non viderunt et crediderunt [n: (Luc. XXIV)] — the text agrees verbatim with John 20:29; the citation does not, Migne printing Luc. XXIV. The note passes through verbatim per the locked register. This is a misattributed locator, not a textual divergence, and takes no [var:].

1.3 · Verified agreements, named by chunk

Everything not listed in §1.1–§1.2 was collated and agrees with the Clementine word for word (fragments agreeing as far as they run). Named so a later reader can tell a checked book from a skipped one:

• Chunk 0 (21 lemmata). Quod fuit ab initio · quod audivimus · quod vidimus · quod perspeximus · Vidimus · Annuntiamus vobis · Et haec est annuntiatio · Annuntiamus · Si dixerimus · Et in tenebris · Si autem in luce · Sicut et ipse · Si dixerimus · Habemus · Mendacem — all agree. (habuimus: is not scripture; it is the contrastive form the gloss coins to set against Habemus.)
• Chunk 1 (48). All agree except the two named in §1.2 (observamus, quicunque negat Christum). Includes the full VERS. 10–11 chain qui odit est in tenebris · et nescit · quo eat; et · nescit quo eat · quia tenebrae obcaecaverunt oculos ejus, which reproduces 2:11 exactly, and the split lemma quia tenebrae … transierunt (2:8), whose two italic spans and the roman jam between them are all reproduced in place.
• Chunk 2 (40). All agree except the four named above (es tu, semen ipsius est, cognoscimus, Majorem enim charitatem, Qui avertit aurem suam) and the clipped possessives. Quotquot autem receperunt eum, dedit eis potestatem filios Dei fieri (Jo 1:12) and Cum autem senueris … quo tu non vis (Jo 21:18) agree verbatim, which is worth saying: three of the work's eight cited proof-texts diverge and two do not, and only collation tells them apart.
• Chunk 3 (18). All agree except the VERS. 7 composite (§2.1). Hic est Filius meus dilectus (Mt 17:5 / 3:17), Deum nemo vidit unquam, Timor non est in charitate, Et nos debemus alterutrum diligere, Ideo de mundo loquuntur, Quoniam in eo manemus — all verbatim.
• Chunk 4 (29). All agree. Including the entire 5:16–21 run (Est peccatum ad mortem · Omnis iniquitas peccatum · Omnis qui natus est · Non peccat · Sed generatio Dei · Et mundus totus · Et scimus · Filioli · Custodite), and Quoniam tres sunt — on which see §2.2.

1.4 · ⚠ THE SCRIPTURE-CONFORMATION HUNT, run separately

This is the fault that beat Ruth and it is this book's named risk, so it was hunted as its own pass rather than folded into the collation. Every place where the familiar AV/Douay wording and Migne's Latin part company was listed first, then checked in the finished English:

Result: zero conformations found in the finished English. The two places the pull was strongest and was resisted are the 1 Cor 10:13 echo at 0695C and Deus charitas est at 0701C.

§2 · CRUCES

2.1 · 0701C · VERS. 7 — the lemma ends with v. 8's clause, and the gloss's own next sentence says so. [var:]

Printed: …in quo eum maxime debemus audire, scilicet: Qui diligit, ex Deo natus est et Deum novit, quia Deus charitas est. Dixerat: Charitas ex Deo est; hic superaddidit: Deus charitas est.

The composite lemma set out as 1 Jo 4:7 ends quia Deus charitas est. But 4:7 ends quia caritas ex Deo est; Deus caritas est is 4:8, and the very next sentence of the gloss says so in as many words — he had said "Charity is of God"; here he added "God is charity." The gloss's argument requires the lemma to end the way the Vulgate ends it, and Migne's lemma does not.

Rendered exactly as printed, with a [var:] naming both clauses and which verse each belongs to. No emendation, and no re-ordering of the lemma to rescue the argument — the mismatch between a lemma and the gloss built on it is precisely what this edition exists to show. The gloss also prints Deum novit where 4:7 has cognoscit Deum; same English, recorded in §1.2 and not marked separately.

⚑ This is a crux that would obviously profit from an editor's note. None is available: the plate spot-check for this work (p. 352 cols 693/694, p. 354 cols 697/698, full width, 300 dpi) found no foot-of-page apparatus and no asterisk-keyed EDIT. note, so no [cn:] exists to settle it. Cols 701–702 were not themselves sampled, and a plate read there is the one cheap thing that could still add evidence.

2.2 · ⭐ 0702D–0703A · VERS. 7 · THE COMMA JOHANNEUM — what Migne actually prints, and it is not the Comma

The brief flags this as the corpus's densest region for 1 Jo 5:7–8. The finding is negative and it is the interesting kind of negative.

Migne lemmatizes VERS. 7 with the bare stump Quoniam tres sunt. and prints no more of the verse at all — not qui testimonium dant in caelo, not Pater, Verbum, et Spiritus Sanctus, not et hi tres unum sunt, and nothing of v. 8's earthly triad either. Verses 8 and 9 are not lemmatized; the next verse address is VERS. 10. Nothing was imported and nothing was omitted on the strength of what the verse is expected to say: the stump was collated against the Clementine (Quoniam tres sunt, qui testimonium dant in caelo), agrees exactly as far as it runs, and is rendered "For there are three."

The Comma is nevertheless presupposed by the gloss, which is the substantive observation:

Et de utroque habemus certum testimonium, de Deitate quidem per Patrem, et Filium, et Spiritum sanctum: de humanitate per animam, aquam, et sanguinem.

Two things worth a later reader's attention, neither of which took a marker:

1. The heavenly triad is named Filium, where the Comma reads Verbum. This is the gloss's own prose, in roman, not a quotation and not a lemma, so Pattern 14 does not reach it (the same call 2 John made on Positus est Adam in paradiso). Recorded here so the substitution is not invisible.
2. The earthly triad is anima, aqua, sanguis where v. 8 reads spiritus, aqua, sanguis — and the gloss has already declared its own equivalence four lines earlier at 0702D (Spiritus. Id est, humana anima). So this is exegesis, not a variant text.

No [var:], no [sic:], no conjecture. The zero here is a checked zero.

2.3 · The [sic:] register — ten sites, and what each one was NOT allowed to swallow

Every marker was closed by stripping it and reading the clause aloud (test 1). Where the carried run took a verb, a subject, a conjunction or an object, the sense is supplied in plain words beside the marker; where it took only a modifier, the marker stands alone and the conjecture lives here.

All ten are Pattern 10/12 non-word type: viucendum, apparel, Filil, aetenam, acciplunt and Creora are single-letter substitutions (n/u, t/l, i/l, r-loss, i/l, b/o); hae eticorum and ten atio are split runs and carry their spacing; Latin and nis are truncations. Whether any of them is Migne's plate or Corpus Corporum's transcription cannot be told from the Latin, and the brief's §4 warning cuts both ways here — see §2.7.

2.4 · 0704A · cui for qui — [cj:], not [sic:], and why

Printed: …et ideo oratione justorum non purgatur: quoniam cui talia agunt, regnum Dei non consequentur.

The clause quotes Gal 5:21, which reads quoniam qui talia agunt, regnum Dei non consequentur. cui is a perfectly real Latin word — the dative of qui — so Pattern 12 cannot reach it ([sic:] may only wrap type that is not a word), and Pattern 7 requires it to be rendered. But cui cannot govern the plural agunt, and the faithful English is broken. That is exactly the gap Pattern 18 was ruled for.

Rendered additively: "since to whom [cj: cui; read qui, "who"] do such things, they shall not obtain the kingdom of God." Migne's word keeps its English, the conjecture stands beside it, and nothing is replaced. The rival reading — treating cui as broken type and marking [sic:] — loses, because a [sic:] asserts the type is not a word, and cui is one; Pattern 12's own boundary paragraph ("a real word that is merely wrong") puts this case on the far side of it.

2.5 · 0699A · iniquitatibus | quitatibus — a doubling ACROSS the column break. [d:]

Printed: Ne nos (qui peccatis et iniquitatibus [0699A] quitatibus carere non possumus) de salute desperemus…

The two runs differ (iniquitatibus / quitatibus), so Pattern 11a governs and the marker goes on the intrusion, wherever it stands. The intrusion is the second: iniquitatibus is the complete word, agrees with peccatis, and is what the sentence needs; quitatibus is its tail re-set immediately after the column break and is not a word at all. Marked [d: quitatibus], which renders as text with the "repeated in Migne's plate" rule, and the reader loses nothing — stripping it leaves "who cannot be free of sins and iniquities", complete.

[sic:] was considered and rejected: the defect here is not that a word is broken, it is that a run is set twice, and .dittog is the mark that says so. Saying it with [sic:] would file a doubling under broken type and make the two classes indistinguishable in a corpus that separates them deliberately.

⚑ This one wants a plate read and is the strongest candidate in the work. A partial re-set of the last word before a column break is the classic shape of a catchword, and it is equally the classic shape of a digitization re-reading one. If the plate has iniquitatibus once, this is Corpus Corporum's doubling and belongs in data/tei-patches/8979.json, not in a [d:] marker. Recorded as unresolved rather than assumed either way. Col. 0699 was not in today's spot-check.

2.6 · Printed oddities rendered literally, with no marker available

• 0698A · Et nunc, filioli., manete. Migne sets a full stop and then a comma inside the lemma. Rendered exactly so: "And now, little children., abide." Pattern 8 governs (marks ride as printed); no marker fits, because the type is not broken, nothing is missing from the digitization, and the Vulgate agrees (2:28 Et nunc, filioli, manete in eo). ⚑ A stray period is also a plausible transcription artifact and this is the second site in the work worth a plate read.
• 0696A · quia tenebris, ejectis desiderium novae lucis infundit. The comma falls inside the ablative absolute (tenebris ejectis) rather than after it. Preserved; the English cannot reproduce a comma in that position and reads "because, the darkness being cast out, it pours in the desire of a new light." Recorded so the displacement is not invisible.
• 0700B · Quia non est perfecta charitas in illo; qui non ponit animam… A semicolon where the sense wants a comma, cutting the relative clause off from its antecedent. Reproduced 1:1 ("perfect in him; who does not lay down his life"), Pattern 8.
• 0704B · non peccavit, ad mortem, sed veniam poenitendo meruit. The comma splits peccavit from ad mortem. Reproduced ("he did not sin, unto death, but by repenting he earned pardon").
• 0702D · contra naturam aqua sanguis vivaciter fluxit. Migne prints the two nouns with no et, four lines after printing the lemma aqua et sanguis in full. Rendered as English asyndeton, "water, blood flowed vigorously" — the conjunction was not supplied. Nothing is broken and nothing is missing, so no marker; but a reader comparing the columns will see it, and this is where it is said.
• 0704B · sic justitia et iniquitas malignus et generatio Dei. The comparison has no verb in its second member and no conjunction between its two pairs. Rendered with the same ellipsis: "so justice and iniquity, the malignant one and the generation of God." No verb supplied — the same discipline the usque ad formula requires.
• 0704C · Nemo sine divina cognitione ad vitam aeternam pervenire, nemo cognoscere sine gratia Dei potest. One potest serves two infinitives across a comma. English cannot carry that ellipsis, so "can" is repeated: "No one can come to eternal life without the divine knowledge, no one can know without the grace of God." Recorded because a word appears twice in the English that appears once in the Latin.
• 0699B · Hoc ipsum quod qui justus est ab illo justus est. A verbless sentence. Rendered verbless: "This very thing, that he who is just is just from him."

2.7 · ⚑ Forms that read suspiciously TIDY, raised per brief §4

The digitization alters spelling in both directions, and a word that reads well may read well because someone upstream fixed it. Two candidates, both raised and neither acted on:

• 0704A · supra in Sed supra alia est peccatum ad mortem. It construes ("but above", i.e. as said earlier) and it also reads as though it might be tidied from super or from a displaced supra dicta. Corpus check: supra is overwhelmingly common across src/latin/, so frequency gives no signal, and the sentence is intelligible as printed. Left alone, and the translation deliberately does not choose between "above" (anaphoric) and "besides" — see §3.2.
• 0703C · In fide et confessione nominis opening a Bede excerpt whose surroundings are rough. Nothing anomalous found; a checked negative, recorded because the other candidates in this section came back inconclusive and a clean one is worth having.

⚑ And the general caution, since this work carries ten [sic:] sites in 4,275 words, which is a high rate: the run of single-letter substitutions (apparel, Filil, viucendum, acciplunt, aetenam) has the profile of OCR, not of a nineteenth-century compositor. Cols 0697–0704 were not sampled today. If a plate read is ever run on this work, run it on those five first: if the plate is clean, all five are data/tei-patches/8979.json entries and the [sic:] markers come out. They are marked as the plate's defects because that is what the evidence on disk supports, not because the question was settled.

§3 · THE SUPPLIED-SUBJECT REGISTER — every pronoun our English prints that the Latin does not

The brief names this the largest defect class corpus-wide and 1 John its worst case, because the referent of ille / hic / ipse oscillates between Father and Son. Each supplied pronoun below carries the evidence that fixes it.

3.1 · Settled, with the evidence

• 0693D, ARGUMENTUM · manifestat · dissertat · comprobet → "he makes plain… he discourses… he proves". Subject = John. Evidence: the ARGUMENTUM is a summary of what the epistle's author does, the same construction as the ARGUMENTUM of 2 John (8984), and every finite verb in it is third singular with no other candidate in the sentence.
• 0695A · monstrat (Hac sententia divinae puritatis excellentiam monstrat) → "he shows". = John. Fixed by the immediately preceding vide quid voluit docere — the one doing the teaching is the writer whose sententia it is.
• 0695A–0695B · ostendit, 0695B Non ait, 0695B vocat / appellat → "he shows", "he does not say", "he calls". = John throughout; Non ait, habuimus is decisive, since only the epistle's author can be said not to have written a word.
• 0696A · promisit (Secundum quod promisit) → "he promised". Two readings both construe and the English keeps both open: God, who is the promiser of the knowledge and love in question, or John, the subject of every other verb in the paragraph. Not resolved, and the reason is in §3.2.
• 0696A · impletum (Christo impletum) → "Fulfilled in Christ, because he obeyed the Father unto death: and therefore he was glorified." Subject = Christ, named in the clause itself. No ambiguity.
• 0698B–0698C · ostenderit Deus … Dedit enim … Deus dedit — subject printed in all three. Nothing supplied. Named here because Dedit. is a lemma and the temptation is to supply a different subject for the lemma than for the gloss; the gloss prints Deus.
• 0698D · viderunt mali, et in judicio sunt visuri → "the wicked have seen him, and in the judgment they shall see him." Object supplied. = the Word made flesh, from the same sentence's Verbum caro factum est. The contrast with videbunt soli justi (of the Word in principio) makes the object identical in both halves.
• 0699B · sicut ille est … in illo … in isto → "as that one is… in that one… in this one." ille = Christ (the ille of 1 Jo 3:7, sicut et ille justus est, which is the verse being glossed); iste = the man who does justice. The English keeps the ille/iste deixis rather than naming either, because the Latin does not name them.
• 0699A · subdit, 0700A supponit, 0701D voluit, 0704D perhibuerit / demonstrat → "he adds / subjoins / wished / had borne / shows". = John in every case; each sits in a sentence about the ordering of the epistle's own argument.
• 0700C · scimus nos esse in conspectu → "we know that we are in his sight." Possessive supplied from the gloss's own next clause, id est promereri ejus conspectum. Migne prints the possessive four words later.
• 0701B · Solvit. → "He dissolveth." Subject = omnis spiritus qui solvit Jesum, the lemma directly above (1 Jo 4:3). The gloss closes hic solvit Jesum, naming the same subject.
• 0702C · Vere per fidem vincit mundum → "he conquers the world." Subject = the believer, from the lemma above (Quis est autem?, 1 Jo 5:5 qui vincit mundum).
• 0702C · Quis est autem? etc. → "But who is he? etc." The one supplied word in a lemma anywhere in this work. Fixed by 5:5, Quis est, qui vincit mundum; "But who is?" is not English and Ruth §1.2 governs. The etc. carries the rest and the expansion stops there.
• 0702D–0703A · the dedit testimonium sequence. Pater dedit … Ipse Filius dedit … Spiritus sanctus dedit — all three subjects are printed. The English supplies only the object "it" for the second and third (dedit absolute), which is the testimony named in the first. This is the paragraph the brief warns about most — Father, Son and Spirit in three consecutive clauses — and Migne names every one of them.
• 0703A · quando super baptizatum … requievit → "when he rested upon him who had been baptized". Subject = the Holy Spirit, printed as the sentence's subject; baptizatum is the substantive participle, i.e. Christ. Neither is a free choice.
• 0703B · quasi dicat → "as if he should say". = John, continuing the paragraph.
• 0704D · Cum testimonium … perhibuerit illis → "Although he had borne them* testimony". illis* is printed; only the subject is supplied, and it is John.

3.2 · Declined calls, said plainly rather than left to look decided

A flagged passage is not a checked passage. Where this file declines, it declines out loud.

1. 0696A · Secundum quod promisit. The subject is genuinely undetermined between God and John. Evidence for God: the content promised is ad ipsius cognitionem et dilectionem perveniri, and the immediately preceding gloss (VERS. 3) says per quae ad cognitionem ipsius venitur — the promise is about reaching God, which suits God as promiser. Evidence for John: every other finite verb of narration in the surrounding chapter has John for its subject. The English prints "he promised" and supplies no name, which is the one rendering that leaves the ambiguity exactly where Migne leaves it. Declined deliberately, not overlooked.
2. 0704A · Sed supra alia est peccatum ad mortem. supra is either anaphoric ("but above", as said earlier in the chapter) or "besides / beyond these". Rendered "But above, the sin unto death is another thing," which carries the printed word and lets the reader see the difficulty. Declined: nothing in the paragraph decides it, and choosing would put a reading into the text that the page does not make.
3. 0702D · aqua sanguis (§2.6). Whether the missing et is Migne's asyndeton or a lost word cannot be told from the Latin, so the English reproduces the asyndeton and no conjecture is offered. If the plate is ever read, this is the third site to look at.

⚠ Both 2 John and Esther were beaten by a translator who saw the crux, set out both readings, declined to resolve, and then supplied the wrong one anyway. The three declines above were therefore re-read after the English was finished, specifically to check that the rendering does not quietly encode a choice. It does not: item 1 prints a bare "he", item 2 prints the printed adverb, item 3 prints the asyndeton.

§4 · NEGATION AUDIT — every printed negative, its host named from STRUCTURE

Test 2 validates the mapping you propose, not the mapping the Latin makes (test 2a), so the population the brief names — a negative resting on a verb with a quantifier, correlative or predicate nearby — was worked first, and each host is proved from the sentence's own structure.

4.1 · The trigger population: every non solum … sed, nisi, omnis, solus, sicut … ita

4.2 · Read clean, and named

Every remaining non · nec · neque · nisi · ne · nemo · nihil · nullus · nolite · noluit · nequeunt in all five chunks was placed on its host and checked in the English. The dense stretches, all clean: 0696B (VERS. 10–11, six negatives in one sentence — nescit · non providet · nescit · non ideo · non sit · non exsequatur); 0697B (VERS. 19, five — non terreat · non putetis · non erant · Non potuissent … nisi · non est); 0704A–0704B (VERS. 17–18, eight). No negative was added and none dropped; the antithetical construction the brief warns of (qui dicit se in luce esse et fratrem suum odit, non diligamus verbo sed opere) is present at 0696A, 0696B and 0700B and each was mapped onto its correlative before it was rendered.

4.3 · The one place English has no exponent, said so rather than hidden

0702A · De hac radice non potest nisi bonum exire. The non is on potest; English cannot keep it there without an indefinite object ("there cannot come forth anything except good"), and the natural rendering moves it onto the object ("nothing can come forth except good"). The two are identical in scope here — there is no quantifier for the negative to interact with — so nothing is at stake, and the reading chosen is the one that is English. Recorded because the class it belongs to is the one that has beaten this pipeline, and a checked instance of it is worth having even when the answer is "no difference".

§5 · CANDIDATES EXAMINED AND REJECTED

• The VERS. n.-- sequence is Migne's and is not out of order. 1, 2, 5, 6, 7, 8, 9, 10 · 1, 3, 4, 7, 8, 9, 10–11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 24, 25, 27, 28, 29 · 1, 2, 3, 4, 5, 6, 7, 8, 9, 12, 14, 15, 16, 17, 18, 19, 20, 22, 23 · 1, 3, 5, 7, 8, 9, 10, 11, 12, 13, 16, 18 · 1, 3, 4, 5, 6, 7, 10, 11, 14, 15, 16, 17, 18, 19, 20, 21. Ascending throughout, with gaps but no inversions — nothing of the Jude kind. Two glosses (Ex ipso. at 0698B, Et haec est vita. at 0703C, Est peccatum ad mortem. at 0704A, Spiritus. at 0702D) stand as their own paragraphs with no verse address; that is Migne's setting and is reproduced.
• The bare one-word lemmata are NOT inconsistent with the expanded fragments (brief, 2026-08-19). Different classes, and the class was named before either was rendered: Habemus. · Filioli. · Adolescentes. · Solvit. · Dedit. · Custodite. · Videte. are complete in themselves and are rendered as they stand. Advocatum. · Justum. · Mendacem. · Quoniam multi. are bare accusatives or bare subordinators whose governing word is elided, and Ruth §1.2 expands them to the phrase they open ("We have an advocate", "Jesus Christ the just", "We make him a liar", "Because many false prophets are gone out"). The rule applied, stated so it can be checked: expand where the English stump is not a well-formed English phrase; leave it where it is. Every expansion tracks Migne's own Latin verse, never an English Bible's wording.
• ⭐ 0694D · Scripturas perhibentes testimonium de ipso Verbo — a mis-parse caught on the fourth-test read-back, and repaired. The first draft read "with much handling of the Scriptures bearing witness", taking Scripturas as a genitive with contrectatione. It is accusative, and perhibeo takes a double accusative ("to bring forward X as Y"). Now rendered "with much handling bringing forward the Scriptures as testimony concerning the Word himself." ⚑ Recorded because the wrong reading was smooth English and passed every count: it is the exact shape test 5 exists to catch, and it was caught by reading the paragraph back against the Latin and nothing else.
• ⭐ 0698D · tolletur impius ne videat gloriam Dei — Isaiah 26:10 in its OLD LATIN form. The Clementine reads Misereamur impio, et non discet justitiam … et non videbit gloriam Domini; Migne's clause matches instead the Vetus Latina/LXX form (ἀρθήτω ὁ ἀσεβής, ἵνα μὴ ἴδῃ τὴν δόξαν κυρίου), the version Augustine quotes. [var:] declined: Migne sets it in roman, with no citation, no note and no unde, as a clause of the gloss's own sentence, and Pattern 14 governs citations. Rendered from Migne — "the wicked man shall be taken away, lest he see the glory of God" — with the negative on videat where he prints it. Logged here because a reader collating against the Clementine will find no such verse, and the reason is worth having.
• 0698C nomine et actu nominaremur filii Dei, et in futuro simus. simus has no complement; it takes filii Dei from the clause before. Ordinary Latin ellipsis, not a defect. Rendered "and that in the world to come we should be so."
• 0701C Non differt charitas in Deo. differt prints no a Deo. Sound as it stands (the point is that in God charity is not a distinct thing), and the rest of the sentence spells it out. No marker.
• 0703B ut bene optando tendat in ipsum habet testimonium. No comma between the ut clause and habet. Punctuation only, Pattern 8; reproduced, and the English reads through.
• 0696C Patres, adolescentes, juvenes against the Vulgate's patres, adolescentes … infantes … juvenes (2:13–14). The gloss's list is its own summary of two verses, not a quotation; it also drops infantes, which it has already covered with filioli. No [var:] — Pattern 14 governs citations, and this is prose. Rendered adolescentes "youths" and juvenes "young men", fixed at first occurrence and held through VERS. 13 and 14, whose glosses depend on the two being distinguishable.
• Gender and attested spellings. No Pattern 9 item was marked anywhere: no gender mismatch has an English exponent, and Charissimi · charitas · unquam · exsecrabilis · quandiu · ocius are all attested forms.
• No guillemets, no Migne ( sic ), no [ed:] hole, no [nt:] prose note. All nineteen [n:] notes are bare attribution sigla or locators — (RAB.)-class material: (Luc. XXIV) · (II Cor. VI.) · (BEDA.) ×7 · (Joan. I) · (ID.) · (Psal. XXVI) · (Joan. XXI) · (Ibid. XV) · (Prov. XXVIII) · (Joan. XII) · (Rom. VIII.). Not one is editorial prose, so Pattern 15 does not fire and the [nt:] count is a checked zero.
• Pattern 4 [f:] was not applied, per runbook false positive 4: this is a lemma-and-gloss commentary, not a florilegium.

§6 · PLATE STATUS

Spot-checked today at p. 352 (cols 693/694) and p. 354 (cols 697/698), full width, 300 dpi: no foot-of-page apparatus and no asterisk-keyed EDIT. note. So no [cn:] is expected for this work and no conjecture of Migne's is available to settle any crux above. That is a sample, not a proof, and it is consistent with the PL 114 sampling recorded in 8984's cruces (leaves n356–n357, cols 703–706, also zero).

Four sites in this work would profit from a plate read, in priority order:

1. 0699A — the iniquitatibus | quitatibus doubling (§2.5). If the plate has the word once, this is a tei-patch, not a [d:].
2. 0697B–0703B — the five single-letter substitutions acciplunt · apparel · Filil · viucendum · aetenam (§2.3, §2.7). Their profile is OCR, not compositor.
3. 0698A — filioli., the stray full stop inside a lemma (§2.6).
4. 0701C — VERS. 7, the crux that most obviously wants an editor's note and has none (§2.1).

§Adjudication — plate read + blind polarity read (2026-08-19)

### ⭐⭐⭐ ELEVEN OF TWELVE SUSPECTED PLATE DEFECTS WERE OURS — see data/tei-patches/8979.json The translator raised its ten [sic:]/[d:] sites for a plate read, observing that five had the profile of OCR damage rather than compositor error. It was right eleven times and wrong once, and nothing in our files could have told us which — only the plate. pp. 353–357 (cols 695–704) read at 400 dpi, every reading taken from a full-resolution crop of the whole line. Patched and re-chunked; ten markers removed from the English. Three sub-kinds:

• line-break damage, where Migne hyphenates and the transcription mangles the join: ten-|tatio, vin-|cendum, and the ini-|quitatibus duplication, which spans a COLUMN break (698D→699A) — our source kept the rejoined whole word AND its tail. No dittography of Migne's; he sets it once.
• single-letter OCR substitutions a compositor does not make: acciplunt · Creora · apparel · Filil · aetenam · nis.
• a lost final letter at an unhyphenated line end (Latin for Latini) and a spurious full stop inside a lemma (filioli., manete). ⭐ THE SURVIVOR IS WHY THE READ WAS NECESSARY: hae eticorum at 0696D IS MIGNE'S — a full word-space where the r belongs, mid-line, no line break to blame, no broken sort. Its [sic:] stands, and the [sic:] on every other site would have been a public false claim about the primary witness. ⚑ And one of the removals exposed a swallowed word: "By [sic: Creora] repetition" had the bracket standing in the adjective slot, so deleting it left "By repetition". Crebra = "frequent"; supplied. ⚠ Contrast with 8959 the same day, where the suspicion ran the other way and the plate said the defect was Migne's own. The profile of a defect never decides it.

### Blind read — 3 sites, all repaired Reader fenced from this file, all five chunks read against the CORRECTED Latin, with the eleven orchestrator-edited sites named for particular attention (they had been edited by an adjudicator, not by a translator, and nobody had read them). It reported them faithful, and checked every negative against its host individually.

1. REPAIRED · 0704A · a [cj:] STANDING IN THE SLOT OF THE WORD IT MARKS — the swallow class again, twice in one work. quoniam cui talia agunt had been rendered "since to whom do such things, they shall not obtain…", which is not a readable clause: the corrupt cui was translated and the conjecture logged, so the conjecture never reached the English. Now "since they who [cj: cui; read qui] do such things shall not obtain the kingdom of God."
2. REPAIRED · 0697A · both relatives re-hosted, and a mis-parse. qui … impugnet, qui … exstirpet are masculine and pestis is feminine, so they attach to Antichristi, not to the plague — and hujus nequissimi capitis downstream requires exactly that personal antecedent, which our English had left dangling. Also praeconia is neuter plural of praeconium, "proclamations", not "heralds" (praecones). Both fixed.
3. DECLINED as proposed, RESOLVED THE OTHER WAY · the lemma-expansion cluster (0701A Quoniam multi. · 0695C Justum. and Advocatum.). The reader proposed contracting them to the printed catchword. Declined: expansion is never the defect, CONFORMATION is, and convention 3 renders a fragment as the phrase it stands for. ⚑ Its control was mixed, for the third time in two batches: of the three "fragments kept fragmentary" it cited, Filioli. → "Little children." and Ex ipso. → "Of him." are complete phrases with nothing elided, not clipped lemmata. Only Non quasi nos. was the same class — so the work held three expansions against one stump, and the stump was expanded to "Not as though we had loved God" (1 Jo 4:10, non quasi nos dilexerimus Deum), confirmed by the gloss's own next words, Non prius dileximus. Same resolution as 8998's Nolite. this morning.

⚑ Reader's most valuable rejection: 0699B Hoc ipsum quod qui justus est ab illo justus est — a sentence asserting nothing, which is the dropped-line signature found in 8962 today. It checked the Latin on disk, found it equally verbless, and flagged the site without proposing words. Correct on both counts.