Commentaria in Cantica

Merged 2026-07-31 from four agent batches, split on canticle boundaries (0000–0007 · 0008–0014 · 0015–0021 · 0022–0026). 27 chunks, 32,975 Latin words, 205 notes. verify-english OK. First English translation.

⚠ WHAT THIS WORK IS — read before anything else

It is NOT a commentary on the Song of Songs. The bare title Commentaria in Cantica says otherwise and the corpus misfiled it until 2026-07-30. It expounds the ten canticles sung at Lauds, written for King Louis from the cell at Ratestorph: Isaiah 12 · Hezekiah (Isa 38) · Anna (1 Sam 2) · the Song of the Sea (Exod 15) · Habakkuk 3 · Deuteronomy 32 · the Three Children (Dan 3) · Benedictus · Magnificat · Nunc dimittis. Caught during translation prep when its lemma vocabulary came back with four dilectus and a great deal of Theman and Edom. The doubled form Cantica canticorum is the reliable signal for the Song; the bare form is not — watch list in data/commentary-overrides.json.

⚠ THE METHOD — Rabanus prints TWO Latin editions on purpose

His preface states it: utramque editionem simul posui; illam videlicet juxta quam Romana psallit Ecclesia, et hanc quam ex Hebraica veritate translatam caeterae Occidentis ex maxima parte tenent Ecclesiae, utriusque editionis, ubi inter se discrepare videbantur, sensum exponens.

Expounding where the two editions differ IS THE BOOK'S METHOD, so levelling the two Englishes would not lose a nuance — it would delete the reason half these expositions exist. Fixed formulae: juxta Romanam → "according to the Roman [edition]" · ex Hebraica veritate → "according to the Hebrew verity" · Alia editio → "Another edition" (42×) · editio LXX interpretum → "the edition of the LXX interpreters" — never "the Septuagint", which is not what the page prints (one instance swept at merge). Worked pair: Hab. 3:2 turns on auditio "hearing" vs auditus "report", and that single difference is what the whole column argues from.

Work-wide conventions (reconciled at merge)

• canticum → "canticle" (28 in the Latin), carmen → "song" (13). Kept apart; "song" is never used for canticum, which would collide with the Song of Songs.
• feria kept, never converted to weekday names: secunda feria → "the second feria". Rabanus assigns each canticle to its day and the numbering carries his argument. Dominica die → "the Lord's day"; Sabbato → "on the Sabbath".
• Names: Ezechias → "Hezekiah" (established English), but Anna → "Anna", NOT "Hannah" — two expositions turn on Anna, quae interpretatur gratia, and the English form breaks the gloss. One "Hannah" swept at merge.
• christos kept as "christs", because the gloss unctos suos fecit depends on it.
• Etymologies translated, the name kept as printed: Theman … interpretatur Auster · Edom qui interpretatur terrenus · Philisthiim, id est cadentes populo · Reseph … reptans ventre interpretatur.

Pattern 7a″ — this work is the reason the rule exists, and it fired everywhere

The Benedictus, Magnificat and Nunc dimittis are the most familiar texts in the Latin liturgy: verses a translator can recite, which is exactly when remembered English arrives before the printed Latin is read. Eleven material divergences marked [var: …] in the final batch alone, ~30 immaterial ones logged. The ones that would have been lost:

• Luke 1:50 — a progenie in progenies, not the Vulgate's in progenies et progenies, and neither is the remembered "from generation unto generation".
• Ps. 86:5 — homo et homo factus in ea for natus est; the gloss then rests on factum explicitly, so "born" would have gutted the christological argument.
• Dan. 3:66/69 — aestas (summer), not aestus (heat) — and the heat is the gloss nine words later, so levelling collapses it into a tautology.
• Ps. 38:5 — ecce veteres posuisti dies meos for mensurabiles.
• Exod. 15:16 @1109C — Irruat super eos fortitudo et pavor for formido; the exposition hanging off it is entirely about fear (Quid timent daemones?), so Rabanus expounds dread out of a word meaning strength.
• Isa. 55:10 — dat semen ferenti for serenti; the remembered "him that soweth" would remove the sowing from an agricultural figure.
• Wisd. 5:4 — aestimabamus infamiam for insaniam.
• The whole Nunc dimittis agrees with the Vulgate word for word — verified twice, precisely because it is the verse most likely to be supplied from memory. Recording the negative result is part of the check.

7a″ extended to APPARATUS, not just lemmata: at 1139D the two-edition formula itself is broken — Alia edito — and was rendered "Another [sic: edito]:" rather than normalized, on the reasoning that a quietly conformed instance of the book's own method is 7a″ transposed. Keep that extension.

Rendered as printed, NOT repaired

• @1156D — ut omnes renascamur in Christo et non aqua baptismatis nostra nobis peccata donentur, in a passage expounding Zech. 13:1 as the laver of baptism. The page therefore says our sins are not forgiven by the water of baptism, against the clause it depends on. Carried, conjecture logged.
• @1136C — Heb. 12:9, and it needs Wilson. Migne prints et reverebamur eos: non multo magis obtemperabimus patri Spirituum et vivemus with a colon and no interrogative. The words are identical to the Vulgate; only the pointing lapsed, and Heb. 12:9 is a rhetorical question. Rendered flat per Pattern 8, so the English asserts what Paul denies. Tested empirically, as on 11065: Migne DOES point questions in this work — five of six Nunquid clauses carry their mark; the low density (1.36 per 1,000 words against 11065's 6.11) is genre, not carelessness. So the flat rendering stands. But this is the second instance of the class and it is sharper here, since it is a quoted scripture whose sense is not in doubt. See the open question below.
• @1092A — Ecce Deus Salvator mens (for meus). mens is a real word, so Pattern 12's boundary forces "Behold, God the Saviour, mind." A third specimen for the real-word-typo convention Wilson still owes (with 11553 @0540B Ei/Et and 11081 @0030B munde/mundo).
• Silent-repair candidates all declined and logged, never quietly fixed: descendunt for defendunt (1125C — "descend from a wound", which reads oddly and should), in fortitudine for in formidine, infirmorum for infimorum, manent for manant, ducta for dicta, abysso for ab ipso.
• Plate irregularities mirrored, not tidied: nine unclosed or inverted guillemet passages across the work (and one apparent case correctly REJECTED at 1116C–D, where the pair balances across a paragraph break), several missing terminal stops, a question closed with a full stop, the 1165A → 1166A column jump. One supplied » was caught by a parity check and removed. Expect a proofreader to want to "fix" these.

▶ OPEN for Wilson — when the pointing lapses on a scripture citation

At @1136C Migne's words match the received text exactly and only his question mark is missing, and the result is English that asserts the opposite of the verse it quotes. Pattern 8 says punctuation follows the plate, never supplied — which is what shipped, here and at 11065 @0150A. The alternative is a narrow exception for scripture citations where the words are identical and only the mark is lost. Recommendation on file: keep Pattern 8 unchanged and mark these on the page instead, so the reader meets the flat sentence and a note that the plate lost the interrogative — the "we never emend Migne" promise stays intact and nobody is misled. Not acted on.

## For the indexer Non-resolving [n: …] references reproduced as printed and itemized per batch: six in 0000–0007 (e.g. (Rom. VI) on a Colossians 3 quotation, (Isa. V) on Isa 53), five in 0015–0021 (Matt. 5:19 as XIX; Ezech. 33:11 as Psal. XL; Matt. 18:10 as Matth. XII; Exod. 32:4 as Exod. XII; Matt. 16:27 as Rom. II). Four inline citations set as running text with a column anchor inside — see data/inline-citations/BRIEF.md, and note (Psal. II) at 1121C is wrong as printed (it is Ps. 23:2) and is left wrong.

Per-batch logs

The four batch logs follow verbatim, in chunk order.

Range: the dedicatory preface to King Louis; the Canticle of Isaiah (Isa. 12); the Canticle of Hezekiah (Isa. 38); the Canticle of Anna (1 Reg. 2); and the Canticle of Exodus (Exod. 15) through 1111D — cols 1089–1111D. Translated 2026-07-31. First English translation.

Work-wide conventions set in this range

Terminology decisions not already fixed by the prompt. Later agents on 8930 and merge-time reconciliation should follow these.

• Title. The ## Commentaria in Cantica head of chunk 0000 is rendered "Commentaries on the Canticles." Not "on the Song of Songs" — the work expounds the ten canticles of Lauds, and the plural Cantica here is the liturgical set, as Rabanus's own preface states (cantica quae in matutinis laudibus sancta psallit Ecclesia).
• The two editions. illam … juxta quam Romana psallit Ecclesia → "that one … according to which the Roman Church sings"; quam ex Hebraica veritate translatam → "translated from the Hebrew verity" (prompt). In the canticles themselves Rabanus most often signals the second text by edition rather than by name, and those formulae are kept fixed: juxta aliam editionem → "according to the other edition" where the pair has just been named and only two are in play (1096A), otherwise "according to another edition" (1104A, 1104C-note, 1105A, 1105D, 1108D–1109A, 1111B); Alia autem editio habet / Alia autem editio sic habet → "But another edition has" / "But another edition has it thus"; Alia editio: → "Another edition:"; Septuaginta … ita posuerunt → "the Seventy set it down thus"; pro quo LXX posuerunt → "for which the Seventy set down"; juxta LXX → "according to the Seventy"; aliqui interpretum → "some of the interpreters"; ut alii interpretati sunt → "as others have translated". "Roman" and "Hebrew verity" are never used as loose synonyms for "another edition" — Rabanus does not label the pair that way inside the expositions, and importing the labels would assert an identification he leaves open.
• Names. Ezechias → Hezekiah (established English form). Anna → Anna, NOT "Hannah": the exposition at 1098D turns on the etymology (Anna, quae interpretatur gratia) and again at 1101B (cujus nomen interpretatur gratia ejus), and "Anna" is also the Douay form, so the printed name stands. Fenenna → Phenenna; Simon at 1099A is kept as printed (the plate's form for Simeon, Luke 2); Manasses → Manasseh; Moyses → Moses; Amalech → Amalec; Philisthiim, Edom, Moab, Chanaan, Sion, Ratestorph, Rasni, Ia, PIDIA all kept exactly as printed.
• feria kept (prompt): "the second feria" … "the sixth feria"; Dominica die → "the Lord's day"; Sabbato → "on the Sabbath". matutinis laudibus → "the morning lauds"; vespertino officio → "the evening office".
• canticum → "canticle" throughout (prompt); carmen → "song"; hymnus → "hymn"; psallitur / canitur / decantatur → "is sung" (no attempt to distinguish; the plate uses them interchangeably of the same act).
• confessio / confiteri → "confession" / "to confess" throughout, never "praise" or "thanksgiving", because both canticles gloss the word's ambiguity explicitly (1091D aequivocum nomen est; 1097B and 1097C confessio … pro laudatione ponitur). Rendering it by its glossed sense would delete the gloss.
• adinventiones (Isa. 12:4) → "devices", kept identical in lemma and exposition (Adinventiones enim Domini praecepta sunt divina).
• gratiarum actio → "thanksgiving"; scriptura as the title of Isa. 38 → "writing" (so that the three competing titles — writing / prayer / thanksgiving — stay three).
• ascensor → "rider", equitatus → "riding", quadrigae → "four-horse chariots", currus → "chariot". The Exodus exposition runs on the horse/rider figure for two columns and the words must not collapse into each other.
• statores (Exod. 15:4 in the second edition, electos ascensores ternos statores) → "attendants", with terni/trini both rendered "threefold". The gloss depends on the number, not on the office (triplex est hominibus peccandi via), so an English word was preferred to carrying the Latin; the choice is flagged here because stator has no settled English equivalent and a later agent meeting it again should follow this rendering rather than invent a second.
• traduces (of vine-layering, 1110C) → "layerings", kept for both the verb-phrase (traduces fieri) and the noun, so Rabanus's viticultural figure survives.
• sanctuarium / sanctimonium (1110D–1111A) → "sanctuary" / "holy place", kept distinct because the sentence sets them side by side as alternatives (Sanctuarium sive sanctimonium dicitur) and then builds on sanctimonium alone for the rest of the paragraph.
• conflatorium → "smelting-place"; conflat → "smelteth" (1107C).
• Italics. The Latin of this range carries italic markup only in the two canticle sub-titles (Regis Judae., Matris Samuelis prophetae.), which are kept italic in the English heads. The English adds no other italics except inside [sic: …] and for Ia / alleluia at 1095A, where the plate is discussing the syllable as a word. Straight quotes " " are used where English needs a quotation mark Migne does not print (Pattern 6); guillemet counts are 1:1 with the Latin twin in all eight chunks.

Defects carried through — [sic: …]

• Head of 0001 — INCIPIT CANTICUM ISALAE PROPHETAE. ISALAE for ISAIAE (the same head's own text names Isaiah in the first line). Carried untranslated in italics: "HERE BEGINS THE CANTICLE OF [sic: ISALAE] THE PROPHET." Pattern 9's section-label exemption was considered and rejected: this is not a structural label with a normalized English exponent (QUAESTIO → QUESTION) but a proper name, and normalizing it would silently mend a misprint in the one place a reader of the English would never think to check.
• 1093C — « Confitemini Domino et invocate nonem ejus, annuntiate inter gentes opera ejus. » nonem is not a word; conjecture nomen, which the same clause carries correctly nine lines earlier (invocate nomen ejus). Carried: "call upon his [sic: nonem]".
• 1095D — Dominus enim poenitens est uper malitiis. uper is not a word; conjecture super. Carried: "For the Lord is one who repents [sic: uper] evils." The [sic:] governs no negation or conjunction (Pattern 7a′ checked).
• 1099B — Distinctione utitur, non dizit, nisi Dominus, sed nullus est… dizit is not a word; conjecture dixit. Carried, and — this is the 7a′ point — the printed non stands outside the marker and is rendered: "he does not [sic: dizit], 'unless the Lord,' but there is none…". The quarantined fragment is two letters; the negative it sits under governs translated text and therefore had to survive into the English.
• 1100A — Saturati prius pro pane se rocaverunt et famelici saturati sunt. (1 Reg. 2:5.) rocaverunt is not a word; conjecture locaverunt, which is the Vulgate's verb (pro panibus se locaverunt, "hired themselves out for bread"). Carried: "They that were filled before have [sic: rocaverunt] themselves for bread". The English lemma consequently has no finite verb — that is the cost of Pattern 12 here, and it is the right cost: supplying "hired" would put the Vulgate's word on the page in place of the plate's.
• 1105A — Notandum vero, quod juxta aliam editiotem legitur. editiotem is not a word; conjecture editionem, which this work prints correctly a dozen times. Carried: "according to another [sic: editiotem]".

Scripture divergence — [var: …]

• 1109C, Exod. 15:16 — Migne prints Irruat super eos fortitudo et pavor in magnitudine brachii tui where the Vulgate has formido. Rendered as printed, "Let strength and dread fall upon them", with a [var: …] giving the received reading. This one is material rather than merely divergent: the exposition attached to the lemma is entirely about fear (Quid timent daemones, quid tremunt? … Timor ergo et tremor cadet super eos), so on the printed text Rabanus expounds dread out of a word that means strength. The divergence is left standing and marked; no emendation.

Cruces: printed readings rendered literally (Pattern 7), no edit made

Each is a real word or a construable phrase, so it is rendered as ordinary English (Pattern 12's boundary) and only logged here.

• 1092A, Isa. 12:2 — « Ecce Deus Salvator mens: fiducialiter agam, et non timebo. » Conjecture meus ("Behold, God my Saviour"), which is the Vulgate and which the following sentence assumes. But mens is a real word, so under Pattern 12's boundary it is rendered rather than carried: "Behold, God the Saviour, mind". This is the one line in the range where the literal rendering is actively strange, and it is deliberate — a reader who meets "mind" and the crux learns that the plate is defective, whereas "my" would have concealed it. Flagged for a later sweep as a candidate for the 11081-class reopening of the marker boundary.
• 1093B, Isa. 12:5 — « Cantate Domini quoniam magnifice fecit. » Genitive where the sense wants the dative Domino. Rendered on the printed case, "Sing ye of the Lord, for he hath done magnificently." Both the parallel at 1093C (Confitemini Domino) and the Vulgate have the dative; English has a case exponent here, so the printed form can be rendered and is.
• 1094A, Isa. 11:11 — ita duntaxat, ut adjiciat secundo manum tuam ad possidendum. The Vulgate and the sense want manum suam; the plate's second person is inside a sentence otherwise addressed to Sion, so "thy hand" is not impossible. Rendered as printed; the possessive was NOT silently switched to "his".
• 1096B, Isa. 38:14 — the lemma prints Domine, vim patior, responde pro me, and Rabanus's paraphrase eight lines later prints tu sponde pro me. Two different verbs, rendered two different ways ("answer thou for me" / "be thou surety for me"). Levelling them to the Vulgate's responde was the obvious 7a″ trap here and was refused; the paraphrase may well be Rabanus glossing rather than misquoting.
• 1096B, Isa. 38:15 — quid respondebit mihi, cum ipse fecerim: first person where the Vulgate has fecerit. Rendered "since I myself have done it"; the exposition immediately below construes it the other way (quia fecit ipse quod voluit, "since he himself did what he willed"), which is rendered as printed too. The plate contradicts itself across ten lines and the English reproduces the contradiction.
• 1095B, Ps. 114:9 — Placebo Domino in regione viventium — Vulg. vivorum. Rendered "of the living"; the whole paragraph turns on terra viventium, so the divergence is doing work and is not marked.
• 1096A — atque spatium vitae, quod inopinate foris est: additum sine augmento praescientiae fuit intus statutum — the sentence has no terminal punctuation on the plate and its clauses do not resolve. Rendered as closely as the words allow ("and the span of life, which unexpectedly is outside: it was added without increase of foreknowledge, having been appointed within") and, per Pattern 8, no full stop has been supplied. The English chunk therefore ends that paragraph unpunctuated, as the Latin does.
• 1097D — Ab hodie enim filios faciam, qui annuntiabunt justitiam tuam, causalis, quae conjunctio sequentia cum superioribus copulat. The bare adjective causalis stands without its noun (conjunctio follows in the relative clause). Rendered "a causal [particle], which conjunction couples what follows with what goes before", the supplied noun bracketed per the supplied-word convention.
• 1099A — Quia et in angustiis pressurarum sermo Dei non est alligatus, nec in praeconibus alligatis. Rendered "nor is it bound in bound heralds" — the jingle is the plate's (2 Tim. 2:9), and both negatives are carried.
• 1100D — quamvis commodius possit intelligi eosdem vivificare, quos mortificaverit, id enim vel repetivit addendo. Elliptical and probably corrupt (vel for velut?). Rendered as printed, "for this he either repeated by adding."
• 1100C–D — Cernimus etiam quae multa in filiis, nunc infirmatam Jerusalem terrenam: no finite verb in the relative clause, and quae multa in filiis is a formula the chunk uses four times. Rendered "We also behold her who was many in sons, the earthly Jerusalem, now made weak."
• 1101B — hoc totus habet sermo, cujus nomen interpretatur gratia ejus. totus … sermo with habet is odd and gratia ejus answers no antecedent cleanly. Rendered literally, "this the whole discourse holds, whose name is interpreted 'his grace.'" The etymology of Anna (1098D) is what it points back to; not emended.
• 1102A–B, 1 Reg. 2:10 — the ejus/suum alternation is the whole point of the paragraph (Dominum formidabunt adversarii ejus … Potest ex ambiguo graeco intelligi et adversarium suum) and is kept distinct throughout: "his adversary" vs. "his own adversary".
• 1102C — « Et dabit imperium regi suo, hoc est, populo christiano: … Sequitur: Et sublimabit cornu Christi sui. » Migne opens the guillemet at Et dabit and does not close it until after Christi sui, so that his own exposition and the word Sequitur: fall inside the quotation. Reproduced 1:1 (Pattern 5); not re-punctuated.
• 1102C — Quomodo Christus exaltabit cornu Christi sui? Sense wants Dominus (the lemma's subject). Rendered as printed, "How shall Christ exalt the horn of his Christ?" — which is not nonsense on Rabanus's own next sentence (head and body are one Christ), so no conjecture is pressed.
• 1103B, Isa. 53:2 — et vultus ejus ignobilis super filios hominum glorificatus est. The final verb belongs to no clause of the quotation and looks like contamination from the glorificatus/honorificatus discussion running around it. Rendered as printed, "he was glorified", inside the quotation where it is printed.
• 1103A–C — the three verbs of Exod. 15:1 are three. Migne prints gloriose enim magnificatus est (1103, the incipit), gloriose enim honorificatus est (twice, 1103A and 1103B), and then argues with glorificatus est / gloriose glorificatus est. Rendered "magnified" / "honoured" / "glorified" respectively, and at 1111D Cantabo Domino, gloriose enim honorificatus est keeps both the first-person Cantabo (against Cantemus at the opening) and honorificatus. The whole exposition is built on the difference between being glorified and being gloriously glorified; levelling any of the four would have destroyed it.
• 1103C — tunc non solum glorificatur. Cum omnes ita honorificant Filium, sicut honorificant Patrem. The non solum has no answering sed etiam, and the following Cum clause stands as its own sentence with the plate's full stop before it. Rendered as printed, negation and punctuation both ("then he is not only glorified. When all so honour the Son as they honour the Father."), per Patterns 7a and 8.
• 1103C — et post adventum humanitatis secundum nobis in gloria ostendit adventum: secundum read as the adjective agreeing with adventum, "shows to us the second coming in glory".
• 1103C, 1104B — equum et ascensorem projecit (1103, 1103D) vs. equum et ascensorem dejecit (1103C): the same verse quoted twice with different verbs, rendered "cast" and "cast down". Not levelled.
• 1105A — plurimos simul plagis torti verberibus cogit. torti has no head noun. Rendered as closely as the words allow, "many at once he compels with strokes and twisted lashes"; conjecture tortis, agreeing with verberibus. Not emended.
• 1107A — Demersi sunt, inquit, tanquam plumbum in aquam validissimam (accusative), against in aqua validissima (ablative) at 1107B and 1107D in the same argument. Rendered "in the mightiest water" throughout, since English has no exponent for the case difference here; logged so a later sweep knows the plate is inconsistent and the English is not.
• 1108A, 1 John 3:2 — « Filioli, nondum scimus quid futuri sumus; » — Vulg. nondum apparuit quid erimus. Rendered as printed, "we do not yet know what we shall be." Not marked [var: …]: the argument that follows is about likeness by grace, not about knowing or appearing, so the sense does not turn on it.
• 1108B–C, Exod. 15:12 — the lemma is printed Extendisti manum tuam et devoravit eos terra, and Rabanus re-quotes it eight lines later as Extendisti dexteram tuam et devoravit eos terra. Rendered "thy hand" and "thy right hand" respectively. This was the sharpest 7a″ trap in the range — the second form is the one the argument uses (the Lord's right hand on the cross), and the temptation is to make the lemma match it.
• 1109B — invenies quia Philisthiim, id est, cadentes populo. Jerome's etymology is cadentes poculo / ruina bibentes. Rendered on the printed word, "those falling by a people"; the name itself is kept as printed per the etymology convention.
• 1109A–B — the two editions of Exod. 15:14–15 differ at every clause (Ascenderunt populi / Audierunt gentes; conturbati sunt principes Edom / festinaverunt duces Edom; obriguerunt / tabuerunt), and Rabanus's own recapitulation then mixes them (vel festinasse, ut dicit, vel iratae esse, festinasse, with festinasse printed twice). All rendered as printed, including the doubled festinasse, which is not marked [d: …]: the two occurrences sit in different syntactic slots (the list and its resumption), so it is Rabanus's clumsiness rather than a compositor's doubling.
• 1110A — Manent ergo lapides nunc, donec transeat populus tuus hic, quem acquisisti, against quem possedisti in the lemma (1109D) and quem acquisivit at 1110A. Three verbs for the same clause within one column, rendered "hast possessed" / "acquired" / "hast acquired". Not levelled.
• 1111A — the column anchor falls inside a word: ne tibi ali [1111A] quid desit. This is a page break through aliquid, not split type (Pattern 10), so no [sic: …]. The anchor is reproduced in the English at the nearest equivalent point in the clause ("that nothing [1111A] be lacking to thee"); exact intra-word placement is not reproducible across the languages.
• 1111B — aliquis sine dubio longior quidem, tamen potitur finis. Rendered "some end, longer indeed without doubt, nevertheless takes hold." potitur normally takes the ablative and has no object here; construed intransitively rather than emended.
• 1107C — the note and the column anchor invert: Migne prints « Modicae fidei, quare dubitasti? » [1107C] [n: (Matth. XIV.)]. Both are reproduced in that order, anchor before note.

Migne's [n: …] references that do not resolve

Reproduced verbatim per the verbatim rule; listed only so the indexer's unparsed/mismatch bucket has a reason attached.

• 1089C [n: (Matth. VII)] on In qua mensura mensi fueritis remetietur vobis — correct (Matt. 7:2).
• 1092A [n: (Rom VII)] — printed without the full stop after Rom; reproduced as printed.
• 1092B [n: (Psal. CXVII)] on Dominus mihi adjutor est, et ego videbo inimicos meos — Ps. 117:7, correct.
• 1098D [n: (Psal. LXX)] on Exaltabuntur cornua justi — Ps. 74:11.
• 1099C [n: (Eccli. III)] on Altiora te ne quaesieris — Ecclus. 3:22, correct; the neighbouring [n: (Eccle. V)] on Dominus in coelo et tu super terram is Eccles. 5:1, correct.
• 1100D [n: (Rom. VI)] on Si mortui estis cum Christo, quae sursum sunt quaerite — Col. 3:1, not Romans. The whole paragraph (through vita vestra abscondita est cum Christo in Deo) is Colossians 3.
• 1103B [n: (Isa. V)] on vidimus eum, et non habebat speciem neque decorem — Isa. 53:2.
• 1105D [n: (Job XLI)] on Aestimabit abyssum quasi senescentem — Job 41:23 (Vulg. 41:22), correct; printed without the full stop after Job.
• 1108A [n: (Joan. X)] on Ego dixi: Dii estis — quoted by John 10:34 from Ps. 81:6; the following Deus stetit in congregatione deorum (Ps. 81:1) is given no note at all.
• 1109C [n: (Rom. X)] on Tota die expandi manus meas — Rom. 10:21, correct (quoting Isa. 65:2).
• 1110B [n: (Psal. LXXIX)] on Vineam ex Aegypto transtulisti — Ps. 79:9–11, correct.
• 1111A [n: (I Cor. III.)] — printed with the full stop inside the parenthesis; reproduced as printed. Note also that this [n: …] falls inside the closing guillemet in the Latin (« Dei agricultura, Dei aedificatio estis [n: (I Cor. III.)] »), and is reproduced there.

Migne's scripture wording, divergent from the Vulgate but rendered as printed

Not marked [var: …] because the sense does not turn on the divergence.

• 1092A, Acts 4:12 — non est aliud nomen super terram datum hominibus — Vulg. sub caelo. Rendered "upon the earth". Rabanus is paraphrasing in his own prose (no guillemets), so no marker.
• 1092A, Eph. 2:5 — « Gratia enim sumus salvati » — Vulg. estis salvati. Rendered "we are saved"; the first person is what makes the sentence Rabanus's own confession.
• 1093B, Phil. 2:10 — omne genu flectitur — Vulg. flectatur. Rendered "is bent" (indicative), not "let every knee bow."
• 1093C, Ps. 104:1 — Confitemini Domino et invocate [nonem] ejus, annuntiate inter gentes opera ejus — Vulg. notas facite in populis adinventiones ejus. The plate's own Isaiah lemma four lines earlier carries the adinventiones wording, so the two are deliberately being compared.
• 1093D, Luke 6:46 — Quid enim vocatis me — Vulg. Quid autem.
• 1094B, Ps. 50:20 — Benigne fac in bona voluntate tua, Sion — the vocative Sion is displaced from the Vulgate's Benigne fac, Domine, in bona voluntate tua, Sion; Domine is simply absent. Rendered as printed.
• 1099A, Luke 2:29–30 — Nunc dimittis, inquit, Domine, servum tuum in pace — the inquit is inside the quotation and stays there (Pattern 5).
• 1104C, Exod. 14:28–29 — per medium sicci maris — Vulg. per medium sicci maris; agrees. quasi pro muro for Vulg. quasi pro muro; agrees. Noted only because the surrounding conflation invites suspicion.
• 1104D, 1 Cor. 10:1–2 — Omnes quidem mare transierunt, et omnes quidem baptizati sunt — the doubled quidem is the plate's; both are rendered ("all indeed … all indeed"). Not marked [d: …]: they govern two different verbs, so this is not a doubled run.
• 1106C, Exod. 15:10 — the chunk quotes the verse twice within four lines, as flavit spiritus tuus et operuit eos mare and then as Misisti spiritum tuum et cooperuit eos mare, and the second form is what the fivefold-Spirit argument needs. Both rendered as printed ("thy spirit blew" / "Thou didst send thy spirit"); operuit and cooperuit kept apart as "covered" (the plate uses the compound only in the second form).
• 1106D — Per spiritum irae tuae divisa est aqua against the lemma's in spiritu furoris tui congregatae sunt aquae (1106B). Rabanus says he is quoting what "he said above"; he is not, and it is left uncorrected.

Negations checked and carried

Every printed negative in the range appears in the English, and none appears that is not printed. Those most at risk of smoothing, checked individually:

• 1089B non cito reprehendat, sed prius consideret utrum majorum sensui concordet et veritati contrarium non sit — the second negative is the one the clause turns on and reads oddly in English ("is not contrary to the truth" after "accords with"); carried.
• 1090A agnoscat hoc non frustra factum — carried.
• 1092B non timet adversarium, quia nocere illi non potest, nisi permissus — all three carried.
• 1092C sed ne putaretur sola pericula declinare quem Domini dona servabant — the negative purpose clause is rendered as such ("lest he … should be thought"), per Pattern 7a′.
• 1092C–D creationem … non semper conditionem eorum quae non erant sed interdum gratiam … intelligamus — both carried; the sentence is about what creation is not.
• 1092D ne videretur alter esse praeter eum / 1092D nequaquam de aquis fluminis Aegypti … neque de aquis fluminis Rasni — carried, including the double nequaquam … neque.
• 1095C Timet ergo, ne … non habitet in quiete, ne non videat Deum … ne … dissolvatur, ne … praecidatur, et nequaquam … oriatur — six negatives in one sentence, including the doubled ne non, all carried. This is the single densest negation site in the range.
• 1095D–1096A nec propheta igitur fallax … nec Dominica statuta convulsa sunt — carried.
• 1096B nec se putat prae doloris magnitudine esse ultra victurum — carried.
• 1097A nec patitur quae refert de humano statu — carried.
• 1097B quia non infernus confitebitur tibi neque mors laudabit te — carried, and the exposition's infernus enim et mors non confitebuntur neque laudabunt te likewise.
• 1097C et non est pulchra laudatio in ore peccatoris. Cumque infernus et mors non confiteantur, nec laudent Deum — carried.
• 1097D quod certe suae non erat potestatis, neque enim per Prophetam illi liberi repromissi sunt … et non benedixit, sed maledixit Deo — all carried; the last is the crux of the Manasseh sentence.
• 1099B Neque enim est alius extra te, et non est fortis sicut Deus noster … quia nemo fit, nisi abs te … nullus tamen eorum naturaliter est … non est, non est* (the shadow and smoke) — carried.
• 1099C Nolite multiplicare / Altiora te ne quaesieris / Omnis sermo malus … ex ore vestro non procedat — carried.
• 1099C–D ignorantes Dei justitiam … justitiae Dei non sunt subjecti … non de Dei posse placere — carried.
• 1100A sine Dei dono atque adjutorio — no printed negative; none supplied.
• 1100A–B ne crederent in Salvatorem / quibus lex illa non erat data / quia in eis non terrena, sed coelestia sapuerunt — carried.
• 1100B In Hebraeo et in Septuaginta translatione non plurimi, sed septem leguntur / Anna autem non plus quam quinque filios habuerit — carried.
• 1100D quoniam littera est, et spiritus non est — carried.
• 1101A Nam qui proprio filio non pepercit / Non derelinques animam meam in inferno — carried.
• 1101B de nullo melius quam de illo intelligo / ut caro ejus non videret corruptionem, nec illud ab illo alienabo — carried.
• 1101C nec solum detrimenta, verum etiam stercora — carried.
• 1102A non viribus nostris, quia non in virtute propria potens est vir … ac per hoc non glorietur potens … et non glorietur prudens … nec dives in divitiis suis — five carried in one period.
• 1102B non parva ex parte intelligit — carried ("In no small part does he understand").
• 1102B Nam quis ascendit … nisi et qui descendit — carried.
• 1102C quoniam non judicabuntur, quae in melius, vel in deterius medio tempore commutantur — carried.
• 1103A tanquam qui sciant, non sua virtute, sed Dei gratia victoriam paratam — carried.
• 1103B quasi non sufficeret glorificatus est / non tamen gloriose glorificatus est (twice) / et non habebat speciem neque decorem / humiliter, non gloriose glorificatus est — all carried; see the three-verbs crux above.
• 1104B nec est ullus qui possit illi resistere. Nolo putes, quia tantum visibiles pugnas conterit Dominus … quae nobis sunt non adversum carnem et sanguinem / non est ulla creatura cujus Dominus non sit — carried.
• 1104C quod nec unus quidem superfuerit ex eis / quia mundi illecebris non permiscetur — carried.
• 1105A etiamsi videntur rogare Jesum, ne eos interim mittat in abyssum / non illi satis est uni equo ascendere — carried.
• 1105D quia non erant tales lapides, de quibus suscitari possent filii Abraham — carried; the whole comparison depends on it.
• 1106D quibus profuit, quod aqua divisa est, non fuit ille spiritus irae Dei? — carried as a question, negative intact. Smoothing this to a statement would have reversed Rabanus's point about the one Spirit.
• 1107A timor incussus est eis qui gratiam nondum intelligebant / conscripta tantummodo jubens sine adjutorio gratiae — carried.
• 1107B sancti autem non demerguntur … et peccati pondere non gravantur / Ipse enim est, qui vere peccatum nescit / non enim tantus erat et talis, qui nihil omnino de specie plumbi in se haberet admistum — all carried, including the difficult nihil omnino clause, which is rendered "as to have nothing at all of the nature of lead mingled in himself" and is immediately contradicted by Habuit, licet parum — the plate's own tension, preserved.
• 1107D–1108A non simulacris gentium comparat Deum, nec daemonibus / nullus tamen Deo similis invenitur / Similitudo tamen haec, non ad naturam, sed ad gratiam revocatur / nondum scimus quid futuri sumus / Nullus ergo in diis similis Domino, nullus enim invisibilis, nullus incorporeus, immutabilis, nullus sine initio et sine fine, nullus creator omnium, nisi Pater — the five-fold nullus plus nisi carried entire.
• 1108C qui ad coelum non respicit, qui futura non cogitat, qui judicium Dei non metuit, nec beata ejus promissa desiderat / in quo animus nihil valet — carried.
• 1108D Nec tamen penitus desperandum est — carried.
• 1109A–B nullum ex his gentibus interfuisse mirabilibus … constat — carried.
• 1109C Non te ergo aliter timebunt, nec aliter tremor tuus veniet super eos, nisi videant in te crucem Christi, nisi et tu poteris dicere / absit gloriari, nisi in cruce — six carried in two sentences.
• 1110A Fieri tanquam lapidem, non est natura esse lapidem; non enim fieret, nisi quod non erat / et non ex arbitrii libertate in haec esse deductos / Non enim dixit, fiant tanquam lapis, et tacuit / Quo scilicet post transitum populi non sint tanquam lapides / Verum non eatenus ut in lapidis natura permaneat — carried; the opening argument of 0007 is nothing but negations and any dropped one inverts it.
• 1110B quae sua esse non dubium est — carried.
• 1110B–C Non vult nos Deus in Aegypto plantare, nec in dejectis … non plantat in vallibus … non vult eos iterum in humilibus collocare … non vult nos per terram repere, nec ultra vult vineam suam … non in quibuscunque humilibus arboribus — seven carried in one period; 1110D* omnis arbor, quae non facit fructum bonum* likewise.
• 1110D non vult te inducere ad laborem, non vult ut ipse tibi facias habitaculum — carried.
• 1110D–1111A hoc non dicit manu hominis factum, sed manu Dei / ne tibi aliquid desit / quod non manu hominum factum est / non enim manu hominum factum est, id est, non opere humano / sine manibus, id est, absque opere hominum — carried.
• 1111B nullum sensum termini alicujus aut finis reliquit — carried.
• 1111C–D ut in medio peccatorum incedentem te non infundat humor peccati … nulla libidinis unda respergat, nullus cupiditatis aestus verberet / non declinat ad dexteram, neque ad sinistram — carried. # Cruces — 8930 (Rabanus Maurus, Commentaria in Cantica), chunks 0008–0014

Range: the Canticle of Habakkuk (Hab. 3), complete — cols 1112A–1132B. Translated 2026-07-31. First English translation.

Work-wide conventions set in this range

Terminology decisions not already fixed by the prompt. Later agents on 8930 and merge-time reconciliation should follow these; where a decision only concerns Hab. 3 it is marked (local).

• The two editions. juxta Hebraicam veritatem / ex Hebraica veritate → "according to the Hebrew verity" (fixed by prompt). Distinguished from juxta Hebraicum → "according to the Hebrew" and from juxta Hebraeos → "according to the Hebrews"; the three are kept apart wherever Migne prints them apart. juxta Septuaginta → "according to the Septuagint"; juxta LXX / bare LXX: → "according to the LXX" / LXX: (the abbreviation is left standing, as Migne prints it). juxta aliam editionem → "according to another edition"; Alia editio: → "Another edition:"; juxta quaedam exemplaria → "according to certain copies"; quidam codices habent → "certain codices have"; juxta nos → "according to us"; juxta historiam → "according to the history"; Solam nostram editionem posuimus → "we have set down our own edition alone."
• auditio vs auditus — the two editions differ at Hab. 3:2 (Hebrew auditionem tuam, LXX auditum tuum) and the exposition then runs on auditus for a full column. Rendered "thy hearing" for auditio and "report" for auditus throughout (auditus = "report" has Vulgate warrant at Isa. 53:1 / Rom. 10:16, auditui nostro). Levelling them would have erased the version difference the passage is built on.
• arcum tendere/intendere/extendere/suscitare — four verbs, kept four ways, because the two editions differ precisely here: suscitans suscitabis → "raising up thou wilt raise up"; intendens → "bending"; tendens → "drawing"; extendes → "thou wilt stretch forth."
• canticum → "canticle" throughout (prompt); carmen → "song"; Canticum canticorum → "the Canticle of Canticles."
• christos (lower-case, Hab. 3:13 ut salvos facias christos tuos) → "christs", not "anointed ones". The gloss depends on the word: quos … christos, id est, unctos suos fecit is rendered "whom he made christs, that is, his anointed ones," and only survives if christos keeps a form of its own.
• diapsalma, Sela — kept in italic Latin with Rabanus's own gloss ("always") beside them, as untranslatable liturgical/textual terms.
• jumentum → "beast of burden" (consistently, incl. Ps. 72 Jumentum factus sum apud te).
• tentoria / tabernacula / pelles (Hab. 3:7, where the two editions differ) → "tents" / "tabernacles" / "skins", each kept distinct.
• prae laboribus → "by reason of the labours" (both in the lemma and in the exposition, so the echo holds).
• commatice per capitula → "clause by clause, section by section" (local).
• Etymological glosses keep the printed name and translate only the gloss: Theman … Auster → "Theman … the South"; Pharan os videntis → "Pharan … the mouth of one seeing"; Reseph reptans ventre → "Reseph … creeping on the belly"; Madian … ex judicio → "Madian … from judgment"; Jesus … Salvator → "Jesus … Saviour."
• Italics. The Latin of this range carries no italic markup at all; the English therefore adds none for repeated lemmata (Rabanus re-quotes his lemma constantly), reserving … for carried type inside [sic: …] and for Latin words deliberately left in Latin (diapsalma, Sela). Straight quotes " " are used where English needs a quotation mark Migne does not print (Pattern 6).

Defects carried through — [sic: …]

• 1124A — receperunt lumen suum et pristmum habuere fulgorem. pristmum is not a word; the conjecture is pristinum ("their former brightness"), which the sentence requires and which differs by one letter-group. Carried untranslated in italics per Patterns 7/12: "had their [sic: pristmum] brightness."
• 1125D — et inferendo fororem condemnabis eos in perpetuum. fororem is not a word; conjecture furorem, which the lemma four lines above supplies (In comminatione tua minorabis terram et in furore tuo). Carried: "and by bringing in [sic: fororem] thou wilt condemn them for ever."
• 1128D — « Et introivit tremor in ossa mea. » Sicu carnis nomine aliquoties Scriptura… Sicu is not a word; conjecture Sicut, required by the answering ita in the same sentence. Carried at the head of the clause, so the English sentence opens "[sic: Sicu] by the name of flesh Scripture sometimes designates…" — the correlative "so" is left standing because ita is printed.

Scripture divergence — [var: …]

• 1131C, Hab. 3:19 — Migne prints deducet me victori in Psalmis canentem where the Vulgate has the nominative victor. Rendered as printed, "he will lead me, singing in the Psalms, to the victor," with a [var: …] giving the received reading. The dative changes who is doing what: on Migne's text the prophet is led to a victor; on the Vulgate's he is himself the conqueror. Rabanus's own exposition a few lines later (Canam autem victoriam ejus et triumphum) sits comfortably with either, so no conjecture is offered.

Cruces: printed readings rendered literally (Pattern 7), no edit made

Each of these is a real word or a construable phrase, so it is rendered as ordinary English (Pattern 12's boundary) and only logged here.

• 1122B — mentem a coelestibus desideriis in appetitum demergunt infirmorum: rendered "into the appetite of infirm things." Sense wants infimorum, "of the lowest things," which is the standard antithesis to coelestibus and which this same work uses at 1129B–C (relicto infimorum appetitu, in infimis deprimitur). One letter apart; not emended.
• 1122C — ad ejus quae sursum est Jerusalem, quae est mater omnium nostrum videnda regna ducuntur: the accusative plural has no governing preposition and no clear head. Rendered as closely as the words allow, "are led through the apostles to the kingdoms that are to be beheld of that Jerusalem which is above." Unconstruable as printed; no conjecture proposed.
• 1123A — quae terrae concussione rumpantur et manent flumina: the indicative manent after subjunctive rumpantur looks like manant, "flow." But "remain as rivers" construes and fits the geological anecdote (hidden waters burst out and stay as rivers), so it is rendered "and remain as rivers." Not marked.
• 1125C — jaculis quippe bellantes adversarios feriunt, armis muniendo sese a vulnere descendunt: rendered literally, "with arms, by fortifying themselves, they descend from a wound." The near-certain reading is defendunt ("they defend themselves from a wound"), which is what the following sentence's contrast (darts wound the enemy, arms protect the bearer) requires. Rendering "defend" would be exactly the silent repair Pattern 7a forbids: descendo is a real word and can be rendered, so it is.
• 1120C — quod metu judicii aeternarumque poenarum semper in fortitudine sint, et cruciatus, quos se sciunt quotidiano pavore sustinere: rendered "they are always in strength, and the torments which they know themselves to endure with daily dread." The context is dread, not strength; conjecture in formidine. The second limb (et cruciatus) has no verb of its own on the printed text and is carried as an appended accusative.
• 1130C — ductaque prophetae veracia falso ore resonaret dicens: rendered "and re-echoed with a false mouth the truthful things drawn from the prophet." ducta is a real neuter plural participle and construes with veracia, so it is rendered; the expected reading is dictaque, "the truthful sayings of the prophet," which the quotation that follows (Ps. 51) supports.
• 1130A — eo quidem pereunte electus sum populo: rendered "while that one indeed perishes I have been chosen for the people." The dative populo after electus sum is unusual and the sentence is elliptical; rendered as printed.
• 1123C, Hab. 3:13 — Migne prints the bare participle egressus in salutem populi tui where the Vulgate has egressus es. Rendered "having gone forth for the salvation of thy people," not "thou wentest forth": supplying the finite verb would import a word the plate does not carry, and the same verse reappears at 1125D in the finite form Migne does print there (Existi in salutem populi tui), which the English keeps distinct.
• 1119B — Stant autem in pedes et movetur terra: accusative after stare; rendered "But they stand upon feet and the earth is moved." Cf. the lemma it glosses, Pedes ejus steterunt et mota est terra.
• 1119C — Denique Saulus et Matthaeus montes erant, hic de sapientia litterae carnalis, ille de mammona iniquitatis elatus: on the ordinary force of hic/ille the nearer term (Matthew) takes the wisdom of the carnal letter and the further (Saul) the mammon of iniquity, which reverses the historical fit. Rendered with the same ambiguity — "this one … that one" — so the English neither repairs nor resolves the printed order.
• 1128C — Ventrem suum more prophetis consueto animum suum dicit: the dative prophetis with more … consueto; rendered "after the manner customary to the prophets."
• 1128D — omne quod agitis, cujus sincera intentio est Domino placendi: a second-person plural in a third-person sentence; rendered as printed, "everything that you do."
• 1129A — et quo altior efficitur lumine contemplationis, eo se imperfectionem conspicit: se stands without an obvious construction beside the accusative imperfectionem; rendered "the more does he behold his own imperfection."
• 1131A — sed quia pabulum coelestis intellectus in his non sapiunt, qui suave jugum Evangelii bajulent absunt: rendered "those who would bear the sweet yoke of the Gospel are wanting," taking absunt with the relative clause as its subject.
• 1131C — ut calcem super aspidem et basilicum: basilicum for basiliscum (Ps. 90:13). Rendered "the basilisk" and not marked [sic:]: basilicus for the basilisk is attested in medieval Latin, so this falls under Pattern 9 (attested spelling), not Pattern 12. Logged because the printed form is also a distinct word (basilicum, basil) and a later sweep may want to revisit the call.
• 1131C — extraham serpentem Judam de colubro: rendered as printed, "draw out the serpent Judas from the snake." No conjecture; the phrase has no obvious biblical antecedent and Judam may be a proper name Rabanus intends. mittam in foramen has no object; "[my hand]" is supplied in brackets per the supplied-word convention.
• 1115D — cum duodenis in medio doctorum residens: read as duodennis, "at twelve years old" (Luke 2:42–46, which the sentence is narrating). A single n for a double is orthographic, not a defect: Pattern 9, rendered normally.
• 1114D — credo quod otius misericordiam … consequar: otius for ocius; attested spelling, rendered "the sooner," Pattern 9, no marker.
• 1130C — oleo adolationis impinguaret capita: attested spelling of adulationis; rendered "of flattery," Pattern 9, no marker.
• 1117D — ut nec ipsa cornu eorum, regum videlicet insolentium potentia posset infringi: cornu read as ablative (4th decl.), with ipsa picking up charitas: "so that it could not be broken even by their horn, namely by the power of insolent kings." Reading cornu as nominative subject would make the sentence say the horn could not be broken, which contradicts the argument. Rendered on the ablative reading; noted because it is a construal, not a certainty.

Quotation marks — unclosed « in Migne (Pattern 5 corollary)

Three quotations are opened and never closed on the plate. The English mirrors the plate exactly; no » has been supplied. These are the reason the guillemet counts are 28/27, 22/21 and 16/15 in those chunks — they are not omissions.

• 1121D — Unde et Moyses loquitur ad Deum: « Ego autem sum irrationabilis; et David: « Jumentum factus sum apud te…, » — two « and one ». The Moses quotation is never closed and the David quotation opens inside it.
• 1125A — the 2 Cor. 6:11–13 quotation (« Os nostrum patet ad vos, o Corinthii …dilatamini et vos.) runs straight into Rabanus's own sentence Unde autem vel quo ordine provenerit… with no closing mark.
• 1130D — the Ps. 51 quotation (« Ego autem sicut oliva fructifera…) runs into Propter quod tempore ultimae retributionis… with no closing mark.

A fourth case is not a defect and should not be "fixed": at 1116C–D the Luke 2 quotation « Gloria in excelsis Deo… is left open at the end of one paragraph and its » falls at the end of the Hab. 3:4 lemma in the next (…Ibi abscondita est fortitudo ejus. »). The pair is balanced across the paragraph break, and the English reproduces it in the same two places. The column anchor [1116D] is set hard against the following word in the Latin ([1116D]Splendor) and is reproduced hard against it in the English.

Migne's inline citations printed as italic parentheses, not as notes

At 1121C and 1121D two scripture references are set in the running text as italic parentheses with a column anchor inside them, instead of as [n: …] notes: (Psal. [1121C] II) and (Cant. [1121D] I). Both are reproduced verbatim, italics, anchor and all. No [f: …] tag was applied: this work's citing practice is [n: …] (nine notes in this chunk alone), and two stray instances are not a practice (Pattern 4's trigger). Worth flagging to the indexer, since neither will be harvested as a note.

Both references are also wrong as printed, and are left wrong (Pattern 7 / refDisplay principle): (Psal. II) heads Ipse super maria fundavit eam, which is Ps. 23:2; (Cant. I) is correct for Cant. 1:8.

Migne's [n: …] references that do not resolve

Reproduced verbatim, unaltered, per the verbatim rule; listed here only so the indexer's unparsed/mismatch bucket has a reason attached.

• 1114A [n: (Matth. III)] on Hic est Filius meus dilectus … ipsum audite — the ipsum audite clause is the Transfiguration (Matt. 17:5), not Matt. 3:17. At 1118A the same words without ipsum audite are cited [n: (Matth. III)] correctly.
• 1126C [n: (Act. V)] on Stephen's Dura cervice et incircumcisi — Acts 7:51.
• 1129B [n: (Rom. X)] on Spe gaudentes, in tribulatione patientes — Rom. 12:12.
• 1131C [n: (Psal. XXVIII)] on Qui perfecit pedes meos quasi cervi — Ps. 17:34.
• 1112B [n: (Habac. I)], 1112B [n: (Ibid.)] — Hab. 1:2 and 1:13; the book-level citation is correct and is left as printed.

Migne's scripture wording, divergent from the Vulgate but rendered as printed

Not marked [var: …] because the sense does not turn on the divergence; logged so that a later sweep is not surprised by them.

• 1112B respicere ad iniquitatem non potes (Hab. 1:13) — Vulg. poteris.
• 1113A apparebit in fine vs 1114C apparebit in finem — the same verse (Hab. 2:3) quoted twice by Migne with different case. Both rendered as printed ("at the end" / "unto the end"). The divergence is internal to the plate and is exactly the kind of thing this edition exists to show.
• 1112A Oratio Habacuc prophetae pro ignorationibus — Vulg. ignorantiis. Rendered "for ignorances" wherever it recurs (the phrase is the canticle's title and returns at 1128D), so the title reads identically in both places.
• 1121B fluent de ventre ejus flumina aquae salientis in vitam aeternam [n: (Joan. VII)] — a conflation of John 4:14 with John 7:38, rendered as printed and cited as printed.
• 1130C in loco pascuae ibi me collocavit (Ps. 22:2) — Vulg. in loco pascuae ibi me collocavit agrees; the preceding Dominus regit me et nihil mihi deerit agrees also. No divergence; noted only because the verse is quoted in a stretch dense with divergent quotation and reads as if it might be.

Negations checked and carried

Every printed negative in the range appears in the English. The ones most at risk of being smoothed away, checked individually:

• 1115A revelatur ut terreat, et territis non inferatur — the closing clause of 0008 turns entirely on this second negative.
• 1115A qui autem irascitur, interdum non percutit — rendered "sometimes does not strike," not "does not always strike": the negation governs percutit, and interdum is not part of it.
• 1119C nequaquam homini praedicanti, sed illustranti est gratiae tribuenda.
• 1122B ut in talibus ira, quam merentur, nunquam solvenda perseveret.
• 1126D haud dubium quin impii — the haud is carried ("there is no doubt but that they are the impious one's").
• 1130A Et nisi Dominus reliquisset nobis semen (Rom. 9:29).
• 1131C nullam etiam aliam, nisi in Christo, habebo virtutem — both nullam and nisi carried.
• 1128C ne forte corde peccarem, neve aliis praedicans, ipse reprobus efficerer.
• 1129B qui sine peccato intraret in mundum, non sine poena peccati exiturum — the second negative is what the whole paragraph is for. # Cruces — 8930 (Rabanus Maurus, Commentaria in Cantica), chunks 0015–0021

The Canticle of Deuteronomy (Deut. 32), complete, cols 1132C–1151D.

Work-wide conventions established here

Terminology decisions taken in this batch, beyond the pre-set vocabulary. A later agent working other chunks of 8930 should follow these or record a reason not to.

• The two editions. Alia editio → "Another edition", invariably, and always introducing the second of the pair (Rabanus prints the Roman lemma first, then the other). juxta Hebraicam veritatem / ex Hebraica veritate → "according to the Hebrew verity" (1137A, 1150C, 1151A); editio LXX interpretum / illa quae ex LXX interpretibus → "the edition of the LXX interpreters" (1150C, 1151A) — never "the Septuagint," which is not what the page prints. septuaginta interpretes spelled out at 1134C → "the seventy interpreters," as printed. The two Englishes are never levelled: where the Latins differ the renderings differ, word for word, even in formulaic verses (see the checklist below).
• carmen → "song"; canticum → "canticle." Both appear in 1132C–1132D (verba carminis hujus, principium hujus carminis) against the head's CANTICUM. The distinction is Rabanus's and is kept throughout.
• theoria stays Latin/Greek — ad illam coelestem theoriam (1132D) is glossed by Rabanus himself (hoc est, ad inspectionem secretarum rerum), so an English substitute would make the gloss explain nothing.
• Etymologies keep the name as printed: Basan pinguedo interpretatur (1140D) → "Basan is interpreted fatness."
• Sense-labels: secundum litteram → "according to the letter"; secundum historiam / juxta historiam → "according to the history"; mystice → "mystically"; allegorice → "allegorically"; tropologia → "tropology" (all adverbs where the plate prints adverbs, per Pattern 9).
• Verbs of the two-edition method: subjungitur → "it is subjoined"; Sequitur: → "There follows:" — note 1151B prints Sequitur. with a full stop where every other instance has a colon; kept as printed (Pattern 8).

Pattern 12 — carried type ([sic: …])

Pattern 14 — scripture divergence ([var: …])

• 1133C, Isa. 55:10. Migne prints et dat semen ferenti et panem comedenti; the Vulgate has det semen serenti. "Him that beareth" vs "him that soweth" — material, because the whole comparison is agricultural and the sowing is the point of the figure. Rendered as printed, marked [var: Vulg. serenti…].
• 1151C, Wisd. 5:4. Migne prints nos insensati vitam illorum aestimabamus infamiam; the Vulgate has aestimabamus insaniam. "A disgrace" vs "madness." Marked. This is a 7a″ near-miss: the familiar English ("we esteemed their life madness") arrives before the printed word is read.

Divergences logged but not marked

Citations that differ from the received text without changing the argument, and misattributed [n: …] locators. The notes pass through verbatim per the sacred-marker rule; nothing is corrected in the text.

• 1132D, [n: (Matth. XIX)] — the quotation (Qui solverit unum de mandatis istis minimis…) is Matt. 5:19, not 19. Locator kept as printed.
• 1133D, Gen. 2:1–2 — printed complevitque Deus die septimo opus suum; the Vulgate has complevitque Deus die septimo opus suum quod fecerat, which the plate also gives. No divergence; noted only because the sabbath argument turns on it.
• 1134A, Ps. 18:8 — printed Lex Domini irreprehensibilis, Vulg. immaculata. Rendered "blameless." Not marked: the exposition (quia omnia mandata ejus vera) rests on the following clause, not on this adjective.
• 1134A, Ps. 18:10 — printed Judicia Dei vera, Vulg. iudicia Domini. Divine name only.
• 1134B–C, Joan. 8:44 — printed ille homicida fuit ab initio et latro. The Vulgate has et in veritate non stetit; latro has no counterpart in the received text at all. Rendered "and a robber." Not marked because Rabanus builds nothing on the word, but it is the most substantial free citation in the batch.
• 1134D, Jer. 14:7–8 — printed Tibi peccavimus patientia Israel, Domine, conflating Jer. 14:7 (peccavimus tibi) with 14:8 (expectatio Israhel). Rendered literally, "O patience of Israel." Rabanus's argument is about the dative tibi, which is intact.
• 1135A, [n: (Psal. XL)] on Nec vult mortem peccatoris — the sentence is Ezech. 33:11, not a psalm; and the same locator was used two lines earlier for a genuine Ps. 40 quotation. Kept as printed (a mis-bound locator, not our repair).
• 1135B, Ps. 77:39 — printed Quoniam caro sunt et Spiritus ambulans et non revertens; Vulg. spiritus vadens et non rediens. Synonym-level.
• 1135B, Matt. 12:39 — printed Generatio prava et perversa signum quaerit; the Vulgate has mala et adultera. The substitution is almost certainly deliberate — Rabanus is adducing the verse precisely because it echoes Deut. 32:5's generatio prava atque perversa, the lemma under exposition. Rendered as printed; not marked, because here the divergence IS the argument rather than a threat to it.
• 1137B, [n: (Matth. XII)] on Angeli eorum in coelis… — the verse is Matt. 18:10.
• 1136A, [n: (Exod. XII)] on Hi sunt dii tui, Israel — Exod. 32:4.
• 1150A, [n: (Rom. II)] on Filius enim hominis venturus est… — the verse is Matt. 16:27; Rom. 2:6 carries only the closing clause (reddet unicuique secundum opera sua), which is presumably how the locator arose.
• 1148A, Et in servis tuis miserebitur — Rabanus re-quotes his own lemma, which read in servis suis, with a second-person possessive. Rendered "on thy servants," against the "his servants" eleven lines above. Not levelled: the shift is on the plate and the two readings sit in plain sight of each other.

Pattern 5 / 6 — quotation marks reproduced as printed

Three passages where Migne's guillemets are defective. All mirrored 1:1; none supplied, none removed.

• 1133A, unopened close. The Isaiah frame opens « at Audi, coelum and does not close until …verba oris mei, » — nine lines later, after an intervening block of Rabanus's own prose and a second (unmarked) quotation of Deut. 32:1 in the singular (Attende, coelum). One «, one », spanning both. Reproduced as one span.
• 1134D, unclosed quotation. « Si in Deum peccaverit, quis orabit pro eo [n: (I Reg. II)]. never closes. The » is not supplied (Pattern 5 corollary); this is the single guillemet-count mismatch in chunk 0015 and it is Migne's.
• 1142B, doubled and inverted marks. The plate prints « Ne projicias me a facie tua [n: (Psal. L)]. « Et item: » Ne avertas, inquit, faciem tuam a me. » — an opening « where a close belongs, and a closing » before the quotation it should follow. Reproduced exactly, marks and all.
• 1143D, stray close inside a citation. …quae qui agunt, » ut Paulus testatur, » « regnum Dei non consequentur… — two » in five words. Reproduced.
• 1151C, unclosed quotation. The Wisd. 5 block opens « at Gementes enim and the chunk ends with the note and no ».
• 1149B: the plate prints « Alia editio: Ego occidam… — the guillemet opens before the edition formula, enclosing it inside the quotation. Kept.

Pattern 8 — punctuation and sentence boundaries as printed

• 1132D: Ideo ipse Dominus in Evangelio ait - « — a spaced hyphen where the plate everywhere else has a colon. Kept.
• 1137C: « Funes ceciderunt mihi, in praeclaris - etenim haereditas mea… — the same spaced hyphen inside a psalm quotation. Kept.
• 1139C: the paragraph on the unity of God ends …quot erroribus, tot idolis obnoxii sunt with no terminal stop. None supplied; the English stops where the Latin stops.
• 1143A–B: …et id quod alias scriptum est. Jesus populum de terra Aegypti salvans… — a full stop breaking what is grammatically one sentence (the second object of non attendunt). Kept, and the et rendered "and," not "nor": smoothing the connective under a governing negative is Pattern 7a mechanism 3.

Cruces of construction

• 1136C, Hebr. 12:9. Migne prints …et reverebamur eos: non multo magis obtemperabimus patri Spirituum et vivemus… — with a colon, no interrogative mark. The Vulgate's clause is a rhetorical question (shall we not much more be in subjection…?), which is how the non earns its sense. Rendered flat and negative as printed: "we shall not much more obey the Father of spirits, and we shall live." No ? supplied (Pattern 8) and the non carried (Pattern 7a), with the consequence that the English asserts what Paul denies. That is the printed reading; the conjecture is that Migne's colon stands for a lost question mark.
• 1141C, ex beneficiis largissimus pejores facti. largissimus is nominative singular where the ablative plural largissimis is required by ex beneficiis. Pattern 9: a case mismatch has no English exponent — "made worse out of most bountiful benefits" is the only available rendering either way. No emendation to make or undo.
• 1145A, pro peccato Israel indignantem et eas disperdere volentem. eas (feminine plural) for a masculine collective. Pattern 9 again — no exponent in English; rendered "them."
• 1143A, duplici confusione contere eos. Jer. 17:18 reads duplici contritione contere eos. confusione is a real word wrongly set, not carried type, so it is rendered as ordinary English ("with a double confusion crush them") per Pattern 12's boundary. The sentence is also syntactically broken — an imperative quotation spliced into the objects of non attendunt — and is rendered as closely as the printed words allow.
• 1143C, a foris sine filiis privabit eos gladius. Double construction: privabit eos wants an ablative of the thing removed, and sine filiis supplies a preposition instead. Rendered "the sword shall deprive them, leaving them without sons"; the Roman column's foris vastabit eos gladius is untouched by it.
• 1146A, quod etiam non solum nostri, verum etiam nostrorum approbant adversarii. Compressed to the point of ambiguity: nostri (our own men) against nostrorum adversarii (the adversaries of our own). Rendered as the words stand.
• 1147A, Qui non est Deus noster, sicut dii illorum. The other edition's rendering of Deut. 32:31 opens with a relative where the Roman column has Non est enim Deus noster ut Deus eorum. Rendered literally ("Who is not our God, as their gods are") rather than conformed to the Roman reading beside it — the whole point of the parallel is that they differ.
• 1151C, quando viderunt gloriosam apparere… — perfect indicative under a future main verb (Laudabunt omnes gentes), where viderint is expected. Rendered "when they have seen," which is available for the printed perfect; no emendation.

Negation checklist

Every printed negative in the batch was checked into the English individually. The dense sites, all carried: 1134B peccantibus et non poenitentibus; 1134C fidem ejus recipere nolentibus; 1134C quem prophetarum non sunt persecuti … et non custodistis; 1134D qui peccat, non Deo peccat, id est, non Deo nocet; 1134D ne peccares in me; 1135A condignam poenitentiam … non agunt; 1135A Nec vult mortem peccatoris; 1135B nec poenitentiam agere vellent; 1135C nec ei pro hoc condignas laudes … rependebat; 1136B patres non quilibet hebetes; 1136C non multo magis obtemperabimus (see above); 1136D numerositas … comprehendi non possit; 1136D non quia cunctas gentes numero vincebatis; 1137A ultra non differat; 1137D ubi non erat aqua; 1138A mandata Dei servaverunt … nulla mentio Dei; 1139C non idola falsa … non alius, quam idem Deus; 1139D quia non sterilis; 1140A nusquam … nusquam … nusquam (three, all rendered "nowhere"); 1140D nec sufficiebant haec dona, nisi adderentur; 1141A non enim satis est … nisi poenitentiam agant; 1141D nequaquam fecerunt; 1142B ejus misericordiam non merentur; 1142B non conturbato animo; 1142C qui non erat Dominus … qui non est populus … in quibus non est fides … et non in Deo … et non in gentem; 1142D si pro malis suis non poenituerint; 1143A Percussisti eos, nec doluerunt … renuerunt accipere disciplinam; 1143A Non vindicat Deus bis in idipsum; 1143A non attendunt; 1143A quamlibet culpam bis ferire non posse; 1143D regnum Dei non consequentur; 1144A sine ulla miseratione; 1144B Vermis eorum non morietur, et ignis non exstinguetur; 1144B nulli aetati … esse parcendum; 1144B a peccatis non cohibet; 1144C memoria eorum … non numeretur; 1144D ne forte superbirent … et non Deus fecit haec omnia (both editions); 1145A Ne, quaeso, Domine, dicant Aegyptii; 1145A ne faceret malum; 1145A nolens perire quemquam; 1145B Ignorans quoniam; 1145B impoenitens cor; 1145C sibi providere nesciunt; 1145C disciplinam non sapuerunt; 1145C non exaudiam … non invenient me … timorem Domini non susceperint; 1145D non ex infirmitate potentiae … nullo modo fieri posset … nisi Dominus eos tradidisset; 1146A Non est enim Deus noster … Qui non est Deus noster; 1146C recipere noluerunt; 1146D nisi se inde liberaverit; 1146D ostendit non sanae esse mentis; 1147A ubi praeteritum et futurum non est; 1147A nec oblivioni tradentur; 1147B si se a peccatis non correxerint; 1147C nec diuturnum esse potest; 1147C nunquam perfecte destiterunt; 1148B nec unquam in tristitiae moerorem commutatur; 1148D ultra non praevalebunt … non habebunt ulterius potestatem; 1148D nec opus, nec ratio, nec sapientia, nec scientia; 1148D nullum adjutorium … sed nec mundi amatores; 1149A nihil secum conferunt; 1149A non sit alius Deus praeter me / non est praeter me; 1149B non est, qui possit de manu mea eruere / non est qui eripiat de manibus meis; 1149C nullo Dei amore sauciantur … aerumnam non sentiunt … nec quolibet minimo affectu; 1149D regnum ejus, quod non corrumpetur; 1150C quae in Hebraica veritate non inveniuntur; 1151C gloriam sanctorum … quam non speraverunt.

No negation was added where the plate prints none. # Cruces — 8930 (Rabanus Maurus, Commentaria in Cantica), chunks 0022–0026

Range: PL 112, cols 1152A–1166A. Four canticles: the Hymn of the Three Children (Dan. 3), the Canticle of Zechariah (Benedictus), the Canticle of Mary (Magnificat), the Blessing of Simeon (Nunc dimittis).

Work-wide conventions set in this range

Terminology decisions not already fixed by the batch prompt. Later agents on this work should follow these or record a reasoned departure.

• in saecula → "unto the ages"; in saecula saeculorum → "unto the ages of ages"; in saeculum → "unto the age". Not "for ever" / "world without end" — the refrain is the load-bearing phrase of the Dan. 3 hymn and Rabanus glosses it (ut ipse super omnia exaltatus in saecula regnans praedicetur, 1153B), so it is kept lexically transparent.
• superexaltate / superexaltatus / superexaltemus → "exalt … above all" / "exalted above all". "Superexalt" is not English; "exalt above all" keeps the intensive force and matches the gloss at 1153B (super omnia exaltatus). The full refrain Laudate et superexaltate eum in saecula therefore renders "Praise and exalt him above all unto the ages" throughout.
• Benedicite, X, Dominum → "Bless ye the Lord, O X." See the accusative note below.
• canticum → "canticle" (prompt-fixed); hymnus → "hymn"; carmen (1162C) → "song" — Rabanus distinguishes them and carmen there points back to the Magnificat as a composed piece.
• Vetus Instrumentum (1159D) → "the Old Instrument", not "Old Testament". Rabanus uses Testamentum separately (1164D, Novi … Veteris Testamenti) and the two must stay distinguishable.
• oriens (Luke 1:78; Zach. 6:12) → "the Orient", capitalized, since the gloss turns on it as a title (recte oriens vocatur).
• prelati … subditi → "prelates … subjects" (1157C, 1158A).
• plebs / populus / gens → "people / people / nation" respectively; omnis Israel → "all Israel".
• ἅγιον (1162D) kept in Greek, verbatim, unaccented-normalization not applied.
• Etymologies: the name is kept as printed and the etymon set in italic Latin — Aevae … vae → "the name of Eve (Aeva) — that is, of vae, woe"; Idithun, hoc est, transiliator → "Idithun, that is, the overleaper".

⚠ THE HEADLINE ITEM — printed lemma vs. Vulgate wording

This is the list the batch prompt asked for, and it is the most valuable output of the range. Every scripture lemma in cols 1152A–1166A was read word by word against the printed Latin before and after translating. Divergences are grouped by whether they were material enough to earn a [var: …] (Pattern 14).

A. Divergences MARKED with [var: …] in the English

B. Divergences LOGGED but NOT marked — no material change of sense

Recorded here so a later reader is never told by silence that a lemma agrees with the Vulgate (Pattern 14's standing caution).

• The whole Dan. 3 refrain is accusative, not dative. Migne prints Benedicite … Dominum at every one of the ~30 occurrences in 0022–0023, and Benedicite Dominum omnia opera ejus in the Ps. 102:22 citation at 1153A, where the Vulgate has Domino. This is systematic, not sporadic — it is the text-form Rabanus is expounding. It has no English exponent ("bless the Lord" either way), so it is logged as a Pattern 9-type case and left unmarked. A later agent must not "fix" a chunk to the dative or think it a defect.
• 1152C Dan. 3:51 benedicebant Dominum in fornace — Vulg. Deum.
• 1153A Ps. 102:22 Benedicite Dominum omnia opera ejus — see above.
• 1153D Gen. 1:7 Divisitque aquas quae erant super firmamentum — Migne quotes only the second half of the verse; not a variant, an ellipsis.
• 1154A Ps. 102:21 omnes virtutes ejus et ministri ejus — Vulg. has no et.
• 1154B Ps. 148:3 — agrees with the Vulgate exactly, including et lumen.
• 1155B 2 Cor. 4:17 in sublimitatem for in sublimitate; Id, inquit, for id enim (the inquit is an interruption, kept inside the guillemets per Pattern 5).
• 1155C Eccli. 3:17 Sicut in sereno glacies, ita solventur peccata tua — ita added.
• 1156A Ps. 23:1 qui habitant in ea — Vulg. in eo. No English exponent (both render "in it"/"in her" ambiguously); left unmarked.
• 1157A Gen. 1:10 Congregationes aquarum for congregationesque; Eccle. 1:7 intrant in mare for intrant mare.
• 1157A 2 Cor. 5:10 unusquisque propria corporis recipiat for referat, and quatenus for ut. Rendered "receive"; the sense is unchanged.
• 1157B Dan. 3:79 ceti for cete (spelling); 1152D Rabanus's own prose has coetos — both rendered "whales", both attested variants (Pattern 9).
• 1157C Dan. 3:81 omnes bestiae et pecora for bestiae et universa pecora.
• 1158B Matt. 11:29 — agrees exactly.
• 1158C Dan. 3:88 Anania, Azaria et Misael — et added.
• 1159B Luke 1:68 Benedictus Dominus Deus Israel … redemptionem plebis suae — Vulg. Benedictus Deus Israhel … plebi suae. The added Dominus is the liturgical form and is what Rabanus expounds; the genitive for the dative changes nothing in English ("the redemption of his people").
• 1160B Luke 1:74 de manibus inimicorum nostrorum — Vulg. de manu.
• 1160C Gen. 12:3 atque in te benedicentur for et benedicentur in te.
• 1160C 1 Par. 17:11–13 — agrees exactly, including stabiliam regnum ejus.
• 1161C Eph. 5:8 Fuistis aliquando tenebrae for eratis enim.
• 1161D Ps. 33:15 Inquire … pacem et sequere eam for persequere eam.
• 1162A Eccli. 3:20 — agrees exactly.
• 1162C Ps. 34:9 et delectabitur super salutare ejus for delectabitur super salutari suo — a case-and-pronoun shift with no English exponent.
• 1163A 1 Pet. 1:16 Sancti estote, quia et ego sanctus sum — Vulg. 1 Pet. 1:16 sancti eritis quia ego sanctus sum (after Lev. 11:44).
• 1163C Matt. 23:12 quia omnis qui se exaltat, humiliabitur — Vulg. qui autem se exaltaverit; present for future-perfect, omnis added.
• 1163C Luke 1:52–53, 1163D Ose. 11:1, 1164A Gal. 3:29 — all agree exactly.
• 1163D Ps. 53:6 et Dominus susceptor est animae meae — two words added.
• 1163D Matt. 18:4 Quicunque autem for quicumque ergo; hic est major for maior.
• 1163D Luke 1:54 memorari misericordiae suae — suae added.
• 1164B–1166A: the entire Nunc dimittis (Luke 2:29–32) agrees with the Vulgate word for word. Verified twice, precisely because it is the verse in this range a translator is likeliest to supply from memory. Nunc dimittis servum tuum, Domine, secundum verbum tuum in pace is printed exactly so.
• 1166A Ps. 97:2 ante conspectum gentium for in conspectu gentium.

Pattern 7 — printed readings rendered literally, with conjectures

None of these was repaired in the text. Each is rendered as closely as the printed words allow (Pattern 7 / 7a), and the conjecture lives only here.

1. 1156D — a printed non that inverts Rabanus's own argument. Migne prints ut omnes renascamur in Christo et non aqua baptismatis nostra nobis peccata donentur. The sentence is expounding the fountain of Zach. 13:1 as baptismi lavacrum; the negation makes it say that our sins are not forgiven by the water of baptism, which contradicts the clause it depends on. Carried (Pattern 7a: every printed negation appears in the English, "and that our sins be not forgiven us by the water of baptism"). Conjecture: the non is intrusive, the intended reading being et aqua baptismatis nostra nobis peccata donentur. This is the one place in the range where fidelity makes the page argue against itself; it is left standing deliberately.
2. 1156A — abysso for ab ipso. Nam licet omnia abysso sint condita. abyssus is a real word, so no [sic:] (Pattern 12 boundary); rendered literally, "though all things have been founded by the abyss". The sentence wants "by him" (ab ipso) — the contrast is with quod eum veneratur auctorem in the next clause. Near-certain word-division defect.
3. 1152C — Deus ergo nostrorum. A bare genitive plural with no head noun; the lemma two lines above is Deus patrum nostrorum, so patrum has dropped out. Rendered literally, "the God of our own"; conjecture Deus ergo patrum nostrorum.
4. 1153A — benedictionis atque laudabilis. An adjective where a genitive noun is wanted. Rendered "this order of blessing stands historical and praiseworthy"; conjecture benedictionis atque laudationis ("of blessing and of praise"), which is what the parallel with ad benedicendum et ad laudandum two lines earlier requires.
5. 1159D — de enim ille scripsit (John 5:46). The object of de has dropped; the Vulgate is de me enim ille scripsit. Rendered as closely as the printed words allow — "for of, he wrote" — rather than silently restoring me, since restoring it would make the proof-text supply a word the plate does not print. Conjecture: de me enim ille scripsit.
6. 1162B — cujus me caro temporali conceptione laetatur. laetatur is deponent-intransitive and cannot govern the accusative me. Rendered literally, "whose flesh rejoices me by a temporal conception". Conjecture: laetificat ("gladdens me"), the reading the parallel spiritus meus … laetatur in the same sentence makes necessary.
7. 1164B — ut facerent secundum legis pro eo (Luke 2:27). A genitive with no head; the Vulgate has secundum consuetudinem legis. The English supplies the head in square brackets per the corpus's supplied-word convention — "according to the [custom] of the law" — so that the reader can see the supply is ours. Conjecture: secundum consuetudinem legis.
8. 1152C — hortatur angelos et coelos, aquas atque, virtutes. The comma after atque is printed and is carried into the English ("the waters and, the powers") per Pattern 8. It is almost certainly a compositor's stop; it is not deleted.

Pattern 10 / 12 — carried type, marked [sic: …]

Pattern 8 — punctuation carried as printed

• 1158B, sed toto corde inclinemur — the paragraph ends with no terminal stop; the next line begins a new lemma. The English ends without a period ("but that we be bowed down with the whole heart"). Not supplied.
• 1164D, habituros esse perpetuam Denique et Idithun — a sentence boundary with no stop before a capitalized Denique. Carried: "…perpetual rest in the bosom of Abraham Finally Idithun also…". Not supplied.
• 1164C, quis enim nesciat opera per manus et brachia solere figurari. — a question closed with a full stop. Carried as a full stop.
• 1161A, Verum quaeritur quomodo octo dierum alloquatur infantem? sed qua ratione … audivit? — both question marks printed and kept, including the lower-case sed opening after the first.

Pattern 5 / 6 — guillemets mirrored 1:1, including the plate's failures

Verified mechanically against the Latin twin: opening and closing counts are identical in every chunk of the range. Three of Migne's own irregularities were mirrored rather than tidied:

1. 1155C (0022) — Benedicite, glacies et nives, Dominum. « In gelu et frigore… : an opening guillemet where a closing one belongs. Reproduced as an opening guillemet. The run then closes 90 words later after the Eccli. 3 citation, so the chunk is left with 24 « and 23 » — matching the plate.
2. 1156C–D (0023) — the Zach. 13:1 quotation opens and never closes; the commentary runs straight on out of the quotation. No » supplied. (An earlier draft of this file's English did supply one; it was removed, and the guillemet parity check is what caught it.)
3. 1162A (0025) — the Eccli. 3:20 quotation « Quanto magnus es, humilia te in omnibus opens and never closes; the narrative continues inside it. Mirrored.
4. 1163A (0025) — dicente Domino: » Sancti estote : a closing guillemet where an opening belongs, and it is never balanced. Reproduced as printed.

Interrupting inquit kept inside the quotation at every occurrence (1155B, 1160A, 1161C, 1161D, 1162B, 1163A, 1164D, 1165A) per Pattern 5.

Citations

• 1156C [n: (Zach. XII)] is Zech. 13:1, not 12. The note is reproduced verbatim (refDisplay keeps what Migne prints); no correction in the text.
• 1153A — juxta finem centesimi et secundi Psalmi is Ps. 102 (Gallican), and the two verses quoted are 102:22 and 102:21, in that reversed order across 1153A and 1154A. Rendered as printed.
• 1162C — the Ps. 34:9 citation is printed inline, in italic parentheses ((Psal. [1162C] XXXIV)) rather than as an [n: …] note, and a column anchor falls inside it. This is a single stray instance, not a citing practice, so Pattern 4 [f: …] does not fire (known false-positive 4). Reproduced verbatim, anchor in place.
• 1155A Unde Apostolus exhortatur: « ut in omnibus glorificetur Deus per Jesum Christum » is 1 Pet. 4:11 — Peter, not Paul. Migne prints no note; nothing added.
• 1157C qui fecit coelum et terram, maria et omnia quae in eis sunt (Ps. 145:6) and 1158D Coelum plicabitur, sicut liber involutus (Isa. 34:4) and firmamentum est Dominus timentibus eum (Ps. 24:14) are quoted without guillemets and without notes. Left unquoted and unnoted, as printed.

Structure

• Heads. INCIPIT HYMNUS TRIUM PUERORUM. → HERE BEGINS THE HYMN OF THE THREE CHILDREN.; the (cont.) on 0023's repeated head is kept (known false-positive 7). INCIPIT CANTICUM ZACHARIAE Patris B. Joannis Baptistae. → HERE BEGINS THE CANTICLE OF ZECHARIAH Father of the Blessed John the Baptist. — the abbreviation B. is expanded to "Blessed" because it is a title inside a translated subtitle, not a locator; flagging it here in case a merge-time pass prefers Father of B. John the Baptist.. INCIPIT BENEDICTIO SIMEONIS PROPHETAE. → HERE BEGINS THE BLESSING OF SIMEON THE PROPHET.
• Column anchors. All reproduced verbatim, in order, in place; counts checked against the Latin twin before and after writing (15 / 14 / 11 / 9 / 5). 0026 runs 1165A → 1166A with no 1165B–D; that is the plate, not a loss. 0026 also opens [1164B]Narrat with no space after the anchor; the English mirrors the spacing.
• Notes. [n: …] counts and order identical to the twin (8 / 10 / 6 / 7 / 2). No note in this range is editorial prose, so no [nt: …] conversion (Pattern 15).
• No [ed: …] — nothing in this range is lost from the digitization.
• Word ratio EN:LA ≈ 1.49 / 1.52 / 1.59 / 1.55 / 1.63, within the expected ~1.5×.

Four Opus units, blind, 27/27 chunks. 11 sites: 4 source-side, 7 ours. Units in data/sweeps/polarity-8930/. Boundaries straddled the canticle batches.

⚑ THE PROMPT-DESIGN FINDING — all 7 ours-side defects fell in ONE unit

Same rule, same model, same work, comparable familiarity of text. Naming the words and making the divergence list an explicit deliverable is what moved 7 → 0. A general instruction to be careful did not. Fold the stronger form into the runbook's prompt template for every commentary range, not only the ones that look risky.

Fixed (ours)

• @1098B — two printed readings levelled into one. Migne prints the verse TWICE: « Laetata sum in salutari meo » and, further down, « Quia laetata sum in salutari tuo ». Our English gave "thy" both times. The passage is Rabanus sorting which words of the canticle are actually prayer — the first-person salutari meo against the second-person quia non est Deus praeter te — so the English pre-supplied the very person the argument is testing for and the distinction collapsed. Now "in my salvation" / "in thy salvation".
• @1100C — Prov. 9:1 conformed. Migne prints et suffulsit columnas septem ("propped up"); we gave the Douay's "hath hewn out" (excidit), a word not on the plate, in the proof-text adduced for the septenary. Now "hath propped up".
• @1106D — the English construction inverted the expected answer. Printed filiis vero Israel, quibus profuit… non fuit ille spiritus irae Dei? was rendered "was that not the spirit of God's wrath?" — which in English idiomatically expects yes, asserting the spirit WAS wrath toward Israel. Rabanus's next clause (spiritum Dei dissimiliter appellari, cum sit unus atque idem) requires the opposite: wrath to the Egyptians, not-wrath to Israel, which is what makes the naming dissimilar. Now "— that was not the spirit of God's wrath?", which keeps Migne's question mark (Pattern 8) and the negation without supplying the wrong expectation. ⚠ This column had a crux that actively DEFENDED the old rendering ("carried as a question, negative intact"). Presence of the negative was never the issue; the English construction was. Third instance of "a defended entry deserves more scrutiny".

Recorded, NOT fixed — flagged for a reading

• @1101B Suscitat a terra pauperem / de stercore erigit inopem — our lemma reverses the pairing against the plate. Rabanus himself then says the poor man is the same as the needy, so no sense is reversed and the fix is fidelity, not meaning. Left pending.
• @1094B Ps. 50:20 rendered "Deal kindly, O Sion" — Sion is accusative and the continuation ut acceptabile fiat in ea treats her as the place acted on. The crux logs this column but only notes the dropped Domine and says "rendered as printed" — it is not. Another logged-but-unaudited column.
• @1100C the Hebrew/LXX variant multa in filiis infirmata est rendered with the Vulgate lemma's "had many children", levelling the columns; our own English renders multa in filiis correctly in three other places.
• @1101D lemma prints non in fortitudine roborabitur vir with no sua; English adds "by his own strength", so Rabanus's gloss hoc est, non in fortitudine propria explains a word he is in fact supplying.

## Source-side, new

• @1117C — quod Deus Pater idcirco aperuit caelos recapping a lemma printed a column earlier as Operuit coelos virtus ejus. Antonyms.** Every other element of the verse survives in order; only the verb flips. Rendered literally, correct — the column needed a crux and now has one.
• @1120C — in fortitudine where the subject is metu judicii and the continuation quotidiano pavore sustinere; conjecture in formidine. This is the SAME substitution as the already-marked @1109C (fortitudo for formido, Exod. 15:16). Twice in one work looks like a habit of the setting, not a slip.

## Controls @1109C was planted in the WRONG UNIT (it lies in chunk 0006 = u1, not u2). The u2 agent reported controlFound: false honestly and said why, then found the identical substitution mechanism at @1120C in its own range unprompted — better evidence than the control would have given. @1136C, @1156D and @1092A were blind and all were confirmed.

## Also worth a house-style ruling @1160D coram homnibus tantum, et non etiam coram Domino → the [sic: homnibus] marker passes 7a′ (the negation keeps its force) but "men" disappears entirely, leaving "before ___ only, and not also before the Lord." A marked word whose own meaning is load-bearing is quarantined out of existence — a different problem from 7a′, and a fourth specimen for the real-word-typo convention Wilson owes (with 11553 @0540B, 11081 @0030B, and 8930 @1092A).