Expositio in Abdiam

Merged from two agent batches (chunks 0000–0005 and 0006–0010), 2026-07-28.

Work-wide conventions (reconciled at merge)

• Sense labels normalized per Pattern 9. Migne prints LITTERAL., MORAL., MORALIS., ALLEGORICE., ALLEGOR.; the English prints LITERAL. / MORAL. / ALLEGORICAL. throughout. The two batches disagreed on the third label (ALLEGORICALLY. vs ALLEGORICAL.) and were unified to the adjective at merge, so that the three labels read as one series; the Latin column preserves Migne's own variation, which is what Pattern 9 relies on.
• The prophet's name is rendered Obadiah in running prose. Abdias stands verbatim wherever it is inside a Latin string (frontmatter incipits, [n:] contents, carried type).
• Auditum audivimus = "We have heard a hearing," not "a report" — Hugh's gloss at 0375A–B turns on the phrase being an un-Latin Hebraism (locutio quidem minus est Latina, sed ex Graeco fideliter expressa), so an idiomatic English would destroy the thing being glossed.

PL 175, cols 0371C–0389D. Chunks 0006–0010 are logged by the peer agent; the two files are merged afterward.

Conventions: Pattern 7 (render the printed Latin literally, never emend), Pattern 10 (carry the whole broken run, spacing included), Pattern 12 (carried type is wrapped [sic: …]; a real word that is merely wrong is rendered as ordinary English and gets a crux but no marker).

Carried type — marked [sic: …] in the English

### 0374A (chunk 0001) — in cordo for in corde Printed: dum virus fovit in cordo latitans. Carried, [sic: cordo]. cordo is no form of cor, cordis (dat./abl. cordi / corde); the only Latin word it could be is the ablative of the adjective cordus ("late-born"), which is unconstruable here. Intended: in corde, "in his heart" — near-certain. Because there is nothing renderable, it is carried rather than translated (Pattern 12's "real word merely wrong" exemption does not reach it).

### 0375B (chunk 0001) — sifignicat for significat Printed: Inculcatio vero verbi intentionem sifignicat eloquii. Carried, [sic: sifignicat]. Metathesis/foul type for significat. The sentence is the verb, so the English reads "The repeating of the word, however, [sic: sifignicat] the intention of the utterance."

### 0378A (chunk 0002) — irerverberate for irreverberate Printed: dum aciem mentis in ipsum veritatis radium irerverberate figunt. Carried, [sic: irerverberate]. Adverb irreverberate, "unblinkingly," modifying figunt — the eagle-topos of gazing at the sun without flinching.

### 0380C (chunk 0003) — invaseres for invasores Printed: Deinde eosdem invaseres, illusores, insidiatores nominat. Carried, [sic: invaseres]. The correct invasores stands eleven lines later in the same chunk (0380D, invasores per obedientiam), which is the evidence that this is compositor's type and not a variant form.

### 0386A (chunk 0004) — laetitio for laetitiae Printed: Relinquamus ubique signa laetitio nostrae (Wisd. 2:9). Carried, [sic: laetitio]. Note: the run sits inside an italic scripture lemma, so per Pattern 10 it takes no second italic marker; the [sic: …] content is bare.

### 0387B (chunk 0005) — influmine for in flumine (split/run-together type) Printed: qui convertis mare in aridam, ut influmine sicco pertranseant pede. Carried, [sic: influmine], per Pattern 10 (word-division defect; the whole run is carried, including the real word in). The carried run stands where the prepositional phrase stands: "so that they pass over [sic: influmine] with dry foot."

### 0389A (chunk 0005) — in humanitatem for inhumanitatem Printed: Historialiter ostendit in humanitatem Idumaeorum, et crudelitatem eorum. Carried, [sic: in humanitatem], per Pattern 10 (split type; in stays inside the run). The paired crudelitatem immediately following makes inhumanitatem ("inhumanity") certain — and note that the split reading inverts the sense into "humanity," which is exactly the class of damage Pattern 7 exists to expose rather than repair.

Rendered literally, no marker (real words, wrong readings)

### 0376D (chunk 0002) — vanitatem for unitatem Printed: Per haec enim, quae distinguo, personarum proprietatem ostendo, non naturae vanitatem confundo. Rendered as printed: "I do not confound the emptiness of nature." The argument requires unitatem — the sentence sits between "lest he confound the trinity of the persons, or divide the unity of the divine essence" and "as the Trinity is one in nature or essence." A Trinitarian theologian does not concede an emptiness of the divine nature. High-value defect: this is a doctrinal statement inverted by one letter, and the literal English keeps the witness intact.

### 0377D (chunk 0002) — exsultantem for exaltantem Printed: per exsultantem solium, praedicatio malitiae — but the lemma four lines above (0376D–0377A) and the parallel gloss at 0385C both read exaltantem solium ("that exaltest thy throne," Obad. 3). Rendered as printed, "by the throne exulting." The gloss-list otherwise tracks the lemma word for word, so this is a slip in the gloss, not a variant lemma.

### 0380D (chunk 0003) — subtilitatis … praedicationis Printed: In linteo sindonis, subtilitatis designatur praedicationis. The sentence has no nominative subject for designatur; sense requires subtilitas. Rendered with the printed double genitive: "In the cloth of fine linen is designated of the subtlety of preaching." Deliberately awkward — the doubled "of" is the defect, and smoothing it would hide a case error inside an otherwise unremarkable Gregorian borrowing (Prov. 31:24).

### 0381C (chunk 0003) — induxerunt without an object Printed: Haeretici raptores spirituales terram sanctam invadentes, et Scripturas canonicas pervertentes induxerunt; linguas suas acuerunt… The verb governs nothing. Rendered "have brought them in," supplying only the pronoun English requires. Possibly a dropped noun (errores? schismata?) or a compositor's transposition; nothing in the sentence lets one choose. Flagged for a human eye.

### 0382A (chunk 0003) — cum for eum Printed: Cui hoc Dominus dedit, quando cum talem, judicio fieri permisit, etc. The parallel at 0377C reads dedi te, hoc est justo judicio fieri permisi, i.e. "I permitted him to become such." Rendered with the printed cum as a preposition: "when with such a one he permitted it to come to pass by judgment."

### 0382A (chunk 0003) — lemma variants expellit / deducet Printed: Superbia expellit te … Quis me deducet in terram? against the lemma as quoted at 0376D–0377A and 0385C: Superbia cordis tui extulit te … Quis me detrahet in terram? — and detrahet returns three lines later at 0382A itself ("Quis me detrahet in terram? Dicere istud…"). Both printed readings rendered as they stand ("Pride driveth thee out"; "Who shall lead me down"). Whether Hugh is quoting loosely from memory here or Migne's compositor drifted cannot be settled from the plate.

### 0383A (chunk 0003) — unconstruable clause on the Fathers Printed: quorum nonnulli cum eis foedus iniere pacem, aliquando tenuere panem quoque cum eis comedere. foedus iniere pacem has two accusatives and no conjunction; tenuere … comedere has no construction. Rendered as closely as the printed words allow: "some of whom entered into covenant with them, peace they sometimes held, to eat bread also with them." The sense wanted is plainly the three Obadiah terms Hugh has been working (covenant / peace / eating bread) applied to Augustine among the Manichaeans. A verb has probably dropped out. Flagged.

### 0385B (chunk 0004) — corruptibilis for contemptibilis Printed: parvulus est, quia momentaneus; corruptibilis, quia inhonestus, in the middle of a rhetorical chain that otherwise pairs parvulus with contemptibilis nine consecutive times (0385B–0385C), following the lemma "Ecce parvulum dedi te … contemptibilis tu es valde." Rendered as printed ("corruptible, because dishonorable").

### 0384A–0384B (chunk 0004) — the vice list and the Canaanite-nations list disagree The catalogue at 0384A assigns contra mundum, avaritia; contra negotium acedia (fourth = avarice, fifth = sloth); the interpretation of the seven names at 0384B assigns per quartum acedia; per quintum avaritia (fourth = sloth, fifth = avarice). Both rendered as printed. The etymologies favor the second list (Amorrhaeus "embittered" → acedia; Gergesaeus "driving out the settler" → avarice), so the first list is likelier the slip — but Hugh hands the "fitting of each one" to the reader in the very next sentence, so he may simply not have checked. No emendation attempted.

Migne's own printed apparatus (reproduced, not ours)

### 0378B (chunk 0002) — bracketed variant [nunquid] Printed: nonne [nunquid] saltem racemum reliquissent tibi? Migne's own alternative reading, reproduced verbatim and untranslated in the English, on the same footing as [ al., …] elsewhere in the corpus.

### 0386C–D (chunk 0004) — a citation the TEI did not mark as a note Printed inside the italic Job lemma: et hic confringes tumentes fluctus tuos (Job XXXVIII). Every other citation in this work carries an [n: …] marker; this one is running text in Corpus Corporum's transform. Reproduced verbatim as printed text, in the italic run, so that noteCount (19) still matches the Latin twin. Not patched: the words are Migne's and the markup question is cosmetic, but a later harvest may want to normalize it.

Punctuation and typography carried per Pattern 8

• 0379C (chunk 0002) — dum tubis sonuerunt praedicationum - dum lampade fulserunt miraculorum: a hyphen/dash where the parallel members take commas. Reproduced.
• 0378C (chunk 0002) — De tertiis. Potentes potenter…: a full stop where the two preceding parallel members (De primis legitur:, De secundis:) take a colon. Reproduced. Same at 0379B, Hinc Salvator de seipso.
• 0386D–0387A (chunk 0005) — carnalium cogitationum amputare student superflua Unde [0387A] Job:: no stop between the sentences. Reproduced; no mark supplied.
• 0388B (chunk 0005) — etc. Sequitur with no terminal stop, where every other instance in the work reads Sequitur:. Reproduced.
• 0374B (chunk 0001) — hoc, inquam, Verbum illuminat omnem hominem: the sentence opens with a lowercase h after a full stop. The English capitalizes ("This Word, I say…"), on the ground that sentence-initial capitalization is an orthographic convention of the receiving language rather than a word or a mark of punctuation, and the defect stays visible in the facing Latin column (Pattern 9's section-label reasoning). Noted so a later sweep does not churn it.

Column-anchor anomaly (worth a human's eye)

Chunks 0000–0001, cols 0371C → 0372C → 0371 → 0371D → 0372D. The anchors run out of numerical order across the chunk boundary. The Praefatio (chunk 0000) carries [0371C] and then [0372C], skipping 0371D–0372B; chunk 0001 then opens with a band-less [0371], followed by [0371D], then [0372D]. This is the signature of PL 175 printing the Praefatio and the opening of the Expositio in parallel columns on the same opening — the transcription reads down one column and then the other. Reproduced verbatim and in the printed order, per the sacred- marker rule; not reordered. Worth confirming against the plate before the citation resolver (migne.app/pl/175/371c) is pointed at these chunks, since a reader arriving at 371c will land in the Praefatio while 371d is in the Expositio.

Migne's citation numbers (kept verbatim in [n: …], listed for the indexer)

These are printed as they stand; no citation-corrections.json entry is proposed here, only a record that the printed locator and the actual locus differ.

• 0371C [n: (Eccle. XX)] — Sapiens verbis innotescit paucis is Ecclesiasticus (Sirach) 20:5, not Ecclesiastes. Migne's Eccle. is used for both books in this work (cf. [n: (Eccle. II)] at 0384D, which is Ecclesiastes 2:3).
• 0374B [n: (Psal. CXVI)] — Misit verbum suum et sanavit eos is Ps. 106:20.
• 0376B [n: (Isa. XII)] — Ecce puer meus, quem elegi is Isa. 42:1.
• 0378C [n: (Psal. LXXVI)] — Confundantur omnes qui adorant sculptilia is Ps. 96:7.
• 0380A [n: (Job. XIV.)] — Nunquid ligabis eum ancillis tuis is Job 40:24.
• 0384D [n: (Psal. XIV)] — In terra deserta, invia et inaquosa is Ps. 62:3.
• 0384C [n: (Gal. V)] — the quotation conflates Gal. 5:24 with Rom. 8:9 (vos autem in carne non estis); Migne cites only Galatians.
• 0385A [n: (Thren. IV)] — the italic run tagged with this note runs past the end of Lam. 4:5 and swallows Hugh's own exhortation (Surgite ergo, amici mei, ad ultionem hostis…), which is not scripture. The italics are Migne's and are reproduced as printed; flagged because the indexer will otherwise attribute Hugh's sentence to Lamentations.

Translation conventions fixed for this work (for the merge)

• Lemma rendering. Obadiah is translated from Migne's Latin throughout, never from a conventional English version. In particular Auditum audivimus a Domino is rendered "We have heard a hearing from the Lord," because Hugh's gloss at 0375A–B turns on the phrase being an un-Latin Hebraism ("locutio quidem minus est Latina, sed ex Graeco fideliter expressa"). "We have heard a report" would silently destroy the comment.
• No guillemets anywhere. Migne marks every quotation in this work by italics, never by « ». The English carries zero guillemets, matching the Latin twin 1:1 (Pattern 6); quoted lemmata repeated inside the gloss are set unitalicized exactly where Migne sets them unitalicized.
• Sense labels. LITTERAL. → LITERAL.; ALLEG. → ALLEG.; MORAL. and MORALIS. → MORAL.; ALLEGORICE. → ALLEGORICALLY. (the abbreviated and spelled-out forms are kept distinct, since the variance is abbreviation, not a misprint).
• Proper names follow the plate: Hebrom (not Hebron), Dothain, Sichem, Jezrahel, Jezabel, Idumaea, Chanaanite, Manichaeus.
• gastrimargia is left in Latin (unitalicized, as Migne prints it) — Hugh glosses it himself with seu gula, so it is a technical term, not a defect.

PL 175, cols 0390A–0406A. Chunks 0000–0005 were translated concurrently by a second agent and carry their own crux file; these two files are to be merged.

Rendering policy: translation-style.md patterns 7 (render the printed Latin literally), 10 (carry the whole broken run), 12 ([sic: …] on carried type). Conjecture lives here, never in the running text.

Carried type — [sic: …] in the English (Patterns 7, 10, 12)

### 0391B — accipistis Printed: Ego veni in nomine Patris mei, et non accipistis me (John 5:43). accipistis is not a form of accipio; the perfect is accepistis, which the Vulgate prints and which the sense requires. A compositor's slip (i for e). Carried untranslated inside the scripture lemma, so it takes no second italic marker (Pattern 10): you [sic: accipistis] me not.

### 0396B — Achartumnam Printed: quae vergit ad scorpionem, id est omnem Achartumnam. No such Latin or transliterated word. Hugh is abbreviating Jerome In Abdiam on v. 19, where the territory Judah holds is the Daroma — the Hebrew darom, "south," which Jerome elsewhere transliterates and glosses. Achartumnam looks like a badly set transliteration of that (or of another Hebrew regional name) rather than a Latin word, so it is carried, not translated. Worth a human's eye — a reader with Jerome's text at hand could probably fix this in one look, and if it is recoverable it belongs in data/tei-patches/11063.json with provenance, not here.

Related, same clause: vergit ad scorpionem ("slopes toward the scorpion"). This is real Latin and is rendered as printed under Pattern 7, no marker. It is nonetheless suspect: a compass bearing is wanted here, and scorpio is not one. Possibly a garbling of a wind-name (Africum, Notum) in Jerome's geography. Flagged, not emended.

### 0397C — confragrosos Printed: inter confragrosos scopulos nostram naviculam reximus. The word is confragosus ("broken, rugged"); confragrosos has an intrusive r and is not attested. Carried under Pattern 7. (Contrast the Pattern 9 exemption: this is not a period spelling, it is a slip unique to this plate.)

### 0400B — Ab dias Printed: Dicit ergo Ab dias: — the prophet's name split across a word space. Pattern 10 split type; the whole run is carried, including the fragment Ab which is a real word. The name is Abdias and is set correctly everywhere else in the work (e.g. 0394A, 0397D).

### 0402A–B — hisistere Printed: Sed neque hisistere potest semel coepta corruptio. Not a Latin word. The sense requires subsistere ("come to a stand, stop") — the argument is that corruption once begun does not halt but breeds the next vice; insistere is the other candidate but gives the wrong sense. Carried under Pattern 7.

Rendered as printed (English on the page is ours — no marker)

### 0390B — ut tu Hierusalem with no verb Printed: Et haec idcirco vindicta procedet, ut tu Hierusalem: te Babylonius, Babylonium Medus ac Persa consumat. The first limb of the tricolon has lost its verb; the shared verb consumat arrives only in the third. Rendered with a supplied bracketed verb — "that thou [consume] Jerusalem" — on the existing [it]/[them] convention for words we supply. The bracket is ours; the gap is Migne's.

### 0390A — exinanite, Exinanite Printed: qui dicunt, exinanite, Exinanite usque ad fundamentum in ea. NOT dittography: Ps 136:7 genuinely doubles the imperative, and the Vulgate prints it doubled. No [d: …]. The mid-sentence capital E is Migne's and is reproduced ("empty it out, Empty it out"). Pattern 8.

### 0391D — comma before a capitalized new sentence Printed: …superba daemonum tyrannide fuerunt subjugata, Haec autem sunt posse, velle, nosse. A comma where a full stop is wanted, with the following word capitalized. Preserved exactly under Pattern 8 (never delete a printed mark).

### 0392A — full stop before a quotation Printed: Hoc etiam Psalmista breviter insinuat dicens. Cor meum turbatum est… A stop where a colon is wanted. Preserved; the English reads "…intimates, saying. My heart is troubled".

### 0392C–D — Jacob's fourth wrestling assigned to the brother Printed: Ter cum fratre suo, semel cum angelo. Primo in utero; secundo de lentis edulio; tertio de benedictionis paternae oraculo. Cum fratre de Mesopotamia rediens, ut praefatum est. The fourth item is introduced cum fratre, but the schema Hugh has just announced (and the biblical narrative, Gen. 32) makes the return from Mesopotamia the wrestling with the angel; Hugh himself then writes Cum angelo quarto loco… in the next sentence. Probably a slip for cum angelo, possibly a compressed reference to the reconciliation with Esau on the same journey. Rendered as printed.

### 0392D — Hebr. III lemma, usque ad firmum Printed: si tamen initium spei usque ad firmum teneamus. The Vulgate has fiduciam et gloriam spei usque ad finem firmam retineamus; Migne's text has lost finem and left firmum as a bare substantive. Rendered literally — "right up to that which is firm" — per the scripture rule (translate the wording quoted, never a conventional version).

### 0393B — salvabit for salvavit Printed (Isa. 63:5): sed salvabit mihi brachium meum, et indignatio mea ipsa auxiliata est mihi. Future beside a perfect in the same verse. Rendered as printed ("my own arm shall save for me"), which preserves the tense clash.

### 0393A — per te passus est for pro te Printed: quid pro te sustinuit sol justitiae, quid pro te egit Dominus virtutum, quid per te passus est rex gloriae. Two pro te then a per te in a deliberately parallel tricolon. It may be rhetorical variation ("suffered through thee," i.e. at thy hands) rather than a slip — Hugh is berating the reader — so it is rendered as printed, "suffered through thee." A real word, merely possibly wrong: no marker (Pattern 12 boundary).

### 0395C — antequam ad persecutionem quis evadat Printed thus, paired with antequam ad arcem contemplationis transeat. The parallel wants perfectionem, and evadere ad ("get clear to") fits an attainment, not a persecution. Rendered as printed. A near-certain persecutionem/perfectionem confusion, but a real word either way, so no marker.

### 0397B — quem aliud arbitror Printed: Possumus autem locum quemlibet regni Babylonis intelligere, quem aliud arbitror. Unconstruable as it stands (quem … aliud). Rendered as closely as the printed words allow: "which I judge to be something else."

### 0397A–B — Bophorum [Bosphorum] Printed: id est, Bophorum [Bosphorum] vocari non Ephrata. The bracketed correction is Migne's own, and it is reproduced verbatim and untranslated, like [ al., …]. Per Pattern 12, no [sic: …] is added on top of it: the nineteenth-century editor had already caught the misprint and flagged it, and ours marks only defects he passed over.

### 0398A — sic secunda ad quartam, sic secunda ad tertiam Printed: Et sicut prima ad tertiam, sic secunda ad quartam, sic secunda ad tertiam. The third limb repeats secunda and produces a proportion that contradicts the second limb. Not dittography (the two runs differ), so no [d: …]; rendered as printed. The whole proportion passage (0397D–0398A, Unde prima eodem numero sic historialiter…) is muddled in the transmission and Hugh's own arithmetic is hard to recover — he then sends the reader to "arithmetic" for instruction. Worth a human's eye.

### 0400C–D — Fiat voluntas tuas Printed thus in the list of petitions (accusative plural for nominative singular). Latin case has no English exponent — "thy will" is identical either way — so, as with the gender mismatches of Pattern 9, there is no emendation to make and none to undo. Logged; the English is left alone. The same petition is printed correctly as Fiat voluntas tua at 0403C (chunk 0010).

### 0400D–0401A — succedunt quatuor virtutes, then seven are listed Printed: Deinde quarto loco succedunt quatuor virtutes, followed by Prima … septima — seven items, in a passage whose entire architecture is sevenfold (seven vices, seven petitions, seven gifts, seven beatitudes). quatuor is almost certainly caught from the immediately preceding quarto loco. Rendered as printed ("there follow four virtues"), and the seven items follow it on the page, which is the defect made visible. Worth a human's eye — this is the clearest single-word error in the batch.

### 0401A — Avaritia … per montem Esau significatur The list of vice-to-place correspondences begins at the fifth vice (avarice) and gives only two members (avarice/mount Esau, envy/Philistines) before breaking off into the physician simile. The promised seven-part mapping onto Obadiah's seven place-names is never completed. Rendered as printed; noted because a reader will expect the missing five and may suspect our text of a hole. It is not a hole in the transcription — the Latin runs continuously.

### ~~0405A → 0406A — the column anchors jump~~ DO NOT FIX — false alarm Documented do-not-reopen class; no plate check, no [ed: …], no patch. Cols 0405B–D are not "absent": Migne's A–D marks are positional quarter-guides transcribed where they appear in the plate, not four per column, so 0405A followed by 0406A is simply the next mark in the next column. Arithmetic over the band letters manufactures phantom gaps — settled corpus-wide 2026-07-18 (49 such "gaps" examined, all same-letter, zero real). The only real signal is text broken mid-word or mid-construction across an anchor, and the text here is continuous: quae hominem illecebris captum in [0406A] servitutem redigit. colLast 0406A is correct.

Conventions used here, for the merge

• Sense labels. Migne prints LITTERAL. (0391A, 0396B), MORAL. (0391B), ALLEGORICE. (0394A), ALLEGOR. (0397C) — and, in the other agent's chunks, MORALIS. as well. Under Pattern 9 (section labels normalize in English, the misprint staying visible in the facing Latin column) these are rendered LITERAL. / MORAL. / ALLEGORICAL. throughout. If chunks 0000–0005 chose otherwise, normalize the whole work to one set at merge.
• The prophet's name. Abdias is rendered Obadiah in running prose (0394A, 0397D). Align with 0000–0005 at merge.
• [id est sol] at 0395B (Quod de supernis Helios [id est sol] intuens) is Migne's own bracketed gloss and is reproduced verbatim, untranslated, on the same footing as [ al., …] and [Bosphorum].
• super coeleste / super coelestium (0394C, 0396A) are written as two words for supercoeleste / supercoelestium. Treated as period orthography under Pattern 9, not as split type: rendered "the supercelestial", "supercelestial desires", with no marker.
• gastrimargia (0394D) is rendered "gluttony" rather than kept in italic Latin — it has a clean English equivalent, and Hugh uses gula for the same vice at 0400C.
• No guillemets occur anywhere in chunks 0006–0010; the Latin twins have none either (verified 0/0 per chunk). Every quotation mark in the English of these five chunks is a supplied " " under Pattern 6 — there are four pairs, all glosses: "holy" (0391A, dic sanctum absolute), "holy" and "without earth" (0391D, the agios etymology), "seeing" (0397B, Jerome on Sapharat).

<!-- corpus-polarity-sweep 2026-07-28 --> ## Corpus polarity sweep — 2026-07-28

Blind Opus read of the Latin/English pair for sense-reversing defects (one word or one letter asserting the opposite of what the context requires). 6 site(s). Migne's defects are recorded here and rendered literally in the English; our own errors were fixed in place. Full context: data/sweeps/corpus-polarity/REPORT.md.

0376D — sense-reversing-letter (migne, high confidence)

Printed: Per haec enim, quae distinguo, personarum proprietatem ostendo, non naturae vanitatem confundo.

Conjecture: non naturae UNITATEM confundo — "I do not confound the unity of nature"

The immediately preceding sentence states the very rule being applied: "Caveat lector, ne personarum confundat trinitatem, nec divinae essentiae dividat unitatem"; the next sentence continues "sicut Trinitas natura vel essentia UNA." Hugh repeats the same formula verbatim at 0392B with the right word: "Haec dicens non divinae essentiae unitatem divido, nec propria personarum confundo." "Vanitas naturae" (emptiness of nature) is not a Trinitarian category at all and asserts the opposite of the unity he is defending.

0377D — sense-reversing-letter (migne, low confidence)

Printed: Tumor praecedit superbiae, tenebrae sequuntur ignorantiae, infidelitas succedit idololatriae.

Conjecture: infidelitatI succedit idololatriA — idolatry succeeds unbelief (case endings transposed)

The very next sentence gives the sequence in fixed order and puts unbelief BEFORE idolatry: "per petram, infidelitas ignorantiae; per scissuras, schismata idololatriae; per habitantem, consensum immunditiae." As printed, the order of the fall is reversed against its own immediately following exposition.

0382A — sense-reversing-letter (migne, medium confidence)

Printed: Superbia expellit te, habitantem in scissuris petrae; qui dicis in corde tuo: Quis me deducet in terram?

Conjecture: Superbia [cordis tui] EXTULIT te — "the pride of thy heart hath lifted thee up" (Obad. 3, as printed twice elsewhere in this same work)

The lemma is quoted correctly twice in the same commentary — at 0377A/0377C ("Superbia cordis tui extulit te, habitantem in scissuris petrarum, exaltantem solium tuum") and again at 0385C. And the exposition attached to THIS occurrence expounds elevation, not expulsion: "Solium, est haeretica doctrina. Solii exaltatio, perversae doctrinae dilatatio," and the answering divine word two lines later is "Si exaltatus fueris ut aquila ... inde detraham te" — a threat that presupposes pride has raised him up, not driven him out.

0385D — dropped-word (migne, low confidence)

Printed: Hic est decalogus inobedientiae, quem in monte Sinai, quod interpretatur humilitas, Dominus dedit Moysi.

Conjecture: a lost antithetical particle, e.g. "decalogus inobedientiae, CONTRA [eum] quem in monte Sinai ... Dominus dedit Moysi"

The ten items just enumerated are "contemptum Dei ... praedicationem peccati ... odium boni ... amorem mali ... oblivionem Dei" — which God manifestly did not give on Sinai — and the very next sentence names its keepers as Pharaoh and the reprobate: "Hunc Decalogum impleverat, qui dicebat: Nescio Dominum; et Israelem non dimittam." As printed, the sentence makes the Lord the giver of the decalogue of disobedience, the opposite of what the surrounding argument requires.

0389A — dropped-word (migne, high confidence)

Printed: Historialiter ostendit in humanitatem Idumaeorum, et crudelitatem eorum adversus populum Israeliticum

Conjecture: INHUMANITATEM Idumaeorum — the word-division defect splits inhumanitas into in + humanitas

The coordinate noun in the same clause is "et crudelitatem eorum adversus populum Israeliticum," and what follows is the charge that they "not only failed him with help in the time of his distress, but even joined themselves to the fellowship of his enemies." The printed division makes the sentence say the prophet displays the Idumaeans' HUMANITY, the exact opposite of the cruelty it is coupled with. (Cf. the same accusation at 0388C: "adversus populum Israeliticum inique egit.")

0403A — dropped-word (migne, medium confidence)

Printed: Cumque huic limo profundi animus fuerit infixus et luto coinquinationis et immunditiae involutus, evelli nequaquam potest, ad illum clamet, et auxilium ejus postulet, de quo Psalmista loquitur

Conjecture: evelli nequaquam potest, NISI ad illum clamet et auxilium ejus postulet — "it can by no means be plucked out unless it cry to him and beg his help"

The Psalm verse quoted in the same sentence asserts precisely that the extraction does happen on such a cry: "Exspectans exspectavi Dominum, et intendit mihi, et exaudivit preces meas, et eduxit me de lacu miseriae, et de luto faecis." With nisi dropped, the clause flatly denies the rescue that the quotation it introduces affirms, and the two hortatory subjunctives (clamet, postulet) are left hanging with no conjunction — the mark of the lost particle.