These are the edition's working notes, published as they were written. They record, work by work, where Migne's plate is defective or the reading uncertain: the printed form, what the English does with it, and why. They are not a groomed apparatus. Where a note argues with itself, withdraws an earlier decision, or leaves a question open, that is the record of how the reading was actually reached, and it is left standing — a conjecture that is visible can be checked, and one that has been tidied away cannot.
The English never silently repairs the page. A construable but wrong reading is rendered literally and the conjecture is recorded here, never applied to the text. How the English is produced and checked →
Every judgment call is listed individually below. No blanket policy is asserted in place of an entry. Governing rules: translation-style.md Patterns 7 (render the printed Latin literally), 8 (punctuation follows the plate), 10 (split/run-together type is non-word type, carried whole), 11 (dittography). Carried non-word type is set in italics in the English and listed here with the conjectured reading.
Chunk 0000
col 0009A — inline source citation, [f: …] tagging (Pattern 4 judgment). Printed: the Horace tag HORATIUS, Art. poet., 333. stands on its own line immediately after the quoted hexameter Aut prodesse volunt, aut delectare poetae. The chunk's noteCount is 0, so this is body text, not a Migne [n: …] note. It is a locator following its unit, so it is wrapped verbatim as [f: HORATIUS, Art. poet., 333.] rather than translated. Flagged because this is the only inline locator in the work and the pattern is normally triggered by florilegia, not treatises.
col 0011B — clementis for elementis (non-word class, carried). Printed: Opus conditionis est creatio mundi cum omnibus clementis suis. The form clementis is a real Latin word (gen. sg. of clemens) but is grammatically impossible in an ablative-plural slot governed by omnibus … suis, so there is no literal English rendering to give. Carried into the English untranslated in italics: "with all its clementis." Near-certain intended reading: elementis, "with all its elements" — the sense the following sentence's sacramentis suis parallels exactly. Logged rather than emended per Pattern 7.
col 0012A — Greek ἰςορέω (defective type, carried verbatim). Printed: Historia dicitur a verbo graeco ἰςορέω, historeo. The correct Greek is ἱστορέω; the plate sets a final/lunate sigma where στ belongs and drops the rough breathing. Reproduced exactly as printed, untranslated, per the marker-verbatim rule; the transliteration historeo that Migne supplies beside it is likewise left as printed.
col 0009A — supplied quotation marks. quod est video et narro (col 0012A) is rendered with supplied straight quotes — "I see" and "I narrate" — because English needs them and Migne prints no quotation mark of any kind. Per Pattern 6 these are ours, not the plate's; the chunk contains no « ». Same for "other-speech" (alieniloquium, col 0012B).
col 0009A — printed semicolon retained mid-sentence. Facilius quippe intelliges quod dicitur; si bene notum fuerit, unde dicatur. The semicolon falls between main clause and conditional, where English wants a comma. Carried through as printed (Pattern 8): "you will the more easily understand what is said; if it has been well known whence it is said."
col 0012D — printed semicolon retained mid-sentence. Quod et si in multis congrue assignari possit; ubique tamen observare… Same class as the previous entry; the semicolon precedes the tamen apodosis. Kept.
Capitula list vs. chapter heads — divergent wording, both rendered as printed. The capitulum for CAP. IV reads ad hanc triplicem interpretationem; the head of the chapter itself reads ad dictam triplicem interpretationem (and comperta against the body's reperiuntur). Neither has been harmonized to the other; each is translated as it stands ("to this threefold interpretation" / "to the said threefold interpretation").
Chunk 0001
col 0013C — convenentia for convenientia (non-word class, carried). Printed: pro convenentia similitudinis ad significationem propositae. convenentia is not a Latin word in any lexicon (it is not an attested medieval spelling of the Pattern 9 kind — the -ienti- syllable is simply dropped). Carried untranslated in italics: "on the ground of the convenentia of the likeness set forth." Intended: convenientia, "suitability, aptness."
col 0014B — nonita for non ita (Pattern 10, run-together type, carried whole). Printed: quando id quod dicitur, nonita intelligitur ut dicitur. Word-division defect of the sedet / Canetenim class. The whole run is carried: "is nonita understood as it is said." Intended: non ita, "is not so understood as it is said."
col 0014B — quod quid est per litteram non dicitur (unconstruable as printed). The clause hangs between the preceding nec aliud intelligendum ex iis quae dicuntur and the following sed per id quod littera dicit, significatur. As printed, the relative quod has no antecedent that will take quid est. Rendered as closely as the printed words allow — "for that which a thing is is not said through the letter, but is signified through that which the letter says" — with no word supplied or removed. A plausible intention is quod quid est = "the what-it-is (quiddity)," but the text is not emended.
col 0015A — fuges where the sense requires sugas (real word, rendered literally). Printed: quasi ex favo quodam postmodum meditando spiritualis intelligentiae dulcedinem fuges. fuges (2 sg. pres. subj. of fugo) means "you may put to flight / drive away," which reverses the argument: Hugh has just told the reader to press the letter into the mind so as to draw sweetness out of it, as from a honeycomb. Because fuges IS a well-formed Latin word, Pattern 7 requires the literal rendering, not the carve-out: "you will drive away … the sweetness of spiritual understanding." Near-certain intended reading: sugas, "you may suck out."
col 0015A — missing terminal stop after a note marker, not supplied. Printed: caeci oculus ad videndum illuminatur [n: (Joan. IX)] Lege ergo Scripturam. No full stop between the note and the next sentence. Reproduced exactly (Pattern 8, "never supply"): "…is enlightened to see [n: (Joan. IX)] Read, therefore, the Scripture."
col 0014D — comma inside per humilitatem, illuminaris, retained. Printed comma separates prepositional phrase from its verb. Kept: "because through lowliness, you are enlightened unto divinity."
col 0015C — Sextum Jeremias, Septimum Ezechiel. (comma for full stop, retained). Every other item in this list of prophetic volumes is closed with a full stop; here a comma stands where a stop belongs and the following ordinal is nonetheless capitalized. Reproduced: "The sixth Jeremiah, The seventh Ezekiel." The English deliberately looks broken because the plate is.
col 0015C — Octavum Thereasra qui est duodecim prophetarum (no comma, not supplied). The parallel entries all print a comma before qui; this one does not. No comma has been supplied in the English: "The eighth Thereasra which is that of the twelve prophets." Likewise the ordinal- plus-name pairs throughout the list are rendered without a comma after the ordinal, because the Latin has none there either.
col 0015D — In secundo similiter sunt quatuor: followed by a full stop mid-enumeration. Printed: Actus videlicet apostolorum, Epistolae Pauli numero quatuordecim sub uno volumine contextae. Canonicae Epistolae, Apocalypsis. The stop splits a four-item list after the second item. Kept as printed (Pattern 8, "never delete").
col 0013D — significatur where the argument wants an active. Printed: cum leo Christum significare dicit, non nomen animalis, sed animal ipsum significatur. Hugh's point is that the animal, not the noun, does the signifying; the printed passive says the animal is signified. Rendered literally as the passive, and the unattached 3 sg. dicit ("when he says") is likewise left without the supplied subject the sense implies.
Supplied quotation marks for the objectors' speeches. The dialogue at 0013B (Nos inquiunt, Scripturam legimus…), 0013C (Sed littera, inquiunt…) and 0014A–B (Sed non omnia, inquiunt…) is marked in Migne only by the bare inquiunt — no « », no other mark. Straight quotes have been supplied per Pattern 6, and the interrupting inquiunt is handled by ordinary English speech-tag punctuation, which is legitimate precisely because these are our marks and not the plate's. The chunk's guillemet count is 0 in both twins.
Chunk 0002
col 0016C — quorum nomina sunt. (full stop before the list, retained). Printed: quorum nomina sunt. Osee, Joel, Amos… The stop falls where a colon belongs, orphaning the list of twelve names. Reproduced: "whose names are. Hosea, Joel, Amos…"
col 0016D — tectu for textu (carried). Printed: Liber Esdras auctoris sui titulo praenotatur, in cujus tectu ejusdem Esdrae Neemiaeque sermones pariter continentur. tectu is not a well-formed ablative of tectum (that is tecto); as the supine of tego it cannot be governed by in cujus. No construable literal rendering exists, so it is carried in italics: "in whose tectu." Intended: textu, "in the text of which" — which is also what the parallel phrase in textu divinarum Scripturarum at 0016A prints correctly.
col 0018A — [primo] (Migne's own bracket, untranslated). Printed: Usu autem pravo [primo] invalescente. Square-bracketed matter supplied by Migne's editors (a variant for pravo), on the same footing as [ al., …]. Reproduced verbatim and untranslated inside the English sentence.
col 0018B — [n: (Ezeeh. I)] (misprinted citation, verbatim). The note reads Ezeeh. for Ezech. Note contents pass through verbatim by rule, so the typo is preserved without comment in the text.
col 0017A — sacerdotis magni rendered "the great priest," not "the high priest." Printed nepos Jesu sacerdotis magni. Migne does not print summi; the standard designation is rendered literally to keep the printed wording visible, since elsewhere in this same work (0026B–C) Hugh does use pontifex for the high priesthood and it is rendered "high priest" there.
col 0018C — nec patet sanctis Patribus a quibus edita sint (ambiguous case). sanctis Patribus can be read as dative ("it was not clear to the holy Fathers by whom they were published") or as ablative with a quibus ("it is not clear from what holy Fathers they were published"). The second is adopted, because a quibus is otherwise stranded; the first is recorded here as the live alternative.
Chunk 0003
Head of CAP. XIV — septem circumstantiis against the body's sex circumstantiis. The chapter head (0020D) announces seven circumstances; the body at 0021A lists exactly six (res, persona, numerus, locus, tempus, gestum), 0024A recapitulates "six," and the capitulum in chunk 0000 also says six. The head is rendered as printed ("seven"), not harmonized.
Capitulum vs. head, CAP. XIV — divergent titles, both rendered. The capitulum (chunk 0000) reads de sex circumstantiis quibus res significatae discernuntur mystice; et primum de tribus quae sunt res, persona, numeri; the chapter head here reads de septem circumstantiis quibus res significatae discernuntur. Neither has been conformed to the other. Same for CAP. XII (De Bibliothecae interpretatione vs De bibliothecae interpretatione) and CAP. XV/XVI, whose capitula and heads differ in wording throughout this work.
col 0020B — [pravi, ED.] (Migne's editorial bracket, untranslated). Printed: etsi etiam nihil in eo [pravi, ED.] reperiatur. The bracket carries the editor's own siglum ED. and is reproduced verbatim inside the English sentence rather than translated.
col 0020B — Ecclesiasticus and the book of Jesus son of Sirach listed as two books. Printed: Item Ecclesiasticus, liber Sapientiae Salomonis et duo libri Machabaeorum, Tobias, Judith, et liber Jesu filii Sirach apocryphi sunt. These are one and the same book, and Hugh himself identifies them at 0016D (Librum Ecclesiasticum … filius Sirach … composuit). The doubling is rendered as printed. It is a doubling of substance rather than of type, so it is NOT marked [d: …] — Pattern 11's marker is for a repeated run of type, and no run of type repeats here.
col 0021A — hanc … illam with reversed referents. Printed: excellentior valde est rerum significatio quam vocum: quia hanc usus instituit, illam natura dictavit. By normal usage hanc points to the nearer noun (vocum) and illam to the further (rerum), which gives the correct sense (usage makes words, nature makes things) only if the reader tracks the genitives rather than the order of the clause. Rendered neutrally as "usage instituted the one, nature dictated the other" to avoid resolving a pronoun the plate leaves ambiguous.
col 0021C — melos in sono, quod et quem percipimus auditu. Two relative pronouns of different gender for one antecedent (melos, neuter Greek loan, and a masculine quem with no masculine noun at hand — possibly attracted to sono). There is no English exponent for the discord, since English relatives carry no gender; rendered "which and whom we perceive by hearing," which preserves the doubling without inventing agreement. Marked so a later sweep does not churn it.
col 0019A — unde scriptores librarios vocabant (ambiguous object). Either "whence they called writers librarii" or "whence they called librarii writers." The first is adopted as the etymological point being made (liber → librarius); the printed word order is otherwise preserved.
Chunk 0004
col 0022A — the nine-modes table. The plate sets a ruled list of nine italic captions with numerals in a second column. The chunker represents the column as | N. The captions are translated; the numerals and the pipe structure are reproduced 1:1. No caption has been reworded to match the running recapitulation that follows it, though the two differ slightly in Latin as well.
col 0022D — column anchor set flush against the following word. Printed [0022D]Secundum numeri computationem… with no space. Reproduced exactly, with no space inserted: [0022D]According to the reckoning of number….
col 0023A — Quinarius, quinque senius. (non-word class, carried). The entry for the number five in the list of significations by multitude of parts. senius is not a Latin word; every neighbouring entry names what the number's units stand for (two → the two loves, three → the Trinity, four → the four seasons). Carried untranslated in italics: "Five, the five senius." Intended, near-certainly: sensus, "the five senses." Note also that the entry for the number six is missing from this list entirely, between Quinarius and Septenarius.
col 0023A — Ad Adam for addam (Pattern 10, split type inside a scripture lemma). Printed: Ad Adam correptiones vestras septuplum propter peccata vestra (Lev. XXVI). The Vulgate reads addam correptiones vestras septuplum propter peccata vestra — "I will add to your chastisements sevenfold for your sins." The plate has split the verb into a preposition and a proper name. Per Pattern 10 the WHOLE broken run is carried, including the fragment Ad which is itself a real word; because the run sits inside an already-italic lemma it takes no second italic marker. Rendered: Ad Adam your chastisements sevenfold for your sins.
col 0023D — Psalm quotation follows the plate, not a conventional version. Printed: Ne tradas me Domine a desiderio meo peccatori, (Psal. CXXXIX). Rendered "Give me not over, O Lord, from my desire to the sinner," which is what Migne prints; the standard text and every English version restructure the line. Scripture policy: translate what Migne prints.
col 0024B — five ages listed where six are announced. Printed: et sex aetates. Prima aetas ab Adam usque ad Noe … Quinta a transmigratione Babylonis usque ad adventum Christi. The sixth age (from the coming of Christ to the end of the world) is never stated. Rendered as printed; no sixth age supplied.
col 0024C — adesse for ad esse (Pattern 10, run-together type, carried whole). Printed: Status dicuntur, quia adesse hominis pertinent. As printed, adesse is the infinitive "to be present," which cannot govern the genitive hominis nor satisfy pertinent. Same class as sedet for sed et. The whole run is carried: "because they pertain adesse of man." Intended: ad esse hominis pertinent, "because they pertain to the being of man." The English reads awkwardly by design — the fragment is carried rather than smoothed.
col 0024D — no terminal punctuation at the chunk's end. Printed: …ut ab Edom Edumei, a Levi Levitae, a Juda Judaei — the sentence simply stops, with no full stop, at the foot of col. 0024D. No stop supplied (Pattern 8).
Chunk 0005
col 0025A — Arfaxat rex Judaeorum. Judith 1:1 makes Arphaxad king of the Medes; the plate prints Judaeorum, which also contradicts the sentence's own logic (the Jews are besieged by this king's conqueror later in the same paragraph). Rendered literally as "Arphaxad king of the Jews." Intended: Medorum.
col 0025A — jure obtentus belli (case discord). obtentus is nominative singular where the sense wants a genitive or ablative agreeing with belli or jure. Rendered "by the right of the war he had won," which is the closest construable English; the discord itself has no English exponent.
col 0025C — deprehendimur for deprehendimus. Printed: Verum in numeris multa mendacia scriptorum libris inesse deprehendimur. The passive "we are caught / found" cannot govern the accusative-and-infinitive multa mendacia … inesse. Rendered as literally as the printed form allows — "in numbers we are found to have many falsehoods present in the books of the writers" — rather than emended to the active. Intended: deprehendimus, "we find."
col 0025D — regum Syriae for regnum Syriae. Printed: per austrum nihil aliud quam regnum Aegypti; et per aquilonem nihil aliud quam regum Syriae. The parallel regnum Aegypti one clause earlier shows the intended reading. Rendered as printed — "by the north nothing else than the kings of Syria" — so the asymmetry stays visible.
col 0026A — Tryphon quidam Partium Alexandri filium. Capitalized Partium sits between the name and the object. Two readings are live: (a) partium Alexandri, "of Alexander's party," which is how it has been rendered; (b) a corruption of Parthorum or a place-name. Under (a), filium is then Alexander's son, and Migne's capital P is a compositor's habit rather than a proper name. Rendered "a certain Tryphon, of Alexander's party, killed the son whom he had received to be reared."
col 0026B — Postremo sanguine et virtute germanus defuncto Symone successit. The sentence lacks any construction: germanus is nominative with no verb, Postremo stands alone, and Simon — who has not been named before — appears already dead. The historical sequence Hugh means is Judas, Jonathas, Simon, then John Hyrcanus. Rendered as closely as the printed words allow ("Last of all, a brother in blood and in valor, when Simon was dead, John Hyrcanus his son succeeded"), with nothing supplied. This is the most damaged sentence in the work.
col 0026B — Alexandria uxor ejus for Alexandra. Salome Alexandra, widow of Alexander Jannaeus. The plate's form is the city's name. Rendered as printed, "Alexandria his wife," per Pattern 7.
col 0028A — suggéstione (stray acute accent on a Latin word). Printed: suggéstione Herodiadis ipsum Joannem decollavit. The acute is a typographic accident of the plate — Latin does not take it, and the word is otherwise correct. There is no English exponent for a stray accent mark, so the word is translated normally ("at the suggestion of Herodias") and the defect recorded here only. Marked so a later sweep does not churn it.
col 0028A — a patre Herode et ob suspicionem parricidii trucidatum. The et has no second member; either a phrase has dropped out or et stands for etiam. Rendered with the et in place — "by his father Herod and on suspicion of parricide" — rather than deleted.
col 0028C — matresua for matre sua (Pattern 10, run-together type, carried whole). Printed: cum Bernice matresua ad Festum praesidem visendum et salutandum descendisse legitur. Word-division defect. The whole run is carried: "with Bernice matresua." Intended: matre sua — though the historical relation is wrong in any case (Berenice was Agrippa II's sister, not his mother; Migne prints no correction). Both defects are recorded; neither is repaired.
col 0028D — the closing enumeration of the Herods breaks down. Printed: Tertius Agrippa Herodes, cujus avus fuit primus Herodes. Secundus patruus qui Jacobum interfecit. A "second" reappears after the third, and the work ends on it. Rendered as printed ("The third, Agrippa Herod … The second, his paternal uncle, who killed James."), with the printed full stop retained and nothing added to close the work.
col 0027B — supplied quotation marks around Marianne's speech. Printed: Verum (inquit illa) quomodo amas, quam mortuam malles quam vivam? Nam et interfici me praecepisti. Migne marks the speech with parentheses around inquit illa and nothing else — no « ». Straight quotes have been supplied per Pattern 6, and Migne's parentheses are kept inside them: "But (says she) how do you love her whom you would rather have dead than alive? For you even gave orders that I be killed." The chunk's guillemet count is 0 in both twins.
col 0027A–B — historic presents retained where the plate shifts tense mid-narrative. secretum aperit … jubet (0027A), corripit … ostendit (0027B), jubet utrosque occidi (0027B) sit inside otherwise past-tense narration. The shifts are reproduced rather than levelled to a uniform past.
<!-- 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). 2 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.
0015A — sense-reversing-letter (migne, high confidence)
Printed: Lege ergo Scripturam, et disce primum diligenter quae corporaliter narrat. Si enim formam horum secundum seriem narrationis propositae studiose animo impresseris, quasi ex favo quodam postmodum meditando spiritualis intelligentiae dulcedinem fuges.
Conjecture: fuges is a corruption of sugas (Didascalicon VI.3: 'quasi ex favo quodam postmodum meditando spiritualis intelligentiae dulcedinem exsugas'): 'you may SUCK OUT the sweetness of spiritual understanding, as from a honeycomb.'
The clause is the reward-clause of the whole chapter's exhortation, and its own image is a honeycomb — a thing one draws sweetness OUT of, not drives sweetness away from. The immediately preceding sentences promise gain, not loss: 'Lege ergo Scripturam, et disce primum diligenter quae corporaliter narrat' and, just above, 'Primum igitur illarum rerum quas tibi sacrum eloquium proponit, ad mysticam significationem stude legendo comparare notitiam, ut ex iis specie cognitis, postmodum meditando colligas' — meditation COLLECTS. A promise that diligent literal study will make you drive the sweetness of spiritual understanding away is the exact opposite of the argument against the self-styled allegory-teachers that the chapter is prosecuting.
0025A — sense-reversing-letter (migne, medium confidence)
Printed: Quemadmodum hoc quod in libro Judith legitur, Arfaxat rex Judaeorum multas gentes suo imperio subjugasse, ac contra Nabuchodonosor regem Assyriorum pugnasse
Conjecture: Judaeorum for Medorum (Judith 1:1, 'Arphaxad itaque rex Medorum'); the English follows the plate and so repeats the reversal.
Within the same sentence the Jews are the besieged subject people, not an empire subjugating many nations: 'Judaeos rebellare conantes in Bethulia obsedit, atque, Achior principe filiorum Amon narrante, didicit ipsum esse populum qui, nuper a captivitate reversus, eadem montana possedit.' A people that has only just come back from captivity and holds a few hill-towns cannot at the same moment be the power that 'multas gentes suo imperio subjugavit' and takes the field against Assyria; and the chronological puzzle Hugh then poses ('quo tempore haec gesta sint, vel quis fuerit Nabuchodonosor iste') turns entirely on the return from captivity, which presupposes the Jews as the subject party.