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 →
Anselm of Laon and his school. Translated 2026-08-23, one agent, 8 chunks, 7,556 Latin words → 11,521 English (1.52×, the expected Glossa ratio). verify-english.mjs: clean — 8 chunks, columns/notes/sections aligned, no duplicate paragraphs.
Apparatus: 4 [var:] · 2 [cj:] · 0 [sic:] · 0 [d:] · 0 [ed:] · 0 [nt:] · 168 [n:] (exact parity with the Latin twins). Zero guillemets in either language. Per-chunk parity of column anchors, [n:], italic spans, ## heads, ?, !, : and ; was checked band by band against the Latin twin and is exact everywhere (the markers' own colons, semicolons and italics account for every arithmetic difference).
Structure: CAPUT PRIMUM–CAPUT XXV, twenty-five heads, 1:1 with the Latin.
§1 · CONVENTIONS — Ruth's §1 applied, plus what this book had to settle
The Glossa conventions were settled on Liber Ruth (8968) §1 and are not reopened here: VERS. n.-- passes through verbatim; the lemma is Englished; a fragmentary lemma is rendered as the phrase it stands for; attribution sigla ride as ordinary [n:] notes, unanglicized (passing (RAB.) through is not endorsing Migne's attribution). Heads follow Ruth's shipped forms: CAPUT PRIMUM. → CHAPTER ONE., CAPUT II. → CHAPTER TWO.
Four further calls this book forced:
1a. usque ad → as far as, and six of the thirty-seven occurrences are not the formula
The abridgment formula is Englished in its own italic span per shipped 8978 and Tobit (8970 §1b): etc., as far as …. No verb is ever supplied to complete a resumption tail, and none was here.
⚠ The launch brief's warning was live. The Latin prints usque ad 37 times; 31 are the formula and 6 are ordinary prose or the verse's own words, decided one by one from what stands on either side:
| col | occurrence | verdict |
|---|---|---|
| 0617B | in porticibus circa domum Domini usque ad sabbatum manentes | ordinary — "until the sabbath" |
| 0620B | aquis … fervefactis, et usque ad salis firmitatem coquendo perductis | ordinary — "as far as the firmness of salt" |
| 0620C | …terminos Israel, ab introitu Emath usque ad mare solitudinis | the verse's own words (4 Kings 14:25), inside the lemma's italics — exactly the Sap 8:1 trap named in the brief |
| 0620C | per stadia quingenta octuaginta usque ad Zoaros Arabiae | ordinary |
| 0620C | in longitudine centum quinquaginta, usque ad vicinia Sodomorum | ordinary |
| 0620D | eo quod usque ad nubes caput erigat | ordinary — "as far as the clouds" |
⭐ Related and NOT an instance: the lemma Igitur usque. at 0618A is 4 Kings 12:6's own opening (Igitur usque ad vigesimum tertium annum) clipped before ad. Rendered "Therefore until.", not as a formula.
Two formula occurrences carry their resumption inside the italic span rather than after it — usque ad Fides sine operibus mortua est. (0620A) and usque ad: et orietur vobis timentibus sol justitiae. (0627B, with Migne's colon). Both reproduced with the italic boundary and the colon exactly where he sets them.
1b. CAPUT XVI and CAPUT XXV print without the final stop. That is Migne, and it stands.
Twenty-three of the twenty-five heads print CAPUT N.; two do not — chunk 0003's CAPUT XVI and chunk 0007's CAPUT XXV. The English heads reproduce the presence and the absence 1:1 (CHAPTER SIXTEEN, CHAPTER TWENTY-FIVE, no period). Recorded here so a later blind reader does not report it as a dropped stop: it is the plate's own inconsistency, and Pattern 8 forbids supplying a mark he does not print as firmly as it forbids deleting one he does.
1c. The non-word policy — and why NOT ONE [sic:] was fired in this work
Four forms in this work are no Latin word at all: 0618A Paralimopenis (for Paralipomenis), 0618D raeedificabant (for reaedificabant), 0628A cupuditati (for cupiditati), 0629D ulit (for Tulit, opening a lemma at a paragraph head — the TEI reads <p> <hi>ulit Saraiam,</hi>).
Pattern 7's non-word class says carry them; Pattern 12 says mark the carried run [sic:]. Both were declined, and the reason is the 2026-08-23 addendum, clause A: measured ten times out of ten, a suspicion aimed at a NON-WORD came back as OUR digitization's corruption, not Migne's. A [sic:] would therefore be a public claim blaming the plate for a defect the plate almost certainly does not have — the exact failure that withdrew seven false [sic:] in that batch. The governing rule instead is 13a clause 1: words lost mid-sentence with the plate effectively certain are restored silently and logged. So the English reads Paralipomenon · rebuilt · cupidity · He took Saraias, and all four are named in §4 as tei-patch candidates for a plate read.
⭐ Recorded as method, not as a preference: this is the one place where a reader comparing the columns will find an English word with no exactly corresponding Latin form, and it must be said somewhere.
1d. Etymological glosses of names are glossed Latin equivalents, not scripture
This book is thick with them — Elisaeus salus Dei · Moab … de patre · Naaman decus sive commotio · Athalia temporalis Domini · Josabeth … saturitas Domini · Joas … memoria Domini · Joiada … dilectus Domini · Jectehel … coetus Dei vel auxilium Dei · Aphec … continebit vel apprehendet · Ezechias, apprehendens Dominum vel fortitudo Domini · Nohestan … aes eorum · Thopheth … latitudo · Pharao dissipans, Nechao percussio, Mageddo de tentatione · Matthathia … munus sive donum · Sedecias justus Dominus · Nabuzardan … ventilabrum · Saraias … vinctus · Sophonias … absconditus · Sopher … dissipans vel dividens · Reblatha … multitudo scelerum · Elios enim Graece sol dicitur. Each is translated as the Latin gloss it is, never conformed to a received English form of the Hebrew, and none takes a [var:] — a name-etymology is not a scripture citation. Migne's own spellings of the names are kept everywhere, including where they diverge from the Clementine (§2c).
1e. Pattern 15 — a checked zero
All 168 notes were read. Every one is a citation or an attribution siglum — (RAB.), (ID.), (ISID.), (GREG.), (AUG., serm. 211, de temp.), (BEDA in lib. Reg., quaest. 30, tom. II.), (RAB. ex Euch.), (RAB. ex Bed., ibid.), (ID., ibid.), and the scripture references. Not one is editorial prose, so no [nt:] fires. Recorded as a zero that was checked, not skipped. Note that at 0615B Migne prints the marker with a stop outside it ([n: (RAB.)].) — reproduced as printed.
§2 · THE FULL COLLATION — all 225 italic spans, walked against the Clementine
sources/vulgate/clementine-flat.txt, book 4Rg (Migne's numbering, no conversion). Every span below was looked up at its own verse, not merely phrase-searched: the launch inventory's ✓ marks were treated as an ordering device and re-checked in place.
2a. Spans that AGREE with the Clementine exactly (or are clean clips of it) — 190
Listed so that this report can be told from one by an agent that never looked.
Chapter 1–5 (chunk 0000): Ceciditque Ochozias. (1:2) · Factum est. (2:1) · Fiat in. (2:9) · Si videris. (2:10) · Pallium. (2:13) · Mittite. (2:17) · Ascendit autem. (2:23) · Facite. (3:16) · Parumque. (3:18) · Civitatem. · Lignum. (3:19) · Mulier, (4:1) · Crevit autem. (4:18) · Et ille. (4:29) · Filii prophetarum. (4:38) · Concidit. (4:39) · Vir autem. (4:42) · Naaman, (5:1) · Porro de Syria. (5:2) · Tollam onus. (5:17) · Giezi. (5:20) · Vos ex patre diabolo estis (Joh. 8:44, verbatim).
Chapter 6–10 (chunk 0001): Accidit autem. · Hoc ipsum. (6:5) · Heu, heu, (6:5, 6:15) · Qui cum venissent. (6:14) · Plures. (6:16) · Domine, aperi. (6:17) · Percute. (6:18) · Congregavit. (6:24) · Cras modius, (7:1) · Quatuor, (7:3) · Porro. (7:17) · Vade. (9:1) · Diligite inimicos vestros, et benefacite his qui oderunt vos (Luke 6:27; Migne adds only et).
Chapter 11–12 (chunk 0002): Athalia vero. (11:1) · Ad portam. (11:6) · Scutariorum et custodietis excubias domus. (11:6, two halves of the verse joined) · Et posuit super eum diadema et testimonium. (11:12) · inter se: (2 Par. 23:16) · Sartatecta. (12:5) · Igitur usque. (12:6) · Ad instaurationem templi reddite eam, (12:7) · Et tulit. · Et posuit. (12:9) · Scriba regis, (12:10) · Verumtamen non fiebant. · Non fiebant. · Et fuscinulae, · Thuribula. · Tubae omne vas, (12:13, the last a clip across the verse's et) · Pecuniam vero pro delicto, (12:16) · Dirigatur oratio mea, sicut incensum in conspectu tuo (Ps. 140:2, verbatim) · Quasi tuba exalta vocem tuam (Isa. 58:1, verbatim).
Chapter 13–17 (chunk 0003): Elisaeus autem. (13:14) · Superposuit. (13:16) · Aperi fenestram. · Sagitta. (13:17) · Tolle. Qui cum. (13:18, two clips of one verse) · In anno secundo. (14:1) · Et fecit. (14:3) · In valle. · Petram. · Jectehel. (14:7) · Ipse restituit terminos Israel, and the fuller … ab introitu Emath usque ad mare solitudinis. (14:25, both verbatim) · Quae est in, (14:25) · Ipse aedificavit. (15:35) · Consecravit. (16:3) · Musach quoque. (16:18, twice) · Adduxistis. (17:27) · Et unaquaeque. (17:29) · Et cum. (17:33).
Chapter 18–19 (chunk 0004): Anno tertio. Ezechias. (18:1) · Confregitque. (18:4) · Rabsacen. (18:17) · Eliacim. (18:18) · Haec dicit rex magnus, rex Assyriorum. (18:19) · An speras. (18:21) · In Domino. (18:22) · Dabo vobis. (18:23) · Dominus dixit. · Ascende. (18:25) · Precamur. (18:26) · Non vos, (18:29) · Cumque audisset. (19:9) · Virgo filia. (19:21) · Excelsa. (19:23) · Et siccavi. (19:24) · Nunquid. (19:25, Migne's Nunquid for Numquid is an attested spelling, Pattern 9, not a divergence) · Hoc erit signum. · Comede. (19:29) · Et propter David servum. (19:34) · Factum est igitur. (19:35) · Cumque adoraret in templo. · In terram Armeniorum. (19:37) · Vocabo servum meum Eliacim (Isa. 22:20, verbatim) · Ne timeas a facie verborum quae audisti (Isa. 37:6, verbatim — see §3 on the citation).
Chapter 20–22 (chunk 0005): In diebus illis. (20:1) · Qui convertit. (20:2) · Flevit itaque. (20:3) · Afferte massam. (20:7) · Laetatus est autem in adventu eorum Ezechias. · Domum aromatum, et aurum et argentum. (20:13) · De terra longinqua venerunt. · Ad me de Babylone. (20:14) · Dormivitque Ezechias cum patribus, (20:21) · Josiam. (21:24) · In Secunda. (22:14) · Tradidit in captivitatem virtutem eorum. (Ps. 77:61) · Vidi Dominum sedentem super solium excelsum (Isa. 6:1, clipped before et elevatum) · Ululatus a Secunda. (Soph. 1:10).
Chapter 23–24 (chunk 0006): Contaminavit. · In convalle. (23:10) · Abstulit. (23:11) · Excelsa. · Quae aedificaverat. · Et Chamos. · Offensioni Moab. (23:13) · Tulit ossa, (23:16) · Facite phase. (23:21) · Nec enim factum est phase tale, (23:22) · Joachaz filium Josiae. (23:30) · Argentum autem et aurum dedit, (23:35) · Jerusalem. (24:8) · Artificem et inclusorem, nihilque relictum est, exceptis pauperibus populi terrae. (24:14) · Sepultura asini sepelietur (Jer. 22:19) · quia videbitur Deus Deorum in Sion. (Ps. 83:8; Migne's capital Deorum is his) · k · m · ch · n (the four letters discussed as letters at 0628B–C; kept in italics, untranslated).
Chapter 25 (chunk 0007): Et interrupta est, (25:4) · Mense quinto. · Venit Nabuzardan, (25:8) · Columnas autem aereas quae, (25:13, verbatim including autem) · Ollas quoque. (25:14) · Thuribula et phialas. (25:15) · Tulit quoque. · Saraiam, · sacerdotem primum, · Sophoniam, · sacerdotem secundum, (25:18) · Quinque viros de his qui steterant coram rege · Sopher principem. · Et sexaginta viros e vulgo, · Et sexaginta viros. (25:19) · Et Saraia. (25:23).
Plus the 31 usque ad formula spans (§1a), which quote nothing.
2b. Divergences from the Clementine — the whole list, with the verdict on each
Four fired a [var:]. The rest are logged and unmarked, and the reason is stated for each, because Pattern 14's gate is a divergence that changes the sense, and a marker on every dropped enim would bury the four that matter.
| col | Migne prints | Clementine | verdict |
|---|---|---|---|
| 0614B | Ecce securis ad radicem arboris posita est. | Matt. 3:10 / Luke 3:9 Jam enim securis ad radicem arborum posita est | [var:] FIRED. Two changes at once: the opening formula, and singular tree for plural trees. The gloss is about Christ cutting down the Jews tanquam infructuosas arbores — plural — so the number is doing work in the sentence and the divergence is visible to any reader. |
| 0615A | Nolite timere. | 4 Kings 6:16 Noli timere, singular | [var:] FIRED. A number change in the addressee, which Pattern 17 treats as load-bearing throughout this edition; and Migne's plural is exactly what lets the gloss answer it with the plural Gospel word (Nolite timere eos qui occidunt corpus). Rendered "Fear ye not." so the plural is visible without the apparatus. |
| 0618D | Unicuique autem vestrum data est gratia | Eph. 4:7 Unicuique autem nostrum | [var:] FIRED. us → you. This is the prose analogue of the ἡμεῖς/ὑμεῖς class the style file names as a hazard; the whole paragraph is about apostles apportioning gifts to their hearers, so Migne's vestrum changes who is being spoken of, not merely how. |
| 0619B | Et non fiebat ratio iis omnibus, | 4 Kings 12:15 iis hominibus | [var:] FIRED. A substantive word swap: "with all these" for "with those men." The gloss that follows turns on who was trusted (Tam religiosi fuerunt, ut nullus dubitaverit…), so the reader needs to know the plate reads omnibus. |
| 0611A | ut apprehenderent Jesum | Joh. 7:32 ut apprehenderent eum | logged, no marker — the referent is identical and the substitution is the ordinary one an author makes when quoting out of context. |
| 0611B | et quinquaginta viros | 4 Kings 1:9 et quinquaginta qui erant sub eo | logged, no marker — the count is the same and nothing in the gloss (which expounds the number fifty) turns on it. |
| 0611C | qui ascendens in, altum captivam duxit captivitatem, dedit dona hominibus | Eph. 4:8 Ascendens in altum, captivam duxit captivitatem : dedit dona hominibus | see §3 — the qui is Migne's own stitching and the comma is a punctuation crux, not a textual one. |
| 0612C | ut aliquis vos doceat | 1 Joh. 2:27 ut aliquis doceat vos | logged, no marker — word order only. |
| 0612C | Bibent homines et jumenta | no such sequence anywhere in the Clementine | logged, no marker — the words are close to 4 Kings 3:17 (bibetis vos … et jumenta vestra), the very verse being glossed, and to Jonah 3:7–8; the note reads (Joan. III), which is a citation misprint. Citation repair belongs to citation-corrections.json at index time, never to the running text. |
| 0613C | Littera occidit | 2 Cor. 3:6 littera enim occidit | logged, no marker — a dropped connective. |
| 0613C | Prudentia carnis mors est … vita | Rom. 8:6 Nam prudentia carnis, mors est … vita et pax | logged, no marker — head-and-tail clipping, the standard way this gloss quotes. |
| 0613C | de Baalalisa (twice) | 4 Kings 4:42 de Baalsalisa | logged, no marker — and deliberately kept: the gloss is the etymology (Baalalisa enim tertium habens interpretatur), so Migne's own form is the form the argument uses. Rule 2 of the "each commentary stands on its own" ruling. |
| 0614A | quos praescivit, hos et vocavit | Rom. 8:29–30 quos praescivit, et praedestinavit … quos autem praedestinavit, hos et vocavit | logged, no marker — a conflation of two consecutive verses, not a variant reading; nothing in the gloss turns on praescivit against praedestinavit. |
| 0614C | et hoc mutuo acceperam | 4 Kings 6:5 et hoc ipsum mutuo acceperam | logged, no marker — and note Migne prints Hoc ipsum. whole as its own lemma four lines earlier, so nothing is concealed. |
| 0614D | Quod confregit lignum et jactavit, et sic ferrum sustulit. | not the Clementine | not a divergence at all: Migne introduces it in alia translatione, and a text quoted as another version is not a variant of the Vulgate. No marker, by the same logic that keeps [var:] off a chant text (L3). |
| 0615A | Nolite timere eos qui occidunt corpus, et post non habent quid faciant vobis | Luke 12:4 / Matt. 10:28 | logged, no marker — unitalicized in the Latin, i.e. the gloss speaking, not a lemma. Same for Possum rogare Patrem meum… (Matt. 26:53, where the Vulgate's An putas, quia non possum is a question and Migne's is an assertion) — the gloss is answering Plures, and it is not set as a quotation. |
| 0616A | Vae vobis … qui clauditis | Matt. 23:13 Vae autem vobis … quia clauditis | logged, no marker — relative for causal; the negatives (non intratis, nec … sinitis) are identical in both. |
| 0616B | Benedab, rex Syriae, | 4 Kings 8:7 Benadad rex Syriae | logged, no marker — a proper-name spelling, and see §2c. |
| 0616C | Regnavitque. | 4 Kings 9:29 regnavit | logged, no marker — an added enclitic. |
| 0616C | Estne cor tuum, | 4 Kings 10:15 Numquid est cor tuum rectum | logged, no marker — the same question recast; no negative is added or lost. |
| 0616D | Hic est dilectus Filius meus | Matt. 3:17 Hic est Filius meus dilectus | logged, no marker — word order. |
| 0617A | misit Joiada pontifex | 4 Kings 11:4 misit Jojada (no title) | logged, no marker — the title is true of him and the gloss below calls him summus pontifex on its own account; the lemma carries etc. and is a clip. |
| 0617C | Et imposuerunt ei diadema dederuntque… | 2 Par. 23:11 …diadema et testimonium, dederuntque… | logged, no marker — a clip inside a text Migne cites as the parallel (In libro Dierum apertius). |
| 0617D | Pepigit igitur Joiada | 4 Kings 11:17 Pepigit ergo Jojada | logged, no marker — synonymous connective. |
| 0618B | Praedicate Evangelium omni creaturae, infirmos curate … gratis date | Mark 16:15 + Matt. 10:8 (which also has daemones ejicite) | logged, no marker — a conflation of two gospels under one reference, the same class as 0614A. |
| 0619A | Cum complessent | 2 Par. 24:14 Cumque complessent | logged, no marker. |
| 0620A | Tu autem, homo Dei, haec fuge, sectare vero justitiam et pietatem, et apprehende vitam aeternam | 1 Tim. 6:11 (o homo Dei … justitiam, pietatem, fidem, caritatem, patientiam, mansuetudinem) + 6:12 (apprehende vitam aeternam) | logged, no marker — an abridged list spliced to the next verse. |
| 0620D | Anno vicesimo Azarias | 4 Kings 15:1 Anno vigesimo septimo Jeroboam … Azarias | logged, no marker — a clipped lemma; vicesimo/vigesimo is an attested spelling pair (Pattern 9). |
| 0620D | Hepher | 4 Kings 14:25 Opher | logged, no marker — and kept, because the gloss builds on Hepher as a Hebrew word for darkness/cloud. |
| 0621B | dum recordaremur | Ps. 136:1 cum recordaremur | logged, no marker. |
| 0621C | fabricata est Deum suum | 4 Kings 17:29 fabricata est deum suum | logged, no marker — rendered lowercase "god" in the English: the sentence asserts that each nation made an idol for itself, and an English capital G would assert the opposite. Migne's capital is recorded here rather than in the text. |
| 0621C | Viri enim Babylonici | 4 Kings 17:30 Viri enim Babylonii | logged, no marker — both mean "of Babylon". |
| 0622A | Nohestan vocavit, | 4 Kings 18:4 vocavitque nomen ejus Nohestan | logged, no marker — a lemma reordered to open on the name it glosses. |
| 0622D | Non seducat vos Ezechias, et non vobis tribuat fiduciam super Domino Deo vestro. | 4 Kings 18:29–30 Non vos seducat Ezechias : non enim poterit eruere vos … neque fiduciam vobis tribuat super Domino | logged, no marker — word order plus Deo vestro; both printed negatives are carried and neither is relocated. |
| 0623C–D | the long Comede hoc anno… | Isa. 37:30–32 (in anno secundo pomis vescere · seminate et metite, et plantate · fructum earum · de Jerusalem · faciet istud) | logged, no marker — an abridged running quotation Migne himself frames as Isaias ita. |
| 0624A | principemque exercitus Assyriorum | 2 Par. 32:21 principem exercitus regis Assyriorum | logged, no marker. |
| 0625D | Non fuit verbum quod non monstraret | 4 Kings 20:13 has no verbum; Isa. 39:2 does (Non fuit verbum quod non ostenderet) | logged, no marker, and this is the most interesting non-marker in the book: the divergence is the point of the gloss, whose parenthesis (quod apud Hebraeos pro re frequenter accipitur) exists precisely to explain verbum. Firing [var:] against 4 Kings would tell the reader Migne's text is wrong when in fact he is conflating the two canonical parallels on purpose. The second span's ostenderit against Isaiah's ostenderet is the same seam. |
| 0626A | sepultus est Ezechias super sepulcra regum | 2 Par. 32:33 sepelierunt eum super sepulchra filiorum David | logged, no marker — unitalicized, an indirect report. |
| 0627B | et orietur vobis timentibus sol justitiae | Mal. 4:2 orietur vobis timentibus nomen meum sol justitiae | logged, no marker — the resumption tail of an abridgment, clipped like every other. |
| 0627C | Non utique in fermento malitiae et nequitiae | 1 Cor. 5:8 non in fermento veteri, neque in fermento malitiae | logged, no marker — clipping; the negative count is unchanged in what Migne prints. |
| 0627C | planctus Adremon | Zach. 12:11 planctus Adadremmon, and magnus erit planctus | logged, no marker — unitalicized, and a name spelling. |
| 0628B | Regnavitque Joachim, | 4 Kings 24:6 regnavit Joachin | logged, no marker — but flagged, because the gloss contradicts its own lemma: it rules that the son's name is written per ch et n and the father's per k et m, yet the lemma (of the son) prints ch + m. See §3. |
| 0628C | …gladium aut lanceam ad repugnandum | 1 Kings 13:19 ends at lanceam | logged, no marker — an added tail. |
| 0629D | Escae ejus electae sunt [n: (Job XXXIII)] | nowhere in the Clementine; Job 33:20 reads Abominabilis ei fit in vita sua panis, and the nearest thing in the canon is Hab. 1:16 cibus ejus electus | logged, no marker — the source cannot be identified with enough confidence to name a "received reading", so [var:] would be an assertion I cannot support. Rendered as printed; the reference is a candidate for citation-corrections.json. |
| 0630A | Et tres inani, | 4 Kings 25:18 et tres janitores | [cj:] FIRED — see §3. |
2c. The proper-name spellings, gathered — Migne's forms kept everywhere
Elisaeus (Clem. Eliseus) · Baalalisa (Baalsalisa) · Benedab at 0616B against the work's own Benadad at 0615B (Benadad) · Joiada (Jojada) · Hepher (Opher) · Nohestan · Beenon / Beennon / Gehennon / Ennon / Ennom within four lines of each other at 0621A and 0626C (Clem. Ennom) · Nebahan at 0621C against Nebahath at 0621D, and Tharthac against Tharthach in the same pair (Clem. Nebahaz · Tharthac) · Adremon (Adadremmon) · Joachim (Joachin) · Zoaros · Jectehel · Sochoth Benoth (Clem. Sochothbenoth).
None is marked. They are rendered as printed, each occurrence on its own, including where Migne is inconsistent with himself two lines apart — that inconsistency is his and the parallel Latin column shows it. Harmonizing them would be exactly the editorializing the corpus rules forbid.
§3 · CRUCES — column, printed reading, the reading taken, and why the rival lost
0614B · deservit — the printed word is real, and the English it yields is false. [cj:]
Ab eis interveniente passione corpus ipsum deservit: profunda descendit…
deservit is a perfectly good present of deservio, "serves diligently" — so Pattern 12 has nothing to wrap and Pattern 7 requires it to be rendered. But deservio does not take ab, and the allegory is the axe-head flying off the haft: the body leaves them, goes down into the deep, and comes back. deseruit (perfect of desero) is the only form that takes ab eis and the only one the next three clauses continue.
Rendered "the body itself does service", with [cj: deservit; read deseruit, "forsook"] beside it. The rival — leaving it as a crux only — lost on Pattern 18's founding ruling: munde/mundo at 11081 was left as a crux and the page went on asserting what the author did not, "where a reader would not meet it". The marker is additive: Migne's word keeps its English and the conjecture stands next to it. ⚠ It is not a [sic:], because nothing is broken; and I do not claim the plate is at fault — the single-letter difference could as easily be the digitization's (§4).
0630A · Et tres inani — the gloss proves the word it does not print. [cj:]
Et tres inani, qui ostium fidei aperire debuerant, et tenentes clavem scientiae…
4 Kings 25:18 reads et tres janitores, and the gloss expounds them as door-keepers in its own next breath — they "ought to have opened the door of faith" and hold "the key of knowledge". inani is a real word (dat./abl. sg. of inanis) and so is Migne's under addendum A; but it does not agree with tres and yields no clause.
Rendered "And three empty," with [cj: inani; read janitores, "doorkeepers"]. The rendering is as literal as the printed words allow: English adjectives carry no case, so the number and case mismatch has no English exponent (Pattern 9) and "empty" is what the word says. The rival readings: (a) rendering "door-keepers" silently — rejected, that is the conformation Pattern 7 exists to forbid, and it destroys the evidence; (b) a [sic:] — rejected, it would blame the plate for a defect I have not seen; (c) crux only — rejected for the same reason as 0614B, and more strongly here, because the surrounding gloss makes the false English ("three empty ones" who hold the key of knowledge) actively misleading.
0616B · ultione juxta perimeret — crux only, no marker
…quem Dominus destinavit ut sacrilegam civitatem … ultione juxta perimeret…
justa is the obvious reading: it agrees with ultione and gives the sentence its point. But juxta is a real word with a real adverbial sense ("alike, equally"), and addendum A is explicit that a real word reading oddly is Migne's until a plate says otherwise. Rendered "should destroy with vengeance alike".
Why this one did not earn the [cj:] that 0614B and 0630A did, and the line is worth stating because it is the same class: at 0614B the construction (ab eis) forbids the printed word, and at 0630A the gloss's own next clause names the conjecture. Here neither holds — juxta construes, however oddly, and nothing in the sentence requires justa except taste. Reading well is what the wrong host gives you (test 2a), and it is what the wrong conjecture gives you too. Crux, no marker; plate candidate in §4.
0611C · qui ascendens in, altum captivam duxit captivitatem — a comma English cannot host
Migne (Eph. 4:8) sets the comma between the preposition and its noun. English has no position corresponding to that; a comma between "on" and "high" is not a mark, it is a typo. Rendered who ascending on high, led captivity captive, gave gifts to men — the mark is kept, one word later, where English can carry it, on exactly the reasoning of 8b (the English refuses the mark because of English grammar, not because we improved his sentence). Nothing is supplied and nothing deleted; the chunk's colon/semicolon/comma bands still compare 1:1. ⚠ The qui opening the quotation is Migne's own stitch into his sentence, not part of Eph. 4:8.
0614B · Dominum suum with a capital D — rendered "his own lord"
…et super eam consistens, Dominum cujus erat terra, adoraret, ne offenderet Dominum suum si pariter non adoraret.
The first Dominum is God, and Migne capitalizes the second identically. Rendered lowercase, because 4 Kings 5:18 — the verse the gloss is expounding — is Naaman asking pardon for bowing in the house of Remmon when my lord (dominus meus) leaneth on my hand. The rival (both capitals, both God) makes the sentence say Naaman feared to offend God by not worshipping God alongside himself, which is not a sentence. Latin capitalization is Migne's compositor's; the crux records it and the facing column shows it.
0611C · non esse discipulus super magistrum postulavit — the nominative kept, the negative kept
discipulus is nominative where the infinitive wants an accusative. No English exponent (Pattern 9): rendered "he asked not to be a disciple above his master", with the negative on esse, where Migne prints it. The gloss is alluding to Matt. 10:24 Non est discipulus super magistrum, and the alternative parse — "he asked that the disciple should not be above the master" — is not excluded and would need the nominative. Both readings construe and both assert the same thing, which is why nothing is marked; recorded because a blind reader will stop here.
0628B · Regnavitque Joachim — the gloss's own rule contradicts its lemma
The gloss rules that the name is written per k et m for the father and per ch et n for the son, then prints, as the son's lemma, Joachim — ch + m, neither. The Clementine reads Joachin at 24:6. Rendered as printed and the four discussed letters kept in italics, untranslated, as the mentions of letters that they are. No marker: the defect, if it is one, is in the plate's own consistency and is fully visible in the facing column.
0620C · quod hanc vel coetus populi Dei vel Deo adjuvante ceperit — an ambiguity kept
The two limbs of the vel … vel are not parallel: a nominative (coetus populi Dei) against an ablative absolute (Deo adjuvante), both governed by one singular ceperit whose subject is Amasias. Rendered "in that he had taken it either as the assembly of the people of God, or with God helping" — the one construction that lets the printed singular verb stand. Nothing emended; the rival (making coetus the subject of ceperit) breaks the number.
0622B · quem tradunt Hebraei … prodidisse Jerusalem — the antecedent is Sobna, not Eliacim
The paragraph is headed Eliacim, and the bare quem four words after post Sobnam invites the wrong reading. The sentence settles it itself: it ends quem etiam Sobnam cum Eliacim ad Rabsacen exisse quidam autumant. The English keeps the relative rather than naming the man, exactly as Migne does, so nothing is supplied — but this is the unexpressed-referent class the brief names as dominant, and a reader who takes it as Eliacim will have the betrayal on the wrong man.
0626B · reduxitque eum in regnum suum — a supplied subject, named
The verb's subject is God (2 Par. 33:13, reduxitque eum Jerusalem in regnum suum, of the Lord), but the nearest English noun is Manasses, and a bare "he" would make Manasses restore himself. Rendered "and the Lord brought him back" — the one supply of a noun in the whole work, logged here because the rule requires it to be logged.
0626A · quia princeps Sodomorum appellavit eos — a nominative for an accusative plural
Isa. 1:10 reads principes Sodomorum. Migne prints the nominative singular princeps against a plural object eos. Rendered literally, "because he called them prince of Sodom" — the English is odd because the Latin is. Not marked: a case/number slip is Pattern 9's class and a real word is Migne's until the plate says otherwise. Plate candidate, §4.
0623D · comede quae iterum germinare, — an italic boundary falling mid-construction
potuerunt stands only in the second italic span and governs both. English cannot leave the first span without a modal, so "eat what could again sprout," carries "could" that the Latin span does not print. The alternative — leaving the first span verbless — is not English and would read as a dropped word of ours. The italic boundary is 1:1 with Migne.
0625C · latrocinantium spiritum incursione — accusative for genitive plural
spirituum is what the sense wants. Rendered "by the onset of robber spirits". Case slip, Pattern 9, no marker.
0620B · Valli Salinarum erat — a dative/locative for a nominative
Rendered "It was at the Valley of the Saltpits". No English exponent for the case, no marker.
Sentences Migne leaves without a terminal stop — three, all kept open
0612C …doctrinam accipient, ingeniosi et simplices · 0627B …Deo retribuente secundum merita · 0628C …componere possit vel resarcire. Pattern 8: never supply a mark the plate lacks, a terminal stop included. All three English sentences stop where his do.
0621B · sicut alia multa Achaz sibi tribuente. quo inveniret gratiam illius.
A full stop mid-sentence followed by a lowercase quo. Reproduced exactly — stop, lower case, and all. This is the 11542/11436 class where the mark had been silently swallowed; it is not swallowed here.
0628D · quae significatur denario, numero.
A comma between denario and numero that leaves numero stranded. Rendered "which is signified by ten, the number." Preserved as printed; no marker, nothing is broken.
0626C · Josias Thopheth … dispergens, quatenus … appareret.
A sentence with a participle and no main verb. Rendered as it stands. Migne's, not ours; no marker exists for "a sentence that will not close", and inventing one for a single site is worse than saying so here (the 8968 @0534A precedent).
0623B · [n: (Isa. XLI)] on Ne timeas a facie verborum quae audisti
The text is Isa. 37:6 verbatim (and 4 Kings 19:6 in Jerome's other wording, Noli timere a facie sermonum quos audisti). The reference is a misprint. Rendered as printed; citation repair belongs to citation-corrections.json at index time, never to the running text. Same class: [n: (Joan. III)] on Bibent homines et jumenta (0612C), [n: (Job XXXIII)] on Escae ejus electae sunt (0629D), and [n: (Matth. V)] on Quinque viros de his qui steterant coram rege (0630B), which is 4 Kings 25:19.
0611D · Lethi fluvium dirupit — "the river of death"
Lethi is read as the genitive of lethum/letum, death, against the proper name Lethe. The clause continues nobisque transitum ad vitam praeparavit and the whole gloss is Elias dividing the Jordan as a figure of the Incarnation; a classical underworld river would be an intrusion in a sentence whose own antithesis is death/life. Not marked.
§4 · PLATE CANDIDATES — READ AND SETTLED (orchestrator, 2026-08-24)
⭐⭐ The plate was read the same session, PL 113 patrologiaecurs04migngoog on disk, page map calibrated against the printed corner numbers before any crop, each crop taken wider than its target and the whole line read.
| site | our file | the plate at 450 dpi | verdict |
|---|---|---|---|
| 0618A | Paralimopenis | Paralipomenis | OURS — patched |
| 0618D | raeedificabant | reædifi- / cabant | OURS — patched |
| 0628A | cupuditati | cupiditati | OURS — patched |
| 0629D | ulit Saraiam | — page clipped — | UNREADABLE in this scan |
Three of the four non-words are ours and are patched in data/tei-patches/8955.json; the work is re-chunked and verify-english passes with no boundary moved. §1c's policy is vindicated as written: not one [sic:] was fired at any of them, so this work never asserted a defect in Migne's plate, and the repair belonged in our Latin.
⚠ 0629D — UNREADABLE, and not guessed
This copy of PL 113 clips the left margin on some pages. PDF p. 320 (cols 629/630) is one of them — the left column's opening characters are cut off, and the line under test reads t Saraiam, qui interpretatur vinctus, sacer- with whatever preceded the t lost with the margin. The reading the context points to is Tulit Saraiam (4Rg 25:18 Tulit quoque princeps militiae Saraiam sacerdotem primum, and the gloss two lines above lemmatizes Tulit quoque) — but a contextual reconstruction is a conjecture, and a conjecture entered as a patch would wear Migne's own voice. It stays exactly as our source has it. ⚑ I Esdrae has one candidate waiting on the same defect (0697A magistorum); a second PL 113 scan settles both in two pages.
The candidates as the translator filed them
Cheap, and named explicitly because three of the four normalization-class finds in this series surfaced exactly this way. Every entry below is a request, not an assertion; none fires a marker.
Non-words (addendum A says these have been OUR corruption ten times out of ten):
| col | our Latin reads | almost certainly | class |
|---|---|---|---|
| 0618A | Paralimopenis | Paralipomenis | metathesis, the Nephtlali class |
| 0618D | raeedificabant | reaedificabant | doubled/transposed vowel |
| 0628A | cupuditati | cupiditati | vowel |
| 0629D | ulit Saraiam | Tulit Saraiam | lost initial capital at a paragraph head (TEI: <p> <hi>ulit Saraiam,</hi>) |
Real words that read oddly — Migne's until the plate says otherwise, and rendered as printed: 0614B deservit (for deseruit) · 0616B ultione juxta (for justa) · 0620D sive nubea (a form that is no standard Latin adjective; rendered "or cloudy") · 0626A princeps Sodomorum for principes · 0630A inani (for janitores).
A "suspiciously tidy" check, per brief §4 and addendum C, run on the passages that are otherwise rough: the four Beenon / Beennon / Gehennon / Ennon spellings at 0621A, and Nebahan/Nebahath · Tharthac/Tharthach at 0621C–D, are inconsistent, not tidy — which is evidence that the digitization has not normalized them, and is worth recording as a checked negative. Nothing else in the work read as suppressed roughness.
Migne's foot-of-page apparatus: none is available. data/plate-notes/ carries no entry for 8955, so no [cn:] reaches the Latin chunks, and no crux above could be settled from his own notes. This is a gap, not a checked zero, and it is stated as one — unlike Ruth, whose plate pages were read and whose zero is real.
§5 · READ CLEAN, AND THE CANDIDATES THAT WERE REJECTED
Read clean
- Column anchors: 80, all verified for presence, order AND position (test 2b) — each was checked against the word it stands beside in the Latin, not merely that it is present. Two are deliberately split mid-construction to match Migne: sacerdotem [0630A] primum and et [0620B] potius ibi quam hic. ⚠ At 0630A the anchor stands one word earlier in the English (the first [0630A] priest) than in the Latin (sacerdotem [0630A] primum), because English fixes the order of adjective and noun; there is no rendering that puts it between them and remains English.
- Negation: 110 printed negatives across the eight chunks, every one present, none invented. Each was checked by test 2's method — name the single Latin word it attaches to, then find that word in the English and check the negative is on it.
- Test 2a's trigger population was worked explicitly — every negative resting on a verb with a quantifier, correlative or predicate nearby. Twenty sites, worked one by one: 0611C non nisi … nisi … non potest · 0614C non solum … sed etiam · 0615B non eos occidi permisit, sed · 0616B non invasione hostium fuisset factum (the 8960 @0740B model: the non stands before the verb, and et … fuisset factum shows it governs the phrase) · 0619A non de eadem pecunia … sed de reliqua · 0619B non minus fideles … quam · 0619C non proprio cujusque, sed sacerdotum judicio · 0620A non debet … sed magis · 0620B sed non corde perfecto (twice) · non civitatem dicit, sed magnam petram (0620C) · 0622A non solum … quin · 0622C non contra Deum, sed pro Deo · 0622C non de imbecillitate populi … sed quia · 0622C non sua voluntate, sed Dei · 0624A non suo merito, sed Dei clementia · 0624B nec angelico perit gladio … sed parricidio · 0624B non ad montes generaliter Armeniae, sed ad montes Tauri · 0625D non solum thesauros … sed et templi · 0627B non solum pecora … sed et seipsos · 0629D Non solum populares, sed et fortes. In every one the host is the phrase or the quantifier, not the adjacent verb, and the English puts it there.
- Question marks: 2, both in the Augustine extract at 0615D (Cur famem, o miserrime, pateris? Cur tanta penuria laceraris?), both direct, both carried. No indirect question closed with
?anywhere in this work, so Pattern 8b never fires — a checked zero. One exclamation, Proh nefas! at 0615D, carried; one more, Quomodo sedet sola civitas! at 0627C, carried. - Pattern 17: the work's second persons are Augustine's address to a single sinner at 0615D (pateris · laceraris · vita · compesce · convertere · relinque · poteris · in te), rendered thou throughout, and the plural scripture addresses (Vos ex patre diabolo estis · Nolite timere … vobis · Vae vobis … non intratis · Diligite inimicos vestros · gratis accepistis, gratis date · Non seducat vos Ezechias · Unicuique autem vestrum), rendered you. The lemma Estne cor tuum and the darts of 0620A (Tu autem, homo Dei) are singular and take thou.
- Pattern 4 does not apply (runbook false positive 4): this is a lemma-and-gloss commentary, not a florilegium, and no
[f:]is tagged. - Column-band arithmetic (runbook false positive 1): the bands run 0611A→0630D with the ordinary irregularities and three anchors —
[0621C],[0623D],[0630A]— sitting inside an italic span. Nothing is reported.
Rejected, with the reason
- A
[sic:]on any of the four non-words. Rejected on addendum A + B: a[sic:]is a public claim about the plate, and the measured base rate says these four are ours. §1c. - A
[var:]on Benedab (0616B) or any other proper-name spelling. Rejected as a class: Migne diverges on a dozen names, is inconsistent with himself on four of them within two lines, and marking each would put a maroquin note beside every third proper noun while telling the reader nothing he cannot see in the facing column. Gathered in §2c instead. - A
[var:]at 0625D on Non fuit verbum.** Rejected because the "received text" is ambiguous here — the word is absent from 4 Kings 20:13 and present in Isa. 39:2, and the gloss is expounding precisely that Hebraism. A marker would assert a divergence where Migne is quoting a parallel. §2b. - A
[var:]on Possum rogare Patrem meum… (0615A), where the Vulgate's Matt. 26:53 has An putas, quia non possum. Rejected because the sentence is not italicized in the Latin: it is the gloss answering the lemma Plures, not a quotation set as one. The polarity difference is real and is logged; the marker is reserved for quoted scripture. - A
[cj:]on ultione juxta (0616B). Rejected — §3. The conjecture is attractive and nothing in the sentence compels it. - Reporting the missing stops at 0612C, 0627B, 0628C. Rejected: Pattern 8. They are Migne's and they are now on record.
- Reporting the two heads without final stops. Rejected as a defect, recorded as a fact: §1b.
- Harmonizing sartatecta. The word appears as a lemma at 0618A and again in the gloss at 0618D; rendered "the repairs" in both, which is reconciliation within one work — required — and not the cross-work harmonizing that is forbidden.
- Reading Regnavitque Joachim as Joachin to agree with the gloss's own rule (0628B). Rejected: that is conformation, and the mismatch between a gloss and its lemma is precisely the kind of evidence this edition exists to keep. §3.
- Em-dashes. Not marked, not reported, and kept rare: the whole work carries none.
§6 · THE BLIND POLARITY READ (2026-08-24) — three sites, two acted on
A reader fenced from this file read all eight chunk pairs and reported before opening it. Its twenty-site walk of the test-2a trigger population matched this file's own site for site with no host misassigned — which is the evidence that the two passes were independent rather than one pass repeated.
6.1 · 0628A — an indirect reflexive that English pulled onto the wrong man (repaired)
…nam Nechao vinxit eum, eo quod populus contra suam voluntatem constituisset eum regem…
The English read "the people had made him king against his own will", which lands on the nearest person — Joachaz — and says he was crowned unwillingly. That is not what the sentence is for. Suam cannot be reflexive to populus, who are the ones doing the making; it is the indirect reflexive, referring to the subject of the governing clause, Nechao. The people crowned Joachaz without Pharao's leave, which is precisely the reason (eo quod) Nechao bound him — and 4 Rg 23:30–34 with 2 Par 36:1 is that story. Now "against Nechao's will."
⚑ English has no indirect reflexive, so the referent must be named; the name is a supply, and it is recorded here as one.
6.2 · 0630A — ⭐⭐ SETTLED AT A SECOND WITNESS: the broken letter is MIGNE'S, and the marker changed
Read this first; what follows it is the reasoning as it stood before the plate was in hand.
Our archive.org copy of PL 113 clips the left margin of p. 320, so this word could not be read there and was recorded as unreadable. A second witness settles it: Gallica bpt6k5505319w, f315 (f = (column + 1) / 2, calibrated against the printed corner numbers first), which carries the full margin. It prints:
⁻ ulit Saraiam, qui interpretatur vinctus, sacer-
Migne's own plate prints ulit — nothing but a broken fragment of a raised stroke where the T should stand. ⭐ The contrast is on the same page: VERS. 18. — Tulit quoque. prints its T cleanly two lines above. So the missing letter is a failed sort in Migne's forme, not a loss of our source's: Corpus Corporum transcribed him faithfully, and there is nothing to patch.
⚑ The marker therefore changed from [cj:] to [sic: ulit]. The reasoning below was right that the English must stop supplying the letter silently and wrong about which instrument to use — [cj:] was chosen precisely because a [sic:] asserts a defect in the plate and the clipped page could not support that claim. With the plate in hand it can. The marker-strip test passes: delete it and He took Saraias still has its verb and its object.
⛔ The general lesson, and it is the one worth carrying: "unreadable" was a fact about OUR COPY, not about the record. One volume can have two scans with different margins, and the second one answered in a single page fetch what the first could not answer at all.
The reasoning as it stood before the second witness
The sharpest catch of the four blind reads, and it is about consistency between our two columns.
Our Latin prints ulit Saraiam, — the lemma's initial is missing, and §4 above records that the plate could not settle it (PDF p. 320 is clipped at the left margin) and therefore deliberately did not patch it. That decision stands and is right. But the English was printing "He took Saraias" — supplying the very letter the Latin column had refused to supply, silently, while this same work's two other conjectures (0614B deservit, 0630A inani) both ride [cj:] markers. One column declined; the other asserted; nothing told the reader.
Now: He took Saraias, [cj: ulit; read Tulit, "he took"]. The gloss is additive in the Pattern 18 form — the rendering stands, the conjecture stands beside it, and the printed word is quoted so verify-english can check it against the twin.
⚠ The judgement call, stated plainly for Wilson. Pattern 18's founding case is a word that is REAL and misleads; ulit is not a real word, so [sic:] would be the shape — and [sic:] is exactly what must not fire here, because it asserts a defect in Migne's plate that the clipped page cannot support. [cj:] was chosen because it claims only read Tulit, which is a claim about the text as our own Latin column prints it, checkable by any reader with both columns. ⚑ The reading is not in doubt: 4 Rg 25:18 prints Tulit quoque princeps militiae Saraiam sacerdotem primum, and Migne's own lemma two lines above, in our own file, reads Tulit quoque. What is in doubt is whose loss the missing letter is — and the marker deliberately says nothing about that.
6.3 · 0619B — qui for quin: declined, and why
Tam religiosi fuerunt, ut nullus dubitaverit qui pecuniam Dei sine fraude tractarent. The reader noted that the English ("no man doubted that they handled…") reads the printed qui as quin, and asked whether a [cj:] was owed on the model of 0614B and 0630A.
Declined. Qui for quod/quin after a verb of doubting is attested medieval usage, not a defect — Pattern 9's population exactly — and the reader itself observed that the assertion is identical either way. A [cj:] here would be a public claim that Migne's word is wrong when it is merely late. ⚑ This is the addendum's rule applied to a function word: a real, attested form is Migne's until a plate says otherwise, and here there is nothing for a plate to settle.
6.4 · Migne's own, found rendered literally — no action, now on the record
0614B deservit · 0630A inani · 0616B juxta · 0626A princeps Sodomorum · 0620C the Dead Sea's reversed breadth and length · 0621B the mid-sentence full stop. All reproduced as printed, all confirmed correct by the blind reader.