Liber Judicum

Translated 2026-08-19, one translator, 6 chunks (4,698 Latin words), on Opus. verify-english.mjs clean: columns, notes and sections aligned, no duplicate paragraphs. Apparatus: 23 [var:] · 2 [cj:] · 3 [sic:] · 0 [d:] · 0 [ed:] · 0 [nt:] · 116 [n:] (exact parity with the Latin twins, 18/26/25/11/28/8). Structure: PRAEFATIO, CAPUT PRIMUM–CAPUT XXI (21 chapter heads across six chunks). Zero guillemets in either language; zero em-dashes introduced.

Conventions are inherited from Liber Ruth (8968) §1 and were not re-derived: 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, usque ad → as far as (the shipped Glossa/OT rendering, 18 instances in 8968/8970).

§1 · PLATE STATUS — A CHECKED ZERO, AND WHERE IT DOES NOT REACH

Two pages of this work were read at 300 dpi before translation (p. 266, cols 521/522; p. 268, cols 525/526, full width): no foot-of-page apparatus. Consistent with Ruth's five pages and five further pages across two other runs of PL 113. So there is no [cn:] to expect in this book, and no conjecture of Migne's is available to settle any crux below.

⚠ That is a sample, not a proof, and three sites below would be settled instantly by a plate read that has not happened: 0522B seniorunt, 0525B periculam, 0530A In icat (§3). All three are corpus-wide hapaxes; all three are handled as [sic:] on 8960's precedent (subju averit, mise unt, solli itudo, qui ibet, corre ti, fallaci — same volume, same batch, same shape) and each is flagged here as a candidate for a plate read, not a finding. Two further sites want the plate for a different reason: 0532B CAPUT XVIII and 0532C CAPUT XX, the only two of twenty-one chapter heads printed without a final period (§4.3).

§2 · THE VULGATE COLLATION — the deliverable

Collated against sources/vulgate/clementine-flat.txt (Jdc = Judges). Every italic span in the launch list was looked up. Divergences are marked [var:] in the English and named here; agreements are named here too, individually, because a list of findings alone cannot be told from a list by someone who did not check.

2.1 · Divergences marked [var:] (23)

2.2 · Divergences CHECKED and deliberately NOT marked, with the reason

• 0523C Ego sum via, veritas et vita ‖ Jo 14:6 Ego sum via, et veritas, et vita. One conjunction. No sense turns on it; [var:] fires on sense (Pattern 14), not on orthography.
• 0529D ubi erit fletus et stridor dentium ‖ Mt 8:12 ibi erit. ubi for ibi is syntactic accommodation to the host sentence (aut omnes gehennae praedestinati: ubi erit…), not a claim about a different text.
• 0530D Exinde mos increvit ‖ Jdc 11:39 Exinde mos increbruit. Cognate verbs, same sense ("grew" / "grew prevalent"). Named here so a blind reader knows it was seen.
• 0531A Orietur vobis sol justitiae, et sanitas in pennis ejus ‖ Mal 4:2 Et orietur vobis timentibus nomen meum sol justitiae…. This is clipping, not divergent wording; every word Migne prints stands in the Vulgate in that order.
• 0526C si noster es, an adversariorum? ‖ Jos 5:13 Noster es, an adversariorum? The si is Migne's own subordination under inquirit; the question mark he prints is kept by taking the question direct behind the colon (Pattern 8b).
• 0523A Fuit in eo Spiritus Domini ‖ Jdc 3:10 Fuitque in eo spiritus Domini. The dropped -que is recovered in Migne's own second lemma on the same verse (Fuitque in., 0523B). Both lemmata are rendered as the phrase they stand for; the difference is visible in the facing Latin column and nowhere lost.
• 0523B Servieruntque Israel ‖ Jdc 3:14 Servieruntque filii Israel. Clipping.
• 0527C Qui manu et lingua ‖ Jdc 7:5 Qui lingua lambuerint aquas. Migne's lemma anticipates the LXX form he is about to discuss; the discussion is explicit, so no marker is needed for the reader to see it.
• 0525B Auribus percipite, omnes satrapae ‖ Jdc 5:3 auribus percipite, principes. Gloss prose, not a lemma, and Migne's next sentence glosses satrapae as such. Recorded, not marked.

2.3 · Lemmata collated and found to AGREE with the Clementine — named individually

Chunk 0000: Dixitque Dominus: Judas ascendet (1:2) · Dixitque (1:12) · Nec potuit (1:19) · Jebusaeum (1:21) · Manasses (1:27) · Arctavitque (1:34) · Ascenditque angelus (2:1) · Quamobrem nolui delere (2:3, Vulg. Quam ob rem, word-division only) · Dimisit ergo Josue (2:6) · Mortuus est (2:8) · Iratusque Dominus (2:14).

Chunk 0001: In medio (3:5) · Et tradidit eos Dominus (3:8) · Salvatorem (3:15) · Qui utraque manu pro dextera utebatur (3:15, verbatim) · Ingressus est Aod (3:20) · Pinguissimo (3:22) · Erat autem Eglon crassus nimis (3:17, verbatim) · Tulerunt (3:25) · Post hunc (3:31, twice) · Qui percussit (3:31) · Euge, serve bone (Mt 25:21) · Mihi autem absit gloriari (Gal 6:14, verbatim) · Addideruntque filii (4:1) · In manus Jabin regis Chanaan (4:2) · Erat (4:4) · Et sedebat sub palma (4:5) · Si venis (4:8) · Non ascendam, nisi tu ascenderis mecum (4:8 LXX-side, Vulg. si nolueris venire mecum, non pergam — Migne's is the positive form of the same refusal; recorded, not marked, because the two are the same speech act) · Egressa igitur (4:18) · Quae aperuit (4:19) · Egressaque (4:22) · Debbora (name).

Chunk 0002: Cecineruntque Debbora et Barac (5:1) · Audite (5:3) · Montes fluxerunt a facie Domini (5:5, verbatim) · In (5:6) · Cor meum diligit (5:9) · Qui propria voluntate (5:9) · Qui ascenditis super nitentes asinos (5:10, verbatim) · Collisi sunt (5:11) · Ibi narrentur (5:11) · Surge, surge (5:12) · Diviso (5:15) · Quare habitas inter duos terminos? (5:16, verbatim, question mark included) · Traxit cadavera eorum (5:21) · Cumque sevisset (6:3) · Da mihi signum (6:17) · Portans (6:18) · Ingressus (6:19) · Igitur (6:33) · In valle Jezrael (6:33) · Si salvum facis (6:36) · In area (6:37) · Si ros in solo vellere (6:37) · Oro (6:39) · Qui formidolosus (7:3) · Adhuc populus (7:4) · Fuit itaque numerus eorum (7:6) · Percussit illud atque subvertit (7:13) · Divisitque (7:16) · Et hydrias confregissent (7:20).

Chunk 0003: Fecitque ex eo Gedeon ephod (8:27, verbatim) · Postquam (8:33) · Remansitque Joatham (9:5) · Dixeruntque (9:10 and 9:14) · Ego sum vitis vera (Jo 15:1, verbatim) · Misitque (9:23) · Surge (9:32) · Post Abimelech (10:1).

Chunk 0004: Habitavit in terra Tob (11:3) · In illis diebus (11:4, verbatim) · Dixeruntque principes (11:8) · Et pugnes (11:8) · Ego ero vester (11:9) · Nonne ea quae possidet (11:24) · Et circumiens Galaad (11:29, Vulg. circuiens, a spelling variant of the period, Pattern 9) · Si tradideris filios Ammon (11:30) · Et plangam (11:37) · Et fecit ei (11:39) · Exinde mos increvit (11:39) · Judicavit itaque Jephte (12:7) · Sicut juravit Domino, votum vovit Deo Jacob (Ps 131:2, verbatim) · Sed concipies (13:3) · Nec immundum (13:4) · Tulit itaque (13:19) · Quod (13:20) · Descendit itaque Samson (14:5) · Percussitque eos (15:8) · Jurate (15:12) · In maxilla (15:16).

Chunk 0005: Abiit quoque (16:1) · Vocavitque tonsorem (16:19) · Multoque (16:30) · Et fecit ephod et theraphim (17:5, verbatim; the gloss's teraphim is the same word unaspirated, Pattern 9) · Venerunt Lais (18:7) · Eduxit ad eos concubinam suam (19:25, verbatim) · Consuluerunt (20:18) · Juraverunt quoque filii Israel in Maspha (21:1, verbatim).

§3 · THE THREE HAPAXES — carried, marked, and offered for a plate read

All three are corpus-wide hapaxes in src/latin/ and none is a Latin form. The corpus frequency check is the evidence, and it is given so a later reader can weigh it.

3.1 · 0522A · Amorrhaeos — a real word, and the faithful English says the opposite. [cj:]

Migne: Et contribulavit Amorrhaeos filios Dan in monte, quoniam non permisit eos descendere. Rendered as printed: "And he crushed the Amorrhites, the sons of Dan, in the mountain" — which makes Dan's sons Amorites and leaves contribulavit with two accusatives in apposition. The Vulgate lemma standing three words above it (Arctavitque) has Amorrhaeus, nominative, and the Amorite is the one doing the confining throughout Jdc 1:34 and Jos 19:47.

Nothing is broken type here, so [sic:] cannot reach it and Pattern 7 must render it; the English then asserts what Augustine did not. That is precisely Pattern 18's bar, so: [cj: Amorrhaeos; read Amorrhaeus, "the Amorite crushed the sons of Dan"]. ⛔ The rival was to leave it unmarked with a [var:] alone. It lost because the [var:] records a divergence from the Vulgate and says nothing about the sentence being unconstruable as printed; a reader would take "the Amorrhite sons of Dan" as the gloss's own claim. Both markers therefore stand, and they answer different questions.

3.2 · 0525B · periculam — broken type inside a scripture quotation. [sic: periculam]

qui sponte obtulistis animas vestras ad periculam, for periculum (Jdc 5:2, and periculum stands 81 times elsewhere in src/latin/; periculam once, here). Rendered and marked: "your souls to peril [sic: periculam]". The [sic:] sits inside the italic span with its content unitalicized, per the corpus's handling of a carried run inside an italic lemma (nested italics are not expressible; cf. [sic: interra]).

3.3 · 0530A · In icat — split type. [sic: In icat]

In icat enim se impeditum, ne illud quod cogitaverat impleret, for Indicat: a dropped d shown as a space. This is 8960's exact shape (subju averit, mise unt, solli itudo, qui ibet, corre ti) and takes the same treatment. Rendered and marked: "For he shows [sic: In icat] himself to have been hindered" — the clause keeps its verb, its subject, its object and its negation. ⚠ Whether the defect is Migne's plate or Corpus Corporum's transcription cannot be told from the Latin; 8960's whole class was handled as [sic:] and this follows it, but if the plate is ever read at p. 270 this is the first word to check.

3.4 · 0522B · seniorunt — the lemma word itself is not a word. [sic: seniorunt]

Et seniorunt., for seniorum (Jdc 2:7 et seniorum, qui longo post eum vixerunt tempore): u/n confusion, the commonest of all. seniorum stands six times elsewhere in src/latin/; seniorunt once. The gloss's own closing words settle the sense beyond doubt — dies vero seniorum sanctorum longaevi sunt. Rendered and marked: And of the elders [sic: seniorunt]. ⚠ Same plate/transcription ambiguity as §3.3, and the same request: p. 266, col. 522B.

§4 · MIGNE'S OWN, RENDERED LITERALLY, NOT TIDIED

4.1 · 0530A · a specula Galaad transeundi et transiens — a genitive gerund governing nothing

Rendered "from the watch-tower of Galaad of passing over and passing over to the children of Ammon". transeundi has nothing to govern it and no ad to hang on; the participle transiens beside it is what the Vulgate has. No conjecture is offered: this is Ruth's effoetandi class (8968 §2, 0533D), the fault is legible, it is Migne's, and Pattern 7 governs. A reader who wants the difficulty has the Latin opposite.

### 4.2 · 0530B · ut omnino evitetur, sed tantum ut differatur, laborat — a truncation tail ### that has evidently lost a non, AND NO non IS SUPPLIED

The sense the sed tantum demands is "he labours not that it should be altogether avoided, but only that it should be deferred." Migne prints no negative. The usque ad formula removes words; it does not license us to put one back. Rendered as printed: "that it may be altogether avoided, but only that it may be deferred, he labours." ⚠ This is the single site in the book where a supplied negative would have read better than the true text, and it is exactly the class the brief names (omnino + a correlative + a plausible verb). Logged, not repaired.

4.3 · 0532B, 0532C · CAPUT XVIII and CAPUT XX print without a final period

The other nineteen chapter heads in the work carry one. This is 8970's CAPUT XIV class (a digitization drop) or Migne's own inconsistency, and the two cannot be told apart without the plate, which has not been read at p. 271. Reproduced as printed — ## CHAPTER XVIII, ## CHAPTER XX — following 8960, whose shipped English carries ## CHAPTER VII for the same reason. Pattern 9's label-normalization was considered and rejected: it licenses normalizing a Latin typo in the label word, which has no English exponent; a full stop has an exact English exponent, and supplying one would be Pattern 8's forbidden supply. If the plate is read, this is a data/tei-patches/8959.json entry and the English follows the Latin.

4.4 · 0521A · [n: (ATHANAS., in (Synopsi. )] — an unbalanced parenthesis in the note

Migne's note as our TEI carries it opens two parentheses and closes one, with a stray space. Notes pass through verbatim, contents untranslated, so it rides exactly as printed. Recorded so a later reader knows it was seen and let alone.

4.5 · 0524A · verudata [ MSS., veruclata] — Migne's own variant bracket

sicut ea quae verudata [ MSS., veruclata] dicuntur. The bracketed variant is Migne's apparatus, not ours, and passes through untranslated per the locked register. verudata is kept in italic Latin as a technical term with no clean English equivalent (Augustine is describing a bolt-fastening). It takes no [sic:]: Migne has annotated it himself, and stacking our marker on his is forbidden (runbook false positive 5).

4.6 · 0526D · Quem pater dolis deputavit — a conjecture WEIGHED AND DECLINED

Isidore on Jdc 6:25. The sense wants idolis: the bull his father had dedicated to the idols. Rendered as printed: "Him whom the father had appointed for deceits". ⛔ The [cj:] was declined, and the reason is the bar, not timidity: dolis is a real word attested six times elsewhere in src/latin/, it construes as a dative, and the resulting English is odd but not false — idolatry is deceit on this author's own account two sentences later. Pattern 18 fires where the English asserts something the author did not; here it asserts something he would grant. The conjecture idolis loses in the text and is recorded here instead. (Contrast §3.1, where the printed word makes the English assert a falsehood, and the marker stands.)

4.7 · 0527D · Gregory puts the trumpet in the LEFT hand — against Jdc 7:20

in sinistra tubam, in dextera tenet lagenam, where Jdc 7:20 has sinistris manibus lampades, et dextris sonantes tubas. Rendered as printed. This is inside an usque ad tail of Gregory's Moralia and is his, not a lemma; no marker, because [var:] marks Migne's scripture citation diverging, and this is his source's exposition, not a quotation.

4.8 · 0531C · Laisan for Lais at 0532C, and other proper-name spellings

Judges carries names in bulk and spelling variation is not a defect (Pattern 9): Jesus Nave, Josue, Aod, Semgar/Samgar (both, in the same book), Debbora, Eglon, Chamos, Sichimorum, theraphim/teraphim, Laisan/Lais, Adamantius as ADAM. and ADAMAN.. All are kept as Migne prints them, each in its own place, and none is normalized to another. ⚠ Sichimorum is a corpus hapax and was checked: it is Augustine's own form for the men of Sichem (the Vulgate has habitatores Sichem), it is a well-formed Latin genitive plural, and it is already covered by the [var:] at 0529A. Not a candidate for a plate read.

4.9 · Migne's citation misprints — left standing

• 0524C [n: (Ibid. X.)] on a quotation of Ps 18:10–11, following [n: (Psal. CXVIII.)]. The Ibid. therefore resolves to Psalm 118, which is not the psalm quoted.
• 0530B [n: (Luc. VII)] on wording that is Lc 8:48's.
• 0532C [n: (AUG. de Civ., cap. 34.)] with no book number.

None is corrected in the running text. Citation repair belongs to data/citation-corrections.json at index time, and ⛔ an Ibid. must never be corrected upstream of its antecedent (runbook false positive 2). The 0524C case is exactly that shape.

4.10 · 0531A · a bare lemma with no gloss closes the work

CAPUT XXI carries Juraverunt quoque filii Israel in Maspha, etc. and nothing else — no VERS. address, no comment. Reproduced as printed. Migne's abridgment simply stops there.

§5 · THE SUPPLIED-SUBJECT REGISTER — every "he / she / it / they" the Latin does not print

Ruth's worst site was this class and it parsed perfectly either way; each supply below names its evidence.

§6 · PUNCTUATION AND QUESTION MARKS — band by band, not by total

The six chunks' column anchors partition both twins identically (9 / 9 / 12 / 5 / 10 / 2, verified 1:1). Colons, semicolons, question marks, exclamation marks and guillemets were compared band by band, and every band now matches. ⚠ The first pass did not: seven bands were off, all in the same direction — six supplied colons and semicolons of mine, in six different places, with nothing destroyed anywhere. Compared by total they would have looked close to parity. They are listed so the method is checkable: 0000 band 6 (; and :), 0001 band 3 (;), 0002 bands 1, 3 and 8 (:), 0003 band 2 (;). All six were reverted to Migne's comma.

Question marks: 12 in the Latin, 12 in the English, 4 / 4 / 4 / 0 / 4 / 0 by chunk, each in its own band.

• 0524D usque ad: id est, facit mecum prospera per angelum? — Migne's ? on a truncation tail. Kept, rendered direct.
• 0526C cur non possit angelus propheta dici? and si noster es, an adversariorum? — both indirect in Latin, both taken direct behind Migne's own colon (Pattern 8b). Neither mark supplied nor deleted.
• 0530A Ex utroque ergo prophetia adaptata est? — reads as a statement and keeps its mark, because it is his (Pattern 8).
• 0531B Quid est quod ad visionem angeli vir est timidus… sed tamen spes praesumit. — an interrogative opening that Migne closes with a full stop. ⚠ The parallel sentences at 0523D, 0531C, 0532A open identically; two of those also end in a full stop and one does not. No mark was supplied here. This is the shape most likely to attract a "correction" and it is Migne's.
• ! — one, at 0524C (…ori meo!), inside the Ps 118 quotation, in the same band in both.

§7 · NUMBERS AND TALLIES — checked against their own neighbours

Judges is dense with counts and this book keeps them consistent everywhere but once.

Verified consistent: thirteen judges and 411 years (0521A, and 13 judges is the count the chapter heads yield); eight years then eighteen years of servitude (0523B–0523C, matching Jdc 3:8 and 3:14, and the gloss supra octo annis: hic decem et octo dicit checks itself); six hundred men slain by Semgar (0524B ‖ Jdc 3:31); three hundred who lapped (0527D ‖ Jdc 7:6); the bull of seven years and the sevenfold Spirit (0526D); four days of lamentation, four parts of the world (0530D ‖ Jdc 11:40); forty years of Philistine servitude (0531B ‖ Jdc 13:1); six years of Jephthah's judging (0531A ‖ Jdc 12:7); six times seven = forty-two (0531A, arithmetic correct).

7.1 · 0531A · propter septuaginta dies — a number that contradicts its own sentence. [cj:]

Sicut illi duo menses propter septuaginta dies senarium numerum sex aetatum significant… Two months are sixty days, and only sixty yields the senary number; seventy yields the septenary. The sentence names senarium and sex aetatum in its own clause and closes sexies enim septem quadraginta duo sunt — the arithmetic is otherwise exact throughout. Rendered as printed and glossed additively: [cj: septuaginta; read sexaginta, "sixty"]. ⛔ The rival reading was to leave it and log it, on the ground that a number is a number and Pattern 7 renders it. It lost because the English then states a false arithmetical claim in the author's voice, with no signal to the reader that anything is wrong — which is exactly the gap Pattern 18 was written to close. The printed septuaginta keeps its English ("seventy") and nothing is replaced.

§8 · THE FOURTH TEST — read back for what the English SAYS

Run last, with counts, bands, parity and attachment already clean. Every paragraph was read back against the Latin asking only does it assert what he asserts. What it changed:

• 0523A non enim Deus homines exterminare poterat, et bestias non poterat. First draft: "for God could not exterminate men, and could not exterminate beasts" — which asserts that God can do neither, the opposite of Augustine's a fortiori. The first non scopes over the whole coordination, so the English now reads "for it was not that God could exterminate men and could not exterminate beasts." Both negatives stay on their own poterat; nothing is added or removed; the scope is carried by the frame. Proved from structure (the et … non correlative), not from what reads well.
• 0529D Quae enim magis terra optima quam terrenum corpus… rendered "For what land is more the best than the earthly body…", keeping optima audible against optimum / Tob two lines above, which is the whole etymological move. A smoother "more excellent" was rejected for breaking that chain.
• 0526B the and in "Who sent to them a prophet, and spoke" is English glue for Migne's asyndeton, not the Vulgate's et; the [var:] beside it names what is actually missing (virum), which is the word the gloss argues about.

Read clean, and named so a blind reader knows where I looked: the whole of chunk 0005 (the six-chapter abridged tail — every usque ad resumption parsed as strict prose, and §4.10 and §5's last three rows are what that produced); the Gedeon fleece paragraph at 0527A–B, whose habentem non in vellere … sed in aperto is the correct-handling model the brief names; the non solum … sed etiam at 0529B; the non habent … sed tantum at 0527D; the non … sed ipsum sacrificium futurum at 0531C; the double negative inulta non remansit at 0530B, kept as printed rather than resolved to a positive.

Candidates raised and rejected, with reasons: dolis (§4.6, declined); Sichimorum (§4.8, a hapax that is Augustine's own form, not a defect); increvit for increbruit (§2.2, cognate); ubi for ibi (§2.2, syntactic); the missing period on two CAPUT heads (§4.3, reproduced, not normalized); Gregory's left-handed trumpet (§4.7, his exposition, not a citation).

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

Blind reader, fenced from this file, all six chunks read: 2 sites, both repaired. It also verified the whole arithmetic chain independently (13 judges / 411 years · 8 vs 18 years · 300 men · sexies septem quadraginta duo = 42 · 40 years · 4 days for 4 parts of the world) and found no number contradicting its clause beyond the septuaginta already carrying a [cj:]; and it confirmed all three [sic:] render their word and all twelve chunk-0005 lemma expansions are convention-3 expansions of Migne's own continuations, none conformed to a Clementine wording.

1. REPAIRED · 0528D · qui veritati … credere noluit — the relative clause had been left sitting immediately after "Antichrist", where English reads it as a description of Antichrist. It cannot be: qui is nominative and Antichristo is dative. Its subject is populus, and the sentence is built on 2 Thess. 2:11–12 (ut credant mendacio … qui non crediderunt veritati). As it stood, our English made Antichrist the one who refused to believe Christ. Re-hosted with dashes so the clause reaches its subject.
2. REPAIRED, AND THE PLATE SETTLED THE OTHER HALF · 0526D · Quem pater dolis deputavit, hunc Gedeon occidit. - The referent was wrong and the gloss proves it: Ipse enim est vitulus septennis two clauses later. The object is the bull, not a man; our "Him whom … Gedeon slew" made Gideon kill a person. Now "That which the father had appointed … this Gedeon slew, and offered another, of seven years, to God." - dolis IS MIGNE'S — verified at the plate, p. 268 (cols 525/526) read at 400 dpi: five letters, one dot, standing over the fourth; mid-line, no hyphenation, and the preceding pater ends in an undamaged r that could not have absorbed an i; the inter-word gap is ordinary, so no sort fell out. No tei-patch. The reading is defective in Migne himself and now carries [cj: dolis; read idolis] — the conjecture recorded beside the printed word, never substituted for it. ⭐ The translator declined idolis for want of a witness and was right to. The blind reader proposed it from Jdc. 6:25 and was also right — but neither could settle it, and the plate could. This is the shape the register wants: raise it, refuse it, then go and look. ⚠ And note the contrast with 8979 the same day, where eleven suspected plate defects turned out to be ours. Here the suspicion ran the other way and the plate said Migne. The profile of a defect never decides it.

⚑ Plate: cols 521/522, 525/526 (and now the 400 dpi re-read of 268) carry no foot-of-page apparatus — no numbered Forte note, no asterisk-keyed EDIT. note. coverage: "spot".