De eruditione hominis interioris

Richard of St Victor (attrib.), De eruditione hominis interioris, PL 196, 1229D–1366A. 58 chunks, twelve agents, two rounds. This section heads the merged cruces.md; read it before changing any word of the English. What follows is what was fixed and why, so a later sweep does not churn it.

The seam this work turns on

mens = mind · animus = spirit · anima = soul · spiritus = spirit. 11536 flattened mens and animus both to "mind" because in that work they never collided. Here they collide inside single clauses — chunk 0030's own chapter head is Quod quidquid animum per cogitationem pulsat, mentis devotionem excitare debeat — so the flattening was not available and the four words are kept three ways (animus and spiritus share "spirit").

animus and spiritus collide in exactly two sentences in the whole work, and both were handled deliberately, by opposite reasoning that reaches the same English:

• 0005 @1244A — ex animi impetu beside ex spiritus sui impetu. Richard's eodem/eamdem assert these are the same impulse; rendering them differently would silently deny the identity his own correlatives assert.
• 0034 @1309A — animum humanum Deo conjungit, et cum eo quasi unum spiritum efficit (1 Cor. 6:17). Here the force is conclusion, not identity: what is joined becomes one spirit, and rendering animus "mind" would leave unum spiritum efficit nothing to attach to.

Ratified vocabulary

Heads: ## BOOK ONE/TWO/THREE. · ## CHAPTER ONE. for CAPUT PRIMUM, the plate's own numeral otherwise (## CHAPTER XV.). Chapter numbering restarts in each book — identical numerals recur across books and that is the plate.

⭐ simultas = pretense — how it was settled, since it overturns a classical dictionary sense

Not by ear. The metals allegory reads in auro charitas, in argento veritas, in aere simultas, in ferro crudelitas, in testa fictili fragilitas — a moral descent in which "enmity" sits oddly between truth and cruelty — and 1258D prints semetipsum in omnibus sine simultate accusat, accusing oneself without dissembling, which "without enmity" does not say. Two agents had rendered the word two ways (0018 "enmity", 0024 "pretense"); the second was right. ⚑ This also WITHDREW a Pattern 18 census candidate at 1269D. That site was filed as the range's strongest candidate on the reasoning that the plate glosses the slot with simulatio — but if simultas already carries that sense, the word is not a defect at all. One agent's crux was answered by another agent's vocabulary call, which is what the merge is for.

⚠ Two ratifications NARROWED at the merge, and it is worth saying why

Both were mine, and both were solving problems this work does not have:

1. intellectus = "intellect". Ruled to keep it apart from intelligentia = "understanding". The two words never once occur in the same sentence — zero collisions in 58 chunks — so the seam needed no defending, and the cost was real: 0038 read "In the intellects of his hands he led them" (Ps. 77:72) and "shall declare intellect" (Sir. 34:9). Reverted to "understanding" there. The agent had followed the ruling and reported the cost rather than quietly reverting, which is how it came to light.
2. scientia = "learning" vs cognitio = "knowledge". They collide in exactly one place in the work — CAPUT XVI's own head, Quibus gradibus mysteriorum cognitio proficit, et ad quantam scientiae perfectionem… — which is preserved distinct ("the knowledge of mysteries … a perfection of learning"). Elsewhere scientia keeps the natural "knowledge". A blanket sweep of 65 sites would have bought nothing.

The general rule both cases teach: a fixed-vocabulary pair earns its rigidity only where the two words actually meet. Check for collisions before ruling, not after.

Declared exception: retained Latin is italicized where the plate prints roman

Seven spans in 0052–0054 (insolitum / non solitum / insolentia; invidens / invidentia; rogare / derogare). These are etymological mentions — the Latin word is retained in English prose and glossed beside it — and the plate prints them roman because in a Latin text they needed no marking. In English a bare Latin word has no exponent at all, which is Pattern 12's own logic (normalize the language; mark the page). Kept, and declared here rather than done silently. ✅ RATIFIED BY WILSON, 2026-08-17. It is the only break from 1:1 italic structure in the work.

The oath exception

1324A cum sacramento asseverant means an oath, in parallel with cum juramento in the previous clause; rendered "under solemn oath" beside "with an oath". sacramentum = "sacrament" stands everywhere else, and 0029's drift to "this secret" for Dan. 2:47's sacramentum hoc was reconciled at the merge (0014 prints the same Dan. 2 phrase and reads "sacrament").

⚑ What this plate does — three habits, each found independently by more than one agent

1. Luke 17:21 prints regnum coelorum for regnum Dei at four sites (1240A, 1243D, 1261A, 1283B), one of them also intra nos for intra vos. Four agents, blind to each other.
2. Dan. 2:34 drops de monte at both its printings (1265B, 1276B), while the exposition turns the stone into the mountain. Two agents, blind to each other.
3. Migne's [n:] notes name the wrong book, repeatedly — (Isai. XL) over Rom. 11:34 twice; (Ephes. I) over Phil. 2:13 and over 1 Cor. 12; (Dan. I) for Dan. 2:41; (Job XXII) for Job 30:7; (I Cor. XII) for 2 Cor. 12:7. All reproduced verbatim. ⚠ INDEX CONSEQUENCE, checked before shipping: index-work.mjs keys the scripture index off these notes, so real citations will file under wrong books.

⭐⭐ The finding of the run: Richard's copy-text, recovered from his own argument

1347C/1353D print Dan. 7:4 as cor ejus where the Vulgate reads cor hominis — and at 1355D Richard expounds it as printed: Cujus ejus? Aut cor hominis, aut cor leaenae, a question the received reading makes impossible to ask. His copy-text read ejus. The beast is kept "it" throughout so the ambiguity survives into English. What makes this evidence rather than a guess is that the entire bear-lemma of 0054 diverges nowhere**: a range that marks everything proves nothing.

⭐ The plate is reliable when it QUOTES and unreliable when it RE-QUOTES

The Daniel 4 tree lemma at 1311D agrees with the Vulgate word for word, all nine clauses — while its re-quotations diverge at 1311B (solutiones), 1312D (arbor magna pulled forward from the next verse, with the gloss built on it), 1319B (vigil sanctus, the et dropped, one subject where two are needed) and 1319D (Praecidite for Succidite, where cap. XXII's scheme depends on the two verbs staying apart). Book I shows the same shape: significet → annuntiet (1257B), ostendisti mihi → nobis (1253D/1254A), Dominus coeli → Deus coeli (1281B/1284A), the lemma dropping deorum at 1305C while both re-quotations restore it. Give the polarity read the re-quotations first.

Non-findings, recorded so nobody re-opens them

• Dan. 4:22 super regnum hominum vs 4:29 in regno hominum looks like the plate contradicting itself on the word CAPUT XLII rests on. The received text makes that distinction itself (threat vs. fulfilment), as it does at fenum ut bos / fenum quasi bos. Migne is correct twice. A later reader who "fixes" either will be wrong.
• The doubled statua at 1265B is not dittography — the Vulgate prints the same words separated by a colon, and 0017 re-quotes the run in two pieces and glosses them consecutively.
• 1289C's Si aut in argento… will not construe: two mutually contradicting negations in one protasis, both carried, the English as self-contradictory as the Latin. Escalated, not repaired.
• Markers withdrawn by their own author after a corpus control: conspergatur and commutentur at 1335D (four PL works print the first, three against one the second — Migne is with the majority), cum bestiis ferisque (corpus split 2–2), Ps. 64:9 delectabis (12 works against 2), Jer. 17:9 cor hominis (11 against 2). Every literary signal said "divergence"; the corpus said otherwise.
• 1302C's qua: is the ligature quæ lost by the DIGITIZATION, not a plate defect (16 files corpus-wide). No [sic:] — a defect in our files is never attributed to Migne. Possible data/tei-patches/ candidate.
• 0016's Ps. 124:1 singular Qui confidit was flagged across ranges as possibly unmarked. It is marked — the flagging agent could not see 0016's markers. Likewise 0054's Ps. 93:21 imperfects.

⚠ Weak markers, named as weak

• [sic: ita que] @1256B sits on a TEI line break, so it is equally consistent with Corpus Corporum failing to rejoin a hyphenation. "ita que " returns 0 files corpus-wide and the TEI carries no line-final hyphens, but withdraw if a plate read shows itaque whole.
• [sic: cruditur] @1346B — 4 PL files, not a hapax; a recurring c-for-e accident. It stands only because the form answers to no lexical entry. The work's own title word.
• [sic: tamem] (38 files) and [sic: torpore] (syntactic slot only, no rarity evidence) are the two weakest in the work and say so in their entries. ### ✅ THE TWO PLATE READS ARE DONE (2026-08-17) — both settled, opposite ways

Witness: the archive.org Google scan of PL 196, patrologiaecurs104unkngoog (metadata volume 196), PDF page 637 = columns 1233/1234, rasterized at 450 dpi and read at the word.

• ✅ [sic: sommia] @1234A CONFIRMED — it is Migne's own type. The scan prints the lemma Præcepit autem rex, ut convocarentur arioli, et magi, et malefici, et Chaldæi, ut indicarent regi sommia sua in italic, plainly double-m. Two independent witnesses now agree. The marker stands.
• ⛔ The angüstiae diaeresis @1233A is OURS, not Migne's — and it is now patched. The plate prints sessio in agnitione angustiæ with the ordinary æ ligature and no diacritic. ⭐ The mechanism is visible in the same image: two isolated specks of offset ink sit in the blank band above the first text line, roughly over ngu — too high to be a diacritic and spaced about three letters apart, where a diaeresis would sit tight above the u at x-height (the dotted i of agnitione, on the same line, shows what tight looks like). A transcriber taking those specks for a diaeresis and attaching them to the nearest vowel produces exactly the reading the TEI carries. The same scan shows a comparable ink blot under convocarentur in col. 1234. With angü a hapax across all 5,276 files and 87M words of the PL TEI, and the same etymology printed twice more in the same two columns as plain angustiae, this is a digitization defect. data/tei-patches/11535.json written and applied through chunk-work.mjs — one file, one character, 540 column marks / 336 notes / 54,433 words conserved. The [sic:] that was correctly declined pending this read is never needed, and the English is unaffected ("anguish" either way). ⚑ The general lesson, and it is cheap to repeat: where our text carries a diacritic or a form that is a corpus-wide hapax, suspect our files first. The plate read cost one PDF and two crops.

⬜ Open for Wilson — nothing downstream is blocked on these

1. ✅ RESOLVED — the [cj:] gloss at 1231B STANDS (Wilson, 2026-08-17). A stranded correlative whose answering member is printed and whose own member is not is inside the bar, and translation-style.md Pattern 18 now says so. The other four [cj:] were always squarely inside it (dropped non at 1239A and 1289D, nisi/nec at 1302B, sit/sic in a Quemadmodum … sic correlative at 1294B). The ~35 declines below are unaffected: the class admitted is narrow and does not reach mood, voice, number, person or prepositions.
2. The retained-Latin italics (declared exception above).
3. ~35 declined Pattern 18 census candidates are named across the twelve cruces files, for WITHDRAWN-NET.md. The four strongest, in the order they should be revisited if the bar ever widens: 1295A minus/imis (contradicts its own chapter head two lines above — the founding 11081 munde shape exactly) · 1246C amittere/admittere (the remedy reads as the opposite of the ill it cures) · 1261A praefectis/perfectis · 1360C ab/ad subdiaconatum (inverts the sentence's own gradient; 17 PL files print ad, one prints ab).

⚠ The ending — belongs in data/work-about.json before deploy

The work is not truncated mid-sentence; it is truncated mid-exposition. The last printed words at 1366A close a complete sentence on Migne's own question mark, but the final paragraph announces the restoration half of the argument (Audiamus quomodo … ad rectitudinis integritatem reformatur), quotes Rom. 5:20 as its programme, and then delivers only two of the Quid est quod…? questions on Dan. 7:9–12. It never reaches the Son of man and never reaches the kingdom that shall not be destroyed — the figure the whole treatise is built toward from Dan. 2:44 in Book I. It sits under CAPUT XIX, a chapter head about despair, with no head of its own, in a chunk of 552 Latin words against a work median near 1,000.

§Z. THE BLIND POLARITY READ (2026-08-17) — 12 OUR-SIDE DEFECTS, AND verify-english PASSED EVERY ONE

Six readers, 58 chunks, none permitted to open cruces.md, cruces-*.md, this file, or the fleet brief. Every finding below was checked against the cruces only AFTER it was formed, and not one of the twelve was logged — the columns appear in the cruces, but for other reasons (a [sic:] one clause away, a Migne-side census candidate, a [var:] on a different word in the same verse). A flagged passage is not a checked passage, for the second work running.

The twelve, all fixed

⚑ What this read teaches, beyond the sites

1. ⭐⭐ The 7a″ protocol worked so well that the failure MIGRATED. Every launch message named its range's lemma words verbatim, and the lemmata held: the Daniel 4 re-quotation chains, the nine negations of 1 Cor. 13, the seven of Ps. 113, Dan. 2:20 printed three times at three lengths — all came through clean, and one reader verified them site by site. The damage is in the exposition's own rhetorical pairs (morning/evening, illud/istud, aula Babylonis, the distributed genitives), where nobody was pointed. ⬜ Next lemma-and-gloss work: name the AUTHOR'S OWN antithesis-words in the prompt too, not only the scripture.
2. Two of the three worst sites were seen and declined by the translating agent. 1356A and 1357A were both in the cruces as reasoned declines. A blind reader with no stake in the earlier decision overturned both on evidence in the same chunk. The cruces cannot audit itself.
3. A reader disclosed self-contamination and voided its own strongest finding. A corpus-wide grep for a word returned matching lines from cruces.md and this file; the reader had seen them, said so, and moved the affected site to "rejected". ⚠ The blind-read prompt must bar grep-ping the work's English directory as well as opening the files — the answer key can arrive through a search result.
4. Yield tracked genre exactly as the runbook predicts: 0 sites in 0000–0009 and 0 in 0030–0039 (both verified positive, with the clean chunks named), against 5 in 0020–0029 and 3 in 0040–0049. The two barren ranges are the ones densest in scripture lemma; the fertile ones are densest in Richard's own argumentative prose.

Migne-side defects the read surfaced that are NOT in the apparatus

All rendered literally, which is correct under Pattern 7 — but a reader meets the contradiction with no cue. Recorded here as crux material, not as text changes: 1309B relevationibus/relevari for revelationibus/revelari (control: 30+ revelatio- in this work against exactly 2 relevatio-, both in this one sentence) · 1267D nunc for non in a Jer. 3:2 / Isa. 49:18 conflation, the negation-destroying letter class, missed because the quotation is not italicised on the plate · 1268D servire for saevire · 1270B a gradation whose first step runs backwards · 1333B Vespertinum defining evening with every mark of morning (same defect zone as 1333A) · 1287A honores for horrores · 1284D negligentius for negligentes · 1283D ea for aes, leaving the metal-list one short · 1282B cui for qui · 1289D uritur for irritatur in a clause-by-clause re-quote of 1 Cor. 13 · 1276A corpore for torpore · 1279C opibus for opus · 1363A implicare for explicare · 1354C amissis aliis for alis · 1356D et for ad.

Corroboration for a merge ruling, found blind

A reader independently flagged simultas → "pretense" as an unmarked normalisation and produced new evidence the merge did not have: chunk 0020 @1274A prints in aere simulatio, rendered "simulation" — Migne himself puts a simulation-word in that slot of the metals allegory.* That corroborates the ruling rather than upsetting it. The ratification stands, now on three legs (the allegory's own position, 1258D's sine simultate*, and 1274A).

⚠ THE "WRONG BOOK" CITATIONS — HALF THE CLAIM WAS WRONG, AND THE HALF THAT SURVIVED IS FIXED

Three separate agents reported that Migne's [n:] notes name the wrong book "repeatedly and reproducibly", and I carried that into the merge as a third plate habit with an index consequence. Checked one at a time before any correction was written, two of the six collapsed.

⛔ NOT corrected — (Isai. XL) at 1296D and 1305A over Quis cognovit sensum Domini, aut quis consiliarius ejus fuit. The agents read this as Rom. 11:34 filed under Isaiah. But Richard's own prose at 1296D introduces the quotation as quod Isaias dixit — the attribution is the author's, not a slip of the plate, and Rom. 11:34 is itself Paul quoting Isa. 40:13. Correcting it would have overwritten a medieval writer's own source-attribution with a modern one. ⛔ NOT corrected — (Ephes. I) at 1362C over Deus est qui operatur omnia pro bona voluntate. Reported as Phil. 2:13. But Eph. 1:11 reads qui operatur omnia secundum consilium voluntatis suae, which matches operatur omnia better than Phil. 2:13's operatur in vobis et velle et perficere does. It is a conflation leaning to the cited chapter. Left as printed.

✅ Corrected in data/citation-corrections.json (4), each verified to have no following Ibid. depending on it — the standing hazard, since a correction applies upstream of anaphora resolution: (Cor. XV) @1236a → 1Cor.15.10 (ordinal-less, so it did not parse at all; the disambiguation is determinate — 2 Corinthians has only 13 chapters) · (Dan. I) @1286b → Dan.2.41 · (Job XXII) @1287c → Job.30.7 (Richard's frame says secundum sententiam Job, so the book is right and only the chapter is wrong) · (I Cor. XII) @1295a → 2Cor.12.7. refDisplay keeps what Migne printed and refKeyPrinted keeps the key he printed, so the record is not overwritten.

Index after: 340 scripture, 117 heads, ZERO unparsed.

⚑ The lesson is the one this whole session keeps teaching. A claim that repeats across three independent agents still has to be checked site by site: agreement between agents is not evidence, because they share a genre expectation. Two of six "wrong book" notes were the author being himself.

Richard of St Victor, De eruditione hominis interioris, Book I, chapters I–VI. Opening range: Daniel 2:1–11 quoted in running order, with fragments re-quoted and glossed.

Conventions set here for the whole work (peer agents follow these)

Heads. ## LIBER PRIMUS. → ## BOOK ONE. · ## CAPUT PRIMUM. → ## CHAPTER ONE. · every numbered head keeps the plate's own numeral (## CAPUT II. → ## CHAPTER II., never renumbered, never normalized to a word). The italic summary is translated and stays italic; the plate's terminal period is kept where it prints one. The three blank lines the plate leaves between LIBER PRIMUS. and CAPUT PRIMUM. in 0000 are mirrored 1:1 (Pattern 8a).

The name. Nabuchodonosor, not "Nebuchadnezzar", throughout. Chapter I turns on two etymologies of the name — prophetans istiusmodi signum (1231D) and sessio in agnitione angustiae (1232A, 1233A) — and an English form leaves both glosses explaining nothing, which is exactly the case the brief's etymology rule and Pattern-7 §2 (stylus/epistylium) cover. The corpus is split (4 shipped works print Nabuchodonosor, 7 print Nebuchadnezzar), so this is a per-work call and it is submitted for ratification. Other names take their received English (Daniel, Chaldeans, Jerome, Gregory, David, Uriah, Babylon, Syria).

Vocabulary added to the fixed table (proposed for the whole work): angustia = anguish — the hook the etymology hangs on (sessio in agnitione angustiae · quam sine angustia nosse non poterat · cum dolore et animi angustia) · conjectura = conjecture, kept distinct from interpretatio = interpretation (Migne prints both in the Daniel lemma and Richard glosses only the second) · publicare = to make public, never "confiscate" (Douay), because Richard's gloss is built on it: Quid autem est sapientium domos publicare, et ad communem omnium usum redigere? · conatus = endeavour · instantia = urgency · pudentia = modesty (see crux 28).

Markers fired (7 sites)

[var: …] — Pattern 14, scripture divergence (3)

1. 1233A · Rom. 13:1 · [var:] FIRED Printed: Quae autem a Deo sunt, teste Apostolo, ordinata sunt [n: (Rom. XIII)]. Vulgate: quae autem sunt, a Deo ordinatae sunt. Migne draws the agent a Deo into the relative clause and prints the neuter ordinata for the feminine ordinatae. Material: the Vulgate says the existing powers are ordained by God; Migne's line says the things that are from God are ordered — and that is precisely the load Richard puts on it, the unjust king's voluntas inordinata against his potestas ordinata. The preceding Non est enim potestas nisi a Deo is printed roman, unmarked, and agrees with the Vulgate. Rendered as printed; the received reading is in the marker.

2. 1234A · Dan. 2:2 · [var:] FIRED Printed: Qui cum venissent, statuerunt coram rege. Vulgate: steterunt. statuerunt is transitive and leaves the clause with no object, so the faithful English ("set up before the king") states something the received text does not. Control run: statuerunt coram rege occurs once in the whole PL TEI (this file); steterunt coram rege also once. But statuerunt itself is a common word — 1,005 occurrences — so this is a real word wrongly printed, not broken type: Pattern 12 has nothing to wrap. Pattern 18 is excluded by its own rule that a divergence inside a quoted scripture belongs to Pattern 14.

3. 1240A · Luke 17:21 · [var:] FIRED Printed: Regnum caelorum, inquit, intra vos est [n: (Luc. XVII)]. Vulgate: regnum Dei intra vos est. Migne's own note cites Luc. XVII, so the substitution is on the plate, not in the reference. Material: Richard is arguing that the interior kingdom is a kingdom, and caelorum pulls the verse into the orbit of the Matt. 5:3 beatitude he quotes four lines later — which prints the other spelling, coelorum, in the same chunk (crux 39).

[sic: …] — Pattern 12, carried type (3)

4. 1229D · eficiat for efficiat · [sic: eficiat] Control run: word-boundary grep of sources/pl/tei/ returns 1 occurrence in 87M words — this file, this site — against 674 files carrying efficiat. Not a word; carried untranslated. ⚠ 7a⁗ q1 applies: it is the only verb of the si clause, so the sense stands in plain words beside the marker — "should effect [sic: eficiat]" — on the which slays [sic: eccidit] model. Conjecture: efficiat, near-certain.

5. 1234A · sommia for somnia · [sic: sommia] (inside Dan. 2:2) Control run: sommia returns 4 occurrences across 4 different works in the whole PL TEI against 863 of somnia — 0.5%, which reads as sporadic broken type (n → m), not as an orthographic convention, so Pattern 9 does not cover it. ⚠ Marked without asterisks: it sits inside an already-italic scripture lemma and nested italics are not expressible (Pattern 10's closing clause). ⚠ 7a⁗ q1: the marker takes the clause's object, so "his dreams" stands in plain words beside it. Honest weakening: 4 occurrences is thinner evidence than a hapax. If a later census reads the class as Migne's habitual mis-set rather than damage, this marker should come off.

6. 1238C · conanimis for conaminis · [sic: conanimis] Control run: hapax — 1 occurrence in 87M words, this file, this site. Carried on the bonorem precedent: the noun-head is carried untranslated and the conjecture (conaminis, genitive of conamen, "effort") stays here rather than in the text. 7a⁗ q1–3 all pass — the clause keeps its verb (docere), carries no negation, and no conjunction is swallowed.

[cj: …] — Pattern 18, real word, faithful English misleads (2)

7. 1231B · quam stranded without its comparative · [cj: quam; read facilius quam, "more easily than"] Printed: contemplativae vitae gratia quam activae elabitur, sed multo difficilius reparatur. As printed the sentence has a quam with nothing to compare — it does not construe at all. Richard supplies the missing word himself, twice, within eight lines: somnii visionem quam tam facile amisit (1231B) and a contemplativa facilius decidimus (1231C). ⚠ Bar note, flagged for ratification. Wilson's 2026-08-15 bar admits a wrongly-printed negation or correlative, and a stranded comparative correlative is the second class; the 7561 @0648 precedent (solum for non solum) establishes that a dropped word is glossable, because the gloss opens on the word Migne did print. I have not widened the bar beyond that. If the ruling is read narrower than I have read it, this is the one marker in my range to withdraw.

8. 1239A · volebant for nolebant · [cj: volebant; read nolebant, "were unwilling"] Printed: Ignorantiam suam dissimulabant, et aperte fateri volebant, et in hoc tempus redimebant. The printed reading contradicts its own sentence twice: they dissembled their ignorance, and they were buying up the time — which is what one does precisely in order not to confess. Richard's whole chapter is quam multis modis animus seipsum decipit. This is the 11547 @0260D defect (nolebant for volebant) with the polarity reversed, and the word that went wrong is the negation itself — squarely inside the ratified bar.

Conjectures DECLINED — named as census candidates

9. 1233A · angüstiae for angustiae — the etymology sessio in agnitione angustiae is printed twice in chunk 0000, and the second printing carries a diaeresis on the u. Control run: the string angü is a hapax across the entire PL TEI — one occurrence in 87M words, this file, and it is the only umlauted vowel anywhere in 11535. Declined a [sic:] on three grounds, and the call is genuinely marginal: (a) the word renders in place perfectly, so what a reader would meet is our English, not Migne's type — the Pattern 12 operational test; (b) English has no exponent for a diacritic, so a marker would put Latin on the page for a mark that means nothing in either language; (c) I cannot check the plate, and a stray diaeresis is at least as likely to be a Corpus Corporum OCR artifact as Migne's own type, which would make it a data/tei-patches/ question rather than a Pattern 12 one. Both printings render identically ("a sitting in the acknowledgement of anguish"). ⭐ Named for the census: a single plate read of col. 1233A settles it either way.

10. 1232D · quae, qualia sint, possunt — singular relative quae (antecedent rei) against plural qualia sint / possunt. Number anomaly with no English exponent; declined per the bar (not a negation or a correlative). Rendered as closely as the printed words allow, with a parenthetic dash marking the break.

11. 1233D · cognitionumque disciplinam where the parallel eleven lines above (affectionibus … cogitationibus) leads one to expect cogitationum. A real word in its slot; declined. Rendered "the discipline of his knowings".

12. 1236C · hanc nobis sollicitudinem primo ingenerant — indicative among the surrounding subjunctives (satagant … studeamus). Mood anomaly; declined per the bar.

13. 1239A · quomodo illud potuissent quo majus erat? — quo for quod, breaking the parallel with quod minus erat four words earlier. Case anomaly; declined. Rendered "that whereby it was the greater".

14. 1239B · Ad tertiam autem, lucem jam palam profitentur. The object should be their ignorance (cf. ignorantiam suam magis minuunt quam loquuntur in the preceding sentence); lucem is a real word in a real slot and yields fluent nonsense. The strongest of the declined sites and the closest to Pattern 18's bar — but it is neither a negation nor a correlative, so it stays here. Rendered "they now openly profess the light".

15. 1239D · secundum malum quod in seipso invenit, alios adjicit. The antithesis (alios omnino non praedicat … alios adjicit) wants judicat. Real word; declined. Rendered "adds others".

Passages that would not construe — rendered as printed, no marker

16. 1233C · the dangling tertium. Primum studium … ut possit dominari affectionibus; secundum ut possit imperare cogitationibus; tertium ergo animus spiritalia quaeque … incipit habere. The third member of the series has no predicate of its own and tertium is left hanging. Rendered as printed. ⚠ Both co-subjects preserved against the singular verb — animus ("the spirit") and anima ("the soul that is victor over its own vices") — per the brief's warning that 11550 lost four co-subjects to one.

17. 1236B · a sentence break the plate does not print. Quidquid peragas, illi, non tibi ascribas hoc utique regi illi … dicendum erat. No stop between ascribas and hoc. Pattern 8 forbids supplying one; the English runs the two clauses together and reads roughly as the Latin reads.

Pattern 8 — punctuation carried, not repaired

18. Questions printed with a period and left that way (no ? supplied): 1231A Quid enim est quod Nabuchodonosor … agnovit. and Quid, inquam, … reparatur. · 1232A quo, quaeso, in loco … non possit videri. · 1238C Quam multa divinorum judiciorum cognoscimus … penetramus. and O quam difficile humana impudentia … reprimitur. · 1238D Quam multa saepe simulamus, quam multa dissimulamus. · 1241B nihil in ea nisi rectum sonare videtur ubi dicitur.

19. 1236C — O infinita hominis stultitia, [1236C] et recte quidem dicis … sine gratia? Migne prints a comma after the exclamation and a ? at the end of the run. Both carried exactly; no ! supplied, though English wants one there.

20. 1236A — the chunk's one ! is Migne's, closing a statement (… interioris hominis vitam recte intelligimus gratiam divinam!). Carried.

21. 1236A — Gratia, Dei, inquit, apostolus (1 Cor. 15:10): Migne prints a comma between the noun and its genitive. Carried — "By the grace, of God," — per Pattern 8's never-delete clause. He also drops the Vulgate's autem; no [var:], no sense change.

Question-mark parity across all five chunks: 9/9, 5/5, 19/19, 6/6, 7/7. Exclamation parity 0/0, 0/0, 1/1, 0/0, 0/0. Column anchors identical in order and value; note counts 1/1, 1/1, 4/4, 1/1, 5/5.

⭐ The internal disagreements — the same verse printed twice, differently, on one plate

These are the evidence the brief asked for, and none of them earned a marker: each is a transposition or a spelling, and Pattern 14 fires only on material divergence. All four are named as census candidates, and where possible the English registers the difference so it survives translation.

22. 1237A vs 1237D · Dan. 2:6 · WORD ORDER, and it is the Vulgate that adjudicates.

• 1237A (italic): praemia, et dona, et honorem multum a me accipietis
• 1237D (italic): praemia et dona, et honorem multum accipietis a me
• 1238A (roman re-quote): praemia, inquit, et dona, et honorem multum accipietis a me
• 1238B (roman, broken off): et praemia, et dona; et honorem multum accipietis.

The Vulgate reads accipietis a me, so the second, third and fourth printings agree with the received text and the first does not. The English registers it: "you shall receive from me rewards…" at 1237A against "you shall receive rewards … from me" at 1237D. The second printing also drops the comma after praemia that the first has — same site, same class.

23. 1240A vs 1240B · Dan. 2:10 · SPELLING. The italic lemma prints verbum hujusmodi; the roman re-quote eleven lines later prints verbum hujuscemodi, which is the Vulgate's form. Registered in the English as "such a word" against "a word of this kind".

24. 1240A vs 1240B · caelorum beside coelorum. The Luke 17:21 lemma prints Regnum caelorum; the Matt. 5:3 lemma four lines later prints regnum coelorum. Pattern 9 orthography; both rendered "heaven". (The Luke lemma's separate and material defect is crux 3.)

25. 1232A vs 1233A · angustiae beside angüstiae — crux 9 above.

⭐ Lemma pulls RESISTED — where Migne agrees and my instinct wanted to improve him

26. 1236A — the most dangerous one in my range, and it is not a defect. … ut in aeternum vivamus, quod tunc quidem fiet, si nihil nisi bonum, et sine bona intentione agamus. Read at speed the second member says "and [if we act] without a good intention", which reverses Richard's entire chapter, and the reflex is to supply a negative or to gloss a dropped nec. It needs neither. nihil is shared across both members: nihil nisi bonum / nihil … sine bona intentione. Rendered "if we do nothing but what is good, and nothing without a good intention" — the second "nothing" is a grammatical ellipsis supply, plain text, no brackets. No [cj:], no repair, no marker.

27. Dan. 2:1 (1233B), 2:3 (1234C), 2:4 (1234D), 2:5 (1236D–1237A), 2:6b (1238B), 2:8 (1238C), 2:9 (1238C–D) — all agree with the Vulgate word for word. Nothing was "improved", nothing familiar was allowed to arrive ahead of the printed Latin. In particular Sermo recessit a me is rendered "The word has departed from me", not Douay's "The thing is gone out of my mind", because Richard re-quotes recesserit a me sermo and glosses it.

28. 1234D — the italic boundary in Dan. 2:4. Migne closes the italics after in sempiternum vive. and prints Dic somnium servis tuis, et interpretationem ejus indicabimus roman. English habit wanted one continuous quotation. The boundary is reproduced exactly. Same at 1240B, where Beati enim pauperes spiritu… splits the Matthew lemma around Richard's own enim: the English reads "for Blessed are the poor in spirit…", not one lifted quotation.

29. Immaterial Daniel divergences, all declined, none marked: arioli for harioli (2:2 — and many Vulgate printings carry the unaspirated form; Pattern 9) · Si igitur for Si ergo (2:9) · Responderunt ergo … et dixerunt for Respondentes ergo … dixerunt (2:10) · Rex somnium servis suis dicat for Rex somnium dicat servis suis, and ejus for illius (2:7) · quem tu, rex, quaeris for quem tu quaeris, rex (2:11).

30. Non-Daniel scripture, all resisted: Matt. 6:23 (1235D) prints Quia, si oculus tuus fuerit nequam for the Vulgate's Si autem oculus tuus — connective substitution only, declined · Matt. 5:3 (1240B) agrees exactly · Ps. 2:6 (1240B) and Ps. 81:6 (1240D) agree and are simply truncated (super Sion montem sanctum ejus and omnes dropped), which is quotation practice, not divergence · John 15:5 (1236A) agrees · 1 Cor. 10:11 (1231D), Phil. 3:20 (1241A) and Luke 18:27 (1240D) are printed roman, unitalicized and unnoted, so they are not set as proof-texts; their trivial divergences (dropped autem; apud Deum possibilia sunt transposed) take no marker.

31. 1238D — non nihil profecit carried as the double negative Migne prints ("profited not nothing"), not smoothed to "profited somewhat". Pattern 7a, mechanism 3.

Vocabulary tensions — followed the table against my own judgment

32. 1233D · humanam virtutem. The table fixes virtus = virtue, and here the sense is plainly "human strength/power": quam sit supra humanam virtutem divinis beneficiis respondere. Rendered "above human virtue" per the table. If the table is to admit an exception, this is the site. Note the same chunk's 1240C nihil de propria virtute praesumunt takes "virtue" naturally.

33. 1236D · pudentia / impudentia. pudentia is not classical but is formed regularly from pudens and is attested in medieval Latin — Pattern 9, no marker. The table fixes impudentia = shamelessness, so pudentia is rendered modesty: "A common modesty; a general shamelessness." The tension: Migne's antithesis is a single root negated (pudentia / impudentia), and modesty/shamelessness does not preserve it as a pair. "Shamefastness / shamelessness" would, at the cost of an obscure word. Following the table; flagged.

34. mens / animus / anima / spiritus held visibly distinct throughout (mind / spirit / soul / spirit), including where two collide in one clause — 1233C animus … et vitiorum suorum victrix anima, 1234D mens humana … animus, 1235A sensus ejus usque ad animum non ingreditur. Neither of the two flagged animus/spiritus collision sentences (0005, 0034) falls in my range; the only spiritus here is Dan. 2:1's conterritus est spiritus ejus and Matt. 5:3's pauperes spiritu, and neither has an animus in the same sentence.

35. Pattern 17 throughout. thou wherever the Latin is singular (ut vides 1231A · tibi videtur 1232B–C · quidquid disponas … peragas 1236B · dicis … gloriaris 1236C · te … tibi ostendimus 1234C, 1235B · sermonem tuum, rex 1239A) and you wherever it is plural (Sine me nihil potestis facere · dii estis · peribitis vos … domus vestrae · intra vos est · narraveritis … accipietis). ⚠ The Daniel dialogue switches inside single sentences — the Chaldeans address the king as tu while he addresses them as vos — and every switch is preserved: "Tell the dream to thy servants" (1234D) against "Unless you declare to me the dream … your houses" (1236D).

Ratios and em-dashes

Word ratios EN:LA — 0000 1.60 · 0001 1.62 · 0002 1.50 · 0003 1.49 · 0004 1.44. The two high chunks are driven by Richard's quanto … tanto / ejusmodi … istiusmodi correlatives, which English can only unpack ("by how much the lower, by so much is it always found worse than itself"), not by padding. Three em-dashes in five chunks, each carrying a break the plate's own syntax makes: 1232D's number mismatch, 1233C's dangling tertium, 1235A's Syriace parenthesis.

Richard of St Victor, De eruditione hominis interioris, PL 196. Agent range: chunks 0005, 0006, 0007, 0008, 0009. Entries give column · printed reading · what was done · why.

A. The animus / spiritus collision — 0005 @1244A (flagged in the launch brief)

Printed: …quod meminit se prius per spiritus sui impetum statuisse? Multa enim … quae prius ex animi impetu, et quasi praecipitanter definivit … et multa saepe ex eodem spiritus sui impetu homo contra conscientiam suam dictat, et tamen eamdem animi sui praecipitationem … temperat.

What was done: both words rendered spirit, per the fixed vocabulary table, with no disambiguating gloss and no marker.

Why — and this is the point: the collision here is benign, and the table is what makes Richard legible. The sentence turns on eodem and eamdem: Richard is explicitly asserting that the spiritus sui impetus and the animi impetus are the same impulse, and that the animi praecipitatio is the same headlong rush. Rendering the two words differently ("mind"/"spirit", or "soul"/"spirit") would silently deny the identity his own correlatives assert, and the reader would meet two faculties where Richard has one. So the one-English-word outcome the table produces is not a loss here; it is the reading. No tension to escalate.

(For the record, the four sites in order: 1244A per spiritus sui impetum → "through the impulse of its own spirit" · ex animi impetu → "out of the impulse of its spirit" · ex eodem spiritus sui impetu → "out of that same impulse of his spirit" · eamdem animi sui praecipitationem → "that same headlong haste of his spirit".)

⚠ Note for the merge: mens is kept visibly distinct as mind throughout this range (0005 studiosa mens, compuncta mens; 0006 devota mens; 0007 mens studiosa), and anima as soul (0008 @1248D in amaritudine animae meae; 0009 @1252B anima diu multumque oppressa). No flattening anywhere in the range.

B. Markers placed

### B1. [sic: doctrinam] — 0005 @1242C Printed: prius saecularium, postea spiritalium doctrinam curam deponit. Done: "first lays aside the care of secular, afterwards of spiritual teachings [sic: doctrinam]". Why: saecularium … spiritalium require a genitive plural; doctrinam is an accusative singular standing beside a second accusative, curam, and has no grammatical slot (Pattern 12, test 1). Corpus control run: doctrinam curam occurs once in 87M words of PL TEI — here; doctrinarum curam occurs in 11535, in this very chapter's own head six lines above (tam saecularium quam spiritalium doctrinarum curam deponit). The plate contradicts itself on the same column-block, which settles it. 7a⁗ check: the clause keeps its verb (deponit) and its object (curam); the lexeme the quarantine removes is supplied in plain words ("teachings") beside the marker, so nothing of the sense sits inside the marker. Declined: a data/tei-patches/11535.json entry. A patch corrects markup, never words, and there is no markup defect here — only a reading. Left to Pattern 12.

### B2. [sic: alios] — 0009 @1252B Printed: In eamdem sententiam currit, quod alios idem Propheta de vitiorum exstirpatione, vel virtutum plantatione cavit. Done: "…that which the same Prophet [sic: alios] took heed of concerning…" Why: alios is a bare masculine accusative plural with no noun to agree with and no verb to govern it; it cannot be rendered in place. The near-certain reading is alias ("elsewhere") — and the plate itself supplies the internal control: one column later, at 1253B, Richard writes exactly the parallel formula with the right word, quod alias psallit. Corpus control: alios idem Propheta and alias idem Propheta both return zero files in the PL TEI, so the corpus is silent; the decisive evidence is the internal one at 1253B. 7a⁗ check: verb (cavit), negation (none) and conjunction (quod) all remain outside the marker.

B3–B9. [var: …] — the Vulgate-divergence list

See section C, which is the deliverable.

C. THE VULGATE-DIVERGENCE LIST (7a″)

Every italic span named in the launch brief was worked through word by word against the received text, plus every unitalicized continuous quotation in the range (Pattern 14 has no formatting precondition). Marked = a [var: …] fired. Logged = a real divergence, rendered as printed, judged immaterial. Resisted = Migne agrees with the received text and the instinct to "improve" was refused.

Marked (6)

(That is eight rows; "6" above counts distinct marker sites — 1253D and 1254A are two markers on one textual crux.)

Logged, rendered as printed, no marker (immaterial or non-scriptural)

Pulls RESISTED (Migne agrees with the received text; the instinct to "improve" was refused)

1. 0006 @1244C — Turbatus est a furore oculus meus. My first reflex bound this to Ps. 30:10 (conturbatus est in ira oculus meus) — which would have been catastrophic here, because the entire chapter Richard has just written distinguishes furor from ira (0005 @1241D: Furor est perturbatio mentis, totius expers rationis. Ira est mentis perturbatio magna, sed a ratione non penitus aliena). Checked: this is Ps. 6:8, where the Vulgate reads a furore exactly. Migne is right, my memory was wrong, and furore is the word the argument needs. No marker, nothing changed. The most instructive site in the range.
2. 0008 @1248D — Isa. 38:15 Recogitabo tibi omnes annos meos in amaritudine animae meae. Exact. The familiar English ("I shall go softly all my years") is a different verse tradition entirely and was not allowed near it.
3. 0008 @1248D — Ps. 115:12 Quid retribuam Domino pro omnibus quae retribuit mihi? Exact, ? and all. Kept the ? Migne prints (Pattern 8).
4. 0008 @1248D — Isa. 49:18 Leva in circuitu oculos tuos. Exact. The pull was to complete the verse (et vide: omnes isti congregati sunt) because Richard's next words are et vide — they are not Isaiah's, they are the Jeremiah half (see C, row 2), and completing the Isaiah verse would have buried the divergence.
5. 0008 @1250C / 1250D — Ps. 33:4 and Ps. 15:7. Magnificate Dominum mecum, et exaltemus nomen ejus in idipsum and Benedicam Dominum qui tribuit mihi intellectum: both exact.
6. 0008 @1250C — Ps. 80:4 Buccinate in neomenia tuba in insigni die solemnitatis vestrae, unitalicized and unnoted. Exact.
7. 0009 @1251B — Dan. 9:5 and Dan. 9:19. Peccavimus, iniquitatem fecimus, impie egimus, et recessimus, et declinavimus a mandatis tuis ac judiciis and Exaudi, Domine; placare, Domine; attende, et fac: both exact, including the semicolons.
8. 0009 @1252C — Dan. 2:21–22, the Dat sapientiam sapientibus block. Exact through et lux cum eo est, which is the clause the next column interrogates.
9. 0009 @1252C — Jas. 1:17. Exact (de sursum / desursum is spacing only).
10. 0009 @1253A / 1253B — Ps. 50:8, Ps. 118:18, Ps. 30:5, Ps. 13:3, Ps. 7:10. All five exact. On Ps. 118:18 the pull was Douay's "Open thou my eyes"; rendered "Unveil mine eyes" instead, because Revela is the very word the whole column is expounding (Revelat profunda et abscondita six times over) and "open" shares no root with it. Pattern 17: tuos/tua singular → thine/thy throughout.
11. 0008/0009 — Dan. 2:20–21 printed TWICE (0008 @1250B, 0009 @1251B), as the brief asked be compared. ✅ Word for word identical between the two printings. The only difference is comma placement — 0009 adds a comma after sapientia and after regna. Both English renderings are therefore word-for-word identical and mirror the plate's commas. Also compared against the Vulgate: agrees, except that the Vulgate closes a sentence after ejus sunt where Migne runs on with et ipse mutat (punctuation, carried as printed).

D. Passages that will not construe, rendered as closely as the printed words allow

1. 0007 @1247A — Sed, ut omnes communiter et sollicitudine efficimus, alia orando, et Deo supplicando precibus obtinemus. The correlative is broken: alia … alia is wanted and only the second alia is printed, leaving efficimus without an object. Rendered "But, as we all in common effect things by carefulness too, other things we obtain by prayers, by praying and beseeching God." Declined [cj:] — it is a correlative, which is inside Wilson's 2026-08-15 bar, but the defect is an absent word with no printed word to open the gloss on ([cj:] requires Migne's printed word first, verbatim in the twin, and alia here would be glossing the surviving half of the pair, not the lost one). Flagged as a census candidate: if the convention is ever extended to a missing correlative whose surviving half is a bare alia, this is the site.
2. 0008 @1248C — et de praesentibus se explicare. Governed by possimus, so nos is wanted; se strands a third-person reflexive in a first-person-plural run. Rendered literally ("and extricate oneself from present ones"). Declined [cj:] — a person anomaly, explicitly outside the bar (11064 @1154C precedent).
3. 0008 @1250C — Neomenia, nova luna, nova luna intelligenti innovata, et divinitus aspirata. intelligenti is dative/ablative where the equation the sentence is making wants nominative intelligentia ("the new moon is understanding renewed"). Rendered "the new moon, renewed for the understanding and divinely inspired." Declined [cj:] — case anomaly, not a negation or correlative. Census candidate.
4. 0009 @1252B — cavit for cecinit. cavit is a real word occupying the verb slot, so no [sic:] (Pattern 12: it renders in place); rendered "took heed of". Corpus control: Propheta … cecinit is a common formula in the PL TEI (five hits in the first sample alone); Propheta … cavit returns none. The conjecture cecinit is near-certain and lives here, not in the text. Declined under the bar. See B2 for its neighbour alios, which is marked.
5. 0009 @1251C — nunc harum, illarum virtutum. The second nunc is not printed, though the exactly parallel sentence at 1252A prints it (qui nunc harum, nunc illarum virtutum). Rendered as printed ("now to the exercises of these virtues, of those"). Declined — a dropped adverb, not a negation.
6. 0009 @1253C — Praecipuus ordo charitatis bona communia praeferre privatis. No finite verb; est is not printed. Rendered as the printed ellipsis ("The chief order of charity, to prefer common goods to private ones"), with no bracket of any kind — an ellipsis supply is plain text or nothing, and here it is nothing.

E. Conjectures DECLINED — census candidates, named as the brief asks

F. Vocabulary calls — for ratification across the whole work

1. ⭐ fortitudo = "fortitude", not "strength". It is not in the brief's table but it occurs four times in this range, all inside Dan. 2:20 / 2:23 (quia sapientia et fortitudo ejus sunt; sapientiam et fortitudinem dedisti mihi), which is quoted three times across 0008 and 0009 and which Richard then expounds at length (0009 @1251C, the whole sapientia et fortitudo paragraph). "Fortitude" is also the Douay reading, and it keeps strength free for virtus corporis (below). Please ratify or overrule for the whole work before merge — this word will recur.
2. virtus corporis rendered "the strength of the body" (0009 @1251D), against the table's virtus = virtue. Followed my own judgment here and am logging it as the table instructs: the phrase is physical vigour in a list of the ages of man, no gloss hangs on it, and "the virtue of the body is more robust" would read as a moral claim in a work whose spine word is vir virtutum. Everywhere else in the range virtus is "virtue" (0009 @1252A virtutes regnare facit, diversarum virtutum efficaciam, etc.) — this is the single exception.
3. sacramentum = "sacrament", mysterium = "mystery" (0008 @1248B, @1249B–1249D). They collide if sacramentum is rendered "mystery", and they collide in adjacent sentences: Richard's argument at 1249B is that Daniel calls the matter a sacramentum, whence a mystica understanding must lie hidden in it — and four lines later mysterium revelatum est begins a whole page on mysterii revelatio. Kept apart. The etymological hook (sacramentum → mysticam) is weaker in English than in Latin; the alternative loses the distinction entirely, which is worse.
4. spiritalis / spiritualis both = "spiritual". The plate alternates them freely within one sentence (0005 @1241C spiritalium … spectaculorum, @1242C spiritalium doctrinam, beside @1242D spiritualia exercitia). Pattern 9 orthography; no marker, no crux.
5. Fixed-table words honoured against instinct, for the record: eruditio → instruction (0005 @1242D sacrae eruditionis studia; 0006 @1246A ad nostram eruditionem — "chastening" is plainly the sense at 1246A, sitting in a list with correctionem, correptionem, damnationem, and the table's ruling that the sentence carries it was followed). industria → effort (0007 @1247C supra humanam industriam; 0008 head XIII non nostram industriam). impudentia → shamelessness (0006 @1245C). praesumere → presume (0007 @1248A; 0009 @1252C).

G. Punctuation and structure carried under Pattern 8 — all deliberate, none to be "fixed"

1. 0006 head — ## CAPUT IX Quod quidquid… prints NO period after the numeral, where CAPUT VII, VIII, X, XI, XIII and XIV all do. 0008 head — ## CAPUT XII likewise. Preserved in the English heads exactly (## CHAPTER IX That whatever…, ## CHAPTER XII With what…).
2. 0006 @1246A — Daniel ingressus rogavit regem Quid est autem praedicta illa… No terminal stop between the closing re-quotation and the next sentence. No period supplied; the English runs on identically ("…asked the king But what is that aforesaid…"). This will look like our defect and is not.
3. 0006 @1246B — the chapter ends on a question (Quid est … nisi … agnoscere) that Migne closes with a PERIOD. No ? supplied. Same at 0009 @1253B (Sed quid sibi voluit quod dicit, et lux cum eo est.) — a direct interrogative, closed with a period, carried as a period.
4. 0005 @1243C — (rigor videlicet; et severitas justitiae) — a semicolon inside a parenthetical gloss. Preserved: "(namely rigor; and the severity of justice)".
5. Question-mark and exclamation parity verified per chunk against the Latin twin: 0005 8/0, 0006 3/0, 0007 5/0, 0008 7/0, 0009 8/1. Column anchors 12/8/7/12/12 and [n:] 2/0/0/11/16, all matching frontmatter.

H. Migne's own citations, reproduced verbatim and NOT corrected

Flagged only so a later reader does not mistake them for our binding errors. Three [n:] references in this range do not point where their text lives, and all three are reproduced exactly as printed (the [n:] verbatim rule; and per the runbook's false-positive 2, an antecedent a following (ibid.) depends on is never "corrected"):

1. 0008 @1248D — [n: (Psal. XCVIII)] on Feci judicium et justitiam, which is Ps. 118:121.
2. 0009 @1251C — [n: (I Cor. XIV)] on Quid enim habes, quod non accepisti?, which is 1 Cor. 4:7.
3. 0009 @1253B — [n: (Psal. XII)] on Scrutans corda et renes Deus, which is Ps. 7:10. And @1252B [n: (ibid.)] binds a Ps. 43:3 text to the preceding Ps. LXXIX — the (ibid.) whose divergence is marked in section C.

Richard of St Victor, De eruditione hominis interioris, Book I, CAPUT XV–XIX. Range: src/english/11535/0010.md … 0014.md. Entry form: column · printed reading · what I did · why.

1. Markers placed

### 1254B · secretioria · [sic: secretioria] Printed secretioria aequitatis arcana where the comparative is secretiora. Control run: grep -rlF over the 5,276 PL TEI files — secretioria 2 files (11535, 11787) against secretiora 107 files. A form that occurs twice in 87M words is the plate's, not the language's. Carried untranslated in italics under Patterns 7/10/12; the English reads "the [sic: secretioria] secrets of equity". 7a⁗ check: the marked run is an attributive adjective — it carries no verb, no negation, no conjunction, and the clause's predicate (quaerere) and its object (arcana) are both outside the quarantine. No sense is supplied beside it, and none is needed. Conjecture: secretiora, "the more secret".

### 1256B · ita que · [sic: ita que] — the weakest marker in this range; say so Printed Quidam ita que sunt qui nunquam nisi… — itaque split across a space. Controls run, three of them: (a) "ita que " across all 5,276 TEI files returns 0 files — the string does not exist in the corpus outside this site, where our chunker joined a line break into a space; (b) the TEI carries no line-final hyphens at all (grep -c -- '-$' 11535.xml = 0), so the transcriber's convention is to rejoin hyphenation, and a surviving split is anomalous; (c) a scan of every line-boundary in 11535.xml for a short line-final token joining a following -que/-tur/-tio/-bus suffix returns exactly one hit in the whole work: this one. Unique at this site, against a convention that elsewhere holds. ⚠ The honest weakness: the split lives on a TEI line break (ita↵que), so it is equally consistent with Corpus Corporum having failed to rejoin one hyphenated word. Our source is the TEI, not the plate, and I could not check the plate. If a plate read ever happens on col. 1256B and shows itaque set whole, this marker should be withdrawn and the line rejoined silently — it would be a source artifact (Pattern 13's family), not Migne's type. 7a⁗ check: itaque is a discourse connective and the quarantine is the size of the whole lexeme (7a⁗-b). Its force is therefore supplied in plain words beside the marker: "There are certain men, then [sic: ita que], who…". Nothing else in the clause was inside the quarantine.

### 1259A · scipsos · [sic: scipsos] Printed qui scipsos inglorios vident for seipsos — broken type, the familiar Migne e→c. Control: scipsos = 2 files corpus-wide (11535, 7266); seipsos = 631 files. Carried untranslated. 7a⁗ check: the run is a reflexive pronoun object. The clause keeps its verb (vident → "see"), its predicate (inglorios → "without glory") and its following negation (videri non refugiunt → "do not shrink from being so seen"). The English supplies "themselves" in plain words beside the marker so the clause still has an object — no brackets, per the brief's ten-marker rule.

2. The Vulgate-divergence list — [var: …] fired (6)

Every italic span named in my launch message was read word by word against the received text before the English was written. Six diverged materially.

### 1256A · Dan 2:19 · [var: Vulg. per visum nocte, "by a sight in the night"] Migne prints Tunc Danieli per visionem nocte mysterium revelatum est; the Vulgate has Tunc Danieli mysterium per visum nocte revelatum est. Material, and material to this work specifically: Richard's whole exposition runs on the somnium / visio distinction (fixed vocabulary table: "keep them apart — Richard argues from the difference"), and the sentence he is glossing is the one that gives him mysticae visionis notitiam. The received text's visus would not carry that hook. Word order also differs; not separately marked.

### 1256D · 1 Pet 3:15 · [var: Vulg. de ea, quae in vobis est, spe, …] Migne prints parati semper … ad satisfactionem omni poscenti rationem et ea quae in nobis est fide, et spe. The received text reads ad satisfactionem omni poscenti vos rationem de ea, quae in vobis est, spe — no fide, vobis not nobis, de not et, and vos dropped. Introduced by quemadmodum scriptum est and carrying its own [n: (I Petr. III)], so Pattern 14 fires squarely. Declined alongside it: the printed et for de is a real word in a real slot — ea … fide, et spe construes as an ablative of accompaniment ("and with that faith and hope which is in us"), which is how the English renders it. Under Wilson's 2026-08-15 bar a [cj:] is available only where the wrongly-printed word is a negation or a correlative; et is neither. Declined, named here as a census candidate.

### 1257B · Dan 2:25 · [var: Vulg. annuntiet, "may announce" …] The lemma prints qui solutionem regi significet. The Vulgate has annuntiet — and so does Richard, three times, in his own exposition of this very verse at 1257D (twice) and 1259A. The divergence is therefore internal to the plate as well as external to it, which is what makes it material rather than cosmetic: the word the commentary expounds is not the word the lemma prints. The English keeps significet's "may signify" in the lemma and "may announce" in every re-quotation, exactly as printed.

### 1259D–1260A · Dan 2:28 · [var: Vulg. tibi, rex Nabuchodonosor …] The lemma prints qui indicavit tibi Nabuchodonosor quae ventura sunt novissimis temporibus; the Vulgate has indicavit tibi, rex Nabuchodonosor. Same shape as the entry above: Richard's own re-quotation at 1261A restores the title — Qui indicavit tibi, rex Nabuchodonosor — and then builds a whole argument on the verse's other words (non ait quae ventura sunt tibi, sed quae ventura sunt novissimis temporibus). The same note records the second divergence in the lemma, Migne's plural in coelis against the received singular in caelo; that one is not sense-changing and would not have earned its own marker. Also noted, not marked: Migne sets sapientes, magi, et arioli, et aruspices where the received text has no et after magi. A list connective; no sense at stake.

### 1261A · Luke 17:21 · [var: Vulg. regnum Dei, "the kingdom of God" …] The clearest divergence in the range. Migne prints Regnum, inquit Dominus, coelorum intra vos est with [n: (Luc. XVII)]. Luke 17:21 reads regnum Dei intra vos est; regnum coelorum is Matthew's formula, not Luke's, and the note names Luke. Rendered as printed ("The kingdom … of heaven is within you"), because Richard's surrounding argument is explicitly about regnum coelorum — Neque enim alibi quam apud electos regnum coelorum suscitatur — so conforming the lemma to regnum Dei would have detached the proof-text from the sentence it proves.

### 1262A · Dan 2:29 · [var: Vulg. post haec, "after these things"] Migne prints cogitare coepisti in stratu tuo quid esset futurum post te; the Vulgate has post haec. Material: "what should be after thee" is a statement about the king's succession, "after these things" about the sequence of ages, and Richard's gloss (Quia de futuris studiose quaesisti) leans on the futurity either way. Also in the same lemma, not marked: in stratu tuo for the received in strato tuo — a fourth-declension form of a second-declension noun, no sense at stake (Pattern 9's neighbourhood).

### 1262C · Ps 37:11 · [var: Vulg. lumen oculorum meorum, "the light of my eyes"] Migne prints Et lumen oculorum, et ipsum non est mecum, dropping the possessive. Marked because the English then has to read "the light of the eyes, even that is not with me", which is exactly the reading a translator's ear wants to repair back to "my eyes" — the 7a″ pull, made visible rather than smoothed.

### 1262C · Job 29:2 · [var: Vulg. Quis mihi tribuat, ut sim juxta menses pristinos, "the former months"] Migne prints Quis mihi det, ut sim juxta dies pristinos? Two divergences, one of them material: months → days. Richard adduces the verse for ad pristinam gratiam aspirabat, and the English keeps "the former days" as printed.

3. The pulls I resisted — lemmata where Migne agrees and my instinct wanted to "improve"

Recorded as positive evidence, per the brief.

1. 1254B · Dan 2:24 (Post haec Daniel ingressus ad Arioch …). Agrees with the received text word for word, including introduc me in conspectu regis — the ablative, which reads like a slip for in conspectum and is not one. The pull was real and it has a witness: twenty lines later at 1256A Richard himself re-quotes it as in conspectum regis, accusative, and at 1259C back to in conspectu. Both forms stand as printed in the English; no marker, since the sense is identical and the lemma is the form that agrees with the Vulgate.
2. 1255B · Ps 111:5 (Jucundus homo qui miseretur et commodat…) — exact. Wanted the Douay cadence "Acceptable is the man"; the printed Jucundus is "pleasant", and that is what the English says.
3. 1256B · Job 3:24 (Antequam comedam suspiro…) — exact, tanquam for tamquam being orthography only.
4. 1256C · Ps 38:4 (Et in meditatione mea exardescet ignis) — exact; traditional "shall flame out" allowed under the scripture-echo licence because the Latin matches.
5. 1258A · Dan 9:5 (Peccavimus, iniquitatem fecimus…) — exact, all five verbs and both objects.
6. 1258B · Ps 43:16 and 1258C · Ps 43:18 — both exact. The second carries two printed negations (nec obliti sumus te, inique non egimus) and both are in the English.
7. 1260B · Ps 50:8, Ps 77:2, 1260B · Jer 1:11, Jer 1:13 — all four exact. Ollam succensam wanted the Douay "boiling caldron"; rendered "a pot set burning", which is what succensam says.
8. 1260D · Dan 2:35 (Lapis autem qui percusserat statuam, factus est mons magnus, et implevit universam terram) — exact, and printed in roman, not italic. Resisted the pull to italicize it into a lemma; the plate does not, so neither do we.
9. 1261C/1262A · Dan 2:28b and 2:30 (Somnium tuum, et visiones capitis tui…; Mihi … non in sapientia quae est in me plus quam in cunctis viventibus…) — both exact. 2:30 also carries a printed italic boundary that English wants to close up: Migne sets Mihi autem quoque non in sapientia…, with autem in roman outside the quotation. The two italic runs are preserved 1:1 ("To me moreover too, not in the wisdom…") even though it reads awkwardly.
10. 1262B · Ps 102:8–9 (Miserator, et misericors Dominus…) — exact, both negations (non in perpetuum, neque in aeternum) carried.
11. 1262C · Ps 73:9 (Jam non est propheta, et nos non cognoscet amplius) — exact, both negations carried.
12. 1262C · Ps 118:18 (Revela oculos meos…) — exact.
13. 1263B · Matt 10:8 (Gratis accepistis, gratis date) — exact. Second person plural, so "you", not "thou", per Pattern 17 — inside a chapter that is otherwise addressing one reader as tu throughout. The switch is the source's and is visible on the page.

### Divergences noticed and deliberately NOT marked (sub-material)

• 1258C · 1 Cor 4:4 — Migne prints Nihil mihi conscius sum without the received enim. A dropped connective, no sense at stake.
• 1260C · Jer 1:6 — Migne prints A, a, a, Domine, nescio loqui; the received text has A, a, a, Domine Deus, ecce nescio loqui, quia puer ego sum. Treated as an abbreviated citation rather than a divergent one — Richard quotes only as far as he needs, which is the norm throughout this work — so no marker. Named here as a declined site in case a later census wants it.
• 1262B · Ps 76:8–9 — Migne prints Nunquid in aeternum projiciet Deus, et non apponet…? Aut in finem…? against the received …projiciet Deus? aut non apponet…. Migne fuses the psalm's first two questions into one by swapping the connective and deleting a question mark. Rendered exactly as printed (one question, then a second), both printed ? preserved and none supplied (Pattern 8). Declined: a connective swap plus repunctuation, with the non intact in both witnesses.
• 1263B · 1 Cor 4:7 — Quid enim habes for the received Quid autem habes. Connective only.
• 1259B · 1 Cor 2:15, 1262D · Rev 2:5 + Luke 10:41, 1262D · Ps 9:38 — four scriptural sentences woven into Richard's own prose, printed in roman with no note. The most interesting is 1 Cor 2:15: Migne prints Spiritualis enim dijudicat omnia where the received text has Spiritualis autem judicat omnia — and the verb has been conformed to Richard, whose immediately preceding clause is omnia discutit, omnia dijudicat. Rev 2:5 likewise prints priora opera fac for the received prima opera fac. Both rendered as printed; no [var:], because none of the four is adduced as a proof-text or set off as a quotation — Pattern 14 is for scripture quoted as scripture. Named as declined sites.

4. Conjectures DECLINED — census candidates, by name

Under Wilson's 2026-08-15 bar ([cj:] only where the wrongly-printed real word is a negation or a correlative) every one of these stays out of the text. All are real Latin words in real grammatical slots; all are rendered literally.

• 1254D–1255A · suspicere. Cum locum magisterii sive ex praecepto, sive pro officio suspicere oportet — the sense the sentence needs is suscipere, "to take up" (and Richard writes docendi locum suscipimus eight lines later, and ad tantae gratiae susceptionem at 1257D). suspicere is a real verb, "to look up to", and the English says so: "When the place of mastery must be looked up to…". Neither a negation nor a correlative → declined. Control: officio suspicere occurs in 1 file corpus-wide (this one); officio suscipere in 2. That control is too thin to carry weight either way, and suspicere/suscipere interchange is itself attested in medieval orthography (Pattern 9 territory) — which is a second reason not to mark it.
• 1261A · praefectis. …non uni alicui tantum, sed pluribus, seu etiam praefectis pene omnibus juxta aliquid convenire queat — almost certainly perfectis, "the perfect", which is the class Richard has been describing and which praefectis, "prefects", is not. Rendered literally ("or even almost all prefects"). Declined: a noun substitution, not a negation or correlative. This is the site in my range most likely to be wanted if the bar is ever widened.
• 1257B · Daniel ad Arioch manuducitur. The whole exposition has Daniel led by Arioch (ab Arioch … introducitur, 1257B, twice). Printed ad. Rendered literally ("Daniel is led by the hand to Arioch"). A preposition, declined.
• 1263B · devotione (ablative). Si igitur devotione gratis accipimus — the argument wants the accusative devotionem ("if we receive devotion freely"), which is what the preceding sentence establishes (ipsam devotionem … ex munere largientis accipimus) and what the apodosis needs (id quod ejus merito obtinemus). Rendered on the printed ablative: "If therefore by devotion we receive freely…". Case anomaly, declined.
• 1256C · tam. quando meditationibus nostris tam divina revelatio occurrit — reads as jam. Rendered on tam ("when so divine a revelation occurs"). Declined.

5. Passages that would not construe, and other judgment calls

• 1257B · Optimus autem pene modus compunctionis qui surgit ex consideratione districti examinis. No finite verb printed. The English supplies "is" as plain text with no brackets (a grammatical ellipsis supply, per the brief's ten-marker rule), and nothing else. No marker, no [ed:].
• 1254D · fierique defendit. In devotio saepe iniquum esse invenit, fierique defendit, quod violentus animi impetus … fieri decrevit, defendere has to be the medieval "forbid" (cf. French défendre), against the classical "defend" — which would make devotion defend the very thing it has just found iniquitous, and would contradict the parallel Daniel … fieri prohibuit three lines above. Rendered "forbids its being done". Flagging the call because it is a lexical choice, not a rendering of an obvious word, and a later reader may want to revisit it.
• 1256A · the mid-sentence full stop. Migne prints alii in sua oratione, alii in sua meditatione. alii autem in ipsa sua saepe locutione, revelationem divinam accipiunt — a period inside the sentence, followed by lowercase alii. Preserved exactly, period and lowercase both (Pattern 8: never delete a mark he prints).
• 1258C · no stop after the note. Migne prints …in testamento tuo [n: (ibid.)] Sed cum inter instantia flagella… with no terminal punctuation between the quotation and Sed. Mirrored: the English also runs the note straight into "But when…". Never supplied.
• 1258B · Nonne … gemebat, qui dicebat: …cooperuit me. A Nonne question that Migne closes with a period, not a question mark. Preserved as a period. This is the one place in the range where an English ear insists on a ? and does not get one.
• 1259A · the anchor inside the marked run. Latin sets qui scipsos [1259A] inglorios vident. The English keeps marker and anchor in printed order — see themselves [sic: scipsos] [1259A] without glory — rather than moving the anchor to a tidier English boundary.

## 6. Vocabulary — where I followed the table against my own judgment, and where the work needs a ratification

Followed the fixed table without exception. mens = mind, animus = spirit, anima = soul, cor = heart, somnium = dream, visio = vision, gradus = step, detrimentum = decline, virtus = virtue, industria = effort, studium = zeal, praesumere = presume, impudenter = shamelessly, liberum arbitrium = free will. No spiritus occurs anywhere in 0010–0014, so the animus/spiritus collision the brief warns about (chunks 0005 and 0034) does not arise here; animus is "spirit" throughout with no competitor on the page.

Two places where the table cost something and I followed it anyway:

• 1256D · ad tantam eruditionem profecisse → "have advanced to so great an instruction". The sentence wants "learning" or "attainment"; eruditio is the title word and is fixed as "instruction", so "instruction" it is. It reads better at 1258D (modo ad eruditionem, modo ad probationem … flagellamur → "now for instruction, now for testing"), which is presumably why the table was set that way.
• 1259B · subtilis of a hair. Est autem capillus capitis … longus admodum atque subtilis — natural English is "fine" or "slender", but the etymology gloss for Balthasar hangs on it: subtilis profunditas, et profunda subtilitas mentis, ad omnem subtilitatem exacuit, ad miram subtilitatem, subtilia penetrare, subtiliter disserere. Rendered "subtle" everywhere, including of the hair, so the gloss still explains something. Same reasoning as the stylus/epistylium precedent.

Calls I set for this range and would like ratified for the whole work (all five are words that recur far beyond my chunks):

7. Proper names — one call that a reader will notice

Migne's Judae / Judas is rendered "Judas", not "Juda(h)." Dan 2:25's de filiis transmigrationis Judae is genitive and could be either; but at 1258A Richard prints the nominative outright — Nam et Judas confitens interpretatur — in order to gloss it, and the brief requires that the Latin name an etymology explains stay visible so the gloss still explains something. Following the plate's own nominative, the English reads "the sons of the transmigration of Judas" at all five occurrences and "For Judas too is interpreted confessing". ⚠ It will read oddly to anyone expecting Douay's "the children of the captivity of Juda", and a reader may momentarily think of Iscariot. Flagging it rather than deciding for the whole work: if a later agent finds Richard using Juda elsewhere, this range should be swept to match.

Related, and not a defect: Migne prints both Hierusalem and Jerusalem within four lines at 1258A–B (quasi de Hierusalem ad Babylonem … Hierusalem namque visio pacis … de Jerusalem in Babylonem, and de Jerusalem in Babylonem captivatus at 1257D). Both spellings are carried into the English exactly where they stand. Pattern 9.

Etymology glosses, all three in this range, handled per the brief — translated, with the Latin name kept visible beside them: Judas = confessing (1258A) · Hierusalem = a vision of peace, Babylon = confusion (1258A) · Balthasar = a hair of the head (1259B). confusio is then held to "confusion" for the rest of the chapter — locum confusionis, injustae confusionis opprobrio, illam confusionis suae captivitatem, and Ps 43:16's confusio faciei meae — so the etymology keeps paying out where Richard makes it pay.

8. Left unresolved

1. The ita que marker at 1256B — plate-checkable, and should be checked. See §1.
2. 1261A praefectis — the strongest declined-conjecture site in the range; wanted by name if the [cj:] bar is ever widened past negations and correlatives.
3. 0010's word ratio is 1.63 (verifier warn; corpus target ~1.5). It is not padding: the chunk is a dense chain of quando/quasi/ejusmodi subordination and of ablative abstracts (ex magna animi devotione, intima perscrutatione, rectitudinis zelo) that English cannot carry in fewer words without paraphrase. Recording it so a later reader does not read the number as a symptom.
4. Not mine, but seen while verifying: 0008.md carries [var: Jer. 3:2 vide ubi non prostrata sis …], which fails verify-english check 11 because a [var: …] must open with the witness name (Vulg. | LXX | Gk | Heb. | Douay | Vet. Lat.). Reordering it to [var: Vulg. Jer. 3:2 …] clears it. Flagged for whoever holds 0005–0009.

Richard of St Victor, De eruditione hominis interioris, PL 196. Range: CAPUT XX–XXIV. Format: column · printed reading · what was done · why.

A. Markers fired

1. [sic: exaeguabo] — 1267C (0017), Lam. 2:13

Migne prints cui exaeguabo te, et consolabor te, virgo filia Sion? for exaequabo (g for q). Control run: exaeguabo occurs in exactly one file in the whole PL TEI (11535.xml, this site); exaequabo occurs in 7 files. A form that occurs once in 87M words, here, is the plate's defect. Carried untranslated under Patterns 7/10, marked [sic:].

⚠ 7a⁗ check. The quarantined run is the clause's only verb (cui … te). Following the [sic: eccidit] precedent, the sense is supplied in plain words beside the marker: to whom shall I make thee equal [sic: exaeguabo], and comfort thee. Deleting the marker's contents still leaves a predicate and the clause's interrogative force intact.

2. [sic: commatationem] — 1270B (0018)

O mirabilem (imo detestabilem) commatationem de charitate ad iniquitatem! — for commutationem. Control: commatation occurs in 1 file corpus-wide (this one); commutationem in 264. Plate defect. Carried, marked.

⚠ 7a⁗ check. The marker swallows the exclamation's head noun, so the English supplies it in plain words: O the wonderful (nay, the detestable) change [sic: commatationem] from charity to iniquity! Nothing but the broken type is inside the marker.

3. [sic: meus] — the CHAPTER XXIII head (0018)

The head prints Quod de magno saepe virtutum culmine soleat meus humana gradatim in ima corruere. — meus (masc. possessive) where the feminine humana requires mens. This is Pattern 12 test 1: a real word with no grammatical slot (it can be carried by no English at all, and the clause loses its subject). The chapter's own body is about mens/animus humanus falling by steps, and chunk 0019's head prints animus humanus in the identical frame.

⚠ 7a⁗ check. The quarantine removes the subject, so the English supplies it beside the marker: That from a great summit of virtues the human mind [sic: meus] is often wont to fall by degrees into the depths. No asterisks inside the marker — the head summary is already italic and nested italics are not expressible (Pattern 10's closing clause).

⚑ Head-wording flag for the merge pass: heads are an indexed field. This head is the only one in the range that carries a marker; if a peer's convention differs, this is the site to reconcile.

B. [var: …] — Migne's scripture against the received text (Pattern 14)

Nine markers fired. Every italic span named in the launch message was worked through; the divergences that did not earn a marker are in §C, and the pulls resisted are in §D.

C. Divergences examined and left UNMARKED (with the reason)

Every one of these was read word by word against the received text; none changes the sense, so Pattern 14's threshold is not met. They are logged because a later reader must not take the absence of a marker as evidence of agreement.

• 1265B, Dan 2:33 — quaedam fictilis for Vulg. quaedam autem fictilis: a dropped connective inside a list already asyndetic in Migne's setting.
• 1265B, Dan 2:34 — lapis autem qui percussit statuam for Vulg. percusserat: perfect for pluperfect, no change of fact.
• 1267D, Ps 48:13 — Homo cum in honore esset for Vulg. Et homo: dropped initial connective at the head of a quotation.
• 1268A, Ps 102:12 — Nam quantum distat for Vulg. Quantum distat: Nam is Richard's own hinge into the quotation, not a variant of it.
• 1268A, Rom 8:14 — hi filii sunt Dei for Vulg. ii sunt filii Dei: orthography and word order only.
• 1270C, Ezech 18:24 — Si averterit justus justitia sua for Vulg. Si autem averterit se justus a justitia sua. Looks like two losses, and is one: justitia stands as a bare ablative of separation, which is good Latin, so "turn away from his justice" is the literal rendering of the printed words. No repair was made and none was needed.
• 1272A, Isa 3:12 — qui beatum te dicunt for Vulg. qui te beatum dicunt: word order.
• 1272B, Abd 4 — Sed si exaltatus fueris … et inter sidera posueris: Sed is Richard's connective; the omitted second si is the Vulgate's, not a sense change. See §B for the coeli half of this pair.
• 1272B, Isa 47:2–3 — Et tunc revelabitur ignominia tua for Vulg. Revelabitur ignominia tua: an added temporal connective binding the two verses Migne has run together.
• 1271C, Apoc 2:4 — quod charitatem tuam primam dereliquisti for Vulg. reliquisti. Intensified compound, same sense. ⚑ Worth a note anyway: Richard's own gloss in the very next sentence prints the Vulgate's simple verb (quod charitatem primam reliquit), so the plate carries both forms four words apart. Evidence about the setting, not about the sense.
• 1273A, Eccli 2:16 — Vae illis for Vae his, and diverterunt in pravas for in vias pravas. Ellipsis of the repeated noun; no sense at stake.
• Citation labels. Four [n: …] locators in the range point at the wrong book or chapter — 0017's (Rom. XIV) on Rom. 8:14–17; 0018's (Apoc. VIII.) and 0019's (Apoc. VIII) on Apoc. 2; 0019's (Luc. IX) on Apoc. 2:2; 0016's (Abd. IV) on a one-chapter book. Note contents are Migne's citations and pass through verbatim and untranslated — nothing was corrected. Flagged for the indexer, not for the text.

D. Lemma pulls RESISTED (positive evidence — the brief asks for these by name)

1. ⭐ 1265B — the doubled statua is NOT dittography, and no [d: …] was fired. Migne prints ecce quasi statua una grandis statua illa magna, et statura sublimis stabat contra te, and the obvious reading is Pattern 11. Two controls say otherwise. (a) The Vulgate prints the same words at Dan 2:31 — quasi statua una grandis: statua illa magna — separated by a colon; what Migne has lost is the punctuation, not doubled the noun. (b) The work is its own control: chunk 0017 re-quotes the passage in two pieces and treats the two as consecutive phrases, glossing statua una grandis first and statua illa magna, et statura sublimis second. A [d: …] here would have told the reader the plate repeated itself when it did not. The English carries the run-on exactly as printed, with no supplied colon (Pattern 8).
2. 1264C — Rom 11:33 agrees with the Vulgate here and was left alone. Migne prints O altitudo, divitiarum, sapientiae, et scientiae Dei, quam incomprehensibilia sunt judicia ejus… ⚑ Chunk 0009 (col. 1252D) prints the same verse WITHOUT divitiarum and without sunt, and closes it with ! where 1264C closes with . So the divergent printing is 0009's, not this range's, and the pull to "regularize" the two occurrences to each other was resisted in both directions. This is evidence about the plate and belongs to whoever holds 0009.
3. 1264C — no ! supplied. Rom 11:33 is exclamatory and the Vulgate prints !; Migne prints a period here. Carried as printed (Pattern 8).
4. 1270D — Apoc 2:19 keeps its question mark. Migne closes a sentence that is not a question in the received text with ?. Carried exactly, per Pattern 8 and the launch brief. It is in fact defensible on the page — Richard's frame is Nonne operari tibi videtur in auro, cui dictum est a Domino: …? — but it was carried because it is printed, not because it construes.
5. 1272A and 1273A — two further printed periods on sentences that are questions (Nunquidnam tibi non tale videtur opus illorum… and Quid autem est, quod hujusmodi homines … famae.). Both carried with the printed period; no ? supplied.
6. 1268A, Ps 102:11 — secundum altitudinem coeli terra. The Vulgate has coeli a terra, and the missing preposition looks like a defect. It is not: terra is a well-formed ablative, and an ablative of separation without preposition construes. Rendered "according to the height of heaven from the earth"; no marker, no crux-level conjecture.
7. 1264B, Ps 41:5 and 1266A, Job 30:19 and 1266A/B, Job 7:16 & 13:15 and 1267D, Jer 2:20 and 1268B, Rom 1:28–29 and 1268D, Ps 49:21 and 1270D, 1 Cor 13:4–7 and 1272B, Apoc 3:1 — all verbatim against the Vulgate. Read word by word; nothing to mark. Listed so the negative result is on the record.
8. ⭐ The familiar-cadence pull, twice refused. Dan 2:35's favillam aestivae areae wants to become "the chaff of a summer threshingfloor" (Douay) and Ps 102:12's ortus wants to become "the east." Both were kept literal — "cinders", "the rising" — because 0017's exposition of favilla depends on it being the light ash that flies apart at a breath (ad quemlibet, vel tenuem flatum dissilire), which "chaff" would silently replace with a different image.

E. Conjectures DECLINED (census candidates, named as the brief requires)

None of these clears the [cj: …] bar of 2026-08-15 (negation or correlative only). All are rendered as printed.

1. ⭐ 1269D (0018) — simultas in the metal-vice list. In auro intelligitur charitas, in argento veritas, in aere simultas, in ferro crudelitas, in testa fictili fragilitas. Control: simultas occurs in 66 PL files and means enmity, rivalry (checked in context at 11771: ut nulla se ingerat simultas in opere). It is a real word in a real slot, so Pattern 12 has nothing to wrap and Pattern 7 governs: rendered "enmity." ⚠ But the plate itself glosses it as something else. The very next paragraph expounds the bronze as simulatio (Sic simulatio suavem adulationis tinnitum habet in sermone), and chunk 0019's closing summary prints in aereis simulatio. So the work's own two other statements of this slot say "dissimulation," and only this one says "enmity." This is the strongest Pattern 18 candidate in the range and it is declined only because the bar is scoped to negations and correlatives. If that bar is ever widened, this site should be revisited first. The English reads "enmity" in the list and "dissimulation" in the gloss — the mismatch is the plate's and is deliberately visible.
2. 1268D (0017) — servire for saevire. contra semetipsum per increpationem servire videtur — "seems to serve against itself by rebuke." saevire ("to rage against itself") would construe far better with increpationem. Declined: a real word, a filled slot, no negation or correlative involved. Rendered as printed.
3. 1265C (0016) — modus dilectionis est sine modo. quia modus dilectionis est sine modo, sine mensura est aestimationis suae magnitudo. The clause wants a correlative frame (quia ubi modus … sine modo est, sine mensura est …). ⚑ This one is arguably inside the [cj:] bar, since the word that went wrong is a correlative — but the bar's own wording is that the printed word must make the sentence say something the author is arguing against, and here it merely fails to construe as a period. Declined and named as a census candidate; rendered literally, the two clauses standing in bare parataxis as printed.
4. 1266B (0016) — mittere in mari. dicere huic monti tollere, et mittere in mari — ablative where the Vulgate's in mare (accusative of motion) belongs. Attested medieval usage (Pattern 9); no marker, no conjecture. Rendered "be cast into the sea."
5. 1273A (0019) — quae super et in prosperis per concupiscentiam resolvitur. super stands without an object and may be a setting of supra or a survival of semper. Declined; rendered "which both above and in prosperous things is dissolved."
6. 1269B (0017) — the verbless talem. ne forte oporteat velis nolis, quandoque interius vel exterius talem forte — esse is not printed. The infinitive is supplied as plain text with no brackets ("to be such"), per the brief's rule that a grammatical ellipsis supply takes no marker at all.
7. 1264D (0015) — seipsam for seipsum, of animus. Gender mismatch with no English exponent: Pattern 9, no marker, no edit, and flagged here so a later sweep does not churn it.

F. Vocabulary — calls made, and the ones wanted RATIFIED for the whole work

⭐ The three starred calls below govern the lemma the entire work expounds and are re-quoted in 0017 and 0021 by other agents. They need ratification, not silent inheritance.

1. ⭐ fictilis = "earthenware"; testa = "clay" (the brief's fixed term). The 1265B lemma prints BOTH words, and the exposition later turns on which one Migne prints where — 0019's fictiles plantas against 0017's testeos pedes, and 0018's vasa testea. Giving fictilis its own English word keeps the pair tellable apart on the page. Compound sites read accordingly: in testa fictili fragilitas (1269D) = "in the earthenware clay, fragility."
2. ⭐ favilla = "cinder(s)"; cinis = "ashes". Kept distinct because 0017 expounds favilla specifically (omni turbini cedere … ad quemlibet, vel tenuem flatum dissilire) and Job 30:19 prints the pair favillae et cineri. Douay's "chaff" at Dan 2:35 was refused (§D.8).
3. ⭐ lutum = "mire". Forced by the brief's testa = "clay": the two words are contrasted in the same sentences (inchoatur ex auro, et consummatur in luto, 1265C; velut vile lutum, 1265D) and cannot share an English word. Job 30:19 accordingly reads "I am compared to mire."
4. contritio / conterere = "crushing" / "to crush" throughout, including Dan 2:35 (contrita sunt pariter ferrum, testa, aes…), Lam 2:13 (Magna velut mare contritio tua) and its gloss at 1273A (Contritio propter testae corruptibilitatem). One English root, so the lemma and the gloss stay visibly the same word.
5. intuitus = "look"; aspectus = "aspect". They collide across a single paragraph in 0017 (intuitus ejus erat terribilis … O statuae nostrae aspectum vere terribilem), so the distinction is load-bearing there.
6. sustinentia = "endurance" (Eccli 2:16), kept off patientia = "patience", which stands four lines away in the same chunk.
7. simulatio = "dissimulation" — and see §E.1 for simultas, which is NOT given the same English.

Fixed-table entries followed against my own judgment (logged per the brief)

• virtus = "virtue" even at de Domini virtute praesumere (1266C) and de propria virtute et fortitudine (1269B), where "power"/"strength" reads far better in English. Followed the table; the tension is real and recurs in 0017's Omnem virtutem, omnem fortitudinem, omnem industriam tuam where virtus and fortitudo stand side by side.
• animus = "spirit", mens = "mind", anima = "soul" held apart everywhere. The range's tightest site is 1268D (0017), where animus, anima and conscientia occur in two consecutive sentences (tactus animus conscientiae propriae accusatione confunditur. Et … praevaricatrix anima statuitur). Neither of the two flagged animus/spiritus collision chunks (0005, 0034) is in this range.
• pietas = "mercy" (1266A de Domini pietate desperare, 1266B in sola Domini pietate), after the 9741 anchor's bonitatis et pietatis Dei → "the goodness and mercy of God". misericordia does not occur in the same chunk, so nothing collides; flagged in case a peer needs both words in one sentence elsewhere.
• detrimentum = "decline" at 1265B (per multiplicia detrimenta decurrens), per the table.
• gradus = "step" at 1269C (quanto gradus altior, tanto casus deterior) and in 0019's head.

G. Punctuation and structure carried as printed (Pattern 8)

Question-mark and exclamation parity is exact across all five pairs (0015 1/1 · 0016 4/4 · 0017 11/11 · 0018 4/4 · 0019 14/14 ?; exclamations 0/0 · 4/4 · 5/5 · 4/4 · 0/0). Specific carries worth naming:

• 1266B (0016) — lowercase o after !: O quanta mentis altitudo! o quanta fortitudo. Preserved lowercase.
• 1267C (0017) — Migne's semicolon after filia Hierusalem where the Vulgate prints ?. Carried as a semicolon; no question mark supplied.
• Three questions printed with periods and one statement printed with a question mark: §D.4–5.

Paragraphing follows the Latin twin 1:1 (one body paragraph per chunk, as printed). No verse lineation anywhere in the range. No guillemets in either twin; the few word-as-word quotations in 0017 ("great," … "lofty") use straight quotes, which are ours, per Pattern 6.

H. Within-work drift found against peer chunks — for the merge pass (runbook step 4)

Checked after writing, against the English peers already on disk. Two conflicts, both on the starred calls in §F, and both are within one work, so the "let each stand" ruling does not apply:

1. ⚠ fictilis — chunk 0021 renders it "clay", the same word it gives testa. Its 1276B re-quotation reads "its feet of iron and of clay", and its gloss then runs "the iron and the clay are crushed". That collapses the very pair the launch brief asked to be kept distinct, and 0021 is re-quoting the same Dan. 2:34 lemma my 0016 carries. Recommend 0021 move to "earthenware" for fictilis and keep "clay" for testa, per §F.1. (0026's lutum = "mire" already agrees with §F.3 — no action there.)
2. favilla — chunk 0022 renders it "ash", where 0015–0019 render "cinder(s)". 0022 is the chunk that expounds favilla at length (nine occurrences), so it is the natural place to settle it; but note its "chaff" is for palea, not favilla, so there is no three-way collision, only a two-way one. Whichever word wins, the Dan. 2:35 lemma at 1265B and the exposition at 1269A must use it.

I. Left unresolved

• §E.1 (simultas, 1269D) is the one site in the range where the shipped English will look wrong to a reader who reads the next paragraph. It is deliberate and it is the largest open judgment here.
• §D.2 (Rom 11:33 at 0009 vs 1264C) needs whoever holds chunk 0009 to see this note; the divergent printing is on their column, not mine.
• §A.3 ([sic: meus] in a ## head) is the only head-borne marker in the range and should be confirmed at the merge pass, since heads index as identity.

Richard of St Victor, De eruditione hominis interioris, PL 196. Range: chunks 0020, 0021, 0022, 0023, 0024. Format: column · printed reading · what I did · why.

A. Markers placed

[var: …] — Migne's scripture against the received text (Pattern 14)

1. 1276A–B · Dan. 2:34 · Videbas, ita, donec abscissus est lapis sine manibus… Migne prints the lemma without de monte; the Clementine Vulgate reads abscissus est lapis de monte sine manibus. Rendered as printed; [var: Vulg. adds de monte after lapis: the stone is "cut out of a mountain"]. Material, because Richard's whole argument in 0022 turns on the stone becoming a mountain (lapis non crescit in rupem sed mutatur in montem); the Vulgate's de monte would have the mountain there from the start and would blunt the point he is making. Chunk 0016, which prints the whole dream, likewise lacks de monte, so this is the work's text, not a one-off setting.

2. 1276C · Jer. · Percussi, filios vestros, et non doluerunt, renuerunt recipere disciplinam [n: (Jer. V)]. Migne's note points to Jer. 5, whose Vulgate reads Percussisti eos, et non doluerunt: attrivisti eos, et renuerunt accipere disciplinam — second person, and no sons. The opening words (percussi filios vestros) are Jer. 2:30. Rendered as printed; [var:] names both. Material: Migne's first person makes God the striker in his own voice, which is exactly the point Richard is pressing (melius operatur inspiratione quam percussione). recipere for accipere noted inside the same marker; on its own it would not have earned one.

3. 1276D · Ps. 50:19 · cor contritum et humiliatum Deus non despiciet (unitalicized, but a proof-text and therefore inside Pattern 14's scope). Vulgate: Cor contritum et humiliatum, Deus, non despicies — a vocative and a second person. Migne's third-person despiciet turns an address to God into a statement about him. Rendered as printed ("God will not despise"); [var:] gives the Vulgate. Material under the runbook's mechanism 5 (agency mis-bound) — this is the same class, one step milder.

4. 1278C · Dan. 2:35 · nullusque locus inventus est in eis Vulgate (and this work's own chunk 0016): nullusque locus inventus est eis, dative, "no place was found for them." Migne's in eis is locative. Rendered as printed; [var:] gives both and names 0016 as the internal control. Material and internally attested: Richard's own gloss at 1279D reads Raptis autem vento nullus locus eis invenitur — he glosses the dative, so the lemma at 1278C disagrees with its own exposition nine lines later.

5. 1279A · Eccl. 1:13 · Proposui, in corde meo quaerere… Vulgate: proposui in animo meo quaerere. Rendered "in my heart"; [var: Vulg. in animo meo, "in my mind"]. Material in this work specifically, because cor / animus / mens are the load-bearing seam of the fixed vocabulary and Richard is elsewhere precise about them.

6. 1279A · Phil. 3:8 · Omnia arbitratus sum ut stercora, ut Christum lucrifacerem Vulgate: arbitror ut stercora, ut Christum lucrifaciam — present throughout. Migne's perfect/imperfect subjunctive is rendered as printed; [var:] gives the Vulgate tense. Richard's frame (cum dicebat) is past, so the tense shift is doing work on the page.

7. 1279D–1280A · I John 4:18 · perfecta charitas mittit timorem (unitalicized proof-text). Vulgate: perfecta caritas foras mittit timorem, "casteth fear out." Without foras the sentence says perfect charity sends fear, which is the opposite of what Richard immediately draws from it (Nescit enim perfectio charitatis augere formidinem). Rendered as printed; [var:] supplies the Vulgate. This is the most consequential divergence in the range — a dropped adverb inverts the proof-text.

8. 1281B · Dan. 2:37 · Tu, rex regum es, et Dominus coeli… Vulgate: et Deus caeli regnum et fortitudinem… Rendered "the Lord of heaven"; [var:]. ⚠ The three-way comparison the brief asked for: the verse is printed three times in this stretch — 1281A (Tu, Rex regum es, capitalized, no continuation), 1281B (rex regum … Dominus coeli), and chunk 0025 at 1284A (rex regum … Deus coeli, the Vulgate reading, with the Dan. 2:38 continuation attached). So Migne prints the Vulgate's Deus once and Dominus once, in the same work, seven columns apart. Only the 1281B occurrence is marked, since 0025's agrees with the received text. Recommendation for whoever holds 0025: do not "harmonize" the two; the disagreement is the datum. The capitalized Rex at 1281A is a typographic variant only, and takes no marker — but note that Richard glosses it as a title (Rex itaque regum est rationalis consensus), which is very likely why the compositor capitalized it.

9. 1283B–C · Luke 17:21 · Regnum coelorum, intra vos est Vulgate: ecce enim regnum Dei intra vos est. Rendered "the kingdom of heaven"; [var: Vulg. regnum Dei]. Material and load-bearing: the sentence it proves is Neque enim habet hujus regni imperator sortem in regno coelorum, and the following paragraph runs duo illa praedicta regna pertineant ad regnum coelorum, si omnino regnum coelorum intra nos est. Richard needs coelorum, and Migne's lemma gives it to him where the Vulgate does not. This is the clearest case in the range of a lemma that has been conformed to the argument rather than to the Vulgate.

[sic: …] — carried type (Pattern 12)

10. 1279B · reatus nostri formido omnis atque suscipio absorbetur suscipio is a first-person verb standing in a nominative noun slot coordinated with formido; the intended word is plainly suspicio, "suspicion." Carried as [sic: suscipio]. Corpus control run: atque suspicio occurs in 11 PL files; atque suscipio occurs in 3, and I read all three — 9285 (confiteor, et adoro, atque suscipio sacras… imagines) and 7530 (veneratione complector atque suscipio) are both legitimate first-person verbs. This site is the only one in 87M words where the form stands in a noun slot. 7a⁗ check: the clause keeps its verb (absorbetur, "is swallowed up") and its first subject ("all dread of our guilt"); no negation, interrogative or conjunction is inside the quarantine.

11. 1280C (head) · et quatem potestatem ex gratia accipimus quatem is not a Latin word; qualem is, and the head's own first clause prints Qualem libertatem. Carried as [sic: quatem] inside the head's italic summary (no nested asterisks — Pattern 10's rule for a carried run inside an italic span). Corpus control: qualem potestatem occurs 6×, quatem potestatem once, in this file. 7a⁗ check: the quarantine removes an interrogative adjective, but the parallel Qualem three words earlier is intact and governs the series, so the reader sees the gap in the right place and the head still reads as a list of indirect questions.

12. 1282D · Simile esse Deo, mira ile decus; similem esse pecoribus, miserabile dedecus. A word-division defect (Pattern 10, split direction) for mirabile. Carried whole: [sic: mira ile]. Corpus control: mira ile occurs in one file (this one); mirabile decus occurs in zero, so the control here is negative and I say so rather than overstating it — the conjecture rests on the antithesis, not on a parallel. 7a⁗ check (and this is the 7a⁗-b shape, so it was checked hard): the clause is verbless in the Latin too (decus / dedecus in apposition), so no predicate is lost; no negation, no conjunction. What the quarantine does cost is half the mirabile / miserabile jingle, which is why it is recorded here in full.

B. The pulls I resisted — lemmata where Migne agrees and my instinct did not

These are positive evidence, per the brief.

1. 1277A · Ps. 49:21 · Arguam te, et statuam contra faciem tuam. Douay reads "and set before thy face," and that phrasing arrived before the Latin was read. Migne prints contra, the Vulgate prints contra, and Richard's gloss hangs on it: the very next clause is per ejusmodi percussionem seipsum contra seipsum erexerat. Rendered "will set against thy face." No marker: nothing diverges. ⚑ Note also the pun Richard is working, which an English reader can only be told about: statuam is both "I will set" and "statue," the object of the whole exposition. Logged, not marked; there is no marker for a pun.
2. 1277A · Ps. 50:5 · peccatum meum contra me est semper. Douay's "always before me" is the received English. Migne and the Vulgate both print contra me, and the same contra … contra hook as above depends on it. Rendered "always against me."
3. 1280A · Ps. 124:1 · Qui confidunt in Domino sicut mons Sion. Migne here prints the Vulgate's plural. ⚠ Chunk 0016 of this same work prints the singular Qui confidit in Domino with a [n: (Psal. CXXIV)]. My chunk agrees with the Vulgate and takes no marker; whoever re-reads 0016 should consider one there. Flagged because a later reader who meets both will assume one of them is ours.
4. 1279C · Ps. 30:2 · In te speravi, non confundar in aeternum. Vulgate has In te, Domine, speravi. A pure omission that changes nothing; rendered as printed, no marker.
5. 1279A · Rom. 6:21 · Quem ergo fructum habuistis in quibus nunc erubescitis? Vulgate: habuistis tunc in illis, in quibus… Omission only; no marker. (Also Pattern 17: habuistis / erubescitis are plural, so "you," not "thou.")
6. 1276C · Isa. 55:11 · non redibit ad me vacuum … et prosperabit in his. Vulgate: non revertetur … et prosperabitur. [var:] declined, deliberately. redire / reverti are synonyms, and the active/passive shift in prosperabit has no English exponent ("shall prosper" renders both), so a marker would announce a divergence the reader could not see. Recorded here instead. Named because it is precisely the sort of thing a divergence-hunting agent over-marks.
7. 1279C · Ps. 102:2–3 · Benedic, anima mea, Domino… — exact, including the elision of nothing between v. 2 and v. 3. No marker.
8. 1281C · II Cor. 1:12 and John 15:5 — both exact (the former drops only Nam).
9. 1282D · Ps. 48:13 · Homo, cum in honore esset non intellexit… — exact.
10. 1280A, 1280B · Dan. 2:35 · Lapis autem, qui percussit statuam. Vulgate: qui percusserat. Pluperfect → perfect; no sense change; chunk 0016 also prints percussit. No marker.
11. 1282D–1283A · Eccl. 7:30 · quia fecerit Deus hominem rectum, et ipse immiscuerit se infinitis quaestionibus. Vulgate: quod fecerit … et ipse se infinitis miscuerit quaestionibus. Compound verb + word order only. No marker. Migne's note reads (Eccli. VII) for Ecclesiastes; [n:] contents are verbatim and were not touched.
12. 1276A · Matt. 3:9 · potens est enim Dominus de lapidibus istis… Vulgate: potens est Deus. Declined: same referent, no argument rests on it. (Compare item 8 in section A, where Dominus for Deus was marked — there it stood inside a lemma printed three times with two different readings, which makes it evidence about the setting; here it does not.)
13. 1283D · Dan. 2:40 — Migne prints Et quartum regnum and sic comminuet haec omnia et conteret; the Vulgate has Et regnum quartum and sic comminuet et conteret omnia haec, and chunk 0025 prints yet a third arrangement (Et quartum, inquit, regnum … sic comminuet omnia haec, et conteret). Word order only, three ways, in one work. No marker; recorded because the variation is itself interesting.

C. Passages that would not construe, and conjectures declined

Every one of these is a census candidate. The [cj:] bar (a negation or a correlative) is not met by any of them, so none was glossed in the text; all are rendered as printed.

1. 1275C · quia quo quisque proximum suum minus amat, eo utique minus infirma molestius tolerat. Two comparatives (minus, molestius) in one eo clause; the sense the argument needs is "the more troublesomely he bears weak things," and minus after eo utique contradicts it. Rendered as printed ("by so much the less, assuredly, does he bear weak things the more troublesomely"), which is as close as the words allow. Declined: not a negation, not a correlative — but note it is adjacent to the correlative class (quo … eo), and it is the one site here I would re-put to Wilson if the bar is ever widened.
2. 1275D · dummodo in alios furit, modo in semetipso torpescit. The neighbouring modo makes modo … modo ("now … now") the evident intent; dummodo is a real conjunction and renders ("so long as"), so nothing is marked. Declined.
3. 1277C · usque ad digni vel debiti finis talum non perducitur. talum = "ankle, heel." Rendered literally, "even to the heel of a worthy or due end." Corpus control: finis talum occurs in exactly one file, this one. Declined nonetheless, and deliberately: in a work whose entire subject is a statue read from head to feet, "the heel of an end" may well be Richard's own figure rather than a defect. Naming it here so that nobody silently emends it to terminum.
4. 1278A · Unde utique est, quod jactantiae argentum quam aurum arrogantiae, quo facilius eo et citius atteritur. The quam clause has no comparative to attach to. Rendered as closely as the words allow. Declined.
5. 1279C · Nonne pro favillarum eventilatione Domino supplicabit, cum Domino in oratione proclamaret… Future indicative against imperfect subjunctive, where the two flanking sentences both use past tenses (cum Propheta dicebat, cum … erumperet). supplicabat is the evident reading. Declined — a tense/person anomaly, explicitly outside the bar.
6. 1279C · Vide ergo quia corruptionis opibus in favillam redigitur. "By the riches of corruption" — rendered literally, with no expressed subject. Corpus control: corruptionis opus occurs 2×, corruptionis opibus only here. Declined.
7. 1280C · sed non aequa accepit qualitatem personae agnoscere. Rendered "but he did not receive equal things, to recognize…" aeque would construe. Declined.
8. 1282B · Volucribus coeli potenter imperat, cui cogitationes suas … occupat. The two parallel sentences on either side both use cum + indicative; cui leaves the relative unattached. Rendered literally ("to whom he occupies his own thoughts"). Declined — not a negation or correlative. This is the least satisfying English in the range and I would rather it were seen than smoothed.
9. 1276A · nec propter testeam fragilitatem de suo corpore redargui. The series is tumore … errore … rigore, three ablatives in -ore naming a fault; torpore would complete it and corpore breaks it. Rendered "reproved concerning their own body." Declined.
10. 1275C · Et quidem saepe quaedam … praesumpserunt … vivunt … possunt. Feminine/neuter plural pronoun governing masculine plural participles and verbs. Pattern 9: grammatical gender has no English exponent — rendered "certain ones," no marker, no conjecture, flagged so a later sweep does not churn it.
11. 1277B · Primo itaque in ferri contritionem damnamus mala. Accusative where the parallel next sentence has ablative In testae … contritione. No English exponent (Pattern 9 family); rendered "in the crushing of the iron." Logged, not marked.

D. Punctuation carried against the English ear (Pattern 8)

Each of these is a mark the plate prints that an English editor would remove or replace, and each is preserved. They are listed so that a later reader does not "fix" them.

1. 1274D · per aurum potest intelligi arrogantia. per argentum jactantia, per aes pertinacia… — a full stop in the middle of a five-item list, followed by a lower-case continuation. Carried exactly: "…by gold arrogance can be understood. by silver boastfulness, by bronze obstinacy…"
2. 1279B, 1279C, 1279C · three consecutive Nonne … sentences closed with a PERIOD, not a question mark (…cooperuit me. · …non confundar in aeternum. · …propitiatur omnibus iniquitatibus tuis.). All three end with a period in the English. This is the single largest punctuation-fidelity decision in the range; the ear wants three question marks and the plate gives none.
3. 1281C · Quid enim habes quod non accepisti, nihil enim aliud est… — I Cor. 4:7 printed with a comma, not a question mark. Carried: "For what hast thou that thou hast not received, for to have such power…"
4. 1282A · Quid autem cum et illa suppetunt quae sufficiunt ad perfectionem. — period. Carried. The next sentence (…et sub ditione tua universa constituit?) does print a question mark and keeps it.
5. 1283A · Nonne minus est bonum facere ex sola deliberatione, quam etiam ex magna dilectione seu delectatione. — period after a Nonne. Carried.
6. 1277C · In argenti igitur contritione boni operis imperfectio. — a verbless sentence closed with a full stop; the verb (arguitur) arrives only in the next sentence. Rendered verbless and stopped where the plate stops.
7. 1279A · …de omnibus quae fiunt sub sole [n: (Eccle. I.)] Ecce area…* — no terminal stop at all after the Ecclesiastes lemma. None supplied.
8. 1279B · sed, flante vento, favillae, tolluntur — a stray comma between subject and verb. ⚠ This one I did NOT carry. Latin and English comma conventions are not 1:1 anywhere in this corpus (our English re-commas subordinate clauses throughout), so a preserved subject–verb comma would read as our typo rather than Migne's, which is the opposite of what Pattern 8 exists to achieve. Rendered "the ashes are taken away." Recorded here so the decision is visible; if the project wants internal commas carried 1:1, this is the site to start from.

Question-mark parity: 0020 9/9 · 0021 8/8 · 0022 10/10 · 0023 2/2 (body) · 0024 11/11. Exclamation marks: none in the Latin, none supplied. Guillemets: none in the Latin, none supplied.

E. Vocabulary — the fixed table, and three calls I want ratified

Followed against my own judgement, per the brief:

• animus = spirit / mens = mind, held visibly apart everywhere. The site the brief named — 1281A, in uno eodemque animo quasi unus alium consulit, quando illud quod per unum mentis statum non potuit — reads "in one and the same spirit the one, as it were, consults the other, when what he could not attain through one state of mind he learns through another." The other animus in the range is 1282D, magnamque animi devotionem → "a great devotion of spirit." anima is "soul" throughout (1274C, 1279C, 1279D); spiritus is "spirit" (1278C–1279A, and afflictio spiritus). No animus / spiritus collision occurs in this range — the brief's two flagged sentences are in 0005 and 0034, neither of which I hold.
• detrimentum = decline at 1279D (contra similium corruptionum detrimenta), against a strong pull toward "damage" or "loss" in that sentence.
• gradus = step throughout, including the two chapter heads that use it, where "degree" read more naturally.
• testa = clay and aes = bronze throughout, including plantam testeam → "the clay sole."
• industria = effort at 1276B (quaelibet humana industria).
• praesumptio = presumption: at 1279C, per veniae praesumptionem is rendered "through the presuming upon pardon" rather than "presumption of pardon," because here the word is positive (confident reliance) and the English noun alone would have made Richard say the opposite of his sentence. The root word is preserved; flagging the wording in case a peer prefers strict "presumption."

Three calls I want ratified for the whole work:

1. ⭐ simulatio vs simultas. Both occur in this range and they must not collapse. 1274A/1283B print simulatio ("the good that is done ex simulatione"); 1283B and 1283D print simultas in the rhyming triad charitas · veritas · simultas, where the sense demanded is plainly the same as simulatio, not the classical "grudge, enmity." I have rendered simulatio = "dissimulation", simultas = "pretense", and ficta / simulatum = "feigned" / "simulated". Rendering simultas as "enmity" would have been classically correct and would have broken the argument (bronze = pertinacia and feigned piety everywhere else in the work). Please ratify or overrule before other agents meet simultas.
2. intelligentia = "understanding" at 1278C (spiritalis intelligentia → "spiritual understanding"). ⚠ The fixed table assigns understanding to intellectus. intellectus does not occur anywhere in 0020–0024, so no collision arises in my files, but it will elsewhere. If the work needs them distinguished, intelligentia → "intelligence" is the natural fallback and this is the site to change.
3. theoria. 1282B prints bonis meditationibus spiritualibusque theoriis. "Theories" is actively misleading in modern English and "contemplations" is already spoken for by contemplatio, which occurs three times in the same chunk (1280B, 1280D). Rendered as italic Latin, theoriae, under the register rule for technical words with no clean English equivalent. Ratify or replace.

F. One structural note

1281C · the split lemma with the inquit inside it. Migne prints Gloria, inquit, nostra haec est… — the catchword, the speech verb, then the rest. English word order cannot put "Glory" alone in the first span and still construe, so the first span carries "Our glory" and the second "is this, the testimony of our conscience." The same one-word shift was needed at 1279A (Quem ergo, ait Apostolus, fructum habuistis… → "What fruit therefore, says the Apostle, had you…"). In both cases the number of spans, the position of the inquit inside the quotation, and the [n:] position are unchanged (Pattern 5). Logged as a class, since every chunk of this work will meet it.

Richard of St Victor, De eruditione hominis interioris, Book I, capp. XXXII–XL. Range closes with cap. XL; ## LIBER SECUNDUS. does not fall in this range (my heads frontmatter carries only CAPUT XXXII–XL, all Book One).

Each entry: column · printed reading · what I did · why.

A. Markers fired

A1. [var:] — Migne's scripture against the received text

1286A / 1288A (0026) · Dan. 2:43 · commiscebuntur quidem humano semini Vulgate: commiscebuntur quidem humano semine (ablative). Migne prints the dative semini. Rendered "shall be mingled to human seed"; [var: Vulg. humano semine, ablative, "mingled with human seed"; Migne's dative governs the gloss that follows] placed after the lemma. Why it is material and not cosmetic: Richard's gloss immediately takes up the dative — Ecce quomodo adhaereat humano semini (1288A) — i.e. carnal prudence cleaves to human seed rather than being mixed with it. The exposition rests on the case ending. Rendering the Vulgate ablative here would have detached the gloss from its own lemma: 7a″ exactly. Marker placed after the closing italic rather than inside it, because nested italics are not expressible inside a marker (Pattern 10 note).

1288C (0027) · Rom. 5:20 / Matt. 24:12 · ubi abundavit iniquitas, superabundavit et gratia, [n: (Matth. XXIV)] Vulgate Rom. 5:20: ubi autem abundavit delictum, superabundavit gratia. Migne prints iniquitas — the noun of Matt. 24:12 (abundavit iniquitas) — and his own note points at Matthew, so the plate is quoting a conflation and citing the wrong half of it. Rendered as printed ("where iniquity abounded"); [var: Vulg. Rom. 5:20 reads delictum, "offence", where Migne's noun is that of Matt. 24:12, to which his own note points]. Material because iniquitas is the chapter's hook word: the head is de profundo iniquitatis ad summum fastigium charitatis, and the sentence four lines later repeats de profundo iniquitatis provehit. "Offence" would have cut the proof-text loose from the argument it proves. Note reproduced verbatim per the [n:] rule; the mis-citation is Migne's and is not corrected.

1288D (0027) · Dan. 2:44 · et regnum ejus populo alteri non dabitur Vulgate: et regnum ejus alteri populo non tradetur. Migne prints dabitur (and inverts alteri populo). [var: Vulg. non tradetur, "shall not be delivered over"]. Marked rather than merely logged because the [cj:] at 1289D hangs on this verb — the reader who meets the conjecture there needs to know that the received text has a different verb in the lemma. Also dropped in Migne: autem twice (2:44), unmarked.

1290B (0027) · Dan. 2:45 · ostendit regi quae futura sunt postea Vulgate: quae ventura sunt postea. [var: Vulg. ventura, "what shall come to pass"]. Material because the gloss picks the printed participle straight up: ut cum adhuc futura est, dum adhuc praecaveri potest, sua homini ruina reveletur. The whole of cap. XXXV is about foresight of what is to be, not of what comes.

1292C (0028) · Isa. 42:8 · nec laudem meam sculptili Vulgate: et laudem meam sculptilibus (plural). Migne prints nec for et and a singular sculptili. Rendered "nor my praise to a graven thing"; [var: Vulg. et laudem meam sculptilibus, "and my praise to graven images", plural]. The number is worth marking here because cap. XXXVIII, three sentences on, argues from plural simulacra gentium and simulacra corde gestare: the plate's singular is the odd man in its own context.

1295B (0029) · Isa. 40:6 · omnis gloria ejus quasi flos feni Vulgate: quasi flos agri. [var: Vulg. flos agri, "the flower of the field"]. ⭐ The strongest divergence in the range. Richard adduces Isa. 40:6 solely to license the equation fenum (Ps. 105:20's calf-fodder) = carnal glory: Quid aliud erit vitulus fenum comedens, nisi humani favoris appetitus, carnalem et temporalem gloriam ambiens? With the Vulgate's flos agri the second fenum is simply not in the verse and the syllogism has no middle term. The printed divergence is what makes the argument work; conforming it would have quietly destroyed it.

A2. [cj:] — printed word real, faithful English false (Pattern 18, negation/correlative bar only)

1289D (0027) · O felix regnum quod nunquam dissipabitur, quando populo alteri in aeternum dabitur! The exclamation praises the kingdom for the very thing the lemma denies. Migne's own lemma at 1288D reads regnum ejus populo alteri non dabitur, and his gloss at 1289A reads it as exclusivity (fons ille de quo non communicat alienus). The non has fallen out of the exclamation. Rendered as printed, with [cj: dabitur; read non dabitur, "shall not be given"]. Inside Wilson's bar: the wrongly-printed word is a negation (here a dropped one), which the 7561 @0648 precedent explicitly admits ("a dropped word is glossable under this pattern too, because the gloss opens on the word Migne did print"). The guard is satisfied — dabitur stands verbatim in the twin. Alternative construal considered and rejected: that quando is the corruption and the clause is causal. Either way the negation is the thing missing, so the gloss is placed on the negation.

1294B (0029) · sit postmodum in carnibus commodis... videant The sentence is a Quemadmodum enim ... sic correlative: Quemadmodum enim in spiritualibus ... prius vident quae salubriter facere possint, et tamen facere nolint; sit postmodum ... videant quae utiliter facienda sint. With sic the period construes exactly; with sit the apodosis has no correlative and a stranded subjunctive of sum. Rendered "may it be", with [cj: sit; read sic, "so"]. Inside the bar: correlative, the 11064 @1142D si/sic class precisely.

A3. [sic:] — carried type, with corpus controls

1286D (0026) · somno et otio torpore The clause is a list of infinitives — vestiri velle, madere, torpore, irasci, indignari. torpore is an ablative noun stranded among infinitives and has no slot; this is 11632's sternis-for-sterilis shape, not fuges/munde. to grow numb [sic: torpore] with sleep and idleness. 7a⁗ discharged: the marker took the limb's only predicate, so the sense ("to grow numb") stands in plain words beside it. Conjecture: torpere. Control: torpore is common (327 files) — the frequency control does not support this marker, and I say so: it rests entirely on the syntactic-slot test, not on rarity. This is the weakest of my four [sic:].

1287A (0026) · in agnoque labore acquire in [sic: agnoque] labour. Control: 1 file in the whole PL TEI, and it is ours (grep -rlw agnoque sources/pl/tei/ → 11535.xml). A hapax at this site is the plate's defect. Conjecture uncertain — magnoque is the nearest shape and gives "in great labour too", parallel to multo tempore; angore quoque is possible. I do not claim either. 7a⁗ check: the quarantine takes a modifier only; the clause keeps its verb, has no negation, no conjunction. Safe.

1294B (0029) · temporalibis lucris in temporal [sic: temporalibis] gains. Control: 1 file in the PL TEI, ours. Non-word for temporalibus. 7a⁗ discharged — the adjective's sense stands in plain words beside the marker so the phrase is not left headless.

1294D (0029) · et tamem idololatriae simile and yet [sic: tamem] what he practised in deed was like to idolatry. Non-word for tamen. Control is weaker and I flag it as such: tamem occurs in 38 PL TEI files, so it is a recurring compositor/transcription slip across the corpus rather than a hapax. I still marked it, because Pattern 7's non-word class turns on the form not being a word at all, and tamem is attested nowhere as a Latin word; but the corpus-control argument that settled agnoque is not available here. 7a⁗ discharged: the marker swallowed the sentence's adversative, so "yet" stands in plain words beside it — without that, the page would have said Richard thinks soundly and his deed resembles idolatry, with no contrast.

B. Conjectures DECLINED — census candidates, named with reasons

These are all real printed words that make a false or unconstruable claim, and all fall outside Wilson's 2026-08-15 bar (negation or correlative only). I have not widened the bar. Each is a Pattern 18 census candidate.

1. ⭐ 1295A (0029) · inde minus desideriorum turba corrumpatur — the most consequential declined site in the range. The clause as printed says the crowd of desires is less corrupted by the source that illuminates the understanding. Migne's own chapter head, two lines above, says the opposite: CAPUT XL. Quod saepe desiderium corrumpitur unde intelligentia illuminatur. The likely reading is imis, answering in summis in the same sentence (unde in summis intelligentia ... illuminatur, inde in imis desideriorum turba corrumpatur) — a clean summis/imis antithesis that the printed minus destroys. Declined because minus is a comparative adverb, not a negation or a correlative; but it is the founding 11081 munde shape exactly (real adverb, fluent English, false assertion), and if the bar is ever widened this site should be the first one revisited. Rendered as printed: "the crowd of desires is the less corrupted."

1. 1293D (0029) · Quartum quod quasi varium sit — the seven-fold scheme is os · oculi · aures · nares · manus · pedes · guttur (Richard sets it out in that order at 1293A), and the fourth item's content is prudenter discernere, which he has just assigned to the nostrils (ad nares discretio). narium for varium is near-certain. Declined: not a negation or correlative. Rendered "which is as it were a various thing."

1. 1287A (0026) · Quot timores, quot honores, quot algores hostis, noctis et frigoris — the three genitives pair off timores/hostis, ?/noctis, algores/frigoris, so honores is almost certainly horrores. "Honours of the night" is fluent and false. Declined on the bar; rendered "how many honours".

1. 1284D (0025) · imperium, quia negligentius urgere — in a series of quia + object + potes clauses (bonos disponere, malis resistere, rebelles revincere), the fourth has a comparative adverb where the pattern wants an accusative object; negligentiores ("the negligent") would restore it. Rendered literally, "because thou canst urge the more remissly". Declined.

1. 1289D (0027) · nemo uritur! — the exclamation runs through the 1 Cor. 13 list (nemo aemulatur · non agitur perperam · nemo inflatur · nemo ambitiosus quaerit quae sua sunt) and then prints nemo uritur, "no one is burned", where the list's next item is non irritatur. irritatur is the obvious reading. Declined; rendered "no one is burned!" The negation itself is intact, so 7a is not at issue.

1. 1294B (0029) · in carnibus commodis — the parallel is in spiritualibus et animae lucris ‖ in [carnalibus] commodis et temporalibus lucris. carnibus (fem.) cannot agree with commodis as a noun, but commodus is also an adjective, so the phrase does construe — absurdly — as "in advantageous fleshes", and that is how I rendered it. No [sic:], because it renders in place and the absurdity leaves the defect visible (Pattern 12's own test). Conjecture: carnalibus commodis.

1. 1289D (0027) · ad debiti finis talum non perducitur — ad calcem perducere is the idiom, and calcem would be expected. But I decline even the conjecture: this is a treatise on a statue read from its golden head down to its feet of iron and clay, and talus ("ankle-bone", "heel") may well be deliberate. Rendered "to the heel of its due end". Recorded here so a later reader does not "fix" it.

C. Passages that will not construe — rendered as closely as the words allow

1. 1289C (0027) · Si aut in argento illud opus intelligitur quod fit ex sola deliberatione, non autem ex dilectione, non fit tam ex dilectione quam ex veritate, quid tam consentaneum charitati quae congaudet veritati? The protasis contains two clauses that contradict each other (non autem ex dilectione … non fit tam ex dilectione quam ex veritate), and Si aut is itself odd (Si autem?). Both negations carried per 7a, with the result that the English is as self-contradictory as the Latin. No marker fired: nothing here is a non-word, and the defect is not localizable to one word, so no [cj:] guard could be satisfied. Escalated rather than repaired.

1. 1294D (0029) · Ecce quam sane, et de divinis sentit, et tamem idololatriae simile quod per opus exercuit. The second limb has no finite verb (est must be understood) and tamem is carried under [sic:] (A3). Rendered with the ellipsis supplied as plain text, no brackets. The sentence remains loose in English because it is loose in Migne.

1. 1286A (0025) · Caetera autem quae sequuntur utramque lectionem in eamdem sententiam concurrunt — utramque lectionem is a bare accusative with no governing verb (concurrunt is intransitive). Rendered as an accusative of respect, "run together, for both readings, into the same sense." No marker; Pattern 9 territory at worst.

D. Punctuation carried against the ear (Pattern 8) — all verified by parity check

Question-mark and exclamation parity across all five pairs is exact (7/0, 16/1, 6/5, 7/1, 5/2). Three sites are worth naming because the mark is wrong and was kept:

1. 1292A (0028) · …suavis odor latius spargitur Quasi igitur viva, sed bruta caro mactatur… — Migne prints no full stop between the two sentences, and a capital Quasi follows the bare word. The English reproduces the run-on exactly: "…is spread more widely As therefore living but brute flesh is slain…". Not supplied. This is the clearest Pattern 8 datum in the range.
2. 1294A (0029) · Carnis prudentia est ista profecto, et quae prorsus inimica sit Deo? — a flat declarative closed with a question mark. Kept.
3. 1287B (0026) · quanta, quaeso, rixa vel contentio sit frequenter in corde illius… avaritia jubeat reservare. — an unmistakable rhetorical question closed with a period. Kept; no ? supplied. Same class at 1295A (0029): Quis enim nesciat… closes with a period after the 2 Cor. 12:7 quotation. Kept.
4. 1290A (0027) · cum pro qualitate animorum; vel morum… — a semicolon inside a noun phrase. Reproduced.
5. 1290D (0027) · …nubes transierunt [n: (Psal. XVII)], Domini respicere, est illuminare. — a comma, not a stop, after the psalm citation. Reproduced.

E. The pulls I RESISTED — lemmata where Migne agrees with the received text

Positive evidence, per the brief. In each case the familiar English arrived first and the printed Latin was checked word by word before it was trusted.

1. 1289A (0027) · 1 Cor. 13:4–6, the nine-clause Charity passage. Migne agrees with the Vulgate word for word, including congaudet autem veritati. All eight printed non are in the English (envieth not · dealeth not perversely · is not puffed up · is not ambitious · seeketh not her own · is not provoked · thinketh no evil · rejoiceth not upon iniquity), and none is added. (The launch note said nine; the span prints eight non, and nunquam excidit at 1289A/1289D is the ninth negation of the Charity material but sits in a separate italic span.) No marker.
2. 1289A (0027) · 1 Cor. 13:7 omnia suffert, omnia credit, omnia sperat, omnia sustinet — exact. No marker.
3. 1289B (0027) · Rom. 5:5 Charitas Dei diffusa est in cordibus nostris… — exact.
4. 1290D (0027) · Ps. 17:13 Prae fulgore in conspectu ejus nubes transierunt — exact, truncated before grando et carbones ignis. Truncation is not a variant.
5. 1293A (0028) · Ps. 113:13–15, the idols passage. Migne agrees with the Vulgate throughout. All seven printed non carried (shall not speak · shall not see · shall not hear · shall not smell · shall not feel · shall not walk · neither shall they cry out); none added. I resisted the pull to render nares… non odorabunt as Douay's "noses" — "nostrils" is kept because the gloss at 1293A assigns nares the office of discretio, spiritual smelling.
6. 1293D (0029) · Ps. 113:16 Similes illis fiant qui faciunt ea… — exact.
7. 1291A (0028) · Dan. 2:46 Tunc rex Nabuchodonosor cecidit in faciem suam… — exact. I resisted "offer sacrifice" for ut sacrificarent ei and kept the printed word order (hostias, et incensum praecepit).
8. 1284A (0025) · Dan. 2:37–38 — see F1 below. Deus coeli is the Vulgate reading and I did not conform it to the Dominus coeli printed in chunk 0023.
9. 1287D (0026) · Dan. 2:41–42 and 1288D (0027) · Dan. 2:44–45 — the standing pull here is to restore pedum after digitos and regnum before erit solidum from the Vulgate. Both are absent in Migne and both stay absent in the English.
10. 1292C (0028) Qui gloriatur, in Domino glorietur (1 Cor. 1:31) and Maledictus homo qui confidit in homine (Jer. 17:5) — both exact, both left as printed.

F. Divergences noted and deliberately NOT marked (sense unchanged)

Recorded so a later sweep does not read the absence of a [var:] as agreement.

1. ⭐ 1284A (0025) · Dan. 2:37 · Deus coeli — and the three-way comparison the brief asked for. - Chunk 0023 @1281A prints the verse twice in short form: Tu, inquit, Rex regum es and then Tu, inquit, rex regum es, et Dominus coeli, regnum et fortitudinem, et imperium, et gloriam dedit tibi — with Dominus, and with the bestias agri half unitalicized and turned into a question at 1282A. - My chunk 0025 @1284A prints the verse in full and with Deus coeli — which is the Vulgate reading. Chunk 0023's Dominus coeli is the divergent printing, not mine. - The gloss in my chunk hangs on Deus, and twice: quod de Dei munere potes, noli tibi ascribere: Deus enim coeli, regnum et fortitudinem… dedit tibi (1284C) and non est quod tibi ascribas: Deus enim coeli dedit omnia ista tibi (1284D). The whole argument of cap. XXXII is that the power is God's gift, so the divine name in the lemma is the hook. Answer to the brief's question: yes, the gloss hangs on Deus, and here Migne and the Vulgate agree. No [var:] in my range. ⚠ The [var:] question belongs to whoever holds 0023, where Dominus is printed and the gloss there (intellige a quo potestas haec datur) also turns on the divine name. Flagged for the merge. - Secondary, unmarked: Migne prints bestias agri volucresque coeli (accusative) where the Vulgate has bestiae agri: volucres quoque caeli (nominative, a Hebraizing anacoluthon). The accusative tidies the syntax and the Douay renders the Vulgate as though it were accusative anyway; the sense is unchanged and no marker fired. But it is the same divergence in both printings (0023 and 0025), which suggests Migne's copy-text, not a compositor.
2. 1285A–C (0025) — a string of small ones, none material: initium omnis peccati superbia drops the Vulgate's est (Ecclus. 10:15); Qui autem facit peccatum for John 8:34's omnis qui facit peccatum; Tu caput aureum es for Dan. 2:38's Tu es ergo caput aureum; regnum aliud tertium aereum for regnum tertium aliud aereum; sic comminuet omnia haec, et conteret for sic comminuet, et conteret omnia haec.
3. 1286B/1287A/1287B (0026) — Migne prints partem testeam figuli (adjective) for the Vulgate's partem testae figuli (genitive), at all three printings of the lemma; and he varies his own re-quotes between quia vidisti and quod vidisti. The sense is identical either way. Also: [n: (Dan. I)] at 1286B is Migne's own mis-citation — the verse is Dan. 2:41. Reproduced verbatim under the [n:] rule, uncorrected.
4. 1286C (0026) · avaro nihil est scelestius — Ecclus. 10:9 has Avaro autem nihil est scelestius. Particle only.
5. 1286D (0026) · Viscera avari crudelia — Prov. 12:10 reads viscera autem impiorum crudelia. Migne (or Richard) has substituted avari to fit the chapter on avarice. Deliberately unmarked: this is printed as a bare tag with no inquit, no note, no italics, and reads as the author's own adaptation rather than a quotation offered as proof — the 8a′ "rhyme-driven divergence takes no [var:]" logic, applied to an allusion. Recorded here because it is a judgment call and a reasonable reader could mark it.
6. 1287C (0026) · esse sub sentibus delicias computat [n: (Job XXII)] — the phrase is Job 30:7, not Job 22. Migne's citation, reproduced verbatim.
7. 1295A (0029) · [n: (I Cor. XII)] on Ne magnitudo revelationum extollat me — the verse is 2 Cor. 12:7. Migne's citation, reproduced verbatim. The quotation itself drops only the Vulgate's opening Et.
8. 1295B (0029) · Ps. 105:20 · Et mutaverunt gloriam… — the Vulgate has gloriam suam. The possessive is dropped. Left unmarked because the referent is unambiguous from context; but note that Richard's gloss does trade on possession (gloriam suam constituere, illam quam apud Deum habere consueverant gloriam veram), so this is the borderline case of my range and a later sweep may reasonably promote it.
9. 1295A (0029) · Dan. 2:47 — Migne drops et before revelans mysteria and tu before potuisti. Particles only.
10. 1294A (0029) — servus sciens voluntatem domini sui, et non faciens quin potius agens indigna, plagis vapulabit multis is a loose weave of Luke 12:47–48, not a quotation; printed roman, no note. Rendered as printed, unmarked.

G. Fixed vocabulary — where I followed the table, and one tension

• mens = mind · animus = spirit · anima = soul, kept visibly distinct throughout. They collide inside single clauses twice in my range and both were handled consciously:
• 1287C (0026) cum praevaricatrix mens primo ad saecularia desideria revertitur, ad tolerandos mundi labores animus obduratur → "when the transgressing mind first turns back … the spirit is hardened".
• 1288B (0026) dum interna animi anxietate … desaeviunt → "by the inward anxiety of the spirit". Neither of the two animus/spiritus collision sentences named in the brief (0005, 0034) falls in my range; spiritus occurs here only as Spiritus sanctus (1289B), so no collision arose.
• aes = bronze throughout, including simulationis aes (1289C) and aeream pertinaciam (1289C). testa = clay; lutum rendered mire to keep the pair distinct in ferrum mistum testae ex luto ("iron mingled with clay of mire") and in Quasi enim lutum in testam duratur ("mire is hardened into clay"). Flagging this for the merge as a work-level call I would like ratified: the table fixes testa but is silent on lutum, and the two words are glossed against each other here.
• gradus = step (1284A head corruptionis gradibus → "steps of corruption"; 1288C quibusdam profectuum gradibus → "certain steps of advance").
• virtus = virtue · gratia = grace · devotio = devotion · discretio = discretion · industria = effort (1292C in aliqua industria → "in any effort") · praesumere = to presume · obduratio/obdurare = hardening / to harden · superbia = pride.
• Tension logged, not decided: the four-fold vice scheme of cap. XXXII needs four fixed English words and the table names none of them. I used arrogance (arrogantia) · boasting (jactantia) · obstinacy (pertinacia) · savagery (saevitia), and held them through cap. XXXIV where all four recur against the four metals. These four recur across the whole of Book I; if a peer agent has fixed different words, mine should yield to whichever came first — but the set must be uniform, because Richard's argument at 1289C is a point-by-point matching of the four vices to the four clauses of 1 Cor. 13.
• detrimentum does not occur in my range.

H. Structural notes (not defects)

• 1293C–1294A (0029): the chapter promises septem ista and enumerates only Primum–Sextum. The seventh is the unnumbered clause that follows — ad omnia haec divinum auxilium implorare debemus — matching ad guttur supplicatio in the list of seven offices at 1293A. Not a lacuna; nothing supplied.
• 1288C (0027) eductas de lacu miseriae — feminine accusative plural with no expressed antecedent (animas understood). Pattern 9: gender has no English exponent, no marker, no churn.
• Chapter numerals follow the plate exactly (XXXII–XL, arabic nowhere). No head printed without its closing period in this range.
• Em-dashes: zero in all five chunks.

Richard of St Victor, De eruditione hominis interioris. Range spans the end of Book One (capp. XLI–XLIV, closing at 1298D with Amen.), the head ## LIBER SECUNDUS. which stands at the foot of 0030 with no text after it, and then Book Two, capp. I–X. The running lemma changes inside the range: Dan. 2:47–49 finishes Book One, and Book Two opens at Dan. 4:1 with the dream of the tree.

Each entry: column · printed reading · what I did · why.

A. Markers fired

A1. [var:] — Migne's scripture against the received text (11 sites)

1296D (0030) · Quis novit sensum Domini, aut quis consiliarius ejus fuit? [n: (Isai. XL.)] Rendered as printed ("Who hath known the sense of the Lord, or who hath been his counsellor?"). [var: Vulg. Isa. XL:13 opens Quis adjuvit spiritum Domini, "Who hath helped the spirit of the Lord"; the wording printed here is that of Rom. XI:34, Quis enim cognovit sensum Domini]. Material: the plate cites Isaiah and prints Paul. Isa. 40:13 in the Vulgate asks who helped the spirit of the Lord; the clause Richard needs — who has known the Lord's sensus — exists only in Rom. 11:34. The note is Migne's and is reproduced unchanged. ⚠ Paired site: 1305A (0032) prints the same verse as Quis cognovit sensum Domini, again under [n: (Isai. XL)]. The two printings differ from each other (novit / cognovit) and both differ from Isaiah, so both are marked, each var naming the other. This is the "Migne re-quotes his own lemma against himself" habit the brief names, here compounded by a mis-citation that is constant across both printings.

1302B (0032) · Job 29:16 · Causam quam nesciebam, diligenter investigabam Vulgate: causam quam nesciebam diligentissime investigabam. [var: Vulg. diligentissime, "most diligently"]. Material because the proof-text is adduced for the superlative: the sentence that introduces it is Ecce quam diligenter, ecce quam instanter quaerit quod nescit. Migne's positive flattens exactly the degree the argument is proving.

1303C (0032) · Ps. 118:18 · Illumina oculos meos, et considerabo mirabilia de lege tua [var: Vulg. Ps. CXVIII:18 Revela oculos meos, "Uncover mine eyes"; this work prints Revela at 1253A and at 1262C]. The work is its own control here, and it settles the case. The same verse, with the same [n: (Psal. CXVIII)], is printed three times in this work: Revela at 1253A (0009) and 1262C (0014), Illumina here. Two printings agree with the Vulgate and one does not, so the divergence is local to this setting — Illumina oculos meos is the opening of Ps. 12:4, i.e. a contamination from a neighbouring verse, not a variant copy-text running through the work. Rendered "Enlighten mine eyes." ⚑ Merge note for the reconciliation step: the two earlier printings were Englished "Unveil mine eyes" (0009) and "Reveal mine eyes" (0014). Those two render the same Latin two ways and should be reconciled to one form; mine is deliberately different because the Latin is different, and must not be swept into agreement with them.

1304A (0032) · Ps. 115 · Vota mea reddam in conspectu omnis populi ejus [var: Vulg. Ps. CXV:14 Vota mea Domino reddam coram omni populo ejus, "I will pay my vows to the Lord before all his people"]. Migne drops Domino and prints in conspectu omnis populi for coram omni populo — the wording of Ps. 21:26 (vota mea reddam in conspectu timentium eum) bleeding into Ps. 115. Rendered as printed. Material because the addressee (Domino) is the thing the sentence is about: Richard adduces it as the exsecutio of the vow sworn to God.

1305A (0032) · Isa. XL / Rom. XI:34 — see the paired entry at 1296D above.

1305A–B (0032) · 2 Pet. 1:20 · quia prophetia Scripturae propria interpretatione non fit [var: Vulg. quod omnis prophetia Scripturae, "that no prophecy of Scripture"]. Migne drops omnis. Rendered as printed: "that prophecy of Scripture is not made by private interpretation." Material and it is a scope change, not a stylistic one: with omnis … non the Vulgate says no prophecy is so made — a universal negative, which is exactly the universal restraint Richard is citing it for (quomodo pariter omnes ab ejusmodi praesumptione compescat). Without omnis the printed sentence is a weaker, unquantified claim. Per 7a the negation printed (non) is carried and nothing is supplied; the quantifier's absence is recorded in the marker, not repaired in the text.

1305B–C (0033) · Dan. 4:5 · qui habet spiritum sanctorum in semetipso [var: Vulg. Dan. IV:5 has deorum before sanctorum, "the spirit of the holy gods"; Richard restores the word in both of his own re-quotations below]. The strongest site in the range, and it is a within-plate diff. The lemma as printed drops deorum; then the very next sentence glosses it as Ecce spiritum deorum sanctorum habere dicitur (1305C), and the re-quotation at 1305D–1306A prints Qui habet spiritum deorum sanctorum in semetipso. The whole chapter turns on the word: Richard's argument is that dii sancti excludes omnes dii gentium, daemonia, and that the sancti here are men called gods (Ego dixi: Dii estis). Rendered as printed ("the spirit of the holy ones") with the divergence marked; the two re-quotations are rendered with "gods", because that is what they print. Only the lemma is marked, per the brief's rule that the divergent printing takes the marker and not both.

1305D (0033) · Rom. 8:26 · qui interpellat pro nobis gemitibus inenarrabilibus [n: (Rom. VIII)] [var: Vulg. Rom. VIII:26 reads postulat, "asketh"; interpellat, "maketh intercession", is the verb of Rom. VIII:34]. A conflation of Rom. 8:26 and 8:34 under a chapter-level note that is correct for both. Marked although the quotation is not in italics — Pattern 14 has no formatting precondition, and this is a proof-text carrying its own [n:].

1308D (0034) · 1 John 3:2 · Fratres, filii Dei sumus, et nondum apparet quid erimus [var: Vulg. I John III:2 opens Carissimi, nunc, "Dearly beloved, now"; and it reads apparuit, "hath appeared", for the present apparet]. Three divergences in eight words: the vocative is changed, nunc is dropped, and the perfect becomes a present. Rendered as printed.

1309A (0034) · 1 Cor. 6:17 · Qui adhaeret Deo, unus spiritus est [var: Vulg. I Cor. VI:17 reads Domino, "to the Lord", where the plate prints Deo]. Material because the sentence it proves says Deo: quae animum humanum Deo conjungit, et cum eo quasi unum spiritum efficit. The proof-text has been assimilated to the clause it proves. Rendered as printed. (See §C1 for the animus/spiritus handling in this same sentence.)

1309C and 1310A (0034) · Num. 24:4, printed twice and divergently First printing (1309C, in italics, with [n: (Num. XXIV)]): qui visionem omnipotentis Dei intuitus est … et sic aperientur oculi ejus → [var: Vulg. Num. XXIV:4 has no Dei after omnipotentis, and reads aperiuntur, "are opened", for the future aperientur]. Second printing (1310A, unitalicized, closing the chapter): qui visionem Omnipotentis videt, qui cadit, et sic aperiuntur oculi ejus → [var: Vulg. Num. XXIV:4 has the perfect, "hath beheld", where this second printing prints the present videt, "seeth"; the printing at 1309C keeps the perfect]. Both are marked because both diverge, in different places — the brief's exception to "mark one, not both". Note that the divergence is load-bearing in one direction: the gloss between them reads quod cadendo oculi aperiuntur (present), which agrees with the second printing and against the first printing's future aperientur. Both are rendered exactly as printed, the first with "shall be opened" and the second with "are opened", so the reader can see the plate disagreeing with itself across five lines.

A2. [sic:] — carried type (3 sites), each checked against the whole PL TEI

1298C (0030) · quia perfores aditus ad interiora patet Carried: through the doors [sic: perfores] the entrance to the inner parts lies open. Control run: perfores occurs in exactly two files in the 87M-word PL TEI — 11535 (here) and 7132. It is a real form (2 sg. subj. of perforare, "thou mayest bore through") and therefore not a non-word, but it has no grammatical slot: the clause needs per fores, "through the doors", which is what the whole passage is about (Daniel … in regis foribus observat, Danielem regis foribus praesidere, civitatis nostrae foribus praesidentem — fores four times in fifteen lines). This is Pattern 10's fused direction, and Pattern 12's "renders only by hiding the defect" test. 7a⁗ compliance: the marker swallows a preposition and its object, so the sense is supplied in plain words beside it ("through the doors"). Without that the clause would have lost both.

1303A (0032) · multam ibi moram invectere Carried: weaves in [sic: invectere] much delay there. Control run: invectere occurs in one file in the whole PL TEI — this one. A form unique to this plate at this site is the plate's defect (the standard this work has been using). Conjecture: innectere, "to weave in / entangle" — moram innectere is idiomatic and gives exactly the required sense of protracted, unnoticed delay. 7a⁗ compliance: the marked run is the only infinitive of the clause (antequam advertat quid intendat, multam ibi moram invectere), so the English supplies the verb in plain words beside the marker. The conjecture stays here, not in the text.

1303B (0032) · quasi per somnium cornere Carried: to discern [sic: cornere] the divine judgments. Control run: cornere occurs in one file in the whole PL TEI — this one. Conjecture cernere, near-certain: the antithesis is aliud est … cornere, et aliud … perpendere … dijudicare, and the parallel passage the sentence points back to (sicut in alterius visionis expositione jam diximus) uses seeing-verbs throughout. 7a⁗ compliance: again the marked run is the limb's only verb, so "to discern" stands in plain words beside the marker.

A3. [cj:] — one site, and it is offered for ratification, not asserted

1302B (0032) · quaerit quod nescit, nisi prius desistit, donec somnii sui solutionem inveniat Rendered: he seeks what he knows not, unless [cj: nisi; read nec, "nor does he first desist"] he first desists, until he finds the solution of his dream which he seeks. The printed nisi is a real word and renders in place, so Pattern 12 has nothing to wrap; and the English it yields asserts the opposite of the sentence's own point, which is that Nabuchodonosor did not leave off (Ecce quam diligenter, ecce quam instanter quaerit, and at 1302A videamus quam instanter Nabuchodonosor quaesivit quod nescivit). The idiom wanted is nec prius desistit donec … inveniat. Why I fired it rather than declining it: the word that went wrong is a negation (nec), which is the class Wilson's 2026-08-15 ruling explicitly brought inside the bar even where the sentence is incoherent rather than merely false — the 7561 @0648 solum site is the same shape (a printed word that survives, a negation that fell out). The verbatim-in-the-twin guard is satisfied by nisi. ⚠ If the bar is read narrowly — that nisi is itself in 7a's negation list and so is a substitution rather than a loss — this site should be demoted to a declined census candidate. It is the only [cj:] in the range and nothing else depends on it.

B. Divergences RECORDED and deliberately NOT marked — the pulls I resisted

Every italic span named in the launch message was worked against the received text. The list below is the residue: places where Migne and the Vulgate differ, or where my ear wanted a familiar English, and no marker fired. Positive evidence, per the brief.

1. 1296B (0030) · Rom. 13:1 · Non est potestas nisi a Deo — Vulgate has Non est enim potestas. A connective dropped when a clause is lifted out of its sentence is not a divergence of wording; no marker. Both negations carried.
2. 1296D (0030) · 1 Cor. 2:11, woven unitalicized into Teste namque Apostolo. Richard reshapes quis enim scit hominum quae sint hominis into nemo novit quae sunt hominis, i.e. paraphrase in his own voice, not a quotation. No marker; and note that his spiritus hominis is rendered "spirit", the same word as animus, per the fixed table.
3. 1297A (0030) · Dan. 2:48 · munera magna, et multa dedit ei — Vulgate munera multa et magna. Order reversed. Not marked: no sense turns on it, and Richard's own gloss keeps the plate's order (munera magna et multa devotionis nostrae Daniel accipit), so the plate is internally consistent.
4. 1297D (0030) · Dan. 2:49 · Sidrac against the Vulgate's Sidrach — and the exposition twelve lines later prints Sidrach. Orthography of a proper name, Pattern 9; both are rendered "Sidrac"/"Sidrach" as printed, no marker.
5. 1297C (0030) · 1 Thess. 5:17 · Sine intermissione orate! — agrees with the Vulgate word for word. The exclamation mark is Migne's and is kept (Pattern 8); the English keeps his word order too ("Without ceasing pray!"), which is also what holds the [1297C] anchor in its printed position inside the lemma.
6. 1299B (0031) · 2 Cor. 1:12 · Gloria nostra, haec est testimonium conscientiae nostrae — Vulgate opens Nam gloria nostra haec est: testimonium…. A dropped Nam and Migne's comma after nostra. No sense change; the comma is reproduced.
7. 1299C–D (0031) · 2 Tim. 4:7–8 · de reliquo reposita est mihi corona justitiae — Vulgate In reliquo. A real variant, but de reliquo and in reliquo both mean "for the rest"; nothing in the gloss touches it. Declined. ⚠ Separately: Migne's note reads [n: (II Tim. VI)] and 2 Timothy has only four chapters. Reproduced verbatim per the [n:] rule; the mis-citation is his.
8. 1300A (0031) · Dan. 4:2 · cogitationes meae in stratu meo — Vulgate in strato meo. Fourth-declension stratus against second-declension stratum; same sense ("upon my bed"), and Richard's gloss (Quid hic intelligitur per stratum) takes the accusative, which is compatible with either. Declined.
9. 1301A (0031) · Eccl. 9:12 — Migne drops the second sicut (et aves laqueo comprehenduntur for et sicut aves). Declined: a dropped correlative particle inside a simile with the first sicut still standing. This one I looked at twice, because a dropped correlative is inside Pattern 18's class; but the correlative here is the second half of a doubled simile, not of a negation or an antithesis, and the English ("as fishes… and birds…") carries the comparison without asserting anything false. Named here as a declined site.
10. 1300C (0031) · Ps. 35:7 · judicia Domini abyssus multa — Vulgate judicia tua abyssus multa. Not marked: the phrase is inside Richard's reported clause (Viderat utique Propheta, quia judicia Domini abyssus multa), where the second person of the psalm could not stand; converting tua to Domini is the grammar of reporting, not a divergent copy-text.
11. 1302D (0032) · Ps. 76:6–7 — Migne drops the Et that opens v. 7. Declined. scopebam spiritum meum is rendered "I swept my spirit", which is the Vulgate's own odd verb and the traditional English; this is an "already-literal is not an error" case, not a conformation.
12. 1306A (0033) · Eccl. 1:17 · Dedi cor meum — Vulgate Dedique. The -que has no place once inquit is inserted after the first word. Declined.
13. 1308C (0034) · John 17:1 · Pater, clarifica Filium tuum — the Vulgate's venit hora between vocative and imperative is simply not quoted. An ellipsis in a citation is not a divergence of wording; declined. (Had I marked it, every abbreviated proof-text in the work would need one.)
14. 1309A (0034) · Rom. 5:5 · Charitas autem Dei diffusa est — Vulgate quia caritas. autem for quia is the connective the host sentence needs. Declined.
15. 1306A (0033) · 1 Cor. 2:15 · Spiritualis autem dijudicat omnia — Vulgate iudicat. Unitalicized, uncited, and Richard's own compound dijudicare is a fixed term of this work (it stands in the chapter head). Declined as authorial usage, not plate divergence.
16. 1308C (0034) · Heb. 1:5 / Ps. 2:7 · Filius meus es tu, ego hodie genui te — agrees with both. Recorded because the pull to write "Thou art my Son, today I have begotten thee" in the modern cadence was real; the printed Latin supports the traditional wording exactly and it stands.
17. 1308D (0034) · Matt. 6:9 · Pater noster qui es in coelis — agrees. The pull was to continue the prayer; the plate stops there and so does the English (the same discipline as L3's chant cues).

C. Vocabulary: the collisions this range was flagged for

C1. animus / spiritus at 1309A (0034) — the second of the work's two collisions

Printed: quanta sit virtus verae devotionis … quae animum humanum Deo conjungit, et cum eo quasi unum spiritum efficit. Qui enim adhaeret Deo, unus spiritus est. The same collision stands in the chapter head (quae animum humanum Deo conjungit).

What I did: rendered both "spirit" — "which joins the human spirit to God, and makes it with him as it were one spirit." The chapter head follows suit.

Why, and why my site differs from 0005's. The 0005 agent met ex animi impetu beside ex spiritus sui impetu and ruled both "spirit" because Richard's eodem/eamdem there assert the two are the same impulse, so differentiating would silently deny an identity the text states. Here the logic runs the other way: the two words name two different things — a human animus and the spiritus that results — which are joined into one. So the argument for one English word is not identity but the conclusion: the sentence's whole force is that what is joined becomes unus spiritus, and it is 1 Cor. 6:17 that licenses the claim. If animus were rendered "mind" here (as 11536 would have), the proof-text would be about spirit and the claim about mind, and the unum spiritum efficit would have nothing to attach to. The union is the point, and one English word is what carries it. ⚠ Note that the same sentence is where the plate has assimilated the proof-text to the claim (Deo for Domino, §A1 above). The two observations reinforce each other: this is a sentence in which quotation and gloss have been drawn together, and both moves are recorded rather than smoothed.

C2. animus / mens — kept visibly distinct at all 14 sites in the range

The range holds chunk 0030's chapter head, the sentence the brief names as forcing the ruling: Quod quidquid animum per cogitationem pulsat, mentis devotionem excitare debeat → "That whatever strikes the spirit through thought ought to arouse the devotion of the mind." The body repeats the collision verbatim at 1298C (omne quod animum per cogitationem pulsat … mentis devotionem excitat) and again at 1298D (devotio mentis), and 0031 has quando eam animus jam ad manum habet immediately after quae mentis arbitrio subjacet (1301D). Every one is rendered spirit / mind. No site in the range resisted the distinction; the head and its own body sentence are the proof that the plate is using the two words deliberately, three lines apart, of the same act.

C3. sacramentum — I followed the ratified table against a shipped neighbour

Dan. 2:47's quia potuisti aperire sacramentum hoc falls twice in 0030 (1296B, 1296D) and is rendered "this sacrament", per ratified item 5 (sacramentum = sacrament, kept apart from mysterium = mystery). ⚠ This disagrees with the immediately preceding chunk. 0029 renders the same phrase "this secret". The rest of the work is with me: 0008 twice ("concerning this sacrament", "calls that matter a sacrament"), 0011 ("the divine sacraments and mysteries"), and 0014 twice — where the very same Dan. 2 phrase sacramentum hoc is already Englished "this sacrament". So 0029 is a single drift against five sites plus the ruling, and it is a merge-step fix in 0029, not in my chunks. Flagged here rather than acted on because 0029 is committed and outside my range.

C4. Other fixed-vocabulary calls followed, one of them against my ear

• sensus = sense throughout, including inside the quotation Quis novit sensum Domini (1296D, 1305A), where every English Bible says "the mind of the Lord". Followed the table against my own judgment, and it is the right call here for a reason beyond consistency: Richard immediately glosses the verse with nemo novit quae sunt Dei nisi Spiritus Dei, and the range's other sensus sites — ex intimo animae sensu (1295D–1296A), intellectualis sensus (the head of Book II cap. II, and 1300D, 1309C), sensu exteriori / interiori sensu (1302C) — are one continuous thread about the faculty of sense. Rendering the lemma "mind" would have imported mens into a sentence that does not contain it, in a work whose whole seam is keeping those apart.
• scientia = learning (1305B ad divinorum judiciorum scientiam), cognitio = knowledge (1301C ad plenam sui cognitionem) — they stand four lines apart in 0031.
• enodatio = unravelling (heads of capp. V and VI) vs solutio = solution (the Dan. 4 lemma throughout 0032) — they stand in the same sentence at 1304C.
• magisterium = mastery, at 1297B (magisterii arcem), 1298B twice, and 1305B (sine ejus magisterio, qui eas inspiravit) — the last is the head-word of cap. VI.
• mentis excessus = excess of mind (1300A, 1300C — the only two in the range).
• intelligentia = understanding (1297A, 1302C, 1303C, 1307C, 1307D); intellectualis = intellectual (1300D, 1309C). No intellectus in the range.
• conjectura/conjicere = conjecture (1296B), kept apart from interpretatio.
• arbitrium alone = "the will", after 0023's shipped arbitrii potestas / arbitrii libertas → "the power of the will" / "the liberty of the will"; liberum arbitrium = free will (1301D).
• Proper names in Migne's Latin form: Nabuchodonosor, Balthassar, Sidrac / Sidrach, Misach, Abdenago, Balaam, Urias. Balthassar matters twice over in 0034, where the whole chapter is the etymology capillus capitis; the Latin name is kept beside the translated gloss ("Balthassar is interpreted hair of the head") so the gloss still explains something.
• Babylon … confusio (1297D) — etymology gloss, translated ("interpreted confusion") with the name it explains standing beside it.

D. Passages that would not construe, and other plate observations

1305C (0033) · Sed profecto qui hoc loco designantur. Spiritum sanctum habere creduntur; — Migne's full stop falls inside the sentence, stranding a relative clause with no main verb and leaving creduntur without its subject. Rendered exactly so: "But assuredly they who are designated in this place. They are believed to have the Holy Spirit;". Pattern 8: the mark is not deleted and nothing is supplied. The sense the two fragments carry between them is plain and is not in doubt; only the punctuation is.

1309A (0034) · vide quam recte Balthassar per effectum dicitur. Per quem quilibet in filiorum libertatem adoptatur. — the same shape: a full stop breaking a relative clause off from its antecedent. Kept, and the English breaks in the same place.

1300A (0031) · quia per somnium corporis sensus corporeus sopitur — somnium ("dream") stands where the sense wants somnum ("sleep"): it is sleep that lulls the bodily sense, and the very next clause distinguishes them (Prius dormimus quam somnium videamus), while 1300C of this same chunk prints post excussum somnum. Rendered "dream" per the fixed table and Pattern 7 — "Because through the dream the bodily sense of the body is lulled" — awkward, and deliberately so. ⭐ The work's own control settles it as authorial usage, not a defect, and I record the control because I nearly called it one. 0014 argues the identical point in the identical words — Quia per somnium sopitur sensus corporeus, recte per illum mentis excessum intelligimus — so Richard uses somnium of sleep in both places, and the shipped 0014 English renders it "by the dream" exactly as mine does. No marker, no conjecture, and NOT a census candidate. Two agents independently reaching the same rendering on the same construction is the outcome the fixed table exists to produce.

1302C (0032) · quae sensu exteriori percepi possunt — percepi for percipi. Control run: percepi possunt occurs elsewhere in the PL TEI, and percepi itself in five-plus files, so the e/i interchange in the passive infinitive is an attested medieval orthography, not this plate's defect. Pattern 9: rendered normally ("can be perceived"), no italics, no marker.

1302C (0032) · Illa autem velut intus sunt qua: nisi interiori sensu… — the token qua: is quae: it is the ligature quæ misread by the digitization, and qua: occurs in 16 files across the PL TEI, which makes it an OCR artifact class of our source rather than anything Migne set. No [sic:] — a defect in our files is never attributed to his plate. Rendered "which". Recorded here so a later pass can decide whether it belongs in data/tei-patches/.

1298B (0030) · vel communi infirmitati deseruimus — with a dative and in a present- tense series (consulimus … deseruimus), this is deservimus, "we serve", with u for v — the ordinary orthography, Pattern 9. Rendered "serve". Deseruimus read as the perfect of desero cannot govern a dative and would reverse the sense of the clause; it is recorded here so that nobody later "restores" a desertion Richard is not describing.

1309B (0034) · qui frequentibus relevationibus divinitus relevari laetificarique meruerunt — relevatio/relevare are real words ("relief", "to relieve") and render in place, so they are rendered ("relieved and gladdened by frequent relievings"); but the whole paragraph is about revelation (divinae revelationis desiderio, ex multitudine revelationum suarum, lumen revelantis gratiae), and revelationibus … revelari is almost certainly what stood. Conjecture declined (not a negation, not a correlative; the printed reading asserts nothing false, only something tautologous). Named as a census candidate.

1307D (0034) · quid, inquit, per effectum istum David scelestius — effectum where Gregory's sense wants affectum; a real word, renders in place ("through this effect"), declined, named.

1300B (0031) · cum divino lumine tactus mentis intelligentia perspiciat — masculine tactus against feminine intelligentia. Pattern 9: Latin grammatical gender has no English exponent. Nothing to render either way, no marker, entry marked so a later sweep does not churn it.

E. Structural notes

1. ## LIBER SECUNDUS. closes 0030 with no text after it, and the two blank lines the Latin chunk carries after it are reproduced. Rendered ## BOOK TWO. per the ratified head conventions.
2. ## CAPUT PRIMUM. recurs here (0031) as Book Two's first chapter, rendered ## CHAPTER ONE. — the second of the work's three. Not a defect; the plate restarts.
3. ## CAPUT X (0034) prints with no period after the numeral, unlike every other head in the range. Preserved: ## CHAPTER X That most of all…, no period supplied (Pattern 8 applied to heads).
4. Standalone column-anchor lines — [1298C], [1299A], [1301B], [1307C], [1309B] each stand as their own paragraph after a head in the Latin, and do so in the English; [1303B] is run inline at the head of its paragraph in the Latin and is inline in the English. Line structure is 1:1 with the twin in all five chunks.
5. Interpolated particle inside a lemma — Nescit namque homo finem suum… (1301A) and Qui enim adhaeret Deo… (1309A). Both keep the plate's two italic spans with the roman particle between, after the shipped 0027 precedent (Charitas enim omnia suffert → "Charity indeed beareth all things"): "Man knoweth not indeed his own end…", "He who indeed cleaveth to God, is one spirit". Asterisk counts match the twin exactly in all five chunks.
6. Guillemets: one pair in the range, the Gregory quotation at 1307D, reproduced 1:1. Every other quotation Migne marks by italics alone, and no " " were supplied inside lemmata; the two supplied pairs in 0030 ("your" and "our", 1296A) mark words being mentioned rather than used, which is the one place the English needs them.

F. Counts

Question-mark parity (Latin body vs English body, [var:] content excluded): 8/8, 7/7, 8/8, 1/1, 9/9. Exclamation parity: 3/3 in 0030, 0 elsewhere. Guillemet parity: 1/1 in 0034, 0 elsewhere.

Richard of St Victor, De eruditione hominis interioris, Book II, capp. XI–XXII. Range carries the tree lemma itself (Dan. 4:7–9 at 1311D) and its first re-quotations, plus Dan. 4:10–11 (1318D) and the two re-quotes of 4:11 at 1319D and 1320C.

Each entry: column · printed reading · what I did · why.

A. Markers fired — [var: …] (Pattern 14). Eleven in all.

### A1. 1310B (0035) · Dan. 4:6 · quem ego scio quod spiritum … habeas Vulg. Baltassar princeps ariolorum, quoniam ego scio quod spiritum sanctorum deorum habeas in te. Migne prints the relative quem, which leaves "whom I know that thou hast the spirit of the holy gods in thee" without a governing verb. Rendered as printed; [var: Vulg. quoniam ego scio, …]. Material* because the whole chapter hangs on the king's reason for demanding the interpretation of Daniel rather than of the soothsayers — Richard re-quotes the clause at 1311A precisely as a causal premise (si veraciter constat quod per Nabuchodonosor dicitur: Scio quod spiritum … habeas in te). The causal quoniam* is the hook; Migne's relative is not.

### A2. 1311B (0035) · Mark 11:24 · Amen dico vobis, quidquid orantes petieritis … et fiet vobis Vulg. Propterea dico vobis, omnia quaecumque orantes petitis, credite quia accipietis, et evenient vobis. Four divergences in one clause (opening formula, quidquid for omnia quaecumque, future perfect petieritis for present petitis, fiet for evenient). One [var:] covering the run. ⚠ Cross-agent note: the launch brief records that chunk 0044 prints petitis for the same verse in the same work. So Migne's plate quotes Mark 11:24 twice with different verb forms; the 0044 agent should diff against this entry rather than treat it as fresh. Migne's note reads [n: (Marc. VI)] for a verse in Mark XI; the note is Migne's citation and is reproduced verbatim, uncorrected.

### A3. 1311B (0035) · Dan. 4:6 re-quoted · solutiones eorum for the lemma's solutionem eorum Migne re-quotes his own lemma four columns later with the plural. Vulg. and the lemma at 1310B both have the singular. Marked on the divergent printing only, per the brief's rule.

### A4. 1312D (0036) · Dan. 4:7 re-quoted · ecce arbor magna in medio terrae The lemma at 1311D (and Vulg. 4:7) read ecce arbor in medio terrae; magna belongs to v. 8. Migne's re-quote pulls it forward, and the gloss that follows immediately (Quid intelligemus per magnitudinem arboris?) is built on it. Marked; rendered as printed.

### A5. 1313B (0036) · Phil. 4:12–13 · Ubique et in omnibus instructus sum. Scio humiliari… Vulg. Scio et humiliari, scio et abundare (ubique et in omnibus institutus sum) et satiari, et esurire, et abundare, et penuriam pati. Two divergences: instructus for institutus, and the parenthetical clause hoisted to the front of the quotation.

### A6. 1314C (0036) · Phil. 4:12 re-quoted · Scio humiliari, scio et abundare ubique, et in omnibus instructus sum ⚠ The diff the brief asked for. The two printings are NOT the same text:

Both keep instructus and both drop the Vulgate's et after Scio, so the shared divergence is the plate's habit, not a one-off. Both marked, each on its own difference. Migne's second note cites Philippians VI, a chapter the epistle does not have; reproduced verbatim, uncorrected.

### A7. 1313B–C (0036) · Eccl. 1:16 · mens mea contemplata est multa sapienter et didicit Vulg. et didici — first person. Migne's third person makes mens mea the subject of the learning. Rendered "and hath learned"; marked. Material: Richard adduces the verse for the tree's proceritas, i.e. for what the mind reaches, and the third person keeps the mind rather than the man as the agent of the whole clause. Word order praecessi sapientia omnes (Vulg. praecessi omnes sapientia) is cosmetic and took no marker.

### A8. 1314D (0037) · Prov. 15:2 · Os enim sapientis ornat scientiam Vulg. Lingua sapientium ornat scientiam. Migne prints mouth for tongue and the singular wise man for the plural. Marked. Material for the gloss, which is running on verba / folia and on the single tree ("in the leaves, words").

### A9. 1317C (0038) · Sir. 34:11 · Qui non est tentatus, quid scit? Vulg. Qui tentatus non est, qualia scit? Rendered "what doth he know?"; marked. Migne prints the ? and it is kept (Pattern 8).

### A10. 1317C–D (0038) · Sir. 34:9 · Vir in multis expertus cogitavit multa Vulg. cogitabit, future. Perfect for future; marked. Printed roman, not italic — Pattern 14 has no formatting precondition, and the clause is adduced as a proof-text.

### A11. 1318D (0038) · Dan. 4:11 · dispergite fructum ejus Vulg. dispergite fructus ejus, plural. Marked once, at the lemma. The whole exposition of cap. XXII (1321C–D) is built on the singular — fructusque dispergitur, Fructus dispergitur twice — so the singular is the reading the gloss needs, and the re-quote at 1322A repeats it. Deliberately not marked a second time in 0039: it is one divergence, not two.

### A12. 1319A (0038) · Prov. 28:14 + 27:1 · Nescit enim quid futura pariat dies Prov. 28:14 ends at Beatus homo qui semper est pavidus; the second clause is Prov. 27:1, and there it reads ignoras enim quid superventura pariat dies — second person, and a different participle. Migne conflates the two verses, alters both words, and cites only Prov. XXVIII. Rendered as printed; marked.

### A13. 1319B (0039) · Dan. 4:10 re-quoted · Ecce, inquit, vigil sanctus de coelo descendit The lemma at 1318D and Vulg. print vigil, et sanctus — two subjects, "a watcher, and a holy one". Migne's re-quote drops the et, yielding one subject, "a holy watcher". Marked; rendered as printed ("a holy watcher"). Material and immediately so: the next two sentences argue from sanctus and de coelo as two separate warrants (cum sanctum legas…; ex eo quod sanctus et de coelo descendisse perhibetur), and the third printing of the same clause, eight lines later at 1319C, restores the et. The plate is inconsistent with itself inside one column band.

### A14. 1319D (0039) · Dan. 4:11 re-quoted · Praecidite arborem The lemma at 1318D and Vulg. print Succidite arborem, et praecidite ramos ejus — succidere of the trunk, praecidere of the branches. At 1319D Migne re-quotes the command as Praecidite arborem, transferring the branch-verb to the trunk; at 1320C, seven lines later, he prints Succidite arborem again. Marked on the divergent printing only. Material*: cap. XXII's whole scheme depends on the two verbs staying apart (Arbor succiditur et ramus praeciditur, quando virtus in occulto, et virtutis opus deficit in aperto). Rendered as printed ("Cut off the tree") both at 1319D and in its echo two lines later, since Migne prints praecidere* there as well.

### A15. 1319D (0039) · Heb. 1:14 · administratorii spiritus sunt … qui haereditatem capiunt salutis Vulg. Nonne omnes sunt administratorii spiritus … qui haereditatem capient salutis? Present for future, plus the word order of the opening. Marked. Material: Richard is arguing that the angels' guardianship is now and perpetual (perpetuam sollicitudinem impendunt, nec nocte nec die), which the present tense serves and the future does not.

### A16. 1321A (0039) · 1 Cor. 10:13 · Fidelis est Dominus, qui non sinit vos tentari Vulg. Fidelis autem Deus est, qui non patietur vos tentari. Two divergences, one of them the plate's known Dominus/Deus habit (the brief records Dominus coeli → Deus coeli across 1281B/1284A in round 1 — this is the same substitution running the other way). Rendered "the Lord"; marked.

B. Divergences found, examined, and deliberately NOT marked

Recorded so a later reader knows they were read, not missed.

1. 1310B · Dan. 4:6 · spiritum deorum sanctorum for Vulg. spiritum sanctorum deorum.* Word order only; the English is identical either way and no gloss touches it. The same order recurs in the 1311A re-quote, so it is the plate's settled form, not a slip.
2. 1315C · Sir. 3:22, quoted twice in roman · in rebus supervacuis for Vulg. In supervacuis rebus. Word order only.
3. 1315D · Sir. 3:26 · detinuit sensum illorum for Vulg. sensus eorum*. Number and pronoun; the English "held their sense in vanity" carries it, and nothing in the gloss turns on singular vs plural. Declined because I could not satisfy myself which of the two is the Clementine reading at the level of certainty a [var:] asserts.
4. 1317B · Ps. 127:2 · Labores manuum tuarum manducabis — Vulg. interposes quia (Labores manuum tuarum quia manducabis). A dropped conjunction that changes nothing; the sense and the gloss are untouched.
5. 1317D · Heb. 5:14 · boni et mali for Vulg. boni ac mali. Cosmetic.
6. 1313C · 1 Cor. 2:6–7 · two omissions — Migne drops Deus from quam praedestinavit Deus ante saecula and drops in gloriam nostram altogether. Compression inside a quoted run, not a variant reading: the subject of praedestinavit is recoverable and the English ("which he predestined") states no more than the Latin. Note reads [n: (II Cor. XXIX)] for 1 Cor. 2 — wrong epistle and a chapter number 1 Corinthians does not reach. Reproduced verbatim; the mis-citation is Migne's.
7. 1314B · Ps. 18:5 · note reads [n: (Psal. LXXXI)] for Ps. XVIII. Text matches the Vulgate exactly. Citation reproduced verbatim.
8. 1317C · Ps. 77:72 · note reads [n: (Psal. LXVII)] for Ps. LXXVII. Text matches (minus the initial et, which the quotation could not carry). Reproduced verbatim.
9. 1313B · Ps. 26:3 split across two notes. Migne closes the first quotation at non timebit cor meum under [n: (Psal. LV)] (Ps. 55:12 supplies only the first clause, Non timebo quid faciat mihi homo) and opens the second at Si exsurgat under [n: (Psal. XXVI)]. The verse division is Migne's; both texts match the Vulgate, so there is no textual variant to mark and the mis-binding is a citation matter.
10. 1316C · Ps. 36:35 · elevatum super cedros Libani for Vulg. sicut cedros Libani. A real and interesting divergence ("above" for "like"), but it is woven into Richard's own first-person sentence (Quoties ego ipse vidi, vidensque ingemui…), not adduced as a proof-text and not glossed. Per Pattern 18's boundary rule the deciding question is where the divergence sits, and this one sits in the author's own prose. Rendered as printed, no marker. ⚑ Named here in case a later reader wants it ruled the other way; it is the closest call in the range.
11. 1318A · Ps. 1:2 · sit in lege Domini voluntas ejus … meditetur for Vulg. Sed in lege Domini voluntas ejus … meditabitur. Richard has bent the verse into a jussive to serve his own sentence (Qui autem voluerit … sit in lege Domini voluntas ejus). Authorial adaptation, unmarked in the plate and unmarkable by us.
12. 1313D · Prov. 25:27, 1315C · Sir. 3:22 (Altiora igitur te…), 1316D · Ps. 101:11 (elevans allisisti me), 1321D · Ps. 49:16 — all woven into Richard's syntax with a connective of his own; all otherwise agree with the Vulgate.

C. ⭐ The pulls I resisted — lemmata where Migne agrees and the ear wanted otherwise

Positive evidence, per 7a″.

1. 1311D · THE TREE LEMMA, Dan. 4:7–9 · it agrees with the Vulgate word for word. I checked all nine clauses individually against the received text: Visio capitis mei in cubili meo · videbam et ecce arbor in medio terrae et altitudo ejus nimia · Magna arbor, et fortis, et proceritas ejus contingens coelum · aspectus illius erat usque ad terminos universae terrae · Folia ejus pulcherrima, et fructus ejus nimius · et esca universorum in ea · Subter eam habitabant animalia et bestiae · et in ramis ejus conversabantur volucres coeli · et ex ea vescebatur omnis caro. Zero divergences. The pulls, all resisted: - altitudo ejus nimia — Douay's "exceeding great" was the arriving phrase. Rejected: Richard's whole cap. XIV turns nimia into a fault (vide … ne forte sit et nimia; hoc fortassis succisionis causa exstitit quod ejus altitudo nimia fuit), and cap. XVI does the same for fructus nimius against multus and copiosus explicitly. Rendered "excessive" at every occurrence, in lemma and gloss alike. "Exceeding great" would have made the tree's fault a virtue. - aspectus illius — the eye wanted ejus to match the surrounding clauses. Migne prints illius and so does the Vulgate; kept. - esca universorum in ea is verbless in both. English needs a copula, so "and in it the food of all" supplies one as plain text with no brackets (the brief's rule); nothing else is added. - animalia et bestiae — "cattle and beasts" (Douay) was the arriving pair and is wrong here: Richard argues from the pair that bestiae are the wild animals and therefore animalia must be the tame ones (relinquitur ut nomine animalium sola hic mansueta intelligere debeamus). "Cattle" pre-empts an inference he draws in the next sentence. Rendered "the animals and the beasts" throughout, lemma and gloss.
2. 1314D–1315A · Wisd. 7:15–21 · agrees. The long catalogue matches clause for clause; Migne elides v. 16 and runs ipse enim straight on, which is quotation-compression, not a variant. The pull was virtutes elementorum → "powers of the elements": rejected, because virtus is fixed as virtue for this work and the tree's whole allegory is virtus.
3. 1314D · Job 29:21–22 · agrees, including stillabat eloquium meum.
4. 1315A · Rom. 1:14 · agrees. 5. 1313D · Rom. 11:20 and Rom. 1:21 · agree (the non of non sicut Deum glorificaverunt is carried). 6. 1314A–B · 2 Cor. 2:15–16 · agrees.
5. 1317B · Ps. 118:104 · agrees. 8. 1317D · 1 Cor. 9:27 · agrees.
6. 1318A · Col. 3:2 · agrees. 10. 1318B · Eccl. 1:5 · agrees.
7. 1320A · John 13:27 · agrees. 12. 1321A–B · Ps. 29:7–8 · agrees but for the dropped autem of Ego autem dixi.
8. ⭐ 1321B · Ps. 37:11 · Dereliquit me virtus mea, et lumen oculorum meorum et ipsum non est mecum — agrees with the Vulgate, and therefore diverges from Migne's OWN earlier printing. The brief asked for this diff and it comes out the other way round from what one expects: chunk 0014 at 1262C prints the clause WITHOUT meorum and carries a [var:] for it (cruces-0010 §1262C). Here the plate prints the full Vulgate clause. No marker here — the divergent printing is 0014's and it is already marked. Rendered "the light of my eyes" here and "the light of the eyes" there, which is exactly right: the two printings must not be harmonized.

D. Passages that would not construe, and conjectures DECLINED

Named, per the brief; census candidates.

1. 1319D (0039) · Quoniam erga rectitudinis statum subvertere malum esse cognovit. erga is a real word in a real form, so Pattern 12 has nothing to wrap; but with it the clause has no object for subvertere and reads "since he knew that to overthrow, as regards the state of rectitude, was an evil". The near-certain reading is ergo ("since therefore he knew that to overthrow the state of rectitude was an evil"), which is what the sentence's nevertheless… second half requires. [cj:] DECLINED: Wilson's bar reaches a negation or a correlative, and ergo is neither. Rendered as literally as the printed words allow, with erga given its own English ("as regards"); the conjecture lives here. This is the strongest Pattern 18 census candidate in the range.
2. 1319D (0039) · et secundum quod in summis divinitus discunt, et imis humana disponere. et imis for in imis — the preposition is not repeated and the ablative stands bare. Renderable ("in the lowest") without any repair to the sense; no marker, no conjecture.
3. 1314A (0036) · auctoritatesque ejus tam a bonis diligitur, quam a malis timetur. Plural subject, two singular verbs. The intended reading is almost certainly auctoritasque ejus. No marker. English concord is obligatory, so the disagreement has no English exponent that is not simply ungrammatical English of our own making; a [sic:] on auctoritatesque would swallow the -que (7a⁗ question 3) and the noun itself, leaving a lexeme-sized hole in a clause whose whole point is what authority does. Rendered "and his authorities are alike loved by the good and feared by the evil". Flagged here for the merge in case the work's other agents meet the same shape.
4. 1313C (0036) · Proceritas ejus coelum tangebat, qui veraciter dicere poterat. Masculine relative for feminine proceritas — Richard has slid from the tree to the man. Pattern 9: grammatical gender has no English exponent, the English relative is "who" either way, nothing to emend. Marked here so a later sweep does not churn it.
5. 1319C–D (0039) · Alia itaque sunt quae boni angeli per seipsos faciunt, alia quae fieri permittunt nulla omnino sine eorum dispositione uspiam fiunt. Two clauses fused with no connective and no punctuation between them. Rendered as printed, with a comma where Migne has one; the negation nulla omnino is carried in full. No marker: Pattern 8 forbids supplying the stop the sentence wants.

E. Corpus controls run

1. 1320B · volontatis (a patris summi beneplacito volontatis depravatione elongatus), beside voluntatis in the same chunk twelve lines later. Control: grep -rl "volontat" sources/pl/tei/ returns 10 files across the PL corpus — an attested medieval orthography, not this plate's defect. Pattern 9: translated normally ("of his will"), no italics, no [sic:], no crux beyond this line. Had it been unique to this site the answer would have been the opposite.
2. fortitudo checked against all of Books I–II before rendering the tree's fortis: the shipped chunks render fortitudo "fortitude" uniformly (Dan. 2:20/2:23 at 0008, 0011, 0016). Followed; see F2.

F. Fixed vocabulary followed against my own judgment, and tensions logged

1. ⚠ intellectus = "intellect" (ratification 3) produces two hard sentences in 0038, and the brief predicted both. - 1317C · Ps. 77:72 · In intellectibus manuum suarum deduxit eos → "In the intellects of his hands he led them." The sense of the verse is manual skill; "intellects of his hands" is opaque English. Followed the ratification, because Richard's gloss immediately makes the phrase mean cognitive advance won through works, and the pairing with A mandatis tuis intellexi two lines earlier is the point of the whole paragraph. - 1317D · enarrabit intellectum → "shall declare intellect". Same call, same reason. - The seam holds where it matters: intelligentia is "understanding" throughout (ad majorem intelligentiam proficit 1317C, intelligentiae sol 1318B, sensum spiritalis intelligentiae 1311C), and the verb intellexi/intelligit is "understood / understands", which the ratification does not govern and which keeps the paragraph readable. No tension in the noun/verb split; the tension is only in the two Psalm phrases above.
2. fortitudo = fortitude (ratification 4) forces robur = "strength" here, since the tree's robur is the trunk-word: immensum robur virtutis (1312D), robur fortitudinis (1317B, rendered "the strength of fortitude"), arboris illius robur succisum (1321B). The lemma's adjective fortis is "strong" and its abstract fortitudo is "fortitude"; the pair is visible on the page and nothing collides with virtus corporis, which does not occur in the range.
3. scientia = learning / cognitio = knowledge (ratification 5) carries two chapter heads here (XVI twice) and the whole of 0037; notitia = "cognizance" at 1314B (notitiam virtutis). Followed without difficulty.
4. studium. The fixed table gives "zeal", and the shipped chunks use both "zeal" (the abstract: summo studio, studia devotionis) and "studies" (the concrete plural: a nostris studiis, secular studies). I followed the shipped split rather than flattening it — "spiritual studies" in the cap. XI head and body, "zeal" for the abstract (studium meditationis → "the zeal of meditation", studii genus → "kind of zeal"). Naming it for the merge: if the work is to have one English word, "zeal" everywhere makes the cap. XI head read "subservient to spiritual zeals", which I judged worse.
5. mens / animus kept visibly apart everywhere, as required: mens "mind" (mens mea contemplata est, circumspectiones mentis, mens humana), animus "spirit" (animus humiliatur 1310C, virtute animi deficiente 1321C, animi arbitrium 1311B → "the judgment of the spirit"). They collide inside single clauses twice in 0035 alone (1310C–D) and the distinction is carried in both.
6. sacramentum = sacrament (ratification 5), never "secret", in Dan. 4:6 and its three glosses — including where Douay's "no secret is impossible to thee" is the arriving phrase. The gloss depends on it: Fit itaque summae devotioni nullum sacramentum impossibile, and mysterium stands separately at 1313C (Dei sapientiam in mysterio) and 1311C (de divinis mysteriis).
7. solutio = solution (ratification 5) in the lemma and both re-quotes; enodatio does not occur in the range.
8. eruditio = instruction (plena eruditio omnium 1314D, multitudinem eruditionis 1319A) and erudire "to instruct" (hominem ad discretionis … perfectionem erudiunt, ad discretionem erudimur), against the chastening sense the sentences carry by themselves.

G. Negation and punctuation notes

• Every printed negation is carried, including the awkward ones: et omne sacramentum non est impossibile tibi → "and every sacrament is not impossible to thee" (both omne and non, not the smoother "no sacrament is impossible"), while the gloss's own nullum sacramentum impossibile is rendered "no sacrament … impossible" because that is what it prints. non timebit cor meum, nihil audebant, nulla omnino, nec nocte nec die, non sicut Deum, nemo posset, nil … praesumunt, nec gregem porcorum … ausi sunt, all carried.
• Question marks: every ? Migne prints is kept and none supplied. Three nonne sentences in the range close with a period in the plate (1314B de quibus veraciter dicitur: …; 1314C …veraciter asseverat.; 1321B …quae dixit: …) and they close with a period in the English. Two more (1315A …qui in hoc ipsum se debitorem cognovit, cum dixit: …; 1313D …aut gratias egerunt:) end on a colon and a quotation; likewise kept.
• 1313B · …et proceritas ejus contingens coelum, Magna arbor fuit… — Migne prints a comma before a capitalized new sentence. Preserved exactly (Pattern 8); the English reads "…reaching unto heaven, A great tree it was".
• Em-dashes: zero in all five chunks.

H. For ratification, if the work wants it

• nimius = "excessive", in lemma and gloss alike, throughout Book II. It is not a preference: cap. XIV and cap. XVI both argue that the word is a fault, and cap. XVI contrasts it with multus and copiosus in one sentence. Chunks 0040+ re-quote altitudo ejus nimia; they should not print "exceeding great".
• animalia = "animals" / bestiae = "beasts", never "cattle", for the same reason (1316A's argument from mansueta vs fera). Re-quoted at 1316B, 1316D, 1317A and again in the later Daniel chapters.
• vigil = "watcher", vigilare "to be watchful", vigilia "watch" — Dan. 4:10's word, and cap. XX turns it into an angelology (utrique recte satis vigiles dici queant).
• proceritas = "tallness" kept distinct from altitudo "height" and sublimitas "sublimity". Cap. XIV runs all three in one sentence (Proceritas itaque ejus alta fuit, et altitudo ejus procera fuit), which collapses into nonsense if two of them share a word.

Richard of St Victor, De eruditione hominis interioris, Book II, capp. XXIII–XXXIV. The range runs from the flight of the beasts and birds (Dan. 4:11) through the whole of the second half of the tree-lemma — germen radicum … alligetur vinculo ferreo et aereo … et septem tempora mutentur super eum … In sententia vigilum — to the king's demand for the interpretation (Dan. 4:15). No book division falls in this range.

Each entry: column · printed reading · what I did · why.

A. Markers fired

A1. [var:] — Migne's scripture against the received text

1322C (0040) · Ps. 24:21 · Quia innocentes, et recti adhaeserunt mihi? Vulgate begins the verse Innocentes et recti adhaeserunt mihi, quia sustinui te. Migne's italic run opens with a quia that is Richard's own subordinator after protestetur ("protests … that the innocent…"), while the Vulgate's quia stands later in the verse and governs a clause Migne does not print. Rendered as printed ("that the innocent and the upright have cleaved to me?"), with [var: Vulg. Ps. 24:21 begins Innocentes et recti adhaeserunt mihi; the quia Migne prints inside the quotation is the conjunction binding it to protestetur]. Marked rather than merely logged because the italic boundary makes the conjunction look like part of the verse, which it is not. The ? is Migne's and belongs to the framing Nonne conversabantur…; carried per Pattern 8.

1322C (0040) · Ps. 87:9 · posuerant me abominationem sibi Vulgate posuerunt (perfect). Migne prints the pluperfect posuerant. Rendered "they had made me an abomination to themselves"; [var: Vulg. posuerunt, perfect, "they have made"]. Material here because the tense is doing work in Richard's frame: he is asking whether the birds had already flown from the branches at that time (Nonne et tunc temporis … aufugerant), and the pluperfect of the proof-text is what answers the pluperfect of the question.

1323D (0040) · Ps. 101:9 (Migne: Psal. CII) · Non in aeternum irascetur, neque in aeternum comminabitur? Two divergences at one site, and this is the range's most valuable entry.

1. Vulgate: Non in perpetuum irascetur, neque in aeternum comminabitur. Migne prints in aeternum twice, flattening the verse's own variation of phrase.
2. Migne closes it with a question mark on a clause the received text states flatly. The work's own control settles that both are this printing's, not the verse's: chunk 0014 prints the same verse at 1262B as … non in perpetuum irascetur, neque in aeternum comminabitur — with in perpetuum, and with a full stop. The committed English there reads "he will not be angry for ever, nor will he threaten unto eternity." Rendered here as printed and with both negations intact: "He will not be angry unto eternity, neither will he threaten unto eternity?" — i.e. the ? is carried and the Non … neque is carried, even though the two together make the English read as a question expecting the answer Richard does not want. [var:] records the Vulgate reading and names 0014's printing. ⚠ Note for the merge: do not "reconcile" 0040 with 0014. The plate disagrees with itself and that disagreement is the finding.

1323D (0040) · Hab. 3:2 · Cum enim iratus fuerit, misericordiae recordabitur Vulgate: cum iratus fueris, misericordiae recordaberis — second person, addressed to God. Migne prints the third person throughout. Rendered "For when he shall have been angry, he will remember mercy"; [var:] gives the second-person Vulgate reading. Material: the third person is what lets the clause serve as the object of audeat, et dicat in Richard's sentence, i.e. as something the penitent says about God rather than to him.

1324C (0041) · 2 Cor. 4:18 · quae autem non videntur aeterna Vulgate closes quae autem non videntur, aeterna sunt. Migne drops the second copula. [var:] marked; rendered with the ellipsis carried ("but the things which are not seen, eternal"). Small, but it sits inside the chunk's four-term antithesis (invisibilia/visibilia · aeterna/temporalia) and the elision is visible in the English.

1326D (0041) · Prov. 12:10 · Viscera avari crudelia Vulgate: viscera autem impiorum crudelia, "the bowels of the wicked". Migne prints avari, "of the covetous man". Rendered as printed; [var:] gives the Vulgate noun. Material, and the reason it is marked rather than logged: the sentence it proves is ex avaritiae vitio … obduratur — Richard adduces the verse for avarice specifically, and the received impiorum would leave his proof-text proving something more general than the vice he named. Whether the substitution is Migne's compositor, his copy-text, or Richard quoting from memory is not adjudicated here (Pattern 14 forbids speculating). Note the printed full stop after secundum illud. before the quotation: carried, Pattern 8.

1328B (0042) · Dan. 4:14 · quod dominatur Excelsus … constituet super eo Vulgate: donec cognoscant viventes quod dominetur Excelsus in regno hominum … et humillimum hominem constituet super eum. Migne prints the indicative dominatur and the ablative super eo. [var:] on the lemma; the re-quote eleven lines later at 1328C takes a short second [var:] per the round-1 rule that each occurrence is marked. Self-diff, and it runs the other way for once: chunk 0043's opening line (1329A, non-italic) prints dominetur — agreeing with the Vulgate — and then at 1329D the same chunk prints dominatur again. So within four columns the plate gives subjunctive, indicative, indicative. 0043's dominetur takes no marker (it agrees with the received text) and is logged here instead. The gloss is indifferent to the mood, so nothing in the exposition adjudicates it.

1328C (0042) · Phil. 2:13 · qui operatur in nobis velle et perficere pro bona voluntate Vulgate: qui operatur in vobis et velle et perficere. Migne prints in nobis and drops the first et. [var:] marked. This is the plate's known habit — round 1 recorded intra nos for intra vos in Luke 17:21 at 1240A — so it is reported as another instance, not as a discovery.

1330A and 1330B (0043) · Ps. 85:16 · Da imperium puero tuo Vulgate: Da imperium tuum puero tuo. Migne drops tuum at both printings; both marked, per the Luke 17:21 precedent. ⚠ Vocabulary hook: imperium is rendered "rule", not the Douay "command", because Richard's gloss immediately equates it with a kingdom — Pete ergo et tu tale imperium, vel regnum — and glosses the whole passage as the interior kingdom over one's own vices. "Command" would leave tale imperium, vel regnum explaining nothing.

1330B (0043) · Rom. 9:16 · Non est enim volentis, neque currentis, sed miserentis Dei Vulgate: Igitur non volentis, neque currentis, sed miserentis est Dei. Migne prints enim for igitur and moves the copula forward, leaving miserentis Dei without it. Diffed against the plate's own earlier printing, as the brief requires: 0040 @1323D prints Non enim volentis neque currentis, sed miserentis est Dei — i.e. the enim substitution is common to both printings, and only the copula moves. I marked 0043 (the divergent printing) and left 0040 unmarked, per the round-1 rule "mark the divergent one, not both". The 0040 site's enim-for-igitur is logged here and is deliberately not marked: it is a connective swap of the kind a citation formula makes, and it changes no sense.

1331A (0044) · Dan. 4:15 · Hoc somnium vidi ego rex Nabuchodonosor … nequeunt Vulgate: Hoc somnium vidi ego Nabuchodonosor rex … omnes sapientes regni mei non queunt solutionem edicere mihi. Migne inverts title and name, and prints nequeunt. [var:] records both; the English follows the printed order ("I, king Nabuchodonosor"). nequeunt is the same verb written solid and would not on its own be worth a marker; it rides along in the same note rather than taking a second. Self-diff: the re-quote at 1331D drops rex altogether and adds ejus (interpretationem ejus narra festinus). Both printings rendered as printed; the re-quote takes no marker, since truncation and a resumptive pronoun are Richard's citation practice, not a divergent text.

1332C (0044) · Mark 11:24 · Amen dico vobis, quidquid orantes petitis … et fiet vobis Vulgate: Propterea dico vobis, omnia quaecumque orantes petitis, credite quia accipietis, et evenient vobis. Migne prints Amen dico vobis, quidquid, and fiet vobis. [var:] marked. The diff the launch message asked for: chunk 0035 @1311B prints this same verse with petieritis (future perfect) where 0044 prints petitis — and it is 0044 that agrees with the Vulgate here, while 0035 diverges. 0035 also carries a different note, (Marc. VI) against 0044's (Marc. XI); both notes are reproduced verbatim and neither is corrected. So the verse's frame (Amen … quidquid … fiet) is divergent at both sites and its verb is divergent at only one, which is the opposite of what the re-quoting habit usually produces.

A2. [sic:] — carried type

1323C (0040) · ut in effectum crumpat crumpat is not a word. Corpus control: grep over the full PL TEI (5,276 files, 87M words) returns crumpat in exactly one file — this one, while erumpat is common and this author uses it (11325/0001, surculus erumpat in arborem). One occurrence, this file, this site: the plate's own defect, per the corpus-is-its-own-control rule. Carried untranslated, that it may [sic: crumpat] into effect. Conjecture: erumpat, "break forth" — near certain, and it lives here, not in the text.

1325A (0041) · ab assuetis concupiscentiis a vertitur Split type for avertitur (Pattern 10). Control: word-bounded a vertitur occurs in 2 of 5,276 TEI files, this one and 6956; avertitur in 113. Carried as the whole broken run. ⚠ 7a⁗ applied deliberately: the run is the relative clause's only verb, so quarantining it alone would have left "which … from its accustomed concupiscences?" with no predicate — the 1917A shape exactly. The English therefore says the sense beside the carried run: "which is turned away [sic: a vertitur] from its accustomed concupiscences". Pattern 10 decides what is carried; 7a⁗ decides what must still be said.

1325D (0041) · tantam saepe incurrit caligmem errorum caligmem is not a word. Control: one file corpus-wide (this one); caliginem in 385 files, and Richard prints errorum caliginem himself nine lines later at 1326A. Carried with the sense supplied beside it: "so great a darkness [sic: caligmem] of errors". Conjecture caliginem, effectively certain.

B. Passages that would not construe — rendered as printed, no marker

1325B (0041) · quasi contradictionem suarum portas rebellionum vectibus munire The accusative contradictionem and the feminine genitive plural suarum cannot both be governed here; the sense the sentence needs is contradictionis suae portas, "the gates of their contradiction", which is exactly what the next sentence spells out (porta aerea, nisi contradictio pertinax … vectis ferreus, nisi rebellio audax). Rendered by the sense the printed nouns supply, "to fortify the gates of their contradiction with the bars of rebellions", and logged. No [cj:]: the case anomaly is neither a negation nor a correlative, and Wilson's bar stands. Census candidate, declined and named.

1326C (0041) · cum crudelitatis ministris, tabulisque feris partem habere ⭐ The range's one genuinely unresolved reading. tabulisque is a perfectly good ablative plural of tabula, so Pattern 12 has nothing to wrap and Pattern 7 governs: rendered literally, "with the ministers of cruelty, and with wild planks". It is nonsense in place. Control: the string tabulisque feris occurs once in the whole PL TEI — here; every other tabulisque in the corpus sits with carpentry or records (cratibus tabulisque tecti, documentis tabulisque populi). Conjecture: talibusque feris, "and with such wild beasts" — and the sentence's own answering clause is the evidence, nisi cum talibus sociis, five words later, with feris already established as the chapter's word for the cruel (Per feras … recte satis crudeles quique intelligi possunt). Declined as a [cj:]: talis is a correlative pronoun, which brushes the bar, but the printed tabulis is not a wrongly-printed negation or correlative pair in the sense Wilson ruled — the correlative is in the answering clause and is printed correctly. Named here as a census candidate; I have not widened the bar. If the census reopens it, this is the strongest site in the range.

1329D (0043) · justum quaelibet et coelestem vitam ducentem justum (masc. acc. sg.) and quaelibet (neut. nom./acc. pl.) cannot agree, and neither agrees with vitam. Read as justam quamlibet … vitam, "a just life of whatsoever sort"; rendered that way and logged. No [cj:] — number/gender anomaly only, and Pattern 9 notes that English has no exponent for the gender. Census candidate, declined.

1330D (0043) · talem hominem, forte qualis designatur in arbore forte read as the adverb "perhaps" (with the talem … qualis correlative), not as fortem agreeing with hominem — which the chapter's own vocabulary would also support (vide quam fortis et potens … invenitur, three lines above). Rendered "such a man, perhaps such as is designated in the tree" and flagged here as genuinely ambiguous: the alternative reading is "such a man, strong, such as is designated in the tree." Nothing in the sentence decides it.

C. Lemma pulls RESISTED — positive evidence, per the brief

These are places where Migne agrees with the received text and the remembered English wanted something else, or where a familiar cadence invited a smoothing that was refused.

1. 1322B–C (0040) · Job 30:1–2, the range's longest lemma. Migne prints it word for word as the Vulgate, including et vita ipsa putabantur indigni. The pull was toward the Douay's "were thought unworthy of life itself" as a smoothing of an awkward Latin construction — which is in fact what the Latin says, so it stands, but it was checked word by word rather than recognized. No divergence: recorded as clean.
2. 1322C (0040) · Ps. 87:19, Elongasti a me amicum, et proximum, et notos meos a miseria. Exact. The pull was to render a miseria as "in my misery"; kept as "because of misery", which is what the preposition prints.
3. 1323C (0040) · Gen. 4:13, Major est iniquitas mea quam ut veniam merear. Exact — including the fact that this is the Vulgate's reading and not the Hebrew's ("my punishment is greater than I can bear"). No marker: Pattern 14 compares against the received Latin, not the Hebrew.
4. 1324B (0041) · Dan. 4:12–13, the whole alligetur vinculo ferreo et aereo … cor ferae detur ei fragment. Diffed against 0038's printing of the preceding verses and against 0046's printing of the interpretation, as instructed: exact against the Vulgate, all of it. 0046's germen radicum ejus in terra dimittite … vinciatur ferro et aere … rore coeli conspergatur … cum feris sit pabulum ejus … donec septem tempora commutentur is not a divergent re-quote of my text — it is Dan. 4:20–22, Daniel's interpretation, whose wording differs in the Vulgate itself. Do not merge these two as one lemma.
5. 1324B and 1327D · Et septem … tempora mutentur super eum. Exact at both printings.
6. 1326C (0041) · Ps. 56:5, Filii hominum, dentes eorum arma, et sagittae, et lingua eorum gladius acutus. Exact.
7. 1326C–D (0041) · Ps. 25:5 and Ps. 100:4, Odivi Ecclesiam malignantium… and Non adhaesit mihi cor pravum. Both exact. ⚠ Migne's notes read (Psal. C) and (ibid.) — the first verse is Ps. 25, not Ps. 100; only the second is. The notes are reproduced verbatim and the mis-citation is not corrected ([n:] rule).
8. 1326D (0041) · Ps. 5:7, Virum sanguinum … et dolosum abominabitur Dominus. Exact.
9. 1330B (0043) · Ps. 2:6, Ego autem constitutus sum rex ab eo. Exact; the Vulgate continues super Sion montem sanctum ejus, which Migne simply stops short of. Truncation is not divergence and takes no marker.
10. 1329D (0043) · Ps. 21:29, quoniam Domini est regnum, et ipse dominabitur gentium, woven into the prose unmarked and unitalicized. Exact.

D. Proof-texts woven into the prose — marking DECLINED

All of these are continuous enough to fall inside Pattern 14's "no formatting precondition", and I read each against the received text before declining. They are named so a later reader knows they were checked rather than missed.

• 1325A (0041) · Ps. 106:16, contrivit portas aereas, et vectes ferreos confregit — exact; nothing to mark, and it is the hinge of the whole chapter's allegory.
• 1325D (0041) · Isa. 5:20, ut dicat malum bonum, et bonum malum, ponens tenebras lucem… — Richard's own paraphrase in indirect discourse, not a continuous quotation; declined.
• 1324C, 1326C (0041) · Isa. 40:6, omnis caro fenum, et omnis gloria ejus quasi flos feni, printed twice, unmarked. Rendered "all flesh is hay, and all the glory thereof as the flower of hay" — following 0029's committed English at 1295B, which carries the [var:] for the plate's flos feni against the Vulgate's flos agri. Since round 1 has already marked this divergence at its italic printing, and here it is unitalicized authorial repetition, I did not re-mark it. Flagged for the merge in case a uniform policy is wanted.
• 1328C (0042) · Heb. 1:14, omnes administratorii spiritus missi sunt propter eos qui haereditatem capiunt salutis — bent into indirect discourse after didicimus quia; declined.
• 1330B (0042/0043) · 1 Cor. 4:7, si nihil habes quod non accepisti, and 1330C · Jas. 4:6, Superbis namque resistit, humilibus autem dat gratiam (Migne drops Deus) — both syntactically absorbed; declined.
• 1330D (0043) · Eccli. 10:15, initium sit omnis peccati superbia, with the subjunctive drawn into Richard's Cum … attestante Scriptura — the Vulgate's initium omnis peccati est superbia. Declined: the mood change is the frame's, not a variant text.
• 1332D (0044) · Rom. 8:26, interpellat pro nobis gemitibus inenarrabilibus; 1331B · Dan. 4:19, Tu es arbor illa quam vidisti (Vulgate Tu es rex … arbor quam vidisti). Both woven and unitalicized; declined, though the second is the closest call in this list.

E. Fixed vocabulary — followed, including where it cost something

• animus = spirit / mens = mind held apart everywhere, and they collide in this range: 1327D–1328A (Vitiorum … dominatio animum paulatim emollit) sits in a chapter whose whole subject is the mens, and 1331D–1332A runs animo … animus … animus through a passage about what the mens cannot reach. Rendered "spirit" for animus throughout. spiritus and animus never stand in the same clause anywhere in this range, so the brief's two-site collision is not touched here: spiritus appears only as the malignant spirits at 1326A, the ministering and unclean spirits at 1328B–1328A, and in 0044 as the Holy Spirit and the Dan. 4 spiritus deorum sanctorum — every one of them unambiguous in context, and none beside an animus.
• aes = bronze, uniformly, against the ear's "brass" for portas aereas (Ps. 106:16) — the ratified five-materials rule wins over the familiar psalm cadence, and it has to, because the chapter's allegory runs ferrum → audacia / aes → pertinacia across four columns and back into the Dan. 2 statue.
• pertinacia = obstinacy (the four-vice scheme) and saevitia = savagery (1327B, primus gradus saevitiae) followed exactly; audacia, which is not in the scheme, is rendered audacity throughout so the pair pertinaciter / audaciter stays visible.
• detrimentum = decline at 1327A (intellige detrimenta bonorum) — "losses" read better against incrementa malorum and was refused.
• industria = effort at 1327B–C, twice, where "industry" was the pull.
• simulatio = dissimulation at 1329C (absque simulatione) — and note this is simulatio, not the ratified simultas; the two do not collide anywhere in this range.
• notitia = cognizance in CAPUT XXXIV's own head; solutio = solution and interpretatio = interpretation kept apart through 0044, where Nabuchodonosor asks for both.
• eruditio/erudire = instruct at 1329D (ad circumspectionis cautelam erudiuntur), the title word, where the sense is plainly chastening — carried by the sentence, per the table.

⚠ One tension logged, per the brief's instruction to follow and report: 1324A (0040) · se a suis voluptatibus non posse cohibere cum sacramento asseverant. The ratified table fixes sacramentum = sacrament, kept apart from mysterium — a ruling made for the Dan. 2 material, where sacramentum is the mystery Daniel unfolds. Here the word plainly means an oath: it stands in parallel with etiam cum juramento affirmant in the preceding clause, and no sacrament is in view. I rendered cum juramento "with an oath" and cum sacramento "under solemn oath", so the parallel survives and neither word is flattened onto the other. I did not follow the table into nonsense, and I am not deciding the work's usage: this is a request for ratification — either the table gains an exception for the oath sense, or it is narrowed to the Daniel material where it was made.

F. Punctuation carried (Pattern 8) — the sites worth naming

• 1323D, the ? on Ps. 102:9 — see A1.
• 1322B–C (0040), the long Nonne … nonne, inquam, … abominabantur … cum de se veraciter protestabatur: [Job XXX lemma] — Migne closes this rhetorical question with a full stop, not a ?. No question mark supplied. The one in my first draft was removed.
• 1324C (0041), Quid enim nisi quaedam herbae terrae sunt bona ista terrena, quae nobis sunt cum bestiis communia. — a Quid enim question closed with a period. Carried.
• 1326D (0041), the full stop inside secundum illud. Viscera avari crudelia — carried.
• Question-mark parity was checked mechanically across all five pairs after writing: 8/8, 10/10, 1/1, 8/8, 8/8. Anchors and [n:] counts likewise match the Latin twin exactly and in order.

G. Note on the Job citation printed inside the italics

1322C (0040) prints its locator as bare text inside the lemma — … et vita ipsa putabantur indigni (Job XXX) — rather than as a trailing [n: …] marker, which is what every other citation in this range does. Reproduced exactly as printed, inside the italic run and untranslated, on the same principle that governs [n:] contents. It is not a defect and takes no marker; noted only so that an indexer knows this locator will not be found by scanning [n:] alone.

Richard of St Victor, De eruditione hominis interioris, PL 196. Range: chunks 0045, 0046, 0047, 0048, 0049. Format: column · printed reading · what I did · why.

Content of the range: Daniel's silence and the office of devotion (0045); Daniel's interpretation of the tree dream, Dan. 4:16–24 (0046); the twelve months, Nonne haec est Babylon magna, and the voice from heaven, Dan. 4:25–29 (0047); the fulfilment, the hair like eagles' and the claws, and the raising of the eyes, Dan. 4:30–31 (0048); the doxology and the restoration, Dan. 4:31–33 (0049).

A. Markers placed

[sic: …] — carried type (Patterns 10 + 12)

Each was controlled against the whole PL TEI (5,276 files, 87M words) before it was set.

1. 1335B · Dan. 4:17 · Arborem … quam vidisti sub limem atque robustam Migne sets sublimem as two words. sub limem occurs once in the entire PL TEI — this file, this site; sublimem occurs in 310 files. A word-division defect, Pattern 10: the whole run is carried, spacing included, and marked [sic: sub limem]. No second italic marker, since the run sits inside an already-italic lemma (Pattern 10's nesting rule). 7a⁗ check: the quarantined run is an attributive adjective; the clause keeps its verb (vidisti), carries no negation and no conjunction, so nothing is supplied beside the marker. Conjecture: sublimem, "lofty" — near-certain, and the parallel at 1335B's own cujus altitudo pertingit ad coelum carries the sense anyway.

2. 1337C · divinae pietatatis viscera A dittographic syllable; not a word. Corpus control: pietatatis in 2 files against 2,148 for pietatis. Decisive against the plate. Carried and marked. 7a⁗: the marker takes a genitive noun only; the governing phrase ("the bowels of the divine …") and the sentence's verb survive outside it. The work's own parallel at 1342A prints Quis divinae pietatis viscera digne pensare sufficiat? — same sentence, correct form, five columns later. Conjecture: pietatis, certain.

3. 1337C · quod contemplum suum non statim vindicat contemplum is not a Latin word. Corpus control: 3 occurrences in 87M words (11296, 21385, and this file), i.e. plate-level rarity everywhere it appears. Carried and marked. ⚠ 7a⁗ was checked at this site with care, because the marker sits in a negative clause: the English retains both the verb (vindicat, "avenge") and the negation (non) — "he does not at once avenge his own [sic: contemplum]". Only the object is quarantined. Conjecture: contemptum, "the contempt shown him", which is what the antithesis (sed correptionis nostrae satisfactionem exspectat) requires.

4. 1338A · quem de fieta humilitate sua conscientia accusando fatigat fieta for ficta. Unique to this file in the whole PL TEI. Carried and marked. Conjecture: ficta — certain, because the same sentence's neighbours print ficta humilitas three times in the preceding six lines (1338A, twice).

5. 1341A · et vetut unguibus debacchantur vetut for velut. Corpus control: 3 bare occurrences of vetut in the whole TEI, plus 4 of vetutissim- (itself defective for vetustissim-) — i.e. a known compositor's confusion of l/t, not a word. Carried and marked, untranslated (so "as it were" is not also rendered; that would double-count the run). 7a⁗: velut is a comparative particle, not a negation, conjunction or predicate; the clause keeps debacchantur and unguibus. Conjecture: velut — certain, and the same phrase prints correctly as et velut unguibus desaeviunt eleven lines later at 1341B.

[var: …] — Migne's scripture against the received text (Pattern 14)

6. 1334D · Dan. 4:16 · somnium, et interpretationes ejus non conturbent te Vulgate: interpretatio eius, singular. Corpus control: interpretationes ejus non conturbent te occurs nowhere else in the PL TEI; the singular is attested (7197). Material and internally decided: Migne's own next lemma, eight lines below, prints et interpretatio ejus hostibus tuis sit — singular. Rendered as printed ("the interpretations thereof"); [var:] names the singular and points at the next lemma.

7. 1334D–1335A · Dan. 4:16 · somnium his quae te oderunt Vulgate: his qui te oderunt, masculine. Corpus control: somnium his quae te oderunt occurs nowhere else; somnium his qui te oderunt in two other works (11412, 7197). Decided inside the chunk: Richard re-quotes the same clause three lines later as Somnium his qui te oderunt — so the plate contradicts itself and the second printing holds the received form. Rendered as printed with English which (the relative that can carry a non-personal antecedent), which is not an accident of style: Richard's gloss immediately reads Vitia veri animae hostes sunt, quae eam ad aeternum interitum impellunt — vices, neuter. The printed quae is the reading his gloss picks up. Also noted inside the marker: Migne adds sit, which the Vulgate lacks.

8. 1336C · Dan. 4:24 · forsitan ignoscet Deus dilectis tuis ⭐ The most consequential divergence in the range, and the one the launch message asked to be adjudicated between Patterns 14 and 18. Migne prints dilectis, "thy beloved ones", where the sense requires delictis, "thy sins" — the whole verse is about redeeming sins with alms.

What settled it. Corpus control: forsitan ignoscet Deus dilectis tuis occurs once in 87M words — here; forsitan ignoscet delictis tuis occurs in four other PL works (11412, 10900, 7197, 10885). And the work is again its own control: at 1337B, inside my own chunk, Richard re-quotes the clause as Forsitan, inquit, ignoscet delictis tuis and then expounds it as pardon for offences. So delictis is both the received reading and this author's reading; dilectis is this setting of this column.

Pattern 14, not Pattern 18, and why. Pattern 18's own boundary is explicit and is about where the divergence sits, not how it reads: "[var:] inside a quoted scripture, [cj:] in the author's own prose" — the worked case being 7561 @0664's modio for medio inside Deut. 13:1, which has exactly this shape. This word stands inside a continuous, italicised, [n:]-cited lemma of Dan. 4. It is therefore [var:]. Independently, Wilson's [cj:] bar would decline it anyway: dilectis is neither a negation nor a correlative. Both routes give the same answer, which is why I am confident in it. Note the divergence from ignoscat (Vulgate subjunctive) and the added Deus were not marked: the four corpus witnesses all print ignoscet, so the indicative is the circulating reading of the verse, not this plate's. The marker names delictis only.

9. 1340C · Dan. 4:30 · et ungues quasi ungues avium Vulgate: et ungues eius quasi avium. Migne sets ungues twice and drops ejus. Corpus control: ungues quasi ungues avium occurs nowhere else in the PL TEI; ungues ejus quasi avium is attested (7197). Rendered as printed ("his claws like the claws of birds"). Not a [d: …]: the two occurrences are not an adjacent doubled run but a well-formed comparative construction, so Pattern 11 does not fire. Material, because the whole of CAPUT XLIV is a gloss on this clause word by word (Quid itaque est ungues velut avium habere), and because the plate itself restores ejus when it re-quotes at 1341B (et ungues ejus* quasi ungues avium) and abbreviates again at 1341D (ungues ut avium accepit) — three forms of the clause inside one chunk.

10. 1342D · Rom. 14:5 · Alius judicat inter diem et diem Vulgate: alius iudicat diem plus inter diem, "one esteemeth one day above another". Rendered as printed. Material: Richard's exposition two lines above turns on the printed form and its contrast (qui judicant inter diem et diem … qui judicant omnem diem*), so the Vulgate's comparative would put the argument on a phrase not on the page. ⚠ Honest weakness of this one, stated so a later sweep can reverse it: the corpus control is ambiguous rather than supporting. judicat inter diem et diem is printed by two other PL works (9442, 11074) besides ours, and judicat diem plus inter diem by none — so Migne's form is a circulating quotation of the verse, not this plate's slip. I kept the marker because Pattern 14 records the divergence and expressly forbids speculating about its cause; but of the four [var:] in this range this is the one I would drop first.

11. 1345A · Eccli. 3:19 · in mansuetudine fac opera tua Vulgate: in mansuetudine opera tua perfice, "perfect thy works". Corpus control: in mansuetudine fac opera tua occurs nowhere else in the PL TEI; opera tua perfice in two other works (8200, 8902). A verb substitution, unique to this setting. Rendered as printed; [var:] gives the received verb.

B. ⚠ Markers I placed and then WITHDREW, and the corpus control that killed them

These are recorded in full because the withdrawals are evidence, and because three of them would have been false markers — a [var:] asserting a divergence that is not there is worse than a missed one, since nothing downstream can catch it.

W1. 1335D · Dan. 4:20 · rore coeli conspergatur … donec septem tempora commutentur — WITHDRAWN. I first marked conspergatur against the Vulgate's tinguatur and commutentur against mutentur, on the assumption (which the launch message shared) that this run is a fourth printing of the passage that 0038, 0040 and 0041 print, and could be diffed against them. It is not. Dan. 4:20 is the interpretation, and it differs from the vision of Dan. 4:11–13 in the received text itself: dissipate illam for praecidite ramos ejus, dimittite for sinite, vinciatur ferro et aere for alligetur vinculo ferreo et aereo, sit pabulum ejus for pars ejus in herba terrae. Migne reproduces every one of those Vulgate-internal differences correctly. Corpus control, run afterwards and decisive: rore coeli conspergatur is printed by four independent PL works (11412, 7197, 21346, 11535); donec septem tempora commutentur by three (11412, 21346, 11535) against one for mutentur (7197). Migne is with the majority at both sites. No divergence. No marker.* The coordinator's correction arrived independently and agrees.

W2. 1335A · Dan. 4:22 · cum bestiis ferisque erit habitatio tua — WITHDRAWN. I marked this against a Vulgate cum bestiis feris ("with the wild beasts", one class), arguing it was material because the exposition at 1339C expounds two classes (bestialiter viventibus … feraliter agentibus). Corpus control is a 2–2 split: cum bestiis ferisque erit habitatio in 10877 and here; cum bestiis feris erit habitatio in 7197 and 8585. With the witnesses evenly divided and no plate at hand, I cannot establish which is the received reading, so the marker comes off. Recorded here as an open question, and it is a real one, because the plate prints the clause three ways inside my two chunks: ferisque at 1335A and 1335D, atque feris at 1339C, et feris at 1339C's own gloss twelve lines later.

W3. 1333B/1333D · Ps. 64:9 · Exitus matutini, et vespere delectabis — WITHDRAWN. This looked like the strongest find in 0045: Migne's active delectabis against the Gallican delectaberis, with Richard's gloss hanging on exactly that (Tu, inquit, delectabis, non ipsa hora — "Thou shalt delight, not the hour itself"). Corpus control killed it: matutini … delectabis is printed by twelve PL works against two for delectaberis. The active is the dominant PL reading of the verse, not this plate's. Rendered active (which the gloss requires), no marker*. ⭐ This is the single best argument in the range for running the corpus control before setting a marker rather than after: every literary signal pointed at a divergence and there is none.

W4. 1333D · Jer. 17:9 · Pravum siquidem est cor hominis, et inscrutabile — WITHDRAWN. Vulgate: cor omnium. Corpus control: Pravum est cor hominis in eleven works, Pravum est cor omnium in two. Dominant PL reading. Rendered as printed, no marker.

C. Conjectures and markers DECLINED — census candidates, named

Wilson's [cj:] bar (a conjecture is glossed only for a negation or a correlative) declines every one of these. They are named so the census does not have to find them again.

C1. 1333B · Istud est tempus ridendi. Printed with no antithetic clause. The argument requires the pair tempus flendi / tempus ridendi — the next two sentences ask Cur non sit tempus ridendi in matutino … Cur non sit tempus flendi in vespertino — and Istud, following considerandi infirma nostra, attaches "laughing" to the wrong member. Either a clause has dropped out or ridendi stands for flendi. Rendered as printed ("That is the time of laughing"), no marker. Declined: the missing element is a correlative member, not a correlative particle, and supplying it would be inventing a clause, not glossing a word.

C2. 1333B · Vespertinum tempus est quando ante contemplantis oculos lumen revelantis gratiae emergit… The sentence describes the morning in this chapter's own scheme (mane … ab interioribus ad exteriora exire, et vespere ab exterioribus ad interiora redire; Delectant … exitus matutini, contristant autem introitus vespertini). Matutinum is what the argument needs. Rendered as printed, no marker. Declined: a real, correctly-formed word in its slot, and neither a negation nor a correlative — the Pattern 18 bar as written. The nearest ruled analogue is 11064 @1154C, likewise declined.

C3. 1333C · Quis autem digne miremur quam sit dulce… Interrogative singular Quis with a first-person plural verb. The sentence will not construe in either language. Rendered as closely as the printed words allow ("But who may we worthily wonder how sweet it is…"), which is awkward English on purpose: I would not silently drop either quis or the plural, since a mid-run person switch is a named invisible-defect class. Declined — an explicit person/number anomaly, the class the 2026-08-15 ruling scoped out.

C4. 1333C · saepe extra semetipsum per mentis excelsum rapit against 1333C's own humanae mentis excessus eight words later, and against the work's fixed term mentis excessus = "excess of mind" (ratified item 5). excelsum is a real noun ("a height") and renders in place ("through the height of the mind"). Rendered as printed, no marker. Declined: not a negation or correlative. Flagged for the merge pass, because a reader will meet "the height of the mind" and "the excesses of the human mind" in consecutive clauses and they are the same technical phrase in the plate's intention.

C5. 1337C · Post finem, inquit, mensium duodecim … Respondit rex, et ait against the Vulgate's Responditque rex et ait — and against 1338B, where Richard re-quotes it as Responditque rex, et ait, with the -que. A genuine self-diff, logged; no [var:], because a lost connective particle does not change the sense and Pattern 14's bar is materiality. mensium for mensuum is orthographic and takes nothing.

C6. 1339A · Matt. 5:16 · ut videant opera vestra bona for the Vulgate's vestra bona opera. Word order only, no gloss turns on it. Declined — but recorded, because the 7914 @1079B precedent shows a reversed order can* be material when the gloss picks up the sequence, and here it does not.

C7. 1339B–C · Dan. 4:29 · et cum bestiis atque feris erit habitatio tua for the Vulgate's et feris; and at 1339C Richard's own gloss then prints et cum bestiis et feris erit habitatio tua, the Vulgate form. Connective only. Declined, logged with W2 above as part of the same open question.

C8. 1339D · Isa. 40:6 / 1 Pet. 1:24 · omnis caro fenum, et omnis gloria ejus quasi flos feni — Isaiah reads flos agri, 1 Peter flos faeni. Migne's form is 1 Peter's, and it is woven into Richard's own sentence with no italics and no [n:]. Declined: this is a conflated allusion, not a quotation adduced as proof, so it sits outside Pattern 14's scope. Same treatment for ante ruinam exaltabitur cor* at 1334C, which conflates Prov. 16:18 and 18:12.

C9. 1342A · Dan. 4:31 · et regnum ejus in generatione et generationem for the Vulgate's in generationem et generationem. Declined on the corpus control, decisively: in generatione et generationem is printed by 42 PL works, in generationem et generationem by 40. Both forms circulate massively; this is Pattern 9 territory, not a divergence. The same form recurs at 1343B* in 0049 and takes nothing there either.

C10. 1342C–D · Isa. 5:20 · ponentes amarum dulce et dulce amarum where the Vulgate (and chunk 0041 at 1325D) prints ponens amarum in dulce et dulce in amarum. Woven, unitalicised, uncited. Declined; logged because it is another instance of the plate re-quoting itself with a difference.

C11. 1340B · Dan. 4:30 · sermo completus est super Nabuchodonosor, ex hominibus abjectus est — the Vulgate has et ex hominibus. Dropped connective, no marker.

D. The resisted pulls — lemmata where Migne agrees and my instinct did not

Positive evidence, per the brief. Each of these I checked word by word and left alone.

R1. ⭐ 1336A vs 1339D–1340A · Dan. 4:22 super regnum hominum / Dan. 4:29 in regno hominum. This is the range's headline non-finding, and it is exactly the shape the brief warned would be a divergence. At 1336A Richard quotes Donec scias quod dominetur Excelsus super regnum hominum; at 1339D–1340A, three columns later, the same formula prints quod dominetur Excelsus in regno hominum. The plate appears to contradict itself on the word the argument rests on, and CAPUT XLII's whole point is that una eademque sententia is repeated three times. Both printings are correct. The received text distinguishes them: Dan. 4:22 (the threat, super regnum) from Dan. 4:29 (the fulfilment, in regno) — Douay preserves the distinction as "ruleth over the kingdom of men" against "ruleth in the kingdom of men", and Dan. 4:14, quoted at 1328B–1329D in the peer range, is in regno again. Migne has quoted the right verse in the right place both times. No marker at either site, and a later reader who "fixes" one of them will be wrong.

R2. 1339C · Dan. 4:29 · Fenum quasi bos comedes against 1335D's Dan. 4:22 fenum ut bos comedes. Same trap, same answer: the Vulgate itself has ut at 4:22 and quasi at 4:29. Migne is right twice. No marker.

R3. 1335D · Dan. 4:20's whole watcher-decree. See W1. Every clause that looks like a variant of 0038/0040/0041 is the received text's own difference between the vision and its interpretation. The temptation to harmonise here is very strong because the four printings sit within twenty columns of each other. Nothing marked.

R4. 1336B · Dan. 4:23 · Quod autem praecepit, ut relinqueretur germen radicum ejus, id est arboris: regnum tuum tibi manebit postquam cognoveris potestatem esse coelestem. Verbatim the received text, including the abrupt id est arboris. No marker. Checked because the clause reads like a gloss that has fallen into the lemma; it has not.

R5. 1333D · Ps. 41:7 · Ad meipsum, anima mea conturbata est, propterea memor ero tui — verbatim. 1333D · Ps. 76:4 · Memor fui Dei, et delectatus sum — verbatim. 1338A · Ps. 121:3 · 1338D–1339A · Joan. 15:5, 1 Cor. 1:31 · 1341B · Ps. 49:19 · 1342A · Ps. 36:6 · 1342C · Ps. 76:6 · 1343B · Philipp. 3:20 · 1343C · Ps. 134:6 · 1344B · Ps. 33:6, Ps. 35:10 — all checked word by word against the received text, all agreeing, nothing marked. The Migne [n:] locators at several of these are off by one psalm-numbering convention (Psal. XXXVII for Ps. 36:6, Psal. LXXVIII for Ps. 76:6, I Cor. II for 1 Cor. 1:31); those are Migne's citations and pass through verbatim under the [n:] rule, uncorrected and unmarked.

R6. 1341B · Ps. 93:21 · Captabunt in animam justi, et sanguinem innocentem condemnabunt — futures, verbatim the received text. No marker. ⚠ Flag for the agent holding chunk 0054, which the launch message reports prints the same verse with imperfects (Captabant … condemnabant): on the evidence of my site, 0054 is the divergent printing and 0045–0049's is the sound one. The marker belongs there, not here.

E. Fixed vocabulary — followed, and one place it strained

Followed without incident: mens mind · animus spirit · anima soul · cor heart · intelligentia understanding · intellectus intellect (only intellectualiter occurs here, 1339C) · fortitudo fortitude (load-bearing in this range: in robore fortitudinis meae recurs five times at 1337C–1339A and is glossed ecce arrogantia, so "in the strength of my fortitude" keeps robur and fortitudo visibly apart) · cognitio knowledge · scientia learning (does not occur here) · discretio discretion · magisterium mastery (1338D) · arrogantia arrogance, the four-vice scheme intact · eruditio/erudire instruction/instruct (1334A divinitus eruditur) · praesumere presume · liberum arbitrium free will (1344C) · mentis excessus excess of mind (1333C) · proper names in Migne's Latin form (Nabuchodonosor throughout; Balthasar at 1334D and 1335A — note the plate spells it with one s in 0046 where 0035 at 1310B prints Balthassar; both kept as printed in their own chunks).

Strain, logged as the brief requires: pietas. Held to piety everywhere, against my own judgment at two sites. At 1337C divinae pietatatis viscera and 1342A divinae pietatis viscera the sense is God's tenderness, for which "the bowels of the divine piety" is stiff; and at 1337D tam justa quam pia pairs pius with justus. "Compassion" was the obvious alternative and is unavailable, because compassio/compati occurs in the same range (1337A–B, alienae afflictioni compati) and would collide. misericordia holds mercy, miseratio compassion (1338D, 1342A–B). Piety stands; flagged for ratification rather than decided here.

F. Vocabulary calls I am proposing for the whole work

These are new to this stretch of Daniel and will recur in 0050–0057 and in the peer ranges 0035–0044. They need to be one word across the work or the tree allegory loses its hooks.

1. bestia = beast · fera = wild beast · animal = animal. Non-negotiable in this range: 1339C glosses bestialiter degens et feraliter agens as two distinct vices (carnality and cruelty), and 1340D–1341A turns on ferae preying on animalia while aves prey on aves. Collapsing any two is the testa/fictilis error in another key.
2. ungues = claws (not "nails"). Carries both the birds of Dan. 4:30 and 1341D's dente et ungue in omnes saevit, "with tooth and claw", where "nails" would be unreadable.
3. fenum = hay · herba = grass. Both are in the Dan. 4 lemma at once (fenum ut bos comedes beside in herbis foris / in herba terrae), so they must not share a word. "Hay" also carries omnis caro fenum … quasi flos feni at 1339D.
4. germen radicum = the shoot of its roots. Used at 1335D and 1336B; the peer range 0040 (1323B–1324A) expounds germen at length as propositum correctionis, so whatever that agent chose must win if it differs — flagged for the merge pass.
5. decor = comeliness · decorare = to make comely. Dan. 4:33's ad honorem regni mei decoremque perveni is expounded at 1344C–1345A as a matched pair with honor, so the two English words must stay distinct and stable.
6. aula = hall · palatium = palace. Richard distinguishes them explicitly at 1337D–1338A (Forte palatium … humilitas vera against aula Babylonica humilitas ficta).
7. gloriatio = glorying · glorificatio = glorification. CAPUT XLVII's head is built on the opposition and it recurs through 1344A.

G. Left unresolved

• W2 / C7, the bestiis ferisque / atque feris / et feris question at Dan. 4:22 and 4:29. Three forms in two chunks, corpus witnesses split 2–2, no marker set. A plate read of 1335A, 1335D and 1339C would settle it in one pass.
• Marker 10 (Rom. 14:5) is the weakest [var:] in the range; the reasoning for keeping it and for dropping it is set out above, and I would not defend it hard.
• C4, mentis excelsum against mentis excessus in the same column, needs the merge pass to see it beside however 0030–0044 rendered mentis excessus.

Range: 0050 (1345C–1348B, end of Book Two + the LIBER TERTIUS head) · 0051 (1347C–1350D) · 0052 (1351A–1353C) · 0053 (1353D–1356A) · 0054 (1356B–1358C). Lemma across the range: Dan. 7:2–5, the four winds and the beasts from the sea — the lioness with eagle's wings (0051–0053) and the bear with three rows of teeth (0054).

Format: column · printed reading · what I did · why.

A. Markers fired

[var: …] — Pattern 14 (Migne's scripture against the received text)

Policy applied, stated so it can be checked: a [var:] fires where the English on the page would assert something different — a different word, tense, number, direction, or a dropped word that carries sense. Pure abridgement of a longer verse (clauses elided, no wording altered) and connective-level substitutions that leave the English unchanged are logged in §B instead, not marked. Twelve fired.

1. 1345D · Job 4:6 — Ubi est timor tuus, et fortitudo tua, patientia tua, et perfectio viarum? Vulg. Ubi est timor tuus, fortitudo tua, patientia tua, et perfectio viarum tuarum? Migne adds et after timor tuus and drops tuarum: the English loses "of thy ways" and the verse stops indicting Eliphaz's hearer. Marked; rendered as printed.
2. 1345D · Apoc. 2:4–5 — Habeo adversum te aliquid quod primam charitatem reliquisti. Memento unde excideris… Vulg. Sed habeo adversum te, quod caritatem tuam primam reliquisti. Memor esto itaque unde excideris. Three divergences at once: Sed dropped, aliquid intruded, primam charitatem for caritatem tuam primam (the possessive gone), Memento for Memor esto itaque. Diffed against the work's own other printing at 1271C (chunk 0019), which reads Sed habeo adversum te, quod charitatem tuam primam dereliquisti — i.e. much closer to the Vulgate, and with dereliquisti for reliquisti. The two printings of one verse disagree in four places. Marked here (the more divergent).
3. 1346D · Dan. 4:34 — Nunc ego Nabuchodonosor… Vulg. Nunc igitur ego. The inferential connective is the hinge of the chapter (Nabuchodonosor praises because he has been restored), and it is gone. Marked. ⚠ Migne also prints a comma inside the lemma, et viae ejus, judicia, which Richard's own exposition four lines later does not have (viae ejus judicia). Punctuation, so carried under Pattern 8, not marked.
4. 1348B · Prov. 16:18 — Scimus quia ante ruinam exaltabitur cor. Vulg. ante ruinam exaltatur spiritus. Both noun and tense differ; unitalicized but carried by a [n:] and functioning as a proof-text, so Pattern 14's "no formatting precondition" applies. Marked.
5. 1347C · Dan. 7:4 — et cor ejus datum est ei. Vulg. et cor hominis datum est ei. ⭐ The most consequential divergence in the range. Marked. The plate's reading is not a slip that can be shrugged off, because Richard expounds it as printed: at 1355D he asks Cujus ejus? Aut cor hominis, aut cor leaenae — "Whose, its? either a man's heart or the lioness's" — a question the Vulgate's cor hominis makes impossible. His copy-text read ejus. Rendered "its heart was given to it", and the beast is "it" throughout the range so that the ambiguity Richard exploits survives into English.
6. 1353D · Dan. 7:4 (re-quotation) — same cor ejus, marked again at its second printing, per the Luke 17:21 precedent in the brief (mark each occurrence).
7. 1348D · Ps. 75:12 — Vovete… qui in circuitu ejus affertis munera. Vulg. omnes qui in circuitu ejus. The dropped omnes narrows the address. Marked. (The note reads (Psal. CLXXV); the Psalter stops at 150. Note kept verbatim — see §D.)
8. 1348D · Eccl. 5:4 — Alioquin melius est non vovere, quam vota non reddere. Vulg. Multoque melius est non vovere, quam post votum promissa non reddere. Marked.
9. 1349A · Ecclus. 2:1 — Fili, accedens ad servitutem Dei sta in timore… Vulg. sta in justitia et timore. A whole member of the precept is gone. Marked.
10. 1350D · Ps. 13:5 + Ps. 73:19 — ⭐ the double printing named in the launch brief, and it is worse than a doublet. At 1350C Migne prints Ps. 73:19 as Ne tradas bestiis animas confitentes tibi (= Vulgate, unmarked); eight lines later, at 1350D, he prints it again as Non tradas… bestiae animas confitentes tibi, joined under one [n:] with Ps. 13:5 Illic trepidabant timore (Vulg. trepidaverunt). So the second printing diverges twice over — negative particle and the number/case of the beasts — and the note conflates two psalms. The plate adjudicates itself three lines later*: Richard paraphrases sed ne tradas bestiis animas confitentes tibi*, i.e. back to the first printing. Marked on the second printing only; rendered "Do not deliver … unto the beast" against the first printing's "Deliver not up unto the beasts", so the difference is visible on the page and not only in this file.
11. 1351C · Ps. 130:1 — qui ambulant in magnis et mirabilibus sub se. Vulg. neque ambulavi in magnis, neque in mirabilibus super me. Marked. ⚠ The plate contradicts itself within two columns: at 1353A the same tag is quoted ambulare in magnis et mirabilibus supra se, which is what the argument needs (presumption reaching above itself). sub se at 1351C makes the first class of bestial men walk in great things beneath themselves, which is the opposite of the vice being described. Rendered as printed at both sites; marked at 1351C.
12. 1354A · Abd. 4 — ⭐ the third printing of this verse in the work, and the three disagree. 1265D (0016): Si exaltatus fueris ut aquila, et si inter sidera coeli posueris nidum tuum, inde detraham te. 1272B (0019): Sed si exaltatus fueris ut aquila, et inter sidera posueris nidum tuum: inde detraham te. 1354A (mine): Si exaltatus fueris, ut aquila, et inter sidera coeli posueris nidum tuum, inde traham te. Vulgate: …et si inter sidera posueris… inde detraham te. So coeli is the plate's own addition (twice of three), the second si survives once of three, and only this printing loses the prefix of detraham. That loss is directional: detrahere is to drag down from the stars, trahere merely to draw — and the whole chapter is about wings being plucked off and the beast falling. Marked, and rendered "thence will I draw thee" as printed.
13. 1357C · Ps. 43:16 — …cooperuit me a facie exprobrantis et obloquentis. Vulg. a voce exprobrantis et obloquentis. Marked. Note that the divergence runs against Richard's own use: the chapter is about the tongue that speaks against, so a voce is the reading his argument wants and the plate gives him a facie. Diffed against 1258B (chunk 0012), which prints the same verse without the a facie/voce tail at all — so the work quotes Ps. 43:16 twice, at two different lengths, and the longer printing is the divergent one.
14. 1357C · Ps. 93:21 — Captabant… condemnabant (imperfects). Vulg. Captabunt… condemnabunt (futures). Marked. Diffed against 1341B (chunk 0048), which prints the same verse with the Vulgate's futures — the work disagrees with itself, and this is the aberrant printing.

(Numbered 1–14 for reference; twelve distinct [var:] markers, two of them the second occurrence of an already-marked divergence.)

[sic: …] — Pattern 12 (carried type)

Two fired, both settled by corpus control before firing.

• 1346B · cruditur — ex ipso lapsu suo, homo saepe ad veram humilitatem cruditur. There is no verb crudire; the sense required is eruditur, and it is the work's own title word (De eruditione hominis interioris, and the ratified table's erudire = "to instruct"). Marked [sic: cruditur] and, per 7a⁗ question 1, the sense is supplied in plain words beside the marker — "is often instructed [sic: cruditur] unto true humility" — because the marked run was the clause's only predicate. ⚠ Control run, and it is honest rather than flattering: grep -rlw cruditur sources/pl/tei/ returns four files, not one — 9324 (ipsorum locorum testimoniis cruditur), 11703 (secundo cruditur ignorantia), 8486 (Israel ergo, reservatis gentibus cruditur) and this one. Every one of the four is a context that requires erudi-. So this is not a hapax defect but a recurring c-for-e accident in Migne's plates, and the marker is correspondingly weaker than the 11550-style hapax cases. It still stands, because the form corresponds to no lexical entry and therefore cannot be rendered at all (Pattern 7's non-word class); Pattern 9 exempts attested medieval spellings, and this is not one.
• 1351D · erudelitas — in tertio calliditas, in quarto erudelitas. Control: hapax in the whole PL TEI (grep -rl erudelitas returns only 11535). The plate corrects itself four lines later — In quarto regnat crudelitas — and the vice-list is fixed by ratified item 6's neighbourhood. Marked; no plain-word supply needed, since the clause's verb (regnat) is carried from the previous member and no negation is involved.

B. Divergences and oddities logged, deliberately not marked

Recorded because the brief asks for the whole divergence list, and because a later sweep should not have to re-derive why each was declined.

• 1348C · Eccl. 5:3 — displicet enim Deo infidelis, et stulta promissio for Vulg. displicet enim ei. An explicitation; the English is identical either way.
• 1349B · Jas. 1:14 — Unusquisque enim tentatur concupiscentia sua for Vulg. Unusquisque vero tentatur a concupiscentia sua. Connective swap; the ablative without a renders the same.
• 1349C · 1 Mach. 7:50 — Siluit… terra Juda diebus paucis for Vulg. Et siluit terra Juda dies paucos. Ablative for accusative of duration; no sense at stake.
• 1350D · Deut. 32:24 — Dentes bestiarum immittam in eis for Vulg. in eos. Same. (The note reads (Deut. XXXVIII); Deuteronomy has 34 chapters — see §D.)
• 1352A · Eccl. 4:10 — non habet sublevantem for Vulg. non habet sublevantem se.
• 1352A · Ecclus. 2:14 — Vae duplici corde, et terram ingredienti duabus viis against the Vulgate's fuller Vae duplici corde, et labiis scelestis, et manibus malefacientibus, et peccatori terram ingredienti duabus viis. Elision of clauses, not variation of wording — declined by the stated policy, though peccatori is a real loss.
• 1352B · 1 Cor. 13:8 — Nam charitas est, quae nunquam excidit recasts Charitas nunquam excidit as a relative clause to fit Richard's sentence.
• 1351B/C, 1351C · Ps. 9:29(10:8), Luke 18:2, Matt. 24:12 — all pluralized or recast into Richard's syntax (qui sedent… ut interficiant; qui nec Deum timent nec homines verentur; abundante iniquitate for quoniam abundavit iniquitas). Author's weaving, not the plate's divergence.
• 1354D · 1 John 2:19 — Exierunt a nobis for Vulg. Ex nobis prodierunt. Different verb, identical English ("they went out from us"); declined.
• 1357D · Ps. 26:2 — Appropiant super me nocentes drops the Vulgate's opening Dum, turning a subordinate into a main clause. No assertion changes.
• 1358A · Ps. 5:11 + Ps. 13:3 — the two half-verses are conflated under one note (Psal. V). Diffed against 1253B (chunk 0009), which prints the second half alone (Linguis suis dolose agebant, venenum aspidum sub labiis eorum) under (Psal. XIII), and against 1272-ish in 0019, which quotes the same half unitalicized. The conflation is the plate's, the wording of each half is the Vulgate's; nothing marked.
• 1355C — Deus superbis resistat, humilibus autem det gratiam is noted (Luc. IV); the text is Jas. 4:6 / 1 Pet. 5:5. Note kept verbatim (§D).

C. Passages that would not construe, rendered as printed

• 1345C · qualis ejusmodi fuerit — Debet autem anima aliquando diligenter quaerere, qualis ejusmodi fuerit. qualis and ejusmodi stack two "of what sort" expressions on one clause. Control: "ejusmodi fuerit" returns zero hits in the PL TEI outside this site. Rendered flatly — "what sort of thing of this kind it was". No conjecture proposed; quondam would give good sense but I have no evidence for it.
• 1346B · Amplificatio magnitudinis, ex virtute humilitatis adjuncta, divinitus gratia confirmationis. — verbless. Rendered with the copula supplied as plain text (no brackets, per the brief), reading adjuncta with virtute: "The amplification of greatness, from the virtue of humility adjoined, is divinely the grace of confirmation."
• 1348A · quod tertius in seipso proprio experimento probavit — "which he, a third, proved in his own self by his own trial". tertius has no antecedent: the subject is God, the experiencer Nabuchodonosor. Rendered literally. Likeliest conjecture ipse; offered here with low confidence and not in the text.
• 1348A · Ante forte illud… voluit… imprimere… ? — the sentence opens with Ante and closes with a question mark that only an interrogative particle would license; An forte would give exactly the printed punctuation. Declined as a [cj:] — see §E. Rendered "Beforehand, perhaps, he wished…?", question mark carried per Pattern 8.
• 1349B–C · animus ad semetipsum redit atque redarguit acriterque reprehendit — the two latter verbs have no expressed object. Rendered without supplying one ("returns unto itself, and confutes, and sharply rebukes"); the next clause (Sibi irascitur, sibi indignatur) supplies the sense for the reader without our supplying a word.
• 1354C · Sic utique, sic saepe quod ex cordis levitate praesumitur, post in multam mentis fluctuationem in amaritudine. — no main verb, and in amaritudine is ablative where the motion-construction wants accusative. Rendered as the printed words allow, verbless.
• 1354C · de coelo amissis aliis dejecta — "cast down from heaven, the others having been lost". alis (wings) is almost certainly meant: the sentence is entirely about the beast's alae, which were plucked off eight lines earlier. A single-letter intrusion, a real word, not a negation or correlative — declined, §E. Rendered literally.
• 1356D · animus nondum et apertam malitiam effrenatur — et where ad is wanted (effrenari ad takes the goal). Rendered "not yet unbridled even to open malice", carrying the negation intact. Declined as a conjecture (§E).
• 1358B · Qui enim loquitur, inquam, non potest latere. — the argument needs "he who detracts"; loquitur renders in place and is left alone.

D. Notes ([n:]) that misfire — all kept verbatim

Four citation numerals in the range are impossible or wrong, and all four are reproduced exactly as Migne prints them, per the standing rule that [n:] contents are his citations and the index's refKey/refDisplay pair resolves them: (Psal. CLXXV) for Ps. 75 (1348D) · (Deut. XXXVIII) for Deut. 32 (1350D) · (Psal. XXV) for Ps. 45 (1354D) · (Luc. IV) for Jas. 4:6 / 1 Pet. 5:5 (1355C). Also (Prov. XXIII) for Prov. 25:14 (1347D), (Eccle. III) for Eccl. 5:3, (Eccle. II) for Ecclus. 2:14, (II Cor. XIII) for 1 Cor. 13:8, and (Psal. XCIII) for Ps. 93:21 in the Vulgate numbering Migne is using elsewhere. Not defects of ours; not marked; listed so a later indexing pass is not surprised.

E. Conjectures declined — census candidates, named as the brief requires

Wilson's bar (a [cj:] is glossed only where the wrongly-printed real word is a negation or a correlative) is applied and not widened. Five sites:

1. 1348A · Ante for An (0050). The nearest to the line in this range: an interrogative particle governs the mood of the whole sentence, which is 7a⁗'s own question 3, and the printed ? is otherwise unexplained. Declined because Ante is neither a negation nor a correlative, and because 7a⁗ governs markers, not bare Pattern 7 renderings. If the bar is ever widened to interrogative particles, this is a candidate.
2. 1354C · aliis for alis (0053). Noun; declined.
3. 1356D · et for ad (0054). Preposition/conjunction; declined. (Note it is not a swallowed conjunction in the 7a⁗ sense — nothing is quarantined here, the negation nondum survives intact in the English.)
4. 1348A · tertius (0050). No antecedent; declined, and no confident conjecture.
5. ⭐ 1357A · redere for rodere (0054) — the one I most nearly marked, and the control is what stopped me. The clause reads proprie proprium sit invidiae alienam vitam redere et morsibus obtrectationum lacerare, and nine lines later the same idea is qui alienam vitam libenter rodunt — so rodere, "to gnaw", is certain from the plate itself. I was about to fire a [sic:] on the non-word class. The corpus control forbids it: grep -rlw redere sources/pl/tei/ returns four files, and the other three (8611 redere et confirmare, 8744 necesse est redere rationem, 11578 debent redere coronas) all use it as an attested medieval single-d spelling of reddere. So redere is a real word (Pattern 9 territory), merely the wrong one (Pattern 12's "fuges/munde" class), and it renders in place. Rendered "to render another's life", which is odd but not false, and logged here. This is a Pattern 18 census candidate on the noun-slip side of the bar.

Also declined and worth naming for the same census: 1356A · ei non agnus sed ut sus succedit (0053). Read ursus: the very next chapter opens bestia alia, similis urso and closes Sic pardus urso succedit — the identical X <dative> succedit frame. So "not a lamb but as it were a sow" almost certainly conceals "not a lamb but a bear", i.e. Pattern 10 split type (ur sus → ut sus), the same shape as the corpus's known ur sam for ursam. Declined as a marker because both fragments are real words forming a grammatical phrase that renders in place — Pattern 12's stated test — and declined as a [cj:] because a noun is neither a negation nor a correlative. Rendered "there succeeds it not a lamb, but as it were a sow". ⚠ This is the site in the range where the faithful English most misleads, and it is here rather than in the text by the rules as they stand; if the [cj:] bar is ever widened, read this one first.

F. Lemma pulls resisted (positive evidence, per the brief)

Places where Migne agrees with the received text and the instinct to "improve" was live:

• 1345C · Dan. 4:33(36) — Et optimates mei, et magistratus mei requisierunt me… is exactly the Vulgate, doubled et and all. The pull was to smooth the doubled conjunction; resisted, and the doubling is in the English.
• 1346C · Ps. 67:29 — Migne prints Confirma Deus hoc quod operatus es in nobis where the Vulgate has Confirma hoc Deus, quod…. Word order only; no marker, and the English follows the printed order.
• 1346C · 1 Kgs (1 Sam.) 2:6 and 1346B · Ps. 50:14 — verbatim Vulgate; the pull here was toward the received English ("bringeth down to hell, and bringeth back") which happens to be right, so it was allowed under the familiar-verse licence.
• 1350C · Ps. 73:19, first printing — verbatim Vulgate. Marking it because its neighbour eight lines later diverges was the live temptation; resisted. Only the divergent printing carries the marker.
• 1353A · Prov. 30:30 (Leo, fortissimus bestiarum, ad nullius pavebit occursum), 1354D · Ps. 45:3, 1355C · Matt. 20:15, 1356B · Dan. 7:5, 1357C · Ps. 37:21, 1358A · Joan. 13:27 and Eccl. 10:11, 1350D · Ps. 103:20, 1347D · Prov. 13:4 and Prov. 25:14, 1349B · Ps. 4:5 — all verbatim Vulgate, all checked word by word, none marked. ⭐ Dan. 7:5 is worth naming: the bear-lemma, the whole basis of chunk 0054, diverges from the Vulgate nowhere at all — which is what makes the cor ejus divergence in the lioness-lemma credible rather than a scanning artefact of ours.

G. Vocabulary — calls made, and one I would like ratified for the whole work

Following the ratified table without exception; mens mind / animus spirit / anima soul are kept visibly distinct throughout (they stand in adjacent clauses repeatedly in 0051 and 0053, never in one clause), fortitudo is "fortitude" at 1345D, scientia is "learning", cognitio "knowledge", intelligentia "understanding", discretio "discretion", detrimentum "decline", gradus "step", praesumptio "presumption", pertinacior "more obstinate", saevum "savage", simulatio "dissimulation" and simulare "to dissemble", impudentia "shamelessness".

Requested for ratification — Book Three's own vice-chain, which is a second four-term scheme running alongside the ratified arrogantia / jactantia / pertinacia / saevitia, and which will recur in every remaining chunk of Book Three (one beast per vice, and 0054 already hands the leopard to the next agent):

And the three degrees of the bear's teeth at 1357A–B, which must also stay distinct because Richard defines each against the others: oblocutio = speaking-against · derogatio = derogation · detractio = detraction. (oblocutio has no one-word English; the hyphenated noun was chosen over "reproach", which is already exprobratio in the Ps. 43 lemma printed in the same paragraph.)

Two further pairs held apart deliberately: infrenata = uncurbed vs effrenata = unbridled malice (1351D, printed four words apart in one list), and praecipitatio = precipitancy vs praesumptio = presumption (the two eagle-wings, 1353B).

⚠ One tension logged against my own judgement, as the brief requires. At 1351A the Latin is zelum rectitudinis habeant, sed secundum scientiam non habent — an unmarked echo of Rom. 10:2, where every English ear wants "not according to knowledge". The ratified table gives scientia = learning and cognitio = knowledge, and the quotation is neither italicized nor noted, so the scripture-echo licence does not reach it. Followed the table: "not according to learning". Flagging it in case the table wants an exception for Rom. 10:2 specifically.

H. Italics added where the plate has none — declared

Three sites where Richard glosses a Latin word as a word and the plate prints it in roman: 1352D (insolitum / non solitum / insolentia), 1356A (invidens / invidentia), 1357B (rogare / derogare). In each the Latin word is kept in the English and set in italics, on the mention-vs-use principle; without the word the gloss explains nothing, and without the italics it reads as a typesetting error. This is the only place in the range where italic structure is not 1:1 with the Latin twin. Declared here rather than done silently; if a verifier enforces strict italic parity, these are the three.

Richard of St Victor, De eruditione hominis interioris, Book III, CAPUT XI–XIX. This range holds the third and fourth beasts of Daniel 7 (the leopard; the fourth beast with iron teeth and ten horns) and the little horn with the eyes of a man — and it closes the work.

A. ⭐ The ending of the work — read this first

The treatise does not close its argument. It breaks off mid-exposition.

• The last printed words, at 1366A, are: Et quid est quod bestiarum potestas aufertur, nisi, quod per visitantem gratiam, vitiorum tyrannide sopita, regnum peccati destruitur? → "And what is it that the power of the beasts is taken away, unless that through visiting grace, the tyranny of the vices being lulled to sleep, the kingdom of sin is destroyed?"
• It is not truncated mid-sentence. That sentence is grammatically complete and closes with the question mark Migne prints. Nothing is broken at the plate's edge.
• It IS truncated mid-exposition, and demonstrably so from the text's own promise. The final paragraph opens by announcing the restoration half of the argument — Audiamus quomodo, vel unde eadem ipsa quandoque, post multa et ultima mala, ad rectitudinis integritatem reformatur — quotes Rom. 5:20 as its programme, and then delivers only two of the Quid est quod …? questions on Dan. 7:9–12 (the throne set and the Ancient of days; the beasts' dominion taken away). It never reaches the Son of man, and it never reaches the kingdom that shall not be destroyed — which is the figure the whole work is built toward, expounded at length from Dan. 2:44 in Book I (chunk 0027, col. 1288D) and promised again here. The chapter head standing over this material is still CAPUT XIX. Quomodo mens cum in profundum malorum venerit, saepe etiam in desperationem cadit. — a chapter on despair, under which the restoration paragraph sits with no head of its own.
• Chunk 0057 is 552 Latin words against a work median near 1,000, consistent with a final chunk that stops early rather than one that fills its columns.
• ⚑ This is worth a sentence in the site's "On this text" note: the work as Migne prints it ends unfinished, and a reader who reaches 1366A expecting the restoration Richard has just promised will otherwise think our text is defective.

Column-band note, checked and NOT reported as a finding. The anchors run … [1364D] [1365A] [1366A], skipping 1365B–D, and only ~55 Latin words stand between 1365A and 1366A. I checked it against runbook false-positive 1 (column-band arithmetic) and the text does not break mid-word or mid-construction across the boundary — quod in Apostolo legimus: Quoniam ubi abundavit [1366A] iniquitas, superabundavit et gratia reads continuously. So: no [ed:], no tei-patch, no plate check. Recorded here only so a later reader does not re-open it.

B. Vulgate divergences marked [var: …] (9 markers)

Divergences noted and deliberately NOT marked:

• 1364D, Rom. 9:16 — Migne Non est utique volentis neque currentis sed miserentis Dei against Vulg. Igitur non volentis, neque currentis, sed miserentis est Dei. Particle (utique for Igitur) and the position of est; no change of sense, so no marker. Logged because the list is the deliverable.
• 1361D, 1362A — Richard's own re-quotations of Dan. 7:7 drop autem (Dissimilis erat, inquit, caeteris bestiis). Trivial; no marker. Per the brief's rule I diffed the two printings before translating either: they otherwise agree word for word.
• 1366A, Dan. 7:9 — posito throno, antiquus dierum sedit against Vulg. throni positi sunt, et antiquus dierum sedit. This is not a quotation; it is woven into Richard's own Quid enim est quod … construction (singular throno because he is naming one figure). Pattern 14 stays reserved for scripture quoted as scripture.
• 1364D — Oportet, inquiunt, esse … quis Dei voluntati … resistere potest? echoes Rom. 9:19 but is put in the mouth of the objectors and is not quoted as scripture. No marker.

C. ⭐ The two lemma diffs the launch message named

C.1 Jer. 5:3 — 1276C (chunk 0021) vs 1363B (mine). They differ, and mine is the divergent printing.

• 0021 @1276C: Percussi filios vestros, et non doluerunt, renuerunt recipere disciplinam.
• 0056 @1363B: Percussi eos, et non doluerunt.
• Vulgate Jer. 5:3: Percussisti eos, et non doluerunt; attrivisti eos, et renuerunt accipere disciplinam.

So the two printings diverge from each other in the object (filios vestros / eos), and my printing is the one that agrees with the received text there. But both carry the first-person Percussi for the Vulgate's second-person Percussisti — the received verse is addressed to God ("Thou hast struck them"), and Migne's makes God the speaker. That is the material divergence and it is what my [var:] names; the marker also cross-refers 1276C so the reader can see the work printing the same verse two ways. 0021's shipped [var:] already covers its own printing, correctly; I have not duplicated its content.

C.2 Isa. 47:2–3 — 1272B (chunk 0019) vs 1361A (mine). ⚠ The divergence is in 0019, not in mine.

• 0019 @1272B: … transi flumina. Et tunc revelabitur ignominia tua, et videbitur opprobrium tuum.
• 0055 @1361A: … transi flumina, revelabitur ignominia tua, et videbitur opprobrium tuum.
• Vulgate Isa. 47:2–3: … transi flumina. Revelabitur ignominia tua, et videbitur opprobrium tuum.

My printing agrees with the received text; 0019's adds Et tunc. I therefore fire no [var:] here — this is the brief's "mark the divergent one, not both" rule, and the divergent one is not mine. ⚑ But 0019 is shipped and its Et tunc is unmarked. That is a live, unmarked Pattern 14 site in a peer range; flagged for the merge step, not fixed by me. (0019 also prints a full stop where mine prints a comma; I have kept my plate's comma per Pattern 8, which is why the two Englishes are punctuated differently — that is deliberate and should not be "reconciled" at merge.)

I have kept 0019's English wording for the verse (Uncover thy shame, lay bare thy shoulder, make bare thy legs, pass over the rivers) so the work speaks with one voice where the Latin agrees.

D. Lemma pulls I RESISTED (positive evidence — Migne agrees with the received text here)

Every one of these is a verse whose familiar English arrived before the printed Latin, and in each the printed Latin turned out to be exactly the Vulgate. No marker fired.

1. Dan. 7:7 @1361B, the fourth beast — Post haec aspiciebam in visione noctis, et ecce bestia quarta terribilis atque mirabilis et fortis nimis. Dentes ferreos habebat magnos, comedens atque comminuens, et reliqua pedibus suis conculcans. Dissimilis autem erat caeteris bestiis quas videram ante eam, et habebat cornua decem. Word-for-word the Vulgate, all 41 words. I checked mirabilis (the pull was toward Douay's "wonderful" as if it were a gloss), fortis nimis, and reliqua … conculcans individually. Rendered "terrible and wonderful, and exceeding strong" — the traditional cadence is allowed here precisely because the Latin matches.
2. Dan. 7:8 @1364A, the little horn — Considerabam cornua, et ecce cornu aliud parvulum ortum est de medio eorum, et tria de cornibus primis evulsa sunt a facie ejus. Et ecce oculi quasi oculi hominis erant in cornu isto, et os loquens ingentia. Exact Vulgate. The strongest pull in the range was to render parvulum as the familiar "little horn" and move on without checking evulsa sunt a facie ejus; checked, it stands.
3. Isa. 52:2 @1363A — Solve vincula colli tui, captiva filia Sion. Exact.
4. Ps. 115:16–17 @1363A — Dirupisti vincula mea, tibi sacrificabo hostiam laudis. Exact for the words quoted. (Migne stops before et nomen Domini invocabo; a quotation that simply stops is not a divergence.)
5. Ps. 82:14 @1363C — Deus meus, pone illos ut rotam, et sicut stipulam ante faciem venti. Exact.
6. Ps. 31:9 @1363B — sicut equus et mulus quibus non est intellectus. Exact for the fragment quoted. See §F.3 for the vocabulary tension this one raises.
7. Prov. 10:9 @1360D — Qui ambulat simpliciter, ambulat confidenter, qui autem depravat vias suas manifestus erit. Exact. I resisted Douay's "walketh sincerely" — see §F.1, the word is a gloss hook.

E. Markers fired for defective type

### E.1 [sic: exerescit] @1358D — fraus in fraudulentiam exerescit Non-word. Control run: across the whole PL TEI (5,276 files, 87M words) exerescit occurs in one file — this one, at this site — against 113 files printing excrescit. That is the corpus's own verdict: plate defect, not a period spelling. 7a⁗ check: the marked run is the clause's only verb, so the English says the sense in plain words beside the marker — "fraud grows [sic: exerescit] into fraudulence". Conjecture: excrescit.

### E.2 [sic: eum] @1359A — Certe eum volucres volant A real word with no grammatical slot: an accusative pronoun in a clause whose verb (volant) is intransitive and which already has its subject. Pattern 12's test 1. The next sentence is the parallel Certe cum volare volunt, which makes cum certain. 7a⁗ check, question 3: the marked run carries the clause's conjunction, so the English supplies it — "Certainly when [sic: eum] birds fly". Without that the sentence loses its subordination entirely.

### E.3 [sic: desidorium] @1363B — Aliud est desidorium suum cohibere non posse Non-word. Control run: desidorium occurs in one file corpus-wide (this one) against 1,377 files printing desiderium. 7a⁗ check, question 1: the marker swallows the clause's object, so the English supplies it — "not to be able to restrain one's own desire [sic: desidorium]". Conjecture: desiderium.

F. Vocabulary calls, including one I want ratified for the whole work

F.1 ⭐⭐ simulatio vs dissimulatio — the ratified equivalence CANNOT hold in this chunk, and I want a work-wide ruling

cruces-0020.md §7 fixed simulatio = "dissimulation" (beside the brief's ratified simultas = "pretense"), and chunks 0018–0024 ship it that way. CAPUT XII of my range breaks it, because at 1359C Migne sets the two nouns side by side and makes the distinction between them the whole point of the chapter:

Simulatio et dissimulatio est de occultis. Sed simulatio de occultis bonis, dissimulatio vero de occultis malis.

Four wings, four vices — simulationis, dissimulationis, ostentationis, excusationis — and simulatio is feigning a good one does not have while dissimulatio is hiding an evil one does. Under the ratified equivalence, dissimulatio has nowhere to go and the chapter's four-fold scheme collapses to three. This is the testa/fictilis situation exactly: two words printed side by side, the exposition turning on which is which, so they must never land on one English word.

What I did, and why: in this range simulatio = "simulation", dissimulatio = "dissimulation", simulator = "simulator", simulatrix mens = "simulating mind". English's own simulation/dissimulation pair carries Richard's distinction precisely (feigning what is not / concealing what is), which is more than any substitute would.

The ratification I am asking for (merge-step decision, not mine to take): sweep simulatio → "simulation" work-wide, freeing "dissimulation" for dissimulatio, and leaving simultas = "pretense" untouched. That touches six shipped sites, all rendering simulatio as "dissimulation" — 0018 (1270A), 0019 ×2 (1272B area, 1273B), 0020 (1274A), 0024 (1283B), 0027 (1289C) — plus the simulatum / simulati participles at 1283D and 1289C, which already read "simulated" and would need no change. The alternative — leaving the work with "dissimulation" meaning two opposite things in two places — is worse. If the ruling goes the other way, CAPUT XII still cannot use one word for both, and whatever is chosen for dissimulatio here has to be recorded in the brief.

### F.2 New terms fixed in this range (peer agents on 0030–0054 have not been visible to me) | Latin | English | note | |---|---|---| | fraudulentia / fraus | fraudulence / fraud | the vice-name of CAPUT XI–XIV; not attested in 0000–0029 | | impudentia | shamelessness | brief-fixed; impudens "shameless" | | contumacia | contumacy | ⚠ NOT "obstinacy" — the brief reserves obstinacy for pertinacia, which occurs in my 0057 (diffidentiae pertinacia). The two stand within two columns of each other | | extollentia | self-exaltation | keeps the attollit hook of its own gloss and stays clear of elatio | | arrogantia | arrogance | brief-fixed | | affectatio | the coveting of | ambitio = ambition, ambire = to court (0029 precedent). "Affectation" now means pretense in English and would collide with §F.1 | | obligatio / obduratio / effrenatio mentis | binding / hardening / unbridling of the mind | obduratio brief-fixed; effrenatio follows 0020's "unbridling of superfluity" | | abjectio, objectio, dejectio, subjectio | abjection, objection, dejection, subjection | see §F.4 | | diffidentia | diffidence | 0057, beside desperatio = despair (brief-fixed) | | magisterium | mastery | brief-fixed, and it lands at 1363D | | detrimentum | decline | brief-fixed; its last occurrence in the work is 1366A |

### F.3 Followed the fixed vocabulary against my own judgment 1363B, Ps. 31:9 — sicut equus et mulus quibus non est intellectus. The brief's ratified item 3 gives intellectus = intellect* (reserving "understanding" for intelligentia), so the psalm reads "as the horse and the mule which have no intellect". Every instinct wanted the traditional "understanding" here, and the scripture-echo licence would arguably have allowed it. I followed the table, because the psalm is quoted inside the definition of mentis effrenatio* — a psychology passage, which is exactly the seam the fixed table exists to protect. Logging it as a followed-against-judgment call in case the merge wants the licence applied instead.

### F.4 The four -jectio horns @1363D cornu abjectionis, objectionis, dejectionis, subjectionis — Richard builds a four-member figura on one root and glosses each member on the spot. I rendered them abjection, objection, dejection, subjection, which reproduces the figure exactly. ⚠ "Abjection" is being used in its etymological sense (casting off — his gloss is quando jam dedignatur obedire), not the modern English sense of degradation. The gloss follows immediately in every case, so the reader is not left to guess; recorded here because a later reader might otherwise think it a mistranslation.

G. Passages that will not construe (rendered as literally as the words allow)

1. 1361B — …ad omnia se praecipitem dare, et conspectum divinum vel vultus hominum, nec in erubescendis erubescere… The accusatives conspectum divinum vel vultus hominum have no governing verb; something like non revereri is wanting. Rendered with the hole visible: "and the divine gaze or the countenances of men, not to blush even at things to be blushed at". No marker — nothing here is defective type, and a [sic:] would have to swallow a whole clause, which §"Scripture and garbled type" of the brief forbids.
2. 1360C — Docet nos ille illius Achitopel loquens quid in quibusdam possit ambitio auctoritatis. ille illius will not construe with a genitive proper name and a participle. The likely intention is illud illius Achitophel consilium ("that counsel of that Achitophel"), which would match the balancing clause docet nos proditor et parricida Absalon. Rendered as printed: "That man of that Achitopel, speaking, teaches us…". Conjecture recorded here only.
3. 1360D — majora sunt pene quae hypocritae facere possunt, et faciunt et districtioris vitae observantias custodiunt, quam hi qui se propter Deum affligunt. The comparative majora … quam has two finite clauses shoved between its members, so the quam clause has no clean term to compare with. Rendered as printed.
4. 1363A — ut se a peccati consuetudine implicare, etiam cum velit, non valeat. The verb should be explicare ("disentangle"); as printed the clause says the mind cannot entangle itself from the custom of sin, which is the opposite of the sentence's point (and the printed implicat four words earlier is what explicare would answer). Rendered as printed — "cannot … entangle itself from the custom of sin". Declined as a [cj:]: see §H.
5. 1362B — cornua a carne excrescunt, et supra carnis sensibilitate obdurescunt. supra + ablative. Rendered "harden above the sensibility of the flesh". Pattern 9 territory at most; no marker.
6. 1361D — …qui verbis et operibus in proximos saevit, De alia bestia dictum est… Migne prints a comma and then a capital D beginning a new sentence. Pattern 8: preserved exactly, comma and capital both. The English therefore reads "…against his neighbours, Of another beast it was said…". Not our punctuation.

H. Conjectures DECLINED — census candidates, named as the brief asks

Wilson's [cj:] bar (2026-08-15) glosses only a negation or a correlative. None of these is either; all four are named here rather than widened into the running text.

1. ⭐ 1360C — aspirare ab subdiaconatum. A real word, wrongly chosen: ab governs the ablative, subdiaconatum is accusative, and the two parallel members that follow in the same sentence both print ad (subdiaconos ad diaconatum, diaconos ad presbyteratum). Control run: ad subdiaconatum stands in 17 PL files; ab subdiaconatum in one — this. So ad is near-certain. I rendered the printed word: "those who are of a lower grade aspiring from the subdiaconate", which is fluent and says the opposite of the sentence's own gradient. This is the nearest thing to a Pattern 18 site in my range — it is exactly the munde/mundo shape, a real word whose faithful English asserts something the author is arguing against. It is declined because a preposition of direction is neither a negation nor a correlative, and the bar is not mine to widen. If the census ever reaches for a preposition class, this site should be its worked example.
2. 1363A — implicare for explicare (§G.4). A prefix, not a negation. Declined.
3. 1362C/1363B — mentis obduratio announced, cordis obduratio delivered. CAPUT XVII sets out the three steps as mentis obligatione, mentis obduratione, mentis effrenatione and then defines the second as Cordis autem obduratio est…; 0057 @1364B looks back at all three as cordis obligatione, cordis obduratione, et effrenatione. Rendered as printed each time — "of the mind" where Migne writes mentis, "of the heart" where he writes cordis. ⚠ This one matters for the fixed vocabulary, since mens and cor are separately fixed terms in this work and the plate is inconsistent inside a single scheme. Not a defect I can mark; recorded so the inconsistency is attributed to Migne.
4. 1362C — desertus … praevaricator efficitur, and 1362B tribus apud semetipsum. Masculine agreement where the subject is mens/anima (feminine). Pattern 9: Latin grammatical gender has no English exponent. No emendation to make or undo; flagged so a later sweep does not churn it.

I. Smaller notes

• 1361D — the claws Daniel does not give. The lemma at 1361B has only dentes ferreos; Richard's exposition twice adds ungues (Dentes autem et ungues ferreos habere dicitur; sicut in ferreis unguibus conculcat inferiores @1363D). Dan. 7:19 does supply ungues eius aeneos — bronze, not iron. This is Richard's own conflation inside his exposition, not a quotation, so no [var:]; noted because a reader checking the lemma will find no claws in it.
• [n: (Ephes. I)] twice, and neither time is it Ephesians. 1362C prints pro bona voluntate (Phil. 2:13) and 1364D prints omnia in omnibus … prout vult (I Cor. 12:6, 12:11), both under a note reading (Ephes. I). Two occurrences in three chunks looks like a habit rather than an accident; both are marked, and the pair is offered to whoever is keeping the plate-habit list beside the regnum coelorum and de monte entries.
• Rom. 5:20 printed twice in the work, both times with iniquitas (0027 @1288C with a note pointing to Matt. XXIV, mine @1366A with no note at all). Consistent within the work; both marked.
• 1364B wordplay — animum non tam tumidum quam timidum reddit. Rendered "not so much swollen as timid", which keeps the sense and loses the jingle. No English pair carries both; recorded rather than forced.
• Pattern 17 — Vultis adhuc plenius nosse (1359B) is plural and takes "you", matching 0019's Vultis autem nosse. The singular addresses in this range (Attende, obsecro 1361B; Habes ergo 1359D; Ut te excuses 1364D) take thou.
• Guillemets: Migne prints none in this range; none supplied. Question marks: 8 / 3 / 8 across 0055–0057, matching the Latin twins exactly; the ! at 1362A (oppugnat superiores!) is Migne's and is kept. The ? closing Deum blasphemas? @1364D and the period closing the Nonne … question at 1364D are both the plate's and both kept.
• Ratios: 1.46 / 1.46 / 1.48 (Latin words → English words, [var:] content excluded).

Richard of St Victor, De eruditione hominis interioris, Book I, chapters I–VI. Opening range: Daniel 2:1–11 quoted in running order, with fragments re-quoted and glossed.

Conventions set here for the whole work (peer agents follow these)

Heads. ## LIBER PRIMUS. → ## BOOK ONE. · ## CAPUT PRIMUM. → ## CHAPTER ONE. · every numbered head keeps the plate's own numeral (## CAPUT II. → ## CHAPTER II., never renumbered, never normalized to a word). The italic summary is translated and stays italic; the plate's terminal period is kept where it prints one. The three blank lines the plate leaves between LIBER PRIMUS. and CAPUT PRIMUM. in 0000 are mirrored 1:1 (Pattern 8a).

The name. Nabuchodonosor, not "Nebuchadnezzar", throughout. Chapter I turns on two etymologies of the name — prophetans istiusmodi signum (1231D) and sessio in agnitione angustiae (1232A, 1233A) — and an English form leaves both glosses explaining nothing, which is exactly the case the brief's etymology rule and Pattern-7 §2 (stylus/epistylium) cover. The corpus is split (4 shipped works print Nabuchodonosor, 7 print Nebuchadnezzar), so this is a per-work call and it is submitted for ratification. Other names take their received English (Daniel, Chaldeans, Jerome, Gregory, David, Uriah, Babylon, Syria).

Vocabulary added to the fixed table (proposed for the whole work): angustia = anguish — the hook the etymology hangs on (sessio in agnitione angustiae · quam sine angustia nosse non poterat · cum dolore et animi angustia) · conjectura = conjecture, kept distinct from interpretatio = interpretation (Migne prints both in the Daniel lemma and Richard glosses only the second) · publicare = to make public, never "confiscate" (Douay), because Richard's gloss is built on it: Quid autem est sapientium domos publicare, et ad communem omnium usum redigere? · conatus = endeavour · instantia = urgency · pudentia = modesty (see crux 28).

Markers fired (7 sites)

[var: …] — Pattern 14, scripture divergence (3)

1. 1233A · Rom. 13:1 · [var:] FIRED Printed: Quae autem a Deo sunt, teste Apostolo, ordinata sunt [n: (Rom. XIII)]. Vulgate: quae autem sunt, a Deo ordinatae sunt. Migne draws the agent a Deo into the relative clause and prints the neuter ordinata for the feminine ordinatae. Material: the Vulgate says the existing powers are ordained by God; Migne's line says the things that are from God are ordered — and that is precisely the load Richard puts on it, the unjust king's voluntas inordinata against his potestas ordinata. The preceding Non est enim potestas nisi a Deo is printed roman, unmarked, and agrees with the Vulgate. Rendered as printed; the received reading is in the marker.

2. 1234A · Dan. 2:2 · [var:] FIRED Printed: Qui cum venissent, statuerunt coram rege. Vulgate: steterunt. statuerunt is transitive and leaves the clause with no object, so the faithful English ("set up before the king") states something the received text does not. Control run: statuerunt coram rege occurs once in the whole PL TEI (this file); steterunt coram rege also once. But statuerunt itself is a common word — 1,005 occurrences — so this is a real word wrongly printed, not broken type: Pattern 12 has nothing to wrap. Pattern 18 is excluded by its own rule that a divergence inside a quoted scripture belongs to Pattern 14.

3. 1240A · Luke 17:21 · [var:] FIRED Printed: Regnum caelorum, inquit, intra vos est [n: (Luc. XVII)]. Vulgate: regnum Dei intra vos est. Migne's own note cites Luc. XVII, so the substitution is on the plate, not in the reference. Material: Richard is arguing that the interior kingdom is a kingdom, and caelorum pulls the verse into the orbit of the Matt. 5:3 beatitude he quotes four lines later — which prints the other spelling, coelorum, in the same chunk (crux 39).

[sic: …] — Pattern 12, carried type (3)

4. 1229D · eficiat for efficiat · [sic: eficiat] Control run: word-boundary grep of sources/pl/tei/ returns 1 occurrence in 87M words — this file, this site — against 674 files carrying efficiat. Not a word; carried untranslated. ⚠ 7a⁗ q1 applies: it is the only verb of the si clause, so the sense stands in plain words beside the marker — "should effect [sic: eficiat]" — on the which slays [sic: eccidit] model. Conjecture: efficiat, near-certain.

5. 1234A · sommia for somnia · [sic: sommia] (inside Dan. 2:2) Control run: sommia returns 4 occurrences across 4 different works in the whole PL TEI against 863 of somnia — 0.5%, which reads as sporadic broken type (n → m), not as an orthographic convention, so Pattern 9 does not cover it. ⚠ Marked without asterisks: it sits inside an already-italic scripture lemma and nested italics are not expressible (Pattern 10's closing clause). ⚠ 7a⁗ q1: the marker takes the clause's object, so "his dreams" stands in plain words beside it. Honest weakening: 4 occurrences is thinner evidence than a hapax. If a later census reads the class as Migne's habitual mis-set rather than damage, this marker should come off.

6. 1238C · conanimis for conaminis · [sic: conanimis] Control run: hapax — 1 occurrence in 87M words, this file, this site. Carried on the bonorem precedent: the noun-head is carried untranslated and the conjecture (conaminis, genitive of conamen, "effort") stays here rather than in the text. 7a⁗ q1–3 all pass — the clause keeps its verb (docere), carries no negation, and no conjunction is swallowed.

[cj: …] — Pattern 18, real word, faithful English misleads (2)

7. 1231B · quam stranded without its comparative · [cj: quam; read facilius quam, "more easily than"] Printed: contemplativae vitae gratia quam activae elabitur, sed multo difficilius reparatur. As printed the sentence has a quam with nothing to compare — it does not construe at all. Richard supplies the missing word himself, twice, within eight lines: somnii visionem quam tam facile amisit (1231B) and a contemplativa facilius decidimus (1231C). ⚠ Bar note, flagged for ratification. Wilson's 2026-08-15 bar admits a wrongly-printed negation or correlative, and a stranded comparative correlative is the second class; the 7561 @0648 precedent (solum for non solum) establishes that a dropped word is glossable, because the gloss opens on the word Migne did print. I have not widened the bar beyond that. If the ruling is read narrower than I have read it, this is the one marker in my range to withdraw.

8. 1239A · volebant for nolebant · [cj: volebant; read nolebant, "were unwilling"] Printed: Ignorantiam suam dissimulabant, et aperte fateri volebant, et in hoc tempus redimebant. The printed reading contradicts its own sentence twice: they dissembled their ignorance, and they were buying up the time — which is what one does precisely in order not to confess. Richard's whole chapter is quam multis modis animus seipsum decipit. This is the 11547 @0260D defect (nolebant for volebant) with the polarity reversed, and the word that went wrong is the negation itself — squarely inside the ratified bar.

Conjectures DECLINED — named as census candidates

9. 1233A · angüstiae for angustiae — the etymology sessio in agnitione angustiae is printed twice in chunk 0000, and the second printing carries a diaeresis on the u. Control run: the string angü is a hapax across the entire PL TEI — one occurrence in 87M words, this file, and it is the only umlauted vowel anywhere in 11535. Declined a [sic:] on three grounds, and the call is genuinely marginal: (a) the word renders in place perfectly, so what a reader would meet is our English, not Migne's type — the Pattern 12 operational test; (b) English has no exponent for a diacritic, so a marker would put Latin on the page for a mark that means nothing in either language; (c) I cannot check the plate, and a stray diaeresis is at least as likely to be a Corpus Corporum OCR artifact as Migne's own type, which would make it a data/tei-patches/ question rather than a Pattern 12 one. Both printings render identically ("a sitting in the acknowledgement of anguish"). ⭐ Named for the census: a single plate read of col. 1233A settles it either way.

10. 1232D · quae, qualia sint, possunt — singular relative quae (antecedent rei) against plural qualia sint / possunt. Number anomaly with no English exponent; declined per the bar (not a negation or a correlative). Rendered as closely as the printed words allow, with a parenthetic dash marking the break.

11. 1233D · cognitionumque disciplinam where the parallel eleven lines above (affectionibus … cogitationibus) leads one to expect cogitationum. A real word in its slot; declined. Rendered "the discipline of his knowings".

12. 1236C · hanc nobis sollicitudinem primo ingenerant — indicative among the surrounding subjunctives (satagant … studeamus). Mood anomaly; declined per the bar.

13. 1239A · quomodo illud potuissent quo majus erat? — quo for quod, breaking the parallel with quod minus erat four words earlier. Case anomaly; declined. Rendered "that whereby it was the greater".

14. 1239B · Ad tertiam autem, lucem jam palam profitentur. The object should be their ignorance (cf. ignorantiam suam magis minuunt quam loquuntur in the preceding sentence); lucem is a real word in a real slot and yields fluent nonsense. The strongest of the declined sites and the closest to Pattern 18's bar — but it is neither a negation nor a correlative, so it stays here. Rendered "they now openly profess the light".

15. 1239D · secundum malum quod in seipso invenit, alios adjicit. The antithesis (alios omnino non praedicat … alios adjicit) wants judicat. Real word; declined. Rendered "adds others".

Passages that would not construe — rendered as printed, no marker

16. 1233C · the dangling tertium. Primum studium … ut possit dominari affectionibus; secundum ut possit imperare cogitationibus; tertium ergo animus spiritalia quaeque … incipit habere. The third member of the series has no predicate of its own and tertium is left hanging. Rendered as printed. ⚠ Both co-subjects preserved against the singular verb — animus ("the spirit") and anima ("the soul that is victor over its own vices") — per the brief's warning that 11550 lost four co-subjects to one.

17. 1236B · a sentence break the plate does not print. Quidquid peragas, illi, non tibi ascribas hoc utique regi illi … dicendum erat. No stop between ascribas and hoc. Pattern 8 forbids supplying one; the English runs the two clauses together and reads roughly as the Latin reads.

Pattern 8 — punctuation carried, not repaired

18. Questions printed with a period and left that way (no ? supplied): 1231A Quid enim est quod Nabuchodonosor … agnovit. and Quid, inquam, … reparatur. · 1232A quo, quaeso, in loco … non possit videri. · 1238C Quam multa divinorum judiciorum cognoscimus … penetramus. and O quam difficile humana impudentia … reprimitur. · 1238D Quam multa saepe simulamus, quam multa dissimulamus. · 1241B nihil in ea nisi rectum sonare videtur ubi dicitur.

19. 1236C — O infinita hominis stultitia, [1236C] et recte quidem dicis … sine gratia? Migne prints a comma after the exclamation and a ? at the end of the run. Both carried exactly; no ! supplied, though English wants one there.

20. 1236A — the chunk's one ! is Migne's, closing a statement (… interioris hominis vitam recte intelligimus gratiam divinam!). Carried.

21. 1236A — Gratia, Dei, inquit, apostolus (1 Cor. 15:10): Migne prints a comma between the noun and its genitive. Carried — "By the grace, of God," — per Pattern 8's never-delete clause. He also drops the Vulgate's autem; no [var:], no sense change.

Question-mark parity across all five chunks: 9/9, 5/5, 19/19, 6/6, 7/7. Exclamation parity 0/0, 0/0, 1/1, 0/0, 0/0. Column anchors identical in order and value; note counts 1/1, 1/1, 4/4, 1/1, 5/5.

⭐ The internal disagreements — the same verse printed twice, differently, on one plate

These are the evidence the brief asked for, and none of them earned a marker: each is a transposition or a spelling, and Pattern 14 fires only on material divergence. All four are named as census candidates, and where possible the English registers the difference so it survives translation.

22. 1237A vs 1237D · Dan. 2:6 · WORD ORDER, and it is the Vulgate that adjudicates.

• 1237A (italic): praemia, et dona, et honorem multum a me accipietis
• 1237D (italic): praemia et dona, et honorem multum accipietis a me
• 1238A (roman re-quote): praemia, inquit, et dona, et honorem multum accipietis a me
• 1238B (roman, broken off): et praemia, et dona; et honorem multum accipietis.

The Vulgate reads accipietis a me, so the second, third and fourth printings agree with the received text and the first does not. The English registers it: "you shall receive from me rewards…" at 1237A against "you shall receive rewards … from me" at 1237D. The second printing also drops the comma after praemia that the first has — same site, same class.

23. 1240A vs 1240B · Dan. 2:10 · SPELLING. The italic lemma prints verbum hujusmodi; the roman re-quote eleven lines later prints verbum hujuscemodi, which is the Vulgate's form. Registered in the English as "such a word" against "a word of this kind".

24. 1240A vs 1240B · caelorum beside coelorum. The Luke 17:21 lemma prints Regnum caelorum; the Matt. 5:3 lemma four lines later prints regnum coelorum. Pattern 9 orthography; both rendered "heaven". (The Luke lemma's separate and material defect is crux 3.)

25. 1232A vs 1233A · angustiae beside angüstiae — crux 9 above.

⭐ Lemma pulls RESISTED — where Migne agrees and my instinct wanted to improve him

26. 1236A — the most dangerous one in my range, and it is not a defect. … ut in aeternum vivamus, quod tunc quidem fiet, si nihil nisi bonum, et sine bona intentione agamus. Read at speed the second member says "and [if we act] without a good intention", which reverses Richard's entire chapter, and the reflex is to supply a negative or to gloss a dropped nec. It needs neither. nihil is shared across both members: nihil nisi bonum / nihil … sine bona intentione. Rendered "if we do nothing but what is good, and nothing without a good intention" — the second "nothing" is a grammatical ellipsis supply, plain text, no brackets. No [cj:], no repair, no marker.

27. Dan. 2:1 (1233B), 2:3 (1234C), 2:4 (1234D), 2:5 (1236D–1237A), 2:6b (1238B), 2:8 (1238C), 2:9 (1238C–D) — all agree with the Vulgate word for word. Nothing was "improved", nothing familiar was allowed to arrive ahead of the printed Latin. In particular Sermo recessit a me is rendered "The word has departed from me", not Douay's "The thing is gone out of my mind", because Richard re-quotes recesserit a me sermo and glosses it.

28. 1234D — the italic boundary in Dan. 2:4. Migne closes the italics after in sempiternum vive. and prints Dic somnium servis tuis, et interpretationem ejus indicabimus roman. English habit wanted one continuous quotation. The boundary is reproduced exactly. Same at 1240B, where Beati enim pauperes spiritu… splits the Matthew lemma around Richard's own enim: the English reads "for Blessed are the poor in spirit…", not one lifted quotation.

29. Immaterial Daniel divergences, all declined, none marked: arioli for harioli (2:2 — and many Vulgate printings carry the unaspirated form; Pattern 9) · Si igitur for Si ergo (2:9) · Responderunt ergo … et dixerunt for Respondentes ergo … dixerunt (2:10) · Rex somnium servis suis dicat for Rex somnium dicat servis suis, and ejus for illius (2:7) · quem tu, rex, quaeris for quem tu quaeris, rex (2:11).

30. Non-Daniel scripture, all resisted: Matt. 6:23 (1235D) prints Quia, si oculus tuus fuerit nequam for the Vulgate's Si autem oculus tuus — connective substitution only, declined · Matt. 5:3 (1240B) agrees exactly · Ps. 2:6 (1240B) and Ps. 81:6 (1240D) agree and are simply truncated (super Sion montem sanctum ejus and omnes dropped), which is quotation practice, not divergence · John 15:5 (1236A) agrees · 1 Cor. 10:11 (1231D), Phil. 3:20 (1241A) and Luke 18:27 (1240D) are printed roman, unitalicized and unnoted, so they are not set as proof-texts; their trivial divergences (dropped autem; apud Deum possibilia sunt transposed) take no marker.

31. 1238D — non nihil profecit carried as the double negative Migne prints ("profited not nothing"), not smoothed to "profited somewhat". Pattern 7a, mechanism 3.

Vocabulary tensions — followed the table against my own judgment

32. 1233D · humanam virtutem. The table fixes virtus = virtue, and here the sense is plainly "human strength/power": quam sit supra humanam virtutem divinis beneficiis respondere. Rendered "above human virtue" per the table. If the table is to admit an exception, this is the site. Note the same chunk's 1240C nihil de propria virtute praesumunt takes "virtue" naturally.

33. 1236D · pudentia / impudentia. pudentia is not classical but is formed regularly from pudens and is attested in medieval Latin — Pattern 9, no marker. The table fixes impudentia = shamelessness, so pudentia is rendered modesty: "A common modesty; a general shamelessness." The tension: Migne's antithesis is a single root negated (pudentia / impudentia), and modesty/shamelessness does not preserve it as a pair. "Shamefastness / shamelessness" would, at the cost of an obscure word. Following the table; flagged.

34. mens / animus / anima / spiritus held visibly distinct throughout (mind / spirit / soul / spirit), including where two collide in one clause — 1233C animus … et vitiorum suorum victrix anima, 1234D mens humana … animus, 1235A sensus ejus usque ad animum non ingreditur. Neither of the two flagged animus/spiritus collision sentences (0005, 0034) falls in my range; the only spiritus here is Dan. 2:1's conterritus est spiritus ejus and Matt. 5:3's pauperes spiritu, and neither has an animus in the same sentence.

35. Pattern 17 throughout. thou wherever the Latin is singular (ut vides 1231A · tibi videtur 1232B–C · quidquid disponas … peragas 1236B · dicis … gloriaris 1236C · te … tibi ostendimus 1234C, 1235B · sermonem tuum, rex 1239A) and you wherever it is plural (Sine me nihil potestis facere · dii estis · peribitis vos … domus vestrae · intra vos est · narraveritis … accipietis). ⚠ The Daniel dialogue switches inside single sentences — the Chaldeans address the king as tu while he addresses them as vos — and every switch is preserved: "Tell the dream to thy servants" (1234D) against "Unless you declare to me the dream … your houses" (1236D).

Ratios and em-dashes

Word ratios EN:LA — 0000 1.60 · 0001 1.62 · 0002 1.50 · 0003 1.49 · 0004 1.44. The two high chunks are driven by Richard's quanto … tanto / ejusmodi … istiusmodi correlatives, which English can only unpack ("by how much the lower, by so much is it always found worse than itself"), not by padding. Three em-dashes in five chunks, each carrying a break the plate's own syntax makes: 1232D's number mismatch, 1233C's dangling tertium, 1235A's Syriace parenthesis.

Richard of St Victor, De eruditione hominis interioris, PL 196. Agent range: chunks 0005, 0006, 0007, 0008, 0009. Entries give column · printed reading · what was done · why.

A. The animus / spiritus collision — 0005 @1244A (flagged in the launch brief)

Printed: …quod meminit se prius per spiritus sui impetum statuisse? Multa enim … quae prius ex animi impetu, et quasi praecipitanter definivit … et multa saepe ex eodem spiritus sui impetu homo contra conscientiam suam dictat, et tamen eamdem animi sui praecipitationem … temperat.

What was done: both words rendered spirit, per the fixed vocabulary table, with no disambiguating gloss and no marker.

Why — and this is the point: the collision here is benign, and the table is what makes Richard legible. The sentence turns on eodem and eamdem: Richard is explicitly asserting that the spiritus sui impetus and the animi impetus are the same impulse, and that the animi praecipitatio is the same headlong rush. Rendering the two words differently ("mind"/"spirit", or "soul"/"spirit") would silently deny the identity his own correlatives assert, and the reader would meet two faculties where Richard has one. So the one-English-word outcome the table produces is not a loss here; it is the reading. No tension to escalate.

(For the record, the four sites in order: 1244A per spiritus sui impetum → "through the impulse of its own spirit" · ex animi impetu → "out of the impulse of its spirit" · ex eodem spiritus sui impetu → "out of that same impulse of his spirit" · eamdem animi sui praecipitationem → "that same headlong haste of his spirit".)

⚠ Note for the merge: mens is kept visibly distinct as mind throughout this range (0005 studiosa mens, compuncta mens; 0006 devota mens; 0007 mens studiosa), and anima as soul (0008 @1248D in amaritudine animae meae; 0009 @1252B anima diu multumque oppressa). No flattening anywhere in the range.

B. Markers placed

### B1. [sic: doctrinam] — 0005 @1242C Printed: prius saecularium, postea spiritalium doctrinam curam deponit. Done: "first lays aside the care of secular, afterwards of spiritual teachings [sic: doctrinam]". Why: saecularium … spiritalium require a genitive plural; doctrinam is an accusative singular standing beside a second accusative, curam, and has no grammatical slot (Pattern 12, test 1). Corpus control run: doctrinam curam occurs once in 87M words of PL TEI — here; doctrinarum curam occurs in 11535, in this very chapter's own head six lines above (tam saecularium quam spiritalium doctrinarum curam deponit). The plate contradicts itself on the same column-block, which settles it. 7a⁗ check: the clause keeps its verb (deponit) and its object (curam); the lexeme the quarantine removes is supplied in plain words ("teachings") beside the marker, so nothing of the sense sits inside the marker. Declined: a data/tei-patches/11535.json entry. A patch corrects markup, never words, and there is no markup defect here — only a reading. Left to Pattern 12.

### B2. [sic: alios] — 0009 @1252B Printed: In eamdem sententiam currit, quod alios idem Propheta de vitiorum exstirpatione, vel virtutum plantatione cavit. Done: "…that which the same Prophet [sic: alios] took heed of concerning…" Why: alios is a bare masculine accusative plural with no noun to agree with and no verb to govern it; it cannot be rendered in place. The near-certain reading is alias ("elsewhere") — and the plate itself supplies the internal control: one column later, at 1253B, Richard writes exactly the parallel formula with the right word, quod alias psallit. Corpus control: alios idem Propheta and alias idem Propheta both return zero files in the PL TEI, so the corpus is silent; the decisive evidence is the internal one at 1253B. 7a⁗ check: verb (cavit), negation (none) and conjunction (quod) all remain outside the marker.

B3–B9. [var: …] — the Vulgate-divergence list

See section C, which is the deliverable.

C. THE VULGATE-DIVERGENCE LIST (7a″)

Every italic span named in the launch brief was worked through word by word against the received text, plus every unitalicized continuous quotation in the range (Pattern 14 has no formatting precondition). Marked = a [var: …] fired. Logged = a real divergence, rendered as printed, judged immaterial. Resisted = Migne agrees with the received text and the instinct to "improve" was refused.

Marked (6)

(That is eight rows; "6" above counts distinct marker sites — 1253D and 1254A are two markers on one textual crux.)

Logged, rendered as printed, no marker (immaterial or non-scriptural)