Epistolae et decreta

Governed by translation-style.md Pattern 7 and "Papal letters — Innocent I" (P0–P3). The English renders what Migne prints (PL 20, cols 463D–612C), which he reprints from Coustant's Epistolae Romanorum Pontificum (1721). Where Migne prints a Greek version after a Latin letter, the English translates the Greek as printed (Epp. V, VII, IX, XII) and the differences are logged in the stint files.

Merge, chunks 0000–0031 (2026-09-24)

This section is a merge layer. The six stints' own records stay beside it, unedited: cruces-0000.md (cols 0463D–0488A), cruces-0006.md (0488B–0519A), cruces-0012.md (0519B–0544A), cruces-0018.md (0544C–0576A), cruces-0024.md (0576B–0594A) and cruces-0028.md (0594B–0612C). What follows is only what had to be settled across them and after the blind read.

Coustant's foot notes

537 letter-keyed notes were swept from the plate (one Sonnet sweep, then read again page by page by the translating stints). 528 stand as [cn:] in the Latin with a translated [nt:] twin in the English; 9 are Coustant's Prolegomena keys (before Epistola I) and are out of scope. The sweep put about 30% of its keys one word early and mis-transcribed or truncated some 100 note texts; every correction was made against the plate and is recorded in data/briefs/7090-NOTE-EDITS-.md and in data/plate-notes/7090.tsv. The blind read found sixteen more slips in the Latin note layer (e.g. 581 and 580 for 381 and 380 at 0478B and 0487A, clanculo for clangulo at 0479A, ABFERRI for AUFERRI* at 0530A) and all sixteen were corrected at the plate. A ] inside a note (0574 d) is written fullwidth ［ ］.

Markers fired, every one on a column read at the plate

[sic:] at 0466B asbcessurus, 0477A nom, 0483B patres (added after the blind read), 0491A editiotiones, 0525A Macedodonas, 0525C χειρονουμένους, ~0586A beattiudini, in quam, 0598A fratrer, 0599A Cathaginensi, 0606A epicopis (added 2026-09-24, below). [var:] at 0470B (II Thess. tradidi), 0529A (Prov. 19:14), 0578A (Jn 8:56), 0591A (Ps 51:9), 0592B (Lk 18:16). No marker sits on a Coustant reading.

TEI patches — 29 applied

data/tei-patches/7090.json: 29 plate-read corrections from the stints, each a place where the plate is sound and our digitization is not. Two are HELD for want of a second witness (worn glyph): 0545A impletas and 0575B misimus. The English followed the plate at every site.

Blind read (7 readers, Opus, cruces withheld) — findings and verdicts

Seven readers (R1–R6 on 7090, R7 on 7089 and 7091) filed 77 numbered findings; reports at data/briefs/INNOCENT-BLIND-READ-R*.md. Every one was judged against the Latin, and against the plate where it rested on a printed form. 73 applied (the English, or the Latin note layer where the Latin column was the one that departed from the plate), 1 logged with no change to the English, 2 rejected, 1 held. The readers' minor points were also taken up where they held.

Polarity and who-did-what (the class that matters), all applied: 0479B nulla ratione dissolvi (a contract is dissolved on no account) · 0583B inemundabili (mire that cannot be cleansed) · 0579C "can, not undeservedly" · 0537A note a (no one ever alleged it) · 0544A the letters were sent to Alexander, to be delivered to Acacius · 0546A Alexander gathered John's clergy · 0574A Christ, not "us", ascended · 0582A the pope, not Pelagius, judges the acts · 0588B the jussive damnent … submittant · 0590A the others see the heresiarchs cut off · 0031 conjunxisse is active.

Silent repairs found: 0522A Novatianos had been turned into the misprint "Navatians" in the English · 0483B Migne's patres for partes is now [sic:] · 0515A maxime is adverbial, "chiefly", where the English had silently adopted Coustant's maximis.

Contested, decided on the evidence (Wilson may overrule)

• 0570B quae fidelia corda sustineant. The stint had "must endure", which adds an obligation the Latin does not carry. Now "might endure": the subjunctive as printed. Coustant's note j says Garnier's added non is unnecessary, so the sentence probably carries a negative sense through its rhetorical form; the plate has no question mark, so no negative is supplied. Render as printed, no [cj:].
• 0595A deprehendi aliquis facile … nec alicubi possit agnosci. The English had supplied a negative ("no one could easily be detected") that only the inferior manuscripts and Garnier carry (note c). Now the printed reading, "someone could easily be detected", with the variant in the note. The sentence wants a negative and does not have one in Coustant's text.
• Saragossa 381 and 380 (0478B, 0487A). Both plates read the flat-topped 3 of Migne's face and the chronology settles it (Majorian and Agde come "afterwards"; the Priscillianists were condemned in 380). The earlier undecided cruces entries are closed.
• 7091 0633B nequidquam [Forte nil quidquam]. Now "not anything", reading ne quidquam; Migne's own bracket supports it and "in vain" left the clause with the opposite polarity. No [cj:].
• 7091 Greek οὔτε οὐ δυνήσεται (0632C). The pleonastic οὐ is not carried: "nor will worldly rank be able to help anyone". A double negative in English would cancel and reverse the sense.
• ἀρχιερεύς is "high priest", not "archpriest" (three Greek heads in 7091).

Rejected

• 7091 0630C cum istud … non fugisset (R7 F7). Read at the plate (p. 320): Migne prints cum, and the sentence construes with an elided object. "Him" is supplied in English as the Latin's own ellipsis, not a repair. No marker.
• 7089 0628B inergumeno (R7 F9). The plate (p. 319) prints iner-|gumeno, hyphenated at the line end. Inergumenus is an attested variant of energumenus, so Pattern 9 applies: no [sic:].
• 0606A epicopis Apulis (R6). First kept unmarked as ruled by the stint; reversed 2026-09-24 for consistency with fratrer and Cathaginensi (see "Held items settled").

Held

• R2 B7, 0497B note c; 0519 note a; Dionysius's number in Ep. XVIII's note g: moved to "Held items settled" below (2026-09-24). The Dionysius digit is still open.
• The second [cn: a] in 0527 (Ep. XVII inscription) is keyed before [0527A] in our anchors, but on the plate the key belongs to p. 527/528. Anchor placement, not a translation matter.
• Coustant's own figures, rendered as printed and not corrected: the Monitum's "Baronius … 308" and "430 aut 431" (plate p. 275), Cod. Th. lib. XIX, Pagi's anni 310, Arles I anno 514.

The 7091 English is two texts per letter

Migne prints each letter twice (Greek block, Latin block, each with its own column anchors), and the English translates each from its own language. The Greek-versus-Latin differences are in src/english/7091/cruces-0000.md.

Held items settled (plate, 2026-09-24)

Scan raw/scans/pl020/pl020.pdf. Each page was rendered on its own, and the printed corners were read before anything else. The disputed glyphs were read at 600 dpi against other sorts of the same face on the same page.

1. 0606A epicopis Apulis: [sic: epicopis] added in 0030 at "bishops of Apulia", in the same form as fratrer and Cathaginensi. The plate (p. 308, corners 605/606) was already read and recorded in data/plate-reads.json, and it prints Migne's own misprint for episcopis. plate-gate passes.
2. 0497B note c: the plate reads manendum (p. 254, corners 497/498). The sort after -mend- has two arms with an open notch at the top, which is the italic u of officium in note a on the same page. The italic a on the page (ad, sacrificia, fas) has a closed, rounded top. Fixed in data/plate-notes/7090.tsv (propter manendam → propter manendum) and re-injected. The Latin diff is that one word in 0007. The English twin was fixed to match.
3. 0519 note a: the plate reads sibi (p. 265, corners 519/520): Spiritales divitias nullus sibi praeter te conferre poterit, clean italic, with the initial the same sort as the s of nullus. The TSV and the Latin already had sibi. The English had rendered tibi ("confer on thee"). It now renders the printed sibi: "confer on himself". That is a real word and it construes, so no marker (Pattern 7; P0, no marker on a Coustant reading).
4. Ep. XVIII note g, Dionysius's number: UNRESOLVED. The note is on p. 274 (corners 537/538), col 0537 left foot, not 0526D as logged above: Apud hunc inter Innocentii decreta locum ?5 obtinet. The second digit is a 5: it has the left upright of the known 5 in the corner "537". The first digit is not settled: its top bar stands detached from the bowl at both ends, and the joint is the one feature that tells this face's flat-topped 3 (joined top right, like the known 3 in "(13 Decemb." = Idibus Decembris, same page) from its 5 (joined top left). Second witness: volumes.json lists for PL 20 only the archive item patrologiaecurs75unkngoog (an Augustine volume, per the style file) and Google Books IDs with no calibrated fetch method, so no second witness was read. The Latin keeps 55 and the English "55th" as transcribed. A reading of 35 or 55 stays open.
5. TEI patches 0545A impletas and 0575B misimus: NOT APPLIED, still held. - 0545A (p. 278, corners 545/546): imple?as. The sort is a short thin stroke in the upper half of the x-height and nothing more. On the next line a worn r in detersas is the same kind of thin upright, and every t on these lines is a full sort. So the plate does not tell t from r. - 0575B (p. 293, corners 575/576): misi-|mu?. The last sort is an upper hook plus a detached dot. That fits a broken s (spine lost) or a broken r (stem lost) equally well. Not clear. Neither is clear at the plate, so data/tei-patches/7090.json is unchanged and there was no re-chunk. The English ("have been fulfilled", "We have sent both") is unchanged. Both want a second witness.

Every page in range was opened at the plate once (raw/scans/pl020/pl020.pdf, 300 dpi, corner numbers checked first), text collated and every foot note read in the same visit. The English renders what the plate prints. Notes placed or corrected by hand in the Latin are listed under "Notes" so that data/plate-notes/7090.tsv can be brought into line before any re-chunk (a re-chunk from the TSV as it stands would erase them).

Standing decisions

• MONITUM IN EPISTOLAM SEQUENTEM → "NOTICE ON THE FOLLOWING LETTER." (corpus precedent: MONITUM → NOTICE).
• Coustant's letter notes: Latin [cn: a | …] (Latin only; the verifier forbids [cn:] in the English), English [nt: …] at the same point. The brief's wording ("keep each verbatim in the English … AND add an [nt:] twin") conflicts with verify-english.mjs check 9a and with Pattern 19; the corpus precedent (8950, 9198) and the verifier were followed: the English carries the [nt:] only.
• Mentioned words (the consular formula Honorio aug. VI et Aristeneto conss., post consulatum, P. C., misi, conss.) stay Latin italic: Coustant's argument turns on their exact form.
• Addressee singular (Anysius, Victricius) → thou/thee (Pattern 17); plural addressees (the Toledo bishops) → you.

Notes (placement and correction by hand)

• p. 237, col 0463, note (a) on the head EPISTOLA I: recovered by the sweep but filed among the out-of-scope Prolegomena notes. It is keyed (ª) to the head of Epistle I itself, so it belongs to this work. Placed in the Latin head of 0000 before EPISTOLA I. Plate reads colligamus (TSV had colligimus) and Holstenius (second letter damaged on the plate; the TSV's Helstenius is a misreading; Lucas Holstenius is meant).

• p. 240 (cols 469–470), notes a–k, sequence complete. Placed by hand in 0002: (a) at X. of the Dionysian title list (unplaced), (b) at XI (unplaced), (k) before usura poenarum (unplaced). Corrected: (c) the Latin [cn: c] carried the ANCHOR words (clericus factus fuerit, (c) uxorem non ducat) instead of the note; replaced with the plate's note Ita purum Dionysii exemplar Regium. Alia vero subinde aucta, quibus Isid. et Hisp. coll. concinunt, propositum suum servare debeat. Moved to the printed key: (d) after virginibus (was before it), (g) after normam (was before it), (j) after namque (was before it). Completed: (f) the note runs over into the right column: …Etsi tibi nota sunt omnia docendi ac vivendi praecepta, quae regula ecclesiastica continentur. Note (a) prints cleritatus (for clericatus) and convocare (the decree reads convolare); both Coustant's type, carried in the note, no marker.
• p. 241 (cols 471–472), notes a–i, sequence complete. Placed by hand: (d) after dicente Domino, (e) after nullus audeat ordinare, (i) before cui in omnibus causis (all unplaced). Corrected: the Latin [cn: h] (the long note on vel inter laicos) stood at the key of note (i), after Romanae Ecclesiae,; the plate keys (h) after inter clericos and (i) before cui in omnibus. (h) moved, (i) placed. Moved to the printed key: (b) before disciplinam deificam, (c) before Exstiterunt. Completed: (e) from the plate (…hoc item Gallis injungat.); (i) runs over to p. 242, completed from there (…agendum sibi negotium in urbe Roma.). In (i) the plate prints quamvis id Oxoniensi Codicis … exemplari non deessent (read at 400 dpi): id is stray (for in?); the dative construes without it and the English follows the dative ("not wanting to the Oxford exemplar").
• p. 242 (cols 473–474), notes a–g, sequence complete. Placed by hand: (a) after ad sedem apostolicam, in §6 (the unplaced list gave the anchor as regionibus propriis ecclesiis praesident, regularum hunc librum, which is col 470A; the note is plainly on beata consuetudo exigit and the plate keys it at apostolicam, ª sicut synodus statuit); (b) after Ut in §7; (e) after Unius uxoris virum in §9 (anchor occurs 6×; the plate key is §9's); (g) after satis errat a regula,. (a) completed from the plate (the list truncated it at Ant); (g) runs over to p. 243, completed from there (…ac Leo, epist. 45.). In (a) and (i) the MS number is read 3368 (p. 241 clearly; p. 242's digit is ambiguous 3/5; the unplaced list had 5368). Ambrose's letter to Vercelli is printed epist. 63 in (g) (p. 243) and in p. 243 (h); the 3/5 digit is not decidable at 400 dpi in this face, and the section numbers (n. 63 in g, n. 62 in h) likewise. Recorded as read.
• p. 243 (cols 475–476), notes a–h, sequence complete. Placed by hand in 0003: (e) after cum uxoribus suis in §12 (anchor occurs 2×; the plate key is §12's, not the title list's), (g) after exaudiri (anchor 2×), (h) after Unius uxoris virum: in §12 (anchor not found by the injector: the phrase occurs in 0002 and 0003). Moved to the printed key: (f) after praeter (was before it).

• p. 244 (cols 477–478), notes a–g, sequence complete. All at printed keys. (e) plate prints licitorum.... gloriam (an elision in Paulinus's quotation), restored in the Latin. (g) is truncated at the page foot (cum et an-); the Latin carried a placeholder note of ours inside the [cn:] ("(illeg. — note is cut off …)") and a stray opening quote: both removed, and the note completed from the top of p. 245's foot area, where it continues: …cum et antea puellis virginitatis propositum licuerit suscipere. So this "truncated" note is NOT truncated: it runs over.
• p. 245 (cols 479–480; left corner misprinted 379), notes a–h, sequence complete. All at printed keys.
• p. 246 (cols 481–482), notes a–b, both at printed keys. p. 247 (cols 483–484): no notes (checked zero).
• p. 248 (cols 485–486), notes a–e, all at printed keys. In (b) the plate reads sit ?equendum (first letter damaged; sequendum is Patruinus's word); left as the TSV has it (aequendum).
• p. 249 (cols 487–488), notes a–e. (c) our Latin ended in a placeholder (continues past assigned crop, held on the plate); completed from the right foot column (… statuimus communioni nostrae esse reddendum.). (d) corrected suppressi → supprimi (plate). (e) keys §3 Num quod non aliquid at 0488A, which is in chunk 0006 (T2's): not placed here. T2 should place it: Sirm. cum mss. Corc. et Colb., Numquidnum aliquid. Codex Bonaevallis, Num quod non aliquid. Ex duobus primis revocato num, et ex postremo quod non, integra et perspicua jam habetur lectio. … (full text in the TSV, p. 249 e).
• Digits. This face sets 3 and 5 nearly alike. Years in the notes are left as the TSV reads them where the plate is not decisive: p. 244 (g) Anno 581 (plate looks like 381; the Saragossa council is 380), anno 397; p. 249 (c) anno 580 (plate looks like 380); MS not. 3368 (p. 241 clear, p. 242 ambiguous); Ambrose epist. 63 (p. 243). The English follows the Latin as it stands. A reader with a second witness should settle them.

Text cruces

• 0466B Roma asbcessurus: the plate prints asbcessurus (p. 238, right col., band B), a non-word for abscessurus "about to depart". Migne's type, read at the plate. [sic: asbcessurus], sense rendered beside it.
• 0465A par fuit recognoscere with Migne's inline (f. te cognoscere): rendered as printed ("it was fitting to recognize me ordained"); Migne's own conjecture stays in his [n:].
• 0468B juxta legem dudum latam tradant curiae facultates: the plate (p. 239, right col.) prints tradant; our TEI has tradunt. The English follows the plate ("or … to hand over their means"). Patch filed.
• 0474A (§9) Unius uxoris virum. Et iterum: the plate (p. 242, right col.) prints a full stop after virum (the key ᵉ stands before it); our TEI has none. The English follows the plate. Patch filed.
• 0470B II Thess. II, 14 printed quas tradidi vobis sive per verbum, sive per epistolam; Vulgate quas didicistis, sive per sermonem, sive per epistolam nostram. Innocent's point is tradition handed down, and tradidi carries it; [var:] fired (col read at the plate, p. 240). Coustant's note (j) records tradidimus in the old editions.
• 0475A Prov. XIX, 14 printed A Deo praeparabitur viro uxor (LXX-type); Vulgate a Domino autem proprie uxor prudens. Sense not changed for the argument; no marker.
• 0474A Sacerdotes mei semel nubant / Sacerdotes mei non nubent amplius, the second cited by Migne as (Tit. I, 6): neither is Vulgate text; Coustant's note (e) traces both to Tertullian's "Leviticus". Rendered as printed; the reference is Migne's.
• 0470C Luke XII, 48 is quoted loosely (Cui multum enim creditur, plus ab eo exigitur); Ps. XLIX, 18 Videbas furem et currebas cum eo lacks the Vulgate's Si. Not material; no marker.
• 0477A Rom. VIII, 8–9 Vos autem jam nom estis in carne: the plate prints nom (p. 244, left col., 400 dpi). A non-word for non; [sic: nom] with the negation rendered beside it ("now not [sic: nom] in the flesh", 7a⁗ q. 2). The marker is unitalicized because it sits inside an italic scripture run.
• 0476A Lev. XI, 14 cited for Sancti estote, quoniam et ego sanctus sum Dominus Deus vester (the text is Lev. XI, 44 / XIX, 2): Migne's reference, kept verbatim in its [n:]. 0477A Rom. VIII, 8 is cited for 8–9.
• 0486B a pace omnium discederunt: plate prints discederunt (p. 248), which Coustant's note (e) says is the manuscripts' reading (Ita mss. Corb. et Colb.) and would emend to discesserunt. A Coustant reading, not a type defect: no marker (P0). Rendered "have withdrawn".
• 0485B indulgetiam (TEI) is indulgen-|tiam on the plate, a line-end hyphen our TEI lost a letter at; rendered "indulgence". Patch filed. 0486A haectenus (TEI) is hactenus on the plate; patch filed. 0481B hauc (TEI) is hanc; patch filed.
• 0485A adeo que / necue: plate glyphs damaged; rendered "and thus" / "or not"; no patch.
• 0481B Jac. Sermondus: the plate prints Sermondus (Sirmond meant); carried as printed.
• 0483C Incipit regula fidei … et cum praecepto papae urbis Romae Leonis ad Balconium episcopum Galliciae transmiserunt: rendered as printed.
• 0486B Nam fratres nostri coepiscopus Hilarius, et Elpidius presbyter: plural fratres with singular appositives; rendered "our brethren, our fellow bishop Hilarius and the presbyter Elpidius".
• 0487B ( Seu discordant et in simultates abeunt): Migne's own bracketed variant, translated in place ("Or disagree and go off into enmities"), parentheses kept.

## Totals (T1)

• Pages read: 237–249 in full (13), 250 for note run-over only.
• Letter notes now in the Latin (each with an English [nt:]): 0000: 2, 0001: 0, 0002: 28, 0003: 24, 0004: 0, 0005: 9 = 63. Of these, 14 were placed by hand (1 from the out-of-scope list, all 13 of the unplaced list in this range), 7 moved to the printed key, 1 replaced (0002 c), 7 completed from the plate. One note owed to T2 (p. 249 e).
• [sic:] 2 (0466B asbcessurus, 0477A nom); [var:] 1 (0470B); [cj:] 0.

Every page in range was opened at the plate once (raw/scans/pl020/pl020.pdf, 300 dpi full width, 600 dpi crops where a letter or digit was in doubt, corner numbers read first on every page: 249 = 487/488 … 265 = 519/520). Text collated and every foot note read in the same visit. The English renders what the plate prints. Hand edits to the Latin [cn:] layer are listed under "Notes" so that data/plate-notes/7090.tsv can be brought into line before any re-chunk (a re-chunk from the TSV as it stands would erase them).

Standing decisions

• Coustant's lettered notes: Latin [cn: a | …], English [nt: …] only (T1's decision, followed: the brief's "keep each verbatim in the English" conflicts with verify-english.mjs check 9a and Pattern 19).
• Greek parallel texts (Ep. V, VII, IX, XII: Migne prints Palladius's/Sozomen's Greek after the Latin) are TRANSLATED from the Greek as printed, not carried in Greek and not harmonized with the Latin. Where the Greek says something the Latin does not, the English of the Greek says it (see Greek cruces below). This roughly doubles those letters in the English; the Greek is Migne's text on the page. No [lat:] markers (a PG device).
• Pattern 17: Exsuperius, Theophilus, Arcadius (Ep. IX), John (Ep. XII), Rufus, Aurelius (Ep. XIV), Juliana → thou. The Toledo bishops, the Constantinopolitan clergy and people, Aurelius+Augustine (Ep. X), Honorius's vos/vestra to Arcadius in Ep. VIII → you.
• Migne's bracketed matter kept and rendered: [ Forte expolitam] → [ Perhaps expolitam], [ Baron. addit. ecclesiae], [ Supple, quis] → [ Supply, quis], [ Antiquo more pro spectat] → [ By the ancient usage for spectat], [Baron. consortem], [Baron. praemissa], [ut compertum habeas] (translated: it is Coustant's supplement, note c says so), [Lab. παραμυθήσεται], [Add. ἢ] kept as printed.
• CAP. → CHAP.; EPISTOLA → EPISTLE; Migne's -- kept.

## Plate readings that differ from our TEI (English follows the plate; patches filed in ## data/briefs/7090-PENDING-TEI-PATCHES-0006.md)

• 0489A TEI tamquam singulis scripta, sint, accipiant — plate scripta sint, accipiant. Punctuation only.
• 0490B TEI reprebendimus — plate reprehendimus (600 dpi: the ascender is h). Rendered "reprehend".
• 0492B TEI elegendi — plate eligendi. Rendered "chosen".
• 0498A TEI remittetur. Ita ut — plate remittetur, ita ut. English runs the two as one sentence.
• 0501 (Ep. VII head) TEI ipsorummel — plate ipsorummet; TEI aecumenica — plate oecumenica.

Plate defects (Migne's)

• 0491A editiotiones — the plate prints editio-|tiones, a doubled syllable across the line break. Carried [sic: editiotiones] with the sense ("shows", editiones) beside it; Coustant's note (c) on the word (supple ludorum vel munerum) shows he read editiones. Not a hyphenation artifact of ours: the plate sets the extra tio in type.
• 0494A (Ep. IV summary list, 0007) reprëssione ( pro reprehensione) — Migne's own gloss on his own form; rendered "repression" with his parenthesis kept. No marker (he flagged it himself).
• 0508B quo pacto extra summum … [ Supple, quis] putaret esse discrimen — the sentence needs Migne's supplied quis; rendered with his bracket, as printed.

Notes (placement and correction by hand)

Sequence check per page: every page in range runs complete (249 e [a–d are T1's]; 250 a–i; 251 a–i; 252 a–d; 253 a–k; 254 a–g; 255 a–d; 256 a–g; 257 a–c; 258 a–d; 259 a–f; 260 a–f; 261 a–e; 262 a–e; 263 a–g; 264 a–m; 265 a is mine, b–j are chunk 0012's). No gap found; no note missed by the sweep.

Placed from 7090-UNPLACED-NOTES.md (10 of the 11 on my pages):

• p. 250 (h) after ne perturbationes (0490B) — the plate reader's anchor "nisi perturbationes" abbreviates nisi perpenderemus ne perturbationes.
• p. 253 (c) on the head EPISTOLA VI.; the note runs on in the right column of the same page (…ut sententiam suam adversus eos, qui aut scientes aut ignorantes peccassent, aequo discrimine moderaretur.), completed. (f) after --III. in the Dionysian title list, before De administratoribus.
• p. 254 (c) after Non (0497B, Non permanentem in concupiscentia); (d) after et before habentem filios (0498A).
• p. 261 (a) in the head EPISTOLA IX. (plate: EPISTOLA ª IX.); (e) in the head EPISTOLA X. (plate: EPISTOLA ᵉ X.).
• p. 262 (c) in the head EPISTOLA XII. (plate: EPISTOLA XII ᶜ).
• p. 263 (e) after in minus bonis, before censio (0516A).
• p. 264 (b) in the head EPISTOLA XIV. (plate: EPISTOLA ᵇ XIV.). Plate reads sed ab anno 453 certior fieri (the TSV had 455, which makes Leo inform himself in the year itself); corrected to 453.
• Not placed: p. 265 (b), keyed to EPISTOLA XVI, which opens chunk 0012 (another stint's). Left for T3; its text in the TSV reads clean against the plate.

Moved to the printed key:

• p. 256 (e) after Martias (was before it): the plate raises the key after Martias, before the comma.
• p. 261 (c) after scriptae, before sunt (was after quae ad me): plate quae ad me scriptæ ᶜ sunt.
• p. 265 (a) after gratiae (was after nobilitatem): plate gratiæ ª, nobilitatem.

Completed / corrected in the Latin:

• p. 250 (i) the note runs over to p. 251 top-left; completed from additur, exstat quidem in decurtatis… to …gratiae Christianae sacramentum appellare familiare est. (the sweep stopped at addi-).
• p. 251 (d) runs on into the right column of the same page: …seu summarium capitulum Innocentii explicans textum. In iisdem libris mox pro creatoribus substitutum est ordinatoribus. Completed.
• p. 253 (a) plate opem ferre citius (TSV obesse ferre cutius), ad synodum (TSV synodam); (b) plate frustra niti (TSV viti). Corrected. (a) prints γενουμένη and italic cogitur: kept as printed.
• p. 254 (e) ends …altera ad eos qui scientes: at the foot of the page and does NOT continue on p. 255 (the top of p. 255 carries only the run-over of (g)). Truncated as printed; the sweep's "continues at top of p.255" was wrong. Translated as printed, nothing supplied.
• p. 254 (g) sweep's parenthetical removed (note ends …conversionem ac mutationem petit.). ⚠ It prints Leo epist. 100 (the plate figure is worn: 100 or 108); left as the TSV had it.
• p. 255 (a) the unkeyed paragraph in the right-hand apparatus (Cui Ambrosius n. 2 respondet: …ut in voluntate magis abstinentis, quam in necessitate sit legis.) is the run-on of (a) itself (it continues the Ambrose ad Studium quotation (a) introduces); appended to (a). The sweep had recorded it as a keyless excursus.
• p. 257 (c) sweep's parenthetical and stray quotation marks removed. The raised d inside the quoted canon (locum d fensionis) is Migne's stray sort; carried in the note, rendered "place of d efence".
• p. 258 (a) the sweep dropped a clause by eye-skip: plate reads Sardicenses canones ut a Nicaenis distinctos, non ut pro Nicaenis habitos eorumque nomine vestitos, ab eo receptos esse; restored.
• p. 261 (c) runs over to the top of the right column on p. 261/262 apparatus: completed …cum varias epistolas accepisset, ad Honorium scripsisse, ὑποτάξαντα κατὰ μέρος τῶν γραμμάτων δύναμιν, … simul ad ipsum fuisse destinata.
• p. 263 (a) the TSV's tail (quintum, Eo ipso forte, qui a Siricio…) splices two notes. Plate: (a) ends on p. 263/264 …quintum Theodosii, non quartum aut tertium, ac vicissim cum quinto Theodosii nonum Honorii, non septimum aut alium quemvis componerent. Corrected. Plate anno 413, quo frumentum Heraclianus… (TSV 417; 600 dpi, the last digit is a 3; Heraclianus's embargo is 413). Plate ut ne me Rufi ordinatio (TSV nec me); the type is broken, left as the TSV has it; English "neither Rufus's ordination nor".
• p. 263/264 (g): the plate prints under key (g) at Senecione the words In edit. Concil. 11. Quae autem 14 erat, nunc 25. and the note runs on at the head of p. 264's apparatus Eo ipso forte, qui a Siricio epist. 9, n. 5, Basso episcopo in consortium regendae ecclesiae datus dicitur. Both carried as (g), as printed. The first sentence is the usual heading-note formula and belongs by sense to the head of EPISTOLA XIII (whose own note (a) lacks it); Migne's (or Coustant's) misplacement, NOT moved.
• p. 264 (h) prints ante Paschu / ante hoc Paschu (the sweep normalized to Paschae); plate form restored.
• p. 264 (k) prints Arelatensis I synodus anno 514 habita (Arles I is 314). Coustant's/Migne's figure; translated as printed.

Greek-text cruces (translated from the Greek as printed)

• 0504A (Ep. VII Gk) τῆς ἐπιστολῆς τῆς ἡμετέρας ἀγάπης — "the letter of our charity" where the Latin has litterarum charitatis vestrae. Rendered as printed.
• 0506B (Ep. VII Gk) τοῦ Χριστοῦ αὐτῶν — "the Christ of them" (Latin Christi ipsius). As printed.
• 0506A (Ep. VII Gk) TEI βεῳαιότητα — plate βεβαιότητα (a digitization error; patch filed); rendered "firmness".
• 0512A (Ep. IX Gk) ἀποστῆναι τῆς αὐτῶν κοινωνίας — "their communion" where the Latin has ab ejus communione (John's). As printed. ἐκ πασῶν τῶν ἐπισκοπῶν — "all the bishoprics" (Latin epistolis; ἐπιστολῶν would match). As printed. The note (b) spells the word ἐθελοχακοῦντας, the text ἐθελοκακοῦντας.
• 0514B (Ep. XII Gk) Migne's [Add. ἢ] supplies the comparative particle; rendered with it kept.
• 0514C (Ep. XII Gk) καθαρισθεῖσα … στήσεται — "having been cleansed, will stand" (Latin pura consistit).

Other

• 0507B Honorius cur ista vos detrimenta — Coustant's note (d) says the sense requires nos; rendered "why you", as printed.
• 0515A consolationibus — Coustant's note (b) conjectures consultationibus; rendered "consolations", as printed; his note carries the conjecture.

Cols 0519B–0544A, PL 20 PDF pp. 265–277 (raw/scans/pl020/pl020.pdf, 300 dpi, corners read first on every page; spot zooms at 400/600 dpi where noted). Every page in the range was opened once, text collated and the foot notes read in the same visit.

Standing decisions

• Coustant's lettered notes. Latin [cn: <letter> | …] at the key; the English carries a translated [nt: …] at the same point and no [cn:]. The stint brief asked for the Latin [cn:] to be repeated verbatim in the English with the [nt:] after it, but Pattern 19, P0 ("English: an [nt: …] opposite every note") and verify-english.mjs check 9a (an English [cn:] is a hard error, and it would also break note-count parity) all require the [nt:] alone. Followed P0 and the verifier; if the brief's form is wanted it is a mechanical copy.
• Bare variant words inside a note stay Latin italic in the case printed; sigla expanded at first use in each note (Merlin, Crabbe, Quesnel, Garnier, the Roman edition, the Isidorian collection); Dion./Hadr. → "Dionysius"/"Hadrian" (the collections); manuscript sigla (Colb., Corb., Reg., Pith., Navar.) rendered "the Colbert / Corbie / Royal / Pithou / Navarre manuscript".
• Pattern 17: Innocent's singular addressee (Marcianus, Ep. XVI) takes thou; the Macedonian bishops (Ep. XVII) take you.

Notes: corrections to the recovered note text (plate read; Latin [cn:] corrected in place)

These change the text of [cn:] markers already in the Latin chunks. The same corrections are owed to data/plate-notes/7090.tsv / the sweep TSV, or a re-chunk will revert them.

Notes placed by hand (from 7090-UNPLACED-NOTES.md)

• p. 265 col 0519 (b) — at EPISTOLA ᵇ XVI (0012, in the head, as the other 7090 heads do).
• p. 265 col 0520 (i) — at aut ille sic ⁱ peccari; the phrase occurs twice in our Latin (again in the Monitum §5 quotation, 0013), the plate key is in the letter, 0012.
• p. 268 col 0525 (a) — at EPISTOLA ᵃ XVII (0014).
• p. 268 col 0526 (b) — at INNOCENTIUS ᵇ RUFO (0014). The sweep described the key as sitting in "an illustrative quotation"; it does not — it is the letter's own inscription, which our Latin has.
• p. 268 col 0526 "(c)" — NOT A NOTE. No key c is printed on p. 268 (the page's keys are a and b only). The sweep's "c" (dici concinit et alter Colb. Unde liquet…) is the middle of note b, split off at the word Eidem codici in the note's own text. Folded into b; nothing to place.

Cruces in the text

• 0521B/0522B Navatianos, Anisii, Bonosio (Monitum §3, p. 266) — the plate prints Navatianos (beside Novatianos elsewhere), Anisii (beside Anysii two lines up) and qui Bonosio adhaeserant (beside Bonosus/Bonoso throughout). All read at the plate, all as our TEI. Rendered as printed ("Navatians", "Anisius", "Bonosius"); variant name-forms, no marker.
• 0522C ὅστε χειροθετουμένους — plate prints ὅστε (for ὥστε); carried as printed, Greek.
• 0525A Macedodonas (0014, §7) — Migne's type. The plate sets Macedo- at a line end and donas at the next line's head (p. 268, left col), so the syllable doubling is his, not a digitization fusion. [sic: Macedodonas] with the sense ("the Macedonians") beside it.
• 0525C χειρονουμένους (0014, §8) — plate read at 400 dpi: Migne prints χειρονουμένους, which is not the word Coustant's argument needs (χειροτονουμένους, set correctly two lines above). Carried in Greek inside [sic: χειρονουμένους].
• 0525B/C χειροθετουμένους, : — our TEI has a stray colon; the plate prints only the comma. English follows the plate; logged in the pending-patches file.
• 0526A peceaverat — our TEI; the plate prints peccaverat. English renders the plate ("had sinned"); pending patch filed.
• 0526A occassione — our TEI; the plate's glyph between occ and sione is worn and looks like one letter (so occasione), but not certain. Either is an attested spelling (Pattern 9); no marker, no patch.

Notes, pp. 269–277: corrections, placements, additions

• p. 269 (d) — the sweep marked it cut off at ante bap-. It is not cut off: it runs on in the next column of the foot (…ante baptismum. Probabile autem est eum, Siricii exemplo curasse … quae subinde tractantur.). Completed in the Latin; the sweep's own comment "(note cut off … not visible …)" had been injected into the [cn:] text and is removed.
• p. 269 (b) — plate reperiremus (sweep reperimus).
• p. 270 (j) — ADDED; the sweep missed it. Key ʲ after in dante (0530C), note Duo mss. in dante remanserat. Ad imitationem hujus loci … conferri non potuisse ordinationem. Now [cn: j | …] in 0015 with its [nt:]. The page's run is a–j with no gap.
• p. 270 (i) — the injected text ended with the sweep's "(note continues onto p.271)"; removed.
• p. 271 (c) ab Apostolis apostolicisve viris (sweep dropped Apostolis); (e) Jam e Siricii (sweep Jamne).
• p. 272 (c) — completed from the foot (ne qui diu convincitur fuisse perversus, per speciem remeantis non remedium salvationis inquirat, sed facultatem propagandae pravitatis inveniat., italic, Simplicius quoted); the injected text carried a stray quote mark and the sweep's commentary, both removed. (d) placed by hand at transierunt ᵈ damnatum (0016); the phrase also occurs in the Monitum (0012), the key is in the letter.
• p. 273 (c) Docet quippe / Et hic igitur (sweep De caet quippe / Et haec igitur); (d) placed at inultum ᵈ solet; the sweep read the anchor as inultum sit transire, the plate and our TEI both read inultum solet transire; disciplinam for discipinam; (e) placed at rediit; ᵉ et vel veniam; text quem reum non esse contestamini (sweep dropped non esse, which inverts the note); (g) initium ducit; (h) quod in altero Colb.
• p. 274 (a) — WRONG NOTE INJECTED. The Latin carried note e's text (Editi post Dion. et Isid., reparatus mercatur pacem…) at key a as well as at e. The plate's a is Verbum, aestimatur, Quesn. praetermisit … vel jactavit vel finxit. Replaced.
• p. 274 (e) — moved: the plate keys it after reparatus (cum suis reparatus ᵉ, profundam pacem), not after non fictam. Italic span and charitas/castigantur read from the plate.
• p. 274 (f) — placed by hand at (13 Decemb. ann. 414) ᶠ Flavio Constantio.
• p. 274 (g) — completed from the right-hand foot (…tractaretur, adeoque circa ann. 414 scriptam esse censeamus … moram in Urbe majorem quam sperabatur, facerent.).
• p. 274 (c) corriguntur; (h) nec omnes; (l) Hadrianus II; (m) Illius scilicet.
• p. 275 (b) ne talibus socientur.
• p. 276 (h) — moved: the plate keys it Subscripserunt ʰ viginti, not before Subscripserunt. (a) exprimat, duas in partes, successit; (b) precationibus vacantes.
• p. 277 (a) placed at EPISTOLA XX ᵃ, (c) at EPISTOLA XXI ᶜ (0017, in the head). (c) Beroeae, in collectionibus, teste Palladio. (d) interfuerunt; one word worn: Quae ro….i in synodo — not read, carried as ro...i in the Latin note and said so in the English twin. Do not guess it. (f) cognoscat, apud Nicephorum (sweep Niceph. rom.).
• p. 277 (b) retragantibus is the plate's form; carried.
• Key sequence, every page: 265 a–j · 266 a · 267 none (a checked zero) · 268 a–b · 269 a–h · 270 a–j · 271 a–g · 272 a–f · 273 a–h · 274 a–n · 275 a–c · 276 a–h · 277 a–g. No gaps. Keys owed to other stints: p. 265 (a) (0011), p. 277 (g) (0018).

Cruces in the text, pp. 269–277

• 0529A axoris, 0530A cripuerit — our TEI; the plate prints uxoris, eripuerit. English renders the plate ("wife", "rescued"); pending patches filed.
• 0529A Prov. 19:14 Etenim a Deo praeparatur viro uxor — Vulgate a Domino autem proprie uxor prudens. [var:] placed (column read at the plate). Lev. 22:13–14 at 0528A is cited as printed; no marker.
• 0531A honoram — plate honor.m, the vowel worn; English "honour" either way; no marker.
• 0540A expelere (Monitum §2, TEI) — rendered "ask for" by sense. The plate at p. 276 top reads expetere (clear t), so the TEI form is our digitization's; no marker. Patch owed: find communionem expelere festinavit → communionem expetere festinavit.
• 0544A subcisivum — the plate's form (= subsicivum, "left over"); rendered "residual".
• 0525D Macedodonas, 0525C χειρονουμένους — see above; the only two [sic:] in the range, both whole-syllable or whole-word, both read at the plate.

Cols 0544C–0576A, PL 20, PDF pp. 277–293 (raw/scans/pl020/pl020.pdf, 300 dpi, corners read first on every page, with spot zooms). Every page in the range was opened once: the text was collated and the foot notes were read on the same visit. Page reads: data/briefs/7090-PLATE-READS-0018.json. Note edits: data/briefs/7090-NOTE-EDITS-0018.md. TEI patches: data/briefs/7090-PENDING-TEI-PATCHES-0018.md.

Standing decisions

• Coustant's lettered notes. Latin [cn: <letter> | …] at the printed key. The English carries a translated [nt: …] at the same point and no [cn:] (Pattern 19, P0, verifier check 9a). This follows the launch correction and T3.
• Bare variant words inside a note stay Latin italic, in the case printed. Editors are expanded (Merlin, Crabbe, Quesnel, Garnier, Labbe, the Roman edition, the Isidorian collection, the edition of Augustine). Manuscript sigla stay as printed (Colb., Corb., Germ., Pith., Vatic., Reg., Navar.), per P0.
• Pattern 17. Innocent's singular addressees (Maximianus, Boniface, Alexander, Decentius) take thou. So do the African councils and the five bishops writing to Innocent (sanctitas tua, te). Plurals take you (ad vestram synodum, 0544C; litteris vel commonitorio vestro, 0547A).
• Heads. CAP. I. etc. are left as printed, and the Monitum head is "MONITUM ON THE FOLLOWING EPISTLES", both after T3. In the bishops' lists the names keep their Latin form, except AUGUSTINUS → AUGUSTINE and the addressee INNOCENTIUS → INNOCENT.
• Nested italics. Where a note sits inside an italic run (inscriptions, titles, scripture), the italics are closed before the [nt:] and reopened after it, as T3 did.
• Brackets inside a note. A marker cannot contain ]. In 0023 d the Roman editor's square brackets are written as fullwidth ［ ］ in both columns.

Text: plate against twin

• 0545A impleras (twin) against the plate's imple·as (read as impletas; the glyph is thin and has no r shoulder). English: "have been fulfilled". A patch is filed, flagged as a worn glyph.
• 0546B (Vide epist. 19, n. 1), etiam: the plate shows a point with no tail. It could be a period or a worn comma. The English keeps the comma, and no patch is filed.
• 0558A anctoritatem (twin) against the plate's clean auctoritatem. English: "authority". Patch filed.
• 0560A orent super cum (twin) against the plate's orent super eum. English: "pray over him". Patch filed.
• 0571A auctoritate sanctitatis tuae (twin) against the plate's auctoritati (clean). English: "yield to the authority of thy holiness". Patch filed.
• 0575B misimur (twin): the last sort on the plate is broken, but it looks more like s than r (there is no stem). English: "We have sent both". Patch filed, flagged as worn. A second witness is recommended before the patch is applied.

Readings rendered as printed

• 0548B–0549A non ʰ esse vere visum est … Dei Ecclesiam commutari: the double infinitive is rendered "it has truly seemed that the Church of God is not to be changed". The negation is carried. Coustant's note h gives the variants (non vere visum, non ergo visum, and the Royal exemplar's non esse visum).
• 0548A nisi quod illa in transitu meruit, ista susceptum apud se consummatumque gauderet → "except that the former merited him in passing, while the latter rejoiced that he was received and brought to his consummation in her". The "him" comes from susceptum. Nothing is added.
• 0553A–0556B (Ep. XXV): the rendering is literal throughout. "leaven" is kept for fermentum, as Coustant's note f requires.
• 0555A Verba vero dicere non possum: rendered "But to say the words I cannot". The note sits after "say" to match dicere ᵃ non possum.
• 0566A Rom. 7:23 is printed captivantem me sub lege peccati (Vulg. in lege). It is rendered "under the law of sin". No [var:] is fired, because Coustant's own note e already gives in lege (Pattern 18 ❌: the edition already glosses it).
• 0569C, 0574A Ne nos inferas in tentationem → "Do not bring us into temptation", not conformed to "lead us not" (Vulg. inducas). 7a″.
• 0570B haec interim duo … quae fidelia corda sustineant has no question mark on the plate, so it is rendered as a statement ("which faithful hearts must endure"). No mark is supplied (Pattern 8). Garnier's non sustinent is only in note j.
• 0574C Da mihi intellectum, ut discam mandata tua (Psal. CXVIII, 125): this is the wording of Ps. 118:73 under the locator of 118:125. It is rendered as printed. Not a [var:]: the locator is Migne's, and the wording construes.
• 0574C–0575A Jubetur ut habeamus continentiam, ubi dicitur, Cum scirem …: the command half is missing, and Coustant's note d (placed at the plate) says so and gives the restoration from Aug. ep. 147. It is rendered as printed.
• 0575A Col. 1:10 is printed in omni opere et sermone bono (Vulg. in omni opere bono fructificantes). It is rendered "in every good work and word". This is a [var:] candidate, not fired: the words are added, but the sense is not reversed.
• 0572B Ps. 3:3 Non est salus illi (Vulg. ipsi): same sense, no marker.

Coustant's notes: readings on the plate

• 0548A note c: the Jerome quotation reads regulus quaeris (Jerome has regulas). The italic vowel is worn and could be a or u. The Latin keeps the sweep's regulus, and the English renders "rules".
• 0558A note e: Coustant prints quando aberat, ac nec necessitas urgebat ("and not even necessity pressed"), which sits oddly with the argument (the bishop permitted this in necessity). It is rendered literally, with the negation carried (7a). nec may be a slip for ne or for nothing.
• 0545A note b: ut·ue was read as utque. The text's own utque illi at 0545A confirms it.
• 0545B note d: Merlin's corrupt reading is printed eamdem quum/quam praest·loris; it is written quam praestaturis. Where letters are worn, see the note-edits file.
• 0544 note g: moram, pati nesciit, quin …. The commas are on the plate and have no English exponent.
• p. 287 note h is not truncated, although the brief lists it among the truncated notes. It runs from the foot of the left column to the head of the right foot and ends Sufinianus.
• p. 283 note b has a second sentence at the head of the right foot, which the sweep missed. It is now added.
• p. 284 note d (0558A) goes on for two long paragraphs on the right foot, which the sweep missed. It is now added.
• p. 283 note d (0554B) prints Meldense ann. 843; c. 44, while the same council is ann. 845 in note a of p. 281. Both are rendered as printed. The discrepancy is Coustant's (or the compositor's).

Sequence checks (per page)

277 g (mine; a–f T3) · 278 a–h + tail of 277 g · 279 a–h · 280 a–g · 281 a–f + tail of 280 g · 282 a–d · 283 a–f · 284 a–e · 285 a–h · 286 a–b · 287 a–p · 288 a–g · 289 a–f · 290 a–l · 291 a–g · 292 a–d · 293 a–d (c, d are T5's). No gaps. No note was missed by the sweep as a whole. The two missing continuations (283 b, 284 d) are listed above.

Cols 0576B–0594A, PL 20 PDF pp. 293–302 (raw/scans/pl020/pl020.pdf, 300 dpi, corners read first on every page; zooms at 300/400 dpi where noted). Every page in the range was opened once, the text collated and the foot notes read in the same visit. Content: the close of the five bishops' letter to Innocent (Ep. XXVIII, §§8–17 = Augustine Ep. 177), Ep. XXIX to the council of Carthage, Ep. XXX to the council of Milevis, and the head and inscription of Ep. XXXI.

Standing decisions

• Coustant's lettered notes. Latin [cn: <letter> | …] at the key; the English carries a translated [nt: …] at the same point and no [cn:] (Pattern 19, P0, verifier 9a), as T3 did.
• Sigla as T3: Merlin, Crabbe, Quesnel, Garnier, Labbe, "the Roman edition", "the Councils" / "the edition of the Councils", "the Isidorian (collection/manuscripts)", "the Augustine edition" (Aug., August., the Benedictine Augustine), manuscripts as "the Colbert / Corbie / Saint-Germain (Germ.) / Saint-Remi (Remig.) / Pithou / Navarre / Vatican manuscript". Bare variant words stay Latin italic in the printed case.
• Pattern 17. The five bishops write to Innocent in the singular (thou); Innocent's rhetorical opponent in Ep. XXIX §3 (te negare… consequeris) is singular (thou); the councils of Carthage and Milevis are plural (you).
• liberum arbitrium = "free choice" throughout; possibilitas (naturae) = "possibility (of nature)"; gratis mortuus est = "died in vain"; adjutorium = "help", auxilium = "aid".
• Migne's inline [Forte leg. praeciperent] (0583B) is his apparatus and is carried verbatim, untranslated.

The sweep's key positions (see data/briefs/7090-NOTE-EDITS-0024.md)

32 of the injected notes in this range sat one word (or one clause) before the key the plate prints, two of them inside a ## section head (0582 h, 0586 i). All were moved to the printed key. Seven unplaced notes were placed by eye (0577 a, 0580 h, 0581 a, 0586 g, and the heads 0588 k, 0593 f; 0592 n with its truncated text completed). Per-page key sequences are all gapless; no note was missed by the sweep. Eleven note texts corrected at the plate (list in the edits file).

Cruces in the text

• 0576D evacuatum (Ergo evacuatum est scandalum crucis) — the plate glyph reads like evacuutum, a worn a on the 1-bit scan. Single letter, one witness: no patch, no marker; rendered "made void".
• 0579C beattiudini (tuae beattiudini potius credidimus dirigendam) — Migne's type: the plate prints the transposed non-word, and our TEI carries the same. [sic: beattiudini] beside the sense "thy blessedness". (Whole-word transposition, legible on the plate; not a single worn letter.)
• 0587 margin — the gutter letter prints C twice on p. 299 (at Quis, rogo / praecipitiis and again at sentire / sit proficere), no D. Our doubled [0587C] and [0588C] are the plate's. Anchors carried verbatim.
• 0589B–0590A in quam (honoris, in quam, illius) — the plate prints inquam split. Pattern 10: I say [sic: in quam].
• 0583B quid praecipere … quos abluere — the infinitives have no governing verb; Migne's own bracket [Forte leg. praeciperent] and Coustant's note b (supply deberent) both mark it. Rendered literally: "what to command … whom to wash".
• 0586A Si ergo hoc ille ita magnus scit ut assidue dicere, ita necessarium confessus est, ut doceret — does not construe (Coustant's note h: "we should prefer ita magnus scivit assidue dicere, et ita"). Rendered as closely as the words allow: "If then he, so great, knows so to say this continually, so confessed it necessary as to teach it".
• 0586B Liberum enim arbitrium olim ille perpessus — Coustant (note i) says the sense wants permissus; the printed perpessus is rendered "having undergone free choice".
• 0586B hujus ruinae jacuisset oppressu — oppressu (abl. noun), rendered "would have lain under the crushing of this ruin"; note j discusses the variants.
• 0592A celeri — plate prints celeri veritatis … ratione; our TEI has the non-word ceferi. English follows the plate ("by the swift reason of truth"); TEI patch filed.
• 0583B regulis. — the plate has no stop after regulis; the English has none; patch filed.
• 0582B laudat, Praetextato — plate forms (TEI laudet, PRAETENTATO); English follows the plate; patches filed.
• 0593A nolle se mortem morientis — "the death of the dying": a paraphrase of Ezek 33:11 (Vulg. mortem impii), not set as a quotation; no marker.

Cruces in the notes (rendered in the [nt:])

• 0582 e (head of Ep. XXIX) Quae autem 29 orat, nunc 7 — the plate prints orat (for erat); carried in italics: "And what orat 29 is now 7".
• 0587 a unde matilus judicetur — the plate prints matilus (for mutilus, Garnier's mutilum just above); carried in italics.
• 0581 b corriguntur ex Quesn. Auguts. — printed Auguts.; the [nt:] renders "the Augustine edition" (abbreviation, no sense at stake).
• 0582 d verum etiam no ter licet exiguus — broken type for noster, kept inside the italic Latin.
• 0583C c parete — the plate glyph is ambiguous between parete and parere; left as the sweep read it.
• 0593 f (head of Ep. XXXI) Quae autem erat 59, nunc 5 — numeral worn (59 or 39); left 59.
• 0592 o tuari (for tutari, quoting the letter's opening) — rendered "protect".
• 0592 n digni sunt poena. haud scio — the plate's stop before lower-case haud is kept.

Scripture as printed (7a″ deliverable)

[var:] fired where the printed wording changes the sense:

• 0578A Jn 8:56 Abraham concupivit videre diem meum — Vulg. exsultavit ut videret: "desired to see" vs "rejoiced that he might see". [var:].
• 0591A Ps 51:9 Ecce homines, qui non posuerunt Deum adjutorem sibi — plural; Vulg. Ecce homo qui non posuit. [var:].
• 0592B Lk 18:16 Sinite infantes … talium est enim regnum coelorum — Vulg. pueros … regnum Dei. [var:].

Logged, no marker (same sense, or paraphrase not set as quotation):

• 0576B Gal 2:21 Non irritam facio (Vulg. Non abjicio): "I do not make void".
• 0576C Rom 4:5 deputatur (Vulg. reputatur).
• 0578A Gal 3:19 Transgressionis gratia posita est … dispositum (Vulg. Propter transgressiones posita est … ordinata): "set for the sake of transgression … ordained".
• 0578D 1 Cor 15:56 Aculeus (Vulg. Stimulus), 0580A 1 Cor 15:55 aculeus: both "sting".
• 0580B Lk 1:6 lacks Vulg. mandatis et before justificationibus: rendered as printed.
• 0581A 1 Jn 3:9 Qui natus est ex Deo, non peccat (Vulg. Omnis qui … peccatum non facit).
• 0582A Mt 6:13 Ne nos inferas (Vulg. inducas): "Bring us not into temptation".
• 0592A Jn 6:54 in the third person (manducaverint … non habebunt vitam in semetipsis; Vulg. second person): the argument concerns the infants, so the third person is Innocent's own adaptation; rendered as printed, no marker.
• 0592B Gal 5:12 + 1:7 conflated (Abscidendi sunt enim, qui vos conturbant, et volunt convertere Evangelium Christi), not set in italics: rendered as printed.

Cols 0594A–0612C, PL 20 PDF pp. 302–312 (raw/scans/pl020/pl020.pdf, 300 dpi, corners read first on every page; zooms at 1.6–3× where noted). Every page in the range was opened, the text collated and the foot notes read. p. 311 (611/612) carries no notes; p. 312 (613/614) is the Notitia appendix, outside the work. The work ends at 0612C (Laus Deo in aeternum! Amen.).

Standing decisions

• Coustant's lettered notes. Latin [cn: <letter> | …] at the printed key; the English carries only a translated [nt: …] at the same point, no [cn:] (Pattern 19, P0, verifier 9a). Bare variant words stay Latin italic in the case printed; sigla expanded (Merlin, Crabbe, Quesnel, Garnier, Labbe, Baronius, the Roman edition, the edition of the Councils, the Isidorian collection); manuscript sigla rendered "the Colbert / Corbie / Pithou / Vatican / Royal / Navarre manuscript". Greek words in notes (τῷ ποιμνίῳ ἑαυτοῦ, τὴν εὐσέβειαν) carried in Greek script.
• Quotations inside notes. The Latin notes set Augustine's words in << >>; the English twins use straight " " so that the body guillemet parity check (Pattern 6) still compares the text only.
• Pattern 17. Ep. XXXI (five bishops) you; Ep. XXXII (Aurelius) thou; Ep. XXXIV (Jerome) thou, but multis saepe narrastis is plural and stays you have often related; Ep. XXXV (John) thou; Ep. XXXVII (Felix) thou, with the plural oblitos vos esse dicimus → you; Epp. XXXVIII–XXXIX you; Epp. XL–XLI thou (apud vos in XLI → among you); Pelagius to Innocent, apud beatitudinem tuam → thy beatitude; Augustine to John (Monitum) thou, then Si diligitis you.
• Greek versions (brief item d). Migne prints no Greek text of any letter in this range. Coustant argues (Monitum §4 and notes d, f to Ep. XXXV) that Ep. XXXV was translated from Greek; that is his argument, carried in the notes, not a printed Greek version. Ep. XLIII survives in Mai's Latin from an Arabic codex. No Greek-vs-Latin cruces arise.
• Heads. CAP. → CHAP. as in 0006–0011 (0015 keeps CAP.; left for merge reconciliation).

Plate vs TEI (English follows the plate; patches filed in data/briefs/7090-PENDING-TEI-PATCHES-0028.md)

Migne's own type (plate read, TEI agrees)

• 0598A fratrer charissime (p. 304, clear). Non-word, carried as [sic: fratrer] beside "dearest brother". Intended frater.
• 0599A Cathaginensi (p. 305, clear). Carried as [sic: Cathaginensi] beside "Carthage"; intended Carthaginensi.
• 0601A head Joannem Hieronymum Hierosolymitanum (p. 306, italic, clear). Rendered literally "That John Jerome of Jerusalem". The letter is to John alone; Hieronymum looks like an intrusion from the neighbouring Jerome letters, but the plate prints it. No marker (a real word).
• 0602A conqueruntur after plus … quam (p. 306): indicative where Coustant's own quotation of the passage in note i to Ep. XXXIV has conquerantur. Both render "complain"; no English exponent.
• 0606A epicopis Apulis (p. 308): Migne's spelling; rendered "bishops of Apulia", Pattern 9 (no marker: the head's facing Latin shows it).
• 0600B ut ratio non tractandi consulendique sit — rendered literally "that there is reason for not treating and consulting", which says the opposite of the context; Coustant's note f says a word is wanting and proposes ut nulla ratio. The note carries the conjecture; no [cj:] (P0/ Pattern 19: Migne's apparatus already answers it).
• 0601A–B Ep. XXXV, tua lacessit negligentia admittere — rendered "which … thy negligence provokes to admit into the flock of the Lord"; construes only loosely, as Coustant's notes d–f say of the whole letter (Greek idiom in Latin). Rendered as printed.

Notes: text as printed that looks wrong (kept, not emended)

• 0594B note m: quinque e, i-covorum n. 14 — broken type for episcoporum (p. 302, re-read at 0.9× and 2×). Carried in italics in the twin.
• 0594A note i: pos int, idoneo dato cuiris — the readings Coustant reports; printed so.
• 0595B note g: qui judicarerunt — printed so in the note; carried.
• 0602A note j (Jerome to Demetrias): renumdare — carried untranslated in the twin (Jerome's text has venundare, "to sell"); the twin italic run is left roman inside the italic quotation.
• 0607A note d: the plate sets sedulioremin ea without a visible space in a tight line; entered in the Latin note as seduliorem in ea (logged in data/briefs/7090-NOTE-EDITS-0028.md).

Note placement, by page (full list of every edit in data/briefs/7090-NOTE-EDITS-0028.md)

• Sequences checked per page, all gapless: 302 a–n (f, g are chunk 0027's), 303 a–l, 304 a–i, 305 a–i, 306 a–l, 307 a–j, 308 a–k, 309 a–g, 310 a–h; 311 none.
• Placed by hand (were unplaced): 302 i; 304 d, e, i; 305 h; 306 a, i. All seven of the unplaced-notes list that fall in chunks 0028–0031. None left unplaced.
• Not mine: 302 f (EPISTOLA XXXI head, unplaced, "In edit. Concil. 26. Quae antem erat 59, nunc 5.") keys the head of Ep. XXXI, which is in chunk 0027. Left for that stint/the merge.
• 35 auto-placed keys sat one or more words from Migne's raised letter (the injector put most keys BEFORE the anchor word the sweep recorded after the key); moved to the printed position.
• Completed or corrected note texts: 0028 d; 0029 g, h, i (spliced text replaced), e, l; 0029 a; 0030 g (p. 307), b, g (p. 308), k, a, b, c, d (p. 309); 0031 e, g.

Word ratio

EN/LA (body words, apparatus stripped): 0028 1.50, 0029 1.49, 0030 1.47, 0031 1.39.