MIGNE.APP

Cruces

Dialogi de processione Spiritus Sancti (dial. I + fragments II–VI)

Nicetas of Maroneia — read the work →

These are the edition's working notes, published as they were written. They record, work by work, where Migne's plate is defective or the reading uncertain: the printed form, what the English does with it, and why. They are not a groomed apparatus. Where a note argues with itself, withdraws an earlier decision, or leaves a question open, that is the record of how the reading was actually reached, and it is left standing — a conjecture that is visible can be checked, and one that has been tidied away cannot.

The English never silently repairs the page. A construable but wrong reading is rendered literally and the conjecture is recorded here, never applied to the text. How the English is produced and checked →

Merged 2026-08-01 from cruces-0000.md (Fable anchor), cruces-0001.md (chunks 0001–0005) and cruces-0006.md (chunks 0006–0009). The three source files follow verbatim below; this header is the authority where they disagree. Runbook step 4: the merge is a required step, and reconciling convention drift is its substance — concatenation is the easy half.

WORK-WIDE CONVENTIONS (authoritative)

Carried from the anchor, uncontested: speaker sigla ΓΡ./ΛΑΤ. → GR./LAT., one paragraph per turn · section letter-numerals as arabic + period · οὐσία = essence · φύσις = nature · ὑπόστασις = hypostasis · πρόσωπον = person · ἐκπόρευσις = procession · ὑπερ- compounds take "super-" · mentioned particles stay in Greek script, quoted credal clauses translated in quotation marks · a heading the plate carries but Calfa drops is restored as inline [ed:], never as a ## head.

Uncontested additions from the two halves: ἀγένητον = "ingenerate" vs ἀγέννητον = "unbegotten" (with the orthographic parentheses translated, the letter kept in Greek) · πρόοδος = "going-forth", never "procession" · ἐκπορευτόν = "proceeding" / "that which proceeds" · ἄμεσος/ἔμμεσος = immediate/mediate · ἀρχή = "principle" (hence ἄναρχος = "without principle", ὑπεράρχιος = "super-principial") · αἴτιον = "cause", kept distinct from ἀρχή · ὕφεσις = "diminution" · ὑπεράριθμος / ὑπερφυής / ὑπερταγής = super-numerical / super-natural / super-ordered · πατριά = "fatherhood" on the authority of Migne's own note (y) · homoousion carried untranslated in italics · γηγενάρχης / ἀγγελάρχης = head of the earth-born / chief of the angels · παρόμοιον = "nearly-like" (semi-Arian term, not "similar").

⚑ CONFLICTS RULED AT MERGE — the text was edited to match

The two halves settled the same vocabulary differently, and both renderings were live in the shipped English until this merge. This is the 11551 stylus/epistylium failure class: neither agent could see the other.

  1. The προβολ- family — RULED "put forth". - 0001–0005 had: προβάλλεσθαι = put forth · προβολεύς = putter-forth · προβλητικόν/πρόβλησις = putting-forth. - 0006–0009 had: προβάλλεσθαι = produce · προβολεύς = producer. - Ruling: the "put forth" family, throughout. Not by majority (20 vs 6) but on principle: Latin producere flattens three distinct Greek verbs here, and "produce" is simply the English of producere. Adopting it would import Allatius's flattening into our English through the back door — the precise thing PLAN.md rule 4 forbids. "Put forth" answers to no single Latin word. - Edited in the text: 0006.md — "producer/producer" → "putter-forth/ putter-forth" (the τίς ἡ διαφορὰ προβολέως καὶ προβολέως question, where the repetition IS the argument), "producing…produces" → "putting forth…puts it forth", and "not to beget or to produce" → "…or to put forth" (Greek: οὐ τὸ γεννᾷν ἢ προβάλλεσθαι).
  1. The προϊέναι family — RULED "send forth". - 0001–0005 had "emit" (2 sites); 0006–0009 had "send forth" (1 site). - Ruling: "send forth". Same principle: Allatius's own word here is emittere, and "emit" mirrors it. "Send forth" also keeps the author's deliberate near-synonymy with προβάλλεσθαι visible without collapsing the two verbs (0206 sets them in the same sentence: the Father puts forth mediately, the Son immediately sends forth the same). - Edited in the text: 0001.md "to put forth or to emit" → "…or to send forth"; 0002.md "he emits the Spirit" → "he sends forth the Spirit".

⚠ Trap recorded for future merges. cruces-0006.md states the προΐεσθαι convention citing col 0201 — but chunks 0006–0009 contain no προΐε- form under that spelling; the site is προῖεται (circumflex iota, no diaeresis), and both textbook προΐε- forms are in the peer's chunks. A merge that trusted the cruces prose without checking the Greek would have "reconciled" two correct renderings into a convention declared for a word the declaring agent never actually translated. Check the Greek, not the cruces claim.

Verified after merge

node scripts/verify-english-pg.mjs nicetas-maroneia-dialogi → 10 chunks, anchors + frontmatter + ratios + dedupe clean. No producer/produce or emit survives in the work; the προβολ- family reads 36 instances of the put-forth root, προϊέναι 4 of send-forth.

Work-wide conventions (anchor session, 2026-07-31 — merge into the work-wide section)

Latin-twin pass (pilot §6): result for chunk 0000 — FINDINGS (not clean)

Read against src/pg-latin/nicetas-maroneia-dialogi/0000.md (cols 0170–0175) after drafting. Sites follow. No chunk in this batch was clean; this batch is chunk 0000 only.

Ladder row 1 — Calfa OCR damage, corrected silently in the English, Latin/plate cited

Pattern 13 — [ed:], digitization lost text the plate carries

Borderline Pattern-16 case — logged, NO marker (§8 Q3 calibration datum)

Plate-apparatus notes carried by the twin (not ours to mark; future fontes/scripture harvest)

The plate's notes for this span, visible in the twin and on leaf renders: (2) Laur. ἐπιεικείᾳ · (3) Laur. διαλύειν · (4) Sic Laur., Maius omittit τοῦ λόγου · (5) Apud Maium: καὶ ἀλληλότερα · (6) I Cor. II, 10 · (7) Laur: περί · (8) Cod. V: λέγε; fort. λεγεις · (9) Isa. XL, 4 · (10) Psal. LI (Vulg. L.) 10 · (11) Isa. l.c. · (b)–(e) editorial (Maius, Etherianus cross-refs). Calfa strips all note anchors from the Greek, so the English carries no [n:] markers — the identifications survive in the committed twin.

Work-wide conventions of cruces-0000.md are in force and are not restated. This file adds only what chunks 0001–0005 forced. Scan used for every plate check below: raw/scans/pg139/patrologiaecurs63migngoog_djvu.xml, leaf index = position in the file's <OBJECT> sequence; for this span leaf n carries Migne cols (2n+3)/(2n+4) (leaf 89 = 181/182 … leaf 97 = 197/198), verified against the running heads. ⚠ Formula corrected 2026-08-01 — this file originally wrote it as (2n−7)/(2n−6), which contradicts its own worked examples (2·89−7 = 171, not 181). The examples were right, the formula wrong; a blind reader caught it while using it. Arithmetic in a method note gets checked against its own examples before it is trusted.

Conventions this batch had to decide (anchor did not cover)

[lat:] markers fired — 2

  1. 0185 (twin col 0186) — a NEGATION the two columns do not share. Greek: ἀλλ᾿ οὐ σύνθεσις ἐκεῖ, ἀλλ᾿ ἁπλότης ὑπὲρ πᾶσαν ἁπλότητα, ὅτι μέρη ταῦτα τῆς θεότητος (ἀμερὴς γὰρ ὑπὲρ πᾶσαν ἀμέρειαν), ἀλλ᾿ ἰδιότητες… Latin: "Verum non est compositio ibi, sed simplicitas super omnem simplicitatem, quia nec partes Divinitatis ista sunt quum ea sit super omnem indivisibilitatem indivisibilis, verum proprietates…" PLATE-VERIFIED on the Greek side (leaf 91): the scan's independent OCR reads ὅτι µέρη ταῦτα τῆς θεότητος (ἀμερὴς γὰρ ὑπὲρ πᾶσαν ἀμέρειαν) — no οὐ, parenthesis present. So the Greek plate prints the sentence without the negative that its own ἀλλ᾿ ἰδιότητες requires; the Latin column supplies it. English follows the printed Greek ("because these are parts of the Godhead") and marks the Latin. Corroboration that the omission is the plate's and not a Calfa drop: the exactly parallel sentence at col 0188 (chunk 0003) does print the negative — ἀλλ᾿ οὐκ οὐσία ταῦτα … οὐδὲ μέρη τῆς οὐσίας· ἡ θεότης γὰρ ἀμερής — so the author's formula normally carries it. This is the dropped-negative class Pattern 7 exists to expose.
  2. 0189/0190 (twin col 0190) — the reciprocal-ratio terms. Greek (plate-verified, leaf 93): νῦν μὲν γινομένη τὸ ἥμισυ καὶ τὸ ὑποδιπλάσιον, νῦν δὲ τρίτον καὶ ὑποτριπλάσιον, νῦν δὲ τέταρτον καὶ ὑποτε[τρα]πλάσιον. Latin: "nunc quidem fit dimidium et duplum, nunc vero tertium et triplum, nunc quartum et quadruplum." Allatius drops the ὑπο-, which turns each pair from two names for the same ratio (half = subduple) into two opposite ratios (half and double). A reader of the Latin column takes away a different mathematical claim, so this clears Pattern 16's materiality bar. Numeral-class, so the Greek was plate-checked before the marker fired, per pg-paired-pilot.md §4.

Logged, NO marker — the plate's own apparatus or column concurrence settles it

Ladder row 1 — Calfa OCR damage, corrected silently in the English

Negations and lost parentheses (the load-bearing class — every one checked against the plate before it was restored):

Lost speaker tags, section marks and lost text (structural):

Intrusive-Κ artifact class (the anchor's ΠνενύΚματος class, still running): 0177 Κτὴν ὑπεράγαθον→τὴν · 0184 Κγεννητὸς→γεννητός · 0189 περὶ Καὐτήν→αὐτήν, τῷ λόγῳ Κδιακέκριται→διακέκριται, Κγινομένη→γινομένη, σχεΚδόν→σχεδόν, ἄγνωστόν Κτε→τε · 0193 Κφῶς→φῶς · 0200 Κφαινόμενα→ φαινόμενα.

Routine OCR, bundled (no sense at issue): 0176 Οὔτω γὰρἔδοξε, ἐοτιν, ταύτόν, Πατοὶ καὶ ἰἵῷ, υσικά, κατ᾿ οὐείαν, ἁλλά, ἑπί, ἀτδιον, ἀπερίγοαπτον, φυσικἀ, κοινῷς, Ὑπόσασιν, δηλαόή, ἀνυπἑλάκτως→ ἀνυπαλλάκτως, Υἱνῦ, ἷέννησιν, Πατοὸς, τοῦΓἰνεύαατος, Ἑνος→μόνος (Lat. solus … Unigenitus), ἔκν τινος, Υὸς, καθ᾿ δν, Υἱλν, εἰαί→εἰμί, ἐξεκάλυύε→ἐξεκάλυψε, Υλοῦ, Ινεύματος · 0184 χεννῶν, πρυνθάλλεται, Πεῦμα, ἐκπορεύόται, Ποῦ→Υἱοῦ, Γίοῦ, μόναν→μόνον, Κυρίον, λατήρ→Πατήρ, ἐχόντοιν οὐδ᾿ Ἀν ἀνομάτων→οὐδ᾿ ἐν ὀνομάτων, Κἰ γάρ→Εἰ γάρ, παναγίευ, ἐσὶ→ἐπί, Σὲ→δέ, Ἰδιότης · 0188–0192 σύνθισιν, λἱός, ἐκπορετόν→ἐκπορευτόν, παριληπτικον→περιληπτικόν, ψποστάσεων, ἰὸιότητας, προῖοῦτοα Κἐ ἐογειαν→προϊοῦσαν ἐνέργειαν, ἀλλδμως ὅλλο→ἀλλ᾿ ὅμως ἄλλο, ἀριθμοῖς, τρίο ναὶ ὑποτοιπλάσιον→τρίτον καὶ ὑποτριπλάσιον, τετράγωνες, ποντάγωνος, φεινομένη, ἄπλουστάηη, ὑποσπτάσεων, ἐνυΤαρχούσης, συνθςτοις, τὸ!] εῦμα→τὸ Πνεῦμα · 0193–0196 ἐκπρρελόμενον, Πεὖμα, Ιρῶτον, ῦτο γάρ→τοῦτο γάρ, Ιἱός, Θ γάρ→Ἢ γάρ, ἐκτοῦ ΥΙοῦ, ακρων, τριων, νεύματος→Πνεύματος, κ τοῦ ΙΙατρὸς→ἐκ τοῦ Πατρός · 0197–0200 ἡ οὲ→ἡ δέ, ατελής, ἐκπύρευσις/ἐκπίρευσιν, τοῦΠατρός, διαροράν, παρεμφαίνουσ, Ὁμὲν, Ιατρός/Γατρός, Οίοῦ, χωοίζονται, ἐκ τοῶ Πασοός, Ὥοπερ, ἕκαστυν, Υἱλς, δτι, συνορῷ, Ἱσως, Πνεῦρα το ἅγιον, εἴῷ, λἱοῦ, υἱωνὸςὲ→υἱωνὸς δέ, φθέγγεσθνι, ἐκατερα, ἐν τῷ αὐτῳ, αὐτων, ποὸς.

One correction taken on plate authority against Calfa: 0196 (chunk 0004) Calfa Τῆς γὰρ μονᾶς ὁποτερασοῦν οὔσης τελείας; plate (leaf 96, verified) Τῆς γὰρ μιᾶς ὁποτερασοῦν οὔσης τελείας, Latin "Si enim una…". Rendered "the one of them, whichever it be". Left unmarked because the plate settles it; worth noting only because μονάς is a live technical term four lines earlier and a reader might expect it here.

Pattern 7 non-word type carried into the English

Scripture quoted in these chunks — divergences from the received text: NONE

Two only, both verbatim against the received Greek text, both matching the Latin column:

Latin-twin pass (pilot §6) — per-chunk result

Every chunk got the deliberate post-draft pass against its twin. Outcome:

No chunk in this batch came back clean. Reporting that plainly, since the pilot asks for it: 0/5 clean, and in four of the five the finding was a defect in our file rather than a divergence between Migne's columns — which is the result the standing PG attribution rule predicts and the reason the twin plus the plate were both needed.

Plate apparatus visible in the twin for this span (not ours to mark)

Numeric notes (Greek column): (15) Ita scriptum. Quod tamen mendum exhibere videtur. [on μέση, 0184] · (16) Cod. Vat. f. 5,6, mendose, ὑπόθεσις · (18) Cod. Vat. Ἐκπορεύεται · (19) marginal, Gk: Ἐνταῦθα εἰσαγωγὴ ἑτέρου ἀτόπου ὅτι διπλῆ ἔσται ἡ τοῦ Πνεύματος ἐκπόρευσις… · (21) Frequen[ter]… · (23) marginal, Gk: Ἕτερον ἄτοπον ὅτι ἐκ τοῦ Πατρὸς καὶ ἐκ τοῦ Υἱοῦ ἐκπορευομένου τοῦ Πνεύματος, ἢ ἀτελὴς ἡ ἐκπόρευσις ἑκατέρα ἢ περιττὴ ἡ ἑτέρα · (24) Cod. ἀδύνατον · (26) Vat. εἰ · (28) Vat. τελείων · (29) marginal, Gk: Ἕτερον ἄτοπον, ὅτι εἰ ἐκ τοῦ Υἱοῦ τὸ Πνεῦμα υἱὸς ἔσται τοῦ Υἱοῦ. Letter notes (Latin column): (f) Elias of Crete on Greg. Naz. or. 11 + Arist. Metaph. IV · (k) S. Maximus, Opusc. theol. et polem. II p. 155 ed. Combefis; ps.-Justin Expos. rectae fidei n. 3 · (m)/(4) on προβάλλεσθαι being middle, not passive · (n)/(o) the two marginal aliud absurdum notes, (o) adding the cross-reference to Photius, De Spiritus sancti mystagogia c. 7 and Nicholas of Methone c. 11 for the Greek's fourth argument · (p) Haec sane non satis accurate dicta sunt + the attack on mediate / immediate · (q), (r) (imitatur Gregorii Nazianzeni responsum or. 27) · (s) Quintum argumentum. Calfa strips all note anchors from the Greek, so the English carries no [n:] markers; the identifications survive in the committed twin and are listed here for the eventual fontes/scripture harvest.

Second half of the work (chunks 0006–0009 of 10). Chunks 0001–0005 were held by a peer agent who could not see this file; the merge is a required later step, not done here. Conventions below are the anchor's (cruces-0000.md) unless flagged as decided fresh.

Method note — what "plate-verified" means in this batch

The Latin twin was cropped from raw/scans/pg139/patrologiaecurs63migngoog_djvu.xml by x-range split. The same leaves carry Migne's GREEK column in that XML, and its OCR is independent of Calfa's. Every reading marked plate-verified below was checked against the Greek column of the named leaf (leaves 98–108 = cols 0201–0221), so the evidence is a second witness to the plate, not a second reading of our own file. Where the scan OCR is itself illegible at the point at issue, that is said.

Conventions carried from the anchor (unchanged)

Speaker sigla ΓΡ./ΛΑΤ. → GR./LAT., one paragraph per turn · section letter-numerals as arabic + period · οὐσία = essence, φύσις = nature, ὑπόστασις = hypostasis, πρόσωπον = person, ἐκπόρευσις = procession · ὑπερ- compounds take "super-" · mentioned particles stay in Greek script, quoted credal clauses translated in " ".

Conventions decided fresh (flag for the merge)

These terms first fall due in this half of the work; the peer's half may have settled some of them differently.

Latin-twin pass (pilot §6) — result by chunk

Performed after drafting, against src/pg-latin/nicetas-maroneia-dialogi/000[6-9].md.

No chunk in this batch was clean of Calfa OCR damage. The twin pass found no "Greek made no sense and the Latin explains why" case that the plate's own Greek column did not settle more directly.

Pattern 16 — [lat: …] fired: ONE

Pattern 16 — candidates LOGGED, marker NOT fired

Pattern 14 — [var: …] fired: TWO (the divergence list)

Both are quotations Migne prints in a form the received text does not carry. These are the most valuable findings in the batch, and the search for them was word-by-word against the plate, not against memory.

Pattern 13 — [ed: …], text the DIGITIZATION lost that the plate carries

This work is Hergenröther's set of fragments of dialogues II–VI, and its structural furniture is exactly what Calfa strips. Five centered division headings and two editorial bridges are gone from our Greek. All are restored as [ed: …], never as ## heads (marker parity with the Greek chunk), following the anchor's ΛΟΓΟΣ ΠΡΩΤΟΣ precedent.

Fragment boundaries as found: dialogue II opens at col. 0201/0202 and the work runs to a stop at col. 0221 in mid-argument, where PG 139 turns to the next author (Joannes of Citrus). Ellipses (…) inside the Greek are Hergenröther's own and are reproduced as printed; no gap anywhere was bridged with invented connective prose.

Ladder row 1 — Calfa OCR damage, corrected silently in the English

Itemized inline these would drown the text (register rule). Grouped by column. Unless marked otherwise, the correction is confirmed by the plate's Greek column on the leaf named in the section head above.

Polarity- and sense-bearing (these are the ones that mattered):

Speaker sigla mangled by Calfa (all restored, all plate-verified): 0201 ΠΡ. → ΓΡ.; 0201 ΑΑΤ. → ΛΑΤ.; 0201 ΤΡ. → ΓΡ.; 0204 ΛΛΤ. → ΛΑΤ.; 0213 λΑΤ → ΛΑΤ.; 0216 ΕΡ. → ΓΡ.; 0216 ΛΑΤ.θ → ΛΑΤ.; 0212 ΑΑΤ. → ΛΑΤ. 0213: Calfa drops the ΓΡ. tag entirely before Πάραγε δή μοι καὶ ταῦτα, silently handing the Greek's speech to the Latin; the plate (leaf 104) and the twin (GR. Profer jam mihi) both carry it. Restored.

Section marks: 0205 νʹ → γʹ ("3.") — plate leaf 99 reads γ'., twin reads 3.; Calfa's ν is an OCR of γ, and left standing it would have printed "50." in the English. 0201 αʹ restored (lost with the heading, above).

The intrusive-Κ class (already named in the anchor's cruces for chunk 0000, and still running at this end of the volume): 0209 Κθεοί → θεοί · 0209 Κγὰρ → γάρ · 0213 Κἵκα → ἵνα · 0213 Κἐμφαῖνον → ἐμφαῖνον · 0217 Κῥνωθεν → ἄνωθεν · 0217 Κἐτράνωσεν → ἐτράνωσεν · 0220 Κσυμβόλῳ → συμβόλῳ.

Dropped opening parenthesis class (same defect as the anchor's 0172 ἀδύνατονὐδὲ): 0209 ἁπασῶνπὲρ φύσιν → ἁπασῶν (ὑπὲρ φύσιν · 0209 ἁπασῶνκὲρ τάξιν → ἁπασῶν (ὑπὲρ τάξιν · 0209 διαίρεσιν παρεισάγουσαα γὰρ ἡ φύσις → διαίρεσιν παρεισάγουσα (μία γὰρ ἡ φύσις · 0209 θέσινπερίγραπτος → θέσιν (ἀπερίγραπτος (see above) · 0209 τὸ ἅμαν οἷς γὰρ → τὸ ἅμα (ἐν οἷς γὰρ.

Routine letter-level damage, bundled (correction ← Calfa): 0201 ἀναλαβόντες kept (Calfa's reading; the scan's Greek OCR gives only λαβόντες, but the twin's disputationem resumentes wants the ἀνα-) · 0201 ἑτοπα → ἄτοπα · 0201 ἐκτοῦ → ἐκ τοῦ · 0201 ατρὸς → Πατρὸς · 0201 ἐχόντοιν kept (genitive dual, both witnesses — not an error) · 0201 Πατοὸς → Πατρὸς · 0204 θΗατήρ → Ὁ Πατήρ · 0204 εἴῃνε → εἴληχε · 0204 φατὲ/φατέ as printed · 0205 ὁ ὸς → ὁ Υἱὸς · 0205 Ἀ δὲ Πατὴρ → ὁ δὲ Πατὴρ · 0205 Ιει → ἕξει · 0205 Μακεἐόνιος → Μακεδόνιος · 0205 ποοὸν → ποσόν · 0208 ποότερον → πρότερον · 0208 λέ, εται → λέγεται · 0208 τὰ κῶ γράμματα → τὰ κδʹ γράμματα (twenty-four; plate κδ', twin Viginti quatuor — a letter-numeral, verified on both sides before rendering) · 0209 Δδάμ → Ἀδάμ · 0209 Ἀιχαὴλ → Μιχαὴλ · 0209 ἀν οὐρανοῖς → ἐν οὐρανοῖς · 0209 θεαρχκῆς → θεαρχικῆς · 0209 όῦτως → οὕτως · 0209 Τίοῦ → Υἱοῦ · 0212 τίθεαμεν → τίθεμεν · 0213 Ἰ δὲ → Ἡ δὲ · 0213 Παῆρὸς → Πατρὸς · 0213 δονματιζόμενον → δογματιζόμενον · 0213 Ιατέρα → Πατέρα · 0213 ἀυδρῶς → ἀμυδρῶς · 0213 Γοαφῆς → Γραφῆς · 0213 ἐξ αὐτῆς τῆς Ὃν πραγμάτων → ἐξ αὐτῆς τῶν πραγμάτων καὶ νοημάτων φύσεως · 0213 συσκεόώνεθα → συσκεψώμεθα · 0213 Αέγω → Λέγω · 0213 καυσις → καῦσις · 0213 βύλει → βούλει · 0213 ἠ ἐνέργεια → ἡ ἐνέργεια · 0213 Γρας ῆς → Γραφῆς · 0216 ἕτερος. ὧ Λατήα τοῦ δόγματος → ἕτερος, ὦ Λατῖνε, τοῦ δόγματος · 0216 ἀσθενής ἐστε → ἐστι · 0216 παραοειγμάτων → παραδειγμάτων · 0217 Τκῦτα → Ταῦτα · 0217 τοὐτὰ → ταὐτὰ · 0217 πρὸ τούτν → πρὸ τούτων · 0217 σύνοδ ν → σύνοδον · 0217 τὸ δὲ ἄγιος σύμβολον → τὸ δὲ ἅγιον σύμβολον · 0220 προΐντο → προΐεντο · 0220 συνεγράψαντότε → συνεγράψαντό τε · 0221 βίόλοι → βίβλοι · 0221 ιοοονοῦντες → φρονοῦντες · 0221 Γνεῦμα → Πνεῦμα · 0221 οὕτιο → οὕτω · 0221 κλέος πὸ → κλέος τὸ · 0221 λένετε → λέγετε · 0221 ἐκπορευεσίλαι → ἐκπορεύεσθαι · 0221 συγγοαφῶν → συγγραφῶν · 0221 ικπόρευσιν → ἐκπόρευσιν · 0221 Καμασκηνός → Δαμασκηνός · 0221 δίᾳ → ἰδίᾳ · 0221 τὸ κ τοῦ → τὸ ἐκ τοῦ · 0221 τὸ νεῦμα → τὸ Πνεῦμα · 0221 νοῦμεν → νοοῦμεν · 0221 Πιτρός → Πατρός · 0221 προσθρκν → προσθήκην · 0221 τὸ ἐκ τοῦ ἰατρὸς καὶ ἐκ τοῦ λἱρῶ ἐν σορευόμενον → τὸ ἐκ τοῦ Πατρὸς καὶ ἐκ τοῦ Υἱοῦ ἐκπορευόμενον (plate leaf 108; twin Qui ex Patre Filioque procedit) · 0221 Πἱῷ → Υἱῷ.

One reading kept against the plate: 0221 δρῶντας (Calfa) where the scan's Greek column OCRs δρῶντος. The twin construes plural (Patrum vestigia sequentes … in symbolo … pronuntiare) and the whole address is to the Latins in the plural; the scan's singular is very likely its own OCR slip on a final letter. Rendered plural, logged so the choice is visible.

Negation audit (7a) — the doctrinal core

Every οὐ/οὐκ/οὐχ/μή/μηδέ/οὔτε/μηδὲ ὅλως in these four chunks was checked one by one against Calfa, the plate's Greek column and the twin before drafting, and again after. The dense sites, all carried through unaltered:

Pronoun-itacism audit (ἡμεῖς/ὑμεῖς). Every first/second-person plural in these chunks was checked across all three witnesses before drafting, because Calfa confuses the pair and this dialogue's whole rhetoric turns on who is being accused. All agree: 0204 κατὰ τὸν ὑμέτερον λόγον … οὐχ ἡμῖν μᾶλλον … ἀλλ' ἑαυτοῖς συνηγάγετε (twin vestram … non magis nobis, quam potius vobis ipsis); 0212 τῆς μεθ' ὑμῶν … συζητήσεως; 0217 ἡμῖν τε ἐναντιόφρονες … ὑμεῖς; 0221 καὶ ἡμῖν καὶ ὑμῖν ἔχει τὸ ἀναμφίβολον (twin et apud nos et apud vos), ἀρκέσει δὲ ὑμῖν καὶ ἡμῖν. No [lat:] candidate in this class.

Not repaired to make sense: nothing. The one place where the printed Greek does not construe on its own is 0221, Ἐν γὰρ τῇ ἀποδόσει τῆς πίστεως τὸ ἐκ τοῦ Πατρός, οὐκ ἐκ τοῦ Υἱοῦ ἐκπορευόμενον, ἰδίᾳ δὲ ἐπισημαίνεται — the first limb has no finite verb (the twin supplies exprimit). Rendered with a bracketed supplied "[it is]" rather than borrowing Allatius's verb into the English; the crux is here.

Ellipses, anchors, ratio

The Greek's own … marks (Hergenröther's elisions) are reproduced 1:1; column anchors are 1:1 with the Greek chunks ([0201][0204][0205] · [0208][0209][0212] · [0213][0216][0217][0220] · [0221]). EN:GK ratio 1.39 / 1.36 / 1.41 / 1.36 — at the dialogue anchor's 1.37×.

For the merge

  1. Diff the fresh-decided vocabulary above against chunks 0001–0005, especially ἀρχή/αἴτιον, προβολεύς/προβάλλεσθαι/προΐεσθαι, and the ὑπερ- coinages.
  2. Do not harmonize the two forms of the creed's glorification clause (anchor 0000 doubled, 0221 single) — see the [var:] note; that divergence is Migne's.
  3. The anchor's [ed:] for ΛΟΓΟΣ ΠΡΩΤΟΣ and the five in this half should end up phrased alike; they were written to the same template but not against each other.

Work-wide conventions (anchor session, 2026-07-31 — merge into the work-wide section)

Latin-twin pass (pilot §6): result for chunk 0000 — FINDINGS (not clean)

Read against src/pg-latin/nicetas-maroneia-dialogi/0000.md (cols 0170–0175) after drafting. Sites follow. No chunk in this batch was clean; this batch is chunk 0000 only.

Ladder row 1 — Calfa OCR damage, corrected silently in the English, Latin/plate cited

Pattern 13 — [ed:], digitization lost text the plate carries

Borderline Pattern-16 case — logged, NO marker (§8 Q3 calibration datum)

Plate-apparatus notes carried by the twin (not ours to mark; future fontes/scripture harvest)

The plate's notes for this span, visible in the twin and on leaf renders: (2) Laur. ἐπιεικείᾳ · (3) Laur. διαλύειν · (4) Sic Laur., Maius omittit τοῦ λόγου · (5) Apud Maium: καὶ ἀλληλότερα · (6) I Cor. II, 10 · (7) Laur: περί · (8) Cod. V: λέγε; fort. λεγεις · (9) Isa. XL, 4 · (10) Psal. LI (Vulg. L.) 10 · (11) Isa. l.c. · (b)–(e) editorial (Maius, Etherianus cross-refs). Calfa strips all note anchors from the Greek, so the English carries no [n:] markers — the identifications survive in the committed twin.

Work-wide conventions of cruces-0000.md are in force and are not restated. This file adds only what chunks 0001–0005 forced. Scan used for every plate check below: raw/scans/pg139/patrologiaecurs63migngoog_djvu.xml, leaf index = position in the file's <OBJECT> sequence; for this span leaf n carries Migne cols (2n+3)/(2n+4) (leaf 89 = 181/182 … leaf 97 = 197/198), verified against the running heads. ⚠ Formula corrected 2026-08-01 — this file originally wrote it as (2n−7)/(2n−6), which contradicts its own worked examples (2·89−7 = 171, not 181). The examples were right, the formula wrong; a blind reader caught it while using it. Arithmetic in a method note gets checked against its own examples before it is trusted.

Conventions this batch had to decide (anchor did not cover)

[lat:] markers fired — 2

  1. 0185 (twin col 0186) — a NEGATION the two columns do not share. Greek: ἀλλ᾿ οὐ σύνθεσις ἐκεῖ, ἀλλ᾿ ἁπλότης ὑπὲρ πᾶσαν ἁπλότητα, ὅτι μέρη ταῦτα τῆς θεότητος (ἀμερὴς γὰρ ὑπὲρ πᾶσαν ἀμέρειαν), ἀλλ᾿ ἰδιότητες… Latin: "Verum non est compositio ibi, sed simplicitas super omnem simplicitatem, quia nec partes Divinitatis ista sunt quum ea sit super omnem indivisibilitatem indivisibilis, verum proprietates…" PLATE-VERIFIED on the Greek side (leaf 91): the scan's independent OCR reads ὅτι µέρη ταῦτα τῆς θεότητος (ἀμερὴς γὰρ ὑπὲρ πᾶσαν ἀμέρειαν) — no οὐ, parenthesis present. So the Greek plate prints the sentence without the negative that its own ἀλλ᾿ ἰδιότητες requires; the Latin column supplies it. English follows the printed Greek ("because these are parts of the Godhead") and marks the Latin. Corroboration that the omission is the plate's and not a Calfa drop: the exactly parallel sentence at col 0188 (chunk 0003) does print the negative — ἀλλ᾿ οὐκ οὐσία ταῦτα … οὐδὲ μέρη τῆς οὐσίας· ἡ θεότης γὰρ ἀμερής — so the author's formula normally carries it. This is the dropped-negative class Pattern 7 exists to expose.
  2. 0189/0190 (twin col 0190) — the reciprocal-ratio terms. Greek (plate-verified, leaf 93): νῦν μὲν γινομένη τὸ ἥμισυ καὶ τὸ ὑποδιπλάσιον, νῦν δὲ τρίτον καὶ ὑποτριπλάσιον, νῦν δὲ τέταρτον καὶ ὑποτε[τρα]πλάσιον. Latin: "nunc quidem fit dimidium et duplum, nunc vero tertium et triplum, nunc quartum et quadruplum." Allatius drops the ὑπο-, which turns each pair from two names for the same ratio (half = subduple) into two opposite ratios (half and double). A reader of the Latin column takes away a different mathematical claim, so this clears Pattern 16's materiality bar. Numeral-class, so the Greek was plate-checked before the marker fired, per pg-paired-pilot.md §4.

Logged, NO marker — the plate's own apparatus or column concurrence settles it

Ladder row 1 — Calfa OCR damage, corrected silently in the English

Negations and lost parentheses (the load-bearing class — every one checked against the plate before it was restored):

Lost speaker tags, section marks and lost text (structural):

Intrusive-Κ artifact class (the anchor's ΠνενύΚματος class, still running): 0177 Κτὴν ὑπεράγαθον→τὴν · 0184 Κγεννητὸς→γεννητός · 0189 περὶ Καὐτήν→αὐτήν, τῷ λόγῳ Κδιακέκριται→διακέκριται, Κγινομένη→γινομένη, σχεΚδόν→σχεδόν, ἄγνωστόν Κτε→τε · 0193 Κφῶς→φῶς · 0200 Κφαινόμενα→ φαινόμενα.

Routine OCR, bundled (no sense at issue): 0176 Οὔτω γὰρἔδοξε, ἐοτιν, ταύτόν, Πατοὶ καὶ ἰἵῷ, υσικά, κατ᾿ οὐείαν, ἁλλά, ἑπί, ἀτδιον, ἀπερίγοαπτον, φυσικἀ, κοινῷς, Ὑπόσασιν, δηλαόή, ἀνυπἑλάκτως→ ἀνυπαλλάκτως, Υἱνῦ, ἷέννησιν, Πατοὸς, τοῦΓἰνεύαατος, Ἑνος→μόνος (Lat. solus … Unigenitus), ἔκν τινος, Υὸς, καθ᾿ δν, Υἱλν, εἰαί→εἰμί, ἐξεκάλυύε→ἐξεκάλυψε, Υλοῦ, Ινεύματος · 0184 χεννῶν, πρυνθάλλεται, Πεῦμα, ἐκπορεύόται, Ποῦ→Υἱοῦ, Γίοῦ, μόναν→μόνον, Κυρίον, λατήρ→Πατήρ, ἐχόντοιν οὐδ᾿ Ἀν ἀνομάτων→οὐδ᾿ ἐν ὀνομάτων, Κἰ γάρ→Εἰ γάρ, παναγίευ, ἐσὶ→ἐπί, Σὲ→δέ, Ἰδιότης · 0188–0192 σύνθισιν, λἱός, ἐκπορετόν→ἐκπορευτόν, παριληπτικον→περιληπτικόν, ψποστάσεων, ἰὸιότητας, προῖοῦτοα Κἐ ἐογειαν→προϊοῦσαν ἐνέργειαν, ἀλλδμως ὅλλο→ἀλλ᾿ ὅμως ἄλλο, ἀριθμοῖς, τρίο ναὶ ὑποτοιπλάσιον→τρίτον καὶ ὑποτριπλάσιον, τετράγωνες, ποντάγωνος, φεινομένη, ἄπλουστάηη, ὑποσπτάσεων, ἐνυΤαρχούσης, συνθςτοις, τὸ!] εῦμα→τὸ Πνεῦμα · 0193–0196 ἐκπρρελόμενον, Πεὖμα, Ιρῶτον, ῦτο γάρ→τοῦτο γάρ, Ιἱός, Θ γάρ→Ἢ γάρ, ἐκτοῦ ΥΙοῦ, ακρων, τριων, νεύματος→Πνεύματος, κ τοῦ ΙΙατρὸς→ἐκ τοῦ Πατρός · 0197–0200 ἡ οὲ→ἡ δέ, ατελής, ἐκπύρευσις/ἐκπίρευσιν, τοῦΠατρός, διαροράν, παρεμφαίνουσ, Ὁμὲν, Ιατρός/Γατρός, Οίοῦ, χωοίζονται, ἐκ τοῶ Πασοός, Ὥοπερ, ἕκαστυν, Υἱλς, δτι, συνορῷ, Ἱσως, Πνεῦρα το ἅγιον, εἴῷ, λἱοῦ, υἱωνὸςὲ→υἱωνὸς δέ, φθέγγεσθνι, ἐκατερα, ἐν τῷ αὐτῳ, αὐτων, ποὸς.

One correction taken on plate authority against Calfa: 0196 (chunk 0004) Calfa Τῆς γὰρ μονᾶς ὁποτερασοῦν οὔσης τελείας; plate (leaf 96, verified) Τῆς γὰρ μιᾶς ὁποτερασοῦν οὔσης τελείας, Latin "Si enim una…". Rendered "the one of them, whichever it be". Left unmarked because the plate settles it; worth noting only because μονάς is a live technical term four lines earlier and a reader might expect it here.

Pattern 7 non-word type carried into the English

Scripture quoted in these chunks — divergences from the received text: NONE

Two only, both verbatim against the received Greek text, both matching the Latin column:

Latin-twin pass (pilot §6) — per-chunk result

Every chunk got the deliberate post-draft pass against its twin. Outcome:

No chunk in this batch came back clean. Reporting that plainly, since the pilot asks for it: 0/5 clean, and in four of the five the finding was a defect in our file rather than a divergence between Migne's columns — which is the result the standing PG attribution rule predicts and the reason the twin plus the plate were both needed.

Plate apparatus visible in the twin for this span (not ours to mark)

Numeric notes (Greek column): (15) Ita scriptum. Quod tamen mendum exhibere videtur. [on μέση, 0184] · (16) Cod. Vat. f. 5,6, mendose, ὑπόθεσις · (18) Cod. Vat. Ἐκπορεύεται · (19) marginal, Gk: Ἐνταῦθα εἰσαγωγὴ ἑτέρου ἀτόπου ὅτι διπλῆ ἔσται ἡ τοῦ Πνεύματος ἐκπόρευσις… · (21) Frequen[ter]… · (23) marginal, Gk: Ἕτερον ἄτοπον ὅτι ἐκ τοῦ Πατρὸς καὶ ἐκ τοῦ Υἱοῦ ἐκπορευομένου τοῦ Πνεύματος, ἢ ἀτελὴς ἡ ἐκπόρευσις ἑκατέρα ἢ περιττὴ ἡ ἑτέρα · (24) Cod. ἀδύνατον · (26) Vat. εἰ · (28) Vat. τελείων · (29) marginal, Gk: Ἕτερον ἄτοπον, ὅτι εἰ ἐκ τοῦ Υἱοῦ τὸ Πνεῦμα υἱὸς ἔσται τοῦ Υἱοῦ. Letter notes (Latin column): (f) Elias of Crete on Greg. Naz. or. 11 + Arist. Metaph. IV · (k) S. Maximus, Opusc. theol. et polem. II p. 155 ed. Combefis; ps.-Justin Expos. rectae fidei n. 3 · (m)/(4) on προβάλλεσθαι being middle, not passive · (n)/(o) the two marginal aliud absurdum notes, (o) adding the cross-reference to Photius, De Spiritus sancti mystagogia c. 7 and Nicholas of Methone c. 11 for the Greek's fourth argument · (p) Haec sane non satis accurate dicta sunt + the attack on mediate / immediate · (q), (r) (imitatur Gregorii Nazianzeni responsum or. 27) · (s) Quintum argumentum. Calfa strips all note anchors from the Greek, so the English carries no [n:] markers; the identifications survive in the committed twin and are listed here for the eventual fontes/scripture harvest.

Second half of the work (chunks 0006–0009 of 10). Chunks 0001–0005 were held by a peer agent who could not see this file; the merge is a required later step, not done here. Conventions below are the anchor's (cruces-0000.md) unless flagged as decided fresh.

Method note — what "plate-verified" means in this batch

The Latin twin was cropped from raw/scans/pg139/patrologiaecurs63migngoog_djvu.xml by x-range split. The same leaves carry Migne's GREEK column in that XML, and its OCR is independent of Calfa's. Every reading marked plate-verified below was checked against the Greek column of the named leaf (leaves 98–108 = cols 0201–0221), so the evidence is a second witness to the plate, not a second reading of our own file. Where the scan OCR is itself illegible at the point at issue, that is said.

Conventions carried from the anchor (unchanged)

Speaker sigla ΓΡ./ΛΑΤ. → GR./LAT., one paragraph per turn · section letter-numerals as arabic + period · οὐσία = essence, φύσις = nature, ὑπόστασις = hypostasis, πρόσωπον = person, ἐκπόρευσις = procession · ὑπερ- compounds take "super-" · mentioned particles stay in Greek script, quoted credal clauses translated in " ".

Conventions decided fresh (flag for the merge)

These terms first fall due in this half of the work; the peer's half may have settled some of them differently.

Latin-twin pass (pilot §6) — result by chunk

Performed after drafting, against src/pg-latin/nicetas-maroneia-dialogi/000[6-9].md.

No chunk in this batch was clean of Calfa OCR damage. The twin pass found no "Greek made no sense and the Latin explains why" case that the plate's own Greek column did not settle more directly.

Pattern 16 — [lat: …] fired: ONE

Pattern 16 — candidates LOGGED, marker NOT fired

Pattern 14 — [var: …] fired: TWO (the divergence list)

Both are quotations Migne prints in a form the received text does not carry. These are the most valuable findings in the batch, and the search for them was word-by-word against the plate, not against memory.

Pattern 13 — [ed: …], text the DIGITIZATION lost that the plate carries

This work is Hergenröther's set of fragments of dialogues II–VI, and its structural furniture is exactly what Calfa strips. Five centered division headings and two editorial bridges are gone from our Greek. All are restored as [ed: …], never as ## heads (marker parity with the Greek chunk), following the anchor's ΛΟΓΟΣ ΠΡΩΤΟΣ precedent.

Fragment boundaries as found: dialogue II opens at col. 0201/0202 and the work runs to a stop at col. 0221 in mid-argument, where PG 139 turns to the next author (Joannes of Citrus). Ellipses (…) inside the Greek are Hergenröther's own and are reproduced as printed; no gap anywhere was bridged with invented connective prose.

Ladder row 1 — Calfa OCR damage, corrected silently in the English

Itemized inline these would drown the text (register rule). Grouped by column. Unless marked otherwise, the correction is confirmed by the plate's Greek column on the leaf named in the section head above.

Polarity- and sense-bearing (these are the ones that mattered):

Speaker sigla mangled by Calfa (all restored, all plate-verified): 0201 ΠΡ. → ΓΡ.; 0201 ΑΑΤ. → ΛΑΤ.; 0201 ΤΡ. → ΓΡ.; 0204 ΛΛΤ. → ΛΑΤ.; 0213 λΑΤ → ΛΑΤ.; 0216 ΕΡ. → ΓΡ.; 0216 ΛΑΤ.θ → ΛΑΤ.; 0212 ΑΑΤ. → ΛΑΤ. 0213: Calfa drops the ΓΡ. tag entirely before Πάραγε δή μοι καὶ ταῦτα, silently handing the Greek's speech to the Latin; the plate (leaf 104) and the twin (GR. Profer jam mihi) both carry it. Restored.

Section marks: 0205 νʹ → γʹ ("3.") — plate leaf 99 reads γ'., twin reads 3.; Calfa's ν is an OCR of γ, and left standing it would have printed "50." in the English. 0201 αʹ restored (lost with the heading, above).

The intrusive-Κ class (already named in the anchor's cruces for chunk 0000, and still running at this end of the volume): 0209 Κθεοί → θεοί · 0209 Κγὰρ → γάρ · 0213 Κἵκα → ἵνα · 0213 Κἐμφαῖνον → ἐμφαῖνον · 0217 Κῥνωθεν → ἄνωθεν · 0217 Κἐτράνωσεν → ἐτράνωσεν · 0220 Κσυμβόλῳ → συμβόλῳ.

Dropped opening parenthesis class (same defect as the anchor's 0172 ἀδύνατονὐδὲ): 0209 ἁπασῶνπὲρ φύσιν → ἁπασῶν (ὑπὲρ φύσιν · 0209 ἁπασῶνκὲρ τάξιν → ἁπασῶν (ὑπὲρ τάξιν · 0209 διαίρεσιν παρεισάγουσαα γὰρ ἡ φύσις → διαίρεσιν παρεισάγουσα (μία γὰρ ἡ φύσις · 0209 θέσινπερίγραπτος → θέσιν (ἀπερίγραπτος (see above) · 0209 τὸ ἅμαν οἷς γὰρ → τὸ ἅμα (ἐν οἷς γὰρ.

Routine letter-level damage, bundled (correction ← Calfa): 0201 ἀναλαβόντες kept (Calfa's reading; the scan's Greek OCR gives only λαβόντες, but the twin's disputationem resumentes wants the ἀνα-) · 0201 ἑτοπα → ἄτοπα · 0201 ἐκτοῦ → ἐκ τοῦ · 0201 ατρὸς → Πατρὸς · 0201 ἐχόντοιν kept (genitive dual, both witnesses — not an error) · 0201 Πατοὸς → Πατρὸς · 0204 θΗατήρ → Ὁ Πατήρ · 0204 εἴῃνε → εἴληχε · 0204 φατὲ/φατέ as printed · 0205 ὁ ὸς → ὁ Υἱὸς · 0205 Ἀ δὲ Πατὴρ → ὁ δὲ Πατὴρ · 0205 Ιει → ἕξει · 0205 Μακεἐόνιος → Μακεδόνιος · 0205 ποοὸν → ποσόν · 0208 ποότερον → πρότερον · 0208 λέ, εται → λέγεται · 0208 τὰ κῶ γράμματα → τὰ κδʹ γράμματα (twenty-four; plate κδ', twin Viginti quatuor — a letter-numeral, verified on both sides before rendering) · 0209 Δδάμ → Ἀδάμ · 0209 Ἀιχαὴλ → Μιχαὴλ · 0209 ἀν οὐρανοῖς → ἐν οὐρανοῖς · 0209 θεαρχκῆς → θεαρχικῆς · 0209 όῦτως → οὕτως · 0209 Τίοῦ → Υἱοῦ · 0212 τίθεαμεν → τίθεμεν · 0213 Ἰ δὲ → Ἡ δὲ · 0213 Παῆρὸς → Πατρὸς · 0213 δονματιζόμενον → δογματιζόμενον · 0213 Ιατέρα → Πατέρα · 0213 ἀυδρῶς → ἀμυδρῶς · 0213 Γοαφῆς → Γραφῆς · 0213 ἐξ αὐτῆς τῆς Ὃν πραγμάτων → ἐξ αὐτῆς τῶν πραγμάτων καὶ νοημάτων φύσεως · 0213 συσκεόώνεθα → συσκεψώμεθα · 0213 Αέγω → Λέγω · 0213 καυσις → καῦσις · 0213 βύλει → βούλει · 0213 ἠ ἐνέργεια → ἡ ἐνέργεια · 0213 Γρας ῆς → Γραφῆς · 0216 ἕτερος. ὧ Λατήα τοῦ δόγματος → ἕτερος, ὦ Λατῖνε, τοῦ δόγματος · 0216 ἀσθενής ἐστε → ἐστι · 0216 παραοειγμάτων → παραδειγμάτων · 0217 Τκῦτα → Ταῦτα · 0217 τοὐτὰ → ταὐτὰ · 0217 πρὸ τούτν → πρὸ τούτων · 0217 σύνοδ ν → σύνοδον · 0217 τὸ δὲ ἄγιος σύμβολον → τὸ δὲ ἅγιον σύμβολον · 0220 προΐντο → προΐεντο · 0220 συνεγράψαντότε → συνεγράψαντό τε · 0221 βίόλοι → βίβλοι · 0221 ιοοονοῦντες → φρονοῦντες · 0221 Γνεῦμα → Πνεῦμα · 0221 οὕτιο → οὕτω · 0221 κλέος πὸ → κλέος τὸ · 0221 λένετε → λέγετε · 0221 ἐκπορευεσίλαι → ἐκπορεύεσθαι · 0221 συγγοαφῶν → συγγραφῶν · 0221 ικπόρευσιν → ἐκπόρευσιν · 0221 Καμασκηνός → Δαμασκηνός · 0221 δίᾳ → ἰδίᾳ · 0221 τὸ κ τοῦ → τὸ ἐκ τοῦ · 0221 τὸ νεῦμα → τὸ Πνεῦμα · 0221 νοῦμεν → νοοῦμεν · 0221 Πιτρός → Πατρός · 0221 προσθρκν → προσθήκην · 0221 τὸ ἐκ τοῦ ἰατρὸς καὶ ἐκ τοῦ λἱρῶ ἐν σορευόμενον → τὸ ἐκ τοῦ Πατρὸς καὶ ἐκ τοῦ Υἱοῦ ἐκπορευόμενον (plate leaf 108; twin Qui ex Patre Filioque procedit) · 0221 Πἱῷ → Υἱῷ.

One reading kept against the plate: 0221 δρῶντας (Calfa) where the scan's Greek column OCRs δρῶντος. The twin construes plural (Patrum vestigia sequentes … in symbolo … pronuntiare) and the whole address is to the Latins in the plural; the scan's singular is very likely its own OCR slip on a final letter. Rendered plural, logged so the choice is visible.

Negation audit (7a) — the doctrinal core

Every οὐ/οὐκ/οὐχ/μή/μηδέ/οὔτε/μηδὲ ὅλως in these four chunks was checked one by one against Calfa, the plate's Greek column and the twin before drafting, and again after. The dense sites, all carried through unaltered:

Pronoun-itacism audit (ἡμεῖς/ὑμεῖς). Every first/second-person plural in these chunks was checked across all three witnesses before drafting, because Calfa confuses the pair and this dialogue's whole rhetoric turns on who is being accused. All agree: 0204 κατὰ τὸν ὑμέτερον λόγον … οὐχ ἡμῖν μᾶλλον … ἀλλ' ἑαυτοῖς συνηγάγετε (twin vestram … non magis nobis, quam potius vobis ipsis); 0212 τῆς μεθ' ὑμῶν … συζητήσεως; 0217 ἡμῖν τε ἐναντιόφρονες … ὑμεῖς; 0221 καὶ ἡμῖν καὶ ὑμῖν ἔχει τὸ ἀναμφίβολον (twin et apud nos et apud vos), ἀρκέσει δὲ ὑμῖν καὶ ἡμῖν. No [lat:] candidate in this class.

Not repaired to make sense: nothing. The one place where the printed Greek does not construe on its own is 0221, Ἐν γὰρ τῇ ἀποδόσει τῆς πίστεως τὸ ἐκ τοῦ Πατρός, οὐκ ἐκ τοῦ Υἱοῦ ἐκπορευόμενον, ἰδίᾳ δὲ ἐπισημαίνεται — the first limb has no finite verb (the twin supplies exprimit). Rendered with a bracketed supplied "[it is]" rather than borrowing Allatius's verb into the English; the crux is here.

Ellipses, anchors, ratio

The Greek's own … marks (Hergenröther's elisions) are reproduced 1:1; column anchors are 1:1 with the Greek chunks ([0201][0204][0205] · [0208][0209][0212] · [0213][0216][0217][0220] · [0221]). EN:GK ratio 1.39 / 1.36 / 1.41 / 1.36 — at the dialogue anchor's 1.37×.

For the merge

  1. Diff the fresh-decided vocabulary above against chunks 0001–0005, especially ἀρχή/αἴτιον, προβολεύς/προβάλλεσθαι/προΐεσθαι, and the ὑπερ- coinages.
  2. Do not harmonize the two forms of the creed's glorification clause (anchor 0000 doubled, 0221 single) — see the [var:] note; that divergence is Migne's.
  3. The anchor's [ed:] for ΛΟΓΟΣ ΠΡΩΤΟΣ and the five in this half should end up phrased alike; they were written to the same template but not against each other.