Errata
What was asserted where the print left it open, and how to correct it.
The corrections log
What this page is. A standing record of the three places an error is most likely: a reading the library supplied where no witness settled the place; a reading that rests on ONE instrument and no second reader corroborates it — merged, so it can be seen and checked, and flagged for human review; and a page image a reading rested on that is not held with the edition. It is a worklist in the second sense only: the first and third lists record what the library asserted, not what someone is being asked to do. Every rule served here has been read and classified — OCR, punctuation, transliteration or conjectural. The REVIEW FLAG is a second thing, and this page does not conflate it with the second list: it is true for every reading that rests on one instrument, and also for every rule whose TYPE fell to the taxonomy's conservative default (model §4.1 rule 6) — 726 in On the Theology of Plato. Each rule's flag is carried in its apparatus.json; the list below is the footing, not the flag. An edition whose repair has not begun records no rules yet, and its own block below says so rather than reading as a clean bill of health.
A correction is a new rule. Where a place is wrong the library does not edit the served text in silence: the report becomes a NEW RULE in a NEW VERSION, the old version stands at its own address, and the change is citable (see the versioning promise). The rules are not reworded to fit the software, and nothing is dropped to keep a count tidy.
Readings supplied without a witness
The Elements of Theology — Proclus, tr. Thomas Taylor
None. Every rule in this version is typed OCR, punctuation or transliteration — no reading here was supplied without a witness.
On the Theology of Plato — Proclus, tr. Thomas Taylor
None. Every rule in this version is typed OCR, punctuation or transliteration — no reading here was supplied without a witness.
On the Cave of the Nymphs — Porphyry, tr. Thomas Taylor
porphyry-on-the-cave-of-the-nymphs-taylor-1917:r0109· conjectural · reading
“fountains and river;” → “fountains and rivers;”
The plural 's' is missing from 'rivers' — a wrong word with no damage character, invisible to the census (class 2). (The space before the semicolon is normalized away before the rules run.)
no witness · no page image is held for this reading
Readings that rest on one instrument
The Elements of Theology — Proclus, tr. Thomas Taylor
None. Every reading in this version is corroborated by a second reader, or rests on a witness that is not a single instrument.
On the Theology of Plato — Proclus, tr. Thomas Taylor
146 readings in this version rest on ONE instrument. Each rationale names the reader and states that no second reader corroborates it.
proclus-theology-of-plato-taylor-1816:r11689· transliteration
“T *( nUraf, nanii 0w, to»; pty anayjrtnc ay/jv juaXa, ro*< h xoretfcso/4f|0«. Dissert. I. Edit. PrUlC” → “τ' ἔθνη πρῶτον, πάντα Θεόν, τὸν ἐπὶ πάντων ἀγαθὸν μάλα, τοῖς τὰ καὶ σύμπαντα. Dissert. I. Edit. Prasc”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "τ' ἔθνη πρῶτον, πάντα Θεόν, τὸν ἐπὶ πάντων ἀγαθὸν μάλα, τοῖς τὰ καὶ σύμπαντα. Dissert. I. Edit. Prasc" where the transcription has "T *( nUraf, nanii 0w, to»; pty anayjrtnc ay/jv juaXa, ro*< h xoretfcso/4f|0«. Dissert. I. Edit. PrUlC" (page image, archive n7). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11690· transliteration
“Ta*f zfwrai; ovtnai;, tu fxi rn m* ovwnotTt avrwi ttyvfxtfai, xai oio11* WjdufMif avmf inri* #«(wu” → “πας ζωηαις, ουπηαις, τη οὐκ μη ην οὐωνοιαπΐ᾽ αὐτῷ ττγυπιτιιαι, και οιοι11* ὡρδωαοις αὐτηί ἱπηι* #εων”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "πας ζωηαις, ουπηαις, τη οὐκ μη ην οὐωνοιαπΐ᾽ αὐτῷ ττγυπιτιιαι, και οιοι11* ὡρδωαοις αὐτηί ἱπηι* #εων" where the transcription has "Ta*f zfwrai; ovtnai;, tu fxi rn m* ovwnotTt avrwi ttyvfxtfai, xai oio11* WjdufMif avmf inri* #«(wu" (page image, archive n22). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. 1.0.9 LEAF CORRECTION (2026-10-08). THE EVIDENCE LEAF OF THIS RULE IS MOVED, from n22 to n21: the passage stands on archive leaf n21 (front matter) and not on n22. MEASURED over every transliteration rule that cites an evidence leaf (189 of them) by reading each cited plate BLIND — 123 page transcriptions, the reader shown no candidate text — and locating each page’s own first printed line in our source.txt: this rule’s find stands before n22’s own first body line (source line 1019) and is therefore on the page before it, whose first body line is 965. THE READING IS NOT TOUCHED — only the leaf the evidence is drawn from, so that a corroboration pass reads the page the passage is actually printed on.proclus-theology-of-plato-taylor-1816:r11691· transliteration
“wonesf and thd transcendency of Cherubim and Seraphim,” xott yog eyto ov xad’ o, n BiBe/xcu, xat” → “æones, 4 and the transcendency of Cherubim and Seraphim,” και γαρ εγω ου καθ' ο, τι δεδεμαι, και”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "æones, 4 and the transcendency of Cherubim and Seraphim,” και γαρ εγω ου καθ' ο, τι δεδεμαι, και" where the transcription has "wonesf and thd transcendency of Cherubim and Seraphim,” xott yog eyto ov xad’ o, n BiBe/xcu, xat" (page image, archive n18). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11692· transliteration
“ianoia, avotot, from whence dianoelic, is the discursive energy of reason; (8ft£oSjxi) too Xoyoo” → “IANOIA, διάνοια, from whence dianoetic, is the discursive energy of reason; (διεξοδικὴ τοῦ λόγου”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "IANOIA, διάνοια, from whence dianoetic, is the discursive energy of reason; (διεξοδικὴ τοῦ λόγου" where the transcription has "ianoia, avotot, from whence dianoelic, is the discursive energy of reason; (8ft£oSjxi) too Xoyoo" (page image, archive n74). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. WITNESS 1.0.4: Project Gutenberg’s human-proofread transcription of the 1816 print reads ["dianoia","from","whence","dianoetic","is","the","discursive","energy","of","reason"] at this point (vol1, slid). the witness confirms the reading, but the placement is SLID and the confirmation rests on the CHANGED fragment alone (1–2 common words) — the whole reading is what anchors a slid placement, so the flag standsproclus-theology-of-plato-taylor-1816:r11694· transliteration
“ai trtr*p£»£»pti/*a* yxyoiviaf ivyu^tw; } w; ttxuoi oiqriov itrct, xeu xar» T*T i DC Ctt” → “και τρτρ*ρξομενις'αί γεγοναται ἵυχαρωω;, ὡς ττκύοι οιητοιν εἴπηι, και κατα ταυτι DC Cœ”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "και τρτρ*ρξομενις'αί γεγοναται ἵυχαρωω;, ὡς ττκύοι οιητοιν εἴπηι, και κατα ταυτι DC Cœ" where the transcription has "ai trtr*p£»£»pti/*a* yxyoiviaf ivyu^tw; } w; ttxuoi oiqriov itrct, xeu xar» T*T i DC Ctt" (page image, archive n22). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. 1.0.9 LEAF CORRECTION (2026-10-08). THE EVIDENCE LEAF OF THIS RULE IS MOVED, from n22 to n21: the passage stands on archive leaf n21 (front matter) and not on n22. MEASURED over every transliteration rule that cites an evidence leaf (189 of them) by reading each cited plate BLIND — 123 page transcriptions, the reader shown no candidate text — and locating each page’s own first printed line in our source.txt: this rule’s find stands before n22’s own first body line (source line 1019) and is therefore on the page before it, whose first body line is 965. THE READING IS NOT TOUCHED — only the leaf the evidence is drawn from, so that a corroboration pass reads the page the passage is actually printed on.proclus-theology-of-plato-taylor-1816:r11695· transliteration
“xcn t^s ** ijxfj $c$i)A^Kfv, cijAi 8* ryo> /tAtcnj rc ftwit citouta fwnj r’ aftigan rcov” → “καὶ τὴς ἀπειρία* ἤτοι φάσεως ἢ ἀποφάσεως, οὐκ ἔστιν εἰ μὴ ὅταν μῖξιν ἔλαβε τὴγ ἀπὸ ἀμφοῖν τὼγ”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "καὶ τὴς ἀπειρία* ἤτοι φάσεως ἢ ἀποφάσεως, οὐκ ἔστιν εἰ μὴ ὅταν μῖξιν ἔλαβε τὴγ ἀπὸ ἀμφοῖν τὼγ" where the transcription has "xcn t^s ** ijxfj $c$i)A^Kfv, cijAi 8* ryo> /tAtcnj rc ftwit citouta fwnj r’ aftigan rcov" (page image, printed page 185 = archive v2-n690). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11696· transliteration
“K*4 xf* 1 TOf tvt ra( aya^airoyra fnruy, i< ivrarot tirat n xpurrry mg vmnBumg **'” → “XLIV. Tοῦτ' ἔστι τὰς ἀγαθότητας ἁπτέσθαι, ἡ ἀνωτάτω εἴη τὸ πρωτότυπον τῆς ὑποθέσεως”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "XLIV. Tοῦτ' ἔστι τὰς ἀγαθότητας ἁπτέσθαι, ἡ ἀνωτάτω εἴη τὸ πρωτότυπον τῆς ὑποθέσεως" where the transcription has "K*4 xf* 1 TOf tvt ra( aya^airoyra fnruy, i< ivrarot tirat n xpurrry mg vmnBumg **'" (page image, archive n28). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11697· transliteration
“row TfpaTo^ and vpo rr}$ a*upiu$ } it is necessary to read 7rpo$ tow Tcfoer©^ anck” → “τοῦ Τριττοῦ and προ τῆς ἀκραιφνοῦς, it is necessary to read πρὸς τὸν Τρίτον and”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "τοῦ Τριττοῦ and προ τῆς ἀκραιφνοῦς, it is necessary to read πρὸς τὸν Τρίτον and" where the transcription has "row TfpaTo^ and vpo rr}$ a*upiu$ } it is necessary to read 7rpo$ tow Tcfoer©^ anck" (page image, printed page 191 = archive v1-n266). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11699· transliteration
“u>$ wri SijXov ex 6eoyovia$. For aXXa o fogxu$, therefore, I read aXXa o Jlxeavp$\” → “ὡς τῇ δῆλον ἐκ θεογονίας. For ἀλλὰ ὁ Φιλκυς, therefore, I read ἀλλὰ ὁ Ὠκεανός·”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ὡς τῇ δῆλον ἐκ θεογονίας. For ἀλλὰ ὁ Φιλκυς, therefore, I read ἀλλὰ ὁ Ὠκεανός·" where the transcription has "u>$ wri SijXov ex 6eoyovia$. For aXXa o fogxu$, therefore, I read aXXa o Jlxeavp$\\" (page image, printed page 205 = archive v2-n710). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11700· transliteration
“owroic as in the original, which is nonsense, it is necessary tq read, as in tfi” → “τέλος έχειν as in the original, which is nonsense, it is necessary to read, as in th”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "τέλος έχειν as in the original, which is nonsense, it is necessary to read, as in th" where the transcription has "owroic as in the original, which is nonsense, it is necessary tq read, as in tfi" (page image, archive n33). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11701· transliteration
“mt* iSouXijf, I read>rr«/3oX>if. For die mundane Gods are in no part of the Tima” → “βουλῆς, I read ὑπερβολῆς. For the mundane Gods are in no part of the Timæ”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "βουλῆς, I read ὑπερβολῆς. For the mundane Gods are in no part of the Timæ" where the transcription has "mt* iSouXijf, I read>rr«/3oX>if. For die mundane Gods are in no part of the Tima" (page image, printed page 124 = archive v2-n629). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11702· transliteration
“yog xx xgwxtxxx ran vQirrctfumv, t>jv thxv u fiogfyv, it is necessary to read I” → “γὰρ καὶ ἀχράντων, τῶν ὑφιστάμενον, τὴν ἰδίαν ἔφη μορφὴν, it is necessary to read”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "γὰρ καὶ ἀχράντων, τῶν ὑφιστάμενον, τὴν ἰδίαν ἔφη μορφὴν, it is necessary to read" where the transcription has "yog xx xgwxtxxx ran vQirrctfumv, t>jv thxv u fiogfyv, it is necessary to read I" (page image, printed page 295 = archive v1-n370). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11703· transliteration
“xanx^ou yxg xx xgmtrrx xw» vfitfinm, xm xgovQtrrafiwm, t»jv iSixy t%tt fugf^y” → “καταχοῦ γὰρ καὶ ἀχράντων, τῶν ὑφισταμένων, τὴν ἰδίαν ἔχει μορφὴν”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "καταχοῦ γὰρ καὶ ἀχράντων, τῶν ὑφισταμένων, τὴν ἰδίαν ἔχει μορφὴν" where the transcription has "xanx^ou yxg xx xgmtrrx xw» vfitfinm, xm xgovQtrrafiwm, t»jv iSixy t%tt fugf^y" (page image, printed page 295 = archive v1-n370). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11704· transliteration
“<and one, through unconfosed union with^Ae one itself." Hr yag av tommy Xotr” → “and one, through unconfused union with the one itself." Ἧν γὰρ ἂν τουτων λα-”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "and one, through unconfused union with the one itself.\" Ἧν γὰρ ἂν τουτων λα-" where the transcription has "<and one, through unconfosed union with^Ae one itself.\" Hr yag av tommy Xotr" (page image, archive n21). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11706· transliteration
“&av(jL*vra, it is requisite to read Oav/iavro$; and for wovrov, tvovtco” → “θαυμάζοντα, it is requisite to read θαυμαστὸν; and for ἐνομομον, ἔνονον”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "θαυμάζοντα, it is requisite to read θαυμαστὸν; and for ἐνομομον, ἔνονον" where the transcription has "&av(jL*vra, it is requisite to read Oav/iavro$; and for wovrov, tvovtco" (page image, printed page 205 = archive v2-n710). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11708· transliteration
“rm ptra in the original immediately before rc vo>jtov 0«ov yno(” → “πῶς ἔχω in the original immediately before τὸ νόμιμον Θεοῦ γήπω”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "πῶς ἔχω in the original immediately before τὸ νόμιμον Θεοῦ γήπω" where the transcription has "rm ptra in the original immediately before rc vo>jtov 0«ov yno(" (page image, printed page 182 = archive v1-n257). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11709· transliteration
“ j for he says that the divine multitude avtxfotrryrof too wa$” → “; for he says that the divine multitude ἀνεκφοίτητος τοῦ ἑνός”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "; for he says that the divine multitude ἀνεκφοίτητος τοῦ ἑνός" where the transcription has " j for he says that the divine multitude avtxfotrryrof too wa$" (page image, archive n20). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. 1.0.8 ADJUDICATION (2026-10-08). THE READING IS UPHELD on a blind read of the page it stands on: archive leaf n20 (this rule’s own evidence leaf, MEASURED correct) reads “the divine multitude ἀνεκφοίτητος του ἑνος, i. e. does not depart from, but abides in the one” — the rule’s letters. THE EVIDENCE IS MOVED BY THE RE-FLAG: a witness entry now records the blind read that put the rule back on the worklist. THE REVIEW FLAG IS RESTORED. It was cleared at 1.0.4 on a witness verdict that cannot see Greek: the witness locator’s token stream is ASCII letters and digits ONLY, so for a transliteration rule its “the witness carries the rule’s whole `after` reading” test is decided on the English AROUND the Greek and the Greek the rule exists to supply is never compared.proclus-theology-of-plato-taylor-1816:r11710· transliteration
“tijv h £fgoo/9i/x tp.$aweiv, rXrfiot yvcoctoos, ij ymn eoQutf*” → “τήν δὲ χερουβὶμ ἐμφαίνειν, πλῆθoς γνώσεως, ἢ χύσιν σοφίας.”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "τήν δὲ χερουβὶμ ἐμφαίνειν, πλῆθoς γνώσεως, ἢ χύσιν σοφίας." where the transcription has "tijv h £fgoo/9i/x tp.$aweiv, rXrfiot yvcoctoos, ij ymn eoQutf*" (page image, archive n21). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11711· transliteration
“QfifLcm itw, ro pi) cx $<%jvo/x*va>y tci /Skswofitpct yeyowai” → “ῥήματι θεοῦ, εἰς τὸ μὴ ἐκ φαινομένων τὰ βλεπόμενα γεγονέναι”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ῥήματι θεοῦ, εἰς τὸ μὴ ἐκ φαινομένων τὰ βλεπόμενα γεγονέναι" where the transcription has "QfifLcm itw, ro pi) cx $<%jvo/x*va>y tci /Skswofitpct yeyowai" (page image, archive n17). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11712· transliteration
“vxtgougcmon in this place, it is JteceMtfy tD reed vwovf*vkw” → “ἀγαθοεργόμον in this place, it is necessary to read ἀγανόφυκον”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἀγαθοεργόμον in this place, it is necessary to read ἀγανόφυκον" where the transcription has "vxtgougcmon in this place, it is JteceMtfy tD reed vwovf*vkw" (page image, printed page 329 = archive v1-n404). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11715· transliteration
“nfo yog rot uMa irgocXiow* ns <p xeu rip r«ftv «auri \$ 9” → “Νοητογ ἐστὶ τὸ μὲν πατρικὸν ἀφιέναι τὴγ πρόσχυσιν· ὃς ἔχει τήν τε τῶν ἑαυτοῦ ὄντων.”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "Νοητογ ἐστὶ τὸ μὲν πατρικὸν ἀφιέναι τὴγ πρόσχυσιν· ὃς ἔχει τήν τε τῶν ἑαυτοῦ ὄντων." where the transcription has "nfo yog rot uMa irgocXiow* ns <p xeu rip r«ftv «auri \\$ 9" (page image, printed page 185 = archive v2-n690). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11716· transliteration
“' For voijt»v, it is obviously necessary to read voepoov” → “For νοήτων, it is obviously necessary to read νοερῶν”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "For νοήτων, it is obviously necessary to read νοερῶν" where the transcription has "' For voijt»v, it is obviously necessary to read voepoov" (page image, printed page 415 = archive v1-n490). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11717· transliteration
“For Tot wagayoyra, it is necessary to read rot nctpovrou” → “4 For τοῖς παραγωγοῖς, it is necessary to read τοῖς παρόντοις”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "4 For τοῖς παραγωγοῖς, it is necessary to read τοῖς παρόντοις" where the transcription has "For Tot wagayoyra, it is necessary to read rot nctpovrou" (page image, printed page 82 = archive v2-n587). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11718· transliteration
“vtik* is erroneously printed in the original for fxamxot” → “ytix* is erroneously printed in the original for ἐξαίρετοί”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ytix* is erroneously printed in the original for ἐξαίρετοί" where the transcription has "vtik* is erroneously printed in the original for fxamxot" (page image, printed page 174 = archive v2-n679). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11719· transliteration
“xar f£o%Y)v by way of eminence, the Platonic Successor ” → “κατ᾽ ἐξοχήν by way of eminence, the Platonic Successor”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "κατ᾽ ἐξοχήν by way of eminence, the Platonic Successor" where the transcription has "xar f£o%Y)v by way of eminence, the Platonic Successor " (page image, archive n42). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. 1.0.8 ADJUDICATION (2026-10-08). THE READING IS UPHELD. The rule’s letters are the print’s: a blind read of archive leaf n42 (this rule’s own evidence leaf, MEASURED correct) reads “κατ’ ἐξοχὴν by way of eminence, the Platonic Successor”, and “κατ’ ἐξοχήν” is the phrase the English gloss “by way of eminence” translates. The final accent is CONTESTED (the blind read gives ἐξοχὴν, the pass’s own draws give both) and the rule keeps the standard form ἐξοχήν. THE EVIDENCE IS MOVED BY THE RE-FLAG: a witness entry now records the blind read that put the rule back on the worklist. THE REVIEW FLAG IS RESTORED. It was cleared at 1.0.4 on a witness verdict that cannot see Greek: the witness locator’s token stream is ASCII letters and digits ONLY, so for a transliteration rule its “the witness carries the rule’s whole `after` reading” test is decided on the English AROUND the Greek and the Greek the rule exists to supply is never compared.proclus-theology-of-plato-taylor-1816:r11720· transliteration
“tov xctr^ov, it is necessary to read u* tq tw xcxr/*«■/” → “τοῦ κατώτῳ, it is necessary to read ἐκ τῷ κατώτῳ”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "τοῦ κατώτῳ, it is necessary to read ἐκ τῷ κατώτῳ" where the transcription has "tov xctr^ov, it is necessary to read u* tq tw xcxr/*«■/" (page image, printed page 65 = archive v2-n570). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11721· transliteration
“irapxrxevetZovra it is requisite to read weptrxia%ovr*” → “ἀπαρχομένον it is requisite to read περιέχοντα”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἀπαρχομένον it is requisite to read περιέχοντα" where the transcription has "irapxrxevetZovra it is requisite to read weptrxia%ovr*" (page image, printed page 61 = archive v1-n136). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. 1.0.9 LEAF CORRECTION (2026-10-08). THE EVIDENCE LEAF OF THIS RULE IS MOVED, from v1-n136 to v1-n135: the passage stands on archive leaf v1-n135 (printed page 60) and not on v1-n136. MEASURED over every transliteration rule that cites an evidence leaf (189 of them) by reading each cited plate BLIND — 123 page transcriptions, the reader shown no candidate text — and locating each page’s own first printed line in our source.txt: this rule’s find stands before v1-n136’s own first body line (source line 7145) and is therefore on the page before it, whose first body line is 583. THE READING IS NOT TOUCHED — only the leaf the evidence is drawn from, so that a corroboration pass reads the page the passage is actually printed on.proclus-theology-of-plato-taylor-1816:r11722· transliteration
“xai yag ovgaviav xai njv (TfArjvTjv OgQsvs Tgwqyogtwiy” → “καὶ γὰρ οὐρανίαν καὶ τὴν Σελήνην Ὀρφεύς τρωγῴγοιῶν”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "καὶ γὰρ οὐρανίαν καὶ τὴν Σελήνην Ὀρφεύς τρωγῴγοιῶν" where the transcription has "xai yag ovgaviav xai njv (TfArjvTjv OgQsvs Tgwqyogtwiy" (page image, printed page 195 = archive v2-n700). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11723· transliteration
“y*l<ny, «nt0T0 g ayrmrrov, rpig rovro **if *fxi {*ng,” → “γνῶσιν, καὶ ἀπερίγραπτος πανταχοῦ, τρὶς τοῦτο δεῖ ἔκφημαι ξένης.”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "γνῶσιν, καὶ ἀπερίγραπτος πανταχοῦ, τρὶς τοῦτο δεῖ ἔκφημαι ξένης." where the transcription has "y*l<ny, «nt0T0 g ayrmrrov, rpig rovro **if *fxi {*ng," (page image, archive n28). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11724· transliteration
“Kvfut!ti gaivefricv itotytgw /3otgur)%tTog aXfirjg.” → “οἱ θεουργοὶ χεῖρας ὀνομάζουσιν”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "οἱ θεουργοὶ χεῖρας ὀνομάζουσιν" where the transcription has "Kvfut!ti gaivefricv itotytgw /3otgur)%tTog aXfirjg." (page image, printed page 155 = archive v2-n660). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11725· transliteration
“x For xoei t w Bvatiog, read, xai onto nj$ tvahog*” → “For αἰὼν ἐνεργείᾳ, read, ἀεὶ τῶν ἐνεργείᾳ.”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "For αἰὼν ἐνεργείᾳ, read, ἀεὶ τῶν ἐνεργείᾳ." where the transcription has "x For xoei t w Bvatiog, read, xai onto nj$ tvahog*" (page image, printed page 45 = archive v1-n120). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11726· transliteration
“avofMtor^Tot, it is necessary to read o/Aoionjra.” → “ἀρχὴ δευτέρα, it is necessary to read ἀρχαιοτέρα,”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἀρχὴ δευτέρα, it is necessary to read ἀρχαιοτέρα," where the transcription has "avofMtor^Tot, it is necessary to read o/Aoionjra." (page image, printed page 39 = archive v1-n114). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11727· transliteration
“*v(tXuovT»f, it is necessary to read avaxoXo emj.” → “ἐν δὲ τοῖς νοητοῖς καὶ νοεροῖς, it is necessary to supply καθ᾽ ὑπαρξιν, ἐν δὲ τοὺς”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἐν δὲ τοῖς νοητοῖς καὶ νοεροῖς, it is necessary to supply καθ᾽ ὑπαρξιν, ἐν δὲ τοὺς" where the transcription has "*v(tXuovT»f, it is necessary to read avaxoXo emj." (page image, printed page 418 = archive v1-n493). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11728· transliteration
“xttTct twro> it is necessary to read xara t««tok*” → “διακρινον, it is necessary to read διατεινον.”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "διακρινον, it is necessary to read διατεινον." where the transcription has "xttTct twro> it is necessary to read xara t««tok*" (page image, printed page 10 = archive v2-n515). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11729· transliteration
“For T*A<Mrar«, 1) is necessary to read T»Xn#njr«” → “Instead of διὰ τήνων ἐγκοσμίων, it is necessary to read διὰ τι τῶν ἐγκόσμιον”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "Instead of διὰ τήνων ἐγκοσμίων, it is necessary to read διὰ τι τῶν ἐγκόσμιον" where the transcription has "For T*A<Mrar«, 1) is necessary to read T»Xn#njr«" (page image, printed page 127 = archive v2-n632). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11730· transliteration
“eat ro to ov it is necessary to read auno to ov” → “ἔσται τὸ ὄν it is necessary to read αὐτὸ τὸ ὄν”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἔσται τὸ ὄν it is necessary to read αὐτὸ τὸ ὄν" where the transcription has "eat ro to ov it is necessary to read auno to ov" (page image, printed page 115 = archive v1-n190). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11731· transliteration
“auroettierw it is requisite to read avyo fiSwiv” → “αὐτοπέρσην it is requisite to read αὐτοῦ φάσιν”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "αὐτοπέρσην it is requisite to read αὐτοῦ φάσιν" where the transcription has "auroettierw it is requisite to read avyo fiSwiv" (page image, printed page 157 = archive v1-n232). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11732· transliteration
“rtKinovpya, it is necessary to read yivta’impya” → “πκινοῦργα, it is necessary to read γιγαντοῦργα”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "πκινοῦργα, it is necessary to read γιγαντοῦργα" where the transcription has "rtKinovpya, it is necessary to read yivta’impya" (page image, printed page 380 = archive v1-n455). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11733· transliteration
“iv «AAoi$; it is requisite to read tv oAAi)Aoi$” → “ἐν ἄλλοις; it is requisite to read ἐν ἀλλήλοις”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἐν ἄλλοις; it is requisite to read ἐν ἀλλήλοις" where the transcription has "iv «AAoi$; it is requisite to read tv oAAi)Aoi$" (page image, printed page 383 = archive v1-n458). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11734· transliteration
“fur avrov, it is necessary to read pur avrri $.” → “φῦρ αὐτῶν, it is necessary to read φῦρ αὐτῇς”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "φῦρ αὐτῶν, it is necessary to read φῦρ αὐτῇς" where the transcription has "fur avrov, it is necessary to read pur avrri $." (page image, printed page 205 = archive v2-n710). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11735· transliteration
“Tonu/xtvov, it is necessary to read xoiouptvsf” → “Τονύμενον, it is necessary to read κοιούμενος”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "Τονύμενον, it is necessary to read κοιούμενος" where the transcription has "Tonu/xtvov, it is necessary to read xoiouptvsf" (page image, printed page 365 = archive v1-n440). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11736· transliteration
“•*/!« tty* 0 wtoi;, aruf c vtt ttoinv #/xc u»” → “οὐδὲ ἱερὰ θνητοῖς, ἀταρ οὐδὲ θεοῖσιν ὁμοία”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "οὐδὲ ἱερὰ θνητοῖς, ἀταρ οὐδὲ θεοῖσιν ὁμοία" where the transcription has "•*/!« tty* 0 wtoi;, aruf c vtt ttoinv #/xc u»" (page image, archive n22). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11737· transliteration
“■nu oAijf, it is necessary to read tpij oAo»*” → “nu ολιγῆς, it is necessary to read τῆς ὅλης”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "nu ολιγῆς, it is necessary to read τῆς ὅλης" where the transcription has "■nu oAijf, it is necessary to read tpij oAo»*" (page image, printed page 21 = archive v2-n526). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11738· transliteration
“Oivou f«VT» ftcAy xoffbco kajk? xou jxoi mix*” → “Οἴνου πάντα μέλη κόσμῳ λάβε, καὶ μοι ἔνεικε”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "Οἴνου πάντα μέλη κόσμῳ λάβε, καὶ μοι ἔνεικε" where the transcription has "Oivou f«VT» ftcAy xoffbco kajk? xou jxoi mix*" (page image, printed page 217 = archive v2-n722). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. WITNESS 1.0.6: Project Gutenberg’s human-proofread transcription of the 1816 print reads ["as","for","instance","orpheus","i","e"] at this point (vol2, slid). NO EVIDENCE EITHER WAY: the witness reads words of its own at the point and our own `before` is the transcription’s GARBLE, so the witness’s words are not evidence about this rule at allproclus-theology-of-plato-taylor-1816:r11739· transliteration
“roXuf r* $n* t rt pmg rov **t(xqv, r* it nBt” → “τολuφ r* $n* τ rt pmg rov **t(xqv, r* it nBx”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "τολuφ r* $n* τ rt pmg rov **t(xqv, r* it nBx" where the transcription has "roXuf r* $n* t rt pmg rov **t(xqv, r* it nBt" (page image, archive n21). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. 1.0.9 LEAF CORRECTION (2026-10-08). THE EVIDENCE LEAF OF THIS RULE IS MOVED, from n21 to n20: the passage stands on archive leaf n20 (front matter) and not on n21. MEASURED over every transliteration rule that cites an evidence leaf (189 of them) by reading each cited plate BLIND — 123 page transcriptions, the reader shown no candidate text — and locating each page’s own first printed line in our source.txt: this rule’s find stands before n21’s own first body line (source line 965) and is therefore on the page before it, whose first body line is 912. THE READING IS NOT TOUCHED — only the leaf the evidence is drawn from, so that a corroboration pass reads the page the passage is actually printed on.proclus-theology-of-plato-taylor-1816:r11740· transliteration
“rrw-reAaj, it >9 necessary to read nri|SoAa$” → “μόνον ἁρμοστεον, it is necessary to read εἶδος ἀναρμόστεον”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "μόνον ἁρμοστεον, it is necessary to read εἶδος ἀναρμόστεον" where the transcription has "rrw-reAaj, it >9 necessary to read nri|SoAa$" (page image, printed page 23 = archive v1-n98). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11741· transliteration
“avcthjetoif it is necessary to read mmAoyiaf” → “ἀνθαιρέσεως it is necessary to read ἀνθολογίας”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἀνθαιρέσεως it is necessary to read ἀνθολογίας" where the transcription has "avcthjetoif it is necessary to read mmAoyiaf" (page image, printed page 151 = archive v1-n226). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11742· transliteration
“tvrovo 15 it is necessary to read tv ropvoi$” → “τυραννοις it is necessary to read ἐν τυραννισι”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "τυραννοις it is necessary to read ἐν τυραννισι" where the transcription has "tvrovo 15 it is necessary to read tv ropvoi$" (page image, printed page 162 = archive v1-n237). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11743· transliteration
“wtgi ourov, it is necessary to read mgi camp” → “περὶ οὔρων, it is necessary to read περὶ κάμπης”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "περὶ οὔρων, it is necessary to read περὶ κάμπης" where the transcription has "wtgi ourov, it is necessary to read mgi camp" (page image, printed page 57 = archive v2-n562). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11744· transliteration
“mtw $wiy, it is necessary to read tutrm fwjy” → “μὲν φύσιν, it is necessary to read τοῦτον φύσει”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "μὲν φύσιν, it is necessary to read τοῦτον φύσει" where the transcription has "mtw $wiy, it is necessary to read tutrm fwjy" (page image, printed page 118 = archive v2-n623). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11745· transliteration
“hu e r ro ne o u sly ptv*s instead of jmvov” → “πρὸς erroneously prints instead of μόνον”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "πρὸς erroneously prints instead of μόνον" where the transcription has "hu e r ro ne o u sly ptv*s instead of jmvov" (page image, printed page 122 = archive v2-n627). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11746· transliteration
“T£ioSi)y, it is necessary to read rpteiSaif” → “Τειδίαν, it is necessary to read τρισάειφ”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "Τειδίαν, it is necessary to read τρισάειφ" where the transcription has "T£ioSi)y, it is necessary to read rpteiSaif" (page image, printed page 197 = archive v2-n702). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11747· transliteration
“- 1 For rjyoj it is necessary to read nvij” → “For ἐξήρηται it is necessary to read ἐξήρτηται”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "For ἐξήρηται it is necessary to read ἐξήρτηται" where the transcription has "- 1 For rjyoj it is necessary to read nvij" (page image, printed page 166 = archive v1-n241). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11748· transliteration
“xpovxmif it is necessary to read *gt*x»nts” → “χρονισμῷ it is necessary to read ἄτμητος”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "χρονισμῷ it is necessary to read ἄτμητος" where the transcription has "xpovxmif it is necessary to read *gt*x»nts" (page image, printed page 207 = archive v1-n282). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11749· transliteration
“' For xvtov, it is necessary to read «otwv” → “For χύτον, it is necessary to read κοινόν”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "For χύτον, it is necessary to read κοινόν" where the transcription has "' For xvtov, it is necessary to read «otwv" (page image, printed page 401 = archive v1-n476). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11750· transliteration
“For wpwrmy it is necessary to read wpo T«y” → “. For πρώτην it is necessary to read πρὸ τὴν”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads ". For πρώτην it is necessary to read πρὸ τὴν" where the transcription has "For wpwrmy it is necessary to read wpo T«y" (page image, printed page 415 = archive v1-n490). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11751· transliteration
“x*gi***> >t >* necessary to read x»girror” → “περιόδους, it is necessary to read προόδους”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "περιόδους, it is necessary to read προόδους" where the transcription has "x*gi***> >t >* necessary to read x»girror" (page image, printed page 65 = archive v2-n570). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11752· transliteration
“qfgiaw tAanjga xvvaov yimwv rt xcu uygoov” → “ἠερίων ἐλατήρα κυνῶν χθονίων τε καὶ ὑγρῶν”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἠερίων ἐλατήρα κυνῶν χθονίων τε καὶ ὑγρῶν" where the transcription has "qfgiaw tAanjga xvvaov yimwv rt xcu uygoov" (page image, printed page 271 = archive v2-n776). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. WITNESS 1.0.6: Project Gutenberg’s human-proofread transcription of the 1816 print reads ["serial","governors","whence","also","the","chaldean","oracle","says","i","e"] at this point (vol2, slid). NO EVIDENCE EITHER WAY: the witness reads words of its own at the point and our own `before` is the transcription’s GARBLE, so the witness’s words are not evidence about this rule at allproclus-theology-of-plato-taylor-1816:r11753· transliteration
“Tgcor>jv it is necessary to read rgijT)y” → “Θεωρίαν it is necessary to read τρίτην”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "Θεωρίαν it is necessary to read τρίτην" where the transcription has "Tgcor>jv it is necessary to read rgijT)y" (page image, printed page 271 = archive v1-n346). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11754· transliteration
“fftpaftw, it is necessary to read wpo«tw” → “στρατῷ, it is necessary to read προσώτῳ”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "στρατῷ, it is necessary to read προσώτῳ" where the transcription has "fftpaftw, it is necessary to read wpo«tw" (page image, printed page 144 = archive v2-n649). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11755· transliteration
“uxmg cnti tm toXAoi, xou xugioi aroXAoi” → “ὥσπερ εἰσὶ θεοὶ πολλοί, καὶ κύριοι πολλοί”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ὥσπερ εἰσὶ θεοὶ πολλοί, καὶ κύριοι πολλοί" where the transcription has "uxmg cnti tm toXAoi, xou xugioi aroXAoi" (page image, archive n0). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11756· transliteration
“xitrra. xetfovra, I read xavrct t* w r*” → “πάντα. ὄντα, I read πάντᾳ τῷ ὄντι”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "πάντα. ὄντα, I read πάντᾳ τῷ ὄντι" where the transcription has "xitrra. xetfovra, I read xavrct t* w r*" (page image, printed page 128 = archive v1-n203). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11757· transliteration
“JivTipa it is necessary to read iiunpvf” → “Δῐνθῃρα it is necessary to read Ἡμῖπρηφ”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "Δῐνθῃρα it is necessary to read Ἡμῖπρηφ" where the transcription has "JivTipa it is necessary to read iiunpvf" (page image, printed page 194 = archive v1-n269). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11758· transliteration
“ngioiou it is necessary to read *goo8ev” → “νοεξων, it is necessary to read νοητων”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "νοεξων, it is necessary to read νοητων" where the transcription has "ngioiou it is necessary to read *goo8ev" (page image, printed page 311 = archive v1-n386). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11759· transliteration
“ev rw ijXiep tiero to tnojveojua avrov” → “ἐν τῷ ἡλίῳ ἔθετο τὸ σκήνωμα αὐτοῦ”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἐν τῷ ἡλίῳ ἔθετο τὸ σκήνωμα αὐτοῦ" where the transcription has "ev rw ijXiep tiero to tnojveojua avrov" (page image, archive n8). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11760· transliteration
“ette it is necessary to read «*' *vtwv” → “εἶπε it is necessary to read «*' ἐντῶν”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "εἶπε it is necessary to read «*' ἐντῶν" where the transcription has "ette it is necessary to read «*' *vtwv" (page image, printed page 33 = archive v2-n538). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11761· transliteration
“epm/uvos here, it is necessary to read” → “συνέκτεινατο, it is necessary to read συνετεκταίνετο.”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "συνέκτεινατο, it is necessary to read συνετεκταίνετο." where the transcription has "epm/uvos here, it is necessary to read" (page image, printed page 287 = archive v2-n792). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11763· transliteration
“v^ncfix xat /3«8o$) with printipali-'” → “ὕψωμα καὶ βάθος) with principali-”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ὕψωμα καὶ βάθος) with principali-" where the transcription has "v^ncfix xat /3«8o$) with printipali-'" (page image, archive n20). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. WITNESS 1.0.6: Project Gutenberg’s human-proofread transcription of the 1816 print reads ["with","principalities","and","powers","and","is","one","of","the","aeones"] at this point (vol1, slid). NO EVIDENCE EITHER WAY: the witness reads words of its own at the point and our own `before` is the transcription’s GARBLE, so the witness’s words are not evidence about this rule at allproclus-theology-of-plato-taylor-1816:r11764· transliteration
“mrm after the words to s» mijtov «Soj” → “ἀμερῆ after the words τὸ μεριστὸν εἶδος”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἀμερῆ after the words τὸ μεριστὸν εἶδος" where the transcription has "mrm after the words to s» mijtov «Soj" (page image, printed page 219 = archive v1-n294). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11765· transliteration
“r»i$ «Am$, it is necessary to read nn” → “τῆς ἀκλινῶς, it is necessary to read τιι”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "τῆς ἀκλινῶς, it is necessary to read τιι" where the transcription has "r»i$ «Am$, it is necessary to read nn" (page image, printed page 21 = archive v2-n526). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11766· transliteration
“««x«rov, k » necessary to read t^waty” → “κακὸν, is necessary to read τὸ κοινόν”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "κακὸν, is necessary to read τὸ κοινόν" where the transcription has "««x«rov, k » necessary to read t^waty" (page image, printed page 62 = archive v2-n567). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11767· transliteration
“itgooiovs is omitted in the original” → “In the original it is παραπετα instead of παραπετάσματα”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "In the original it is παραπετα instead of παραπετάσματα" where the transcription has "itgooiovs is omitted in the original" (page image, printed page 212 = archive v2-n717). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11768· transliteration
“Ka.ga.xeTa instead of xagarna^fuiiTa” → “Καθάπερ instead of καταπηγνυνται”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "Καθάπερ instead of καταπηγνυνται" where the transcription has "Ka.ga.xeTa instead of xagarna^fuiiTa" (page image, printed page 213 = archive v2-n718). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11769· transliteration
“ling"^’ “^ ^ °" e ° f,he sU b °° ks” → “ting."^" Οὐδὲ γὰρ εὗρον ἐγώ τινα τῶν βιβλίων”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ting.\"^\" Οὐδὲ γὰρ εὗρον ἐγώ τινα τῶν βιβλίων" where the transcription has "ling\"^’ “^ ^ °\" e ° f,he sU b °° ks" (page image, archive n46). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11770· transliteration
“urf^xoer^t»> as in the translation.” → “ὑπάρξεως, as in the translation”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ὑπάρξεως, as in the translation" where the transcription has "urf^xoer^t»> as in the translation." (page image, printed page 47 = archive v1-n122). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11771· transliteration
“Thi, h’ h d DameS b * accuratol >'” → “Τίς ὁ δὴ Δάμασις b' ἀκριβῶς”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "Τίς ὁ δὴ Δάμασις b' ἀκριβῶς" where the transcription has "Thi, h’ h d DameS b * accuratol >'" (page image, archive n46). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. WITNESS 1.0.6: Project Gutenberg’s human-proofread transcription of the 1816 print reads ["powers","this","however","is","not","done","by","him","in","any","one","of","the","six","books"] at this point (vol1, slid). NO EVIDENCE EITHER WAY: the witness reads words of its own at the point and our own `before` is the transcription’s GARBLE, so the witness’s words are not evidence about this rule at allproclus-theology-of-plato-taylor-1816:r11772· transliteration
“?rpa£*V| it is necessary to read r” → “ αἰτιον it is necessary to read αἰτίου.”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads " αἰτιον it is necessary to read αἰτίου." where the transcription has "?rpa£*V| it is necessary to read r" (page image, printed page 134 = archive v1-n209). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11773· transliteration
“it is necessary to read *£>ign}Tai” → “γονίμου, it is necessary to read μονίμου”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "γονίμου, it is necessary to read μονίμου" where the transcription has "it is necessary to read *£>ign}Tai" (page image, printed page 30 = archive v2-n535). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11774· transliteration
“faityw it is necessary to read *f” → “ὑποθέσεως I read ὑποστάσεως.”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ὑποθέσεως I read ὑποστάσεως." where the transcription has "faityw it is necessary to read *f" (page image, archive n51). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. WITNESS 1.0.6: Project Gutenberg’s human-proofread transcription of the 1816 print reads ["contents","of","this","belong","to","the","preceding","chapter","69","for","it","is","necessary","to","read","70","in","the","greek"] at this point (vol1, slid). NO EVIDENCE EITHER WAY: the witness reads words of its own at the point and our own `before` is the transcription’s GARBLE, so the witness’s words are not evidence about this rule at allproclus-theology-of-plato-taylor-1816:r11775· transliteration
“T«y>i£ia* it is necessary to read” → “ἰδρυαν, it is necessary to read ἰδρυσιν.”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἰδρυαν, it is necessary to read ἰδρυσιν." where the transcription has "T«y>i£ia* it is necessary to read" (page image, printed page 134 = archive v2-n639). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11776· transliteration
“ov gym xaprov erexov ijAio* rymro” → “οὐ γὰρ καρπὸν ἔτεκον ἥλιος ῥητὸς”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "οὐ γὰρ καρπὸν ἔτεκον ἥλιος ῥητὸς" where the transcription has "ov gym xaprov erexov ijAio* rymro" (page image, printed page 169 = archive v2-n674). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11777· transliteration
“n ouv; raw fiopiwv sxartpov rot/” → “ἢ οὖν; τῶν μορίων ἑκάτερον τοῦ”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἢ οὖν; τῶν μορίων ἑκάτερον τοῦ" where the transcription has "n ouv; raw fiopiwv sxartpov rot/" (page image, printed page 213 = archive v1-n288). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11778· transliteration
“<nj/xau»»v rov #«ov xai met pour” → “σημαίνων τὸν οἷον καὶ τινα νοῦν”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "σημαίνων τὸν οἷον καὶ τινα νοῦν" where the transcription has "<nj/xau»»v rov #«ov xai met pour" (page image, printed page 218 = archive v2-n723). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. WITNESS 1.0.6: Project Gutenberg’s human-proofread transcription of the 1816 print reads ["such","as","and","particular"] at this point (vol2, slid). NO EVIDENCE EITHER WAY: the witness reads words of its own at the point and our own `before` is the transcription’s GARBLE, so the witness’s words are not evidence about this rule at allproclus-theology-of-plato-taylor-1816:r11779· transliteration
“ouig, xgog rag xocrfLoxgarogag” → “σιας, προς τας κοσμοκρατορας”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "σιας, προς τας κοσμοκρατορας" where the transcription has "ouig, xgog rag xocrfLoxgarogag" (page image, archive n18). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. WITNESS 1.0.6: Project Gutenberg’s human-proofread transcription of the 1816 print reads ["i","e","the","seven","planets","augustin","17"] at this point (vol1, slid). NO EVIDENCE EITHER WAY: the witness reads words of its own at the point and our own `before` is the transcription’s GARBLE, so the witness’s words are not evidence about this rule at allproclus-theology-of-plato-taylor-1816:r11781· transliteration
“t/wrm, it is necessary to read” → “ἀποστασεως, it is necessary to read ὑποστασεως.”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἀποστασεως, it is necessary to read ὑποστασεως." where the transcription has "t/wrm, it is necessary to read" (page image, printed page 30 = archive v1-n105). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11782· transliteration
“epfiaivovT* read ex/S cuvovrot” → “παρασκευαζοντα it is requisite to read περισχιαζοντα”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "παρασκευαζοντα it is requisite to read περισχιαζοντα" where the transcription has "epfiaivovT* read ex/S cuvovrot" (page image, printed page 59 = archive v1-n134). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11783· transliteration
“xgog rag ag%ag, rgog rag tgov” → “προς τας αρχας, προς τας εξου”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "προς τας αρχας, προς τας εξου" where the transcription has "xgog rag ag%ag, rgog rag tgov" (page image, archive n18). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. WITNESS 1.0.6: Project Gutenberg’s human-proofread transcription of the 1816 print reads ["i","e","the","seven","planets","augustin","17"] at this point (vol1, slid). NO EVIDENCE EITHER WAY: the witness reads words of its own at the point and our own `before` is the transcription’s GARBLE, so the witness’s words are not evidence about this rule at allproclus-theology-of-plato-taylor-1816:r11784· transliteration
“XX' x«t« tov vm r/#» voyroy,” → “ἀλλ᾽ κατὰ τὸν ὑπὲρ πρὸς νοητῶν,»”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἀλλ᾽ κατὰ τὸν ὑπὲρ πρὸς νοητῶν,»" where the transcription has "XX' x«t« tov vm r/#» voyroy," (page image, printed page 187 = archive v1-n262). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11785· transliteration
“TO S' oAov toutq QvXotccoov” → “τὸ δ᾿ ὅλον τοῦτο φυλάσσων”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "τὸ δ᾿ ὅλον τοῦτο φυλάσσων" where the transcription has "TO S' oAov toutq QvXotccoov" (page image, archive n13). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11786· transliteration
“rpwm apx 1 ** orog v*tp mtv” → “τρὶς ἀρρητος σκότος ὑπὲρ πᾶν”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "τρὶς ἀρρητος σκότος ὑπὲρ πᾶν" where the transcription has "rpwm apx 1 ** orog v*tp mtv" (page image, archive n28). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11787· transliteration
“stead of aMqAo*;, I read .” → “Instead of ἀπάγων ἡμᾶς ἀπὸ τῶν νοητῶν ἐπὶ τὴν αἰσθητὴν φύσιν, it is necessary to read ἐπάγων ἡμᾶς”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "Instead of ἀπάγων ἡμᾶς ἀπὸ τῶν νοητῶν ἐπὶ τὴν αἰσθητὴν φύσιν, it is necessary to read ἐπάγων ἡμᾶς" where the transcription has "stead of aMqAo*;, I read ." (page image, printed page 298 = archive v1-n373). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11788· transliteration
“umgavco xa<rrig agxyg xcu” → “ὑπερανω πασης αρχης και”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ὑπερανω πασης αρχης και" where the transcription has "umgavco xa<rrig agxyg xcu" (page image, archive n18). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. WITNESS 1.0.6: Project Gutenberg’s human-proofread transcription of the 1816 print reads ["and","in","the","6th","chapter","and","12th"] at this point (vol1, slid). NO EVIDENCE EITHER WAY: the witness reads words of its own at the point and our own `before` is the transcription’s GARBLE, so the witness’s words are not evidence about this rule at allproclus-theology-of-plato-taylor-1816:r11789· transliteration
“x«i <rt$*Qfp to{ wregox*{” → “ἀλλὰ καὶ ὑλαίοις ὅσα δουλεύει Συνοχεῦσι”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἀλλὰ καὶ ὑλαίοις ὅσα δουλεύει Συνοχεῦσι" where the transcription has "x«i <rt$*Qfp to{ wregox*{" (page image, archive n19). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. 1.0.9 LEAF CORRECTION (2026-10-08). THE EVIDENCE LEAF OF THIS RULE IS MOVED, from n19 to n18: the passage stands on archive leaf n18 (front matter) and not on n19. MEASURED over every transliteration rule that cites an evidence leaf (189 of them) by reading each cited plate BLIND — 123 page transcriptions, the reader shown no candidate text — and locating each page’s own first printed line in our source.txt: this rule’s find stands before n19’s own first body line (source line 863) and is therefore on the page before it, whose first body line is 808. THE READING IS NOT TOUCHED — only the leaf the evidence is drawn from, so that a corroboration pass reads the page the passage is actually printed on.proclus-theology-of-plato-taylor-1816:r11790· transliteration
“ncuroi <rff fJLeX.ir.Bi” → “πᾶσα σε μελπει”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "πᾶσα σε μελπει" where the transcription has "ncuroi <rff fJLeX.ir.Bi" (page image, archive n11). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11791· transliteration
“oXiya /xsv, «AA* txuvwv” → “ὀλίγα μὲν, ἀλλ᾿ ἐκείνων”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ὀλίγα μὲν, ἀλλ᾿ ἐκείνων" where the transcription has "oXiya /xsv, «AA* txuvwv" (page image, archive n13). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11792· transliteration
“vevtfjuiiitvum ftogfoi>” → “νενεμημεναισι μορφαι,”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "νενεμημεναισι μορφαι," where the transcription has "vevtfjuiiitvum ftogfoi>" (page image, archive n13). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11793· transliteration
“xutos ougavoov tXunrw” → “κύτος οὐρανῶν ἑλίσσει·”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "κύτος οὐρανῶν ἑλίσσει·" where the transcription has "xutos ougavoov tXunrw" (page image, archive n13). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11796· transliteration
“tv vq axjonjn raviy” → “ἐν τῷ ἀξιώματι ταύτῃ”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἐν τῷ ἀξιώματι ταύτῃ" where the transcription has "tv vq axjonjn raviy" (page image, printed page 286 = archive v1-n361). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11797· transliteration
“xa/A/Mg<pov ayax/ta” → “καλλίμορφον ἀγαθὸν”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "καλλίμορφον ἀγαθὸν" where the transcription has "xa/A/Mg<pov ayax/ta" (page image, printed page 228 = archive v2-n733). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11799· transliteration
“xa rrtpgowf ff$oy>” → “κατέχουσι πέδον,”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "κατέχουσι πέδον," where the transcription has "xa rrtpgowf ff$oy>" (page image, archive n13). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11800· transliteration
“upovoovmw, axnrtp'” → “φρονοῦσι, ἀκίνητοι”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "φρονοῦσι, ἀκίνητοι" where the transcription has "upovoovmw, axnrtp'" (page image, printed page 37 = archive v2-n542). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11801· transliteration
“fax* r *x\iyrj$,'” → “ψυχα τ' ακλινης,”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ψυχα τ' ακλινης," where the transcription has "fax* r *x\\iyrj$,'" (page image, archive n12). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. 1.0.9 LEAF CORRECTION (2026-10-08). THE EVIDENCE LEAF OF THIS RULE IS MOVED, from n12 to n11: the passage stands on archive leaf n11 (front matter) and not on n12. MEASURED over every transliteration rule that cites an evidence leaf (189 of them) by reading each cited plate BLIND — 123 page transcriptions, the reader shown no candidate text — and locating each page’s own first printed line in our source.txt: this rule’s find stands before n12’s own first body line (source line 514) and is therefore on the page before it, whose first body line is 441. THE READING IS NOT TOUCHED — only the leaf the evidence is drawn from, so that a corroboration pass reads the page the passage is actually printed on.proclus-theology-of-plato-taylor-1816:r11802· transliteration
“tx tow vo«g«v Uow” → “ἐκ τῶν νοητῶν ὕλων”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἐκ τῶν νοητῶν ὕλων" where the transcription has "tx tow vo«g«v Uow" (page image, printed page 49 = archive v1-n124). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11803· transliteration
“yorifioy Ogfxtjg” → “γονιμον Θρηκης”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "γονιμον Θρηκης" where the transcription has "yorifioy Ogfxtjg" (page image, archive n12). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11804· transliteration
“ymot (JLaxagoov” → “γενεά μακάρων”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "γενεά μακάρων" where the transcription has "ymot (JLaxagoov" (page image, archive n11). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11805· transliteration
“o< iregi HoafMV” → “οἱ περὶ κοσμον”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "οἱ περὶ κοσμον" where the transcription has "o< iregi HoafMV" (page image, archive n11). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11808· transliteration
“tfntova’t yvt»$” → “ἐφέπουσι γυίας”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἐφέπουσι γυίας" where the transcription has "tfntova’t yvt»$" (page image, archive n13). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11809· transliteration
“igya /3 Igonix” → “εργα βροτεια”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "εργα βροτεια" where the transcription has "igya /3 Igonix" (page image, archive n12). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. 1.0.9 LEAF CORRECTION (2026-10-08). THE EVIDENCE LEAF OF THIS RULE IS MOVED, from n12 to n11: the passage stands on archive leaf n11 (front matter) and not on n12. MEASURED over every transliteration rule that cites an evidence leaf (189 of them) by reading each cited plate BLIND — 123 page transcriptions, the reader shown no candidate text — and locating each page’s own first printed line in our source.txt: this rule’s find stands before n12’s own first body line (source line 514) and is therefore on the page before it, whose first body line is 441. THE READING IS NOT TOUCHED — only the leaf the evidence is drawn from, so that a corroboration pass reads the page the passage is actually printed on.proclus-theology-of-plato-taylor-1816:r11810· transliteration
“IxtTOVT* itOV$” → “Ικετευσα θεους”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "Ικετευσα θεους" where the transcription has "IxtTOVT* itOV$" (page image, archive n12). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11811· transliteration
“Of r* arrtngat” → “οἱ τ᾿ ἀντίπεραν”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "οἱ τ᾿ ἀντίπεραν" where the transcription has "Of r* arrtngat" (page image, archive n13). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11813· transliteration
“xtovmf oyxov$” → “χθονίους όγκους”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "χθονίους όγκους" where the transcription has "xtovmf oyxov$" (page image, archive n12). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. 1.0.9 LEAF CORRECTION (2026-10-08). THE EVIDENCE LEAF OF THIS RULE IS MOVED, from n12 to n11: the passage stands on archive leaf n11 (front matter) and not on n12. MEASURED over every transliteration rule that cites an evidence leaf (189 of them) by reading each cited plate BLIND — 123 page transcriptions, the reader shown no candidate text — and locating each page’s own first printed line in our source.txt: this rule’s find stands before n12’s own first body line (source line 514) and is therefore on the page before it, whose first body line is 441. THE READING IS NOT TOUCHED — only the leaf the evidence is drawn from, so that a corroboration pass reads the page the passage is actually printed on.proclus-theology-of-plato-taylor-1816:r11814· transliteration
“Bppryg*S otoi” → “δρηστηρας όσοι”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "δρηστηρας όσοι" where the transcription has "Bppryg*S otoi" (page image, archive n12). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11815· transliteration
“t# aqrm *gwov” → “τῷ ἀρμονικῷ”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "τῷ ἀρμονικῷ" where the transcription has "t# aqrm *gwov" (page image, printed page 292 = archive v1-n367). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11817· transliteration
“^aAxijSovia*” → “Χαλκηδονίας”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "Χαλκηδονίας" where the transcription has "^aAxijSovia*" (page image, archive n13). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11819· transliteration
“wtrxtg (xinn” → “ὥσπερ (φύσιν”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ὥσπερ (φύσιν" where the transcription has "wtrxtg (xinn" (page image, printed page 254 = archive v1-n329). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. 1.0.9 LEAF CORRECTION (2026-10-08). THE EVIDENCE LEAF OF THIS RULE IS MOVED, from v1-n329 to v1-n328: the passage stands on archive leaf v1-n328 (printed page 253) and not on v1-n329. MEASURED over every transliteration rule that cites an evidence leaf (189 of them) by reading each cited plate BLIND — 123 page transcriptions, the reader shown no candidate text — and locating each page’s own first printed line in our source.txt: this rule’s find stands before v1-n329’s own first body line (source line 16662) and is therefore on the page before it, whose first body line is 16598. THE READING IS NOT TOUCHED — only the leaf the evidence is drawn from, so that a corroboration pass reads the page the passage is actually printed on.proclus-theology-of-plato-taylor-1816:r11820· transliteration
“ecS*v»T»s i»” → “ἑξῆς ἥμισυ is”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἑξῆς ἥμισυ is" where the transcription has "ecS*v»T»s i»" (page image, printed page 129 = archive v2-n634). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11822· transliteration
“* «raptn; m” → “ «τὸ πρῶτον»”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads " «τὸ πρῶτον»" where the transcription has "* «raptn; m" (page image, printed page 103 = archive v1-n178). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11823· transliteration
“ro tvyecro*” → “το τυγέσθαι”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "το τυγέσθαι" where the transcription has "ro tvyecro*" (page image, printed page 403 = archive v1-n478). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11824· transliteration
“Se /uwAA ov” → “δὲ ἀμελεῖ οὖν”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "δὲ ἀμελεῖ οὖν" where the transcription has "Se /uwAA ov" (page image, printed page 82 = archive v2-n587). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11825· transliteration
“rtrajj-tvov” → “τραπ-ενου”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "τραπ-ενου" where the transcription has "rtrajj-tvov" (page image, printed page 172 = archive v2-n677). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11826· transliteration
“aad W/tusv” → “ead ὥσπερ”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ead ὥσπερ" where the transcription has "aad W/tusv" (page image, printed page 236 = archive v2-n741). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11827· transliteration
“$ipo/ytrw” → “δι᾽ ὕγροῦ”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "δι᾽ ὕγροῦ" where the transcription has "$ipo/ytrw" (page image, archive n24). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11828· transliteration
“xeu Auroi” → “καὶ αὐτοὶ”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "καὶ αὐτοὶ" where the transcription has "xeu Auroi" (page image, printed page 118 = archive v2-n623). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 2 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11829· transliteration
“figo/nym” → “φῐγονυμ”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "φῐγονυμ" where the transcription has "figo/nym" (page image, printed page 359 = archive v1-n434). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11830· transliteration
“svewm* ” → “ἕνωσιν˙”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἕνωσιν˙" where the transcription has "svewm* " (page image, archive n21). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11832· transliteration
“an expc” → “ἀλεξάνεμος, an expe”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἀλεξάνεμος, an expe" where the transcription has "an expc" (page image, archive n44). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. WITNESS 1.0.6: Project Gutenberg’s human-proofread transcription of the 1816 print reads ["an","expeller","of","wind","62"] at this point (vol1, slid). NO EVIDENCE EITHER WAY: the witness reads words of its own at the point and our own `before` is the transcription’s GARBLE, so the witness’s words are not evidence about this rule at allproclus-theology-of-plato-taylor-1816:r11834· transliteration
“txpetmi” → “ἐκπλέω”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἐκπλέω" where the transcription has "txpetmi" (page image, printed page 182 = archive v1-n257). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11835· transliteration
“ihorvpa” → “ἱδρῦσθαι”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἱδρῦσθαι" where the transcription has "ihorvpa" (page image, printed page 191 = archive v1-n266). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 3 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11836· transliteration
“vwftios” → “κατὰ τὴν προσήκουσ̈αν ἐκείνην”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "κατὰ τὴν προσήκουσ̈αν ἐκείνην" where the transcription has "vwftios" (page image, printed page 252 = archive v1-n327). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11837· transliteration
“taides” → “ ἀβίδες”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads " ἀβίδες" where the transcription has "taides" (page image, archive n55). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. WITNESS 1.0.6: Project Gutenberg’s human-proofread transcription of the 1816 print reads ["eternity","abides","is","the","summit","of","intelligibles","chapter","xviii"] at this point (vol1, slid). NO EVIDENCE EITHER WAY: the witness reads words of its own at the point and our own `before` is the transcription’s GARBLE, so the witness’s words are not evidence about this rule at allproclus-theology-of-plato-taylor-1816:r11839· transliteration
“o/tuov” → “οὔτεον”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "οὔτεον" where the transcription has "o/tuov" (page image, printed page 53 = archive v2-n558). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11840· transliteration
“Siot#” → “διότι”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "διότι" where the transcription has "Siot#" (page image, archive n75). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist. 1.0.9 LEAF CORRECTION (2026-10-08). THE EVIDENCE LEAF OF THIS RULE IS MOVED, from n75 to n74: the passage stands on archive leaf n74 (front matter) and not on n75. MEASURED over every transliteration rule that cites an evidence leaf (189 of them) by reading each cited plate BLIND — 123 page transcriptions, the reader shown no candidate text — and locating each page’s own first printed line in our source.txt: this rule’s find stands before n75’s own first body line (source line 4033) and is therefore on the page before it, whose first body line is 3987. THE READING IS NOT TOUCHED — only the leaf the evidence is drawn from, so that a corroboration pass reads the page the passage is actually printed on.proclus-theology-of-plato-taylor-1816:r11895· transliteration
“fartf%e npn, the second honours next to the Gods, as Celsus calls those t«; *\wwo**e npatf” → “δευτέρας τιμάς, the second honours next to the Gods, as Celsus calls those τὰς προσηκούσας τιμάς”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "φάρσχε πρῶ, the second honours next to the Gods, as Celsus calls those τὰς δευτερείας τιμᾶσ" where the transcription has "fartf%e npn, the second honours next to the Gods, as Celsus calls those t«; *\\wwo**e npatf" (page image, archive n22). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.THE READING THIS RULE SERVED WAS FABRICATED. It read “φάρσχε πρῶ, the second honours next to the Gods, as Celsus calls those τὰς δευτερείας τιμᾶσ”; BOTH Greek phrases are refuted — “φάρσχε” is not a Greek word, and “δευτερείας” is a different word from the print’s. The print reads, per the human-proofread word-level transcription of this print (Project Gutenberg ebook 77393 / 78800, Distributed Proofreaders) and a BLIND read of the page image (the candidate text never shown to the model, google/gemini-2.5-flash, the whole page and an archive crop of the same line): “we give δευτέρας τιμάς, the second honours next to the Gods, as Celsus calls those τὰς προσηκούσας τιμάς the due honours that belong to the lower demons” — and the human-proofread transcription reads the letters “δευτερας τιμας” and “tας προσηκουσας τιμας” at that point. Both match the English glosses the print sets immediately beside them (“the second honours”, “the due honours”). THE EVIDENCE LEAF IS CORRECTED TOO: the passage stands on archive leaf n22 (printed page xxvii), NOT n23 (page xxviii) — MEASURED by a blind read of both leaves, n23 carrying the Panchoa passage and n22 this one. THE REVIEW FLAG IS RESTORED. It was cleared at 1.0.4 on a witness verdict that cannot see Greek: the witness locator’s token stream is ASCII letters and digits ONLY, so for a transliteration rule its “the witness carries the rule’s whole `after` reading” test is decided on the English AROUND the Greek and the Greek the rule exists to supply is never compared. 1.0.8 ADJUDICATION (2026-10-08). THIS RULE REPLACES r11693, WITHDRAWN WITH THIS VERSION (the id is spent, never reused; a corrected reading is a NEW id, not an edit beneath a spent one). 1.0.9 NOTE (2026-10-08). THE LEAF NAMED IN THIS RULE'S OWN PROSE IS CORRECTED from “archive n23” to “archive n22”, which is the leaf the rule's own `evidence` and `witness` have named since 1.0.8: the prose still carried the DRAW's leaf while the record carried the measured one. MEASURED: a blind read of both plates puts the passage on archive n22, and no rule of this version may name two different leaves for one reading.proclus-theology-of-plato-taylor-1816:r11896· transliteration
“exi /mv to*$ rou ayxiov vt/v ij&ij xgoivpois tpwraveui) and Dionysius says 1” → “ἐπὶ μὲν τοῖς τοῦ ἀγαθοῦ νῦν ἤδη προθύροις ἐφεστάναι) and Dionysius says ¹”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ἐπὶ μὲν τοῖς τοῦ ἀγαθοῦ νῦν ἤδη προθύροις ἐφε̂στάναι) and Dionysius says ¹" where the transcription has "exi /mv to*$ rou ayxiov vt/v ij&ij xgoivpois tpwraveui) and Dionysius says 1" (page image, archive n21). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.THE READING IS UPHELD AND ONE CHARACTER IN IT IS CORRECTED. “ἐφε̂στάναι” carried U+0302 COMBINING CIRCUMFLEX ACCENT over epsilon, which Greek cannot set (a circumflex needs a long vowel or a diphthong); it is a rendering artifact, not a reading. The print sets “ἐφεστάναι”, per the human-proofread word-level transcription of this print (Project Gutenberg ebook 77393 / 78800, Distributed Proofreaders) and a BLIND read of the page image (the candidate text never shown to the model, google/gemini-2.5-flash, the whole page and an archive crop of the same line) — the blind read of archive leaf n21 gives “(ἐπὶ μὲν τοῖς τοῦ ἀγαθοῦ νῦν ἤδη προθύροις ἐφεστάναι) and Dionysius says”, and the human-proofread transcription reads “προθυροις εφεσταναι”. THE REVIEW FLAG IS RESTORED. It was cleared at 1.0.4 on a witness verdict that cannot see Greek: the witness locator’s token stream is ASCII letters and digits ONLY, so for a transliteration rule its “the witness carries the rule’s whole `after` reading” test is decided on the English AROUND the Greek and the Greek the rule exists to supply is never compared. 1.0.8 ADJUDICATION (2026-10-08). THIS RULE REPLACES r11705, WITHDRAWN WITH THIS VERSION (the id is spent, never reused; a corrected reading is a NEW id, not an edit beneath a spent one).proclus-theology-of-plato-taylor-1816:r11897· transliteration
“aones of the Valentinians were mg, /9 viog, ciyy, aktfina, <ropj«” → “æones of the Valentinians were νοῦς, βύθος, σιγή, ἀλήθεια, σοφία”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "æones of the Valentinians were σιγη, βυθος, νους, αληθεια, σοφια" where the transcription has "aones of the Valentinians were mg, /9 viog, ciyy, aktfina, <ropj«" (page image, archive n17). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.THE LETTERS WERE RIGHT AND THE ORDER WAS WRONG, AND THE EVIDENCE LEAF WAS OFF BY ONE. This rule served “σιγη, βυθος, νους, αληθεια, σοφια”; the print sets “νοῦς, βύθος, σιγή, ἀλήθεια, σοφία”, per the human-proofread word-level transcription of this print (Project Gutenberg ebook 77393 / 78800, Distributed Proofreaders) and a BLIND read of the page image (the candidate text never shown to the model, google/gemini-2.5-flash, the whole page and an archive crop of the same line) — the human-proofread transcription reads exactly “νους, βυθος, σιγη, αληθεια, σοφια”, with the print’s own gloss immediately after it, “i.e. intellect, a profundity, silence, truth, and wisdom”, which fixes the order; the blind read of archive leaf n17 gives the same order. THE EVIDENCE LEAF IS CORRECTED: the passage stands on archive leaf n17 (printed page xxii), NOT n18 (page xxiii) — MEASURED by blind reads of both. THE REVIEW FLAG IS RESTORED. It was cleared at 1.0.4 on a witness verdict that cannot see Greek: the witness locator’s token stream is ASCII letters and digits ONLY, so for a transliteration rule its “the witness carries the rule’s whole `after` reading” test is decided on the English AROUND the Greek and the Greek the rule exists to supply is never compared. 1.0.8 ADJUDICATION (2026-10-08). THIS RULE REPLACES r11707, WITHDRAWN WITH THIS VERSION (the id is spent, never reused; a corrected reading is a NEW id, not an edit beneath a spent one). 1.0.9 NOTE (2026-10-08). THE LEAF NAMED IN THIS RULE'S OWN PROSE IS CORRECTED from “archive n18” to “archive n17”, which is the leaf the rule's own `evidence` and `witness` have named since 1.0.8: the prose still carried the DRAW's leaf while the record carried the measured one. MEASURED: a blind read of both plates puts the passage on archive n17, and no rule of this version may name two different leaves for one reading.proclus-theology-of-plato-taylor-1816:r11898· transliteration
“opoum and d«<»” → “ὀρούω and θείω”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "ορουμ and θεῖα" where the transcription has "opoum and d«<»" (page image, printed page 227 = archive v2-n732). The transcription and the second transcription are both OCR passes over this print and flatten the Greek the same way, so neither can settle it. The leaf was drawn 4 time(s) over 2 model(s) (google/gemini-2.5-flash, anthropic/claude-haiku-4-5); the 2 draw(s) that read these characters were google/gemini-2.5-flash — ONE instrument repeated, not two, so this reading rests on that model alone. (decided by the page image itself.) ONE INSTRUMENT'S READING: these letters rest on ONE reader, google/gemini-2.5-flash — no second reader's own draws corroborate them. The rule is merged with that fact stated, the review flag stands on it, and the rule is listed in the edition's review worklist.THE READING THIS RULE SERVED WAS FABRICATED. It read “ορουμ and θεῖα”, and NEITHER WORD ENDS IN ω WHILE THE PRINT’S OWN SENTENCE ON THAT LINE SAYS THE LETTER DISCUSSED “is here an ω” — and the OCR renders the print’s standalone ω in that very clause as “m” (“which letter is here an m”), so the rule made the same μ/ω misdecode twice. The print reads “from ὀρούω and θείω, according to the Attic custom of adding a letter at the end of a word, which letter is here an ω”, per the human-proofread word-level transcription of this print (Project Gutenberg ebook 77393 / 78800, Distributed Proofreaders) and a BLIND read of the page image (the candidate text never shown to the model, google/gemini-2.5-flash, the whole page and an archive crop of the same line) — the human-proofread transcription reads “from ορουω and θειω”, and the blind read of a crop of that line on archive leaf v2-n732 gives “ὀρούω and θείω”. Both are real Greek verbs of rushing, which is the derivation the sentence states. THE REVIEW FLAG IS RESTORED. It was cleared at 1.0.4 on a witness verdict that cannot see Greek: the witness locator’s token stream is ASCII letters and digits ONLY, so for a transliteration rule its “the witness carries the rule’s whole `after` reading” test is decided on the English AROUND the Greek and the Greek the rule exists to supply is never compared. 1.0.8 ADJUDICATION (2026-10-08). THIS RULE REPLACES r11812, WITHDRAWN WITH THIS VERSION (the id is spent, never reused; a corrected reading is a NEW id, not an edit beneath a spent one).proclus-theology-of-plato-taylor-1816:r11899· transliteration
“ctpx” → “αρχαι”
the print sets GREEK here, which the scan read as a Latin lookalike: it reads "αρχαι" where the transcription has "ctpx" (page image, archive n21). 1.0.9 ADJUDICATION (2026-10-08). THIS RULE REPLACES r11842, WITHDRAWN WITH THIS VERSION (the id is spent, never reused; a corrected reading is a NEW id, not an edit beneath a spent one). THE READING THIS RULE SERVED WAS FABRICATED. It read “πραξις”; the print reads “αρχαι”, and the print’s own sentence on the same line fixes the sense: “For the supermundane Gods are called by Proclus in the sixth book of the following work αρχαι Principalities, or rulers, which is the word employed by Dionysius and Paul” — the Greek word is the one the print’s own English gloss “Principalities, or rulers” translates, and αρχη is “rule”. πραξις, “action”, is a one-letter misdecode of the same letters and translates nothing in that sentence. MEASURED by a BLIND read of the page image (the candidate text never shown to the model, google/gemini-2.5-flash, the whole page): the model returns “the supermundane Gods are called by Proclus in the sixth book of the fol- lowing work αρχαι Principalities, or rulers, which is the word employed by Dionysius and Paul”, and this unit’s own read of the plate gives the same word. THE EVIDENCE LEAF IS CORRECTED TOO: the passage stands on archive leaf n21 (printed page xxvi), NOT n22 — MEASURED by blind reads of both leaves, n22 carrying the Valerius footnote and no part of this sentence. ONE INSTRUMENT'S READING: these letters rest on ONE reader’s own draw, google/gemini-2.5-flash, so the review flag stands on it and the rule is listed in the edition's review worklist.proclus-theology-of-plato-taylor-1816:r11902· transliteration
“e$«rou<n, rofoi” → “ἐφέπουσι, σοφοί”
the print reads "ἐφέπουσι, σοφοί" here, where the transcription has "e$«rou<n, rofoi" (page image, archive n11). 1.0.11 CORRECTION (2026-10-08). THIS RULE REPLACES R11806, WITHDRAWN WITH THIS VERSION (the id is spent, never reused; a corrected reading is a NEW id, not an edit beneath a spent one). THE READING THIS RULE SERVED IS THE GREEK OF A DIFFERENT PRINTED LINE. It served "προχυν σειρά" — the Greek of the FIRST verse line of the NEXT page — where the transcription's "e$«rou<n, rofoi" is the scan of "ἐφέπουσι, σοφοί", the last verse line of THIS page. MEASURED (this unit, H2b-iii). The leaf a rule's `find` stands on is taken from the scanned item's OWN per-leaf text layer (thomastaylor _djvu.xml): this edition's transcription IS the first 28,444 non-empty lines of that layer IN ORDER (28,444 of 28,444 equal, 0 mismatches), so which leaf a line stands on is the layer's own page element and not an inference from a photograph. The rule's `after` was then read against the PRINTED LINE that line carries, on instruments independent of the rule: a BLIND full-page transcription of the plate by google/gemini-2.5-flash (prompt carries no candidate text of any kind), the human-proofread Project Gutenberg transcription of the same 1816 print, the two other scans of vol. I of that print (their own _djvu.xml), and the letter-for-letter correspondence between the rule's own `find` garble and the reading it claims. FOR THE PLATE READ ITSELF THIS IS ONE INSTRUMENT'S READING: these letters rest on ONE reader of the image, google/gemini-2.5-flash — no second reader's own draws of the plate corroborate them. The corroboration this version added is a DIFFERENT KIND of instrument, named above.proclus-theology-of-plato-taylor-1816:r11903· transliteration
“etfi^tfiuntges,” → “ἀμφιβατῆρες,”
the print reads "ἀμφιβατῆρες," here, where the transcription has "etfi^tfiuntges," (page image, archive n11). 1.0.11 CORRECTION (2026-10-08). THIS RULE REPLACES R11807, WITHDRAWN WITH THIS VERSION (the id is spent, never reused; a corrected reading is a NEW id, not an edit beneath a spent one). THE READING THIS RULE SERVED IS THE GREEK OF A DIFFERENT PRINTED LINE. It served "το, τε κυδήεν" — the second verse line of the NEXT page — where the transcription's "etfi^tfiuntges," is the scan of "ἀμφιβατῆρες,". MEASURED (this unit, H2b-iii). The leaf a rule's `find` stands on is taken from the scanned item's OWN per-leaf text layer (thomastaylor _djvu.xml): this edition's transcription IS the first 28,444 non-empty lines of that layer IN ORDER (28,444 of 28,444 equal, 0 mismatches), so which leaf a line stands on is the layer's own page element and not an inference from a photograph. The rule's `after` was then read against the PRINTED LINE that line carries, on instruments independent of the rule: a BLIND full-page transcription of the plate by google/gemini-2.5-flash (prompt carries no candidate text of any kind), the human-proofread Project Gutenberg transcription of the same 1816 print, the two other scans of vol. I of that print (their own _djvu.xml), and the letter-for-letter correspondence between the rule's own `find` garble and the reading it claims. FOR THE PLATE READ ITSELF THIS IS ONE INSTRUMENT'S READING: these letters rest on ONE reader of the image, google/gemini-2.5-flash — no second reader's own draws of the plate corroborate them. The corroboration this version added is a DIFFERENT KIND of instrument, named above.proclus-theology-of-plato-taylor-1816:r11905· transliteration
“yovof ygwwv,” → “γένος ἡρώων,”
the print reads "γένος ἡρώων," here, where the transcription has "yovof ygwwv," (page image, archive n12). 1.0.11 CORRECTION (2026-10-08). THIS RULE REPLACES R11816, WITHDRAWN WITH THIS VERSION (the id is spent, never reused; a corrected reading is a NEW id, not an edit beneath a spent one). THE READING THIS RULE SERVED IS THE GREEK OF THE NEXT PRINTED LINE. It served "εργα τα θνητων" (the following verse line) where the transcription's "yovof ygwwv," is the scan of "γένος ἡρώων,". MEASURED (this unit, H2b-iii). The leaf a rule's `find` stands on is taken from the scanned item's OWN per-leaf text layer (thomastaylor _djvu.xml): this edition's transcription IS the first 28,444 non-empty lines of that layer IN ORDER (28,444 of 28,444 equal, 0 mismatches), so which leaf a line stands on is the layer's own page element and not an inference from a photograph. The rule's `after` was then read against the PRINTED LINE that line carries, on instruments independent of the rule: a BLIND full-page transcription of the plate by google/gemini-2.5-flash (prompt carries no candidate text of any kind), the human-proofread Project Gutenberg transcription of the same 1816 print, the two other scans of vol. I of that print (their own _djvu.xml), and the letter-for-letter correspondence between the rule's own `find` garble and the reading it claims. FOR THE PLATE READ ITSELF THIS IS ONE INSTRUMENT'S READING: these letters rest on ONE reader of the image, google/gemini-2.5-flash — no second reader's own draws of the plate corroborate them. The corroboration this version added is a DIFFERENT KIND of instrument, named above.proclus-theology-of-plato-taylor-1816:r11906· transliteration
“xgufiatrt* o So if” → “κρυφίαισιν ὁδοῖς”
the print reads "κρυφίαισιν ὁδοῖς" here, where the transcription has "xgufiatrt* o So if" (page image, archive n12). 1.0.11 CORRECTION (2026-10-08). THIS RULE REPLACES R11798, WITHDRAWN WITH THIS VERSION (the id is spent, never reused; a corrected reading is a NEW id, not an edit beneath a spent one). THE READING THIS RULE SERVED IS THE GREEK OF THE NEXT PRINTED LINE. It served "διανισσομενον," (the following verse line) where the transcription's "xgufiatrt* o So if" is the scan of "κρυφίαισιν ὁδοῖς". MEASURED (this unit, H2b-iii). The leaf a rule's `find` stands on is taken from the scanned item's OWN per-leaf text layer (thomastaylor _djvu.xml): this edition's transcription IS the first 28,444 non-empty lines of that layer IN ORDER (28,444 of 28,444 equal, 0 mismatches), so which leaf a line stands on is the layer's own page element and not an inference from a photograph. The rule's `after` was then read against the PRINTED LINE that line carries, on instruments independent of the rule: a BLIND full-page transcription of the plate by google/gemini-2.5-flash (prompt carries no candidate text of any kind), the human-proofread Project Gutenberg transcription of the same 1816 print, the two other scans of vol. I of that print (their own _djvu.xml), and the letter-for-letter correspondence between the rule's own `find` garble and the reading it claims. FOR THE PLATE READ ITSELF THIS IS ONE INSTRUMENT'S READING: these letters rest on ONE reader of the image, google/gemini-2.5-flash — no second reader's own draws of the plate corroborate them. The corroboration this version added is a DIFFERENT KIND of instrument, named above.proclus-theology-of-plato-taylor-1816:r11907· transliteration
“*§ fJA\XVUVyn$” → “ἐς μελαναυγεῖς”
the print reads "ἐς μελαναυγεῖς" here, where the transcription has "*§ fJA\\XVUVyn$" (page image, archive n12). 1.0.11 CORRECTION (2026-10-08). THIS RULE REPLACES R11821, WITHDRAWN WITH THIS VERSION (the id is spent, never reused; a corrected reading is a NEW id, not an edit beneath a spent one). THE READING THIS RULE SERVED IS THE GREEK OF A DIFFERENT PRINTED LINE. It served "προχτει σειρα" — the FIRST verse line of the page, twenty lines earlier — where the transcription's "fJA\XVUVyn$" is the scan of "ἐς μελαναυγεῖς". The find is widened to include the "*§" that stands before it, which the old find left in the served text. MEASURED (this unit, H2b-iii). The leaf a rule's `find` stands on is taken from the scanned item's OWN per-leaf text layer (thomastaylor _djvu.xml): this edition's transcription IS the first 28,444 non-empty lines of that layer IN ORDER (28,444 of 28,444 equal, 0 mismatches), so which leaf a line stands on is the layer's own page element and not an inference from a photograph. The rule's `after` was then read against the PRINTED LINE that line carries, on instruments independent of the rule: a BLIND full-page transcription of the plate by google/gemini-2.5-flash (prompt carries no candidate text of any kind), the human-proofread Project Gutenberg transcription of the same 1816 print, the two other scans of vol. I of that print (their own _djvu.xml), and the letter-for-letter correspondence between the rule's own `find` garble and the reading it claims. FOR THE PLATE READ ITSELF THIS IS ONE INSTRUMENT'S READING: these letters rest on ONE reader of the image, google/gemini-2.5-flash — no second reader's own draws of the plate corroborate them. The corroboration this version added is a DIFFERENT KIND of instrument, named above.proclus-theology-of-plato-taylor-1816:r11909· transliteration
“xara njv irgov^xovrav $xuvy, it seems requisite to read x*t« ti)v xgomjxowav nfiaw” → “κατὰ τὴν προσήκουσαν ἐκείνῃ, it seems requisite to read κατὰ τὴν προσήκουσαν ἀξίαν”
the print reads "κατὰ τὴν προσήκουσαν ἐκείνῃ, it seems requisite to read κατὰ τὴν προσήκουσαν ἀξίαν" here, where the transcription has "xara njv irgov^xovrav $xuvy, it seems requisite to read x*t« ti)v xgomjxowav nfiaw" (page image, printed page 253 = archive v1-n328). 1.0.11 CORRECTION (2026-10-08). THIS RULE REPLACES R11698, WITHDRAWN WITH THIS VERSION (the id is spent, never reused; a corrected reading is a NEW id, not an edit beneath a spent one). THE READING THIS RULE SERVED IS NOT THE PRINT'S. It served "κατὰ τὴν προσηκούσας ἔχουσιν" and "φύσιν", neither of which the plate carries, where the footnote reads "κατὰ τὴν προσήκουσαν ἐκείνῃ, ... κατὰ τὴν προσήκουσαν ἀξίαν". MEASURED (this unit, H2b-iii). The leaf a rule's `find` stands on is taken from the scanned item's OWN per-leaf text layer (thomastaylor _djvu.xml): this edition's transcription IS the first 28,444 non-empty lines of that layer IN ORDER (28,444 of 28,444 equal, 0 mismatches), so which leaf a line stands on is the layer's own page element and not an inference from a photograph. The rule's `after` was then read against the PRINTED LINE that line carries, on instruments independent of the rule: a BLIND full-page transcription of the plate by google/gemini-2.5-flash (prompt carries no candidate text of any kind), the human-proofread Project Gutenberg transcription of the same 1816 print, the two other scans of vol. I of that print (their own _djvu.xml), and the letter-for-letter correspondence between the rule's own `find` garble and the reading it claims. FOR THE PLATE READ ITSELF THIS IS ONE INSTRUMENT'S READING: these letters rest on ONE reader of the image, google/gemini-2.5-flash — no second reader's own draws of the plate corroborate them. The corroboration this version added is a DIFFERENT KIND of instrument, named above.proclus-theology-of-plato-taylor-1816:r11910· transliteration
“9nrte ttwnv it aiuvarov, xat von oat yt ovln fMtteov paitor” → “θέντες· εἰπεῖν δὲ ἀδύνατον, καὶ νοῆσαι γε οὐδὲν μᾶλλον ῥᾴδιον”
the print reads "θέντες· εἰπεῖν δὲ ἀδύνατον, καὶ νοῆσαι γε οὐδὲν μᾶλλον ῥᾴδιον" here, where the transcription has "9nrte ttwnv it aiuvarov, xat von oat yt ovln fMtteov paitor" (page image, archive n29). 1.0.11 CORRECTION (2026-10-08). THIS RULE REPLACES R11713, WITHDRAWN WITH THIS VERSION (the id is spent, never reused; a corrected reading is a NEW id, not an edit beneath a spent one). THE READING THIS RULE SERVED IS NOT THE PRINT'S. It served "ζητοῦσι πάντων αἰωνίου, καὶ νοῦ σοφίας οὐδὲν βέλτιον", a sentence the plate does not carry at this place; the line reads "θέντες· εἰπεῖν δὲ ἀδύνατον, καὶ νοῆσαι γε οὐδὲν μᾶλλον ῥᾴδιον" — the close of the Damascius footnote. The find is widened by the "9" that stands before it, which the old find left in the served text. MEASURED (this unit, H2b-iii). The leaf a rule's `find` stands on is taken from the scanned item's OWN per-leaf text layer (thomastaylor _djvu.xml): this edition's transcription IS the first 28,444 non-empty lines of that layer IN ORDER (28,444 of 28,444 equal, 0 mismatches), so which leaf a line stands on is the layer's own page element and not an inference from a photograph. The rule's `after` was then read against the PRINTED LINE that line carries, on instruments independent of the rule: a BLIND full-page transcription of the plate by google/gemini-2.5-flash (prompt carries no candidate text of any kind), the human-proofread Project Gutenberg transcription of the same 1816 print, the two other scans of vol. I of that print (their own _djvu.xml), and the letter-for-letter correspondence between the rule's own `find` garble and the reading it claims. FOR THE PLATE READ ITSELF THIS IS ONE INSTRUMENT'S READING: these letters rest on ONE reader of the image, google/gemini-2.5-flash — no second reader's own draws of the plate corroborate them. The corroboration this version added is a DIFFERENT KIND of instrument, named above.proclus-theology-of-plato-taylor-1816:r11911· transliteration
“wrt t» ayaXuaT* xpooxwuv *0»xw; v, «xx» nin «t/rot* Toi/f” → “οὔτε τὰ ἀγάλματα προσκυνεῖν ἐθέλων, ἀλλ᾽ ἤδη ἐπ᾽ αὐτοὺς τοὺς”
the print reads "οὔτε τὰ ἀγάλματα προσκυνεῖν ἐθέλων, ἀλλ᾽ ἤδη ἐπ᾽ αὐτοὺς τοὺς" here, where the transcription has "wrt t» ayaXuaT* xpooxwuv *0»xw; v, «xx» nin «t/rot* Toi/f" (page image, archive n29). 1.0.11 CORRECTION (2026-10-08). THIS RULE REPLACES R11714, WITHDRAWN WITH THIS VERSION (the id is spent, never reused; a corrected reading is a NEW id, not an edit beneath a spent one). THE READING THIS RULE SERVED IS NOT THE PRINT'S AT THREE POINTS. It served "φησὶ" for the plate's "οὔτε", an "οὐκ" the plate does not set, and "ἤτοι" for the plate's "ἤδη". MEASURED (this unit, H2b-iii). The leaf a rule's `find` stands on is taken from the scanned item's OWN per-leaf text layer (thomastaylor _djvu.xml): this edition's transcription IS the first 28,444 non-empty lines of that layer IN ORDER (28,444 of 28,444 equal, 0 mismatches), so which leaf a line stands on is the layer's own page element and not an inference from a photograph. The rule's `after` was then read against the PRINTED LINE that line carries, on instruments independent of the rule: a BLIND full-page transcription of the plate by google/gemini-2.5-flash (prompt carries no candidate text of any kind), the human-proofread Project Gutenberg transcription of the same 1816 print, the two other scans of vol. I of that print (their own _djvu.xml), and the letter-for-letter correspondence between the rule's own `find` garble and the reading it claims. FOR THE PLATE READ ITSELF THIS IS ONE INSTRUMENT'S READING: these letters rest on ONE reader of the image, google/gemini-2.5-flash — no second reader's own draws of the plate corroborate them. The corroboration this version added is a DIFFERENT KIND of instrument, named above.proclus-theology-of-plato-taylor-1816:r11912· transliteration
“Oi iroLgoL xkeivovg” → “οἱ παρὰ κλεινοὺς”
the print reads "οἱ παρὰ κλεινοὺς" here, where the transcription has "Oi iroLgoL xkeivovg" (page image, archive n11). 1.0.12 CORRECTION (2026-10-08). THIS RULE REPLACES R11904, WITHDRAWN WITH THIS VERSION (the id is spent, never reused; a corrected reading is a NEW id, not an edit beneath a spent one). R11904 served THESE SAME LETTERS with an ACUTE on the ultima ("οἱ παρὰ κλεινούς"). MEASURED: that acute has no instrument anywhere in the record — the rule states no witness for the mark, and the unit that wrote it recorded that its own draw was truncated before this word ("οἱ πᾰρᾱ κλεινο..."), so the mark was written without a plate read of the word it sits on. THE GRAVE is what the plate gives, and the print's own convention agrees. THE PLATE. Read blind at the item's maximum resolution (archive.org page/n11_w3000.jpg, 5367x7017 — four times the stored n11.jpg), google/gemini-2.5-flash returned "οἱ παρὰ κλεινοὺς" — GRAVE — in 3 of 4 independent draws (the fourth read no marks at all), and Qwen/Qwen3-VL-235B-A22B returned "οἱ παρὰ κλεινοὺς" in 2 of 2. The 955 review's own two draws of the stored leaf returned the grave on both. NO DRAW OF ANY INSTRUMENT HAS RETURNED THE ACUTE AT THIS WORD. THE PRINT'S OWN CONVENTION, MEASURED ON THIS PAGE. A final acute is set as a GRAVE before a following word. On the two plates of this hymn the same construction is read with the grave wherever the phrase continues — οἱ περὶ κόσμον, οἱ κατὰ κόσμον, ἐφέπουσι, σοφοὶ, οὓς ἀγγελικὰ — while a line-final word before no following word keeps its acute (κόσμου μοίρας). The word here is followed by οἰηκοφόρους, the phrase running on with no punctuation between them, which is exactly the position the convention graves. AND INDEPENDENTLY OF BOTH, THE RECEIVED TEXT: the hymn Taylor quotes is Synesius, Hymn 3, whose received text prints "οἱ παρὰ κλεινοὺς / οἰηκοφόρους· / οὓς ἀγγελικὰ / προχέει σειρά·" — the grave. It is cited here as an outside authority on the reading, not as the print: the print is the plate. THE PLATE, READ BLIND. The stored leaf is 1342x1755; the item serves the same leaf at FOUR times that (archive.org page/n12_w3000.jpg, 5367x7017), and the higher resolution is the instrument used here — no candidate text entered any prompt. Because the endpoint is deterministic for an identical request, every draw carries its own nonce and temperature 0.8, so the draws are independent: google/gemini-2.5-flash was drawn four times on the band and Qwen/Qwen3-VL-235B-A22B-Instruct twice, and the draws differ from each other elsewhere on the band, which is how a reader can see they are not one draw served repeatedly. The instruments that CANNOT read this italic Greek are recorded rather than counted: anthropic/claude-haiku-4-5 returns nonsense on these leaves (its n11/n12 draws are not Greek words), google/gemini-2.5-pro split on the line (one draw an acute, one no mark at all, and both mangled a neighbouring word), and DeepSeek-v4-flash-vision returned "κλεῖους"/"χλεινους" with a circumflex or nothing on the n11 line. PG HERE MEANS the human-proofread Project Gutenberg transcription of the same 1816 print (eBooks 77393 vol. I / 78800 vol. 2), which this edition holds as a witness and which is named in this RATIONALE — the `witness` field of a rule names the PLATE ("page image, archive nNN"), never the transcription. FOR THE PLATE READ ITSELF THIS IS ONE INSTRUMENT'S READING: the letters served here rest on ONE reader of the image, google/gemini-2.5-flash, which is the instrument this edition's own Greek pass CLASSED for this line — the sentence the footing is stated in, and the field `instruments` carries the same one model. The further draws named above (Qwen/Qwen3-VL-235B-A22B, and the 955 review's own two) are recorded here as an ADDITIONAL check of the mark, not as a classed instrument of this edition: they were taken after the pass and are not in its draw cache.
On the Cave of the Nymphs — Porphyry, tr. Thomas Taylor
None. Every reading in this version is corroborated by a second reader, or rests on a witness that is not a single instrument.
Leaves cited but not held
The Elements of Theology — Proclus, tr. Thomas Taylor
None. Every page image this edition’s rules cite is held with the edition.
On the Theology of Plato — Proclus, tr. Thomas Taylor
None. Every page image this edition’s rules cite is held with the edition.
On the Cave of the Nymphs — Porphyry, tr. Thomas Taylor
None. Every page image this edition’s rules cite is held with the edition.
The editions, and where their apparatus is
The Elements of Theology — Proclus, tr. Thomas Taylor
The version served here. 1.0.1 — 415 repair rules in the version, 105 of them read off a page image; 141 leaf images are stored with the edition. repaired — finished: every damaged word and every standalone marker the scanner left has been cleared against the printed page. The reading view shows no damage character at all — every repair the print called for is applied against the 1816 print, its Greek restored from the page images where the letters had been read as Latin lookalikes. One note stands as the transcription has it, the footnote to Proposition XXVII on printed page 321, because the page itself will not read: repeated reads of that one line disagree, so it is left VISIBLE rather than guessed. A second whole-text read has been run over the whole volume.
The whole record, and the page images. Every one of the 415 rules is in the edition’s apparatus at /texts/proclus-elements-of-theology-taylor-1816/apparatus/, a page of its own, browsable BOTH ways: the edition’s whole scan forms an index in leaf order, and selecting any leaf opens its page image together with the readings decided from it — each with its id, its type, the words it changes, its reason and the witness it rested on, anchored at #repair-<id> so a single reading can be cited — while the readings that carry no page image (310 of 415 here) are listed in their own right, and a list of all 415 readings is paged 25 at a time for a reader who has no leaf in mind. The whole log is served with the edition as data at /texts/proclus-elements-of-theology-taylor-1816/apparatus.json, so it can be read, filtered and searched without the page. That is where the scans are: the page images are stored with the edition and served beside its text, one image per leaf, at /texts/proclus-elements-of-theology-taylor-1816/scans/ — its whole scan, 141 leaf images for this edition.
On the Theology of Plato — Proclus, tr. Thomas Taylor
The version served here. 1.0.12 — 3255 repair rules in the version, 203 of them read off a page image; 800 leaf images are stored with the edition. in repair — this text is READABLE BUT STILL GARBLED. The transcription it is built from is a good reading of the 1816 print, not a faithful one, and where nothing could decide what the print actually says, the reading view shows the transcription’s own wrong characters, marked as damage — so a reader will meet passages that read as garble, and this page does not pretend otherwise. The whole text has been read, and the witnesses were the second transcription of the same 1816 print that stands beside this one on the shelf, together with the other copies of that same print held here — their own scans, their own page images, their own reading of the page. Every reading those could settle is recorded as a rule and applied by the reading view: each rule gives the words it changes, why, and the witness it was decided from, and where two copies of the print read the same words at the same line, the rule says that, because two independent readings of the same print agreeing is what makes the reading worth trusting; where the witness was a copy’s own page rather than this edition’s, the rule names that copy and the leaf of it the words stand on. What could not be settled is left standing and named, not hidden: the questions still open keep their reasons — a reading nothing places on the line it was asked about, a page line two answers came back for, a word the page itself will not read — and the reading view shows the transcription’s characters there rather than a guess. The edition is therefore not finished and not clean: what was settled is settled and recorded, what was not is visible, and the work of settling the rest stands open.
The whole record, and the page images. Every one of the 3255 rules is in the edition’s apparatus at /texts/proclus-theology-of-plato-taylor-1816/apparatus/, a page of its own, browsable BOTH ways: the edition’s whole scan forms an index in leaf order, and selecting any leaf opens its page image together with the readings decided from it — each with its id, its type, the words it changes, its reason and the witness it rested on, anchored at #repair-<id> so a single reading can be cited — while the readings that carry no page image (3052 of 3255 here) are listed in their own right, and a list of all 3255 readings is paged 25 at a time for a reader who has no leaf in mind. The whole log is served with the edition as data at /texts/proclus-theology-of-plato-taylor-1816/apparatus.json, so it can be read, filtered and searched without the page. That is where the scans are: the page images are stored with the edition and served beside its text, one image per leaf, at /texts/proclus-theology-of-plato-taylor-1816/scans/ — its whole scan, 800 leaf images for this edition.
On the Cave of the Nymphs — Porphyry, tr. Thomas Taylor
The version served here. 1.0.1 — 384 repair rules in the version, 4 of them read off a page image; 72 leaf images are stored with the edition. repaired — every damaged place a witness could settle has been settled — against the transcription’s own context, the 1823 parallel of the same translation, or the printed page itself, read one page at a time. What remains is not the book’s text: the page furniture at one page break (a recorded decision to leave it visible) and this copy’s own marks. A second whole-text read has not been run, so the claim is that the found damage is settled, not that nothing further is findable.
The whole record, and the page images. Every one of the 384 rules is in the edition’s apparatus at /texts/porphyry-on-the-cave-of-the-nymphs-taylor-1917/apparatus/, a page of its own, browsable BOTH ways: the edition’s whole scan forms an index in leaf order, and selecting any leaf opens its page image together with the readings decided from it — each with its id, its type, the words it changes, its reason and the witness it rested on, anchored at #repair-<id> so a single reading can be cited — while the readings that carry no page image (380 of 384 here) are listed in their own right, and a list of all 384 readings is paged 25 at a time for a reader who has no leaf in mind. The whole log is served with the edition as data at /texts/porphyry-on-the-cave-of-the-nymphs-taylor-1917/apparatus.json, so it can be read, filtered and searched without the page. That is where the scans are: the page images are stored with the edition and served beside its text, one image per leaf, at /texts/porphyry-on-the-cave-of-the-nymphs-taylor-1917/scans/ — its whole scan, 72 leaf images for this edition.
How to report an error
How to report an error. This is a static site: there is no form and no server to receive one, so a report is an email. The library records no address for reports yet, so this page names none rather than invent one; when the record carries one it will be printed here. What a report carries is the same either way — name the edition, the repair id above or the section and printed page, and what the print or the page image shows instead. A correction that names a witness is the kind this library is built from — the transcription’s own context, a parallel edition, or the page image — and a report that names one is acted on as a new rule in a new version, never as a silent edit to the served text.