Mashrabaمشربة

About this text: Alcibiades I: Know Yourself

Ἀλκιβιάδης α

Plato (c. 428–348 BCE) · Alcibiades I, 103a–135e

The whole Alcibiades I, 103a–135e (whether Plato wrote it is disputed): Socrates' approach to the young Alcibiades, his double ignorance, what the self is that we are told to know, the soul as the user of the body, and the eye that sees itself only in another eye; with 133c8–17, absent from the medieval manuscripts, marked, and the ancient commentaries of Proclus and Olympiodorus beside the text.

162 sections · ~175 min · commentaries · published translation · 10,259 words

Read it →

Versions

28 layers, each with its licence

The original

Burnet's Oxford text (Platonis Opera II)John Burnet · 1901Public domain
By
John Burnet (editor) · 1901
Source:
J. Burnet, Platonis Opera II (Oxford: Clarendon Press, Oxford Classical Texts, 1901) · link ↗
In the reader:
In full, the whole dialogue, 103a–135e, with the speakers; 133c8–17, which the medieval manuscripts lack, between ⟨ ⟩ as Burnet prints them (supplied from Eusebius)
Licence:
Text public domain
Notes
Burnet's apparatus at 133c: 'c 8–17 Ἆρʼ οὖν … Ναί Eusebius : om. B T (sed talia fere legerunt Iulianus Stobaeus)' (scan page n362–n363). Denyer (Cambridge 2001) is in copyright, not held, and was not used.
Perseus Digital Library digitisation of BurnetPerseus Digital Library · UNVERIFIEDOpen licence
By
Perseus Digital Library (digitiser (canonical-greekLit)) · UNVERIFIED
Source:
Perseus Digital Library, canonical-greekLit tlg0059.tlg013.perseus-grc2 (EpiDoc; Stephanus milestones) · link ↗
In the reader:
The Greek as shown, with its paragraphing and speakers
Licence:
CC BY-SA 4.0 (the digitisation; Burnet's text is public domain). Share-alike: this page and the derived data files carry the same licence
Notes
Speakers and an <add> for 133c8–17 encoded.
UNVERIFIED:
The date (release or commit) of the Perseus canonical-greekLit file used is not recorded in the repo.
Eusebius, Praeparatio evangelica XI.27.5 (the source of 133c8–17)Eusebius of Caesarea, W. Dindorf, First1KGreek · 1867Public domain
By
Eusebius of Caesarea (author); W. Dindorf (editor (1867)); First1KGreek (digitiser) · 1867
Source:
Eusebius, Praeparatio evangelica XI.27.5 (Viger p. 551b–c), ed. Dindorf (Leipzig 1867), vol. 2, pp. 56–57; First1KGreek tlg2018.tlg001.1st1K-grc1 · link ↗
In the reader:
Consulted (2 Oct 2026): quotes 133c1–17 continuously, with the lines where Burnet prints them; not shown
Licence:
Text public domain; First1KGreek digitisation CC BY-SA 4.0
Notes
Also checked in Gifford's translation (Oxford 1903, vol. III.2, pp. 595–596, the lines in round brackets; note at 551 b 8: 'The passage in brackets is not in the MSS. of Plato'), archive.org eusebius-gifford-1903-preparation-for-the-gospel_202311. Mras's edition (GCS 43.2, 1956) was not consulted (lending-only scans).
Stobaeus, Anthologium III.21.24 (a second witness to 133c8–17)Stobaeus, O. Hense · 1894Public domain
By
Stobaeus (anthologist); O. Hense (editor) · 1894
Source:
Stobaeus, Anthologium III.21.24 (ch. 'Περὶ τοῦ γνῶθι σαυτόν'), ed. O. Hense, Wachsmuth–Hense vol. III (Berlin 1894), pp. 573–577, the lines at p. 576.4–11 · link ↗
In the reader:
Consulted (2 Oct 2026); not shown. Taylor (1804) follows its placement
Licence:
Public domain
Notes
Excerpts 132b–134b and has a version of the lines at a different point: after 133c18–19, opening 'Ἆρʼ ὥσπερ κάτοπτρα' (no οὖν, ὅθʼ), with 'καθαρώτερά τε', and with 133c18–19 then repeated. Julian, named by Burnet, was not checked (locus not given).

Our edition

Our Greek: GLAUx tokens with Perseus's speakersscripts/plato/build_greek.py · 2026-10-02Open licence
By
scripts/plato/build_greek.py (build) · 2026-10-02
Source:
scripts/plato/build_greek.py alcibiades-1-mirror; method in docs/texts/alcibiades-1-mirror.md §§3–5
In the reader:
162 passages, one per Stephanus section (#s-103a … #s-135e); speakers (ΣΩ., ΑΛ.) at each turn; 133c8–17 marked ⟨ ⟩
Licence:
CC BY-SA 4.0 (share-alike, as derived from Perseus and GLAUx)
Notes
Tokens walked back through the Perseus text for its 903 speaker turns and its <add> in 133c. GLAUx's artificial SPEAKER= and elliptic E tokens dropped. The reader note on 133c8–17 (checked on the page images, 2 Oct 2026): These lines are not in Plato's medieval manuscripts. Burnet prints them, in angle brackets, from Eusebius (Praeparatio evangelica XI.27.5), and notes that Julian and Stobaeus read something like them; Stobaeus (Anthologium III.21.24) gives a version of them a little later in the conversation, after "we agreed that knowing oneself is self-control". The reader marks them ⟨ ⟩ as Burnet does: words supplied from an indirect witness, not an editor's deletion. Neither Jowett nor Lamb translates them; Taylor (1804) does, in square brackets, from Stobaeus and at Stobaeus' place in the conversation (see the Sydenham–Taylor layer). The English below is AI.
Corrections to the automatic analysisClaude (AI) · 2026-10-02Our work
By
Claude (AI) (found by matching every lemma against LSJ and reading the misses) · 2026-10-02
Source:
app/src/data/alcibiades-1-mirror.json (`fixes`), listed on the About page with links to each word
In the reader:
17 corrections, each with its reason, on the About page; the corrected lemma in the word panel
Licence:
Ours; marked AI
Made by:
Claude (Opus; exact version not recorded), 2026-10-02 (first commit of app/src/data/alcibiades-1-mirror.json)
Notes
Made by AI in this edition and open to review; the Greek text itself is unchanged. GLAUx's reported accuracy is about 97% for morphology and 99% for lemmas, so more errors should be expected. Some were found by the translation triangulation (§7b of the text record): 104b γένεσιν, 106c παριέναι, 106c ὅτι, 113b ἦ, 119c περιέσομαι, 123c ἐξῃρημένους, 126e ἐπισταμένῃ, 128d ἄλλῃ, 129d ἀρτίως. Two lemmas called errors were checked and kept (131d ἄπεισιν, 133c ὅθʼ).

Published translations

Sydenham–Taylor (1804): The First AlcibiadesFloyer Sydenham, Thomas Taylor · 1804Public domain
By
Floyer Sydenham (translator (1767–80)); Thomas Taylor (reviser and annotator (1804)) · 1804
Source:
The Works of Plato, viz. His Fifty-Five Dialogues and Twelve Epistles, translated from the Greek; nine of the dialogues by the late Floyer Sydenham, and the remainder by Thomas Taylor (London: Thomas Taylor, 1804), vol. 1, pp. 16–96. Taylor's notes signed T., Sydenham's S. · link ↗
In the reader:
In full, by section, beside the Greek or in columns (Translations and Compare in the reader); only Taylor translates 133c8–17
Licence:
Public domain: published 1804; Sydenham died 1787, Taylor 1835.
Notes
Date: Sydenham's translation (first published 1767-80) as revised and annotated by Taylor, 1804. Text: Transcribed by AI (Claude) from two archive.org OCRs of the 1804 printing (in.ernet.dli.2015.39263 and thwworksofplatov0001thom), page by page; long s modernised, 1804 spelling and punctuation kept. Not yet checked word by word against the page images. Sections: Automatic and approximate (a clause either way), as for Jowett: each clause is placed in a section by a monotonic best path over its word overlap with Lamb's Loeb English and Jowett's English of that section. Note: Taylor alone of our translators prints 133c8–17, in square brackets: his note names Stobaeus (Serm. 21, p. 183) as the source, and he places them where Stobaeus does, after 'to know oneself we acknowledge to be wisdom' (133c18–19). Notes kept apart (see notes). Taylor's (T.) and Sydenham's (S.) notes are kept apart (91 notes; Taylor's on 133b–c quote Proclus, Platonic Theology I.3); 180 places the transcription could not settle are listed in the file (uncertain), mostly punctuation. Pages 23, 38, 50, 58, 68, 69 were not located in the alignment, so their notes' sections are approximate.
Jowett (1892): Alcibiades IBenjamin Jowett · 1892Public domain
By
Benjamin Jowett (translator) · 1892
Source:
The Dialogues of Plato, translated into English with analyses and introductions by B. Jowett, 3rd ed. (Oxford: Clarendon Press, 1892), vol. 2 · link ↗
In the reader:
In full, by section, beside the Greek or in columns (Translations and Compare in the reader); only Taylor translates 133c8–17
Licence:
Public domain: published 1892; Jowett died 1893. Project Gutenberg text: not copyrighted in the US.
Notes
Date: 3rd edition, 1892. Text: Project Gutenberg #1676 (Jowett's translation), compared word by word with the archive.org OCR of the 1892 volume 2; 3 Gutenberg readings restored to the 1892 wording. Sections: Automatic and approximate (a clause either way): Jowett's margins give Stephanus pages but not the letters, so each clause is placed in a section by a monotonic best path over its word overlap with Lamb's Loeb English (1927) of that section. Where his wording seems to straddle two sections, read the neighbour too. Jowett does not translate 133c8–17 (absent from the medieval manuscripts); his speaker names are printed as in Gutenberg. Note: Divided into sections automatically; approximate. In Jowett's vol. 2, Appendix I, pp. 457 ff. (apart from the dialogues he took to be Plato's).
Lamb (Loeb 1927): Alcibiades IW. R. M. Lamb · 1927PD-US
By
W. R. M. Lamb (translator) · 1927
Source:
Plato in Twelve Volumes, vol. 8: Charmides, Alcibiades I and II, Hipparchus, The Lovers, Theages, Minos, Epinomis, translated by W. R. M. Lamb (Loeb Classical Library; London: Heinemann / Cambridge, Mass.: Harvard University Press, 1927) · link ↗
In the reader:
In full, by section, beside the Greek or in columns (Translations and Compare in the reader); only Taylor translates 133c8–17
Licence:
Public domain in the US (published 1927; the US term of 95 years ended on 31 December 2022). Perseus digitisation CC BY-SA 4.0. Lamb died 1946: in copyright in life+70 countries until the end of 2016, so public domain there too. (UNVERIFIED outside the US: the repo also gives Lamb's death as 1961)
Notes
Date: published 1927. Text: Perseus Digital Library, canonical-greekLit tlg0059.tlg013.perseus-eng2 (TEI, sectioned by Stephanus number). Sections: By the Stephanus sections Perseus marks in the translation (placed at the nearest sentence, so a sentence either way). Note: Loeb footnotes omitted. Lamb, like Jowett, does not translate 133c8–17.
UNVERIFIED:
Lamb's death year: the Alcibiades translation file gives 1946, the Symposium About page gives 1961; the repo does not settle it, so whether Lamb is public domain outside the US is unverified.

AI layers

English of 133c8–17 (AI)Claude (AI) · 2026-10-02Our work
By
Claude (AI) (translation) · 2026-10-02
In the reader:
Under section 133c, in the note on the lines the medieval manuscripts lack
Licence:
Ours; marked AI
Made by:
Claude (Opus; exact version not recorded), 2026-10-02 (first commit of app/src/data/alcibiades-1-mirror.json (insertion.en))
Notes
An AI translation from Burnet's Greek of the 71 tokens neither Jowett nor Lamb translates; checked against Taylor's English (no contradiction). There is no AI translation of the rest of the dialogue.
Word glossesClaude (AI) · 2026-10-02Our work
By
Claude (AI) (glosses) · 2026-10-02
In the reader:
One gloss per lemma (1,072 lemmas)
Licence:
Ours; marked AI
Made by:
Claude (Opus; exact version not recorded), 2026-10-02 (first commit of app/src/content/ai/alcibiades-1-mirror.json (338 new glosses in scripts/plato/ai/greek4-glosses-5.tsv, by headless Claude))
Notes
Shared by lemma with the other Greek texts; 338 written for 103a–127e on 2 October 2026 from LSJ's first senses and one occurrence each. The glosses flagged by the triangulation's AI-English check (15 in the first run, 34 + 4 in the second) were fixed in our own words; a third run of the affected sections returned no flags.
Word familiesClaude (AI) · 2026-09-28Our work
By
Claude (AI) (word-families) · 2026-09-28
In the reader:
In the word panel: stems grouped from the etymologies
Licence:
Ours; marked AI
Made by:
Claude (Opus; exact version not recorded), 2026-09-28 (the families were made for the Symposium (first commit of app/src/content/ai/symposium-diotima.json); this text shows those of its lemmas)
Notes
AI groupings of lemmas by stem (e.g. ἐρ-: ἔρως, ἐράω, ἐραστής …), made from the etymologies for the Symposium; each text shows the 24 Symposium stems filtered to its own lemmas.
Translators differ (triangulation)Claude (AI) · 2026-10-02Our work
By
Claude (AI) (triangulation) · 2026-10-02
In the reader:
The serious divergences (63 in 54 sections) as 'Translators differ' notes under the section
Licence:
Ours; marked AI
Made by:
Claude (Opus; exact version not recorded), 2026-10-02 (generatedOn in app/src/content/triangulation/alcibiades-1-mirror.json; scripts/triangulate.py (PROTOCOLS §17), run twice and --redo/--revalidate)
Notes
Compared: Jowett (1892), Lamb (1927), Sydenham–Taylor (1804) on all 162 sections; commentaries weighted by lineage: Olympiodorus, Proclus, Ficino's argumentum, the curriculum entries, Taylor's and Sydenham's notes. 257 divergences (35 interpretation, 12 term of art, 11 syntax, 5 variant among the serious). No copyrighted translation is held (no O'Neill, Hutchinson, Denyer), so none was compared privately. Descriptions are ours (AI); quotations are the translators' own words, cited.
The commentators' Greek, corrected by AI from the page imagesClaude (AI) · 2026-10-02Our work
By
Claude (AI) (corrected-ocr) · 2026-10-02
In the reader:
The Greek of each Olympiodorus and Proclus comment, with its uncertain readings listed
Licence:
Ours; marked AI
Made by:
Claude (Opus; exact version not recorded), 2026-10-02 (first commit of app/public/data/plato/commentary/alcibiades-1-mirror-olympiodorus.json and -proclus.json; scripts/plato/ocr_creuzer.py)
Notes
For each page of Creuzer's print, Tesseract (grc) and archive.org's OCR as witnesses, and Claude (headless) reading the page image in two strips; Greek only (Creuzer's Latin notes left out), lemmata marked, ϑ written θ. Olympiodorus: the whole commentary, 1,193 readings the model could not settle listed; spot checks by eye (pp. 10, 11, 84, 192, 198) found no wrong words beyond those. Proclus: 134 lemma pages, 578 uncertain readings listed (234 in the excerpts). Not yet proofread page by page.
English of the readers through the ages (AI)Claude (AI) · 2026-10-02Our work
By
Claude (AI) (translation) · 2026-10-02
In the reader:
Beside each commentary entry, marked AI
Licence:
Ours; marked AI
Made by:
Claude (Opus; exact version not recorded), 2026-10-02 (first commit of app/src/content/plato/alcibiades-1-mirror-braid.json (scripts/plato/ai/alcibiades-1-mirror-braid-*.json))
Notes
Olympiodorus per lecture, Proclus per excerpt, Ficino's argumentum, the Prolegomena and the Arabic bibliographers' notices: all English is AI (Claude).

Automatic analysis

Lemmas and morphology (GLAUx)Alek Keersmaekers (KU Leuven) · UNVERIFIEDOpen licence
By
Alek Keersmaekers (KU Leuven) (GLAUx corpus: automatic annotation) · UNVERIFIED
Source:
GLAUx corpus, document 0059-013 (A. Keersmaekers, KU Leuven): automatic lemma, morphology and syntax · link ↗
In the reader:
Per word, marked as automatic; corrections listed on the About page
Licence:
CC BY-SA 4.0
Notes
Built on the same Perseus text, so its tokens carry their Stephanus sections. Reported accuracy: morphology 97.2%, lemmas 98.8%.
UNVERIFIED:
The GLAUx release or commit used is not recorded in the repo.
Liddell–Scott–Jones (dictionary)Liddell–Scott–Jones, Perseus Digital Library, H. Dik (Logeion) · 1940Open licence
By
Liddell–Scott–Jones (dictionary, 9th ed. (1940)); Perseus Digital Library (digitisation); H. Dik (Logeion) (edited digitisation) · 1940
Source:
Liddell–Scott–Jones, A Greek–English Lexicon, 9th ed. (1940), Perseus digitisation as edited for Logeion (H. Dik) · link ↗
In the reader:
The entry (first 1,500 characters), its opening senses, and where LSJ cites this passage; 1,020 of 1,072 lemmas matched (the misses mostly proper names and a few particles LSJ files elsewhere)
Licence:
CC BY-SA 4.0 (the Perseus/Logeion digitisation)
Notes
Matched by headword, as for the Symposium.
LSJ's own derivationsscripts/plato/lsj_etym.py · 1940Open licence
By
scripts/plato/lsj_etym.py (automatic extraction from the LSJ entry) · 1940
Source:
Liddell–Scott–Jones, 9th ed. (1940): the derivation given after the headword and occasional cognate notes, extracted automatically (scripts/plato/lsj_etym.py) · link ↗
In the reader:
Where LSJ gives one (about 280 lemmas)
Licence:
CC BY-SA 4.0
Notes
LSJ's <etym> tags were checked and are too inconsistent to use, so the derivation is taken from the text of the entry.
Boisacq, Dictionnaire étymologique (the primary etymology)Émile Boisacq, Tesseract (tessdata_best grc+fra), scripts/plato/boisacq_index.py · 1916Public domain
By
Émile Boisacq (author); Tesseract (tessdata_best grc+fra) (OCR); scripts/plato/boisacq_index.py (automatic matching) · 1916
Source:
É. Boisacq, Dictionnaire étymologique de la langue grecque (Heidelberg/Paris 1916), archive.org dictionnairety00bois: OCR of the page scan (Tesseract, Ancient Greek + French models, 300 dpi), matched to the lemmas automatically · link ↗
In the reader:
The entry as machine-read, marked unchecked, with the page image one tap away (the image is authoritative); 505 lemmas found
Licence:
Public domain
Notes
Method in docs/texts/symposium-diotima.md §8: entries found by the dictionary's alphabetical order, then matched accents aside (the lemma's own entry, a line or paragraph it opens, the entry that treats it, or, when LSJ derives the lemma from a base word, Boisacq's entry for that base, labelled as such). Printed page = PDF page − 34. Known weaknesses: OCR noise in small Greek; occasionally the matched paragraph belongs to a neighbouring entry.
Wiktionary etymologies (via kaikki.org)Wiktionary contributors, kaikki.org (Wiktextract), scripts/plato/fetch_wiktionary.py · UNVERIFIEDOpen licence
By
Wiktionary contributors (etymologies); kaikki.org (Wiktextract) (extraction); scripts/plato/fetch_wiktionary.py (lookup) · UNVERIFIED
Source:
Wiktionary, Ancient Greek entries, via kaikki.org (Wiktextract) JSONL · link ↗
In the reader:
Secondary, labelled: the etymology and Proto-Indo-European forms, credited and linked (collapsed when Boisacq has the word); 908 lemmas have an entry
Licence:
CC BY-SA 4.0 and GFDL (kaikki states no licence of its own; Wiktionary's is inferred)
UNVERIFIED:
The date on which the kaikki.org extracts were fetched is not recorded in the repo.
Beekes, Chantraine and Frisk (cited only)R. Beekes, P. Chantraine, H. Frisk · UNVERIFIEDCopyrighted: cited only
By
R. Beekes (Etymological Dictionary of Greek (2010)); P. Chantraine (Dictionnaire étymologique de la langue grecque (DELG)); H. Frisk (Griechisches etymologisches Wörterbuch (GEW)) · UNVERIFIED
Source:
Beekes, Etymological Dictionary of Greek (2010); Chantraine, DELG; Frisk, GEW
In the reader:
A citation s.v. in the word panel
Licence:
In copyright: cited only ("see Beekes s.v. …"), never quoted
UNVERIFIED:
Publication years of Chantraine's DELG and Frisk's GEW are not given in the repo (Beekes: 2010).

Commentaries

Proclus, In Alcibiadem (to 116b)Proclus, F. Creuzer · 5th c.; ed. 1820Public domain
By
Proclus (author); F. Creuzer (editor) · 5th c.; ed. 1820
Source:
Proclus, In Platonis Alcibiadem priorem commentarii, ed. F. Creuzer, Initia philosophiae ac theologiae ex Platonicis fontibus ducta, vol. 1 (Frankfurt: Brönner, 1820) · link ↗
In the reader:
An excerpt per lemma (the lemma and the opening of the comment, about 1,800 characters): 65 excerpts, 103a–116a, plus the prologue's opening (p. 1) and its passage on Iamblichus' course (p. 11), under each section with the commentator's lineage
Licence:
Public domain (Creuzer's edition, 1820). Greek corrected by AI (Claude) from the archive.org scan against the page images, with Tesseract and archive.org's OCR as witnesses; uncertain readings listed
Notes
Greek: Creuzer's text corrected by AI against the page images (scripts/plato/ocr_creuzer.py); typographic ϑ written θ; Creuzer's footnotes left out. English: AI (Claude), marked. Proclus spends some 340 pages on 103a–116b; the lemma pages were found by matching Burnet's Greek against archive.org's hOCR. Westerink (1954) and Segonds (1985–86) are in copyright and were not used; O'Neill's English (1965) is not held.
Olympiodorus, In Alcibiadem (the whole dialogue)Olympiodorus, F. Creuzer · 6th c.; ed. 1821Public domain
By
Olympiodorus (author (lectures)); F. Creuzer (editor) · 6th c.; ed. 1821
Source:
Olympiodorus, In Platonis Alcibiadem priorem commentarii, ed. F. Creuzer, Initia philosophiae ac theologiae ex Platonicis fontibus ducta, vol. 2 (Frankfurt: Brönner, 1821) · link ↗
In the reader:
370 entries: the prologue, each of the 28 lectures' θεωρία ('approximate'), and 352 comments on lemmata, each at the Stephanus section its lemma quotes ('exact')
Licence:
Public domain (Creuzer's edition, 1821). Greek corrected by AI (Claude) from the archive.org scan against the page images, with Tesseract and archive.org's OCR as witnesses; uncertain readings listed
Notes
Greek: Creuzer's text corrected by AI against the page images (scripts/plato/ocr_creuzer.py); typographic ϑ written θ; Creuzer's footnotes left out. English: AI (Claude), marked. Archive.org leaves n16–n247 = pp. 1–232; each entry links its first page.
Anonymous Prolegomena to Platonic philosophy, ch. 26Anonymous (school of Olympiodorus), C. F. Hermann · 6th c.; ed. 1853, repr. 1887Public domain
By
Anonymous (school of Olympiodorus) (author); C. F. Hermann (editor) · 6th c.; ed. 1853, repr. 1887
Source:
Platonis dialogi secundum Thrasylli tetralogias dispositi, ed. C. F. Hermann, vol. 6 (Leipzig: Teubner, 1887 reprint of 1853), Appendix Platonica, Prolegomena, ch. XXVI, p. 219 · link ↗
In the reader:
1 entry at 103a: Iamblichus' course, 'first the Alcibiades, because in it we come to know ourselves', with AI English
Licence:
Public domain (1853/1887)
Notes
Read from the page image (n262). The print names eleven dialogues; the Sophist and Statesman supplied by modern editors are reported, not checked.
Ficino, Argumentum in Alcibiadem primum (1484)Marsilio Ficino · 1484Public domain
By
Marsilio Ficino (author) · 1484
Source:
Plato, Opera, Latin by Marsilio Ficino, with his argumenta (Florence: Laurentius de Alopa, 1484[–85]), the argumentum to Alcibiades I (Biblioteka Jagiellońska copy, archive.org jbc.bj.uj.edu.pl.NDIGSTDR006716, images n22–n23) · link ↗
In the reader:
1 entry at 129e (its definition of man), with AI English
Licence:
Public domain (1484)
Notes
Checked against the Florence copy of the same printing (ita-bnc-in2-00000298-001, n55); collated with Lyon 1590 (p. 755) and Paris 1641 (Opera II, pp. 106–107). Contractions expanded; paragraph breaks ours; 6 uncertain readings listed.
Ibn al-Nadīm, al-Fihrist (the title only)Ibn al-Nadīm, OpenITI · c. 987Open licence
By
Ibn al-Nadīm (author); OpenITI (digital text) · c. 987
Source:
Ibn al-Nadīm, al-Fihrist, OpenITI text 0385IbnNadim.Fihrist.Shamela0007642-ara3, l. 7274 (printed vol. 1 p. 304) · link ↗
In the reader:
1 entry at 103a, with AI English
Licence:
Text public domain; OpenITI transcription CC BY-NC-SA 4.0
Notes
'Two discourses he called Alcibiades, on the beautiful' (Theon's list); no translator named. Documented: the title only; no evidence was found of an Arabic translation of the dialogue's content.
Ibn Abī Uṣaybiʿa, ʿUyūn al-anbāʾ (the title only)Ibn Abī Uṣaybiʿa, N. Riḍā, OpenITI · c. 1245–1268Open licence
By
Ibn Abī Uṣaybiʿa (author); N. Riḍā (editor); OpenITI (digital text) · c. 1245–1268
Source:
Ibn Abī Uṣaybiʿa, ʿUyūn al-anbāʾ, ed. N. Riḍā (Beirut), OpenITI text 0668IbnAbiUsaybica.CuyunAnba.Shamela0006687-ara1, l. 2421 · link ↗
In the reader:
1 entry at 103a, with AI English
Licence:
Text public domain; OpenITI transcription CC BY-NC-SA 4.0
Notes
Names 'the first Alcibiades … the second Alcibiades'. Line checked 2 Oct 2026.
Taylor's and Sydenham's notes (1804)Thomas Taylor, Floyer Sydenham · 1804Public domain
By
Thomas Taylor (annotator (T.)); Floyer Sydenham (annotator (S.)) · 1804
Source:
The Works of Plato (London: Thomas Taylor, 1804), vol. 1, pp. 16–96, notes signed T. and S. · link ↗
In the reader:
With the Sydenham–Taylor translation, kept apart from it, by section
Licence:
Public domain: published 1804; Sydenham died 1787, Taylor 1835
Notes
Part of the tr-sydenham file (notes); weighed as commentators in the triangulation (Taylor reads Plato through Proclus and Olympiodorus; Sydenham's notes are philological).

✦ marks AI that no person has reviewed yet; ✧ marks AI reviewed by a named person. Copyrighted works are cited, never shown. All sources · How texts are established

Share-alike: the Perseus, GLAUx, LSJ and Wiktionary layers are CC BY-SA, so this page, and anything made from those layers, is shared under the same terms, with credit. Nothing in copyright is shown. Audio is off for this text; background music is available. All sources · How texts are established

Stephanus sections

Each links to its place in the reader, and ↗ to the same section of Burnet's Greek in the Scaife reader (Perseus).

Lines the medieval manuscripts lack

⟨ 133c8–17 ⟩ These lines are not in Plato's medieval manuscripts. Burnet prints them, in angle brackets, from Eusebius (Praeparatio evangelica XI.27.5), and notes that Julian and Stobaeus read something like them; Stobaeus (Anthologium III.21.24) gives a version of them a little later in the conversation, after "we agreed that knowing oneself is self-control". The reader marks them ⟨ ⟩ as Burnet does: words supplied from an indirect witness, not an editor's deletion. Neither Jowett nor Lamb translates them; Taylor (1804) does, in square brackets, from Stobaeus and at Stobaeus' place in the conversation (see the Sydenham–Taylor layer). The English below is AI.

SOC. Then, just as mirrors are clearer than the mirror in the eye, and purer and brighter, is the god also purer and brighter than the best part of our soul? — ALC. It seems so, Socrates. — SOC. Then it is by looking at the god that we would use the finest mirror, and, among human things, at the excellence of the soul; and so, most of all, we would see and know ourselves. — ALC. Yes.

Published translations

Public-domain English, section by section, to read beside the Greek or compare in columns (Translations and Compare in the reader).

Sydenham–Taylor (1804) · English · The First Alcibiades

Floyer Sydenham, revised by Thomas Taylor, Sydenham's translation (first published 1767-80) as revised and annotated by Taylor, 1804. The Works of Plato, viz. His Fifty-Five Dialogues and Twelve Epistles, translated from the Greek; nine of the dialogues by the late Floyer Sydenham, and the remainder by Thomas Taylor (London: Thomas Taylor, 1804), vol. 1, pp. 16–96. Taylor's notes signed T., Sydenham's S.

Text: Transcribed by AI (Claude) from two archive.org OCRs of the 1804 printing (in.ernet.dli.2015.39263 and thwworksofplatov0001thom), page by page; long s modernised, 1804 spelling and punctuation kept. Not yet checked word by word against the page images.

Sections: Automatic and approximate (a clause either way), as for Jowett: each clause is placed in a section by a monotonic best path over its word overlap with Lamb's Loeb English and Jowett's English of that section.

Note: Taylor alone of our translators prints 133c8–17, in square brackets: his note names Stobaeus (Serm. 21, p. 183) as the source, and he places them where Stobaeus does, after 'to know oneself we acknowledge to be wisdom' (133c18–19). Notes kept apart (see notes).

Page scan / sourcePublic domain: published 1804; Sydenham died 1787, Taylor 1835.

Jowett (1892) · English · Alcibiades I

Benjamin Jowett, 3rd edition, 1892. The Dialogues of Plato, translated into English with analyses and introductions by B. Jowett, 3rd ed. (Oxford: Clarendon Press, 1892), vol. 2

Text: Project Gutenberg #1676 (Jowett's translation), compared word by word with the archive.org OCR of the 1892 volume 2; 3 Gutenberg readings restored to the 1892 wording

Sections: Automatic and approximate (a clause either way): Jowett's margins give Stephanus pages but not the letters, so each clause is placed in a section by a monotonic best path over its word overlap with Lamb's Loeb English (1927) of that section. Where his wording seems to straddle two sections, read the neighbour too. Jowett does not translate 133c8–17 (absent from the medieval manuscripts); his speaker names are printed as in Gutenberg.

Note: Divided into sections automatically; approximate.

Page scan / sourcePublic domain: published 1892; Jowett died 1893. Project Gutenberg text: not copyrighted in the US.

Lamb (Loeb 1927) · English · Alcibiades I

W. R. M. Lamb, published 1927. Plato in Twelve Volumes, vol. 8: Charmides, Alcibiades I and II, Hipparchus, The Lovers, Theages, Minos, Epinomis, translated by W. R. M. Lamb (Loeb Classical Library; London: Heinemann / Cambridge, Mass.: Harvard University Press, 1927)

Text: Perseus Digital Library, canonical-greekLit tlg0059.tlg013.perseus-eng2 (TEI, sectioned by Stephanus number)

Sections: By the Stephanus sections Perseus marks in the translation (placed at the nearest sentence, so a sentence either way).

Note: Loeb footnotes omitted. Lamb, like Jowett, does not translate 133c8–17.

Page scan / sourcePublic domain in the US (published 1927; the US term of 95 years ended on 31 December 2022). Perseus digitisation CC BY-SA 4.0. Lamb died 1946: in copyright in life+70 countries until the end of 2016, so public domain there too.

Corrections to the automatic analysis

  • 104b γένεσιν · γένεσιν is the dative plural of γένος 'race, nation' ('among many great nations'); GLAUx gave γένεσις 'coming to be' (found by translation triangulation, §17)
  • 106c ὅτι · ὅτι here is ὅ τι, the neuter of ὅστις, 'what' ('so that I may know what you will say'); GLAUx gave the conjunction ὅτι (found by translation triangulation, §17)
  • 106c παριέναι · παριέναι is the infinitive of πάρειμι (from εἶμι) 'come forward' to speak in the assembly; GLAUx gave the suppletive παρέρχομαι (found by translation triangulation, §17)
  • 113b ἦ · ἦ is the Attic first person imperfect of εἰμί, 'I was (the one asking)'; GLAUx gave ἠμί 'say' (found by translation triangulation, §17)
  • 116e ἀτεχνῶς · the adverb ἀτεχνῶς 'simply, really' (LSJ's own headword); GLAUx gave the adjective ἄτεχνος 'unskilled'
  • 119c περιέσομαι · περιέσομαι is the future of περίειμι (from εἰμί) 'be superior, prevail'; GLAUx gave περιέρχομαι 'go about' (found by translation triangulation, §17)
  • 120b τρίχα · τρίχα is the accusative of θρίξ 'hair' ('with the slave's haircut in their soul'); GLAUx made it an adverb τρίχα 'in three'
  • 123a ἀτεχνῶς · the adverb ἀτεχνῶς 'simply, really' (LSJ's own headword); GLAUx gave the adjective ἄτεχνος 'unskilled'
  • 123c ἐξῃρημένους · ἐξῃρημένους is the perfect participle of ἐξαιρέω 'take out, set apart'; GLAUx gave ἐκλέγω (found by translation triangulation, §17)
  • 123c ὑός · ὑός is the Attic genitive of υἱός 'son' ('Deinomache's son'); GLAUx gave ὗς 'pig'
  • 126e ἐπισταμένῃ · ἐπισταμένῃ is the participle of ἐπίσταμαι 'know', like ἐπιστάμενον just before; GLAUx gave ἐφίστημι 'set over' (found by translation triangulation, §17)
  • 128d ἄλλῃ · ἄλλῃ here is the dative of ἄλλος with τέχνῃ understood ('by another art'), as in the same section's ἄλλῃ μὲν ἄρα τέχνῃ; GLAUx gave the adverb ἄλλῃ 'elsewhere' (found by translation triangulation, §17)
  • 128d ἄλλῃ · ἄλλῃ here is the dative of ἄλλος with τέχνῃ understood ('by another art'), as in the same section's ἄλλῃ μὲν ἄρα τέχνῃ; GLAUx gave the adverb ἄλλῃ 'elsewhere' (found by translation triangulation, §17)
  • 129d ἀρτίως · the adverb ἀρτίως 'just now' ('what I was asking just now'), LSJ's own headword; GLAUx gave the adjective ἄρτιος 'complete, even' (found by translation triangulation, §17)
  • 132c ἐπιμελητέον · ἐπιμελητέον is the verbal adjective of ἐπιμελέομαι, 'one must take care'
  • 132d ἐννοῶμεν · ἐννοῶμεν is from ἐννοέω 'consider'; GLAUx gave a non-existent ἐννοόω
  • 133e κατιδεῖν · κατιδεῖν is the aorist of καθοράω 'look down on, discern'; GLAUx gave κατοράω

Made by AI in this edition and open to review; the Greek text itself is unchanged. GLAUx's reported accuracy is about 97% for morphology and 99% for lemmas, so more errors should be expected.

The text record

Our working notes on this edition: sources, decisions and what AI supplied. Show the notes

Added to the Greek reader on 2 October 2026 (127e–135e) for the theme path "Seeing with the heart" (docs/research/heart-sight.md), by the pipeline of the Symposium pilot; extended the same day to the whole dialogue, with the ancient commentaries as readers through the ages (§7c). The id alcibiades-1-mirror is kept so that links (#s-133a, the heart-sight path) stay valid. Method: docs/EDITORIAL-METHOD.md.

Work / author / datesAlcibiades I (Ἀλκιβιάδης α), transmitted with Plato's works. Its authenticity has been doubted since Schleiermacher and is still debated (reported); it was the first dialogue of Iamblichus' course in the Neoplatonic schools: verified in Proclus, In Alc., Creuzer p. 11; Olympiodorus, In Alc., Creuzer pp. 10–11; the anonymous Prolegomena, ch. 26 (Hermann vol. 6, p. 219) (§7c)
Part addedThe whole dialogue, 103a–135e (127e–135e first; 103a–127e added 2 Oct 2026)
Reader route/read/alcibiades-1-mirror · about /read/alcibiades-1-mirror/about · sections #s-103a … #s-135e
Statusworking edition · reviewed by: —

1. Witnesses

SiglumWitnessKindLicenceLocation
BurnetJ. Burnet, Platonis Opera II (Oxford Classical Texts, 1901)standard editionpublic domainas digitised by Perseus
PPerseus Digital Library, canonical-greekLit tlg0059.tlg013.perseus-grc2 (EpiDoc; Stephanus milestones, speakers, <add> for 133c8–17)digitalCC BY-SA 4.0github.com/PerseusDL/canonical-greekLit
GGLAUx corpus, document 0059-013 (A. Keersmaekers)digital, analysedCC BY-SA 4.0github.com/alekkeersmaekers/glaux

Consulted, not used: Denyer (Cambridge 2001) is in copyright (not held locally).

2. Base and why

Burnet via Perseus, as for the Symposium; GLAUx's tokens carry their Stephanus sections.

3. Collation

  • Script: scripts/plato/build_greek.py alcibiades-1-mirror. 162 sections; tokens walked back through P for its speaker turns (903 labelled turns, ΣΩ. or ΑΛ. as P labels them) and for P's <add> in 133c. Every token was found apart from 135 halves of compounds (‑δέ, ‑τε, crasis) and quotation marks.
  • GLAUx's artificial SPEAKER= tokens and elliptic E tokens are dropped.

4. Decisions

  • 133c8–17 (from ἆρʼ οὖν, ὅθʼ ὥσπερ κάτοπτρα to ναί: 71 tokens) are not in Plato's medieval manuscripts. Verified 2 October 2026 on the page images:
    • Burnet, Platonis Opera II (1901), prints them between angle brackets ⟨ ⟩ (a supplement), and his apparatus at 133c reads: "c 8–17 Ἆρʼ οὖν … Ναί Eusebius : om. B T (sed talia fere legerunt Iulianus Stobaeus)". So his text of the lines is Eusebius's; his two manuscripts B and T omit them; Julian and Stobaeus are said only to have read "roughly such things". Scan: archive.org platonisopera02platuoft, page n362 (the page headed 133a; the lines continue on n363 to 133c17 ⟨ΑΛ. Ναί.⟩).
    • Eusebius, Praeparatio evangelica XI.27.5 (Viger p. 551b–c) quotes 133c1–17 continuously, with the lines in exactly the place where Burnet prints them, as proof that Plato learned from Moses. Checked in Dindorf's text (Leipzig 1867, vol. 2, pp. 56–57; First1KGreek tlg2018.tlg001.1st1K-grc1, GitHub) and in Gifford's translation (Oxford 1903, vol. III.2, pp. 595–596, the lines in round brackets, note at 551 b 8: "The passage in brackets is not in the MSS. of Plato"; archive.org). Mras's edition (GCS 43.2, 1956) was not consulted: its archive.org scans are lending-only.
    • Stobaeus, Anthologium III.21.24 (chapter "Περὶ τοῦ γνῶθι σαυτόν"; ed. O. Hense, Wachsmuth–Hense vol. III, Berlin 1894, pp. 573–577, lemma "Ἐν ταὐτῷ (p. 132 B — p. 134 B)"; the lines at p. 576.4–11; archive.org adw8682.0003.001.umich.edu, page n655) excerpts 132b–134b and has a version of the lines at a different point: after "Τὸ δὲ γιγνώσκειν αὑτὸν ὡμολογοῦμεν σωφροσύνην εἶναι; — Πάνυ γε" (133c18–19), opening "Ἆρʼ ὥσπερ κάτοπτρα" (no οὖν, ὅθʼ), with "καθαρώτερά τε", and with 133c18–19 then repeated. Hense's apparatus (p. 576, on line 4) prints Eusebius's form for comparison and notes that Wex wanted Stobaeus's text put after οὐκ ἔχομεν (133c3) with τῷ θεῷ … φαίνεται deleted.
    • Julian: named by Burnet; the locus is not given there and was not checked.
    • So: "transmitted via Eusebius and Stobaeus" was loose. Burnet's text is Eusebius's; Stobaeus is a second, differently placed and slightly different witness. The claim that most editors take the lines as a late-antique addition (in the reader note) is not verified and is dropped from the corrected wording below.
    • Burnet's ⟨ ⟩ marks a supplement; ⟦ ⟧ conventionally marks words an editor deletes. Changed 2 Oct 2026: the reader now marks the lines ⟨ ⟩ as Burnet does, and its note says they are supplied from an indirect witness (Eusebius), not deleted.
    • Taylor (1804) prints them in square brackets; his note names "Stobæus, Serm. 21. p. 183" and he places them where Stobaeus does, after "to know oneself we acknowledge to be wisdom" (133c18–19). So his source is Stobaeus (from his note and his placement; his English cannot show which Greek form he followed word by word). The words carry ins in the data. Neither Jowett nor Lamb translates them, so an AI English is given in the note, marked AI. The path "Seeing with the heart" (content/paths.json, stop 133a, "ground") now carries the verified wording.
  • Speakers are shown as P labels them (ΣΩ., ΑΛ.), at the start of each turn.
  • Corrections to the automatic analysis (17, listed on the About page, each with its reason): 132c ἐπιμελητέον (ἐπιμέλω → ἐπιμελέομαι), 132d ἐννοῶμεν (ἐννοόω → ἐννοέω), 133e κατιδεῖν (κατοράω → καθοράω); for the whole dialogue: ἀτεχνῶς (116e, 123a; ἄτεχνος → ἀτεχνῶς), 120b τρίχα (→ θρίξ 'hair'), 123c ὑός (ὗς 'pig' → υἱός); and from the triangulation (§7b): 129d ἀρτίως (ἄρτιος → ἀρτίως), 128d ἄλλῃ ×2 (adverb → dative of ἄλλος with τέχνῃ), 104b γένεσιν (γένεσις → γένος), 106c παριέναι (παρέρχομαι → πάρειμι 'come forward'), 106c ὅτι (→ ὅ τι, ὅστις), 113b ἦ (ἠμί → εἰμί 'I was'), 119c περιέσομαι (περιέρχομαι → περίειμι 'prevail'), 123c ἐξῃρημένους (ἐκλέγω → ἐξαιρέω), 126e ἐπισταμένῃ (ἐφίστημι → ἐπίσταμαι).
  • Two lemmas the first triangulation called errors were checked and kept: 131d ἄπεισιν (lemma ἄπειμι is right for both homographs; the shared gloss now gives both, 'be away' from εἰμί and 'go away' from εἶμι), and 133c ὅθʼ (= ὅτʼ, ὅτε elided and aspirated before ὥσπερ; ὅθεν cannot elide; the gloss of ὅτε now adds its causal sense 'since').

5. Segmentation and word IDs

162 passages, one per Stephanus section (#s-103a … #s-135e); word IDs <section>.<n> (unchanged for 127e–135e when the text was extended).

6. Analysis layer

  • LSJ: 1,020 of 1,072 lemmas matched (the misses are mostly proper names, and a few particles LSJ files elsewhere).
  • Etymology: Boisacq 505, LSJ's derivations ~280, Wiktionary 908.
  • Glosses: AI, shared by lemma with the other Greek texts; 338 written for 103a–127e on 2 October 2026 (scripts/plato/ai/greek4-glosses-5.tsv, from LSJ's first senses and one occurrence each, by headless Claude), and the doubtful ones corrected after triangulation (§7b).

7. Translation

  • Jowett, 3rd ed. (1892), vol. 2, Appendix I, pp. 457 ff. (in an appendix, apart from the dialogues Jowett took to be Plato's). Text: Project Gutenberg #1676, compared word by word with the OCR of the 1892 volume (archive.org dialoguesofplat02plat): the wording agrees; Gutenberg's two inline cross-references "(Compare Arist. Pol.)" (footnotes in 1892) are removed and "Erechteus" restored to the printed "Erechtheus". Divided into sections automatically against Lamb's sectioned English (a clause either way).
  • Lamb, Loeb vol. 8 (1927), via Perseus tlg013.perseus-eng2, sectioned by Perseus. Public domain in the US (95-year term ended 2022); Lamb died 1946, so public domain in life+70 countries too.
  • Sydenham–Taylor: Floyer Sydenham's translation (1767–80) as revised and annotated by Thomas Taylor, The Works of Plato (London 1804), vol. 1, pp. 16–96 (the whole dialogue; pp. 78–96 first), with Taylor's (T.) and Sydenham's (S.) notes kept apart (91 notes; 15 on 127e–135e; Taylor's on 133b–c quote Proclus, Platonic Theology I.3). Taylor alone of our translators prints 133c8–17, in square brackets, from Stobaeus (§4). Text: transcribed by AI (Claude) from two archive.org OCRs of the 1804 printing (in.ernet.dli.2015.39263, thwworksofplatov0001thom), page by page (for pp. 16–77 the pages are cut by the archive.org page map of thwworksofplatov0001thom), long s modernised, 1804 spelling kept; 180 places the transcription could not settle are listed in the file (uncertain), mostly punctuation; not yet checked word by word against the page images. The sectioning of a few 127e–135e clauses moved by a clause when the whole text was aligned (129a–b, 130e–131a, 133b–c). Sectioned like Jowett (scripts/plato/tr_sydenham_alcibiades.py). Public domain (1804; Sydenham d. 1787, Taylor d. 1835).

7b. Translation triangulation (scripts/triangulate.py alcibiades-1-mirror, PROTOCOLS §17), 2 October 2026

First run (127e–135e, morning). Jowett, Lamb, Sydenham–Taylor; Olympiodorus from the uncorrected OCR. 66 divergences, 22 serious. Our AI English of 133c8–17: no contradiction (checked against Taylor). 15 glosses flagged "doubtful"; fixed (shared by lemma, so every Greek text gets them; build_lexicon.py rerun for all six): φέρω (127e, 128d) + "(imperative φέρε) come now"; πολλάκις (127e) + "(after εἰ or μή) perhaps, by chance"; ᾗ (128c) + "in which way (dative of ὅς)" (the word itself was already analysed as ὅς); ἄλλῃ (128d) lemma corrected to ἄλλος (§4); ἀρτίως (129d) lemma corrected (§4), gloss "just now, lately"; ἔχω (129e, 133c) + "(with infinitive) be able"; ἕως (131d) + "while, as long as"; ἄπειμι (131d) "be away, be absent; (from εἶμι) go away" (lemma kept, §4); ἀπολείπω (132a) + "leave behind, abandon"; ὅτε (133c) + "since, seeing that" (lemma kept, §4); ἀρχή (134c, 135a) + "rule, office"; σχῆμα (135d) + "bearing, role".

Second run (the whole dialogue, afternoon; --redo). Compared: Jowett (1892), Lamb (1927), Sydenham–Taylor (1804) on all 162 sections; commentaries, weighted by lineage (the braid entries carry it): Olympiodorus (whole, corrected), Proclus (to 116b, excerpts), Ficino's argumentum, the curriculum entries, Taylor's and Sydenham's notes. No copyrighted translation is held (no O'Neill, Hutchinson, Denyer), so none was compared privately.

  • 257 divergences; 63 serious (shown as "Translators differ") in 54 sections: 35 interpretation, 12 term of art, 11 syntax, 5 variant. Commentators bear on 43 of the serious ones: Olympiodorus 39, Proclus 10, Sydenham's notes 4, Taylor's 1.
  • The 133c8–17 divergence's wording was corrected by hand in the log and republished (--revalidate): Burnet prints the lines from Eusebius; Stobaeus has a version placed later, where Taylor puts them.
  • AI-English check: 34 glosses flagged (30 doubtful, 4 contradiction), then 4 more on a rerun of those sections; all fixed, and a third run of the affected sections returned no flags. Lemma errors corrected in the data (§4): 104b γένεσιν, 106c παριέναι, 106c ὅτι, 113b ἦ, 119c περιέσομαι, 123c ἐξῃρημένους, 126e ἐπισταμένῃ. Glosses widened in our own words: ἀπαλλάσσω (middle 'depart'), φρονέω (μέγα φρονεῖν 'be proud'), θαυμαστός ('surprising'), ἐπιτίθημι (middle 'set about'), ἐφίημι ('allow'), ἐκεῖ ('there, in that case' first; the Neoplatonic 'intelligible world' second), σκευάριον ('gear, costume'), χαίρω (χαίρειν ἐᾶν 'let be'), ἔσχατος ('utmost'), διαφερόντως ('differently' first), τελέω ('pay'), ἑκάστοτε ('on each occasion'), ὄγκος ('dignity, pride'), δύναμις (εἰς δύναμιν), ἀκόλουθος ('consequent on'), οὗ ('where'), πλέθρον (as a measure of land), παρά (with accusative 'to'), κάμνω ('be ill'), συμβάλλω ('make contracts'), μή (μὴ οὐ), καλός (adverb 'well, rightly'), ὁστισοῦν ('anything'), ἀνδρεῖος ('of men'), πάρειμι ('come forward'), περίειμι (new), ὀρτυγοκόπος (a player), μήν (καὶ μήν 'and moreover'), ἐξαιρέω (new).

7c. Readers through the ages (scripts/plato/commentary_alc.py; app/src/content/plato/alcibiades-1-mirror-braid.json)

442 entries, every one with its edition, licence and a link to the page scan; each states the reader's lineage. Long texts are fetched when a section's readers are opened (app/public/data/plato/commentary/alcibiades-1-mirror-{olympiodorus,proclus}.json).

Olympiodorus, *In Alcibiadem*: F. Creuzer, Initia philosophiae ac theologiae ex Platonicis fontibus ducta II (Frankfurt: Brönner, 1821), public domain; archive.org creuzer-initia-philosophiae-ac-theologiae-ex-platonicis-fontibus-ducta-gr-lat-t.-2-1821, leaves n16–n247 = pp. 1–232 (page images https://archive.org/download/<item>/page/n<leaf>.jpg; each entry links its first page).

  • OCR corrected for the whole commentary (lectures 1–28, not only 22–28) by scripts/plato/ocr_creuzer.py: for each page, Tesseract (grc) and archive.org's OCR as witnesses, and Claude (headless) reading the page image in two strips at full width; Olympiodorus' Greek only (Creuzer's Latin notes left out), lemmata marked, ϑ written θ, nothing else normalised. Cached per page in data/raw/plato/olympiodorus/ai-pages/ (gitignored); full corrected text olympiodorus-corrected.txt. 1,193 readings the model could not settle are listed (each comment shows those in its own text, "Uncertain readings in the print"); 24 pages were flagged for low agreement with Tesseract, which on inspection is mostly Creuzer's Greek-heavy footnotes in the OCR, not misreadings. Spot checks by eye against the images (pp. 10, 11, 84, 192, 198): no wrong words found beyond the listed uncertainties (p. 192 has ταύτην where the print may read ταὐτὴν). Not yet proofread page by page.
  • Structure: 28 lectures (Πρᾶξις σὺν θεῷ …; p. 170 prints θ̅ for the 19th, read as 19). 370 entries: the prologue (lectures 1–2 opening, pp. 1–9), each lecture's θεωρία (opening with the passage discussed; spans the sections of the lecture, "approximate"), and 352 comments on lemmata, each placed at the Stephanus section its lemma quotes ("exact": the lemma matched against Burnet's words; the lemmata are the largest run of quotations that never goes backwards, so Plato quoted from elsewhere stays in the text). A lemma Creuzer prints in two spaced pieces is one comment.
  • English: AI (Claude), per lecture, marked; cached in data/raw/plato/olympiodorus/ai-en/.

Proclus, *In Alcibiadem* (to 116b): Creuzer, Initia I (Frankfurt 1820), public domain; archive.org creuzer-initia-philosophiae-ac-theologiae-ex-platonicis-fontibus-ducta-gr-lat-t.-1-1820. Westerink (1954) and Segonds (1985–86) are in copyright and were not used; O'Neill's English (1965) is not held.

  • Proclus spends some 340 pages on 103a–116b, so the layer gives an excerpt per lemma: the lemma and the opening of his comment (about 1,800 characters), with the page on which the next excerpted lemma starts. The lemma pages were found by matching Burnet's Greek against archive.org's hOCR of every page; 134 pages were corrected as for Olympiodorus (578 uncertain readings listed; 234 fall in the excerpts). 65 excerpts, 103a–116a; plus the prologue's opening (p. 1) and its passage on Iamblichus' course (p. 11), both checked by eye.
  • English: AI, marked.

The course (lineage). Proclus p. 11 (Iamblichus gave it "the first place among the ten dialogues"), Olympiodorus pp. 10–11 (to be placed first of all; the dialogue is the propylaea of the temple, the Parmenides its sanctuary), and the anonymous Prolegomena, ch. 26 (Hermann's XXVI), in Platonis dialogi ed. K. F. Hermann vol. 6 (Teubner 1887 reprint of 1853), p. 219, archive.org platonisdialogi06plat n262: Iamblichus' twelve dialogues, "first the Alcibiades, because in it we come to know ourselves", last the Philebus (the print names eleven; the Sophist and Statesman supplied by modern editors are reported, not checked). All three read from the page images.

Ficino: Argumentum Marsilii in Alcibiadem primum de natura hominis, in Plato, Opera, Latin by Ficino (Florence: Laurentius de Alopa, 1484[–85]), unpaginated: Kraków copy, archive.org jbc.bj.uj.edu.pl.NDIGSTDR006716, images n22–n23 (the leaf foliated "16" by hand), checked against the Florence copy of the same printing (ita-bnc-in2-00000298-001, n55). Collated with Lyon 1590 (p. 755) and Paris 1641 (Opera II, pp. 106–107): differences are of spelling, plus 1590/1641 "animam ut ideam / ut idem" for 1484 "animam in idea", 1590 "A quo" for "A qua", 1641 "omnino" for "nomino". The Basel Opera (1561/1576) vol. 2 was not found on archive.org (both local OCRs are vol. 1). Contractions expanded; paragraph breaks ours; 6 uncertain readings listed. English AI. Shown at 129e (its definition of man).

The Arabic transmission (graded note). Checked in OpenITI texts and the scholarship, 2 Oct 2026:

  • Documented: the title only. Ibn al-Nadīm, Fihrist (Theon's list of Plato's works): «قولان سماهما القيبادس في الجميل» "two discourses he called Alcibiades, on the beautiful" (OpenITI 0385IbnNadim.Fihrist.Shamela0007642-ara3, l. 7274, printed vol. 1 p. 304; no translator named, unlike the Republic, Laws, Timaeus and Sophist nearby); Ibn al-Qifṭī copies it garbled (الفيناذس); Ibn Abī Uṣaybiʿa, ʿUyūn, «القيبيادس الأول … القيبيادس الثاني» (OpenITI 0668IbnAbiUsaybica.CuyunAnba.Shamela0006687-ara1, l. 2421). Both lines checked.
  • No evidence found of an Arabic translation, epitome or gnomological excerpt of the dialogue's content (the self as the soul that uses the body; the eye in the eye), nor of an Arabic Proclus or Olympiodorus on it (the Fihrist's Proclus list has none), in Ibn Abī Uṣaybiʿa's, al-Shahrazūrī's and Miskawayh's sayings collections (keyword searches). Not checked directly: al-Mubashshir, the Ṣiwān al-ḥikma, Ḥunayn's Ādāb (not in OpenITI); Gutas in DPhA Va (paywalled).
  • Possible echo only: al-Masʿūdī's "saying of Plato" on the Ḥarrān temple, "whoever knows his essence becomes divine" (Murūj, OpenITI 0346Mascudi.MurujDhahab.JK010344-ara1 ll. 12454–56); Rosenthal (Islamic Culture 14, 1940, p. 409 n. 3) compares Alc. I 130a–131a, without claiming transmission. Altmann ("The Delphic Maxim", 1963, pp. 198–200, 224) traces the Arabic "know yourself" to a late Greek source and Proclus' commentary, not to an Arabic Alcibiades. The bare maxim (man ʿarafa nafsahu) is not evidence for the dialogue.
  • In the layer: two entries at 103a (Ibn al-Nadīm, Ibn Abī Uṣaybiʿa), title only, with AI English.

8. Rights

Share-alike as for the other Perseus texts. Nothing in copyright is shown.

9. Open questions / to review

  1. ~~Check Burnet's presentation of 133c8–17 and the Eusebius and Stobaeus references~~: done 2 Oct 2026 (§4). Taylor's source: Stobaeus, from his note and placement (§4). Still open: Julian's locus; Mras's apparatus to PE XI.27.5.
  2. Collate with Denyer (2001).
  3. ~~A readers layer~~: done (§7c). Open: proofread the corrected Creuzer Greek page by page, starting with the listed uncertain readings; correct the rest of Proclus (only the lemma pages are done); Taylor's "Additional notes … from the MS. commentary of Proclus" (1804, vol. 1, pp. 481 ff.) could give a public-domain English beside some Proclus excerpts.
  4. Check the Sydenham–Taylor transcription against the 1804 page images (pp. 16–96), starting with its 180 uncertain places; 6 pages (23, 38, 50, 58, 68, 69) were not located in the alignment, so their notes' sections are approximate; p. 79 prints "Soc." where the sense needs Alcibiades ('I do not apprehend your meaning').

9b. Pending for the data: applied (2 Oct 2026)

The corrected 133c8–17 wording is in the reader note (build_greek.py → insertion.note), the docstring, the triangulation file (via the log, §7b) and the Sydenham note (Taylor's source shown from his note and placement); the reader marks the lines ⟨ ⟩.

10. Change log

  • 2026-10-02: the whole dialogue (103a–135e), id kept; Sydenham–Taylor pp. 16–96; Olympiodorus (whole) and Proclus (excerpts to 116b) corrected from Creuzer's page images as readers through the ages, with the curriculum entries, Ficino's argumentum (1484) and the Arabic bibliographers (§7c); 133c8–17 marked ⟨ ⟩ with the corrected note (§9b); triangulation rerun and all glosses it flagged fixed (§7b); 12 lemma corrections.
  • 2026-10-02: 133c8–17 claim verified on the page images (Burnet's apparatus; Eusebius PE XI.27.5; Stobaeus III.21.24, Hense p. 576) and corrected in §4, in paths.json and in docs/research/heart-sight.md; wording for the data under §9b.
  • 2026-10-02: Sydenham–Taylor (1804) added as a translation layer; translation triangulation run (§7b).
  • 2026-10-02: added: Greek with speakers, 133c8–17 marked, lemmas (three corrected), LSJ, Boisacq, Wiktionary, AI glosses, Jowett and Lamb, AI English of 133c8–17; the "Seeing with the heart" stop now opens at 133a.