Charmides

Note on this Edition

How the volume was made

The Greek was taken from the Perseus Digital Library’s digital transcription of Burnet and cut into Burnet’s lines by reading every line-end from the photographs of his page, in the 1909 impression described in the Editorial Note. The section letters and line numbers in the margin were computed from the Stephanus divisions rather than copied, and were then compared with Burnet’s: all of them, across the 1,067 lines of the dialogue, fall where he prints them. From the second block on, doubtful accents, breathings and superscript numerals were read on the sharper 400-dot text layer that lies beneath each photograph in the scan; in the first block they were read on the photographs at up to fourteen times’ magnification. Burnet’s apparatus was transcribed from the same images.

The work was done in five blocks of six to nine pages, each finished and checked before the next was begun. Each block was checked by four independent reviewers: two compared the Greek and the apparatus with the page images letter by letter, one compared the English with the Greek clause by clause, and one checked every fact and reference in the notes. Each finding was examined again against the evidence before it was adopted, and not all were adopted. What each block’s check caught is recorded at the end of this Note.

When the last block was done the volume was read through as one text, since a check made page by page cannot see some kinds of error: a word rendered one way in one block and another way in the next, a sentence broken badly at a page turn, a step of the argument stated in terms different from those of the step it relies on. The translator read the whole volume, and four further reviewers read it independently: one for the consistency of terms across the whole book, two for sense, omissions and continuity in its two halves, and one for the front matter and the abbreviations of the apparatus. The Introduction, the Key Words of the Dialogue, the Outline and the Glossary were then written from the finished text, and their claims and references were checked against it.

What the final reading caught

It found some fifty faults that the five block checks had passed, most of them small and several substantive. They fall into a few classes, and the classes are worth a reader’s knowing, because they are the kinds of error that an edition checked page by page is likely to contain.

The most serious was words that Burnet deletes surviving in the English. At 169a4 he brackets πλὴν ἐπιστήμης, “except knowledge”, as spurious, but the English still translated them, in square brackets and without a note, so that Socrates’ question named knowledge as the one exception before asking whether knowledge is one. The words are now left untranslated, as the edition promises, and a note records the deletion.

The largest class was terms rendered two ways, with the change falling between blocks. At 162a Socrates was made to speak of “making and doing” one’s own, where the Greek has “working and doing”, so that at 163b Critias attacked a word Socrates had not used. Medicine was “knowledge of health” at 165c but “of the healthy” from 170a, although 171a refers back to the definition at 165c. The knowledge that makes us happy was “of good and evil” at 174b–d, while the same Greek pair was “good and bad” at 156e and in Critias’ definition at 163e. Critias’ “usefully” at 163c hid that Socrates’ “beneficially” on the next page was the same word, and “useless” at 175d hid that the verdict answered the question of benefit. Socrates’ “You rascal!” at 161b had become “You villain!” at 174b; the verb “examine” of 167a and 170d had become “test” at 172b; the ship’s “captain” at 173b was a “pilot” at 174c; and at 176b Charmides “believed” Socrates where the Greek resumes the play on persuading and obeying from 156a and 157b–c. At 161b “consider” hid that Charmides was turning Socrates’ own “look into” back on him. All are now rendered one way, and the Glossary and Key Words record the rulings.

The next class was sense errors in hard constructions, sentences that read plausibly on their own. At 164e the English had Critias tell people to urge one another not to rejoice; he says we should not bid one another “Rejoice!” but “Be temperate”. At 161d “or teaches you boys only yours” added words not in the Greek. At 172c “we have been searching for nothing useful” made temperance the useless thing, where the Greek says the search was to no purpose, as Socrates repeats at 175a. At 175b–c the recapitulation now says “by this knowledge one knows the works of the other kinds of knowledge”, in the terms of the step it recalls (170b–c). The tentative question at 174a, the plural formula at 171d and an overstatement at 167c were also corrected, and three times (154a, 154e, 171a) a Greek idiom meaning “is in fact” had been rendered “happens to be”, against the edition’s own ruling.

At two page turns (174a and 174d) the English page carried words that belong to the Greek page following; the breaks now match. Nested quotation marks on the page of Critias’ Delphic speech (164d–165b) were set as double instead of single, and at 165c an opening quotation mark was turned the wrong way by the typesetting program, which has been fixed. Small additions (“of itself” at 159a, “for hire” at 163b) and an omission at 175e (“a charm for a thing worth nothing”) were set right.

In the notes, three cross-references were wrong or loose (161c–162b, where Socrates speaks of a riddle three times, not twice; the reference at 166a; the return of the charm, but not the remedy, at the close); two textual notes that the edition’s working record said had been written, at 156e and 157d, had never been written, and have now been supplied; and a note on the play on persuading and obeying has been added at 176b.

In the front matter, the Editorial Note had said that the Greek was checked against “the 1903 printing”; the copy is in fact the second impression of 1909, corrected in places, and the Note now says so. It named two passages that Stobaeus quotes where he excerpts three, left Priscian and many of Burnet’s abbreviations unexplained, and called Burnet’s text the one in which most readers meet the dialogue, a claim it could not support and has withdrawn.

What remains undone

The identity of Burnet’s witness Θ is not settled; he does not list it among his sigla or discuss it in his preface. Its agreement with B’s gaps in the Protagoras suggests that it was copied from B, and it may be the Vaticanus gr. 226 that Bekker and Stallbaum cite as Θ, which contains this tetralogy, but that has not been confirmed.

A few diacritics and superscripts in Burnet’s small apparatus type lie at the limit of what the scan can show, and are printed as read: the breathings under a circumflex at 163a11, 167b10, 167c8 and 167d4; the damaged superscript at 168e9, read as T³, and the heavily inked one at 169a3, read as T²; and the two breathings at 174b1. They should be checked against a better copy of the 1909 impression, or of the first impression of 1903, which may differ from it in places.

The apparatus is Burnet’s alone. It has not been collated against the manuscripts, and readings proposed since 1903 are not reported. The translation follows Burnet’s text even where later scholars would print something else.

The notes are a reader’s aid, not a commentary. They were chosen to explain what the English cannot carry and to mark the dramatic irony, and they label interpretations as such; a reader who wants the scholarly debate on the argument of 165–175 will need a commentary.

The checks described above were made by independent reviewers working from the same files and page images. The volume has not been read by a specialist outside the project, and a reader who finds an error is asked to report it.

The record of the blocks

The notes that follow were written as each block was finished, and are kept as a record of what each check found. Where the final reading changed something they report, the text above says so.

Block I (153a–159b)

How it was made. The Greek was taken from the Perseus digital transcription of Burnet and cut into Burnet’s lines by reading every line-end from the photographs of his page (in the 1909 impression, as the final reading established). The section letters and line numbers in the margin were then computed from the Stephanus divisions rather than copied, and all of them, across the 243 lines of these nine pages, fell exactly where Burnet prints them. The paragraph breaks read from the photographs agreed with the digital text at every change of speaker. Burnet’s apparatus was transcribed from the photographs at magnifications of up to fourteen times.

What verification caught. The block was then checked by four independent reviewers. Two compared the Greek and the apparatus with the page photographs letter by letter, one compared the English with the Greek clause by clause, and one checked every fact and reference in the notes and in the Editorial Note. Each finding was examined again against the evidence before it was adopted.

In the Greek there were three corrections. At 156d6 the digital text’s Θρᾲξ, with a grave accent, is what Burnet actually prints, although the expected form is Θρᾷξ; the draft had “corrected” it, wrongly, and Burnet’s reading is restored. In the apparatus at 155b9, B and W read ἥκει, not ἧκει, and at 157c4 the correcting hand is B³, not B². Two doubtful breathings, at 155c2 αὑτῷ and 157b3 αὑτοῦ, were confirmed as rough.

In the English, six errors of sense were corrected. At 155c, Socrates’ boldness rested on his expectation of an easy conversation, not the other way round. At 157b, the mistake is one made “in treating human beings” by some doctors, not by people at large. At 157c, “this weakness in the head” would have had Critias call his cousin simple; it is now “this trouble with his head”. At 158a, the comparison with “your uncle Pyrilampes” had been lost. At 158b, the remedy “should” be given, an obligation, not a possibility. At 158c, Charmides looked more beautiful “for one thing”, not “at first”. Smaller omissions and additions were set right at 153a, 153b, 154c and 155b. At 154b the pun on “measure” and “measuring-line” is now carried in the English, and at 154e Critias’ phrase is rendered “beautiful and good” so that its joke shows.

In the notes, a speaker was corrected: at Phaedrus 270c it is Phaedrus, not Socrates, who cites Hippocrates. T’s reading at 158b1 was restated as “to disgrace none of your forebears”, not “disgraced”. Critias’ genealogy was brought into line with the modern reconstruction, the Symposium reference narrowed to 219e–220e, and the lovers of 154a identified as young men rather than grown men. Cydias’ fragment is now called his only secure one, and several interpretive remarks are labelled as interpretation.

What remains open. The manuscript Burnet calls Θ is not on his Sigla page and has still to be identified; the Editorial Note names it only as he does.

Block II (159b–162e)

How it was made. As for Block I: the line-ends of Burnet’s pages 10–15 were read from the photographs, the margin numbers computed and found to agree with Burnet’s at every one of the 180 lines, and the paragraphing read from the photographs agreed with the digital text throughout. Block II adds Burnet’s first displayed verse (Homer at 161a) and his first italic Latin in the apparatus (160a).

What verification caught. The same four checks were run. In the Greek text itself nothing was wrong. In the apparatus three readings were corrected: at 159b2 the reading written above the line in W is οἱ, with a breathing only; at 159d4 Cobet’s conjecture is τἆρα, with a circumflex; and at 162e1 Burnet prints re vera in roman, not italic. A reviewer found that the scan carries a sharper 400-dot text layer beneath the photograph, and every doubtful accent and superscript in this block was settled on it; two superscripts were confirmed as B³ and one as B².

In the English, the idiom pantos mallon at 162b had been taken as “more than ever”; it means “most certainly”. The simile at 162d now keeps “the poet’s own verses”, which matters because the definition being mangled is Critias’ own. The play between “like a man” and “very much like a man” at 160d–e is now audible, and Homer’s line at 161a keeps “present”, which the argument picks up two lines later. Smaller corrections were made at 160d, 161b, 161c and 162b, and a section letter was moved at 162e.

In the notes, the account of the second half of the argument at 160b–d was corrected; the claims about Critias as a dramatist were reduced to what the ancient evidence supports; the note on Homer’s line now says that Plato adapts it; and the notes on hēsychē and on Charmides’ second answer now label their interpretations as such.

Block III (162e–167e)

How it was made. Burnet’s pages 16–23. From this block on, every doubtful accent, breathing and superscript is read on the 400-dot text layer of the scan rather than on the photograph. All 233 lines, margin numbers and paragraph breaks agree with Burnet’s page. One defect in the digital text was removed: a stray numeral keyed into it at 163c.

What verification caught. The Greek text itself was again free of error. In the apparatus ten readings were corrected, all single accents, breathings or superscripts in Burnet’s small type: at 163a11 ἦ B; at 165a6 φῇς B; at 165b2 αὑτὸν B; at 165e3 ταύτῃ B, where a correction made in drafting had wrongly removed the iota subscript; at 166a1 B³; at 166d4 κοινων B, printed without accent; at 167b2 ὴ B, with no breathing; and at 167b10, c8 and d4 the smooth breathing under a circumflex. These last four are read at the limit of what the scan shows and are recorded as open questions in the register.

In the English, the doctor who “does not know, of himself” (164c) could have been read as “does not know unaided”; it is now “does not know himself — how he has acted”, which keeps the Greek and prepares “know thyself”. At 163d the construction of Socrates’ summary of Critias was made unambiguous, and Critias’ two words “appropriate” and “one’s own” now echo each other as in the Greek. Smaller omissions were restored at 163a, 163d, 165a, 165c, 166a, 166b and 166c–d, a missing section letter was added at 166d, and three page-turns were brought into line with Burnet’s.

In the notes, the account of logistikē was wrong: the contrast between practical reckoning and the theory of number is later than Plato. The maxim “Pledge, and ruin is near” is now given its ancient interpretations and ascriptions, the notes on Hesiod and Xenophon were made exact, and four interpretive remarks are labelled as such. The Editorial Note now defines every abbreviation used in the apparatus so far.

Block IV (167e–172d)

How it was made. Burnet’s pages 24–31, read on the 400-dot text layer. All 238 lines, margin numbers and paragraph breaks agree with Burnet’s page. The builder now also checks every line’s length against the rest, which caught one line-end matched too early before the proofs were made.

What verification caught. In the Greek, a space had crept between a bracket and a comma at 169a4 (and would have at 172c8); the fault was in the handling of the digital text and is fixed throughout. In the apparatus, three readings were corrected: at 169a3 T² reads αὐτὴν, with a smooth breathing; at 171b11 T reads ἦ, with a circumflex; and at 168e9 the damaged superscript is read as T³, not T². Burnet’s apparatus gives “e8” for πᾶν at 171e, which stands on line e7; the slip is his and is kept. A stray accent printed above ἀναγκαῖον at 172a2 is a displaced sort, not a reading, and is not reproduced.

In the English, the question at 170a10 now follows Burnet’s text, which asks whether knowledge of knowledge is the same as knowledge of the healthy and of the just; the draft had followed the manuscripts. At 170e the question-tag had reversed the doctor’s answer, and at 172c the draft had added a clause (“is what it is”) that Plato does not write. At 171c “his own fellow craftsman” is now the temperate man’s, and “do well” is kept both times at 172a, so that the reader can see the phrase slide from acting well to faring well. Smaller corrections were made at 168a, 168e, 169a, 170b, 171e and 172b.

In the notes, the suspicion of “forced” at 169c is Badham’s, not Burnet’s; the claims about love of the beautiful, about Homer’s megalōsti, and about the Republic’s division of labour were made exact; and three interpretations are now labelled as such.

Block V (172e–176d)

How it was made. Burnet’s pages 32–37, the end of the dialogue, read on the 400-dot text layer. All 173 lines, margin numbers and paragraph breaks agree with Burnet’s page, and the build runs to the last word of the text. Three new cases were taught to the builder: a line made only of the end of a divided word (τευτικόν; at 174b), the end of the text itself, and a raised point set against the following word at 175e6.

What verification caught. The Greek text itself was free of error. In the apparatus five readings were corrected. At 174b1, b5, b9 and b10 Burnet reports the manuscripts’ own forms, which the draft had normalised to ᾗ: he prints ἢ τί B, ἥ τι T; ἀλλ’ ἡ B; ἠ τι B, ἡ τί T; and ῆ τὸ T, with no iota subscript anywhere. At 176b5 Madvig’s conjecture is δρᾶ, without the subscript that Hermann’s δρᾷς has beside it.

In the English, the draft had translated at 174d7 the words [ἡ ὠφελίμη], which Burnet deletes, and in the form of Schleiermacher’s conjecture, which he rejects; they are now left untranslated. At 175c6 the draft had made Socrates’ point a flat contradiction (“he knows the very things he doesn’t know”); the Greek says that our agreement has him know, of what he does not know, that he does not know it. At 172e Socrates no longer gives a reason the Greek does not give, and at 174d8 “if it really is, in the fullest sense” now matches the same idiom at 172e and 173a. Two small jokes were restored at 175c: the ironic “if you please” and “simple-minded” for the self-mocking euēthikoi. Smaller corrections were made at 174c9 and 176b4, and the section letter was added at the head of page 32.

In the notes, two statements were wrong. Plato does not treat draughts as a trivial pursuit; he counts it a real skill, and at Republic 487b–c Adeimantus compares Socrates’ own questioning to it. And Critias and Charmides did not take Athens by force: the Thirty were voted in, under Spartan pressure, and then ruled by force, and Charmides was not one of the Thirty but one of the Ten set over the Piraeus; both died at Munychia. The note on the oath “by the dog” now gives the ancient explanation, the “oath of Rhadamanthys”, with Porphyry’s denial that it was a joke. The parallels in the Laches and Euthydemus are now stated exactly, and the notes on alazones and on the insolent inquiry label their interpretations as such.