The Orphic Collection, Volume I: The Hymns

A Note on This Edition

How it was made

In fourteen blocks of three to five of Abel's pages — forty-six pages in all — over some six days. Each block was transcribed from the page images at magnification, translated, annotated, typeset, and then checked by three independent verification passes — one reading the Greek and the apparatus against the images, one checking the English against the Greek line by line, one testing every claim in the notes against the other eighty-eight pieces of the collection. Every finding about what is on the page was then re-cropped and confirmed before it was adopted. Each block's corrections were written up in full, as a separate document, and delivered with it; those fourteen files are the edition's audit trail.

Then the final run: the whole translation read through as one text, terms checked across the block boundaries where nothing had been checking them, and the volume rebuilt from the data rather than stapled from the block proofs.

What the final run caught

— Rendering settled and applied backwards over the whole volume — about a hundred and fifty lines changed. The largest single item was the five words for holy: ἁγνός, ἱερός and ὅσιος had all three been collapsing into English holy through the first eleven blocks, and forty-seven lines had to be repaired. Beyond that: εὐδύνατος, βουλή, γνώμη, ἠπιόχειρ, κράντειρα, ὀβριμόθυμος, ἀγλαόμορφος, ἀγλαότιμος, ἀγλαόφημος, ὑγροκέλευθος, πολύολβος, πολυώνυμος, μεγαλώνυμος, πολύσεμνος, πολύμορφος, κουροτρόφος, τιθήνη, ὀλβοδότης, ἀμεμφής, ἀκάματος, ἀτειρής, κόσμος, ἀνάγκη, φαεσφόρος, παμβασίλεια, σωτήριος, νυκτέριος, παννυχίς; the transliteration of the Charites, the Horai, the Moirai, Gaia, Ouranos and Nyx; δαίμων, which was translated at four of its twenty-nine places; ὄργια, translated at two; and one spelling for the Bacchic cry, which had been euhoi, euhoe and euoi in the same book.

— Two inconsistencies in the rubrics. Eight of the ten rubric-less hymns printed an italic (no fumigation named) and two printed nothing; all ten now print it. And two of Abel's thirteen Greek rubrics were being Englished two ways each — θυμίαμα μάνναν as both manna and frankincense-powder, θυμίαμα λιβανομάνναν as both frankincense-manna and frankincense and manna — which had turned a census of the offerings into a wrong statement about the collection.

— A typesetting fault invisible in the proofs. A twelve-line hymn split 11 + 1 across a spread, leaving a page carrying one verse. The volume balances its spreads instead: twelve lines go 6 + 6, seventeen go 9 + 8.

— Abel's praefatio, read for the first time at the final run, which settled the identity of Wiel as far as the book allows, explained Sanctamandus, confirmed that the hymn-numbering is Hermann's, produced a second and milder statement of Abel's doubt about LXXXVIII — and, most important, established that his apparatus is by his own statement a selection, which weakened one of this edition's own notes.

What it did not catch, and what remains

Stated plainly, because an edition that claims to be finished and is not is worse than one that says where it stops.

— Wiel and Dinner are not identified. Wiel is third in the apparatus and is named once, in a Latin parenthesis, with six dates and no first name. Dinner is not named outside the apparatus at all. Neither can be taken further from this volume.

— One apparatus entry was never read from the page. At XLV.4 the lemma and its numeral are covered in the copy used. Abel's own note says so — lemma et numerus in exemplari adhibito obtecti — and the edition prints the lemma restored between angle brackets and flagged, rather than silently. It is the one place in the volume where something is printed that no one has seen.

— Slothouwer's place of publication was not established.

— The Hekate / proem division is followed on Hermann's and Abel's authority, and the alternative — a proem of fifty-four lines, which is what Thomas Taylor translated in 1787 — is recorded in the note at Hymn I but not printed.

— The collation of I.1 left one disagreement standing: whether the reading Abel rejects there is ἐραννήν or the manuscripts' ἐραυνήν. Recorded rather than resolved.

— Quandt's numbering is not concorded. Abel's is followed throughout.

What is delivered with it

The decision register and the running observations ledger, which are not bound into the volume but are its working record: every ruling with the date it was taken and the evidence behind it, every claim that was made and withdrawn, and the standing lessons — of which the one that cost most was learned twice. A claim about what is on a page is worth exactly what the image behind it is worth; a correction is a claim like any other, so it must be tested on the case that would have falsified it — this edition got a margin rule wrong, corrected it, and got the correction wrong in the same way before a third reading settled it; and a ruling taken at block thirteen is worth nothing until it has been carried back to block one, which is the whole reason the final run exists and is what it mostly spent its time doing.

Volume II will be Musaeus.