Pseudo-Josephus, “A Discourse to the Greeks concerning Hades” – an investigation

There is a text floating around the web under the title of “Josephus’s Discourse to the Greeks concerning Hades.”  The full title is “An extract out of Josephus’s discourse to the Greeks, concerning Hades: wherein are contained the souls of the righteous and the unrighteous.”  Bill Thayer has the most reliable version of the piece on his site, here.

This item is from William Whiston’s 1737 version of the complete works of Josephus, and in the original edition it was printed as Dissertation V in volume one.  (Dissertation VI defended its authenticity).  While omitted in most reprints, these dissertations can be found in the 1741 reprint.^[1] But the text is plainly not by Josephus, not least because chapter 8 starts with an apocryphal saying of Jesus.  This is found in other patristic texts and reads, “In whatsoever ways I shall find you, in them shall I judge you entirely.”  So we may call it pseudo-Josephus.

We need to find the Greek text for this, before we can discuss it.  In his original edition, Whiston gives a note about this, which leads us down some quite interesting rabbit holes.  As we shall see, it leads to a number of what appear to be unresolved issues.

In the reference literature today, the text is assigned to Hippolytus of Rome, and given a title such as “Adversus Graecos” or “Oratio ad Graecos de inferno”.  It is hesitantly classified as fragment 1 of the lost “De universo,” περί τοῦ παντός, (Clavis Patrum Graecorum CPG 1898).  This classification is based on Photius, Bibliotheca 48, which mentions such a work and ascribes it to Josephus, just as the manuscripts do today.  But Whiston, despite his cranky ideas, was right when he noted that the basis for the proposed identification with De universo was unsound, as Alice Whealey recently discussed.^[2]

The text is printed in the Patrologia Graeca 10, cols. 796-801, with a Latin translation copied from Etienne Le Moyne, of whom more anon. The first half of the same text appears in PG 96, 541-544, labelled as the “Rupefucaldinum” version of John Damascene’s Sacra Parallela (CPG 8056).  There is a modern edition of our own text in K. Holl Fragmente vornizanischer Kirchenvater aus den Sacra Parallela (TU 20, 2), Leipzig (1899), pp.137-143.

Whiston’s translation predates all of these editions.  Ordinarily that would not matter, but in this case, as we shall see, it does.

Whiston encountered the text in the Addenda of the 1726 Havercamp edition, the basis for his translation of the works of Josephus.  This piece was in vol. 2, Addenda, p.145-7.^[3]

But rather than using Havercamp’s text, he preferred that printed by David Humphreys, The Apologeticks of the Learned Athenian Philosopher Athenagoras, (1714).  This had a loose English translation on pp.292-9 and the Greek on pp.302-307.  Humphreys also collated it with a text printed in the end notes of David Hoeschelius, Photii Myriobiblon (1611).  The Notae at the end restart the page numbering from 1, so its on columns 9-12 at the back of the book.  Whiston mentions also Etienne Le Moyne, Varia sacra, ceu Sylloge variorum opusculorum graecorum, vol. 1 (1694), pp.53-62, whose Latin translation was adopted by Migne in the PG10.  Le Moyne attributes the text, without manuscript authority according to Whiston, to Hippolytus, Sermone contra Graecos, cuius titulus, contra Platonem de universi causa.

But here the mysteries begin.  Whiston prints the text in 8 chapters – the division and numbering are his own –, based upon the text of Humphreys.  But although Whiston himself states, “All the four copies … very nearly agree, till towards the latter end of § 6,” after that we have a problem.

I have only skimmed the texts, but it looks as if all the editions seem to agree as far as the last-but-one sentence in chapter 6.  Hoeschelius, the PG10 text, Le Moyne, and Holl also print a section of text equivalent to chapter 7.  The chapter 7 of Humphreys is rather different.  The chapter 8 of Humphreys is not present anywhere else.

Inevitably we have to ask what manuscript evidence each of these editions is based on.

The Pinakes database lists 10 manuscripts of our text:  Barocci MS 26 (9^th c.) in the Bodleian Library in Oxford, BNF Coislin 131 (14^th c.) in Paris, one in the Lavra on Mt Athos (K 113, 16^th c.), Vatican gr. 723 (13-16^th), and 6 more recent MSS also in the Vatican.  This from the online scan of a microfilm of Vatican gr. 723:

So we can now look at the editions.

The Humphreys edition is based on MS Oxford Barocci 26.  Whiston tells us that this becomes nearly unreadable toward the end, but that this is the source of the material beyond chapter 6.  Sadly it is not online. He adds that the Barocci copy “is much the most valuable; because it is about a forth part larger than the other; and yet appears equally genuine.”^[4]

Whiston also refers to the Coislin manuscript, which is online here.

Hoeschelius used an Italian manuscript, without specifying it.  He says only:

“Eius fragmentum ut ex Italia missum est pridem mihi a M.M. ita edo, pseudepigraphum, nec ne (non iniuria enim dubites) iudicent eruditi.  Equidem homini Christiano adscripserim.”

“I publish this fragment, just as it was sent to me some time ago from Italy by M.M., pseudepigraphical or not (for you would not be wrong to doubt), let the scholars decide. For my part, I would attribute it to a Christian author.”

So this was probably one of the Vatican manuscripts.  Whiston in “Dissertation 6” p.clxxxv tells us that “M.M.” was a certain Max Marguntius.

As far as I could tell from his rambling preface, Le Moyne does not identify the manuscript that he used for “Hippolytus”.

Pinakes also lists 39 manuscripts of the Sacra Parallela, from the 9^th century onwards.  This work exists in various different recensions, and I am unclear whether pseudo-Josephus is present in all of these manuscripts.  Paris BNF gr. 923 (9^th c.) and Venice Marciana gr. Z. 138 (10 med.) were used for the Holl edition, and there are three other manuscripts listed of recension “PMLb”, whatever that is.  But the mysteries of the Sacra Parallela will have to await another blog post.

It looks as if we have at least two recensions, and possibly three; the Barocci text via the Humphreys edition, the manuscript itself nearly unreadable at the end; the PG10 / Hoeschelius / Le Moyne text; and perhaps the Holl and PG96 text as a third group similar to the second.  Someone needs to collate these, collate the manuscripts, and establish a stemma.

I notice that even in the first sentence of the work, we find a separating variant.  Humphreys prints “Καὶ οὗτος μὲν ὁ περὶ δαιμόνων τόπος,” “And this is the place of demons.”  But other witnesses read “Καὶ οὗτος μὲν ὁ περὶ δαιμόνων λόγος,” translated by Whiston as “And this is the discourse concerning daemons.”

• “topos” is the reading of Humphreys (and so, presumably, the Barocci MS), Hoeschelius, Le Moyne, and the PG10. It is also the reading of the Coislin MS, from which Whiston must have taken his reading.
• “logos” is the reading of Holl, and also PG96, both based on the Sacra Parallela.

So there is editorial work to be done on the “ad Graecos”, or whatever we call the text.  The text, its authorship, and relationships between the text witnesses, all this would make an interesting subject for publication.  But not by me!

1. ^[1]William Whiston, The Genuine Works Of Flavius Josephus, The Jewish Historian. Translated from the Original Greek : according to Havercamp’s accurate Edition, vol. 1 (1737), pp.clxxix-clxxxiv.↩
2. ^[2]A. Whealey, “Hippolytus’ lost ‘De universo’ and ‘De resurrectione:’ Some new hypotheses”, in: Vigiliae Christianae 50 (1996), p.244-256.  JSTOR.↩
3. ^[3]S. Havercamp, Flavii Josephi quae reperiri potuerunt, 2 vols (1726). BSB: Vol. 1. Vol 2.  The text begins on p.676 of the PDF download of vol. 2.↩
4. ^[4]In “Dissertation VI”, p.clxxxvi.↩