Textual issue in Philo of Alexandria

Greek Learning Greek
1

φιλόλογοι·

I am working on textual issues in Philo of Alexandria. I would value your thoughts on/gut reactions to a problem in his Life of Moses.

Background:

Philo is concluding his symbolic interpretation of the high priest’s clothing (cf. LXX Exodus 28). He understands the outfit as a representation of the κόσμος: the full-length robe is a μίμημα of the air, the pomegranates of the water, etc. The textual issue is in bold. Mss read μίμημα or μιμημάτων; one editor conjectured μιμήματα. What do you think is the best reading?

τοῦτον τὸν τρόπον ὁ ἀρχιερεὺς διακοσμηθεὶς στέλλεται πρὸς τὰς ἱερουργίας, ἵν’, ὅταν εἰσίῃ τὰς πατρίους εὐχάς τε καὶ θυσίας ποιησόμενος, συνεισέρχηται πᾶς ὁ κόσμος αὐτῷ δι’ ὧν ἐπιφέρεται μίμημα/μιμημάτων/μιμήματα, ἀέρος τὸν ποδήρη, ὕδατος τὸν ῥοΐσκον, γῆς τὸ ἄνθινον, πυρὸς τὸ κόκκινον, οὐρανοῦ τὴν ἐπωμίδα, καὶ κατ’ εἶδος τοῖν δυοῖν ἡμισφαιρίοιν τοὺς ἐπὶ τῶν ἀκρωμίων σμαράγδους περιφερεῖς, ἐφ’ ὧν καθ’ ἑκάτερον γλυφαὶ ἕξ, τοῦ ζῳοφόρου τοὺς ἐπὶ τῶν στέρνων δώδεκα λίθους ἐκ τριῶν κατὰ τέτταρας στοίχους, τοῦ συνέχοντος καὶ διοικοῦντος τὰ σύμπαντα τὸ λογεῖον.

2

Original μίμημα αὐτῶν (but elided) motivating the nonsense μιμημάτων reading. (Maybe τούτων, if the δἰ ὧν isn’t strong enough to bring in the mental presence of the genitive plural for πᾶς ὁ κόσμος.)

3

δἰ’ὧν ἐπιφέρεται μιμημάτων is definitely not a nonsense reading; in fact, it’s pretty straightforward, and I think it’s probably right. This is the common syntactic structure known as “attraction”: the relative ὧν is “attracted” into the case of its antecedent, instead of διὰ μιμημάτων ἃ ἐπιφέρεται. δι’ needs a genitive complement, and ὧν needs a plural antecedent. “the entire universe enters with him through the symbols he wears . . .”

I can see μίμημα, with ὧν indefinite, but still genitive by “attraction” into the case of the understood complement of δι’, διὰ [τῶν] ὧν [instead of τῶν ἃ] ἐπιφέρεται, μίμημα ἀέρος τὸν ποδήρη, [μίμημα] ὕδατος τὸν ροίσκον, κτλ – “the entire universe enters through the things he wears: the symbol of air, the full-length robe; of water, the pomegranate; . . .” etc. The punctuation would be different: a comma after ἐπιφέρεται instead of after μιμημάτων.

On balance, I prefer μιμημάτων because it gives slightly more salience to the fact that the high priest is wearing symbols, rather than simply enumerating them.

I find μιμημάτα harder to swallow.

The accusatives that follow, ποδήρη, etc., are direct objects of ἐπιφέρεται.

4

That does makes sense. Surprisingly enough, I have come across attraction once or twice (what are the odds? It’s astounding to imagine). Anyway, I was reading without attraction “the entire cosmos enters with him because of the things of which he bears an image”. “Through” as you have it does make sense, but the successive ἀέρος, ὕδατος, γῆς, etc., seem to make the genitive a bit awkward following μιμημάτων. Not impossible, just a bit busy.

5

Thank you for the discussion! Here’s my current thinking, with a few responses to comments above.

(1) To nitpick, it would be an instance of “incorporation” (where attraction is also required)—i.e. an internally headed relative clause. The unincorporated form would be διὰ τῶν μιμημάτων ἃ ἐπιφέρεται. Note the def article. Internally-headed relative clauses have to be both definite and restrictive (as your translation suggests). The shift from genitive to accusative objects of ἐπιφέρεται, in loose apposition to μιμημάτων, would be slightly awkward but understandable. He had to avoid an endless cluster of genitives—ἀέρος τοῦ ποδήρους ὕδατος τοῦ ῥοίσκου κτε.

(2) Alternatively, one could construe it as an instance not of incorporation but of omission of the antecedent, ie *δι’ ἐκείνων ὧν ἐπιφέρεται μιμήματα (“through the things whose representations he wears”). In that reading the accusatives that follow are less awkward.

(3) Construing μίμημα with ἀέρος (and so punctuating δι’ ὧν ἐπιφέρεται, μίμημα ἀέρος…) would leave an unacceptable hiatus. It is also perhaps easier to imagine the loss of -τα/-των than their addition. This reading is the least likely.

(4) Finally, you could try to take μιμήματα as predicate accusative. The relative pronoun ἅ is attracted to the case of an omitted antecedent. Without attraction and omission it would be *δι’ ἐκείνων ἃ ἐπιφέρεται μιμήματα (“through the things that he wears as representations). The retained accusative μιμήματα would be most awkward, but not unimaginable in Post-Classical prose, when there were all sorts of changes to internally-headed relative clauses, including mismatch between case markers. Still, this seems improbable for Philo.

So, the choice is between μιμημάτων (1) and μιμήματα (2). At the moment, I prefer μιμημάτων. I’m happy Hylander [edit: and mwh] think so too. Unlike the other readings, δι’ ὧν ἐπιφέρεται μιμημάτων is perfectly straightforward. Philo just wrote himself into a minor syntactical hole with the accusatives that follow, leaving the sentence less elegant than he had hoped for. Still, μιμήματα may well be right.

Any corrections to the above are most welcome. I’m still thinking this through.

6

For what’s it’s worth, I agree with Hylander. μιμηματων seems the best reading by far. I hesitated for a moment over the succeeding ἀέρος τὸν ποδήρη etc., but only for a moment (the accusatives revert to the normal case for direct objects, and the genitives are objective genitives with μιμηματων). My only quibble with Hylander’s analysis is his addition of the definite article, which I think must be a simple slip.
[This posted before seeing phalakros’ response to the various suggestions.]

7

exactly

8

With regard to μιμήματα, taking it as a predicate accusative, as you note, seems very awkward. The alternative, “through the things of which he wears representations,” seems to make less sense to me. It’s the representations — the robe, the pomegranates, etc. — that enter, worn by the high priest, not the things represented themselves — air, water, etc. — that enter. The things represented are the cosmos. The whole cosmos is entering symbolically through the representations, but the whole cosmos is not entering actually. So μιμημάτων makes more sense to me.

But perhaps strict logic isn’t apposite here, and the idea is that the represented things are entering actually, not just symbolically, by virtue of their representations.

9

Yes, that’s roughly how I would understand it under this reading (my #2). He would be emphasizing the signified (the universe) rather than the signifiers (the garments). It’s not strictly logical—hardly surprising in a context like this. Leopold Cohn proposed μιμήματα in his edition of Mos (still by far the best ed. and unlikely to be fully replaced—he was a very good philologist). I always hesitate to depart from his decisions, though he of course made mistakes.

Incidentally, the widely used Loeb edition of this text suffered from the heavy influence of our infamous Rouse (one of the original editors of the LCL). There are a number of outlandish emendations attributed to him.

Thank you for lending your keen eyes, Hylander and others!

10

Rouse was the original editor-in-chief of the Loeb series, personally selected by James Loeb. Some of the original Loebs are pretty atrocious. But the effort and expense of completely redoing them would seem prohibitive, except in the case of major authors. One of my college professors, G.P. Goold played a role in reviving the series as editor-in-chief in the 1980s. He was an outstanding Latinist and a great teacher. He revised Catullus and Tibullus, and Ovid’s Amores and Heroides, and produced his own new Propertius.

11

Though I think that you could take ὧν as a not-attracted accusative and not need the unlikely punctuation, I’ve been trying to figure out what is meant here about hiatus. I don’t know anything about Hellenistic hiatus rules, but Philo seems to do words like μίμημα followed by a vowel pretty often, doesn’t he? What is particular about this one?

12

Indeed. And even Chariton! Did you study with him at Harvard or Yale?

I’m not sure I follow. Do you mean genitive?

In practice—regardless of how it is written—a short -α can be presumed to be elided in pronunciation before another vowel to avoid hiatus (so μίμημ’ ἀέρος). This is the usual practice in Philo and roughly similar prose writers like Plutarch and Philodemus (the latter is a bit stricter; none of them as scrupulous as Isocrates). Scratch the point about hiatus from my post above. It was a slip–there is no hiatus.

13

Scratch the point about hiatus from my post above.

I guess the utrum in alterum point still counts against μίμημα. μιμήματα is awkward, and, as I wrote previously, I don’t think it makes makes good sense, though, again, perhaps strict logic isn’t to be expected in this context. That still leaves μιμημάτων the front-runner.

Why don’t you post more of these if you have the time?

Did you study with him at Harvard or Yale?

I would be exaggerating if I said I studied with him As an undergraduate at Harvard (and that’s as far as I went academically in Classics), I took two courses from him in 1967-8: Advanced Latin Composition and Roman Elegy. He was very accessible to everyone. I still have readings he suggested penciled in my copies of Ovid’s Amores and Propertius. I’ve been going through Propertius, triangulating between his Loeb, Heyworth’s OCT and companion, and occasionally the old Barbers old OCT with some of Goold’s readings. In his Loeb, Goold adopted many more readings that differ from Barber. Propertius is a puzzle. I don’t think any Greek or Latin text I’ve seen is as wretchedly preserved as Propertius – even Catullus.

14

I will try to. Thanks again for your input!

15

I’m not sure I follow. Do you mean genitive?

My understanding is that if ὧν is a genitive object of μίμημα here, referring to the discrete elements of the universe, air/water/earth/fire/heaven/god then we have either:

διὰ [τούτων] ὧν ἐπιφέρεται μίμημα, ἀέρος τὸν ποδήρη…
διὰ [ταῦτα] ὧν ἐπιφέρεται μίμημα, ἀέρος τὸν ποδήρη…

by these that he wears symbol of, of air the robe…
because of these that he wears symbol of, of air the robe…

6 to 1, half dozen to the other in this context, but I personally find understanding “διὰ [ταῦτα] ὧν” nicer because it makes the flow of cases so nice. The genitive is, of course, not at all busy. In both though, we have the nice flow of ὧν μίμημα repeated in the following examples ἀέρος τὸν ποδήρη, etc. The word order is easy to understand.

Against διὰ [ταῦτα] ὧν, though, I read through examples of δι’ ὧν in Philo, but found only examples of δι’ ὧν = διὰ τούτων ὧν, and no examples that were clearly διὰ ταῦτα ὧν. On the other hand, these forward-leaning reletives were somewhat rare, with the unstated τούτων/ταῦτα antecedent really referring to something about to be introduced.

μίμημα (as opposed to μιμημάτων) would motivate ἀέρος τὸν ποδήρη as the first listed term, as Philo often seeks alpha-words to follow it (see below).

The difficulty, to me, with μιμημάτων isn’t just the busy genitive, but also that the word order of the terms of the following list seem backwards. With μιμημάτων we understand “by means of the symbols he bears”, and the word order of the following becomes ever so slightly hard to justify:

…ἀέρος τὸν ποδήρη, ὕδατος τὸν ῥοΐσκον, γῆς τὸ ἄνθινον…

The symbols were the object of the last phrase, but now it is as if he has begun going through an exhaustive list of elements one by one, and naming the symbol for each. That just seems a little backwards. Not impossible, but another awkwardness along with the very busy genitive, as well as the attraction (μιμημάτων), unattraction-back-to-accusative (ποδήρη, ῥοΐσκον, …) dance going on for the objects of the verb.

μίμημα vowel-junction in Philo.

There appear to be 52 occurrences of “μίμημα” in Philo, that is this particular nom./acc. singular form.

As usual (at least it’s usual in Homer), elision before the same vowel is more frequent than other kinds:

4: μίμημα ἀρχετύπου
2: μίμημα αἰσθητὸν
1: μίμημα αἰῶνος
1: μίμημα, ἀλλ’ αὐτὸ

Versus:

1: μίμημα ἐνδυόμενος
1: μίμημά ἐστιν ἡ ἱερὰ
1: μίμημα ὁ ποιῶν

We have vowel-vowel junction (resulting in elision) 21% of the time, with α-α junction occurring in 73% of those cases.

16

Just a quick note. I follow what you were saying now: not that “ὧν is a non-attracted accusative” but that it’s a gen relative pronoun with an omitted antecedent in the acc, διὰ (ταῦτα) ὧν. That strikes me as grammatically impossible.

17

it’s a gen relative pronoun with an omitted antecedent in the acc

Smyth says it’s normal, not impossible, and that when the omitted antecedent is nominative or accusative the relative retains its own case. Ex. from Smyth: στυγῶν ἥ μ’ ἔτικτεν = στυγῶν [ἐκείνην] ἥ μ’ ἔτικτεν.

Following this pattern, antecedent ταῦτα is omitted, with relative ὧν retaining its genitive. HOWEVER–

It’s around prepositions that things get squidgy. I mean, something like “ἐν ὧν” would obviously be too disconcordant to fly (though a search gives it once in Anabasis: ἐν ὧν τῇ χώρᾳ ἐσμέν). But διά, taking accusative or genitive as it does, may behave differently. Like I said, I couldn’t find any examples. As it doesn’t matter, I agree on taking the other branch of my argument for μίμημα: διὰ [τούτων] ὧν.

18

Yes, the preposition in διὰ (ταῦτα) ὧν is what I was referring as probably not grammatical–gen relative pronoun with omitted antecedent in the acc is totally normal otherwise. ἐν ὧν τῇ χώρᾳ ἐσμέν is unrelated and unproblematic (ἐν just governs τῇ χώρᾳ)

19

Again, μιμηματα doesn’t make as good sense as μιμηματάτων. It’s through the symbols the high priest is wearing that the whole universe enters with him, not through the things the symbols represent, which are not entering with him. δι’ὧν with understood accusative ταῦτα or ἐκεῖνα sandwiched in between is impossible. Genitive ὧν must be the object of δι’, and δι’ must mean “through,” not “because of/on account of.”

20

Indeed. I hope this was clear in what I said above, hastily.

I think we’ve done this one to death. I’ll post another textual problem when I can.