In the last fifty years or so, the editing of Valerius Flaccus’ Argonautica has seen remarkable progress: if Courtney's 1970 edition was based on the assumption that the only manuscript with independent value was V, Ehlers's 1980 edition also recognized the independence of S and L, and finally Liberman's 1997–2002 edition re-established the (mostly) lost codex Carrionis as yet another independent witness.Footnote 1 Taylor-Briggs concludes her masterful account of Valerius’ manuscript tradition with the following statement: ‘The work of Valerian textual critics seems to be drawing to a close: the twentieth century opened with only one manuscript considered to be of any value, but closed, much sweat and ink later, with a much more sophisticated understanding from which to reconstruct an accurate text.’Footnote 2 The implication appears to be that, once we have established the paradosis (which more or less equals reconstructing the archetype), there is only so much left for us to do. The 2017 edition by Mariné Isidro seems to confirm this impression: while it is based on the same stemma as Liberman's moderately radical edition, it mostly reverts to Ehlers's conservative text, with only a handful of differences.
But is this impression accurate? There is no doubt that the archetype of the Argonautica was already gravely corrupt, and although reconstructing it is a necessary task, it does not bring us all the way back to Valerius’ autograph. If so, why is it that we might consider our job all but completed? Two answers seem possible, but neither is convincing. First, one could submit that, over the centuries since the 1474 editio princeps (and even before) during which Valerius’ text was persistently subjected to scrutiny, most errors of the paradosis have already been eliminated. While it is true that many good corrections have been found and universally adopted, it is unlikely that there only remain a few unsolved problems—if for no other reason than that we have only had an accurate idea of the paradosis for less than fifty years. The alternative (or perhaps complementary) answer is that the remaining problems are unlikely to find satisfactory solutions ever at all. Again, one could respond that, with less than fifty years of critical engagement with the paradosis, it is perhaps too early to give up.Footnote 3
In what follows, I propose to discuss about a dozen passages from the eighth book of the Argonautica, a selection which in no way is intended to exhaust all of the book's remaining textual difficulties. While I realize that individually some of my proposals may be less successful than others, cumulatively I hope they do show that there is still ample scope for finding new solutions and identifying new problems.Footnote 4 At the same time, I do believe that, especially when dealing with a text as uncertain as Valerius’, it may be more profitable to ask, before considering textual intervention in any given instance, not whether we can be certain that the paradosis is corrupt, but whether we can be certain that it is intact; and unless one can give a positive answer to the latter question, one should be open to forming a hypothesis as to what Valerius may be likely to have written instead.
For each passage I consider, I first quote a version of the current ‘vulgate’, with a minimal apparatus; γ is the common source of V and L (S does not exist in Book 8), ω is the common source of γ and the codex Carrionis (Δ is its surviving fragment).Footnote 5
ARGONAUTICA 8.20–1
Book 8 opens with Medea deciding to abandon her home and flee to Greece with Jason and the Argonauts. After taking leave of her absent father (10–15), Medea gathers her magical drugs (16–19) and makes haste to meet Jason (20–2):
One well-known problem concerns ω's eiecta: while not absolutely impossible, eiecta has too physical a meaning (‘thrown out, expelled’), and given the ease of the corruption, modern editors tend to prefer erecta from the editio princeps. Yet erecta is not without weaknesses either: on the one hand, taken literally, it confusingly conveys the wrong sense (‘raised’: Medea must get up from her bed, but before she collects her drugs); on the other, the psychological sense (‘roused, excited’) seems usually to have more positive connotations (note, for example, 3.631–2 tali mentem pars maxima flatu | erigit, of the Argonauts regaining confidence after Jason's admonition). Superior to both is, I believe, euecta, cited by Carrion from a Leiden edition but apparently forgotten by modern editors: it has just the right sense (OLD s.v. eueho 3: ‘[of an emotion or other impulse] To carry away [to excessive or extreme action]’, note especially Sen. Phaedr. 1070–1 quacumque rabidos [sc. equos] pauidus euexit furor, | hac ire pergunt), and palaeographically is barely less likely than erecta; at the very least it should feature in the apparatus criticus.Footnote 6 A second problem, usually unrecognized as such, is the repetition of prosilit at line 21, with different subjects (first Medea, then Ino) and in different senses (first ‘rushes forth’, then ‘springs forth’). Liberman finds the repetition effective, but to me it rather seems to destabilize the text, by misleadingly suggesting that Medea may be actually jumping, just like Ino.Footnote 7 I suggest that the first prosilit may be an error for proruit (ru → ſıl, not without input of course from the following prosilit), though the verb is rare in the sense ‘to rush forward’ (OLD s.v. proruo 2; cf. TLL 10.2.2166.20–32).Footnote 8 Alternatively, one might propose proripit (in imitation of Virgil's absolute usage, note Aen. 5.741 quo deinde ruis? quo proripis?).
ARGONAUTICA 8.83–4
Medea takes Jason to the sacred grove in which the golden fleece is guarded by the dragon; after failing to subdue it by invoking Sleep (69–82), she turns to hardcore magic (83–7):
Two difficult moments here. First, spumare at line 83: foaming is a typical effect of uenena (cf., for example, 6.447 Atracio lunam spumare ueneno), but it is difficult to see how ‘foaming with poison’ can plausibly be construed to refer to employing magical drugs.Footnote 9 Heinsius's cumulare and sputare may be mentioned as diagnostic conjectures, but much likelier would be spirare, ‘to blow drugs (at)’, which can especially be supported with 6.157 paribus spirans Medea uenenis, as well as with 7.327 magicis spirantia tecta uenenis and Luc. 9.679–80 quanto spirare ueneno | ora rear (of the Gorgon): evidently in our context spirare will imply a manner of spreading magical substances with one's breath.Footnote 10 The corruption would be quite easy in minuscule script (spırare → spūare), and can in fact be paralleled, in the opposite direction, at Claud. Rapt. Pros. 1.283 aegra soporatis spumant [spirant] obliuia linguis; scribal reminiscence of 6.447 may have been an additional factor. The second weak point is cuncta at line 84: it is rather vague (what are these ‘all silences’?), while its positioning makes it unlikely that it can have a predicative force (‘to the last drop’), which would also run contrary to the imperfective aspect of the construction with perstat. Liberman's tincta, however, is hard to construe.Footnote 11 I suggest muta: it is a fitting epithet for silentia (cf. Ov. Met. 4.433, 7.184, 10.53, but especially Stat. Theb. 10.92–3 pressisque Silentia pennis | muta sedent), and the corruption, perhaps by way of multa, would be fairly easy (for muta → multa, cf., for example, Lucr. 4.1057; for multa → cuncta, for example Anth. Lat. 286.83; cf. also muti → cunctis at Juvencus, Evang. 1.111; one may also suspect influence from contra right above it).
ARGONAUTICA 8.89
Finally the dragon succumbs (88–91):
Line 89 hosts a well-known (if underestimated) problem: on the one hand, ceruix lacks an expressed predicate; on the other, extra sua uellera is an odd and imprecise expression (what exactly does it mean that the dragon's neck is—goes?—‘beyond’ the fleece?). Liberman's itque for atque addresses the first issue but not the second.Footnote 12 It seems natural to suspect that extra conceals the missing verb, and Löhbach's laxat—not even mentioned in any of the editions and commentaries from the last hundred years or so—is exactly what we need: the dragon's neck lets go of the fleece (OLD s.v. laxo 4).Footnote 13 The corruption may not be the most straightforward one, but it is not difficult to imagine laxat losing its initial l after the final ſ of ingens, following which axat would have a fair chance of being interpreted as an error for extra (or perhaps for an abbreviation thereof, such as ext̄).
ARGONAUTICA 8.127
Jason returns, with Medea and the fleece, and joins the rest of the Argonauts at a pre-appointed place (127–9):
The deixis of tunc seems awkward (there is no need to specify that the action was taking place ‘then’, as there is no reason to assume that the context may refer to some other time), and I fail to see why Valerius would not have written iam; the confusion is, of course, quite easy (cf., for example, tunc for iam: Ov. Met. 6.52; tum for iam: Verg. Aen. 3.531, 12.239; dum for iam: Ov. Met. 6.467; iam for tum: Verg. Aen. 5.382; iam for cum: Ciris 513, Ov. Am. 2.11.25).
ARGONAUTICA 8.158
Having heard of Medea's elopement, her mother breaks out in an exasperated monologue, first blaming Jason and then her daughter (158–9):
The most obvious problem here is metrical: in classical Latin poetry ego never forms an iambus; the easiest solution is Müller's o added after ego, which, of course, is but a metrical filler. The second problem, insufficiently appreciated, is the indefinite quemquam: it could mean something like ‘why am I ready to blame anyone (but my daughter) with random accusations’, but Jason is surely not just ‘anybody’.Footnote 14 The old correction by Columbus—sed quid ago quemue (quemque Habenicht)Footnote 15—solves both these problems, but brings out a third one: it is rather magnanimous of Medea's mother, and not quite in tune with her preceding words, to claim that Jason does not really deserve her reproaches (immeritis). Courtney seems generally right about the expected sense: ‘this is the line by which Medea's mother turns from accusation of Jason to accusation of her daughter, and what we want is something meaning “I am blaming the wrong person”’.Footnote 16 I propose to make three minor changes and read: sed quid ego hunc, quamquam meritis, incuso querellis? ‘but why am I blaming him, even if he deserves my reproaches?’—implying that, though Jason is far from innocent, Medea should be held no less accountable for her own actions (ipsa fugit). The omission of hunc would be quite easy, especially in elision (for a part of hic after quid ego, cf., for example, Verg. Aen. 2.101 sed quid ego haec autem nequiquam ingrata reuoluo?); the misreading of quamquam as quemquam can hardly surprise; the negative prefix in immeritis is no doubt just a reinterpretation of the last letter of quamquam, read twice (m → ın).
ARGONAUTICA 8.165, 167–8
Medea's mother continues (165–9):
First, I am worried by prodita: the verb seems normally to emphasize the role of one revealing, rather than that of one learning, a secret, and I find it odd that Medea's mother should be complaining that no one told her of Medea's infatuation rather than that she failed to recognize it herself; cognita would be much more natural.Footnote 17 A much more serious problem is, of course, the articulation of the whole passage: I think Liberman is right to adopt Koestlin's aut at line 166 and treat consideret and fuisset as two coordinated alternatives (had she known, she would either make Jason stay or follow him together with Medea).Footnote 18 I find, however, the placement of the second aut at the beginning of line 168 suspicious (a postponed conjunction should as a rule be prosodically attached to the fronted phrase): the sentence structure will become much clearer if we read aut in front of commune instead, while replacing aut at line 168 with hoc (omne nefas might more naturally mean ‘all kind of disaster’, whereas the demonstrative adds desired specificity: ‘all this disaster’).Footnote 19 Monosyllables are generally mercurial, but in this case one might speculate that the first aut was omitted (in line 167), then written in the margins, and finally restored in the wrong place, ousting hoc in line 168.
ARGONAUTICA 8.176
The Argonauts sail away (175–7):
I fail to see how line 176 can possibly be correct (pace OLD s.v. transcurro 2b, among others): even if transcurrere could be taken as a historic infinitive, it is difficult to imagine across what it could be said that lands are moving, even from the perspective of the Argonauts.Footnote 20 I think it necessary that transcurrere should have the Argonauts as its subject, which means that notae … terrae should be converted to the accusative (notas … terras), and I think it likely that Minyis has ousted a finite verb. The sense should continue the idea of the previous clause: the wind that propels the Argonauts homewards is more welcome (than the one that carried them away), and they are glad—gaudent—to be sailing past the lands they had already seen (that is, to be sailing back).Footnote 21 Minyis could be an interpolated gloss, originally intended to clarify redeuntibus (or perhaps notas?); the change of notas … terras to the nominative would be a subsequent development, probably meant to harmonize the phrase with aura.
ARGONAUTICA 8.212
The landscape the Argonauts are passing by even seems sympathetic to Medea's plight (209–12):
As the apparatus criticus shows, the second half of line 211 is textually unstable, and the construal of the clause is uncertain. The usual approach is to accept V's ponunt, but it is open to two objections: first, it is odd that the Argonauts should take the climactic place, accentuated by quoque, in a list of natural features that (by implication) would be unlikely to be sympathetic to Medea; second, it is not clear what should be understood as the direct object of ferre.Footnote 22 Castelletti favours L's ponti, which addresses the second objection, but the first remains in force.Footnote 23 Assuming the corruption is not more serious, I suggest we should take ipsi quoque murmura ponti—which would be a fitting climax to the list of natural features—as the subject of uolunt, while writing Minyas to be construed as the object of ferre: even the rumbling sea is now willing to carry the Argo (out of compassion for Medea); for murmura ponti used as a metonymic periphrasis for the sea as endowed with moral agency, compare Lucr. 3.1032 et contempsit equis insultans murmura ponti [sc. Xerxes].Footnote 24
ARGONAUTICA 8.230–1, 233
Landed at the island of Peuce and, for the time being, safely out of reach of the pursuing Colchians, Jason decides to marry Medea (228–33):
In the simile comparing Jason to Hercules the central, well-known, problem is sustinet, which implies a Hercules who can barely stand on his feet and needs to be supported by his new wife—not a very flattering comparison for Jason. Baehrens's solution—fessum Iuno iam destinet Hebae—may indeed be too drastic, but Liberman's defence of sustinet fails to convince: in a context like this, the verb can only mean ‘to support (physically)’, not ‘to invigorate’.Footnote 25 Heinsius's long-forgotten suscipit—‘receives into her home, welcomes’ (cf. OLD s.v. suscipio 5b)—seems exactly the right word, especially as the previous line refers to Hercules’ joining the gods’ banquets.Footnote 26 That line does not seem to have attracted critical attention before, but I wonder if Valerius would not have avoided the unnecessary elision by writing uisere instead of inuisere.Footnote 27 The second nexus of problems affects lines 232–3, the key difficulty here being that, if one accepts the transmitted adsunt at line 232, suscitat at line 233 lacks a subject; I agree with Liberman that Meyncke's adnuit unanimis is an attractive solution, though it is mostly irrelevant for the present argument whether or not it is the right one.Footnote 28 What seems not to have worried editors and commentators much is that affixus ‘intent on, absorbed in (a study, occupation, or the like)’ (OLD s.v. affigo 5b), having as it does no negative connotations, is not a suitable term to pair with curis in a description of emotional distress: Medea is not analysing, but simply suffering from, her ‘worries’. The expected term is of course afflictam, ‘cast down’, which will be nicely matched by suscitat, ‘arouses, reinvigorates’.Footnote 29
ARGONAUTICA 8.251
Yet the wedding celebration is not unclouded—Mopsus has a presentiment about Medea's future (250–1):
The passage does not seem to have troubled editors and commentators, but there are two conspicuous features. One oddity is the unmotivated intrusion of a perfect form (optauit) in a sequence of historic presents; the other, the pointless tum.Footnote 30 There may be more than one way to eliminate these oddities, so what I offer here is intended more or less exempli gratia. The second issue can easily be solved by writing non ullos: non ullus → nullus is a frequent error, usually caused by non being abbreviated as n̄, and tum may simply be a metrical filler.Footnote 31 The first problem may involve a more complex course of corruption: I suspect optauit is a distortion, one way or another, of optattibi (perhaps bi was omitted before ba, after which tti was read as ui, though other scenarios may also be possible), while the transmitted tibi is a misplaced correction, ousting fore; in sum: et fore non ullos optat tibi, barbara, natos.Footnote 32 For tibi immediately preceding vocative barbara, compare especially 8.148 quis locus Inachias inter tibi, barbara, natas? Footnote 33 For optare with fore, compare Cic. Fam. 1.7.11 speroque et opto nobis hanc coniunctionem uoluptati fore.
ARGONAUTICA 8.264
In the meantime, Absyrtus exhorts the Colchians to pursue the Argonauts (261–6):
The central difficulty here is that bare atque cannot adequately introduce a direct speech. One line of approach has been to supply a uerbum dicendi at the beginning of line 246 by writing either atque ait (Watt) or hanc ait (Liberman).Footnote 34 The alternative is to take the direct speech as dependent on clamore, as for instance Baehrens did (heia agite).Footnote 35 I think the second approach is to be preferred: clamor—‘war-cry, battle-cry’ (OLD s.v. 3)—is precisely what Absyrtus’ speech is. Although a decisive solution may be difficult to obtain, I suggest that we should read hanc, hanc: atque could easily be either a mechanical corruption of hanc (perhaps by way of ac), or a metrical filler remedying the omission of one hanc by haplography. If I am right, I believe that the repeated hanc is intended to evoke Hypsipyle's invective against Medea at Ov. Her. 6.129–34:Footnote 36
The passage has three points of contact with the immediate context of the Valerian line: Colchis; Absyrtus; Jason and Medea's union. Now, Ovid's text at the beginning of line 131 is likewise uncertain: hanc, hanc is a conjecture by Palmer for what is transmitted as either hanc or hanc tamen; in other words, we evidently have the same kind of corruption here: first the omission of hanc, then an attempt at conjectural supplement (tamen).
ARGONAUTICA 8.357, 360–1
To help the Argonauts, Juno creates a storm, but Styrus—Medea's former fiancé—is no less eager to go after them (356–62):
Having suggested that prosilit is corrupt at line 21, I cannot avoid the suspicion that it may be corrupt here as well.Footnote 37 Spaltenstein cites Aen. 5.139–40 inde ubi clara dedit sonitum tuba, finibus omnes, | haud mora, prosiluere suis (describing the start of a ship race) as a parallel, but it may not be a perfect equivalent;Footnote 38 as Fratantuono and Smith comment, ‘V[irgil’]s choice of words makes the ships seem like horses, as the verb is normally used to describe a horse, other animal, or a person capable of leaping forth’.Footnote 39 Since the verb is naturally used of a live being, it is rather confusing in the Valerian context: its usage suggests that it should be taken literally (‘leaps forth’), of Styrus as an individual (especially as it is coordinate with dixit), rather than by metonymy of the ship he is commanding, but that, of course, cannot be the intended sense. I suspect that, again, proruit—probably in the transitive sense of OLD s.v. proruo 1 ‘To impel violently forward’ (with nauem as the implied object)—may be the original reading, though, again, proripit might be an option too. A more clear-cut case is ibat. Commentators claim that ire can be used of swimming or sailing, but that claim is misleading: the verb can denote movement as such (even in water) but not the manner of movement (at least, not in reference to swimming); the point here, however, is not that Styrus kept moving even after he fell in the water, but that he was floating, and still would not let go of his shield and sword.Footnote 40 Heinsius's nabat is, I think, as necessary as it is easy (line-initial N could easily be omitted, after which abat had a high chance of being misinterpreted as ibat).Footnote 41 Yet this is not the only difficulty here: another, apparently unnoticed, concerns the relation of imperfect nabat (ibat) to the historic presents surrounding it. A further conspicuous feature is the double et construction repeated in two consecutive lines (60–1). Both these issues can be eliminated by writing (nabat) ut: as he was floating in the water, Styrus begins to—what?Footnote 42 In and of itself, et remos et quaerere transtra—‘to search for oars and thwarts’—seems unobjectionable, but it is rather striking that the same word is used, likewise of Styrus, a mere two lines above but with a different object: the repetition ought to be pointed, but the point can only be bathetic (Styrus gives up his ambition to reach the shore for the ambition of reaching a piece of wreckage that could support him, as it were), which would undermine the tragedy of the moment. Heinsius's prendere deserves serious consideration (a tired eye could easily mistake prēdere for querere, especially when the latter verb was fresh in the scribe's mind from line 359).Footnote 43
ARGONAUTICA 8.374
After Styrus drowns, Absyrtus gives up (374):
The line has a twofold problem: on the one hand, it is inelegant to have two verbs of the same root; on the other, their difference in tense is unmotivated. The usual solution is to adopt Caussin's resedit in place of recedit, but it has two weaknesses: first, it is not clear if residere can be construed with ab (contrast Verg. Aen. 6.407 tumida ex ira tum corda residunt); second, the predominant narrative tense in the context is the historic present, so the two perfects would appear unjustified. Far superior is Löhbach's absistit for abscessit, completely ignored in recent editions and commentaries; the corruption may have been facilitated by scribal reminiscence of line 368 tandem … cessit.Footnote 44
Open access
