18. “Drownded in the Tide”: The Nauagika and Some “Problems” in Augustan Poetry [1]

Richard F. Thomas, Harvard University
Phlebas the Phoenician, a fortnight dead,
Forgot the cry of gulls, and the deep sea swell
And the profit and loss.
               A current under sea
Picked his bones in whispers. As he rose and fell
He passed the stages of his age and youth
Entering the whirlpool.
               Gentile or Jew
O you who turn the wheel and look to windward,
Consider Phlebas, who once was handsome and tall as you.
T.S. Eliot, The Waste Land 312–321
In “The New Posidippus and Latin Poetry,” G. Hutchinson began the task of considering “how this big accession of new material enlarges our appreciation of the Hellenistic background to Latin literature, and our understanding of Latin elegy in particular.” [2] The Milan papyrus contains one six-line and five four-line epigrams of the type ναυαγικόν, a title here attested (XIV 2) for the first time. Hutchinson notes a parallel between XIV 12 (AB 91.2) and Ovid’s Tristia IV 4.55–58 (dangers of the Euxine), and otherwise on this group of 26 lines notes: [3]
xiv.4 θεοῖϲ μέμφεται οἷ᾿ ἔπαθεν (the empty tomb laments the death of Lysicles at sea); xiv.23-4 τὸν νέκυν, ὠϲ χρή, πατρώηι, πόντου δέϲποτα, γὴ ἀπο̣δ̣ο̣ϲ̣. Both passages give relevant context for the speech of the drowning Paetus, Prop. 3.7.57–64; the whole poem is like a huge expansion of an epigram in the class ναυαγικά. The speech begins with the complaint to di maris Aegaei, mentions Neptune (caeruleo … deo) in 62, and ends asking the gods (cf. mandata 55) at saltem Italiae regionibus evehat aestus; | hoc de me sat erit, si modo matris erit.
Again, Hutchinson makes no claims to comprehensiveness, and is to be applauded for initiating this important inquiry into the reception and intertexts of the papyrus. We could, of course, already see the role played by Hellenistic ναυαγικά in Propertius III 7, since we already had two runs of the sub-genre in the Palatine Anthology (AP VII 263–279, 282–294), though most of the second group are post-Augustan. [4] Incidentally, as noted by Bastianini and Gallazzi, one of these poems (AP VII 267) is by Posidippus, but is absent from the ναυαγικά of the papyrus—one of the reasons for skepticism about Posidippan authorship of all the new poems. [5] As for these pre-existing ναυαγικά, AP VII 263, ascribed to Anacreon, and certainly Hellenistic or earlier, [6] would seem to have a connection to Propertius, with each conflating erotic and sepulchral: AP VII 263.3–4 (narrator) ὑγρὰ δὲ τὴν σὴν κύματ᾿ ἀφ᾿ ἱμερτὴν ἔκλυσεν ἡλικίην = Prop. III 7.57–9 (Paetus) “di maris Aegaei quos sunt penes aequora, venti, | et quaecumque meum degravat unda caput, | quo rapitis miseros primae lanuginis annos?”
Nevertheless, the six new poems preserved by the papyrus permit us to see just how intensely a poem like Propertius III 7 may indeed seem to represent a “huge expansion” of epigram in the process of creating a genre, Roman elegy, that is rooted in Hellenistic epigram. In this respect, the elegist may be seen to be working in the tradition of expansion observable in Catullus 68, which takes sepulchral epigram (already “translated” in Catullus 101) as a starting-point and expands that genre into a complex proto-elegy. [7] These new poems may be usefully viewed in relation to two other Roman poets, who are similarly involved in generic appropriation and renovation. In all three cases we will see that readers and critics have found something odd or strange about the lines in question. I believe that the Milan papyrus allows and encourages us to see this oddness as part of the compositional condition of these epigrams, and also allows us to reconstruct the peculiar essence of that corpus of Roman poetry that most looked to Alexandria for much of its inspiration and innovation.

The Archytas Ode (Horace, Odes I 28)

The new papyrus does not at first sight seem to provide a transformative lens through which to view Odes I 28, the Archytas ode. As Nisbet-Hubbard already noted in their introduction, “the poem includes many traditional elements from Greek sepulchral epigram.” [8] We can however, see certain important elements of Horace’s poem coming into sharper focus through the appearance of the new ναυαγικά, when we consider them as a group. It may be worth noting at the outset that the first of the Milan epitaphs (AB 89) records the grief of Polemon at the loss by drowning of his friend Lysicles. That Polemon, like Archytas, was a philosopher may not be irrelevant, since the epigram and the ode may both be seen as being rooted in the tradition of literary epitaphs for philosophers, to be found at AP VII 94–132 (VII 103 is for Crates and Polemon).
As for shared themes and actual intertexts, the new epigrams offer one detail that, although already available to Horace in part from Meleager’s anthology and elsewhere, is closer than examples we already had. Horace’s poem ends with the injunction that the traveling merchant pause to give burial to the unburied speaker (35–6 quamquam festinas, non est mora longa: licebit iniecto ter pulvere curras). Nisbet-Hubbard (ad I 28.35) refer to various texts, including Asclepiades 33 GP (= AP XIII 23), dealing with the necessity to interrupt one’s business in order to mourn the dead. Their examples, however, are not specifically from shipwreck poems, the context of the plea of Horace’s ναυαγός. But now we have a model in the final poem of our series in the Milan papyrus (AB 94):
ναυηγόν με θανόντα καὶ ἔκλαυϲεν καὶ ἔθαψεν
     Λεώφαντοϲ ϲπουδῆι, καὐτὸϲ ἐπειγόμενοϲ
ὡϲ ἂν ἐπὶ ξείνηϲ καὶ ὁδοιπόροϲ· ἀλλ᾿ ἀποδοῦναι
     Λεωφάντι μεγάλην μικκὸϲ ἐγὼ χάριτα.
“I died in a shipwreck and was mourned and buried
     hurriedly by Leophantus, as too he was hastening
like a traveler in a foreign land. But I am too small
     to give great thanks in return to Leophantus.” [9]
Although the ending is either obscure or lame, [10] the poem gives us a closer situation to that of the ending of Odes I 28 than we hitherto had. Both texts represent the voice of the shipwrecked in the first person, each either gives thanks for or requests burial by a trader, and both refer to the haste of the trader. The differences are merely ones of aspect and temporal progression: while the Latin poem urges burial, the Greek one records burial that has been granted. [11]
Perhaps more important than this intertext is the fact that we now have the Milan epigrams in their original, Hellenistic context, not just as six poems preserved in a later redaction, but as a run of 26 lines of related and varied poems, so placed by Posidippus or the third-century BCE anthologist. Nisbet-Hubbard have rightly defended Horace’s 36-line poem against those who would see it as a “chaotic youthful experiment.” Whatever the oddities, as they note, “anybody who likes this poem has discovered something about poetry.” [12] Nisbet-Hubbard have also designated it “undeniably bizarre in conception.” [13] The basis for finding Odes I 28 “chaotic,” “bizarre” or otherwise odd, has purely to do with voice and identity. Who speaks the opening lines (1–4 Te … cohibent, Archyta, pulveris exigui … munera)? The corpse of a drowned man? The passing nauta of line 23? Horace himself? And what is the relationship of the te of line 2 to the me of line 21? Is the poem a monologue or a dialogue—or neither? These are the questions that have vexed all critics of this poem ever since it has been read. Frischer has a good survey of recent work on the question, citing the “standard interpretation” of Nisbet-Hubbard: [14]
the poem is spoken by a corpse of a drowned man. First the dead man apostrophizes the great fourth-century BCE Pythagorean Archytas of Tarentum, as he lies buried in his grave. Then at 23 he turns to a passing nauta and asks for burial himself. The structure of the poem causes perplexity because we do not know till 21 the speaker is not Horace but a corpse.
Frischer’s own view is that the opening speaker is Horace, who sarcastically addresses the dead Archytas (1–20): [15]
Horace’s purpose in attacking Archytas and his beliefs is to create a situation in which Archytas would not fail to reply, if he could, to Horace’s abuse. In 21–36, the epitaph on Archytas’ tomb is quoted as the only response that Archytas can make after Horace’s address. The wit of the poem consists in the fact that Archytas’ epitaph in no sense responds to the issues Horace raises. All the text of the epitaph can do is endlessly repeat the same empty threats and promises, the irrelevance of which reinforces the point of Horace’s attacks on Archytas’ beliefs in the survival of consciousness in the afterlife and in the possibility of communication between the living and the dead at the tomb.
There are attractions to such a reading, but it too calls for an act of faith, since the dramatic setting and the movement from the first part to the second is by no means explicit or even implicit in the text. Frischer’s reading is possible, but it hardly resolves the issue of identity and voice in any compelling or final way. Other possibilities remain, as does the question of coherence.
Let us return briefly to the Milan poems, and to the issue of voice. They are quite varied in this respect:
AB 89     B      Third-person narration by text on stone or reader
AB 90     A1     Third-person narration, no text/stone; focalization of swimmer (ὁρώμενον)
AB 91     C1     Address to reader/passerby by text/stone; deceased in third person [16]
AB 92     A      Third-person narration, no text/stone
AB 93     D      Address to earth/sea to cover lightly/return to shore
AB 94     E      First-person narration by deceased
Horace’s poem, on the other hand, falls into the following categories:
1-20      F       Address to deceased (Frischer)
             E1     Apostrophe by deceased (Nisbet-Hubbard)
21–22   E       First-person narration by deceased
23–36   C       Address to passerby by deceased
In other words, Horace’s poem demonstrates a combination of voices, such as one finds in the epigrams—only there across poem-boundaries. Nisbet-Hubbard noted of I 28.21ff.: “Horace now weaves in themes from the other type of Greek epitaph, where the dead man does the speaking himself.” [17] Horace cannot give us an epigram in the Odes (though IV 10 will come close), but, particularly with the impulse from a string of thematically connected but varied epigrams such as we find on the Milan papyrus, he can combine and conflate, and in the process produce a lyric poem of some mystery, a poem that looks like two or three epigrams, and yet still somehow works as a poem, if only we listen to, and appreciate, its disparate voices.

Propertius III 7: Epigrams or Elegy?

At a number of points the third book of Propertius’ elegies seems to respond to the recently published collection of Odes 1–3. [18] It may therefore be worth considering Propertius III 7, in connection with Odes I 28, but also in the light of the new epigrams. Again the words of Hutchinson on this poem: “the whole poem is like a huge expansion of an epigram in the class ναυαγικά.” There is some truth to that, particularly if we accept the multiple transpositions, and claims for interpolation, that have been made by editors in a frantic attempt to create a coherent poem of 3.7. Housman is representative, as he reads the poem thus: 1–10/43–66/ 17–18/11–16/67–70/25–32/37–38/35–36/19–20/33–34/21–24/39–42/71–72. [19] But as Butler and Barber note, “While few are likely to regard any of these suggestions as wholly acceptable, and fewer still will venture to explain how such dislocations could have been brought about, all four schemes are fundamentally inspired by the desire to eliminate three genuine difficulties.” [20] All three difficulties have to do either with confusion of adressees (as in the case of Odes I 28) or with logical problems in the narrative unity of the poem.
Butler and Barber record the attempt of Vahlen to explain the poem without these transpositions. [21] His argument is one “which may be summarized as a plea that Propertius’ method is not to exhaust any one motive once introduced, but to recur to it and add new touches; and that consequently all wholesale transpositions are based on a misconception.” Butler and Barber concede some ground, but while they preserve the order of the MSS, as does Barber’s OCT, they ultimately resist Vahlen’s suggestion: “Such a defense … is an assertion that a rambling style and a love of parenthesis is characteristic of the Propertian elegy. And while there is some truth in such an assertion, the present elegy provides an extreme example of the poet’s particular art or lack of it … [T]he most cursory analysis at once reveals the eccentricity of the structure and arouses strong suspicions of the soundness of the tradition.” [22]
The experimental and innovative aspects of Propertius’ third book are well acknowledged. What if, with the help of the Milan papyrus, we were to revive Vahlen’s hypothesis that the “rambling” style is intended, and is indeed part of the art? When we accept and consider the poem as it is preserved in the MSS, we are left, not in the words of Hutchinson with a “huge expansion of an epigram,” which might describe the text produced by Housman and others, but rather with a reduction of elegy to its constituent component string of epigrams, all of them ναυαγικά for Paetus (some naming him, others not). This is what leads to an appearance of “rambling style and a love of parenthesis.” When the parenthesis becomes a poem in its own right it is no longer a parenthesis, rather an epigram.
Such a hypothesis could have been proffered before the appearance of the Milan papyrus, on the basis of the pre-existing runs at AP VII 263–79, 282–94. But what we can now for the first time do is make two further, notable observations: 1) since scholars have already begun to see the thematic groupings of the papyrus as having some unitary status within the group, [23] the Greek poems give us a direct or indirect model, precisely for a poem such as III 7, read as Vahlen proposed to read it; 2) the once popular question of the origin of Roman elegy comes back into play, since the papyrus gives us a book of at least ten “poetic units” ranging in length from 26 to 126 verses (excluding col. XVI.18ff., the last grouping, of 8 lines). Propertius III 7 looks very much like one of those “poetic units”, but it is unique among the poems of Book 3 in doing so. To this extent the Latin poem may be seen as reversing the process apparent in much elegy (but not in III 7), wherein the genre can indeed be viewed as a “huge expansion” (both in size and poetic complexity). This expansion is heralded by the opening of the Monobiblos (Prop. I 1 1–4), translating and transgendering as it does the homoerotic Meleager 103 Page (= AP XII 101), which serves only as the opening intertext. Propertius’ string of ναυαγικά at III 7, on the other hand, shows us where Roman elegy largely came from, precisely by giving us a poem that is to be read as epigrams, as the disiecta membra that are the basis for the “poem’s” making sense. It is truly experimental precisely in its revisionism, which especially comes into focus and appears familiar now that we have the Milan papyrus.
Here then is Propertius III 7 (Nauagica Paeti), with individual “epigram titles” supplied: [24]
Perils of the marine merchant
Ergo sollicitae tu causa, pecunia, vitae!
     per te immaturum mortis adimus iter;
tu vitiis hominum crudelia pabula praebes;
     semina curarum de capite orta tuo.
tu Paetum ad Pharios tendentem lintea portus5
     obruis insano terque quaterque mari.
nam dum te sequitur, primo miser excidit aevo
     et nova longinquis piscibus esca natat.
Cenotaph and a mother’s grief
et mater non iusta piae dare debita terrae
     nec pote cognatos inter humare rogos,10
sed tua nunc volucres astant super ossa marinae,
     nunc tibi pro tumulo Carpathium omne mare est.
Cruelty of Neptune and the North Wind
infelix Aquilo, raptae timor Orithyiae,
     quae spolia ex illo tanta fuere tibi?
aut quidnam fracta gaudes, Neptune, carina?15
     portabat sanctos alveus ille viros.
Too late a son’s plea and cry for his mother
Paete, quid aetatem numeras? quid cara natanti
     mater in ore tibi est? non habet unda deos.
nam tibi nocturnis ad saxa ligata procellis
     omnia detrito vincula fune cadunt.20
Even Agamemnon suffered such loss, by water and on land
sunt Agamemnonias testantia litora curas,
     quae notat Argynni poena Athamantiadae.
hoc iuvene amisso classem non solvit Atrides,
     pro qua mactatast Iphigenia mora.
Bring his body to the shore, to be a sailors’ warning
reddite corpus, aquae! posita est in gurgite vita;25
     Paetum sponte tua, vilis harena, tegas;
et quotiens Paeti transibit nauta sepulcrum,
     dicat ‘et audaci tu timor esse potes.’
Human skill brings human loss
ite, rates curvate et leti texite causas:
     ista per humanas mors venit acta manus.30
terra parum fuerat fatis, adiecimus undas:
     fortunae miseras auximus arte vias.
Stay at home or suffer this fate
ancora te teneat, quem non tenuere penates?
     quid meritum dicas, cui sua terra parum est?
ventorumst, quodcumque paras: haud ulla carina35
     consenuit, fallit portus et ipse fidem.
Nature more cunning than culture hero
natura insidians pontum substravit avaris:
     ut tibi succedat, vix semel esse potest.
saxa triumphalis fregere Capherea puppes,
     naufraga cum vasto Graecia tracta salo est.40
paulatim socium iacturam flevit Ulixes,
     in mare cui soliti non valuere doli.
Better poverty at home than wealth across the sea
quod si contentus patrio bove verteret agros,
     verbaque duxisset pondus habere mea,
viveret ante suos dulcis conviva Penates,45
     pauper, at in terra nil nisi fleret opes.
noluit hoc Paetus, stridorem audire procellae
     et duro teneras laedere fune manus,
sed thyio thalamo aut Oricia terebintho
     effultum pluma versicolore caput.50
The sum of his evils: wave-torn, night-blinded, drownded in the tide
huic fluctus vivo radicitus abstulit ungues:
     Paetus ut occideret, tot coiere mala.
hunc parvo ferri vidit nox improba ligno,
     et miser invisam traxit hiatus aquam.
Melodious tears, last words: “Look homeward angel”
flens tamen extremis dedit haec mandata querelis55
     cum moribunda niger clauderet ora liquor:
‘di maris Aegaei quos sunt penes aequora, venti,
     et quaecumque meum degravat unda caput,
quo rapitis miseros primae lanuginis annos?
     attulimus longas in freta vestra manus.60
ah miser alcyonum scopulis affligar acutis!
     in me caeruleo fuscina sumpta deo est.
at saltem Italiae regionibus evehat aestus:
     hoc de me sat erit si modo matris erit.’
subtrahit haec fantem torta vertigine fluctus;65
     ultima quae Paeto voxque diesque fuit.
“Where were ye nymphs?”
o centum aequoreae Nereo genitore puellae,
     et tu, materno tacta dolore, Theti;
vos decuit lasso supponere bracchia mento:
     non poterat vestras ille gravare manus.70
Sea of love for me!
at tu, saeve Aquilo, numquam mea vela videbis:
     ante fores dominae condar oportet iners.
The antepenultimate and penultimate titles’ references to Milton’s Lycidas (formally a ναυαγικόν for Milton’s friend, Edward King, like Paetus lost at sea and dead “ere his prime”), are not casual. Like Propertius III 7, that great poem has regularly been criticized for its excessive art in engaging so openly and emphatically with the intertexts in its tradition. Its fractured state may be parallel to that of Propertius III 7, precisely because the two poets are engaged on parallel confrontations with their traditions.
Rather than transposing sequences of lines in Propertian elegy, we might do better to reflect on the sort of sequences that the Milan papyrus gives us, and to see in a poem like Prop. III 7 a radical encounter with the tradition reperesented by the papyrus. The Roman elegy is still a poem—just—but it is a poem with looser narrative logic and progressions, and more propensity to disunity, than we are traditionally capable of tolerating. Perhaps the papyrus will increase our tolerance, and also give a basis for similar findings elsewhere in the corpus of Roman poetry, that of Propertius in particular.

Palinurus (Virgil, Aeneid V 833–71, VI 337–83)

Virgil’s Palinurus is a character of some complexity in terms of literary genealogy and intertextuality. As has long been noted, his Homeric counterpoint is Elpenor, like Palinurus the first familiar character encountered by the hero in the course of his katabasis (Od. XI 51–83 ~ Aen. VI 337–83), just as he shared with Aeneas’ navigator the fate of being the last to die in the upper world (Od. X 552–60 ~ Aen. V 833–71). [25] At the same time, Virgil draws from Od. III 278–85, Nestor’s account of Menelaus’ loss (to the arrows of Apollo) and subsequent burial of the helmsman Phrontis. And Aeneas’ burial of Misenus (Aen. VI 149–53) further complicates things. [26] That parallel establishes the Odyssean moments as genre models for Virgil, but as so often his artistic genius is most on display and at its most subtle precisely when he departs from the details of those models. The greatest distance between the two characters lies in the manner of their death, which, however, both involve a fall, Elpenor’s from the roof of Circe’s house.
Throughout both Palinurus passages Virgil engages intensively with the tradition of the shipwreck epigram, embedding it into his larger narrative, and so creating an expansion and appropriation of the genre, as he blends it into the Homeric genre model. This is, again, something we could know before the discovery of the Milan epigrams, but these latter help to strengthen our appreciation of the Hellenistic overlay on the Homeric foundation. [27] Given that the narrative draws attention emphatically to Palinurus’ struggles and death at sea, one could even conclude that the epigrammatic intertext assumes a greater impact in its new, epic context.
What does the papyrus give us, then, as readers of Virgil’s Palinurus? First it demonstrates the fact, unsurprising perhaps, that the ναυαγικά were even more prevalent than the Meleagrian examples would suggest. The particularly powerful grief and pathos they evoke, since the deceased in such cases potentially or in fact go without burial, were available for Virgil to activate, and he indeed invokes these emotions throughout both passages. But there are also significant intertextual observations to be made. When Palinurus encounters Aeneas in the Underworld, he gives details of his fate:
tris Notus hibernas immensa per aequora noctes
vexit me violentus aqua; vix lumine quarto
prospexi Italiam summa sublimis ab unda.
paulatim adnabam terrae; iam tuta tenebam,
ni gens crudelis madida cum veste gravatum
prensantemque uncis manibus capita aspera montis
ferro invasisset praedamque ignara putasset.
nunc me fluctus habet versantque in litore venti.
(Aen. VI 355–62)
The lines convert to the first person the narrative account of Odysseus’ being washed ashore on Phaeacia (Od. V 388–435), as Knauer and others have noted. But Palinurus’ less happy ending (362 nunc me fluctus habet versantque in litore venti), take us in a different direction, to the language of the ναυαγικόν. The self-contained one-line sentence creates epigrammatic effect. [28] It is true that this line also has a (syntactically distinct) Homeric intertext, in the anxieties Telemachus expresses about his father’s fate at Od. I 161–2 (ἀνέρος, οὗ δή που λεύκ᾿ ὀστέα πύθεται ὄμβρωι | κείμεν᾿ ἐπ᾿ ἠπείρου, ἢ εἰν ἁλὶ κῦμα κυλίνδει), [29] but the actual language is more reminiscent of sepulchral epitaph. So, at Euripides Hec. 28–30, the ghost of Polydorus similarly refers to its fate: κεῖμαι δ᾿ ἐπ᾿ ἀκταῖς, ἄλλοτ᾿ ἐν πόντου σάλῳ, | πολλοῖς διαύλοις κυμάτων φορούμενος, | ἄκλαυτος ἄταφος·. [30] And Euripides’ and Virgil’s language have reminiscences in two of the Milan ναυαγικά:
AB 89.5–6 τὸν δέ που ἤδη ἀκταὶ καὶ πολιὸν κῦμα [θανόντ᾿ ἔλαχον (Austin and Bastianini) | κῦμ᾿ {α} [ἐπέχουϲι ἁλόϲ (Bastianini and Galazzi)
AB 91.13–14 κενεὸν Δώρου τάφον, ὃν Παριανῶν τῆλέ που εἰκαῖαι θῖνεϲ ἔχουϲιν ἁλόϲ.
Virgil’s verb habent (cf. ἐπέχουϲι; ἔχουϲιν) is particularly notable, and is paralleled by our Horatian ναυαγικόν discussed above (Odes I 28.9–10 habentque Tartara Panthoiden). [31]
Aen. VI 362 functions as a pivot from the narrative that precedes to what could almost be an actual sepulchral epigram. Except for the fact that it is addressed to Aeneas, it looks in large part like the conventional address to the passerby, requesting at least a perfunctory burial (365–366 aut tu mihi terram inice):
“quod te per caeli iucundum lumen et auras,
per genitorem oro, per spes surgentis Iuli,
eripe me his, invicte, malis: aut tu mihi terram365
inice, namque potes, portusque require Velinos;
aut tu, si qua via est, si quam tibi diva creatrix
ostendit (neque enim, credo, sine numine divum
flumina tanta paras Stygiamque innare paludem),
da dextram misero et tecum me tolle per undas,370
sedibus ut saltem placidis in morte quiescam.”
The model, of course, is Od. XI 66–78, Elpenor’s request that Odysseus burn him with his armor and set up a σῆμα for him, when Odysseus returns to Aeaea. But Palinurus’ speech is modified to suit the Roman sepulchral tradition, and also to suit the new situation (death at sea). And a significant detail is added: since Aeneas is in the Underworld and about to cross the Styx, Palinurus assumes he has the divine aid of Venus (367–369), and that he may in fact be able to lift the helmsman from the waves and take him to the other side for his final rest (370–371, “da dextram misero et tecum me tolle per undas, | sedibus ut saltem placidis in morte quiescam”). The least well-preserved of the Milan ναυαγικά is to be found at AB 92, a poem apparently dealing with a divine rescue of a swimmer, following the loss of his ship and crew:
νηὸϲ ἀπολλυμένηϲ ϲυναπώλετο πᾶϲ ἁμαεργὸϲ
     ναύτηϲ, νηχ̣ο̣μ̣έ[νωι δ᾿ ἦν    ] ε̣ν̣τ̣ι φυγή·
τὸν γὰρ ἐπαμ[                               ]α̣ δαίμων
     νηχόμενον [                                 ̣]̣ϲ.
Here is Austin’s translation:
As the ship was sinking, down with it went the whole crew
on board but [ … ]noeis escaped by swimming.
[Coming to his rescue] a god wrapped [him up as he gently] swam
[and saved him from the ice-cold sea]. [32]
From Aen. VI 362 (nunc me fluctus habet versantque in litore venti), we are to imagine that Palinurus is still amidst the waves of the sea when he pleads with Aeneas to rescue him; whether or not there is any direct relationship, what we find in Palinurus’ plea for rescue from the divinely empowered Aeneas is a (non-Homeric) equivalence to the situation of the epigram, so supplemented.
There is another feature of the Virgilian treatment that warrants comment in the light of the Milan papyrus. As occurred in the case of Prop. III 7, so Virgil’s two passages on Palinurus have been convicted of inconsistency on a number of grounds. [33] Although these inconsistencies are part of the larger fabric of ambiguities and oddities in the poem, no other instances rise to the level of the Palinurus sections, which Horsfall proposes may have been the only sections that Virgil would have changed had he had those three years to polish up the poem. He well summarizes the major inconsistencies (100–101):
Aeneid V Aeneid VI
• Somnus throws P. overboard • No divine actor present
• Calm sea • Stormy sea
• Aeneas thinks it an accident • Aeneas speaks of divine cause
• Event occurs between Sicily and Italy • Virgil speaks of Libyco cursu
• The trip lasts just one night • P. tells of swimming for three nights
It is as if the two sections tell different stories. And even within each of the two sections there is a sense of episodicity, which may be seen as a result of the presence of sepulchral epigram throughout. Likewise there is some play with voice, some of the confusion between narration and focalization that one finds in the tradition of such epigram. [34] The fifth book ends with what is most immediately an address by Aeneas to his dead helmsman:
     ipse ratem nocturnis rexit in undis
multa gemens casuque animum concussus amici:
“o nimium caelo et pelago confise sereno,               870
nudus in ignota, Palinure, iacebis harena.”
Aen. V 868–71
The words belong to Aeneas, but the absence of a verb of speaking and the gnomic nature of the final couplet, along with its sepulchral content, identify it as a cenotaph, whose voice is not just Aeneas’, but also that of the narrator of Aen. V, and of us, readers of the inscription and of the poem. Indeed the inscription may even be seen as attaching to the story that proceeds, whose narrative ἐνάργεια, beginning with ecce at V 854, functions almost as a narrative image, to which the closing epigram lends its epigrammatic commentary.
We have similar variety in Aen. VI, where the Palinurus episode again begins with the arresting marker, ecce:
ecce gubernator sese Palinurus agebat,
qui Libyco nuper cursu, dum sidera servat,
exciderat puppi mediis effusus in undis.
Aen. VI 337–9
The ensuing exchange between Aeneas and the shade of Palinurus (340–71) may be reduced to the following: A: “How did you die?” P: “This is how I died.” Such a conversation is possible only in two situations, in katabasis and in sepulchral epigram (also in dreams, presumably). Ultimately the model for this exchange is to be found in the genre model, at Od. XI 57–80, where Odysseus asks Elpenor how he came to Hades (57 “Ελπῆνορ, πῶς ἦλθες ὑπὸ ζόφον ἠερόεντα;”), and is given a reply, along with the request for burial, to which Odysseus assents, after which the two converse. Significantly, in Virgil’s version there is no reply by Aeneas, and no conversation, just the question and response, and to that extent the passage exhibits similarities to sepulchral epigrams with similar exchanges [35] —unimaginable in the Homeric version, where the text narrates a conversation between Odysseus and Elpenor following the exchange (Od. XI 81–83).
The culmination of the whole episode has no Homeric intertext. The Sibyl, and not Aeneas, replies to Palinurus, first rebuking him and denying him passage across the Styx, but then consoling him by predicting the fame that will come to Cape Palinurus, and that will be spread precisely by the presence of a cenotaph:
“nam tua finitimi, longe lateque per urbes
prodigiis acti caelestibus, ossa piabunt
et statuent tumulum et tumulo sollemnia mittent,380
aetern umque locus Palinuri nomen habebit.”
Aen. VI 378–81
The culmination of the Palinurus story in Aen. V was a virtual epigram, while that of Aen. VI is a prediction of the cenotaph to which such a ναυαγικόν, already elaborated through the complex narrative of Aen. V and VI, would be affixed.


In Horace Odes I 28, Propertius III 7, and Virgil Aen. V 833–71, VI 337–83 we have three roughly contemporary Latin engagements with the ναυαγικόν. Whether or not any of these poets was specifically aware of any of the epigrams on the Milan papyrus, those poems nevertheless first of all add to the Hellenistic context of the Augustan texts, a context whose presence Latinists so oddly resist, but that becomes less resistible whenever a new literary papyrus adds to our store of Hellenistic poetry. More importantly, the “problems” of episodicity, of unity, of voice, of consistency, shared by all of these Latin poems, perhaps become less problematic if we see that each poet is in different ways engaging with thematically ordered epigrams. Whether Virgil, Horace and Propertius found their ναυαγικά in the Garland of Meleager or on a text such as P. Mil.Vogl. VIII 309, we cannot be sure. What cannot be doubted, however, is the fact that they read and integrated such texts into the eclectic genres of Augustan poetry, just as Eliot, in the epigram with which we began, integrated the same sub-genre into the larger, innovative project of The Waste Land.


[ back ] 1. I am grateful to David Petrain for helpful comments and suggestions.
[ back ] 2. Hutchinson 2002:1–10.
[ back ] 3. Hutchinson 2002:6.
[ back ] 4. See Gutzwiller 1998:313 for the proposition that the shipwreck poems in the Cephalan section at VII 263-73 were in last section of the Meleagrian book with VII 494–506, 650–654, 738–739.
[ back ] 5. Cf. BG 216. In other words the absence of a Posidippan ναυαγικόν from the ναυαγικά of the papyrus may argue as much for a multi-authored anthology as the presence of two possibly Posidippan poems argues for a single-authored book. (Gow-Page say of one [APl. 199 = Pos 18 GP = AB 65] there are “reasons for questioning [with Schott] the ascription to Pos.,” of the other [ap. Tzetz. Chil. 7.660 = Pos 20 GP = AB 15] “its authenticity seems open to considerable doubt”). This is especially true given the variation in quality of the new poems, the oddity of the collection (whose lemmata seem to be the organizing principle), and the absence of more familiar Hellenistic subgenres (straightforward sympotic, and erotic, in particular). I would be prepared to accept Posidippan authorship; the case, however, given the slightness of the evidence, has yet to be made.
[ back ] 6. Cf. GP I xxii.
[ back ] 7. On this see Thomas 1998:214–216.
[ back ] 8. Nisbet and Hubbard 1970:318.
[ back ] 9. Translation C. Austin in Austin and Bastianini 2002:119.
[ back ] 10. See BG:221 for discussion, and for the possibility (ultimately rejected) that there is a play on the name Μίκκος. Without some such play, the poem is rather feeble.
[ back ] 11. BG: 220–221 give no reference to Horace, but they do point to the parallel with Callimachus, Ep. 58 Pf. = AP VII 277 (the model for Eliot’s lines at the beginning of this article), which records Leontichus’ (cf. Leophantus) burial of an unnamed ναυαγός. A far superior poem to that of the Milan papyrus, Callimachus’ epigram is however distinct from it and from Odes I 28 in that it addresses the deceased and narrates the burial, while they are addresses by the deceased. David Petrain points out that the opening of Callimachus’ poem (τίς ξένος …) raises questions of voice that bring it closer to Odes I 28 but not to the Milan poem.
[ back ] 12. Nisbet and Hubbard 1970:320.
[ back ] 13. Nisbet and Hubbard 1970:319.
[ back ] 14. Frischer 1984:71–102; cf. also Nisbet-Hubbard 1970–1978: 317–20.
[ back ] 15. Frischer 1984:73.
[ back ] 16. Following AB. The edition of BG (14 καὶ <μ᾿> αἱ θῖνεϲ ἔχουϲιν ἁλόϲ) would have the deceased addressing the reader (C).
[ back ] 17. Nisbet and Hubbard 1970:319.
[ back ] 18. For the connections between the two poets see Solmsen 1948:105–109; Flach 1967: passim.
[ back ] 19. Butler and Barber 1933:275 record the interventions of four critics: Scaliger, Housman, Postgate, and Richmond. Fedeli’s 1984 Teubner and Goold’s 1990 Loeb (rev. 1999) have slightly different arrangements.
[ back ] 20. Butler and Barber 1933:275.
[ back ] 21. Vahlen 1883:36–90.
[ back ] 22. Butler and Barber 1933:276.
[ back ] 23. Notably, in this volume, Fantuzzi for the ἱππικά, Baumbach and Trampedach for the οἰωνοϲκοπικά.
[ back ] 24. Many of these “epigrams” demonstrate the presence of antithesis or paradox that is so marked a feature of the Hellenistic genre.
[ back ] 25. See Knauer 1979: 135–9 for Homeric aspects, particularly the connection with Elpenor.
[ back ] 26. See Knauer 1979:136–137; 137n1 for the possibility that Naevius’ account of the death of Prochyta would have provided another doublet. On this see Mariotti 1955:40–47; also Horsfall 1991:101.
[ back ] 27. See Norden 1970: 230–231 for reference to shipwreck epigrams in connection with Palinurus.
[ back ] 28. For similar gnomic or epigrammatic effects (I exclude instances where forms of hic, etc. connect to what precedes, or sic, etc. to what follows), as a clausula, just from Aen. 1–2, see I 33 tantae molis erat Romanam condere gentem; 334 multa tibi ante aras nostra cadet hostia dextra; 401 perge modo et, qua te ducit via, derige gressum; 630 non ignara mali miseris succurrere disco; II 49 quidquid id est, timeo Danaos et dona ferentis; 354 una salus victis nullam sperare salutem.
[ back ] 29. See Knauer 1979:ad loc.
[ back ] 30. See Norden 1970:ad loc.
[ back ] 31. Also Aen. IX 490–1 (sepulchral, of the whereabouts of Euryalus’ headless corpse) quae nunc artus avulsaque membra | et funus lacerum tellus habet ? Cf. ThLL s.v. habeo 2431.53–69.
[ back ] 32. Following his exemplary supplementing of 3–4: τὸν γὰρ ἐπαμ[πίϲχων ϲωτήριοϲ ἠρέμ]α̣ δαίμων | νηχόμενον [ψυχρῆϲ ἐξεϲάωϲεν ἁλ]ο̣ϲ. Unless Austin is wrong in this general reconstruction, this poem must number among the most miserable on the papyrus. The poems given as parallel by BG (e.g. AP VII 289, 290, 550; 9.269) all involve some sort of paradox, with the even-tual death (but not by drowning,) of the shipwrecked man. Our poem would be better if, for instance, the last two lines involved some ill fortune (?κακὸϲ δαίμων) coming up against him (ἐπαμβαίνει), or enfolding him (ἐπαμπίϲχων), in the form, say, of a shark, as occurs in an epigram of Leonidas of Tarentum (AP VII 506.10 = Leon. Tar. 75 GP ἥμισυ δὲ πρίστις ἀπεκλάσατο). Or perhaps some god (δαίμων) sent a man-eating fish (κῆτοϲ) against him while he was swimming, one of the primal fears of the swimming ναυαγός, as at Od. V 417–22:
εἰ δέ κ᾿ ἔτι προτέρω παρανήξομαι, …
δείδω μή μ᾿ ἐξαῦτις ἀναρπάξασα θύελλα
πόντον ἐπ᾿ ἰχθυόεντα φέρηι βαρέα στενάχοντα,            420
ἠέ τί μοι καὶ κῆτος ἐπισσεύηι μέγα δαίμων
ἐξ ἁλός, οἷά τε πολλὰ τρέφει κλυτὸς Ἀμφιτρίτη·
[ back ] 33. See Horsfall 1991:Ch. 6 “Incoerenze.” J. O’Hara will treat the topic in Virgil and other Latin poets in a forthcoming monograph.
[ back ] 34. For this topic see the forthcoming Harvard PhD dissertation of Michael Tueller.
[ back ] 35. E.g. AP VII 163–5; 317 (Callimachus); 470 (Meleager); 503 (Leon. Tar.)