Some imitations of Pindar and Sappho by Horace

2015.12.31 | By Gregory Nagy

"The Singing Swan," Reinier van Persijn (circa 1655) [Public domain], via Wikimedia Commons
“The Singing Swan,” by Reinier van Persijn (circa 1655) [Public domain], via Wikimedia Commons


§0.1. This posting of 2015.12.31 is based on an article of mine entitled “Copies and Models in Horace Odes 4.1 and 4.2,” which was originally published over 20 years ago in Classical World 87 (1994) 415–426. Concurrently with this posting, I am republishing an online version of that article.

§0.2. The 30 paragraphs that comprise the online version of “Copies and Models in Horace Odes 4.1 and 4.2” replicate almost word for word the content of the original printed version, except that

(1) I numbered the 30 paragraphs: §1, §2, §3 … §30.

(2) I added my own translations of the quoted Latin and Greek

(3) I signaled, within double-square brackets (“[[” and “]]”), some additional observations that needed to be made. All these additional observations date from 2015.12.31 and are further explained in the present posting.

§0.3. In the paragraphs §1, §2, §3 … §30 that follow the paragraphs §0.1, §02, §03, §04, §05 of this introduction to the present posting of 2015.12.31, I will expand some aspects of the original content while compressing or even eliding other aspects. The compressions and expansions result in a new essay that is meant to be simpler than the old one—or, at least, easier to read. But this new essay is not meant to be a second edition: rather, it is a rethinking.

§0.4. Wherever I expand the original content of a given paragraph, I will indicate the expansion by adding a, b, c, and so on to the original number of the paragraph: for example, the original paragraph §2 of the article will be expanded as the new paragraphs §2a, §2b, §2c in this posting.

§0.5. I have kept the bibliographical references here at a minimum, since readers will now have the option of consulting the original article, which tracks all of the secondary sources that I had used in putting together my original argumentation.

Imitations of Pindar in Horace Ode 4.2

§1. When Horace imitates Pindar in Ode 4.2, beginning with Pindarum quisquis studet aemulari ‘whoever is eager to imitate [aemulārī] Pindar’ at line 1, he presents his composition not so much as an act of copying within a genre but as a model, even an archetype, in its own right. Key words are aemulārī ‘to imitate’ at line 1 and imitātus ‘imitating’, from imitārī ‘to imitate’, at line 57, towards the very end of the poem.

§2a. These two Latin words aemulārī and imitārī, along with a third, imāgō, are conceptually related to the Greek word mīmēsis. [[In n1 of the article, I drew special attention to the meaning of Latin aemulus as ‘striving to equal, understudy’ and of Latin imāgō as ‘death-mask of an ancestor’ (Nagy PH 349n58 = 12§24, especially with reference to Pliny Natural History 35.6).]]

§2b. In terms of my argument, the primary meaning of mīmēsis was ‘re‑enactment’ in a dramatic sense, and the secondary meaning of ‘imitation’—which is a built‑in aspect of re‑enactment—became the new primary meaning of this word only after its older dramatic sense became destabilized. [[In n1 of the article, I referred to an updating of my argument in the posting Nagy 2015.10.15, “Homo ludens in the world of ancient Greek verbal art.”]]

§2c. Here I will argue that Horace’s own dramatic sense of re-enactment through poetry brings to life the older and more traditional idea of mīmēsis.

§3. This idea of mimesis, which I write without italics from here on, works on the principle that those who re-enact something are not only imitating a model: by re-enacting they also become models in their own right, to be imitated by a series of successors who perpetuate, moving forward in time, the chain of re-enactment. I suggest that Horace has achieved an imaginative repetition of this idea in the poetics of Ode 4.2—and of Ode 4.1 as well.

§4. I have introduced the word repetition in precisely this context in order to evoke a 1843 work of Kierkegaard, entitled Repetition. Just as ancient Greek philosophy teaches, it is claimed, “that all knowledge is a recollecting,” so also “modern philosophy will teach that all life is a repetition” [p. 131 of the translation: see the Bibliography]. To quote further: “repetition and recollection are the same movement, except in opposite directions, for what is recollected has been, is repeated backward, whereas genuine repetition is recollected forward” [again, p. 131].

§5. Mimesis is like Kierkegaard’s repetition. When you re-enact an archetypal action in drama, you imitate those who re-enacted before you and who served as your immediate models. But the ultimate model is still the archetypal figure that you are acting out or re-enacting, who is coextensive with the whole line of imitators who had re-enacted the way in which their ultimate model acted, each imitating each one’s predecessor. When it is your turn, your moment to re-enact something in this forward movement of mimesis, you become the ultimate model in that very moment. As a working definition, I will equate this moment of mimesis with the poetic occasion.

§6a. Horace’s poetic occasion in Ode 4.2 centers on the imitation of Pindar, beginning with Pindarum quisquis studet aemulari ‘whoever is eager to imitate [aemulārī] Pindar’ at line 1. The poet’s strategy is characteristically complex, matching the complexity of his predecessor. It looks at first as if Horace does not want to have Pindar as a model. Whoever seeks to imitate Pindar, the poem starts to say, is like the doomed Icarus who botched his imitation of Daedalus: ceratis ope Daedalea | nititur pinnis ‘he relies on wings crafted by Daedalus and fastened with wax’ at lines 2–3. After all, the wings of Icarus, unlike those of his model Daedalus, were defective, even bogus.

§6b. In my posting for 2015.12.31 here, I have deleted “of Daedalus” and “Daedalus” from the wording of the original article. That is because, as I will argue in §29g, Icarus was ultimately imitating not Daedalus but someone even more exalted.

§6c. So, we may ask, with Giorgio Pasquali (1920:782), why Horace should be trying to imitate Pindar precisely at the moment when he advises against imitating Pindar? Gregson Davis offers an answer, and I agree with him that this much is relatively straightforward: “the disavowal [recūsātiō] itself,” he writes, “by virtue of its form [emphasis mine], reveals the speaker’s actual competence to undertake precisely what he claims to be incapable of doing” (Davis 1991:134). This way, the model who is Pindar becomes the foil for the new model, who is to be Horace. So maybe the wings of this new Icarus are not defective at all. We shall return in the end to this image of the winged poet, but for the moment I draw attention simply to the pride of the artist in achieving an imitation—or, better, let us call it a mimesis in the older sense of the word, in that this particular ‘imitation’ claims to become a model in its own right.

§7. The occasion of a poet, which I have just defined as the moment of mimesis, can become absolutized by his or her composition when this composition is performed. A striking example is the epinician moment as dramatized in any given song of Pindar, which comes to life in the context of its performance (Nagy PH 381 = 12§81, with further references). Seth Schein (1987: 246–247) has analyzed Pindar’s Pythian 6 as an illustration of that moment, and he quotes in this context the remarks of Hans-Georg Gadamer (1975:129), who has this to say about the element of the occasional in the epinician songs of Pindar:

The occasional in such works has acquired so permanent a form that, even without being realised or understood, it is still part of the total meaning. Someone might explain to us the particular historical context, but this would be only secondary for the poem as a whole. He would only be filling out the meaning that exists in the poem itself.

§8. Just as any given composition of Pindar is meant to absolutize the occasion of its performance, so also Horace, in composing an imitation of Pindar, imitates also the absolutization of the occasion. But in this case, following the pattern of Alexandrian poetics, the occasion may not be “real”—to the extent that there need not be a real performance. Still, the occasion is presented by the poet as absolute, and it is indeed “real” on the strength of that presentation. Provided that Horace follows the tradition, that tradition will uphold the absoluteness of the given occasion. In terms of mimesis in the older sense of the word, that is, re-enactment, the occasion of Horace is indeed real. The occasion becomes “unreal” only if mimesis is restricted to its newer sense, that is, mere imitation.

§9. I am arguing that mimesis in the older sense of the word requires that the speaker’s identity merge with that of his role as speaker, just as the identities of those who are spoken to and spoken about must merge with their respective roles. If the merger is successful, then the model has not been merely copied, that is, imitated. It has been remodeled, that is, re-enacted. What is remodeled can continue to be a model. What is merely copied cannot become a model. The paradox here is that a model implies no change, whereas whatever is remodeled does indeed imply change. That is to say, an explicit idea of unchangeability through time subsumes an implicit idea of change in the here-and-now of the occasion of performance.

A poetics of repetition in the songmaking of Sappho

§10. I maintain that the premier metaphor for this paradox of re-enactment is repetition. This metaphor is ideally expressed by adverbs meaning ‘again’, such as Greek dēute in the fourth and fifth stanzas of Song 1 of Sappho, where Aphrodite is being addressed. I quote here the full text of the song as I analyze it in the posting Nagy 2015.11.05, “Once again this time in Song 1 of Sappho”:

|1 ποικιλόθρον’ ἀθανάτἈφρόδιτα, |2 παῖ Δίοc δολόπλοκε, λίϲϲομαί ϲε, |3 μή μ’ ἄϲαιϲι μηδ’ ὀνίαιϲι δάμνα, |4 πότνια, θῦμον, |5 ἀλλὰ τυίδ’ ἔλθ’, αἴ ποτα κἀτέρωτα |6 τὰc ἔμαc αὔδαc ἀίοιϲα πήλοι |7 ἔκλυεc, πάτροc δὲ δόμον λίποιϲα |8 χρύϲιον ἦλθεc |9 ἄρμ’ ὐπαϲδεύξαιϲα· κάλοι δέ ϲ’ ἆγον |10 ὤκεεc ϲτροῦθοι περὶ γᾶc μελαίναc |11 πύκνα δίννεντεc πτέρ’ ἀπ’ ὠράνωἴθε|12ροc διὰ μέϲϲω· |13 αἶψα δ’ ἐξίκοντο· ϲὺ δ’, ὦ μάκαιρα, |14 μειδιαίϲαιϲ’ ἀθανάτωι προϲώπωι |15 ἤρε’ ὄττι δηὖτε πέπονθα κὤττι |16 δηὖτε κάλημμι |17 κὤττι μοι μάλιϲτα θέλω γένεϲθαι |18 μαινόλαι θύμωι· τίνα δηὖτε πείθω |19 βαῖϲ᾿ ἄγην ἐc ϲὰν φιλότατα; τίc ϲ’, ὦ |20 Ψάπφ’, ἀδικήει; |21 καὶ γὰρ αἰ φεύγει, ταχέωc διώξει, |22 αἰ δὲ δῶρα μὴ δέκετ’, ἀλλὰ δώϲει, |23 αἰ δὲ μὴ φίλει, ταχέωc φιλήϲει |24 κωὐκ ἐθέλοιϲα. |25 ἔλθε μοι καὶ νῦν, χαλέπαν δὲ λῦϲον |26 ἐκ μερίμναν, ὄϲϲα δέ μοι τέλεϲϲαι |27 θῦμοc ἰμέρρει, τέλεϲον, ϲὺ δ’ αὔτα |28 ϲύμμαχοc ἔϲϲο.

stanza 1||1 ‘You with pattern-woven flowers, immortal Aphrodite, |2 child of Zeus, weaver of wiles, I implore you, |3 do not dominate with hurts and pains, |4 Mistress, my heart! stanza 2||5 But come here [tuide], if ever at any other time |6 hearing my voice from afar, |7 you heeded me, and leaving the palace of your father, |8 golden, you came, stanza 3||9 having harnessed the chariot; and you were carried along by beautiful |10 swift sparrows over the dark earth |11 swirling with their dense plumage from the sky through the |12 midst of the aether, stanza 4||13 and straightaway they arrived. But you, O holy one, |14 smiling with your immortal looks, |15 kept asking what [otti] is it once again this time [dēute] that has happened to me and for what reason [k’ōtti] |16 once again this time [dēute] do I invoke you, stanza 5||17 and what [k’ōtti] is it that I want more than anything to happen |18 to my frenzied [mainolās] heart [thūmos]? “Whom am I once again this time [dēute] to persuade, |19 setting out to bring [agein] her to your love? Who is doing you, |20 Sappho, wrong? stanza 6||21 For if she is fleeing now, soon she will be pursuing. |22 If she is not taking gifts, soon she will be giving them. |23 If she does not love, soon she will love |24 even against her will.” stanza 7||25 Come to me even now, and free me from harsh |26 anxieties, and however many things |27 my heart [thūmos] yearns to get done, you do for me. You |28 become my ally in war.

Song 1 of Sappho = Prayer to Aphrodite

§11. At the beginning of this song, the female speaker invokes Aphrodite, the archetype of love, in the form of a prayer (first stanza). The goddess is then represented as flying down from Olympus, but this narration happens not in a third-person diegesis but still in the second person, so that the potential diegesis is subsumed by the syntax of prayer (second, third, and fourth stanzas, lines 5–18). Then, as the goddess arrives all the way from her distant celestial realm, she is quoted by the speaker as speaking directly in the first person to this speaker, who is now suddenly switched into the second person (fifth and sixth stanzas, lines 18–24). Aphrodite’s first question is: what is wrong with you this time (line 15)? And she is addressing a woman whom she calls Sappho (line 20). So, we learn that the speaker who had started speaking at the beginning of the song was Sappho. But now the speaker Sappho is speaking in the first person of Aphrodite (lines 18–24): she is in effect re-enacting the goddess. Moreover, at the end of Sappho’s prayer (lines 25–28), she asks to be the goddess’s equal partner, a summakhos ‘fellow fighter’ in the warfare of love. As I have argued elsewhere (Nagy GM 259–260), the active telessai ‘to get done, to get someone to experience’ as used in place of the expected passive telesthēn ‘to experience’ toward the end of the poem (line 26) suggests that the controlling plan is meant to be in the mind of Sappho, as if she were equivalent to Aphrodite herself.

§12. The re-enactment of Aphrodite as the archetype of love is made manifest by the adverb dēute ‘again’ (lines 15, 16, 18), which refers to the onset of love in the speaker’s heart. It is reinforced by the repetition of this adverb denoting repetition—three times at that (to repeat, lines 15, 16, 18). And there is further reinforcement in the triple repetition of otti/k’ōtti ‘what’ (line 15 twice, line 17 once). Yet, in this paradox of repetition, the more you hear ‘again’ or ‘one more time’, the more changes you see. It is all an archetypal re-enactment for the archetypal goddess of love, but for the humans who re-enact love it becomes a vast variety of different experiences by different people in different situations. This paradox of repetition brings to mind the words of Kierkegaard (p. 149 of the translation: see the Bibliography): “The dialectic of repetition is easy, for that which is repeated has been—otherwise it could not be repeated—but the very fact that it has been makes the repetition into something new.”

§13. The variety of erotic situations suggested by dēute ‘again’ and highlighted by the instances of amor versus ‘reverse love’ toward the end of Sappho’s song (lines 21 to 24) can also be illustrated by the strikingly plentiful set of examples that we may find in the relatively few surviving fragments of Anacreon:

σφαίρηι δηὖτέ με πορφυρῆι | βάλλων χρυσοκόμης Ἔρως | νήνι ποικιλοσαμβάλωι | συμπαίζειν προκαλεῖται

Golden-haired Eros throws at me once again this time [dēute] a purple ball, prompting me to frolic with a young girl who wears multicolored [poikila] sandals.

Anacreon PMG 13.1–4

ἀρθεὶς δηὖτ’ ἀπὸ Λευκάδος | πέτρης ἐς πολιὸν κῦμα κολυμβῶ μεθύων ἔρωτι

Lifting off once again this time [dēute] from the white rock I dive down into the gray eddies below, intoxicated with eros.

Anacreon PMG 31.1–2

μνᾶται δηὖτε φαλακρὸς Ἄλεξις

Bald Alexis is wooing once again this time [dēute].

Anacreon PMG 49b.1

παρὰ δηὖτε Πυθόμανδρον | κατέδυν Ἔρωτα φεύγων.

Once again this time [dēute] I went down, slinking, to Pythomandros, trying to get away from Eros.

Anacreon PMG 55.1–2

μεγάλωι δηὖτέ μ’ Ἔρως ἔκοψεν ὥστε χαλκεὺς | πελέκει, χειμερίηι δ’ ἔλουσεν ἐν χαράδρηι.

Once again this time [dēute] Eros struck me with a great double-axe [pelekus] like a coppersmith [khalkeus] and washed me in a wintry torrent.

[[For objections to the interpretation of pelekus as a smith’s ‘hammer’, see Goldhill 1987:9–11. He argues that, as we see from Odyssey 9.391–393, a khalkeus ‘coppersmith’ is a craftsman who has mastered the craft of manufacturing a pelekus or ‘axe’, and that such an axe is a product of the craft, not an instrument used by the craftsman in his activity as a coppersmith. I should add, though, that a coppersmith can actually be pictured as wielding his own axe: see for example Diogenes Laertius 1.76.]]

Anacreon PMG 67.1–2

ἐρέω τε δηὖτε κοὐκ ἐρέω | καὶ μαίνομαι κοὐ μαίνομαι

I passionately love [erô] once again this time [dēute]—and don’t love [erô].

Also I am mad—and I’m not mad.

Anacreon PMG 83.1–2

§14. Surveying these and other instances of dēute in Greek love lyric, Anne Carson (1986:118–119) remarks about the constituents ‘now’ and aute ‘again’: “The particle marks a lively perception in the present moment: ‘Look at that now!’ The adverb aute peers past the present moment to a pattern of repeated actions stretching behind it: ‘Not for the first time!’ places you in time and emphasizes that placement: now. Aute intercepts ‘now’ and binds it into a history of ‘thens’.”

§15. I could go on with other illustrations, but the point has already been made. Every time I say to myself, “here I go again,” I am repeating the pattern of Aphrodite, but each time it is a different experience for me. No wonder Aphrodite is invoked as poikilothronos in the first word of Song 1 of Sappho. This epithet, if indeed it is derived from throna ‘pattern-woven flowers’ rather than from thronos ‘throne’, can be translated ‘with varieties of pattern-woven flowers’. For those who re-enact her, the goddess of love is as limitlessly varied as the limitless varieties of flowers that are pattern-woven on her exterior. [[In the original printed version, I used the word ‘embroidered’ where I now use ‘pattern-woven’. For background, see MoM 2§§57–147, especially 2§70.]]

Imitations of Sappho in Horace Odes 4.1

§16. The metaphor of re-enactment as repetition, expressed in Song 1 of Sappho by way of Greek adverbs such as dēute meaning ‘again’, is in turn repeated in Horace’s use of Latin preverbs and adverbs such as re-, rursus, iterum, and so on, all meaning ‘again’. We come now to Ode 4.1 of Horace. After a long interruption of warfare in love (line 1), Venus is being addressed once again in prayer—precor, precor ‘I pray, I pray’— as she begins the battle rursus ‘once again’ (line 2). The doubling of precor, precor ‘I pray, I pray’ reinforces the repetitiveness already marked by rursus ‘once again’.

§17a. Venus the warrior reminds us of Aphrodite the summakhos ‘fellow-warrior’ invoked by Sappho in the final line of her Song 1. But now the Roman poet does not want any part of such warfare: Horace’s poem is now the opposite of Sappho’s song of invocation, since the poet here prays that the goddess be absent, not present. ‘Come here, right now!’—Sappho had prayed to Aphrodite: ἔλθε μοι καὶ νῦν (line 25). ‘Go away from here!’—Horace now prays to Venus: abī (line 7). Go instead to the young men whose prayers do indeed invoke you: quo blandae iuvenum te revocant preces ‘to where the pleasing prayers [preces] of young men are re-calling [re-vocant] you’ (line 8). The re of revocant preces reinforces the repetitiveness of their prayers invoking the goddess—the same kind of insistent repetitiveness that we saw in Sappho’s dēute ‘once again this time’.

§17b. In the original article, I compared at this point (n17) Horace Ode 4.13, which begins audivere, Lyce, di mea vota, di | audivere, Lyce ‘the gods have heard, Lycus, my prayers, | they have heard, Lycus’. The repeated sound of di reinforces the repeated words of the prayers, mea vota, which are in fact repeated prayers. This poem mentions Cinara (line 21), as does the poem that we are now considering, Ode 4.1 (line 4). More on Cinara in what follows.

A poetics of repetition in Horace Ode 4.1

§18a. In Ode 4.1, Horace shows that he too, like Sappho, had once made repetitions in his own invocations when he was a young love poet—but supposedly no more, now that he is old: non sum qualis eram, he declares: ‘I’m not the man I used to be’ (4.1.3). Then I was the love slave of the girl Cinara (line 4), and I was the young love poet of Odes Books 1–3. But now this is Book 4: I am past fifty years of age (line 6), and I am a love poet no more. Remarkably, Horace’s denial of repetition happens precisely at the moment when he is repeating the words of a love poet—just as his denial of imitating Pindar happens precisely at the moment when he replicates and even re-enacts the sum total of Pindar’s poetics. At the very moment that Horace says non sum qualis eram ‘I’m not the man I used to be’ at line 3 of Ode 4.1, as if he were no longer the love poet of Books 1–3, he not only repeats the words of a love poet, but that love poet is none other than his old self, back in his old love poems: the expression dulcium | mater saeva Cupidinum ‘savage mother of sweet passionate loves’ that follows at lines 4–5 is an imitation, even more, a repetition, of mater saeva Cupidinum ‘savage mother of passionate loves’ at line 1 of Ode 1.19 (see Putnam 1986:42; also Davis 1991:65–66). So, Horace is repeating his words of love poetry, repeating himself, repeating his self, at the very moment when he prays never again to be a love poet. He is up to his old tricks of poetic reversal. In that same earlier poem, Ode 1.19, the image of in me tota ruens Venus ‘Venus, rushing into me with her entire essence’ at line 9, so reminiscent of the erotic melt-down in Song 31 of Sappho, happens to occur in the context of an ultimate symbol of poetic reversal, the picture of Parthians shooting versis … equis ‘with horses turned around’ (Ode 1.19.11). We may compare the image of amor versus in Sappho 1.21–24, as noted above.

§18b. When I refer to an “erotic melt-down” in Song 31 of Sappho, I have in mind especially lines 5–17:

|1 φαίνεταί μοι κῆνοc ἴϲοc θέοιϲιν |2 ἔμμεν’ ὤνηρ, ὄττιc ἐνάντιόc τοι |3 ἰϲδάνει καὶ πλάϲιον ἆδυ φωνεί-|4ϲαc ὐπακούει |5 καὶ γελαίϲαc ἰμέροεν, τό μ’ ἦ μὰν |6 καρδίαν ἐν ϲτήθεϲιν ἐπτόαιϲεν, |7 ὠc γὰρ ἔc ϲ’ ἴδω βρόχε’ ὤc με φώναι-|8ϲ’ οὐδ’ ἒν ἔτ’ εἴκει, |9 ἀλλὰ κὰμ μὲν γλῶϲϲα ἔαγε λέπτον |10 δ’ αὔτικα χρῶι πῦρ ὐπαδεδρόμηκεν, |11 ὀππάτεϲϲι δ’ οὐδ’ ἒν ὄρημμ’, ἐπιρρόμ-|12βειϲι δ’ ἄκουαι, |13 κάδ δέ μ’ ἴδρωc κακχέεται τρόμοc δὲ |14 παῖϲαν ἄγρει, χλωροτέρα δὲ ποίαc |15 ἔμμι, τεθνάκην δ’ ὀλίγω ’πιδεύηc |16 φαίνομ’ ἔμ’ αὔται·

|1 He appears [phainetai] to me, that one, equal to the gods [īsos theoisin], |2 that man who, facing you |3 is seated and, up close, that sweet voice of yours |4 he listens to, |5 and how you laugh a laugh that brings desire. Why, it just |6 makes my heart flutter within my breast. |7 You see, the moment I look at you, right then, for me |8 to make any sound at all won’t work anymore. |9 My tongue has a breakdown and a delicate |10 —all of a sudden—fire rushes under my skin. |11 With my eyes I see not a thing, and there is a roar |12 my ears make. |13 Sweat pours down me and a trembling |14 seizes all of me; paler than grass |15 am I, and a little short of death |16 do I appear [phainomai] to myself.

Sappho Song 31

§18c. I see a comparable “erotic melt-down” in Horace’s Ode 1.19, picturing ‘Venus rushing into me with her entire essence’ at line 9: in me tota ruens Venus.

§19. The virtuoso display of repetitions in Horace Ode 4.1 reinforces Michael Putnam’s conclusion about this poem (1986:40): “Looked at, therefore, as an epigone to Sappho’s masterpiece, Horace’s ode is a prayer for the reritualization of Venus.” Such a reritualization, I would add, reflects Horace’s success in mimesis, and again I mean it in the older sense of the word.

A revitalization of love poetry in Horace Ode 4.1

§20. The reritualization in Horace Ode 4.1 turns out to be a revitalization. The poet’s denial of interest in love poetry, expressed as mollibus | iam durum imperiis ‘hard [durus] in responding to soft commands’ at lines 6–7, is suggestive. The word imperiis ‘commands’ in this expression picks up on the word regno ‘régime’ in the earlier expression bonae | sub regno Cinarae ‘under the régime of good Cinara’ at lines 3–4, which was the actual context for non sum qualis eram ‘I’m not the man I used to be’ at line 3. Who, then, is this Cinara, the girl who had once made the young love poet Horace become a slave to her régime? And what does it mean that the poet has by now grown ‘hard’ to the ‘soft’ imperia ‘commands’ of Venus?

§21. Let us start with Cinara. The schoolbook from which I first studied Horace many years ago (Bennett 1934:367) taught me that she was a “real” person, “alone among Horace’s loves.” Her name, however, suggests that there is more to it. The cinara is a kind of thistle, now known as Cinara scolymus in the Linnaean taxonomy. Particularly suggestive of its implications is a passage in Columella 10.235–236 where we learn that ingesting this plant was good for ‘a drinking Bacchus’ but ‘bad for a singing Apollo’: hispida ponatur cinara, quae dulcis Iaccho | potanti veniat nec Phoebo grata canenti ‘let the bristly cinara be planted, which comes across as sweet to Bacchus when he drinks but displeases Phoebus [Apollo] when he sings’. The Greek word from which the Latin cinara was borrowed is κινάρα (and variant κυνάρα), which is synonymous with σκόλυμος (skolumos). This other Greek word for ‘thistle’ is prominently featured in a drinking song of Alcaeus (fr. 347), as also in a Hesiodic passage (Works and Days 582–588; on the relationship between the Alcaic and Hesiodic passages, see Nagy PH 462–463 = Appendix §§35–36). Gregson Davies (1991:68–69) was the first to notice the sympotic connotations of the name Cinara by making the connection with the Alcaic and Hesiodic passages; he links Horace’s reference to Cinara in Ode 4.1 to the very idea of the symposium—and, by extension, to the idea of sympotic poetry. Going further, Michèle Lowrie (1991) notes another detail in both these Greek passages: the blossoming of the thistle marks the time in the Dog Days of summer when men are least potent sexually—and women are most desirous and lascivious (for more on the Dog Days, I cite Oliensis 1991, Ceragioli 1992, Petropoulos 1994). Lowrie combines this detail with the observations of Pliny Natural History 20.262 and 22.86–87, to the effect that the thistle is an aphrodisiac when it is ingested with wine. I agree therefore with Lowrie’s conclusion that “the reign of Cinara was erotic as well as sympotic.”

§22. Let us return, then, to the passage that I cited from Columella, which says that the cinara is good for ‘a drinking Bacchus’ but ‘bad for a singing Apollo’. If Horace was following a régime—let us say a diet—of cinara, on the grounds that the thistle is an aphrodisiac when ingested with wine, then the poet’s declaration that he is now ‘hard’ to the ‘soft’ régime of love may well be yet another example of Horatian affirmation through denial. The poet may say that he has lost interest in Venus, and yet he is now ‘hard’ to her ‘soft’ régime. What, then, do we make of the ancien régime of Cinara? Evidently it is not after all a thing of the past, if indeed the poet is now ‘hard’ in response to Venus’s ‘soft’ commands.

§23. Here I invoke the important observations of Ellen Oliensis (1991) concerning the negative imago of the Horatian female love-partner, the notorious Canidia, especially as she figures in Epodes 5 and 17 as well as in Satire 1.8. In Epode 17, as Oliensis shows (1991:120), the poet in fact prays to Canidia in what amounts to a sinister mirror image of the prayer to Venus in Ode 4.1:

‘On my knees I pray, … leave off (parce) casting spells, and loose, loose the magic wheel (solve, solve turbinem) so it may run backward.’ [= Epode 17.2, 6–7] In both poems, Horace pleads unfitness for love—decorously formulated in the late ode (non sum qualis eram … [‘I’m not the man I used to be’]), but spelled out in ugly detail in the epode (21–22): ‘My youth has fled, the modest blush has abandoned my bones in their wrapper of sallow hide …’ (fugit iuventas et verecundus color | reliquit ossa pelle amicta lurida …). If Venus is ‘looking for a suitable liver to scorch’ (si torrere iecur quaeris idoneum [Ode 4.1.]12), Canidia already has Horace on the coals ([Epode 17.]30, 33–35): ‘O sea and earth, I’m on fire.  … Are you going to keep up the heat, you workshop of Colchian drugs, until I’m nothing but arid ashes scattered by the injurious winds?’

§24. As Oliensis observes (1991:116n20), the mark of Canidia is the Dog Star, and, in all three of the Horatian poems where Canidia plays a central role, Epodes 5 and 17 and Satire 1.8, “an overheated Canidia is viewed through the eyes of an impotent male” (p. 121). The Horatian image of male impotence goes back to a central idea in the ancient Greek drinking song as enacted in Alcaeus fr. 347 and as paraphrased, as it were, in Hesiod Works and Days 582–588. That central idea, as formulated by Oliensis (1991:121; also Detienne 1977: 121–122), is this: “it is the woman’s heat that saps the man’s strength.”

§25. According to Oliensis (1991:121), the Horatian image of impotentia is two-sided, conveying the “twinned meanings” of “lack of power, or weakness,” and “lack of self-control, or violence.” From the standpoint of the earlier Greek poetics, however, I would suggest more simply that impotentia has a double meaning, one for male and one for female, conveying simultaneously ‘impotence’ on the part of the male and ‘loss of self-control’ on the part of the female, as is made explicit in the drinking song about the Dog Days. At Epode 16.53–54, there is a reference to the aestuosa inpotentia ‘loss of potentia in the heat-wave’ of the Dog Star. I see both meanings implicit here.

§26. The very notion of a female ‘loss of self-control’ is of course subjective, defined by a male point of view. Negatively portrayed as it is, female sexuality becomes a reflection of male impotentia. And here we come to a paradox: as Oliensis (1991:125) observes about Roman love poetry, the very state of being in love is presented by the male poet as a matter of impotentia, given that “the traditional opposition of the erotic lifestyle to military heroism sets up an opposition of ‘soft’ to ‘hard’, feminine to masculine, disarmed to armored, amorous complaint to aggressive invective”. The “solution” perpetuates the paradox: “to regain his potency,” the male lover in Roman love poetics “must either borrow the attributes of the soldier—hence the pervasive motif of the ‘soldier-lover’, especially in Roman elegy—or else he must assert his masculine independence from his mistress, adopting the tough stance of invective.” Such “toughness” or “hardness” requires a rejection of love. As an example of this paradox, Oliensis (1991:125) cites Catullus 8, where the idea of being in love with Lesbia is the same thing as being impotens, that is, ‘powerless’: tu quoque impotens noli ‘you also, powerless, must decide not to be willing’ (line 9); so, the poet keeps telling himself to ‘stay hard’ in order to escape his domina ‘mistress’: obdura (line 11), obdurat (line 12), obdura (line 19).

§27. So also in Ode 4.1 of Horace, we see that the poet has now become ‘hard’ to the ‘soft’ domination of Venus: mollibus | iam durum imperiis ‘hard [durus] in responding to soft commands’ (lines 6–7). Moreover, he addresses as ‘O you hard one [durus]!’ that last object of his desire, what we may call that “sweet bird of youth” (I recall here the title of a 1959 play by Tennessee Williams). The ‘you’ who is addressed by the poet eludes him endlessly in his sleep as he now reaches the end of his poem:

nocturnis ego somniis | iam captum teneo, iam volucrem sequor | te per gramina Martii | Campi, te per aquas, dure, volubilis

In my nighttime dreams, I | at one moment catch up with you and hold on to you, while at the next moment you take wing, flying off, | and I pursue you as you speed over the grounds of Mars’ | field and over the swirling waters, O you hard one [durus]!

Horace Ode 4.1.37–40

The winged boy and the swan

§28a. Who, then, is that winged boy, that last object of the poet’s desire? What do we know about him, besides the description volucris ‘winged, taking wing’ at line 38 and durus ‘hard’ at line 40 of Ode 4.1? A key is his name, Ligurinus:

sed cur heu, Ligurine, cur | manat rara meas lacrima per genas?

But why, I say with regret, O Ligurinus, why | is there this tear, so rare, that flows down my cheeks?’

Horace Ode 4.1.33–34

§28b. I am persuaded by the suggestion of Michael Putnam (1986:44–46) that Ligurinus, the “Ligurian” of line 33 in Ode 4.1, reflects an Italic version of Cycnus the “Swan” (Greek κύκνος ‘swan’). The story of the hero Cycnus and how he changed into a swan is well known to us from Ovid Metamorphoses 2.367–380. Also, Cycnus was a lover of that hero with a solar name, Phaethon, as we see from the reportage of Servius on Aeneid 10.189: fuit etiam Ligur, Cycnus nomine, dulcedine cantus ab Apolline donatus, amator Phaethontis ‘there was also a Ligurian by the name of Cycnus who was given the sweetness of song by Apollo and who was a lover of Phaethon’. In n39 of my original article, building on the connection of Cycnus with Phaethon as noted by Putnam (1986:44), I compared the theme of being in love with the sun itself, as made explicit at lines 25–26 of Fragment 58 of Sappho. These two lines, as we now know, turn out to be the last two lines of Sappho’s Tithonos Song, which I translated and analyzed in my posting for 2015.11.12:

|1 […words missing… ἰ]ο̣κ[ό]λ̣πων κάλα δῶρα, παῖδεϲ, |2 […words missing… τὰ]ν̣ φιλάοιδον λιγύραν χελύνναν· |3 […words missing…] π̣οτ̣’ [ἔ]ο̣ντα χρόα γῆραϲ ἤδη |4 […words missing…]ἐγ]ένοντο τρίχεϲ ἐκ μελαίναν· |5 βάρυϲ δέ μ’ ὀ [θ]ῦμο̣ϲ̣ πεπόηται, γόνα δ’ [ο]ὐ φέροιϲι, |6 τὰ δή ποτα λαίψηρ’ ἔον ὄρχηϲθ’ ἴϲα νεβρίοιϲι. |7 τὰ ⟨μὲν⟩ ϲτεναχίϲδω θαμέωϲ· ἀλλὰ τί κεν ποείην; |8 ἀγήραον ἄνθρωπον ἔοντ’ οὐ δύνατον γένεϲθαι. |9 καὶ γάρ π̣[ο]τ̣α̣ Τίθωνον ἔφαντο βροδόπαχυν Αὔων |10 ἔρωι φ̣ ̣ ̣α̣θ̣ε̣ιϲαν βάμεν’ εἰϲ ἔϲχατα γᾶϲ φέροιϲα[ν], |11 ἔοντα̣ [κ]ά̣λ̣ο̣ν καὶ νέον, ἀλλ’ αὖτον ὔμωϲ ἔμαρψε |12 χρόνωι π̣ό̣λ̣ι̣ο̣ν̣ γῆραϲ, ἔχ̣[ο]ν̣τ̣’ ἀθανάταν ἄκοιτιν. |13 […words missing…]ιμέναν νομίϲδει |14 […words missing…]αιϲ ὀπάϲδοι |15 ⸤ἔγω δὲ φίλημμ’ ἀβροϲύναν, …⸥ τοῦτο καί μοι |16 τὸ λά⸤μπρον ἔρωϲ ἁελίω καὶ τὸ κά⸥λον λέλογχε.

|1 [… ] gifts of [the Muses], whose contours are adorned with violets, [I tell you] girls [paides] |2 […] the clear-sounding song-loving lyre. |3 […] skin that was once tender is now [ravaged] by old age [gēras], |4 […] hair that was once black has turned (gray). |5 The throbbing of my heart is heavy, and my knees cannot carry me |6 ­—(those knees) that were once so nimble for dancing like fawns. |7. I cry and cry about those things, over and over again. But what can I do? |8 To become ageless [a-gēra-os] for someone who is mortal is impossible to achieve. |9 Why, even Tithonos once upon a time, they said, was taken by the dawn-goddess [Eos], with her rosy arms |10—she felt […] passionate love [eros] for him, and off she went, carrying him to the ends of the earth, |11 so beautiful [kalos] he was and young [neos], but, all the same, he was seized |12 in the fullness of time by gray old age, even though he shared the bed of an immortal female. |13 […]|14 […]|15 But I love delicacy [(h)abrosunē] […] this, |16 and passionate love [erōs] for the Sun has won for me its radiance [tò lampron] and beauty [tò kalon].

Sappho, Tithonos Song

§28c. The idea of conceiving a passionate love for the sun is connected in the poetics of Sappho with the idea of her own passionate love for a beautiful imaginary boy named Phaōn, whose name is morphologically parallel to Phaethōn. As I have argued at length elsewhere (Nagy GM 260–262), both these names Phaōn and Phaethōn were epithets of the sun-god, meaning ‘radiant’, and both these names were associated with myths about the setting and the rising of the sun. And now, as we just saw at §28b in the case of the mythical boy named Phaethon, his lover was none other than the mythical swan-boy Cycnus, who had been given the voice of a virtuoso singer by the solar god Apollo himself.

Horace and the swan

§28d. Having made the point that Ligurinus is connected with Cycnus as a stylized swan, Putnam (1986:45) goes on to argue that Ligurinus is a self-extension of Venus. I raise the question, further, whether Ligurinus is a self-extension of Horace. Putnam himself says (again, 1986:45): “Ligurinus is very much the speaker-poet’s former self.”

§29a. Earlier (§28a), I quoted line 33 from Odes 4.1, featuring the boy named Ligurinus. Still earlier (§27), I quoted line 38 from the same ode, where this Ligurinus is described as volucris ‘winged, taking wing’. Putnam (1986:45) notes that the swan, as evoked by the name Ligurinus at line 33 and by the word volucris ‘winged, taking wing’ at line 38, is of course none other than the bird of Apollo. Perhaps the most memorable passage about the swan as Apollo’s songbird is Plato, Phaedo 84d–85b, where Socrates speaks of his own impending death and compares the swan song, that most beautiful of all the songs that the bird sings in its lifetime, which is to be sung when it prophetically foresees the impending moment of its own death. I quote, translate, and analyze this passage from the Phaedo in Nagy H24H 22§§34–35.

§29b. The swan is also the archetypal poet, precisely because he is the bird of Apollo, and Pindar himself is called ‘the swan [cycnus] of Dirce’ in Ode 4.2 (Dircaeumcycnum, line 25)—that is, in the very same poem where Pindar becomes the prime object of Horace’s imitation and even replication.

§29c. Even more, as Putnam has noted (1986:45; also Davis 1991:135), Horace imagines himself as turning into a swan in another ode:

iam iam residunt cruribus asperae | pelles et album mutor in alitem | superne nascunturque leves | per digitos umerosque plumae

Any moment now, my rough skin is sloughed off, down to my legs, while in my upper half I am transformed into a white winged bird, and a light plumage is growing over my fingers, my shoulders.

Horace Ode 2.20. 9–12

§29d. In this same poem, Ode 2.20, the vates ‘poet’ (line 3)—or swan—is seen flying through the aether at line 2. At lines 15–16, he is described as a canorus | ales ‘winged singer’, canorus | ales, who at lines 13–16 is seen flying over vast stretches of water and land. At line 16 I note the usage of campos for ‘land’, which may be compared with the campi of per gramina Martii | campi ‘over the grounds of Mars’ | field’ at line 39 of Ode 4.1.

§29e. In this context, I cite the work of Jaeger (1995), who makes clear the importance of the Campus Martius or ‘Mars’ Field’ as the poetic space that brings to life, as it were, such images as the athletic boy of Ode 4.1.37–40. In Jaeger’s interpretation, as I stress at n37 of my original article, this youthful athlete is figured as a training partner or pacesetter, as it were, for the aging poet.

§29f. I propose, then, that the ever-elusive object of the poet’s ultimate desire is his own self as the ultimate poet, as that archetypal bird of song, the swan of Apollo. And, as the swan of Apollo, the poet describes himself at line 13 of Ode 2.20 as the rival of Icarus, son of Daedalus.

§29g. So, I come back to the rivalry of Horace with Pindar in Ode 4.2. And I repeat what I noted at §6a… Whoever seeks to imitate Pindar, as we read at lines 2–3 of Ode 4.2, is like the doomed Icarus who botched his imitation: ceratis ope Daedalea | nititur pinnis ‘he relies on wings crafted by Daedalus and fastened with wax’. After all, the wings of Icarus, unlike those of his model, were defective, even bogus. So, I finally come back to the question: who is the model of Icarus? The answer should by now be clear. Surely the ultimate model is the solar god Apollo himself. But the wings of Icarus were artificial, and so he failed to live up to the standards of the sun when he got too close and the wax melted. And who is the model of the poet? Surely the ultimate model is, again, the solar god Apollo. Apollo is the model for the poet Pindar, that swan of Dirce, who in turn is the model for the poet Horace. But now the wings of the poet are not artificial: they are natural, because they are the wings of the swan. The poet has turned into a swan, and thus the poet himself is natural, not artificial.

It’s now Apollo’s turn to sing

§30a. We have seen from the verses of Columella that the cinara is bad for the poet’s voice, bad for the singing Apollo. But the poet in Ode 4.1 says: non sum qualis eram bonae | sub regno Cinarae ‘I’m not the man I used to be under the régime of good Cinara’. Suddenly, I am not the man I used to be—when I used to be under the régime of Cinara. I am no longer the love poet who sings at symposia. The regimen of cinara is over. The aphrodisiac diet is finished. Bacchus must stop drinking and Apollo can start singing again. So too his bird of song can start to sing. And so too the poet Pindar can now sing, replicated by Horace. The poet replicates the poet who replicates the swan who replicates Apollo. But the song of the swan, bird of Apollo, threatens to become the swan song for Horace. As Putnam remarks (1986:46), “the call of the swan elicits what is in fact a moving swan song to his human and imaginative past.” And the winged boy remains, alas for the poet, durus ‘hard’ at the end of Ode 4.1. That is much to be regretted by one who is now durus to the régime of Venus, as we had heard him say at the beginning of Ode 4.1—unless the durus man and the durus boy can become replicas of each other.

§30b. While thinking about the “winged boy” and the ancient connections between Pindar and Horace, I find it helpful to make a typological comparison with two pieces of modern visual art: “Evening (The Fall of Day)” by William Rimmer (circa 1869–1870) and the logo for Swan Song as created by graphic artist Joe Petagno for the rock band Led Zeppelin in 1974. The visual imagery in these two pieces is comparable to the verbal imagery at lines 13–16 of Horace Odes 2.20 as highlighted at §29d above. Claudia Filos has pointed out to me the evocative picturing of the waters and the land in the color version by Petagno, which resembles Horace’s own picturing in Odes 2.20.

Appendix: About an allusion to Sappho 31.9 in Horace Ode 4.1.35–36

[[From n19 in the original article…]]

cur facunda parum decoro | inter verba cadit lingua silentio

Why does my eloquent tongue in less-than-seemly…|—between one word and the next—fail…silence?’

Horace Ode 4.1.35–36

The effect of an unconventional elision necessitated by the juxtaposition of o and i in decoro | inter is used here as a symbol for the actual meaning of the two lines. In my translation, the wording ‘less-than-seemly silence’ is interrupted by an ellipsis that I symbolize as “…”): so, ‘less-than-seemly…silence’. What interrupts, as marked by the ellipsis, is the wording ‘—between one word and the next—fail’. (Cf. Putnam 1986:38.) Such use of a sound-effect to match meaning may be compared with the “gagging” effect of the hiatus in the sequence γλῶσσα ἔαγε (‘[my] tongue broke down’) of Sappho 31.9, where the sound of gagging conveyed by hiatus matches an expression that designates the sensation of gagging. See Nagy 1974:45 and 2009:69–72. See also Davis 1991:255 for Horatian imitations, besides Ode 4.1, of Sappho 31 (especially Ode 1.13.5–8, Epodes 11.8–10, 14.1–8).



Bennett, C. E. 1934. Horace: Odes and Epodes (revised by J. C. Rolfe). Boston.

Carson, A. 1986. Eros the Bittersweet: An Essay. Princeton. 118–119.

Ceragioli, R. C. 1992. Fervidus ille Canis: The Lore and Poetry of the Dog Star in Antiquity. Harvard Ph.D. dissertation.

Davis, G. 1991. Polyhymnia: The Rhetoric of Horatian Lyric Discourse. Berkeley and Los Angeles.

Detienne, M. 1977. The Gardens of Adonis: Spices in Greek Mythology (translated by J. Lloyd). Hassocks, Sussex.

Gadamer, H.-G. 1975. Truth and Method (translation edited by G. Barden and J. Cumming). New York. Originally published 1960 as Wahrheit und Methode. Tübingen.

GM. See Nagy 1990b.

Goldhill, S. 1987. “The Dance of the Veils: Reading Five Fragments of Anacreon.” Eranos 85: 9–18.

H24H. See Nagy 2013.

Jaeger, M. 1995. “Restructuring Rome: The Campus Martius and Horace, Ode 1.8.” Arethusa 28:177–191.

Kierkegaard, S. [1983.] Fear and Trembling. Repetition (edited and translated, with introduction and notes, by H. V. Hong and E. H. Hong; Princeton 1983). The original dateof Kierkegaard’s Repetition is 1843.

Lowrie, M. 1991. Review of Davis 1991. Bryn Mawr Classical Review 02.07.02,

MoM. See Nagy 2015.

Nagy, G. 1974. Comparative Studies in Greek and Indic Meter. Harvard Studies in Comparative Literature 33. Cambridge MA.

Nagy, G. 1990a. Pindar’s Homer: The Lyric Possession of an Epic Past. Baltimore.

Nagy, G. 1990b. Greek Mythology and Poetics. Ithaca NY. Revised paperback edition 1992.

Nagy, G. 2009. “The Fragmentary Muse and the Poetics of Refraction in Sappho, Sophocles, Offenbach.” Theater des Fragments: Performative Strategien im Theater zwischen Antike und Postmoderne (ed. A. Bierl, G. Siegmund, Ch. Meneghetti, C. Schuster) 69–102. Bielefeld. In Nagy 2012 v1.

Nagy, G. 2013. The Ancient Greek Hero in 24 Hours. Cambridge MA.

Nagy, G. 2015. Masterpieces of Metonymy. Hellenic Studies Series 72. Washington and Cambridge MA.

Oliensis, E. 1991. “Canidia, Canicula, and the Decorum of Horace’s Epodes.” Arethusa 24:107–138.

Pasquali, G. 1920. Orazio lirico. Florence.

Petropoulos, J. C. B. 1994. Heat and Lust: Hesiod’s Midsummer Festival Scene Revisited. Lanham MD.

PH. See Nagy 1990a.

PMG = Page, D. L., ed. 1962. Poetae Melici Graeci. Oxford.

Putnam, M. J. 1986. Artifices of Eternity: Horace’s Fourth Book of Odes. Ithaca.

Schein, S. L. 1987. “Unity and Meaning in Pindar’s Sixth Pythian Ode.” MHTIC: Revue d’Anthropologie du Monde Grec Ancien 2:235–247.