Erotic Magic, Elegy, and Iambic in Horace’s Epodes

DOI : 10.54563/eugesta.1264


This paper explores Horace’s engagement with Latin love elegy through the motif of magic in Epodes5 and17. I argue that Horace’s Canidia and the elegiac puellae are constructed in tandem as part of an ongoing dialogue between Horace’s Epodes and early love elegy. Horace literalises the elegiac lovers’ metaphorical characterisation of their mistress’ beauty as magically enchanting to construct Canidia as a witch who controls her lovers with erotic magic. Through Canidia and the theme of magic in the Epodes, Horace uses elegy to define his iambic poetics and to demonstrate that love elegy is central to his development of the Roman iambic tradition.



The narrators of Latin love elegy enthusiastically praise the beauty of their puellae: they present it as the reason for their servitium to their mistresses and they illustrate its power to captivate them by claiming that it is more enchanting than magic carmina1. By contrast, Canidia, the old and grotesquely ugly witch of Horace’s Epodes and Satires, is forced to use practical erotic rituals to attract lovers. On the surface, Canidia and the elegiac puellae appear to be polar opposites but they share a fundamental similarity: they all embody the text and the poetics of their respective genres. Canidia’s identity as a witch is integral to her metapoetic roles as Horace’s “anti-Muse” and the embodiment of his iambic poetry2. In this article, I argue that Canidia and the elegiac puellae are constructed in tandem as part of an ongoing dialogue between Horace’s Epodes and the developing genre of Latin love elegy, as exemplified by the first books of Propertius and Tibullus3.

After a survey of elegy in the Epodes, my argument falls into two parts. In the first part, I suggest a new reading of the function of magic in Latin love elegy, both within the fictional narrative of the elegiac affair and on the level of genre. I argue that, within the elegiac narrative, the lovers collapse the distinction between magic and beauty which they use to flatter their puellae and imply that their mistresses use erotic magic to control and dominate them. The suggestion that their passion is the result of erotic magic shifts the elegiac lovers’ responsibility for their servitium to their puellae and presents it as something less desirable than they would like to suggest that it is. On a metapoetic level, I argue that the connection of the puella’s beauty with magic suggests that she also embodies elegy’s enchanting power over its audiences and over the elegiac lover himself. Overall, the connection of magic with the elegiac puellae and with the poet’s text suggests that this theme is central to Latin love elegy’s self-construction and self-representation as a genre4.

In the second part of my argument, I explore Horace’s engagement with elegy through the motif of magic in Epodes 5 and 17. I argue that Horace parodies the elegiac lovers’ metaphorical attribution of magic to their puellae by representing it as practical magic used by Canidia to control male desires. Through this, Horace foregrounds the self-serving nature of the elegiac lovers’ claim that their mistresses have enchanted and enslaved them by magic. As magic plays a metapoetic role in the Epodes as much as in Latin love elegy, I argue that this use of magic in Epodes 5 and 17 also allows Horace to make a statement about his own Roman iambic: the elegiac element in Canidia’s construction illustrates the multiplicity of genres with which Horace enriches his Epodes. Identifying the elegiac element in Epode 5, in particular, indicates that love elegy is central to Horace’s poetics throughout the Epodes, not simply in the second half of the book, and that the generic experimentation and flexibility which comes to the fore in the metrical and thematic variety of Epodes 11-16 is already embodied in Horace’s iambic anti-muse5.

I. Elegy in the Epodes

Horace engages with elegy throughout his poetic career. He regularly caricatures the genre – particularly the lover’s slavish devotion to a single beloved – and uses it as a foil for his own poetics6. Epode 11 is a well-established example of this in his iambic collection: the narrator enacts the elegiac exclusus amator’s (“locked-out lover’s”) submission to a single unrequited love before the final lines undercut this with the admission that not only has his passion for Lyciscus freed him from his love for Inachia but that another pretty boy or girl would cure his current obsession7. Previous scholarship, despite using parallels from Tibullus and Propertius to illustrate the elegiac nature of Epode 11, has largely considered Horace to be reacting to Cornelius Gallus. This is because the traditionally accepted chronology of publication for the Epodes, Propertius Book 1, and Tibullus Book 1 dates Propertius’ first book to 29-28 B.C., Tibullus’ to 27 B.C., and the Epodes to late 31 or early 30 B.C.8. This dating should not, however, preclude dialogue between the early work of Propertius, Tibullus, and Horace. The three poets will have been composing their works contemporaneously: Horace compiled the Epodes over approximately ten years between 42-30 B.C.; Propertius and Tibullus likely worked for comparable periods on their books9. The mingling of contemporary circles of poetic patronage, the release of individual poems ahead of the completed books, and the pre-publication circulation of work through recitations and early drafts would have provided ample opportunities for contact between the three authors, and for their continual access to and mutual creative involvement with one another’s work10. It is likely that Horace will have been aware of the key motifs and language of love elegy, including, as I will argue, the metapoetic function of magic in the genre and its connection with the elegiac puellae. These forms of literary interaction are regularly cited as, for example, the means for Horace and Tibullus alluding to Vergil’s Georgics and Aeneid before their publication11. What, then, should prevent the same means of interaction between Horace’s Epodes and the first books of Propertius and Tibullus?

Recent scholarship has begun to trace such interaction between Horace’s Epodes and Propertius Book 1. Lyne and Heslin have demonstrated that Epode 11 may be responding to Propertius Book 1 as well as to Gallus’ Amores12. Heslin reads Epode 11 as the culmination of a poetic exchange between Horace’s Satires and Epodes and Propertius Book 1, in which Horace responds, in particular, to Propertius 1.4 and 1.1. Most notably for my purposes, Heslin highlights the irony of “Bassus”, the iambographer whom Propertius’ narrator addresses in 1.4, praising the beauty of other women to the elegiac lover considering the vitriolic treatment of women in the Epodes and comments that: “the ugliness of Canidia has almost as prominent a role to play in Horace’s early work as the beauty of Cynthia has in Propertius; they might even be considered mirror-images”13. Damer and Skinner have recently built on Heslin’s work to argue that the women in Epodes 8 and 12 respond to early love elegy, particularly to Propertius. Damer reads the women of Epodes 8 and 12 as embodiments of Horace’s iambic poetics who engage with the elegiac characteristics of mollitia (“softness and effeminacy”) and inertia (“inactivity”) to define Horace’s work against the developing genre of love elegy14. Skinner argues that the old woman of Epode 12 responds to Cynthia’s attack on the elegiac lover in Propertius 1.3 to parody elegiac gender dynamics, incorporating polemic with Propertius’ early work into Horace’s self-definition against Catullus as part of the Latin iambic tradition15.

Collectively, these readings highlight a persistent pattern in the Epodes: women and their bodies are used as a focus for parodying the gender dynamics and the rhetoric of the elegiac lover in early Propertian and Tibullan love elegy. This use of female figures contributes to the gender dynamics of the Epodes, which share with love elegy a concern with weakened masculinity and disrupted identity at the end of the Republic16. Horace’s engagement with elegy in connection with these themes allows him both to assert his iambic aggression in contrast to the elegists’ subservience to their puellae and to underline the iambic narrator’s mollitia and impotence throughout the Epodes.

I suggest that it is possible to gain a fuller understanding of Horace’s engagement with early Propertian and Tibullan elegy by investigating the relationship between the elegiac puellae and Horace’s Canidia. To do this, I focus on the connection of Canidia, Propertius’ Cynthia, and Tibullus’ Delia and Pholoe with erotic magic, a motif which is central to the narrative and metapoetic roles of these women. As outlined above, I begin by exploring the connection between beauty and magic in the first books of Tibullus and Propertius and the narrative and metapoetic functions of this connection. I then move on to Horace’s expansion of the elegiac magical subtext into a main narrative element in Epodes 5 and 17 to caricature the power which the elegiac puellae are presented as wielding over their lovers and the same lovers’ self-serving rhetoric.

II. Puellegy

Erotic magic features prominently in the first books of Propertius and Tibullus. Both poets construct their elegy in terms of magic carmina which reflect the characteristic themes of the elegiac affair – flattering and seducing girls, opening doors, and deceiving husbands and guards – and which illustrate the enchanting power of their verses for their extratextual audiences. Within the narrative of the elegiac affair, the representation of poetry as magic carmina expresses the lover’s attempts to seduce his mistress in terms of magical enchantment. The alignment of poetry with magic highlights that the lover aims to deceive his puella with his poetry; the doubts which the lover casts on the power of magic, however, foreshadow his failure to seduce his mistress and indicate that, as his verses continue to offer the hope of successfully persuading his mistress to love him, their deception works more effectively on him than on her17.

This connection between magic and elegiac carmina is expressed most explicitly in Propertius’ programmatic elegy 3.3, where Calliope explains to Propertius what his poetry should include (47-50):

quippe coronatos alienum ad limen amantis
    nocturnaeque canes ebria signa morae,
ut per te clausas sciat excantare puellas
   qui volet austeros arte ferire viros.

For you will sing about garlanded lovers at another’s doorway
    and the drunken signs of nocturnal delays,
so that he who would wish to trick strict husbands
   may learn from you how to charm out locked-in girls with magic.

Excantare (“to charm out”, 49), which is used as early as the Twelve Tables for the magic effect of words, implies that Propertius’ elegy has an influence on his puella that is similar to the effect of magic18. 3.3 is a polite rejection of epic poetry (recusatio) in which Calliope and Apollo advise Propertius to return to the elegiac subject matter with which he launched his career; the poem highlights that Book 3 will take Propertius’ elegy in a new direction from that of his previous books19. The characterisation of Propertius’ elegy as a form of magic enchantment here suggests that this is a construction of earlier elegy which readers would have been familiar with.

Indeed, Propertius uses magic programmatically to characterise his poetry in the first elegy of Book 120. In the centre of 1.1, the lover appeals to witches, asking them to change Cynthia’s mind and make her love him even more than he loves her (19-24):

at vos, deductae quibus est fallacia lunae
   et labor in magicis sacra piare focis,
en agedum dominae mentem convertite nostrae,
   et facite illa meo palleat ore magis.
tunc ego crediderim vobis et sidera et amnes
   posse Cytaeines ducere carminibus.

But you, who know the trick of the drawn-down moon
    and who perform sacred rites on magic hearths,
come on, change the mind of my mistress
    and make her face paler than my own.
Then I may believe that you can lead stars
    and rivers with Colchian songs.

The expression deductae … lunae (“drawn-down moon”) in the opening line of the lover’s appeal (19) suggests that magic is used to represent poetic carmina in this passage: deducere (“to draw down”) is used programmatically by Roman poets for the composition of refined Callimachean verse, indicating the connection between magic carmina and poetry here21. This connection is made explicit in lines 23-24, where carmina are introduced emphatically as the final word of the passage and deductae (19) is echoed in ducere (“lead”, 23)22. Propertius explicitly calls the effect of these carmina a “trick” (fallacia, 19), highlighting deception as a defining feature of elegy – one which is connected similarly with magic in 3.3.49-50. Medea’s introduction alongside carmina in line 24 underlines this by alluding to Medea’s ability to control the Moon with “deceitful spells” at Apollonius Rhodius’ Argonautica 4.59-60 (δολίῃσιν ἀοιδαῖς, 59)23. The suggestion that fallacia characterises Propertius’ elegy in 1.1 is also reinforced by the programmatic 4.1, which opens his final book. The astrologer Horos, attempting to dissuade Propertius from his new program of aetiological elegy, reminds him of Apollo’s past order that Propertius compose love elegies: at tu finge elegos, fallax opus (“but you, create elegies, deceitful work”, 135)24. In 1.1, the appeal to carmina with the power to trick indicates the lover’s attempts to deceive Cynthia with his poetry and illustrates the enchanting power of Propertius’ elegy. At the same time, the subjunctive crediderim (“I may believe”, 23) and the mention of Medea, who does not use erotic magic to win love, hints at the failure of the elegiac lover’s poetry to deceive Cynthia and suggests that he is the one tricked into hoping that he will be able to win her with his elegies25.

Tibullus connects his poetry with magic in his first paraclausithyron (“lament outside a locked door”), 1.2. The lover claims that an honest witch has written a spell that will enable Delia to deceive her husband: nec tamen huic credet coniunx tuus, ut mihi verax / pollicita est magico saga ministerio (“nevertheless, your husband will not believe it, as an honest witch has promised me with her magic aid”, 43-44); haec mihi composuit cantus quis fallere posses (“she has composed me a spell which will enable you to deceive”, 55). To convince Delia of the witch’s abilities, the lover catalogues the powers of her magic (45-54):

hanc ego de caelo ducentem sidera vidi;
    fluminis haec rapidi carmine vertit iter;
haec cantu finditque solum manesque sepulcris
    elicit et tepido devocat ossa rogo.
iam tenet infernas magico stridore catervas;
    iam iubet aspersas lacte referre pedem.
cum libet, haec tristi depellit nubila caelo;
    cum libet, aestivo convocat orbe nives.
sola tenere malas Medeae dicitur herbas,
    sola feros Hecatae perdomuisse canes.

I have seen her drawing the stars from the sky;
   she turns the course of the rapid river with her song;
with her song she splits the earth and calls forth ghosts
   from the tombs and calls down bones from the warm pyre.
Now she holds back the infernal hoards with her magic shriek;
     now she orders them, sprinkled with milk, to retreat.
When she wants, she expels the clouds from the sad heaven;
     when she wants, she summons snow in the summer sky.
She alone is said to understand Medea’s evil herbs,
    she alone to have tamed Hecate’s wild dogs.

Tibullus attributes the witch’s power to her words, foregrounding speech: carmine (46), cantu (“song”, 47), devocat (“calls down”, 48), stridore (“shriek”, 49), convocat (“summons”, 52). This focus on words, together with the use of compono (“compose”) for the composition of the witch’s spell (55), aligns her erotic magic with the lover’s – and Tibullus’ – composition of elegy in 1.2 which is designed to flatter and seduce Delia: blandaque compositis abdere verba notis (“and hide seductive words with pre-arranged signals”, 22), and et sibi blanditias tremula componere voce (“and rehearses seductions with a trembling voice”, 93)26. The urban paraclausithyron of 1.2 disrupts both the pastoral fantasy and the poetic programme established by Tibullus 1.1, developing and clarifying the image of the lover chained outside Delia’s door at 1.1.51-56 to reassert a new and complementary programme for Tibullus’ book, one which highlights the power of elegy to enchant its readers and to create fantasies27. The framing of the passage with questions of belief and deception (1.2.43-44 and 55) and the inclusion of Medea at 1.2.53 implies that the lover’s poetry will be incapable of seducing Delia and that he is the one who is deceived by the false hope of his elegy28.

Magic is also associated with the elegiac puellae, specifically with their beauty and its effect on their lovers, in both Propertius Book 1 and Tibullus Book 1: the lovers illustrate the irresistible force of their attraction by favourably comparing their mistresses’ appearances with magic. At first glance, this favourable comparison of beauty and magic compliments the girls. I argue, however, that the terms in which the lovers draw these comparisons collapse the distinction between magic and beauty, implying that their puellae bewitch them magically after all29. Metaphorically equating the effects of magic and beauty suggests that the power the elegiac mistress holds over her lover originates from a source external to her. This justifies the lover’s servitium and removes his responsibility for it by implying that it stems from a supernatural power beyond his control, rather than from his physical attraction to a pretty girl. The elegiac lover’s implicit characterisation of his beloved, her attractiveness, and his love for her in terms of magic betrays an ambivalence towards his mistress and his passion for her which lurks beneath his flatteries and his apparently willing servile devotion which is, ironically, self-imposed and from which he rarely attempts to remove himself. Connecting his mistress with an illegitimate, artificial, and potentially harmful source of power suggests a resentment of his situation and his awareness of a need to justify it to himself and to his peers30. The introduction of magic into his relationship ultimately reveals more about the lover than about his beloved, highlighting his capacity for self-deception and for justifying his actions to himself and to his readers – both of which extend to constructing the image of his mistress best suited to his current needs – and illustrating his fallibility and untrustworthiness as a narrator31.

Tibullus 1.5 and 1.8 include the most explicit examples of this use of magic. In 1.5, the lover, separated from Delia after a quarrel, repeatedly attempts to make love to another woman. When he is unable to perform, the woman accuses Delia of using a binding spell to control his body and his desires (39-44):

saepe aliam tenui sed iam cum gaudia adirem
admonuit dominae deseruitque Venus.
tunc me discedens devotum femina dixit,
   heu pudet, et narrat scire nefanda meam.
non facit hoc verbis: facie tenerisque lacertis
    devovet et flavis nostra puella comis.

Often I embraced another: but as soon as I came to the point of joy
   Venus reminded me of my mistress and deserted me.
Then, departing, the woman called me bewitched –
   oh, the shame! – and she says that my girl knows unspeakable arts.
She does not do this with words; my girl bewitches with her appearance
    and soft arms and golden hair.

Tibullus initially appears to contrast Delia’s beauty with the effects of magic, saying that she has no need to use this because her beauty is enough to captivate her lover. The use of devoveo (“bewitch”) for the effects of both magic and beauty draws attention to the similarity of their effects on the lover, collapsing the difference between the two even as he attempts to stress it and implying that magic is the source of his obsession with Delia – and of his impotence – after all32. This repetition causes the explicit characterisation of magic and its effects as unspeakable (nefanda, 42) and shameful (pudet, 42) to be carried over to the effects of Delia’s beauty, too. By associating Delia’s beauty with a force which is described as shocking and harmful, the Tibullan lover presents his experience of loving Delia as something which is similarly disgraceful but which is beyond his control, removing his responsibility for both his infatuation and his inability to perform with another woman.

The narrator of Tibullus 1.8 uses a similar technique to illustrate the power of Pholoe’s beauty over her young lover, Marathus, who is also the narrator’s own beloved (15-26):

illa placet, quamvis inculto venerit ore
  nec nitidum tarda compserit arte caput.
num te carminibus num te pallentibus herbis
    devovit tacito tempore noctis anus?
cantus vicinis fruges traducit ab agris,
    cantus et iratae detinet anguis iter,
cantus et e curru Lunam deducere temptat,
    et faceret si non aera repulsa sonent.
quid queror heu misero carmen nocuisse, quid herbas?
    forma nihil magicis utitur auxiliis;
sed corpus tetigisse nocet sed longa dedisse
    oscula sed femori conservisse femur.

That girl pleases, even though she arrives with no make-up on her face
    and her glittering hair arranged by no measured art.
Has some old woman, in the quiet hour of the night,
    cursed you with songs or with pallor-inducing herbs?
Song draws the crops from a neighbouring field,
    song also halts the path of the angry snake,
song attempts, too, to draw down the Moon from her chariot
    and it would do so if the bronze did not ring when it was struck.
Why do I complain, alas, that spells have harmed the wretched boy, why herbs?
     Beauty uses no magic aids,
but the touch of a body harms, giving long kisses harms,
    thigh pressing close to thigh harms.

The narrator introduces the question of whether magic is the source of Marathus’ obsession with Pholoe in line 17, immediately after emphasising Pholoe’s natural beauty (15-16). After cataloguing the powers of magic spells (19-22), he rejects this as the reason for Marathus’ attraction and returns to the idea of beauty being more powerful than magic (24). As in 1.5, however, Tibullus uses the same verb to express the harmful effects of magic and physical contact with a beautiful girl, repeating noceo (“harm”) in lines 23 and 2533. Tibullus also uses the structure of 1.8.15-26 to equate Pholoe’s beauty with magic. In lines 19-22, the anaphora of cantus (“song”) and the ascending tricolon of cantus … cantus et … cantus et e emphasises the power of spells. This is echoed in lines 25-26, where the anaphora of sed (“but”) and the repeated perfect infinitives (tetigisse (“touch”), 25; dedisse (“give”), 25; conservisse (“press close”), 26) connect the catalogue of the effects of beauty with the powers of magic. The perfect infinitives in lines 25-26 also pick up nocuisse (“harm”), which is used of the harmful effects of magic in line 23. Through these features, Tibullus collapses the opposition between magic and beauty to suggest that Marathus’ infatuation is caused by magic after all. As in 1.5, the focus on harm presents the experience of elegiac love as something negative and debilitating, despite the overt praise of Pholoe’s beauty.

A similar connection between magic and beauty occurs in the opening couplet of Propertius 1.1:

Cynthia prima suis miserum me cepit ocellis,
     contactum nullis ante cupidinibus.

Cynthia was the first who captured me, wretched, with her eyes,
     though I had been seized before by no desires.

As Duncan Kennedy has highlighted, cepit (1) has a range of meanings: it can characterise the experience of love as capture, in both military and hunting settings, and it can equally suggest magical enchantment. Similarly, contactum (2) can mean both touched and infected by disease. These senses are all active in Propertius 1.1.1-2 and they are all expanded as the elegy progresses: military capture is developed immediately in the image of Amor pressing his foot on the lover’s head (3-4); hunting is central to the narrative of Milanion and Atalanta (9-16); the lover appeals to magic in the centre of the poem (19-24) and then to friends for medical assistance in the following lines (25-28)34. The subtext of magic in 1.1.1 is underlined by Cynthia’s name which, as the feminine of “Cynthius”, the Callimachean epithet for Apollo, associates her with chthonic Hecate and celestial Luna, the aspects of the triple goddess which complement Apollo’s sister, the earthly Diana35. The association of Cynthia and her appearance with magic at the very beginning of his elegiac corpus indicates that Propertius wishes his audience to recall this connection between beauty and erotic magic throughout the narrator’s relationship with his beloved.

Propertius connects Cynthia and her effect on her lovers with magic again in 1.5. The narrator warns his friend Gallus (31) against pursuing Cynthia: his beloved is angry, insensitive to prayers, and tortures her lovers (1-10); insomnia, disorientation, pallor, and emaciation result from devotion to the jealous and possessive girl who will slander unfaithful men (11-26). Propertius can offer his friend no cure, only the promise of comfort and a warning to stop enquiring about his mistress (27-32). The symptoms of elegiac love listed in 1.5 correspond with the effects wished on the targets of erotic attraction spells (ἀγωγαί)36. These symptoms also reflect those of lovesickness in antiquity. The narrator of 1.5, however, compares his infatuation with Cynthia to the effects of drinking Thessalian potions in the opening lines of the elegy: et bibere e tota toxica Thessalia (“and to drink all the poison brews of Thessaly”, 6)37. By prefacing the physical and mental torments of elegiac love with this comparison with magic, Propertius implies that magic is the cause of the lover’s physical and mental torments and characterises Cynthia, metaphorically, as a practitioner of erotic enchantment. Cynthia’s beauty is not mentioned in 1.5, but it plays a prominent role in the preceding elegies, including, and significantly for my purposes, 1.4, with which 1.5 has close thematic and structural connections. Both of these elegies are addressed to poets: the iambographer, Bassus (1.4), and a “Gallus” (1.5), who evokes Propertius’ elegiac predecessor Cornelius Gallus; the delayed introduction of Gallus’ name until the end of 1.5 elides the division between this poem and 1.4 and encourages the reader to understand 1.5 as still addressing Bassus, making him the victim of her erotic magic38. The initial implicit continuation of 1.4 carries over the previous emphasis on Cynthia’s attractiveness (1.4.5-14) as the source of the narrator’s love, associating the effects of this beauty with those of the Thessalian potions39.

By this point, I hope to have demonstrated that Tibullus and Propertius collapse the distinction they create between beauty and magic, justifying their amatory servitium and removing their responsibility for it by implying that it stems from a supernatural power beyond their control rather than from their own weakness for pretty girls. This metaphorical association of the mistress’ beauty with magic has, I argue, a corresponding metapoetic dimension which is a natural extension of her role as the Muse and embodiment of the elegiac text: the enchanting power of the girl’s beauty reflects the same feature of the poet’s carmina40. This level to the bewitching force of the puella’s beauty adds humour to the narrator’s self-serving association of his beloved with magic, as the magic which enables the puella’s hold over him is none other than his own poetry. Propertius 1.1 provides an example which suggests that this connection is programmatic in early elegy. As I noted above, Cynthia’s name associates her with Callimachean poetics, while deducere evokes the composition of fine, Callimachean verse: it is thus possible to read Propertius 1.1.19 as addressing witches whose carmina have the fallacia “of the drawn-out moon”, with Cynthia’s enchantment symbolising the composition of the elegiac text and its effect as much as the intended seductive effect of the elegiac lover’s poetry on his beloved41. If elegy is constructed as an erotic spell, then Cynthia, as a scripta puella (“written woman”), also embodies the magic inherent in the elegiac text. This connection at the centre of Propertius’ first elegy prompts us to re-read the opening effect of Cynthia’s beauty at 1.1.1-2 as expressing the elegiac poet’s enchantment by his own poetry and his own poetic creation.

III. Through the Looking-glass

Horace’s Canidia provides a contemporary precedent for this relationship between magic, poetry, and female appearance. I suggest that Canidia is constructed as a direct parody of the elegiac puellae: her advanced age, white hair, and grotesque appearance invert the metaphorically bewitching youth and beauty of the elegiac mistresses, leaving her with only practical magic to bind suitors to her. By presenting Canidia as engaging in erotic magic in Epodes 5 and 17, Horace literalises the elegiac metaphor of erotic enchantment and presents an iambic vision of what the elegiac lover’s situation would be if his puella were a real witch rather than possessed of magical beauty, highlighting the hypocrisy of the elegiac lover’s voluntary submission to a woman whom he presents as responsible for controlling his desires and his actions42. Focusing this parody through the figure of Canidia enables Horace to use elegy as a foil for defining his poetics in the Epodes while simultaneously highlighting the variety of genres he incorporates into his Roman iambic.

Epode 5 presents Canidia and her accomplices – Sagana, Veia, and Folia – attempting to produce a love-philtre capable of attracting back Canidia’s lover, Varus. They construct an ἐπίθυμα – a burnt-offering of trees, animal parts, and herbs from Iolchos and Hiberia (17-24) – and purify the area with water from Avernus (25-28), preparing for the key component of their magic: the burial and starvation of a Roman citizen boy (32-40). This process will imbue his organs with an intense desire to be transferred to Varus via Canidia’s philtre. The poem is framed by the boy’s words to his captors: his initial attempts to soften Canidia’s heart (1-10) and his final curses on the witches (87-102), which are inspired by Canidia’s central prayer to Nox and Diana (49-82).

In addition to the poetic aspect of Canidia, scholars agree that Epode 5 incorporates a comment on Horace’s iambic. The contrasting descriptions of the boy and his words at either end of the epode – his pitiful appearance (11-13) and his soft words (mollire, “to soften”, 14; mollibus / lenire verbis impias, “to mollify the impious women with soft words”, 83-84), and his vengeful curses (Thyesteas preces, “Thyestean prayers”, 86; humanam vicem, “human retribution”, 88)43 – have been interpreted metapoetically by Oliensis and by Johnson. The curses’ stated aim of revenge on the witches and the emphasis on their active effect (diris agam vos, “I will pursue you with curses”, 89) mark them as iambic44: Oliensis reads the child’s speeches as illustrating the “origins of invective in impotence”, with Epode 5 enacting the “symmetrical progression” between Canidia’s carmina and those of the iambic poet as “one vengeful speech (Canidia’s against Varus) begets another (the little boy’s against Canidia)”. Johnson charts a similar progression, though he associates the rage and violence of the boy’s reaction to Canidia with the “Archilochean-Lykambid invective” which Horace later denies for his Epodes (Epistles 1.19.19-25)45. Epode 5 ends, however, with the suggestion that the boy’s revenge will take effect only after his death at the hands of his captors; the implicit failure of his curses to change his fate suggests that Horace’s iambic may be as practically ineffective as any other poetry46.

The clear association of the child’s curses with iambic invites a similar literary reading of his opening appeals. I suggest that their repeated description in terms of “softening” the witch’s heart (mollire, 14; mollibus / lenire verbis impias, 83-84) evokes elegy. mollis (“soft, effeminate”) is already a key term in early elegiac self-definition, particularly in Propertius Book 1. In 1.7, for example, the Propertian narrator polemically contrasts his “soft verse” (mollem … versum, 19) with Ponticus’ epic47. By evoking elegy in the child, Horace incorporates an element of self-definition against this genre into the boy’s words: rather than maintaining the futile appeals which will lead to a death through tantalisation and desire, the boy fights to avenge himself with “active” curses. The implicit failure of these curses collapses the surface contrast with elegy, emphasising the questionable nature of poetry’s power to influence the world and, at the same time, illustrating that contemporary elegy is one of the many genres with which Horace enriches his Roman iambic.

It is possible to develop this elegiac element in the boy’s construction further and to identify parallels between the erotic ritual in Epode 5 and the elegiac relationship. Canidia tortures the boy, arousing and manipulating his desire as part of her ritual, while she pursues Varus with her magic. This dynamic evokes the elegiac ‘triangle’ of the lover, his mistress, and his rival, particularly the situation of Tibullus 1.8, where Pholoe – whose dazzling beauty needs no magic aids to bewitch suitors (1.8.24), but who will find it difficult to attract any lovers when she ages (39-46) – torments the puer delicatus (“effeminate boy”) Marathus while she courts an older lover (canus amator, 29)48. In Epode 5, Canidia’s seizure of the boy’s emblems of Roman citizenship, the toga praetexta and bulla (purpurae decus, 7; insignibus, 12) symbolises, I suggest, his removal from society and demotion to the status of a slave, dramatising the elegiac lover’s social detachment through his emasculating servitium to his puella. The boy’s youth and sexual purity also develops the documented use of young boys in magic – for example, the non-fatal use of sexually pure boys as divinatory mediums – by combining it with popular Roman superstitions regarding the victimisation of children by magic-workers. This is evident in Cicero’s accusations in his invective in Vatinium that Vatinius honoured the gods with the entrails of young boys, and in a gravestone for a girl which records her abduction by witches; equally relevant for Epode 5 is the belief that screech-owls fed on the blood of male children49. By drawing on these superstitions and, as I will argue, the realities of magic rituals, Horace exaggerates the elegiac lover’s youth and mollitia, and the puella’s cruelty and offers a grimly literalised and suitably iambic image of the lover’s experience in his description of the boy’s torture.

This torture is central to Canidia’s ritual in Epode 5, and the narrator describes the planned burial and its effects in detail (32-40):

   quo posset infossus puer
longo die bis terque mutatae dapis
    inemori spectaculo,
cum promineret ore, quantum extant aqua
    suspensa mento corpora,
exsecta uti medulla et aridum iecur
    amoris esset poculum,
interminato cum semel fixae cibo
   intabuissent pupulae.

    In which the boy, placed in the earth,
could die slowly amid the sight, over a long day,
    of the meals changed twice three times,
with his head standing out, his chin prominent
    as one whose body is suspended in water,
so that his bone marrow and dried out liver, having been cut out,
    could be made into a love potion,
when his eyeballs, fixed once and for all on the forbidden food,
     had withered away.

Horace’s use of contemporary magic for the imagery in Epode 5 has been explored by previous scholarship on the poem50. Eitrem, for example, highlights PGM XII.15-95 – a ritual for producing an “Eros assistant” to serve the practitioner – as a parallel for the length of the boy’s tantalisation (Epode 5.33); while this is not a specifically erotic ritual, it does aim to make targets love the practitioner and submit to his or her will51. It is possible to extend this parallel. The spell instructs the practitioner to consecrate the figurine by placing before it offerings which include fruit, birds, sweetmeats, and honey-wine, as well as inedible objects (21-24), “perfecting” it and preparing it for use as a ritual object52. These meals correspond to those repeatedly placed before the boy in Epode 5. Canidia’s treatment of her victim, which prepares his organs for use in her philtre, parallels that of the waxen figurine at PGM XII.15-95; strikingly, a live human replaces the inanimate statue here.

I suggest that Horace combines and exaggerates details of magic rituals in Epode 5 to construct the boy as a tendentious portrait of the elegiac lover. The description of the boy as someone whose body is suspended in water with his head above the surface (35-36) so that he can see the forbidden food as he dies evokes Tantalus in the Underworld, an allusion Horace explicates at Epode 17.65-66 when Canidia compares her victim with Tantalus (optat quietem Pelopis infidi pater / egens benignae Tantalus semper dapis, “Tantalus, the father of treacherous Pelops, prays for / rest, forever desiring the rich meals”, 65-66)53. Tantalus is used as a paradigm for the elegiac lover’s unattainable desires in both Tibullus Book 1 and Propertius Book 1, most explicitly at Tibullus 1.3.77-7854:

Tantalus est illic et circum stagna sed acrem
    iam iam poturi deserit unda sitim.

And Tantalus is there, and around him the pools, but
    now already the wave slides away from his bitter thirst.

Tibullus depicts Tantalus in the pool, unable to drink the water he craves. Propertius 1.9 alludes to the Titan’s punishment in the context of literary polemic with the epicist Ponticus: nunc tu / insanus medio flumine quaeris aquam (“now you, madman, are searching for water midstream”, 14-15)55. While Horace emphasises the starvation of Canidia’s victim, the desiccation (aridum, 37) of his organs parallels the thirst which the elegists foreground56. I suggest that we can read the boy’s torture – physically trapped in the ditch and reduced to an implement for Canidia’s spell – as encapsulating the elegiac lover’s amatory servitium and his expression of his love and fidelity to his puella in terms of his eroticised death and burial57. Horace literalises and caricatures the elegiac lover’s connection of his enslavement to a single beloved with magical enchantment and expresses it in terms which are more fitting to his lower iambic invective.

So far, I have highlighted parallels between Canidia’s treatment of her victim and rituals for consecrating Eros assistants, figurines which are created to perform a practitioner’s bidding. A separate use of wax, lead, or clay figurines occurs in erotic binding spells. The effigies in these rites, which are found bound, pierced, or arranged in pairs alongside written spells or curse-tablets, represent the target of the magic and the intended effects upon them – namely, the torturous symptoms of passionate love58. I suggest that Horace combines these two distinct uses of figurines in Epode 5 so that Canidia’s victim embodies the physical effects of elegiac love as though they were the result of magical enchantment, within the framework of a modified ritual for consecrating an Eros assistant.

The final lines of the boy’s torture describe his wasted eyes and his fixation upon the food placed in front of him as he dies (Epode 5.39-40). I suggest that these features evoke the physical emaciation and sleeplessness which are characteristic of the elegiac lover. At Propertius 1.13.15, the narrator states that love conquers and enfeebles its victims (vinctum languescere, “languishing conquered”); at 1.9.27, he stresses that love does not permit its victim to rest his eyes (quippe ubi non liceat vacuos seducere ocellos)59. Both features are listed in Propertius 1.5 among the torments resulting from the narrator’s unrequited love for Cynthia – he declares that love has caused his entire body to be reduced to nothing (aut cur sim toto corpore nullus ego, “or why my whole body is reduced to nothing”, 22) and claims that Cynthia controls his eyes and denies him sleep (non tibi iam somnos, non illa relinquet ocellos, “Now you will not sleep, she will not relinquish your eyes’, 11). As discussed above, Propertius 1.5 characterises the experience of elegiac love as equivalent to drinking magic potions (et bibere e tota toxica Thessalia, “and to drink all the poison brews of Thessaly”, 6) – a statement made early in the elegy, which colours the subsequent narrative with the suggestion of magical enchantment60. Horace expands this metaphorical comparison between witchcraft and overwhelming physical attraction into an overtly magical scene, exploiting the similarities between the elegiac symptoms and the effects wished on targets of erotic spells to represent them as caused by a practical magic ritual.

The denial of food, drink, and sleep, as well as physical wasting, were torments regularly wished upon victims of amatory spells. PGM IV.1496-1595 (εἰ πίνει, μὴ πινέτω, / εἰ ἐσθίει, μὴ ἐσθιέτω, “if she is drinking, let her not keep drinking, if she is eating, let her not keep eating”, 1515-1516; and εἰ κοιμᾶται, μὴ κοιμάσθω, “if she is sleeping, let her not keep sleeping”, 1521), and PGM XXXVI.134-160 (πεινῶσαν, διψῶσαν, ὕπνου μὴ τυγχάνουσαν, “let her be hungry, thirsty, let her not find sleep”, 149-153) are two examples of curses which deny victims both nourishment and rest. One second century A.D. erotic curse (Preisendanz Ostrakon 2 = Gager 35) specifically demands that the victim be tormented by “starvation” (ἀσιτῳ, 35); the spirit invoked at PGM XVI.1-75 is repeatedly commanded to make the target “pine and melt away” with passion (ποίησον / φθείνειν καὶ κατατήκεσθαι  ἐπὶ / τῷ ἔρωτι, 11-12). In Epode 5, the torments of starvation and thirst are graphically expressed by the wasting of the boy’s eyeballs (intabuissent pupulae, “eyeballs withered away”, 40), and the dehydration of his organs (medulla et aridum iecur, “bone marrow and dried out liver” 37); the unrelenting fixation of his eyeballs upon the food (fixae cibo /  pupulae, “eyeballs … fixed on the food”, 39-40) evokes the insomnia and the single-minded focus on the practitioner of the spell commonly wished upon victims of erotic magic. By highlighting the physical deterioration of the child, Horace lingers upon the grim, visceral reality of magic, creating an iambic adaptation of the elegiac metaphor of magical enchantment.

Erotic deprivation curses aim to cause the target’s “isolation from the land of the living”61. This is particularly notable for my purposes, as asserting his detachment from the normal course of society is central to the rhetoric of the elegiac lover and he illustrates this through his willing enslavement to his beloved62. In Epode 5, as I suggested above, the child’s entrapment in the ditch and loss of his emblems of citizenship physically illustrate the lover’s isolation and his domination by his beloved. Horace exploits these similarities between the effects of deprivation curses and the elegiac self-presentation to expose the lover’s self-serving rhetoric and parody his servitium in overtly magical terms.

In addition to literalising the metaphorical use of magic in elegy, I suggest that the child’s burial and torture evokes the eroticisation of the narrator’s death and burial in the first books of Tibullus and Propertius. The elegiac lovers repeatedly imagine their death and burial, anxious that they should occur during their love for their mistresses, who will faithfully perform their funeral rites. The lovers use death as a metaphor for their amatory experiences and to illustrate the strength of their passion and fidelity to their beloveds. Canidia’s live inhumation of her victim reworks these elegiac concerns in a representation of literal magic practice, dramatising the elegiac narrator’s romanticised equation of love with death and his vision of dying in and through love. I highlight some examples of how love and death are treated in Tibullus’ and Propertius’ early work before looking at Horace’s adaptation of the motif more closely.

Tibullus 1.1 offers an extended vision of the narrator’s death (57-68) in which he hopes to see Delia as he dies in her arms (te spectem suprema mihi cum venerit hora; / te teneam moriens deficiente manu, “Let me gaze on you when my final hour comes; / as I am dying, let me hold you with my failing hand”, 59-60) and imagines her tears and kisses as she places him on his funeral bier (flebis et arsuro positum me, Delia, lecto, / tristibus et lacrimis oscula mixta dabis, “You will weep for me, Delia, when I have been placed on the funeral bier, / and you will give me kisses mixed with sad tears”, 61-62); echoes of his earlier wish to hold Delia in his arms in bed by the fire (45-48) underline the narrator’s eroticisation of his funeral63. The Tibullan lover’s envisioned death provides the framework for 1.3: the elegy begins with the narrator lamenting the absence of his mother, his sister, and Delia at his funeral (5-10); the epitaph he recites (55-56) introduces his catabasis into an elegiacised Underworld (57-82). This eroticisation of death features more frequently in Propertius Book 1: the Propertian lover treats this theme in 1.6, 1.14, 1.17, and 1.19. The lover’s earliest expression of his devotion to Cynthia as death occurs in 1.4: after stressing Cynthia’s beauty, the narrator asserts that she has many more attributes for which he would willingly perish: haec sed forma mei pars est extrema furoris; / sunt maiora quibus, Basse, perire iuvat (“But this beauty is the smallest part of my frenzy: / she has more qualities, Bassus, for which I would gladly perish”, 1.4.11-12)64. This may provide a jumping off point for Horace’s literalisation of the elegiac lover’s equation of love and death in the eroticised burial of Canidia’s victim.

Propertius also expresses the elegiac lover’s wish for his love to last until his death. In 1.6, the narrator envisions his death and burial in the earth during his continuing devotion to love, equating his demise with his desire through periere: multi longinquo periere in amore libenter, / in quorum numero me quoque terra tegat (“many have gladly died in a long-lasting love, may the earth cover me, too, as one among these”, 27-28). Propertius creates the same impression in 1.17, where the lover twice specifies his burial in the ground: firstly, separated from Cynthia following a shipwreck (haecine parva meum funus harena teget?, “will this small tract of sand cover my corpse?”, 8); secondly, concluding his vision of Cynthia attending his funeral in Rome at 1.17.19-24 (ut mihi non ullo pondere terra foret, “that the earth may not weigh on me at all”, 24). At the beginning of this passage, he imagines his burial taking place during his love for Cynthia (illic si qua meum sepelissent fata dolorem, “there if any fates might have buried my love”, 19)65. The narrator expresses the wish to remain in love at his death, without reference to burial, at 1.14.14 (quae maneant, dum me fata perire volent, “and may they remain until the fates wish that I should perish”). The burial of Canidia’s victim in Epode 5 – whose slow demise centres around the desire he experiences for the food before him (longo die … mutatae dapis / inemori spectaculo, 33-4) – dramatises elegy’s eroticisation of death. It also evokes the lover’s liminal status in this theme, simultaneously dead and holding the position of poet and narrator66. Canidia’s attendance at the boy’s burial, meanwhile, distorts the elegiac lover’s desire for his mistress’ presence at his funeral as an indication of her fidelity and expands his fears on this score as Canidia plunders the child’s remains in pursuit of another man.

The boy’s murder will cause his desire to become distilled into and endure in his remains, echoing the elegiac narrator’s claims of his eternal love and fidelity to his puella. Propertius 1.19 is the most developed treatment of this idea in early elegy. The narrator declares that his “great love will cross over the shores of fate” into the Underworld (traicit et fati litora magnus amor, 12), and claims that his passion is so strong that it will endure even in his ashes (non adeo leviter nostris puer haesit ocellis, / ut meus oblito pulvis amore vacet, “not so lightly has the boy fastened on my eyes / that my ashes should be free from forgotten love”, 5-6). Lines 5-6 evoke the physicality of the lover’s desire which stems from his mistress’ beauty, underlining this through the emphasis on his eyes67. Horace adapts the elegiac lover’s emphasis on desire lasting in physical remains to the context of a magic ritual. The boy’s starvation deflates the lover’s eternal fidelity by substituting forbidden and desired food for the mistress, distorting the elegiac expression of fidelity into a magical practice and mocking the lover’s desire by indicating that it can be felt as strongly for frequently changed plates of food as for the woman the lover professes to hold above all others.

Propertius 1.19 provides a further parallel for the Horatian boy’s torture and imagined return as a vengeful ghost through the narrartive of Protesilaus and Laodamia (9-10):

sed cupidus falsis attingere gaudia palmis
    Thessalus antiquam venerat umbra domum.
But desiring to touch his delight with false hands,
    the Thessalian came as a shade to his former home.

Propertius stresses Protesilaus’ carnal desire for Laodamia (9), whom his ethereal form renders him unable to touch. The returning ghost acts as a model for the narrator, illustrating Propertius’ passion for, and separation from, Cynthia, as well as his eternal love for her68. In Epode 5, Horace presents the child in the ditch as one similarly stranded between life and death and unable to reach the object of his desire (39). The continuation of the child’s spirit after death as an ἄωρος or a βιαιοθάνατος – those who have died prematurely or violently – for Canidia to control also echoes the returned Protesilaus; unlike the Propertian ghost, who remains desirous of the woman he cannot touch, the child’s closing curses threaten that his spirit will return and assail the witches in their sleep: petamque voltus umbra curvis unguibus/ quae vis deorum est Manium, / et inquietis assidens praecordiis / pavore somnos auferam (“as a ghost I will attack your faces with my curved claws, / – such is the power of the gods below – / and perching on your anxious hearts / I will snatch away your sleep with fear”, 93-96). The boy’s resistance to this fate in his closing curses highlight the difference between slavish elegiac devotion and iambic aggression.

To close this section, I would like to return briefly to Propertius 1.4 and 1.5. I have suggested that the child’s torture in Epode 5 literalises the symptoms of elegiac love which the narrator of Propertius 1.5 expresses in magical terms, picking up on the lover’s emaciation, sleeplessness, and fixation on his beloved. I have also highlighted the close link between Propertius 1.5 and 1.4 which creates the initial impression that “Bassus” is the imagined victim of Cynthia’s erotic magic in 1.5. In 1.4, the narrator threatens Bassus with unbridled defamation by Cynthia, who will adopt the role of the iambist, and with his consequent exclusion from the girls of Rome (17-22). At lines 17-18, Propertius’ narrator highlights the expansive range of Cynthia’s potential abuse: haec insana puella / … tibi non tacitis vocibus hostis erit (“this crazy girl / … will be an enemy to you with no quiet words”). Instead of offering Cynthia’s speech, however, the narrator outlines his mistress’ extensive efforts to retain his love before praying that she will never change and reaffirming his devotion (27-28). I suggest that we can read Canidia’s speech in the Epodes as filling the gap in Propertius 1.4 by giving the Propertian lover – and the poet himself – a taste of a real iambic puella at home in her own genre. The Horatian narrator’s introduction of Canidia’s speech in Epode 5 – quid dixit aut quid tacuit? … (“what did she say, or what did she keep silent? …”, 49) – picks up the verbal cue at Propertius 1.4.17-18, introducing the witch’s monologue as simultaneously a response to his elegiac contemporary and a characterisation of his iambic. Canidia’s demand that Night and Diana direct their divine wrath towards keeping Varus, her enemy (hostilis, 53), from the doors and the beds of other women shows her attempting to carry out the slanderous actions of Cynthia only imagined in 1.4; at the same time, Canidia’s incapability of barring doors to Varus responds to Propertius’ characterisation of his poetic magic as perennially ineffective in love, showing this failure in action. Canidia’s uncensored words also reflect both the tendency of iambic to concentrate on low subject matter that would otherwise be unspeakable and the particular concerns with impotence in the Epodes, which are illustrated in the narrative of Epode 5 by the inability of her magic to control Varus and in the final curses of the boy69. While Epode 5 begins to fill the gap of the elegiac puella’s unwritten invective, it is Epode 17, in which Canidia takes over as speaker in the second half of the poem and ultimately closes the book of Epodes, which develops this engagement with contemporary love elegy.

Epode 17 is a dialogue between a male poet and Canidia: in the first half, the poet prays for release from Canidia’s magic and offers to recant his earlier insults of her character (1-52); in the second, Canidia rejects his appeals and promises that she will torment him for eternity with her magic in revenge for his slander (53-81). Previous scholarship has proposed a range of metapoetic readings of Epode 17. Barchiesi argues that Epode 17 treats the “principles of iambic poetry and its effects”, among which magic is prominent: the poet and Canidia employ iambic spells against one another and both embody the genre – in the narrator’s case, thanks to Canidia’s magic. Through the symmetry of the poet and the witch, Horace demonstrates the tendency of iambic verse to be as harmful to the practitioner as to its victim. Barchiesi also suggests, in keeping with the theme of reversability which he traces through the poem, that the introduction of the Cotyia (Epode 17.56) – a festival which involved male transvestism – raises the possibility that Canidia is Horace in drag, a figure through whom the iambic poet speaks but who also has the power to problematise his words70. Spina’s metaliterary interpretation draws out a similar chiastic relationship between Horace and Canidia and suggests that the content and structure of Epode 17 emphasises the variety of genres, styles, and traditions – Archilochean, Alexandrian, and Neoteric – which Horace experiments with in the Epodes71. The power and effectiveness of poetry is central to Epode 17, which presents the poet’s carmina as magic incantations – a performative category of verse which, like iambic, aims to have tangible influence on the world – before raising doubts about their capabilities in the closing lines72.

Johnson’s metapoetic reading acknowledges the significance of elegy in the second half of the Epodes. Johnson argues that Horace’s speech in the first half of Epode 17 tries to engage Canidia through epic, lyric, and elegy, reflecting the variety of genres in Epodes 11-16 and concluding the book by anticipating Horace’s lyric. Johnson reads the narrator’s mental and physical torments at Epode 17.21-26 as characterising him as an elegiac lover addressing Canidia as his beloved; Johnson, however, focuses only on the first half of Epode 17 and does not comment on elegiac elements in Canidia’s response73. Bushala also identifies elegiac resonances in the poem: after arguing that the male narrator is Canidia’s lover and the victim of her erotic magic, Bushala closes his article by quoting E. K. Rand’s suggestion that Epode 11 “laughs prophetically forward at” Propertius and Tibullus and proposing that Epode 17 ridicules the “enclosed, absurd, and morbid world” of Roman elegy and the lover’s anguished relationships with an “uncanny, voracious Charybdis-Cimaera”74. I develop these readings of elegy to argue that Horace literalises the metaphorical use of magic in elegy to parody the dynamics of the elegiac relationship across Epode 17 and, by making this a framework for the final poem of the book, to illustrate the integration of love elegy into his iambic.

In Epode 17, I suggest, Horace draws on the connection between age and poetic immortality in early elegy which is expressed through the elegiac lover’s idealised future with his puella. The lover wishes for a faithful relationship with his puella which will last until their old age, even though his mistress rarely reciprocates his affection and is rarely loyal to him. This concern with love lasting into old age features in both Tibullus Book 1 and Propertius Book 1. In Tibullus 1.6, the narrator concludes his lament over Delia’s deceitfulness and infidelity (1-36), his promises to submit to punishments from his mistress if he mistreats her (43-74), and his predictions of an old age of loneliness, poverty, and mockery for unfaithful women (75-84) by wishing that “these may be curses for other couples” and that he and Delia “may be a model of love in our old age”: haec aliis maledicta cadant. nos, Delia, amoris / exemplum cana simus uterque coma (85-86). Propertius expresses a similar sentiment in the closing lines of 1.8B. The narrator celebrates Cynthia’s positive response to his elegiac prayers and seductions and the fidelity to him that she has proven by refusing to accompany a rival to Illyria (1.8B.1-45) as she had threatened to do in 1.8A. He concludes by declaring that nobody will steal his love, “a boast which will last into my old age”: ista meam norit gloria canitiem (46). On a metapoetic level, the theme of age and lasting love comments on poetic immortality. Propertius 1.8A and B dramatise the power of Propertius’ elegiac seductions by showing that they can successfully persuade Cynthia to stay with him in Rome, and the imagery of old age and the poet’s own catasterism (nunc mihi summa licet contingere sidera plantis, “now I may touch the highest stars with the soles of my feet”, 1.8B.43) predict the lasting fame of Propertius’ elegy75. Tibullus 1.6 concludes the Delia elegies of Book 1: ending this sequence on the theme of love lasting into old age again gives the lover’s wish an undertone of literary achievement and longevity. It is, of course, ironic that although the lovers imagine sharing their old age with their mistresses, the fame that Propertius, Tibullus, and their work will enjoy through the ages will depend on, and preserve, the immortal youth and beauty of the fictional lover and his beloved.

Horace, I argue, exploits this irony to parody the dynamics of the elegiac relationship in Epode 17 in terms which reflect the characteristics of his iambic and the enduring power and fame of his Epodes. Through Canidia, Horace grants the elegiac lovers’ wish for a faithful mistress to share their old age, but he gives them one whose old age and ugliness compel her to use magic to bind them to her: Epode 17 inverts elegiac dynamics by reversing the exclusus amator’s pleas for admittance and presenting the narrator as begging for release from Canidia’s spells which prolong his amatory torture. Through this, Horace again literalises the elegiac metaphor of magically enchanting beauty, playing on erotic curses which wish that their targets should love the practitioners for the rest of their lives and using this to illustrate the potency and longevity of the Epodes and the importance of elegy to his iambic76.

The opening lines of Epode 17 introduce this connection between elegy and magic (1-7):

iam iam efficaci do manus scientiae
supplex et oro regna per Proserpinae,
per et Dianae non movenda numina,
per atque libros carminum valentium
refixa caelo devocare sidera,
Canidia, parce vocibus tandem sacris
citumque retro solve, solve turbinem.

As I suppliant, I surrender now to your powerful
knowledge, and by the kingdom of Proserpina,
and by the divine authority of Diana, which is not to be provoked rashly,
and by the books of spells which have the strength
to call down the loosened stars from the sky, I beg you,
Canidia, stop your sacred words at last
and reverse, reverse your swift top.

Supplex (“suppliant”, 2) characterises the iambic narrator in elegiac terms, evoking the exclusus amator prostrate at his mistress’ threshold77. The reason for the narrator’s supplication, however, is not to seduce his puella but to beg Canidia to release him from her erotic magic: her carmina and her turbo (“top”)78. Tibullus uses the image of a spinning top to metaphorically illustrate the servitium and madness of elegiac love in the opening lines of elegy 1.5, where the narrator recants his previous harsh words towards Delia and his desire to end their relationship (1-8):

asper eram et bene discidium me ferre loquebar,
   at mihi nunc longe gloria fortis abest;
namque agor ut per plana citus sola verbere turben
    quem celer assueta versat ab arte puer.
ure ferum et torque, libeat ne dicere quicquam
   magnificum posthac: horrida verba doma.
parce tamen per te furtivi foedera lecti
    per Venerem quaeso compositumque caput.

I was harsh and I said that I was bearing our separation well,
   but now my courageous pride is long gone;
for I am driven like a swift top whipped across level ground,
   which a nimble boy spins with his accomplished skill.
Burn and torture the wild beast, so that he will not want to say anything
    proud in future: tame his savage words
but spare me, by the pacts of our stolen bed,
    by Venus, and, I beg, by the head that lay together with mine.

I suggest that Horace evokes the elegiac imagery of madness and servile punishment in the context of magic in Epode 17, literalising the elegiac lover’s metaphorical illustration of his mental and physical suffering and parodying this by having his iambic narrator beg for release from Canidia’s spells rather than demanding further punishment79. Horace also suggests the limitations of Canidia’s power here: Epode 17.1 echoes Canidia’s declaration that a stronger witch must be controlling Varus at Epode 5.71-72 (a, a, solutus ambulat veneficae / scientoris carmine!, “Ah, ah, he walks free by the spell of a more knowledgeable witch!”). By recalling the weakness of Canidia’s magic, Horace undermines the seriousness of the narrator’s surrender in Epode 17 and brings out the humour in the elegiac lover’s submission to a metaphorical magic which he himself attributes to his mistress80.

The physical and emotional effects of Canidia’s magic on the narrator of Epode 17 continue to literalise the elegiac metaphor of magic (21-27):

fugit iuventas et verecundus color,
relinquor ossa pelle amicta lurida,
tuis capillus albus est odoribus;
nullum a labore me reclinat otium,
urget diem nox et dies noctem neque est
levare tenta spiritu praecordia.

My youth and my boyish complexion have fled,
I am left as bones clothed in sallow skin,
my hair is white thanks to your magic odours;
no respite releases me from my pain,
night presses on day and day on night and it is not possible
relieve my tight chest with breath.

The narrator’s pallor (21-22), emaciation (22), insomnia (24-25), wretchedness (27), and light breathing (26) evoke the elegiac lover; in the following line (27), the narrator describes himself as miser (“wretched”), a term which is virtually programmatic of the lover in early elegy81. In addition to the effects of the curses I discussed above in connection with Epode 5, erotic curses frequently demand that their victims become pale and suffer torments day and night82. While extant curses focus on their targets’ breathing only occasionally83, light breathing is a characteristic of the elegiac exclusus amator which Horace also parodies in Epode 11: latere petitus imo spiritus (“the breath sought by my deepest lung”, 10)84. spiritus is used for breath only in Epodes 11 and 17 and this echo reinforces the element of elegiac parody – this time in connection with the metaphorical use of magic in the genre – in the narrator’s symptoms in Epode 17. As I highlighted above, the narrator’s appearance in these lines also embodies the characteristics of iambic poetry. The victim of Canidia’s magic in Epode 17, therefore, blends the elegiac and iambic elements which were kept separate in the words of the boy which framed Epode 5 and embodies the enrichment of Horace’s iambic with contemporary elegy.

Epode 17.21-27 contains further echoes of Epode 11 which reinforce this reading of Canidia’s victim. The narrator of Epode 17 laments the loss of his verecundus color (“boyish complexion”, 21) which recalls inverecundus deus (“shameless god”) at Epode 11.13. Although inverecundus describes Bacchus, it also characterises the effect of the undiluted wine on the narrator (fervidiore mero arcana promorat loco, “coaxed out my secrets with fiery wine”, 11.14). The repetition of verecundus in Epode 17.21, the sole parallel in the Epodes, prompts the reader to recall Horace’s earlier recasting of elegy in iambic terms in Epode 1185. I suggest that the significance of these adjectives for engaging with contemporary elegy can be pressed further. verecundus and inverecundus connote moral as well as physical characteristics: the narrator’s loss of his youth and verecundus color since coming into contact with Canidia evokes the shamelessness (nequitia) which characterises the elegiac lover and his experiences. The Gallus fragment from Qasr Ibrîm suggests that nequitia was central to love elegy from the genre’s earliest appearance86. nequitia and shamelessness feature prominently in the first elegies of Propertius and Tibullus. The narrator of Propertius 1.1, for example, claims that love for Cynthia has caused him to lose his moral purity and propriety (tum mihi constantis deiecit lumina fastus, 3) and to live aimlessly and senselessly (et nullo vivere consilio, 6)87. In Tibullus 1.1, the lover’s nequitia is indicated by his desire to remain “lazy and idle” in love with Delia (segnis inersque, 58)88. In Epode 17, the iambic narrator’s loss of modesty literalises the elegiac metaphor of magic enchantment by echoing the demand of erotic spells that their targets come “without shame” to the practitioners (PGM XVIIa.21), shifting the emphasis from the effect of the woman’s beauty to her erotic magic and amplifying the elegiac narrator’s characterisation of his experience as negative. In addition to reinforcing the elegiac element in Epode 17, the echoes of Epode 11 in the final poem of the Epodes draw together Horace’s engagement with elegy across the book to cement this genre’s role as a constituent element of his iambic.

Canidia’s reply continues to literalise the metaphorical application of magic to the elegiac mistress. It also, I argue, finally fills the gap left by Cynthia’s suppressed iambic words in Propertius 1.4 by presenting, at the climax of the Epodes, the fully-fledged iambic curses of Horace’s anti-elegiac puella89. This elegiac element to Canidia is established at the beginning of her monologue, where she proclaims that she is deaf to her victim’s pleas: quid obseratis auribus fundis preces? (“Why do you pour forth prayers to bolted ears?”, 53). The description of her ears as “bolted” evokes the elegiac paraclausithyron, creating a witty reversal of the elegiac beloved by refusing to listen to pleas for escape rather than for entry. Canidia goes on to illustrate her unresponsiveness by comparing herself to a rock which will never hear the prayers of sailors battered by Neptune (54-55), drawing out the hard characters of the elegiac puellae implicit in their etymological links with the mountains Cynthus and Pholoe90.

After assuring the narrator that magic will not help him (60-61), Canidia promises that his punishments will last for eternity (62-69):

sed tardiora fata te votis manent:
ingrata misero vita ducenda est in hoc,
novis ut usque suppetas laboribus.
optat quietem Pelopis infidi pater
egens benignae Tantalus semper dapis,
optat Prometheus obligatus aliti,
optat supremo collocare Sisyphus
in monte saxum: sed vetant leges Iovis.

However, a slower death waits for you than you would wish:
your wretched life is to be drawn out by this misery,
so that you will be constantly available for new sufferings.
Tantalus, the father of treacherous Pelops, prays for
rest, forever desiring the rich meals,
Prometheus, bound for the bird, prays,
Sisyphus prays to erect his rock
on the mountain top: but Jove’s laws forbid it.

Barchiesi has highlighted that the sinners’ punishments which Canidia lists here regularly symbolise insatiable desire91. Tantalus’ prominent position at the beginning of this catalogue (65-66) recalls the burial and starvation of the boy at Epode 5.35-36, particularly as, here, Horace chooses the alternative tradition of Tantalus’ torment and represents him as unable to reach “food” (dapis (66), echoing dapis at Epode 5.33) rather than water. This is underlined by the reference to Prometheus (67), where the mention of the bird alludes to its consumption of his liver: Canidia’s future punishment of the narrator of Epode 17 recalls her harvesting of the boy’s liver for her erotic ritual and her revenge against Varus in Epode 5. This echo of the boy’s tantalisation in Epode 5 reinforces Horace’s literalisation of the elegiac lover’s experience through magic in Epode 17 and continues to draw together the threads of Horace’s engagement with elegy across the Epodes.

Epode 17.65-69 also plays on the elegiac motif of love extending beyond death. I suggest that Canidia’s list of eternal torments at lines 62-69 evokes the repeated wishes of erotic spells that the practitioners may control their targets for eternity (ποίησον τὴν δεῖνα / ἀγρυπνοῦσάν μοι διὰ παντὸς [αἰῶνοσ], “make her lie awake for me through all eternity”, PGM IV.2965-2966) or for a lifetime (ὑποτεταγμένην εἰς τὸν ἅπαντα χρόνον τῆς ζωῆς μου, “submitting herself [to me] for the entire time of my life”, Suppl. Mag. 47.26). PGM XVI.1-75 includes a formulation of this wish which is particularly pertinent for Epode 17: the practitioner demands that their victim conform to their will and love them “until he reaches Hades” (ἕως ὅτ[αν εἰ]ς [Ἅδ]ην ἀφίκηται, 16)92. This evocation of contemporary magic ritual continues to literalise the elegiac lovers’ passion and their eroticisation of death in ways that are similar to those which I have identified in Epode 5, reinforcing the elegiac parody in both poems.

The eternal punishments in these lines pick up the thread of poetic immortality in Epode 17, developing the paradox of this theme in love-elegy in terms suited to Horace’s work. As I highlighted above, the narrators of Tibullus 1.6 and Propertius 1.8B predict that their happy, faithful relationships with their mistresses will last into their old age; while the fame of each book will grow as the years pass, however, the lovers and their puellae will remain young and beautiful and their idealised future relationship will remain forever out of the lover’s reach. Canidia, by contrast, is already old and unattractive; she has accelerated her victim’s age to match her own (Epode 17.21-27) and her spells will keep the aged poet alive forever in eternal torture. Canidia’s focus on punishment expands the element of revenge which remains suppressed in the elegiac puellae of Propertius Book 1 and Tibullus Book 1 but which is a central component of the iambic genre. This reversal also allows Horace to use the same imagery of old age and immortality to illustrate the longevity of his iambic collection and its fame.

Poetic power is the focus of the final lines of Epode 17 (76-81):

an quae movere cereas imagines,
ut ipse nosti curiosus, et polo
deripere lunam vocibus possim meis,
possim crematos excitare mortuos
desiderique temperare pocula,
plorem artis in te nil agentis exitus?

Or will I, a woman who manipulates wax figurines,
as you discovered by prying, and can
tear the moon down from the heavens with my words,
can rouse the cremated dead
and can concoct love-potions,
lament because the results of my skills are useless against you?

Canidia concludes the poem, and the Epodes, by cataloguing the powers of her magic. Her list reflects the rituals that she has performed in Epode 5, in Epode 17, and, beyond the Epodes, in Horace’s Satires 1.893. She also claims for herself the power to draw down the moon (78), which, in Epode 5, was attributed to her accomplice, Folia (Epode 5.45-46). This expresses the power of Horace’s poetry in terms which are suited to a genre that is rooted in magic, revenge, and attack, as underlined by Candia’s control of the moon94. In elegy, the ability to draw down the moon (deducere lunam) represents the composition of the elegiac text. Unlike deduco, deripio (“tear”, Epode 5.46 and 17.78) has no literary connotations. It does, however, reflect the speed and violence associated with the iambic metre and subject matter95. The balanced and chiastic structure of Epode 17, alongside the theme of magic which runs through it, tightens the association of poet and witch; by the final lines, they speak as one voice, an impression underlined by the iambic poet’s lack of response to Canidia’s speech96. At the same time, Canidia’s last question and the repeated subjunctive possim (“can”, 78-79) destabilise the power of her erotic magic, recalling the limitations which she acknowledges at Epode 5.71-72 and which are implied by the echo of these lines at the beginning of Epode 17.

I suggest that the evocation of the elegiac connection between magic and poetry in Canidia’s final lines has two functions. First, it literalises the metaphorical connection between magic, the puella, and the poetic text in early elegy to define Horace’s iambic poetics in contrast to this genre through the figure of Canidia. Second, it concludes the Epodes with the blending of elegy, magic, and iambic that I have traced through Epodes 5 and 17. This is underlined by the emphatically placed plorem (“lament”, 81) which brings to mind both elegy’s connection with unsuccessful lament and the iambic narrator’s question at the end of Epode 6: an si quis atro dente me petiverit, / inultus ut flebo puer?, (“and if anyone attacks me with a black tooth, I will weep like a boy unavenged?”, 15-16). Horace evokes both the enchanting power and the failure of erotic magic in elegy to underline the questions of poetic power, impotentia, and masculinity that run through the Epodes and which remain unresolved, and literally unanswered, at the climax of the book, illustrating that elegy’s treatment of magic and gender is fundamental to defining and constructing his unique brand of Roman iambic.

IV. Conclusion

I conclude by drawing together the implications of my discussion for reading early love elegy and for Horace’s Epodes. In the narrative of the elegiac affair, the lover uses the metaphor of magic to simultaneously compliment his puella’s beauty and imply that she has used magic to cause his obsession with her. This self-serving insinuation highlights the ambivalence in his feelings towards her and towards his voluntary submission to elegiac servitium, characterising him as a hypocritical lover and as a fallible and untrustworthy narrator. The expression of the puella’s attractiveness in terms of magical enchantment in Propertian and Tibullan elegy also functions metapoetically: the attribution of magic power to the girls who embody the elegiac text complements elegy’s self-construction as an erotic spell which is designed to seduce the puella. Connecting magic with the puella’s beauty gives a physical dimension to the enchanting effects of poetry and its influence over the lover and poet as well as the extratextual audience. This three-way relationship between magic, elegy, and the puella in the earliest extant collections of Latin love elegy provides new evidence for the importance of magic as a central metaphor of the genre.

Recognising the connection of the puella’s beauty with magic in early elegy also opens a new avenue for reading the figure of the scripta puella and the motif of magic as means of communication between poetic genres – dramatising literary polemic through female personifications of the poetic text who are associated with witchcraft97. My interpretation of Horace’s Epodes 5 and 17 as responding to Propertius’ and Tibullus’ first books illustrates this use of the motif and its importance from the very beginning of the elegiac genre as we have it, and testifies to a contemporary awareness of the metaliterary dimension to the puella’s magical attractiveness. As an inversion of the elegiac puellae, Canidia replaces their metaphorically bewitching beauty with literal erotic rituals, embodying the negative view of the elegiac mistresses which lies beneath their lovers’ flattery. Horace’s exaggeration of this subtext indicates that it was an aspect of early Tibullan and Propertian love elegy which was not only recognisable to a contemporary audience, but which was central enough to the presentation of the puellae to merit him targeting it as part of his engagement with this genre. Readers returning to the elegies in light of Horace’s work would no doubt bring an extra awareness of this subtext and would perceive more keenly the irony and humour in the elegiac lover’s character and relationship with his puella.

So far, I have focused on the implications of this dialogue for reading magic and the mistress in Propertius and Tibullus. The question now is: what impact does this engagement with elegy have on Canidia and on Horace’s Epodes more broadly? I suggest an answer to this by posing a second question: why is Canidia a woman, rather than a male personification of genre or victim of iambic aggression in the vein of Lycambes and Bupalus, the targets of Horace’s Greek predecessors Archilochus and Hipponax, or even Ovid’s later Ibis? Several answers to this present themselves: misogyny is a prominent element of iambic; laughter provoked by the lewd behaviour of elderly women is associated with the roots of the genre; the contemporary political climate associated Rome’s chaotic state with licentious, “masculine” women98. Based on the interaction I have traced between the Epodes and Tibullan and Propertian elegy, it is possible to add a more specific, literary reason: that Canidia, as much as being a positive embodiment of Horatian iambic, is composed as a pointedly anti-elegiac puella. Horace’s Canidia and her companions vividly expose the grim “reality” behind the elegiac appeal to witchcraft and de-romanticise the metaphorical application of magic power to the puella, using this parody to define and enrich the iambic poetics of the Epodes. Instead of detracting from Canidia’s independence and originality as a literary entity, reading her as constructed symbiotically with the female beloveds of Propertius and Tibullus adds a new dimension to Horace’s witch and, by extension, to his Epodes. Canidia’s inversion of the elegiac puella necessarily incorporates the generic characteristics she reverses and rejects, uniting both traditions to embody the literary variety of Horace’s iambic.


Ahl F. M., (1974), “Propertius 1, 1”, Wiener Studien 8, 80-98.

Allen A. W., (1950), “Elegy and the Classical Attitude Toward Love: Propertius I, 1”, Yale Classical Studies 11, 255-277.

Anderson W. S., (1972), “The Form, Purpose, and Position of Horace’s Satire I, 8”, The American Journal of Philology 93.1, 4-13.

Anderson R. D., Parsons P. J. and Nisbet R. G. M., (1979), “Elegiacs by Gallus from Qasr Ibrîm”, Journal of Roman Studies 69, 125-155.

Andrisano A. M., (2012), “Multiple Models and Horace: Epodes 1, 3 and 5”, in Cairns F. (ed.), Papers of the Langford Latin Seminar Fifteenth Volume, ARCA 51, Cambridge, 285-302.

Baker R. J., (1970), “Laus in amore mori: Love and Death in Propertius”, Latomus 29.3, 670-698.

— (ed.) (tr.), (1990), Propertius I, Armindale.

Ball R. J., (1994), “Albi, Ne Doleas: Horace and Tibullus”, Classical World 87.5, 409-414.

Barchiesi A., (1994a), “Alcune difficoltà nella carriera di un poeta giambico. Giambo ed elegia nell’ epodo XI”, in Tovar R.C. and Corte J. C. F. (eds.), Bimilenario de Horacio [ponencias del Congreso celebrado en Salamanca en 1992], Salamanca, 127-138.

— (1994b), “Ultime difficoltà nella carriera di un poeta giambico: L’epodo XVII”, in Comitato Nazionale per le Celebrazioni del Bimillenario della Morte di Q. Orazio Flacco (ed.), Atti dei convegni di Venosa Napoli Roma novembre 1993, Venosa, 205-217.

— (1995), “Poetica di un mito sessuale: la strega giambica”, in Raffaelli R. (ed.), Vicende e figure femminili in Grecia e a Roma: atti del convegno, Pesaro, 28-30 aprile 1994, Ancona, 335-342.

— (2001), “Horace and Iambos: The Poet as Literary Historian”, in Cavarzere A., Aloni A. and Barchiesi A. (eds.), Iambic Ideas: Essays on a Poetic Tradition from Archaic Greece to the Late Roman Empire, Lanham, 141-164.

— (2002), “Palingenre: Death, Rebirth and Horatian Iambos”, in Paschalis M. (ed.), Horace and Greek Lyric Poetry, Rethymnon, 47-69.

Bather P. and Stocks C. (eds.), (2016), Horace’s Epodes: Contexts, Intertexts, and Reception, Oxford.

Bassi K., (1994), “Desired Silence: Amor and Mors in Tibullus 1.1”, Syllecta Classica 5, 53-61.

Beard M., North J. and Price S., (1998), Religions of Rome Volume 1: A History, Cambridge.

Betz H. D. (ed.), (1986), The Greek Magical Papyri in Translation: Including the Demotic Spells, Chicago.

Bicknell P. J., (1984), “The Dark Side of the Moon”, in Moffat A. (ed.), Maistor: Classical, Byzantine and Renaissance Studies for Robert Browning, Canberra, 67-75.

Boyle A. J., (1974), “Propertius I, 19: A Critical Study”, Latomus 33.4, 895-911.

Bright D. F., (1978), Haec Mihi Fingebam: Tibullus in his World, Leiden.

Bushala E. W., (1968), “Laboriosus Ulixes”, Classical Journal 64.1, 7-10.

Butler H. E. and Barber E. A. (eds.), (1933), The Elegies of Propertius, Oxford.

Cairns F., (1974), “Some Observations on Propertius 1.1”, Classical Quarterly 24.1, 94-110.

— (1979), Tibullus: A Hellenistic Poet at Rome, Cambridge.

— (1983), “Propertius 1,4 and 1,5 and the ‘Gallus’ of the Monobiblos”, in Cairns F. (ed.), 1984: Papers of the Liverpool Latin Seminar Fourth Volume, Liverpool, 61-103.

— (1995), “Horace, Odes 3.7: Elegy, Lyric, Myth, Learning, and Interpretation”, in Harrison S. (ed.), Homage to Horace: A Bimillenary Celebration, Oxford, 65-99.

— (2006), Sextus Propertius: The Augustan Elegist, Cambridge.

Camps W. A. (ed.), (1961), Propertius: Elegies Book I, Cambridge.

Carruba R. W., (1969), The Epodes of Horace: A Study in Poetic Arrangement, The Hague.

Ciraolo L. J., (1995), “Supernatural Assistants in the Greek Magical Papyri”, in Meyer M. and Mirecki P. (eds.), Ancient Magic and Ritual Power, Leiden, 279-295.

Clausen W., (1964), “Callimachus and Latin Poetry”, Greek, Roman and Byzantine Studies 5.3, 181-196.

— (1976), “Cynthivs”, American Journal of Philology 97.3, 245-247.

Collins D., (2008), Magic in the Ancient Greek World, Malden.

Commager S., (1962), The Odes of Horace: A Critical Study, New Haven.

— (1974), A Prolegomenon to Propertius, Norman.

Coutelle E., (2005), Poésie et métapoésie chez Properce: de l’ars amandi à l’ars scribendi, Leuven.

Courtney E., (1969), “Three Poems of Propertius”, Bulletin of the Institute of Classical Studies of the University of London 16, 70-87.

Damer E. Z., (2016), “Iambic Metapoetics in Horace, Epodes 8 and 12”, Helios 43.1, 55-85.

Daniel R. W. and Maltomini F. (eds.) (trs.), (1990-1992), Supplementum Magicum I-II, Papyrologica Coloniensia Vol. XVI.1-2, Opladen.

Davis G., (1991), Polyhymnia: The Rhetoric of Horatian Lyric Discourse, Berkeley.

Deremetz A., (1987), “Le carmen deductum ou le fil du poème: à propos de Virgile, Buc., VI”, Latomus 46.4, 762-777.

Eitrem S., (1933), “Exkurs: Die 5. und 17. Epode des Horaz”, Symbolae Osloenses 12, 30-38.

Elder J. P., (1962), “Tibullus: Tersus Atque Elegans”, in Sullivan J. P. (ed.), Critical Essays on Roman Literature: Elegy and Lyric, London, 65-105.

Enk P. J. (ed.), (1946), Sex. Propertii Elegiarum: Liber I (Monobiblos). Pars Prior: Prolegomena et Textum Continens, Leiden.

Fabre-Serris J., (2010), “La réception des Amores de Gallus dans un lusus horatien: lectures de l’Epode 11”, in Alvarez Moran C. and Iglesias Montiel R. (eds.), Perfiles Grecia y Roma II (Actas del XII Congreso Español de Estudios Clásicos), Madrid, 881-895.

Faraone C. A., (1999), Ancient Greek Love Magic, Cambridge.

Fauth W., (1980), “Venena Amoris: Die motive des Liebeszaubers und der Erotischen Verzauberung in der augusteischen Dichtung”, Maia 32, 265-282.

Fedeli P., (1978), “Il V epodo e i giambi d’Orazio come espressione d’arte Alessandrina”, Museum Philologum Londiniense 3, 67-138.

— (ed.), (1980), Sesto Properzio: il primo libro delle elegie, Firenze.

Fitzgerald W., (1988), “Power and Impotence in Horace’s Epodes”, Ramus 17.2, 176-191.

Flaschenreim B., (1997), “Loss, Desire, and Writing in Propertius 1.19 and 2.15”, Classical Antiquity 16.2, 259-277.

Freudenburg K., (1993), The Walking Muse: Horace on the Theory of Satire, Princeton.

Fulkerson L., (2013), “Seruitium Amoris: the Interplay of Dominance, Gender and Poetry”, in Thorsen T. S. (ed.), The Cambridge Companion to Latin Love Elegy, Cambridge, 180-193.

Ezquerra A. A., (1997), “Los Epodos eróticos de Horacio y los inicios de la elegía Latina”, Estudios clásicos 111, 7-26.

Gager J. G. (ed.), (1992), Curse Tablets and Binding Spells from the Ancient World, Oxford.

Gardiner H. H., (2013), Gendering Time in Augustan Love Elegy, Oxford.

Goold G. P. (ed.) (tr.), (1990), Propertius: Elegies, Cambridge.

Gordon R., (1999), “Imagining Greek and Roman Magic”, in Ankarloo B. and Clark S. (eds.), Witchcraft and Magic in Europe Volume 2: Ancient Greece and Rome, London, 161-275.

Gowers E., (2016), “Girls will be Boys and Boys will be Girls, Or What is the Gender of Horace’s Epodes?”, in Bather P. and Stocks C. (eds.), Horace’s Epodes: Contexts, Intertexts, and Reception, Oxford, 103-130.

Graf F., (1997), Magic in the Ancient World, Cambridge.

Grassmann V., (1966), Die erotischen Epoden des Horaz: Literarischer Hintergrund und sprachliche Tradition, Monographien zur klassischen Altertumswissenschaft 39, München.

Houghton L. B. T., (2007), “Tibullus’ Elegiac Underworld”, Classical Quarterly 57.1, 153-165.

Harrison S., (2007), Generic Enrichment in Vergil and Horace, Oxford.

Heslin P., (2010), “Virgil’s Georgics and the Dating of Propertius’ First Book”, Journal of Roman Studies 100, 54-68.

— (2011), “Metapoetic Pseudonyms in Horace, Propertius and Ovid”, Journal of Roman Studies 101, 51-72.

Heyworth S. J., (1993), “Horace’s Ibis: On the Titles, Unity, and Contents of the Epodes”, in Cairns F. and Heath M. (eds.), Papers of the Leeds International Latin Seminar Seventh Volume, Leeds, 85-96.

— (2007), Cynthia: A Companion to the Text of Propertius, Oxford.

Hutchinson G. (ed.), (2006), Propertius Elegies Book IV, Cambridge.

Ingallina S. S., (1977), Orazio e la magia: Sat. 1.8, Epodi 5 e 17, Palermo.

Ingleheart J., (2010), A Commentary on Ovid, Tristia, Book 2, Oxford.

Johnston S. I., (2001), “Charming Children: The Use of the Child in Ancient Divination”, Arethusa 34.1, 97-117.

— (2002), “Sacrifice in the Greek Magical Papyri”, in Mirecki P. and Meyer M. (eds), Magic and Ritual in the Ancient World, Leiden, 344-358.

Johnson T. S., (2012), Horace’s Iambic Criticism: Casting Blame (Iambikē Poiēsis), Leiden.

Kennedy D. F., (1993), The Arts of Love: Five studies in the discourse of Roman love elegy, Cambridge.

Kidd D. A., (1979), “Propertius Consults His Astrologer”, Greece & Rome 26.2, 169-180.

King J. K., (1980), “The Two Galluses of Propertius’ Monobiblos”, Philologus 124.2, 212-230.

Knox P. E., (2005), “Milestones in the Career of Tibullus”, Classical Quarterly 55.1, 204-216.

Leo F., (1960), “De Horatio et Archilocho”, in Fraenkel E. (ed.), Ausgewählte kleine Schriften II, Rome, 139-157.

Lowrie M., (1997), Horace’s Narrative Odes, Oxford.

— (2009), Writing, Performance, and Authority in Augustan Rome, Oxford.

Luck G., (1962), Hexen und Zauberei in der römischen Dichtung, Zürich.

— (1976), “An Interpretation of Horace’s Eleventh Epode”, Illinois Classical Studies 1, 122-126.

Luther A., (2003), “Ein Terminus ante quem für die Monobiblos’ des Properz”, Latomus 62.4, 801-806.

Lyne R. O. A. M., (1979), “Servitium Amoris”, Classical Quarterly 29.1, 117-130.

— (1998), “Love and Death: Laodamia and Protesilaus in Catullus, Propertius, and Others”, Classical Quarterly 48.1, 200-212.

Macleod C. W., (1976), “Propertius 4.1”, in Cairns F. (ed.), Papers of the Liverpool Latin Seminar First Volume, Liverpool, 141-153.

Maltby R. (ed.), (2002), Tibullus: Elegies, Text, Introduction and Commentary, Cambridge.

Mankin D. (ed.), (1995), Horace: Epodes, Cambridge.

Martinez D., (1995), “‘May She Neither Eat Nor Drink’: Love Magic and Vows of Abstinence”, in Meyer M. and Mirecki P. (eds.), Ancient Magic and Ritual Power, Leiden, 335-359.

McNamee K., (1993), “Propertius, Poetry, and Love”, in DeForest M. (ed.), Women’s Power, Man’s Game: Essays on Classical Antiquity in Honor of Joy K. King, Illinois, 215-248.

Michels A. K., (1955), “Death and Two Poets”, Transactions and Proceedings of the American Philological Association 86, 160-179.

Miller P. A., (2004), Subjecting Verses: Latin Love Elegy and the Emergence of the Real, Princeton.

Murgatroyd P. (ed.), (1980), Tibullus I: A Commentary on the First Book of the Elegies of Albius Tibullus, Pietermaritzburg.

Murgia C. E., (1989), “Propertius 4.1.87-88 and the Division of 4.1”, Harvard Studies in Classical Philology 92, 257-272.

Mutschler F.-H., (1985), Die poetische Kunst Tibulls: Struktur und Bedeutung der Bücher 1 und 2 des Corpus Tibullianum, Frankfurt am Main.

Nethercut W. R., (1976), “Propertius’ ‘Fallax Opus’ (4.1.135)”, Rheinisches Museum für Philologie 119.1, 30-38.

Nikoloutsos K. P., (2011), “From Tomb to Womb: Tibullus 1.1 and the Discourse of Masculinity in Post-Civil War Rome”, Scholia N.S. 20, 52-71.

Nisbet R. G. M. and Hubbard M., (1970), A Commentary on Horace: Odes Book 1, Oxford.

— (1978), A Commentary on Horace: Odes Book 2, Oxford.

Ogden D., (2001), Greek and Roman Necromancy, Princeton.

Oliensis E., (1991), “Canidia, Canicula, and the Decorum of Horace’s Epodes”, Arethusa 24.1, 107-135.

— (1998), Horace and the Rhetoric of Authority, Cambridge.

O’Neil E., (1958), “Cynthia and the Moon”, Classical Philology 53.1, 1-8.

O’Neill K., (1998), “Symbolism and Sympathetic Magic in Propertius 4.5”, Classical Journal 94.1, 49-88.

Papanghelis T. D., (1987), Propertius: A Hellenistic Poet on Love and Death, Cambridge.

Parca M., (1988), “Tardus Amor and Tardus Apollo in Propertius’ Monobiblos”, Latomus 41.3: 584-588.

Pasoli E., (1977), “Poesia d’amore e ‘metapoesia’: aspetti della modernità di Properzio”, in Bigaroni M. and Santucci F. (eds.), Colloquium Propertianum Assisi, 26-28 marzo 1976: Atti, Assisi, 101-121.

Postgate J. P. (ed.), (1903), Selections from Tibullus and Others, London.

Prince M. D., (2002), “Magic Love and the Limits of Power: The Figure of Medea in Latin Love Elegy”, PhD Diss., Duke University, North Carolina.

— (2003), “Medea and the Inefficacy of Love Magic: Propertius 1.1 and Tibullus 1.2”, The Classical Bulletin 79.2, 205-218.

Putnam M. C. J., (1972), “Horace and Tibullus”, Classical Philology 67.2, 81-88.

— (1973), Tibullus: A Commentary, Oklahoma.

Quinn K., (1982), “The Poet and his Audience in the Augustan Age”, in Aufstieg und Niedergang der römischen Welt II.30.1, 75-180.

Richardson Jr. L. (ed.), (1977), Propertius: Elegies I-IV, Norman.

Richlin A., (1983), The Garden of Priapus: Sexuality and Aggression in Roman Humor, New Haven.

— (1984), “Invective Against Women in Roman Satire”, Arethusa 17.1, 67-80.

Rimell V., (2005), “Facing Facts: Ovid’s Medicamina through the Looking Glass”, in Ancona R. and Greene E. (eds.), Gendered Dynamics in Latin Love Poetry, Baltimore, 177-205.

Rives J. B., (1995), “Human Sacrifice among Pagans and Christians”, Journal of Roman Studies 85, 65-85.

Ross Jr. D. O., (1975), Backgrounds to Augustan Poetry: Gallus, Elegy, and Rome, Cambridge.

Rothstein M. (ed.), (1966), Propertius Sextus Elegien Erster Teil: Erstes und zweites Buch, Dublin and Zürich.

Sandbach F. H., (1962), “Some Problems in Propertius”, Classical Quarterly 12.2, 263-276.

Schlegel C., (2005), Satire and the Threat of Speech in Horace’s Satires, Book 1, Madison.

Schmidt E. A., (1990), Notwehrdichtung: Moderne Jambik von Chénier bis Borchardt (mit einer Skizze zur antiken Jambik), München.

Scibilia A., (2002), “Supernatural Assistance in the Greek Magical Papyri: The Figure of the Parhedros”, in Bremmer J. N. and Veenstra J. R. (eds.), The Metamorphosis of Magic from Late Antiquity to the Early Modern Period, Leuven, 71-86.

Shackleton Bailey D. R., (1949), “Propertiana”, Classical Quarterly 43.1/2, 22-29.

— (1967), Propertiana, Amsterdam.

Sharrock A., (1994), Seduction and Repetition in Ovid’s Ars Amatoria II, Oxford.

— (1995), “The Drooping Rose: Elegiac Failure in Amores 3.7”, Ramus 24.1, 152-180.

— (2013), “The Poeta-Amator, Nequitia and Recusatio”, in Thorsen T. (ed.), The Cambridge Companion to Latin Love Elegy, Cambridge, 151-165.

Skinner M. B., (2018), “Lesbia as Procuress in Horace Epode 12”, EuGeStA 8, 131-144.

Smyth W. R., (1949), “Interpretationes Propertianae”, Classical Quarterly 43.3/4, 118-125.

Solmsen F., (1975), “‘Catullus’ Artistry in c. 68: A pre-Augustan ‘subjective’ Love-Elegy”, in Lefèvre E. (ed.), Monumentum Chiloniense: Studien zur augusteischen Zeit: Kieler Festschrift für Erich Burck zum 70. Geburtstag, Amsterdam, 260-276.

Spina L., (1993), “Orazio nell’Ade (Lettura del’epodo XVII)”, Lexis 11, 163-188.

Stahl H.-P., (1985), Propertius: “Loveand “War. Individual and State Under Augustus, Berkeley.

Stratton K. B., (2007), Naming the Witch: Magic, Ideology, and Stereotype in the Ancient World, New York.

Stroh W., (1971), Die Römische Liebeselegie als werbende Dichtung, Amsterdam.

Suits T. A., (1976), “The Iambic Character of Propertius 1.4”, Philologus 120.1, 86-91.

Syme R., (1978), History in Ovid, Oxford.

Syndikus H. P., (1972), Die Lyrik des Horaz: Eine Interpretation der Oden. Band 1: Erstes und zweites Buch, Darmstadt.

Tupet A.-M., (1976), La Magie Dans La Poésie Latine. Vol. I: Des origines à la fin du règne d’Auguste, Paris.

Ullman B. L., (1912), “Horace and Tibullus”, The American Journal of Philology 33.2, 149-167.

Veyne P., (1988), Roman Erotic Elegy: Love, Poetry, and the West, Chicago.

Voutiras E., (1998), ΔΙΟΝΥΣΟΦΩΝΤΟΣ ΥΑΜΟΙ: Marital Life and Magic in Fourth Century Pella, Amsterdam.

Wallis J., (2018), Introspection and Engagement in Propertius: A Study of Book 3, Cambridge.

Watson L. C., (2003), A Commentary on Horace’s Epodes, Oxford.

White P., (1993), Promised Verse: Poets in the Society of Augustan Rome, Cambridge.

Wimmel W., (1968), Der frühe Tibull, München.

— (1987), “Zur Rolle magischer Themen in Tibulls Elegie 1.5”, Würzburger Jahrbücher für die Altertumswissenschaft, 231-248.

Winkler J. J., (1991), “The Constraints of Eros”, in Faraone C. A. and Obbink D. (eds.), Magika Hiera: Ancient Greek Magic and Religion, Oxford, 214-243.

Wyke M., (1987), “Written Women: Propertius’ Scripta Puella”, Journal of Roman Studies 77, 47-61.

— (1989a), “Mistress and Metaphor in Augustan Elegy”, Helios 16.1, 25-47.

— (1989b), “Reading Female Flesh: Amores 3.1”, in Cameron A. (ed.), History as Text: The Writing of Ancient History, London, 113-143.

Zetzel J. E. G., (1996), “Poetic Baldness and Its Cure”, Materiali e discussioni per l’analisi dei testi classici 36, 73-100.


1 Servitium here refers to the elegiac lover’s representation of his relationship with his puella as a form of enslavement, both to the girl as his mistress and to his passion for her. For an overview of servitium amoris (the enslavement of love) in elegy, see: Fulkerson 2013. Carmina is a multivalent term which can mean poems, songs, incantations, or prayers: the Latin elegists play on this polysemy to characterise their poetry as magic incantations. For this use of carmen in elegy: Sharrock 1994: 50-86. Return to text

2 For Canidia as Horace’s iambic “anti-Muse”: Oliensis 1991: 110-119. For Canidia personifying Horatian iambic: Barchiesi 2002: 51-52 and 63-64; see also Andrisano 2012: 294-297. Return to text

3 Barchiesi 1994b: 212-213 and 216-217 and Barchiesi 1995: 339-342 identify Lesbia, Lycoris, and Cynthia as metaliterary constructions who parallel Canidia. Damer 2016: 56 argues that the women in Horace Epodes 8 and 12 who embody Horace’s iambic text are evidence of a “two-way influence between the poets of Roman iambic and elegiac erotic poetry”. Return to text

4 Rimell 2005: 177-205 notes a poetic element in the connection between the puella’s beauty and magic in Ovid’s erotodidactic work; see also Sharrock 1994: 74-76. As I aim to demonstrate here, this connection is fundamental to Latin love elegy from the beginning of the genre as we have it. Return to text

5 For the metrical variety of Epodes 11-16 and the enrichment of Horace’s iambic with contemporary elegy, including Propertius and Tibullus, in the second half of the Epodes: Harrison 2007: 120-135; Bather and Stocks 2016: 12. The interaction between early love elegy and Horace’s Epodes that I suggest may, of course, be read from the opposite direction; it seems more probable, however, that Horace comically amplifies this subtextual element of early elegy. For suggestions that the lover’s curses on the lena (“procuress”) at Tibullus 1.5.49-56 draw on Epode 5.83-102: Luck 1962: 50-51; Wimmel 1987: 239-241. Magic becomes more prominent in later love elegy, particularly in the figure of the lena in Propertius 4.5 and Ovid Amores 1.8, two figures who are very likely responding to Horace’s Canidia: see, for example, Myers 1996: 6. Return to text

6 For Horace’s critique of the elegiac fixation on a single beloved in Odes 1.5, 2.8, and 2.9, and Epistles 1.4: Davis 1991: 29-60; Lowrie 1997: 77-93 and 266-297 (Odes 3.7 and 3.11). See also: Putnam 1972: 81-88; Ball 1994: 409-414. Return to text

7 Leo 1960: 139-157 at 146-153; Luck 1976: 122-126. Lyne 1979: 117-130; Fabre-Serris 2010: 881-895. For Epode 11 engaging with Alexandrian sources shared with Latin elegy rather than with the contemporary Roman genre: Grassmann 1966: 34-46; Fedeli 1978: 117-118; Ezquerra 1997: 7-26. Return to text

8 Recent re-evaluations of the criteria for dating Propertius Book 1 and Tibullus Book 1 have also suggested that these books were composed, and perhaps published, before Actium. Heslin 2010: 54-61 dates Propertius Book 1 to “most probably” early 33 B.C.; Luther 2003: 801-806 dates Propertius Book 1 to before 30-29 B.C. Knox 2005: 204-216 dates Tibullus Book 1 to before Actium; Ingleheart 2010: 346 and 356 convincingly critiques Knox’s arguments. Even if they are not entirely convincing, Knox’s arguments nevertheless demonstrate the flexibility of the criteria for dating Tibullus’ book. Return to text

9 For the composition period of Horace’s Epodes: Watson 2003: 1 and Carruba 1969: 15-17. Return to text

10 For poetic interaction and circulation in Rome: Quinn 1982: 75-108; White 1993: 35-63. Return to text

11 For example, Maltby 2002: 39-40 (“Possible echoes of Virgil’s Aeneid in Tib. 2.5 need not mean that book two was composed after the publication of the Aeneid, but simply that T. had heard pre-publication recitations of parts of the work”) and Watson 2003: 76-77 (“That the Georgics were published after the Epodes proves nothing … it seems virtually certain that Horace, as a fellow poet and close friend, was party to [Vergil’s] occasional recitations … and that in consequence Horace was acquainted with the Georgics long before their formal publication”). Return to text

12 Lyne 1979: 117-130; Heslin 2011: 51-66. For Epode 11 and Propertius 1.4, see also Barchiesi 1994a: 127-138, at 133-134. On Propertius 1.4 as an iambic elegy: Suits 1976: 86-91. Return to text

13 Heslin 2011: 59. Richlin 1984: 75: “The ludicrous monster in Hor. Epod. 12 reverses the norms of elegy”. Oliensis 1991: 115-116 and 1998: 71 describes Canidia as “a debased version of Catullus’s Lesbia”; in the extant Catullan corpus, however, Lesbia herself is not connected with magic. Canidia’s magic suggests that she is more likely to be commenting on the elegiac puellae. Return to text

14 Damer 2016. Return to text

15 Skinner 2018. Reading this dialogue from the opposite direction, Gowers 2016: 120-121 identifies parallels between the relationship dynamics of Epode 12 and Propertius 1.5: “Binding, poison, anger: elegiac male entrapment in female toils is easily translated into the terms of iambic debilitation. Indeed, Cynthia is pushing at her elegiac limits”. Return to text

16 For masculinity and political upheaval in the Epodes: Fitzgerald 1988; Oliensis 1991; Oliensis 1998: 64-101. For these concerns in early elegy: Wyke 1989a: 41-43; Nikoloutsos 2011. Return to text

17 Sharrock 1994: 50-86. Return to text

18 O’Neill 1998: 63. Return to text

19 Willis 2018: 41-45. Return to text

20 For the parallel between magic carmina and elegiac carmina at 1.1.19-24: Ahl 1974: 91 n. 24 and 92-93; Commager 1974: 34; Zetzel 1996: 97; O’Neill 1998: 74; Prince 2003: 210-211. Return to text

21 Vergil employs deducere in Eclogue 6 in his adaptation of Callimachus’ Aetia prologue: Apollo tells Tityrus that “a shepherd should rear a fat flock but sing a fine-spun song” (pastorem, Tityre, pinguis / pascere oportet ovis, deductum dicere carmen, 4-5). deducere also occurs in Eclogue 6 during the elegist Gallus’ poetic initiation: Linus, handing Gallus the pipes of the Muses, recalls how Hesiod “was accustomed to draw trees down from the mountains with their singing” (ille solebat / cantando rigidas deducere montibus ornos, 70-71). For Eclogue 6.5: Ross 1975: 19 and 26-27 with 27 n. 1; Deremetz 1987: 764-770; Zetzel 1996: 78 and 97. Parca 1988: 585 n. 6 notes that Propertius links his elegy to the Neoteric tradition through deducere. Return to text

22 On the simple verb expressing the same meaning as its compound: Ross 1975: 65-66. For the “chiastic” arrangement of 1.1.19-24: Cairns 1974: 100. Return to text

23 Cytaeines (“Colchian” 1.1.24) is an emendation of the MSS †Cythalinis†. Enk 1946, Rothstein 1966, and Richardson 1977 adopt Cytaeines; Butler and Barber 1933, Camps 1961, and Heyworth 2007 adopt Cytinaeis (“Thessalian”); Fedeli 1980 retains the MSS reading. For discussion: Butler and Barber 1933: 155-156; Enk 1946: 15-16; Fedeli 1980: 82; Prince 2002: 64-65. Shackleton Bailey 1949: 22 highlights Argonautica 4.59-60 as a parallel for Propertius 1.1.19 but does not develop the significance of this parallel for magic in 1.1. Return to text

24 Most editors retain the dominant reading fallax at 4.1.135, as most editors retain fallacia at 1.1.19. Stroh 1971: 108, discussing Propertius 4.1.135, notes that fallacia fingere (“to create tricks”) concisely expresses carmina fingere (“to compose poems”) and the seductive, bewitching effects of elegy. On fallax: Butler and Barber 1933: 332; Shackleton Bailey 1967: 224-225; Stroh 1971: 107-108 n. 169; Nethercut 1976: 30-38; Kidd 1979: 177; Coutelle 2005: 533. Murgia 1989: 268 and Goold 1990 adopt pellax (“deceitful”). For elegos (4.135) denoting love elegies specifically: Butler and Barber 1933: 332; MacLeod 1976: 147 and 148; Hutchinson 2006: 84. For the likelihood that Propertius and his audience did not distinguish amatory from non-amatory elegy: Sandbach 1962: 268 and Courtney 1969: 75. Return to text

25 For Medea alluding to the failure of love elegy in Propertius 1.1.23: Prince 2003: 209-211. Return to text

26 Putnam 1973: 68. Bright 1978: 147, with n. 54, calls compono a “key-word” of 1.2, highlighting 1.2.93 as “a reflection of the poetic process” alongside 1.2.22 and 1.2.55. Return to text

27 For this relationship between 1.1 and 1.2: Bright 1978: 133-148; Mutschler 1985: 50 and 63-64; Lee-Stecum 1998: 71 and 100; Maltby 2002: 50 and 153. Return to text

28 For Medea suggesting the failure of love elegy at Tibullus 1.2.53: Prince 2003: 213-215. Return to text

29 For the elegiac “pseudo-opposition” between magic and love, rather than magic and beauty: Sharrock 1994: 58-61. Return to text

30 For the social function of accusations of magic for rationalising or justifying behaviour in Greco-Roman culture: Gordon 1999: 194-204. Return to text

31 For accusations of magic revealing more about the accuser than their target: Winkler 1991: 215. Return to text

32 Elder 1962: 65-105 at 77-78 highlights Tibullus’ equation of beauty and magic in 1.5.41-44; see also Fauth 1980: 274-275. Return to text

33 Sharrock 1994: 76 highlights that Tibullus uses the repetition of noceo to equate magic and love in 1.8.23-26. For Tibullus 1.8.17-26 opposing beauty and magic: Luck 1962: 51-52; Wimmel 1968: 59-61; Putnam 1973: 130; Tupet 1976: 345; Bright 1978: 242-243; Cairns 1979: 140; Maltby 2002: 307. Return to text

34 Kennedy 1993: 47-48. Sharrock 1994: 57 interprets Propertius 1.1.1 as implying that Cynthia has performed a binding spell on Propertius with her eyes, describing Cynthia as “both the essence and agent” of the magic she uses to enchant Propertius. Fauth 1980: 265-282 discusses the opposition between “internal” physical magic and “external” magic in Propertius 1.1.1-4. Return to text

35 For Cynthia’s etymological connection with luna in 1.1.21-22: O’Neil 1958: 2-3; Ahl 1974: 81 and 91; Commager 1974: 33-34; Bicknell 1984: 69. For “Cynthius” as an epithet for Apollo: Clausen 1976: 245-247; Wyke 1987: 59; Wyke 1989a: 33. Return to text

36 Zetzel 1996: 95 compares the symptoms of love in 1.5 with those in an ἀγωγή spell, PGM IV.1508ff. and 350ff. Gowers 2016: 120-121 suggests that the narrator of 1.5 constructs Cynthia as a “heartless Thessalian witch”. Return to text

37 For Propertius 1.5.6 illustrating the narrator’s experience of love: Luck 1962: 39; Fauth 1980: 279-280. For 1.5.6 contributing to Cynthia’s characterisation as a “magical spirit” in 1.5: Zetzel 1996: 92-97, though Zetzel argues that Propertius opposes magic and Cynthia’s enchanting effect on her lovers to his poetic enchantment. Return to text

38 For the delayed introduction of Gallus as the addressee in 1.5 and for 1.4 and 1.5 as a pair: Cairns 1983: 61-103. For identifying “Gallus” in 1.5 with Cornelius Gallus: King 1980: 212-230; Cairns 1983: 79-88; Miller 2004: 60-94; Cairns 2006 Chapters 3-7. Contra: Syme 1978: 99-103; Anderson et al. 1979: 154-155; Fedeli 1980: 153. Return to text

39 I return to this connection between the two poems below. Return to text

40 For Cynthia illustrating Propertius’ poetics, as well as being a fictional beloved in the narrative: Wyke 1987: 47-61; Veyne 1988: 3-14 and 50-63; Wyke 1989a: 25-47; McNamee 1993: 215-248; for Delia and Pholoe in Tibullus Book 1: Bright 1978: 99-123; Veyne 1988: 50-66. Commentators do not relate the metapoetic role of the elegiac beloveds to the motif of magic as I do here. Return to text

41 Wyke 1987: 58-60 argues that Cynthia’s “stupefaction” by poetry represents the composition of elegy at Propertius 2.13.3-8, highlighting that the effect of Propertius’ elegy on Cynthia “is expressed in the same vocabulary as the spellbinding of natura” (sed magis ut nostro stupefiat Cynthia versu: / tunc ego sim Inachio notior arte Lino, “but rather that Cynthia should marvel at my poetry: / then I would be more famous in my art than Inachian Linus”, 2.13.7-8) and that this illustrates “an analogous yet favoured form of poetic production” to that expressed in Eclogue 6.71 (cantando rigidas deducere montibus ornos). Return to text

42 Horace targets the power of the elegiac puella’s beauty to reduce her lovers to a servile status in his Odes, most notably through Barine in Odes 2.8. For Barine’s elegiac characteristics: Syndikus 1972: 388; Nisbet and Hubbard 1978: 123. Return to text

43 For Thyesteas preces (86) and revenge: Mankin 1995: 133-134; Watson 2003: 243; Johnson 2012: 108-109. On humanam vicem (88): Mankin 1995: 134; Watson 2003: 243-244. Return to text

44 For agere characterising Horatian iambic: Barchiesi 2001: 145-146; Barchiesi 2002: 44 and 64; Lowrie 2009: 104-105 and 108-110. Return to text

45 Oliensis 1998: 73-77 and 95-96, citing (96 n. 79) Maurizio L. (1989), “Engendering Invective”, American Philological Association meeting, Boston, December 1989, for “the ‘educational’ value” of Epode 5; for the “origins of invective in impotence” in Epode 5 and the mirroring of witch and poet: Oliensis 1998: 73-77. Johnson 2012: 101-109. For the boy as an iambic poet: Andrisano 2012: 298. Return to text

46 For the boy’s curses suggesting the “pragmatic” ineffectiveness of iambic: Lowrie 2009: 110-111. Epode 6 reinforces this impression: the narrator undercuts the effectiveness of his iambic tirade by asking if he will weep like an unavenged boy (an si quis atro dente me petiverit, / inultus ut flebo puer?, “and if anyone attacks me with a black tooth, I will weep like a boy unavenged?”, 15-16): Fitzgerald 1988: 185-187; Oliensis 1998: 76-77; Lowrie 2009: 110; Johnson 2012: 106-108. Pace Watson 2003: 187-190. Return to text

47 On Propertius 1.7.19: Stroh 1971: 18-21. Horace Odes 2.19.17-18 characterises Valgius’ elegies as “soft complaints” (mollium … querelarum), defining Horace’s lyric against elegy’s excessive love and grief: Commager 1962: 239-241; Putnam 1972: 81-88; Davis 1991: 39-60 and 184-186; Lowrie 1997: 77-93. Return to text

48 The older, rather than richer, rival in Tibullus 1.8 is unique in extant love elegy and Horace’s Odes 1.33 suggests that Horace was alert to Tibullus’ distinctive use of age: the lyric narrator comforts the mourning elegist, Albius (Albi, 1) – whom commentators identify with Tibullus – over losing Glycera to a younger rival (cur tibi iunior / laesa praeniteat fide, “why, when faith has been broken, a younger man outshines you”, 3-4). For Horace Odes 1.33 ridiculing Tibullus’ exaggeration of age: Cairns 1995: 72-73 with n. 12. For debate over the problem which this focus on age creates for equating “Albius” with Tibullus: Postgate 1903:183; Ullman 1912: 153. Watson 2003: 186 compares Tibullus 1.8.24 (“an argument widely canvassed in love-poetry”) with witchcraft as “Canidia’s only hope of holding onto Varus”. Return to text

49 Cicero in Vatinium 14. On Cicero’s allegation as evidence of public attitudes towards magic rather than of real events: Tupet 1976: 206-208; Beard-North-Price 1998: 155-156; Rives 1995: 72-74. Ovid Fasti 6.101 testifies to the superstition regarding screech-owls and young boys: Ingallina 1977: 116-119. For the use of young, chaste boys in magic: Watson 2003: 197. On “boy-mediums”: Ogden 2001: 196-201; for analysis of male and female children as divinatory mediums: Johnston 2001: 97-117. Sexually pure boys are specified most explicitly at PDM xiv.68 and PDM xiv.805-40. Return to text

50 Magic was practised throughout the Greek and Roman worlds in all periods. Though much of the extant evidence is of a late date – the majority of the texts compiled in the PGM date to the third-fourth centuries A.D. – it transmits material with a far longer heritage. For the validity of using the PGM as evidence for earlier practices: Faraone 1999: 33-36. Return to text

51 Eitrem 1933: 30-88 at 37. On the connection between the attraction desired in PGM XII.15-95 and the “more focussed” passion of erotic spells: Winkler 1991: 214-43 at 220. For “Eros assistants” in the PGM: Ciraolo 1995: 279-295; Scibilia 2002: 71-86; Collins 2008: 97-101. Return to text

52 PGM XII.21-24: π[άντα ­ταῦτ]α ­ἀποτ[ε]λέσας ­ἀφιέρωσον /ἡμέρας ­γ’­. παραθήσεις ­δὲ ­αὐ[τῷ] ­παντοῖα ­γένη ­καρ[πῶν ­νέω]ν ­πόπ[α]νά ­τε ­ζ’, στροβ[ί]λους ζ’, τραγημάτων πᾶν γ[]νος, λύχνους ἀμιλτώ[τους ζ’] καὶ [τρί]α μικρὰ δίπα[λ|τα, πινακίδας, τόξα, μῆλα φο[ι]νίκια, κρατῆρα κεκρ[α]μένον ο[]ν[ο]μέλιτι (“When you have completed all this, prepare a three-day consecration. Present to it [Eros] fresh fruits of every variety and seven round-cakes, seven pinecones, every kind of sweetmeats, seven lamps not painted red; also [three] small double-edged knives, votive tablets, a bow and arrows, Phoenician apples, a bowl mixed with honey-wine”). For the meals in PGM XII.14-95 “perfecting” the figurine: Johnston 2002: 344-358 at 355-356 with n. 30. Johnston’s discussion of “perfecting” objects through sacrifices focuses on PGM IV.26-51, in which the practitioner’s body is “The ‘tool’ that [he] must perfect”. Mankin 1995: 120 compares the boy’s meals with those given to Greek scapegoats; contra, Watson 2003: 212-213. Return to text

53 For Tantalus at Epode 5.35-36: Mankin 1995: 120; Watson 2003: 213. Return to text

54 On Tibullus 1.3.77-78: Houghton 2007: 158-160. For Tantalus’ applicability to elegiac poetics: Sharrock 1995: 155-156. Return to text

55 Smyth 1949: 123; Baker 1990: 100. Propertius’ allusion to Ixion at 1.9.20 (et magis infernae vincula nosse rotae, “and rather be bound to the infernal wheel”) underlines the allusion to Tantalus. Return to text

56 Horace also uses Tantalus in the pool to illustrate misplaced or excessive desires in Satires 1.1.68-69 and 1.2.107-108. Freudenburg 1993: 195-197 highlights that Horace connects Tantalus with parody of early Roman love elegy in Satires 1.2, where he is mentioned in the context of adulterous love and “vain amorous pursuits” (1.2.101-120) and shortly before an explicit reference to Cornelius Gallus (illam “post paulo”, “sed pluris”, “si exierit vir” / Gallis, hanc Philodemus ait …, “that woman [who says] ‘after a short time’, ‘but with more’, ‘if my husband is absent’, she, Philodemus says, is fit for the Galli”, 120-121): “As a plural form of Gallus, couched in a distinctly elegiac context, the expostulation would be heard not only as a reference to the castrated priest of Cybele but also as an allusion to Gallus himself, Rome’s most famous elegiac poet to date and a famous emulator of Callimachus. ‘For the Gauls’, then, becomes ‘For the Galluses’, as a type of woman becomes a type of poetic pursuit”. I am grateful to the anonymous reader for highlighting the significance of Satires 1.2 for my argument. Return to text

57 For comparison see Horace Odes 2.8.19-20, where Barine’s elegiac suitors are physically unable to leave her house despite their threats. Return to text

58 Gager 1992: 15; Graf 1997: 136-141; for the argument that the treatment of statues in amatory magic generated, by means of “persuasive analogy”, the same effects in the targets they represented: Faraone 1999: 41-42 and 51-53. For an alternative interpretation of the function of figurines in erotic binding spells and the differences between these figurines and those designed as magical assistants: Collins 2008: 92-103. Return to text

59 Languescere is synonymous with intabescere (TLL s.v. intabesco, 2066, 60), used of the disintegration of the child’s eyeballs at Epode 5.40; the applicability of both verbs to the physical condition of the elegiac lover reinforces Horace’s gory representation of elegiac conventions in the murder of the child. Return to text

60 Zetzel 1996: 73-100 at 92-6 demonstrates the prevalence of magical language throughout Propertius 1.5, illustrated by parallels from the PGM; see also Fauth 1980: 279-280. Return to text

61 Martinez 1995: 358. Return to text

62 Wyke 1989a: 41-3 discusses the elegiac lover’s presentation of his social isolation by emphasising his domination by his mistress; Allen 1950: 264-70 examines the programmatic presentation of a universal experience as unique to the elegiac lover of Propertius 1.1. Return to text

63 Bright 1978: 129-130; Papanghelis 1987: 53; Bassi 1994: 56-57. For Tibullus’ “conflation” of love and death in 1.1: Bassi 1994: 53-61. It is perhaps significant for my elegiac reading of Canidia that the Tibullan narrator’s imagined funeral introduces his encouragement to Delia to love him before death comes or old age makes love and elegy unsuitable (1.1.69-74): iam subrepet iners aetas neque amare decebit / dicere nec cano blanditias capite (“now useless age will steal up, and it will not be fitting to love, nor to flatter with white hair”, 71-72). Return to text

64 On perire (1.4.12) meaning “to die from love” and “to be in love”: Fedeli 1980: 133; Baker 1990: 54. Return to text

65 Baker 1990: 178 notes the equation of dolor with elegiac love in 1.17.19. Return to text

66 Flaschenreim 1997: 266 suggests that death metaphorically enables Propertius’ narrator “to have it both ways: to be present in the poem’s discourse and absent in its governing fiction”. Return to text

67 Papanghelis 1987: 12-13; Michels 1955: 175 cites 1.19.6 as exemplifying the physicality of Propertius’ vision of death. Return to text

68 Papanghelis 1987: 11-12; see also Lyne 1998: 210 and Boyle 1974: 903. Return to text

69 Oliensis 1998: 97 highlights that quid dixit aut quid tacuit (Epode 5.49) could equally well characterise Horace’s invective in Epodes 8 and 12 and connects the “inability to repress speech” with impotentia in the Epodes. Return to text

70 Barchiesi 1994b: 216-217; Barchiesi 1995: 341. Spina 1993: 181 expresses reservations about a similar interpretation. For the link between Horace and Canidia ensuring their “mutual destruction”: Heyworth 1993: 92-93. Return to text

71 Spina 1993: 163-188. Return to text

72 On Epode 17, magic, and poetic power: Barchiesi 1994b: 205-217; Lowrie 2009: 108-110. Return to text

73 Johnson 2012: 163-179. Return to text

74 Bushala 1968: 9-10, citing Rand E. K. (1937), “Horace and the spirit of comedy”, The Rice Institute pamphlet 24.2, 51-52; Johnson 2012: 167-168. Return to text

75 Stroh 1971: 36-54; Pasoli 1977: 101-111; Zetzel 1996: 97-99. On 1.8B.46: Pasoli 1977: 109; Coutelle 2005: 222-225. Return to text

76 One explicit example of this wish exists in a fourth-century A.D. curse-tablet from Pella, in which the practitioner requests that she and her beloved “grow old together” (συνκαταγηρᾶσαι, Voutiras 1998 line 5). Return to text

77 For the exclusus amator as supplex in the first collections of Propertius and Tibullus: Propertius 1.9.3 and 1.16.4; Tibullus 1.2.87, 1.4.72. Return to text

78 For the turbo used in place of rhombus or ἴυγξ (both terms for a magic wheel) here: Watson 2003: 545-546. For a survey of the debate over the identification and function of the rhombus, turbo, and ἴυγξ: Faraone 1999: 63 n. 102. Return to text

79 There are several correspondences between Epode 17.1-7 and Tibullus 1.5.1-8. turbinem (Epode 17.7) echoes the unusual form turben at Tibullus 1.5.3; both passages emphasise the speed of the top using a form of citus (“quick”) (Tibullus 1.5.3; Epode 17.7); both narrators appeal for mercy in similar terms (1.5.7; Epode 17.6); both passages use the anaphoric per (“for”) to evoke hymns (1.5.7-8; Epode 17.2-5). Cairns 1978: 546-552 and Cairns 1979: 168-171 highlight parallels between Epode 17 and Tibullus 1.5, focusing on the palinode element in both poems. For turben: Murgatroyd 1980: 162; Maltby 2002: 242. Return to text

80 For Epode 17.1 recalling Epode 5.71-72: Johnson 2012: 165. Return to text

81 On miser indicating an amatory context at Epode 17.27: Bushala 1968: 8. miser appears elsewhere in the Epodes only of Maecenas’ amatory misery at 14.13, underling the erotic meaning in Epode 17. For miser and the elegiac lover: Allen 1950: 258-560. Return to text

82 Suppl. Mag. 42.16; PGM XVIIa.10-12. Barchiesi 1994b: 214-215 n. 31 quotes PGM 4.1496 and Audollent [DT] 270 as parallels for the narrator’s torments in Epode 17. Watson 2003: 534-584 argues for a non-erotic reading of magic in Epode 17. Return to text

83 PGM IV.149 (“I will bewitch her breath … until she comes to me”); PDM xiv.655-65 includes the lungs of the victim among the places of her body to be burnt (Johnson H. J. (tr.), in Betz 1986: 231). Breathing occurs in Jordan 1985 7 and 8, erotic curses for separation: Jordan 1985: 223-227; Jordan 1985: 251-255 Inv. No. 1737 is a possible curse for erotic attraction which targets the victim’s lungs and complexion. For non-erotic curses affecting a target’s breathing: Watson 2003: 556. Return to text

84 Tibullus 1.8.57-58 (ut lenis agatur / spiritus, “how breath is drawn softly”); Propertius 1.9.32 (spiritus iste levis, “light puff of breath that you are”). Mankin 1995: 198, citing Propertius 1.16.32 and Horace Epode 11.10, notes that Propertius and Horace are the earliest extant poets to use spiritus in this way. Contra: Watson 2003: 556. Return to text

85 For correspondences between Epode 17.30 and Epode 11.4 and 11.27 underlining the amatory, though not elegiac, theme of Epode 17: Schmidt 1990: 158 n. 116. Return to text

86 Anderson et al. 1979: 140; Sharrock 2013: 151. Return to text

87 OLD s.v. constans 1b and 4a. Stahl 1985: 33-34 discusses the emphasis on moral purity at 1.1.3-4. Return to text

88 Gardiner 2013: 88-91. Return to text

89 I am grateful to the anonymous reviewers for their comments on the relevance of Epode 17 for Cynthia’s suppressed iambic speech and the significance of Epode 17 for my argument. Return to text

90 Spina 1993: 176. For Cynthia, Delia, and Cynthus: Maltby 2002: 43; for Pholoe: Nisbet and Hubbard 1970: 373. Return to text

91 Barchiesi 1994b: 214-215 n. 31. Return to text

92 For a non-amatory reading of Epode 17.62-69: Watson 2003: 536-537 and 578. Return to text

93 Epode 17.76 and 79 allude to Candia calling forth the spirits of the dead at Satires 1.8.28-29 and her manipulation of wax figurines as part of her erotic ritual at Satires 1.8.30-33 and 43-44. In Satires 1.8, Horace uses Canidia and Sagana to define his satiric programme, embodying the invective attack of Lucilian satire against which Horace characterises the focus of his own satire on laughter, defence, and community whilst also highlighting – through Priapus’ invective against Canidia – the similarity between the satirist, his targets, and the tradition of Lucilian invective which he tries to distance himself from. For discussion of Satires 1.8 and Horace’s poetic programme in Satires 1: Anderson 1972; Schlegel 2005: 98-107; Oliensis 1998: 68-72, in connection with Horace’s poetics in the Epodes. Alluding to Satires 1.8 in the conclusion of the Epodes underlines the simultaneous use of elegy for Horace’s definition of his iambic and his innovation of the Roman genre by incorporating elegy. In light of the emphasis on raising the dead and lament in Epodes 17.76-81 and Satires 1.8.25 (ululantem, “wailing”), 28-29, and 41 (triste, “sad”) and Horace’s parody of Gallan elegy in Satires 1.2, it may also be possible to read Horace as engaging with elegy through magic and Canidia in Satires 1.8, too. If so, the echoes of Satires 1.8 in Epode 17.76-81 would reinforce Horace’s incorporation of elegy into his poetic programme in the Epodes. The setting of Satires 1.8 on Maecenas’ gardens on the Esquiline may also be picked up in Propertius’ later elegies 3.23 (dominum Esquiliis scribe habitare tuum, “write that your master lives on the Esquiline”, 24), a poem in which he imagines Cynthia accusing him of slandering her (an tu / non bona de nobis crimina ficta iacis?, “or are you hurling wicked slanders against me?”, 13-14), and 4.8, in which Cynthia attacks Propertius after finding him on the Esquiline in Maecenas’ gardens (Esquilias … aquosas / … vicina novis … agris, “the lush Esquiline / … in the neighbourhood of the new gardens” 1-2) in the company of other women. Return to text

94 Lowrie 2009: 108-109 notes that Epode 17.81 relates Canidia’s abilities to “iambic effectiveness” but does not relate the witch’s powers to Horace’s poetry. Return to text

95 For Horace’s unique use of deripio at Epode 5.46: Ingallina 1977: 136-138 and Watson 2003: 220, noting that the verb introduces “a violence … normally absent from the procedure”. Return to text

96 Oliensis 1998: 95-96 reads Horace’s silence at the end of Epode 17 as the reassertion of masculine self-control after the impotentia of invective throughout the Epodes. Return to text

97 For a similar use of female personifications of genre in Ovid Amores 3.1: Wyke 1989b: 113-143 (esp. 118-124). Return to text

98 For old women as the targets of invective in Roman satire: Richlin 1983 and 1984. For the association between the health of the state and female conduct influencing witch-figures in Augustan literature, with reference to Horace Epodes 5 and 17: Stratton 2007: 71-105; for Canidia as a personification of Rome: Mankin 1995: 301. Return to text


Electronic reference

Zara Chadha, « Erotic Magic, Elegy, and Iambic in Horace’s Epodes », Eugesta [Online], 12 | 2022, Online since 01 décembre 2022, connection on 23 février 2024. URL :


Zara Chadha

University of Glasgow