• 1 Woodman 2020, reiterated and summarized at Woodman 2022: 12–14.
• 2 Heyworth 1995; Griffiths 2002; Porph. ad Hor. Carm. 3.1. As Woodman 2020: 278 n7 correctly observes (...)
• 3 The variable numbering of the five asclepiadic schemes used by different scholars is notoriously co (...)
• 4 The same phenomenon could occur if an ode in the first asclepiad (stichic asclepiad lines) were jux (...)

1Tony Woodman has recently challenged two critical claims about the first six poems of Horace’s third book of Odes, one a near-universal commonplace, the other more niche.1 His arguments against the aptness of the ubiquitous designation ‘Roman Odes’ are persuasive, though his alternative suggestion, ‘Alcaic Hexad’, is not especially appealing, either aesthetically or critically. The balance is finer between his case and that of scholars such as Stephen Heyworth and Alan Griffith, who contend that the absence of a change in metre corroborates Porphyrio’s (probable) conception of the eighty-four continuous alcaic stanzas as a single poem.2 One of Woodman’s arguments against this position—and against Heyworth’s wider principle that only a change of metre marks a change of poem—is that there are cases where readers may mistakenly think that the metrical scheme has not changed even though it has, because the stanza of the new metre begins with the same metrical lines as the old. In all three of the examples that Woodman offers, an ode in the third asclepiad is followed by one in the second.3 Both schemes begin with two asclepiadic lines, but these are followed in the third asclepiad by a pherecratean and a glyconic, in the second by another asclepiad and a glyconic. Thus, it is only when the reader reaches line three of the new poem that she realizes the metre has changed and hence (on Heyworth’s principle, which Woodman is of course disputing) that it is indeed a new poem.4

• 5 Woodman 2020: 280 = 2022: 14.

2Woodman makes an ingenious and surely correct observation here, but his assumption about the reader’s probable response to such a blurring of boundaries is more debatable. He declares that the reader (as ever, in the counterfactual scenario that Heyworth is right) ‘will think that the author has performed on them a cruel and pointless trick’.5 A trick it certainly would be, but, quite apart from the dubiously emotive notion of an author’s being cruel to their reader, it is far from self-evident that it would be pointless. This article will explore the possible effects that could be produced by encouraging the reader to think that, in two of these three cases (1.14–15, 1.23–24), the ode was continuing, only to force her to reconfigure her response on realizing that a new poem has begun. It will also examine why she might not have a similar response to the two other pairs of poems in the Odes (1.5–6, 4.12–13) where the metre alone might encourage her to do so. This move can be situated within Roman poets’ wider practice of springing surprises, twists, and even tricks on their readers, and in particular Horace’s tendency in the Odes to make use of false closure and shifts of direction to powerful thematic effect.

• 6 Some of the extensive work on the persona in Horace includes Coffta 1998, Gowers 2003, Wittchow 200 (...)
• 7 Sutherland 2002.

3Before beginning, it is worth clarifying some points of terminology and methodology. All of Horace’s works, taken individually and as a complete oeuvre, feature a particularly complex construction of poetic persona(e)—or ‘faces’—, especially as it or they interact with what can (tentatively) be reconstructed about the life and circumstances of the historical poet.6 For the purposes of this article and in the interests of brevity, I shall generally refer to poet and persona simply as ‘Horace’ (without scare-quotes), but without making naïve assumptions about the autobiographical nature of any first-person statements or believing that the poet was in love with a girl named Chloe. In similar fashion, though with greater significance for my methodology, I shall make suggestions about Horace’s aims. This is not to revert to outmoded beliefs in the knowability of authorial intention, but it does also reject excessive insistence on the death of the author. These poems were written by a human being and written with considerable skill and sophistication. Meaning is indeed generated at the point of reception, but that uncontrollability of effect does not preclude the poet’s attempts to shape his work’s reception or the validity of trying—with due humility—to reconstruct those attempts. Finally, I shall refer throughout to ‘the reader’, drawing (lightly) on reader-response theory, which Sutherland so successfully applied to Horace.7 My reader is hypothetical and constructed as one who responds to the cues I detect in the text itself. She is not the only possible reader and the responses I assign to her should not be taken as ruling out any of a wide range of other possible meanings generated at the point of reception. Finally, a coinage. The phenomenon—indeed, the trick—of false closure, where a text appears to end owing to one or more closural markers but is unexpectedly revealed to be continuing, is a familiar one and will be briefly discussed in the next section. The phenomenon which I am proposing is that in which a text—in this case an ode—appears to be continuing but is unexpectedly (and retrospectively) revealed to have ended. It feels appropriate to term this false non-closure.

• 8 Cic De or. 2.284 (glossing the Greek as praeter exspectationem), Demetr. Eloc. 152, Hermog. Meth. 3 (...)
• 9 Gerlach 1911, Cugusi 1982, Mindt 2019: 198–199.
• 10 Laurens 1989: 317, Sullivan 1991: 242, Williams 2004: 58.

4Tricks, twists, and surprises abound in Roman poetry, taking many forms and serving many ends. While it would be rash to rule out the possibility of more complex effects, many appear to serve as an end in themselves, or at most to produce a relatively straightforward response of pleasure at novelty or amusement at incongruity. The very rhetorical concept of παρὰ προσδοκίαν (‘contrary to expectation’) or ἀπροσδόκητον (‘the unexpected’) codifies the authorial practice of tricking the reader into expecting that a poem is moving in a particular direction, only to surprise them with a twist that runs counter to that expectation.8 Though present in the works of many poets—including Horace, and we shall glance at Carm. 1.16 in the next section, to which 3.26 could be added—in Martial’s epigrams, παρὰ προσδοκίαν is raised to the status almost of a defining principle.9 To offer just one example from many, Epigram 2.11 depicts the apparently grief-stricken appearance and behaviour of Sellius, before revealing in the final two words that the cause of his gloominess is the absence of a dinner invitation: he dines at home (domi cenat, 2.11.10).10 There is no sense here that the reader might feel cheated or disgruntled, or that the potential for such a reaction can be used, as Woodman suggests, as a reason to rule out the possibility of the ‘trick’.

• 11 ‘A coniunctum is an inseparable property: to separate it from the body with which it is conjoined w (...)
• 12 Hinds 1987: 452.
• 13 ‘Lucretius’ analogy has daringly flirted with confusion precisely so as to teach us how to recognis (...)
• 14 Hinds 1987: 453.

6An excellent example of this operation is Stephen Hinds’ exquisite exegesis of the tmesis seque gregari (‘and to be separated’) at Lucretius 1.452. Hinds first shows how the separation of prefix and root to produce the meaningless gregari is verbally mimetic of the principle that Lucretius is expounding, that a body cannot be separated from its property (coniunctum).11 Yet he proceeds to note that, while gregari cannot exist on its own, and the same applies to the prefix se, the latter can easily be mistaken for the homonymous pronoun. Moreover, this pronoun is the one that Lucretius has repeatedly employed to emphasize the impossibility of a property’s existing ‘by itself’ (per se). Hinds expresses the reader’s potential aporia in terms strikingly preminiscent of Woodman’s cruel and pointless trick: ‘A se represents what cannot exist per se: the analogy appears now not so much untidy in its expression as perversely misleading.’12 The reader has been tricked into misreading se as a pronoun, but the trick has a point. As a good and attentive student of Epicureanism, she both realizes her mistake and recognizes it as the sort of error that it is easy for the unenlightened to fall into. The acts of misreading, re-reading, and correcting encourage her to reflect on the necessity and the difficulty of identifying and differentiating bodies and properties.13 This is a trick, but it is very far from being a pointless one, and if it is cruel, that is precisely because it is mimetic of the mercilessly (if impersonally) tricksy nature of the universe. As Hinds sums up, ‘Appearances can be deceptive: the path to the discovery of the nature of things is not always easy. Are we good enough readers to follow the analogies of Lucretius? Are we good enough physicists to follow Epicurus?’14

• 15 On the absence of quotation marks, see Feeney 2011: 57–60.
• 16 Heyworth 1988, Fowler 1994: 240–242, Oliensis 1998: 84–87, Koster 2000: 157–165, Harrison 2007a: 11 (...)
• 17 ‘The poem is both a description of the joys of country life … and a satirical attack on the individ (...)
• 18 ‘On this reading, the concluding gesture of Epode 2 does not help rehabilitate the poem’s seriousne (...)
• 19 Johnson 2012: 88.

7Lucretius’ trick is sophisticated and complex in its operation, but simple—or perhaps better single—in its effect. Only an extreme deconstructionist or a dedicated seeker of l’Anti-Lucrèce chez Lucrèce would expect the reader, once she has realized her mistake, to cling onto any aspects of her misreading and to retain unEpicurean ideas about bodies and properties. However, many tricks and twists do exploit the reader’s inability to ‘unsee’ what she has mistakenly seen in her misreading and encourage her to let that misreading colour her corrected re-reading. This is the case with perhaps the most famous of all twists in Latin poetry, which occurs in Horace’s own Epodes. The extended praise of rustic life that fills the first sixty-six lines of Epodes 2 and that the reader (especially the reader for whom modern quotation marks do not provide a spoiler) has no reason to think is being delivered by anyone other than the poet’s primary persona is suddenly revealed, in the last four lines, to have been the oft-repeated, empty musings of the moneylender, Alfius, who will never abandon his urban lifestyle.15 This trick is by no means pointless, as its complex effect and thematic significance has been extensively discussed, including its probable engagement with a similar move in Archilochus fr. 19 West, revealing Charon the carpenter as the speaker.16 Even a relatively straightforward response to the twist would involve the reader’s retrospective reinterpretation of everything she had read in the previous sixty-six lines, not merely contextualizing them as insincere—or at best self-deludingly aspirational—musings, but reflecting on what that entails for the validity of the praise of rustic life. The rural ideal may be left intact and the iambic attack be only on Alfius’ unwillingness to make the break and embrace it, or shown in retrospect to have been always already tainted and distorted by his mercenary priorities and preoccupations.17 Both (re-)readings are prompted by the disjuncture between the poem’s close and what went before, by the correction of the misreading that lines 1–66 were spoken and vouched for by ‘Horace’. Yet the reader cannot—and perhaps should not— completely unsee what she has (mis-)seen. Misreading the words of Alfius as those of Horace opens up the possibilities of parallels between the two, even of their virtual identification.18 The laus ruris (‘praise of the countryside’) then becomes neither a pure ideal that Alfius has cannot bring himself to attain nor a perverted distortion that is as worthy of ridicule as Alfius himself, but an inherently problematic concept with which Horace, proud owner of a Sabine farm and proponent of rustic simplicity, is himself associated. As Johnson puts it, ‘Finding Horace’s “professed” simple lifestyle in a loan-shark’s wishful daydreams implicates the iambist in his own mockery.’19 The tension between assimilating and contrasting Horace and Alfius, between reading continuity and discontinuity from the laus ruris to the coda, is generated by the process of misreading, correcting, and re-reading, but re-reading in a manner that is coloured by the initial misreading.

• 20 On the effect of this metrical transgression, see Cowan 2014; on Horace’s manipulation of the alcai (...)
• 21 ‘When the implications of Augustus’ resemblance to Hercules change from pessimistic to optimistic, (...)

8Among parallels and precedents for tricks in Horace’s asclepiads, it is particularly relevant that there are also other instances where it is specifically the reader’s metrical expectations which poets manipulate to telling thematic effect, tricking her, but with a point. Some of these effects are very localized, as at Odes 4.14.17. There, for the first time since four marked occasions at Odes 1.16.21, 1.37.5, 1.37.14 and 2.17.21, Horace breaks his own self-imposed rule of placing a caesura after the fifth syllable of the alcaic hendecasyllable.20 Horace’s well-trained reader is so pre-conditioned to expect the regular word-break that she initially (mis)divides the sequence of syllables spectandusincertaminemartio to produce (with elision) spectandus incerta (‘to be seen uncertain’), only for the word to continue beyond the phantom caesura and resolve the sounds into spectandus in certamine Martio (‘to be seen in a martial contest’). Horace benignly and pointedly tricks his reader into fleetingly glimpsing the shadow of a doubt here, only to compel her to revise her interpretation of the word, the line, and perhaps the entire ode. As I have argued elsewhere, the potential effects of this trick are complex, diverse, but always meaningful. The phantom word incerta (‘uncertain’), once evoked, may linger in the reader’s consciousness, unsettling her response to the ostensibly triumphalist depiction of the young Tiberius’ victory over the Raeti. Alternatively, the very process of dramatized misreading followed by correction may—as with Morgan’s analysis of the similar process through which the reader of Odes 3.14 goes in defining and refining its incipit Herculis ritu (‘in the manner of Hercules’)—stage the parallel process of experiencing misplaced doubt followed by well-founded reassurance.21 Finally, it may dramatize, not the journey from doubt to assurance, but the impossibility of completing such a journey without some shadow of that doubt—self-reflexively embodied in the uncertain word ‘uncertain’—remaining in the reader’s mind. A trick is played, and played on the basis of the reader’s metrical expectations, but it is neither cruel nor pointless. Rather, it creatively shapes and problematizes the reader’s response to the ode, by means both of its intrinsic content and of its very status as a trick.

• 22 Tarrant 1982: 351 n35, Barchiesi 1997: 21, O’Hara 2007: 106, Morgan 2010: 350.
• 23 O’Hara 2007: 106.

9Sometimes, the metrical trick has a more wide-ranging impact on the reader’s interpretation and reinterpretation of the poem as a whole. A famous example is the opening of Ovid’s Metamorphoses. As Tarrant first observed in passing, and O’Hara, Morgan, Barchiesi, and others have explored further, the reader’s expectation that the single-mindedly elegiac poet of the Amores, Heroides, and Ars amatoria would also be writing his latest work in the same metre is self-reflexively corrected by an announcement of metrical and generic change at the very moment when what could have been a pentameter is revealed to be a second, stichic hexameter.22 The parenthesis (nam uos mutastis et illa) (‘for you [sc. gods] have changed them [sc. my undertakings] too’) appears after the penthemimeral word-break, springing on the reader the surprise that that word-break is not the anticipated medial diaeresis of the elegiac couplet’s pentameter but the strong caesura of a hexameter. The trick, once again, is far from pointless, as it signals the shift, not only in metre, but in genre, from elegy to epic, and in all the connotations that those genres carry in subject-matter and ethos, from the small-scale, humble, personal, and erotic to the grandiose, elevated, civic, and martial. As O’Hara memorably puts it, ‘this is the precise “Bob Dylan plugs in at Newport” moment of the poem, where Ovid becomes a writer of epic’.23

• 24 O’Hara 2007: 106.
• 25 On the self-conscious stereotype of ‘essential’ epic as ‘all-male, all-war, all the time’, see Hind (...)

10It reverses the metrical and generic shift of Amores 1.1.1–2, where the apparently epic hexameter arma graui numero uiolentaque bella parabam (‘Arms and violent wars in weighty metre I was preparing…’) is revealed to be the first half of an elegiac couplet. O’Hara is right to call it (with significant phrasing for this current discussion) ‘a similar trick’.24 However, there is an important distinction. The fiction of the Amores allows Ovid’s reader to be correct in reading Am.1.1.1 as a genuine epic hexameter, but one that the external intervention of Cupid distorts into an erotic elegiac couplet. In the Metamorphoses she is forced to reassess and correct her (mis)interpretation of Met. 1.1 and realize that it had been an epic hexameter all along. In Am. 1, the poem changes, in Met. 1 only the reader’s perception. Like the reader of Epode 2 who must retrospectively reinterpret the extended laus ruris as Alfius’ fantasy rather than the iambic persona’s beliefs, so the reader of Met. 1.2 must—on a rather smaller scale—retrospectively reinterpret Met. 1.1 as the incipit of an epic rather than the opening of an erotic elegy. While the reassessment of Met. 1.1 is on a small scale, it has massive implications for the fifteen books to come. On the surface, the surprise of the hexametric conclusion of Met. 1.2 sets up expectations of an essential epic to come.25 However, just as the reader of Epode 2 cannot unsee ‘Horace’ as the speaker of the laus ruris so that that misreading prompts her to detect parallels between him and Alfius, in the same way the Met. reader’s misreading of the poem’s opening as elegiac colours her reading of the rest of the poem—with considerable further encouragement—as distinctly elegiac. We are not quite at the point of the Derridean ‘all readings are misreadings’, but these tricks force the reader to blur the line between the two.

• 26 Nisbet and Hubbard 1970: 202–203.
• 27 Woodman 2018, with a convenient list of adherents to the communis opinio (though only listing comme (...)

12Horace’s own exploitation of παρὰ προσδοκίαν twists has already been amply illustrated by Epode 2, but it is worth noting their presence in the Odes and the complexity of their effect. 1.16 has long been the object of divergent interpretations. The traditional view is that Horace’s disquisition on anger in lines 4–21 relates to the invective iambi he wrote, as mentioned in lines 1–4, as part of a Stesichorean palinode for angrily attacking the fairer daughter of a fair mother. If so, then the exhortation to the addressee ‘compose your mind’ (compesce mentem) at 1.16.22 acts as a surprise revelation that the ‘grim rages’ (tristes…irae) belong, not to the poet, but to his beautiful addressee. Nisbet and Hubbard argue against this interpretation, on the debatable grounds that the suggestion that she dispose of the iambi proves that ‘the lady must already be angry’, and—tellingly—that ‘the abruptness of compesce mentem, which ex hypothesi is the first mention of the lady’s anger, is surely intolerable’.26 It is the παρὰ προσδοκίαν surprise that cannot be reconciled with preconceptions about Horatian decorum. Woodman has gone further in his attempt to smooth the flow of the ode, arguing—against the other assumption of the communis opinio—that it is the addressee and not Horace who is the author of the ‘slanderous iambi’ (criminosis…iambis, 2–3).27 Woodman’s case for the authorship of the iambi is convincing, but less so his assertion that the audience realizes it from the first stanza. Even if we are not prejudiced by apparent parallels Catullus 36, the invitation to ‘put an end to those scurrilous iambics however you wish, whether in the fire or in the Adriatic sea’ surely misleads the reader into assuming the iambi are written against rather than by the addressee. If we cease to force on the ode a logic, smoothness, and predictability that it does not possess, we can instead enjoy and unpack the implications of the surprise twist. Yes, the anger is the addressee’s and she is the author of the iambi, but the reader does not realize this until line 22. The familiar process of misreading the poem as a palinode, then correcting and rereading it as an exhortation makes especially potent use of the inability to unsee what has been mistakenly seen. For the misreading of the first twenty-one lines as referring to Horace reinforces his employment of himself (me quoque, ‘me too’, 22) as an exemplum of the perils of anger and iambography.

• 28 Harrison 2004.
• 29 Notably Fowler 1989, 1994, chapters in Roberts, Dunn, and Fowler 1997, and many studies of specific (...)
• 30 Generally: Grewing, Acosta-Hughes, and Kirichenko 2013. In the Odes: Oliensis 1998: 139, Harrison 2 (...)
• 31 Although the archilochian metre is epodic, Meineke’s law encourages the reader to divide the ode in (...)
• 32 Harrison 2004: 98.

13These twists and tricks occur within clearly demarcated poems, but the boundaries—real and apparent—of poems are a privileged site of such misdirections. Stephen Harrison has demonstrated Horace’s tendency to change direction in the middle of an ode.28 All such shifts challenge the reader’s conception of what constitutes a unified ode and what criteria determine its boundaries. However, this challenge is particularly strong in instances of false closure. Closure in Classical literature—including Horace’s Greek lyric models—has been much discussed in recent decades and the Odes have received their fair share of attention.29 False closure is another well-established trick in poets’ repertoire and one that undermines the reader’s (and the viewer’s) strong sense of when, where, and how a text—be it a play, an epic, a satire, or an ode—ends.30 In poetry books, it concomitantly problematizes her sense of when, where, and how a new poem begins. Probably the best-known example in the Odes, 1.28, the so-called Archytas Ode, combines false closure with a twist, doubly challenging the reader’s conception of the boundaries, identity, and nature of the ode. The stanza consisting of lines 17–2031 is strongly closural, as Harrison notes: ‘The idea in 19–20 that death takes all commonly appears at the end of sepulchral epigrams… The poem really ought to be complete’.32 This, however, is a trick, and it is combined with another trick: ‘The speaker of the poem, who up to now has looked like a traditional sepulchral epigrammatist addressing the dead person, now turns out to be dead himself’. False closure encourages the reader to see 1.28.20 as the end of a poem and the unexpected, apparently new speaker of 1.28.21 encourages her momentarily to see that line as the beginning of a separate poem. But the quoque, the metre, and the continuity of subject matter force her to revise this misreading and see continuity instead of—and in tension with—discontinuity. This is precisely the effect that false non-closure has in the odes discussed below, except that there it is continuity which is illusory and closure that is real.

• 33 See esp. the items in n. 2 above.
• 34 On these odes, see Griffith 2002: 72, Barber 2012.
• 35 The argument for division is Perotti 2016.

15Horace’s ludic manipulation of the boundaries between odes can be demonstrated by the fact that scholars have wished to combine traditionally separate poems into one and to divide what manuscripts and editors have transmitted as one poem into two.33 It is not important for our purposes—though of course it is for other reasons—whether such poems are or are not single odes. Rather it is the empirical evidence that skilled readers have responded to the (real or illusory) boundary between odes in diametrically opposite ways. This difference of opinion could be ascribed to the incompetence of one of the greatest and most meticulous of poets or to the wrong-headedness of some of the most learned and sensitive of critics, but it is surely more plausible to interpret it as a reasonable response to Horace’s sophisticated problematization of the boundaries between odes and of the nature and unity of the individual ode. Carm. 1.34 and 1.35 may indeed be a single ode with a delayed address to Fortuna, or two closely related but discrete odes, but we can surely not ascribe it to incompetence or inadvertence that they are both in alcaics and both deal with Fortuna, but shift radically from anecdotal to hymnic form, with a long-delayed address.34 The confusion caused is surely by design. Carm. 1.7 may be two odes (breaking Meineke’s law of quatrains) or there may be false closure and a radical shift in the middle at line 15.35 Again, Horace is challenging how we read the unity and division of poems and his (non-hypothetical) readers have demonstrably responded to that challenge. The question of how to divide Carm. 1.14–15 and 1.23–24 (as well as 1.5–6 and 4.12–13) is not left open in the same way, but the reader’s temporary uncertainty—and its thematic ramifications—fit perfectly into this wider play with the relationship between poems.

• 36 Van Sickle 1980, Krevans 1984, Gutzwiller 2005, Hutchinson 2008, Wulfram 2008, Prodi 2017, Battezza (...)
• 37 Notably Fuqua 1968, Dettmer 1983, Santirocco 1980, 1986, Porter 1987, Minarini 1989, Schwindt 2004, (...)

16Even where there is no possibility of juxtaposed poems’ being read as a single poem, the relationship between such odes is important for the present discussion. The careful design of the Hellenistic Gedichtbuch, both original compositions and editions of earlier poems, including archaic lyric, and its reception in Roman poetry books has been much discussed.36 In the Odes, as well as more elaborate and large-scale structures, studies of the juxtaposition of individual poems and sequences of related odes has revealed complex interplays of continuity and discontinuity, responsion and contrast.37 It is no coincidence that both the pairs of odes which I argue feature false non-closure have been identified as prime examples of such significant juxtaposition. The diachronic reader of the Odes already experiences each poem sequentially as a complex combination of similarity and contrast, continuity and discontinuity. In the following two cases, that experience is further intensified and complexified by the phenomenon of false non-closure, the illusion of continuity, and the process of dispelling that illusion.

• 38 The text of Horace is Shackleton Bailey 2001, but with consonantal u; translations are from West 19 (...)

17The first instance that we shall examine—though not the first in the Gedichtbuch—comes at the transition from 1.14, the so-called ‘ship ode’, to 1.15, the Bacchylidean prophecy of Nereus to Paris (Hor. Carm. 1.14.17–20, 1.15.1–5):38

• 39 ‘Die Sätze drängen über die Strophen- und Versgrenzen, die Strophen selbst bilden keine Einheiten, (...)
• 40 Carrubba 2003: 612.
• 41 Quinn 1980: 152.

18It is not only the opening pair of asclepiads that trick the reader into thinking that the new ode is a continuation of the previous ode. The first four stanzas of 1.14 are a maelstrom of mimetically tempestuous enjambment, where ‘the sentences push across the stanza and verse boundaries, the stanzas themselves do not form units, new sentences begin in their middle. Even sentence structure and word order serve in the poem as a means of increasing the turbulence.’39 By strongly marked contrast, line 16 brings sentence, stanza, and (illusorily) ode to a convergent climax. The apparent closure is marked, not only linguistically and metrically, but thematically and structurally: the ode ‘could well have ended with line 16 and left the reader satisfied with its completeness: a Horatian ode of tripartite structure whose beginning and end correspond to one another and appropriately frame a larger central thematic’.40 Ode 1.14 thus presents a classic case of false closure. As such, it implants in the reader an attitude of ‘fool me once, shame on you’ and primes her to be open to the possibility that the next closural moment—especially one as ‘oddly anticlimactic’ as ‘the mildly worded warning against the deceptive waters of the Cyclades’41—may also prove to be false.

• 42 Woodman 2020: 280 notes the shared reference to ships but dismisses it as ‘an additional complicati (...)

19This negative suspicion that 1.14 may not really have concluded is reinforced by positive hints that the subsequent two lines are a continuation of the ode. On the simplest level, there is clear continuity of subject-matter. Ode 1.14 is, of course, about a ship (at least on the literal level), opening with the address o nauis, and 1.15 opens with an antonomastic reference to Paris dragging someone in a (plural for singular) ship or ships (nauibus).42 The sharp asyndeton and shift from the present to the mythological past could encourage the reader to sense a break, but even this fits within Horace’s programme of misdirection and false continuity.

• 43 Venus: Anderson 1966: 66–67, Traill 1979, Kruschwitz 2007: 172–173. Achilles: Felgentreu 2011: 332– (...)
• 44 Fair wind: Cypria fr. 14 West apud Hdt. 2.117; storm and Sidon: Procl. Chr. Cypria arg. 2, Apollod. (...)

20The specificity of the Cyclades as the islands whose waters are to be avoided at the end of 1.14 has puzzled critics from Porphyrio onwards, and Nisbet and Hubbard (ad loc.) largely content themselves with labelling it ‘a little surprising’ and ‘the mark of a literary man’. The islands may be chosen as an allegorical warning for Brutus simply because ‘narrower seas are more dangerous’ (maria … angustiora periculosiora sunt, Porph. ad loc.), or because their associations with the cult of Venus, more explicitly alluded to at 3.28.14, fit an erotic interpretation of the ode, or, in a subtle connection with 1.5, to evoke Achilles’ concealment on Scyros.43 Whatever interpretation—if any—the reader ultimately privileges, the reference to the Cyclades actually appears (deceptively) less surprising when immediately followed by a description of the abduction of Helen. The route that Paris took from Sparta to Troy was much contested in antiquity. However, whether he completed the voyage in three days with a fair wind and calm sea, or visited (deliberately or storm-driven) any or all of Egypt, Phoenicia, and Cyprus, in any of these scenarios, after sailing south-south-east from Sparta and rounding Cape Malea, his route north-east to Troy or east across the Cretan Sea must have taken him through or at least towards the Cyclades.44 If it was there that Hera sent a storm against him, then his relevance to the warning of 1.14.19–20 is even clearer. If anything, mention of Paris promises (again, misleadingly) to make the Cyclades more comprehensible.

• 45 On Horatian mythic exempla: Pöschl 1981, Basta Donzelli 1994, Lowrie 1997, Breuer 2008: 195–197.

21There remains the abrupt shift from present-tense, personal address to third-person mythological narrative, which could signal the genuine discontinuity from one ode to another. However, even this parallels one of Horace’s established patterns of introducing a mythological exemplum.45 Ode 1.7, another which, as we have seen, plays with ideas of continuity and discontinuity, offers a strikingly close parallel in its introduction of the Teucer exemplum (1.7.15–22):

• 46 Perotti 2016.
• 47 Of course, since the Teucer exemplum begins after the penthemimeral caesura of the first archilochi (...)

22As with the close of 1.14, lines 15–21 of 1.7 are preceded by a strong sense of false closure—indeed so strong that some have taken it as genuine closure46—and constitute a single sentence preceding the exemplum.47 As at 1.15.1, the main character of the exemplum (pastor~Teucer) is introduced in asyndeton and brought forward to the start of the temporal clause in hyperbaton. As at 1.15.2, that temporal clause consists of cum with the imperfect subjunctive (cum traheret~cum fugeret). Although it does not manifest itself in 1.15 until after the metre has disambiguated that a new poem has begun, both exempla consist of a brief, scene-setting preamble and an extended speech, which concludes the ode without the resumption of either the inner or the outer frame. On the basis both of generic formal expectations and of the specific precedent she had read only a little earlier in Odes 1, Horace’s reader is not only unfazed by the asyndetic introduction of the pastor, but positively encouraged to interpret him as the subject of a mythological exemplum that will continue and conclude the current ode.

• 48 Barber 2012: 507.
• 49 Fraenkel 1957: 188.
• 50 Parade of lyric predecessors: Lowrie 1995.
• 51 Bradshaw 2008: 37 argues against the uniqueness of 1.15’s form, claiming it as ‘a dramatic scene, a (...)

23Misdirecting the reader to understand Paris as a mythological exemplum is not merely a mechanism of this benign and pointed trick. It is a large part of its point. The trick forces the reader repeatedly to revise and reassess her interpretation of 1.15, of 1.14, and of the relationship between the two. 1.15 is almost unique in the Horatian lyric corpus, as a mythological narrative with no explicit reference to the poet’s contemporary personal or political world. Formally, it is one of only two poems in Odes 1 in which, in Barber’s words, ‘the speaker, employing his customary first person, gives no hint of the direction in which he is speaking.’48 Barber is actually describing 1.34, using this anomaly as part of his argument that it and 1.35 are a single poem, with the addressee (o diua, gratum quae regis Antium, ‘O goddess, who rule over lovely Antium’) ‘delayed’ until 1.35.1, which would be line 17 of the combined ode. Yet he curiously overlooks 1.15, which also has no addressee but further lacks a comparable means of providing one. Of course, we must beware of making Horace’s infinite variety blandly homogenized and of believing with Fraenkel’s straw men that ‘no form or setting of a poem can be tolerated if there exists only one instance of it.’49 However, that does not mean that Horace cannot play with his reader’s expectations about generic norms. If that reader is (mis)led by the mechanisms we have seen into (initially) reading 1.15 as a mythological exemplum within 1.14, then that will also satisfy her normative expectations that it is a component part of an ode that reassuringly already possesses an addressee (o nauis) and an explicit, if elusive, connection to Horace’s contemporary world (nuper sollicitum quae mihi taedium…, ‘Not long ago you were a worry and a weariness for me…’). When she reaches line 3, she will be forced to reassess, to realize that 1.15 is a separate poem, that 1.15 can be a separate poem, that a Horatian ode can be a free-standing, mythological narrative without addressee or contemporary reference, perhaps that, in this parade of lyric predecessors, Horace can be the Roman Bacchylides as well as the Roman Alcaeus.50 Yet she cannot unsee the connection, the figurative relationship of Paris to the ship. If the Paris myth is not a mythological exemplum within 1.14, then it is a freestanding exemplum within book 1 and within the reader’s wider, contemporary frame of reference, in other words, an allegory, but one still closely bound to 1.14.51 Again, in deference to Fraenkel, this is not the only possible interpretation of 1.15, but it is one which Horace’s structural sleight-of-hand encourages the reader to entertain.

• 52 Draper 2017: 656.
• 53 Santirocco 1986: 49.

24The reader’s journey of interpreting the Paris narrative first as an exemplum within 1.14, then as a completely separate poem, then as an independent but closely related allegory inevitably impacts upon her reading, not only of 1.15 itself, but of the relationship between the two odes. Misreading two poems as one and then rereading them as two establishes in the reader’s mind a complex and shifting interplay of continuity and discontinuity, similarity and difference. Quite apart from the blurring of the metrical lines between them, critics have regularly noted the close relationship of 1.14 to 1.15, ‘another poem portraying a dangerous sea voyage that invites allegorical interpretation’.52 Indeed, Santirocco, pairing the odes as a particularly clear example of the significant juxtaposition of odes, goes as far as to write that the ‘relationship between these two poems is so close … that any reasonable allegorical interpretation of the first will harmonize with the second.’53 He shows how both can be read as political allegories, or 1.14 as an erotic allegory corresponding the more literal erotic narrative of 1.15, or 1.14 as a poetic allegory of epic endeavour, of which 1.15 is an illustrative example. These harmonizings of interpretations are encouraged by the reader’s initial belief that they were a single poem.

• 54 E.g. Commager 1962: 215–219, Kraggerud 1978, West 1995: 76–77, Cresci Marrone 1999.
• 55 Erotic readings of 1.14: Anderson 1966, Traill 1979, Woodman 1980, Jocelyn 1982, Knorr 2006, Krusch (...)
• 56 Metapoetic readings of 1.15: Zumwalt 1977–78, Davis 1989.

25Nevertheless, her realization that they are in fact two poems, that there is discontinuity between them, amplifies the dissonances and mismatches. Yes, both poems feature ships, but the nauis that is the central subject of 1.14, described in immense detail and susceptible of multiple allegorical interpretations, is very different from Paris and Helen’s incidental mode of transport, granted the evocative but solitary epithet Ideaeus, mentioned once in line 1 and never referred to again, and very hard to allegorize. Yes, both poems are susceptible of political allegoresis, but the ship of state for which Horace expresses such concern and affection is an awkward match for any polity represented by Antonius and Cleopatra, the most frequently touted palimpsests for Paris and Helen.54 Yes, both poems can be read as erotic, but if 1.14 is primarily erotic, expressing concern for an old flame, then the very lack of emphasis on love—or even on Helen herself—in 1.15 marks it as a very different poem.55 If 1.14 is read as a metapoetic reflection on a generic aspiration that 1.15 goes on to fulfil, then that in itself sets the latter apart as a related but distinct poem.56

• 57 It is worth noting, though it does not affect the current argument in any way, that the final glyco (...)
• 58 A third asclepiadic line could also fit within a system of stichic asclepiads (the ‘first asclepiad (...)
• 59 As noted above, the nature of the word-break also changes from (apparent) elegiac diaeresis to hexa (...)

27Before turning to our second example, it is worth looking a little more closely at the precise moment when the reader realizes that she is now reading a new poem. Woodman’s formulation ‘it is … not until readers reach the third line … that they realize that the meter has changed’ is true as far as it goes. However, it could reasonably be interpreted as suggesting that the moment of realization would occur at the very start of the third line, which is not the case. Both the asclepiad and the pherecratean (the third line of the third asclepiad) begin identically with a spondaic base and a choriamb.57 It is only when that choriamb is followed, not by the anceps which closes the pherecratean, but by a second choriamb, that the reader realizes that the line is another asclepiad and hence the stanza is a second asclepiad.58 She may already be a little surprised to find a word-break after the first choriamb—mandatory in the asclepiad but rare in the pherecratean—but she has been conditioned to expect even this by the very first pherecratean of 1.14: portum! nonne uides ut (‘…harbour! Do you not see how…’, 1.14.3). This delay of the moment of recognition is closely comparable to that at the opening of Ovid’s Metamorphoses, discussed above. The reader of Met. 1.2 does not know that she is reading, not a pentameter, but a hexameter from the start of the line but only after the penthemimeral word-break.59 As we saw, Ovid self-consciously marks that moment of realization that the metre (and genre) has changed with a parenthetic reference to the gods’ having changed his undertakings too (nam uos mutastis et illa). Does Horace make a comparable move after the first choriamb of Carm. 1.15.3?

• 60 ‘eine kühne Metapher: was ‘verschüttet’ wird, kann sich nicht regen.’ Kießling-Heinze 1930: ad loc.

28The key line is ῑngrᾱtō cĕlĕrῑs || ōbrŭĭt ōtĭō (‘subdued the swift [winds] with unwelcome calm’) and the key word is obruit. Admittedly, we are not dealing here with the overtly metapoetic conceit of gods changing a poet’s undertakings, but the image of ‘overwhelming’, especially overwhelming a potentially metapoetic sea voyage, is an appropriate one to mark the moment when a third asclepiad becomes a second and one poem becomes two. Except it is not the ship that is being overwhelmed by the winds, but the winds by Nereus. Commentators generally seem curiously unperturbed by this surprising use of obruit. Only Kießling and Heinze note that it is ‘a bold metaphor’, rationalizing it with the explanation that ‘what is “buried” cannot move’.60 A clear indicator of its boldness is the fact that OLD does not feel able to classify this instance of obruo under any of its senses, but places it, pis aller, alongside passages where it means ‘to overcome, overpower (with sleep; with wine, etc.)’ (4c), offering an aporetic and almost apologetic cf.

• 61 Her. 7.78, Met. 9.594, 11.569, Tr. 1.2.106. 5.11.13, Pont. 3.6.29, 4.8.28. Propertius employs a for (...)

29The usage is the more startling because obruo, in comparable contexts, is generally used of winds overwhelming ships or the sailors in them: Virgil’s Juno commands Aeolus to let loose the winds’ violence ‘and sink and overwhelm the ships’ (incute uim uentis submersasque obrue puppis, Verg. A. 1.69), his Aeneas sees his drowned comrades whom ‘the South Wind overwhelmed’ (obruit Auster, 6.336), while a storm (procella) repeatedly overwhelms its Ovidian victims (Ars 3.584, F. 1.488, Tr. 3.2.26, Pont. 2.7.54). Ovid often has the stormy sea overwhelm sailors and is clear that the agency is shared between winds and waves, since he aetiologizes the dedication of a temple to personified Tempestas as occurring ‘when [L. Cornelius Scipio’s] fleet was almost overwhelmed by the Corsican waters’ (cum paene est Corsis obruta classis aquis. F. 6.194).61 Indeed, his Penelope expresses a wish that Paris himself had been overwhelmed by the crazed waters, albeit on the voyage to rather than from Sparta (o utinam tum, cum Lacedaemona classe petebat, | obrutus insanis esset adulter aquis! Her. 1.5–6, ‘If only the adulterer had been overwhelmed by the crazed waters then, when he was making for Sparta with his fleet!’).

• 62 Since the fearful man who ‘is overwhelmed by [the pursuit of?] wealth’ (obruitur re) at Epist. 1.16 (...)

30Most significantly of all, Horace himself uses obruo just a little later in Odes 1, in the same form and in an equivalent sedes within a dactylic tetrameter (following word break after the sixth syllable and preceding a final spondee), to describe how ‘the South Wind, raging companion of the setting Orion overwhelmed’ a sailor ‘with the Illyrian waves’ (me quoque deuexi rapidus comes Orionis | Illyricis Notus obruit undis, 1.28.21–2). These are the only two occurrences of obruo in the Odes, and there is only one other in the entire corpus.62 Of course, our diachronic reader, who alone can think that 1.15 is a continuation of 1.14, has not yet read 1.28.22 and so cannot use the sense of obruo there to shape her reading of 1.15.3. Nevertheless, at the very least, even as an unmarked parallel, 1.28.22 is a clear indicator of Horatian usage, especially since it harmonizes so closely with wider poetic usage, in sharp contrast to the startling anomaly of 1.15.3. Among the endless interconnections within the Gedichtbuch, however, it would be rash to rule out a more marked intratextual allusion between the moment of false non-closure when 1.15 is revealed to be a new poem and the moment following the false closure when the speaker of 1.28 is revealed to be an unburied corpse.

• 63 ‘Das otium, bei Sturm vom Schiffer ersehnt (II 16, 1), ist hier unerwünscht ingratum, dem Wind wie (...)

31It is not only the presence of the word obruit that misleads the reader into thinking that Paris’ ship has been hit by a storm. As we have seen, most of the diverse traditions about the voyage from Sparta to Troy involved a storm and, although the Cypria’s smooth three-day journey and Paris’ evasion of the Dioscuri offered alternative scenarios, the dominant version would further reinforce the reader’s inclination to take obruit as ‘overwhelm’ and to expect a wind or even Juno as its subject. Finally, the reader who thinks that she is reading a continuation of—and mythological exemplum within—1.14 would expect a narrative that picks up that ode’s anxious references to the threat of winds (Africo, 5; uentis, 15), waves (fluctus, 2; aequor, 10), and the dangerous waters between the Cyclades. She is expecting something like (purely exempli gratia) ingrato celeris obruit aequore | puppis… | Auster (‘the South Wind overwhelmed the swift ships with the unwelcome sea’). The surprise, verging on paradox, of winds, not overwhelming ships with billows, but being overwhelmed by of all things calm (otio), is compounded by a similar paradox in the callida iunctura of ingrato and otio. As Kießling and Heinze again acutely point out, ‘otium, longed for by the sailor in the storm (Carm. 2.16.1), is here unwelcome, ingratum, to the wind as to the sailor’.63 The paradox of obruit, the word which reveals that this is a new poem, perfectly dramatizes and expressed the tension between continuity and discontinuity in these two odes. Its expected sense represents shipwreck, death, closure, discontinuity at the same time as it indicates continuity with the prominence of those motifs in 1.14. Its unexpected, actual sense represents calm, peace, stillness, the absence of violent rupture, but that very absence marks a rupture with the subject-matter of 1.14 and with 1.14 itself. Even on the level of the individual line, obruit does not overwhelm the metre by bringing the (apparent) pherecratean to its immediate close, but rather extends it to the end of the asclepiad. However, the pacific import of obruit is itself, if not quite illusory, at least strictly limited: the peaceful calm of otium, a paradoxically unwelcome calm, serves only to provide the opportunity for Nereus to predict the war that will destroy Paris and Troy. Horace refrains from using storm imagery to depict the Greek onslaught, but the relationship between the literal war of 1.15 and the allegorical tempest of 1.14 maintains thematic continuity between the two poems, even at the moment when obruit reveals their discontinuity.

• 64 Ring composition: Estevez 1979-80, Ronnick 1993: 157. Cf. Roche 2013: 349, ‘the first two words of (...)
• 65 Feeney 1991: 145–6.
• 66 For this principle of mimetic word-order, with many (mainly Ovidian) examples, see Lateiner 1990, e (...)

33Odes 1.23 does not share with 1.14 the boundary-blurring enjambment across stanzas that makes the strong pause before the final stanza feel so much like false closure and hence primes the reader to be skeptical about the genuine closure at the end of that stanza. Instead, it consists of three self-contained, end-stopped quatrains, each of which feels neither strongly open nor closural, so that the reader is predisposed to be surprised neither if the poem ends at 1.23.12 nor if it continues into a fourth stanza. In this way, she is even more sensitive to the ambiguity of the transition between odes and even more alert to the appropriateness of both continuity and discontinuity. It is true that the third stanza of 1.23 includes several closural motifs, such as the words tandem and desine, and ring composition with the ode’s opening.64 However, Horace’s reader is by now well familiar with the possibility of false closure and 1.24.1 seems positively to encourage such a reading by questioning, conversely, the possibility of genuine closure. Virgil’s Jupiter self-consciously comments to his wife Juno, not only on the teleological fulfilment of Aeneas’ manifest destiny, but on Aeneid 12’s refusal, through a series of false closures, to produce an ending to the epic: quae iam finis erit, coniunx? (‘What end will there be now, wife?’ Aen. 12.793).65 Similarly, Horace appears to question whether there can be any limit, any end-point, any closure (quis…sit…modus), not only to desire (desiderio), but to his expression of it in 1.23, with the problematized modus mimetically positioned at line-end.66 pudor too was used earlier in the book to express the sense of restraint that Horace claims prevents his writing about Augustus’ and Agrippa’s achievements (dum Pudor | imbellisque lyrae Musa potens uetat…, ‘While Modesty and the Muse who commands the unwarlike lyre forbid us…’ 1.6.9–10). Such metapoetic restraint from writing, from continuing the love poem to Chloe, is—or appears to be—absent here.

• 67 All studies of 1.23 inevitably discuss its eroticism, but see Bannon 1993 for the pervasiveness of (...)
• 68 This follows Nisbet’s (1978: 92) reading of Catul. 2.5, comparing Anacr. fr. 444 PMG: πόθῳ στίλβων (...)

34Of course, as the reader will come to realize, the desiderium is not for Chloe (or Chloe’s for a uir), and neither is it (on the surface) erotic. The false non-closure of 1.14, which opened o nauis, is reinforced by the presence of actual ships (nauibus) in the first line of 1.15, even if the status and nature of all the ships involved remains contested and ultimately unresolved. The false non-closure of 1.23 appears to be reinforced by a similar continuity of theme with the first line of 1.24. The erotic desire that, though never quite explicit, is unmistakably the dominant motif of the first ode, through its imagery, its Anacreontic intertext, and the unambiguous implications of the phrase temptestiua … uiro, seems to be made explicit in the desiderium of 1.24.1 and its object, ‘so dear a person’ (tam cari capitis) in line 2.67 However, while Paris’ ships remain ships, the process of misreading, correcting, and retrospectively re-reading turns desiderium from erotic desire into the grief-fuelled longing for a friend. It is not the desire for Catullus with which Lesbia gleams (desiderio meo nitenti, ‘gleaming with desire for me’, 2.5) nor that which Horace himself feels for the ship, whether it be literal lust for a metaphorical woman or metaphorical lust for the personified state (nunc desiderium, ‘now a [source of] desire’, C. 1.14.18).68 When the reader reaches the word lugubris at the end of 1.24.2—a positioning to which we shall return—she will begin to realize that the theme has shifted from the erotic to the funerary, and when she reaches the second choriamb of line 3 with liquidam, she will understand that this is because a new poem has begun. As she reassesses what she has already read in this light, she will reinterpret desiderium, not as the sensual desire for the budding, nubile Chloe or the vivaciously gleaming Lesbia, but as the bereft yearning for the incomparable, irreplaceable Quintilius or Calvus’ Quintilia (quo desiderio, ‘the longing which…’, Catul. 96.3), snatched by untimely death. Realization of the change of metre encourages an emphasis on discontinuity and contrast, from the level of the whole ode to that of the individual word.

• 69 Esp. Putnam 1993: 124–126, 2006: 101–106, Lida-Tarán 2014.

35Nevertheless, the reader cannot totally unsee the erotic connotations of desiderium, and the encouragement to detect continuity between the two poems is reinforced by several factors. Catullus 96 is not merely an arbitrary parallel for the grieving sense of desiderium, but an important, privileged model and intertext for Odes 1.24.69 Catullus’ consolatio to Calvus concerns the latter’s grief for and ἐπικήδειον (‘funeral lament’) to his beloved Quintilia. Whether wife or lover, she is the object of a complex and perhaps indivisible blend of grief and erotic desire. Indeed, Calvus’ desiderium stands poised in word-order, syntax, and sense between pain and love (Catul. 96.1–4):

• 70 Dating individual Horatian odes is a fraught and largely futile exercise, but the generally accepte (...)
• 71 lumina sis oculis etiam bonus Ancu’ reliquit, | qui melior multis quam tu fuit, improbe, rebus (‘Ev (...)
• 72 Cf. Syndikus 1972–3: 236 ‘Aber für Horaz wie ja wohl für Virgil war die alte Sage nicht anders als (...)

37The other cue to read an erotic dimension into 1.24 and hence continuity with 1.23 also derives from a privileged intertext. Horace gently insists that Quintilius will not come back to life, even if Virgil were to surpass Orpheus with the power of his song: quid si Threicio blandius Orpheo | auditam moderere arboribus fidem? (‘What if you were to tune a sweeter lyre than Thracian Orpheus and trees came to listen?’ 1.24.13–14). Though some dissent on the stereotypical grounds of Horatian tact, it is widely acknowledged that this allusion to Orpheus as a comparand for recalling the dead to life in an ode addressed to Virgil must allude to Proteus’ narrative in book four of the latter’s recently published Georgics.70 Yet the allusion is a subtle and complex one. Horace does not mention Eurydice but leaves the reader to supply the unstated fourth term in the double comparison: Virgil:Quintilius~Orpheus:Eurydice. Even with this fourth term supplied, the comparison is a slippery one. The reader expects Horace’s argument to be a fortiori, along the lines of Lucretian ‘even good Ancus died, who was better than you in many ways, wretch’:71 even Orpheus, who was of course an even greater poet than you, Virgil, could not resurrect the dead, so how much less will you be able to. But Horace rarely does the expected. Here he boldly—and flatteringly—allows for the possibility that Virgil could surpass Orpheus, matching his achievement of the adynaton that is moving trees, but doing so with an even more persuasive ‘lyre’. But even then, Virgil could not—surely he could not, though the num-question does not shut down the possibility absolutely—return lifeblood to an insubstantial shade. The argument in fact rests on a different, unstated a fortiori position: if a superior Virgil (surely) could not raise the dead, then the inferior Orpheus must axiomatically have been unable to do so. Yet bringing Eurydice back from the dead was precisely what the Virgilian Orpheus famously could do (albeit abortively). Is Horace gently—one could hardly say polemically—correcting Virgil? ‘Such wish-fulfilment fantasies are all very well in their place but they are no way to come to terms with the emotions of bereavement in the real world.’72 Or is he drawing the parallel with the Virgilian Orpheus’ ultimate failure, his inability to return his beloved to life? ‘Yes, as you yourself so achingly depicted, the power of song can work wonders but even it is finally incapable of bringing back those we love.’ Horace leaves his reader to decide between these alternatives or to keep both in tension, but every possibility adds to the strong parallelism between Orpheus’ emotions towards his dead wife, Eurydice, and Virgil’s towards the dead Quintilius. The intertextuality encourages the reader to colour Virgil’s desiderium for Quintilius with the erotic as well as the grieving quality of Orpheus’ longing for Eurydice, just as of Calvus’ for Quintilia. If she has followed the cues to misread 1.24 as a continuation of the erotic 1.23, her inability to unsee that continuity can only be compounded by these intertextual pressures.

• 73 Nadeau 2008: 123–134; ‘more or less arbitrarily’, Holzberg 2010: 126; ‘not all will agree … in seei (...)
• 74 On responses to Nisus and Euryalus’ homoeroticism, see esp. Makowski 1989.

38How then might the reader interpret this erotic colour in 1.24? Nadeau takes the relationship between Virgil and Quintilius as that of ἐραστής (‘[older, active] lover’) and ἐρώμενος (‘[younger, passive] beloved’), though his reading has not gained many adherents.73 There is undeniably a risk of producing naïve or (consciously or unconsciously) prejudiced interpretations that are oblivious to or wilfully suppress homoerotic elements in Graeco-Roman texts. However, it can be equally naïve and tendentious to impose a sexual interpretation—whether homo- or hetero- —on every interpersonal relationship and especially to read subtle imagery in an overliteral manner. We may be rightly skeptical and perhaps even suspicious of those who insist that Virgil’s Nisus and Euryalus are ‘just good friends’, but that does not necessarily mean that every friendship in Classical literature must be read as an actual or potential sexual relationship.74 It is impossible to prove a negative and hence categorically to refute Nadeau’s thesis, but neither is there a great deal of evidence in its favour and what there is can be more plausibly and (arguably) more interestingly interpreted in other ways. Nadeau’s suggestion stands as clear evidence for the strength of the erotic imagery deployed in depicting Virgil’s feelings for Quintilius, strong enough to suggest such a literal interpretation to a Horatian scholar: the Orpheus-Eurydice parallel that he notes, to which we can add the Calvus-Quintilia parallel, and also the (false) sense of continuity with the erotic 1.23. However, that imagery can be read as exactly that, imagery, paralleling the strength of Virgil’s love for his friend with that of a lover for a beloved, but without suggesting that they were literal lovers. Its very doubleness—like erotic love, but not erotic love—is reinforced by the reader’s contradictory experience of the transition from 1.23 to 1.24, misreading them as a single, erotic ode, then revising that reading to accommodate the discontinuity between two discrete poems, one erotic, one not, but never quite able to unsee the erotic traces that prompted the original misreading.

• 75 Santirocco 1986: 59. Cf. Fuqua 1968: 45: ‘Carm. 1. 23–25 may be considered a carefully structured s (...)
• 76 This reading takes the ode’s depiction of time at face value, but the interplay of form and theme i (...)

39The reader’s persistent (mis)perception of continuity between the two odes colours her view of 1.23 as well as of 1.24. If her memory of Chloe colours Virgil’s grief for Quintilius with the intensity of erotic desire, then Horace’s consolatory reflections on the transience of human life and the necessity of coming to terms with mutability expands the seducer’s insistence on the young girl’s sexual maturation into a universal truth about the overwhelming importance of timeliness. This juxtaposition of the moment of sexual prime and the eternity of death is reinforced by their blending in 1.25 in the figure of Lydia, whose moment of erotic desirability has passed and who is now descending—unwillingly and resistently—into old age and death. As Santirocco puts it, ‘[w]hen read in conjunction with the two poems that surround it… C. 1.24 takes on added meaning and intensity. The coyness of Chloe, the excessive mourning of Vergil, and the haughtiness of Lydia are now revealed as attempts to cling to the past, to deny the passage of time.’75 Yet that very passage of time partakes of a doubleness, a tension between continuity and discontinuity that is mimetically and vividly expressed by the formal ambiguity as to whether 1.24.1–2 constitutes a continuation of or disjuncture from 1.23. Many of the closural features of the latter’s final stanza relate to and dramatize the necessary end of Chloe’s girlhood. She must ‘finally stop’ (tandem desine) the pattern of behaviour that has been the focus of the entire ode, bringing the poem and her adolescence to a simultaneous end, rather than unnaturally continuing both, in perpetual immaturity, like a Hippolytus or an Atalanta.76 However, her hoped-for recognition that she is ‘at the right time for a man’ (tempestiua uiro) also entails a willingness to continue along the natural course of life rather than to stop, frozen at the dead end of an eternal virginity. Chloe’s maturation and the ode must continue beyond her girlish trembling in the wood, into a life-stage and an (apparent continuation of the) ode, where there can be no limit (modus) to or compunction (pudor) about desire (desiderium).

• 78 As noted above with respect to 1.15.3, the asclepiad and pherecratean are identical (spondee plus c (...)
• 79 It is intriguing that the other instance—eheu Postume Postume (‘Alas, Postumus, Postumus,’) at 2.14 (...)

44Heyworth’s point is, of course, that adjacent odes in the same metre (often) are a single ode and so the reader’s fusion of them is not—as focalized through ‘[t]hose who maintain the traditional numeration’—a ‘false amalgamation’. However, it rests on the premiss that the preventative measure of using an addressee would forestall the reader’s error. Reading 4.12 and 4.13 as a single poem would be a ‘false amalgamation’, even if would be revealed as such on purely metrical grounds by the beginning or at the latest the end of the third line.78 Horace does prevent such a (temporary) false amalgamation precisely by using a distinct addressee and does so with striking and almost unique emphasis. Lyce is the second word of the ode, leaving virtually no time for the reader even to entertain the notion that 4.13.1 is a continuation of 4.12. As if self-consciously to remove all doubt, Horace repeats the address—and, for good measure, the verb preceding it—in the second line, one of only two occasions in all the odes where the addressee’s name is given twice.79 Whatever Horace’s purpose was in following a second asclepiad with a third only here in his entire lyric oeuvre, it was not to trick the reader with an effect of false non-closure. Yet his careful, emphatic, and perhaps self-conscious forestalling of this possibility is in itself a signal that he acknowledged its existence, either simply on general principles, or perhaps also because he had set the precedent in 1.14–15 and 1.23–24. Virgil and Lyce are very much the exception that proves the rule.

• 80 Harrison 2017: 115.

46We have already seen, with 1.23 and 1.24, that the blurring of the boundary between an erotic and a homosocial ode can generate in the reader’s mind interesting responses to both poems. Here, however, there is no desiderium or carum caput to trick the reader into misreading the recusatio to Agrippa as a continuation of the detached reflection on Pyrrha’s charms. It is just conceivable that the addressee—not identified as Agrippa until line five—could be Pyrrha, depicted as the conquering beloved-as-general of militia amoris (‘the soldiery of love’), like Barine in 2.8, for whom, in Harrison’s gloss, ‘a whole host of young men is growing up’ (pubes tibi crescit omnis, 2.8.17).80 Indeed, Horace will, later in this ode, define the subject matter of his lyric as proelia uirginum (‘battles of maidens’, 1.6.17), in contrast to the literal epic battles to which Varius is better suited. However, there is nothing in 1.6.1–2 that would cue the diachronic reader to such a metaphorical interpretation. The shift from the imagery of the sea of love and the shipwrecked lover to the beloved as general is harsh in the extreme. We have, of course, seen that changes of direction are characteristic of the Odes, but—quite apart from subjective judgments as to when such a swerve is jarring rather than merely surprising—if the poet is to trick the reader with false non-closure, he must encourage her to do so with illusory continuity rather than betray the truth with a disjuncture that is evident as well as actual. The reader would have to work hard to make 1.6.1–2 a change of direction in the Pyrrha ode and it is much easier for her to read it as the start of a new poem that it actually is.

• 81 Unlike auctor, which can be feminine (e.g. Ov. Fast. 5.192), uictor is always masculine. Notably, H (...)

47Alongside the absence of positive reasons to misread continuity, there are the strong negatives discouraging it. The introduction of the notion of writing (scriberis)—and that without an emphatic tu to mark the change of focus—strongly suggests a new topic, especially when the writer is specified as Varius Rufus. Mention of the genre of Homeric epic (Maeonii carminis) marks the subject matter as literal warfare rather than the metaphorical militia amoris of lyric or elegy. However, by this point, the reader already knows categorically that the addressee is not Pyrrha (and hence this is not the same ode) because of the masculine uictor at the start of line 2.81 As with 4.12 and 4.13, there is no danger of the reader’s being tricked into believing that the second poem is a continuation of the first, but in contrast to the anaphoric address of Lyce, there is no emphatic or potentially self-conscious underscoring of the discontinuity. This requires some explanation, if we are to believe that the other examples in book 1 are deliberate.

• 82 Woodman 2020: 280.

49Of the four places in the Odes where Woodman’s ‘cruel and pointless trick’ could be played, in two Horace carefully forestalls any confusion, and in the other two he fully exploits the trick’s potential to enrich the reader’s response to each ode, to the relationship between them, and to the very act of reading. The cases of Pyrrha and Lyce suggest that Woodman is right to insist that ‘readers needed more than meter to tell them that one poem had finished and another one was about to begin’.82 However, the care with which Horace tells those readers that 1.5 and 4.12 have finished and 1.6 and 4.13 have begun, as well as the blurring of the boundaries at 1.14–15 and 1.23–24, suggest that Heyworth is also right to insist that metre is a powerful indicator of continuity or discontinuity between poems. Honours are even on the question of segmentation, but the more interesting question is not where you draw the line, but how and why Horace manipulates your decision.