1. ANTONIUS AND RECUSATIO

Odes 4.2 ostensibly looks forward to two public events lying at some indeterminate point in the future, Augustus' return from campaign in Gaul, and a triumph over the Sygambri. The celebrations anticipated for these occasions frame the second half of the ode; but they do not supply its dramatic setting or timing, and the latter is evidently associated with the period following Augustus' departure for Gaul in summer 16 b.c., or at any rate with a time when the Sygambri were felt still to present an armed threat.Footnote ² The dramatic date need not however be identical with the time of writing, and it should not be assumed that Odes 4.2 was composed in 16. Indeed, as will be seen (below, Section 6), it is most likely to be the retrospective product of a later period, around the time of Augustus' eventual return in July 13. On this view, the poem does not look forward to future events, but back to the time of Augustus' departure.

The dramatic setting presented in Odes 4.2, whatever it was (below, Section 7), brings together the poet and Iullus Antonius. The son of two enemies of Augustus, the triumvir and his third wife Fulvia, Antonius was the husband of Augustus' niece Marcella, an acknowledged favourite of the Princeps, and ‘among the most resplendent nobiles not yet of consular years’.Footnote ³ Born in 43 (the praenomen advertised loyalty to the dead Caesar), he was praetor in 13 at the age of 30, and then consul in 10, a regular progression that suggests tenure of the aedileship in 16, around the likely dramatic date of our poem.Footnote ⁴ As praetor, Antonius stage-managed events around Augustus' fiftieth birthday, following his return to Rome; and if the fourth ode-book was published in 13, then the prominence within it of Odes 4.2 may mark Antonius' spectacular emergence into public life in that year.Footnote ⁵ In addition, as a pupil of the critic Crassicius Pansa and the attested author of an epic Diomedeia, Antonius evidently took an active interest in poetry;Footnote ⁶ but in the background, as will be seen, stands a more substantial figure in the history of Roman letters, M. Antonius (cos. 99), great-grandfather of Iullus, participant in Cicero's de Oratore, and himself one of the two leading oratores of his generation.

Antonius' activity as poet is adduced in the pseudo-Acronian commentary to explain the poetic celebration anticipated in lines 33–44, in these terms: cantabis nobiscum, tu Antoni, ... meliori opere victorem Caesarem ... (‘you, Antonius, will sing with us of victorious Caesar in a better work’). This interpretation, which has prevailed at least since Late Antiquity, underpins the modern view of the poem as a recusatio, similar in essentials to what we find in Horace's ‘refusals’ to sing of Octavian/Augustus and Agrippa in Sat. 2.1.10–20, Odes 1.6 and 2.12, and Ep. 2.1.250–9.Footnote ⁷ On this reading, Horace responds to a prior suggestion from Antonius that he compose a Pindarizing song in honour of Augustus' anticipated return and triumph;Footnote ⁸ but he declines the request and counter-proposes that Antonius himself should ‘sing of Caesar’;Footnote ⁹ and his praise of Pindar is revealed as a defensive device to parry an unwelcome commission.

The recusatio-reading involves a number of serious difficulties. In the first place, there is no mention of Augustus until the second half of the poem; and his putative triumph (which of course was never sought or granted) is placed at some indeterminate point in the future (quandoque), so that it can scarcely be the object of a song-commission representing the settled requirements of state celebration.Footnote ¹⁰ Again, while a recusatio might put forward a classic poet as an ideal exemplar (in Odes 1.6, Varius is presented as alter Homerus), the Pindar-passage does not fulfil that function in relation to praise of Augustus.Footnote ¹¹ Rather, it constitutes a completed statement about Pindar's incomparable style and status, rounded off by a return in ring composition to Antonius (26). The recusatio-reading thus requires us to accept that for more than half the poem, Horace ignores the object of the supposed request, Augustus' return and (putative) triumph, and focusses instead on the desiderated Greek model for its treatment, in what turns out to be the pretext for pushing the main project back to Antonius.

Yet the figurative description (54–60) of the sacrificial vitulus implies that a poem of some sort will be forthcoming from Horace, or indeed that it is already doing so through the medium of Odes 4.2 itself.Footnote ¹² No scholarly account has satisfactorily related this metapoetic sacrifice (together with the parade of Callimachean principles in stanzas 7–8 [25–32]) to the commonplace encomium anticipated in stanza 12 (canam, 47), to the speaker's identification with the state in stanza 13 (49–52) or to the apparent hyperbole of the poem-outline in stanzas 9–11 (32–44).Footnote ¹³ The focal point for much of the difficulty is line 33, concines maiore poeta plectro, where commentators' puzzlement over the apparent concession of the maius plectrum by the publicly inscribed poet of the Carmen Saeculare (and their uncertainty as to whether Antonius is to compose an epic poem or a Pindarizing ode) is symptomatic of the deep-seated implausibilities that reside within the traditional reading of Odes 4.2.Footnote ¹⁴

In short, the recusatio-reading simply does not hang together, and in my view cannot be satisfactorily defended. In what follows, a new reading of the poem is put forward: Antonius' imagined prior request will be reassessed (Section 2); the central sections (3–5) will deal with dramatic dynamics (including the neglected issue of personae loquentes) and generic identity; and the two final sections (6–7) will consider dating and occasion. As the title indicates, it will be argued that the poem is a Roman version of the Greek dithyramb. This will entail a generic approach which has had a mixed scholarly reception over recent years, namely the categorization as ‘dithyrambs’ of certain Latin poems containing Dionysiac/Bacchic material and celebrating, or anticipating, triumphs.Footnote ¹⁵ In reopening the matter here, I hope to clarify the sacral, literary and political contexts in which these and other works for Liber/Bacchus were composed, and to show how Odes 4.2, read as an Augustan dithyramb, relates and responds to them.

2. ALTER PINDARUS?

At the start of the poem, the reader is evidently in mediis rebus but has no clear signposts to assist in identifying the situation being presented. The opening stanza is parainetic in tone, from which we may infer that, notwithstanding his public status, ‘Iullus’ is addressed either on equal terms or as a younger man. This does not exclude the possibility that ‘Horace’ is responding to encouragement to emulate Pindar;Footnote ¹⁶ but how might such encouragement have been framed? The assumption that it involved an Augustan poem-commission is simply a back-reading from concines maiore poeta plectro | Caesarem (33–4). If, however, Antonius might be supposed to have requested a carmen in the Pindaric manner appropriate to the dramatic occasion (whatever it was), then ‘Horace’ might be fending off an invitation to parade as alter Pindarus.Footnote ¹⁷ There were well-known precedents: Varius, nominated to write about Agrippa, had been puffed by Horace as alter Homerus (above, Section 1); and Propertius (4.1.64) had dubbed himself Callimachus Romanus. Horace touches on the subject of Pindaric imitation in Epistles 1.3, addressed to Iulius Florus (closely related to our poem: above, n. 8); and he goes on to consider poetic nomenclature in the later literary epistle to the same addressee, Ep. 2.2.99–101 (two poets engage in a duel of compliments): discedo Alcaeus puncto illius: ille meo quis? | quis nisi Callimachus? si plus apposcere visus, | fit Mimnermus et optivo cognomine crescat (‘I come away as Alcaeus on his mark. And he on mine? Who but Callimachus? And if he seems to require still more, Mimnermus he becomes; and let him thrive with his chosen cognomen’). The model here is the commander who acquires a new cognomen from conquered territory (a topic deployed also in Odes 4.8).Footnote ¹⁸ Poetic nomenclature features more than once in Odes 4, as Horace explores his own public status in the period following performance of the Carmen Saeculare. The first title to which he lays claim, Romanae fidicen lyrae (‘player of the Roman lyre’), foregrounds the Roman lyricus (4.3.23). Later (4.6.30), in a passage bearing on the term poeta at 4.2.33, Horace ascribes to Apollo the gift of his nomen poetae, in the formal context of the Carmen Saeculare.

Horace's interest in nomenclature is underlined by the framing of the Pindar-section with the two names of his addressee, Iullus (2) and Antonius (26). This adds point to daturus | nomina ponto (4.2.3–4): the poet fears to deposit his own (plural, accumulated) names, ‘Q. Horatius Flaccus Pindaricus’ as it might be, in the sea. The centrality of nomina, and their acquisition, to Horace's purposes is accentuated by the anagram Pindarus | pinnis ... daturus, taken together with the part-acrostic formed by the opening letters of the first four lines, P-I-N-N:Footnote ¹⁹ for Horace to rival Pindar, and as a corollary to be linked with his name, would be to reach for the skies on pinnae made of wax. By the same token, however, it engages the ‘nominal’ aspirant in the act of ‘giving’ (daturus, 4), a motif that proves to be programmatic for the poem as a whole (donandus [9, Pindar], donat [20, Pindar], donavere [38, gods], dabunt [39, gods] and especially dabimus ... tura [51–2, Roman citizens]): ‘Pindarus’, then, is evidently ‘he who will give’; and perhaps also, in an inspirational context (see below, with Section 5 [b]), ‘he who will give wings’.

The nature of Antonius' imagined request may best be approached through the ‘dithyrambs’ (10) placed first within the catalogue of Pindaric genres. This is the only song-eidos (genre) to be specified, and it is further emphasized within a metaphorical identification of Pindar and river (7–12) which itself develops the torrent-simile describing Pindaric style (5–6); moreover, the mountain-torrent has a striking Dionysiac parallel in Euripides' account of bacchants on Cithairon, ‘maddened by the god’ and leaping through a ‘torrent-flowing valley’ (Bacchae 1093–4), a clear analogue between landscape and mental state. The prominence of dithyrambi would be consistent with a prior request from Antonius that specified (or, in the occasional context, implied) the performance of a dithyramb in the Pindaric style. There is support for this reading in some ‘dithyrambic’ features of the preamble (stanza 1). Thus, bilingual interplay between nomina (4) and nova ... verba (10–11) deftly connects the accumulated ‘poet's nomenclature’ (above) with the synthetic, compound, vocabulary deployed by dithyrambists, through Greek ὀνόματα (both ‘names’ and ‘words’).Footnote ²⁰ As will be seen (below, Section 4 [b]), the ‘Pindarizing’ poet's feathered wings (pinnis) are reminiscent of self-reflexive statements by dithyrambic singers (and parodies thereof);Footnote ²¹ and Daedalea evokes poikilia, a particular characteristic of the dithyramb.Footnote ²² The ‘Icarian’ poet's watery denouement is reminiscent of katapontismos (‘diving into the depths’) as a feature of dithyrambic myth.Footnote ²³ The hazards of Pindaric emulation are thus given a markedly dithyrambic spin in the preamble, implying that the genre is associated with the dramatic occasion (and that it featured in Antonius' imagined initiative). Propertius (3.17.39–40) had pointed the way in hymning Bacchus: haec ego non humili referam memoranda coturno, | qualis Pindarico spiritus ore tonat (‘these things will I recount, fit to be memorialized in no humble garb, an inspired breath such as thunders in the mouth of a Pindar’). Here, surely, was a challenge to the Roman lyric poet most obviously associated with Pindar.

The first half of the poem comprises the one-stanza preamble, the five-stanza evocation of Pindar and his poetry (5–24) and then the two-stanza (25–32) comparison of swan (Pindar) and bee (‘Horace’). A central panel is thus framed by two apologetic programme-statements addressed to Antonius, thereby offering a complete, three-part, Pindar-poem.Footnote ²⁴ Pindar himself is projected as a quasi-divine song-power (analogous to a river-god), and the survey of song-types is structured along hymnic lines, as a disjunctive catalogue in er-stil, but with marked elements of ring composition.Footnote ²⁵ The ‘hymn’ brings the poet-god to life (present tense verbs), and celebrates his continuing efficacy: he creates belief and emotional engagement in the historicity of the ‘just’ fate of compound mythological monsters (14–16, central stanza);Footnote ²⁶ he grants monumental status to agonistic victors (17–20; the fourth stanza [palma, donat] recalls the second [laurea donandus]); and in immortalizing the heroic dead, he conveys the sobbing emotion of the flebilis war-widow (21–5; the fifth stanza [with metrical overruns, below Section 4 (b)] recalls the first). Strikingly, ‘Horace’ moves into a higher, impersonal, register, quite different from the two framing personal statements, and suggesting something akin to inspiration by his exemplar.Footnote ²⁷ Yet, the catalogue is set out along the lines of a scholarly edition, book-eidos neatly packaged within Sapphic stanza;Footnote ²⁸ thus, the juxtaposition of Pindaric ingenium and Horatian ars (exemplified in stanzas 2–6 and paraded in the apologetic comparison of swan and bee), owes something to the Library at Alexandria.Footnote ²⁹

Pindar's ‘inspirational’ impact on ‘Horace’ has been in operation through the medium of reading and book-roll.Footnote ³⁰ Again, in awarding Pindar the crown for supremacy among the lyrici (laurea donandus, 9), ‘Horace’ himself has taken on the role of κριτής (‘judge’, ‘critic’), surveying the past masters (a role that will be clarified below, Section 7).

Horace's undoubted admiration for Pindar is not incompatible with wit and humour. To use a term for the literary misrepresentation of contemporary poets, he deploys deformazione in attributing to Pindar a dithyrambic style more characteristic of later practitioners.Footnote ³¹ In this connection, we might ponder whether the verbal play between aureos (23) and aura (25), allied with the prosaic regularity (tendit ... quotiens, 26) of Pindar's expeditions to the clouds, hints at the role of ‘gold’ in transporting the ‘Dircaean swan’ to the sublime heights.Footnote ³² The Athenians paid him well for a dithyramb (fr. 76 Ma.), which however, according to the biographical tradition, also brought a cash fine (compare Horace's multa?) by Thebes.Footnote ³³

3. DRAMATIC DIALOGUE

With carmina fingo (32), the ‘Pindar-poem’ might be thought to be complete. But of course there is more to come, and matters take a new turn in stanza 9, with the unheralded statement around which the poem as a whole hinges: concines maiore poeta plectro | Caesarem ... (‘You will join singing for Caesar as a poet with a greater plectrum ... ’, 33–4). On this translation, ‘Horace’ turns from his own profile as a poet (stanza 8; 29–32) to Antonius and proposes that the latter, a poet with a greater plectrum, should sing of Caesar. But the absence of tu with concines deprives this supposed shift of the emphatic marking we should expect.Footnote ³⁴ By contrast, an emphatic ego (27) marks the comparison of Pindar and ‘Horace’; and, again, meae ... vocis (45–6) marks a contrast between that speaker and the subject of (the second) concines in the preceding stanza (41). The omission of a pronoun marker at line 33 is significant, and it is this anomaly which points to the need for an alternative reading.

The alternative to be considered here is that Antonius enters the poem as persona loquens at line 33. Antonius, I will argue, responds to the ‘Pindar-hymn’ and to its framing stanzas; it is Antonius who addresses ‘Horace’, and ‘Horace’ who is the subject of concines, so that no rhetorical requirement arises for pronoun-marking. It may be acknowledged that the suggestion of an unmarked (or, rather, a negatively marked) change of speaker in the middle of a Horatian ode is likely to strike some readers as involving greater difficulties than the anomaly it seeks to address. Aside from the dramatic dialogue between Horace and Canidia in Epodes 17, and from Odes 3.9, where an ‘I’ figure and ‘Lydia’ speak in alternate stanzas without explicit marking, there is no generally agreed Horatian instance of such a phenomenon. Yet there is some evidence, particularly in contexts where Horace appears to be working within the conventions of Greek choral lyric, to suggest his active interest in speaker-variation, in addition to his experimentation with the variety of individual singer-personae offered by the Greek canon.Footnote ³⁵ At all events, few Horatian critics would claim that the subject of dramatic voice in the Odes is fully understood as things stand, and 4.2 would appear to offer a welcome opportunity to explore its bearing on a major poem in the late book-collection. The analysis that follows is put forward on the assumption that two conditions will need to be satisfied if speaker-change is to be accepted: it will need to remove the anomalies outlined earlier, and yield something superior to the traditional reading it replaces; and it will need to embrace a new context, both generic and dramatic, in which changes of speaker can be understood to take place. The first requirement will be tackled here, and the second in the treatment of dithyramb and drama in Section 5.

The obvious place to start discussion of the persona loquens in 4.2 is at the series of exclamatory cries quoted in stanzas 12 and 13 (45–52). The former contains a play on ‘voice’ (meae ... vocis, 45–6), the owner of which (canam, 47) is different from the singer envisaged in (the second) concines (41), and will address laudes to Augustus. In the latter, the speaker is identified with the ‘whole state’, which will address the processing Triumphus (see below). The parallel double cries o sol | pulcher, o laudande and io Triumphe ... io Triumphe are explicitly signalled as such (canam, 47; dicemus, 50); yet they represent different ‘voices’, in the first instance Antonius and in the second omnis civitas. The anticipation of the phrases to be uttered on Augustus' return focusses attention on the shifting reference of the speaking voice (vox) in that context, and it can, I believe, help the reader understand implicit shifts of persona loquens on the present dramatic occasion.

We turn now to stanza 9 (33–6):

On the working assumption that the Antonius does indeed enter as persona loquens at this juncture, the stanza can be translated as follows: ‘To the accompaniment of a greater plectrum, poet, will you sing of Caesar such time as he draws the ferocious Sygambri the length of the sacred hill, resplendent in well-merited wreath …’. Concines can be given its syntactical (con-) connection with maiore ... plectro (see below, Section 4 [c], on ‘accompaniment’). This leaves poeta, punctuated as shown, as a vocative (not as nominative with suppressed tu), and vocative poeta offers a discreet, but positive, textual signal of a change of addressee and speaker: in addressing ‘Horace’ by the ‘nomen poetae’ which the ‘poet’ goes on to claim for himself in Odes 4.6, Antonius acknowledges his professional status and responds to his concern about nomina (Section 2).Footnote ³⁶ The thrust of Antonius' intervention at line 33 lies in the prediction that ‘Horace's’ diffidence, and claimed deficiencies vis-à-vis Pindar (stanzas 7 and 8), will fall away in the presence of Augustus, and with a Sygambrian triumph as his subject. Antonius is not, however, commissioning a triumph poem, still less reformulating a prior request: he is making an entirely new point that leads on from what ‘Horace’ has said about Pindar and about himself in the first half of the poem. On the evidence of the higher, ‘inspired’, register of the Pindar-hymn he has just heard (above, Section 2), Antonius responds to ‘Horace's’ subjective claim to ‘smallness’ (parvus, 31) by presenting Caesar as a song-subject analogous to Pindar (hence the two laurel wreaths), one that will be equally or more ‘inspiring’. In the presence of the victorious princeps at Rome, he claims, ‘Horace’ will indeed be capable of sublimity.

If Antonius does indeed enter the poem in stanza 9 (33–6), at what point does he cease to act as persona loquens? In stanza 14 (53–6), the contrast of great and small sacrifices makes it clear that ‘Horace’ (me, 54) addresses Antonius (te, 53). A satisfactory speaker-sequence is obtained, I suggest, if Antonius' intervention is understood to include the preceding stanza (49–52), and to conclude at that point. A particular attraction of this speaker-division is that it makes sense of the received text, in particular the respective referents of MS te (49) and te (53) in successive stanza-openings. The first te is addressed by Antonius to ‘Triumphus’, and the apostrophe echoes Epodes 9.21–2, io Triumphe, tu moraris aureos | currus ...? The second te (53) marks ‘Horace's’ re-entry into the text, responding to Antonius as previous speaker. This re-entry is further marked by neglected features in the sacrificial sequence opening dabimusque divis | tura benignis (51–2): the rejoinder deals with votive sacrifices but is syntactically disengaged from the predicted offering of tura; and decem tauri represents a repost to Antonius (‘not just tura, but tauri will be required to discharge your obligations’).Footnote ³⁷ The tura will be offered to all gods as a component of a state thanksgiving supplicatio (below, Section 5), while decem tauri totidemque vaccae will evidently be a state sacrifice (perhaps to take place on the Capitoline [sacrum clivum, 35]) and are the rejoinder to Antonius' prediction of the song ‘Horace’ is to sing about the triumphal procession. The wit of this sequence, and of the greater/lesser sacrificial contrast, is sharpened by reference to tus (a standard representative of the ‘humble offering’) as a component of the state offerings.Footnote ³⁸ With this bantering compliment to Antonius' status, the poet rounds off the dialogue.

On this reading, quotation marks should be placed before concines (33) and after benignis (52). Stanzas 9–13 (33–52) present Antonius' response to the ‘Pindar-hymn’ and its personal frame, while stanzas 14–15 (53–60) present ‘Horace's’ rejoinder to Antonius. This yields an overall stanza-breakdown as follows: 1 (preamble), 5 (‘Pindar-hymn’), 2 (comparison of Pindar and Horace), 5 (Antonius' response), 2 (‘Horace's’ rejoinder). Parallel to the ring-structure of the ‘Pindar-hymn’, Antonius' hypothesized response will unfold in ring composition in five distinct parts, each matched to a stanza, thus:

A¹ (stanza 9; 33–6): ‘Horace’ will sing of Caesar when he leads triumph-procession;

B¹ (stanza 10; 37–40): gods' benevolence, as component of Antonius' laudes of Caesar;

C (stanza 11; 41–4): ‘Horace’ will celebrate festivities vowed for Augustus' return;

B² (stanza 12; 43–8): felix Antonius will sing personal salutation, laudes of Caesar;

A² (stanza 13; 49–52): civitas will salute processing ‘Triumphus’; supplicatio to gods.

The hyperbole of the prescribed content in stanzas 9–11 (33–44) can thus be attributed, not without humour, to Antonius; and the apparent inconsistency between the avowal of Callimachean principles in stanzas 7 and 8 (25–32) and ‘Horace's’ supposed identification with the crowd in stanza 13 (49–52) falls away. Especially striking within the proposed dramatic dialogue is the progressive identification of Antonius with the state. Thus, his sacrificial obligations will arise from state responsibilities and may well (with Quinn) be referred to his tenure of the praetorship in 13.Footnote ³⁹ Again, there is a good deal of official-sounding language in the passage; and in si quid loquar audiendum (45), following the cessation of judicial activity (43–4), we catch the legal tone of the praetor who ex officio ‘hears’ court cases and expects at that juncture himself to say something which ‘requires to be heard’; this is Antonius the orator, speaking of the encomiastic branch of oratory, as distinct from the forensic (and, in bona pars, of laudatio as distinct from vituperatio).Footnote ⁴⁰ Then, identification with the ‘whole city’ in salutation of the processing ‘Triumphus’, dicemus ... | civitas omnis (51–2) recalls Epodes 16.36, eamus omnis ... civitas, and suggests that here, as there, we are hearing a magisterial persona speaking for the community. In taking his interlocutor back to the era of civil war (the iterated io Triumphe, as already noted, echoes Epodes 9.21–3), Antonius reminds ‘Horace’ of his advocacy of escape to a tempus aureum (Epodes 16.64), renouncing reditus to Rome (25–6, 35–6); and in that context, anticipating Augustus' reditus, he asserts that there can be no higher blessing, quamvis redeant in aurum | tempora (‘even if the times return to gold’, 4.2.39–40).

This alternative account of the poem, as I see it, helps to show that if the hypothesis of speaker-change is accepted, solid gains in sense and coherence emerge, and the incongruities outlined earlier (Section 1) fall away. Moreover, a clearer sense of ‘personality’ emerges, as also a pronounced element of personal humour. There is a further gain in the poem's sacral dynamics: speaker-variation creates a break between Antonius' enthusiastic assumption that return and triumph are assured and ‘Horace's’ salutary reminder of the requirement for sacrificial offerings. It is the act of vowing the latter that will secure the benign gods' engagement in Augustus' victorious return; thus, divine benevolence, though noted in general terms by Antonius (stanza 10) and assumed as a given by him in the matter of Augustus' return (stanzas 12–13), is ‘Horace's’ primary focus. The poeta suggests that his (metapoetic) votum will be efficacious in bringing about the desiderated return (at which point he will of course be voti compos, and the vitulus will be sacrificed in fulfilment of the vow). The votum, its metapoetic status and its quasi-sacral function will be considered further in Sections 4 and 5; but the key point lies in ‘Horace's’ culminating claim, speaking as poeta, to be uttering a song-votum that will help ensure Augustus' success and return. Speaker-variation (with the distinct personae of ‘Horace’ and Antonius) provides the key to identifying this central, sacral, focus.

4. SOME ‘DITHYRAMBIC’ FEATURES

It was observed some 40 years ago by Francis Cairns that the prominence given to the Pindaric dithyramb in Odes 4.2 reflects a Roman view that poems about triumphs were themselves a form of that genre.Footnote ⁴¹ Cairns identified Odes 4.2 itself as a dithyramb, a conclusion which will be restated here, in different terms and with additional argumentation. In this section, as a basis for the identification, some features of the poem which can be considered (in some sense) ‘dithyrambic’ are reviewed.

5. DITHYRAMB, VOTUM AND DRAMA

6. THE ODE-BOOK AND THE TIME OF WRITING

My aim in the final two sections is to reach a closer understanding of the poem's time of writing and of the occasion at which it was first performed. The dramatic date appears to be 16, for that is when the publica vota pro reditu will have been decreed (above, Section 1). The date of composition is a more complex matter: it might in principle be late 16 or early 15, but there are good grounds for thinking that the poem is the product of a later period, in or around 13. The first pointers in this direction appear in Antonius' intervention, and in phraseology that seems to reflect his appointment as praetor urbanus in that year (above, Section 3). The true starting point for discussion, however, lies in the aggregated linkages between Odes 4.2 and other poems in Odes 4: its closest partners are 4.3, 4.5 and 4.8, and there are meaningful links, offering material assistance in clarifying or confirming points of interpretation, with 4.1, 4.6, 4.14 and 4.15;Footnote ¹²² that amounts to half the total, and while taken in isolation they might be reckoned the product of a process of composition stretching over several years, the cumulative impact is undeniably impressive and would be consistent with composition of our poem as a component of a planned ode-book.Footnote ¹²³ For that reason alone, the possibility that 4.2 was written closer to 13 than to 16 merits serious consideration.Footnote ¹²⁴

The proposition that Horace could compose poetry about near-contemporary events with a dramatic date substantially earlier than the time of writing is not problematic. Recent work on Epodes 9 has underlined the need to distinguish its dramatic or fictional chronology from its actual time of composition; and a strong case has been made for putting the date of composition of Odes 3.6, set in the uncertain world of the early 20s, in 25 or later.Footnote ¹²⁵ Within Odes 4 itself, the dramatic date of the other reditus poem, 4.5, clearly precedes news of Augustus' return, but again there appear to be good grounds for thinking that it was written to celebrate that event.Footnote ¹²⁶ From time to time, then, Horace seems to locate the perspective of the persona loquens in the period preceding some decisive event that has already taken place, in order to catch the mood of that period, be it uncertainty, hope, despair or bullish exuberance. This kind of dramatic retrojection is effective precisely because it carries the appeal of direct recall for readers who had been participants in the events described; and it skilfully extends the range of techniques that Horace had at his disposal for indirect praise of Augustus and the regime. This, I suggest, is what the poet is doing in Odes 4.2: from a later perspective, he is recreating the period following Augustus' departure from Rome in 16, the publica vota pro reditu, and the sincerely held feelings evoked by that event.Footnote ¹²⁷ Within the generic framework of the dithyramb, Horace fictionalizes a dramatic dialogue between himself and Antonius; and through that medium, he presents afresh some of the issues of public policy that arose around the Gallic expedition.

A further pointer to the poem as the product of 13 is Apollinari (9), alluding to the Ludi Apollinares: these were staged annually from 6 to 13 July, and in the year 13, they will have overlapped with the supplicationes in Augustus' honour following the assumed date of his return to Rome, 4 July.Footnote ¹²⁸ It is difficult to suppose that such a reference is a matter of coincidence, given the onward connection to urbis publicum ludum (41–2). Indeed if Antonius' praetorian post in 13 was as praetor urbanus, as seems very likely (above, Section 3), he will have had charge of the Ludi Apollinares, in line with responsibilities laid down when the festival was established in 212;Footnote ¹²⁹ and as presiding magistrate, Antonius will have led the festival pompa, processing on the high chariot, thereby yielding a resonance with ‘the triumph that never was’.Footnote ¹³⁰ The games cohere with the triumphal and ‘dithyrambic’ scenario outlined earlier in other significant respects: they were vowed to secure victory over Hannibal, and were dedicated to Apollo victoriae, non valetudinis ergo (Livy 25.12.15); they comprised ludi circenses (including horse/chariot racing), and ludi scaenici, in which winners were awarded the palma; and perhaps of greatest immediate relevance, they were observed by the populus laureatus, each individual wreathed in laurel.Footnote ¹³¹Odes 4.2 thus brings together the Roman/Apolline and the Athenian/Dionysiac (as represented, for example, in Pindar, fr. 75 Ma.: above, Section 5 [d]); and in that context, it seems very likely that Horace was aware of Antonius' praetorship in 13 when he wrote the poem, as also of his role in the Ludi Apollinares and the proximity of the latter to the date of Augustus' return.

Of course, there remains the fact that the Sygambrian triumph did not take place. By the time of Augustus' return, the Sygambrian threat was consigned to history; and indeed Augustus' decision not to stage further imperial triumphs will by then have been known to Horace and his audience (Dio 54.24.7–8, under the year 14). The appearance of the phantom triumph is not, however, inconsistent with the separation of the dramatic date from the date of composition, for if Odes 4.2 does present a dramatic reconstruction of a dialogue that notionally took place in 16, then what we have is not the genuine prospective view of the campaign as recorded by Horace in 16, but something altogether more complex. We should be compelled to read the poem as a reconstruction of the campaign perspective of the year 16 composed from the later standpoint of the year 13.

To appreciate why Horace should represent Antonius as anticipating a triumph in a dramatic reconstruction written when it was known there would be (or had been) no such triumph, we might recall the testimony of Dio (54.20.6) to the effect that the Sygambri retired and sued for peace when they learned that Augustus himself was taking the field. That account foregrounds the personal impact of the princeps on a hostile barbarian tribe which had challenged the imperial order; and as such it resonates, from a later century, with a major theme of Odes 4.14, Augustus' capacity to resolve tribal challenges through his presence in the field (41–52). That poem's closing reference to the Sygambri represents the theme (51–2): te caede gaudentes Sygambri | compositis venerantur armis (‘to you the Sygambri whose pleasure is in slaughter pay loving obeisance, their arms laid up’). This culminating hymnic compliment explicitly credits Augustus with effecting, through his presence, the voluntary accommodation of a tribe which was given, by its nature, to the business of slaughter;Footnote ¹³² and of course, the peaceful outcome of a frontier war which at its inception had seemed to require a formidable enemy to be crushed in armed conflict must redound mightily to Augustus' credit as a peacemaker. In Roman contemporary terms, the achievement is more praiseworthy still when considered against the conventional expectation of victory and triumph among authoritative observers at the time of the revolt. For that reason, we need to address the proposition that the two Sygambri-passages, the anticipated triumph (4.2.34–6) and the tribal submission (4.14.51–2) were not after all written in isolation, the product of successive impressions recorded at the time they entered circulation; rather, they constitute the twinned record of the start and conclusion of the campaign from a later perspective, operating together to throw into relief the magnitude of Augustus' achievement as imperial pacifier. Putnam has commented that ‘[t]he sequence of odes devoted to the emperor is a major revocatio, which ... brings Augustus home’.Footnote ¹³³ That is consistent with an ode-book composed as a unitary design; and while it cannot prove that 4.2 was written later than its dramatic date, it certainly encourages the view that the poem is part of a larger design for the period leading up to the senatorial decree of the Ara Pacis Augustae on 4 July 13.

Antonius himself would not, of course, be in any way diminished by a ‘retrospective’ reading of Odes 4.2. On the contrary, it would have been essential for the purposes of dramatic fiction that the ‘forward view from 16’ (that is, of a conventional war followed by triumph) should be placed in the mouth of an authoritative public figure, and a loyalist member of the imperial house. Only thus could it serve as an effective ‘contemporary’ foil for Augustus' actual achievement of a peaceful settlement. In addition, we need to recall the parallel that was being drawn in late 15 between the European campaign culminating in that year and the fall of Alexandria on 1 August 30, fifteen years earlier to the day (Odes 4.14.34–6). The shade of Mark Antony, and his identification with Dionysus-Osiris (and Hercules), hovers in the background; and as a corollary, the regime's consistent efforts to ‘reintegrate them [sc. Dionysus and Hercules] into the post-Actium consensus’ are drawn into the political frame.Footnote ¹³⁴ Antonius' request to Horace to compose a Pindarizing dithyramb (if that is indeed what we are to suppose), something the poet had done once before, in Odes 1.37, to celebrate the Actium triumph, would thus act as a tribute to his commitment to erasing his father's imprint on the cult of the two gods. At the same time it would compliment Augustus' capacity for reconciliation, extending even to the son of his greatest internal enemy. Internal reconciliation would thus complement external peacemaking, and that may be the true significance of the exuberant exclamations of joy and loyalty placed in Antonius' mouth as he envisages the happy scene; Augustus, we infer, has left a united capital, and a governing elite very much at one with itself. Moreover, and by way of counterbalance to an infamous father, the engagement with Cicero's De oratore (above, Section 4 [c]), and the address to Antonius himself at the critical juncture (26), must evoke the great consular ancestor, and his role in a classic literary dialogue. And there, as again in stanza 12, the poet associates rhetoric and poetry, while complimenting Iullus on his status not as poeta, but as orator.Footnote ¹³⁵

It is the extraordinary impact of the praesens princeps, the leader in person, on his loyal subjects that underpins Antonius' prediction that an inspired Horace will rise to the occasion when Augustus comes home. The returning emperor is presented as a deus praesens modelled on Dionysus in epiphany; and his quasi-divine impact on those who come into his presence reflects the powers of the god himself. Horace had developed this same theme in Odes 1.37, where Octavian exercises ‘positive’ Dionysiac powers over enemy forces infected with an orgiastic travesty of maenadism, thereby reducing them to blind panic and putting them to flight.Footnote ¹³⁶ Dionysus, as Versnel observed, ‘is the victorious god, whose weapon is not the lance or the sword, but trance and μανία’ (original emphasis).Footnote ¹³⁷

The god's eastern conquests had lain through the advancement of his choruses and his teletai, and it is the capacity of the choric god to subdue enemies through the advancement of cult that Horace had celebrated in Odes 2.19, idem | pacis eras mediusque belli (‘but in the same way, you were a mediator of peace and midmost in war’, 27–8). Here, to be sure, is a powerful divine model for Augustus' achievement in reducing warlike tribes to peaceful behaviour by his presence alone, as celebrated at the end of Odes 4.14; and while we do not meet those powers in their external application in our poem, we do observe their positive domestic impact on omnis civitas.

For Augustus' epiphanic and triumphal reditus to Rome, a tradition of Greek processional models reaches back to the epiphanic processions of Dionysus' returning image during the Athenian City Dionysia.Footnote ¹³⁸ This processional tradition, itself triumphal in character, later embraced the epiphanic/Dionysiac city-arrivals of Hellenistic rulers, such as that of Demetrius Poliorcetes at Athens (Ath. 535f–536a), and ultimately of Alexander himself. To conclude: the conception of imperial reconciler and peacemaker, and the Dionysiac dimension sketched here, are likely to have matured in Horace's mind towards the time of Augustus' return. The considerations advanced in this section make it probable that Odes 4.2 is the product of the year 13.

7. LIBER AND THE POETS

The matter of dramatic setting, though not explicitly defined in the poem, must nevertheless be addressed in that it bears on the context of Antonius' imagined request to Horace and on the dramatic dialogue that follows. The question, then, is whether a setting can be identified such as to provide a sacral context for a dithyramb that records a poetic votum to Liber for Augustus' return, and offers a technical discussion on the subject of poetry together with critical adjudication on the best on the Greek lyrici.Footnote ¹³⁹

One official occasion that would satisfy most, perhaps all, of these criteria, is known from later evidence to have been staged each year at a shrine of Liber/Bacchus on the Palatine.Footnote ¹⁴⁰ This was a kind of Roman festum poetarum, the primary evidence for which lies in Ovid's celebration of the occasion (perhaps an extension of the Liberalia on 17 March; or else the temple natalis).Footnote ¹⁴¹ This is how Ovid opens his poem (Tristia 5.3.1–4):

The importance of this remarkable testimony lies in the certainty that Ovid is not speaking in figurative terms of poetic sacra but is describing a genuine calendar event, an annual gathering of poetae at the Palatine temple of Liber. Ovid depicts this company of poetae as functioning on collegiate lines, the named members sharing, as cultores, in sacral-poetic activities in the vicinity of Liber's statue and altar, and constituting, at least for the purposes of the immediate occasion, the god's chorus.Footnote ¹⁴² This is the context of Ovid's plea to his fellow poets, notwithstanding his exclusion by exile, to preserve his nomen within their company.

Ovid claims to have participated on many occasions (pars ego saepe fui, Tr. 5.3.6), suggesting an extended period prior to his exile in a.d. 8. In addition, Pausanias (8.46.1, 5) records the location, within a Roman shrine to Bacchus standing ‘in the emperor's gardens’, of a tusk from the Erymanthian boar; that trophy, Pausanias adds, had been taken by Augustus as booty from Tegea following the defeat of Antony. If this hieron is identical with the Palatine temple, its natalis will likely be datable to the period after Actium; it will have been closely associated with the defeat of Antony and will have represented Augustus' personal identification with both Dionysus/Liber and Hercules.Footnote ¹⁴³ The temple of Liber would thus take its place alongside the nearby temple of Apollo as part of Octavian's post-Actium sacral strategy, on his private grounds on the Palatine.Footnote ¹⁴⁴

The event described by Ovid will most probably have been instituted as part of the same programme and at roughly the same time, in the early 20s. Thus, Propertius' hymn to Bacchus (3.17), the dramatic setting of which is the god's shrine adjacent to the Palatine Magna Mater (Bacche, tuis ... aris, 1; iuxta dea magna Cybebe, 35–6; ante fores templi, 37), might well have been presented in the temple on the annual occasion under discussion.Footnote ¹⁴⁵ More important, as Propertius well knew (but chose to ignore on this occasion), was the temple of Palatine Apollo. Ovid refers to the proximity of Apollo in his concluding, collegiate, prayer to his fellow poetae (Tr. 5.3.57–8): sic igitur dextro faciatis Apolline carmen: | quod licet, inter vos nomen habete meum (‘Thus, therefore [sc. given Ovid's candid (candore, 53, an ‘Apolline’ term) approbation of contemporary poetry] may you compose songs with Apollo in your favour; as to what is licit [sc. not illicitum for the ‘collegium’], keep my name among you[rs]'). The burden of Ovid's conditional prayer is that the poets' new work should be acceptable to ‘Apollo’, i.e. at the library adjoining the Palatine temple;Footnote ¹⁴⁶ the prayer is appended to the suggestion (53–6) that some form of critical activity, involving individual iudicia on the work of fellow poets, formed part of the annual proceedings at the temple of Liber; and the passage as a whole implies some association (figurative or otherwise) between those iudicia and admission of books to the library. The realia here remain obscure, but I would suggest that Ovid's closure offers a Palatine temple and ‘library’ context that helps make sense of Horace's ‘Apolline’ iudicium on the ‘dithyrambic’ Pindar, laurea donandus Apollinari (9).Footnote ¹⁴⁷

The existence of the Palatine celebration of Liber may account for other early-Augustan poems addressed to the god, or containing a high level of Dionysiac subject matter. An obvious candidate is Odes 1.37: this ode (in my view, indebted to Pindar's dithyrambic treatment of a ‘Dionysiac’ Hercules' capture of Cerberus) again conjoins Dionysus and Hercules. Tibullus 1.7 also enters the frame for consideration, as do Horace's two hymns to Bacchus (2.19 and 3.25).Footnote ¹⁴⁸ An intriguing feature of these Augustan poems for Liber/Bacchus is the incidence of cross-referencing. Most obvious is Ovid's quotation, in the poem cited at the beginning of this section, of the opening lines of Tibullus 1.7 hunc cecinere diem Parcae, fatalia nentes | stamina non ulli dissoluenda deo (‘the Parcae sang of this day, weaving the threads of fate which are not to be released by any god’); compare Tr. 5.3.256–7 scilicet hanc legem nentes fatalia Parcae | stamina bis genito bis cecinere tibi (‘of course, the Parcae, weaving the threads of fate, twice sang of this law for you [sc. Liber] twice-born’), where Ovid wittily turns the forecast by the Parcae of Messalla's birthday into the Parcae as singers of the ‘first dithyramb’ for Dionysus' double birth. Some of Horace's back-references to his own work have already been noted. Propertius' ‘dithyrambic’ echoes in 3.17 of Odes 3.25 (Bacchus) have been analysed by John Miller;Footnote ¹⁴⁹ and Odes 4.2 in turn echoes Propertius 3.17.Footnote ¹⁵⁰ And while these various resonances could perhaps be explained by the fact that contemporary poets were writing about the same god within the same literary milieu, they are certainly consistent with a common sacral focus and performance scenario in the poets' annual celebration of Liber/Bacchus. The matter calls for a fuller treatment of Tristia 5.3, its relationship to the corpus of Augustan poetry for Liber/Bacchus, and the interconnections within that same body of work than can be offered here; but the connections outlined above may suffice to show that the temple of Liber, and the annual gathering of poets, offer a plausible dramatic setting for Odes 4.2 as a votive dithyramb, as also for the ‘dithyrambic’ assessment of the Pindaric corpus. Antonius' role in the ode may perhaps be understood as analogous to a collegiate patronus, an extension of his emerging profile as poet-impressario that yields, in context, a relationship with Horace not unlike that between P. Papinius Statius and L. Arruntius Stella, in a later era.Footnote ¹⁵¹

The hypothesis of ‘fictitious retrojection’ outlined earlier (Section 6) leaves open the question of the poem's actual place of performance. Perhaps the most economical proposition is that it was presented in the temple of Liber as part of the celebration of 13 (or 12), and that that celebration itself took the form of a kind of cena adventicia in Augustus' honour. There is no direct evidence on the point, but there is perhaps a pointer within Horace's treatment of the vitulus: as was noted earlier (Section 5 [b]), the growing calf is promised (sc. to Liber) in a votum pro reditu. Yet in figurative terms, the calf also represents Odes 4.2 itself, and so additionally functions as a metapoetic dithyramb-sacrifice. The deft combination of (literal) vitulus votivus with (metapoetic) vitulus as poem-sacrifice in effect bridges the space between the dramatic timing and setting (in 16) and the actual timing and setting, in 13 or 12. On this reading, in the personality of poet-devotee within the Palatine cult of Liber, Horace has written a poem-sacrifice in which the dramatic setting is the same annual occasion, but located three or four years earlier. The carmen projected by Antonius is Odes 4.2 itself; and Horace's claim to compose ‘efficacious songs’ (operosa carmina) is vindicated by the fact of Augustus' return (and, for all we know, by his presence in the temple of Liber at Horace's performance).

Odes 4.2 offers a glimpse of a community of poets in a sacral setting; and if the linkage with the temple of Palatine Liber proposed here is acceptable, Antonius and Horace are brought together by common commitment to studia and by common allegiance to Liber as a god of poetry reclaimed from Antonius' orientalizing father for Italy and the Augustan consensus.