II THE POETICS OF OLDER LATIN: COMPLEMENTARY PAIRS AND PHONETIC PLAY

What effects did the author notice in the older style of poetry which he put to use in his own epigram? The relevant techniques can be briefly recalled by two famous texts. The well-known prayer for the suovetaurilia in Cato's de Agricultura reveals the importance of underlying structure, especially binary oppositions, in making surface expressions intelligible.Footnote ¹¹ The prayer asks Mars to ward off ‘morbos visos invisosque, viduertatem vastitudinemque, calamitates intemperiasque’ ‘diseases seen and unseen, unproductiveness and destruction, disasters and unseasonableness’ (Agr. 141). Visos invisosque are obviously opposites. But, after a fashion, so are viduertas and vastitudo: the former describes reproductive failure, the loss of what might have been;Footnote ¹² the latter, destruction and the loss of what already exists. That is, the nouns describe potential and actual loss. Perhaps calamitas and intemperiae can be analysed similarly, if calamitas referred originally to sudden disasters, like hailstorms,Footnote ¹³ and intemperiae to unseasonable or unbalanced weather.Footnote ¹⁴ That is, one refers to the raw power of weather (weather as a Behemoth, as it were) and the other to its variability (weather as a shape-changer). The expectation that pairs will fall into sets like visos invisosque almost forces the other sets to become opposites or complements of some kind. That was certainly a likely response for a second-century or later reader, schooled in the dualisms and complementations of rhetorical amplificatio, especially in its epideictic version.Footnote ¹⁵

Visos invisosque and viduertatem vastitudinemque reveal another important aspect of prayer style, the use of phonetic figures to knit sets together. Such phonetic play was not merely imitated, but expanded into a significant literary device by Latin poets. The most striking examples, perhaps an accident of attestation, come from Ennius. For Ennius, particularly in tragic choruses, phonetic echo is not simply a linking device,Footnote ¹⁶ still less a decoration added after the fact, but appears to be purposefully chosen to highlight the structure and encapsulate the central ideas of a passage. For example, Andromache's famous lament for the destruction of Troy ends ‘haec omnia vidi inflammari, | Priamo vi vitam evitari, | Iovis aram sanguine turpari’ ‘All these things I saw enflamed, | Priam's lifeforce by force unforced, | Jupiter's altar by blood befouled’ (106–108W), with echoing infinitives in –ari. This is no empty flourish. The morphological equivalence sets parallel three stock elements of the sack of a city — destruction of property, murder of persons, and desecration of temples. That triplet, in turn, recalls an earlier part of Andromache's lament: ‘O pater, o patria, o Priami domus, | saeptum altisono cardine templum’ ‘O father, O fatherland, O house of Priam, | temple bound by hinge sounding above’ (101–102W):

Thus, of the opening vocatives, only patria does not have a correspondent. But here another piece of verbal play has its effect. Destruction and murder and desecration directed towards a king and his palace makes the death of Andromache's father (pater) the like of the death of the fatherland (patria). The pun is a kind of gloss on Priam, turning him into the embodiment of his state. That sharpens the piquancy of yet another play on words: Priam is not simply ‘killed’; rather his life (vitam) is ‘unlifed’ (evitari) from him by violence (vi). First, the word order is iconic: ‘violence’ collides with ‘life’ and cleaves Priam from it:

Second, the phonetic similarity (vi vi-), like that of viduer- ~ vasti-, indexes a pair that is linked but opposed — here, death and life, perhaps the ultimate such pair. Andromache for her part did not have in mind any cosmic balance between life and death: that shows in her neologism evitari ‘unlife’, in which ‘life’ (-vit-) is utterly negated (e-). Whether Cicero, in whom these passages are preserved, read them quite this way cannot, of course, be known; but he was moved by their effects: ‘Outstanding poetry! It embodies grief in content and word and metre’ (‘praeclarum carmen! est enim et rebus et verbis et modis lugubre’, Tusc. 3.46); and he recognized that their style did not match the contemporary fashion for Alexandrian poetry: ‘What an excellent poet! — even if these singers of Euphorion reject him’ (‘o poetam egregium! quamquam ab his cantoribus Euphorionis contemnitur’, Tusc. 3.45).Footnote ¹⁷

That Ennius meant his verse to be read this way perhaps requires further illustration, which is beyond the immediate purposes of this article. Here it suffices to note that certain later readers might well have had the response just described. Ennius' poetic technique recalls the method of the ancient etymologists. Two principles of that method are relevant — principles that typically strike moderns as fanciful but which are perfectly natural before the theory of the arbitrariness of the sign took hold. First, similarity of sound was taken to mean similarity of sense, as is clear from almost every ancient etymology.Footnote ¹⁸ Second, phonetics can be iconic. Nigidius Figulus, in service of an argument that nouns were not positiva (‘arbitrary’) but naturalia, pointed out that the outward motion (emovere) of the lips in pronouncing vos ‘you (pl.)’ is like pointing to the addressees, whereas the restraint of the lips in pronouncing nos ‘we’ has an inward, self-referential quality (ap. Aul. Gel., NA 10.4). In Ennius these two principles are elevated to art: similarities of sound are used to suggest not etymologies, but nonce patterns of meaning; and iconicity is expanded beyond phonetics to include word order. A later reader schooled in the particulars of ancient linguistics would have been alert to those very effects.

III TECHNIQUE IN NAEVIUS' EPITAPH: DIVINE AND MORTAL

In the poetics of older Latin, then, relatively simple techniques — phonetic figures and binary (or ternary) sets — were used to create sophisticated patterns and layers of meaning. The sets need not be obvious from the semantics of the component words, but the pressure of the structure may draw out, and even create, the required oppositions or complementations; and the phonetics, far from being mere surface decorations or chance frills, may establish structural patterns and even encapsulate the deepest and most essential ideas of a passage.

These are the very techniques that inform the epigram for Naevius. The entire epitaph rings changes on the opposition between divine and mortal. That opposition is established in 1a, in three distinct ways — semantically, lexically and syntactically. Semantically, there is of course the meaning of the two words, immortalis and mortalis. But divos or deos or some other word might have been chosen for the former, and homines or humanos for the latter.Footnote ¹⁹ The particular lexemes immortalis ~ mortalis give a figura etymologica. That not only sharpens the contrast; it also sets up ‘death’ as a major motif of the poem. Obvious as the contrast may be, it is not in fact all that commonFootnote ²⁰ (probably because each of the words almost automatically implies its opposite).Footnote ²¹ Third, there is a syntactic figure: the opposed pair is fronted around the subordinating conjunction si. That gives the pair an emphatic position. It also produces a metrical effect: the fronting lets the opposed pair fill its half-line, one word nestling on either side of the caesura Korschiana.

There was a reason for making the opposition of divine and mortal prominent: it is the main motif of the epigram, continuing through the rest of it in various forms. So in 1b: the ‘lawful’ (fas) is contrasted to ‘weeping’ (flere). Unlike immortalis and mortalis, fas and flere are not natural opposites. But they, too, contrast divine and mortal. Conventionally gods did not weep. Weeping is thus properly a human action. Fas, by contrast, is ‘divine law’, coming from and proper to the gods. Whatever the origin of the stricture against divine tears, it seems to have been imagined as reflecting the condition of gods as immune to vicissitude. Accordingly, if a god weeps for a mortal, he abases himself. Therein lies the nefas. A god must be true to his own superior nature. The exceptions are instructive. If gods do cry or come close to crying, it is typically when those especially dear to them suffer or die — that is, when their high indifference has already been breached by affection.Footnote ²²

The opposition between fas and flere has another aspect especially relevant in a poem about gods and artful speech. Latin has a cluster of words in fā- derived from PIE *bheH₂- ‘speak’: fārī ‘speak’, fāma ‘saying; fame, rumour’, fābula ‘story’, etc. (the root is also the source of Greek ϕημί ‘say’).Footnote ²³ The prominence of that set induced ancient etymologists to refer other words in #fā- to the act of speaking. That is, the syllable #fā- acquired phonaesthemic associations with speech. Varro, for example, finds a way to connect the unrelated fānum ‘shrine’:Footnote ²⁴ ‘“Shrines” (fana) take their name from “speaking” (fando), because, in dedicating a shrine, priests “declared” (fati sunt) a boundary’ (‘hinc [sc. a fando] fana nominata, quod pontifices in sacrando fati sunt finem’, L.L. 6.54).Footnote ²⁵Fās was etymologized in the same way: ‘“Charm” (fascinum) and “lawful” (fas) are derived from “speaking” (fando)’ (‘fascinum et fas a fando nominantur’, Paul. ex Fest. 88).

The connection is not a mere fantasy of the etymologists. When Roman writers wanted to derive forms of fas and nefas, which are defective nouns without oblique stems, they turned, not to other nouns in -ās,Footnote ²⁶ nor to the putative stem fār-,Footnote ²⁷ which, if it ever existed, seems not to have survived, but to fand- and nefand-, the gerund(ive)s of fari, e.g. ‘at sperate deos memores fandi atque nefandi’ ‘regard ye the gods, mindful of the just and the unjust’ (Verg., Aen. 1.543).Footnote ²⁸ The connection must have been that much easier because of fātum ‘fate’, another kind of divine declaration identical in form to another verbal noun of fari, its (remade) *-to- participle.Footnote ²⁹ The semantic history of fari may also have played a rôle. By the time of attested Latin it was no longer the unmarked word for ‘speak’ or for ‘say’; loqui and dicere, respectively, had assumed these rôles. As a marginalized and therefore marked variant, fari came to be used for special kinds of saying (like ‘prediction’ or ‘declaration’) or in marked contexts (standing for dicere in high poetry).Footnote ³⁰ Divine pronouncement is, of course, also a highly marked kind of speech.Footnote ³¹

In short, fari ‘speak’ and fas were felt to be connected, and the author of the epigram doubtless felt the same connection. And that reveals another aspect to the contrast between fas and flere in 1b. Whereas speech is formed of articulated vocal sounds, weeping retards clear articulation. Thus weeping and speaking are typically exclusive. That requires no illustration from ancient literature, though of course there are such (e.g. ‘atque ego ut primum fletu represso loqui posse coepi’, Cic., Repub. 6.15). That contrast is especially sharp in 1b. Where flere is a failure of speech, the replacement of words by sounds, fas is, not only speech, but the ultimate speech, divine declaration: the clear and timeless speech of the gods. The contrafactuality helps this effect: the goddesses ‘would have wept’, but lamentation and godhead belong to different worlds, mortal and divine. The alliteration (f- fl-, made more prominent still by foret) therefore indexes both sameness (two kinds of vocal expression) and difference (coherent vs. incoherent) — just like viduertas and vastitudo, both losses but of different kinds, potential and actual. To sum up:

The effect, of course, is to elevate the magnitude of Naevius' death. The stark opposition of the first hemistich comes near to collapsing: the order of things was almost torn; the goddesses of speech almost could not speak.

The second line is like the first: each half-line is informed by the polarity of divine and mortal established in 1a. In 2a half of the polarity is overt: the Camenae, Italic spring-goddesses used in the early Latin literary tradition to represent the Muses (on whom more presently), are explicitly divae. The other half of the polarity is covert, depending on 1b: there weeping was established as a human activity, gaining its meaning partly through the pressure of the structure. A phonetic figure assures that that meaning is carried over. 2a–b provide the apodosis to the protasis of 1a–b. The lines are, as such, logically linked. But they are also bodily linked, in the form of a grammatical figure, polyptoton, flere flerent:

That link is particularly close here because the infinitive is, as it were, incorporated into the imperfect subjunctive, so that the beginning of the second line quite literally echoes the ending of the first. That must have been a noticeable effect: such a collation, where an infinitive is juxtaposed to an imperfect subjunctive of the same verb, is vanishingly rare in classical Latin.Footnote ³² Plautus has one instance, Cicero another, in formulations broadly similar to that in the epitaph, setting a general possibility (expressed by the infinitive) beside a specific possible instantiation (expressed by the subjunctive).Footnote ³³

What merits the grief of the Camenae? They would have wept over the poet Naevius (2b). Here, too, the polarity between mortal and divine continues: Naevius is a mortal but his profession was a poet, and that, if not deifies, certainly sanctifies him. In the context of the epigram, in which goddesses come near to weeping for a lost creative genius, poeta is plainly meant to signal not simply a profession but an elevated status. The poet is conventionally inspired by the gods — an old idea, already in Homer (θɛῖος ἀοιδός, Od. 4.17) and attested in Roman literature as far back as Ennius, who called poets sancti (op. incert. fr. 16 Skutsch). In short, our polarity appears in both halves of the second line:

At the same time, the author of the epigram falls short of calling Naevius sanctus in so many words; it is, rather, the structure not only of the line but of the whole epigram that elevates Naevius — a delicacy to which we will return. Here it is worth noticing another point of structure. 2a–b can be viewed as opposing not only the mortal to the divine but also the general to the specific: the epithet of the Camenae precedes their name, that of Naevius follows, giving a chiastic order:

The chiasmus is important, because it gives a significant phonetic figure, another echo across a verse boundary like flere ~ flerent. Here it is the syllables across the caesura that match: Camenae Naevium. That echo, too, must have been conspicuous. The sequence -nae nae- is rare in classical Latin — found, in fact, only in this passage, if the databases can be trusted. The rarity of the sequence as such is not surprising. Many words may end in -nae, but only a few begin that way, and rare words at that: naevus ‘mole’ and its derivatives, including Naevius; and naenia, a variant of nēnia ‘song of lamentation’; and nae, a generally late variant of nē ‘truly’. The sequence -nae nae- might thus have made itself felt in any case; but its presence here, I suggest, is meant to give a particular effect. It certainly seems to have been otherwise avoided: no Saturnian of Naevius or of Livius Andronicus features such an echo of identical syllables across the caesura. Here that echo bodily links the Camenae to Naevius:

The phonetics symbolize the special relationship of the Camenae to Naevius. Ennius' use of phonetic figures, rather than Cato's, is illuminating here. Just as vi vitam was iconic — ‘force’ struck ‘life’ — so here Camenae Naevium is iconic: the Camenae touch or overlap or somehow come into contact with Naevius; somehow something of each is in the other. Furthermore, rather as the special position of patria in the exclamation ‘O pater, o patria, o Priami domus’ produces an identification of pater and patria, so here the juxtaposition Camenae Naevium implies, if not identity, then at least contact between the divine realm and the mortal, a spark bridging the gap rigidly established in 1a and bent, but not broken, in 1b.

This is not too much to take from a phonetic figure: the logic of the poem bears out this very implication. The surface argument of the poem is illogical: ‘The Camenae would have wept for Naevius; therefore when he died the Romans forgot how to speak.’ What is the force of the ‘therefore’ (itaque, 3a)? A middle step is implied: ‘The Camenae would have wept for Naevius [since he was a kind of Muse himself]; and therefore …’ etc.Footnote ³⁴ This is the very sort of implied middle step entirely familiar from Pindar's and Horace's lyric technique. Indeed, here it is more transparent than in many Horatian instances, because of the overt logical particle (itaque). Not only the phonetics, then, but also the logic of the poem, make Naevius out as a kind of Muse.Footnote ³⁵ That is why, when he dies, the Romans forget how to speak.

The echo -nae nae- may have another iconic effect. In the very middle of a line that describes mourning is a syllable pattern that resembles the word for a ‘song of lamentation’: nē-n-ia or nae-n-ia. That word, like ul-ul-āre ‘wail’ or βάρ-βαρ-ος ‘*incomprehensible speaker; foreigner’ or, more precisely, like λαλέω ‘chatter’ (as it were ‘bab-babbler’) or bū-b-ō ‘owl’ (‘*hoot-hooter’), was doubtless felt to contain, and perhaps very well created by, a (partial) reduplication to represent a repeated sound.Footnote ³⁶ The various meanings of nēnia — ‘lamentation’, ‘ditty’, ‘magical spell’, ‘nonsense’ (nēnior ‘vana loquor’, Gloss.; nēniōsus ‘βαττολόγος, garrulus, nugax, nugator’, Gloss.) — certainly support a core sense like ‘rhythmic speech with repeated syllables’.Footnote ³⁷ That very effect of repeated syllables would be more pronounced in the epitaph if the composer did not speak standard late first-century b.c. urban Latin.Footnote ³⁸ In a poem which describes an (avoided) act of lamentation, the sound of lamentation is allowed to echo distantly in the background — the sound of lamentation specifically associated with women.Footnote ³⁹

At all events, the first two lines share a structural pattern: each half-line contains both elements of the polarity. But the order of the elements differs: whereas the first line sets up pairs in the order divine–mortal, the second line sets them up in the order mortal–divine, swinging as it were on the ‘hinge’ of flere–flerent. The polarity is worked out still differently in the last two lines, not only in structure but also in focus: each element of the polarity is given a whole line; and the focus is not on beings but places. The third line describes a divine place; the fourth, a mortal place:

The variation in the deployment of the polarity is not merely variation for its own sake (though in that, too, there is artfulness). Rather, the variation signals a difference in the distribution and dynamics of the polarity. In the first two lines divine and mortal are interwoven in every half-line. In the person of Naevius, they even touch (Camenae Naevius). But when it comes to the two locales, each described in its own line, mortal and divine are kept firmly apart. When Naevius was alive, the divine was in the mortal world; after his death, the worlds again split asunder.

Lexical details reinforce that idea of separate worlds. First, itaque postquam ‘and thus, after …’ introduces, with a certain prosaic exactitude, a strict temporal layering: Naevius enters the underworld; ‘then and thereupon’ the mortal world begins to suffer.Footnote ⁴⁰ Second, Naevius' death is described as his having been ‘passed on’ (est … traditus) to the realm of Orcus. Trado is not commonly associated with death or the like and is not otherwise used with thesaurus. But it is a common word for ‘transferring’ or ‘delivering’ a person into ‘shelter, protection [or] imprisonment’ (L&S s.v. B.1). When Naevius died, he was, as it were, ‘turned over to’ or ‘remanded into the custody of’ Orcus. His death was not an annihilation but a passage or a transfer. Mando is occasionally used in a similar way.Footnote ⁴¹ That is a vivid way of stressing, not only the finality of death, but the gulf between — and therefore the existence of — two distinct worlds. This is exactly the (to my ear, now somewhat faded) English metaphor of ‘pass on (sc. into the next world vel sim.)’ for ‘die’.

Thesauro augments this effect of est … traditus. Thesaurus is literally a ‘treasure-chest’, ‘storehouse’, etc. Here it is metaphorical for the House of Death — a very rare image elsewhere only in Ennius (252W; cf. below). In what way is death a ‘treasure-chest’? The metaphor depends upon the function and the form of a thesaurus. In form the thesaurus was a locked container of heavy stone — the very like of a Roman sepulchre.Footnote ⁴² In function the thesaurus was a secure place of long-term deposit. What goes into a thesaurus drops out of circulation. That function is drawn out by the words obliti sunt … loquier ‘people forgot … how to speak’: when Naevius passed into Orcus' treasure-chest, he fell out of circulation. In short, the phrases itaque postquam, est … traditus thesauro and obliti sunt … loquier underscore the cleft of mortal and divine implied by the third and fourth lines.

The last hemistich, 4b, plays a special rôle in this imagery of separate worlds. The line alliterates (allomorphs of) the liquid consonant l (loquier lingua Latina). The smooth sound represents not a beauty that is but a beauty that is no more. Elaborating a lost grace, a regular affect of death epigrams, in this epigram contributes to the sense of a gulf between worlds — all the more so in the presence of thesaurus, which also served to sequester valuables. There is something still more in the last hemistich: it is metapoetic. It describes ideal speech and is itself meant to example the very techniques that the poet plainly idealized; the line embodies the aesthetic ideal it describes. There is a poignancy in the poet's attempt probably not apparent to him: octosyllabic second hemistichs are not found in Naevian Saturnians.Footnote ⁴³ The graces of antique speech really had crossed to the other side.

IV THE STANCE OF NAEVIUS' ‘EPITAPH’: GREEK vs. ROMAN

In sum, the Naevius ‘epitaph’ is deeply indebted to the techniques of older Latin poetry. Underlying polarities give order to surface expressions, above all, the polarity between mortal and divine, signalled by the first half-line, which informs the entire rest of the poem — in the first distich, within lines; in the second, across lines. Phonetic and syntactic figures emphasize and themselves express the central ideas of the epitaph — the contact of human and divine (immortales mortales, Camenae Naevium); the near breach of the cosmic order because of an inspired poet's death (si foret fas flere); and the — now lost — ideal form of speech (loquier lingua Latina). But despite what, by the standards of later classical poetry, is antiquated technique in antiquated metre, it would be a serious mistake to see the epigram as purely ‘Latin’ or ‘Italic’ in spirit. On the contrary it is thoroughly infused with Hellenism — but with a carefully modulated Hellenism. Greek precedents and parallels are recalled but modified to reflect, not only a Roman, but a very particular kind of Roman, sensibility.

That sensibility is implicated in a literary polemic started by Ennius: how best to Hellenize in Latin poetry. Livius Andronicus had rendered the Odyssey into Latin in Saturnian verse, and that was the metre of Naevius' historical epic, the Bellum Punicum. But Ennius, pioneer of the dactylic hexameter, famously rejected Naevius' efforts as the verse of Fauni ‘Fauns, rural goat-gods’ and vates ‘seers’ or ‘soothsayers’,Footnote ⁴⁴ and claimed to have met Homer himself, and perhaps the Muses, in a dream (Ann. 4–13W).Footnote ⁴⁵ Ennius also, it appears, rejected Livius' use of Camenae for goddesses of poetry, preferring the transliterated Musae. Footnote ⁴⁶ The author of our epigram — naming Naevius, writing in Saturnians, using Camenae — must have had this polemic in mind. Assigning true Latin speech to Naevius, the author pointedly rejects Ennius' claim that his texts were the new beginning that Ennius felt them to be.Footnote ⁴⁷

But the author is more than a champion of Naevius. His own poetic practice takes on the central issue of the polemic by offering itself as a model of how to Hellenize. The technique of the epigram is integral to this message; for the poem, far from being a specimen of unreconstructed nativism, puts forward a distinctive approach to Hellenism — a Hellenism clearly present but by the same token deliberately muted. That restraint appears unmistakably in the depictions of the Camenae and Orcus and to some extent even in the mention of Naevius. In the epigram the Camenae plainly stand for the Muses. But the act of cultural translation is, so to speak, partial. The idea that the Muses should cry at the death of an artist has the feel of a pretty Hellenistic conceit.Footnote ⁴⁸ And indeed it appears in the Greek Anthology from the third century forward. An epigram attributed to Damagetus (late third century b.c.) relates that at the death of Orpheus, ‘The Pierian Muses themselves, with Apollo, the master of the lyre, burst into tears, mourning the singer’.Footnote ⁴⁹ Antipater of Sidon (second half of the second century b.c.) has a similar image (7.8). Alcaeus of Messene (late third to early second century b.c.) has the Muses weeping at the death of Pylades, an actor.Footnote ⁵⁰ An imperial epitaph follows the lead of these poems: ‘When the Camenae wept for your premature death, the people of noble Ostia wept. Julius Nicephorus, unhappy father, built this tomb’ (‘ut te, Palladi, raptum flevere Camenae, | fleverunt populi, quos continet Ostia dia. | Iulius Nicephorus, pater infelix, fecit’, CIL 6.20152 = CLE 606). Divine beings may do more than weep. Alcaeus of Messene imagines Homer's funeral rites performed by the Nereids (7.1); Hesiod's, by the Nymphs (7.55). Dioscorides (second century b.c.) imagines a satyr charged by the Muses with guarding Sophocles' tomb (7.37) and Sappho joining Aphrodite to mourn Adonis (7.407).

Such conceits plainly inspired the epigram; but in comparison to them, the Camenae of the epigram preserve a certain dignity. The crying of the goddesses is contrafactual: they would have cried, but they did not forget divine law (fas); Naevius' status, special though it was (poetam), did not make them forget theirs (divae). In short, the epigram appropriates the Greek idea of a patron god moved by mortal deaths but also maintains a stricter division between god and man than that emotion might imply. Sentiment is allowed to appear but then kept at bay as inappropriate to the depiction of immortals. The contrafactual condition is a kind of symbol for that manoeuvre: such a condition may express a scene, and vividly so, even as it denies that the scene ever took place.

The depiction of Orcus and his thesaurus illustrates the same point. Thesaurus, as I suggested above, keeps its economic meaning. That image implies a god of the underworld with economic concerns. In that there is an intimation of Πλούτων and his Latin equivalent Dīs pater. Πλούτων was originally a separate god but became an epithet of Hades and eventually his chief title. Popularly linked to πλοῦτος ‘wealth’, the god was recognized in Roman state cult in 249 b.c. as Dīs pater, a name that expresses that very connection, Dīs coming from the same root as dīves ‘wealthy’, dīvitiae ‘riches’.Footnote ⁵¹ In short, the metaphor of thesaurus signals an act of syncretism: the epigram's Orcus, a distinctly Italic figure, is, as it were, secretly Pluto. Ennius' Mors has been seen the same way: ‘Acherontem obibo, ubi Mortis thesauri obiacent’ ‘I shall meet Acheron, in the face of the treasure-houses of Death’, avers his Iphigeneia (252W), where thesauri may obliquely allude to Dīs pater.Footnote ⁵²

But again the syncretism is not thoroughgoing. If the phrase Orci … thesauro hints at Pluto, it is only a hint: the fullness of Pluto's godhead is not present. Orcus is not here involved, as Pluto sometimes is, in the cycle of death and rebirth. Perhaps paradoxically for a god of the underworld, Pluto is often associated with life, since all life comes from the earth and returns to it. In this regard Pluto is typically linked to Prosperina and may even be depicted with a cornucopia. When Dis Pater was added to the Roman state cult (249 and 207 b.c.), it was together with Prosperina. But in the epigram there is no hint of that pair or of the cycle of life and its associated folk-tale elements. The flow of wealth goes straight down to Pluto, and there it stays. Indeed, the very metaphor of thesaurus, if it suggests wealth, also suggests a flow in one direction only. Here is a god of death without rebirth — Hades proper, the Orcus of the regular Latin expression (de)mittere Orco ‘send off to Orcus, kill’ and the Orcus of Catullus, whose ‘malae tenebrae | … omnia bella devora[nt]’ ‘dark shadows consume every pretty thing’ (3.13–14).

Thus neither Orcus nor the Camenae are simple equivalents, by way of variatio or syncretism, for Πλούτων or Μοῦσαι. It is, rather, almost as if the old Latin names are emblematic of a restrained and partial Hellenization. One more possible example of restraint may be noticed. The epigram's Naevius, dear to the Muses and named poeta, is evidently an inspired poet of the Greek type. But the poem is much less bold on this point than it might have been.

The poem certainly does defend the poet's elevated status, by way of an intertext of a quite Alexandrian kind. A work of Naevius contained the line ‘fato Metelli Romae fiunt consules’ ‘By fate the Metelli become consuls at Rome’. The line is said to have spurred a rejoinder from the Metelli's side, ‘malum dabunt Metelli Naevio poetae’ ‘The Metelli will make trouble for Naevius the poet’.Footnote ⁵³ The second line of the Naevius epitaph, ending with a form of Naevius poeta, beginning with the verb phrase, and setting a proper name before the caesura, is a deliberate echo of the Metelli's rejoinder. The poetics of the two lines are of interest here. In the original line the august name Metelli, pressing down on Naevius, and the fronted threat malum dabunt make the poetae of Naevio poetae into a mark of his insignificance: his crafted words will not protect him from real trouble; the poetae of the end of the line holds up poorly against the consules of the original verse. If malum dabunt affects the exaggerated outrage of comedy,Footnote ⁵⁴ the line scripts Naevius into the rôle of a cheeky slave. The Naevium poetam of the Naevius epigram echoes Naevio poetae, but in that line, the almost weeping Camenae, linked by an echo of sound to the poet's name, instead ennoble his profession. In short, the epigram echoes a line that also expressed a difference of status but reverses the polarity: Naevio poetae marked a weaker party; Naevium poetam marks an honoured craftsman.

But stronger expressions of inspiration there are none. Ennius' claims of inspiration stand in contrast. As we have seen, he had depicted himself in the Annales meeting Homer in a dream. For that there were literary precedents in Hesiod and Callimachus and philosophical analogues in the thought and literature of Pythagoreanism. Ennius' images are more striking than anything that survives until Cicero, in poetry commemorating his consulship, has himself addressed by the Muse Urania (fr. 6 Blänsdorf), instructed by Minerva (fr. 17), and admitted by Jupiter into a council of the gods (fr. 17). Similarly, the epitaph, as I have suggested, makes Naevius into a Muse, whose people forget how to speak when he dies. But he is not called anything like ‘the tenth Muse’ or ‘the mortal Muse’ (θνατὰν Μοῦσαν, Anth. 7.14.2; Antipater of Sidon on Sappho), or ‘dearest companion of the deathless Muses’ (ἀμβροσίῃσι συνέστιɛ ϕίλτατɛ Μούσαις, 7.41.1; Diodorus (?) on Callimachus), or ‘servant of the Pierides’ (ὁ λάτρις | Πιɛρίδων, 7.44.5–6; Ion on Euripides), or ‘equal to the gods’ (θɛοῖς ἴσα, 7.407.9; Dioscorides on Sappho). The epigram implies much but says little.

In short, intertexts expose the particularity of the poem's stance. The conceits of Hellenistic epigram, in particular those associated with epitaphs of poets, and the tales of Greek myth generally show the sensibility of the epigram to be Hellenistic — but not quite. Neither the Muses nor Pluto nor the inspired bard is fully present, even as they make their presence felt.

Why did the composer of the epigram stop short, so to speak? The answer lies in the last line of the epitaph. That, too, is an instance of stopping short — or rather, in this case, taking a different turn. In the epitaphs of the Greek Anthology the world may be profoundly struck by someone's death. When, in Damagetus' epigram, Orpheus dies, not only do the Muses weep but so do the very stones and trees (Anth. 7.10.7–8). Alternately, to speak only of the epigrams on the death of literary figures, the essence of the poet somehow perdures. The poet may be imagined as still alive: Orpheus charms Hades himself (Damag., Anth. 7.9.7–8); Erinna joins the dance of the Muses (Anon., Anth. 7.12.6); Anacreon receives from them immortality (Simonides, Anth. 7.25.1–2). It may be a poet's works that survive: Sophocles' life will last forever on his ‘deathless pages’ (Simias, Anth. 7.21.6); deathless is Sappho's ‘artful diction’ (σοϕαί … ῥήσιɛς, 7.16.2, Pinytos); Euripides, like an epic hero, will enjoy κλέος ἄϕθιτον ‘glory imperishable’ (7.43.3, Diodorus). Another conceit is that the beauty or character of a poet's achievement may be, or ought to be, reflected by his tomb: Simias prays that Sophocles' tomb be adorned with ivy and roses (7.22); Simonides (?), that grapevines cover the tomb of Anacreon, who loved wine (7.24).

The Naevius epigram features no such sentiment or supernatural fantasy. The alleged result in the last line, if it is exaggerated, is entirely realistic in one respect: it sees the death of a prominent man in social terms; his people are affected. Among the epitaphs of poets in the Anthology, only Leonidas of Tarentum's simple couplet for Pindar has anything comparable, recording that he was ‘agreeable to strangers and dear to his countrymen’ (Ἄρμɛνος ἦν ξɛίνοισιν ἀνὴρ ὅδɛ καὶ ϕίλος ἀστοῖς | Πίνδαρος, ɛὐϕώνων Πιɛρίδων πρόπολος, 7.35). Closer is the famous epitaph of Protogenes: ‘Protogenes Cloul[ei] suavei heicei situst mimus | plouruma que fecit populo soueis gaudia nuges’ (= ‘Protogenes Cluli suavis hic situst mimus | plurima qui fecit populo suo gaudia nugis’ ‘Here lies Clulius’ Protogenes, charming mime, | who by his jests provided his people with very many delights', CIL I².1861, IX.4463 = Bücheler CLE 361).Footnote ⁵⁵ Here the skill of an artist, like that of Palladius above, is measured by his favourable public reception.

There is, then, a public turn in Naevius' epitaph: not, as with the Camenae, a reduced version of the literary fantasies of the epigrams in the Anthology but a step away from them — and a step towards the political world. It is to that world that public reception, as opposed to the infusion of divine graces, especially belongs. The oldest elogia, which of course have the political world firmly in mind, regularly feature the attitude of the citizen body, implicit or explicit.Footnote ⁵⁶

Indeed, the last line of the poem not only represents a turn to the political world; it also rounds off a cosmology, another aspect of the poem's exquisite structure. Beneath the poem's two distichs, which focus on persons, then places, is a descending hierarchy, the first line describing a cosmic principle; the second and third lines, divine beings and an extraordinary man; and the last line, ordinary people:

The workings of that cosmic order are clear. The singular men, like Naevius, depend for their inspiration on the gods who specially favour them and with whom they, somehow, interact; ordinary people depend for their inspiration on the extraordinary men. That, of course, is a social vision of a kind perfectly familiar in Roman society, in which the patres saw to the gods and the government, and the people looked to the patres. In short, the aesthetic cosmology of the epigram exactly parallels the political cosmology.

Or rather one particular political cosomology; for this is a very particular one, in which the flow of energy comes from the top down. Here there is no hint of the vox populi or of the lively and even abusive spirit sometimes attributed to the real Naevius. Here, rather, is the air of an aristocracy both confident and realistic. On the one hand, extraordinary men are like the gods — creators and guarantors of the order of things. On the other hand, if gods and extraordinary men occupy analogous positions, that does not mean that the gods and men are equivalent. In particular, death attaches to men, even favoured men like Naevius, as not to gods. The poem is unwilling to void the foedera naturae, as happened, in the rhetoric of praise anyway, for the Hellenistic god-kings — or even for the Roman conquerors of the second century, as eventually for Augustus and, of course, his successors. The epigram is thus closely parallel to another poem both confident and realistic in its aristocratic worldview, Horace's first Roman Ode. Gods and extraordinary men occupy analogous positions: ‘regum timendorum in proprios greges, | reges in ipsos imperium est Iovis’ ‘Fearful kings over their own flocks wield power — and over the kings, Jupiter’ (C. 3.1.5–6). But richer or more popular or more noble though a man may be, necessity cares not; someday his name will fall from the urn and his days be ended (3.1.9–16).Footnote ⁵⁷

In short, the aesthetic cosmology of the poem mirrors quite closely a particular kind of political cosmology. There, then, is the reason for the muted Hellenism of the epitaph: the author has a political ideal in mind, and that affects him as a reader of Greek. For the author, to embrace too fully the habits of expression of the Greek Anthology would be, as it were, to endorse an alien view of politics — virtually, an alien political theology. If, within the context of the well-worked idioms of Greek literature, the precious conceits of the Greek Anthology represent a striving for novelty and surprise, they might also be read, or perhaps misread, as implying a fantastic cosmos — the world of Ovid, to put it anachronistically — where gods mingle with men, the laws of nature are inverted, and passions rend the very order of reality. The composer of the epigram has avoided endorsing any aspect of that cosmos, even as he has borrowed its conceits. A modern scholar's pithy formulation is worth recalling: ‘The epicists’ use of myth and legend is not a descent into fiction but their way of thinking about reality.'Footnote ⁵⁸ The author of the epigram felt this power of myth keenly.

The author's political ideal affects him not only as a reader of Greek but also as a writer of Latin. It is precisely for political, even ethical, rather than strictly aesthetic or poetic reasons that, I submit, the composer was drawn to older poetic forms. Metrically the very choice of the Saturnian, obsolete by the time of the composer, sets Hellenic fantasy in a context of antique gravity. Lexically, the Hellenism poeta, in the epigram an inspired bard, is, so to speak, kept grounded by Camenae and Orcus, who, respectively, love him — but not too much — and perform what is required, regardless of who loves him. Perhaps most important in the ethical stance of the poem are its phonetic, syntactic and structural figures. The divine and mortal may come into contact, somehow (Camenae Naevium), but they do not really mix (si foret fas flere), for ultimately they are separate worlds (itaque postquam). Ennius' bold claims of inspiration are, for the author of the epigram, excessive.

As with the cosomology, the impetus — and perhaps even a model — for this last delicacy lies in the political world. The Roman ruling class did not, as a rule, claim direct individual inspiration by the gods. In the state religion the gods were approached and appeased according to accepted rituals. The pax deorum was like a treaty with a foreign state, overseen by the appropriate priests and conducted on fixed terms. It is natural that witches, seers and soothsayers appear in literature as figures of scorn. It is not merely that they belonged to popular religion; rather, they were, so to speak, the ideological opposite of the civic priest, claiming individual, unmediated contact with the divine — and for immediate personal gain.

All this posed a problem for politicians who wished to claim some special personal favour from the gods. Hence a celebrated habit of Scipio Africanus. He was said to have spent time in the temple of Jupiter Capitolinus before any major decision.Footnote ⁵⁹ The story has been regarded sceptically. But there is no serious reason to doubt that Scipio did exactly as reported.Footnote ⁶⁰ The artfulness of the gesture has not been stressed:Footnote ⁶¹ it is both bold and restrained. Scipio grandly claimed inspiration — but not directly, only by implication. That stands in contrast to the other stories of Scipio's enjoyment of divine favour. His patent excellences attracted the usual sort of story that his birth had been attended by favourable omens (de vir. illustr. 49.1); these are plainly legendary. Less plainly legendary, and perhaps containing a grain of truth, is Scipio's claim, before the assault on New Carthage, to have been visited in a dream by Neptune, who suggested a plan and promised aid.Footnote ⁶² The implication of divine favour by juxtaposition, on the other hand, reflects, if not perfect modesty, still a certain restraint, and for that reason, it seems to me, is very likely to be true. The unknown author of de viris illustribus seems to have appreciated the subtlety: ‘quasi divinam mentem acciperet’ ‘as if he were receiving divine inspiration’. Gellius' account has a similar phrase: ‘quasi consultantem de republica cum Iove’ ‘as if consulting Jupiter about politics’. This is the very balance of the epitaph, which implies close personal contact without being too precise about the mechanism.Footnote ⁶³

The author of the epigram has thus in his own way anticipated the modern discussion. Ennius' confident claims of improvement shaped older literary histories into teleologies, seeing successive generations of poets as producing improvements on the path to some notionally perfect Hellenized poetic form.Footnote ⁶⁴ Recent scholarship has complicated this simple picture. Verse of the vates the Saturnian may have been, but it could be a vivid and subtle metre, its literary version more artful than its inscriptional versions.Footnote ⁶⁵ Even the oldest inscriptional Saturnians, for all their antique morphology, show the distinct influence of Greek epigram, thoughtfully adapted as an instrument of self-presentationFootnote ⁶⁶. Livius' Camenae need not be taken as clumsy nativism but may be seen as an ingenious act of cultural translation:Footnote ⁶⁷ the Camenae, like the Muses, were linked to springs and spoke clear truths.Footnote ⁶⁸

Much of this more recent criticism shares an approach: aiming not to judge texts better or worse adaptations of their Greek counterparts nor to locate them along an evolutionary path to some ideal literary form, but to appreciate the deliberate creative choices of individual poets in rendering Greek texts and considering prior Latin ones. For there were always choices to be made, on every level of language — choices of syntax, semantics, lexicon, and even of morphology and phonology, to say nothing of metre, genre, and theme. To aim to appreciate these choices is to sympathize with a poet in the act of creation — an act of imagination that, if it has only recently gained broader currency in the critique of older poets, is the stock in trade of critics who read Vergil against Homer and the Alexandrian poets and later hexameter poets against Vergil.

The author of the Naevius epigram strikingly anticipates the stance of modern critics. He did not see Naevius as an ‘archaic’ author whose technique was to be avoided as being out of fashion; he did not see poetry as tending towards some perfection. Instead he attempted to appreciate archaic poetry on its own terms. He did not reject its phonetic play as the doggerel of seers but understood that patterns of sound, far from surface decoration, were a part of the way such poetry produced meaning, even as in his own time that technique was doubtless falling out of favour. He saw that the act of translation was a matter not only of translating, still less transliterating, but of considering how the translated product fits in a different cultural context. The implicit critique he thus produces is not literary in the modern sense; rather, as I have argued, he sees literature through the lens of ideal political life. Still, in my view, his sensitivity to literary choice is quite that of a modern critic's, and some measure of sensitivity to his own choices is required to understand his short but very rich reflection on the nature of Latin poetics.