I.1.1 The Alexandrian Footnote

By far the most commonly attested marker of allusion in Latin poetry is the so-called ‘Alexandrian footnote’, a device which assimilates literary allusion to the transmission of talk and hearsay. General appeals to tradition (such as ferunt, ‘they say’, audivi, ‘I’ve heard’, or ut fama est, ‘so the story goes’) frequently signal an allusion to specific literary predecessors, despite their apparent vagueness and generality. A famous example of this device occurs in the opening of Catullus’ epyllion, Carmen 64, where the dicuntur (‘they are said’) in the second line flags the poem’s polemical interaction with numerous other treatments of the Argonautic voyage.Footnote ²³ A simpler example, however, is that of fertur in Virgil’s description of the two gates of horn and ivory in Aeneid 6, which points back to Penelope’s famous description of these very same gates in the Odyssey (Aen. 6.893–6 ~ Od. 19.562–7). In addition to the verbal and thematic echoes of the Odyssean passage,Footnote ²⁴ Virgil’s vague appeal to tradition invites his audience to ask where these details have been ‘reported’ before, an extra spur to recall the legitimising authority of Homer.Footnote ²⁵

For Stephen Hinds, who has done more than any other to publicise this phenomenon,Footnote ²⁶ such ‘footnotes’ are ‘a kind of built-in commentary, a kind of reflexive annotation, which underlines or intensifies their demand to be interpreted as allusions’: dicuntur and similar expressions can mean not only ‘“are said [in tradition]”, but also, more specifically, “are said [in my literary predecessors]”.’Footnote ²⁷ But it is also worth stressing the variety of nuances that the device can bear in Latin texts. Far from simply marking an allusive debt, it can also highlight a particularly contentious point of tradition. When Virgil claims that Enceladus allegedly lies beneath Etna (fama est, Aen. 3.578), he acknowledges a literary debate about the precise identity of the giant beneath the mountain. In Pindar’s Pythian 1, Virgil’s main model for this passage (Aen. 3.570–87 ~ Pyth. 1.13–28), the giant was Typhon, but in Callimachus’ Aetia Prologue, Enceladus took his place (Aet. fr. 1.36) – an inconsistency that was already noted by the Pindaric scholia.Footnote ²⁸ In this case, Virgil’s fama est gestures not only to a single literary source, but rather to a plurality of competing alternatives, highlighting the contestability of tradition.Footnote ²⁹

In other cases, meanwhile, hearsay is invoked at points of apparent innovation, where inherited tradition is creatively reworked or completely rewritten. When Virgil claims in the Georgics that Aristaeus’ bees were lost through sickness and hunger (amissis, ut fama, apibus morboque fameque, G. 4.318), he seems to be lending the authority of tradition to what is in all likelihood his own invention, further reinforced by the aural jingle of fama and fame.Footnote ³⁰ In the Aeneid, meanwhile, Sinon prefaces an untraditional account of Palamedes’ genealogy and pacifism with an emphatic assertion of the hero’s famous reputation (Aen. 2.81–3):

This insistence on Palamedes’ fame lends a legitimising veneer to Sinon’s (and Virgil’s) untraditional account, but it also invites an audience to challenge the claims that follow, to zero in on their innovations and to dwell on their significance.Footnote ³¹ Such ‘faux footnotes’ as these are ‘a kind of poetic smoke and mirrors’,Footnote ³² a means for a poet to mark his own creative ability and unique place in tradition. By presenting such innovations as ‘traditional’, the poet implies that his work is coextensive with the literary tradition: any word he utters is immediately incorporated into the larger web of authoritative fama.

The ‘Alexandrian footnote’, then, is not simply a shortcut to mark literary debts and sources. It is also a polemical signpost of contested tradition and an authorising signal of literary innovation. At its heart, it is a tool of literary self-representation, a means for a poet to position himself against what his predecessors have said and what his audiences have heard – a valuable feature of any Roman poet’s allusive repertoire.

I.1.2 Troping Allusion

Besides the ‘Alexandrian footnote’, Latin scholars have also identified a host of other tropes which figure, model and mark allusive interactions. Foremost among these are embedded references to memory, repetition and echo. Ovid’s Mars, for example, reminds Jupiter in the Metamorphoses of a prophecy he had previously made in Ennius’ Annals (Met. 14.812–15):

The war god’s emphatic juxtaposition of memoro memori and his overt appeal to the past in quondam signal the verbatim quotation of Jupiter’s former words: the god explicitly recalls the earlier Ennian poem (Met. 14.814 = Ann. 54 Skutsch).Footnote ³³ Similar, if a little more implicit, is Ovid’s description of Narcissus’ death in the Metamorphoses: Echo’s repetition of the egotist’s words (dictoque ‘vale’ ‘vale’ inquit et Echo, Met. 3.501) self-consciously highlights Ovid’s own ‘echoing’ of Virgil’s ‘fading doubled vale’ in the Eclogues (‘vale, vale’ inquit, ‘Iolla’, Ecl. 3.79).Footnote ³⁴ The inconspicuous et further reinforces the echoing effect: Echo speaks these words ‘as well’ as Virgil, Menalcas and Phyllis.

The most famous example of this phenomenon in modern scholarship, however, is the speech of Ariadne in the Fasti (Fast. 3.471–6):

Ariadne’s words here assert a strong sense of literary déjà vu.Footnote ³⁵ Abandoned by Ovid’s Bacchus, she recalls the similar mistreatment she received from Catullus’ Theseus (64.116–206) – a short time previously in her fictional timeline, but several decades ago in terms of Roman literary history. She even quotes her former literary self directly: periure et perfide Theseu (Fast. 3.473) draws on Catullus’ perfide … perfide … Theseu (64.132–3) and periuria (64.135), while nunc quoque ‘nulla viro … femina credat’ (Fast. 3.475) is lifted largely verbatim from Catull. 64.143 (nunc iam nulla viro … femina credat). Together, these repetitions strengthen Ariadne’s and our own sense of déjà vu: ‘how often’, she goes on to ask, ‘must I speak these very words?’ (quotiens haec ego verba loquar?’, Fast. 3.486). Yet besides these verbal reminiscences, it is the accumulation of temporal markers (en iterum, en iterum, nunc quoque) and the language of repetition and similarity (similes, eadem, relata … est) which cue us to see this scene as a self-conscious repeat, alongside the pointed memini that precedes her self-quotation: she actually ‘remembers’ her earlier literary appearance.Footnote ³⁶

There is, of course, considerable irony in this remembrance. As Hinds notes, Ariadne has ‘the very quality of mindfulness (memini) so signally lacking in her earlier lover at his moment of perjury’ (immemor, 64.135, cf. oblito, 64.208),Footnote ³⁷ and – we might add – the same quality allegedly lacking in all men (dicta nihil meminere, 64.148).Footnote ³⁸ Yet in addition to this reversal, Ovid also manipulates the temporality of the scene, undermining the Catullan narrator’s authority by ironically challenging his version of events. The Catullan poem, it turns out, did not present her final lament after all (extremis … querellis, 64.130), since she repeats similar complaints now (similes … querellas, Fast. 3.471). This temporal paradox becomes even more acute when we add several earlier Ovidian scenes into the mix: in the Ars Amatoria, Ariadne is also pictured bewailing her abandonment on Dia and accusing Theseus of being faithless (perfidus, 1.536). There too she beats her breast ‘again’ (iterum, 1.535), yet she also speaks ‘brand new words’ (novissima verba, 1.539) – a claim that already undermines the truth of Catullus’ ‘final’ lament and plays provocatively with tradition. Ovid’s retelling is a peculiar mix of tradition (iterum) and innovation (novissima).Footnote ³⁹ In Heroides 10, Ariadne again laments her lot, appeals to her memory (memini, 10.92) and accuses Theseus of perjury (periuri, 10.76, cf. perfide … lectule, 10.58), while the rocks echo back the name of Theseus (‘Theseu!’ | reddebant, 10.21–2), troping the poet’s repeated ‘echoing’ of the literary tradition (~ Theseu, Catull. 64.133; Thesea clamabat, A. A. 1.531; Theseu, Fast. 3.473).Footnote ⁴⁰ These Ovidian lines, in the Fasti, Ars Amatoria and Heroides, self-consciously highlight their interaction with Catullus and each other by envisaging this engagement through a series of intertextual metaphors: allusion keyed as memory, echo, iteration, similarity and novelty. Amassed together, these motifs proclaim Ovid’s allusive debts and departures. Like the Alexandrian footnote, they are a crucial tool of literary self-representation.

We could spend much time surveying further examples of such self-consciously figured allusions in Roman poetry – indeed, a comprehensive catalogue of the phenomenon, though a Herculean enterprise, would be an extremely useful resource.Footnote ⁴¹ For now, however, it suffices to note that a range of other self-reflexive tropes have been read in a similar manner in Latin literature.Footnote ⁴² Besides report, echo and memory, scholars have explored the allusive potential of other metaphors, including footsteps, grafting, prophecy, recognition, succession and theft.Footnote ⁴³ Any trope, in short, which suggests a relation of dependence or the voice of authority can easily be co-opted as a metaphor of allusive relationships. And even mere temporal adverbs can evoke diachronic literary relationships, as when Ovid’s Achaemenides is ‘no longer’ roughly clad, as he had been in Virgil’s Aeneid (iam non hirsutus amictu, Met. 14.165 ~ Aen. 3.590–4),Footnote ⁴⁴ or when Statius’ Achelous ‘still’ behaves as he had in Ovid’s Metamorphoses, hiding his mutilated forehead (adhuc, Theb. 4.106–9 ~ Met. 9.96–7).Footnote ⁴⁵

Taken together, these phenomena form a nexus of interrelated tropes for figuring and marking allusion. In general terms, they fit into a broader category of metaliterary ‘marking’, standing alongside signals of generic affiliation, etymological play, acrostics and anagrams.Footnote ⁴⁶ But in their range, variety and adaptability, they stand apart. They may not be as explicit as a modern philologist’s footnotes, but as Jeffrey Wills notes, they ‘function much as quotation-marks do in modern scripts, alerting the reader that some reference is being made, the specific source of which must be deduced in other words’.Footnote ⁴⁷ They offer a useful supplement to the ‘grammar’ of ancient allusion, boosting the intertextual signal. It is thus no wonder that they have been taken up with such scholarly vigour in recent decades.

I.1.3 Problems and Limitations: Terminology and Assumptions

For all this vigour, modern scholarship’s engagement with the phenomenon of allusive marking is not without its problems. First among these is the indiscriminate and uncritical labelling of examples. Ever since Hinds opened his seminal Allusion and Intertext with these devices, the ‘Alexandrian footnote’ and other allusive markers have become a familiar concept in classical scholarship. They now proliferate in discussions of not just Latin, but also later Greek authors.Footnote ⁴⁸ Yet like a commentary’s ‘cf.’, the identification of footnotes and markers can all too often mark the end of the interpretative process, rather than its beginning. These terms have become a convenient shorthand, avoiding the need for closer engagement with the details of a specific allusion. What was once an exciting and liberating insight into the self-consciousness and reflexivity of Latin poets now seems a banal cliché.

The uncritical acceptance of these allusive markers is also visible in the very sobriquet which the ‘Alexandrian footnote’ has received. Given the apparent intellectual demands triggered by such tags, one can understand why Hinds adopted David Ross’ ‘Alexandrian footnote’ to describe the phenomenon. As he argues, the footnoting which we find in Catullus and elsewhere figuratively portrays the poet ‘as a kind of scholar, and portrays his allusion as a kind of learned citation’, ‘encod[ing] a statement of alignment with the academic-poet traditions of Callimachus and the Alexandrian library’.Footnote ⁴⁹ In this, he resembles the views of earlier and later scholars: Geoffrey Kirk argues that φασίν in the Michigan Alcidamas papyrus ‘smacks of post-Alexandrian scholarship’; Adrian Hollis regards fama est as ‘an indication that we are in the world of learned poetry’; Andrew Morrison explores how ‘they say’ statements in Hellenistic poetry form part of the creation of a scholarly and learned narratorial persona; and Jason Nethercut treats Lucretius’ use of the device as evidence of his neo-Callimacheanism.Footnote ⁵⁰ Eduard Norden, moreover, distinguishes between earlier Greek and later Hellenistic/Latin appeals to tradition, arguing that only the latter suggest a reliance on a source, whereas the former are simply earnest assertions of the truth of tradition.Footnote ⁵¹ And Gian Biagio Conte, last of all, has seen in Ovid’s allusive signposting the ‘capacity of Alexandrianism to mirror its art in itself and to revel in its skill’, a means for the poet to highlight ‘the artifice and the fictional devices underlying his own poetic world’.Footnote ⁵² Allusive ‘footnoting’ is regarded as something distinctively Hellenistic, learned and artificial.

Indeed, such a view can be traced back at least as far as the Homeric scholia. When Achilles’ horse Xanthus claims that he and Balius ‘could run swift as the West wind’s blast, which they say [φασ’] is the fleetest of all winds’ (Il. 19.415–16), the A-scholia complain that it is ‘not believable that a horse would say φασίν as if he were a man of much learning’ (ἀπίθανον ἵππον λέγειν φασίν ὥσπερ ἄνδρα πολυίστορα, Σ A Il. 19.416–17 Ariston.). The underlying assumption is that this footnoting tag only befits an erudite scholar, such as Callimachus himself, who is elsewhere described with the very same adjective by Strabo (πολυίστωρ, 9.5.17 = test. 68 Pf.) and in a Life of Aratus (Καλλίμαχου πολυίστορος ἀνδρὸς καὶ ἀξιοπίστου, Achill. vit. Arat. 1 = test. 79 Pf.). Such scholarly baggage is also apparent in another scholiastic note, when the Homeric narrator claims that the eagle, ‘they say’ (φασίν), ‘has the keenest sight of all winged things under heaven’ (Σ Il. 17.674–5 ex.|D):

Here, too, the scholiast associates the use of φασί with erudite, scholarly activity, in this case the careful and precise checking of one’s facts and references (ἀξιοπίστως – the other quality of the Aratean Vita’s Callimachus: ἀξιοπίστου, test. 79 Pf.). Yet it is the following citation which is especially illuminating: the scholiast refers to a passage from Aristotle’s History of Animals to corroborate Homer’s statement on the eagle’s sharp-sightedness (Hist. an. 9.34.620a1–5). Séverine Clément-Tarantino has read this under-appreciated passage as the scholiast’s appropriation of Homer’s generalised φασί ‘to transform it into a “reference” to a precise observation of Aristotle’.Footnote ⁵³ Of course, this does not mean that the scholiast would have interpreted Homer as himself having intended this Aristotelian link: any ancient scholar would have been well aware of the chronological impossibilities of such a view, and we know of other cases where scholiasts provide cross references to later parallels of a specific detail, rather than to earlier sources.Footnote ⁵⁴ Rather than showing that the Alexandrians regarded Homer as a scholiast avant la lettre, it is better to see this scholiastic comment as a reflection of Alexandrian reading practices. When coming across a φασί in a text, the scholiast’s first inclination was to ask ‘who says?’ and find an appropriate source for the fact under discussion – not necessarily Homer’s original ‘source’, but another piece of external evidence to confirm that this is indeed what ‘people say’. The evidence of the Homeric scholia, therefore, suggests that already in antiquity φασί was considered an emblem of erudite scholarship and a spur for readers to go source-hunting. The concept of the ‘Alexandrian footnote’ has a considerable pedigree.

However, this lingering perception of the ‘Alexandrian’ nature of such ‘footnoting’ relies on engrained assumptions about a dichotomy between archaic/classical and Hellenistic/Roman literary cultures.Footnote ⁵⁵ Yet as we noted at the outset, this is an area of considerable contestation, and any literary history (of continuity or change) must be argued for, not assumed. In the case of allusive markers, there is little evidence or argument to restrict the phenomenon a priori to Alexandria and Rome. To support the Hellenistic connection, Hinds notes how an ‘Alexandrian footnote’ mimics ‘very precisely … the citation style of a learned Latin commentary’. But the example he cites (Servius on Aen. 1.242) differs significantly from the ‘Alexandrian footnote’: Servius explicitly names his source (Livy), whereas poetic ‘footnotes’ do not.Footnote ⁵⁶ Despite highlighting the presence of an allusion, they do not point to the specific source – they leave the audience to fill in the gaps themselves. Other Latinists, meanwhile, cite individual lines of Callimachus to prove the ‘Alexandrian’ nature of Roman ‘footnoting’, including the famous μῦθος δ’ οὐκ ἐμός, ἀλλ’ ἑτέρων (‘the tale is not mine, but comes from others’, hAth. 56) or the fragmentary τὼς ὁ γέγειος ἔχει λόγος (‘so the ancient tale has it’, fr. 510 Pf.) and ἀμάρτυρον οὐδὲν ἀείδω (‘I sing nothing unattested’, fr. 612 Pf.).Footnote ⁵⁷ When they are taken out of context, however, it is unclear whether these lines function in the same allusive manner as Hinds’ ‘footnotes’. Nor is it clear why scholars should not cite earlier comparanda: the famous remark from Callimachus’ fifth Hymn is closely modelled on a line from Euripides’ Melanippe the Wise (κοὐκ ἐμὸς ὁ μῦθος, ἀλλ’ ἐμῆς μητρὸς πάρα, ‘the tale is not mine, but comes from my mother’, fr. 484 TrGF),Footnote ⁵⁸ and we can already find similar sentiments elsewhere in fifth-century Greece, such as Pindar’s φαντὶ δ’ ἀνθρώπων παλαιαί | ῥήσιες (‘Ancient tales of men say’, Ol. 7.54–5) or Euripides’ παρὰ σοφῶν ἔκλυον λόγο[υ]ς̣ (‘I have heard stories from wise men’, Hypsipyle, fr. 752g.18 TrGF). These phrases appear to gesture to tradition in a similar manner to Latinists’ Hellenistic and Roman examples, but it would be anachronistic to call them ‘Alexandrian’ or to treat them as scholarly ‘footnotes’. Without further investigation, there seems little immediate justification for considering these markers to be distinctively scholarly, post-classical or (just) self-consciously fictionalising.

Yet this is precisely how the phenomenon is constantly presented. Numerous scholars frame the device in terms that stress its apparent artificiality and self-consciousness: Conte’s ‘reflective allusion’, Hinds’ ‘reflexive annotation’, Alessandro Barchiesi’s self-reflexive ‘tropes of intertextuality’ and Christos Tsagalis’ ‘meta-traditionality’.Footnote ⁵⁹ Others, meanwhile, use the term ‘Alexandrian footnote’ as a catch-all title for every case of allusive signalling, even beyond plain appeals to tradition, making the whole process an archetype of learned and scholarly behaviour.Footnote ⁶⁰ And Matthew Wright has coined ‘metamythology’ as an umbrella term to define ‘a type of discourse which arises when mythical characters are made to talk about themselves and their own myths, or where myths are otherwise presented, in a deliberately self-conscious manner’, a phenomenon which he considers specifically intellectual and destabilising, emphasising ‘the fictionality of myth’.Footnote ⁶¹ The most neutral term that I have encountered is Wills’ ‘external markers’ of allusion,Footnote ⁶² but even this risks making these markers sound too detached, undermining how integral they are to the process of poetic interpretation.

In the face of such terminology, bound up with anachronistic or misleading associations, I will use a new term in this study to describe allusive signposting, namely ‘indexicality’. Amid the mass of pre-existing terms, this is not a gratuitous neologism, but rather a means for us to focus on the essence of this signposting phenomenon: by looking back to the original associations of the Latin index (‘pointer, indicator’), it foregrounds the device’s signposting role.Footnote ⁶³ Rather than seeing such marking as the self-aware technique of a terribly clever and bookish poet, this term instead focuses on the ‘pointing’ function of allusive markers: ‘what’s the point?’, we are invited to ask, and ‘what are we being pointed to?’

Of course, ‘indexicality’ itself is not a new term. It is commonly used in linguistics and the philosophy of language to refer to the manner in which linguistic and non-linguistic signs point to aspects of context (an overarching category that embraces ‘deixis’, a concept more familiar to Classicists).Footnote ⁶⁴ The term is ultimately derived from the American philosopher Charles Peirce’s trichotomy of signs, in which the ‘index’ is a sensory feature that denotes and draws attention to another object with which it regularly co-occurs: smoke indexes the presence of fire, dark clouds index impending rain and a weather vane indexes the direction of the wind.Footnote ⁶⁵

Given the term’s prior usage, some caution is required before introducing it into a new field of study, but I believe that doing so here has numerous advantages. First, Peirce’s index offers an apt analogy and broader context for allusive indexicality: an allusive marker signals the presence of allusion, just as smoke signals the presence of fire. In both cases, it is the frequent co-occurrence of signified and signifier which allows the connection to be perceived and understood.Footnote ⁶⁶ Besides this theoretical background, the term also has valuable thematic and semantic associations in its own right. We have already noted its core etymological connection with ‘pointing’, but there is a further association of ‘index’ which makes it particularly fruitful for this study. In modern English, an ‘index’ most often refers to the catalogue at the back of a book which lists specific words or phrases alongside the page numbers where they can be located (as in this very monograph). Such literary road maps are an apt analogy for allusive marking: an allusive ‘index’ similarly points to a specific element of a larger mythical and literary whole, moving from a single passage back to the larger pathways of myth.Footnote ⁶⁷

Finally, the term ‘indexicality’ also has a practical benefit. It is a convenient and flexible term that can be readily adapted to different parts of speech: the noun ‘index’ (pl. ‘indices’), adjective ‘indexical’, adverb ‘indexically’ and verb ‘to index’. No other neutral word (marker, pointer, annotation, signpost) has such a degree of flexibility. The term thus allows us to discuss this phenomenon with greater nuance and precision. In what follows, I will be studying the allusive ‘indexicality’ of early Greek poets.

I.1.4 The Path Ahead

As we have seen, the ‘Alexandrian footnote’ and other indices of allusion are frequently considered the preserve of Hellenistic and Roman poetic cultures, one of the key attributes that distinguish archaic Greece from later centuries. But a close inspection of many early Greek examples reveals a more complex picture. From Homer onwards, indices were already employed to signpost allusion and to position a poet against their larger tradition. From the very start of the (visible) Greek tradition, indexicality was a well-established phenomenon.

Thankfully, this argument is supported by recent scholarship on early Greek poetry which has already begun to take significant steps in this direction. Archaic epic and lyric have long been read in self-conscious and metapoetic terms.Footnote ⁶⁸ But in more recent years, several scholars have already suggested specific moments in these texts that can be read as knowing indices of allusion, especially in epic. A selective review of examples may help set the scene: stories are explicitly acknowledged as familiar to an audience, as when Circe advises Odysseus in the Odyssey to avoid the path of the ‘Argo known to all’ (Ἀργὼ πᾶσι μέλουσα, Od. 12.70), highlighting Homer’s debts to, and divergences from, the Argonautic saga,Footnote ⁶⁹ or when Odysseus similarly designates Oedipus’ woes and crimes as ‘known to men’ (ἀνάπυστα … ἀνθρώποισιν, Od. 11.274).Footnote ⁷⁰ The transfer of specific individuals’ property appears to signal cases of allusive role-playing: ‘in borrowing Aphrodite’s girdle’ to seduce Zeus in Iliad 14, Hera ‘metapoetically dons Aphrodite’s mantle’, replaying the love goddess’ seduction of Paris and Anchises (Il. 14.188–223),Footnote ⁷¹ while Patroclus adopts both Achilles’ armour and persona in the Iliad (Il. 16.130–44),Footnote ⁷² just as the hero’s son Neoptolemus symbolically succeeds his father by taking his armour in the Little Iliad.Footnote ⁷³ Epic characters’ tears have also been read as presaging future woes which only an audience could know from the larger literary tradition,Footnote ⁷⁴ while catalogues too appear to have been loaded sites for incorporating and contesting other traditions.Footnote ⁷⁵ Even the whole divine framework of Greek literature seems to involve a significant indexical element: what is ‘fated’ is often shorthand for what is (or is at least claimed to be) traditional; counterfactuals explore narrative alternatives that go against tradition; major gods act as figures for the poet; and heroes are often saved because they are ‘destined’ to play a role in future episodes of the tradition.Footnote ⁷⁶

In addition, other specific indices have been identified in these early texts, including cases of echo and family relations. For the former, we could cite the Homeric Hymn to Pan, which pointedly ‘echoes’ a famous nightingale simile from the Odyssey (Hh. 19.16–18 ~ Od. 19.518–21);Footnote ⁷⁷ the ‘echoing cicada’ of the Hesiodic Aspis, which recalls its earlier appearance in Hesiod’s Works and Days (ἠχέτα τέττιξ, Scut. 393 ~ Op. 582);Footnote ⁷⁸ and the presence of ‘Echo’ in Aristophanes’ Thesmophoriazusae, which self-consciously tropes the dramatist’s extensive rewriting of Euripides’ Andromeda (Thesm. 1056–97).Footnote ⁷⁹ As for family relations, we may note the intertextual relationship between specific Homeric Hymns (the sibling rivalry of Hermes and Apollo in HhHerm.; the father–son relation of Pan and Hermes in Hh. 19);Footnote ⁸⁰ Aristophanes’ figuring of Philocles’ Pandionis tetralogy as a derivative ‘descendant’ of Sophocles’ Tereus (Ar. Av. 281–3);Footnote ⁸¹ and Theognis’ substitution of the Hesiodic Αἰδώς (‘Respect’) with her daughter Σωφροσύνη (‘Restraint’), marking his debts to his Hesiodic ‘parent text’ (Thgn. 1135–50 ~ Op. 200).Footnote ⁸² In Attic tragedy more generally, Isabelle Torrance has also argued for a wide range of ‘metapoetically loaded terms’ which are ‘used as triggers for audience recognition of novelties or continuations in relation to earlier sources’: δεύτερος (‘second’), δισσός (‘double’), καινός (‘new’) and μῦθος (‘myth’/‘story’).Footnote ⁸³

These recent approaches give an idea of how fruitful a fuller exploration of allusive marking in early Greek poetry may prove to be. Yet despite these first steps, no previous scholar has offered a comprehensive study of allusive marking in any period, let alone early Greek poetry. Individual examples are normally adduced in support of a specific argument for a specific allusion, which leaves the larger picture remarkably hazy. The scholar who has offered the fullest catalogue to date is Bruno Currie, who concludes his discussion of ‘pregnant tears’ with a list of some allusive markers in Homer and Attic tragedy, focused primarily on ‘poetic memory’.Footnote ⁸⁴ This forms part of his broader argument for continuities in allusive practice across Mesopotamian, Greek and Roman literature.Footnote ⁸⁵

In this book, I intend to provide a more holistic and analytical study of these allusive markers across archaic Greek poetry: I will explore a wider range of examples, incorporating both hexameter and lyric traditions, and I will study them in greater depth, examining their purpose and function, as well as their development across time. I have chosen to focus on the development of three specific indices of allusion in archaic epic and lyric poetry (including iambus, elegy and melos), from Homer to Pindar.Footnote ⁸⁶ The three I have selected represent the indices most commonly identified in literature of later times: first, appeals to tradition and report (the ‘Alexandrian footnote’ proper); second, the allusive force of characters’, narrators’ and audiences’ memories and knowledge; and third, the manipulation of temporality to evoke both former and future literary events. We have already seen all three in Phoenix’s introduction to his Meleager exemplum, but I will demonstrate that they are all deeply embedded in our earliest archaic Greek poetry.

In each chapter, I will explore these indices’ comparable and complementary usages. Due to limitations of space, I cannot cover every example, but the impression I have gained is that a very high percentage of examples of the language of hearsay, memory and time are indexical – a far higher percentage than one might initially suspect.Footnote ⁸⁷ Rather than offer a dry catalogue, I will focus on a selection that illuminates the range of ways in which each index was used in archaic epic and lyric. Every reader will no doubt find some examples more compelling than others. Indeed, we may think of indexicality as a scalar issue – some cases seem to me undeniable, while others may be more open to debate – but the latter are still worth exploring since they open up a range of interesting further possibilities (an issue to which I will return: §iii.3). Nevertheless, taken as a whole, the evidence and interpretations that I advance here will show that all three of these indices were an integral part of the literary tradition from the very start.

Before we turn to each index in turn, I will first outline my methodological approach to allusion in archaic Greek poetry as a framework for this study. This is a controversial topic, and one that raises some different questions to those which face scholars of Hellenistic and Roman texts. It is thus worth spending some time addressing the issues involved.

I.2.1 Mythological Intertextuality

Crucial to Burgess’ case for an ‘oral, intertextual neoanalysis’ is the recognition that there are limits to the formulaic nature of early Greek poetry. As he remarks, ‘typology does not overwhelm the distinctiveness of individual characters and their stories’; otherwise, ‘a myth-teller would be free to gather together a new collocation of motifs every time the story is told. Achilles could wear a lion skin and brandish a club, Odysseus could command the Argo, and Agamemnon could put out his eyes after marrying his mother.’Footnote ¹¹¹ Such a humorous counterfactual highlights the limits of typology, limits which were already recognised in antiquity. Aristotle remarks in the Poetics that one cannot break up ‘transmitted stories’ (παρειλημμένους μύθους), such as Clytemnestra’s death at Orestes’ hands or Eriphyle’s at Alcmaeon’s (Arist. Poet. 14.1453b.22–6). Individual myths and stories clearly contained a steady core of specific elements which did not depend on any particular instantiation. It is to specific motifs of this ‘stable skeleton of narrative’,Footnote ¹¹² Burgess contends, that other songs and performances could allude, even within the traditional and typological context of early Greek epic. For archaic epic, some of these mythological traditions would have doubtless been epic in form; indeed, as Tsagalis notes, the shared performance context ‘would have channelled mythical allusion towards other epic songs performed under similar conditions’.Footnote ¹¹³ Yet they would have also embraced other media, including non-epic storytelling, other kinds of poetry and artistic representations.Footnote ¹¹⁴ The plausibility of this model is reinforced by comparative oral traditions in which we can identify similar allusions to other stories.Footnote ¹¹⁵

Of course, despite the limits of typology, mythological traditions were never entirely static and unchanging, and some have questioned whether any definitive and stable version of past myths ever existed.Footnote ¹¹⁶ If multiple conflicting versions were in circulation, even within the very same poem, and if poets were free to add innovative elements to mythical paradeigmata to fit their immediate contexts, how can we determine to which version of a myth poets might be alluding in any given case, or even which of many potential versions their original audiences might have been familiar with or considered ‘canonical’?Footnote ¹¹⁷ This is a pressing concern, and one which is too often glossed over by neoanalysts. Yet one must equally be wary of exaggerating the significance of such discrepancies in the archaic mythological record. Where differences occur, they tend to be minor and superficial for the overall narrative trajectory, and it is often only the instigator of an action which changes, not the action itself: Thetis is still given to Peleus, whether by the gods (Il. 18.84–5), Zeus (Il. 18.432) or Hera (Il. 24.60); Coroebus, a suitor of Cassandra, is still killed, whether by Neoptolemus (in the ‘majority version’, ὁ πλείων λόγος) or Diomedes (according to the poet ‘Lescheos’, Λέσχεως: Paus. 10.27.1 = Il. Parv. fr. 24 GEF); Polyxena still dies, whether through wounds inflicted by Odysseus and Diomedes in the sack of Troy (Cypria fr. 34 PEG) or as a sacrifice on Achilles’ tomb (Il. Pers. arg. 4c GEF); and Astyanax is still thrown from the city walls, whether by Neoptolemus (Il. Parv. fr. 29 GEF) or Odysseus (Il. Pers. arg. 4a GEF).Footnote ¹¹⁸ In all four of these cases (Thetis’ marriage and the deaths of Coroebus, Polyxena and Astyanax), we have a fixed, unalterable event of the Trojan war narrative, even if its precise details varied. As Burgess has remarked, ‘While it would be mistaken to insist that the details of any one manifestation of a myth were always present in every telling of that myth, it is also clear that Greek myth was remarkably stable in the presentation of the sequences of major actions that constituted any given story.’Footnote ¹¹⁹ The same view was also apparently dominant in antiquity. When Sophocles has Agamemnon die in the bath (El. 445) rather than at the table as in Homer (Od. 4.535), the scholia dismiss the inconsistency (Σ S. El. 446):

Whether Agamemnon was killed in the bath or at a feast, it ultimately does not matter: he died either way, and that is the fixed element of the myth.Footnote ¹²¹ It is thus possible, with appropriate care and caution, to reconstruct the core details of a mythological narrative, what Kullmann would call a Faktenkanon or Burgess a fabula, a constellation of fixed narrative events with which the Homeric and other later poems could allusively engage.Footnote ¹²²

Given the typological oral environment of early Greek epic, we should largely expect allusions to such fabulae to be based around repeated key themes and motifs, rather than extensive verbal repetition. The foremost example of such motif-based allusion is the Iliad’s evocation of the ‘death of Achilles’ fabula, which lies at the heart of the second half of the poem and has been extensively studied by numerous scholars. The allusion is not based primarily on verbal correspondence, but rather on large-scale motif transference, as a whole series of episodes from the fabula of Achilles’ death are redeployed in another context.Footnote ¹²³ On a larger scale, moreover, the whole myth of the Trojan war appears to be constructed around an extensive chain of such interlocking fabulae: the sack of Andromache’s Thebe foreshadows and parallels that of Troy; Paris’ return from Sparta to Troy with Helen is mirrored by the itinerary of Menelaus’ own nostos after reclaiming his wife; and the whole war is framed by a chilling pair of human sacrifices, Iphigenia’s sacrifice at Aulis (Cypr. arg. 8 GEF) paralleling Polyxena’s at Troy (Il. Pers. arg. 4c GEF): in both cases, a king’s unmarried daughter is sacrificed as the prelude to the Greek fleet’s departure.Footnote ¹²⁴

As another example of how to conceive of such fabula-based allusion, we could cite the famous ‘Nestor’s cup’ inscription, our earliest known case of Greek intertextuality. A Rhodian kotyle, discovered in a late eighth-century Ischian cremation burial, bears the following inscription in Euboean script (SEG 26.1114 = CEG 454):

Νέστορός : ἔ̣[ε̄ν τ]ι̣ :Footnote ¹²⁵ εὔποτ[ον] : ποτριον.

hὸς δ’ ἂν το͂δε πίε̄σι : ποτε̄ρί[ō] : αὐτίκα κε͂νον

hίμερος hαιρσει : καλλιστε̣[φά]ν̣ō : Ἀφροδίτε̄ς.

These verses, composed of a likely iambic trimeter and two dactylic hexameters (a metrical mixture typical of parody),Footnote ¹²⁶ set up a humorous and pointed opposition between archaic epic and the world of the symposium.Footnote ¹²⁷ The humble, small clay kotyle that bears the inscription is contrasted with the epic Nestor’s large and elaborately wrought drinking vessel familiar to us from the Iliad (Il. 11.632–7). The precise nature of the contrast depends on how we supplement the first line, a lacuna which continues to vex scholars. With ε̣[ἰμ]ί̣ (‘I am the cup of Nestor’), the kotyle identifies itself as Nestor’s cup, a humorous incongruity given its small scale and modest nature.Footnote ¹²⁸ With ἔ̣[ε̄ν τ]ι̣ (‘Nestor had a cup’), the kotyle explicitly differentiates itself from its epic predecessor, self-consciously aligning itself with sympotic erotics in place of epic heroics.Footnote ¹²⁹ In either case, however, humour emerges from the disparity between the humble Ischian cup and the epic Nestor’s grand goblet, which only he had the strength to lift (Il. 11.632–7):Footnote ¹³⁰

Many scholars have suspected a precise allusion to this Iliadic scene in the Ischian inscription, taking it as evidence that our version of the Iliad was already well known in the Greek world of Euboea and its colonies in the late eighth century.Footnote ¹³¹ Given our limited evidence for eighth-century literary culture, such a direct intertextual relationship cannot be ruled out, but it should be stressed that the cup’s allusion is not based on any verbal correspondences with our Iliadic passage, and its diction departs significantly from Homeric usage.Footnote ¹³² In reality, the parallel depends only on similarities of theme and topic: the knowledge required for the allusion to work is simply that Nestor possessed a large and ornate cup, awareness of which could derive from many other sources besides our Iliad.Footnote ¹³³

Indeed, scholars have not refrained from proposing other potential epic ‘sources’ for the cup’s allusion: Stephanie West suggests epic poetry on the exploits of Nestor’s youth,Footnote ¹³⁴ while Georg Danek proposes the scene from the Cypria in which Nestor hosted Menelaus (Cypr. arg. 4b GEF) and apparently encouraged him to drink wine to scatter his ‘cares’ (Cypr. fr. 18 GEF).Footnote ¹³⁵ It would be misleading, however, to pinpoint any of these as the specific ‘source’ of the cup’s allusion, given that Nestor appears to have been associated with lavish hospitality, plentiful drinking and a large, ornate cup in many texts and traditions, especially in his capacity as an adviser and strategist. Of course, drinking vessels, like many other material objects, were highly prized in the world of Greek epic as a source of prestige and authority,Footnote ¹³⁶ and elaborate descriptions of them were a traditional feature of not just Greek, but also Near Eastern poetic traditions.Footnote ¹³⁷ Yet Nestor’s association with drinking ware transcends such typological norms. In addition to the Iliad and Cypria, we could cite Odyssey 3, where Pylos is presented as a place of feasting and merriment (Od. 3.32–66). Nestor’s son Peisistratus presents Telemachus and Athena-Mentor with a beautiful golden cup for prayer (χρυσείῳ δέπαϊ, Od. 3.41; χρύσειον ἄλεισον, 3.50, 53; καλὸν δέπας ἀμφικύπελλον, 3.63), a cup which Peter Bing has suggested could be the very same as in the Iliad, given that the goblet there is said to have been brought from home (ὃ οἴκοθεν ἦγ’ ὁ γεραιός, Il. 11.632).Footnote ¹³⁸ Athenaeus’ later mention of a ‘cup of Nestor’ dedicated to Artemis in Capua, not far from Ischia, might also suggest a local tradition surrounding the heroic Nestor’s cup which could have already been circulating in the region in archaic times.Footnote ¹³⁹ Nestor was thus closely associated with a large, ornate cup throughout early Greek epic, symbolising his panache for hospitality, storytelling and advice-giving – a traditional association that Iliad 11 itself presupposes.

Rather than detecting a precise engagement with the Iliad or any other specific text in the Pithecusan inscription, it is thus better to see an allusion to an established feature of the fabula of the hero’s life.Footnote ¹⁴⁰ The inscription evokes not the specific Nestor of the Iliad, but rather the Nestor of tradition at large, known for his many instances of hospitality and feasting. In so doing, it situates its humbler self within the literary tradition, setting its brief epigrammatic form against the grandeur of epic.Footnote ¹⁴¹ This allusion can be taken as an archetype of what we might usually expect in archaic Greek epic itself: an engagement with the themes, motifs and narrative events of other mythological traditions (fabulae), rather than precise verbal echoes of another specific poem.

Nevertheless, although the majority of archaic mythological allusions would function in this manner, an oral poetic environment does not entirely preclude the possibility of verbal allusion and quotation, even when we are talking of mythological traditions, not fixed poems. As Burgess has again demonstrated, certain phraseology could become associated with specific fabulae, characters or narrative contexts and then be allusively redeployed in other settings. As Homeric examples, he offers the phrase μέγας μεγαλωστί, which appears to be connected with the fabula of Achilles’ death (Od. 24.40, Il. 18.26), and the language associated with Astyanax’s fate, which is proleptically evoked in the Iliad (Il. Parv. fr. 29.3–5 GEF; Il. 6.467–70, 24.735).Footnote ¹⁴² As a further example, we could cite the Iliadic description of the hundred-hander Briareus as ‘greater in strength than his father’ (ὁ γὰρ αὖτε βίῃ οὗ πατρὸς ἀμείνων, Il. 1.404), a phrase which seems to allude to the fabula of Achilles’ birth and the prophesied supremacy of Thetis’ offspring.Footnote ¹⁴³ These are not cases of one text quoting another, but rather instances in which the use of certain phrases and language may evoke specific episodes and characters from the fixed fabulae of the mythological tradition.Footnote ¹⁴⁴ Such examples still face the usual challenges encountered by any neoanalytical interpretation (especially the questions of priority and direction of influence: see §i.2.2 below),Footnote ¹⁴⁵ but Burgess’ arguments offer an attractive framework for exposing the allusive potential of some early epic repetitions. Most repetitions in epic poetry are, of course, likely to be typological in character, so most of these cases of pointed repetition will involve rarely attested phrases which have come to be associated with specific and identifiable contexts or individuals.Footnote ¹⁴⁶

Early Greek poetry, therefore, should be regarded as able to engage allusively with specific mythological traditions on the levels of both motif and phraseology. In a fluid oral poetic environment, where specific episodes would have been repeatedly re-performed, such engagements were likely multidirectional, as various traditions and story patterns came to influence one another,Footnote ¹⁴⁷ but we are no longer in a position to discern such intricacies. Currie has objected that this model restricts us to ‘an impersonal and anonymous model of allusion’, in which we cannot conceive of ‘individually authored compositions’ setting themselves apart from others.Footnote ¹⁴⁸ But this is far from the case. Many of the interpretations that follow will show just how sophisticated and agonistic the Homeric poems were in setting themselves apart from the whole tradition.Footnote ¹⁴⁹ Even if they are not always alluding to a specific poem, this does not deny their own poetic integrity. Nor is this approach designed in principle to rule out the possibility of direct interaction between texts at an early date (see further §i.2.3 below). Rather, it prevents us from thinking anachronistically of a mass of neat, self-contained, easily distinguishable epics interacting with each other as the norm in the archaic period.Footnote ¹⁵⁰ Instead, when dealing with the lost poetic traditions of early Greek poetry, the framework of fabula-based allusion and mythological intertextuality best accounts for the fluid and flexible nature of oral traditions. It is the default paradigm that I will apply in this study.

In what follows, I will employ the language of both allusion and intertextuality to describe this phenomenon, following the flexible practice I outlined above (§i.1). This is not unusual,Footnote ¹⁵¹ but some scholars will doubtless object to one or both of these terms. Some would prefer to restrict ‘allusion’ to precise connections between fixed poems; while for others even ‘mythological intertextuality’ may sound a little misleading or paradoxical, especially since we are not talking here about interaction with specific ‘texts’. Nevertheless, I believe there are good reasons for retaining these familiar nouns. First, ‘allusion’ foregrounds the design that I see and interpret in early Greek poetry’s engagements with traditional fabulae.Footnote ¹⁵² Second, the idea of ‘mythological intertextuality’ is in many respects closer to and thus authorised by Julia Kristeva’s original conception of ‘intertextuality’, in which any cultural product, and not just a literary work, could be considered a ‘text’.Footnote ¹⁵³ And third, this familiar nomenclature is extremely useful, since it highlights the considerable similarity between this kind of fabula-based allusion and the text-based allusion with which Classicists are more familiar. Both involve a reference to another external source (in contrast to intratextuality: allusion within the bounds of a specific poem or corpus). By employing the terms here, I thus acknowledge this essential continuity: in both ‘mythological’ and ‘textual’ intertextuality, the underlying allusive process is the same, even if the target of the allusion is different in each case.Footnote ¹⁵⁴

I.2.2 Reconstructing Lost Traditions

Despite its methodological advantages, this framework of mythological intertextuality still has to deal with one crucial obstacle that faces any neoanalytical undertaking: namely, our limited access to the rich range of traditions and poems that once populated archaic Greece. Given how little we now have, either in full or in fragments, our gaze is extremely blinkered. In the case of the Iliad and Odyssey, our earliest extant Greek texts, this limitation is particularly pressing: how can we talk of allusion in these poems if we have no clear window onto what came before them?Footnote ¹⁵⁵

To escape this paucity of evidence, some scholars have recently looked beyond the Greek canon to Near Eastern (and especially Mesopotamian) narratives as a possible ‘source’ of interaction. Numerous parallels of technique, motif and theme have long suggested some kind of connection between Greek and Near Eastern texts, but it remains hotly debated how best to frame the relationship.Footnote ¹⁵⁶ A growing recent trend, however, is to see Homer and Hesiod ‘directly’ and ‘intentionally’ alluding to the likes of Gilgameš and the Enuma Eliš.Footnote ¹⁵⁷ This is an exciting possibility, but there is need for caution at the very least. Archaic epic is attentive to non-Greek cultures and foreign languages (e.g. Il. 2.803–4, 4.433–8; Od. 1.183; HhAphr. 113–16), but as Johannes Haubold has noted, the genre (unlike fable) does not advertise itself as engaging with Near Eastern traditions – indeed, the Homeric conception of the world mentions no human society east of Cilicia and the PhoeniciansFootnote ¹⁵⁸ – and historical Greeks, even if they were aware of such traditions, were apparently not concerned with spotting references to them.Footnote ¹⁵⁹ Nor, we might add, were they even interested in mentioning them: Γίλγαμος appears only once in extant Greek literature, and only then nearly a millennium after Homer at the turn of the second/third centuries ce, in a context divorced from his Mesopotamian epic adventures.Footnote ¹⁶⁰ Despite the broad cultural influence of the Near East on archaic Greece, it is very difficult not to take the general silence of Greek audiences and writers as a sign of disinterest in (or ignorance of) these foreign myths. Moreover, many of the underlying Greek–Mesopotamian literary parallels are often not ‘sufficiently compelling’ (Currie’s own criterion: (Reference Currie2016) 174) or close enough to necessitate or even encourage a direct and/or allusive connection. Although it is ultimately a subjective matter, alternative explanations for similarity often seem more plausible, usually involving closer and more meaningful parallels within a Greek context.Footnote ¹⁶¹ The converted would of course respond that allusion always works through creative adaptation and reworking, so we should not expect precise similarity.Footnote ¹⁶² But differences can eventually become so overwhelming that it simply becomes misleading to continue postulating direct allusion.Footnote ¹⁶³

More fundamentally, however, this allusive model struggles to give a convincing account for such direct reception of the Mesopotamian poems across time and space. Undoubtedly, ‘historical connections and cultural influence are abundantly attested between archaic Greece and the ancient Near East’, visible in the archaeological, iconographic and inscriptional records, as well as in the Semitic origins of the Greek alphabet.Footnote ¹⁶⁴ And within these broader channels of interaction, it is inevitable that Near Eastern stories would have had some influence on Greek narratives and thought over centuries of contact.Footnote ¹⁶⁵ But the ‘Near East’ is not one monolithic whole: it is a conventional term to describe a wide range of different cultures, languages and traditions, with varying proximity to the Greek-speaking world. Given the vast distance between Mesopotamia and the Greek-speaking world, I consider it implausible that Greek audiences would have been directly and intimately familiar with Mesopotamian texts such as Gilgameš and able to recognise and detect allusive reworkings of them in performance.Footnote ¹⁶⁶ Scholars have hypothesised the schooling of Greek poets in the East, the arrival of bilingual bards to Greece, interactions in a festival context and even Greek translations of Mesopotamian poetry, all of which are certainly not impossible.Footnote ¹⁶⁷ But given the silence of our epic sources, any of these ‘solutions’ requires a rather large leap of faith – one which I am not currently prepared to take. I thus side with those who view parallels with Mesopotamian texts as the result of long-term interaction and evolution,Footnote ¹⁶⁸ extremely valuable for tracing the distant prehistory of Greek poetic motifs – and for identifying the distinctive and unique ‘narrative choices’ made by each individual text or traditionFootnote ¹⁶⁹ – but less so for those interested in allusion and intertextuality as a phenomenon of performance and reception.Footnote ¹⁷⁰

In that case, our evidence for the earlier traditions with which Homer and Hesiod were engaging remains severely restricted. We have no definite knowledge of what tales pre-existed them, or of what specific versions of these tales were in circulation. We are thus compelled to follow the common neoanalytical approach of reconstructing the contours of pre-existing myths and traditions (but not poems, cf. §i.2.1 above) from the scraps we have: internal evidence within our extant poems, alongside later artistic, poetic and prose sources. Considerable caution is required in this endeavour, however – and much more than most neoanalytical scholars acknowledge. In particular, we should note two significant caveats.

The first is the post-Homeric date of our evidence, which raises the possibility that these later texts are simply reacting to and shaping their narratives against the Homeric poems themselves. Later poems may allusively rework a Homeric motif or simply add meat to the narrative bones of a passing Homeric reference – in which case, they cannot reliably provide us with secure, unmediated access to the coveted pre-Homeric tradition.Footnote ¹⁷¹ This is especially true of the Epic Cycle, our evidence for which is late and limited, based on scattered fragments and the summaries of Proclus from the second or fifth century ce.Footnote ¹⁷² It is striking how much early Homeric neoanalysis failed to acknowledge this problem and simply assumed as ‘fact’ that the poems of the Epic Cycle reflect pre-Homeric tradition.Footnote ¹⁷³ Recent attempts to treat evidence of any date as an authentic ‘multiform’ are equally problematic, since they collapse chronology and impugn later storytellers’ potential for invention.Footnote ¹⁷⁴ In reality, the later our sources date in time, the greater our problems become. Attempts to reconstruct the traces of a pre-Odyssean Argonautic tradition from Apollonius’ Argonautica are extremely problematic given how heavily steeped that epic is in the reception and study of both Homeric poems,Footnote ¹⁷⁵ while the content and attributions of prose mythographers cannot always be taken at face value.Footnote ¹⁷⁶ Similar difficulties arise, moreover, when the Iliad and Odyssey are mined for evidence of earlier traditions with which they might interact, where there is a latent danger of circularity.Footnote ¹⁷⁷ The chronological limitations of our evidence are thus a major obstacle, and one which must be taken seriously.

The other major challenge faced by neoanalysts is the subjectivity of their arguments for motival priority: the claim that the non-Homeric instance of a motif must be the original and primary one because it is more natural, suitable and appropriate than its Homeric counterpart.Footnote ¹⁷⁸ Not only must the parallel motif in question prove to be more than just typological, but these arguments for fittingness frequently lack any objective, clearly defined criteria. In particular, are we justified in assuming that a motif’s original use will be more suitable and better-fitting than later adaptations, or could a later poet not adapt and improve the application of a pre-existing motif in a new context?Footnote ¹⁷⁹ Arguments for a motif originally ‘belonging’ to one specific myth or story must thus be treated with considerable circumspection.

Neither of these issues is insurmountable, however, especially when handled with due caution. In the case of using post-Homeric evidence, we should be wary of unduly exaggerating the primacy of Homer, at least at an early date. Among many scholars, Burgess has noted that early Greek artists reflected non-Homeric cyclic themes ‘much earlier and much more often than they reflected Homeric themes’, suggesting that it was not until the late sixth century that the Iliad and Odyssey came to dominate the tradition. In that case, ‘post-Homeric evidence for the pre-Homeric tradition is not necessarily contaminated by Homeric influence, at least not at an early date’.Footnote ¹⁸⁰ Of course, early epic chronology is a disputed field of research, but this observation at least offers the opportunity for us to see in other sources evidence of traditions that may well have developed before the Homeric poems rose to pre-eminence.Footnote ¹⁸¹ More generally, given the limited possibilities for the diffusion of epics at an early date, both through performance and literary circulation, Burgess has also noted that ‘relatively late poems are not necessarily influenced by relatively early poems’ and that chronologically ‘“late” poems may well represent mythological traditions that precede “early” poems’.Footnote ¹⁸² Given this situation, it would be overly reductive and dogmatic to preclude the possibility that some post-Homeric evidence might reflect pre-Homeric traditions.

In that case, neoanalytical arguments of priority remain our best tool for identifying such potential pre-Homeric traditions. A degree of subjectivity is impossible to escape (as indeed it is in any allusive interpretation of poetry), but there are some cases in which it would be difficult to deny the transfer of motifs from one character or situation to another. This is especially the case when a motif is particularly rare, or when we encounter a uniquely shared combination of motifs which we can plausibly argue is more appropriate in one context than another. A commonly cited intratextual example within the Iliad is the relationship of Diomedes and Achilles. The pair share numerous similarities, from their Hephaestan armour (Il. 8.195 ~ Il. 18.369–617, 19.10–23) and the supernatural fire that surrounds their heads (Il. 5.4–8 ~ 18.205–14, 225–7) to their theomachic pretensions (Il. 5.330–54, 841–63 ~ Il. 21.212–382) and support from Athena during their respective aristeiai (Il. 5.121–33, 290–1, 793–859 ~ Il. 20.438–40, 21.304, 22.214–99).Footnote ¹⁸³ All these traits ‘fit’ Achilles better, relating to the poem’s central protagonist at the climax of the narrative. And such connections even extend beyond the strict narrative confines of the Iliad, since Diomedes also appears to foreshadow Achilles’ impending death: the Trojan women pray that he might die at the Scaean gates (Il. 6.305–7), the site of Achilles’ future demise (§iii.2.4), and he is injured in the foot by Paris (Il. 11.369–83), suffering the same injury from the same Trojan that would eventually prove Achilles’ undoing (Il. 22.359–60; §ii.2.4).Footnote ¹⁸⁴ Diomedes is thus an ‘anticipatory doublet’, or altera persona, of Achilles, displaying elements that ‘belong’ primarily to the Phthian hero. In a case such as this, arguments for priority are extremely plausible and enrich our interpretation of the poem. Diomedes exhibits these traits in the Iliad first, but they prove more at home when later repeated of Achilles. In the same way, we can detect cases of motival priority between texts: instances of a motif that appear to us first in Homer may rework other pre-existing traditions or fabulae, even if they are only attested for us at a later date.

Of course, each individual case of such motif transference will have to be assessed on its own merits and treated with extreme care. In some cases, priority might not always be discernible, and we may sometimes suppose that different examples of a motif developed simultaneously through mutual interaction. But in at least some instances, this approach will help us exploit later evidence as a guide for potential earlier literary traditions with which Homer and later poets could interact. After all, as Jim Marks has observed, ‘even if the non-canonical evidence … is “post-Homeric,” it still offers our best approximation of the kinds of stories that would have been known to poets … and to their audiences’.Footnote ¹⁸⁵ Certainty is impossible, but it would be overly defeatist and far less interesting to ignore categorically the hints and clues we have from later sources.

I.2.3 From Myth to Text

The question remains, however, when and how we should transition from this framework of mythological intertextuality to one of full textual intertextuality. And more generally, to what extent can we detect a development in allusive practices between the eighth and fifth centuries bce?

Again, there are no simple answers to this question. But when we turn to Greek lyric poetry of the seventh to fifth centuries bce, we find an increasingly clear sense of authorship, literary history and engagement with specific texts and authors over time.Footnote ¹⁸⁶ This is manifested above all in poets’ direct naming of themselves and their predecessors.Footnote ¹⁸⁷ Numerous testimonia attest to a growing phenomenon of citing other poets by name. Already in the mid-seventh century, Archilochus (fr. dub. 303) and Callinus (fr. 6) are said to have ascribed the Margites and Thebais respectively to Homer, while we are told that Alcman in the late seventh century made explicit mention of the poet Polymnestus of Colophon (fr. 145). In the sixth century, a poem of Sappho was apparently composed in response to Alcaeus (fr. 137), while Stesichorus is said to have blamed Hesiod and Homer (fr. 90.1–6), attested that Xanthus pre-dated him (fr. 281) and ascribed the Shield of Heracles to Hesiod (fr. 168). At the dawn of the fifth century, Bacchylides apparently called Homer a native of Ios (fr. 48); Simonides is said to have compared Hesiod to a gardener and Homer to a garland-weaver (Gnomol. Vatic. Gr. 1144 = T91b Poltera) and to have mentioned a Corinthian poet called Aeson (fr. 609 PMG);Footnote ¹⁸⁸ Timocreon of Rhodes allegedly composed a lyric poem of abuse against Simonides (Suda τ 625 = T1 Davies); and Pratinas reputedly made direct mention of a number of his musical predecessors: Olympus, Thaletas and Xenodamus (713 PMG). Olympus apparently featured again in Pindar (fr. 157), who is also said to have mentioned Sacadas of Argos (fr. 269), called Homer a Chian and Smyrnaean (fr. 264) and ascribed the Cypria to him (fr. 265). Alongside literary critics’ and philosophers’ engagement with Homer from the late sixth century onwards (e.g. Theagenes of Rhegium, Xenophanes, Heraclitus), this evidence suggests an increasingly strong awareness of distinct and recognisable poetic predecessors.Footnote ¹⁸⁹

Of course, these examples are largely based on indirect testimonia and may thus only reflect the inferences and biographical fantasies of later readers.Footnote ¹⁹⁰ Chamaeleon’s claim that Stesichorus ‘blamed’ both Homer and Hesiod (fr. 90.1–6), for example, could have simply been extrapolated from the poet’s general criticism of the epic tradition and its myths (e.g. fr. 91a), rather than being based on any direct naming of either poet in Stesichorus’ poetry.Footnote ¹⁹¹ In some cases, too, potential textual corruption complicates our assessment of the evidence.Footnote ¹⁹² Yet despite these problems, it would be excessively sceptical to dismiss every single one of these testimonia. Not only are some independently confirmed by other evidence,Footnote ¹⁹³ but the general picture they paint is reinforced by numerous examples from our extant texts and fragments in which poets do directly name their forebears.

Alcman may again offer an early example from the seventh century: his description of apparent poetic novelties ([σαυ]μαστὰ δ’ ἀνθ[ρώποισ(ι) …] | γαρύματα μαλσακὰ̣ […] | νεόχμ’ ἔδειξαν τερπ̣[, Alcm. 4 fr. 1.4–6) has plausibly been interpreted as a reference to poetic predecessors, potentially including Terpander (τερπ̣[, 4 fr. 1.6) and Polymnestus (cf. Alcm. fr. 145).Footnote ¹⁹⁴ Yet it is in the sixth and fifth centuries that extant examples proliferate: Alcaeus explicitly attributes the maxim that ‘property makes the man’ to Aristodemus, one of the seven sages (Ἀριστόδαμον, fr. 360) and seems to have addressed Sappho directly (ἰόπλοκ’ ἄγνα μελλιχόμειδε Σάπφοι, fr. 384).Footnote ¹⁹⁵ Solon explicitly quotes and criticises a verse of Mimnermus, whom he identifies directly by his patronymic (Λιγιαστάδη, fr. 20).Footnote ¹⁹⁶ Hipponax directly names Bias of Priene, another of the seven sages (Βίαντος τοὺ Πριηνέως, fr. 123). Xenophanes criticises Homer and Hesiod by name for their portrayal of the gods (Ὅμηρός θ’ Ἡσίοδός τε, D8 L–M; cf. Ὅμηρον, D10). Epicharmus quotes Ananius (fr. 51 K–A) and names Aristoxenus of Selinus as the first to introduce a certain type of iambus (fr. 77 K–A). Bacchylides cites a saying of Hesiod (Βοιωτὸς ἀνὴρ … Ἡσίοδος, Bacchyl. 5.191–4). Corinna explicitly finds fault with Myrtis for competing with Pindar (Μουρτίδ’ … Πινδάροι, fr. 664a).Footnote ¹⁹⁷ Simonides quotes Pittacus’ saying that it is difficult to be good (τὸ Πιττάκειον, fr. 542), critiques Cleobulus’ epigram on Midas’ tomb (Κλεόβουλον, fr. 581), acknowledges Homer and Stesichorus as sources for his account of Meleager (Ὅμηρος ἠδὲ Στασίχορος, fr. 564) and even attributes to the ‘man from Chios’ a hexameter line from the famous leaves simile of Iliad 6.146–9 (Χῖος … ἀνήρ, fr. eleg. 19.1–2, cf. ἀν̣[δρός], 11.15–18; Ὅμηρ̣[ος], 20.14).Footnote ¹⁹⁸ Yet it is Pindar who refers to the greatest range of predecessors, including Archilochus (Ol. 9.1–2, Pyth. 2.54–6), Hesiod (Isth. 6.66–8), Homer (e.g. Pyth. 4.277–8, Nem. 7.20–1, Isth. 4.37–9, Pae. 7b.11–12), Polymnestus of Colophon (fr. 188), Terpander (fr. 125)Footnote ¹⁹⁹ and perhaps also Alcman,Footnote ²⁰⁰ Arion (Ol. 13.18–19) and Xenocritus of Locri ([Λο]κρῶν τις, fr. 140b.4).Footnote ²⁰¹ In some cases, these Pindaric references can even be traced to specific lines of other extant poems (e.g. Isth. 6.66–8 ~ Op. 412; Pyth. 4.277–8 ~ Il. 15.207; Nem. 7.20–1 ~ Od. 1.4).Footnote ²⁰² And to all these examples we could also add instances of poets’ self-naming (e.g. Ἡσίοδον, Theog. 22; Ἀλκμάν/Ἀλκμάων, Alcm. frr. 17.4, 39.1, 95b; Ψάπφ’/ Ψάπφοι, Sapph. frr. 1.20, 65.5, 94.5, 133.2; Ἱππῶναξ etc., Hipponax frr. 32.4, 36.2, 37, 79.9, 117.4) and especially Theognis’ assertion of his personal ownership of his collection of verses in his seal poem (Θεύγνιδός ἐστιν ἔπη | τοῦ Μεγαρέως, Thgn. 22–3).Footnote ²⁰³ Alongside the increasing evidence for the use of writing and literacy throughout the sixth and fifth centuries,Footnote ²⁰⁴ all these examples suggest that we are very much justified in seeing increasingly greater intertextual engagement with specific texts in lyric poetry.Footnote ²⁰⁵

In practice, however, any discussion of allusion in Greek lyric still faces many of the same issues that we have already encountered above, not least whether to prioritise engagement with the limited range of texts we have access to, and how we should negotiate the boundaries of the typological and the specific.Footnote ²⁰⁶ When Archilochus describes his seduction of Neoboule in the first Cologne epode (fr. 196a), for example, should we conceive of this as a pointed rewriting of Hera’s seduction of Zeus in Iliad 14 or a broader engagement with the epic type-scene of seduction?Footnote ²⁰⁷ Similarly, does Mimnermus fr. 2 allude to the leaves simile of Il. 6.146–9 or to a traditional analogy that is found frequently elsewhere, both in Homer and later texts?Footnote ²⁰⁸ So too with the Lesbian poets: does Sappho fr. 44 evoke a patchwork of Iliadic passages or a wider range of Trojan traditions, including not just Hector and Andromache’s wedding, but also that of Paris and Helen?Footnote ²⁰⁹ And does Alcaeus fr. 347 closely rework Hesiod’s description of summer in the Works and Days (Op. 582–96) or draw independently on a traditional body of seasonal song, attested elsewhere by a parallel description in the Hesiodic Aspis (Scut. 393–7)?Footnote ²¹⁰ In all these and other cases, we should be wary of unduly privileging the few texts that we still possess over the broader tradition, but this should not stop us from arguing for direct allusion when the context and content of the passages justify it. In the case of Alcaeus’ summer scene, for example, the parallels between the Alcaean and Hesiodic passages are so numerous and precise that a merely indirect connection seems improbable. On closer examination, the arguments for a traditional motif are also not particularly compelling: the Aspis parallel passage is more likely another ‘echo’ of the Works and Days (even self-consciously marked as such through the recurrence of the ‘echoing’ cicada: ἠχέτα τέττιξ, Scut. 393),Footnote ²¹¹ a means to increase its own ‘Hesiodic’ texture, rather than an independent manifestation of a recurring motif. In this case, it is plausible to read Alcaeus’ fragment as a pointed appropriation of Hesiod’s paraenetic posturing, marking his generic difference to and distance from Hesiod’s far longer didactic epic.

In recent years, however, several scholars have attempted to restrict the origins of extensive textual intertextuality to the time of Stesichorus in the sixth century, a poet whom they perceive as marking a particularly significant watershed in the development of poetic allusion.Footnote ²¹² It is true that Stesichorus does offer us several plausible cases of precise engagement with Homeric epic, often with apparently rarer moments of Homeric narrative: the comparison of Geryon’s drooping head to a poppy echoes the Iliad’s similarly poignant description of Gorgythion’s head (Geryoneis fr. 19.44–7 ~ Il. 8.306–8); Geryon’s mother baring her breast recalls Hecuba’s same action before Hector (Geryoneis fr. 17 ~ Il. 22.79–83); and Telemachus’ departure from Sparta replays events from the Odyssey (Nostoi fr. 170.1–11 ~ Od. 15.1–184).Footnote ²¹³ Such precise engagement can also be traced in Stesichorus’ successors, not only in the three famous epinician poets (Simonides, Bacchylides and Pindar), but also Ibycus, whose Polycrates Ode (S151) plausibly makes sophisticated use of the Iliadic Catalogue of Ships and Hesiod’s Works and Days.Footnote ²¹⁴

However, to posit Stesichorus as a dramatic point of change overplays the novelty of such precise references and underplays the significance of earlier Stesichorean predecessors such as Alcaeus.Footnote ²¹⁵ We have already noted his precise verbal engagement with Hesiod, but we could also cite his fr. 44, which appears to evoke the key theme of the Iliad: in its fragmentary state, we see a son call to his Naiad mother, who then supplicates Zeus on the subject of her son’s wrath (μᾶνιν, fr. 44.8 ~ μῆνιν, Il. 1.1). It is difficult to deny a reference to our Iliad or at least an Iliadic tradition here, especially given that poem’s unusual and loaded use of the noun μῆνις (cf. §i.2).Footnote ²¹⁶

Moreover, scholars’ sceptical arguments about earlier texts can also be turned against their own Stesichorean examples. In the case of Geryoneis fr. 19, for example, Adrian Kelly himself notes that flower similes are common in early Greek epic, while the image of each poppy simile is considerably different: in Stesichorus, the flower sheds its leaves; while in Homer, it is weighed down by the weight of fruit and rain.Footnote ²¹⁷ In addition, we could add that arrows likely played a larger role in other epic material, especially in traditions featuring Philoctetes and Heracles, so the shared instrument of death in these two scenes need not be particularly distinctive or marked. And Kelly’s argument that the Iliadic model is a rare and obscure episode, in comparison to earlier lyric poets’ engagement with more mainstream, marquee episodes, is undermined by its simile form – it is a far more vivid and memorable moment than Kelly supposes. All this is not enough, I believe, to dismiss this Stesichorean allusion, but it goes some way to highlighting the subjectivity inherent in any argument for or against allusion in early Greek poetry.

It is not possible, therefore, to pinpoint a specific watershed moment at which we can start talking of precise intertextual engagements rather than allusion to more general mythological traditions. We may be able to discern a gradual increase in the quantity and verbal specificity of allusions over time, but there is no sudden step change. Indeed, returning to the world of archaic epic, we should perhaps not entirely rule out the possibility of direct textual intertextuality even in our earliest extant texts. Scholars have long noted the elaborate intratextual connections within individual epic poems, especially in the Iliad and Odyssey’s large-scale repetitions of speeches and similes, even over vast distances (Il. 15.263–8 = 6.506–11; Od. 17.124–46 ~ 4.333–50, 4.556–60; Od. 23.157–61 = 6.230–4).Footnote ²¹⁸ It is difficult to deny Currie’s conclusion that ‘each poet knows his own poem as a fixed text, and recalls part of it by quoting specific lines’.Footnote ²¹⁹ And if such fixity and ‘sense of text’ is possible within an individual work, it is indeed hard to resist extending it to a poet’s ‘engagement with other poems’.Footnote ²²⁰ This alone does not permit us to reconstruct a host of lost ‘fixed’ archaic epics, for the reasons we have discussed above (§i.2.1). But when exploring the relationships of our extant texts, it would be overly restrictive to deny the possibility of direct contact at some points. And this, indeed, is what a number of scholars have found. The Hesiodic corpus, for example, is marked by a number of close connections, especially between the Theogony and Works and Days, whose relationship borders on ‘deliberate cross-referencing’:Footnote ²²¹ not only do both poems feature autobiographical accounts of Hesiod’s relationship to the Muses from Mount Helicon (Theog. 22–35, Op. 654–9) and both treat the myths of Prometheus and Pandora in a complementary diptych with numerous verbal parallels (Theog. 507–616, Op. 47–105),Footnote ²²² but the beginning of the Works and Days also appears to self-consciously ‘correct’ the Theogony’s claim that there was only one Strife (Op. 11–26 ~ Theog. 225–6).Footnote ²²³ Similar intertextual links have also been identified in the wider canon of archaic Greek epic, both between the Hesiodic Catalogue of Women and Homer and between the Homeric Hymns and a number of other early Greek hexameter poems.Footnote ²²⁴ Admittedly, in some cases, these connections may still be better explained as instances of mythological intertextuality or traditional referentiality.Footnote ²²⁵ Yet these examples – especially Hesiod’s intertextual diptych – are extremely suggestive for an early sense of (relatively) fixed textuality in the poetic world of archaic Greece.

The most controversial case, however, remains the relationship of the Iliad and Odyssey. There are many parallel passages between the two epics,Footnote ²²⁶ and a number of scholars have made plausible cases for seeing allusive connections between their structure, language and motifs.Footnote ²²⁷ In particular, it has often been argued that the fraught relationship of Achilles and Odysseus in both poems self-consciously reflects the competition between their respective epics, as each hero is defined against the other: the figure of βίη against that of μῆτις – certainly an attractive, if at times reductive, hypothesis.Footnote ²²⁸ It is understandable that some might shrink from arguing for direct allusion between these poems, given the apparently oral setting of archaic epic. And there is, after all, no smoking gun. Yet by reading the pair in dialogue, I believe that already here we can gain a richer understanding of both poems.

To contemplate such a relationship, however, we must tackle the remarkable fact that neither poem directly mentions any event from the other, a phenomenon customarily known as ‘Monro’s Law’.Footnote ²²⁹ Only the mixing of Achilles’ and Patroclus’ bones may offer an exception to this phenomenon (requested by Patroclus’ shade at Il. 23.82–92 and recalled by Agamemnon’s at Od. 24.73–84), but even this is an event that strictly lies outside the main narrative of both poems.Footnote ²³⁰ Denys Page once concluded from this absence that the Iliad and Odyssey developed in complete isolation from each other,Footnote ²³¹ but given the length and similar subject matter of both, it is difficult not to interpret the complete avoidance of each other’s narrative content as deliberate.Footnote ²³² After all, the monumental scale of both poems sets them apart from all other known early Greek epics.Footnote ²³³ In addition, the pair display an unusually high degree of complementarity: we can trace numerous contradictions and differences of detail between the Cyclic epics and Homer, but the contents of the Iliad and Odyssey are strikingly consistent and compatible.Footnote ²³⁴ Indeed, Foley and Arft have argued that ‘overlap and even contradiction’ are ‘natural and expectable’ in a multiform, pre-textual tradition.Footnote ²³⁵ The absence of both in this case is extremely telling. Moreover, when taken as a pair, the Iliad and Odyssey appear to offer an extremely convenient survey of the whole Trojan war: in its main narrative and cross references, the Iliad treats the first sack of Troy to the death of Achilles, while the Odyssey picks up from that point until the end of Odysseus’ story. This complementarity was already recognised in antiquity: Homer in the Odyssey was said to have filled out what was left out of the Iliad (τὰ λελειμμένα).Footnote ²³⁶ But given how seamlessly and coherently the two epics cover the whole Trojan war narrative, this unity certainly seems intentional and premeditated.

Of course, those who remain sceptical could still argue that the Odyssey is merely familiar with many episodes of the fabula of Achilles and the Trojan war, and the Iliad similarly with the fabula of Odysseus’ return,Footnote ²³⁷ but – in my view – the extent of the connections encourages something greater in this case: that the poet of the Odyssey was familiar with the Iliad as a distinctive text, or at least with the distinctive contours of an Iliadic tradition.Footnote ²³⁸ Such fixity would not necessarily depend on writing, but it would equally not preclude it: the excavation of the cup of Acesander at Methone has recently provided further evidence that poetry was recorded in writing by the mid-eighth century bce.Footnote ²³⁹ We should not, however, take this relationship as the norm for the Homeric texts’ engagement with other material, or as sufficient justification to reconstruct a host of distinctive, now lost poems as sources for the Iliad and Odyssey: indeed, our foregoing discussion has highlighted the limitations of that approach. In their shared length and scope, the Iliad and Odyssey clearly stand apart from the larger epic tradition. The strong links between them show that both mythological and textual intertextuality could coexist at an early date – much as specific and generic allusion coexist in later Latin poetry.

In my discussion of Greek epic and lyric that follows, therefore, I will be exploring cases of both mythological and textual intertextuality. My instinct is to assume engagement with mythical fabulae, rather than texts, especially when dealing with the lost traditions underpinning both Homeric poems, unless a particularly strong case can be made for direct textual interaction. But as we proceed to Greek lyric, potential cases of direct allusion will become more numerous. The indexing of such allusions (to fabulae and/or texts) will be the main focus of this study, but I will also stay attuned throughout to the traditional referentiality of individual words and phrases (cf. §i.2). In this way, we will best be able to appreciate the rich texture of archaic Greek allusion.

I.2.4 Agents and Audiences

The foregoing discussion has been rather abstract, focused on the interrelations of texts and traditions, with little focus on the people behind the process – the poets and audiences who comprise the agents of literary interaction. Indeed, this study fits into a growing trend of modern scholarship which focuses on the literary aspects of archaic Greek poetry.Footnote ²⁴⁰ But such a focus should not ignore the excellent progress that has been made in understanding the cultural and social contexts of archaic literature, especially in the lyric tradition.Footnote ²⁴¹ I will thus close this Introduction by addressing three issues of context which are all central to this book: audiences and performance, poetic agonism, and authorial self-consciousness.