We have lost much more early hexameter poetry than we have intact. We possess a certain amount of information about the lost poems, and even quotations from them, but only in a minority of cases: for the most part, we do not possess any fragments of their text. In several cases where a poem is attested by only a single source, we cannot even be certain that it really existed.
Even when a poem is lost, if it can be partially reconstructed by compiling pieces of fragmentary evidence from disparate sources, we may end up with enough of a basis for thinking about it as a poem. There are special problems, however. Often we have no way of knowing where and how a given fragment fits into the poem; even if we do, there is no guarantee that the fragment is representative of the poem as a whole. The fewer and sparser the fragments, the more hazardous it is to guess at what the poem was really like.
So it regularly happens that fragments convey little information in their own right. However, the surviving fragments of some hexameter poems are full enough for us to make intelligent guesses about how the poems worked, and even how they influenced other writers. This chapter offers some suggestions on how to approach these partially salvageable poems: the Catalogue of Women and the epics in the Epic Cycle. Together, their fragments amount to well over 1,500 lines. The surviving fragments of the Catalogue alone are much more substantial than, say, Hesiod's Works and Days, even though the latter has survived intact.
1. Introduction to fragments
When we open an edition of a fragmentary poem, we will find three basic types of fragments: (a) testimonia, (b) quotations, and (c) papyrus fragments.Footnote 1 A very few editions will also include (d) ‘vestigial fragments’, where there is inadequate evidence, or only indirect evidence, to tie a fragment to the poem with certainty.
(a) Testimonia are pieces of information about the poem gleaned from surviving ancient authors. For example, if Pausanias' Periegesis (Guide to Greece) happens to report something about the Cypria in passing, that becomes a testimonium to the Cypria.
(b) Quotations occur when surviving ancient authors give us the exact wording of the lost poem. There are many reasons why an author might quote a few lines; for example, an ancient scholar writing a commentary on a surviving text may quote passages for parallels or for contrasts. One particularly rich source of quotations is Athenaeus' Deipnosophistae (Scholars at Dinner), which portrays a long conversation between dinner guests who often quote old poems to each other.
(c) Papyrus fragments are the most direct of all, since they are actual ancient editions of the lost poems; but they are also the most tantalizing, since they tend to be very badly damaged. A few papyrus fragments of the Catalogue of Women contain more than a hundred lines each, but much of the text is missing because the papyrus is eroded or torn. Very many papyrus fragments consist of only a few letters. However, even when a papyrus is badly damaged, it can often be supplemented in various ways. If it overlaps with a quotation in a surviving ancient author, gaps can be filled in with certainty. Otherwise, modern editors have some deduction work to do; as this type of supplement is open to disagreement, different editions will often print different supplements.
It is important to recognize straightaway that, while the Catalogue of Women is very richly represented in papyrus fragments, few other hexameter poems are – at least, so far as we can currently tell. We have pieces of almost as many separate papyrus copies of the Catalogue as of the Theogony and Works and Days put together, showing that the poem was extremely popular in Roman-era Egypt.Footnote 2 There are also many hexameter papyrus fragments that have not been reliably identified. In contrast, we do not have a single papyrus of any Cyclic epic.Footnote 3 That does not necessarily mean that Roman-era readers had stopped reading the Cyclic epics altogether, but it does indicate a very great disparity in popularity.Footnote 4
(d) Indirect fragments and vestigial fragments. I use these terms here to refer to cases where a text preserves information that may be derived from a given lost text, but where there is no direct evidence for attribution. Indirect fragments are ones where there is a close enough match with a well-attributed fragment that a reliable link is forged. Even then there can still be dangers, but these are usually fairly secure. Indirect fragments are usually testimonia (direct quotations tend to have better attributions in ancient authors); but a large proportion of the papyrus fragments of the Catalogue of Women also relies on indirect links.
In the case of vestigial fragments, no external evidence survives to complete the link, meaning that there is no definite evidence for an attribution. As the name suggests, vestigial fragments do not normally appear in critical editions: careful editors tend to include fragments only when there is good evidence for the attribution. However, there can be exceptions, depending on a variety of factors. We shall turn to the problem of vestigial fragments in the last part of this chapter.
A well-constructed edition of a fragmentary text will present the following information.
(a) There will usually be two sections headed Testimonia and Fragments. In this case, ‘Testimonia’ means reports about the poem's author and about the poem as a whole; this information is the main basis for our conclusions about authorship, performance history, date, and so on. There will usually be testimonia in the ‘Fragments’ section too, but those will relate to specific episodes within the poem.
(b) Each testimonium and fragment will be headed with a number by which to cite it. Readers and scholars may cite items in the ‘Fragments’ section as fr. 1, fr. 2 (or F 1, F 2), etc.; and the ‘Testimonia’ section as test. 1 (or T 1), etc. An exception is the fragments of Xenophanes, Parmenides, and Empedocles: for them, the standard Diels–Kranz numbering uses A 1, A 2, etc. for testimonia, and B 1, B 2, etc. for fragments. The arrangement and numbering of fragments is the editor's responsibility and changes with each new edition, either because of new discoveries about the poem's structure or because of scholarly disagreement.
The Catalogue of Women and the Epic Cycle both exist in several different editions, each with a different numbering, and this naturally causes confusion. So editions will normally contain an index – the comparatio numerorum – that cross-references the fragment numbers in the present edition to those of older editions. If confusion is possible, one should cite the name of the editor for maximum clarity: thus ‘Cat. fr. 204 M–W’ is shorthand for the Catalogue of Women, fragment 204 in the edition by Merkelbach and West. Sometimes it is desirable to cite multiple editions: for example ‘Cat. fr. 204 M–W = 155 M’, where M refers to the Loeb edition edited by Most.
(c) The fragment's source will appear next to the fragment number. In the cases of testimonia and quotations, this will be a surviving ancient text; in the case of papyrus fragments, it will be a papyrus catalogue number. Papyrus collections are usually designated by the place where they are currently held, or by the archaeological site where they were found: ‘p.Lond.’ refers to papyri held in London, while ‘p.Oxy.’ refers to papyri found at Oxyrhynchus in Egypt.
(d) The context for the fragment will normally appear in small type, especially if the fragment is a quotation. This is done to give the reader an idea of the ancient writer's knowledge of the passage, and intent in quoting it. Papyrus fragments, by their nature, have no context, so there will be no small type. Testimonia are effectively all context and no text, so they may be printed entirely in small type.
(e) Finally, the surviving text from the poem (if any) will be given in full-size type. In papyrus fragments, editors try to supplement gaps in the papyrus where it is possible to do so reliably; such text should be printed in square brackets, both in the Greek text and in the translation. There are other elaborate conventions that editors follow in the Greek text for papyri, but we do not need to go into those here.Footnote 5
2. Examples of fragments
To put all of this in some context, let us look at a couple of examples. First, a fragment from West's Loeb edition of epic fragments. The following comes from the Little Iliad (West 2003a: 128–9):
5 Schol. (T) Il. 16.142, ‘ἀλλά μιν οἶος ἐπίστατο πῆλαι Ἀχιλλεὺς’
οἱ δὲ πλάττονται λέγοντες ὡς Πηλεὺς μὲν παρὰ Χείρωνος ἔμαθε τὴν χρῆσιν αὐτῆς, Ἀχιλλεὺς δὲ παρὰ Πηλέως, ὁ δὲ οὐδένα ἐδίδαξεν. καὶ ὁ τῆς Μικρᾶς Ἰλιάδος ποιητής·
5 Scholiast on the Iliad, ‘only Achilles knew how to wield it [Achilles’ spear]’
Some tell the fictitious tale that Peleus learned the use of it from Chiron, and Achilles from Peleus, and that he taught nobody else. The poet of the Little Iliad says:
As the heading indicates, this is Little Iliad fr. 5 in West's numbering; and the fragment comes from scholia – ancient glosses and commentaries – on the Iliad. The parallel Greek text gives detailed information on its source: the fragment is from the T collection of scholia, and the gloss relates to Il. 16.142, ‘only Achilles knew how to wield it’. Below the heading is the context, in smaller print, where the ancient scholiast explains the story of how Achilles learned to use the spear; finally, the scholiast quotes a line and a half from the Little Iliad.
From this we gain the following information:
(a) The fragment is a quotation.
(b) From the context we can infer that Achilles' spear appeared in the Little Iliad; and that the lines quoted refer to the spear.
(c) Both the context and the quoted lines give some information about the spear as depicted by the Little Iliad poet.
(d) As to the lines' context in the original epic: there we enter the realm of deduction and conjecture. In an ideal world, we would consult a modern commentary to see what the editor has to say about the fragment. But modern scholars have not written many commentaries on fragmentary texts – there is not a huge market for them – and, where commentaries do exist, they are very often in languages other than English. In the following section we shall look at some tactics that a diligent reader might follow.
For a second example, we turn to a fragment from Most's edition of Hesiodica. The following comes from the Catalogue of Women (Most 2007: 110–11):
47 [73 MW; *2 H] P. Lond. 486c, P. Oxy. 2488B
47 London papyrus
So we may see that this is Cat. fr. 47 in Most's numbering. It comes from two papyri, one held in London, the other found at Oxyrhynchus. The translation mentions only the London papyrus: the Oxyrhynchus papyrus adds little new text, but matters for other reasons, as we shall see below. In the Greek text, Most cross-references the fragment to fr. 73 in the critical edition by M(erkelbach) and W(est) and fr. *2 in H(irschberger)'s critical edition.
After the heading we plunge directly into the text: papyrus fragments do not come with any context. To compensate for this, Most has provided a convenient heading prior to the fragment (not reproduced above), which introduces the present group of fragments: Atalanta, Daughter of Schoeneus, Athamas' Other Son by Themisto (Book 2, Lines 1ff.).
Within the text we see several square brackets: some encircle phrases, others mark off blank areas. These represent the edges of the papyrus scrap. Where the text has been supplemented, it is not normally mere guesswork: missing text can often be deduced from our knowledge of the stories and characters involved (hence the supplement of ‘Schoeneus’ in the first line); by grammatical necessity; or by repeated formulaic phrases (in the Greek text, ‘possessing the Graces’ radiance’ is missing parts of ‘Graces’ and ‘possessing’, but the formula is a common one and there is no doubt about the wording). But different editors are more or less adventurous about supplementing the text. In comparison with other editions of the Catalogue, Most is relatively daring (while still keeping his feet planted firmly on the ground), on the principle that an expert opinion is better than none at all.
So we may infer the following:
(a) The fragment is a papyrus fragment from an ancient edition of the poem.
(b) Most's headings, and the reference to Atalanta in line 2, tell us that this is part of the story of Atalanta.
(c) We may infer some points about the Catalogue's Atalanta narrative, if we look at the text in conjunction with the subsequent fragments and the story as known from other sources. This fragment tells us about Atalanta shunning human company and hoping to escape marriage – rather like Euripides' Hippolytus. This in turn sets the scene for the athletic competitions mentioned in other accounts: Atalanta's hopeful suitors fail to defeat her in a race and so are executed, until Hippomenes comes along and uses a trick to defeat her and marry her.Footnote 6
Going beyond this point requires more in-depth research. In this case, comparison of critical editions will show that Merkelbach and West, Hirschberger, and Most all print different texts for this fragment. For the purposes of gaining an acquaintance with the poem such differences can often be ignored: close comparison is only necessary if one wants to put weight on the fragment as evidence in an argument. But these differences sometimes have an impact on how and where the editor places the fragment. Here, Most has accepted a conjectural connection between the end of fr. 46 and the start of fr. 47; that is his basis for identifying this as the start of Book 2, and for his inclusion of the text ‘Or like her…’ at the start of the fragment (the other editions leave the first half of line 1 blank).Footnote 7 Moreover, Most assigns the fragment to the Catalogue partly because of its content (the story of Atalanta) and partly because one of the two papyri – p.Oxy. 2488B – comes from the same manuscript as p.Oxy. 2488A, which contains part of Cat. fr. 82; in this he follows Merkelbach and West. This is therefore an ‘indirect fragment’, according to the definition given above. Hirschberger, however, relegates the fragment to ‘fragments of unknown poems’ (ἀδήλων ἐπῶν): partly because some details lead her to doubt that p.Oxy. 2488A and 2488B are parts of the same text, and partly because of her beliefs about other fragments that deal with the Atalanta story.Footnote 8 The balance of probabilities favours Most, but new evidence could in principle prove Hirschberger right one day.
If these details seem tiresome, we should be the more thankful for these scholars' efforts to piece all this information together. The whole point of their editions is to make sure that other readers do not need to go into as much depth.
3. How to approach a fragmentary text
The second example above illustrates that a fragment taken in isolation is often very unclear. The best guide to understanding a fragment and its place in a poem is an acquaintance with the other fragments in the poem. This naturally means a great deal more labour, and it does not always work: sometimes the information is just too sparse to make anything of it. That is unfortunate but unavoidable. Knowledge does not come cheap. Sometimes no amount of research will help: the information may be lost forever. The following suggestions may make the process slightly easier for apprentice researchers.
(a) Although older editions are often freely available online, make sure to use a recent edition where possible. For the Catalogue, in particular, no-one should ever use the old Loeb edition of the Hesiodica by Evelyn-White (1914): so much new information has come to light in the last century that it is utterly obsolete in every way.
(b) If your edition contains an introduction and subheadings, these will give a good impression of how the editor imagines the fragmentary text, and will help explain the decisions he or she has made. An introduction may help explain why edition A includes such-and-such a fragment while edition B does not.
(c) Go through the fragments making rough notes on any structures and patterns that you can see. This is especially important where subheadings are not provided. For example: in West's edition of the Returns, several of the fragments report on various mythological and genealogical details that have nothing to do with the Greek heroes returning from the Trojan War. Fr. 3 is about Tantalus, fr. 4 about Clymene and her father, Minyas, fr. 5 about Maira and her father, Proitos, and so on. At fr. 10 we start to see references to the story of Agamemnon's homecoming; so the problematic sequence is frs. 3 to 9. The reader should make notes on this pattern and then examine this group in the light of the other fragments to try to infer why the editor has arranged the fragments in this way.
The answer is not long coming. Prior to this group, fr. 1 informs us that the Returns, like the Odyssey, contained ‘mention of Hades and the terrors in it’; and fr. 2 mentions that ‘the Cyclic poets’ used the term nekades to refer to ‘the souls of the dead’. Careful note-taking on this material should lead straight to the appropriate conclusion: the Returns contained an Underworld episode, like the one in Odyssey 11, which incorporated a catalogue of the souls of legendary figures of the past, as witnessed by some hero (we do not know who); frs. 3 to 9 must come from this catalogue.Footnote 9
(d) In the specific cases of the Epic Cycle and the Catalogue, extra tools are available to assist the reader in reconstructing the shape of the poems. The idea is to draw up a masterplan on a separate sheet: with this in hand, the reader will find it easy to annotate the plan with indications of particularly interesting fragments.
In the Epic Cycle, Proclus’ summaries should be the basis for drawing up a masterplan. With a list of episodes laid out down one column, the reader will often find it possible to place fragments in their correct spot. Take, for example, the Little Iliad fragment quoted earlier. Proclus’ summary makes it clear that Achilles ought to be dead before the start of the Little Iliad; so the question should naturally follow, what is the occasion for mentioning his spear? Further reading will show that the answer lies in one of the following passages in Proclus.
The awarding of the armor takes place, and Odysseus gets it in accord with Athena's wishes…
And Odysseus fetches Neoptolemus from Scyros and gives him his father's armor…
Here we have two contexts in which the spear might easily be described: ‘armour’ is a rendition of hopla (‘arms’, including weapons). It is possible that the lines appeared in some other context, but these two episodes are such a good fit that the reader can be quite confident in this conclusion. West's ordering of the fragments indicates that he thinks that the spear description belongs most easily in the first episode; but an intelligent reader could perhaps find grounds for debating the point.
In the case of the Catalogue, Most's helpful section headings can serve as the basis for the episodes that the reader will list down one column of a separate sheet. The resulting list would look like this:
-
Book 1
-
Proem (Book 1, lines 1ff.) – fr. 1
-
The descendants of Deucalion – fr. 2 to fr. 10.24
-
Deucalion's children: Pandora, etc. – fr. 2
-
Thyia's sons – fr. 7
-
Magnes’ sons – fr. 8
-
Hellen's sons: Dorus, Xuthus, Aeolus – fr. 9 to fr. 10.24
-
The descendants of Aeolus – fr. 10.25 to fr. 71
-
Aeolus’ children – fr. 10.25–34
-
Perimede's children – fr. 10.35–57
-
(etc.)
It would, of course, take several sheets to include all of Most's subheadings. As an alternative, see the much less exhaustive list in Chapter I, which highlights some of the most interesting episodes.
4. Vestigial fragments
Early hexameter poetry, especially the Catalogue and the Epic Cycle, is deeply preoccupied with legendary narratives; but much of what we know about these legends comes from later sources, and often without source attributions. Later reports of legends undoubtedly include information that is derived in one way or another from the lost poems, but we have no way to tell when they are doing so. Commentaries could discuss this material – where the commentaries exist. As things stand, this invisible testimony is doomed to hover outside the margins of our editions as vestigial fragments.
Editors of fragments are on the whole a cautious bunch, and require reasonably high standards of evidence for attribution. Best of all is an explicit attribution; next best is an ‘indirect’ fragment, where there is sufficient overlap with a ‘solid’ fragment to conclude that it is more than a vestige. And there is no doubt that this is the correct course to follow. The other way lies pure speculative madness. However, even the best editors are willing to make exceptions sometimes. M. L. West's Loeb edition of the Homeric Hymns assigns substantial papyrus fragments to the Hymn to Dionysus (Hymn 1) that do not appear in older editions, even though there is no solid evidence to connect them to the Hymn.Footnote 10 And Alberto Bernabé's edition of the Orphic fragments classifies each fragment as T (testimonia), F (fragments of text), or V (vestigial fragments).
The issues are especially pronounced with legendary material in a group of mythographic texts – that is to say, compendia of myth – which mostly date from the first century bce to the second century ce.Footnote 11 (There were other mythographers writing outside this time frame, of course.) One such text, the Library of pseudo-Apollodorus, is especially important: it is our chief guide to the structure of the genealogies in the Catalogue of Women, and its account of the Trojan War is so close to Proclus’ summaries that West, in his edition, splices the two together as a single text.Footnote 12
So pseudo-Apollodorus can be accepted as legitimate testimony, at least sometimes, even though explicit evidence is missing. But what about the Fabulae of pseudo-Hyginus? In the Fabulae some Trojan War episodes are very similar to what we find in Proclus; others are quite different.Footnote 13 Then there are Conon (Augustan-era) and Ptolemy the Quail (‘Ptolemaios Chennos’, first century ce), both of whom are more eclectic and who survive only as summaries in Photius’ Myriobiblon;Footnote 14 Philostratus’ Heroicus (third century); several late lexicographers and encyclopaedias that replicate some of the same material; and, most mysterious of all, the Ephemeris of Dictys of Crete (late first or early second century), a semi-novelistic prose account of the Trojan War that has links to several of the authors already named but departs from the tradition in other respects.Footnote 15
Let us take an example of an indirect fragment as an illustration. It seems that Little Iliad fr. 11 was the basis for a proverbial phrase, ‘Diomedian compulsion’. Pausanias’ Collected Attic Words reports the story as follows:
Others say that Diomedes and Odysseus were on their way back from Troy at night after stealing the Palladion, and Odysseus, who was behind Diomedes, intended to kill him; but in the moonlight Diomedes saw the shadow of his sword, turned round, overpowered Odysseus, tied him up, and forced him to go ahead by beating his back with his sword.
Pausanias does not indicate his source. The mythographer Conon tells a slightly different version of the story, adding a few details, and with one important difference: it is Odysseus who beats Diomedes on their way back to the ships, not vice versa.Footnote 16 Again there is no attribution. It is only Hesychius’ Lexicon that provides a direct link:
‘Diomedian compulsion’: a proverbial expression. …The author of the Little Iliad connects it with the theft of the Palladion.
But Hesychius’ report is also the most sketchy. Which is closer to the Cyclic epic, Pausanias’ account or Conon's? Other late sources side with Pausanias; and our summary of Conon may be in error through missing a pronoun. So, for now at least, Pausanias is the winner. Even so, caution is needed: if we look at the full text of Pausanias we see that he also links the proverb to a totally unrelated story, about the Thracian Diomedes compelling guests to sleep with his daughters.
Let us now consider a true vestigial fragment, a story that has no substantiated link to an early poem. In Book 6 of Ptolemy the Quail's New History, as summarized by Photius, we find the following intriguing snippet:Footnote 17
[Ptolemy writes] that when Priam came to Achilles to supplicate him and beg for the bones of Hector, Andromache and her sons came along with him.
This is obviously non-Homeric: in the Iliad, Priam visits Achilles’ tent unaccompanied, making his daring all the more astonishing to Achilles and his men, and Andromache and Hector have only the one son, Astyanax.Footnote 18 Has Ptolemy made it up? If not, where does it come from?
One candidate is Aeschylus’ lost Lytra or Ransom of Hector. Unfortunately, we know almost nothing about that play. If it belongs to the period when Attic tragedy featured only two actors it is unlikely to be Ptolemy's source, since two speaking parts are already accounted for by Priam and Hermes/Achilles. A more likely possibility is the Cyclic Aethiopis. Proclus’ summary of the Aethiopis opens after the end of the Iliad, with the arrival of Penthesileia; but there is good evidence that the Cyclic epic actually overlapped with the Iliad and included the ransom and funeral. Some Roman-era ‘Homer cups’ depict a triptych of scenes (implying that they belong to a single narrative) showing (a) the ransom, (b) Priam and Penthesileia at Hector's grave, and (c) Achilles and Penthesileia in battle.Footnote 19 Weitzmann has independently interpreted a second-century-ce sarcophagus lid from the Villa Borghese as evidence that the Aethiopis included Hector's funeral.Footnote 20 And Aethiopis fr. 1 (three lines linking Hector's funeral to Penthesileia's arrival) appears to be a quotation from within the Aethiopis; it is certainly not the start of the poem, so again an overlap is indicated.Footnote 21
Things get more complicated when we compare another account of the ransom, found in Philostratus’ Heroicus:Footnote 22
[Homer] says that Achilles died at the hands of Paris and Apollo; he seems to have known the business at the Thymbraean [temple]…As for the sacrifice of Polyxena on his grave, and everything you hear from the poets about his love, it was like this. Achilles fell in love with Polyxena and was negotiating for her marriage to him, on condition that the Achaians withdraw from Ilion. And Polyxena loved Achilles too, and they saw each other at the ransom of Hector; for when Priam came to Achilles, he made his daughter lead him by the hand…
This is a remarkable passage. The story of the human sacrifice of Polyxena at Achilles’ grave goes back to the Cyclic Iliou persis, and was repeated often. But the earliest datable source to refer to a romance as such is Seneca the Younger's Troades, and even there it is not actually enacted but a posthumous union intended by Achilles’ ghost, contradicting the more traditional story of Achilles marrying Medea in the afterlife.Footnote 23
Pseudo-Hyginus includes the Polyxena romance too. However, we do not know the date of his Fabulae; worse, his account of Achilles’ death is closer to an episode in the Iliad than to anything we know of the Aethiopis.Footnote 24 Finally, there is the Latin version of Dictys of Crete, which portrays the ransom with both Andromache and Polyxena alongside Priam; and Andromache brings her two sons, Astyanax and Laodamas.Footnote 25 Dictys therefore lies at the centre of this problem.
It is not certain what has happened here. Philostratus might have adapted the whole thing from Dictys; but Philostratus cites his source as ‘poets’ (Dictys was prose). In view of that, it is conceivable that Polyxena's role has its origins in Sophocles’ lost Polyxena, or something of the kind. My own reading of the situation is that Philostratus refers to ‘poets’ as sources for the romance but not for the ransom. Polyxena's involvement in the ransom would then originate with Dictys. And, indeed, Philostratus matches Dictys in a few other respects: not least, the setting of Achilles’ death at the temple of Thymbrian Apollo, rather than at the gates of Troy. Pseudo-Hyginus, too, has connections to Dictys: both Fab. 110 (the romance with Polyxena) and the following Fab. 111 (Hecuba's transformation into a dog) are closer to Dictys than to any other surviving account.Footnote 26 Note that this interpretation would make pseudo-Hyginus later than Dictys, which is not the conventional view.
But we lack direct evidence, so the question over Andromache is not an open-and-shut case. Her role in the ransom, and the presence of her two children, are attested in Ptolemy the Quail and in Dictys, and nowhere else – though the Borghese sarcophagus, mentioned above, may hint at Andromache's involvement.Footnote 27 It is distinctly possible that Dictys reflects the Aethiopis reasonably closely in these respects; indeed, it would follow that the opening episodes of the Aethiopis – Hector's death, ransom, and funeral – are likely to have gone along much the same lines as Dictys’ account, which is dominated by non-Homeric elements in each of these episodes.Footnote 28 But not with respect to Polyxena: the Achilles–Polyxena romance is best explained as something new, a novelistic trope that did not exist prior to Dictys.
The relationships between all of these sources, mythographic and otherwise, have been explored for some time, but nowadays they are somewhat under-studied. Erich Bethe argued that a single mythographic handbook, dating to the first century bce, was responsible for the legendary material in the historian Diodorus of Sicily, as well as in Proclus, pseudo-Apollodorus, and pseudo-Hyginus.Footnote 29 Albert Hartmann, who explored other mythographic sources relating to Odysseus’ death, claimed to have rediscovered a Hellenistic mythographic text – not necessarily Bethe's handbook – by collating further miscellaneous texts: Hartmann's texts follow a distinctive pattern of referring, in the same sequence, to (a) the Telegonus story; (b) a prophecy from Aeschylus’ lost Psychagogoi; and (c) an Italian folktale where a witch transforms Odysseus into a horse and keeps him until he dies in old age.Footnote 30
Such concoctions are delightful in their own right, but this kind of Quellenkritik (‘source analysis’) is now, regrettably, considered old-fashioned. More recently Davies, following Wagner and Severyns, has fiercely argued against Bethe's thesis. However, he overshoots his aim: his objection is really to the fact that Bethe thought Proclus interfered with his source, whatever that source may have been, and Davies makes a good case that Bethe was wrong about this. However, that has no bearing on the question of whether Proclus was working from a Hellenistic mythographer, as Bethe thought, or directly from the Epic Cycle, as Davies thinks.Footnote 31 Most recently, Alan Cameron has examined the sources for various mythographic or mythographic-style texts; while he accepts that a chain of influences and cross-influences may exist, he is basically agnostic about its recoverability.Footnote 32
We must not be too cynical about the efforts of Bethe and Hartmann. Their work was flavoured with an enthusiasm inspired by the successes of the physical sciences in the late nineteenth century, which gave humanists the sense that they could use scientific methodologies to recover almost anything about the ancient world that they wanted. That is certainly over-optimistic, but cynicism is not the correct response. We can simply be more realistic. There is no doubt that Bethe's and Hartmann's Quellenkritik was onto something. The question is what that something was, and what an improved form of their methodology would look like.
