The Epic Rhapsode and His Craft: Homeric Performance in a Diachronic Perspective

  González, José M. 2013. The Epic Rhapsode and His Craft: Homeric Performance in a Diachronic Perspective. Hellenic Studies Series 47. Washington, DC: Center for Hellenic Studies.

6. Early Homeric Scholarship and Editions

In this section I wish to examine two arguments for an early, sixth-century written fixation of the Homeric poems. One has an Athenian emphasis, as we might expect from the evolutionary model. This model posits a definitive textual stage for the Iliad and the Odyssey, centralized in Athens, from the middle of the sixth century to the middle of the fourth, during which transcripts of growing significance for performance will have been produced. The arguments to be reviewed are: the Homeric scholarship of Theagenes of Rhegion; and the running quarrel about Salamis between Megara and Athens, with the report of Spartan arbitration which in the view of many implies the existence of an authoritative written copy of the Iliad. A third line of investigation focuses on problematic Homeric forms that are sometimes alleged to prove the existence of early copies written in the old Athenian alphabet. These forms are explained as mistaken readings of a text that did not use word separation, had a deficient graphemic repertory for vowels, and simplified sequences of repeated letters (e.g. OO to O). Cassio (2002:109–114 §§2–3, with bibliography) disposes well of this alleged evidence for the early written transmission of the poems. [1] I will not delve into the old chestnut of the Peisistratean redaction, which I consider a myth of origins, other than to point out that the account must be read for what it teaches us about Homeric performance and the assumptions of the ancients about it. [2]

6.1 Ajax and Salamis

Cassio (2002:115) represents the consensus when he writes: “In my opinion the text of the Iliad must have been fixed for some decades by 560 BC, when the lines Aἴας δ’ ἐκ Σαλαμῖνος ἄγεν δυοκαίδεκα νῆας, | στῆσε δ’ ἄγων ἵν’ Ἀθηναίων ἵσταντο φάλαγγες (Il. 2.557f.) were used as evidence by the Athenians against the Megarians in an interstate arbitration over Salamis.” The case, however, is hardly that simple. In fact, even if we assume arguendo its basic historicity (which is beset by serious doubts), the record best supports that the quarrel played out as one between rival performance traditions and did not hinge on a given written text of the poem. [3] Interpreting the various reports and assessing their soundness is not easy and the list of distinguished attempts is long. But before I make what must needs be a rather brief appraisal of my own, it is good to pause and consider what sort of document might have been produced by Solon at the alleged arbitration—if a written document is actually what the texts imply. For the existence of a late eighth-century written Iliad, either in the possession of the Athenian state or else, say, at Khios in the hands of the Homeridai, as the acknowledged, authoritative source of all other copies, should have prevented the Megarians from falsely accusing Solon of forgery and from supplying an alternative of their own to Β 557–558; or else it should have made clear that Solon had in fact inserted Β 558 into the copy, especially if one accepts the notion that an eighth-century calligraphic bookhand had drafted the text in a suitable archaic script (Euboian?) respecting word boundaries and line divisions. The Spartans in turn should have found it all too easy to check their own copy of this alleged ur-text and establish the truth of the competing claims. No further evidence would have been required.The reports of course suggest no such thing (see below). Neither do the Homeric manuscripts. The textual status of Β 558 is dubious. After a careful review of the evidence, Apthorp (1980:165–178) concludes that there is no good reason to suppose that Aristarkhos found Β 558 in more than a few of his manuscripts, whereas there are various reasons to suppose that he did not. West in his Teubner edition agrees: “Ar[istarchus] versum omisit, cum sciret eum non in omnibus ferri libris.” [4] He also lists a number of papyri that do not carry it or (in one case) add it later. [5] That the parodist Matron of Pitana knew it (Supplementum Hellenisticum 259–266 no. 534.95–97) gives us a terminus ante quem of the late fourth century, although, as Apthorp (1980:171) perceptively observes, this does not entitle us to infer that he found the verse in his text of Homer (if he owned one)—nor even, I would add, that it was regularly performed with the passage. For Plutarch in Solōn 10, [6] after reporting that the majority say (οἱ μὲν οὖν πολλοὶ … λέγουσι) that Solon inserted (ἐμβαλόντα) a verse (ἔπος) in the Catalogue of Ships and read it at the trial (Β 557–558 follows), records that the Athenians themselves think that this is nonsense (φλυαρίαν). They assert that Solon proved (pointed out?) to the jury that the sons of Ajax had become Athenian citizens and made over their island to them. Solon is further said to have argued his case by comparing Athenian and Megarian burial customs to the way in which the Salaminian dead were buried. The further support of Delphian oracles is noted. Taking Plutarch at face value, the Athenians put little stock in the alleged textual evidence, whether we think that the other arguments were added to it [7] or (what seems to me a more natural reading) that they disowned the ‘silly story’ of Homer’s help through Solon’s forgery and substituted for it three alternative lines of reasoning. Apthorp (1980:169) points out that the Athenian denial did not “take the form of a claim that the line is genuine” but “that Solon in fact used other, more respectable, arguments.” Although I do not agree with the search for Homer’s ipsissima verba and judgments about the authenticity of lines whose diction is clearly traditional, if Apthorp’s reading is right and Plutarch’s report factual, we arguably detect a certain Athenian embarrassment at Β 558, as if the circumstances surrounding it had made it notorious. Hence, many may have thought it better suppressed from the performance of the passage and, eventually, from the written texts. This notoriety would sufficiently motivate Matron’s parodic reference to it without any need for its textual attestation: “He could well have known the line merely from the story surrounding it. A piquant anecdote like this would certainly have lent itself to an indirect allusion” (Apthorp 1980:171). Notoriety further motivates, as suggested, the line’s suppression even from many an Athenian manuscript, assuming that it was not manufactured for the story, as Bolling (1916:29) thought.

No support for it must be necessarily inferred from Aristotle’s comment in the Rhetoric 1375b29–31 that the Athenians used Homer as witness concerning Salamis. For a text that included the Athenian section and was followed solely by Β 557 would suffice to draw the plausible inference that Ajax and Salamis were an appendix to Menestheus and Athens. As Bolling (1916:29) memorably put it, “Β 546–57 in which Aias is made but a tail to the Athenian kite, would be an ample foundation for this legendary use of poetry as evidence.” I grant that a version that only featured Β 557 for Ajax would be most unsatisfying (if only slightly less than the one we possess). But if we take seriously what the scholia tell us about Aristarkhos’ evidence, this seems the most likely scenario for the mainstream of the Athenian paradosis. Assuming the historicity of his telling, the Athenian attitude that Plutarch hints at might explain this state of affairs. One can only conjecture what other, less Athenocentric versions might have sung instead. Cingano (2005:143–152) is right that the language in fr. 204.44–51 of the Hesiodic Catalogue of Women does not support the assertion by Finkelberg 1988 that Ajax’s domain is portrayed as embracing the cities mentioned there (Troizen, Epidauros, Aigina, Mases, Hermione, and Asine). But it does perhaps suggest what modifications to our Iliad might have given a section to Ajax commensurate with the hero’s relevance and comparable in length to other entries in the Catalogue of Ships. Strabo 9.1.10 provides another glimpse of an alternative version, this time, Megarian. We are told that either Solon or Peisistratos interpolated immediately after Β 557 the offending Β 558 and resorted to Homer as a witness that Salamis had been the Athenians’ from the beginning. We also learn that the critics (οἱ κριτικοί) do not accept the verse, because it is contradicted by other statements in the poem. Nevertheless, the Athenians appear to have alleged in their own support ‘some such testimony from Homer’ (οἱ μὲν δὴ Ἀθηναῖοι τοιαύτην τινὰ σκήψασθαι μαρτυρίαν παρ’ Ὁμήρου δοκοῦσιν), to which the Megarians responded with a rival version: οἱ δὲ Mεγαρεῖς ἀντιπαρῳδῆσαι οὕτως “Aἴας δ’ ἐκ Σαλαμῖνος ἄγεν νέας, ἔκ τε Πολίχνης, | ἔκ τ’ Aἰγειρούσσης Nισαίης τε Tριπόδων τε.” ἅ ἐστι χωρία Mεγαρικά. [8] The rare word ἀντιπαρῳδῆσαι might imply mockery, as if the Megarians were making the point that they too could compose ad hoc forgeries to their benefit. But we would expect a parody to imitate its model, and there is no imitation of syntax or sound such as we find in the parodic Matron. The Megarian verses open with identical diction up to the thesis of the fourth foot and veer subsequently without any resemblance to the Athenian. For this reason, I think it is better to understand ἀντιπαρῳδῆσαι in its etymological sense: they sang their version alongside (παρά) instead or in return (ἀντί). This is the way Higbie (1997:286) understands it: “the Megarians give this version in reply.” There is every reason to think that the Megarian version, substantially different from the Athenian and hardly recognizable as a parody in the ordinary sense of the term, was intended as a genuinely valid multiform. [9] For the more common νῆας, the verses use the comparatively rarer, but by no means unexampled, νέας; [10] and although the quotation does not report the number of boats, there is no reason to suppose that the Megarian version of this entry stopped with the second line quoted. What followed may have stated the missing information. The obvious point of quoting two verses only was to join the poem with the received opening (what in our Iliad is the beginning of verse 557) and then to supplant the offending 558 with material favorable to Megara’s claims. [11] Strabo’s equation of Tripodes with Tripodiskion and his comment that the places mentioned are Megarian suggest that these were not verses composed ad hoc to counter the Athenian version. Special erudition was required to understand them as making Megara’s point. Rigsby (1987b:100) writes that “a Megarian forger motivated by his city’s claim upon Salamis in the face of an Athenian threat should not have spoiled his case by failing to name Megara.” And Figueira (1985:268) believes that the group of locations mentioned by the Megarian version “portrays the actual situation … in the late seventh and sixth centuries.”

Hence, the report, if it has a factual core, is best understood as a rivalry of traditional versions. Bolling (1925:16) had noted that “[fama affertur Β 558] ab oratore Atheniensi interpolatum esse. Ut haec interpolatio per verba sola fieri intelligatur plane necesse est.” And again at 73: “The meaning is clearly that they recited as if genuine extra lines of their own composition, not that they forged and put into circulation copies of the lengthened text.” Apthorp (1980:168) thought that Bolling’s emphasis on verbal performance went too far and cited παρεγγράψαντα (Strabo 9.1.10), ἀναγνῶναι (Plutarch Solōn 10.2), and ἐγγράψαι (Diogenes Laertios 1.48) as making explicit reference to Solon’s writing. But these can be readily understood as anachronistic details and can hardly be given probative value. Our late sources merely retroject to the alleged ‘trial’ before the Spartan umpires elements of their own book cultures and known practices attested for classical court proceedings, at which a clerk would read aloud when requested documentary evidence given to him by the parties in advance. [12] For Strabo, Peisistratos or Solon subjoined Β 558 to the Athenian section after Β 557 so that he may use Homer as a witness (μάρτυρι χρήσασθαι τῷ ποιητῇ). In this way Solon was effectively preparing the written copy to be read at the trial. Plutarch’s version is similar: Solon inserted the text into the Catalogue of Ships and then read it at the trial (ἐπὶ τῆς δίκης ἀναγνῶναι). Diogenes Laertios does not mention a trial before a panel of Spartans. But there may be an echo of it in his account of Solon’s desire to make the case that Athens’ possession of Salamis was just (δίκῃ). [13]

It is important to mark the following methodological observation. Jensen (1980:138) writes that “Solon’s interpolation presupposes an already fixed Catalogue of Ships.” This kind of claim is often found in connection with apparent or alleged ancient references to specific lines in our Homeric texts. The quotation from Cassio that opens this section is another example of the same. But it involves an inadmissible leap of logic, effected under the spell of our regularized modern editions and their unvarying line numbers. Nothing follows for the entire Catalogue of Ships—much less for the entire Iliad—from the fact that a few ancient authors mentioned two verses identical to our own Β 557–558. Even if we conclude that the ultimate source for the quoted text is early, why should we be surprised if some manuscripts embraced lines whose notoriety drew the attention of commentators and placed them in the very sequence reported by them? It would be fallacious to assert that this or that ancient writer quoted Β 558 as if he had one of our modern editions at hand. That is of course not what is usually meant by such a statement. But in the face of sweeping, impermissible inferences like Jensen’s or Cassio’s, it is hard to resist the impression that the modern scholar has embraced unawares a glaring hysteron proteron.

Of somewhat different cloth are Cassio’s remarks about what he deems “an awkward interpolation in the Nekuia … [from] the time of the Peisistratids.” To him, the alleged interpolation—the statement that not Herakles but his eidōlon was in Hades—“shows by implication that the text in which it was inserted was by then fixed and unalterable” (Cassio 2002:116 and n. 52). Once again, there is slippage in the reasoning. Even if the fact that Odysseus had seen Herakles in Hades was traditional by a certain time, say, the rule of Peisistratos, it does not follow that the text of the Nekuia was fixed, nor even that necessarily the lines about Herakles were. [14] I cannot pronounce on the alleged awkwardness of the juxtaposition, which remains a matter of subjective taste. In this judgment, Cassio joins an array of distinguished scholars, ancient and modern. But, for all their skill, their consensus is none the less textually arbitrary and based on a subjective feel for the sequence and thematic material that would make for the best poetry. Focke (1943:228–229) offers a few specimens. He quotes Wilamowitz as saying that “von einer Erhöhung (des Herakles) weiß das Epos nichts und kann es nichts wissen wollen.” He reports the judgment of Rohde as “Verlegenheitsauskunft ältester Harmonistik.” His own is equally damning: “Theologie, nicht Poesie enthalten die eingeschobenen Verse.” The charge of tampering with the text in the interest of harmonizing a perceived divergence can be readily turned around. For if, harried by a scheming Hera, Herakles had achieved apotheosis through self-immolation, would his end not fully agree with Σ 117–119? [15] There is no incongruity with λ 616, pace Heubeck, since Herakles is bewailing Odysseus’ lot in life, not his own in death, as his words show (especially λ 619 and the imperfect εἶχον in 621). From a textual point of view, Bolling (1925:212) demonstrated long ago that, when the scholia are carefully read, they make clear that λ 604 had not been athetized by Aristarkhos or any other ancient scholar. [16] Rather, the addition of this line, identical to Hesiod Theogony 952, must be post-Aristarchean, as its absence from a second-century AD papyrus and some manuscripts confirm. According to the somewhat muddled scholia ad loc., some athetized one or both verses in λ 602–603 and fathered them on Onomakritos. But there is little to recommend this attribution if, as I think, the combination of heroic and divine status for Herakles was already old and deeply traditional in the sixth century.

Indeed, Herakles is hardly an ordinary hero. It has been plausibly argued that, like Helen in Sparta, he was originally a god with a chthonic dimension who was subsequently given an exceptional heroic identity. [17] Burkert reminds us that the complex of immolation and apotheosis recalls Near-Eastern tradition. Cult was offered on Mount Oita at the penteteric fire festival. At Tarsos, in Cilicia, a pyre was prepared yearly for an ancient Anatolian god, Sandes or Sandon in the local language, called Herakles in Greek. Hittite kings were divinized by cremation. And the equation of Herakles and the Phoenician god Melqart, recognized in Herodotos 2.44, explains his cremation at Tyre. [18] Usener (1896:255) once affirmed: “Wir dürfen mit überzeugung den satz aufstellen, dass alle heroen, deren geschichtlichkeit nicht nachweisbar oder wahrscheinlich ist, ursprünglich götter waren.” In its application to Herakles, this statement eerily recalls Herodotos 2.44: “My investigations plainly show that Herakles is an ancient god.” [19] If so, diachronically speaking, his portrayal as a hero and his presence in Hades would be the innovation—certainly very old and likely a part of the Homeric tradition since its earliest stages—not his dwelling on Olympos in the company of the other gods. [20] Hence, the verses often condemned as an intrusive interpolation could actually reflect a very old stratum of Greek religion. The rationale for classifying them as a later addition to an already fixed text would vanish. In principle, it is still possible, if improbable, that a later addition was made to the poem to accommodate older religious material. But this claim would face the burden of proof—why should older material be a late arrival to the poem?—proof that would be hard to come by considering the utter lack of textual support for it. At any rate, the report that Onomakritos added the offending verses to accommodate a recent elevation of Herakles to divine honors would be demonstrably false. [21]

Be that as it may, even if we grant for the sake of argument that, thematically, Herakles’ presence in Hades was chronological prior to the qualification, allegedly added later, that Odysseus had only seen his eidōlon, it does not follow that the text was already fixed when the qualification was made. Cassio is confusing theme and diction. Thematic material felt to be indispensable to a story need not be attended by already established, equally indispensable verbal expression. A performer may not have felt free to suppress the reference to Herakles, even if—to follow Cassio—he thought it theologically objectionable. But this hardly requires us to hold that he had also memorized specific language, sanctioned by tradition, to narrate the encounter. Cassio’s whole argument hinges on his subjective feeling of awkwardness. It leads him to believe that the qualification is clumsy precisely because it does not fit with an already existing, traditional sequence of clauses. [22] Sentencing the whole as “patently absurd” (Cassio 2002:116n52) flies in the face of the uniform textual paradosis, which only challenges λ 604. The scholiast calls the other offending material not ‘absurd’ but νεωτερικόν—i.e. in line with Hesiodic material—and he does not think that its author is Homer. But the surviving copies do not support his judgment. We simply cannot, without further evidence, infer from his comment that he knew of ancient manuscripts that did not contain λ 602–603. That the scholia should ascribe the authorship of λ 604 to Onomakritos is hardly probative, since Herodotos 7.6 made him notorious as a forger and associated him firmly with the editing of religious subject matter.

6.2 Theagenes of Rhegion

Cassio (2002:118) seems especially impressed by the report in the scholia A to Α 381: ⟨εὐξαμένου ἤκουσεν⟩ Σέλευκός φησιν, ἐν τῇ Kυπρίᾳ καὶ Kρητικῇ “ἐπεί ῥά νύ οἱ φίλος ἦεν”· καὶ Θεαγένης δὲ οὕτως προφέρεται· ἀπίθανον γὰρ τὸ †οδένυ? ???? ????? ??.λίαν φίλος ἦν. [25] Note that the lemma, which I have taken directly from the manuscript, is not ⟨ἐπεὶ μάλα οἱ φίλος ἦεν⟩ as Erbse and Dindorf print it. These editors quote for our convenience the portion from our critical text that the scholiast’s comment directly addresses. The scholiast, however, as might be expected for ease of reference on the page, quotes the beginning of the line he is annotating. This detail is informative because much of the force of Cassio’s argument hinges on the impression that Theagenes is offering a relatively minor variant reading to a specific section of a line in our modern editions. As Cassio (2002:118) puts it, “Theagenes quoted a Homeric line (Il. 1.381) with a variant.” And again: “if our source can be trusted, he quoted a rhapsodic variant in a specific line of our Homer” (118, his emphasis). With Müller (1891:36), I ascribe the reference to Theagenes in the scholion ultimately to Seleukos, the Homeric grammarian from Alexandria. Otherwise, as our sole source we are left with an anonymous scholiast who obviously postdated Seleukos and yet, more than half a millennium after Theagenes’ floruit, claimed detailed knowledge of his Homeric views. But crediting Seleukos for the entire scholion only mildly alleviates our problem. For one must wonder, too, how reasonable it is to expect that even Seleukos had access to Theagenes’ writings, of whom little more than a few testimonia have been preserved. [26] The floruit of Seleukos Homerikos, active at the court of Tiberius, is a full five hundred years after Theagenes. [27] According to Müller (1891:30–31), other than with Theagenes, testimonia and fragments allege his engagement with Aristophanes of Byzantion (ca. 257–180 BC), Athenokles from Kyzikos (Ⅲ/Ⅱ BC), Demetrios of Phaleron (b. ca. 350 BC), Timaios (ca. 350–260 BC), Philokhoros (ca. 340–260 BC), Aristotle, Demetrios Ixion (Ⅱ BC), a certain Kroton, Aristoxenos (b. ca. 370 BC), Plato, Diodoros Aristophaneios (ca. Ⅱ BC), and Khamaileon (ca. 350–post 281). Disregarding Plato and Aristotle, whose writings we can assume readily available in much later times, we can see from the list that no authority is older than the end of the fourth century except for Theagenes himself, who predates every other name by two hundred years. This seems legitimate cause for suspicion.

But let us grant, for the sake of argument, that the quotation is genuine and accurate. Does it justify Cassio’s impression that the text of the Iliad was fixed down to the level of line and word by the end of the sixth century? [28] I submit that its implications are rather more diffuse than his emphatic “specific” and “our” convey. For the statement is merely that ‘Theagenes proffers (προφέρεται) it so’. The meaning of προφέρεται (the verb is used both in the active and the middle) has a wide range that follows from its basic sense ‘to bring forth’. Dickey (2007:257) s.v. glosses it as “to utter, pronounce, use, cite (also in the middle).” [29] Here it may mean no more than that Theagenes ‘brings forth in this way’ the clause in question—to what end and in what way this evidence is proffered remains to be determined, whether in his writings, as reported by another who heard him recite or read his work, or in yet some other manner. From the scholion to be reviewed below, Wilamowitz-Moellendorff (1931–1932:2.215n2) inferred that Theagenes must have been a rhapsode himself: “Theagenes von Rhegion gilt als der erste, der die Götter in physikalische Mächte umgedeutet hat (Porphyrios zu Υ 67), und zwar in einem Buche über Homer. Da wird er ein Rhapsode gewesen sein, denn von diesen verlangte man auch Erklärung.” With this, Pfeiffer (1968:1.9–11) agrees. [30] That he wrote is stated both in the second scholion to be reviewed below [31] and in the Suidas 2.688 Θ no. 81 (Adler). [32] Modern studies of Aristarkhos’ editorial practice show that from a statement like this one cannot infer that Theagenes necessarily discussed the relative merits of the two variant readings ἐπεὶ μάλα οἱ φίλος ἦεν and ἐπεί ῥά νύ οἱ φίλος ἦεν, and that he expressed a reasoned preference for the one allegedly attested by the Cyprian and Cretan versions. It is just as possible that a Homeric text associated with his name was known to contain the reported variant. Or that he is reported to have recited the motive for Apollo’s helping Khryses in agreement with the Cyprian and Cretan editions. In other words, this may have been the one and only way he ever knew or sang the line. If he was indeed defending Homeric poetry against the attacks of Xenophanes and other like critics, it is also possible that he knew the two versions and specifically approved of the second. Assuming that Ludwich’s emendation of the crux is correct, he may have objected to the intensification of φίλος by μάλα as incredible or unlikely (perhaps even inappropriate) in view of the gods’ true nature. ῥά νύ, of course, implied no such thing. Seleukos’ interest in Homeric studies and theology (he wrote a Περὶ θεῶν) might explain why he cites Theagenes’ opinion here. [33]

But all of this is mere conjecture. In the end, what do we have? The fact that a late sixth-century scholar-cum-rhapsode apparently knew the episode that begins our Iliad: he knew that Khryses had petitioned Apollo, and that Apollo had answered his request ‘because he was (very?) dear to him’. This is indeed a “specific” line of “our” Iliad, but what is there to be surprised at? It is a well-known narrative law that the beginning, the climax, and the end of a story are its fixed points. If a story is in thematic flux, these are the first elements to be established. One may, in fact, say that only when these are settled do we really have a recognizable story. [34] Theagenes lived at the time when we first have indisputable proof of an interest in a personal author of the Iliad and the Odyssey. Chronologically, he heads the list of those who first investigated Homer’s poetry, family, and time. Although this report may yet be another example of the ubiquitous search for πρῶτοι εὑρεταί, there is in principle no reason to disbelieve it. [35] This implies that the poems had a sufficiently settled thematic shape to be recognizable units of composition that could be ascribed to an individual. It is not surprising that a line from the beginning of the Iliad that touches on the motivation of Apollo in answering Khryses would have a shape similar, though by no means identical, to the one in our vulgate. This last detail, namely, that Theagenes allegedly proffered a variant to the reading in our vulgate (whether he knew of another or not) must be duly stressed. If anything, this demonstrates a remaining degree of fluidity, even at the point in the poem where we first expect rigidity of theme and diction. [36] How reasonable is it to turn this into an argument for the fixity of the text of the Iliad? No necessary implication for the rest of the poem may be legitimately drawn, and, at this relatively late date, the rather limited scope of fixity entailed presents no difficulties to proponents of the evolutionary model.

What of the second scholion, from the scholia B to Υ 67, attributed by Dindorf to Porphyry? [37] We do well to heed the warning in Pfeiffer 1968:10 that, since the scholion “is obviously derived from a Stoic source,” it should be used “with the greatest caution.” And Feeney (1991:10) is doubtless right that “[i]t is in the highest degree unlikely that Theagenes devised the full panoply of allegorical technique which we see exemplified in [it].” Indeed, after saying that Homer’s account of the gods borders on the unprofitable and unseemly, and that he tells inappropriate stories about them, Porphyry (?) observes that some (οἱ μέν) use the poems’ language to solve the difficulty presented by this kind of charge (ἀπὸ τῆς λέξεως ἐπιλύουσιν), [38] holding that everything is said allegorically and stands for the nature of the elements, as e.g. in the disagreements of the gods (οἶον ⟨ἐν⟩ ἐναντιώσεσι τῶν θεῶν). What is striking about the scholion is the utter generality of its references. Not only is Theagenes not identified before the examples of allegorical reading are presented (we meet only with the indefinite ‘some’); but the label that should allegedly draw our attention to the famous theomachy of Iliad 20 is not the distinctive θεομαχία, ‘battle of the gods’, but the ordinary and generic ἐναντίωσις, ‘disagreement’. [39] Only after a list of element-to-god equivalences do we learn that ‘this type of defense, then, being very old and from Theagenes of Rhegion, who first wrote about Homer, proceeds in this manner, from the diction’. Porphyry only establishes a tenuous historical link with the man reputed to be the first exponent of this interpretive strategy. There is not a hint in his presentation that Theagenes’ actual writings stand with any specificity behind its details, as Feeney once noted. [40] The scribe who added this scholion to the Venetus B [41] doubtless thought that it helped to illuminate the word ἔναντα at Υ 67, which it annotates. This fact does not, however, prove that it was drafted by its author (ex hypothesi, Porphyry) with the specific language of our theomachy in view; nor does it substantiate the still weaker claim that the precise backdrop for its treatment was this particular Iliadic episode as opposed to divine confrontations generally, of which the theomachy—in whatever textual shape it was known to the author—may well have been a signal instance. Both the non-specific label, ἐναντίωσις, and the lack of a detailed correspondence between the scholion’s content and the express language of the episode speak against the proposal that this passage distinctly drives the critique. Before we can accept Cassio’s assertion that the work of Theagenes proves the existence in the late sixth century BC of a text of the Iliadic theomachy substantially like our own, we must embrace two tenuous and implausible links: first, that the scholion in its details faithfully represents Theagenes’ thought; and, second, that the scholion addresses itself to the theomachy that we know from our text. I have already shown that the former premise is unlikely. I will now demonstrate that the latter, too, is dubious at best.

Indeed, if we should let the Iliadic theomachy guide us, we would expect Apollo to fight Poseidon (Υ 67–68 Φ 435–469), Ares Athena (Υ 69 Φ 391–414), Hera Artemis (Υ 70–71 Φ 479–496), Hermes Leto (Υ 72 Φ 497–501), and Hephaistos Skamandros (Υ 73–74 Φ 331–382). Of these, Porphyry mentions and arguably pairs as “fire” against “water” Apollo and Poseidon, and Hephaistos and Skamandros. But for the pairing to work we must excise Helios as a gloss for Apollo, a detail that is certainly un-Homeric. And it is perhaps a little embarrassing for the allegorical reading that, at least in our Iliad, Apollo and Poseidon evade their confrontation. There is no attempt to pair Artemis “the moon” and Hera “the air” (?). If σελήνη is in fact a στοιχεῖον, how are these two “elements” supposed to “fight” each other? [42] Still more difficult to relate to the Homeric theomachy are Athena, Ares, Aphrodite, and Hermes. They are equated with ‘dispositions’ (διαθέσεις). One may readily oppose ‘sensibleness’ (Athena) to ‘senselessness’ (Ares). But the pair Aphrodite and Hermes is neither Homeric nor does it make much obvious sense to oppose ‘desire’ (ἐπιθυμία) to ‘speech’ (λόγος). [43] If instead we pair Athena not only with Ares but also Aphrodite, [44] Hermes is left without a partner. In the Homeric narrative, he was supposed to face Leto, a divinity that is conveniently left out of the narrative. [45] In the end, the only arguable good fit between our Homeric theomachy and the scholion is Hephaistos and Skamandros. What does all of this tell us about the stage of textual fixation this account implies for the Iliad? Very little, I submit. We merely learn that gods were involved in the action and that they had disagreements! Absent the poem, we could not guess that Poseidon and Apollo never fight; we would wonder whom Helios (or Hephaistos) had faced; we would assume that Aphrodite had clashed with Hermes; and we would never suspect that Leto was involved at all. Not much to claim, as Cassio (2002:118) does, that Theagenes’ alleged “allegorical interpretation of the battles of the gods in the Iliad … automatically means that a fixed Homer, in which those battles were inescapable, was already in existence.” His inference that Theagenes’ writing on Homer “must surely mean that written copies of his poems existed and circulated by 530–520 BC” (118) is a similar non sequitur. As a rhapsode, Theagenes could surely write (if writing he did) without needing to consult a copy of the poems. Even at a later time when the existence of written excerpts from largely fixed texts are not to be doubted, many who did not make their living performing Homer had an acquaintance with Homeric epic intimate enough to be able to comment on it with specificity and perceptiveness without the need to refer to written copies. There is no reason to question the existence of partial transcripts of performances by Theagenes’ floruit, but the “sure” inference that Cassio draws does not seem well grounded in the realities of Greek archaic culture.

I close this section with two examples from Cassio (2002:119–120), which, he thinks, make the point that the alleged written copies in existence by the late archaic period did not prevent new significant modifications or “interpolations.” This claim, which highlights Cassio’s presuppositions, can help us to understand how existing written versions of one or another of the poems’ episodes must have functioned in the culture of the time. The first example regards Η 334–335, the two-line proposal to convey home the bones of fallen Greek heroes. According to Cassio, “Jacoby established beyond doubt … [that this] is an Attic interpolation not earlier than the fifth c. BC.” [46] Whether one thinks that Jacoby’s argument succeeds or not, [47] these lines are a good example of what must have been true of existing written texts at the time. They will have been considered mere transcripts of individual festival performances: ordinarily, representative of the notional poem but without canonical authority. Rhapsodes may have used them as aids to training and performance, not as scripts to be memorized by rote for subsequent recitation. Individual members of the public may have owned partial texts—perhaps favorite episodes from the larger notional whole, perhaps more or less faithful transcripts of a particularly successful and memorable performance—for private enjoyment, for use at symposia, or, in the case of the few fortunate enough to afford an education, for reading and writing school exercises. [48] Against this understanding of the significance of hypothetical transcripts of the Iliad and the Odyssey, Cassio’s speculations strike me as a distortion of what must have been the cultural reality: “Either the new copies were never checked against the old ones, or, if they were, nobody had the courage to do away with interpolations of central importance to Attic culture” (Cassio 2002:119). Note his presumption: i) that new copies must have exercised a good measure of textual control; ⅱ) that those concerned should have felt a corresponding need to check the new against the old, to defend or enforce a status quo ante; ⅲ) that, in this case, there was either an inexplicable lapse of zeal and neglect of duty or else a failure of courage to preserve the existing text. There was, of course, no such thing. We can certainly expect at this late date a significant measure of enforced agreement with the poem’s central thematic substance and plot sequence, and even with favorite turns of diction for memorable passages. But, ordinarily, there will have been no sense of obligation to whatever detailed thematic boundaries the story happened to have at any one time in a particular telling. The verses in question concern a detail that a rhapsode should have felt free to recompose in performance—no need here to use the wooden notion that underlies the inorganic and deprecated ‘interpolation.’ It is only fair to conclude from their uniformly positive attestation in the manuscript evidence (whatever the suspicions of scholars, ancient and modern) that this recomposition was done to magnificent effect and the enthusiastic embrace of the Panathenaic audience. [49]

The second passage cited by Cassio (2002:119–120), Β 11–15 pertains to the lying dream Zeus sends Agamemnon. This is an example entirely fabricated by the sensitive conscience of the scholar. For, having decided that it was morally objectionable for Zeus to command a dream to deceive Agamemnon—here Cassio joins similarly troubled, morally sensitive men from antiquity, like Hippias of Thasos [50] and Sokrates [51] —he deems a perfectly ordinary example of textual fluidity an innovation that attempted to remove the alleged embarrassment. There was, of course, no such concern in the minds of the rhapsodes who happily recomposed either variant; nor, in all likelihood, in the mind of the average member of the public who listened to their performances. [52] That, after proffering the plausible pretext of Hera’s successful entreaty, instead of saying “and griefs have been fastened on the Trojans” Zeus should assert “and we are granting him to win his victory boast” hardly seems to me sensibly to aggravate Zeus’ responsibility (or his guilt, if one must condemn his words). Without any ancient statement to the effect, which might support the notion of a deliberate change that seeks moral relief, such judgments lie entirely in the eye of the beholder. If only this were the sole “explicit and undeniable” divine moral transgression! Quite another is the ancient case that we can reconstruct from Aristotle’s statements in the Poetics 1461a21–23 and the Sophistical Refutations 166b1–9. It depends on reading what in our text is Β 15, from the masculine caesura to the end of the verse, as West prints it: δίδομεν δέ οἱ εὖχος ἀρέσθαι. Our vulgate reads instead: Tρώεσσι δὲ κήδε’ ἐφῆπται (which West prints at Β 32 69). Aristotle reports in the Poetics that Hippias sought relief for his scruples in a change of ‘prosody’ (κατὰ δὲ προσῳδίαν, i.e. ‘accentuation’). Instead of δίδομεν, he must have read διδόμεν, the infinitive for the imperative, displacing the agency ever so slightly from Zeus to the destructive Dream. [53] Note well that Hippias is reinterpreting what for us is a variant reading to the vulgate. His ‘solution’ does not depend on substituting his own reading for the one in the vulgate. But this is precisely what Cassio’s argument would lead us to believe. Aristotle does not tell us that he himself was, or even that many in his own time were, troubled by δίδομεν (he was no purist like Sokrates!). His point is simply to illustrate alleged ‘solutions’ to alleged ‘problems.’ Nor does he necessarily imply that he was ignorant of the vulgate reading. It has not occurred to modern scholars, for whom every line has one and only one ‘right’ version—Homer’s—that, in a world tolerant of multiforms, Aristotle need only mention the one form that had given rise both to an objection and to a solution that only affected suprasegmentals. His passing in silence over the vulgate reading does not even mean that no one had found it objectionable: it was simply not susceptible of a solution like Hippias’. It is true that Hippias’ strategy betrays a desire to respect the received text to the maximum extent possible. But this is not to deny the reality of multiforms, only to affirm the value placed on traditional variants. As we have learned from the Southslavic oral poetic traditions in performance, and as I explore at length below in chapter 7, Homeric epic, whose rhapsodes recomposed their poems in performance, paradoxically combined a strong claim to a notionally faithful telling (cf. θ 487–491) with the freedom in practice to change themes and diction within a culturally acceptable range. For the Homeric poems, this range grew increasingly narrow with the passing of time. Only when two variants were consciously and polemically opposed did the source of their respective authority come into view and become problematic. This very polemical clash—in this case, played at the level of Panhellenic versus local traditions—is in view in the famous words of the Muses to Hesiod in Theogony 27–28. [54] In the context of irreconcilable, competing claims to truth, a tradition’s pretension to the authority of true inspiration could be reconceived as the authorship of a uniquely inspired author. This is arguably responsible in large measure for the traditional foregrounding of Hesiod’s authorship vis-à-vis the retiring persona of the Homeric narrator. During its formative period, Homeric epic did not overtly face the challenge of competing traditions. Hence, neither the Iliad nor the Odyssey explicitly acknowledges any rivals as such. If faced directly with the choice of variants, I do not doubt that Aristotle, like Hippias, might have thought desirable, if not possible, to determine on various grounds which one was truly Homer’s. And that, once this determination had been made, there would be no recourse, if morally offended by the choice, but to ‘solve’ with only minimal textual changes the perceived transgression. An acceptable solution would proceed on the grounds that the received text had been misconstrued or slightly altered by mistake. This must have been Hippias’ rationale. [55] At any rate, it is obvious to us—and, in all likelihood, to Aristotle—that Hippias’ solution is, to use Lucas’s colorful criticism, “poverty-stricken casuistry,” “even more pathetic” in light of the Dream’s announcement Διὸς δέ τοι ἄγγελός εἰμι (Β 26); [56] and, I may add, in light of the specification of Zeus’ agency (ἐκ Διός) in Tρώεσσι δὲ κήδε’ ἐφῆπται | ἐκ Διός (Β 32–33 69–70). [57]

Only one whose judgment is shaped by the experience of literacy, who defaults to thinking of such variants in terms of a morally offended writer bowdlerizing his edition of the Iliad, would readily embrace Cassio’s conjecture about the origin and rationale for the vulgate reading of Β 15. We are under no obligation to believe with him that it is “later than Aristotle” (2002:120). In a song culture where multiforms were not rare and recomposition was the mode of epic performance—however reduced its scope for change—there is no reason to assume that Aristotle knew, or cared to quote, every existing variant. And, if he actually knew both, it is possible, if not probable, that, unlike Cassio, he may have thought them morally equivalent. Were it not for the aims of his argument in the cited loci, he might well have quoted either one indifferently. That Aristotle’s multiform was not acknowledged by the Alexandrians is doubtless an accident of preservation. We cannot presume to have every statement Aristarkhos ever made about variant lines known to him.

6.3 The Name ‘Homer’

The earliest reliable attestations of the name ‘Homer’ (as it happens, along with Hesiod) come from Xenophanes, whose floruit is in the second half of the sixth century BC (the reference is explicit in DK 21 B10–11; contextual in B12–13). Somewhat later (ca. 500 BC) are Herakleitos’ (DK 22 B42, B56, and B105 = frr. 30, 21, and 63a Marcovich).

The famous fr. 6 (West = 10 Gentili) of Kallinos is rightly classified as dubium by West [58] —and not surprisingly. For, as Davison (1968:81–82) notes, we depend on a double emendation of Kαλαῖνος to Kαλλῖνος (twice) and Θηβαίοις to Θηβαΐς. The second is indifferent to the argument of this section, for any mention of Homer (whatever the connection) would suffice. But I can hardly judge the first emendation sufficiently reliable. Gentili’s “prob[avit] Davison” does not accurately reflect the rather tepid endorsement of the latter, who merely opines that “there is more to be said” for the first emendation than for the second. The statement makes sense as it stands, without any need to emend it: [59] ‘And epic poetry (ἔπη) about this war was also composed for the Thebans. [60] When he mentioned [lit., ‘having come to the mention of’] this poem (τὰ δὲ ἔπη ταῦτα), Kalainos said that Homer was its composer …’. Pausanias uses the periphrasis ‘to come to the mention of’, ἀφικνέομαι ἐς μνήμην, for ‘to remember’ (5.8.7, 7.19.8), ‘to bethink oneself’ (7.4.10), and ‘to mention’ (7.7.4–5). In this last instance, Pausanias uses it to anticipate his future treatment of a subject: ‘I shall again mention Kleomenes in my Arcadian narrative’. [61] This is the sort of prosaic periphrasis we should expect of an author who comes to a topic in the course of his narrative. The metaphor is one of coming upon (arriving at) something in the course of time, while thinking or in the process of recounting. If one must imagine that Kalainos is a poet, Pausanias’ words would seem to require a poem that goes from one topic to another and that, upon reaching the subject of Thebes’ war, would include an explicit statement of Homeric authorship. This, we might expect of an Alexandrian poet, say, Kallimakhos. And, not surprisingly, Kalainos has been emended to Kallimakhos with much greater plausibility than to a seventh-century Kallinos, of whom we hardly expect scholarship in verse and who, as far as we know, did not pen any literary discussions. [62] Once we surrender the notion that Kallinos of Ephesos is implied, as Scott observes, there is little gain in emending an unknown Kalainos to an unknown Kallinos (1922:359). All the same, the inference seems legitimate that epic poetry on the Theban War (whether referred to by the title Θηβαΐς) was current among the Thebans, and that a scholar of some authority (of time and place unknown, but probably in the late archaic or early classical period) had ascribed it to Homer. We may be troubled not to know the precise identity of this early and weighty authority, but this does not give us leave to extract from the text meaning at the cost of three unnecessary emendations and a somewhat forced reading.

One may still argue that an explicit mention of the name ‘Homer’ is not strictly necessary, and that a mere reference to his persona would do. Allusions to ‘the blind man from Khios’ would be the obvious example. The two oldest instances are the Hymn to Apollo 171–173 and a fragment (apud Stobaios 4.34.28) ascribed to a ‘Simonides’ that scholars have variously equated with Semonides of Amorgos or with Simonides of Keos. As to the Hymn to Apollo, most now place its composition in the second half of the sixth century (cf. Burkert 1979 and West 2003b:9–12), a date that gains us no greater antiquity than explicit instances of the name. As to Stobaios’ fragment, P.Oxy. 3965 (fr. 20 West = 7 Gentili) would now seem to confirm that it is not Semonides’, but Simonides’—bringing us, once again, to the late sixth century BC (cf. Davison 1968:73–77 and Sider 2001). Furthermore, links at the level of theme or diction between an archaic author and Homeric poetry are not enough to establish a terminus ante quem for Homeric authorship, for these are satisfactorily explained by the influence of the epic tradition on other competing poetic production (for the case of Stesikhoros see Burkert 1987); my concern here is strictly with the time when the locus of authority for the performance of Homeric epic moves from ‘inspiration’ (the god presiding over the public occasion) to the ‘authorship’ of ‘Homer.’ One final matter that requires attention is the reference to the Meles, the river of Smyrna, in Homeric Hymn 9.3 and the connection (made at least as early as the fifth century BC) with ‘Melesigenes,’ a competing name for ‘Homer.’ Graziosi (2002:72–76) helpfully reviews the evidence, reaching, in my opinion, the wrong conclusion: that the link between ‘Homer,’ the individual poet, and the Meles is at least as old as the hymn, i.e. with a terminus ante quem of ca. 600 BC. It is not with her dating of the hymn that I disagree: that a goddess probably native to Anatolia (cf. Lebrun 1987:251–253) should be hymned in her connection with Smyrna is hardly surprising; but given the destruction of the city by Alyattes ca. 600 BC and its refoundation as an important urban center only three hundred years later, the composition—here I join the consensus—must be at least late seventh-century, if not older (cf. Cassola 1997:303). On the other hand, the connection of ‘Homer’ with the Meles or Smyrna (apart from the hymn itself, which I believe does not establish it) is not attested earlier than sometime before the Peloponnesian war. [63] The Homeric Hymn 9 alone is left: the view that it makes a connection between a ‘Homer’ and the Meles has force solely on the assumption that Melesigenes can only derive from Smyrna’s river. If it could have arisen independently and can be shown to be an apposite choice for a poet’s sobriquet, then the connection with Smyrna would easily follow from the false etymology ‘born from/by the Meles’ (cf. West 2003b:310). Graziosi seems to assume without warrant that, since Smyrna was not a flourishing polis in the classical period, no one would have thought of linking a known ‘Melesigenes’ with an obscure river Meles. But this ignores the role that the hymn itself may have played—why should we assume that it too would have been obscure? its very transmission suggests otherwise—and the keen biographical interest in Homer’s place of origin, [64] which must have been accompanied by what I may call, faute de mieux, an ‘antiquarian’ concern even for the relatively obscure, a concern that the curious must surely have brought to their investigations. If Athens was to own Homer, since it could lay claim (at least in the eyes of some) to the foundation of Smyrna (Vitae 4.16 and 5.34), it is only natural that it should tie the poet’s name and the river. And so we now come to its etymology, about which Marx (1925:406–407) observes that it must derive from the aorist stem of μέλομαι: “Mελησιγένης wurde ein Mann genannt, der für seine Familie, sein γένος zu sorgen weiss.” But, of course, γένος (and the alternative γενεή) need not look forward in time to offspring: it may look backwards to genealogy and race—arguably an adequate generic characterization of the subject matter of Homeric poetry (especially, but not exclusively, its catalogs). To Sokrates’ question in Plato’s Iōn whether Homer had not recounted ‘the origins [i.e. ‘birth’ or ‘descent’] of gods and heroes’ (γενέσεις καὶ θεῶν καὶ ἡρώων 531c8–d1), the rhapsode answers with an unqualified affirmative. Indeed, taking the perspective of a later age Μ 23 describes the Trojan War as one in which ‘the semi-divine race of men fell in the dust’ (κάππεσον ἐν κονίῃσι καὶ ἡμιθέων γένος ἀνδρῶν). And μέλομαι, in turn, is readily used to denote the poet’s engagement with the subject matter of his song—the Odyssey, for example, mentions the Argonautic tradition in an offhanded way simply by noting that the Argo was everyone’s concern (Ἀργὼ πᾶσι μέλουσα μ 70). So I conclude that ‘Melesigenes’ is a sobriquet that describes ‘him whose [poetic] concern is the races [of gods and heroes]’. Applicable to any rhapsode of heroic epic, it hardly proves the antiquity of an individual conception of Homeric authorship.


[ back ] 1. Cf. Cassio 1999:80.

[ back ] 2. For the Peisistratean recension as a myth of origins, see Nagy 1996c:77–80. Note in particular, his acute observation at 80: the various versions of the foundation myth do not aim to prove “the hypothetical existence of some unattested textual transmission of Homer but rather [to explain] the institutional reality of ongoing performances of Homeric song at the Feast of the Panathenaia in Athens. That is, the myth is concerned with the performance of Homer.” A few convenient bibliographical items from the massive literature about the alleged recension are Merkelbach 1952, Davison 1955, Sealey 1957:342–349, and Jensen 1980:128–158. (Others have been noted above, §3 n. 37.)

[ back ] 3. This is also the view of Shear 2000:99. Although there are many things in her fascinating book to disagree with (cf. Burgess 2000), on this matter she is right: “Megara claimed that the Athenians had added a line to the Salamis entry in the Catalogue of Ships in order to bolster their claim to Salamis. The Athenians stoutly denied the Megarian accusation … . If written texts of the Catalogue had existed, then the presence of an alteration would have been made immediately obvious by a comparison of the Athenian text with other existing texts. … The Megarian accusation was only possible in an oral society when the discussion eventually resorted to arguments based on what ‘our bards’ say as opposed to what ‘your bards’ say.”

[ back ] 4. I do not wish to join here the debate whether Aristarkhos omitted or simply marked the verse as suspect, a debate that involves the meaning of Aristarchean atheteses and a comprehensive grasp of his editorial principles. (See Apthorp 1980:167 on this issue as it applies to the scholia A to Γ 230 and Δ 273.) I merely note that West, too, agrees that the Alexandrian scholar did not find the verse attested in all of his manuscripts.

[ back ] 5. Cf. Apthorp 1980:166.

[ back ] 6. Piccirilli 1975:67–73 F 3.

[ back ] 7. That is to say, that the ‘nonsense’ was not the story that Solon had cited Homer but that he had forged the key verse.

[ back ] 8. Cf. Piccirilli 1975:133–134 F 21b.

[ back ] 9. Pace Wilamowitz-Moellendorff 1884:243. Cf. FGH 485 F6n28 (vol. 3b, Noten, p. 232): “Ich teile die bedenken von Wilamowitz … nicht; aus Strabons bezw. Apollodors ausdruck ἀντιπαρωιδῆσαι ist garnichts zu schliessen.” Although Jacoby does not think the Megarian verses might have stood in a Megarian ἔκδοσις, Piccirilli (1975:134) writes, “ciò non è però del tutto impossibile.”

[ back ] 10. νέας is not used in the attested Catalogue of Ships, but it is in Α 487 Ν 96 101 620 Ξ 392 Ρ 612 γ 153 162 299 δ 582 κ 15 91 λ 124 (=ψ 271) ξ 258 (=ρ 427) and Hesiod Works and Days 247, fr. 205.6 MW.

[ back ] 11. One can readily understand why the beginning of Β 557, which states the name and domain of the hero who leads the ensuing contingent, should have arrived at a comparatively more advanced degree of textual fixation sooner than the rest of the section it headed.

[ back ] 12. Bonner 1905:54 and 58–59. A party who did not merely refer in a general way to a law but had the clerk read its text aloud was responsible for supplying a truthful copy upon the penalty of death. For Bonner (1927:195–196), there was no established procedure other than a challenge and a suit for perjury to guarantee that evidence so presented was true and accurate. It is not clear whether Harrison (1968–1971:2.134–135) makes the clerk himself responsible for the copy of the law or decree whose text he was to read aloud. His statement that the laxity of citation was improved by “the rule that evidence … was to be in the form of written depositions read out by the clerk at the trial” (135) implies the view that the clerk either copied the law himself upon request or else checked for accuracy the copy supplied by the litigants. To judge from Demosthenes 19.270, sometimes the clerk was tendered the text only when he was called to read it. For written and oral forensic testimony, see Bonner 1905:46–53; Calhoun 1919; Harrison 1968–1971:2.134–135 and 139; Humphreys 1985b:317 and 321; Todd 1990:29n15; and Pébarthe 2006:315–343. Thomas (1992:148–149) observes that growing confidence in documentary evidence is a fourth-century development (cf. Rubinstein 2000:72–75, with added nuance). Harris (1989:71–73) adds that, even then, “legal practice, like the administration of large-scale business, remained to a considerable extent oral and independent of documents” (at 72). For large domains like the law of sale oral testimony still reigned supreme. Sickinger 2004 offers a complementary perspective, which suggests that “despite the persistence of oral practices, participants in legal actions frequently relied on written texts” (94). He does not, however, directly address the question of the accuracy of the clerk’s texts or their source.

[ back ] 13. Right before he reports on Solon’s textual insertion, he recounts the argument based on the coincidence of Salaminian and Athenian burial customs. For Plutarch, both were arguments presented to the Spartan arbitrators.

[ back ] 14. Whether one thinks that this encounter was traditional or not since the earliest identifiable stages of the Odyssean tradition, it is arguably the climax of the Nekuia. Cf. Heubeck and Hoekstra 1989: “[t]he encounter with Heracles is deliberately placed last” (114 ad 601–627); “[it is] clearly intended as the concluding climax to the episode” (116 ad 630–631.) To that extent, it is indispensable to the architecture of the episode as we know it.

[ back ] 15. In other words, one might argue that a keen awareness of the gruesome manner in which he had attained his apotheosis would render psychologically plausible Akhilleus’ dwelling on Herakles’ death, rather than his divine afterlife, as a paradeigma for his own fate. The familiar account of Herakles’ death (without the detail of the pyre) is at least as old as Hesiod fr. 25.20–33 MW. It is vain to speculate about Akhilleus’ knowledge of Herakles’ afterlife and how this information might affect the tenor of his words. His emphasis falls squarely on the event of dying. Even the Hesiodic fragment, which dwells at length on Herakles’ present divinity (26–33), describes his end in grim familiar terms: δ[εξ]αμένωι δέ ο[ἱ αἶψα τέλος θανάτοι]ο παρέστη· | καὶ] θ̣άνε καί ῥ’ Ἀΐδ[αο πολύστονον ἵκε]το δῶμα (24–25). (Like λ 602–603, the Hesiodic report of the hero’s divine blessedness could not escape the suspicions of ancient athetizers and is obelized in P.Oxy. 2075.) One should consider, moreover, that Herakles’ apotheosis is not entirely unrelated to the poetic immortality through which Akhilleus manages to transcend his own death (Σ 121).

[ back ] 16. His analysis should be consulted.

[ back ] 17. Burkert 1985: “[He] is the greatest of the Greek heroes and yet thoroughly untypical” (208).

[ back ] 18. For these statements and the corresponding references, see Burkert 1985:210 and West 1997:465.

[ back ] 19. τὰ μέν νυν ἱστορημένα δηλοῖ σαφέως παλαιὸν θεὸν Ἡρακλέα ἐόντα.

[ back ] 20. The narrative of his apotheosis would serve to harmonize his original divine character with his heroic identity.

[ back ] 21. On the other hand, that he added them to accommodate old religious thinking might be true, if sheer conjecture without a shred of documentary support.

[ back ] 22. Nothing in the logic of the case requires that we condemn λ 602–603 as an inorganic interpolation into an already fixed text. This is so, even if we were to accept the view that these verses did not belong in older versions of the poem. Why must we preclude the possibility that it was in the normal course of its diachronic development that rhapsodes organically wove into the poetry the (ex hypothesi) later theme of Herakles’ apotheosis? If so, this would be an ordinary instance of the diachronic layering of themes that follows from a live tradition of poetic recomposition in performance. That is of course not what Cassio has in mind: he is convinced that the offending material bears no organic relation either to its context or to the ordinary growth of the poem and that its insertion does violence to the poetic integrity of the original. It bears saying, once again, that his case turns entirely on the subjective feel of the critic for what makes good poetry, and hence for what ‘Homer’ (or the authentic tradition) might have composed.

[ back ] 23. Recent treatments of Theagenes in the context of early allegoresis can be found in Richardson 1975; Svenbro 1976:108–138; Rispoli 1980; Feeney 1991:8–11; Ramos Jurado 1999; Ford 1999; Ford 2002:68–72; Zumbo 2002; Struck 2004:26–29; and Naddaf 2009 esp. 108–114. Among the older treatments, see Wehrli 1928:88–94.

[ back ] 24. Cassio 1999:81–82 and 2002:118–119.

[ back ] 25. On this scholion, besides Erbse 1969–1988 ad loc. see Ludwich 1884:1.192 and Müller 1891:35–36.

[ back ] 26. His section in Die Fragmente der Vorsokratiker (1.51–52) is merely a page long.

[ back ] 27. For general information on Seleukos, see the New Pauly s.vv. “[13] S[eleucus] Homericus.” See also Müller 1891.

[ back ] 28. Cassio, of course, does not put his case in such strong terms. But his emphasis on the “specificity” of the line in question once again suggests a modern text with its standard line numbers. That the variant is about two syllables only, ῥά νύ versus μάλα, further evokes the image of a late sixth-century van Thiel laboring over manuscript collations.

[ back ] 29. Cf. Ludwich 1884:1.113n128.

[ back ] 30. Wehrli 1928:91 disagrees: “Einfach Rhapsode war er darum nicht, denn dann würde er nicht Grammatiker heißen.” Cf. Rispoli 1980:249–250.

[ back ] 31. [Θεαγένης] πρῶτος ἔγραψε περὶ Ὁμήρου.

[ back ] 32. εἰσὶ δὲ καὶ ἄλλοι δύο Θεαγένεις, εἷς μὲν ὁ περὶ Ὁμήρου γράψας … .

[ back ] 33. The elucidation that starts with ἀπίθανον γὰρ need not be Theagenes’ own. As scholia tend to be, the flow of logic is compressed and unclear. To construe the clause with what precedes it one must supply the link. For example, ‘And Theagenes also proffers it so. For [he says that] the exaggeration “he was very dear” is unconvincing’. Or, ‘Seleukos says that in the Cyprian and the Cretan versions [the reading is] “since he was dear to him”. And Theagenes also proffers it so. For [Seleukos says that] the exaggeration “he was very dear” is unconvincing’ (that is, Seleukos is the one who objects to μάλα and he adduces Theagenes and the Cyprian and Cretan versions to show that there was an old reading that did not involve the exaggeration he deprecated). Or again, ‘And Theagenes also proffers it so. [And it is no wonder that there are texts that do not read μάλα] for the exaggeration “he was very dear” is unconvincing’. Or finally, ‘And Theagenes also proffers it so. [This must be the correct reading] for the exaggeration “he was very dear” is unconvincing [and unworthy of Homer]’. I offer these alternatives exempli gratia simply to make the point that on the assumed and supplied link depends whether the explanatory clause is Theagenes’ or Seleukos’ (or even the scholiast’s).

[ back ] 34. The only partial exception is the end. Climax and end are often one and the same. Sometimes the climax is sufficient to round up a narrative, and to the end is left only the tying of loose threads. If so, it may take various forms without affecting the perceived integrity of the story. Tolkien’s narrative of The Lord of the Rings is of that sort. The Hobbits’ return to the Shire (“The Scouring of the Shire”) is strongly overshadowed in the mind of the reader by the climax of Mount Doom and the victory of Gondor. It is not by chance that Peter Jackson’s celebrated movie version does not include this return and ends with the Grey Havens.

[ back ] 35. περὶ γὰρ τῆς Ὁμήρου ποιήσεως γένους τε αὐτοῦ καὶ χρόνου καθ’ ὃν ἤκμασεν προηρεύνησαν πρεσβύτατοι μὲν Θεαγένης τε ὁ Ῥηγῖνος κατὰ Kαμβύσην γεγονὼς καὶ Στησίμβροτος ὁ Θάσιος καὶ Ἀντίμαχος ὁ Kολοφώνιος Ἡρόδοτός τε ὁ Ἁλικαρνασσεὺς καὶ Διονύσιος ὁ Ὀλύνθιος … . (Tatian Oratio ad Graecos §31.2 Goodspeed [p. 31 Schwarz] = DK 8 1).

[ back ] 36. Even granting arguendo the doubtful reliability of the scholion, for Cassio’s point to succeed we must further assume: i) that Theagenes was deliberately proffering a preferred variant (i.e. he knew that his version was one of several); ⅱ) that there was a base text from which his reading was conceived as a departure, i.e. an agreed text broadly received by one and all; ⅲ) that the shape of this text had largely crystallized in a final form, not only as regards the disputed ‘line’ but also elsewhere. Pace Cassio, none of this follows from the scholion.

[ back ] 37. τοῦ ἀσυμφόρου μὲν ὁ περὶ θεῶν ἔχεται καθόλου λόγος, ὁμοίως δὲ καὶ τοῦ ἀπρεποῦς· οὐ γὰρ πρέποντας τοὺς ὑπὲρ τῶν θεῶν μύθους φησίν. πρὸς δὲ τὴν τοιαύτην κατηγορίαν οἱ μὲν ἀπὸ τῆς λέξεως ἐπιλύουσιν, ἀλληγορίαι πάντα εἰρῆσθαι νομίζοντες ὑπὲρ τῆς τῶν στοιχείων φύσεως, οἶον ⟨ἐν⟩ ἐναντιώσεσι τῶν θεῶν. καὶ γάρ φασι τὸ ξηρὸν τῶι ὑγρῶι καὶ τὸ θερμὸν τῶι ψυχρῶι μάχεσθαι καὶ τὸ κοῦφον τῶι βαρεῖ. ἔτι δὲ τὸ μὲν ὕδωρ σβεστικὸν εἶναι τοῦ πυρός, τὸ δὲ πῦρ ξηραντικὸν τοῦ ὕδατος. ὁμοίως δὲ καὶ πᾶσι στοιχείοις, ἐξ ὧν τὸ πᾶν συνέστηκεν, ὑπάρχει ἡ ἐναντώσις καὶ κατὰ μέρος μὲν ἐπιδέχεσθαι φθορὰν ἅπαξ, τὰ πάντα δὲ μένειν αἰωνίως. μάχας δὲ διατίθεσθαι αὐτόν, διονομάζοντα τὸ μὲν πῦρ Ἀπόλλωνα καὶ Ἥλιον καὶ Ἥφαιστον, τὸ δὲ ὕδωρ Ποσειδῶνα καὶ Σκάμανδρον, τὴν δ’ αὖ σελήνην Ἄρτεμιν, τὸν ἀέρα δὲ Ἥραν καὶ τὰ λοιπά. ὁμοίως ἔσθ’ ὅτε καὶ ταῖς διαθέσεσι ὀνόματα θεῶν τιθέναι, τῆι μὲν φρονήσει τὴν Ἀθηνᾶν, τῆι δ’ ἀφροσύνηι τὸν Ἄρεα, τῆι δ’ ἐπιθυμίαι τὴν Ἀφροδίτην, τῶι λόγωι δὲ τὸν Ἑρμῆν, καὶ προσοικειοῦσι τούτοις· οὗτος μὲν οὖν τρόπος ἀπολογίας ἀρχαῖος ὢν πάνυ καὶ ἀπὸ Θεαγένους τοῦ Ῥηγίνου, ὃς πρῶτος ἔγραψε περὶ Ὁμήρου, τοιοῦτός ἐστιν ἀπὸ τῆς λέξεως (Schrader 1880:240–241 = DK 8 2 = MacPhail Jr. 2011:240–242). Although I depart from it at several points of significance to my argument, I print for convenience the translation by Ford 1999: “In general, [Homer’s] account of the gods tends to be worthless and unsuitable, for the myths he tells about the gods are inappropriate. To such charges as this, some reply on the basis of Homer’s way of speaking [lexis], holding that everything is said by way of allegory [allēgoria] and refers to the nature of the elements, as in the passage where the gods square off against one another. For they say that the dry battles with the wet, the hot with the cold, and the light with the heavy. Moreover, water extinguishes fire while fire evaporates water, so that there is an opposition between all the elements composing the universe, which may suffer destruction in part but remains eternal as a whole. In setting out these battles Homer gives fire the name Apollo, Helius, or Hephaestus, he calls water Poseidon or Scamander, the moon Artemis, the air Hera, and so on. In a similar way he sometimes gives names of the gods to human faculties: intelligence is Athena, folly is Ares, desire Aphrodite, speech Hermes, according to what is characteristic of each. Now this kind of defense is very old and goes back to Theagenes of Rhegium, who first wrote about Homer” (35). Cf. also the translations in Struck 2004:27 and MacPhail Jr. 2011:241–243. For Dindorf’s ascription, see Dindorf 1875–1888:3.ⅶ.

[ back ] 38. For the meaning of the expression, see Carroll 1895:40–55 and Combellack 1987 (“the solution based on language,” at 219).

[ back ] 39. That θεομαχία was used specifically for literary narratives of divine battles, including Homer’s at Iliad 20, is clear from Plato Republic 378d5; Protagoras DK 80 A30; [Longinos] Peri hypsous 9 §§5, 8; scholia A ad Ο 212a and bT ad Υ 4a; Herakleitos Allegoriae 52.1; and Proklos In Platonis Rempublicam commentarii 1.87.12 Kroll. The word ‘battle’ (μάχη) does appear a few lines later, but here again it is not the distinctive θεομαχία, which might evoke the specific literary episode, but the plural and seemingly generic μάχας (μάχας δὲ διατίθεσθαι αὐτόν).

[ back ] 40. Feeney 1991: “It is plain that the author of this passage knows nothing at first hand from Theagenes’ writings, and is aware only of a tradition that he was the first to use ‘this sort of defence’” (9). Naddaf 2009:125n37 strikes me as unfounded and historically naive.

[ back ] 41. Venetus graecus 821 (olim Marcianus graecus 453) folio 270 recto.

[ back ] 42. Unless the moon is supposed to be “heavy” and the air “light.” Little help is to be found in Herakleitos the Allegorist, who in Allegoriae 57.2–4 resorted to a fanciful etymology of Artemis (ἀερότεμις from τέμνω) to motivate the confrontation. This work (also known as Homeric Problems) must be first-century BC or later (probably ca. AD 100; cf. Russell and Konstan 2005:ⅺ–ⅻ). Hence, his statements, like those of the later Porphyry and the much later Eustathios, pace Ramos Jurado (1999:54n75), make no more plausible the suggestion that Theagenes may have adopted a similar interpretive strategy in the late sixth-century BC. The same may be said of the assertion that the scholiast’s failure to pair Hermes and Leto explicitly is not problematic because “entre los alegoristas es tradicional” (Ramos Jurado 1999:54). None of the allegorists cited is any earlier than Herakleitos himself. The question is not whether late allegorists, half a millennium or more after Theagenes’ allegedly seminal work, could come up with highly abstract equivalences that would explain the Homeric pairs. After all, something had to be made of the received text. What we must consider is whether Theagenes himself could have established such correspondences in a cultural—and, if he was a rhapsode, professional—milieu that must have been closely wedded to Homeric myth and language. Once again, Ramos Jurado (1999:56) fails to persuade when he adduces Anaximander to recommend the proposal that Theagenes could have offered such an exegesis in his own time. If Theagenes had been of the same intellectual cast as Anaximander, he could hardly have thought it necessary to allegorize in the first place (cf. Naddaf 2009:113–114). It is not a coincidence that the Ionian philosopher played no recorded part in the development of allegory; or that Naddaf (2009:109) begs the question of his relationship to Theagenes by asserting on the basis of the scholion that “[i]t is as if Theagenes were a disciple of Anaximander.”

[ back ] 43. The polarity would work if we translated ‘passion’ and ‘reason’. But the λόγος that is attributed to Hermes is not likely to be ‘reason’, but ‘speech’.

[ back ] 44. Although Aphrodite does not appear in the original list of Iliad 20, Athena ends up confronting her too (Φ 423–433).

[ back ] 45. Like Apollo and Poseidon, Hermes and Leto end up not fighting.

[ back ] 46. Citing Jacoby 1944.

[ back ] 47. His article has received near universal praise. But there are a few dissenting voices, on which see Shive 1996, himself a critic.

[ back ] 48. For the functional differences between transcripts, scripts, and scripture, see Nagy 1996c:110–112.

[ back ] 49. We cannot dismiss the possibility, perhaps likelihood, that Aristarkhos had non-Athenian Homeric manuscripts that did not include the verses. But without testimony to this effect we simply cannot say. Cf. Erbse’s references in his scholia ad loc.

[ back ] 50. Aristotle Poetics 1461a22.

[ back ] 51. Plato Republic 383a8.

[ back ] 52. The moral fabric of ancient thought is often more complex than our own categories allow for. See, for example, 1 Kings 22:19–23. An understanding of divine testing and double determination that confined culpability to the human sphere, however paradoxical to us, was for many ancients an effective theodicy that obviated the need to excise offending passages. This is not to deny that there were then, like today, those who had scruples and who for relief adopted diverse textual and interpretive strategies.

[ back ] 53. Cf. Lucas 1968:242–243 and West 2001:175. Examples of this infinitival form (not used for the imperative) are found in Pindar Nemean 7.97 and Isthmian 8.60. Although not exampled in Homer, διδόμεν would be perfectly regular. Note the comparable τιθέμεν, also for imperative, in Hesiod Works and Days 744. In the Sophistical Refutations Aristotle seems to broaden the authorship of Hippias’ proposal to ‘some’ (ἔνιοι), if λέγοντες, as seems likely, governs the ὅτι-clause.

[ back ] 54. See below, §8.2.

[ back ] 55. It bears repeating that we do not know that Aristotle was similarly offended, that he felt the need to choose between known variants, or that he would even have thought it possible to determine which one was truly Homer’s. Tolerance for multiforms is best lived in scholarly skepticism, and in the reception of a public that either does not have the means or does not feel the need or the desire to hold one performer to the version of another.

[ back ] 56. Lucas 1968:243.

[ back ] 57. West’s decision to print Aristotle’s line at Β 15, despite the lack of manuscript support, but to respect the vulgate at Β 32 69 depends not only on his conviction that “[i]t fits perfectly here” (this and the following two citations are taken from West 2001:175), but presumably also on the belief that Aristotle must not have known the latter in the former’s sedes (“[T]he quoted hemistich … must have stood in line 15”) and that the Dream is unlikely to have repeated verbatim Zeus’ δίδομεν—presumably a ‘royal we’—making it his own (“δίδομεν could not metrically be converted into the third person”). But epic practice is to repeat the lines verbatim to the extent that syntax and the change of speaker allow. Hippias’ proposal, however untenable on other grounds, allows for this repetition, so long as Β 15 uses the infinitive and Β 32 69 the indicative. ἐκ Διός in Β 33 70 makes clear that the authority behind δίδομεν is not the Dream’s, but Zeus’.

[ back ] 58. West 1989–1992:2.245, index verborum s.v. Ὅμηρος. For this witness, Bernabé’s Poetae epici graeci I.21 no. 2 (under Thebais, Testimonia) should also be consulted.

[ back ] 59. Davison 1968: “Θηβαίοις is perfectly intelligible as it stands” (81).

[ back ] 60. For Homer composing poems or passages for various individuals (or cities) in return for livelihood or hospitality, see e.g. Vita Herodotea 196–207 and 346–354.

[ back ] 61. Kλεομένους μὲν δὴ καὶ αὖθις ἐν λόγοις τοῖς Ἀρκαδικοῖς ἀφιξόμεθα ἐς μνήμην.

[ back ] 62. Cf. Scott 1922:359.

[ back ] 63. The Certamen (cf. Graziosi 2002:73n62), at best, will not get us earlier than the classical period. Kritias (DK 88 B50) only mentions a river, not specifically the Meles, as Homer’s father, and, like Stesimbrotos’, his floruit is, at any rate, late fifth-century. The report that Pindar (fr. 264 Sn-M) made Homer hail from Smyrna (as Graziosi 2002:78 shows) is contradictory and unreliable. Thus, we are left with Euagon’s reference in the Certamen 20 (as Eὐγαίων), whose floruit Fowler (2000:102) places “ante bellum Peloponnesiacum.”

[ back ] 64. Starting perhaps with Theagenes of Rhegion, DK 8 1. Cf. Heath 1998:26.