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.

3. The Technology of Writing

Favoring a very early, ninth-century BC origin and writing of Homeric poetry, [1] Ruijgh 1995 defended the view of a commensurately early derivation of the Greek alphabet ca. 1000 BC. [2] Hence, he pronounced the rightly famous argumentum ex silentio by Carpenter 1933 of no probatory value, because early writing substrates were perishable (wood, papyrus, and parchment) and could not be reasonably expected in the archaeological record (Ruijgh 1995:36–38). But this argument has been effectively countered by Burkert (1992:28) and Marek (1993a:33), who note that Carpenter’s thesis has only been strengthened by the persistent failure to turn up ceramic shards from before the eighth century with Greek alphabetic writing despite the enormous increase in the number of archaeological discoveries. [3] Ruijgh (1995:38) also adduced the case of Cyprus to show that there are undeniable examples of continuity in the use of writing throughout the Dark Age. [4] Lastly attested in the eleventh century, it reappears in an evolved form in the second half of the eighth: “Aucun texte datant de la période 1025–725 ne nous est parvenu, mais il est évident que l’écriture est restée en usage pendant les trois siècles en question, sans doute surtout sur des matériaux périssables.” But there are alternatives to Ruijgh’s theories that do not presuppose extensive writing on perishable material between the eleventh and the eighth centuries. Egetmeyer (1998:233) believes that there was a contraction in the Dark-Age use of writing in Cyprus, and that Paphos, the provenience of the oldest two seventh-century inscriptions (of about twenty-five extant), was responsible for the preservation of writing: “In dieser Stadt, welche das religiöse Zentrum der Insel ist, konnte die Schrift durch die politisch unruhige Zeit hinübergerettet werden, und kam dann von hier aus zu einem neuen Aufschwung” (242). [5] The Mycenaean Greeks who arrived in Cyprus at the time of the great twelfth-century BC disturbances did not bring Linear B with them, but adopted the Cypriot syllabary instead. [6] Egetmeyer (1998:233) cites as a probable reason for this paradoxical development that the majority of the emigrants set out from the mainland after Linear B had already fallen into disuse. Similarities between Arcadian and Cypriot that are not present in Mycenaean Greek suggest a not insignificant period of time spent in this area of the mainland before the migration. Linear B probably disappeared from continental Greece ca. 1200 BC with one last round of destructive fires that struck the administrative centers. [7] Given the duration of the upheavals, it is likely that it had fallen into disuse in various parts of the mainland at an earlier time. Karageorghis (2002:71–113) examines the archaeological record of these unsettled times and concludes that it is consistent with migrations and extensive political upheavals in the Aegean from the fourteenth through the eleventh century (71). [8] The foundation of separate settlements by the refugees, for example, at Pyla-Kokkinokremos (Karageorghis 2002:74–78) [9] and Maa-Palaeokastro (Karageorghis 2002:78–81), should have given the immigrants from the mainland opportunity to continue their use of Linear B if it were still current among them when they set out, [10] just as they imported foreign cultural elements like bathtubs, burnished ware, fibulae and gold rivets of Aegean type, and central hearths in communal halls (Karageorghis 2002:78). Therefore, the example of Cyprus does not serve, as Ruijgh intended, to buttress the suggestion of a continuous use of writing during the Dark Age which might justify his early date for the recording of Homeric poetry in writing. On the contrary, the abandonment of Linear B in favor of the Cypriot syllabary suggests that literacy had been largely lost in the mainland by the end of the Mycenaean civilization. [11]

As noted above, some scholars have used Herodotos 5.58 to argue for an eighth-century recording of the Homeric poems on leather. I cannot decisively rule out such a scenario, but it is extremely implausible in the face of the technical complexity and cost of manufacturing durable parchment. [12] That a Dark-Age strong man should have had the technical expertise at hand to prepare the hides of dozens of slaughtered animals for extensive writing stretches credulity. It would take a Persian king to assemble an archive on skins of sheep and goats. [13] That the king of Byblos in the eleventh century should have imported large quantities of papyrus from Egypt despite the attested native use of leather in the countries bordering on Palestine shows how impractical animal hide was felt to be as a writing substrate already at an early time. [14] This sentiment lies behind the rationale in Herodotos that only (‘for scarcity of papyrus’) did the Ionians resort to . [15] Lewis (1974:86) criticizes scholars who, insisting on the need for writing for the composition of the Homeric poems, infer that the material used was papyrus: “Something more substantial will be required before it can be established that papyrus was a writing material used by the Greeks of the Heroic age.” [16] Even so stalwart an advocate of the eighth-century recording of Homer as Powell rejects the use of leather for that purpose. [17] Only in the second half of the seventh century, after Greek mercenaries joined the service of Psammetichus I, and once Greek traders were represented in Naukratis, can papyrus have grown common in the Greek world. [18] Much has been made of the name used by the Greeks to refer to the Phoenician city Gbl (Akkadian Gublu and Hebrew Gəbāl). Some scholars think that is the Hellenized form of Gbl, and from this they infer that the Greeks must have called papyrus after the city that was their primary source for it. They conclude, further, that papyrus was imported from Phoenicia before the Greek presence at Naukratis. [19] But as Heubeck (1979:154) writes, following Masson (cited in 154n787), this implies the rather unlikely derivation of from Gbl; and the similarly dubious associative chain ‘toponym → material → plant.’ [20]

Considering the question of the substrate on which a monumental Iliad might have been written, Robb (1994:256) observes:

Excluded as materials … must be stone, painted (whitened) wooden boards, clay whether fired or not, wax, linen, metals, and bone—all used to preserve writing in the ancient Near East, and some of them in Greece as well. Unless papyrus is assumed (as by Barry Powell), only leather or skins specially prepared would be a plausible candidate. Even so, the motive for that effort—no negligible one even for an experienced scribe in a more literate century—as early as 700 BC frankly eludes me.

I start with the ‘Hawara Homer,’ a well-known papyrus of Book 2 of the Iliad. [22] The column-to-column width is 23.3 cm and the number of lines in a column, 22. [23] Since the Iliad has 15,114 lines without Book 10, and 15,693 with Book 10, [24] the total length required would be 160 and 166.5 meters respectively. [25] To give an easy point of reference, the length of a football field is 91.44 m without, and 109.73 m with, end zones. With the Doloneia, our written text would make its way from end zone to end zone and half-way back to the center of the field. The column height of the papyrus is 14.7 cm. Unfortunately, we do not have the upper margin; the lower margin is no less than 4.8 cm. Excluding the margins, without the Doloneia the surface area would be 23.5 m2 (24.5 m2 with it). Since the Odyssey has 12,110 lines, it is 80% of the length of the Iliad without Book 10, i.e. it would take a bookroll approximately 128 m long (more than the length of a football field). Of course, as Johnson (2004:144) explains, actual Homeric rolls only rarely contained more than one book of Homer. [26] A copy of Philodemos’ On Piety executed in one exceptionally long roll was only about 23 m in comparison. [27] Johnson (2004:148) speculates that MP3 980 (P.Berol. 16985, also a luxury copy) contained Books 19–22 of the Iliad (some 2700 lines) and measured 19 m in length; [28] this corresponds to 106 m for the entire Iliad without the Doloneia. But since this and the Hawara Homer are exquisitely executed copies, it will be helpful to consider humbler extant bookrolls of Homer from which to estimate hypothetical full-length volumes. These can be found in Johnson’s Table 3.7 (2004:217–230). What follows are rough full-volume length estimates (all without the Doloneia) extrapolated from Johnson’s calculations and rounded (up or down) to the nearest half-meter: [29] P.Oxy. 0445, 40 m; [30] P.Oxy. 3155, 32.5 m; P.Oxy. 0687, 95 m; P.Oxy. 1815, 96.5 m; P.Oxy. 3663, 100.5 m; P.Oxy. 3323, 53.5 m; P.Oxy. 0020, 133 m; P.Oxy. 0223, 133 m; MP3 917.1 (olim 919), 40.5 m; MP3 773, 66 m; MP3 855.1, 60 m; MP3 819, 61.5 m; MP3 830, 64 m; MP3 604.1, 62 m; MP3 821 (olim 822), 69.5 m; MP3 991, 79 m; MP3 632, 98 m; MP3 879, 103 m; MP3 962, 144.5 m; MP3 998, 56 m; [31] MP3 699, 50.5 m; and MP3 650, 170.5 m.

To summarize: in my unscientific sample there is a clustering of texts at the approximate lengths of 60 m and 100 m, with outliers as low as 32.5 m (one) and 40 m (two) and as high as 133 m (two), and 144.5 m, 160 m, and 170 m (one each). Under the suggested circumstances of almost experimental writing (i.e. with the alphabet in its infancy and, presumably, scarce training for the amanuensis), the script is unlikely to have been clean, regular, and compact. This would commend a length towards the higher end of the distribution. It is interesting to consider the number of animals it would take to come up with an equivalent writing surface. Since the production of parchment and leather for writing is a rare occupation in the Western world, it is hard to come by the relevant statistical data. I will therefore be using Tanasi et al. 1991, who studied the properties of sheep and lamb skins. Each skin was divided into three or four pieces 21 cm long (along the spine), with a width of approximately 48 cm. Although they do not state this explicitly, I assume that a total length of three (63 cm) is more common than four (84 cm), since they have drawn their figures with three. At any rate, the addition of a fourth would scale down the following numbers by a factor of 3/4. Johnson 2004 gives us the column length of the papyri above, to which the upper and lower margins are added (with a question mark when they are not known precisely, but estimated with a moderately high probability). All these numbers are estimated grosso modo. There are many uncertainties that could affect the outcome. It is, for example, highly unlikely that all the surface of an animal skin (whether as parchment, which is unthinkable without a well developed technique, or simply as leather) would be usable for writing. The complaints of medieval scribes even about the quality of carefully manufactured parchment are common. [32] Areas often had to be skipped because of various imperfections that made writing on them difficult or impossible. There is also the possibility that cows, not goats or lambs, might be used instead. I do not know how the surface area of an ox hide compares to that of a goat or a lamb. But it is safe to surmise that, even as the number of necessary animal skins would be proportionally diminished if oxen were used, so also the expense involved would be significantly magnified, since a cow is likely much more valuable than a goat or a lamb. [33] Whatever animal one assumes, the cost involved would be great. To illustrate this, I offer below a list of papyri followed by the number of goats necessary for the corresponding surface, assuming that the skin in its entirety is adequate for writing. The numbers in parentheses estimate the additional animal skins required if one allows for the upper and lower margins provided by the papyri. These totals should be scaled down by 3/4 for a lower boundary if one assumes that all the animal skins used are of the longer size (i.e. 84 cm long). Thus, for example, P.Oxy. 0020 takes 78 goats (+ 45 with the margins); P.Oxy. 0223, 80 (+ 35); P.Oxy. 3663, 60 (+ 26); P.Oxy. 0445, 32 (+ 8); and MP3 998, 39 (+ 9). P.Oxy. 3663 provides good reference numbers, since its length, about 100 m, is not a high or low outlier in the distribution reviewed above. One would need 86 goats of average size just for the Iliad (with margins, not including Book 10), even if one assumes the complete absence of skin defects that would prevent writing. It should be obvious that this entails a significant sacrifice of personal wealth for the hypothetical recording project.

A major, if not insurmountable, obstacle to dictation is that it is hard to devise a plausible scenario for the transmission of the textual artifact to one of the sites (Athens, Khios, etc.) that arguably played a role in the dissemination of Homeric epic. Rare is the advocate of dictation who confronts this difficulty; and, invariably, the treatment is vague and rather brief. Janko (1998a:13) merely speculates that “[t]he written transcripts were preserved on Chios among the Homeridae … .” Why the cession of the precious original should have happened (or how a copy for Khios would have come about) is not addressed. One might take the position that the original transcript remained marginal and did not affect Homeric performance anywhere in the Greek world until ‘somehow’ it made its way to a center of diffusion (the Homeridai for Janko, Peisistratean Athens for others). [34] This would contradict the notion that the performance of the original poems of Homer (and not of traditions of heroic poetry generally) was widespread during the archaic period and responsible: for the geographically scattered vases that appear to depict Homeric epic; [35] for graffiti like the inscription on ‘Nestor’s Cup’ from Ischia; or for the influence of Homer’s language on various archaic lyric poets. [36] Since the paradosis was decisively channeled through Athens, we would have to assume that Panathenaic performance was controlled by the original or a faithful copy of it. But this would seem to conflict with the myth of the Peisistratean recension, which assumes a collection of once scattered songs. [37] It would also be more than a little curious that a copy of such venerable authority would not have been explicitly mentioned somewhere in the surviving record (literary, documentary, and scholiastic). [38] We might also have expected a report of the Ptolemies’ (real or apocryphal) interest in such a manuscript, like Galen’s story of their seizure by deceit of the official Athenian texts of the canonical tragedians for their library. [39] If the Athenian state had possessed an official, authoritative manuscript of the Iliad of venerable antiquity to regulate Panathenaic performances, we would expect Athenian bookmakers to popularize its readings, and educated residents (Isokrates, Plato, Aristotle, etc.) both to give great deference to its text and to own personal copies of it. In such circumstances, the view that the many quotations that do not match the vulgate must be blamed on a failing memory or on indifference towards verbal accuracy grows implausible. But perhaps such quotations represent the Athenian text? Whence the vulgate, then? [40]

If, on the other hand, one believes in the extensive diffusion of an original transcript, this assumes readily available copies of it throughout the archaic Greek world. And it is not enough that private men embrace it. Since no significantly divergent, alternative text survives, the performers too must have adopted it. This conjures up the image of Greek rhapsodes, far and wide, on pilgrimage to the site of Homer’s original transcript, pen and paper in hand, with the intention and skill to read and copy for themselves the venerable text. Or else we must envision an archaic publishing house at the site of the manuscript, selling copies to rhapsodes willing to give up their life-long training and practice of recomposition in performance, and eager to memorize and recite as derivative performers only the rival version. [41] In other words, we would have to multiply manifold the first miracle of literacy that produced the original manuscript. West is typical of other scholars in his failure to explore how his dictation model can account for the textual diffusion he alleges. A quaint autobiographical anecdote reveals his conviction that ‘Homer’ [42] composed the Iliad at Ilios. [43] He offers no motivation for the production of what “must have existed initially as a bulky collection of leather or papyrus rolls.” Apparently, he believes that this miracle of early literacy was done “only for intermittent reference” [44] and that the collection of rolls stayed in the poet’s possession. He does not tell us why, when, where, and how other oral bards would have adopted this performance script against their regular training and practice; only—with a convenient passive voice—that “occasionally would the written text have been copied.” [45] Powell is more forthcoming. Assuming a Euboian audience, [46] he states that, at first, only his alphabetic adapter could read the poems. He goes on to affirm, without motivating the development, that (possibly partial) copies of the poems circulated among the Euboians (Powell 1991:232). How they are supposed to have learned to read, or else, if they could not, why they should have wished to own such copies, he does not say. Even Homer may have possessed his own copy. Had he, then, learned to read? Did he, too, give up his oral technique in favor of rote memorization and recitation? He further pronounces the Homeridai his descendants, and the eventual owners of the first manuscript or a copy of it. [47] At last, Hipparkhos would have obtained from them the Athenian copy that, as noted above, no extant literary or documentary source explicitly mentions. [48]


[ back ] 1. Cf. Ruijgh 1995:21–26, received favorably by Janko 1998a:1.

[ back ] 2. Cf., for example, Ruijgh 1995:30 and 34.

[ back ] 3. Marek (1993a:30–31) has convincingly answered Naveh’s arguments for a ca. 1100 BC date (see Marek 1993a:30n18 for bibliography). Various recent corpora of geometric inscriptions strengthen Carpenter’s point. The oldest inscribed find at the site of the temple of Apollo Daphnephoros in Eretria (Pfyffer et al. 2005:75 no. 64; cf. 2005:52) consists of three letters on a shard from a middle-geometric amphora (i.e. no earlier than the first half of the eighth century; cf. Verdan et al. 2008:135). In Kommos, the earliest Greek letter is a doubtful qoppa or phi of the late eighth or early seventh century (Csapo et al. 2000:111 no. 3). In Pithekoussai, the oldest inscriptions are from ca. 740 BC (Bartoněk and Buchner 1995:165–166 no. 23 and 171–172 no. 31; add to these several LG I letters on amphoras in Bartoněk and Buchner 1995:174–176). In Lefkandi, a sherd with two letters that seems to date back to ca. 775–750 has been brought to light (Bartoněk and Buchner 1995:195 no. B.8 with Powell 1991:14–15 and n. 34). Naxos contributes an eighth-century inscribed kratēr sherd (Jeffery 1990:466 no. A). Whether the derivation of the alphabet is to be sought in the northern Levant, as Marek 1993a thinks, or in Cyprus, as Woodard 1997a claims, is too complex a matter for me to enter into here (for the bilingual Assyrian-Aramaic inscription of Tell Fakhariyeh or Fekherye, central to Marek, see Woodard 1997a:243n98). But the agreement of both regarding the mid-ninth- to mid-eighth-century derivation of the alphabet should be noted (cf., for example, Marek 1993a:44 and Woodard 1997a:236). The terminus ante quem of 750 must now be revised in light of the graffito from Osteria dell’Osa, securely dated to ca. 770 BC (cf. Watkins 1995a:36–39 and Bartoněk and Buchner 1995:204–205; Sherratt 2003:228n6 advises a cautious assessment). For a recent review of literacy in the archaic period, see Wilson 2009. For a survey more narrowly focused on alphabetic scripts, see Lemaire 2008. Lazzarini 1999 reconsiders the origin of the Greek alphabet with the most important recent scholarship in view.

[ back ] 4. But contrast Palaima (2005:37), who takes issue with a statement by Snodgrass about the “continuous existence of this ancient [Cypro-Minoan] writing system”: “[I] wonder how far we can speak of a ‘continuous existence of [a] writing system,’ when the Cypriote Syllabic script is so different from the Cypro-Minoan in sign repertory and applications.”

[ back ] 5. Cf. Bazemore 2002. See also Sherratt 2003:227–228, although she probably exaggerates the extent of late Bronze-Age literacy on Cyprus and degree to which writing continued in use during the gap that does not archaeologically attest it.

[ back ] 6. Cf. Woodard 1997b:46–51, esp. 47.

[ back ] 7. Shelmerdine (1997:581–582) dates the destruction of the palaces to the transitional LH ⅢB/ⅢC. For Pylos see the detailed analysis of Mountjoy 1997. Especially relevant is Shelmerdine’s observation that, despite elements of continuity at certain sites, “the significant change from LH ⅢB to LH ⅢC is the demise of palatial administration. … [T]he end of Mycenaean bureaucracy meant the end of literacy …” (1997:582). Cf. Driessen 2008 and Morris 1986:121.

[ back ] 8. The focus of Karageorghis 2002:71–113 is the Late Cypriot (=LC) ⅲa, for which he gives the span ca. 1200–1100 BC. Another wave of destruction happened at the transition between LC ⅲa and LC ⅲb, ca. 1100, on which see Karageorghis 2002:115–117 and Catlin 1994. For the three waves of destruction at Enkomi, Mountjoy and Gowland (2005:165) give tentative dates of 1130, 1110, and sometime after 1070 BC.

[ back ] 9. In his unpublished paper, “When did the Greeks first come to Cyprus?,” presented at the Cypriot Archaeology Day held by the Royal Ontario Museum on 6 December 2009, Dimitri Nakassis casts doubt on the Aegean identification of the settlers of Pyla-Kokkinokremos. Further fieldwork and analysis will be required to bring clarity to this question. For updates see

[ back ] 10. It is always possible that Linear B was in fact used in Cyprus on arrival and yet has left no trace in the record. But unless the migrant Greeks soon abandoned it in favor of the Cypriot syllabary—a development that seems implausible—should we not expect sooner or later to discover at least a scrap of Linear B painted or incised on a vase?

[ back ] 11. Cf. Palaima 1991: “We can imagine that one of the Aegean Greek settlers who … shifted from the nearby settlement of Maa-Palaeokastro to Palaepaphos-Skales when Mycenaean Ⅲ C:1b pottery was in use—a Greek settler whose speech already had developed away from the standard Mycenaean South Greek of the 13th century B.C. toward the characteristic historical Arcado-Cypriote dialect, and who, we must stress, had no need to be familiar with the highly restricted Linear B script which had suddenly vanished 150 years earlier with the destruction of the palaces on the Mycenaean mainland …” (454).

[ back ] 12. Gasparri 2001 calls it a “burdensome material” that called for a “long, complex and minute” preparation. Cf. Ruck 1991.

[ back ] 13. Diodoros of Sicily 2.32.4.

[ back ] 14. Cf. Heubeck 1979:155 and Lewis 1974:84. For the use of leather in the Levant see Driver 1976:81–83.

[ back ] 15. Cf. Jeffery 1990:57–58.

[ back ] 16. His further comment that “it seems likely that papyrus was in use in Greece by the time of the composition of the Odyssey” (Lewis 1974:87) is explained by this apparent belief in a mid- or late-sixth-century composition of the poem (see 87n4).

[ back ] 17. Powell 2002: “[N]o doubt Homer’s poems were recorded on papyrus in the eighth century BC (hardly the expensive and impractical leather, as sometimes thought) …” (90).

[ back ] 18. Heubeck 1979:155 and Lewis 1974:87.

[ back ] 19. So Pfeiffer 1968:22, who dated the introduction of papyrus to the early eighth or late ninth century. Not much follows from the reference to the ‘rope of papyrus fibre’ ( ) at  390–391; cf. Lewis 1974:8, 26 and Jeffery 1962:556–557.

[ back ] 20. Cf., further, Heubeck 1979:155 and Lewis 1974:7n7.

[ back ] 21. The date suggested by West (1995:218), between 670–640 BC (“with perhaps a preference for the decade 660–650”) does not suffer from the same degree of implausibility, but it is still too early to assume with any confidence the availability of papyrus.

[ back ] 22. See Petrie 1889:24–37 and Schironi 2010:140–141 no. 28.

[ back ] 23. For the terminology used to describe a bookroll, see Johnson 2004:x.

[ back ] 24. The number varies somewhat depending on the inclusion or exclusion of plus-verses, but, proportionally, the outcome is little affected. On such plus-verses, see the bibliography cited below, §3 n. 40.

[ back ] 25. I have rounded to the nearest half-meter. The error in the calculation is less that one meter, i.e. less than 1%. I do not bother to make these numbers precise, since the uncertainties in the manufacturing of this hypothetical, fantastically long bookroll would exceed the computational error.

[ back ] 26. See Schironi 2010, esp. 41–53, for a recent study of rolls with more than one book of hexametric poetry.

[ back ] 27. Johnson 2004:148. Interesting general observations about bookroll length and roll diameter can be found at 148–151. See also Birt 1882:289–307; Skeat 1982; Lameere 1960:127–147, 166–174; and Schironi 2010:46–47.

[ back ] 28. Schironi 2010:94–95 no. 5.

[ back ] 29. I arrive at these rough estimates of length by simple-minded extrapolation. I do not include the Doloneia, often contested as a late interpolation by scholars who adhere to dictation, to make the case as favorable as possible to my opponents. Nevertheless, the reader should note that MP3 855.1 (olim 857) only contains portions of the very book, the Doloneia, that I am not including in the supposed dictated Iliad.

[ back ] 30. MP3 778 and Schironi 2010:146–147 no. 31.

[ back ] 31. Schironi 2010:108–109 no. 12.

[ back ] 32. Gullick 1991:149.

[ back ] 33. Of course, only calves, sheep, and goats are used in the manufacture of parchment.

[ back ] 34. Regarding Powell’s dictation theory, Woodard (1997a:253) remarks that “there would obviously be no one able to read the epics once they had been penned by Powell’s adapter in his novel script.” Powell (1991:232) agrees: “At first, only he could read them.” Woodard’s remark is valid even if the script was not designed specifically for Homer (or by Homer). Since the alphabet was still in its infancy, what would be the reading public for whom the transcriber expended his considerable efforts?

[ back ] 35. So Powell 1991:210–211; Janko 1982:230; and West 1995:207, with different termini post quos.

[ back ] 36. Once in vogue, the proclivity among scholars to allege instances of this influence and to trace them immediately to the text of our Homeric poems has now fallen out of favor. See Powell 1991:208 (and Appendix Ⅱ); cf. Janko 1982:225–228 and Signes Codoñer 2004:301–341.

[ back ] 37. Allen 1924:224–248; Merkelbach 1952; Davison 1955; Sealey 1957:342–349; Jensen 1980:128–158; Böhme 1983; Aloni 1984; S. West in Heubeck et al. 1988:36–40; Catenacci 1993; Boyd 1995; Nagy 1996c:77–80; Frame 2009:318–328; and Jensen 2011:295–327.

[ back ] 38. For the notion that the Panathenaic rule might be one such reference, see below, §

[ back ] 39. Commentaria in Hippocratis Epidemias Ⅲ, 17a.607 Kühn (CMG See below, §11.1.

[ back ] 40. Apart from oral composition, there is no good explanation for the so-called plus-verses. Cf. S. West 1967b; Nagy 1996c:138–147; and Bird 2010, with bibliography. On plus-verses more generally, see Bolling 1925 and Apthorp 1980.

[ back ] 41. I note in passing that divergent local writing conventions (to which Greek epichoric alphabets witness) make even more problematic the thought that visitors often copied into their own alphabet a text written in an unfamiliar script; or the alternative notion that many acquired for their personal use local copies in the unfamiliar script.

[ back ] 42. West does not believe in the authenticity of the name, as we learn in West 1999b.

[ back ] 43. West 1995:217n43.

[ back ] 44. At the poet’s initiative, then?

[ back ] 45. All the immediately preceding quotations are from West 1998b:97.

[ back ] 46. On which see immediately below, §4.1.

[ back ] 47. These, and the next statement, are from 232–233n32. Powell does not suggest the reason why the descendants of Homer should have moved to, and settled in, Khios.

[ back ] 48. Powell directs us to his chapter 4, n. 156 (at 216), which includes a reference to [Plato] Hipparkhos 228b and to the late and unreliable reports of a Peisistratean recension.