Taida ————— Greek, Roman, and Byzantine Studies 62 (2022) 132–147 Article copyright held by the author(s) and made available under the Creative Commons Attribution License CC-BY https://creativecommons.org/licenses/by/4.0/ ευ or εου in the Homeric Hymn to Pan Ichiro Taida HE LATEST EDITED TEXT of the Homeric Hymn to Pan (hereinafter Pan) is included in the 2003 edition of the Homeric Hymns by Martin West (LCL). Let us focus on line 27, in which West used ὑµνέουσιν, but all existing manuscripts as well as editors print ὑµνεῦσιν: ὑµνέουσιν δὲ θεοὺς µάκαρας καὶ µακρὸν Ὄλυµπον (West) ὑµνεῦσιν δὲ θεοὺς µάκαρας καὶ µακρὸν Ὄλυµπον (MSS. and other editors) The purpose of this paper is to determine which form we should consider the correct one, based on a study of the language of Pan, ὑµνεῦσιν or ὑµνέουσιν; in other words, to determine which form the poet of Pan used. West did not comment on his emendation, as might be ex- pected for a Loeb edition, but he wrote in the preface: “The nature of the Loeb series precludes the provision of the fullest philological detail about variant readings or scholars’ conjec- tures. I have nevertheless tried to ensure that the reader is alerted to the significant textual uncertainties. In places I have made minor orthographical changes without signalling the fact.”1 Thus, West apparently considered it a mere orthographic cor- rection. We can only speculate on why he chose ὑµνέουσιν. In Homer and Hesiod the present third-person plural of ­εω verbs usually ends in -έουσι(ν) because it is metrically required.2 But -εῦσι(ν) 1 M. L. West, Homeric Hymns, Homeric Apocrypha, Lives of Homer (Cambridge [Mass.] 2003) ix–x. 2 The έου of -έουσι(ν) is usually heterosyllabic, so e.g. Il. 2.286 ἐκτελέουσιν (− ◡ ◡ − ◡). T ICHIRO TAIDA 133 ————— Greek, Roman, and Byzantine Studies 62 (2022) 132–147 is also a generally accepted form,3 and some -εω verbs have -εῦσι(ν).4 At Pan 27, -έουσιν is not metrically necessary. We can find the reason for West’s emendation at Pan 27 in the preface to his Iliad: “Contractio vocalium εο (vel εō) saepe per scripturam ευ redditur: […] At hoc ex usu recentiorum Ionum novatum est, qui quarto a.C. saeculo ita scribere coeperant, ut titulorum testimonio constat: antea cum sono tum scriptura distabat εο contractum ab ευ dipthongo primitiva. Ex Homero igitur exterminandum est falsum istud ευ, restituendum εο sive εου […].”5 Thus, according to West, before the fourth century B.C. the contracted εο was distinguished from ευ, but thereafter εο was written as ευ. The form ευ should therefore be excluded from the text of Homeric epics and replaced by εο or εου. As West pointed out, there is general agreement among scholars6 3 For the inflection of -εω verbs (-έουσι[ν]/-εῦσι[ν]) see H. W. Smyth, The Sounds and Inflections of the Greek Dialects: Ionic (Oxford 1894) 528. 4 In the following examples, εου is not metrically necessary and ευ has been generally used: Il. 9.384 ἐξοιχνεῦσι, Il. 20.254 νεικεῦσ’, Od. 3.322 οἰχνεῦσιν, Od. 9.120 εἰσοιχνεῦσι, Hes. Theog. 48 ὑµνεῦσι, 89 τελεῦσι, Op. 227 ἀνθεῦσιν. Except in the Theogony, West used ­έουσι(ν) in his editions Hesiod: Theogony (Oxford 1966), Hesiod: Works and Days (Oxford 1978), Homerus: Ilias I (Stuttgart 1998), and Homerus: Odyssea (Berlin 2017). 5 West, Ilias XXII. West made a similar argument in his Works and Days 62 and Odyssea XX, excluding ευ from his texts of these epics. In addition, see D. B. Monro, Homeric Grammar2 (Oxford 1891) 55: “Verbs in -εω rarely contract -εο or -εω, except in the Participle (-ευµενος for -εοµενος). This rule is con- firmed from New Ionic inscriptions […], as well as the MSS of Herodotus.” Monro said further that for ευ in ποιεύµην (Il. 9.495), θηεῦντο (Il. 7.444), ὀχλεῦνται (Il. 21.261), ἐγεγώνευν (Od. 9.47, 17.161), and a few similar forms we should write εο. Also G. P. Shipp, Studies in the Language of Homer2 (Cam- bridge 1972) 158, argued that the combination e + o is especially resistant to contraction. 6 C. D. Buck, for example, argued that the fourth century B.C. was a turning point (The Greek Dialects [Chicago 1955] 33 and 40). P. Chantraine writes: “Chez Homère se lisent concurremment εο- et -ευ-, mais si l’on songe que la notation -ευ- n’apparaît pas dans les inscriptions ioniennes avant le ive siècle on admettra que ευ a souvent été substitué à -εο- par des Ioniens ; là 134 ΕΥ OR ΕΟΥ IN THE HOMERIC HYMN TO PAN ————— Greek, Roman, and Byzantine Studies 62 (2022) 132–147 that ευ began to appear in Ionic instead of εο and εου7 in the fourth century B.C. He considered that Pan had been composed in the fifth century: “Hymn 19 celebrates the Arcadian god Pan, who did not become widely known until the early fifth century. The language of the Hymn and the romantic imaginative de- scription of Pan dancing with the nymphs suit a fifth-century dating.”8 Since in the fifth century εου was widely used, West must have supposed that the poet of Pan originally used ὑµνέουσιν, but a scribe later changed it to ὑµνεῦσιν. ὑµνεῦσιν would then reflect the transmission rather than the composition of the poem. However, there are objections which can be raised against West. We find examples of ευ in Greek inscriptions before the fourth century.9 Passa argued that the spelling ευ for εο/εου was already used in antiquity,10 and it is impossible to eliminate ευ entirely from the text of Homer.11 Moreover, Passa pointed out that in metrically monosyllabic places in Homeric texts, ευ is always used; εου as varia lectio does not appear in this case.12 même où la contraction est métriquement necessaire, il est probable que ευ remplace un εο monosyllabique” (Grammaire homérique I [Paris 1958] 61). 7 For the contraction of εο/εου into ευ see H. W. Smyth, Greek Grammar2, (Harvard 1956) 21: “Ionic (and less often Doric) may contract εο, εου to ευ: σεῦ from σέο of thee, φιλεῦσι from φιλέουσι they love.” 8 West, Homeric Hymns 18. 9 See Buck, Greek Dialects 40: “But it occurs also Rhodes, Cos, Thera, Cyrene, Megara, Delphi, Corcyra, etc., mostly late but an exceptionally early example is Corinth. Θευγέν ς of the sixth century B.C.” Also, the form Θευγένες appears in an Attic inscription of 408–406: L. Threatte, The Grammar of Attic Inscriptions I (Berlin 1980) 415. E. Passa, “L’antichità della grafia ευ per εο, εου nell’epica: a proposito di una recente edizione dell’Iliade,” RivFil 129 (2001) 385–417, at 391, wrote that Eastern Ionic inscriptions provide a large amount of evidence of ευ since the end of the fifth century B.C. 10 Passa, RivFil 129 (2001) 397, argued that ευ began to be used for εο between the seventh and sixth centuries B.C. 11 Passa, RivFil 129 (2001) 386, 387, 416. 12 Passa, RivFil 129 (2001) 411. ICHIRO TAIDA 135 ————— Greek, Roman, and Byzantine Studies 62 (2022) 132–147 Of course, even though there are persuasive reasons against West’s choice of εο and εου, it is not easy to dismiss it entirely given that εο and εου long predated the fourth century. But must one exclude ευ from the text of Pan? A fifth-century date for Pan is hardly certain and West’s reasoning for it is naïve and subjec- tive. If the date is uncertain and unreliable, West’s emendation would also be unreliable. Before turning to the question of the date of Pan, we must draw attention to the manuscripts of the Homeric Hymns. If ὑµνέουσιν is the original form, we must suppose that scribes accidentally or intentionally changed it to ὑµνεῦσιν; and we would expect similar variants of ευ for εου elsewhere in the manuscripts of the Homeric Hymns. εου, εο, and ευ in the Homeric Hymns To determine whether the scribes often changed εου into ευ, let us look at the manuscripts of the hymns.13 Other than Pan 27, there are only two cases in which both εου (contracted by synizesis) and ευ are metrically possible.14 The two concern ὑµνεῦσιν, as in Pan 27: Ap. 190 ὑµνεῦσίν ῥα θεῶν δῶρ’ ἄµβροτα ἠδ’ ἀνθρώπων Hymn.Hom. 27.19 ὑµνεῦσιν Λητὼ καλλίσφυρον ὡς τέκε παῖδας In these two lines, as in Pan 27, all manuscripts give ὑµνεῦσιν. But three cases are not enough to establish scribal habits. It is evident that scribes knew εου as well as ευ. We can find 13 Existing manuscripts of the Homeric Hymns are divided into five groups: M, f, a, b, and p. I have referred to the manuscripts as follows: M, D ( f ), E (a), Τ (a), L (b), Π (b), P (p), and V (p). For variation in the use of εο/ευ in Homer, one can refer to Chantraine, Grammaire homérique 61–63. We are also aware of similar variations in the manuscripts of the Homeric Hymns: e.g., Ap. 394 Μ, Θ ἀγ(γ)έλλουσι, p ἀγγελλέουσι; Herm. 241 M προκαλούµενος, Ψ προκαλεύµενος; Hymn.Hom. 32.6 Θ χρυσέου, p χρυσοῦ. 14 Passa, RivFil 129 (2001) 413, focused on two similar examples from the Theogony, ὑµνεῦσαι and ὑµνεύσαις, and argued that they are not replaced with ὑµνέουσαι and ὑµνέουσαις in accordance with the basic metric rule. 136 ΕΥ OR ΕΟΥ IN THE HOMERIC HYMN TO PAN ————— Greek, Roman, and Byzantine Studies 62 (2022) 132–147 many examples of εου in the manuscripts, e.g. Ap. 2 τροµέουσιν, 234 κροτέουσιν,15 among others.16 Scribes did not change εου into ευ in these examples. They were aware of both -εου(σιν) and -ευ(σιν). Moreover, if we consider εο/ευ, we find that, although εο is also metrically and grammatically possible, in the following places all manuscripts have ευ: Dem. 181 ἡγεῦνθ’ (ἡγέονθ’ is not attested) and Ap. 196 ὀρχεῦντ’ (ὀρχέoντ’ is not attested).17 On the other hand, although ευ is also metrically and grammatically possible, in the following loci all manuscripts have εο: Dem. 218, Αp. 282 ἵκεο ( ἵκευ is not attested), and Ap. 135 θάµβεον (θάµβευν is not attested).18 It is clear that ευ and εο coexist in the manuscripts. This survey shows that εου, εο, and ευ all survive in the manuscripts. All three are widespread in the hymns and no one spelling has been standardized throughout the collection. There is therefore no evidence that εου/εο were changed by scribal habit into ευ, as West supposed. The reason manuscripts have ὑµνεῦσιν must be that scribes found this form in their sources and copied it faithfully. Moreover, because West rejected ευ at Pan 27 and used ὑµνέουσιν, he was also forced to assume synizesis. However, the synizesis of εου is very rare. In the epics of Homer and Hesiod and the Homeric Hymns we find only a 15 Manuscript M has κρατέουσιν. 16 See also Ap. 236 κοµέουσι, 373 καλέουσιν, 426 κρατέουσιν, Herm. 56 κερτοµέουσιν, Aphr. 211 φορέουσι, 246 στυγέουσι, Hymn.Hom. 33.9 καλέουσι. 17 See also Ap. 268, Aphr. 293, etc. σεῦ (not σέο) and Ap. 311, 337 µευ (not µεο). West, Homeric Hymns, almost always rejected ευ in favor of εο. Ex- ceptionally, West used σεῦ at Hymn.Hom. 30.6. 18 See also Dem. 360, Hymn.Hom. 24.4 ἔρχεο; Dem. 467, Herm. 378 πείθεο; Ap. 261 βάλλεο; Ap. 325a φράζεο; Ap. 399, Herm. 548, etc. (ἔ)πλεον; Herm. 231, Pan 21, etc. οὔρεος; Herm. 381 αἰδέοµαι; Herm. 491, Hymn.Hom. 33.4 ὄρεος; Aphr. 177 ὄρσεο; Hymn.Hom. 30.5, 31.18, etc. σέο. ICHIRO TAIDA 137 ————— Greek, Roman, and Byzantine Studies 62 (2022) 132–147 handful of instances19 of synizesis of εου:20 Il. 2.268 σκήπτρου ὕπο χρυσέου· ὃ δ’ ἄρ’ ἕζετο τάρβησέν τε; Ap. 9 πασσάλου ἐκ χρυσέου· τὸν δ’ εἰς θρόνον εἷσεν ἄγουσα; Dem. 325 αὖτις ἔπειτα πατὴρ µάκαρας θεοὺς αἰὲν ἐόντας. Synizesis of εου is very rare in the later epics too. There are no examples in Apollonius Rhodius, the Orphica (A., H., and L.), or Quintus Smyrnaeus. Therefore, it is hard to justify the assumption of synizesis in West’s emendation at Pan 27. Date of Pan’s composition West substituted ὑµνέουσιν for ὑµνεῦσιν and read it in synizesis. But no manuscript supports this form and the synizesis of εου is too rare to be plausible. The only support left for ὑµνέουσιν is the belief that Pan was composed in the fifth century when εου was widespread. But there is little agreement concerning the date of Pan. Like West, Janko argued that the hymn was composed at a point between the end of the sixth century and the middle of the fifth century.21 Villarrubia supported a somewhat earlier date, at a time not far from the old epic tradition.22 Citing Janko, Fröhder seemed to prefer the earlier date but concluded that it is im- possible to date the hymn exactly.23 Càssola argued that if the 19 As we saw above (n.4), West changed ευ to εου and he had to read εου in synizesis in some cases, including Pan 27. See also Od. 6.157 εἰσοιχνεῦσαν (West, Odyssea: εἰσοιχνέουσαν). 20 Synizesis is also found where εου of χρυσέου unites with the initial vowel of a following word: Ap. 185 χρυσέου ὑπὸ πλήκτρου καναχὴν ἔχει ἱµερόεσσαν; Hymn.Hom. 32.6 χρυσέου ἀπὸ στεφάνου, ἀκτῖνες δ’ ἐνδιάονται. 21 R. Janko, Homer, Hesiod and the Hymns: Diachronic Development in Epic Diction (Cambridge 1982) 185. 22 A. Villarrubia, “Una lectura del Himno Homérico a Pan,” Habis 28 (1997) 7–13, at 13: “Maś bien, habría que considerar este poema un eslabón maś de la cadena literaria que condujo al florecimiento de una nueva época que asumió sin grandes rupturas toda la tradición anterior.” 23 D. Fröhder, Die dichterische Form der Homerischen Hymnen, untersucht am Typus der mittelgrossen Preislieder (Hildesheim 1994) 304–305: “Entstehungszeit 138 ΕΥ OR ΕΟΥ IN THE HOMERIC HYMN TO PAN ————— Greek, Roman, and Byzantine Studies 62 (2022) 132–147 hymn was performed for the first time in Arcadia, where the cult of Pan had been established earlier, the date of Pan’s compo- sition could be prior to the fifth century.24 However, we should remember that the place where Pan was created is uncertain.25 Andrisano held that the poet of Pan was influenced by Sophocles and Euripides, and dated the hymn from the end of the fifth century to the fourth century, or even during the Alexandrian age.26 Thomas wrote that it is yet to be established whether Pan was composed in the fifth century: “Pan’s cult began to spread from Arcadia c.500 BC. This is a secure terminus post quem for the hymn, since we hear nothing about Arcadian rhapsodic traditions before this date, and since the nymphs so prominent here seem to have been unusually unimportant in Arcadia. […] It is equally hard to identify a terminus ante quem.”27 Therefore, it is almost certain that Pan was composed after the fifth century, but its date cannot be specified with greater precision. und -ort des Preislieds lassen sich somit trotz seiner Länge nicht genau bestimmen.” 24 F. Càssola, Inni Omerici6 (Milan 1997) 364–365. 25 Although studies have been made on the place of Pan’s creation, it is still controversial. A. Baumeister, Hymni Homerici (Leipzig 1860) 350–351, suspected that Pan was recited at the Panathenaic Games. Janko, Homer, Hesiod and the Hymns 185: “the use of n-mobile might suggest an Attic poet, but inconclusively.” See also O. Thomas, “The Homeric Hymn to Pan,” in A. Faulkner (ed.), The Homeric Hymns: Interpretative Essays (Oxford 2011) 151–172, at 172: “I have mentioned some regions which are less likely (Boeotia, Attica, Arcadia) and some such as the Argolid which are marginally more likely, given our limited information.” 26 A. Andrisano, “La datazione di [Hom.] Hymn. Pan,” Museum Criticum 13/14 (1978/9) 7–22, at 22. 27 Thomas, in Interpretative Essays 169. See also T. W. Allen and E. E. Sikes, The Homeric Hymns (London 1904) 262: “It is to be regretted that so interesting a poem cannot be dated with any certainty. On one point, however, scholars are substantially agreed—that the hymn is one of the latest in the collection, and that it could hardly have been composed before the age of Pindar at the earliest.” ICHIRO TAIDA 139 ————— Greek, Roman, and Byzantine Studies 62 (2022) 132–147 The language of Pan is very different from the diction of the Homeric epics and those hymns that are usually considered older than the fifth century.28 First, comparing Pan to the longer Homeric Hymns we find that non-Homeric words are more frequently used in Pan:29 one in every 1.89 lines in Pan one in every 5.25 lines in Dem. one in every 7 lines in Ap. (D.Ap. and P.Ap.) one in every 3.2 lines in Herm. one in every 12.75 lines in Aphr. one in every 4.21 lines in Dion. There are thus many non-Homeric words in Pan (26 words). Further evidence suggests that the language of Pan is morpho- logically different from Homeric language. The following words are worthy of attention: Pan Homer line 32 ὢν ἐὼν line 40 χέρα χεῖρα While in Homer the present participle of εἰµί is ἐών (ἐoῦσα, ἐόν), the poet of Pan used the Atticism ὢν at line 32.30 While χεῖρα is used in Homer,31 the poet of Pan used χέρα at 40.32 In Homer, 28 In this paper I cite statistics compiled by previous scholars for the language and meter of the Homeric epics and the four longer Homeric Hymns. Following their methods and using the text of West, Homeric Hymns, I de- termined the corresponding figures for Pan and Dion. 29 Figures for the four longer hymns are drawn from R. Thiele, Prolegomena ad Hymnum in Venerem Homericum quartum (Halle 1872) 43. 30 Cf. Ap. 330 οὖσα and Hymn.Hom. 29.10 ὢν. See P. Chantraine, Morphologie historique du grec2 (Paris 1961) 281, who called ὤν, οὖσα, ὄν “l’attique emploie.” Concerning οὖσα at Ap. 330, Janko, Homer, Hesiod and the Hymns 117, acknowledged that it is an Atticism, but wrote: “It seems to be found in poetry of varied origin, and thus, whether in origin an Atticism, archaism, or replacement for some other form, such as εὖσα (Cos, Rhodes, Hdt.) or ἔσσα (Lesbos, Argolid, Laconia), this occurrence is rather uninformative.” 31 E.g. Il. 1.219 and 2.389. 32 For χέρ- see Soph. Aj. 97, Eur. Alc. 35, Theoc. 11.55, etc. 140 ΕΥ OR ΕΟΥ IN THE HOMERIC HYMN TO PAN ————— Greek, Roman, and Byzantine Studies 62 (2022) 132–147 χεῖρ- is always metrically long.33 For the morphological features, Janko’s study is noteworthy.34 Among his statistics, we focus on the neglect of digamma: Pan 83.3%, Il. 17.2%, Od. 17.9%, Dem. 45.9%, DAp. 34.6%, PAp. 22.4%, Herm. 53.6%, Aphr. 15.9%, and Dion. 28.6%.35 These 33 The following words are also worth noting. At Pan 2, the MSS. have πίση, while πίσεα is used in Homer (Il. 20.9, Od. 6.124). The contraction of -εα to -η is rare in Homer; cf. Smyth, Greek Grammar2 21. G. Hermann, Homeri Hymni et Epigrammata (Leipzig 1806) 168, and A. Gemoll, Die Homerischen Hymnen (Leipzig 1886) 335, held that πίση indicated that Pan was created in a later period. F. A. Wolf, Homeri Odyssea cum Batrachomyomachia, hymnis, ceterisque poematibus (Halle 1784) 609, used πίσεα. For examples of -εα at the end of a line see Chantraine, Grammaire homérique 59: e.g. θεοειδέα (Il. 3.27, 3.450, etc.), Πολυδεύκεα (Il. 3.237, Od. 11.300), and Διοµήδεα (Il. 4.365, 5.881). Janko, Homer, Hesiod and the Hymns 144, wrote that the contracted forms (-η) are common in the late Homeric Hymns (cf. Hymn.Hom. 28.15 τεύχη). See also Càssola, Inni Omerici6 LIV–LV, who commented on ­εα/-η in the Hymns. At Pan 3, Νύµφαις is used, whereas in Homer the plural dative is Νύµφῃσι (Od. 13.350, 14.435), or Νύµφῃς could have been possible. For -ῃς see Chantraine, Grammaire homérique 199. The feminine dative plural ending -αις is rare in Homer, but it is more commonly used later. Cf. Monro, Homeric Grammar2 86: “The Fem. -αις appears only in the forms θεαῖς (Od. 5.119), ἀκταῖς (Il. 12.284), and πάσαις (Od. 22.471). Hence, it is a question whether the forms in -οις, -αις are Homeric.” Also, ῥείθροισιν is used in Pan 9; however, the Homeric form is ῥέεθρ- (Il. 2.461, 21.244, etc.). ῥείθρ- is the Attic contraction: cf. Smyth 20. This seems to justify ascribing a later date to Pan. But one might object that these words reflect the transmission rather than the composition of the poem. For such an argument see Càssola LIV–LV. 34 Janko, Homer, Hesiod and the Hymns 184–185, argued that Pan dates from the end of the sixth century to the middle of the fifth century B.C. and offered evidence suggesting that Pan was composed at roughly the same time as other earlier epics. The use of n-mobile before consonants diminished in later poetry, but there are five such cases in 49 lines of Pan. The percentage of lines containing the n-mobile before a consonant is higher in Pan than in other earlier epics: Pan 10.2%, Il 3.56%, Od 3.72%, Dem 4.44%, DAp 6.08%, PAp 1.37%, Herm 1.55%, Aphr 4.78%, and Dion 3.39%. Cf. Janko, Homer, Hesiod and the Hymns 64–66, 185. 35 Janko, Homer, Hesiod and the Hymns 47, 183–184. ICHIRO TAIDA 141 ————— Greek, Roman, and Byzantine Studies 62 (2022) 132–147 numbers indicate that Pan was composed later.36 From a study of the presence or absence of digamma, Allen and Sikes con- cluded that “H. Pan cannot belong to an early epic date.”37 Sometimes the language of Pan is semantically different from that of Homer. The following words in Pan have different mean- ings from those in Homeric epics. At Pan 15, δονάκων means “reed pipes,” as in Pindar (Pyth. 12. 25), Aeschylus (PV 574), and Theocritus (20.29), while in Homer δόναξ means “a reed” or “the shaft of an arrow.” At Pan 16, µελέεσσιν means “songs,” as in Theognis (761) and Pindar (Ol. 9. 1), whereas in Homer the word always means “limbs.” At Pan 23, λαῖφος (… λυγκὸς) means “(a lynx’s) skin”; however, in Homer it means “shabby, tattered garment.”38 At Pan 38, τιθήνη means “mother,” as in Coluthus (372, V/VI A.D.); on the other hand, in Homer, it means a “nurse.” Thus, the forms and meanings of several words in Pan are different from those in the early epics. Metrical analysis Metrical analysis also suggests that the language of Pan is different from the language of archaic epic. As evidence, I have collected the following statistics on meter. As Thomas points out, the forty-nine lines of Pan might seem too small a number to extract reliable statistics.39 But the numbers show differences between Pan and other epics not recognized to date. My analysis 36 Regarding digamma, Janko, Homer, Hesiod and the Hymns 42: “Traces of it persist because of the conservative nature of the traditional diction; however, if the tradition is at all open to innovation, then we expect these traces to become less frequent in less conservative or later poetry, and we are not disappointed.” However, A. Vergados, The Homeric Hymn to Hermes: Intro- duction, Text and Commentary (Berlin 2013) 143, holds that no firm conclusions can be drawn regarding the date and provenance of Herm. from observances and neglects of digamma. 37 Allen and Sikes, Homeric Hymns lxxii. 38 We do not find any examples in which λαῖφος means animal skin. On this word see Vergados, The Homeric Hymn to Hermes 357–358. 39 Thomas, in Interpretative Essays 170. 142 ΕΥ OR ΕΟΥ IN THE HOMERIC HYMN TO PAN ————— Greek, Roman, and Byzantine Studies 62 (2022) 132–147 includes the Homeric Hymn to Dionysus, which has fifty-nine lines and is as short as Pan. It shows that the language of Pan is also different from the language of Dionysus. Moreover, it will become clear that Pan is actually closer to poems of the Hellenistic and later periods. First, we consider the frequency of hiatus and irregular lengthening40 in the following texts:41 Hiatus Lengthening Pan 0% 0% Il. 0.78% 1.36% Od. 0.96% 1.16% Dem. 2.23% 0.2% DAp. 1.1% 1.1% PAp. 2.47% 2.19% Herm. 0.52% 1.55% Aphr. 0.34% 1.02% Dion. 6.8%42 0% There are no examples of metrical irregularities (hiatus or ir- regular lengthening) in Pan; the data indicate that the hymn was 40 The list shows the percentage of lines containing a hiatus or an irregular lengthening. For instance, there are eleven examples of hiatus in Dem., which is made up of 495 lines, and thus 11/495 = 0.0222. 41 Figures are drawn from Janko, Homer, Hesiod and the Hymns 36. We should not regard the following places in Pan as metrical irregularities. Line 8 διὰ ῥωπήϊα: -ὰ is long, because an initial ῥ- always appears to have the power of lengthening a preceding vowel; cf. Monro, Homeric Grammar2 344. Line 44 πὰρ δὲ Ζηνὶ: -ὲ is long, because Ζ is a double consonant; cf. Smyth, Greek Grammar2 11. Line 48 χαῖρε ἄναξ: it is common for the hiatus to occur after a short vowel at the feminine caesura, as here; cf. M. L. West, “Homer’s Meter,” in I. Morris et al. (eds.), A New Companion to Homer (Leiden 1997) 218–237, at 232. Moreover, we must remember that ἄναξ used to begin with the digamma, and the effect did not disappear in later poetry: e.g. Theoc. 17.135, Callim. Ap. 113 χαῖρε ἄναξ, and Apol. Rhod. 1.908 τοῖο ἄνακτος. For the same reason, we do not admit the hiatus between δέ and οἱ at Pan 31. For the digamma of ἄναξ and οἱ, see Monro 364, 370, and Chantraine, Grammaire homérique 126–127, 145. 42 Dion. 6 ἐϋσσέλµου ἀπὸ; 24 ὄρσῃ ἀργαλέους; 30 ἐρεῖ αὐτοῦ; 38 παρὰ ἱστίον. ICHIRO TAIDA 143 ————— Greek, Roman, and Byzantine Studies 62 (2022) 132–147 composed somewhat later than the other works. See Janko’s comment: “Another of Parry’s insights was that some metrical irregularities in Homer may be attributed to the juxtaposition and adaptation of formulae in oral composition, and might well have been avoided by a literary poet.”43 Second, we look at the bucolic caesura (word-ending after uncontracted fourth-foot biceps).44 Twenty-eight lines of Pan (57.1%) have the caesura, and the frequency is higher than that in other poems:45 Homer 47%, Dem. 48%, DAp. 42.1%, PAp. 52.9%, Herm. 51.2%, Aphr. 51.7%, and Dion. 50.8%. Even West noted that “the ‘bucolic’ caesura, is a conspicuous feature of Alexandrian and later versification,” e.g. Callimachus 79% and Apollonius Rh. 63%.46 Third, the following table shows the frequency of spondees in each foot:47 1st 2nd 3rd 4th 5th Pan 30.6% 28.6% 14.3% 20.4% 0% Il. 39.1% 39.8% 14.9% 29% 5.1% Od. 37.9% 42.4% 16.2% 30.2% 4.7% Dem. 31.4% 37.2% 16.6% 28.9% 12.7% DAp. 42.1% 46% 24.7% 29.8% 11.2% PAp. 36% 45% 21.8% 26.7% 12% Herm. 34.3% 33.3% 13.4% 23.4% 6% Aphr. 44.8% 42% 23.8% 25.5% 10.3% Dion. 32.2% 35.6% 8.5% 33.9% 3.4% 43 Janko, Homer, Hesiod and the Hymns 33. 44 S. D. Olson, The Homeric Hymn to Aphrodite and Related Texts: Text, Translation and Commentary (Berlin 2012) 37. 45 The figure for Homer is drawn from M. L. West, Greek Metre (Oxford 1982) 154, and those for the four longer hymns are from Olson, The Homeric Hymn to Aphrodite 37. 46 West, Greek Metre 154. 47 Figures are drawn from Olson, The Homeric Hymn to Aphrodite 36. 144 ΕΥ OR ΕΟΥ IN THE HOMERIC HYMN TO PAN ————— Greek, Roman, and Byzantine Studies 62 (2022) 132–147 The poet of Pan never used a spondee in the fifth foot. We should remember that the epic formulae are commonly used in the latter part of a line.48 The difference between Pan and other epics concerning the fifth foot suggests that the poet of Pan did not completely rely on Homeric formulae and that his style was different from that of other epics. Like Pan, later poems show none or few spondees in the fifth foot: e.g. Nicander 2.55% and Nonnus 0%.49 The fourth foot is also noteworthy. Pan’s percentage (20.4%) is the smallest in this table. Like Pan, Hellenistic poems only occasionally feature a spondee in this position: e.g. Aratus 19.16%, Callimachus 19.33%, Apollonius Rh. 17.28%, and Dionysius Periegetes (II A.D.?) 20.74%.50 Fourth, Pan shows a clear preference for dactyls. Only 18.78% of its feet are spondees. Compare this with Il. 25.58%, Od. 26.28%, Dem. 25.36%, DAp. 30.76%, PAp. 28.3%, Herm. 22.08%, Aphr. 29.28%, and Dion. 22.7%.51 In the evolution of the hexameter, the use of spondees decreased in favour of the dactyl:52 e.g. Callimachus 21.82%, Triphiodorus (III/IV A.D.) 20.72%, Quintus Smyr. (IV A.D.?) 16.96%, Nonnus 14.95%, Pamprepius (V A.D.) 16.2%, and Musaeus (V/VI A.D.) 48 A. Parry, The Making of Homeric Verse: The Collected Papers of Milman Parry (Oxford 1987) 9. 49 A. M. Pascual, “El hexámetro de Dionisio el Periegeta. I,” Myrtia 28 (2013) 39–55, at 51. 50 Pascual, Myrtia 28 (2013) 51. M. Van Raalte, Rhythm and Metre: Towards a Systematic Description of Greek Stichic Verse (Leiden 1986) 37, 40, also pointed out that in Homer, Hesiod, the Homeric Hymns, Parmenides, and Empedocles, more than 20% of verses have a spondee in the fourth foot, while in Hellenistic hexameters (Aratus, Callimachus, Apollonius) a spondee is less frequent in this position (under 18%). 51 I have calculated the percentage in the Iliad as 25.58%, based on the frequency of spondees in each foot (the average of 39.1%, 39.8%, 14.9%, 29%, and 5.1%). 52 Pascual, Myrtia 28 (2013) 53. ICHIRO TAIDA 145 ————— Greek, Roman, and Byzantine Studies 62 (2022) 132–147 15.39%.53 Fifth, the following percentages also suggest that the com- positional style of Pan is different from that of other epics. In this table, ddddd for example means that the first five feet of the line are made up of five dactyls:54 Pan Il. Od. Dem. DAp. PAp. Herm. Aphr. Dion. ddddd 28.6 19.1 18.6 20.6 12.9 16.9 24.4 14.1 22 dsddd 20.4 14.5 14.9 13.4 13.5 14.4 15.2 14.8 16.9 sdddd 14.3 13.9 12 8.5 10.7 12.5 16 16.2 13.6 ddsdd 8.2 4.2 4.2 5.7 6.2 4.1 4.3 3.4 3.4 sddsd 8.2 6 6.5 5.5 6.2 4.6 5.7 6.6 6.8 dddsd 6.1 8.5 8.4 9.3 7.3 5.4 7.2 5.9 10.2 ssddd 6.1 8.1 8.1 6.6 7.3 6 5.7 7.9 6.8 ddssd 4.1 1.3 1.4 0.8 0.6 2.7 1 1.4 3.4 dsdsd 2 6.1 6.8 6.8 4.5 7.4 4.3 3.8 8.5 sdsdd 2 2.7 2.8 2.1 1.7 2.5 1.2 3.4 1.7 There are thirty-two potential arrangements of dactyls and spondees, and the poet of Pan used only ten. By contrast, Homer used every one of them,55 and the verse forms of the longer hymns are also full of variety in this regard: Dem. 27, DAp. 25, PAp. 26, Herm. 27, Aphr. 26, Dion. 13.56 Like the poet of Pan, later 53 Pascual, Myrtia 28 (2013) 53–54. 54 Figures for the Iliad and Odyssey are drawn from J. L. Lightfoot, Dionysius Periegetes (Oxford 2014) 65; for the four longer hymns, from Olson, The Homeric Hymn to Aphrodite 35, 40–42. 55 S. E. Bassett, The Poetry of Homer2 (Maryland 2003) 142; Lightfoot, Dionysius Periegetes 65. 56 Olson, The Homeric Hymn to Aphrodite 35, 40–42. 146 ΕΥ OR ΕΟΥ IN THE HOMERIC HYMN TO PAN ————— Greek, Roman, and Byzantine Studies 62 (2022) 132–147 poets also tend to use fewer forms: e.g. Callimachus 21 and Nicander 20.57 Moreover, as Pascual pointed out, since the Hellenistic period a gradual reduction in the number of schemes is evident into the time of Nonnus and the late hexameters: e.g. Nonnus 9, Pamprepius 9, Musaeus 10, and Coluthus 15.58 In Pan, the total frequency of the five most common verse forms is 85.8%. Compare this with those of Il. 64.1%, Od. 62%, Dem. 58.6%, DAp. 51.7%, PAp. 57.5%, Herm. 68.5%, Aphr. 59.6%, and Dion. 71.2%.59 It is thus evident that Pan does not have a great diversity of metrical forms and, even considering its length, its composer used fewer than one might expect. Con- versely, as in Pan, the frequency of the five most common verse forms in later works is rather high: e.g. Callimachus 74%, Nicander 73.66%, and Nonnus 93.34%.60 The statistics show that Pan is metrically different from earlier epics and close to later epics. Conclusion West adopted ὑµνέουσιν because he believed that Pan was composed in the fifth century, and εου was widely used before the fourth century. He believed that ὑµνεῦσιν in the manu- scripts was a scribal form. But the scribes of the Homeric Hymns did not habitually change εο/εου to ευ. Moreover, ὑµνεῦσιν is metrically preferable as ὑµνέουσιν presupposes an unusual synizesis. Furthermore, a fifth-century date for Pan is unlikely. Its language and meter are very different from those of Homer’s Iliad and Odyssey and from the other Homeric Hymns. Metrical statistics suggest that Pan 57 Lightfoot, Dionysius Periegetes 65. 58 Pascual, Myrtia 28 (2013) 43–46. 59 Lightfoot, Dionysius Periegetes 65; Olson, The Homeric Hymn to Aphrodite 35, 40–42. 60 Pascual, Myrtia 28 (2013) 47. Lightfoot’s statistics (Callimachus Hymn to Delos 79.3% and Nicander 73.69%) are slightly different from Pascual’s (see Dionysius Periegetes 65). ICHIRO TAIDA 147 ————— Greek, Roman, and Byzantine Studies 62 (2022) 132–147 shares features typical of Hellenistic and even later poems. There is a lack of evidence to support the fifth-century date, and one cannot therefore accept West’s emendation. While the date of Pan remains uncertain, it is undeniable that ὑµνεῦσιν is the manuscript form. There is no sound reason to reject it and editors should print ὑµνεῦσιν.61 January, 2022 Toyo University taida@toyo.jp 61 This paper is based on an oral presentation delivered in 2017 at the 11th International Conference of the Taiwan Association of Classical, Medieval and Renaissance Studies at Tainan National University of the Arts, Taiwan. I wish to thank the anonymous reviewers for their helpful suggestions and comments.