6. Doricisms in the New and Old Posidippus [1]

Alexander Sens, Georgetown University

Posidippus AB 139 (8 GP = AP XII 131) is a prayer for Aphrodite to favor a woman named Callistium:

As it survives in the single manuscript witness, the dialect coloring of the poem is inconsistent. In the opening verse the feminine nominative singular of the relative pronoun appears three times in the Doric form ἅ [2] but in the third verse the manuscript transmits ἥ. The text of the rest of the poem is consonant with the latter rather than the former: thus the second verse has epic/Ionic Συρίης rather than Doric (and Attic) Συρίας; and the third, ἐραστήν rather than ἐραστάν. [3] Gow and Page, who make the sensibly conservative decision to print the paradosis, nonetheless note that they find it “difficult to believe” that the poet could have written both ἅ and ἥ in the same epigram, and incline toward replacing the Doric pronouns on the ground that Posidippus “does not favour Doric.” Such uncertainties about dialect are common in our texts of Hellenistic poetry, and the difficulty of resolving them in any satisfactory way is compounded by a number of considerations. The category “Doric” is not monolithic but includes a variety of widely dispersed local dialects, each with its own phonological peculiarities. By the end of the fourth century, moreover, spoken and literary forms of the dialect regularly combined traditional characteristics of Doric phonology and morphology with elements drawn from other dialects, including the Attic koine (especially in the case of the spoken language) [4] and (in the case of poetry, at least) the complex, predominantly Ionic literary language of the early epic hexameter poems. Hellenistic poets, moreover, often overlaid Doric phonology on epic vocabulary and morphology. As a result of all this, the Doric found in Hellenistic poetry is highly variable. In fact, the evidence that was available even before the publication of P.Mil.Vogl. VIII 309 shows that the earliest texts of some Hellenistic poems had both Doric and epic/Ionic forms side by side, [5] and indeed Callimachus in Iambus 13 seems to justify his own use of precisely such a combination of dialects (fr. 203.17–18 Pf. τοῦτ᾿ ἐμπ[έ]πλεκται καὶ λαλευσ|[ . . ] . . [| Ἰαστὶ καὶ Δωριστὶ καὶ τὸ σύμμικ|το̣ν̣[.). [6]

As studies of the textual tradition for the Theocritean corpus show, this artificial and variable language was easily altered in the process of transmission, [7] so that the manuscript evidence for dialect coloring must be considered highly suspect. Modern readers of Hellenistic poems preserved in medieval manuscripts thus cannot have confidence either that the appearance of dialectal regularity in a given poem does not mask a more complex original picture, or that what at first appears to be a chaotic and unlikely jumble of different dialect forms was not what the poet wrote. The new Posidippus papyrus offers a rare opportunity for exploring the use of dialect in an extensive, mostly continuous and complete Hellenistic text produced not long after the poems it contains were composed and assembled. [8] Given that P.Mil.Vogl. VIII 309 (like any manuscript) naturally contains slips and errors of various sorts, [9] it cannot be taken as providing a wholly accurate image of what its author wrote, but its early date and the care with which it was produced make it likely to give a far better picture of the poet’s practices than the medieval manuscript tradition does for the “old” Posidippus.

For most of the epigrams of Posidippus known prior to the publication of the papyrus, the basic dialectal coloring preserved in the witnesses is Ionic with a heavy admixture of features characteristic of the Attic koine. Thus, for example, the endings of first-declension nouns and adjectives consistently have Ionic η rather than Attic (and Doric) long α even after ρ, ι, and ε (e.g. AB 116.1 [12.1 GP]; AB 117.2 [24.2 GP]; AB 118.17 [SH 705.17]); AB 129.2 [3.2 GP]; AB 148.1 [SH 700.1], but (as in much other early Hellenistic poetry) the endings of nouns and verbs often contract in the Attic manner. The Attic-Ionic modal particle ἄν (AB 15.6 [20.6 GP]; AB 116.4–5 [12.4–5 GP]; AB 120.2 [14.2 GP]; AB 122.8 [17.8 GP]; AB 125.4 [2.4 GP]; AB 135.2 [5.2 GP]) predominates over κε(ν) (AB 115.9 [11.9 GP]), the form characteristic of most other dialects and preferred in early epic. Moreover, the obvious influence of the Attic koine suggests that features characteristic of that dialect but alien to Ionic should be understood as Atticisms, even when the form in question would be equally appropriate in other dialects. A case in point is the use of -αις as the ending of the first-declension dative plural at AB 137.1 (6.1 GP), [10] a passage that conforms to the usual Hellenistic practice of placing these forms only at verse end. [11]

The situation is much the same for the poems contained in the new papyrus. In most of these epigrams η rather than α follows ρ, ι, and ε in the endings of the first-declension (e.g. AB 10.3; AB 12.5; AB 16.5; AB 23.4; AB 24.6; AB 28.4; AB 32.6; AB 37.1; AB 38.3; AB 40.3, AB 42.3, 4; AB 45.5; AB 48.3, 4; AB 51.3; AB 53.2; AB 55.5; AB 60.2; AB 63.4; AB 63.6), although AB 55.1 has the Attic form ἑώιαν [12] and at AB 89.3 the genitive Ἀκαδημείαϲ is used (appropriately) of the Athenian philosophical school. The contraction of verbs and nouns is common (e.g. AB 15.5). The genitive of the second declension is almost always -ου; epic -οιο seems not to occur. Dative -αις is possible at AB 58.5 and secure at AB 74.4 and AB 83.2, all at verse end. The modal particle is consistently ἄν; κε(ν) appears nowhere on the papyrus, although it has been introduced by the editors at AB 20.3. Once again, certain forms that are common to Attic and to other dialects, including Doric, should probably be taken as Atticisms, given that some overtly Attic forms occur in poems otherwise of epic/Ionic coloring but identifiably Doric features (other than proper names; see below) do not. Thus, for example, at AB 37.8 ναοπόλο[υ is equally appropriate to the Attic koine (cf. J. AJ 14.73) as it is to Doric (cf. Pindar fr. 51d), and is best taken as a natural feature of the prevailing Attic-Ionic dialectal coloring.

Despite the predominance of Attic-Ionic, however, both the previously known and the new, expanded corpus of Posidippus contain poems in which at least some “Doric” coloring is in evidence. The paradosis of AB 141 (APl _[A] 68), a short ecphrastic poem in which the speaker claims not to be able to tell whether a portrait he is viewing depicts Aphrodite or Berenice, uses forms with Doric alpha (ἅδ(ε), Βερενίκας and ὁμοιοτέραν), but the epigram may be by Asclepiades, to whom it is alternatively ascribed by the lemma, rather than Posidippus. [13] Two other instances in which the witnesses preserve α rather than Attic-Ionic η for inherited /a:/ in the “old” Posidippus involve references to the Muse. At Posidippus AB 123 (1.3–4 GP), the Doric form of the article in the phrase ἅ τε Κλεάνθους | μοῦσα, occurs in close conjunction with the markedly Attic contraction of ε+ο to ου (Κλεάνθουϲ) [14] and with the Attic-Ionic pronoun ἡμῖν, while at 23.2 GP, a poem also ascribed to Meleager, [15] the speaker refers to Μοῦσαν ἐμὰν ἱκέτιν in the context of complaining of his desire for a woman whose name has the apparently Attic form Ἡλιοδώρας (genitive). [16] The function of the Doric coloring in these epigrams is not immediately clear—Cleanthes’ surviving fragments do not show any trace of Doric coloring—but the passages illustrate an important feature of the majority of the Doricisms in the entire corpus of Posidippus. Whereas the use of α as a reflex of inherited /a:/ is a common feature of the entire Doric group, the various local dialects treat the products of contraction and compensatory lengthening of /e/ and /o/ in different ways. In some dialects, which constitute a group Ahrens called “Doris mitior,” secondary long /e/ and /o/ are represented as ει and ου respectively, but in a handful of others (which Ahrens calls “Doris severior”), they converge with the inherited long vowels and are written as η and ω. [17] The vast majority of instances of secondary long /e/ and /o/ in the corpus of Posidippus are written ει and ου, as in the case of Μοῦσα (rather than Μῶσα [e.g. Alcman PMG 8; 14; 28; 30; 31; 46; 59 Page; scholium to Theocritus 1.9, p. 34 Wendel]) in 1.4 GP and 23.2 GP.

Other than these passages, the only instance of Doric coloring in the “old” Posidippus occurs in a poem also found on the new papyrus. AB 65 contains the text of an epigram ascribed to Posidippus at APl^[A] 119 (= 18 GP):

1 θαρϲαλέη Pl : δαιδαλέη Himerius Or. 48.14 3 χέεϲ Pl 4 ϲυγγνώμη Pl

The poem offers important evidence for the loss of Doric forms in the process of transmission. Alongside forms consonant with Doric (including not only μορφᾶϲ but also ὁρῇ, which shows the standard Doric contraction of α and ει to ῃ rather than to ᾳ), [19] the later witnesses have Ionic θαρϲαλέη/δαιδαλέη and ϲυγγνώμη where the papyrus shows evidence of Doric θαρϲαλέα and ϲυγγνώμα instead. Furthermore, whereas Pl has the unaugmented form χέεϲ in the second hexameter, the papyrus seems to contain ἔθευ. The form raises an interesting point. The contraction of ε and ο to ευ was widespread in Ionic, but it was not a universal feature of Doric, although it occurred in some local dialects, including Rhodian and Cretan. In Hellenistic poetry, however, it seems to have been treated as a generalized Doric characteristic, occurring in the Doric poems of Theocritus [20] and Callimachus (cf. Hymn 6.127) as well as in Posidippus 23.5 GP φωνεῦντ(α), an epigram in which (as we have seen) the presence of ἐμάν suggests a Doric coloring. Nonetheless, the contraction was also a feature of Corinthian, which is closely related to the dialect spoken at Sicyon, and, in an epigram in which Lysippus’ Sicyonian background is explicitly asserted, it may therefore be thought specially appropriate. [21]

In addition to this epigram, a number of other poems on the papyrus show aspects of Doric coloring. As is to be expected, the most obvious and consistent marker of Doric in the epigrams is the presence of alpha rather than eta as the reflex of inherited /a:/. Because this feature was the most recognizable feature of all forms of Doric, literary Doric from an early period sometimes generalized it even to cases where Attic-Ionic eta represents not /a:/ but /e:/, for which Doric properly uses η. Although it is often difficult to tell whether analogically generated forms of this sort are original or the product of transmission, papyrological evidence shows that many such forms were represented in the tradition at a very early stage and may thus be original. The phenomenon, however, seems largely foreign to the new Posidippus. The only plausible example occurs at AB 87.3, where the original copyist seems to have written the alpha of [πο]λυθρύλατον over the eta he produced initially. Adjectives in -τος from verbs in -έω properly terminate in -ητος in Doric (e.g., Pindar Olympian 1.21 ἀκίνητον; 7.25 ἀναρίθμητοι; Theocritus 14.6 ἀνυπόδητος, 48 ἀρίθμητοι), although hyperdoric -ατος is occasionally attested for adjectives of this sort in the manuscript tradition of both choral lyric (e.g. Pindar Isthmian 5.6 ὠκυδίνατος) and Hellenistic bucolic (e.g. Bion fr. 2.15; Adonis 58 τριπόθατος) [22] , and some at least are probably authentic. [23] Posidippus seems to have used a “proper” Doric form of a comparable adjective at AB 64.4, where the editors’ [ἀδό]νητοϲ is highly likely (despite their unnecessary puzzlement about why Posidippus has not written -ατοϲ; cf. above). Although it is impossible to exclude the possibility that Posidippus wrote πολυθρύλατον at AB 87.3, the most economical explanation is that the form is a hyperdoric miscorrection by the copyist after he wrote the proper form πολυθρύλητον. If this is the case, then despite the small size of the sample, it seems reasonable to conclude that Posidippus understood and was attentive to the difference between Doric α = /a:/ and η = /e:/.

Besides those cases in which Doric alpha is used in place of Attic-Ionic eta, the new papyrus contains only a few traces of other distinctly Doric features. The first-person pronoun ἁμές and the first-person plural verbal ending -μες occur several times each, but only in the ἱππικά (AB 87.1 and AB 88.1 ἁμέϲ; [cf. AB 75.1]; AB 88.2 νικῶμεϲ; AB 87.2 ἀγάγομ[ε]ϲ). [24] With two notable exceptions, the secondary long /e/ and /o/ sounds are represented as ει and ου, as in Attic-Ionic and the “mild” Doric dialects, even in forms that show other features of Doric. [25] At AB 64.2, however, the aorist ἠργάϲατο appears in an ecphrastic epigram on a statue of the Cretan hero Idomeneus by the Cretan sculptor Kresilas. Although the augmented first syllable of this verb is at least sometimes treated as ἠ- rather than εἰ- in Attic and the koine, [26] the form—if it is in fact what Posidippus wrote [27] — would also be appropriate to the overall Doric coloring of the epigram, which seems to have γαρύει, Μηριόνα, and δάν. [28] Other than in this poem, the only instances in which η and ω represent secondary long /e/ and /o/ occur in an epitaph in which Menoitios, the dead man who is the purported speaker, explicitly claims to be from Crete. The augmented aorist ἠάϲατε (= Attic-Ionic εἰάσατε) at the end of the first verse shows the “severe” Doric treatment of the first syllable and is in fact attested in an inscribed epitaph of Doric coloring from the first century CE (GVI 969.2). [29] Similarly, Φιλάρχω is the only example in the papyrus of -ω instead of ου, and—despite the tendency of proper names to retain special dialectal characteristics—must be considered a particularly marked Doricism.

The restriction of η for secondary long /e/ and v for secondary long /o/ to contexts involving Cretans is highly suggestive. Bastianini and Gallazzi (21–2; 189) plausibly suggest that the use of Doric in these poems may be generally mimetic of the Doric color of Cretan speech, but the point may be pushed even farther. Although the use of -ω as the genitive singular of second-declension nouns came to be treated as a generalized feature of literary Doric, it occurs regularly only in Cretan, Heraclean, and Laconian. [30] Cretan also ranks among a handful of Doric dialects in which η rather than ει regularly represents secondary long /e/. The presence of “severe” Doric phonology independently and exclusively in poems involving Cretan subject matter may suggest an attempt to capture the particular, regional characteristics of the language spoken on the island. [31] That both poems contain direct speech placed in the mouth of Cretans supports this view, [32] even if in the Cresilas poem ἠργάϲατο forms part of the narrative rather than the speech ascribed to Idomeneus’ statue.

It is important to note, however, that even if the severe Doricisms in these two epigrams are intended to suggest actual features of Cretan speech, any mimetic effect they produce remains somewhat stylized inasmuch as the poet makes no attempt to capture all the peculiarities of Cretan speech. Moreover, the poems also contain forms apparently alien not only to Cretan but to Doric as a whole. Thus the first verse of the Idomeneus-statue poem has, in addition to the appropriately epic form of the adjective χάλκειον, the Attic demonstrative ἐκεῖνον, a word found also in an otherwise Doric context at AB 68.5, rather than κῆνοϲ, the form used in Crete and elsewhere, or τῆνοϲ. [33] Together with severe Doric Φιλάρχω and ἠάϲατε, the Menoetius-epigram contains εἰμι rather than ἠμι, as well as several forms in which η rather than Doric α appears. Indeed, the “Doric” used by Posidippus throughout the papyrus may be understood as an artificial creation influenced by a number of literary traditions. Thus, if (as seems likely) the editors are correct to supplement [εἰν] at AB 85.4, the preposition is best explained as an epicising admixture in a purely literary language rather than as a reflection of the way actual people spoke. Like Pindar and other lyric poets, Posidippus uses πρῶτοϲ (AB 68.5; AB 88.1) rather than Doric πρᾶτος; Theocritus, by contrast, appears to favor the latter in his Doric poems. [34] Characteristic of both Pindar and Theocritus, on the other hand, is the use of the cardinal numbers τρεῖϲ at AB 88.1 and τέϲϲαρεϲ at AB 75.1 (as opposed to τρῖϲ//τρῆϲ and τέτορεϲ) alongside such forms as Doric εἵλομε<ϲ> and ἁμέϲ. These cases are more complicated than that of εἰν, however, since Attic-Ionic elements seem also to have been characteristic of the Doric koine that had developed by the end of the fourth century. [35]

Thus the presence of these forms in Doric contexts could be due to the influence of other literary traditions, but the words are also consistent with the way in which some people actually spoke. [36] This fact has important implications for the way in which we evaluate the “dialect mixing” we find in the epigrams of Posidippus, since the influence of the koine means that even forms like ἐκεῖνος need not be seen as strikingly out of place in a Doric context, though the use of Attic-Ionic η for inherited /a:/ (especially in the endings of first-declension nouns) alongside other clearly Doric features would presumably still have been felt as highly marked, since the distinction between α and η was upheld even in the Doric koine. [37]

The epigrams in which Doric forms occur are not uniformly distributed throughout the papyrus but are instead restricted to the section containing epigrams on statues (ἀνδριαντοποιικά); to the section containing poems on the equestrian victories (ἱππικά); and to the last extant section, labeled τρόποι; elsewhere the only forms with identifiably Doric coloring occur at AB 47.1, where the speaking tomb calls its occupant Ὀναϲαγορᾶτι<ν>, despite using ἥ(τιϲ) as the relative pronoun to refer to her throughout the epigram, and at 94.4 AB, where μίκκος may also be a proper name. [38] The vast majority of these Doric-colored poems occur in the ἱππικά, and these appear to conform to an interesting pattern. Bastianini and Gallazzi may be correct to suggest that the relatively high number of poems in Doric in this section has something to do with the prominence of the Peloponnese and of the projection of the power of the Lagid dynasty at the Peloponnesian athletic competitions, but the association of Doric with epinician lyric also makes it appropriate for poems celebrating equestrian victories. [39] It is crucial to note, however, that only poems for Olympic victories appear in Doric, whereas the dialect of the epigrams on the outcome of Pythian, Isthmian, and Nemean contests is consistently epic/Ionic. The only example of a Doric form among the latter group of poems occurs at AB 86.4, where the narrator refers to himself as Εὐβώταν, but given the use of Ὀναϲαγορᾶτι<ν> at AB 47.1, this sort of dialect coloring in a proper name may have been felt as a special case and thus not strongly marked. [40] Conversely, with two clear exceptions, the poems on Olympic victories for which there are unambiguous markers of dialect have Doric coloring.

The dialectal distinction between poems on Olympic victories and those on all others becomes particularly striking when one considers AB 81.1–2, where the marked reference to the “Dorian” parsley (Δ]ωρικὰ φύλλα ϲελίνων) awarded for a Nemean victory stands in striking juxtaposition to the Attic-Ionic form κεφαλήν. Why the equestrian poems should for the most part divide neatly by dialect is unclear. The only exceptions occur at AB 83 and AB 78. The only clear dialectal marker in the first of these is μνῆμ(α) at AB 83.2. [41] The epigram, in which the supposed speaker is a victorious Thessalian horse, introduces a series of three poems involving Thessalians. The next epigram (AB 85), celebrating an Olympic victory by a Thessalian named Phylopidas, contains no certain dialectal marking, but the following poem, on an Olympic victory by Amyntas of Thessaly, contains a number of Doric forms. More can be said about AB 78, a long epigram that celebrates the Olympic achievements of Lagid royalty, culminating in the Olympic chariot victory of Berenike II. [42] This poem belongs to a sequence of five epigrams, all with Ionic coloring, for victories by the queen at various athletic contests. The victory celebrated in AB 78 seems to have taken place in 248 BCE, and the epigram must thus have been composed relatively late in Posidippus’ life, perhaps at a chronological distance from the other Olympic victory poems, and almost certainly from those celebrating the Olympic victories by other members of the Lagid dynasty.

As far as it goes, the very limited evidence from other poems unambiguously or dubiously attributed to Posidippus may suggest a regular (and natural, given the supposed Dorian ancestry of the Macedonian ruling class) [43] connection between Doric and epigrams on early Macedonian royalty. Thus an epigram attributed either to Asclepiades or Posidippus, APl ^[A] 68, seems to have had a Doric coloring, while the Doricism of the epigram on Lysippus’ statue of Alexander (like Asclepiades’ poem on a statue of the king by the same sculptor) is appropriate both to the artist’s Sicyonian origins and to Alexander’s background. The connection between Doric and the Ptolemaic court was not universal in early epigram (e.g. Asclepiades 64 GP), however, and in any case by the time of Berenike II it may have seemed less important to underscore the alleged Argive ancestry of the Macedonian royal house. [44]

All three Doric epigrams in the ἀνδριαντοποιικά describe statues by sculptors who would themselves have spoken a West Greek dialect. In these poems the narrator is thus represented as a member of the linguistic community of the artist he praises, [45] as is suggested by the narrator’s use of the severe Doric form ἠργάϲατο in a poem praising a statue by a Cretan. [46] In two of the three poems, the narrator also shares a common dialect with the artist’s subject, since both the Macedonian king Alexander in the case of AB 65 and the Cretan hero Idomeneus in the case of AB 64 could be thought to have spoken a form of Doric “in real life.” [47] The placement of Doric in the mouth of Idomeneus is particularly interesting, and has important implications for our understanding of Posidippus’ project in the epigram in which it occurs. I give the full text of AB 64:

ιδομενεια pap.

Like a number of passage of early Hellenistic poetry, the epigram depends for its point on the ecphrastic topos in which the viewer of a work of art asserts that it would be wholly realistic but for its lack of a voice, [48] so that the attribution of direct speech to the statue of Idomeneus suggests that Cresilas’ piece has transcended even this limitation. [49] In this context, as we have seen, the Doric coloring of the words placed in the mouth of the Cretan hero Idomeneus underscores the realism of the statue. By attributing such language to Idomeneus, Posidippus “brings to life” a character whose existence was established for the poet only through his presence in the Homeric poems. [50] Read in this light, the epigram may be seen as engaging in a project of hyper-realism that departs from and implicitly rejects the epic manner of representing individuals, who in the Homeric poems all speak a relatively undifferentiated and elevated epic-Ionic Kunstsprache. [51] This departure from Homeric practice is typically Hellenistic, but may come as a surprise after the first verse, which contains no Doric elements and leaves the initial impression that—as might be expected in a poem about a Homeric hero—the epigram will have an Ionic or (given ἐκεῖνον) at least Attic-Ionic coloring. In any case, the poem’s “realistic” treatment of Idomeneus’ speech makes the epigram a fitting complement to the immediately preceding epigram, in which Hecataeus is praised for producing a realistic sculpture of Philitas in which he mixed in nothing from the form of heroes (AB 63.4 ἀφ᾿ ἡρώων δ᾿ οὐδεν ἔμειξ᾿{ε} ἰδέης): read against that verse, the Idomeneus-statue epigram may be taken as an example of a treatment of “epic” subject matter in which the stylized language is replaced by a more realistic (if naturally still stylized) representation of Idomeneus’ native dialect.

The final Doric poem in the ἀνδριαντοποιικά (AB 68), on the colossal statue of Helios at Rhodes, also seems to use dialect as a marker of realistic speech, although in this case the speech is indirect and implicit rather than direct:

v. 1 π[εριμάκε]α edd. pr., sed π[εριμήκε]α supplere possis v. 4 θηϲεῖν ?

The dialect is notable for several reasons. Although the poem has Doric α rather than Attic-Ionic η after the first verse (τεχνίταν, τέχνα[ϲ]; γᾶϲ), the second word, Ἠέλιον, may naturally lead readers to assume at first that the coloring of the epigram will be epic/Ionic. This use of an apparently “out-of-place” Ionic form early in the poem finds a close parallel in the equestrian poem at AB 73, where the opening words, εὐθὺϲ ἀπὸ γραμμῆϲ, stand at odds with ἁδύ in the third verse. [52] That epigram, the first in the ἱππικά to celebrate an Olympic victory, is thus also the first poem in the section to contain Doric forms. The initial, isolated Ionicism could easily be the product of a slip by a scribe who was not expecting Doric coloring but who later correctly wrote Doric forms without amending his earlier mistake. A similar explanation could also account for the presence of ἠέλιον rather than ἀέλιον at AB 68.1 (contrast GP adesp. 58A.1–3 [AP VI 171.1–3]), but the numerous examples of mixed dialect in both contemporary inscribed epigram and in other early Hellenistic poetry, as well as the care with which the papyrus has been produced in general, make it is reasonable to explore other, less pedestrian solutions.

In the case of AB 68.1, one possibility is that the poet did not wish to use the form ἀέλιον, a common “artificial” Doric form created by analogy with epic ἠέλιος, [53] but this hypothesis, despite the absence of ἀέλιος from the Doric poems of Theocritus (who uses only ἅλιος), requires the improbable assumption that Posidippus placed a greater premium on the “purity” of Doric forms—even in the face of the well established use of ἀέλιος in literary Doric (e.g. Pindar Olympian 1.5; 2.32; Pythian 4.214; Bacchylides 5.161; Theocritus 16.76; Callimachus Hymn 5.89; 6.91)—than on uniformity of dialect in a single epigram. The issue is further complicated by the gaps in the poem, especially that at the end of the first verse. The first editors adopt Austin’s suggestion π[εριμάκε]α, a word well fitted to both the available space and context, but even if this is the adjective Posidippus used, we have no basis for knowing whether he wrote it with the Doric -μακ- or with Attic-Ionic -μηκ-, and the point at which identifiable Doricisms rather than epic/Ionic equivalents first appear in the poem must remain uncertain. [54]

As the first editors suggest, the Doricisms in the epigram seem appropriate to the content of the poem, since the dialect spoken at Rhodes belonged to the Doric family, but the nature of the “ossequio … alla doricità dell’ isola” needs to be explored further [55] The narrator seems initially to be a speaker of Ionic, but he soon adopts the Doricisms of his putative location, since he presents himself as standing in the presence of the statue (cf. τόϲοϲ, τοῦδε). Given the incomplete state of the end of the first verse, we cannot know whether he switches to Doric already in that line. If he did not, the first unambiguously Doric forms appear at the head of the second hexameter in the indirect statement introduced by ὡρίϲατο, “ordained” (LSJ s.v. III). The subject of that verb is Chares, a Rhodian from the city of Lindus, and the Doric coloring of the dependent clause vividly reflects the dialect Chares would have used in establishing a limit to the size of the Colossus.

In this regard, μηθένα in v. 3 is particularly interesting. The forms οὐθείς/μηθείς appear in Attic inscriptions as early as 378/7 BCE and are widespread in inscriptions and papyri throughout the Greek-speaking world by the end of the fourth century. [56] The form with a theta is thus not in itself unusual or necessarily marked, though it is unique in the papyrus, which elsewhere has οὐδείϲ (AB 62.6; AB 63.4; ambiguous at AB 77.4 οὐ[δ]έν). Heraclides of Miletus (fr. 25 Cohn ap. Eust. 452.20–1), however, reports that Doric speakers pronounced delta as theta and gives the examples of ἔθω for ἔδω, μασθόν for μαδόν, and ψύθος for ψεῦδος, and Archestratos of Gela (if the transmitted text is correct) seems to have used ἔθοντες (= ἔδοντες) to parody speakers of Sicilian Doric (fr. 60.11 Olson–Sens). [57] Moreover, the use of ζ to represent original δ (τόζ[᾿ = τόδε) in a sixth-century Rhodian inscription (IG XII 1.737; GDI 4140) suggests that in that local dialect inherited /d/ had come to be pronounced as a spirant. [58] All of this makes it possible, despite the widespread use of μηθείς and οὐθείς in the third century BCE, that the aspirated form of the adjective in its context serves, in conjunction with its referent τεχνίταν, as a marker of a stylized version of Rhodian Doric. [59] Our ignorance about the end of the first verse complicates the picture, but if the phrase μηθένα τεχνίταν is the first recognizably Doric language in the poem, the effect would be especially striking and pointed. If that was the case, the narrator, having begun in epic/Ionic, switched to Doric in order vividly to represent the actual words of the Rhodian sculptor. The influence of that “quotation” then continues throughout the remainder of the poem (τέχνα[ϲ]; γᾶϲ), as if the narrator’s description of Chares’ accomplishments were expressed by the artist himself.

Perhaps the most intriguing juxtaposition of Doric and epic-Ionic forms within a single poem in P.Mil.Vogl. VIII 309 occurs at AB 102, the first epigram in the section of the papyrus labeled τρόποι:

As in AB 64, Doric forms in the speech of a Cretan speaker may be intended as a realistic feature, “come riflesso della parlata dorica del defunto.” [60] We have already observed that the two “severe” Doric forms (ἠάϲατε and Φιλάρχω) that are preserved in the poem may, like ἠργάσατο in the epigram on Cresilas’ statue of Idomeneus epigram, be introduced to suggest specific features of actual Cretan speech. Despite the presence of such specially marked Doricisms, however, the dialectal coloring of the epigram is not consistent, since the poem also contains ϲῆμα (AB 64.3) and ξενίηϲ (AB 64.4), for which the poet could easily have written ϲᾶμα and ξενίαϲ had he wished. As Palumbo Stracca has pointed out, [61] the existence of similar variations in inscribed epigram from an early period cautions against emending away the inconsistency of the transmitted text. If this is preserved, however, the combination of strongly Doric forms with others that are foreign to Doric would seem at odds with any attempt to capture, even in a stylized manner, the speech of the Cretan narrator.

For Hellenistic readers attuned to nuances of dialect, however, the mixture of Doric and non-Doric forms may have formed part of the point. The poem, of which a text was made available in the preliminary publication of select epigrams from the papyrus, has already received a relatively large degree of critical attention, most of it directed to unpacking the significance of the speaker’s explanation for his taciturnity in the final verse. [62] Whatever the precise connection between the speaker’s Cretan background and his reticence, however, the speaker represents his unwillingness to speak at length as a natural consequence of his actual location in a foreign land, [63] and the simplest conclusion is that the speaker, having been buried away from Crete, explains his reticence by reference to the expected behavior of living foreigners. [64] The precise significance of the generic label τρόποι given to the section to which this poem belongs is uncertain, but the epigram and those that follow may well be intended to illustrate certain types of character through first-person speech. In this case, the poem turns on its head the epitaphic convention in which the deceased person or his tomb calls for passersby to stop and pay attention, and in so doing the epigram paints a humorous and vivid picture of a certain type of speaker, the taciturn foreigner.

When viewed in this context, the juxtaposition of Doric and non-Doric forms may be seen as part of a subtle characterization of the speaker, Cretan by birth but resident, in life and death, in an alien land. For people like him, speech serves as a fundamental marker of foreign origin. Implicit in the poem’s suggestion that foreigners are naturally taciturn may be the idea that they are so not only because reticence is expected of them, but also because by speaking they expose their status. In the mouth of an alien like Menoetius, the mixture of Doric and non-Doric forms represents a stylized version of the way in which foreigners actually speak, combining features of their native tongue with elements of the language indigenous to the place in which they find themselves. If the dialectal variation of the poem is understood in this way, the Ionic coloring of the final words, ἐπὶ ξενίηϲ, takes on special significance, since that phrase both announces the speaker’s foreign status and serves as an icon of his position as social and linguistic outsider: the very word which means foreign is itself uttered in a dialect “alien” to the speaker’s own. [65]

The use to which dialect seems to be put in this poem—and elsewhere in the papyrus—provides a helpful backdrop for thinking about the juxtaposition of Doric and Ionic forms of the relative pronoun in the text of the poem with which this essay began, Posidippus AB 139 (8 GP = AP XII 131). On the one hand, the papyrus offers some justification for regularizing the forms in that epigram, since a comparison of the text of the Alexander-statue epigram contained in the single manuscript witness and that preserved on the papyrus suggests that some but not all of the original Doricisms were displaced in the course of transmission. Given the relatively scarcity of Doric forms in the poems of Posidippus that are preserved in the Anthology—and the predominance of Attic-Ionic in the Anthology in general—the Doricisms in the paradosis of AP XII 131 seem more likely to have been subject to alteration, and if one were to regularize, it would (pace Gow-Page) therefore make best sense to replace transmitted East Greek forms with Doricisms. On the other hand, the juxtaposition of Doric and Ionic forms of the relative pronouns in AP XII 131 is not inherently less likely than the use of markedly non-Doric forms like γραμμῆϲ or ξενίηϲ in close conjunction with Doric elements in individual poems of the new papyrus. Let us therefore briefly explore the consequences of assuming that Posidippus did in fact write ἅ three times in the first verse of AP XII 131 but Συρίης and ἥ elsewhere.

To understand what function this variation might have served, we must consider the poem as a whole more closely. The epigram, cast as a prayer to Aphrodite, opens in the elevated manner of hymn, [66] and the language and style of the first couplet are evocative of “high” lyric poetry in general. The verb shared by the opening relative clauses, ἐποιχνεῖς, is an interesting marker in this regard. The word is rare, occurring in literature apparently only in two contextually similar passages: at Bacchylides 10.1–2, where it is a likely supplement in an address to personified Fame (Φή]μα, σὺ γ[ὰ]ρ ἀ[       ἐπ]οιχνεῖς | [φῦ]λα), [67] and in the Delphic paean by Aristonous, where it appears as part of an address to Apollo (verses 11–12, p. 163 Powell μαντοσύναν ἐποιχνεῖς, ἰὴ ἰὲ Παιάν). [68] Posidippus’ use of the word, whether or not it has a specific model in antecedent poetry, thus helps to evoke the tradition of choral lyric. That the other passages in which ἐποιχνέω occurs have a Doric coloring is telling, and the historical association of Doric with lyric poetry in general makes the use of the Doric forms of the pronoun in the invocation of Aphrodite in the first verse of Posidippus’ epigram particularly appropriate. [69]

At first glance, therefore, the generic background of the poem’s opening might offer some justification for regularizing all of the Ionic forms to their Doric equivalents. A closer reading, however, suggests that the dialectal variation might itself be significant. Some of the wit of the epigram, indeed, depends on the disjunction between the elevated manner of the invocation of the goddess in the first couplet and the coarse content of the final description of the human Callistium. The speaker, whether the woman herself or someone else, uses “special” Doric forms, with their marked association to choral lyric, to refer to the goddess in the first line, but reverts to “ordinary” Ionic forms elsewhere. The effect is to create an opposition between the goddess, addressed at the opening of the epigram via Doric forms of the relative pronoun, and the sexually willing Callistium, who is described with the “unmarked” language of Ionic in the final relative clause, ἣ τὸν ἐραστὴν | οὐδέποτ᾿ οἰκείων ὦσεν ἀπὸ προθύρων.

Given our ignorance about the text, such a reading of the dialect forms in the epigram must be provisional, and the fact that the paradosis has Συρίης rather than Συρίας complicates the opposition that I have suggested. Taken cumulatively, however, the evidence offered by the new papyrus cautions against dismissing this and other passages in which the dialect seems to be inconsistent as merely the result of errors in transmission or as products of authorial carelessness. To the contrary, the new papyrus shows that Posidippus was interested in the nuances of dialect, which he uses as a powerful if often subtle tool not only for affiliating his compositions with the various literary traditions on which he draws, but also for creating other, more complex, forms of meaning. In this respect, as in many others, P.Mil.Vogl. VIII 309 offers unparalleled access into the practices of an important figure in the literary history of the third century BCE, and by implication into those of other learned Hellenistic epigrammatists as well.