Analogical ‑εσσι datives in Sicilian Doric
Borrowing, independent development, or both?
p. 209-228
Résumés
The Aeolic dialects and many varieties of West Greek show the analogical extension of the dative plural marker ‑εσσι from s‑stems (ἔπεσσι) to other athematic stems (ἄνδρεσσι). Against the theoretical background provided by classic studies on the origin of ‑εσσι datives, this article focuses on the evidence from classical Sicily, where ‑εσσι datives are traditionally considered to be a Syracusan trait, imported from the dialect of Corinth. A review of the evidence shows that this idea has little factual basis apart from the use of ‑εσσι datives in the literary Doric of Epicharmus and Sophron. An analysis of the epigraphic attestations of ‑εσσι datives in their textual context instead allows for a different interpretation. The ‑εσσι datives, which spread as a remedy to morphonological difficulties, may have been chosen to stylistically mark certain passages through a morphological feature endowed with a poetic pedigree. As in later Doric Koinai, ‑εσσι datives were instrumental to the creation of a high‑register prose language.
Les dialectes éoliens et de nombreuses variétés du grec occidental montrent l’extension analogique du marqueur de datif pluriel ‑εσσι des thèmes en sifflante (ἔπεσσι) à d’autres athématiques (ἄνδρεσσι). Dans le cadre des hypothèses sur l’origine des datifs en ‑εσσι proposées par les études classiques, cet article se concentre sur les données de la Sicile classique, où les datifs en ‑εσσι sont traditionnellement considérés comme un trait syracusain, importé du dialecte de Corinthe. Un examen des données montre que cette idée a peu de fondement réel, à l’exception de l’utilisation des datifs en ‑εσσι dans le dorien littéraire d’Épicharme et de Sophron. Une analyse des attestations épigraphiques des datifs en ‑εσσι dans leur contexte textuel permet au contraire une interprétation différente. Les datifs en ‑εσσι, qui se sont répandus comme remède à des difficultés morphonologiques, ont pu être choisis pour souligner stylistiquement certains passages avec une caractéristique morphologique d’origine poétique. De la même manière que dans les koinai doriennes postérieures, les datifs en ‑εσσι ont joué un rôle déterminant dans la création d’une variété de prose au registre élevé.
Texte intégral
1. Introduction
1A much debated issue in Greek linguistics is the extension of the dative plural marker ‑εσσι from s‑stems (ἔπεσσι) to other athematic stems.1 The phenomenon is typical of the Aeolic group and prominently features in the Homeric epics, but analogical ‑εσσι datives (henceforth simply “‑εσσι datives”) are also attested in various dialects of the West Greek group and in Pamphylian. The origin and diffusion of this trait has engaged scholars for almost two centuries. Based on their work, it is now possible to say that various factors are behind the spread of ‑εσσι to athematic stems:2
- analogy with the plural declension of thematic stems;3
- the speakers’ perception that ‑εσσι was an ending and that everything that preceded it was the “stem”;4
- the pressure of morphonological transparency, which led to the replacement of ‑σι, which often made the root opaque, with ‑εσσι;5
- some specific morphonological developments which led to the partial merger of the inflections of the εὐγενής and ‑εύς stems in some dialects.6
2Against this background, the present paper will focus on the ‑εσσι datives of the Doric dialect of Sicily, where they are attested from the classical age onwards. The core of the paper is concerned with the sociographic context of use of these forms. A case will be made for the interpretation of Sicilian ‑εσσι datives as marked elements that played a role in the style and register of the texts. The first part of the paper looks into the dialectal distribution of ‑εσσι datives in order to assess the supposed Corinthian origin of this trait in the Doric of Sicily.
2. Distribution in the Greek dialects
3The extension of ‑εσσι to all athematic stems is accomplished only in Lesbian and Boeotian.7 In Thessalian ‑σι is preserved in some dental and ‑ντ‑ stems.8 In the West Greek group, ‑εσσι datives are attested in Delphian, Elean, Locrian, Cyrenaean, in the dialects of the Saronic Gulf, and in Sicilian Doric; they often compete with analogical 3rd‑declension datives in ‑oις.9 In Delphian ‑εσσι is pervasive, but not extended to stems in ‑εύς; original forms in ‑σι are also preserved. In Elean we have five occurrences. The first two are Θεσπιέσσιν (Minon, I.dial. éléennes, no. 15, ca 475 BC) and φυγάδεσσι (Minon, I.dial. éléennes, no. 30, before 324 BC).10 The third, Μαντινέσι (Minon, I.dial. éléennes, no. 22, ca 450‑425 BC), shows a graphically degeminated ending.11 The fourth (Πελλανέσι, l. 1, an ethnonym in ‑εύς) and fifth forms (τρισδικασστε̄́ρεσι, l. 1, from a compound hapax in ‑τήρ), both with degeminated endings, feature in a published bronze tablet from Olympia recording the decisions of three judges from Pellana (ed. pr. in Hallof 2021, ca 475‑450 BC).12
4From West and East Locris we have only three forms: πάντεσιν (IG IX, 12, 3, no. 609, Naupactus, ca 500 BC), χρημάτεσσι, and Κεφαλλανέσσι (both in IG IX, 12, 5, no. 1909, 2nd c. BC). In Cyrenaean ‑εσσι probably spread only to ‑εύς stems, and attestations date to the 4th‑2nd c. BC (e.g. Mελιβoέσσι and Μεγαρέσσι in SEG IX, no. 2, 330‑326 BC).13 The evidence from Sicilian Doric, starting in the early 5th c., is discussed in § 4.
5An important point to note is that in prose inscriptions from the Peloponnese ‑εσσι is not only sporadic but also post‑classical. The evidence amounts to πολίεσσι at Epidaurus (IG IV2, 1, no. 74, ca 300‑250 BC),14 ἄνδρεσσιν at Stymphalus, Arcadia (IG V, 2, no. 357, before 234 BC, a decree in Doric Koina), and the ambiguous Ἐπιτελίδεσ[σι] at Argos (IG IV, no. 569, a Hellenistic dedication; it could also be restored as Ἐπιτελιδε[ῦσι]).15 This meagre evidence speaks against the old theory that the ‑εσσι datives belong to an Aeolic substratum.16 On the contrary, they are partly a contact phenomenon (in those areas closer to Aeolic) and partly an independent development, triggered by morphonological concerns in the first instance.17 I shall argue below that Sicilian ‑εσσι datives are an independent development prompted by morphological analogy.
3. Corinth and its colonies
6Studies of the ‑εσσι datives in Sicilian Greek assume that they spread to Sicilian Greek via the dialect of Syracuse, which inherited them from Corinthian.18 This is a popular but problematic view.19 Due to the nature of the epigraphical data from this area, we simply have no evidence for 3rd‑declension plural datives.20 All known attestations of ‑εσσι datives from Corinthian colonies but one are post‑classical. At Corcyra there are only two instances of ἁρμάτεσσι in IG IX, 12, 4, no. 798 (2nd c. BC), recording a donation and a decree on its administration: the text is in Doric, but notice the regular datives χρήμασιν (l. 10 and 45), τοῖς αἱρεθεῖσι (l. 59‑60 and 74) and τοῖς ἐφ<αι>ρεθεῖσιν (l. 93), which – together with features such as εἰ (l. 17, 25, 29, 72, etc.) for Doric αἰ, τετάρτῃ (l. 1) for τετάρτᾳ and ἀρχή (l. 84) for ἀρχά – point to some Koine influence.21 From Epidamnus there are only three participles in a decree from Magnesia on the Maeander (I.Magnesia, no. 46, 3rd c. BC) which may not fully reflect the local dialect of Epidamnus.22
7The oldest attestation is hε̄ρṓεσσ[ι]ν, which occurs on face B (l. 6) of a late‑5th c. text of debated nature from Apollonia in Illyria (Greek Ritual Norms, no. 40), recording a regulation of the cult of Asclepius sanctioned by the oracle of Dodona.23 É. Lhôte has interpreted this dative as an epic form “embedded” in the text to lend the oracular response a solemn tone, an opinion dismissed by Alonso Déniz 2022a, p. 446, who ascribes the dative to the local dialect of Apollonia.24 It is worth recalling that we have no evidence to support the claim that ‑εσσι datives were widespread in the classical dialect of Corcyra. Moreover, it is far from improbable that some ‑εσσι datives – especially those characteristic of forms with a poetic pedigree, like ἡρώεσσι – may have been used for stylistic heightening.25 The very same form, ἡρώεσσι, is found in two Sicilian texts where poetic influences may be detected, as argued in § 4.2. The use of some “poetic” elements should not be seen as proof that the dialect of these texts is not authentic, but simply as evidence that the combination of local and poetic traits was integral to the language of such oracular and sacred texts. To sum up, the evidence is too limited and late to justify the conclusion that ‑εσσι datives developed in the archaic dialect of Corinth before being transferred to its colonies, including Syracuse: there are no ‑εσσι datives in classical Syracusan, except for literary texts, as I will discuss below.26
4. ‑εσσι datives in Sicily
8As shown in tab. 1, in Sicilian prose inscriptions ‑εσσι datives are attested 23 times (no. 24 is doubtful) and always in areas where Doric was spoken.27 Three further instances occur in Epicharmus and Sophron: the nouns ῥίνεσσι and γυναικάνδρεσσι (Epicharmus) and the participle τρηματιζόντεσσι (Sophron). Two attestations in Sophron (ἡρώνεσσι and καθαρμάτεσσιν) are more problematic (see infra § 4.1).
Tab. 1 – ‑εσσι datives in Sicilian prose inscriptions by date.
Form | City | Inscription | Date | Typology | |
Classical period | |||||
1. | Εὐμενίδεσ(σ)ι | Selinous | I.dial. Sicile II, no. 18 | early 5th c. BC | lex sacra |
2. | ἡρώεσσι | Selinous | I.dial. Sicile II, no. 18 | early 5th c. BC | lex sacra |
3. | ποτηρίδεσσι | Selinous | I.dial. Sicile II, no. 18 | early 5th c. BC | lex sacra |
4. | ἄνδρεσ(σ)ι | Gela | I.dial. Sicile I, no. 134.b | 475‑450 BC | defixio |
5. | γυναίκεσσι | Gela | I.dial. Sicile I, no. 134.b | 475‑450 BC | defixio |
Hellenistic period | |||||
6. | ἱππέεσσι | Entella | I.dial. Sicile I, no. 205 | 254‑241 BC? | decree |
7. | γενετόρεσσι | Nakone | I.dial. Sicile I, no. 206 | 254‑241 BC? | decree |
8. | ἄνδρεσσι | Entella | I.dial. Sicile I, no. 207 | 254‑241 BC? | decree |
9. | συνοικιξάντεσσι | Camarina (found at the Asclepieion at Cos) | I.dial. Sicile I, no. 117 | 242 BC | decree |
10. | πολίεσσι | ||||
11. | παναγυρίεσσι | ||||
12. | παναγυρίεσσι | Gela (found at the Asclepieion at Cos) | I.dial. Sicile I, no. 118 | 242 BC | decree |
13. | περιαγγελλόντεσσι | ||||
14. | Τιτάνεσσι | Lilybaeum | I.dial. Sicile I, no. 80.IIa | late 3rd c. BC | defixio |
15. | [ἐ]μβασίεσσιν (?) | Camarina | Dubois, BE, 1990, no. 859 (cf. SEG XLII, no. 1462). | 4th-3rd c. BC | contract |
16. | παθόντεσσι | Halaesa | SEG LIX, no. 1100 | 1st c. BC? | decree |
Late Hellenistic/Early Roman period | |||||
17. | Παίδεσσι | Acrae | SEG XLII, no. 825 | 1st c. BC-1st c. AD | dedication |
18. | [Παίδε]σσι | Acrae | SEG XLII, no. 828 | 1st c. BC-1st c. AD | dedication |
19. | Παίδεσσι | Acrae | SEG XLII, no. 829 | 1st c. BC-1st c. AD | dedication |
20. | Παίδεσσι | Acrae | SEG XLII, no. 830 | 1st c. BC-1st c. AD | dedication |
21. | Παίδεσσι | Acrae | SEG XLII, no. 832 | 1st c. BC-1st c. AD | dedication |
22. | Παίδεσσι | Acrae | SEG XLII, no. 836 | 1st c. BC-1st c. AD | dedication |
23. | Παίδε[σσι] | Acrae | SEG XLII, no. 833 | 45 AD | dedication |
Uncertain | |||||
24. | υἱέεσσι (?) | Syracuse | IG XIV, no. 10 | Date unknown | dedication |
9The early Sicilian evidence allows for the following conclusions, which I shall use as the basis of my working hypothesis. First, analogical ‑εσσι appears to have been extended to a variety of athematic stems already in the earliest attestations, except for those in ‑εύς and those in ‑ι‑, where ‑εσσι occurs only from the Hellenistic Age (in the Selinous lex sacra we still have ‑εῦσι: Τριτοπατρεῦσι, see below tab. 2). It seems unlikely that the primary purpose of ‑εσσι was to avoid undesirable phonetic changes which would lead to paradigmatic opacity.28 Only the participle τρηματιζόντεσσι and the dental stems Εὐμενίδεσσι and ποτηρίδεσσι would be more transparent than the regular datives in ‑σι. The Sicilian evidence would rather support the view that the formal trigger behind the ‑εσσι forms was morphological analogy with the nominative plural (Εὐμενίδες > Εὐμενίδεσ‑σι, etc.).
10Secondly, the chronology of the first ‑εσσι datives does not allow us to reach any firm conclusion about the date of their introduction into Sicilian Greek. It may well be that they are old features in the Doric varieties of archaic Sicily. It is equally possible, however, that they arose during the classical period and hence that in these inscriptions we witness their first appearance.
11Finally, the distribution of these early forms does not warrant the attribution of this trait to any dialectal subvariety – be it the dialect of a continental metropolis or the variety of the dialect that developed in Sicily. The early attestations in Syracusan are limited to Epicharmus and Sophron, while archaic and classical inscriptions from both Corinth and Syracuse have yielded no datives of any kind. The only possible attestation of an ‑εσσι dative at Syracuse, the form υἱέεσσι (no. 24 in tab. 1), is very problematic. IG XIV, no. 10 is an undated inscription from Syracuse which is lost: the edition in IG XIV depends on an earlier transcription.29 Its Doric dialect suggests a date within the 1st c. AD, but it is impossible to ascertain whether it may be older and hence significant for the presence of ‑εσσι datives at Syracuse.
12It is methodologically flawed to use the attestations of ‑εσσι in inscriptions from the Syracusan sub‑colony of Acrae as evidence that the trait was typical of Corinthian and Syracusan since we are dealing with the same form, Παίδεσσι, always featuring in the formula Παίδεσσι καὶ Ἄννᾳ in inscriptions not earlier than the 1st c. BC. Nothing certain can be inferred on the classical Syracusan dialect from this kind of evidence. The same applies to the Hellenistic inscriptions of Akragas and Gela, which were occupied by Syracuse during the 5th c. BC: it is far from certain that their use of ‑εσσι datives (no. 9‑13 in tab. 1) depends on their adoption of the “authentic” Syracusan dialect rather than on the spread of this morphological trait to vast parts of the island – the latter scenario seems more probable.
13The hypothesis that ‑εσσι may have spread to Sicilian Greek from the dialect of Megara Hyblaea is also weak.30 While there are three attestations in the lex sacra from Selinous, there are no other attestations at either Selinous or her metropolis Megara Hyblaea. There are no datives in classical inscriptions from continental Megaris, nor could their early appearence be proven on the basis of later attestations.31 The only certain fact is that already in the 5th c. BC ‑εσσι datives were used in more than one Doric-speaking polis of Sicily.32
14Before turning to the analysis of each dative in context, let us cast a glance (tab. 2) at the evidence for ‑σι datives in classical and post‑classical Sicilian inscriptions (the latter include only inscriptions dated to before the end of the 1st c. BC).33
Tab. 2 – ‑σι datives in Sicilian prose inscriptions by date.
Form | City | Inscription | Date | Typology | |
Classical period | |||||
1. | πᾶσι | Selinous? | I.dial. Sicile II, no. 30 | 5th c. BC | defixio |
2. | Τριτοπατρεῦσι | Selinous | I.dial. Sicile II, no. 18 | early 5th c. BC | lex sacra |
Dubious attestations in this period | |||||
3. | <ΕΡΟΙΣΙΘΕΟΙ> (= (h)ε̄́ρō{ι}σι?, see discussion below) | Megara Hyblaea | Arena, Iscrizioni I2, no. 78 | mid‑7th c. | dedicatory graffito on vase |
4. | πᾶσι, or part of the personal name Πασάρατος | Megara Hyblaea | I.dial. Sicile I, no. 20 | early 6th c. BC | lex sacra |
5. | κυ<ἱ>άσι (= καὶ υἱάσι?) or KYTAΣ (personal name) | Camarina | I.dial. Sicile I, no. 118 | ca 450 BC | defixio |
Post‑classical | |||||
6. | πᾶσι | Syracuse | IG XIV, no. 2 | 269‑215 BC | honorific |
7. | πᾶσι | Nakone | I.dial. Sicile I, no. 206 | 250‑241 BC? | decree |
8. | πᾶσι | Apollonia (San Fratello) | IG XIV, no. 359 | 2nd c. BC | dedication/honorific |
9. | πᾶσι | Halaesa | IG XIV, no. 353 | 200‑100 BC | dedication/honorific |
10. | δρι<σ>ί or δρίε<σ>ι or δρίει? (see infra n. 42) | Acrae | IG XIV, no. 217 | 200‑100 BC | cadastral |
11. | πᾶσι | Akragas | SEG XXXVII, no. 757 | 2nd‑1st c. BC | honorific |
12. | πᾶσι | Tauromenium | Arangio Ruiz, Olivieri 1925, no. 13 passim | 70‑36 BC | accounts |
15At present there are only two secure attestations of ‑σι datives in archaic inscriptions: no. 1 and 2. Both are probably from Selinous, though the provenance of the 5th‑c. defixio (I.dial. Sicile II, no. 30) in which πᾶσι is attested has been disputed by L. Bettarini on account of its letter forms.34 Τριτοπατρεῦσι, attested in the lex sacra from Selinous which also contains the ‑εσσι datives Εὐμενίδεσσι, ἡρώεσσι and ποτηρίδεσσι, shows that around the beginning of the 5th c. the extension of ‑εσσι to all athematic stems had not yet been accomplished. The dative plural πᾶσι is also routinely employed in Hellenistic inscriptions in Doric Koina from Syracuse, Halaesa and Acrae, as well as in later texts. S. Mimbrera Olarte, who assumes that in classical Sicilian Doric ‑εσσι must have been pervasive, explains the preservation of πᾶσι through the fact that high‑frequency words resist analogical change (πᾶς is no. 2 in the list of the one hundred most frequent words of the TLG corpus).35 This may well be true, but it is telling that two frequent words such as ἀνήρ (no. 42 in the TLG) and γυνή (no. 80 in the TLG) do take analogical ‑εσσι in Sicilian Greek.
16A dative πᾶσι may also be attested in the fragment of a sacred law from Megara Hyblaea (no. 4), the first two lines of which are read as Πασαράτō hάδε (“these are the sacrifices of Pasaratos”) in M.T. Manni Piraino’s ed. pr. and by both L. Dubois and Arena.36 However, M. Guarducci proposed reading πᾶσι ἀρὰ το͂ θεο͂ hάδε “this is the imprecation of the god for all”.37 As noted by M. Guarducci, M.T. Manni Piraino’s interpretation overlooks the distribution of word‑division, signaled by two vertical points.38 They clearly mark out the group <ΑΡΑ> as well as <TO>: if Πασαράτō were the correct reading here, we would have a personal name fragmented into three units by word‑dividers which, however, in the remaining lines are always placed at word‑end. Because the stone was slightly cut on its right side, the restoration of theta at the end of l. 1 and of epsilon at the beginning of l. 2 (the inscription is boustrophedon) is possible. If M. Guarducci is right, we would have evidence here that πᾶσι was the preferred form also in archaic Megara Hyblaea (at any rate, as A. Alonso Déniz pointed out to me, the personal name Πασάρατος would also exhibit the regular dative plural in its first member).
17This may lend some weight to another possible attestation of a ‑σι dative from Megara Hyblaea, the partly restored form (h)ε̄́ρō{ι}σι in the dedicatory graffito on a proto‑Corinthian olpe dated to the mid‑7th c. BC (no. 3 in tab. 2). This too is M. Guarducci’s interpretation of the first part of a fragmentary sequence that reads <ΕΡΟΙΣΙΘΕΟΙ>.39 The main problem with this interpretation is the stray iota, which cannot be part of the stem of ἥρως: a spelling mistake may perhaps be assumed. Other restorations have been proposed (ἀμφοτέροισι, ὑπέροισι, ἱεροῖσι).40
18No. 5 is also problematic. The reading κυ<ἱ>άσι, suggested by A. Olivieri in the ed. pr. (Arena, Iscrizioni II2, no. 129) is not accepted by L. Dubois (I.dial. Sicile I, no. 118), who prefers reading the personal name KYTAΣ.41 This would match the four preceding words, which are all names in the nominative case (though feminine: L. Dubois speculates whether KYTAΣ may be a mistake for ΣKYTA). The interpretation “X, Y, Z for (their) sons” is not syntactically impossible, but the spelling κυ<ἱ>άσι instead of the expected χυ<ἱ>άσι also speaks against it.
19Scanty and problematic as the evidence presented here may be, on balance it seems possible that in early Sicilian Doric some ‑σι datives were still in use. As evidenced by no. 6‑12, some forms also survived in post-classical Sicilian Greek (see also infra § 5).42
4.1. ‑εσσι datives in the fragments of Epicharmus and Sophron
20Let us now turn to the spread of ‑εσσι datives in the Doric dialect of Sicily. I will first examine their attestations in the language of Epicharmus and Sophron, which lend a literary dimension to a seemingly only dialectological question:
γυναικάνδρεσσι ποθεινοί
longed‑for by woman‑men (or: women and men) (Epich., frg. 224 K.‑A.; from an unidentified play [personal transl.])
πρᾶτον μὲν αἰ κ’ ἔσθοντ’ ἴδοις νιν, ἀποθάνοις· / βρέμει μὲν ὁ φάρυγξ ἔνδοθ’, ἀραβεῖ δ’ ἁ γνάθος, / ψοφεῖ δ’ ὁ γομφίος, τέτριγε δ’ ὁ κυνόδων, / σίζει δὲ ταῖς ῥίνεσσι, κινεῖ δ’ οὔατα.
First of all, should you see him eat, you would die. His throat thunders inside, his jaw gnashes, his molar cracks, his canine tooth creaks, he hisses from his nostrils, and waggles his ears (Epich., frg. 18 K.‑A. [personal transl.]).
21γυναικάνδρεσσι of frg. 224, from an unidentified play, probably identifies effeminate men (determinative compound) or men and women together (a dvandva compound). The extant two words of the fragment correspond to the second hemistich after the “feminine” caesura of a dactylic hexameter.43 Whatever the original metre may have been, the combination between lexical choice and rhythm makes it likely that this was one of the lines in which Epicharmus imitated epic or lyric poetry. The ‑εσσι dative thus gestures towards poetic style, even if one assumes that it was a common feature in Epicharmus’ Sicilian dialect.44
22A similar interpretation may also be possible for ῥίνεσσι in frg. 18. The lines, pronounced by an unknown character, describe the monstrous way in which Heracles eats.45 The fragment is in iambic trimeters and features three verbs of noise which occur in Homer (βρέμω, ἀραβέω, τρίζω) and are later used mostly in poetic texts. The verb βρέμω is attested in poetry and high-register prose.46 ἀραβέω too belongs to poetic texts, starting from the pseudo-Hesiodic Scutum. σίζει in l. 4 may be a hint at Od., IX, 394 (where σίζ(ε) is used to refer to Polyphemus’ “sizzling” blinded eye). Another term that may evoke Polyphemus is φάρυγξ (l. 1; cf. Od., IX, 373; E., Cyc., 215, 356, 410 and 592). More hypothetical epic hints may be ἔνδοθ(ι), which is markedly poetic, and the hexametric rhythm of the sequence ταῖς ῥίνεσσι, κινεῖ. This is the oldest attestation of the dative of ῥίς in Greek, which is later usually found in the morphologically regular form ῥισί.47 In conclusion, perhaps ῥίνεσσι was employed to bestow an epic flavour to this description of Heracles.
23Sophron is a different case from Epicharmus since he wrote his mimes in prose. Only one of the possible three attestations of ‑εσσι datives in his works is unproblematic. This is the participle τρηματιζόντεσσι of frg. 124 K.‑A., seemingly a description of an after‑dinner party:
δειπνήσας ὠστίζεται τοῖς τρηματιζόντεσσι
after dinner he jostles with the dice‑players (personal transl.)
24The extension of ‑εσσι to participles is already a Homeric feature, later imitated only by Pindar among the classical poets. Prosodically, the last four syllables of the participle τρηματιζόντεσσι find a parallel in the placing of the participle σπευδόντεσσιν at line‑end in Il., XVII, 745. The issue of whether Sophron’s prose uses poetic rhythms is an open one. E. Norden, recalling Aristotle’s judgment of Sophron as an author blending prose and poetry (Arist., Po., 1447b, 10), positively identified Sophron’s prose as “ein sicheres Beispiel solcher nach Kola gegliederten rhythmischen Rede”.48 J.H. Hordern, although in principle open to the idea that “we should expect to find similar passages [i.e. anapaestic and dactylic rhythms] in Sophron”, remains sceptical about Sophron’s use of prose rhythm.49 The issue, indeed, would need to be explored afresh and this fragment is too short to allow for any clear‑cut conclusions. τρηματίζω is not a poetic verb (it is only attested here), while ὠστίζομαι is mostly used in comedy. Perhaps the occurrence of two ‑ίζω formations within three words may have a stylistic motivation (imitation of a given register?), and maybe a potentially epic‑sounding participle (τρηματιζόντεσσι) was used to provide a contrast with the prosaic context, but this remains speculative.
25The other two attestations of ‑εσσι datives in Sophron are highly suspicious on the textual level. I will discuss them only briefly because their testimony is almost useless for the purposes of the present investigation. The first dative is ἡρώνεσσι, attributed to Sophron by Priscian in a likewise suspicious statement concerning the Syracusan pedigree of the metaplasm ἥρων for ἥρως:
[…] Syracusiis enim ἥρων pro ἥρως dicunt. sic Sophron παρ’ ἡρώνεσσι non ἥρωσι dixit.
[…] In Syracuse they use ἥρων instead of ἥρως. Hence Sophron used παρ’ ἡρώνεσσι, not ἥρωσι (Priscian., Inst., 6, 70; cf. Sophr., frg. 151 K.‑A. [personal transl.]).
26The recharacterised accusative ἥρωνα (for ἥρων) is attested once in a 3rd‑c. BC inscription from the Cos Asclepieion (IG XII, 4, 1, no. 72, l. 23) and, based on this, we also have Ἥρωνος and Ἥρωνι showing the extension of ‑ν to other cases of the singular declension (the forms are attested from the 1st c. AD).50 However, except for Priscian’s testimony, there is no independent evidence that the extension of ‑ν concerned the plural. It is also problematic that the dative plural ἡρώνεσσι occurs much earlier than the singular forms from which the analogy is supposed to have spread. In conclusion, it is impossible to tell whether the ἡρώνεσσι that Priscian attributes to Sophron is authentic, and to define its context of occurrence more closely.
27The second dubious instance is the dental‑stem dative καθαρμάτεσσιν in a textually corrupt invocation (possibly to Hecate) preserved in Plutarch with no attribution. Restoring some of the corrupt forms so that they could comply with Doric, U. von Wilamowitz‑Moellendorff ascribed the fragment to Sophron:51
αἴτε κα ἀπ’ ἀγχόνας ἀΐξασα αἴτε κα λεχοῦν διακναίσασα αἴτε κ’ ἂν νεκρὸς μολοῦσα πεφυρμένα ἐσέλθῃς, αἴτε κα ἐκ τριόδων καθαρμάτεσσιν ἐπισπωμένα τῷ παλαμναίῳ συμπλεχθῇς.
Whether you come hastening from a hanging, or from grinding to death a woman giving birth, or coming from ranging about corpses, or whether you have been engaged with a murderer, or drawing him to you from the crossroads by means of the blood he had cleansed from himself (Plu., De superstitione, 10, p. 170 B; cf. Sophr., frg. **8 K.‑A. [transl. Hordern 2004, p. 49]).
28The quotation is problematic and replete with textual problems.52 However, one element that has not been noted so far is that the prose of this fragment is characterised by rhetorical figures and assonances: the four anaphoric αἴτε κα; the rhyming participles ἀΐξασα and διακναίσασα; the rhyming participles πεφυρμένα and ἐπισπωμένα; the rhyming subjunctives ἐσέλθῃς and συμπλεχθῇς. Was καθαρμάτεσσιν perhaps integral to the construction of a certain sound effect?
29Based on the literary evidence reviewed in this section, as a tentative conclusion I wish to suggest that ‑εσσι – whether it was already widespread in Syracusan (as assumed, e.g., by Willi 2008, p. 129) or not – entered the written register when a certain heightening was required, in some cases through overt epic allusions. We shall now turn to the inscriptions to see whether the use of ‑εσσι datives can be thought to be conditioned by the register and style of the text in which they appear, and hence to be a choice dependent on a sociostylistic register.
4.2. The ‑εσσι datives of Sicilian classical inscriptions reconsidered: the lex sacra of Selinous
30The earliest ‑εσσι datives from Sicilian inscriptions come from the lex sacra of Selinous, a complex text that details the rite to be performed to purify murderers.53 The three datives occur in col. A of the text and are written with the ending ‑εσι, which is due to the inscriber’s practice of simplifying double consonants (see too περιρ(ρ)άναντες of l. 12‑13):54
7 το͂ν hιαρο͂ν hα θυσία πρὸ Ϙοτυτ(τ)ίōν καὶ τᾶς ἐχεχε̄ρίας πένπ̣[τōι] ϝέτει hο͂ιπερ hόκα hα Ὀλῡνπιὰς ποτείε̄. το͂ι Διὶ : το͂ι Εὐμενεῖ θύ[ε̄]ν̣ [καὶ] ταῖς : Εὐμενίδεσ(σ)ι : τέλεον, καὶ το͂ι Διὶ : το͂ι Μιλιχίōι το͂ι : ἐν Μύσϙō : τέλεον : τοῖς Τρ‑ 10 ιτοπατρεῦσι ∙ τοῖς ∙ μιαροῖς hṓσπερ τοῖς hε̄ρṓεσ(σ)ι ϝοῖνον hυπολhεί‑ ψας ∙ δι’ ὀρόφō∙ καὶ τᾶν μοιρᾶν ∙ τᾶν ἐνάταν ∙ κατακα‑ ίε̄ν ∙ μίαν. θυόντō θῦμα : καὶ καταγιζόντō hοῖς hοσία. καὶ περιρ(ρ)ά‑ ναντες καταλινάντō : κε̄̓́πειτα : τοῖς κ<α>θαροῖς : τέλεον θυόντō : μελίκρατα hυπο‑ λείβōν ∙ καὶ τράπεζαν καὶ κλίναν κε̄̓νβαλέτō καθαρὸν hε̃μα καὶ στεφά‑ 15 νōς ἐλαίας καὶ μελίκρατα ἐν καιναῖς ποτε̄ρίδε̣[σ(σ)]ι καὶ : πλάσματα καὶ κρᾶ κἀπ‑ αρξάμενοι κατακαάντō καὶ καταλινάντō τ̣ὰς ποτε̄ρίδας ἐνθέντες.
The sacrifice of offerings before the Kotytia and the truce on the fifth year (i.e. every four years) when the Olympiad also takes place; sacrifice to Zeus Eumenes an animal and to the Eumenides an adult animal and to Zeus Milichios in (the sanctuary?) of Myskos an adult animal; to the (10) Polluted Tritopatreis as to the heroes, having poured wine through the roof, of the nine portions burn one; those to whom it is religiously permitted are to sacrifice an animal and perform the consecration (burning); having sprinkled around (with water?), let them anoint (the altar?), and immediately sacrifice an adult animal to the Pure Tritopatreis; pouring down honey‑and‑milk‑mixture, a table is set out and a couch and let them throw over a pure cloth and wreaths (15) and honey‑and‑milk‑mixture in new cups and cakes and meat. Having made first‑offerings, they are to burn them and perform an anointment, having put the cups on (the altar?) (I.dial. Sicile II, no. 18; cf. Greek Ritual Norms, no. 13, col. A, l. 7‑16 [transl. Greek Ritual Norms]).
31Εὐμενίδεσ(σ)ι and ἡρώεσ(σ)ι occur in the same long sentence, which contains provisions for the sacrifices to be made to the most important gods of the Selinuntine pantheon. A certain solemnity is achieved through the opening τῷ Διὶ τῷ Εὐμενεῖ, which is reminiscent of the hymnodic style, and the repetition of divine names and their titles. Some rhythmic effects may also be part of the solemnity of these lines. Εὐμενίδεσσι – interestingly isolated between two punctuation marks in the text – makes up a perfect adonius, while the whole Εὐμενίδεσσι τέλεον corresponds to a hemiepes assuming that the ‑εο‑ of τέλεον was pronounced as one syllable. Similarly, the dative ἡρώεσσι occurs in a sequence that happens to coincide with the collocation of ἡρώεσσιν in the last two feet of epic hexameters.55 ἡρώεσσι also occurs in the Petelia Orphic leaf (Pugliese Carratelli 2001, no. I, A, l. 1), in the Dodona text from Apollonia (Greek Ritual Norms, no. 40), and in a few stone epigrams, while the standard form ἥρωσι is rarer in poetic texts (Pindar, Aeschylus, Aristophanes).
32This is not to say that the lex sacra is a piece of Kunstprosa, nor that, in exploring its style, we should apply the same rules applied to imperial prose, with its careful selection of admissible and forbidden rhythmic cola (including the rule that a colon must end with a long syllable).56 More simply, I suggest that the ‑εσσι allomorph – which perhaps had already started spreading in 5th‑c. Sicilian Doric under the pressure of morphological analogy, as posited in § 4) – was selected also because of the “poetic ring” that it had.57 This would not be out of context in a sacred text that spells out a series of prescriptions for “a recurrent cycle of sacrificial rituals” (Carbon, Peels in Greek Ritual Norms, no. 13) which must have been actually performed.58 The performative aspect of the ritual thus may have influenced some of the linguistic choices of the text: Εὐμενίδεσσι and ἡρώεσσι, gesturing towards poetic style, could serve to heighten the section devoted to the ritual and its deities.59
33Let us consider a parallel from the same text, the ἀθανάτοισι in l. 13 of col. B:
hόκα το͂ι ἐλαστέρōι χρε̄<ί>ζε̄ι θύε̄ν, θύε̄ν hṓσπερ τοῖς vacat
ἀθανάτοισι. σφαζέτō δ’ ἐς γᾶν. vacat
Whenever one needs to sacrifice to the elasteros, sacrifice as to the immortals. But one shall slaughter down towards the earth (I.dial. Sicile II, no. 18; cf. Greek Ritual Norms, no. 13, col. B, l. 12‑13 [transl. Greek Ritual Norms]).
34‑οισι is not the normal dative plural ending of thematic stems in the majority of Doric dialects.60 Mimbrera Olarte 2012a, p. 137, dismisses it as a Megarian trait of the Selinuntine dialect, but this is unlikely since we have no other instances of “long” datives in the texts from Selinous.61 On the contrary, here too the “long” dative is a marked choice, which accompanies a lexeme endowed with a poetic pedigree. ἀθανάτοισι also occurs in a sequence (θύειν ὧσπερ τοῖς ἀθανάτοισι) which may evoke the rhythm of the second half of a hexametric line.
35A similar usage of an ‑οισι dative occurs in a judicial defixio from Lilybaeum dated to the 3rd c. BC where, moreover, we find another ‑εσσι dative, Τιτάνεσσι:
καταδέω Ζωπυρίωνα τᾶς Μυμβυρ παρὰ Φερσε‑
φόναι καὶ παρὰ Τιτάνεσσι καταχθονίοις καὶ παρ’ ἀ‑
π[ε]υ̣χομένοισι νεκρ̣οῖς· vac. <καταδέω δέ νιν> {ἐς τοὺς ἀτελέστους} καὶ παρ‑
ὰ [ἱ]αρίαις Δάματρος <καὶ> παρ’ ἀπευχομέ[ν]α[ισ]ιν· vacat
I bind down Zopyrion son of Mymbyr before Persephone and before (the) underground Titans and before (the) abominating (male) dead. <And I bind him down> also before (the) priestesses of Demeter and before (the) abominating (sc. female dead) (SEG XLVII, no. 1442; text after Jordan 1997; cf. SEG XLIX, no. 1301, part A, l. 1‑4; judicial defixio from Lilybaeum [transl. Jordan 1997, p. 396]).
36Τιτάνεσσι, used in the opening malediction, is the only mention of Titans – rarely found outside cosmogonic poetry – in a defixio.62 Their unique occurrence here, together with Persephone, in the specific form Τιτάνεσσι that is paralleled by the Ionic Τιτήνεσσι of Hesiodic poetry (Hes., Th., 650, 674 and 882), seems to be due to the search for a certain stylistic effect. Jordan 1997, p. 393, finds these “formulaic phrases […] compatible with dactylic poetry, which may well have been their origin”.63 Dactylic “pseudo‑hexametric tesserae” could be the sequence ‑νεσσι καταχθονίοις (which corresponds to a “masculine” hemiepes) and the last clause of the malediction formula, καὶ παρ’ ἀπευχομένοισι νεκροῖς which, with νε.κροῖς syllabification, would correspond to the hexametric part preceding the hephthemimeral caesura: also note that <καὶ> παρ’ ἀπευχομέναισιν of l. 4 corresponds to a “feminine” hemiepes.64 The poetic flavour of the invocation of chthonic deities in this defixio could be part of the “professional” skills of the author who, if coinciding with the person who materially incided the neat layout of the text, was certainly an expert and not an amateur defigens.65
37A similar hexametric tessera (καταχθονίοισι θεοῖσι) is inserted in a slightly later (late 3rd-early 2nd c. BC) defixio from Lilybaeum which curses a series of individuals by appealing to the chtonic gods:
καταδίδημι παρὰ καταχθονίοισι θεοῖσι
τὰν πρᾶξιν τὰν Απιθαμβ.αλ ποτὶ Νυμήριον
καὶ Δαμ[έ]αν, ὅπως [μ]ὴ δύναται ἀντία
λέγειν
I bind down before the chthonic gods the legal action of Apithamb.al against Numerius and Dameas, lest he be able to speak in opposition, lest he be able to speak in opposition to any legal action or to hate (SEG XLVII, no. 1443; text after Jordan 1997; cf. SEG XLIX, no. 1302, side A, l. 1‑3; judicial defixio from Lilybaeum [transl. Jordan 1997, p. 390]).
38The hypothesis that a stylistic effect lies behind the use of ‑εσσι datives in the Selinous lex sacra now needs to be tested against the third attestation, ποτηρίδεσσι, partly restored at l. 15. This form is trickier because it identifies a common object in a sentence which does not conjure up any poetic images (though note that the feminine ποτηρίς is a hapax, one of the several striking forms of this text).66 However, just like the initial sentence which we have analysed, l. 14‑16 have a central role in the description of the ritual. These lines refer to the preparations for the entertainment of supernatural guests (theoxenia), which are likely to have been performed during a public ritual. According to Jameson, Jordan, Kotansky 1993, p. 68, the reference to new cups (ἐν καιναῖς ποτηρίδεσσι), though “consistent with the general practice of using clean and special clothing”, has a special significance in this passage “because the rite for the pure (καθαροί) Tritopatores follows on those for the impure (μιαροί) Tritopatores”. The choice of the ‑εσσι dative could therefore be integral to a wish to underline the importance of this passage and confer a certain “grandeur” to it.
4.3. The ‑εσσι datives of Sicilian classical inscriptions reconsidered: the Gela defixio
39This speculative analysis of the Selinuntine text may receive further support from a closer consideration of the datives ἄνδρεσι and γυναίκεσσι which occur in a defixio from Gela (I.dial. Sicile I, no. 134.b). The text, which I shall approach following the interpretation of Wilson, Favi 2017, contains an imprecation by a certain Αpellis against the rivals of one Eunikos, who is either Apellis’ lover or a close friend. Here I will focus only on two passages of the text in lines 1 and 11‑14:
l. 1 ε̣ὐχά. Ἄπελλις ἐπὶ φιλότατι τᾶι Εὐνίϙō. l. 11 με̄δέν’ Εὐνίκō σπευδαιό‑ [τερο]ν γενέσθαι, με̄́τ’ ἄνδρεσ(σ)ι με̄́τε γυναίκεσσι· ὁ̃ς οὗτōς βόλιμος τṑς τε̄‑ [νōν ? ἐπ]ο⟨ι⟩δ{ι}αὶ τίμαν ἐρύσαιντο. Εὐνίκōι ἀὲ νίκαν παντε<ῖ> ΡΜΟΑΥ…ἐπ‑ [ὶ φιλ]ότᾱτι τᾶι Εὐνίκō γάρφō.
Evil prayer: Apellis, for love of Eunikos […]. May none be taken more seriously than Eunikos either among men or women. As this lead (sc. does), so may the spells against those men draw away their honour. For Eunikos may there be victory always […]. I write on the lead for love of Eunikos (I.dial. Sicile I, no. 134.b; cf. SEG LVII, no. 905.b; text here slightly adapted following Wilson, Favi 2017; defixio from Gela [transl. slightly adapted from Csapo, Wilson 2020, pp. 347‑348]).
40At l. 12 the sequence μήτ’ ἄνδρεσ(σ)ι μήτε γυναίκεσσι is used to express the author’s wish that nobody may be more successful (σπευδαιότερον) than Eunikos. In Od., XIII, 308 the same mention of men and women, in the genitive, is equally preceded by a middle‑passive infinitive (μηδέ τῳ ἐκφάσθαι μήτ’ ἀνδρῶν μήτε γυναικῶν, “do not tell anyone, either man or woman, etc.”).67 The choice of an ‑εσσι dative may attempt to introduce some form of heightening in an expression with anaphora of μήτε that is not very common: there is only one parallel in the accusative (μήτ’ ἄνδρα μήτε γυναῖκα, followed by μήτε παῖδα μήτε κόραν) in a decree from Delphi, dated to 425‑375 BC (F.Delphes III, 1, no. 194, col. i, l. 11), though the positive ἀνδράσι καὶ γυναιξί is attested a few times in prose texts.68
41It may be objected that register heightening is unexpected in a defixio. The structure, mise en page, topic and language, however, identify this text as unique.69 The lexicon in particular abounds in seemingly recherché choices (e.g. the hapax ἀπρακτία, l. 5). The very first word, εὐχά (if the reading is correct), does not have its usual meaning, ‘prayer’, but refers to the defixio itself. This understanding of εὐχή finds some parallels in tragedy, while in defixiones it occurs in a long 4th‑c. AD text from Cos, probably written by a professional.70
42The optative ἐρύσαιντo of l. 13, referring to the “drawing away” of the honour of Eunikos’ opponents, is another remarkable choice.71 ἐρύω is a very common verb in the Homeric epics, but otherwise scarcely attested in prose.72 It is usually employed to refer to a physical act of drawing something or someone out or away, while in the Gela defixio it would have a metaphorical meaning. This extraordinary usage may be accompanied by a quasi‑hexametric clausula, the sequence of a dactyl and a spondee in (τί)μαν ἐρύσαιντο. Both elements may be proof that the ending of this text was characterised by an increasingly grander register. To return to the datives μήτ’ ἄνδρεσ(σ)ι μήτε γυναίκεσσι, if viewed within the overall linguistic context of the defixio, they look like marked choices, which were consciously employed in an expression that hinted at epic language.
5. A look at Hellenistic texts
43The above analysis has attempted to prove a connection between the use of the early ‑εσσι datives from Sicily and the intention to stylistically mark certain textual passages where the lexicon, morphological choices and perhaps prosody are also noteworthy. ‑εσσι datives are also attested in Hellenistic Sicilian inscriptions and it would therefore be useful to extend the analysis proposed for the classical evidence to later texts. Here I shall offer some preliminary observations in preparation for a full study of the evidence.
44The Hellenistic and early Roman evidence (no. 6‑23 in tab. 1) testifies to the full spread of analogical ‑εσσι to athematic stems, which now include vocalic stems (ἱππέεσσι, πολίεσσι, παναγυρίεσσι) and participles (συνοικιξάντεσσι, περιαγγελλόντεσσι, παθόντεσσι). Studies of the Sicilian Doric Koina approach ‑εσσι datives as a Syracusan feature that spread to the Greek of the rest of the island because of Syracuse’s prestige.73 However, as pointed out in § 4, none of the eleven secure forms from the Hellenistic Age come from Syracuse, and attestations crop up in places as different as Entella, Camarina, Gela and Halaesa. The later attestations are limited to the formulaic inscriptions addressed to Anna and the Paides from Acrae, a self‑defined and very standardised set of texts. It is probable that – whatever their origin – these datives became established in the Doric dialect of central‑western Sicily prior to the rise of Syracusan power.
45Even at this later stage the ‑εσσι datives may not necessarily have been the only morphological variant available to speakers. If this is correct, the regular ‑σι datives attested in post‑classical inscriptions (tab. 2) should not necessarily – or at least not always – be dismissed as “Koine features” of no relevance for the development of Sicilian Greek as a whole. Out of the eleven Hellenistic attestations of ‑εσσι datives, one (in the Lilybaeum defixio which we have already discussed, no. 14) is influenced by poetic language; another (in a contract from Camarina, no. 15) is a very problematic reading; and the remaining nine are all found in official decrees in Doric Koina with a few Koine intrusions. These are all honorary and celebrative decrees, whose language is often marked by stylistic heightening.
46One case is particularly telling: the Entella decree I.dial. Sicile I, no. 205 commemorating Entella’s symmachia with Gela/Phintias (no. 6 in tab. 1). The decree consists of one very long sentence, written in formal and lofty‑sounding prose. Here I shall focus on l. 5‑16, which contain the causal clause explaining the motivations of the honorary decree for the citizens of Gela (the decision of the city occurs from l. 16 onwards):
5 ἐπειδὴ οἱ Γελῶιοι ὅκα τε τὸ πρότερον τὰν πόλιν ταύταν ὠι‑ κέομες πολέμου κατασταθέν‑ τος ποτὶ Καρχαδονίους ἐβοαθό‑ ησαν τᾶι πόλει τᾶι ἁμᾶι ἱππέεσ‑ 10 σι καὶ πεζοῖς ἐνόρκιοί τε ἐγένον‑ το καὶ συνεβάλοντο συμμαχίαν, ὁμοίως δὲ καὶ νῦν ἐξ οὗ τὰμ πό‑ λιν οἰκέομες ἐς τὸ ἐμφανὲς ποι‑ οῦντι ὅτι μέμνανται τᾶς φιλίας 15 καὶ εὐνοίας τᾶς ποτὶ ἁμέ, δεδό‑ χθαι ἀναθέμειν εἰς χάλκωμα
Since the Geloans, when previously inhabiting this city, on the occasion of a war against the Carthaginians, came to the aid of our city with knights and foot soldiers, and were faithful to the oath and entered into a military alliance [with us], and similarly even now, since we began to inhabit the city, they make it clear that they are mindful of their friendship and goodwill towards us, let us approve (inscribing it) into a bronze plate (I.dial. Sicile I, no. 205, l. 5‑15; Decree from Entella [personal transl.]).
47The ‑εσσι dative (l. 9‑10) occurs in the long causal clause placed at the beginning: the expression ἱππέεσσι καὶ πεζοῖς ἐνόρκιοί τε ἐγένοντο recalls how the Geloans came to Entella’s rescue. That ἱππέεσσι may be a marked form is suggested, in my opinion, by the fact that it occurs in a phrase where the use of postpositive τε seems to be integral to the creation of a hexametric clausula (‑οί τ(ε) ἐγένοντο), which finds several parallels in Homer.74 In conclusion, it may be that ‑εσσι datives gradually acquired the status of markers of high‑register formal language and came to be employed in passages characterised by a solemn or grandiloquent flavour. In passing, it may be noted that the formulaic expression Παίδεσσι καὶ Ἄννᾳ of the Acrae dedications perhaps feature an ‑εσσι dative not merely because this was a prestigious element, but also to create a rhythmic effect (the sequence ‑εσσι καὶ Ἄννᾳ fills in a hexametric clausula again).
6. Conclusions
48At present we have no evidence to claim that the ‑εσσι datives were imported into Sicily from a particular dialect of continental Greece (see supra § 4). Given the gradual appearance of this allomorph in a variety of Greek dialects, it is likely that in Sicilian Doric it originated independently out of morphological analogy.75 The earliest Sicilian evidence (tab. 1) dates to the beginning of the 5th c. BC and comes from three different Doric areas (Selinous, Gela, and Syracuse – the latter only represented in literary texts). The scanty attestations of regular ‑σι datives in archaic and classical documents (tab. 2), though most of them πᾶσι, may evidence a situation in which the ‑εσσι datives were an “option” for Sicilian speakers of the classical period but not the only available dative formation for athematic stems: this would also be confirmed by the occurrence of Τριτοπατρεῦσι in the lex sacra side by side with three ‑εσσι datives.
49At the same time, as argued in §§ 4.2 and 4.3, some ‑εσσι datives in written documents may represent borrowings from poetic diction, selected to confer a poetic ring to certain passages. This is not to say that the ‑εσσι datives in Sicilian texts are true Aeolisms (and hence that they are the result of dialect contact), but more simply that literary practice may have affected the material insertion of certain ‑εσσι forms in specific parts of a given text. Speakers would not mark out an ‑εσσι dative as “strange” or “artificial”, because analogy made them morphologically transparent. Yet, at an initial stage, their early usage in Sicilian texts may have been encouraged by considerations pertaining to register and style.
50This special flavour of ‑εσσι datives may have continued in the Hellenistic Age, especially in those texts employing other high‑register features, at least judging from the type of text where they occur. This does not rule out that ‑εσσι datives may have spread to more informal registers, but we simply have no evidence for this: all the post-classical occurrences of ‑εσσι are in honorifics, decrees and dedications (Τιτάνεσσι of the Lilybaeum defixio is likely to be influenced by poetry, as argued in § 4.2).76 The line of interpretation proposed here should be extended to other dialectal areas and consider the interplay between ‑εσσι and ‑οις, another analogical ending which arose because of morphological concerns. ‑οις, however, is much more sporadic than ‑εσσι: a distribution which may be due precisely to the literary pedigree of ‑εσσι and its higher status.77
51It is not by chance that ‑εσσι was singled out as a marker of literary and epigraphic Doric prose. This artificial and dialectally mixed language contains other features that, taken together, do not belong to any real Doric variety and which must have been selected because of their literary appeal, probably through the influence of choral lyric.78 One of them is the feminine participle ending ‑οισα which, far from being an ancient Doric trait, is an Aeolism which came to be associated with poetry in “Doric” (Alcman, Pindar, Theocritus) and as such was exported into the language of Doric prose (Archytas, Archimedes).79
52These concluding notes suggest what direction research should take in the future. We need to study the role of ‑εσσι in the construction of literary and official prose languages in Doric (Koinai) in competition with Attic and Ionic. Together with comprehensive and up‑to‑date lists of dialectal attestations, the context of use of each dative must be assessed with an eye to elements such as style and heightening, poetic allusions, and prosodic effects.
Bibliographie
Abbreviations
Abbreviations of editions and of works of reference for alphabetic Greek epigraphy are those of the list of the AIEGL published online: GrEpiAbbr, version January 2022, https://www.aiegl.org/grepiabbr.html (accessed 06/07/2023) and abbreviations of ancient authors and works of the Diccionario griego-español, http://dge.cchs.csic.es/lst/lst4.htm (accessed 06/07/2023).
LSJ: H.G. Liddell, R. Scott, H.S. Jones, A Greek-English lexicon, Oxford, Clarendon Press, 1996 (9th ed. with a revised supplement).
TLG: M.C. Pantelia (dir.), Thesaurus Linguae Graecae Digital Library, http://www.tlg.uci.edu (accessed 10/06/2021).
References
Alonso Déniz 2022a: A. Alonso Déniz, “La copie de l’hymne à Asclépios d’Apollonia d’Illyrie (SEG 65 397)”, Mnemosyne 75, 2022, pp. 441‑466.
Alonso Déniz 2022b: A. Alonso Déniz, “Une innovation divine: l’origine de l’accusatif dorien Ποτειδᾶ / Ποσειδᾶ, attique Ποσειδῶ”, IF 127, 2022, pp. 151‑167.
Arangio Ruiz, Olivieri 1925: V. Arangio Ruiz, A. Olivieri, Inscriptiones Graecae Siciliae et infimae Italiae ad ius pertinentes, Milan, Udalrici Hoepli, 1925.
del Barrio Vega 2021: M.L. del Barrio Vega, “The dative plural ‑εσσι in the Saronic dialects and their colonies: a new analysis of the epigraphical data”, ZPE 220, 2021, pp. 163‑172.
Bechtold, Brugnone 1997: B. Bechtold, A. Brugnone, “Novità epigrafiche da Lilibeo: la tomba 186 della via Berta”, in Atti delle seconde giornate di studio sull’area elima, vol. I, Pisa, Gibellina, 1997, pp. 111‑140.
van Beek 2018: L. van Beek, Dialect borrowing versus internal development in Homeric Greek: reconsidering the dative plural in ‑εσσι, handout of a paper delivered at the conference “Language contact and variation in ancient Indo‑European languages”, Oxford, 17‑18 May 2018 (unpublished).
Bettarini 2005: L. Bettarini, “Una nuova defixio di Selinunte?”, ZPE 151, 2005, pp. 253‑258.
Bopp 1833: F. Bopp, Vergleichende Grammatik des Sanskrit, Zend, Griechischen, Lateinischen, Litthauischen, Gothischen und Deutschen, Berlin, Königliche-Preußische Akademie der Wissenschaften zu Berlin, 1833 (1st ed.).
Brugmann, Thumb 1913: K. Brugmann, A. Thumb, Griechische Grammatik, Munich, Beck, 1913 (4th ed.).
Buck 1905: C.D. Buck, “Notes on certain forms of the Greek dialects”, CR 19, 1905, pp. 242‑250.
Buck 1955: C.D. Buck, The Greek dialects. Grammar, selected inscriptions, glossary, Chicago, Chicago University Press, 1955.
Cabanes 2013: P. Cabanes, “L’hymne à Asklépios à Apollonia d’Illyrie”, in D. Lauritzen, M. Tardieu (dir.), Le voyage des légendes: hommages à Pierre Chuvin, Paris, CNRS Éditions, 2013, pp. 43‑55.
Capano 2020: M. Capano, Il greco di Sicilia tra età ellenistica e tarda antichità: risultati dell’analisi di un corpus epigrafico, PhD, Università degli Studi di Napoli “L’Orientale”, 2020 (unpublished).
Cassio 1989: A.C. Cassio, “Sviluppo della prosa dorica e tradizioni occidentali della retorica greca”, in A.C. Cassio, D. Musti (dir.), Tra Sicilia e Magna Grecia: aspetti di interazione culturale nel IV sec. a.C. Atti del convegno (Napoli, 19‑20 marzo 1987), Naples, Edizione dell’Ateneo, 1989, pp. 137‑156.
Cassio 2018: A.C. Cassio, “Notes on the origin and diffusion of the ‑εσσι datives”, in G. Giannakis, E. Crespo, P. Filos (dir.), Studies in ancient Greek dialects: from central Greece to the Black Sea, Thessaloniki, De Gruyter, 2018, pp. 189‑196.
Chantraine 1958: P. Chantraine, Grammaire homérique, vol. I, Paris, Klincksieck, 1958.
Csapo, Wilson 2020: E. Csapo, P. Wilson, A social and economic history of the theatre to 300 BC, vol. II, Theatre beyond Athens: documents with translation and commentary, Cambridge, Cambridge University Press, 2020.
Curbera 1994: J.B. Curbera, “Consideraciones sobre el dórico de Sicilia”, in Actas del VIII Congreso Español de Estudios Clásicos. Madrid, 23 al 28 de septiembre de 1991, vol. I, Madrid, Edicones Clásicas, pp. 93‑100.
Dimartino 2005: A. Dimartino, “Siracusa. A: fonti epigrafiche”, in M.I. Gulletta, C. Cassanelli (dir.), Bibliografia topografica della colonizzazione greca in Italia e nelle isole tirreniche, vol. XIX, Siti: Siracusa-Surbo, Pisa/Rome/Naples, Scuola normale superiore/École française de Rome/Centre J. Bérard, 2005, pp. 59‑128.
Dimartino 2015: A. Dimartino, “La lex sacra di Selinunte: analisi paleografica e prospettive storico‑religiose di una laminetta iscritta”, in A. Iannucci, F.M. Muccioli, M. Zaccarini (dir.), La città inquieta: Selinunte tra lex sacra e defixiones, Milan, Mimesis, 2015, pp. 135‑163.
Dover 1997: K. Dover, The evolution of Greek prose style, Oxford, Oxford University Press, 1997.
Drew-Bear 1972: T. Drew-Bear, “Imprecations from Kourion”, BASP 9, 1972, pp. 85‑107.
Dubois 1999: L. Dubois, “La nouvelle loi sacrée de Sélinonte”, in A.C. Cassio (dir.) KATÀ DIÁLEKTON: Atti del III Colloquio Internazionale di Dialettologia Greca (Napoli-Fiaiano d’Ischia, 25‑28 settembre 1996), Naples, Istituto Universitario Orientale, 1999, pp. 331‑346.
Ferrari 2011: F. Ferrari, “Oral bricolage and ritual context in the golden tablets”, in A.P.M.H. Lardinois, J.H. Blok, M.G.M. van der Poel (dir.), Sacred words: orality, literacy and religion in the Ancient World, Leiden/Boston, Brill, 2011, pp. 205‑216.
Gallavotti 1975‑1976: C. Gallavotti, “Scritture arcaiche della Sicilia e di Rodi”, Helikon 15‑16, 1975‑1976, pp. 70‑117.
García Ramón 1975: J.L. García Ramón, Les origines postmycéniennes du groupe dialectal éolien, Salamanca, Universidad de Salamanca, 1975.
García Ramón 1990: J.L. García Ramón, “Proportionale Analogie im Griechischen: der Dativ Pluralis der 3. Deklination in den aiolischen und westgriechischen Dialekten”, Glotta 68, 1990, pp. 133‑156.
Guarducci 1964‑1965: M. Guarducci, “Gli alfabeti arcaici della Sicilia”, Kokalos 10‑11, 1964‑1965, pp. 465‑480.
Guarducci 1986‑1988: M. Guarducci, “Epigrafi arcaiche di Siracusa e di Megara Iblea”, ArchClass 38‑40, 1986‑1988, pp. 1‑26.
Hallof 2021: K. Hallof, “Alte und neue Inschriften aus Olympia III”, Chiron 51, 2021, pp. 99‑122.
Hordern 2004: J.H. Hordern, Sophron’s mimes. Text, translation, and commentary, Oxford, Oxford University Press, 2004.
Hutchinson 2018: G.O. Hutchinson, Plutarch’s rhythmic prose, Oxford, Oxford University Press, 2018.
Jameson, Jordan, Kotansky 1993: M.H. Jameson, D.R. Jordan, R.D. Kotansky, A lex sacra from Selinous, Durham (NC), Duke University, 1993.
Jordan 1985: D.R. Jordan, “A survey of Greek defixiones not included in the special corpora”, GRBS 26, 1985, pp. 151‑197.
Jordan 1997: D.R. Jordan, “Two curse tablets from Lilybaeum”, GRBS 38, 1997, pp. 387‑396.
Jordan 2001: D.R. Jordan, “New Greek curse tablets (1985-2000)”, GRBS 41, 2001, pp. 5‑46.
Jordan 2007: D.R. Jordan, “An opistographic lead tablet from Sicily”, in P. Wilson (dir.), The Greek theatre and festivals: documentary studies, Oxford, Oxford University Press, 2007, pp. 335‑350.
Lazzeroni 1988: R. Lazzeroni, “Il dativo eolico in ‑εσσι: un caso di rianalisi”, AGI 73, 1988, pp. 12‑24.
Lupu 2005: E. Lupu, Greek sacred law: a collection of new documents (NGSL), Leiden, Brill, 2005.
Manganaro 2004: G. Manganaro, “Affitto di spazi pubblici per le panegyreis ad Akrai”, ZPE 147, 2004, pp. 115‑122.
Manni Piraino 1975: M.T. Manni Piraino, “Koiné alfabetica tra Siracusa, Megara Iblea e Selinunte?”, Kokalos 21, 1975, pp. 121‑153.
Méndez Dosuna 1985: J. Méndez Dosuna, Los dialectos dorios del noroeste, Salamanca, Ediciones Universidad de Salamanca, 1985.
Mimbrera Olarte 2012a: S. Mimbrera Olarte, Fonética y morfología del dorio de Sicilia (siglos VII‑I a.C.), Madrid, CSIC, 2012.
Mimbrera Olarte 2012b: S. Mimbrera Olarte, “The Sicilian Doric koina”, in O. Tribulato (dir.), Language and linguistic contact in ancient Sicily, Cambridge, Cambridge University Press, 2012, pp. 223‑250.
Minon 2018: S. Minon, “Dialectalisation et pseudoépigraphie philosophique: la dorisation de la koinè littéraire comme marque d’école dans les fragments transmis par Stobée du traité pseudo‑archytéen Περὶ νόμου καὶ δικαιοσύνης”, REG 131, 2018, pp. 1‑47.
Minon 2021: S. Minon, “La langue de la sentence des trois juges de Pellana: une koina diplomatique achéenne faiblement éléisée”, Chiron 51, 2021, pp. 123‑166.
Morpurgo Davies 1976: A. Morpurgo Davies, “The ‑εσσι datives, Aeolic ‑ss‑, and the Lesbian poets”, in A. Morpurgo Davies, W. Meid (dir.), Studies in Greek, Italic and Indo‑European linguistics offered to Leonard R. Palmer, Innsbruck, Institut für Sprachwissenschaft der Universität Innsbrucks, 1976, pp. 181‑197.
Nieto Izquierdo 2008: E. Nieto Izquierdo, Gramática de las inscripciones de la Argólide, Madrid, Universidad Complutense de Madrid, https://eprints.ucm.es/8475/ (accessed 21/07/2022).
Norden 1898: E. Norden, Die antike Kunstprosa vom VI. Jahrhundert v. Chr. bis in die Zeit der Renaissance, vol. I, Leipzig, Teubner, 1898.
Pugliese Carratelli 1956: G. Pugliese Carratelli, “Silloge delle epigrafi acrensi”, in L. Bernabò Brea (dir.) Akrai, Catania, Impr. “La cartotecnica”, 1956, pp. 151‑177.
Pugliese Carratelli 2001: G. Pugliese Carratelli, Le lamine d’oro orfiche: istruzioni per il viaggio oltremondano degli iniziati greci, Milan, Adelphi, 2001.
Rodríguez‑Noriega Guillén 2012: L. Rodríguez‑Noriega Guillén, “On Epicharmus’ literary and philosophical background”, in K. Bosher (dir.), Theatre outside Athens: drama in Greek Sicily and South Italy, Cambridge, Cambridge University Press, 2012, pp. 76‑96.
Ruijgh 1958: C. Ruijgh, “Les datifs pluriels grecques et la position du mycénien”, Mnemosyne 11, 1958, pp. 97‑116 (reprinted in Scripta minora ad linguam Graecam pertinentia, vol. 1, Amsterdam, Gieben, 1991, pp. 3‑22).
Sgarlata 1993: M. Sgarlata, La raccolta epigrafica e l’epistolario archeologico di Cesare Gaetani conte Della Torre, Palermo, Università degli studi di Palermo, 1993.
Thumb, Kieckers 1932: A. Thumb, E. Kieckers, Handbuch der griechischen Dialekte, vol. I, Heidelberg, Winter, 1932.
Vatri 2019: A. Vatri, “Early Greek dactylic prose in the history of Greek prose rhythm”, in E. Passa, O. Tribulato (dir.), The paths of Greek: literature, linguistics and epigraphy. Studies in honour of A.C. Cassio, Berlin/Boston, De Gruyter, 2019, pp. 175‑195.
Wackernagel 1903: J. Wackernagel, “Zur griechischen Nominalflexion: 2. Der Dativ pluralis auf ‑εσσι”, IF 14, 1903, pp. 373‑375 (reprinted in Kleine Schriften, vol. I‑II, Göttingen, Vandenhoeck & Ruprecht, 1969, pp. 967‑969).
Wathelet 1970: P. Wathelet, Les traits éoliens dans la langue de l’épopée grecque, Rome, Edizione dell’ Ateneo, 1970.
von Wilamowitz‑Moellendorff 1912: U. von Wilamowitz‑Moellendorff, Griechisches Lesebuch, vol. I/2, Berlin, Weidmann, 1912.
Willi 2008: A. Willi, Sikelismos: Sprache, Literatur und Gesellschaft im griechischen Sizilien (8.‑5. Jh. v. Chr.), Basel, Schwabe, 2008.
Wilson, Favi 2017: P. Wilson, F. Favi, “Choragic spells in Gela: a textual and exegetical note on Apellis’ defixio, SEG LVII 905b”, ZPE 204, 2017, pp. 138‑140.
Notes de bas de page
1 I thank Julián V. Méndez Dosuna and Alcorac Alonso Déniz for their careful remarks and suggestions, which have much improved the paper. I alone am responsible for the main hypothesis offered here.
2 I acknowledge my debt towards the overviews in van Beek 2018; Cassio 2018 for the purposes of this brief introduction.
3 Wackernagel 1903; followed by Wathelet 1970, p. 253; Morpurgo Davies 1976; García Ramón 1990.
4 Bopp 1833; followed by Brugmann, Thumb 1913, p. 239; Buck 1905, p. 248, Lazzeroni 1988; Cassio 2018.
5 Chantraine 1958, p. 204; García Ramón 1975, p. 84; Méndez Dosuna 1985, p. 481, n. 13. A detailed analysis according to stem class is provided in García Ramón 1990, pp. 138‑141.
6 The best discussion is in García Ramón 1990, pp. 148‑149. See too Wathelet 1970, p. 260; Dobias, Dialecte Cyrène, p. 101; Cassio 2018, p. 193; del Barrio Vega 2021.
7 See the table in García Ramón 1990, p. 135.
8 García Ramón 1975, p. 83, n. 3, disposes of all instances of ‑σι as not authentically dialectal.
9 Overview in Méndez Dosuna 1985, p. 481; García Ramón 1990, pp. 145‑146.
10 See García Ramón 1990, pp. 151‑155.
11 Analysis in García Ramón 1990, p. 151.
12 I thank Sophie Minon for pointing this out to me. She includes these datives among the supradialectal common traits which characterise a “variété d’achéen faiblement éléisée” used as a “koina diplomatique de registre élevé” (see Minon 2021, p. 145, for the datives and pp. 165‑166 for the language of the text).
13 A doubtful case is Εὐεσπερίδεσσι (SEG IX, no. 76; cf. IG Cyrenaica, no. 017000), possibly more a dental stem than an ‑εύς stem: see Ruijgh 1958, p. 101; García Ramón 1990, pp. 150‑151; Dobias, Dialecte Cyrène, p. 101.
14 Discussion in Nieto Izquierdo 2008, pp. 437‑438; del Barrio Vega 2021, p. 163.
15 See Nieto Izquierdo 2008, p. 442.
16 For this hypothesis, see e.g. Buck 1905, pp. 249‑250; Thumb, Kieckers 1932, p. 131; Wathelet 1970, p. 259 (North-West dialects); Lazzeroni 1988, p. 16. García Ramón 1975, pp. 115‑116, considers the possibility of an Aeolic substratum (due to early contact) only for Elean.
17 See Méndez Dosuna 1985, p. 482; García Ramón 1990; Nieto Izquierdo 2008, p. 438.
18 See Buck 1905, p. 250; Curbera 1994, p. 94; more cautiously, Mimbrera Olarte 2012a, p. 153, n. 85. None of these scholars provide a full list of attestations by chronological period.
19 As del Barrio Vega 2021, pp. 167 and 170‑171, now correctly points out.
20 García Ramón 1975, p. 116 assumes that Corinthian developed the ‑εσσι datives prior to the colonisation period.
21 We read the inscription in an early 18th‑c. copy: del Barrio Vega 2021, p. 164, considers the possibility that these ‑εσσι datives result from the wrong transcription of the original text.
22 Analysis in del Barrio Vega 2021, pp. 164‑166; cf. Alonso Déniz 2022a, p. 446, n. 28.
23 Ed. pr. Cabanes 2013: cf. SEG LXIII, no. 408.
24 É. Lhôte’s interpretation featured in an old version of the CIOD record “SEG LXIII (2013) 408” (originally accessed 06/07/2021) which has now been replaced by a new version (dated 03/06/2020) where this interpretation is no longer present (https://dodonaonline.files.wordpress.com/2020/07/ciod_seg_2013_408.pdf [accessed 20/10/2023]).
25 See Buck 1905, p. 250.
26 Pace García Ramón 1975, p. 84: “il ne s’agisse (sauf dans le cas des colonies corinthiennes) que de formes isolées” (italics are mine).
27 Occurrences in stone epigrams are rare and late: πραπίδεσσι in IG XIV, no. 14 (bilingual inscription with a Greek epigram from Syracuse, Roman period) and λαγόνεσ<σ>ιν in IG XIV, no. 124 (Christian epigram, Syracuse).
28 Mimbrera Olarte 2012a, p. 152.
29 See Sgarlata 1993, p. 130, with commentary on p. 172; Dimartino 2005, p. 94. The record in I.Sicily, no. 000830 is still incomplete.
30 Capano 2020, p. 129.
31 See del Barrio Vega 2021, p. 163.
32 Del Barrio Vega 2021, p. 171, proposes to consider the influence of Gela.
33 Later attestations, which are lexically more varied (e.g. δαίμοσι in an epitaph from Catania, I.Mus. Catania, no. 65, 1st‑2nd c. AD, entirely written in Ionic, or πένησι in a Jewish epitaph from Agrigentum, SEG XLIII, no. 618, 4th‑5th c. AD), are of little use because of the clear influence of Koine or poetic language in the inscriptions of Roman Sicily. Occurrences in undated (usually late) inscriptions are here excluded.
34 See Bettarini 2005 for the proposal of a different provenance, based on letter forms.
35 Mimbrera Olarte 2012a, p. 144. She adds that also in other areas where ‑εσσι is attested πάντεσσι is avoided: this is not quite correct, since we find πάντεσσι in Delphi, Boeotia, Thessaly, and Lesbos.
36 See Manni Piraino 1975, p. 141, with pl. XXX A.
37 Guarducci 1986‑1988, p. 15; followed by Lupu 2005, p. 342. L. Dubois (I.dial. Sicile II, p. 37), who attributes this reading only to E. Lupu, rejects it on account of the word order (“Je ne crois rien de cette lecture car pour une telle traduction on attendrait que le groupe ἀρὰ hάδε fût placé en tête”).
38 See M. Guarducci’s drawing and the better photo in LSAG2, pl. 77, no. 6.
39 Guarducci 1964‑1965, p. 474. She considers the graffito to be in the local dialect (see contra Manni Piraino 1975, p. 139).
40 See Gallavotti 1975‑1976, p. 89; Arena, Iscrizioni II2, no. 78.
41 The dative form υἱάσι is attested in Homer and later mostly in hexametric poetry. The common athematic form is υἱέσι.
42 A note may be useful on no. 10. This is a list of individuals assigned market stalls (other interpretations suggest plots of land or public spaces) which are defined through landmarks such as water tanks (ποτὶ φρητίοις) or other plots/stalls. The syntagm ἐν δ̣ρι<σ>ὶ Κακκαρικοῖς (IG) indicates one of such landmarks. The head noun here is δρίος ‘copse’, and this would be the only occurrence of its dative plural, which we would expect to be δρίεσι. In fact, the stone has δρίει and the singular reading is retained in Pugliese Carratelli 1956, p. 153, followed by Dubois (I.dial. Sicile I, no. 219) and Manganaro 2004, p. 117. Τhis would be at odds with the adjective Κακκαρικοῖς (IG) vel Κακκαβικοῖς (Pugliese Carratelli 1956), which is in the plural: G. Manganaro’s translation “nel boschetto delle pernici” does not register the syntactic problem. Perhaps δρίει should be corrected into δρίε<σ>ι.
43 See K.‑A., ad loc., Rodríguez‑Noriega Guillén 2012, p. 78. The dactylic rhythm with ποθεινός at the end of the line, though not Homeric (as ποθεινός is never attested in Homer), is employed in Callinus (frg. 1, 16) and in Pi., I., 5, 7 (ἔν τ᾿ ἀγωνίοις ἀέθλοισι ποθεινόν).
44 See too the analysis of del Barrio Vega 2021, pp. 169‑170.
45 That the description refers to Heracles is confirmed by the testimonium, Ath., X, 411a‑b (concerning Ἡρακλῆς ἁδηφάγος).
46 As a possibly technical term it is used in Hippocr., De affectionibus interioribus, VI, 16 (said of the abdomen).
47 ῥίνεσσι and its compounds are later used by Oppianus (C., I, 477 and IV, 357).
48 Norden 1898, p. 46.
49 Hordern 2004, p. 15.
50 On these recharacterised forms of ἥρως, see Alonso Déniz 2022b.
51 Von Wilamowitz‑Moellendorff 1912, p. 336; Hordern 2004, pp. 144‑145.
52 See the apparatus in Sophr., frg. **8 K.‑A.; Hordern 2004.
53 Though usually dated to the decade before 450 BC, the text is probably earlier (beginning of the 5th c.). See Dimartino 2015.
54 Only col. A, from l. 7 onwards, uses interpuncts (two dots or a simple dot). They do not mark only individual sentences, but smaller units, apparently following a logical – rather than syntactic or lexical – rationale. See Jameson, Jordan, Kotansky 1993, p. 47.
55 Cf. Il., II, 483 and 579, XVI, 144 = XIX, 391 and XXIII, 645. Homer never uses ἥρωσι.
56 For rhythm in early Greek prose, see the classic Norden 1898, pp. 41‑50, and the more cautious approach of Dover 1997, pp. 165‑177. Metrical or quasi-metrical sequences may not always be deliberate, and the quantitative or qualitative tests performed on such sequences to determine their non-randomness encounter several problems, which are discussed by Vatri 2019, pp. 183‑184. For post-classical prose, see Hutchinson 2018, pp. 5‑32.
57 Del Barrio Vega 2021, p. 168, advances a similar hypothesis.
58 As Vatri 2019, p. 178, notes: “metrical patterns are more likely to have been intentionally sought in texts that were meant to be recited to an audience rather than read privately (whether silently or aloud)”. For a loose parallel, see the case of the hexametric Orphic leaves, where the insertion of lyric sequences is integral to the performative nature of the text, as argued by Ferrari 2011, pp. 211‑215.
59 That this did not happen for Τριτοπατρεῦσι does not mean that the Tritopatreis are “less important” entities in the text, undeserving of stylistic marking. The regular dative Τριτοπατρεῦσι owes to the fact that at this chronological stage the analogical extension of ‑εσσι has not spread to vocalic stems yet. In other dialects too vocalic stems admit of ‑εσσι datives more slowly (if at all): see supra § 2; García Ramón 1990, p. 135.
60 According to Buck 1955, p. 88, “long” datives of this kind are only attested in Cretan and archaic Argive. Nieto Izquierdo 2008, pp. 423‑425, notices how the three ‑οισι datives from Argos and the one instance from Mycenae are also the most ancient datives attested for Argive and concludes that this must have been the only form in use in the 6th c. BC.
61 She also mentions the ‑οισι datives of the Lilybaeum defixiones (see here below) which, far from being authentic traits of Megarian (and why should they be? The texts date from the late 3rd c. BC), are more probably influenced by poetic diction.
62 See Brugnone in Bechtold, Brugnone 1997, p. 120.
63 The same interpretation in del Barrio Vega 2021, p. 168.
64 καταχθόνιος is not a frequent word in archaic and classical texts, but it already occurs in Homer (Il., IX, 457) as an epithet of Zeus in the context of Amyntor’s curse against Phoenix, where Persephone is also mentioned.
65 See Brugnone in Bechtold, Brugnone 1997, pp. 115 and 118. Poetic influence is also obvious in the opening formula consisting of four hexameters and invoking chthonic demons (δαίμονες [δέμονες] οἱ κατὰ γῆν καὶ δαίμονες οἵτινες ἔστε κτλ.) of several 3rd‑c. AD defixiones which then develop into prose (Audollent, Defixiones, no. 22, 25, 26, 28, and 30‑35, all from Amathous, Cyprus; more defixiones with the same formula are registered by Jordan 1985, p. 193; a variation on the formula occurs in text no. 52 of Jordan 2001, p. 17). The formula is likely to depend on the same magical poem, probably circulating on papyrus, and shows various Homeric expressions: see Drew-Bear 1972, pp. 87‑95, for a commentary.
66 See Dubois 1999, pp. 343‑346; Jameson, Jordan, Kotansky 1993, pp. 48‑49, for the language.
67 ἄνδρεσι may be a simple spelling mistake, with the omission of the geminate, or a (conscious?) compromise between the regular form ἀνδράσι and the analogical ἄνδρεσσι: see also del Barrio Vega 2021, p. 168. The prosody of γυναίκεσσι would not allow its insertion in the hexameter, a likely reason for the absence of this form from all poetic texts except Sappho (frg. 96, 6‑7). Interestingly, it is also not the preferred form in Aeolic inscriptions: see the datives γύναιξι καὶ παίδεσσι κόραις in IG XII, 2, no. 68 (Mytilene, 2nd c. AD). I thank Julián V. Méndez Dosuna for this reference.
68 See too ἄνδρεσσι καὶ γυναιξί in Pi., P., V, 64.
69 These will be explored in a paper in preparation.
70 See LSJ, s.v. εὐχή 3: ‘prayer for evil’, SEG XL, no. 1291 (εὐχὴ κατακλητικὴ κατὰ Ἑρμία). The alternative reading Τύχα, suggested by Jordan 2007, p. 344, is now rejected by Wilson, Favi 2017.
71 See Wilson, Favi 2017, p. 140; Csapo, Wilson 2020, p. 348.
72 An exception is F.Delphes III, 1, no. 294, col. ii, l. 12, an early 4th‑c. law where the expression τὸν ἀδικηθέντα ϝερυσάτω can be translated as “let him draw the culprit [into court]”.
73 Mimbrera Olarte 2012b, p. 233.
74 Cf. Il., III, 84: μάχης ἄνεῴ τ’ ἐγένοντο; Il., XXIV, 790: ὁμηγερέες τ’ ἐγένοντο#, a formulaic line‑end.
75 The same interpretation in del Barrio Vega 2021, p. 169.
76 Del Barrio Vega 2021, p. 169, concludes that at this stage these datives must have been a typical feature of Sicilian Doric.
77 I owe this suggestion to Sophie Minon. See also Minon 2018, p. 20.
78 For this analysis, against others that see in this dialectal mixture a real local variety, see Cassio 1989, p. 142; Minon 2018, pp. 7‑9 (specifically on pseudo-Pythagoric prose).
79 Cassio 1989, p. 148, n. 34, includes ‑οισα among the “forme palesemente artificiali” imported into the text of the Pythagoreans at a later stage, whereas Minon 2018, p. 7, interprets them as authentic features of poetic origin which characterised the literary Koina of Pythagoric writings.
Auteur

Le texte seul est utilisable sous licence Creative Commons - Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International - CC BY-NC-ND 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.
Dire la ville en grec aux époques antique et byzantine
Actes du colloque de Créteil, 10-11 juin 2016
Liliane Lopez-Rabatel, Virginie Mathé et Jean-Charles Moretti (dir.)
2020
Dérivation nominale et innovations dans les langues indo‑européennes anciennes
Alain Blanc et Isabelle Boehm (dir.)
2021
Stylistique et poétique de l'épigramme latine
Nouvelles études
Daniel Vallat et Florence Garambois-Vasquez (dir.)
2022
Contacts linguistiques en Grèce ancienne
Diachronie et synchronie
Alcorac Alonso Déniz, Julián V. Méndez Dosuna, Enrique Nieto Izquierdo et al. (dir.)
2024
Les représentations du désir féminin dans la poésie latine, de Catulle à Juvénal
Contraintes et libertés de l’espace poétique
Giacomo Dimaggio
2024
Définir l'épopée en Grèce ancienne
Contribution à une histoire de l'idée d'épopée d'Hérodote à Tzetzès
Flore Kimmel-Clauzet
2024