Διόσκουροι, Διόσκοροι, Διοσκόρω
(Phryn. Ecl. 205, [Hdn.] Philet. 44)

D. General commentary

An entry in Phrynichus’ Eclogue (A.1) discusses the correct phonology of the name of the Dioscuri: the lexicographer prescribes Διόσκοροι as ‘more correct’ (ὀρθότερον) and strongly disapproves of Διόσκουροι. The latter illustrates the outcome of the third compensatory lengthening, one of the most notable sound changes that set the Ionic dialect apart from Attic. Interestingly, an entry in the pseudo-Herodianic Philetaerus (A.2) draws a distinction between plural Διόσκουροι and dual τὼ Διοσκόρω: that is, the form without lengthening is associated with the dual number, which was seen as a typical Attic trait, while the plural form shows the non-Attic lengthening. In the Philetaerus, therefore, Attic phonology goes hand in hand with Attic morphosyntax. As a closer look at the literary and epigraphical sources will show, however, both lexicographers sought to deduce a uniform rule from a situation that was far too complicated to be captured by any single rule.

The two forms of the theonymTheonyms Διόσκο(υ)ροι, literally ‘Zeus’ sons’ (still written separately as Διὸς κούρους in h.Hom. 33.1; 9), differ according to the presence or absence of the third compensatory lengthening triggered by the loss of /w/ in the second compound member κόρϝος ‘boy, son’: Attic (and Western Ionic), where /w/ was lost without lengthening, have κόρος, while Central and Eastern Ionic, including the Homeric language, has κοῦρος. The feminine κόρϝᾱ ‘girl, daughter’ similarly gave Attic κόρη and Ionic κούρη. Although both the masculine and the feminine can be traced as far back as MycenaeanMycenaean language (ko-wo, ko-wa; see also Obrador-Cursach 2020, 277 on likely cognates in Phrygian), their frequencies in alphabetical Greek differ: κόρη/κούρη is the unmarked word for ‘girl, daughter’ in Attic-Ionic, while the masculine is rarer than the feminine, possibly because of the competition of synonymsSynonyms like παῖς or νεανίας (see DELG s.v. κόρος; EDG s.v. κόρη). Due to their use by Homer, the Ionic forms also enjoyed a level of prestige in Attic. Mahoney (2017, 48‒53) has recently studied the distribution of κο(υ)ρος and its compoundsCompounds, including Διόσκο(υ)ροι and ἐπίκουροςἐπίκουρος ‘helper, ally’ (which he argues to be a prepositional Rektionskompositum of κοῦρος) in Attic literary texts. In prose, κόρος and κόρη are the dominant forms, while κοῦρος and κούρη are only found in prose quotations of poetry. In tragedy, κόρος is much more frequent than κοῦρος: Euripides, the only dramatist whose plays feature a significant number of occurrences, uses κόρος 22 times and κοῦρος six times; moreover, the former is found in dialogues and the latter in lyric passages, suggesting that it was perceived as a (more) poeticPoetic language form.

The distribution of Διοσκο(υ)ρ- in the same period is quite different from that of the simplex noun. In both prose and poetry, ‘the Attic phonology (or at least the spelling convention: Διοσκορ‑) competed with the lengthened Ionic form, indicating the possibility that both forms were acceptable in Attic’ (Mahoney 2017, 50). The inscriptional evidence points in a similar direction: Threatte (1980, 219) remarks that ‘[t]he spelling of Διόσκουροι for the Classical Period must remain uncertain’; even in inscriptions like IG 2².1932.15 [400‒350 BCE] – which uses <Ω> for /ɔː/ while still using <ΟΥ> for /oː/ – the spelling ΔΙΟΣΚΟΡΩ could refer to both Διοσκόρω and -κούρω in the old Attic alphabet. When used as an anthroponymAnthroponyms, however, the form is always spelled ΔιόσκοροςΔιόσκορος; while ΔιοσκουρίδηςΔιοσκουρίδης was the standard spelling in Attica. Mahoney (2017, 53) further notes that ‘the manuscripts of our latest authors – Xenophon, Demosthenes, and Aristotle – have Διόσκουροι almost exclusively. Perhaps this indicates the date of the Ionic phonology’s final ascendancy in Attic’. In light of the Philetaerus’ prescription, it is also interesting to see how the distribution of lengthened and unlengthened forms overlaps with the distribution of dual and plural forms. EuripidesEuripides uses a majority of dual forms, sometimes with the lengthened stem (e.g. C.1), but more often with the unlengthened stem (Hel. 284, 720; Or. 465); his plays also contain four occurrences of the plural, mostly with the stem unlengthened (El. 1239; Hel. 1644, 1667), but with it lengthened in C.2 (interestingly, together with the use of the Ionic term δισσός to avoid the use of the dual; see Willi 2010, 110). Euripides’ eclectic distribution is further complicated by the existence of variae lectiones for some of these passages (see Mahoney 2017, 50 n. 48). In comedy, however, we find only dual forms, always with the unlengthened stem (Ar. Pax 285 = C.3, Ec. 1069, fr. 316; Amphis fr. 9.4; Men. Dysc. 192 = C.6; see F.2).

While we have noted the variants’ distribution in drama and poetry, this says nothing of their occurrence in prose, where several mismatches between phonology and morphology occur. Where Thucydides gives τῶν Διοσκόρων (C.4), this may be explained by his general avoidance of the dual, and although the varia lectio Διοσκούρων features in this passage, he uses the unlengthened stem elsewhere (4.110.1) in the derivative Διοσκόρειον. Plato, who systematically uses κόρος (28x), nonetheless oscillates between Διοσκούρων (Euthd. 293a.2) and Διοσκόρων (Lg. 769b.5). Xenophon adopts the lengthened stem for both plural (Διόσκουροι Smp. 8.29) and dual (Διοσκούροιν HG 6.3.6). Demosthenes uses the derivative Διοσκουρείου (19.158). And Aristotle, who, like Plato, prefers κόρ- (22x) for the simplex (κουρ- 5x, only in poetic quotations), uses the mixed form Διοσκόρους (but with the vv.ll. -κούρους and Διὸς κούρους) at EE 1246b.33. Aeneas Tacticus, a non-native of Attica who wrote in a language based on Attic that featured several elements from Ionic – in many ways a forerunner of the koine (see Vela Tejada 2018) –, uses the plural form Διόσκουροι twice (24.1 = C.5 and 24.13; on Aeneas’ avoidance of the dual form, see Vela Tejada 1991, 162).

Later authors of the Hellenistic and early imperial ages who avoided using the dual form ‒ an Attic trait that did not enter the koine ‒ could choose either the lengthened (Ionic) or unlengthened (Attic) stem of the word: Diodorus Siculus, for instance, always uses the latter (10x), while Plutarch (who does not completely shun the dual per se: see Vela Tejada 2019, 298) uses the former (12x) twice as much as the latter (6x). Incidentally, it should be noted that the transliterations of this theonym into LatinLatin may also attest to the diffusion of Διόσκοροι in the Hellenistic period (see C.8, F.3). After Atticism had revived the dual form, some authors began to attach the dual endings to the Ionic lengthened stem, producing the opposite kind of mixed form: Arrian, for instance, always uses the stem Διοσκουρ-, both in the dual (3x, all in the Anabasis of Alexander) and in the plural (3x, all in the Periplus of the Euxine Sea), as does Clement of Alexandria (dual 4x, plur. 2x); Διοσκούροιν is also found in the pseudo-Lucianic Charidemus (3.17). Interestingly, such mixed forms were not entirely unknown to classical Attic: Threatte (1996, 716) reports that Διοσκόροι[ν] (IG 1³.521bis.3, 427 BCE) probably stands for Διοσκούροιν, with <Ο> = /oː/. Among those authors who were influenced by Atticism, LucianLucian is interesting because his works show a distribution that closely mirrors the Philetaerus’ prescriptions while hinting at the colloquial status of the dual. In the Lucianic corpus, the plural always occurs (5x) with the stem Διοσκουρ-; the dual, however, occurs twice, without compensatory lengthening, in the oath ὦ Διοσκόρω (Gall. 20.4 = C.9, DMeretr. 14.4), which probably should be seen as an echo of Old Comedy (see F.2). Indeed, according to Lobeck (1820, 235‒6, followed by Rutherford 1881, 310‒1) the greater frequency of the dual in older texts is the reason why Attic vocalism in Διοσκόρω was preserved, even in late authors: Lobeck notes the case of Themistius (C.10), who, writing in the 4th century CE, still includes the nom.-acc. dual form τὼ Διοσκόρω among other Atticising expressions as examples of pretentious archaismsArchaisms used by false philosophers to take advantage of their audiences’ ignorance.

Although Atticist lexicographers generally recommended the use of dual forms, such as δυοῖνδύο (‘two’, gen.-dat. du.; see entry δυσί, δυοῖν, δυεῖν) or νώἐγώ (‘both of us’) (see La Roi 2022, 218–9 and entry ἀκολουθοῦντε), over the corresponding plural forms, in this case both numbers appear to have been admitted: Phrynichus does not discuss the dual forms of the name of the Dioscuri, but by assessing the correct vocalism of the plural he implicitly admits its use; the Philetaerus, however, prescribes different vocalisms for each number but does not proscribe either. Phrynichus’ judgement that the form without compensatory lengthening is ‘more correct’ in the plural may be based on analogy with the dual form, as well as on an awareness of the Ionic origin of the lengthened stem. We can contrast this with Moer. α 147 (B.1), where Διόσκοροι (in the plural, but with Attic vocalism) is attributed to the ‘users of Greek’ (i.e. the cultivated koine speakers of his time) and distinguished not from a phono-morphological variant but from a different lexical option entirely: ἌνακεςἌνακες, a specifically Attic cult epithet of the Dioscuri. While in the extant Atticist lexica the vocalism of the Dioscuri’s name is only discussed by Phrynichus and the Philetaerus (see F.1 for the possible relevance of this parallel), the formal variation in the simple noun κοῦρος/κόρος was discussed in other erudite sources (B.2; cf. also Epim.Hom. in Il. 1.10A.1a ~ 1.10A.2; EM 607.33‒6, which agree in explaining κοῦρος as an Ionic form, derived from κόρος by the addition (πλεονασμός) or the insertion (ἐπένθεσις) of υ, according to theory of the πάθηπάθη (i.e. phonetic changes) developed by ancient grammarians. It is worth noting that these sources, including the Atticists, all reason in terms of the presence or absence of the letter υ, disregarding the nature of <ου> as a ‘spurious diphthong’ denoting the lengthened counterpart of <ο>. The various descriptive and prescriptive statements about κοῦρος/κόρος and Διόσκο(υ)ροι are, therefore, at best ambiguous with regard to orthoepy and orthography. Vessella (2018, 34) considers the similar case of Phryn. Ecl. 22Phryn. Ecl. 22, πιοῦμαι σὺν τῷ υ λέγων οὐκ ὀρθῶς ἐρεῖς (‘Saying πιοῦμαι [‘I will drink’] with υ, you will speak incorrectly’), where ‘the contrast /o ~ u/ is only indicated by the presence or absence of the letter <υ>’ (further similar cases are in Phryn. Ecl. 28Phryn. Ecl. 28, contrasting ποῖ ‘whither?’ and ποῦ ‘where?’, and in the Atticist lexicon contained in P.Oxy. 15.1803, fol. 1, recto.35 (= TM 65081)P.Oxy. 15.1803, fol. 1, recto.35 (= TM 65081) συνεθίζεσθαι διὰ τοῦ ῑ, on which see Favi 2022, 323–7).

E. Byzantine and Modern Greek commentary

In texts from the Byzantine period, the name of the Dioscuri is always found in the plural, usually with the long and, less frequently, the short vocalism. The dual form, which had long since ceased to be part of the living language, is never found outside of quotations from ancient sources. The Modern Greek Διόσκουροι (both as a theonym and in the figurative meaning, ‘two people who are related by close friendship or work closely together’) is a learned word that demonstrates the non-Attic vocalism, possibly influenced by the use of Dioscuri (a pseudo-Latinism, see F.3) and similar forms in modern European languages (see LKN s.v.). Likewise, the Modern Greek κούρος is an archaeological and art-historical termTechnical language that refers to a free-standing statue of a male youth from the archaic period; this usage corresponds with that of the English kouros, the German Kuros, and similar forms in other modern languages, even though it was first proposed by a Greek scholar (Leonardos 1895, 75 n. 1). This modern usage parallels that of kore for archaic female statues, although the use of this latter term had precedents in antiquity (see Holloway 1992, 267). The Attic κόρη, ‘girl, daughter’, survives into Modern Greek, where it is has been borrowed from the learned language, a higher-register synonym of the popular form, το κορίτσι, originally a diminutive (see Méndez Dosuna 2016, 229 n. 1); nonetheless, the presence of κορίτσι itself and other popular forms like κορασιά, κοράσι, etc. (see DELG s.v. κόρος) shows that κόρη continued to be part of the spoken language long enough to give rise to such derivatives. With regard to κοράσι, it is noteworthy that its ancestor, κοράσιονκοράσιον, was mocked by the 4th-/3rd-century-BCE comic poet Philippides as a foreign import (fr. 37) and censored by the Atticists (Phryn. Ecl. 50Phryn. Ecl. 50, Poll. 2.17Poll. 2.17, Phot. π 26; see entry κόριον, κορίδιον, κορίσκη, κοράσιον), possibly because of its Macedonian origin (cf. schol. Hom. Il. 20.404c); it is, however, frequently found in inscriptions from Central Greece from around the 2nd century BCE (see Hartwig 2022, 283‒4).

F. Commentary on individual texts and occurrences

(1)    [Hdn.] Philet. 44 (A.2)

The fact that this entry in the Philetaerus is concerned with the same problem discussed by Phrynichus in the Eclogue (A.1), while giving a different prescription, might be seen as a point of contact between the two lexica. Argyle (1989)    identifies the author of the Philetaerus as CornelianusCornelianus, the dedicatee of Phrynichus’ Eclogue: she argues that the Philetaerus was written in response to the first part of the Eclogue, and that the second part of the Eclogue was, in turn, a response to this lexicon (see entry [Herodian], Φιλέταιρος (Philetaerus)). To bolster her claims, Argyle adduces several parallel passages between the two works: although the prescriptions about the name of the Dioscuri are not included in this list, theoretically, at least, they might point to another instance of disagreement between the two lexicographers. Argyle’s theory has, however, been criticised (see Alpers 1998, 103‒8; Ucciardello 2021, 61‒2) and the authorship of the Philetaerus remains obscure, although Alpers suggests that its author is the 2nd-century-CE grammarian Alexander of CotiaeumAlexander of Cotiaeum (Reitzenstein 1897, 377‒97 defends the attribution to Herodian, and we cannot exclude the possibility that the treatise contains some Herodianic doctrine; see also Dain 1954, 15). Fischer (1974, 43) argues that the Eclogue and the Philetaerus are quite independent of one another and that any parallel between the two merely results from the use of common sources, such as Aelius Dionysius and possibly other early Atticists. These diverging opinions about Διόσκο(υ)ροι in A.1 and A.2 do not, therefore, need to reflect any form of direct disagreement between the authors of these two lexica, since it is likely that such forms were widely discussed in Atticist circles.

(2)    Men. Dysc. 191‒2 (C.6)

This passage in the Dyscolus contains one of only two occurrences of the Dioscuri’s name in Menander; the other occurrence is τοῖν δυοῖν Διοσκόροιν (fr. 491 = C.7). Here, the Bodmer papyrus (P.Bodm. 4 [= TM 61594], 3rd–4th century CE) transmits διοσκουρω with the dual ending and the Ionic vocalism; while this form would not be metrically impossible in this position, following Sandbach (1968, 441), the editors have corrected it to Διοσκόρω, which is metrically guaranteed in all passages of Attic drama where this name appears – with the exception of Eur. IA 769 (C.2), a lyric passage with a non-Attic linguistic characterisation (see Gomme, Sandbach 1973, 166; Kassel, Schröder PCG vol. 6,1, 79; one should also refer to C.1 here, where the tradition is split between Διοσκόροιν and -κούροιν). The oath, ὦ Διοσκόρω, is relatively rare in comedy, occurring only twice in Aristophanes (C.3, Ec. 1069) and once in Amphis (fr. 9.4). As Papachrysostomou (2016, 73) observes, there may be certain thematic reasons for an invocation to the Dioscuri – the patron gods of Sparta – in the Aristophanic passages (on which, see also Olson 1998, 127‒8; Ussher 1973, 223) and perhaps in Amphis’ fragment from the Gynaikomania, a play whose plot likely had much in common with the Assemblywomen (see Papachrysostomou 2016, 66); however, there are no similar thematic reasons to account for such an invocation in Menander. Cartlidge (2014, 213‒5) observes that Menander uses the dual form ‘almost exclusively in oaths to pairs of deities’, a distribution that ‘does not indicate retention of the dual as a morphological category’. The fact that in fr. 491 (C.7) the dual is redundantly reinforced by the numeral δυοῖν also suggests that this linguistic feature was already moribund in Menander’s lifetime. Be that as it may, given the close association of the dual form with Attic vocalism that is observable in comedy, the dual was very probably a trait of colloquial registersColloquial language. Lucian’s reprise of the oath ὦ Διοσκόρω in two dialogic works (in contrast with his use of Διόσκουροι elsewhere) may be indebted to the language of Old Comedy (see C.9).

(3)    Cic. Nat.deor. 3.53 (C.8)

The Latinate form Dioscuri used in English and other modern languages, as if derived from the Ionic Διόσκουροι, does not in fact occur in classical Latin sources (see Edmunds 2016, 66‒7): if the plausible conjectural reading of Dioscoroe in this passage from CiceroCicero is correct, however, it would be a transliteration of the ‘mixed’ form Διόσκοροι. It is likely that Cicero would have chosen the variant that was most common in the cultivated Greek of his time. His testimony therefore constitutes additional evidence that this form, used by authors since Thucydides (C.4), later granted approval by Phrynichus, and assigned to the ‘users of Greek’ by Moeris, possessed a high status. At a much earlier date still, the Ionic vocalism is found in the archaic Latin QVROIS (dat. plur.) – either as a true borrowing or simply a transliteration of the Greek κούροις – which features in a dedication to Castor and Pollux (ILLRP 1271a [= TM 250662], Lavinium, 6th‒5th century BCE).