ἀνυπόδετος, ἀνυπόδητος
(Antiatt. α 118, Phryn. Ecl. 419, Phryn. PS 27.12, [Hdn.] Philet. 149, Moer. α 63, Poll. 2.199)

D. General commentary

The Atticist lexica deal with the two forms of the adjective ἀνυπόδετος/ἀνυπόδητος (‘barefoot’): while the Antiatticist (A.1) appears implicitly to defend the form with ε by mentioning its use by the comic poet Epicharmus (C.1), Phrynichus (A.2, A.3), the pseudo-Herodianic Philetaerus (A.4), Moeris (A.5), and Pollux (A.6) more or less explicitly defend the form with -η- as Attic, and hence correct. These entries were aimed at promoting one spelling – and probably one pronunciationPronunciation – over the other, at a time when η was still pronounced as a mid front vowel with a timbre similar to ε, but when the perception of vowel quantity had been lost; cf. Vessella (2018, 57–8; 79–80) and AGP vol. 2, Phonology, 2.2.2, forthcoming.

The adjective ἀνυπόδετος/ἀνυπόδητος is the negated verbal adjective of the verb ὑποδέωὑποδέω (‘to tie under (the feet)’, i.e. ‘to shoe’). Like δετός, the verbal adjective of δέω, it is formed on the zero grade of the verbal root *deh₁-, (< *dh₁-tos). Attested only late as a simplex, δετός is nevertheless found in many ancient compounds: cf. e.g. χρυσόδετος ‘overlaid with gold’ (Alc. fr. 350.2, referring to a sword’s hilt), μελανδετός ‘bound with black’ (Il. 15.713, referring to a sword). The regular formation of the adjective meaning ‘barefoot’ therefore requires the presence of ε in the root (ἀνυπόδετος). The form with η, by contrast, is an Attic innovation, arising from the assimilation of δέω to other verbs with a present in -έω but of different origin. The assimilation was facilitated by the aorist ἔδησα, which has a lengthened vowel, like other verbs in -έω (e.g. ποιέω – ἐποίησα, κινέω – ἐκίνησα). Since the verbal adjective of verbs in -έω regularly takes the form -ητός (cf. ποιητός, κινητός), the form ἀνυπόδητος was created in Attic by analogy, although it is not attested in other dialects. The presence of the form ἀνυπόδετος in Epicharmus (C.1) must not, therefore, be interpreted as a peculiar feature of DoricDoric dialect: it is rather the regular formation of the adjective, preserved in the Sicilian Doric variety used by the poet and subsequently transmitted to koine Greek; cf. Cassio (2012, 257–8). In addition to the assimilation to other -έω verbs, one might consider whether the long vowel in the Attic adjective was also influenced by the vocalism of the common noun ὑπόδημαὑπόδημα (‘sandal’). Indeed, ὑπόδημα (attested since Homer) follows a regular formation: in Archaic Greek, deverbative nouns in -μα are typically derived from the strong-stem allomorph in verbs exhibiting ablaut (in this case δη-/δε-). Alternatively, an unattested simplex form *δῆμα could represent the Greek reflex of PIE *deh₁-mṇ-. See Gunkel (2011, 79, 84).

After Epicharmus (C.1), ἀνυπόδετος is attested once in Aristotle (fr. 74 Rose, quoted by Macrobius; see below for the presence of ἀνυπόδητος in another passage of the Aristotelian corpus). A further occurrence may perhaps be found in Chrysippus, in a fragment (fr. 177 SVF) quoted by Simplicius, where the adjective appears twice; only the second instance shows the short vowel. It is impossible to determine either the philosopher’s exact wording or the form he himself actually employed. The form with the short vowel then occurs five times in the Septuagint: C.5, Is. 20.3 and 4, 2Re. 15.30, Mi. 1.8. Note that in C.5 the Codex Alexandrinus reports the variant with η (see Thackeray 1909, 80), which is also transmitted in the Septuagint version revised by Lucian of Antioch (3rd–4th c. CE), who adopts it in the two subsequent passages of the Book of Isaiah as well. Indeed, among the textual interventions characteristic of Lucian’s recension is the replacement of koine forms with Attic forms; cf. Fernández Marcos (2000, 228–9). It should be noted that, more generally, the process of Attic ‘normalisation’ in literary texts may have involved many more occurrences of the adjective than are currently identifiable. In epigraphic documents, ἀνυπόδετος is attested in IG 5,1.1390.15 [Andania in Messenia, 92–91 BCE], ID 2529.16 [116–115 BCE] and I.Lindos 487.8 (225 CE).

The form ἀνυπόδητος is instead attested in Aristophanes (C.2, Nu. 363), Plato (6x, cf. C.3), Xenophon (3x), Lysias (32.16), Aristophon (fr. 10.8), Aristotle (PA 687a.24), Theophrastus (Char. 18.4), and Theocritus (C.4; on this occurrence and the related scholium B.5, cf. F.1). All the epigraphic attestations of ἀνυπόδητος date from the post-classical period.

As noted above, the form with η is prescribed by all Atticist lexica, with the exception of the Antiatticist (A.1) – the Eclogue’s entry (A.2) is the only one to offer a linguistic explanation based on analogy with the aorist of the verb ὑποδέω. The same prescription is also found in late antique and Byzantine lexica and etymologica that depend on Atticist erudition (B.1, B.2, B.3, B.4). From an Atticist perspective, this prescription is justified by the lack of attestations of ἀνυπόδετος in the authors considered canonical, who consistently use the form with η, whereas the form with ε is securely attested only from the Septuagint onwards. On the other hand, it cannot be ruled out that a further reason supporting the prescription of ἀνυπόδητος was the general tendency in Atticist lexica to ascribe long vowels to Attic: cf. Vessella (2018, 93) and AGP vol. 2, Phonology, 2.2.2, forthcoming. A good example of this trend is the frequency with which Atticist lexica prescribe nominal deverbative forms ending in -ημα – in this case regularly formed – over the corresponding forms ending in -εμα, which became widespread during the Hellenistic period: cf. ἀνάθεμα vs. ἀνάθημα with Moer. α 57Moer. α 57 and Philemo (Laur.) 354Philemo (Laur.) 354; σύνθεμα vs. σύνθημα with Moer. σ 21Moer. σ 21; εὕρεμα vs. εὕρημα with Phryn. Ecl. 420Phryn. Ecl. 420 and Philemo (Vindob.) 393.22Philemo (Vindob.) 393.22. See also the entry ἔκθεμα; cf. Thackeray (1909, 79–80); Vessella (2018, 146–8); AGP vol. 2, Word formation, 2.6.1, forthcoming.

In conclusion, this is an example in which the Atticists promote and prescribe an innovative formInnovative forms, the formation of which cannot be explained by the regular historical development of the language. This prescription had lasting effects, beginning with authors contemporary with the Atticist movement and continuing into the Byzantine and modern periods (see the passages collected in B. and Section E.). However, it should be noted that there is considerable variation between ἀνυπόδητος and ἀνυπόδετος in the works of Atticising authors (cf. Lobeck 1820, 445; Dindorf, ThLG vol. 1,2, 1049, who argues that the form with η should always be restored): ἀνυπόδητος is used consistently by Dio Chrysostom (3x), Lucian (4x), Aelius Aristides (2x), and Aelian (2x), while ἀνυπόδετος is found in Galen (3x) and Flavius Philostratus (4x; cf. Schmid, Atticismus vol. 4, 341). There are also cases in which both forms are attested in the work of the same author: e.g., in Pseudo-Lucian’s Asinus, ἀνυπόδητος is used twice (Asin. 29 and 43) and ἀνυπόδετος once (Asin. 16; note that in this passage McLeod emends it to ἀνυπόδητος); in Cassius Dio, ἀνυπόδητος is attested 4x, while ἀνυπόδετος appears 2x. Especially in these cases, one may reasonably suspect that one of the two forms was introduced during manuscript transmission: see Vessella (2018, 156), who analyses such alternations in the works of other authors as well, including those arising from modern editorial practice.

E. Byzantine and Modern Greek commentary

The same alternation in the use of the two forms found in Atticising authors continues into the Byzantine era. Even here, however, any assessment is complicated by possible alterations due either to either scribal practice or to modern editorial choices. Nevertheless, the form ἀνυπόδετος appears to prevail over the other: e.g., in Michael Psellus and Nicetas Choniates we find ἀνυπόδετος (respectively, Epistulae 31.136.13 and Orationes 9.88.18), whereas ἀνυπόδητος is not attested. Examples of the use of both forms within the works of the same author are as follows: in the Chronicon by Georgius Monachus, both forms are attested (605.1 De Boor ἀνυπόδετος, 613.13 De Boor ἀνυπόδητος; note that in the redaction of the Chronicon breve (ed. Migne), the form with η is attested in both cases); as for Constantine VII Porphyrogenitus, while ἀνυπόδετος is attested in De administrando imperio (26.50), ἀνυπόδητος is attested 3x in other works. Finally, Eustathius of Thessalonica uses ἀνυπόδετος in Orationes 12.10.5 and ἀνυπόδητος in in Od. 2.311.45 Stallbaum. As regards literature in vernacular Greek, in the Ptochodromica (12th century), with its teacherly style, ἀνυπόδητος is, interestingly, preferred (4.100; 4.146); in the Historia Belisarii, ἀνυπόδητος is a hapax at 375 (rec. Λ, 15th century).

Αs regards Modern Greek, an example of the use of both forms in the same author is provided by the Nέος Παράδεισος by Agapius Landus (17th century), where ἀνυπόδητος is the common form (3x, with the variant spelling ἀνυπόδυτος attested 2x), while ἀνυπόδετος occurs only once (p. 306 l. 3, Venice 1872). In the 18th century, the Athonite monk Caesar Daponte provides an instance of ἀνυπόδητος as a learned form in his work on the war between the Ottoman, Russian, and German Empires (16.6.20 Legrand, in Ephémérides Daces ou chronique de la guerre de quatre ans 1736–1739, vol. 1, Paris 1880). In the 19th century, the connection of γυμνός ‘naked’ with ἀνυπόδητος – already attested in Plato (Prt. 321c.5) and Xenophon (HG 2.1.1) – is still exploited in the Memoirs of General Makriyiannis (309, l. 41 Vlachogiannis, in Ἀρχεῖον τοῦ στρατηγοῦ Ἰωάννου Μακρυγιάννη, vol. 1, Athens 1907). This is noteworthy, as the author writes in pure Demotic Greek. ανυπόδητος survives as a learned form even in Standard Modern Greek, where the most common word to express the same concept is ξυπόλυτος (cf. LKN s.v.), which derives from the prefixed adjective ἐξυπόλυτος/ἀξυπόλυτος attested in Medieval Greek at least since the 14th century (cf. LBG s.v.; Kriaras, LME s.v.). The form with η also survives in the Euboean dialect; cf. Andriotis (1974 s.v.).

F. Commentary on individual texts and occurrences

(1)    Theoc. 14.5–6 (C.4), schol. Theoc. 14.6 (B.5)

The presence of the Attic form ἀνυπόδητος in Theoc. 14.6 (C.4), one of the poems classified by Gow (1952, LXXII) as ‘genuine poems in Doric’, should come as no surprise, given that the use of the ‘Doric’ ἀνυπόδετος (cf. C.1) would not have been possible within the constraints of the dactylic metre. It is nevertheless interesting to note that part of the manuscript tradition of the poem attests the form ἀνυπόδατος: this is the only occurrence of this form (if we exclude B.5), and it can readily be analysed as one of the ‘hyperdorisms’ that have entered Theocritus’ text. It is precisely in light of the intrusion of this variant into the verse that the scholium in B.5, which identifies ἀνυπόδατος as ‘Doric’, should be interpreted. At the same time, the scholiast’s identification of ἀνυπόδετος as ‘Ionic’ is difficult to verify, given the lack of attestations of the adjective in authors representative of the Ionic dialect, even though the form with ε was probably common in dialects other than Attic. Finally, the scholiast’s definition of ἀνυπόδητος as κοινόςκοινός is noteworthy. While Atticist lexicography usually employs this adjective to label koine forms in opposition to Attic ones (for the nuances of this evaluative term, see Monaco 2024 and entry Moeris, Ἀττικιστής, Section F), here it instead refers to the latter. In this context, κοινός is used in its grammatical sense of ‘expected’, ‘regular’, in contrast to the dialectal variants of ἀνυπόδητος mentioned in the scholium. For this specific technical meaning of κοινός – a usage dating back at least to the dialectological theories of Apollonius Dyscolus and Herodian – see e.g. Tribulato (2014, 459–61), Van Rooy (2016, 253–4; 259–61, and 270–1), and Van Rooy (2020, 13–5), with further references.