σκότος
(Moer. σ 34, Phot. σ 377, Thom.Mag. 333.9–10)

D. General commentary

The entries in the lexica of Moeris (A.1) and Photius (A.2) (the latter depending either on Aelius Dionysius or Orus: see F.1) are concerned with the heteroclitic inflection of the noun σκότος ‘darkness’, found both as a masculine o-stemo-stems ὁ σκότος, τοῦ σκότου and as a neuter s-stems-stems τὸ σκότος, τοῦ σκότους (see Egli 1954, 64–9). While Moeris only mentions the neuter form as Attic (which is problematic: see below), Photius claims that both forms are admissible, quoting the 5th-century comic poet Ameipsias. Based on the highly epitomised text of Photius’ entry, it is difficult to say whether Ameipsias (C.6) used both forms, or only one of them; as suggested by Orth (2013, 336–7), in the latter case, it is likely that Ameipsias was quoted as an authority for the use of the more innovative and less acceptable option, i.e. τὸ σκότος. This would make him one of the first witnesses of the neuter variant in Attic together with Thucydides (C.7).

σκότος is one of several cases in which a thematic noun in -ος was reinterpreted as an s-stem neuter (see Meissner 2006, 94–8). While this phenomenon became increasingly common in Roman times, continuing through the medieval and modern periods, some individual cases are very old. For instance, ὄχοςὄχος ‘carriage’ is neuter in the Homeric poems and is first attested as masculine in the Homeric Hymns (H.Hom.Cer. 19 ὄχοισιν), but the o-grade of the root, unusual for s-stem neuters, shows that the form cannot be inherited as such. It is, rather, a contamination between τὸ ἔχος, with the expected e-grade (cf. Hsch. ε 7613 ἔχεσφιν· ἅρμασι), and an o-grade derivative of the same root: either the thematic masculine ὁ ὄχος or an ā-stem feminine *ϝοχᾱ (cf. Mycenaean wo-ka ‘chariot, vehicle’), thanks, perhaps, to the similarity between the Ionic gen. pl. *ὀχέων (< *-ᾱων) and *ἐχέων (Panagl 1982).

Likewise, σκότος, with its o-vocalism, must originally have been a thematic masculine, as also proven by the earlier documentation of the thematic forms, which are the only ones found in Homer (3x in the formula στυγερὸς δ’ ἄρα μιν σκότος εἷλε, Od. 19.339 ποτὶ δὲ σκότον in a non-formulaic context). In this case, the motivation for the metaplasmMetaplasm cannot be sought in the influence of an inherited s-stem neuter, of which no trace is to be found in Greek nor in other Indo-European languages (Germanic forms like Gothic skadus ‘shadow’ reflect a u-stem, while Old Irish scáth ‘shadow’ points to *skōto-; see EDG 1359–60). The adjective σκοτεινόςσκοτεινός ‘dark, obscure’ (Pi.+) points prima facie to an s-stem, as though from *σκοτεσ-νο- (cf. ἄλγος → ἀλγεινός, φάος → φαεινός, ὄρος → ὀρεινός, etc.), but this is hardly an argument for the antiquity of τὸ σκότος: σκοτεινός may have been created on the model of its antonym φαεινός ‘shining, radiant’. Moreover, the suffix -εινός had become productive and could be added to a thematic base; cf. ἐρατεινός ← ἐρατός, κελαδεινός ← κέλαδος. According to Fränkel (1911, 196), the neuter inflection of σκότος arose under the influence of its near-synonyms ἔρεβος and κνέφας, and of its antonym φάοςφάος ‘light’. Egli (1954, 68) argued that the starting point was, on the one hand, the Homeric formula τὸν δὲ σκότος ὄσσε κάλυψε ‘and darkness covered his eyes’ (Il. 12x), where the absence of the article made the gender of the noun ambiguous, and on the other hand the phrase ἐν φάεϊ ‘in light’, which allegedly produced the antonymous and isometrical expression ἐν σκότεϊ ‘in darkness’. Although, as Egli admitted, the phrase ἐν σκότει is not actually found before Herodotus, it is true that σκότος and φάος are frequently contrasted and thus appear side by side (e.g., in Pindar, see C.4). More problematic is Egli’s idea that the neuter inflection of σκότος was a completely artificial creation of the poetic languagePoetic language, initially confined to (ἐν) σκότεϊ and indirectly reinforced by the existence of σκοτεινός, that would have originated in choral lyric, later spreading to Ionic and eventually to Attic literary prose, from which it would have entered the spoken language only in the 3rd century BCE. In fact, early attestations of the neuter in Sicilian comedy suggest a colloquial formColloquial language.

The neuter variant appears relatively early in non-Attic dialects, beginning with Sicilian authors: the first attestation is in Epicharmus (C.1), followed by Empedocles (C.2) and Sophron (C.3). This evidence has led some scholars to see τὸ σκότος as a specifically SicilianDoric, or even Syracusan, innovationInnovative forms; so, recently, Bellocchi (2016, 343), while Redondo (2021) stresses the Sicilian comic poets’ greater openness to innovations originating in spoken language. This would be one of several cases in which an innovation first found in Epicharmus or other Sicilian authors anticipates a development of the koine (on colloquial language and ‘koineisms’ in Epicharmus, see Willi 2008, 147–9). Of course, as Cassio (2012) pointed out, not all agreements between Sicilian Doric and the koine imply borrowing from the former to the latter. Nevertheless, as early as the 5th century, τὸ σκότος does not appear to be confined to Sicily, as it is found in Pindar (C.4) and Herodotus (C.5). The innovative form also becomes increasingly frequent in Attic from the 5th century onwards: in Old Comedy, the masculine predominates (see Ar. Ach. 1168–9, V. 256, 275, and 911, Pax 691, Av. 1483, Lys. 72, Ec. 288, 314, 375, and fr. 156.1), but while Aristophanes includes no unambiguous occurrences of the neuter, his contemporary and rival Ameipsias is quoted as having used it (C.6), possibly in addition to the masculine (see A.1). In New Comedy, the distribution is reversed: the masculine occurs just once in Archedicus (fr. 1.3 ἐν τῷ σκότῳ), whereas the neuter represents the more common form and is attested in Alexis, Diphilus, and Menander (Alex. fr. 222.12 τοῦ σκότους, Men. Dysc. 428 σκότους, Diph. fr. 91.3 ὥστε καὶ ποιεῖν σκότος) as well as in two comic adespota, which are likely to belong to later comedy (com. adesp. frr. 247.9 and 1001.6; on the identification of the latter fragment, see the discussion by Kassel and Austin ad loc.). In tragedy, the neuter is sometimes found as a varia lectio (see, e.g., Aesch. fr. 6.4; Soph. OC 40; Eur. Hec. 831, HF 563 and 1159, etc.), but it is never a confirmed reading; Barrett (1964, 197) argued that ‘since the neut. became normal later, we may assume that the few instances of neut. in mss. of tragedy are the common copyist’s substitution of a late form for an early’. Barrett’s claim here may be excessive, at least as regards Euripides, who (according to Egli’s count) has 12 passages in which a neuter form is transmitted against 34 in which the masculine is virtually certain; see, recently, Francisetti Brolin (2019, 226 and 230) for a defence of the transmitted neuter forms in Eur. fr. 533.1 and fr. 534. The first certain attestation of the neuter in Attic prose is in Thucydides (C.7), followed by Plato (C.8), Xenophon (C.9), Demosthenes and Aristotle, with increasing frequency compared to the masculine (see Egli 1954, 67 for a full list of occurrences). In the Septuagint and the New Testament, σκότος is exclusively neuter, and this is the only inflection that survives in Byzantine and Modern Greek: the masculine only appears as a varia lectio in LXX Heb. 12.18, where σκότῳ and σκότει are transmitted in addition to the generally accepted reading ζόφῳ (Blass, Debrunner 1976, 41).

In light of the distribution of the masculine and neuter σκότος, Moeris’ (A.1) remark that Attic speakers use σκότος in the neuter is rather surprising, in that it entirely overlooks the fact that masculine ὁ σκότος rather than neuter τὸ σκότος is the rule in 5th-century Attic. Moeris’ entry may have been subjected to epitomisationEpitome: for instance, Moeris may have regarded ὁ σκότος as the proper Attic form and τὸ σκότος as late Attic or more generally as a minority form in Attic. The earlier text may have had <ἀρσενικῶς καὶ> οὐδετέρως, or, alternatively, ἑκατέρως, wrongly corrected to οὐδετέρως. One may compare the Atticist source on which Photius bases his entry stating that both forms of the accusative singular, the s-stem (τὸ) σκότος and the o-stem (τὸν) σκότον, are admissible in Attic, quoting Ameipsias. Interestingly, Ameipsias is quoted by erudite sources in relation to at least two further cases of gender metaplasm, one of which is again shared with Epicharmus: see F.4.

In the case of τὸ σκότος, its use by such canonical authors of Attic prose as Plato and Demosthenes as well as by other authoritative writers (Herodotus, Thucydides, Xenophon), guaranteed its acceptance among Atticist lexicographers. In fact, the masculine and the neuter coexist in Atticist prose writers (Schmid, Atticismus vol. 4, 22), with some notable individual differences: Lucian, for instance, clearly prefers the masculine (15 unambiguous occurrences) over the neuter (3x), and a similar proportion is found in Aelius Aristides (masc. 7x, neut. 1x), while Dio of Prusa consistently employs the neuter in his orations (11x).

E. Byzantine and Modern Greek commentary

The neuter form had already established itself at the expense of the masculine by the Hellenistic period and is the only form surviving in the Byzantine period (see Kriaras, LME s.v., citing only one late example of a gen. sing. σκότου from the Ρίμα παρηγορητική of Marinos Phalieros [15th century CE]); this makes it an early forerunner of the transference of masculine nouns in -ος to the neuter s-stem declension, a phenomenon that persisted in Medieval and Modern Greek (see CGMEMG vol. 2, 298, 661–3). At the same time, σκότος found itself in unfavourable competition with younger forms, such as σκοτίασκοτία (Apoll.Rh., LXX, NT), condemned by Moeris. In the New TestamentNew Testament, σκότος (31x) occurs twice as frequently as σκοτία (16x) and is always neuter. Occurrences of σκοτία are concentrated in Johannine writings (see Bishop 1961, who looks for slight differences in meaning between σκοτία and σκότος: at any rate, both words may refer to both physical and metaphorical darkness). The neuter form τὸ σκότος survives as a learned word in Modern Greek and is still used in expressions such as ο άρχοντας του σκότους ‘the prince of darkness’ denoting the Devil (see LKN s.v.). However, as the usual term for ‘darkness’, it has been replaced by its derivative σκοτάδι, descended from Medieval σκοτάδιον (see Kriaras, LME s.v. for the attestations), while σκοτία survives as a technical term of architecture, ‘scotia’ (Vitr. 3.5.2, Hsch. σ 1124).

F. Commentary on individual texts and occurrences

(1)    Phot. σ 377 (A.2)

This lemma in Photius’ lexicon was assigned to Aelius Dionysius by Erbse (1950, 141), who compared Eustathius’ references (B.4) to the discussion of this form in Aelius Dionysius’ lexicon; this identification is accepted by Theodoridis (2013, 376). Alpers (1981, 255), however, traced the source to Orus’ lexicon: see his apparatus for a fuller list of references to ancient grammarians and erudite sources.

(2)    Choerob. in Theodos. GG 4.1,336.18–37 (B.2)

Choeroboscus, in his comment on the canon of Theodosius (GG 4.1,32.15–19) regarding the inflection of neuter nouns derived from masculine nouns, offers the explanation – possibly based on Herodianic doctrine (see B.3, which Lentz included in his edition of Herodian in the belief that Choeroboscus was using the Περὶ κλίσεως ονομάτων) – that derivations such as that of τὸ σκότος from ὁ σκότος and τὸ τάριχος from ὁ τάριχος are not instances of παρασχηματισμόςπαρασχηματισμός or ‘change of (grammatical) form’ but rather denote a change of grammatical gender (the Herodianic doctrine on the gender of τάριχος is preserved by Eust. in Il. 1.117.12–6). His argument is that in neuters derived by παρασχηματισμός, the nominative singular always differs from that of the base noun, while the genitive and dative have the same endings; in cases such as σκότος, the change in gender is not expressed by any modification to the nominative singular, despite the fact that the masculine and the neuter paradoxically belong to different inflectional classes. However, Choeroboscus later adds (GG 4.1,337.2–12) that cases such as ὁ λύχνος/τὸ λύχνον, although the nom. sing. ending does change between the genders, must still be considered changes in gender and not in form, given that the two genders share the same referent (λύχνος and λύχνον both mean ‘lamp’); by contrast, ὁ σοφὸς ἄνθρωπος ‘the wise man’ and τὸ σοφὸν παιδίον ‘the wise child’ refer to different things. On the notion of μεταβολὴ γένους in ancient scholarship, see also e.g. schol. (ex.?) Hom. Il. 1.312 (A), with examples such as ὁ τράχηλος/τὸ τράχηλον ‘neck’. In Apollonius Dyscolus and Herodian, παρασχηματισμός usually denotes a change in grammatical form, as distinct from παραγωγήπαραγωγή, which refers to the derivation of a word from another.

(3)    Sophr. fr. 86 (C.3)

This fragment of Sophron is transmitted by Apollonius Dyscolus (GG 2.1.1,96.15), who quotes it for the Syracusan form ὧν, a rare genitive plural of the anaphoric pronoun equivalent to the Attic αὐτῶν ‘of them’. The fragment’s text has given rise to exegetical difficulties, largely because of the unusual verbal form ζυγαστροφεῖζυγαστροφέω, a hapaxHapax that has been emended in several different ways (see Hordern 2004, 88 and 182). Cassio (2022, 37–41) offers the most recent discussion of this fragment, defending the transmitted text and arguing for a two-layered meaning (both literal and metaphorical) of the expression ‘shooting from the dark’. In any case, editors have rightly left the s-stem gen. sing. σκότεος untouched, as it is consistent with the word’s other attestations in Sicilian Doric (see C.1, C.2).

(4)    Ameips. fr. 38 (C.6)

As stated above, Ameipsias is quoted in relation to at least two other cases of gender metaplasm. Athenaeus (Epit. 2.68b)Ath. Epit. 2.68b discusses the grammatical gender of the word for ‘oregano’, reporting that it is masculine in Anaxandrides (fr. 51) and Ion (fr. 3 Valerio), feminine in Plato Comicus (fr. *169) or Cantharus, and neuter in Epicharmus (fr. 15) and Ameipsias (fr. 36); an entry in the Etymologicum magnum (630.46–51) dependent on Choeroboscus adds that the masculine was used also by Aristophanes (Ec. 1030), and that the neuter is found in elegy as well as in Epicharmus. The neuter also occurs in Ar. fr. 128, and Antiph. fr. 221.4. The noun ὀρέγανος/-ονὀρέγανος is of unknown, probably foreign origin and was possibly folk-etymologised in Greek as though from ὄρος ‘mountain’ and γάνος ‘brightness, joy’ (DELG 820, EDG 1102); the oscillation in gender may be related to the word’s non-Greek origin and is unconnected with any discernible difference in meaning (Arnott 1996, 388). Ameipsias (fr. 39) is also quoted by the Antiatticist in an entry (σ 7)Antiatt. σ 7 discussing the gender of the word στρωματόδεσμος/-ονστρωματόδεσμον, ‘bed-sack’, wherein both the masculine and the neuter are deemed correct (see entry στρωματόδεσμον). Phrynichus (Ecl. 380)Phryn. Ecl. 380 also admits both forms as Attic, while he proscribes the use of στρωματεύςστρωματεύς – properly, ‘blanket, coverlet’ – that in the koine took on the meaning of στρωματόδεσμον (cf. also Poll. 7.79Poll. 7.79, Moer. σ 44Moer. σ 44, Thom.Mag. 332.9Thom.Mag. 332.9). The neuter form is the only form that is well attested in classical Attic, making it likely that Ameipsias is cited for the use of the rarer masculine (Orth 2013, 338). This case recalls that of σκότος, in that Atticist lexicographers apparently admitted both neuter and masculine genders as equally acceptable on the grounds that both were found in Attic authors, despite the fact that one of them (ὁ σκότος and τὸ στρωματόδεσμον, respectively) was clearly older. Instead, the Atticists directed their criticism towards more recent synonymsSynonyms (σκοτία, στρωματεύς) belonging to the koine.