λαγώς, λαγός, λαγωός
(Phryn. Ecl. 156, [Hdn.] Philet. 29, [Hdn.] Philet. 89, Orus fr. A 66)

D. General commentary

Several Atticist sources deal with issues regarding different forms of the word for ‘hare’, namely λαγώς, λαγός, and λαγωός. In particular, the inflectional model of λαγώς, belonging to the so-called Attic declension, is addressed by Phrynichus (A.1), the Philetaerus (A.2), and Orus (A.4). The Philetaerus (A.3) also postulates a semantic distinction between λαγώς and λαγός (the latter indicating a ‘creature found in the sea’ – probably the Lepus marinus or the Aplysia leporina – as also stated by Hsch. λ 69: see e.g. LSJ s.v. λαγώς III). The Atticists possibly also discussed matters of accentuation and etymology, as suggested by other sources (see below; on the derivatives of λαγώς and λαγός in Atticist lexica, see entry λαγῷος, λάγειος, λαγωβολεῖον).

The formal variation between λαγώς, λαγός, and λαγωός is not easy to explain. Numerous parallels strongly suggest that λαγώς was originally a tabooTaboo word from the language of hunters (see Beekes EDG s.v. λαγώς, with bibliography), most probably a compound formation meaning ‘with flabby ears’, with a first component connected to the root of λαγαίω ‘release’ (see Beekes EDG s.v. λαγαίω) and οὖς ‘ear’ as the second component. Schwyzer (1904, 146–7; see also Schwyzer 1939, 438 n. 3) proposed *λαγ(ο)-ω(υσ)-ός as the basis of the Homeric λαγωός (see also Chantraine 1933, 14), from which the Attic λαγώς would result by contraction. On the other hand, Szemerényi (1967, 84–7) concludes that an unattested *λαγώϝης must be the basis of the contracted Attic form λαγώς (cf. perhaps Mycenaean a-no-we /anōwes-/ ‘without handles’ and cf. ἀμφώης ‘with two handles’ in Theoc. 1.28, an equivalent of the Homeric ἄμφωτος). As regards the Homeric λαγωός, Szemerényi explains it as an artificial form due to metrical needs: see e.g. Hom. Il. 10.361 (C.1), where an acc. sing. *λαγώϝεα before a word beginning with a vowel would have posed difficulties. λαγός, in turn, seemingly arose by analogy with o-stem nouns.

A preference for λαγώς is explicit in Phrynichus (A.1), but it may also be postulated in Moeris’ prescription of the adjective λαγῷος instead of λάγειος in Moer. λ 14Moer. λ 14 (see entry λαγῷος, λάγειος, λαγωβολεῖον). Notably, λαγώς is repeatedly attested in Pollux (5.15Poll. 5.15; 5.33Poll. 5.33; 5.67Poll. 5.67 etc.), though never explicitly prescribed. This preference is by no means surprising, since λαγώς is the form normally attested in canonical authors such as Aristophanes (see C.4, C.5; this should not be seen as contradicting Eustathius’ statement in B.10 that the forms in -ώς are typical of later users of Attic: for here Eustathius may be implicitly referring to a three-stage classification of the Attic dialect, probably dating back to Aelius Dionysius; according to this classification, Aristophanes and his contemporaries belong to ‘middle Attic’Attic, middle: on this matter, see entries χρέως, χρέος, ἀπολλύασιν, ἀπολλύουσιν, and other 3rd-person plurals of -νυμι verbs, and χολάς, χόλιξ). λαγώς is also found in the poets of Middle and New Comedy, in Archestratus, in epigrammatic poetry (Asclepiades, Posidippus), and in Polybius. In later prose, it is well attested as a learned form.

λαγός is considered Ionic by Phrynichus and by the source of Ath. 9.400c (see B.1 and F.2). To be sure, apart from the problematic acc. pl. λαγός in [Hes.] Sc. 302 (on which see Edwards 1971, 165), λαγός first appears in Hippon. fr. 26a West² (= fr. 37 Degani). Later it is found in Herodotus. However, it is also attested in Epicharmus (fr. 53.2) and in Attic poets of Old, Middle, and New Comedy such as Cratinus, Alexis, and Philemon (fr. 93, where Meineke writes λαγῴ instead of the transmitted λαγοί – unnecessarily, it seems to me). Notably, Callimachus, who otherwise uses only the Homeric form λαγωός, has λαγός in Iambus 6 (fr. 196.22 Pfeiffer), written in the Doric dialect. Furthermore, derivatives of λαγός in Attic authors are rare and poetic: see λαγοδαίτης (Aesch. Ag. 123, lyr.), λαγοθηρέω (Ar. Lys. 789, lyr.), and λάγινος (Aesch. Ag. 119, lyr.).

λαγός must at some point have become the predominant form in everyday languageColloquial language, although it was largely banned from the written koine. Indeed, while scholarly texts of the most varied sorts mostly have λαγώς and λαγωός, λαγός is the form attested in Modern Greek. If so, the source of B.1 (probably Tryphon, see F.2) is alluding to the everyday spoken koine of his time when he states: τὸ καθ’ ἡμᾶς ἐστὶ λαγός (‘our form of the word is λαγός’). In this regard, Moer. λ 14Moer. λ 14, which proscribes the adjective λάγειος (< λαγός) as ‘Greek’ (see entry λαγῷος, λάγειος, λαγωβολεῖον), is of limited help: firstly, it may not be related to the question of the distribution of λαγός, but only to that of λάγειος; secondly, the label Ἕλληνες in Moeris’ lexicon can carry different meanings (see Moeris, Ἀττικιστής; entry λαγῷος, λάγειος, λαγωβολεῖον).

The treatment of λαγός in Atticist sources also involves semanticSemantics issues. According to [Hdn.] Philet. 89 (A.3, possibly deriving from a synonymic lexicon, see Olson, Seaberg 2018, 299), λαγός differs from λαγώς in that the latter means ‘hare’, while the former refers to a fish (‘sea-hare’): in support of this, Ameips. fr. 17 (C.8) is mentioned. Although this distinction is probably autoschediastic, it may have arisen quite early: cf. Nic. Alex. 465, where the sea-hare is referred to with the gen. λαγοῖο (the usual form in epic poetry is otherwise the Homeric λαγωός, see below). According to his interpretation of λαγός as ‘sea-hare’, the compiler of the Philetaerus (or his source) explains λαγός in Cratin. fr. 466 (C.3) as referring to wicked men, with an allusion to the sea-hare’s deadly poison (see Nic. Alex. 465–73). Nevertheless, Cratinus’ metaphor is perhaps better explained in terms of the hare’s proverbial cowardice: see Olson, Seaberg (2018, 299–300).

λαγωός, proscribed by Phrynichus (see F.1), is already used in Homer (see C.1 and Il. 22.308–10). It is frequently found in epic and epigrammatic poetry. However, λαγωός is widely used in prose as well, including the scientific treatises of the corpus Hippocraticum and the corpus Aristotelicum, most probably because it was perceived as a learned form: as such, it is also well attested in Byzantine literature.

Finally, a look at the distribution of the three forms in Atticising authors makes it quite clear that they generally avoided λαγός, using either λαγώς (limited to Dio of Prusa and Flavius Philostratus) or λαγωός. The latter occurs in Lucian, Aelius Aristides, and Philostratus Major. Aelian has λαγωός only in NA 2.12.1, emended to λαγώς by Hermann; the latest editors, however, retain the transmitted text (note also that in NA 4.27.5, 11.9.3, 13.14.2, and 13.14.44 the dat. sing. λαγῷ, the gen. pl. λαγῶν, and the nom. pl. λαγῷ are written with a circumflex accent in the MSS: on the problems associated with the accentuation of λαγώς see below).

The inflectional model of λαγώς is dealt with by Phrynichus (A.1), the Philetaerus (A.2, A.3), and Orus (A.4). It belongs to the so-called Attic declension (on which see AGP vol. 2, Nominal morphology, forthcoming, and entry νεώς, νεός) and goes as follows: nom. sing. λαγώς; gen. sing. λαγώ; dat. sing. λαγῴ; acc. sing. λαγών; nom. pl. λαγῴ; gen. pl. λαγών; dat. pl. λαγῴς; acc. pl. λαγώς; nom. and acc. dual λαγώ; gen. and dat. dual λαγῴν. The author of the Philetaerus also recommends the acc. sing. λαγώ, a form proscribed by Orus fr. A 66 (A.4; on this lemma, see also entry νεώς, ναός). In this respect, see also Lucian’s Sol. 3, where the acc. λαγώ ‘is a deliberate blunder’ (MacLeod 1967, 13 n. 3). AccusativesAccusative in -ω, analogical to those of the s-stem nouns in -ως, are attested in Attic prose and comedy – albeit only starting from the 4th century BCE (note also that forms like acc. sing. νεώ are ‘probably universal after 350 BCE’ in Attic inscriptions, as Threatte 1996, 39 points out; further discussion of the accusative in -ω in AGP vol. 1, 260 and AGP vol. 2, Nominal morphology, forthcoming). To take one example, the acc. sing. λαγῶ is repeatedly attested in Xenophon (on the accent, see below), an author who is cited more than once by the compiler of the Philetaerus (see entry [Herodian], Φιλέταιρος (Philetaerus)).

Atticist lexicographers do not explicitly address the question of the accentuationAccent of λαγώς and other nouns of the so-called Attic declension. Nevertheless, it is conceivable that they were concerned with it. A possible hint in this sense is Moer. ν 1Moer. ν 1: νεώς τὴν εὐθεῖαν ἑνικῶς καὶ ὀξυτόνως Ἀττικοί· ναός Ἕλληνες (‘Users of Attic [employ] the nominative νεώς (‘temple’), which is singular and oxytone. Users of Greek [employ] ναός’), on which see entry νεώς, ναός. As for λαγώς, Atticist lexica may have relied on earlier scholarship. According to the early 1st-century BCE grammarian Tryphon (see B.1), in Aristophanes’ Banqueters (fr. 218 = C.4) and Danaids (fr. 263 = C.5) λαγώς has an acute accent, while Xenophon in the Art of Hunting (e.g. 3.3 = C.9) has the acc. sing. λαγῶ, ‘without the ν and with a circumflex accent on the final syllable’. At least in the case of Xenophon, Tryphon may have noted λαγῶ as a deviant form (λαγών being the expected one) and argued for its accent by setting an analogy with other accusatives ending in -ῶ – see entry ἱδρῶ, κυκεῶ, Ἀπόλλω, τυφῶ (Tryphon’s statements are hardly based on the authority of the MSS at his disposal, since accents were rarely marked in his time. In any case, the extant MSS of Xenophon have λαγῶ). If so, it would be unnecessary to think that, according to Tryphon, λαγώς regularly takes an acute accent in the nom. and acc. and a circumflex one in the oblique cases, as cautiously supposed by Dieu (2022, 469). As for λαγώς with an oxytone accent in Aristophanes, Tryphon may be relying on earlier scholarship (Aristophanes of Byzantium? Hellenistic treatises on the Attic dialect? On this matter, see e.g. Probert 2006, 24–5). It is also possible that the rule according to which nouns such as νεώς etc. are accented according to their standard equivalents, although explicitly formulated in later sources depending on Herodian (see B.2, B.3), already existed in Tryphon’s time (the evidence provided by B.1 is not entirely conclusive, anyway: see F.2): if so, Tryphon himself may have applied it to the occurrences of λαγώς in Aristophanes.

Herodian, for his part, thought that the Attic form should have a circumflex accent, unlike its standard equivalent (see B.1, B.2). It is not immediately clear why Herodian (or his source), who otherwise takes Attic forms derived from standard ones in -ός to be oxytone, considered λαγώς to be an exception (together with ὀρφώς/ὀρφῶς, on which see Dieu 2022, 469). Perhaps discussions on the etymology of λαγώς played a role: see Orio 94.9–10, according to whom a ‘Doric’ ὦς (‘ear’) is the second component of the compound (cf. Choerob. Epim. in Ps. 173.6–8; Et.Gen. λ 7; EM 554.21–5; [Zonar.] 1280.8–10) and Eustathius (B.9), stating that a compound of ὦς should have a circumflex accent (here Eustathius may or may not depend on an unknown source: see van der Valk in the apparatus to Eust. in Il. 3.90.14). It is also conceivable that the Homeric form λαγωός was regarded by Herodian (or his source) as the ‘original’ one, from which the Attic form was derived through contraction. However, these hypotheses appear to be contradicted by John Philoponus’ words in B.3 (seemingly reflecting Herodian’s own assessment of the matter): here the accentuation is said to be ἄνευ λόγου ‘without reason’. Moreover, John Philoponus also gives Τηλεβῶς < Τηλεβόας as the only case in which -ῶς is the result of a contraction.

E. Byzantine and Modern Greek commentary

In Byzantine scholarly literature, the usual forms are λαγώς and λαγωός, both of which were probably perceived as learned forms because of their literary pedigree, while λαγός is rarely attested. In particular, the latter occurs in Digenis Akritis (cod. E 675) and in an anonymous poem composed in vernacular Greek (Ad vinum 116 Protopapa–Bouboulidou, 17th century). In this regard, it should also be noted that several compounds with a first component λαγο- are attested in Medieval Greek (see entry λαγῷος, λάγειος, λαγωβολεῖον) and that λαγός is also the standard form in Modern Greek, where, however, ‘the old stem λαγω- is used in learned compounds in the medical vocabulary’, as de Tocqueville (2021) notes.

F. Commentary on individual texts and occurrences

(1)    Phryn. Ecl. 156 (A.1)

Phrynichus’ wording probably means that the Homeric λαγωός is not found in canonical Attic authors: cf. a similar formulation in Phryn. Ecl. 58Phryn. Ecl. 58: προκόπτειν λέγουσιν, τὸ δὲ ὄνομα προκοπὴ οὐκ ἔστι παρ’ αὐτοῖς (‘They [i.e., canonical Attic authors] say προκόπτειν (‘to advance’), while the noun προκοπή (‘advance’, ‘progress’, ‘success’) is not attested in their writings’). Nevertheless, in light of what other sources (B.1, B.8) state about the Homeric form, I would not exclude the possibility that the entry actually means that λαγωός does not exist in the real language but is an artificial form created for the sake of metre.

(2)    Ath. 9.400a–d (B.1)

The pericope οὕτως δ’ ἐχρήσατο τῷ ὀνόματι καὶ Ἐπίχαρμος καὶ Ἡρόδοτος καὶ ὁ τοὺς Εἵλωτας ποιήσας. εἶτά ἐστι τὸ μὲν Ἰακὸν λαγός ‘λαγὸν ταράξας πῖθι τὸν θαλάσσιον’, τὸ δὲ λαγώς Ἀττικόν. λέγουσι δὲ καὶ Ἀττικοὶ λαγός, ὡς ὁ Σοφοκλῆς ‘γέρανοι, κορῶναι, γλαῦκες, ἰκτῖνοι, λαγοί’ is apparently problematic. Since Epicharmus and Herodotus both use λαγός, while Eupolis elsewhere has λαγώς (see fr. 174.2–3 = C.7), Schweighäuser (1804, 213) thought that the mention of Epicharmus and Herodotus should be transposed after the following sentence (εἶτά ἐστι τὸ μὲν Ἰακὸν λαγός etc.). However, here Athenaeus’ source may simply mean that Epicharmus, Herodotus, and Eupolis used oxytone forms (either λαγός or λαγώς), so it is safer to retain the text transmitted by the MSS. This is not the only problem with B.1. In 9.400a Athenaeus explicitly cites Tryphon as his source. However, the possibility that Athenaeus used different erudite material (or at least two different passages by Tryphon himself), briefly mentioned by Olson, Seaberg (2018, 299) and Sandri (2023, 86), is strongly suggested by the repetition of Soph. fr. 111 (the second time with κορῶναι instead of χελῶναι; such repetitions are not unusual in Athenaeus, see e.g. Pellettieri 2020, 281 n. 80). If so, it is conceivable that an unknown source (quoting Sophocles, although in a different way from Tryphon) was appended to the preceding material by means of the adverb εἶτα (‘then’, ‘next’) in Ath. 9.400c: εἶτα is frequently used by Athenaeus, e.g. when quoting different passages from the same comic poet. By admitting at least two different sources (the second perhaps starting from εἶτα), the doubts already expressed by Schweighäuser (1804, 213) would find a different solution, without the need to change the transmitted text. With regard to the content of Ath. 9.400c–d, there is also something to be noted. First, there is a distinction between Attic speakers and Ionians. Secondly, the non-Attic λαγός is said to occur in Attic authors too. Furthermore, the Homeric form λαγωός is regarded as artificial, a view also reported by Eustathius (B.8) and possibly lying behind Phrynichus’ puzzling wording in A.1 (see F.1). The adjective λαγῷος is also discussed: this is the form prescribed by Moeris (λ 14)Moer. λ 14 and mentioned by Eustathius (see entry λαγῷος, λάγειος, λαγωβολεῖον). Thus, Athenaeus’ source may not have been strictly Atticist, but may have had an interest in Ionic forms. This source was possibly common to Phrynichus, Moeris, and Eustathius, as well. This reconstructed picture coincides with what we know of Aelius Dionysius’ work (see entry Aelius Dionysius, Ἀττικὰ ὀνόματα): it is conceivable that his lexicon was used by Athenaeus together with Tryphon.