ἡδίων
(Antiatt. η 5)

D. General commentary

Atticist lexicography dealt extensively with comparatives and superlatives (see e.g. entries ἀμεινότερος, ῥᾳότερος; ἀγαθώτερος, ἀγαθώτατος; βελτίους, βελτίονες; βράχιστος, βραχύτατος; μειζόνως), and AGP vol. 2, Nominal morphology, forthcoming). In this case, the Antiatticist’s interest (A.1) in the primary comparative ἡδίων, ἥδιον (‘more pleasant’, from ἡδύς) concerns not morphology but prosodyProsody and, more specifically, the quantity of the iota in the ending -ίων/-ιον (on primary comparatives in general, see entries βελτίους, βελτίονες and βράχιστος, βραχύτατος, with further bibliography. On the etymology and semantics of ἡδύς and its derivatives, see De Lamberterie 1990, 464–70, 483–97. On matters of vowel quantity in Atticist lexica, see Vessella 2018, 64–81).

The Greek n-stem primary comparative ending -(ί)ων/-(ι)ον is innovative compared to the inherited IE comparative suffix that ends in a sibilant: the latter has two allomorphs – one disyllabic, *‑ios, and one monosyllabic, *-i̯os (the distribution of the two appears consistent with Sievers’ law, cf. e.g. Vessella 2007, 133; Barber 2013, 185–6). The ending -(ί)ων/-(ι)ον derives from the zero-grade of the suffix in the sibilant, i.e. *-is-, with the addition of the ‘individualising’ suffix *-on, resulting in *-is-on (cf. Nikolaev 2022, 552). Depending on the root to which the suffix is attached and the ensuing phonological mutations, these comparatives may synchronically end either in -ίων/-ιον or -ων/-ον (cf. e.g. κακίων ‘worse’, from κακός ‘bad’, and ἐλάσσων ‘smaller’ from ἐλαχύς ‘small’), but sometimes the same root may generate two primary n-stem comparatives, one with and one without iota – cf. e.g. θάσσων and ταχίων from ταχύς (‘quick’, on which, see entry θάττων, ταχίων; cf. also βράσσων and βραχίων from βραχύς ‘short’, discussed in entry βράχιστος, βραχύτατος).

In -ίων/-ιον endings, the quantity of the iota is not fixed. In Homer and Pindar, it is almost invariably -ῐ-, while in Attic poetry, it is mostly -ῑ-, and in other poetic texts from Archilochus onwards, it alternates between -ῐ- and -ῑ- (cf. Nikolaev 2022, 552–3). Modern linguistics has offered various explanations for this quantitative oscillation. Among the most recent studies, Barber (2013, 185–6) posits a suffix *-(i)i̯os-/*-īi̯os-, while Nikolaev (2022, 553–4, drawing on Kuryłowicz 1956, 275–6 and Szemerényi 1968, 32) explains the alternation by way of reciprocal analogical remodelling between different categories of primary comparative stems. In particular, he identifies three primary comparative stem categories – namely, (a) the comparatives in *-i̯os-, which survive only in some instances of specific paradigms (e.g. nom. pl. m.-fem. μείζους ‘bigger’, from *meg-i̯os-es); (b) the comparatives in -ων/-ον, which ‘were not fully remade as stems in -ιων but rather became -on- stems’ (Nikolaev 2022, 554), such as κρείσσων and θάσσων; and (c) the comparatives in -ῐων/-ῐον, such as κακίων (from *-is-on, see above). Consequently, the inconsistent quantity of the iota in Greek primary comparatives is explained as an analogical remodelling of the forms belonging to the third category (c) on those from the second (b), given that ‘in Attic the suffix -ων was frequently preceded by a syllable with a long vowel (μείζων, μᾶλλον)’ (Nikolaev 2022, 553): this ‘long vowel plus ων’ pattern was thus applied to the ending -ίων/ιον, producing -ῑων/-ῑον.

This extension of the long vowel to all primary comparatives is thus specific to Attic, while ‘the other dialects seem to have had a mixed system, in that forms with no lengthening (i.e. μέζων, κρέσσων) are usually found side by side with ones with long ā such as θᾶσσον, μᾶλλον’ (Vessella 2007, 134). However, exceptions within Attic – i.e. instances of short iota in comparatives in -ίων/-ιον – are also attested, the most unambiguous being βελτίω in a Eupolidean trimeter (C.4) along with κάλλῐον in Ar. Eq. 1264 (in which, however, the poet is quoting Pi. fr. 89a). Another example in which the iota is unequivocally short – i.e. ἥδιον at the end of the trimeter in C.3 – is problematic in light of the fragment’s debated attribution: while Kassel and Austin consider it an authentic passage from a play by Alexis (PCG vol. 2, 38; cf. also Nesselrath 1990, 69–70), Arnott (1955; 1996, 819–22) holds it to be unquestionably spurious (the short iota would, in fact, be one of the many hints suggesting that the text is later than Alexis, cf. Arnott 1996, 820 and below). Moreover, in some occurrences, the matter remains unclear: for instance, in Ar. V. 297 (C.1), some interpret the iota in ἥδιον as short (cf. MacDowell 1971, 174; Collard 1975, 386; Diggle 1981, 29; Parker 1997, 220) while others are inclined to read it as long (cf. Biles, Olson 2015, 175).

The Hellenistic scholar Aristophanes of Byzantium, in his Attic Words (fr. 347), appears to have acknowledged the long quantity of the iota in -ίων/-ιον comparatives as a specific feature of Attic (see Slater 1986, 120; Callanan 1987, 31–2; AGP vol. 1, 483–5). Indeed, Herodian (B.1) – in formulating the norm that all comparatives in -ίων have a long iota – quotes him as a testimony (ἱστορεῖ) to this phenomenon (cf. Tosi 1997, 173). The same Herodian (in the work On Modifications of Words, quoted in an entry from the Etymologicum Genuinum, B.2) cites a case in Eupolis (C.4) that contradicts the norm expressed in B.1 – namely, a line in which the neuter plural βελτίω scans with a short iota. In light of these two passages (B.1, B.2), the most reasonable explanation for the Antiatticist’s entry on ἥδιον (A.1) is that the lexicographer is participating in the debate concerning the length of the iota in primary comparatives by signalling a case similar to C.3, in which a comparative in -ιον scans with a short iota in an Attic text against the expected -ῑον (cf. already Kock, CAF vol. 2, 354; Schulze 1892, 300 n. 4).

The passage in which the Antiatticist observed this exception to the general Attic tendency to lengthen the iota is derived from an otherwise unknown play by Alexis, the Ὀδύσσευς ἀπονιπτόμενος (or ἀπονιζόμενος, according to Naber’s [1880, 261] and Kock’s [CAF vol. 2, 353] emendation), i.e. the Odysseus Washed Clean (C.2). According to Slater (1987, 120), the Antiatticist may have sourced the reference to C.2 from Aristophanes of Byzantium’s discussion of primary comparatives in his Attic Words (see above), a discussion in which the scholar proved to be ‘aware of exceptions to his general rule’ (Slater 1987, 120). Tosi (1997, 173), however, convincingly asserts that the phrase Ἀριστοφάνης […] ἱστορεῖ (B.1) implies a description of the phenomenon in Attic authors rather than a prescription concerning the correct prosody of primary comparatives more generally (see also AGP 1, 483–5). Therefore, it appears more likely that the Antiatticist independently drew on Alexis as a notable exception to the Attic lengthening of the iota in primary comparatives, possibly in response to a (now lost) prescription by stricter Atticists, who had elevated this general tendency to the status of puristic norm (see Tosi 1997, 173–4).

Irrespective of the source from which the Antiatticist drew the quotation from Alexis, it is nonetheless remarkable that two fragments ascribed to two otherwise unknown plays attributed to this poet – the Odysseus Washed Clean (A.1, C.2) and the Teacher of Prodigality (C.3) – both preserved ἥδιον with short iota. Arnott, however, makes no mention of this parallelism but instead offers a wholly different interpretation of A.1 and of the debate to which the entry may have related (cf. Arnott 1989; 1996, 464–5). According to him, ἥδιον in A.1 is an itacistic error for ἴδιον (‘one’s own’, ‘private’). ἴδιος had become widespread in the koine as a reflexive possessive adjective, replacing the genitive of the reflexive pronoun (i.e. τὰ ἴδια πράττω instead of τὰ ἐμαυτοῦ πράττω, ‘I do my own things’), and was therefore proscribed by Phrynichus (Ecl. 409Phryn. Ecl. 409; see entry ἴδιος). Nevertheless, despite the potential support offered by Phryn. Ecl. 409, Arnott’s interpretation of A.1 has generally been regarded as unlikely by other scholars (cf. Tosi 1997, 173 n. 7; Valente 2015, 177).

A potential confirmation that strict Atticists prescribed the long iota in primary comparatives may come from another entry from Phrynichus’ Eclogue (B.3), in which Phrynichus discusses the forms ἔγγιον (the primary comparative of ἐγγύςἐγγύς, ‘close’) and ἔγγειονἔγγειον (‘of the land’, from the noun γῆ, ‘earth, land’). He altogether discourages the use of the former (and prescribes the secondary comparative ἐγγύτερον instead) and praises the use of the latter in phrases such as ἔγγειον κτῆμα (‘land property’). As Vessella (2018, 196–7) has suggested, the underlying reason for this entry is that – at least, in the pronunciation that Phrynichus promotes – the primary comparative ἔγγιον and the adjective ἔγγειον were homophones, with the <ι> in ἔγγιον and the <ει> in ἔγγειον being both pronounced as a /iː/. Indeed, although the loss of vowel quantity distinction was already well underway in the imperial period (cf. Vessella 2018, 64–73), correct vowel quantity in spelling and pronunciation was characteristic of educated speech and, as such, was discussed at length in Atticist lexica (cf. Vessella 2018, 78–81).

E. Byzantine and Modern Greek commentary

The comparative ἡδίων, ἥδιον remains widely attested throughout Byzantine literature, particularly in texts that are characterised by a high register, cf. CGMEMG vol. 2, 814 (in works dating between the 10th and the 14th centuries, more than 500 cases of the primary comparative ἡδίων, ἥδιον are attested compared to fewer than 200 cases of the secondary ἡδύτερος, -α, -ον). The primary comparative ἡδίων, ἥδιον is employed up until the modern period by authors who use archaising language (cf. e.g. Neophytus Ducas 28x). The adjective ἡδύς has no continuation in Modern Greek.

F. Commentary on individual texts and occurrences

N/A