ἑκὼν εἶναι
(Phryn. Ecl. 239)

D. General commentary

The correct use of the expression ἑκὼν εἶναι (‘willingly, as far as one’s own will is concerned’) is discussed in Phrynichus’ Eclogue (A.1) and in other Atticist lexica, including the Philetaerus (B.1) and Thomas Magister’s lexicon (B.7). These sources, as well as other erudite writings, agree in identifying ἑκὼν εἶναι as characteristic of Attic prose and in warning that it should be used only in negative sentences. This rule accurately reflects the classical distribution of ἑκὼν εἶναι, as noted by modern commentators on Phrynichus (see Lobeck 1820, 273–5; Rutherford 1881, 340–1, against Dindorf in ThGL vol. 3, 653, who extended the rule to include conditional sentences). The linguistic motivation for this distribution will be examined below.

The Greek adjective ἑκών ‘willing’ is etymologically the participle of an unattested present *ϝεκμι (cf. Hittite u̯ekmi, Sanskrit váśmi ‘to want, to desire’), derived from the PIE verbal root *u̯eḱ- ‘to choose, to want’ (see LIV s.v.). The participle *u(e)ḱ-ónt- is also reflected, with a different ablaut grade, in Sanskrit uśánt- ‘willing’, and may already have been lexicalised as an adjective in the parent language. Whereas ἑκών occurs independently 18 times in Homer, the phrase ἑκὼν εἶναι is a post-Homeric coinage, first attested in Herodotus. As already remarked by ancient scholarship, this phrase is used almost exclusively in negative sentences in Classical Greek; see the occurrences in Herodotus (5x), Thucydides (4x, e.g. C.2, C.3), Xenophon (7x), and Plato (15x, e.g. C.4). Out of 33 attestations in 5th- and 4th-century-BCE authors, only two occur in affirmative contexts (see Grünenwald 1888, 2–5). One of these, however, occurs in Herodotus (C.1) and is therefore of limited relevance for Attic usage, while the other, in Antisthenes’ Ajax (C.5), is Winckelmann’s conjecture for the transmitted οὐκ ὄντες, which other editors emend to φάσκοντες. Although the peculiar behaviour of ἑκὼν εἶναι is regularly discussed in the handbooks (see, e.g. K–G vol. 2,2, 14–5, 17–9; Schwyzer, Debrunner 1950, 378–9), its origin and the functional relationship between its components have not been the object of in-depth study until recently. In particular, scholars have struggled to account for the role of the infinitive εἶναι, which seems redundant with respect to ἑκών and is treated as such in ancient sources, where it is classified it among instances of παρολκή ‘redundancy’ (see B.3, B.5, B.6). The view that εἶναι has a final/consecutive function, as in (ὡς) ἔπος εἰπεῖν ‘so to say’ and (ὡς) ἐμοὶ δοκεῖν ‘methinks’ (see Wackernagel 1926, 265–6 = Wackernagel, Langslow 2009, 333–4), has been independently rejected by García Ramón (1984) and Conti (2020), who nonetheless advance markedly different accounts. According to the former scholar, the construction arose by analogyAnalogy with the synonymous ἐθέλων ‘willing’, which in some contexts may alternate with ἐθέλων εἶναι (the latter, however, is only attested relatively late), as well as with other predicative constructions in which εἶναι alternates with zero, such as αἱρέονται τοῦτον στρατηγὸν (εἶναι) ‘they elect him (to be) general’ or σὺ δοκεῖς σοφὸς (εἶναι) ‘you seem (to be) wise’. On this view, the almost complete restriction of ἐθέλων εἶναι to negative contexts reflects the emphatic character of the expression. Conti (2020), by contrast, convincingly argues that εἶναι has a limitative function, deriving from the veridical sense of εἰμί ‘to be real, to be right, to be true’ (cf. τῷ ὄντι/ὄντως ‘really, actually, verily’): the original meaning of the phrase would thus have been ‘really, actually willing’ (on the limitative function of the absolute infinitive, see the many examples already collected by Goodwin 1897, 310–3). Furthermore, Conti identifies a diachronic development in Greek texts: in three passages (Hdt. 7.164.1 = C.1; Thuc. 7.81.3 = C.3; Pl. Phdr. 252a.1), the original limitative meaning is still discernible, as εἶναι functions as a focus marker highlighting ἑκών and contrasting it with other implicit possibilities. In C.1, for instance, Herodotus stresses the remarkable fact that Cadmus relinquished his tyranny over Cos by his own volition, rather than – as might plausibly be assumed – under external pressure. In all the other occurrences, ἑκὼν εἶναι has already developed into a parenthetical expression, used in sentences with a negative polarity to express the speaker’s self-correction in hypothetical or recurrent situations. Conti further observes that ἑκὼν εἶναι underwent grammaticalisation, albeit only partially: on the one hand, the order of its two elements is fixed and εἶναι cannot be replaced by other forms of εἰμί (e.g., the future infinitive ἔσεσθαι); on the other, ἑκών still functions as an adjective in that it agrees grammatically with its referent, which is always the agent of the sentence (usually the subject, but cf. Pl. Grg. 499c.3: οὐκ ᾤμην […] ὑπὸ σοῦ ἑκόντος εἶναι ἐξαπατηθήσεσθαι, ‘I did not think […] that I would have been deceived by you, of your own will’).

The phrase ἑκὼν εἶναι, characteristic of classical Attic(-Ionic) prose, fell out of use in the koine. After Plato, it is not attested again until the imperial period, when it is employed sparingly by authors aiming at an elevated style, beginning with Dionysius of Halicarnassus (Ant.Rom. 11.9.3), followed by Iosephus (2x), Plutarch (2x), Galen (4x), Lucian (2x, e.g. C.7; see also F.2), Aelius Aristides (4x, e.g. C.6; see F.2), Aelian (1x), and Clement of Alexandria (1x). This distribution of attestations, together with the prescriptions of the lexica, suggests that the expression no longer belong to the spoken language, but was consciously adapted from canonical authors. It should also be noted that absolute infinitives in general were perceived as an Attic trait: cf., e.g., Moer. τ 5Moer. τ 5 on τήμερον εἶναι ‘for today’ and Greg.Cor. De dialectis 29.73–4. However, Phrynichus (A.1) makes the interesting observation that an ἰδιώτης ‘layman’, a sociolinguistic label usually denoting unrefined speakers (on ἰδιώτης/ἰδιῶται and ἰδιωτικός as sociolectal markers in Phrynichus’ lexica, see Tribulato 2025, 169–70), would never make the mistake of using ἑκὼν εἶναι in an affirmative clause, whereas this mistake is typical of overzealous hyper-Atticisers. In my view, this does not necessarily entail that ἑκὼν εἶναι formed part of the ordinary usage of imperial-period speakers; rather, such speakers would simply have lacked the motivation to overuse the expression and extend it to contexts in which it was not employed in Classical Greek. This interpretation is consistent with other entries in the Eclogue in which the ἰδιῶται are mentioned in an apparently – but, in fact, ironically – positive light: not because their usage is superior, but because, by not employing the classical forms at all, they avoid misusing them (see F.1). This raises the question of which group of speakers is meant by the expression οἱ νῦν εὖ φρονοῦντες, literally ‘present-day people who have good judgement’ (not a standard label in Phrynichus or other Atticists), who are said to use ἑκὼν εἶναι correctly. In light of the foregoing, it is unlikely that they should be identified with the ἰδιῶται; more plausibly, they are educated contemporary speakers who are well acquainted with the expression from classical texts but refrain from extending it to affirmative sentences in an effort to show off their learning. The condemnation of hyper-Atticist excesses was common at the time of the Second Sophistic, as illustrated by Lucian’s satires (Lexiphanes, Soloecist, Pseudologista). The same prescription found in the Eclogue appears in a highly condensed form in the Philetaerus (B.1), where it is implied that, in affirmative sentences, the simple ἑκών should be used to mean ‘willingly’.

Atticising writers of the imperial period do employ ἑκὼν εἶναι, but they do not consistently adhere to the classical norm, thereby justifying to some extent Phrynichus’ concerns. Lucian, for instance, has one sure occurrence of the phrase in a negative sentence (Scytha 8.8–9: μηδὲ τὸν ἕτερον πόδα ἑκὼν εἶναι ἀπολειπόμενος, ‘not willingly leaving him for a single step’), while in Merc.cond. 2.13 (C.7) ἑκόντες εἶναι appears to occur in an affirmative clause – unless εἶναι is taken to depend on ἐδόκουν, in which case it would have a dubitative force (‘they seemed willing to forget’). The matter is further complicated by Thomas Magister (B.7), who quotes this passage with the variant ἑκόντες ὄντες (see F.2). In the same entry, Thomas also cites one instance of the use of ἑκὼν εἶναι in an affirmative sentence by Aelius Aristides (C.6); elsewhere, excluding spurious or doubtful works, Aristides employs the phrase three times, all in negative contexts (see Schmid, Atticismus vol. 2, 56). Among other 2nd-century CE authors, Aelian has a single occurrence of ἑκὼν εἶναι in a negative context (NA 8.14.13), while Clement of Alexandria has one in an affirmative context (Strom. 7.16.103.5). Galen’s four occurrences are evenly divided between affirmative (De usu part. 10.14 Helmreich = 3.837.9 Kühn; De puls. diff. 8.660.10 Kühn) and negative sentences (De difficult. respir. 7.869.16 Kühn; De praesag. ex puls. 9.370.7 Kühn).

In later scholarship, ἑκὼν εἶναι continued to be mentioned alongside other typically Attic redundant or pleonastic constructions (B.2; B.5, cf. also Eust. in Il. 3.141.5–8; B.8). Photius (B.3) discusses ἑκὼν εἶναι in an epistle to Leo Philosophus, where he seeks to demonstrate that redundant forms of εἰμί in Scripture are not solecisms. Arguing that pleonastic expressions of this kind should be deemed acceptable in Biblical Greek just as they are in good classical authors, Photius makes the interesting parenthetical observation that some Atticists do not agree in regarding the use ἑκὼν εἶναι as a marker of stylistic refinement, thereby hinting at a disagreement not reflected in other extant scholarly sources. Photius himself employs the expression twice (Bibl. cod. 248, 430b.17; Epistulae 220.7) in negative contexts and in accordance with the classical rule.

In the late Byzantine period, the use of ἑκὼν εἶναι is included in some rhetorical treatises among the traits of Attic phraseology, mastery of which was considered useful for those aspiring to a rhetorical style. One such prescription (B.6) is found in a section of the 13th-century anonymous rhetorical treatise On the Four Parts of the Perfect Speech, falsely ascribed to Gregory of Corinth, as transmitted in cod. Par. gr. 2918, under the heading Εἰσαγωγικὸν τοῖς μέλλουσι γράφειν ῥητορικῶς (Introduction for Those Intending to Write Rhetorically). According to Conley (2006, 115–6; see also Hörandner 2012), this treatise is a product of the Palaeologan Revival, datable ‘between 1204 and about 1300’, and Joseph Rhakendytes (ca. 1260–1330) later drew on it in composing his Synopsis of Rhetoric, where a similar prescription concerning ἑκὼν εἶναι is also found. In contrast to the Atticist lexica, these works do not promote the exclusive use of ἑκὼν εἶναι in negative contexts. It is probably not coincidental that authors of the Palaeologan age use the phrase much more frequently than either their classical or imperial-period forerunners, and do not observe the restriction to negative contexts (see E.).

E. Byzantine and Modern Greek commentary

ἑκὼν εἶναι continued to enjoy a notable success among high-register authors throughout the Byzantine period. Its popularity is also attested in contemporary lexica and rhetorical handbooks (see D.), which recommended it as one of several Attic idioms that could be exploited to impart a classicising flavour to one’s prose. Several authors of the Palaeologan age, as a consequence of that period’s Atticist revival, indeed show an extreme fondness for ἑκὼν εἶναι, as is the case with Theodorus Metochites (37x), John Cantacuzenus (28x), and Nicephorus Gregoras (29x), whose enthusiastic use of the expression in all numbers and cases was highlighted by Lobeck (1820, 275). Notably, these authors show a marked preference for the combination ἑκών γε εἶναι, first attested in the 6th century CE (Anastasius of Antioch, Procopius) and otherwise rare (both the construction with and without γε are exemplified in the Suda – see B.4). In this period, the restriction to negative contexts was abandoned, and ἑκών (γε) εἶναι is used indifferently in affirmative, interrogative, and negative sentences. Particularly noteworthy is the case of Thomas Magister, who not only discusses the expression at length in his lexicon (B.7), as noted above, but also employs it (alongside the non-classical variant ἑκόντες ὄντες) in his own writings, including two declamations composed in imitation of Aelius Aristides (see F.2). The latest attestations of the infinitive construction (9x) come from Neophytus Ducas (18th–19th century). The participial construction ἑκων ὤν – inflected across all numbers, genders, and cases – is not uncommon in late Byzantine authors, including Theodorus Metochites (9x), John Cantacuzenus (14x), Manuel II Palaeologus (4x), and Bessarion (3x). Modern Greek preserves the adjectives εκών ‘willing’ and its derivative and synonym εκούσιος (from ancient ἑκούσιος, attested since Sophocles), both learned borrowings. However, the phrase ἑκὼν εἶναι did not survive.

F. Commentary on individual texts and occurrences

(1)    Phryn. Ecl. 239 (A.1)

This entry’s initial statement (absent from the abridged text of family q and cod. T) – that ‘a layman (ἰδιώτης) would not stumble’ over the usage of ἑκὼν εἶναι – while surprising at first, finds a couple of parallels in the Eclogue, although both present some textual uncertainties. The beginning of Ecl. 184Phryn. Ecl. 184 reads as follows: ἀναθέσθαι· καλῶς ὁ ἰδιώτης ἀναβάλλομαί φησιν· οἱ γὰρ ἐπὶ τούτου τάττοντες τὸ ἀναθέσθαι ἁμαρτάνουσιν […] (‘ἀναθέσθαι (‘to put off’): The layman correctly says ἀναβάλλομαι, for those who employ ἀναθέσθαι in this sense are mistaken […]’). Here, Phrynichus criticises the post-classical use of ἀνατίθεμαι in the sense of ἀναβάλλομαι (‘to put off’), whereas he accepts the former verb only in its classical meanings ‘to retract’ and ‘to put on one’s shoulders’ (see Rutherford 1881, 292, who, however, follows cod. V in printing ἀναθέσθαι κακῶς οἱ ἰδιῶται, ‘The laymen [say] ἀναθέσθαι, wrongly’, thus making the ἰδιῶται responsible for the mistake). In Ecl. 214Phryn. Ecl. 214: ἐπαοιδή ἰδιώτης λέγων οὐκ <ἂν> ἁμαρτάνοι. λέγε οὖν ὀρθῶς ἐπῳδή, ἐπεὶ τὸ διαρούμενον ποιητικόν (‘A layman would not make a mistake by saying ἐπαοιδή (‘charm’). Therefore, say correctly ἐπῳδή, since the uncontracted [form] is poetic’), the codd. read οὐχ ἁμαρτάνει ‘is not mistaken’, which de Pauw emended to ἁμαρτάνει (see Lobeck 1820, 245, who attributed the transmitted text to an interpolator’s attempt to tone down Phrynichus’ criticism). Fischer (1974, 81), on the other hand, emended the text to οὐκ <ἂν> ἁμαρτάνοι on the strength of comparison with A.1 itself, which shows that Phrynichus could occasionally praise the ἰδιῶται for avoiding an error. It should be stressed, however, that in none of the three entries discussed above is the laymen’s usage considered intrinsically superior. Rather, by ignoring the Attic forms altogether, they do not incur the risk of using them incorrectly: they do not use ἀνατίθεμαι in the sense of ‘to put off’, because they would say ἀναβάλλομαι; nor would they be tempted to elevate their diction by using the poetic form ἐπαοιδή, which is inappropriate for prose (cf. Phryn. PS 67.10–1Phryn. PS 67.10–1, where the uncontracted ἀοιδή is rejected ‘even if it was used by Homer’). In fact, the latter remark may be interpreted even more sarcastically, as implying that an uncultured individual would do well to employ ἐπαοιδή, since this would be consistent with his (lack of) education. Similarly, the most likely interpretation of A.1 is that an ἰδιώτης would not use ἑκὼν εἶναι incorrectly, because he would not use that phrase at all.

(2)    Thom.Mag. 124.13–125.9 (B.7)

This entry expands on the same Atticist doctrine expounded by Phrynichus (A.1), adding that ἑκὼν εἶναι can also be used in a sentence that is interrogative in form but negative in meaning, and – more controversially – that it may also be used without negation. In support of this claim, Thomas quotes a passage from Aristides (C.6), which indeed violates the classical norm. The subsequent quotation from Herodotus (C.1), exemplifying classical usage, appears only in cod. V (Vat. Pal. gr. 22Vat. Pal. gr. 22, 1342–1343 CE), a representative of the ‘Constantinopolitan’ recension, and is likely to be an addition by Nicephorus Gregoras’ circle (see Gaul 2007, 304). On the other hand, cod. F (Ferrara gr. II 155Ferrara gr. II 155, 1336–1337 CE) follows the quotation from Aristides with the sentence οὐ δεῖ τῷ ἅπαξ εἰρημένῳ χρῆσθαι (‘one should not use what was said once’), which shows both that Thomas’ original stance towards usage in affirmative contexts was decidedly negative and that his criticism was directed specifically against C.6. Lastly, the entry follows family q of Phrynichus’ Eclogue in discussing the variant construction ἑκόντες ὄντες, and illustrates it with two literary references. Both quotations adduced by Thomas Magister, however, are problematic: in Lucian’s passage (C.7), the MSS all read ἑκόντες εἶναι; the other occurrence, though attributed to Aristides’ Panathenaic Oration, is found neither there nor elsewhere in that author’s corpus. By contrast, the expression ἑκόντες ὄντες occurs twice in the two declamations on the theme of Demosthenes’ Against Leptines, attributed to Aristides but shown by Lenz (1942)    to be the work of Thomas Magister himself. It appears, then, that Thomas adhered in his writings to the rule that he himself formulates in his lexicon, even though in this case it was probably based on faulty examples. According to Lenz (1942, 170–1 = 1964, 269), since the Byzantine scholar is unlikely to have fabricated the quotation from Aristides, we should conclude that he made a mistake in quoting him; nor can it be excluded that his text of Lucian actually contained a variant reading ὄντες for εἶναι (perhaps a gloss originally intended to explain the infinitive?). It should also be noted that ἑκὼν εἶναι occurs four times in Thomas Magister’s other works (Pro Cynaegiro et Callimacho 1.8; Laudatio sancti Gregorii Theologi 308.45; Epistulae 9.445.4; De subditorum officiis 525.7). In the first of these, the expression appears in a rhetorical question implying a negative answer, as in the Eclogue entry, whereas the remaining occurrences are in affirmative contexts.