ὄφλω, ὀφείλω, ὀφλισκάνω
(Antiatt. ο 23, Phot. ο 364)

D. General commentary

An entry of the Antiatticist (A.1) attests to the use of ὀφλισκάνω ‘to owe, to be liable to pay’ by Eupolis (C.3), while an entry of Photius’ lexicon, probably based on Aelius Dionysius (A.2), states that ὄφλειν, like ῥόφειν ‘to slurp, to gulp down’ (see entry ῥοφέω, ῥυφέω), should be accented on the first syllable, i.e. as the infinitive of a (non-contract) thematic present. Although these two entries may appear unrelated, both likely derive from an erudite debate concerning the correct form of the present indicative of verbs meaning ‘to owe, be liable’ (see Olson 2016, 225). The problem, as already clearly and succinctly stated by Lobeck (1866, 150 n. 4), is that ὄφλειν and ῥόφειν are, in fact, aoristAorist infinitives – accented ὀφλεῖν, ῥοφεῖν in modern editions – and corresponding to the aorist indicatives ὦφλον, ἔρροφον. Nevertheless, several ancient authorities interpreted – and accented – them as presentsPresent (see, most recently, Vessella 2018, 237; in principle, A.2 could also be arguing against a contract present ὀφλέω, ὀφλεῖν, but traces of such a form are virtually non-existent; see C.2 for ὤφλεε in some codd. of Herodotus, and E. for -οφλέω in Post-classical Greek). To understand how such an interpretation arose, it is necessary to examine the complex morphological and semantic history of these forms.

The zero-grade thematic aorist ὦφλον, attested already in Mycenaean (-o-po-ro, arguably for unaugmented /-opʰlon/) and again in Classical Greek from Aeschylus (C.1) and Herodotus (C.2) onwards, is etymologically related to the present ὀφείλω (Cretan Doric ὀφήλω, Lesbian and Arcadian ὀφέλλω) ‘to owe, to be obliged to’, arguably from *ὀφέλνω (on the phonology see Savić 2018, 26–7). Over time, however, the relationship between ὀφείλω and ὦφλον became increasingly obscure. In addition to the aorist -o-po-ro, Mycenaean also attests a thematic present indicative o-pe-ro-si (3rd plur.), participle o-pe-ro, etc., which can be interpreted phonologically as /opʰellō/ or /opʰēlō/ (corresponding to ὀφείλω/ὀφήλω/ὀφέλλω) or alternatively as /opʰelō/, in which case the form did not survive into 1st-millennium BCE Greek. Subsequently, the distinct tense-aspect stems acquired different semantic specialisations, giving rise to separate paradigms: a new nasal present *ὀφέλνω was formed (perhaps from the old imperfect ὤφελον, reinterpreted as a thematic aorist) with the meaning ‘to owe, to be obliged to’, with aor. ὠφείλησα, fut. ὀφειλήσω, and perf. ὠφείληκα. The zero-grade stem of ὦφλον, in the technical juridical sense of ‘to lose a cause’, formed the present ὀφλισκάνω (the variants ὀφλίσκω and ὀφλάνω are only attested in erudite sources), with fut. ὀφλήσω and perf. ὤφληκα (for this reconstruction, see de Lamberterie 1992; the diachronic semantic developments are described in detail by Ruiz Yamuza 2008, 37–86 and Allan 2013). Meanwhile, ὤφελ(λ)ον (and the unaugmented form ὄφελ(λ)ον in epic language) specialised in expressing counterfactual wishes and, over time, became an invariable modal particle (see Revuelta Puigdollers 2017). Whether ὀφείλω is etymologically related to ὀφέλλω ‘to increase’ (in Homer and later poetry) and/or ὀφέλλω ‘to sweep’ (a hapax in Hippon. fr. 79.19) is still an open question, which can be set aside here. Hamp (1982, 227–8) posited distinct roots, whereas de Lamberterie (1992) derived all three verbs from a single root, *h₃bʰel- (on the possible isogloss with Armenian, the root awel- corresponds to both meanings of the Greek ὀφέλλω, see de Lamberterie 1992, 209–15; Clackson 1994, 156–8). It should be noted that other reconstructions of these verbs’ prehistory imply a different relationship between ὦφλον, ὤφελον, and ὀφείλω: for instance, Kümmel (2023, 41) reconstructs a root *h₃bʰelh₁- ‘to owe’ (distinct from *h₃bʰel- ‘to increase’), with a root aorist *h₃bʰélh₁-/*h₃bʰl̥h₁-, from whose strong and weak stem ὤφελον and ὦφλον were respectively derived, and a nasal-infix present *h₃bʰel-néh₁-, which underlies *ὀφέλνω > ὀφείλω. By contrast, Willi (2018, 339 n. 72) assumes that the root aorist *(e-)h₃bʰel- was analogically thematised to *(e-)h₃bʰel-e-t (> ὤφελε) after the thematisation of root aorists ending in *-h₁- such as *(e-)gʷelh₁-t > *(e-)gʷel-e-t > ἔβαλε (alongside the thematic *(e-)h₃bʰl-e-t > ὦφλε).

Be that as it may, in Classical Greek ὦφλον was separated from the paradigm of ὀφείλω, while its connection with ὀφλισκάνω was not especially close. Indeed, in classical authors, the present is relatively uncommon and is mostly used in the moral sense ‘to deserve, to bring on oneself, to incur (a charge of)’, in collocation with objects such as αἰσχύνην ‘shame’, γέλωτα ‘ridicule’, δειλίαν ‘cowardice’, etc. (a meaning also widespread in the aorist: see e.g. C.5); in tragedy, for instance, this is the only meaning attested for the present stem (Soph. 2x, Eur. 7x). On the other hand, common juridical expressions such as δίκην ὀφλεῖν ‘to lose a case’ (also simply ὀφλεῖν, cf. e.g. C.7), δίκην ἕρημον ὀφλεῖν ‘to let judgement go by default’, etc., are typically found in the aorist (as well as the future and perfect) stem (incidentally, this technical usage is taken over by Phryn. Ecl. 394Phryn. Ecl. 394: οὐ γὰρ περιόψεσθαί σε ἡγοῦμαι ἐρήμην ὀφλόντα σου τὰ παιδικὰ Μένανδρον, ‘for I do not think that you will allow that sweetheart of yours, Menander, to be condemned in absentia’). To take two canonical prose authors as examples, Plato has only a single occurrence of the present stem (Tht. 161e γέλωτα ὀφλισκάνομεν ‘we deserve ridicule’) out of 65 total instances of the verb; in Demosthenes the proportion is 8 out of 99 (including the compound προσοφλισκάνω). This state of affairs readily explains why ancient scholars were uncertain about the relationship between ὦφλον and ὀφλισκάνω, and why they developed different theories to account for it. One approach was to interpret the occurrences of the infinitive ὄφλειν/ὀφλεῖν and the participle ὄφλων/ὀφλών as deriving from an (unattested) present ὄφλω. Parallels cases of thematic aorists interpreted – and accented – as presents in ancient scholarship are discussed by Dieu (2022, 246‒7), e.g. θιγεῖν, θιγών (from θιγγάνω, ἔθιγον ‘to touch’) accented θίγειν, θίγων as if from *θίγω; or πεφνών, πεφνεῖν ‘to slay’, accented πέφνων, πέφνειν after it became formally and semantically detached from its old present θείνω ‘to strike’. In the manuscript tradition, forms of ὀφλεῖν are likewise sometimes accented as if they were presents: cf. e.g. Aesch. Ag. 534 (C.1), Ar. Ach. 689, 691 (C.4), D. 30.28, etc.

The (fictitious) present ὄφλω was etymologically justified as a syncopated form of ὀφείλω: in spite of the paradigmatic and semantic split between ὀφείλω and ὦφλον, these verbs remained close enough in both form and meaning that awareness of their etymological connection was never entirely lost. The former is, in fact, a frequent trivialisation of the latter in the manuscript tradition of both literary works (e.g. D. 21.44, 24.55, etc.) and erudite ones (see B.4, B.5, and F.1). The idea that ὄφλω was derived by syncope from ὀφείλω goes back at least to Herodian (as preserved in Pseudo-Arcadius’ epitome: B.3), who may well be the source of Orion’s lexicon (B.2), and found its way into the Byzantine Etymologica (e.g. B.8; cf. also Et.Gud. 443.43-6). Eustathius (B.7) preserves a fragment of Heraclides of Miletus, who reports that some scholars appealed to syncope to explain various verbal derivations, including ὄφλω < ὀφείλω, although Heraclides himself preferred a different explanation – at least for δάκνω ‘to bite’. Given that Heraclides was active around 100 CE, the first attestations of the theory deriving ὄφλω from ὀφείλω can be projected to at least the 1st century CE. By contrast, the semantic distinction between ὄφλω and ὀφείλω was asserted by the Tiberian-age grammarian Apollonides of Nicaea (on which see Ippolito 2015), quoted in the epitome of Herennius Philo’s synonymic lexicon attributed to Ammonius (B.4). Thomas Magister (B.9) still clearly sets out the difference, although he treats ὀφλισκάνω as being equivalent to ὄφλω in this respect. It should also be noted that two other presents, ὀφλίσκω (attested in the Byzantine period, see E.) and ὀφλάνω (unattested outside of erudite works), are mentioned by some sources as equivalents of ὀφλισκάνω and/or ὄφλω: in addition to B.6, see Hsch. ο 1970: ὀφλάνειν· ὀφλισκάνειν, ὀφείλειν, Phot. ο 715: ὀφλάνειν· ἀντὶ τοῦ ὀφλισκάνειν, Theognost. Can. 839.12: ὀφλίσκω, ἀφ’ οὗ τὸ ὀφλισκάνω.

If the textual tradition is to be trusted, ὄφλω does not appear to have enjoyed much success among Atticising writers, despite the endorsement it received from some Atticist grammarians. Only Dio Chrysostomus, writing at roughly the same time as the lexica, uses forms of the present ὄφλω (31.143 = C.8; 31.153), which Schmid (Atticismus vol. 1, 85; 4, 600) regarded as a concession to the tendency towards simplification and regularity in conjugation in the vulgar language – comparable to other instances of presents remodelled on aorist or perfect stems. However, in light of the erudite sources that expressly argue in favour of ὄφλω, ὄφλειν, Dio’s usage may be better understood as following contemporary Atticist prescriptions rather than reflecting the spoken language (note that ὄφλω does not appear to be attested in papyri). In fact, it is possible that the Antiatticist’s defence of ὀφλισκάνειν (A.1) was intended as a reaction against those grammarians and lexicographers who promoted ὄφλω, although the severe abridgement undergone by the entry leaves other interpretations open. Forms of ὄφλω accented as presents occur sporadically in the tradition of later authors, but they did not ultimately survive into the modern language (see E.).

E. Byzantine and Modern Greek commentary

Throughout the Byzantine age, ὀφλισκάνω and ὦφλον continued to be in use. In the manuscript tradition, the forms in ὀφλ- are frequently accented as presents, even in high-register authors (e.g. 4x in Michael Choniates, 1x in Nicetas Choniates, 3x in Georgius Pachymeres, 15x in Theodorus Metochites, 7x in Nicephorus Gregoras), possibly owing to the influence of the grammatical tradition that labelled them as Attic. While ὀφλάνω is unattested, ὀφλισκομένων (from ὀφλίσκω) was employed by Theodorus Studites (Μεγάλη κατήχησις 42 [303.12] Papadopoulos-Kerameus; Sermones Catecheseos Magnae 19 [55.24] Cozza-Luzi). Neither ὀφλισκάνω nor ὄφλω survived into Modern Greek, while οφείλω continues to be used with the meaning ‘to owe, to be obliged to’, alongside the more common χρωστώ, -άω ‘to owe, to be in debt’, attested as χρεωστέω since the imperial period and often employed in lexica as the interpretamentum of ὀφλισκάνω and related forms (see B.6, B.8). It is worth noting that Medieval Greek attests the compound verbs ἀποφλέω and ἐξοφλέω (see Kriaras, LME s.vv.), whose contract inflection is based on the aorist stem ὀφλη- (see LKN s.v. ξοφλώ), rather than deriving from an ancient variant ὀφλέω. Standard Modern Greek has (ε)ξοφλώ ‘to pay off, to repay’, while ’ποφλώ survives in dialects (Andriotis 1974 s.v.).

F. Commentary on individual texts and occurrences

(1)    Moer. ω 13 (B.1)

Somewhat surprisingly, Moeris reports ὤφειλε δίκην as the Attic expression equivalent to ἡττήθη δίκην ‘he lost a cause’, with ὤφειλε instead of the expected ὦφλε. Pierson (1759, 427), observing that the usual expression in Attic orators is δίκην ὀφλεῖν, proposed to restore ὦφλε δίκην in Moeris’ lemma and οἱ δίκην ὀφλόντες ‘those who lost a cause’ in Harp. υ 7 (~ Phot. υ 115, Su. υ 282), where the codd. have ὀφείλοντες. The latter emendation was accepted by Dindorf, who compared it with Lex.Rhet. 311.29, whereas Hansen’s edition of Moeris still reads ὤφειλε. A similar case occurs in Poll. 8.37Poll. 8.37, where δίκην ὀφλεῖν is transmitted by cod. B, as opposed to the trivialising δίκην ὀφείλειν of the other codd. The emendation to ὀφλεῖν was already proposed by Dindorf, who noted that Pollux himself uses ὀφλεῖν at 3.84Poll. 3.84, but it was overlooked in Bethe’s edition (see Theodoridis 2003, 77).