δεδιακόνηκα, διηκόνηκα, ἐδιακόνουν, διηκόνουν
(Antiatt. δ 1, Antiatt. ε 2, Moer. δ 10, Moer. δ 21)

D. General commentary

Two entries in the Antiatticist (A.1, A.2) concern the position of reduplication and augment in the forms of the verb διακονέω (‘to serve’); the same topic is addressed by Moeris (A.3), Philemon (B.1), and Orus (B.2). The issue at stake is whether διακονέω should be treated as a prefixed verb, taking ‘internal’ reduplication (perf. διηκόνηκα) and augment (impf. διηκόνουν, aor. διηκόνησα), or as a simple verb, taking ‘external’ reduplication and augment (perf. δεδιακόνηκα; impf. ἐδιακόνουν, aor. ἐδιακόνησα).

The verb διᾱκονέω (Ionic διηκονέω) ‘to serve’ is, at face value, a denominal present derived from the noun διάκονος (Ionic διήκονος) ‘servant, messenger, minister’, which in turn appears to be a prefixed agent noun comparable to ἀμφίπολος ‘(female) attendant, servant’ (an old Indo-European term cognate with Latin anculus and Sanskrit abhicara-, from the root *kʷelh₁- ‘to turn, go round’; see, most recently, Olsen 2023, 311). However, it cannot be excluded that the derivational relationship is in fact the reverse, i.e. that διάκονος is deverbal to διᾱκονέω. In that case, διᾱκονέω would be an iterative-causative formation in *CoC-éi̯e/o-, perhaps built on a lost root present (see DELG and EDG s.v.; Tucker 1990, 78; Willi 2018, 259 n. 138). The root may be reconstructed as *ken(h₁)- (see LIV s.v.), attested in the (mostly poetic) verb ἐγκονέω ‘to be quick, to hasten’, perhaps also in Mycenaean ka-si-ko-no if this represents /kasi-konos/ ‘co-worker’ (see DMic s.v.), and in Latin cōnor, -ārī ‘to exert oneself, to try’ (see EDG s.v. ἐγκονέω for other cognates within and beyond Greek). Another possible prefixed verb from the same root is Laconian ἀγκονίω (= Attic-Ionic *ἀνακονέω?), attested in the feminine participle ἀγκονίωαἱ (= Attic-Ionic *ἀνακονέουσαι) in Ar. Lys. 1310 (where the codd. have -έουσαι or -εύουσαι, while -ίωαἱ is Dindorf’s conjecture; see Colvin 1999, 151). This participle, however, is usually interpreted as meaning ‘raising dust’, i.e. as a compound of κονίω ‘to cover with dust’ < κόνις ‘dust’; but cf. Hsch. α 552: ἀγκόνους· διακόνους, δούλους, ‘ἀγκόνους: servants, slaves’. It should be noted that traces of a simple present κονέωκονέω and of its agent noun κονητής ‘servant’ are preserved in lexicography (Hsch. κ 3502: κόνει· σπεῦδε, τρέχε, ‘κόνει [imper. pres. act. 2nd pers. sing.]: ‘Hurry up!’, ‘run!’’; Hsch. κ 3503: κονεῖν· ἐπείγεσθαι. [αἰσθάνεσθαι]. ἐνεργεῖν, ‘κονεῖν: To urge on; [to perceive]; to be active’; Hsch. κ 3507: κονηταί· θεράποντες, ‘κονηταί: Servants’). It is difficult to determine whether this κονέω really is the simplex of ἐγ-/ἀν-/δια-κονέω, or whether it was instead abstracted from the prefixed forms. At any rate, it is unlikely that we are dealing with a derivative of κόνις involving a semantic shift ‘to raise dust’ > ‘to hasten’ (so LSJ s.v. κονέω; for another instance of confusion between these two lexical families, see now Battezzato, Della Rossa 2023 on the adverb ἐγκονητί in Pi. N. 3.36, which they show to be a late itacistic spelling for ἐγκονιτί ‘with dust’ > ‘with effort’, rather than a derivative of ἐγκονέω). Turning back to διάκονος and διᾱκονέω, it is important to note that the lengthVowel length of the ᾱ – which must at least predate the split between Ionic and Attic (cf. Ion. διηκον-) – is not fully explained. It may be due to compositional lengthening, but this process normally affects the initial vowel of the second member of a compound (see Wackernagel 1889). The ancients, too, were apparently at pains to explain the ᾱ in δικονος: Strabo (9.3.5) compared it to the metrically lengthened initial vowel of θάνατος ‘immortal’ and κάματος ‘untiring’, while Choeroboscus (B.5) devised an etymology that allowed him to explain it as the result of contraction.

As a result of its morphological opacity, διακονέω was subject to competing analogicalAnalogy pressures, so that two different developments may be reconstructed. On the one hand, if διακονέω was (or came to be perceived as) a prefixed derivative of κονέω, it would naturally take internal augment and reduplication. Even if it was in fact – or was perceived as – a denominal derivative of διάκονος (which is the more likely scenario, given the near absence of κονέω from literary and documentary sources), it could still have taken internal augment and reduplication by analogy with prefixed verbs (see Van Emde Boas et al. 2019, 276: ‘when a denominative verb is formed from a compound noun which itself includes a preposition […], it is sometimes treated as if it were formed by prefixation’). On the other hand, the length of the ᾱ, together with the apparent absence of κονέω from both literary and documentary evidence, and the lack of an agent noun *κονός, constituted an obstacle to analysing either διακονέω or διάκονος as a prefixed formation with διά (cf. Lautensach 1899, 145). Indeed, while some ancient erudite sources (Orio 39.1, 48.9; Et.Gud. 357.1–2) explain διάκονος as a compound of διά and κονέω, this derivation is expressly rejected by Choeroboscus (B.5), precisely because it fails to account for the ᾱ. Since διακονέω, regardless of its historical derivation, was perceived as not deriving from κόνος through prefixation, the external augment and reduplication that it exhibits in classical Attic are, in fact, the expected treatment. The Attic analogical tendency towards external augmentation in verbs that were originally prefixed but whose simplex had become rare (as observable e.g. in ἀνοίγω/ἀνοίγνυμι ‘to open’, which in Attic prose is much more frequent than the simplex οἴγω/οἴγνυμι: see entry ἀνέῳγoν, ἤνοιγoν) would at most have encouraged a similar treatment in the case of διακονέω. Although some verbs with an (apparent) prefix do exhibit double augmentation in Attic, διακονέω is not among them (as already recognised by Pierson 1759, 122), despite Rutherford’s (1881, 83, 86) argument that it should, and that κἀδιηκόνουν ought therefore to be read in C.1. In fact, forms with double augmentation of the type ἐδιηκόνουν are only sporadically attested in the medieval period (see E.).

The analogy with prefixed verbs in δια-, however, eventually prevailed in the koine, producing an augmented (and later a reduplicated) stem δι-ηκον-. It should be noted that the latter implies an analysis of the present as formed from δι(α)- plus a vowel-initial stem κονε-ε/ο- (for η as the temporal augment of verbal stems with initial ᾱ, cf. ἠρίστησα < ριστάω ‘to have breakfast’ or ἡλισκόμην < λίσκομαι ‘to be caught’). In Post-classical Greek, by contrast, an augmented stem δι-ε-κον- (first attested in Eus. Commentaria in Psalmos 23.285.16) and a reduplicated stem δια-κε-κον- (first attested in Thdt. Qu.in Ex. 29) occasionally occur, presupposing an analysis as δια- + κονε-ε/ο-. The latter analysis would have become much easier after the loss of distinctive vowel quantity; it is therefore unsurprising that it is attested only after this feature had disappeared from the spoken language (a development difficult to date with precision, but probably completed by the 3rd or 4th century CE: see Allen 1987, 26–8; Horrocks 2010, 160–3). Consistently with this analysis, Eustathius (B.7; cf. also in Od. 2.20.8–9 Cullhed–Olson = 1.199.36–7 Stallbaum) explains διηκόνουν as a long-augment variant of διεκόνουν, parallel to the alternations ἤμελλον/ἔμελλον (on which see entry ἤμελλον, ἠβουλόμην, ἠδυνάμην) and ἠσχόμην/ἐσχόμην.

The forms’ distribution in literary and documentary sources shows a clear split between classical Attic and the koine: in the former, διακονέω exhibits exclusively external augment (C.1, C.2, C.4, C.5, C.6) and reduplication (C.3, C.7). As far as Attic inscriptionsInscriptions are concerned, they follow the general rule that ‘augment and reduplication precede a nominal element of a compound but follow any prepositional prefixes’ (Threatte 1996, 496–8), with the exceptions of ἀμφισβητέω, which takes a double augment, and of κάθημαι, καθιζάνω, which take the augment before the prefix. Although forms of διακονέω are rare in Attic inscriptions, a 3rd-person dual active is legible in the honorific decree for Antiochides and Phanosthenes (IG 1³.182.18–9, 430–405 BCE): ἐπαιν[έσαι μὲν Ἀντιοχίδεν καὶ Φανοσθένε]|[ν ὅτι ἐδια]κονεσάτε[ν, ‘to commend Antiochides and Phanosthenes because they both served’. Since the inscription is written in the stoichedon style, the restitution ἐδιακονησάτην (with the expected external augment) is virtually certain.

Only διηκον- is attested for the augmented forms in koine prose, beginning with a fragment of Posidonius (C.8) transmitted by Athenaeus (4.153c). As a reduplicated stem, however, διηκον- is not attested before the 5th century CE. Indeed, the only author of this period who uses the perfect stem at all, namely Josephus, employs δεδιακον- (4x) with external reduplication, even though, in line with general post-classical practice, he uses the internally augmented stem διηκον- (3x). While διακονέω does not occur in the Septuagint, it is well attested in the New Testament, where the augmented stem (11x, e.g. Ev.Matt. 4.11) is always διηκον- (see Blass, Debrunner 1976, 56). In papyriPapyri, verbs derived from prefixed nouns generally take an internal augment by analogy with prefixed verbs, except where segmentation is ambiguous (Mayser, Gramm. vol. 1,2.109–11; Gignac 1981, 248–53). Although διακονέω itself is only scarcely attested, διηκόνησεν in SB 8.9763.24 (= TM 22899) [Hermopolis, 457–477 CE]) shows the expected internal augment. A major development in the history of this lexical family was the emergence, in Christian Greek, of the specialised meaning ‘deacon’ for the noun διάκονος. As a consequence, διακονέω acquired the additional meaningSemantic shift ‘to be a deacon (in the Christian Church)’ and eventually became far more frequent than in the classical period (the innovative meanings of both noun and verb are first attested in the Second Epistle to Timothy, 3.8, 3.10, 3.13; for further semantic developments in Medieval Greek, see E.).

Given what has been said above about the distribution of the various forms, the Antiatticist’s prescription (A.2) appears surprising: the imperfect ἐδιακόνουν, for which he cites the comic poet Alcaeus (C.4), is not an exception but in fact the only form attested in Attic authors. One may therefore suspect, with Orth (2013, 64), that the unabridged entry actually quoted Alcaeus for the use of διηκόνουν, contrasting it with a stricter Atticist norm prescribing ἐδιακόνουν. Nevertheless, closer inspection reveals no marked Atticist preference for ἐδιακόνουν. On the contrary, Moeris (A.4), in an entry dealing with the same verb but focusing on voice, prescribes the middle imperfect διηκονεῖτο over the active διηκόνει (cf. Moer. δ 20Moer. δ 20, opposing the two voices in the present; see also entry διακονοῦμαι, διακονῶ). It is noteworthy that the recommended form – just like the proscribed one – exhibits an internal augment, which runs counter to actual Attic practice but reflects the overwhelming post-classical one. Further confirmation comes from Atticising writers, who do not attest the perfect stem of διακονέω, but regularly employ the internal augment: cf. D.Chr. Or. 7.65 διηκονεῖτο, 7.68 διηκονεῖτο; Luc. Philops. 35 διηκονεῖτο, Tyr. 22 διηκονήσατο; Aristid. 3.196 Lenz–Behr (= 46.233.11–2 Dindorf) διηκόνησαν, 3.230 Lenz–Behr (= 46.247.18 Dindorf) διηκόνησεν, 3.265 Lenz–Behr (= 46.263.20 Dindorf) διηκόνησαν. This usage has been evaluated in different ways by modern scholars: while Deferrari (1916, 12) considered Lucian’s use of διηκον- to be ‘though formed on analogy, […] good Attic’, Schmid (Atticismus vol. 2, 24, on Aristides) judged it contrary to Attic usage. In fact, it seems likely that even the stricter Atticists, such as Moeris, did not enforce the use of ἐδιακον- – despite its classical attestations – perhaps because it ran counter to the analogical criterion that (seemingly) prefixed verbs should take an internal augment, and perhaps also because διηκον- was by then well entrenched in cultivated prose. On the other hand, Moeris (A.3) and possibly Philemon (B.1: only the lemma survives) rejected the perfect διηκόνηκα, which lacked comparable support from literary practice (indeed, Moeris’ proscription constitutes our earliest attestation of this form), and offered the disadvantage of being less clearly marked as a reduplicated stem. Perhaps not by chance, the distribution recommended by Moeris – augment διηκον-, reduplication δεδιακον- – is the same as that found in Josephus. If this is correct, one may hypothesise that the structure of the Antiatticist’s entry was not distorted by abridgement, but that it actually prescribed ἐδιακον- – on the strength of its classical attestations – as an alternative to διηκον-, which was preferred by some Atticists. Admittedly, the Antiatticist’s (A.1) inclusion of the perfect δεδιακόνηκα is more difficult to interpret. This may have been intended as a response to competing theories; but since διηκόνηκα is not recommended as Attic by any extant source, it is also possible that the lexicographer merely wished to signal the existence of the form. Possibly, since the perfect of this verb – unlike the imperfect and the aorist – has very few attestations overall, and mostly in non-canonical authors such as the New Comedy poet Archedicus (C.7) or Josephus, the lexicographer’s aim in this case was to show, by appealing to Demosthenic usage (C.3), that δεδιακόνηκα was an acceptable Attic form (even though the perfect indicative does not in fact occur in Demosthenes’ oration 51, the participle δεδιακονημένων does). Later sources (B.2, B.3, B.4, B.6), adopting a more descriptive approach, simply state that in Attic διακονέω exhibits both external augment and reduplication.

E. Byzantine and Modern Greek commentary

In the medieval period, the augmentation of prefixed verbs displays considerable fluctuation (see CGMEMG vol. 3, 1419 and Psaltis 1913, 206, with examples from the Byzantine chronicles): such verbs may show internal augment, external augment, both, or neither; verbs with consonant-initial prefixes tend to exhibit external or double augmentation more frequently than internal augmentation alone, which is typical of more formal registersRegister. In Modern Greek, this tendency extends to all prefixed verbs, which take the augment before the prefix. That said, in the case of διακονέω the Hellenistic and imperial-period preference for internal augmentation persists throughout the Byzantine period, with διηκον- remaining the predominant form (see Kriaras, LME s.v. διακονώ). External augmentation occurs in late vernacular texts, especially from Crete and Cyprus: cf. Anonyma Cretica, Anonymi versus Creticus de Vetere et Novo Testamento 2606 ἐδιακονηθῆκα, 5020 ἐδιακόνησες (15th–17th century); Marinus Tzanes Buniales, Ὁ Κρητικὸς πόλεμος 2.65.50 ἐδιακονίζουντα (17th century); Leontius Machairas, Ἐξήγησις τήρ γλυκείας χώρας Κύπρου, ἡ ποία λέγεται Κρόνακα, τουτέστιν Χρονικόν 590.35 ἐδιακονοῦσαν (15th century); Neophytus Rhodinus, Paraphrasis vitae patriarchae Ignatii 2.88 ἐδιακόνιεν (17th century). Of course, these late forms should be understood as outcomes of the above-mentioned tendency towards external augmentation, rather than as imitations of the classical Attic norm. Double augmentation also occurs sporadically in Byzantine texts, cf. e.g. Cyr.S. V.Io. 222.7 ἐδιηκόνησεν (6th century), as does double reduplication (Arethas Scripta minora 76.126.31 δεδιηκονημένων).

In Modern Greek, the verb διακονώ has been borrowed from the learned language with the meanings ‘to serve with devotion’ and ‘to be a deacon’ (see LKN s.v.; Andriotis 1974 s.v. collects other dialectal forms). In the colloquial register, by contrast, a new denominal verb διακονεύω ‘to beg’ was derived directly from διάκονος (which in Modern Greek is a learned variant; cf. the demotic form διάκος, created by metaplasm from the late-antique athematic variant διάκων, -ονος, comparable to γέρων > γέρος ‘old man’). The semantic shift underlying διακονεύω was influenced by the noun διακονιά ‘begging’ (cf. διακονιάρης ‘beggar’), itself motivated by the practice of deacons going from house to house to collect alms for the monastery. By contrast, the learned noun διακονία, already well attested as the abstract of διάκονος since antiquity (see DGE s.v.), retains the meanings ‘office and ministry of a deacon; work done with devotion and selflessness’. The respectively popular and learned origins of these forms are also reflected in their phonology: compare the synizesis of /i/ in διακονιά [ðjakoˈɲa] and διακονεύω [ðjakoˈnɛvo] with its absence in διακονία [ðiakoˈnia] and διακονώ [ðiakoˈno].

F. Commentary on individual texts and occurrences

N/A