διδόασι, διδοῦσι
(Phryn. Ecl. 215, Antiatt. δ 8)

D. General commentary

Phrynichus (A.1) condemns Favorinus’ use of διδοῦσι as the 3rd person plural of the present indicative of δίδωμι (‘to give’) and prescribes διδόασι instead. By contrast, the Antiatticist (A.2) appears to favour διδοῦσι (on this entry’s actual level of prescriptiveness, see below) in light of a comic passage of dubious attribution (C.2). A.1 belongs to a broader group of entries in Atticist lexica that prescribe the ending -ᾱσι for the 3rd person plural of the present indicative of athematic verbs (see entry ἀπολλύασιν, ἀπολλύουσιν, and other 3rd person plurals of -νυμι verbs; La Roi 2022, 209–13).

Greek athematic verbs form the 3rd person plural of the present indicative with the ending *-(e)nti (see Latin -(u)nt, Vedic -a(n)ti, Hittite -anzi) ‘with ablaut variants depending on paradigmatic patterns’ (Willi 2018, 4): the full-grade *-enti can be observed in consonantal roots, e.g., *h1es-enti > Attic εἰσί (‘they are’), while the zero-grade *-nti may occur in postvocalic positions. Moreover, some have argued (see Duhoux 2000, 481) that forms such as δεικνύουσι (from δείκνυμιδείκνυμι, ‘to show’) and ὀμνύουσι (from ὄμνυμιὄμνυμι, ‘to swear’) should be explained as resulting from the o-grade of the same ending, i.e. *-onti, although most modern scholars interpret them as thematised forms (see entry ἀπολλύασιν, ἀπολλύουσιν, and other 3rd person plurals of -νυμι verbs and below). Finally, the ending -ᾱσι shown by many athematic verbs (see below) results from postconsonantal zero-grade *-nti, as follows: *-(C)n̥ti > *-(C)ati and, with the analogical insertion of -n-, > *-(C)anti > Att.-Ion. -ᾱσι (see Duhoux 2000, 481, 485; Willi 2018, 4).

In Attic, the reduplicated athematic verbs mostly use this latter ending for the 3rd person plural of the present indicative, resulting in forms such as διδόᾱσι (‘they give’, from δίδωμι), τιθέᾱσι (‘they put’, from τίθημιτίθημι), ἱᾶσι (‘they let go’, from ἵημιἵημι), and ἱστᾶσι (‘they make to stand’, from ἵστημιἵστημι). The same occurs for the verbs in -νυμι-νυμι (see entry ἀπολλύασιν, ἀπολλύουσιν, and other 3rd person plurals of -νυμι verbs) and the verb εἶμιεἶμι (see ἴᾱσι, ‘they go’). However, outside of Attic texts and documents (and particularly in Ionic, see below), the 3rd person plural of these verbs is formed differently (e.g. τιθεῖσι, ἱεῖσι, δεικνύουσι, σβεννύουσι, see below). In particular, the form διδοῦσι is attested already in Homer (Il. 2.255, 19.265; Od. 1.313, 8.167, 17.450, 18.279) and Hesiod (Th. 219, 915; Op. 225), and its interpretation is not univocal. If it is explained as athematic (i.e. from the zero-grade ending *-nti > δίδοντι > δίδονσι, Mycenaean di-do-si), the form should be proparoxytone (*δίδουσι), and the circumflex accent can only be understood as resulting from analogyAnalogy with thematic contract verbs (see e.g. Rix 1992, 252; Wachter 2009, 100). Meanwhile, some scholars interpret the form as belonging to a fully thematic paradigm (as if from **διδόω: see B.5; Schwyzer 1939, 687; Hackstein 2002, 112; Willi 2012, 264−5), thus understanding it as the result of the tendency (observed throughout the history of Greek, especially in Ionic) towards the replacement of athematic presents and imperfects with thematic equivalents (see e.g. Duhoux 2000, 325−6; Horrocks 2010, 76).

On a quantitative level, διδόασι is by far the most widely attested form throughout the Greek literary corpus. Aside from epic poetry (see above), διδοῦσι occurs in authors who write in IonicIonic Greek (Herodotus, 10x) or who have inherited the Ionic features of the epic diction (Theognis, 6x), including one instance in Callimachus (fr. 475); it also occurs in authors of the imperial period (see below). Regarding documentary texts, διδοῦσι is found in inscriptions from Ionic-speaking areas, dating from the 4th century BCE onwards: see SEG 45.1000.4 [Pontic Olbia, 392−380 BCE], I.Magnesia 38.56 [Magnesia on the Maeander, 3rd−2nd century BCE], SEG 34.1213.5–6 [Saetae, Lydia, 199−200 CE]. The form is attested also in papyri: see P.Cair. Zen. 3.59509.12 (= TM 1147) [Philadelphia, 3rd century BCE], O.Strasb. 1.788.7 (= TM 76365) [northern Egypt, 2nd century CE], P.Merton 2.82.13 (= TM 28784) [unknown origin, late 2nd century CE], SB 26.16831.8–9 (= TM 97188) [unknown origin, 4th century CE?), P.Michael. 41.40 (= TM 21416) [Aphrodites Kome, 6th century CE] (for other examples of the transition of athematic presents to the thematic type in papyri, see Gignac 1981, 280−1).

The presence of διδοῦσι in authors from the imperial period (see e.g. Plu. De communibus notitiis adversus Stoicos 1064d; Gal. De sem. 1.15.1 De Lacy (4.563.17 Kühn); Vett.Val. Anthologiarum libri ix 2.39.13; D.C. 65.13.5) demonstrates that the usage condemned by Phrynichus (A.1) was not peculiar to Favorinus but could also be found in other writers. Still, Phrynichus’ criticism is directed at Favorinus because the latter, as a rhetor of the Second Sophistic, was expected to adhere to a higher, more Atticist, linguistic standard (on Favorinus as Phrynichus’ polemical target, openly criticised in 17 of the Eclogue’s entries, see entries εἶμι, ἐλεύσομαι, νῆες, ναῦς, νῆας, υἱεύς, υἱέως, υἱέα). The aim of Phrynichus’ rule is clearly to condemn διδοῦσι and – although this is not explicitly stated in the text – to prescribe διδόασι as the 3rd person plural of the present of δίδωμι. The lexicographer’s pedantic addition that διδοῦσι ‘means something else’ (i.e. that διδοῦσι is the dative plural of the present participle of δίδωμι) was misunderstood by a later scribe, who assumed that the verb δέω (‘to bind’) was intended and added τὸ δεῖν to the text of the entry (see Rutherford 1881, 315).

As far as the Antiatticist is concerned, A.2 may initially appear to prescribe διδοῦσι rather than διδόασι, but, in reality, the negation οὐ is intended ‘to introduce a (rare) possibility also attested in the classical writers’ (Valente 2015, 48): in other words, in A.2, οὐ is in all likelihood equivalent to οὐ μόνον. Regarding the entry’s content, A.2 is the only extant testimony of the form διδοῦσι’s occurrence in Attic comedyComedy (on the authorship of the play Metrophon quoted by the Antiatticist, see F.1). Athematic verbs (especially those in -νυμι) are frequently thematised in the present in Middle Comedy but considerably more rarely in Old Comedy (only two examples are extant, i.e. Pherecr. fr. 152, on which see entry ἀπολλύασιν, ἀπολλύουσιν, and other 3rd person plurals of -νυμι verbs, and Ar. Pl. 718; see also Willi 2003, 248). One may interpret the information provided by the Antiatticist in three ways: first, the occurrence of διδοῦσι in the text of the Metrophon may have been justified by ‘some special context’ (Olson 2022, 201), possibly the linguistic characterisation of an Ionian figure within the play (see F.1); second, διδοῦσι occurred as an unmarked (although clearly non-standard) form; and third (but less likely), διδοῦσι may have been an erroneous varia lectio in the text available to the Antiatticist.

A degree of uncertainty regarding the dialectal affiliation of διδόασι and διδοῦσι emerges from later erudite texts. Theodosius (B.1) ascribes the forms τιθέασι, διδόασι, and ζευγνύασι to the Ionians, while Choeroboscus, commenting on Theodosius, states that Herodian regarded διδόασι as Ionic and criticises him for this (see Choerob. in Theodos. GG 4,2.334.23−32 = Hdn. Περὶ τῶν εἰς μι GG 3,2.833.33−40). Chrysoloras also defines the form as Ionic in his grammar (B.7). Instead, Photius (B.4) correctly indicates διδοῦσι as the form used by ‘the Ionians and Herodotus’. The source of Photius’ entry is conjecturally identified by Theodoridis (1982, 398) with Aelius Dionysius, in light of the numerous entries in which the lexicographer distinguishes between a word’s Attic and Ionic forms (see e.g. α 67Ael.Dion. α 67, α 178Ael.Dion. α 178, α 203Ael.Dion. α 203, β 21Ael.Dion. β 21, δ 12Ael.Dion. δ 12, ζ 4Ael.Dion. ζ 4, η 3Ael.Dion. η 3, κ 6Ael.Dion. κ 6, ν 11Ael.Dion. ν 11, ο 20Ael.Dion. ο 20; see also entry Aelius Dionysius, Ἀττικὰ ὀνομάτα).

E. Byzantine and Modern Greek commentary

As far as popular speech is concerned, the athematic verbs’ shift towards the thematic conjugation is complete by the early Middle Ages (see Horrocks 2010, 303; CGMEMG vol. 3, 1273). However, with regard to written language, and specifically as regards δίδωμι, the Atticists’ proscription of 3rd-person plural διδοῦσι in favour of διδόασι is obeyed in the subsequent centuries. Indeed, although διδοῦσι remains in use into late antiquity, the Byzantine period, and beyond, διδόασι remains the more attested of the two forms up until the 15th century. Vernacular texts from roughly the 12th to the 18th century reveal the gradual replacement of all the forms of διδῶ with their paroxytone equivalents (as a result of the reanalysis of δίδομεν and δίδοτε as thematic): διδοῦσι is thus replaced by δίδουσι. Alternative forms of the verb ‘I give’ are also attested for specific areas: Cretan texts from the 16th and 17th centuries attest to the variant δούδω (3rd-person plural δούδουσι; see CGMEMG vol. 3, 1377); texts from Cyprus dating from the medieval to the early modern period have διδῶ (see CGMEMG vol. 3, 1309, 1377), while Aegean texts dating between the 16th and the 18th centuries have δώνω (3rd-person plural δώνουν; based on the aorist ἔδωσα, which replaced the ancient ἔδωκα), with the nasal suffixSuffixes typical of verbal forms created in this period (the form was probably reinforced by Latin – and Italian – dono: see CGMEMG vol. 3, 1377). The competition between δίδω and δώνω eventually resulted in the Modern Greek δίνω (3rd-person plural δίνουν), which is attested from the 15th century onwards, initially in south-western areas and the Aegean (see Horrocks 2010, 305; CGMEMG vol. 3, 1376).

F. Commentary on individual texts and occurrences

(1)    Antiatt. δ 8 (A.1), Antiph. fr. 154 (C.2)

The MS reads Ἀριστοφάνης Μητροφῶντι (‘Aristophanes in the Metrophon’; regarding the personal name Μητροφῶν, see Olson 2022, 200 and below). Given that no such title is known for Aristophanes, Meineke and Bergk have queried the reading (see Meineke FCG vol. 1, 328; FCG vol 2,2, 899): the former proposed Ἀντιφάνης (accepted by Kassel, Austin PCG vol. 2, 295; the entry is thus excluded from Fiori’s study of Aristophanic quotations in the Antiatticist, see Fiori 2022, 22), while the latter suggested Ἀριστοφῶν. Although both authors wrote plays with personal names as titles (see e.g. Antiphanes’ Cleophanes [fr. 120], Lampon [frr. 136−7], and Leonides [fr. 141], as well as Aristophon’s Babias [fr. 1], Callonides [fr. 6], and Philonides [fr. 13]), neither Antiphanes nor Aristophon is known to have composed a Metrophon. At first glance, the emendation Ἀντιφάνης may appear more plausible given that Antiphanes is among the authors most frequently quoted by the Antiatticist (26x) and that the two names are frequently confused in the textual tradition (for a list of such cases, see Kassel, Austin PCG vol. 2, 313). However, a misunderstanding of the playwright’s name in the entry regarding διδοῦσι would be rather surprising precisely because of the name’s numerous occurrences within the lexicon. Conversely, the emendation Ἀριστοφῶν presupposes a banal misunderstanding of the (presumably abbreviated) desinence, resulting in the replacement of the rare Ἀριστοφῶν (otherwise absent from the Antiatticist) with the common Ἀριστοφάνης (which occurs in 46 other entries of the Antiatticist).

The name Metrophon occurs almost exclusively in inscriptions from areas in which Ionic was spoken, including southern Thrace (see I.Thrake Aeg. E115.1 [Stryme?, 450−400 BCE] and I.Thrake Aeg. E144.1 [Stryme?, 375−350 BCE]), the Aegean islands (see e.g. IG 12,6.1231.2 [Ikaria, 4th century BCE]), and Asia Minor (see e.g. I.Milet 955b.3 [late 3rd century BCE]): on a speculative level, we might suppose that the form διδοῦσι was employed as a means of portraying the eponymous character’s dialect as Ionic. Were this the case, however, the Antiatticist would have to have been unaware of such characterisation in citing the passage as proof that διδοῦσι was correct Attic. Alternatively, the lexicographer did not have access to the play in its entirety and was obliged to work on second-hand material and thus unable to recognise διδοῦσι as an element of a broader linguistic characterisation.