πέπραγα, πέπραχα, πεπραγώς, πεπραχώς
(Phryn. PS 103.12–3, Moer. π 5)

D. General commentary

The active perfect stem of the Attic verb πράττω (Ionic πρήσσω, koine πράσσω) ‘to achieve, to do; to fare in a certain way, etc.’ is discussed in Phrynichus’ Praeparatio sophistica (A.1) and in Moeris’ lexicon (A.2): both entries recommend the unaspirated stem πεπραγ- over the aspirated stem πεπραχ-. The latter form was, like other aspirated perfects, a late Attic innovationInnovative forms that eventually prevailed in the koine. However, the distinction between the two stems was not merely chronological, but also semantic and functional, as recognised by later scholars (B.2, B.4).

The aspirated perfect (on which see K–B vol. 2, 96; Chantraine 1927, 139–40; Schwyzer 1939, 771–2; Chantraine 1961, 197–9; Duhoux 2000, 401–3) is an innovation of post-Homeric date, spreading in the 5th century BCE and usually assumed to be restricted to Attic-Ionic, although several forms are attested in other dialects (see Cassio 2017 on attestations from Sicily). As a rule, the aspirated perfects are transitive and express a resultative meaning, in contrast with the older root perfects, which are intransitive and express a stative meaning. The origin of the aspirated perfect is disputed. Its starting point has been identified in roots ending in an aspirated stop, which underwent assimilation before consonant-initial endings, with aspirations surfacing only before vowel-initial endings, e.g. ind. perf. m.-p. 3rd sing. ἔστραπται, but 3rd plur. ἐστράφαται from στρέφω ‘to turn’; then, ‘the neutralised outcome was reinterpreted as representing an underlying aspirate, and root-final (non-etymological) aspiration generalised as a categorial marker in (new) transitive active perfects’ (Willi 2018, 19–20; see also Christol 1972). According to a different theory (Ringe 1984; Sihler 1995, 574–5; but see the objections raised by Slings 1986), the reanalysis of aspiration as a marker of the transitive perfect occurred when the perfect εἴληφα (originally from a root *labʰ-, on whose possible traces in Greek nominal forms see Batisti 2022) was attracted into the paradigm of λαμβάνω, ἔλαβον (from the root *(s)lagʷ-).

The Attic verb πράττω is probably derived from the root *per(h₂)- ‘to go through’, with a velar enlargement either in -κ- (which would more straightforwardly account for the consonantism -ττ-/-σσ- < *-k-i̯- in the present stem) or in -γ-, as suggested by the s-stem noun πρᾶγος ‘deed, affair’ (Pi., Aesch.+) and by the root perfect πέπραγα itself (see DELG, EDG s.v.; Barber 2013, 266). Like several other verbs whose stem ends in an unaspirated labial or velar stop, πράττω developed an innovative ‘aspirated perfect’ πέπραχα alongside the older root perfect πέπραγα. Indeed, the stem πεπραγ-, first attested in Pindar (C.1), is, down to Xenophon, the only form attested in Attic literature. It occurs with particular frequency in Euripides (14x, e.g. C.4; for other attestations in tragedy, cf. Achae. fr. 41.8, Ion fr. 44, and possibly trag. adesp. fr. 342, but see F.1) and Aristophanes (8x, e.g. C.5, C.6), but is also found in Thucydides – who has two occurrences of the pluperfect ἐπεπράγεσαν, including C.7 – and in Plato (C.10). In this period, πεπραγ- is always constructed intransitively with a modal adverb, expressing the meaning ‘to fare in a certain way’: κακῶς ‘badly’, εὖ ‘well’, etc. As far as Attic literature is concerned, the transitive perfect stem πεπραχ- is first employed by Xenophon (5x, e.g. C.9). However, its Ionic cognate πεπρηχ- may already be attested in Herodotus (C.3), where πεπρηχέναι is transmitted by MSS of the Roman family, while πεποιηκέναι is transmitted by the Florentine family. Interestingly, in another Herodotean passage (C.2), where the verb is constructed intransitively, both branches of the tradition agree in transmitting πεπρηγέναι. If the textual tradition is reliable, Herodotus would thus already exhibit the functional-semantic differentiation between the intransitive unaspirated stem and the transitive aspirated one that only later becomes established in Attic. Indeed, although Xenophon’s text also contains an occurrence (C.8) of the unaspirated πεπραγ-, this instance is transitive: here πάντων, which must be the object of πεπραγότες εἶεν (‘they had obtained everything’), has been attracted to the case of the relative ὧν. From Xenophon onwards, some authors attest only the innovative aspirated stem, usually in a transitive sense, while others attest only the older unaspirated stem, usually in the intransitive sense. Some oscillations do occur, however, and insofar as these are not the result of textual corruption, they may point to a period of uncertainty following the introduction of the innovative variant. In Plato Comicus (C.11), for instance, the transmitted reading πέπραχε is generally corrected by editors to πέπραγε (see F.2). In Aristotle and the Aristotelian corpus, where both stems are attested, πεπραχ- (10x) is twice as frequent as πεπραγ- (5x), but both are used only transitively (often as absolute transitives, in the sense of ‘to have acted’). In the Rhetorica ad Alexandrum (traditionally attributed to Aristotle, but most likely by Anaximenes of Lampsacus), the distribution appears to be reversed, with transitive πεπραγότας (35.2) and intransitive πεπραχότας (34.15; Fuhrmann 2000 ad locc. report no textual variants for either passage). In Attic inscriptionsInscriptions there are only two late attestations of the stem πεπραχ- (IG 2³,1.1329.13 [173/172 BCE] and IG 2².1043.67 [38/37 BCE]; see Threatte 1996, 565), whereas the innovative stem is more widely attested in koine inscriptions (the earliest occurrence is πεπράχασιν in IG 9,2.517.36 [Larissa, 214 BCE], transcribing a letter of King Philip V of Macedonia).

The innovative stem, with its expected transitive sense, prevails in later Attic prose (C.12) and drama (C.13), and above all in koine prose, including Polybius (3x), the Septuagint (Prov. 30.20.2), Diodorus Siculus (7x), Philo (4x), and Josephus (8x). In the New Testament, πέπραχεν and πεπραχέναι occur in similar contexts in the Acts of the Apostles (25.11: εἰ μὲν οὖν […] ἄξιον θανάτου πέπραχά τι, ‘if I have done anything deserving death’; 25.25: μηδὲν ἄξιον αὐτὸν θανάτου πεπραχέναι, ‘to have done nothing deserving death’), in both cases with a transitive sense. Plutarch, by contrast, employs both stems (unaspirated 2x, aspirated 4x) with an intransitive sense (cf. e.g. Sert. 3.10: ἀπαντᾶν εὖ πεπραχόσι φίλοις καὶ πολίταις, ‘to meet successful friends and fellow citizens’). The papyrological evidence confirms that πεπραχ- was the norm in the koine: in Ptolemaic papyriPapyri, there are no occurrences of πεπραγ- (Mayser, Gramm. vol. 1,2, 151). As far as the imperial period is concerned, a search on the website papyri.info returns only six occurrences of the stem πεπραγ- between the 1st and the 3rd centuries CE, whereas πεπραχ- has the same number of occurrences over the same time span (for the spread of aspirated perfects as attested in the papyri, see Gignac 1981, 300–2).

The chronological distribution of the attestations suffices to explain why the Atticists directed their preference towards πέπραγα (it may be noted in passing that προκρίνουσι ‘they prefer’ in A.1 is a hapaxHapax in Phrynichus’ evaluative terminology), while viewing πέπραχα with suspicion. What the Atticists do not appear to have explicitly noted, however, is the semantic and functional distinction between the two stems, which are neither merely formal variants nor diachronic alternatives. The perception of πέπραγα as the genuinely Attic form and of πέπραχα as the non-Attic one was perhaps reinforced by the presence, in classical authors, of unaspirated forms used with a transitive meaning (Pape, Sengebusch 1880 s.v., with reference to C.8). It could be hypothesised that another source of confusion lay in passages such as C.6, where the intransitive πέπραγα is accompanied by an adverbial neuter, which could be liable to be interpreted as the verb’s object, thereby allowing a transitive reading. Thus, for example, πάντ’ ἀγαθὰ πεπράγαμεν could be understood as ‘we accomplished all the good things’, rather than as ‘all is well with us’. Such a hypothesis, however, can be advanced only with caution, given the admittedly uncertain likelihood of such a misunderstanding on the part of native speakers. Be that as it may, the view that the Atticists regarded πέπραγα as the more correct form may also be supported by the fact that Phrynichus (B.1), apparently adapting a line from the pseudo-Aeschylean Prometheus Bound (263) so that the (intransitive) present active participle πράσσοντ(α) is turned into a perfect, uses the unaspirated form πεπραγότι (but see F.1 for alternative interpretations). Likewise, a rigorously Atticising author such as Aelius Aristides confines himself to a strict imitation of classical practice, consistently using intransitive πέπραγα with an adverb in the sense ‘to fare in a certain way’, as noted by Schmid (Atticismus vol. 2, 32). Cassius Dio, by contrast – an author whose linguistic choices were nonetheless influenced by Atticism – adopts yet another approach: he employs both stems, but successfully distinguishes intransitive πέπραγα (4x) from transitive πέπραχα (8x).

Nevertheless, other ancient grammarians, operating with a different focus, did notice the semantic and functional contrast between aspirated and unaspirated perfects. Benedetti (2020) has recently pointed out that Theodosius – possibly drawing on an unknown earlier tradition – postulated a three-way distinction between an ‘active’ perfect τέτυφα, a ‘middle’ perfect τέτυπα, and a ‘passive’ perfect τέτυμμαι (using his example of τύπτω ‘to strike’). Theodosius himself remarks (B.2) that earlier grammarians failed to recognise this tripartite structure because the middle is not always formally distinct from the other voices. While incongruous from a morphological perspective, this distinction captures the fact that root perfects are diathetically indifferent, since they can encode both active and passive-like functions, whereas aspirated (as well as kappatic) perfects and middle perfects are restricted to transitive and intransitive uses, respectively (on ancient analyses of unaspirated perfects as middle, see entry ἀνέῳγεν, ἀνέῳκται). Even though Theodosius’ distinction is formulated in an abstract manner, Benedetti shows that it is borne out by the textual evidence. It is all the more remarkable, therefore, that the Atticists of the imperial period, owing to their different linguistic ideologies and aims, ignored the semantic contrast between πέπραγα and πέπραχα. However, it should be recalled that Phrynichus’ entry (A.1) is transmitted only in the epitomeEpitome of the Praeparatio sophistica; there is thus a strong probability that the original entry contained a fuller discussion.

By contrast, an entry in Thomas Magister’s lexicon (B.4) distinguishes between πέπραχα and πέπραγα, stating that the former has a passive (rather than active) meaning only when used in an adverbial construction, whereas the latter is always passive, with or without an adverb. This rule, not unlike Theodosius’, shows a good acquaintance with the sources. Indeed, πέπραγα is virtually always intransitive in classical authors, with only a few and dubious exceptions, whereas πέπραχα is normally active and transitive, though it can occasionally occur intransitively with modal adverbs in the construction classically reserved for πέπραγα. This latter usage is attested not only in imperial-period authors such as Plutarch (see above), but also in Libanius, an author much admired by Thomas Magister: cf. [Lib.] Progymnasmata 4.2.9: τοῖς Ἕλλησι καίτοι κακῶς πεπραχόσι θαρρῆσαι παρασχεῖν, ‘to provide the Greeks, even though they fared badly, with a reason to take heart’ (a work judged spurious by many modern scholars: see Gibson 2008, 87).

E. Byzantine and Modern Greek commentary

By the beginning of the Byzantine period, the synthetic perfect had largely disappeared from the spoken language and was being replaced by new periphrastic forms. As a result, occurrences of the synthetic perfect, while not infrequent, are mostly confined to higher-register texts, where they alternate freely with the aorist and are often selected on stylistic or metrical grounds (see CGMEMG vol. 3, 1761; Hinterberger 2014). During this period, both the aspirated and the unaspirated perfect stems of πράσσω occur in literary sources, with the former being more abundantly attested. Although several authors employ both stems, a functional distinction between them is seldom discernible. As a rare exception, one may cite Eustathius, who uses the unaspirated stem with modal adverbs in an intransitive sense (5x), and once the aspirated stem in a transitive sense (Orationes 10.57.10 πεπραχὼς δέ τι ‘having done something’).

The functional equivalence between the perfect and the aorist is also reflected in the fact that late antique and Byzantine lexica often gloss the former with the latter (Hinterberger 2014, 190–1): in the case of πράσσω, this equivalence is explicitly stated in a lemma of Hesychius’ lexicon (B.3), where, moreover, the innovative aspirated form πεπράχαμεν is lemmatised instead of the classical πεπράγαμεν. As an example of this equivalence in Byzantine authors, see Philip Monotropus (11th century), who uses ὅσα πέπραχας (Dioptra 1.5.852) to avoid hiatus, but switches to ἅπερ ἔπραξεν a few lines later (1.5.871), with no apparent semantic difference (see Hinterberger 2014, 197).

F. Commentary on individual texts and occurrences

(1)    Phryn. PS 67.5–6 (B.1)

The lemma in this entry of the Praeparatio sophistica – whose aim is to illustrate the meaning ‘easy’ of the adjective ἐλαφρός, literally ‘light’ – is an iambic trimeter strongly reminiscent of [Aesch.] PV 263–5: ἐλαφρόν […] παραινεῖν νουθετεῖν τε τὸν κακῶς | πράσσοντ’ (‘[It is] easy […] to advise and rebuke one who is faring badly’). De Borries (1911, 67) argued that the lemma is, in fact, an abridgement of the Aeschylean lines, rather than a quotation from an unknown comic or tragic playwright, as assumed by Kock (CAF vol. 3, 621) and Nauck (1889, 903–4), respectively. As Snell and Kannicht (TrGFr vol. 2, 106) point out, in Phryn. PS 64.8Phryn. PS 64.8 another tragic line (Eur. Hipp. 161) is similarly abridged for the sake of lemmatisation. West (1979, 141), however, attributed this line to the Prometheus Unbound, which he dated to around 440 BCE, assuming that the author of the Prometheus trilogy (whom he does not identify with Aeschylus) had a ‘habit of repeating himself’. In this case, πεπραγότι would represent one of the earliest attestations of the perfect stem of πράττω in Attic, exhibiting – as expected – the unaspirated variant.

(2)    Pl.Com. fr. 203 (C.11)

These lines from an unidentified play by Plato Comicus are quoted twice by Plutarch (Nic. 11.6, Alc. 13.9), whose codd. transmit either πέπραχε or πεπράχθαι. The transmitted reading was corrected to πέπραγε by Meineke (1823, 221), who observed that while a ‘non-Attic’ form such as πέπραχε is not out of place in Menander (cf. C.13), in Plato the older variant is required. Cobet (1840, 140) likewise argued for the reading πέπραγε, noting, however, that it is not ‘more elegant or more Attic’ than the aspirated variant, but rather the only form used intransitively. Later editors agree in accepting Meineke’s reading; see further Kassel, Austin PCG vol. 7, 520–1 ad loc.