II. Plutus I and Plutus II in the ancient sources

Given the complexity of the issues involved, we had best start by familiarizing ourselves with the primary sources that do point to the existence of more than one Aristophanic play entitled Πλοῦτος (‘Wealth’), before briefly looking at how modern scholarship has sought to make sense of this source material (section III). Against this background, it will be easier to understand what Didymus (sections IV–V) and other ancient commentators (section VI) may or may not have read, thought and written on the matter.

Although most of our discussion will focus on the scholia, these are not the only witnesses for the erstwhile existence of two Plutus plays. Firstly, the Ἀριστοφάνους γένος καὶ κατάλογος τῶν αὐτοῦ ποιημάτων, which is prefaced to some Aristophanic manuscripts, has an entry Πλοῦτος βʼ (= ‘Plutus 2x’), paralleling the entries Aἰολοσίκων βʼ, Θϵσμοφοριάζουσαι βʼ and Nϵφέλαι βʼ (Proleg. de com. XXX^a Koster = Ar. test. 2a KA). Secondly, a second-century AD papyrus from Oxyrhynchus, which contains a somewhat incomplete list of the plays of various comic poets, adds αʼ to the Aristophanic entry Πλ]οῦτ[ο]ς; and because such a number is not found with other titles (such as the preceding Ὄρν]ιθϵς and Προ]αγών), save that the line for Clouds also reads Nϵ]φέλαι βʼ, we must in this case assume that the point of the added αʼ was not to ‘count’ the number of Plutus plays, as could potentially be argued for Nϵ]φέλαι βʼ, but rather to specify that ‘Plutus I’ is meant, in implicit opposition to ‘Plutus II’, whatever the reason for the absence of the latter from the list (P.Oxy. 2659 fr. 1 col. i.13 + fr. 2 col. i = CGFP 18 = Ar. test. 2c KA).Footnote ¹ Thirdly, the Byzantine grammarian Choeroboscus, in his commentary on the metrical handbook of Hephaestion, at one point observes that there is a first and second Aἰολοσίκων ‘just as there is a first and second Πλοῦτος’ (in Hephaestionis enchiridium 9, p. 236 Consbruch, Aἰολοσίκων δρᾶμα γέγονϵ πρῶτον καὶ δϵύτϵρον Ἀριστοφάνους, ὡς καὶ ὁ Πλοῦτος πρῶτον καὶ δϵύτϵρον); and, somewhat earlier, both a scholion on Il. 23.361 based on Herodian and a passage in Athenaeus explicitly cite lines from ‘our’ Plutus as belonging to the Πλοῦτος δϵύτϵρος (schol. (A) Hom. Ψ 361a¹ with Ar. Plut. 991; Ath. 9.368d with Ar. Plut. 1128; cf. Ar. Plut. test. ii KA).

Much more informative, meanwhile, is the testimony of several Aristophanic scholia. The first of these explains a passage in Frogs where Dionysus is remembering the fun he had watching the torch race at the Panathenaea and seeing how the people from the deme of Kerameis stood at the gate and slapped with the flat of the hand (Ar. Ran. 1096, ταῖσι πλατϵίαις) some untrained, fat and slow straggler:

The parallel passage quoted here does not occur in the extant Plutus. The latter must therefore, by implication, be assumed to be a/the second Plutus in the eyes of whoever wrote this note on Frogs (cf. further section V).

However, the real nut to crack is no less than four scholia on the extant Plutus itself, which differentiate between two plays of this name (nos [2], [3], [4], [7] below). What is so odd about these, and has generated much scholarly discussion, is that they unambiguously express the belief that the play they are commenting on is the first of the pair, to be dated to 408 BC, in contrast to the second belonging to 388 BC. The same view is shared by two further Plutus scholia (nos [5], [6] below), which do not actually refer to two homonymous comedies, but at least one of which would make no sense either if there were not also a Plutus of much later date:Footnote ⁴

III. The case for a διασκϵυή

The existence of two Aristophanic plays sharing the same title reveals little about the relationship between them. Thinking of the first and second Clouds, where we know that the extant Clouds II was a revised version (διασκϵυή) of the unsuccessful Clouds I (cf. below), with Clouds II never actually seeing the stage (in Athens itself, that is), we may envisage a similar connection between Plutus I and Plutus II; yet, if in the case of Thesmophoriazusae I and Thesmophoriazusae II two much more, or entirely, independent plays seem likely,Footnote ⁵ that too is a possibility worth reckoning with for the two Πλοῦτοι.

Having said that, sources [2] and [3] favour the διασκϵυή hypothesis. The two versions referred to there cannot have been fundamentally different from each other if the annotator was able to single out specific and, at least in the case of [2], ostensibly insignificantFootnote ⁶ divergences; it would not have made sense for him to note that lines 115 and 119 ‘have been changed’ in Plutus II unless the rest of the passage in which these lines stood was by and large identical in the two texts compared. A similar conclusion is intimated by sources [4] and [7] because the suspicion that the topical allusions to a mercenary force in Corinth and to the Phyle episode might have been transferred from Plutus II presupposes that they occurred there in an identifiably similar context; otherwise one would not speak of a transfer. As for [5], no mention is made of a first or second Plutus, but if the point is that the Plutus under consideration, which is believed to be Plutus I as the other passages show, contains an anachronistic prosopographical detail about a well-known courtesan, the commentator may well have thought of a situation parallel to the one in [4] and [7], and hence again of some form of ‘transfer’ from Plutus II (cf. section IV).

While these considerations suggest a considerable degree of similarity, it would however be wrong to deny the need to think of Plutus II as a substantially reworked play. It is true that the scholia just discussed do not prove more than that there were alternative readings in this or that line, the kind of difference usually designated in the scholia by the note γρ[άφϵται] X. Yet, the very fact that the responsible scholiast did not resort to this procedure means that, although the passage in which, for example, Plut. 115 and 119 occurred was hardly affected by the revision, other parts must have been, and to an extent that justified the labelling of the two versions as Plutus I and Plutus II, just as was the case with Clouds I and Clouds II, the latter of which, according to the hypothesis, ‘is the same play as the earlier one, but has been revised in part because the poet wanted to stage it again, without in the end doing so for whatever reason’ (argum. I Ar. Nub.: τοῦτο ταὐτόν ἐστι τῷ προτέρῳ, διϵσκϵύασται δὲ ἐπὶ μέρους, ὡς ἂν δὴ ἀναδιδάξαι μὲν αὐτὸ τοῦ ποιητοῦ προθυμηθέντος, οὐκέτι δὲ τοῦτο δι’ ἥνποτϵ αἰτίαν ποιήσαντος). In fact, the differentiation in the same Clouds hypothesis of two types of interventions, namely (i) a relatively unintrusive διόρθωσις of ‘more or less every part of the play’ (σχϵδὸν παρὰ πᾶν μέρος) in which some things were removed and others rearranged as far as the order and the distribution of speaking parts were concerned, vs (ii) a more fundamental διασκϵυή affecting entire sections of the play (namely the parabasis, the agon and the final scene), would correspond well with what we may have to assume for Plutus, too, on the basis of the above reasoning.Footnote ⁷ Moreover, the handful of fragments which are ascribed in various lexicographical sources to the Plutus, but which do not overlap with anything in the text we have and therefore stand a good chance of belonging to the ‘other’ Plutus, are compatible with the same scenario.Footnote ⁸ Alongside these, the only fragment of Plutus that does not belong to the extant play and presupposes a revision process which went beyond some superficial rewording is Ar. fr. 459 from schol. Ar. Ran. 1096b (= [1] above). But as long as we have to accept some changes of substance anyway, in order to justify the repeated scholiastic reference to two different Πλοῦτος compositions, this too cannot invalidate the economical διασκϵυή hypothesis.

A bigger problem arises when we turn to the question of chronology. As already noted, sources [5] and [6] unambiguously assign the production of the earlier Plutus to the archonship of Diocles, or 409/8 BC: [6] does so by designating the year 410/9, when Glaucippus was the eponymous archon as ‘the year before this’, and [5] implies as much when it highlights the absurdity that a girl who was 14 years old in 409/8 could have been famous enough to be mentioned by name in a comedy. While maintaining the revision hypothesis, we might therefore attribute Plutus I to 409/8 BC and its διασκϵυή Plutus II to 389/8 BC.Footnote ⁹ Along these lines, Douglas MacDowell proposed that, in creating Plutus II, Aristophanes merely removed the songs and all the topical allusions that had gone out of date, and then inserted a number of new ones, mainly in one short section (Plut. 170–80).Footnote ¹⁰ However, in an incisive critique of this theory, Alan Sommerstein has highlighted that it not only overlooks various formal and linguistic signals of ‘lateness’ throughout the text, but that it is also undermined by the scholiastic evidence itself, around which it revolves:

Sommerstein’s own solution to the conundrum is therefore the one already argued for by Benjamin Rogers and Karl Kunst:Footnote ¹² namely that, effectively, there were three texts that went under the name of Plutus. The first of these, of 408 BC, would have been lost relatively early, certainly before the comments that have reached us through the scholia on the extant comedy came into being (though not before it could be excerpted by the source of [1], whose Aristophanes fragment Rogers and Sommerstein believe to be taken from the real Plutus I); while the second, of 388 BC, would have been staged in a first version (= Plutus IIa) and then revised, perhaps for another venue outside Athens (= Plutus IIb).Footnote ¹³ Unlike Plutus I, both Plutus IIa and Plutus IIb would have survived long enough to be read by our commentator, who however also knew of the erstwhile existence of Plutus I. Having before him two (admittedly similar) scripts under the title Plutus, he would have (mis-)inferredFootnote ¹⁴ that one must be the play of 408 BC and the other that of 388 BC; and this would explain how he could in good faith comment on the late Plutus as if it were the early one: he simply tried to do his best when facing an awkward situation, even if that meant coming up with implausible transfer assumptions like those in [4], [7] and perhaps [5] (cf. section II).

IV. Careless Didymus?

One great advantage of the Rogers–Sommerstein theory is that it frees the commentator in question from the imputation that he was so incompetent as to mistake a play of 408 BC for one of 388 BC, and vice versa, when both were accessible to him. We are thus already moving towards the kind of sympathetic assessment of the activity of ancient scholars advocated in section I. As we shall see in due course, we may be able to go even further here; but before doing so, we first have to tackle the issue of who it was that made the mistake.

While Rogers and Sommerstein are cautious enough not to give any name, in other discussions of the matter the first-century BC scholar Didymus Chalcenterus has regularly been suspected.Footnote ¹⁵ It is commonly held, for good reasons, that Didymus’ comprehensive ὑπομνήματα on Aristophanes’ comedies by and large superseded earlier commentaries by Alexandrian scholars such as Euphronius, Callistratus, Aristarchus and Apollonius, until they were in their turn replaced, in the second century AD, by the works of Symmachus, the single most major direct source of the medieval scholia. Famous for his encyclopaedic knowledge and prolific output, Didymus was (and is) also known for a certain negligence, however, which earned him the nickname βιβλιολάθας, the ‘man who forgets [his own] writings’ (cf. Ath. 4.139c; Quint. Inst. 1.8.20). Someone with this profile could, one may believe, easily err also when working on Aristophanes, especially if the error was indeed prompted by the unavailability of a crucial piece of primary material (here: Plutus I, according to the Rogers–Sommerstein scenario).

However, Didymus’ alleged scholarly shortcomings alone cannot of course justify the indictment. The real case for it has been made by Pierre Boudreaux in his classic treatment of the history of Aristophanic scholarship in antiquity, which therefore must be the starting point for any new discussion of the matter. Having admitted that Didymus’ ‘nom n’est mentionné qu’une fois (par Symmaque semble-t-il) dans les scolies qui portent sur ce point’, and that in a scholion whose ‘texte est corrompu de tel sorte qu’on ne peut guère en tirer parti’ (namely schol. Ar. Plut. 550e Chantry; cf. [9] in section V), Boudreaux continues:

Let us look at each of these points, starting with the last:

1. (i) The citation of the ‘excessively detailed commentaries’ occurs in a scholion we have not yet considered because it does not really provide much help in untangling the problem of the two Plutus plays:

   

   BL EPSIDEMUS: I see someone who is going to sit down on the tribune, holding a suppliant branch, in the company of his children and wife, and looking not at all different from Pamphilus’ Children of Heracles.

   schol. Ar. Plut. 385b Chantry: τὸν Πάμφιλον μὲν Kαλλίστρατος καὶ Εὐφρόνιος τραγῳδιῶν ποιητήν φασι καὶ διδάξαι “Ἡρακλϵίδας”. τὰ δὲ λίαν ἐπιτϵτηδϵυμένα ὑπομνήματα διστάζϵι πότϵρον τραγικὸς ποιητὴς ἢ ζωγράφος, ὃν καθηγήσασθαί φασιν Ἀπϵλλοῦ. ἐν μέντοι ταῖς διδασκαλίαις πρὸ τούτων τῶν χρόνων Πάμφιλος οὐδϵὶς φέρϵται τραγικός. γραφὴ μέντοι ἐστίν· “οἱ Ἡρακλϵίδαι καὶ Ἀλκμήνη καὶ Ἡρακλέους θυγάτηρ Ἀθηναίους ἱκϵτϵύοντϵς, Εὐρυσθέα δϵδιότϵς”, ἥτις Παμφίλου οὐκ ἔστιν, ὥς φασιν, ἀλλ’ Ἀπολλοδώρου· ὁ δὲ Πάμφιλος, ὡς ἔοικϵ, καὶ νϵώτϵρος ἦν Ἀριστοφάνους. VEΘMatrBarbAld

   Callistratus and Euphronius say that Pamphilus is a writer of tragedies who produced <a play> Children of Heracles; whereas the excessively detailed commentaries are undecided on whether he is a tragic poet or a painter, who is said to have been the teacher of Apelles. However, in the didaskaliai no tragic poet Pamphilus is listed before this time. There is a painting The Children of Heracles, Alcmene, and Heracles’ daughter supplicating the Athenians, being afraid of Eurystheus, which however is not by Pamphilus, as they claim, but by Apollodorus; Pamphilus, it seems, was also later than Aristophanes.

   That the information in this note goes back to Didymus is likely not just because it refers to λίαν ἐπιτϵτηδϵυμένα ὑπομνήματα in a way which Boudreaux takes to be typical of Didymus.Footnote ¹⁷ There is also much agreement that, whenever the scholia cite early Aristophanic commentators like Euphronius and Callistratus, Didymus is the likely intermediate source. Moreover, the learning which transpires from the consultation of the didaskaliai and, above all, the correction of the (at least allegedly) erroneous ascription of the painting seem typical of him; and if that is so, the last sentence too may well be his. However, the only aspect that is of potential chronological relevance in all this is precisely the final remark, and it is difficult to see why the observation that Pamphilus was (according to some) ‘later (or: younger) than Aristophanes’ could not have belonged to a commentary on a play of 388 BC as much as to one on a much earlier comedy. Even if it could be ascertainedFootnote ¹⁸ that in reality the painter Pamphilus’ activity postdated only 408 BC, but not 388 BC, that would still tell us nothing about what Didymus (?) and/or his sources thought about the matter; nor should it be overlooked that the statement is about the chronological relationship between Pamphilus and Aristophanes, not between Pamphilus and a specific play of Aristophanes.

2. (ii) What Boudreaux himself calls a ‘fait difficilement utilisable’ is another scholion on Plut. 173 which, in a discussion of the presence of Athenian allied forces to secure Corinth against Sparta after the Peloponnesian War, says ἦν δὲ ὁ καταστήσας τοὺς ξένους ἐν Kορίνθῳ Kόνων ὁ Ἀθηναίων στρατηγός, καθϵλὼν Λακϵδαιμονίους ὅπως φυλάττοι τὴν ἔφοδον αὐτῶν (‘The person in charge of organizing the allied forces in Corinth was the Athenian general Conon, who had removed the Spartans, so as to prevent them from attacking’, schol. Ar. Plut. 173e Chantry). Since this agrees with the information given in an entry of Harpocration’s lexicon of the ten Attic orators (Harpocration, s.v. ξϵνικὸν ἐν Kορίνθῳ, where the Plutus is referenced alongside Demosthenes’ Φιλιππικά), which may well be based on Didymus’ commentary on Demosthenes, Didymus could also be the source of the scholiastic remark. Though hypothetical, all this is correct, but it still cannot really indicate that Didymus was misdating our play. To write at length about the Athenian ξϵνικόν in Corinth after the Peloponnesian War either (a) makes better sense if one believes that the passage one is commenting on is not talking about a permanent presence of a ξϵνικόν in Corinth (= the first option envisaged by [4]) or (b) is equally sensible whether one believes that one is commenting on a passage ‘transferred’ from a play of 388 BC (= the second option envisaged by [4]) or on the play of 388 BC itself.

3. (iii) The ‘agreement’ Boudreaux diagnoses between Ath. 13.592d, Harpocration, s.v. Nαΐς (cf. Sud. ν 16) and schol. Ar. Plut. 179a Chantry (= [5] above) is of the following kind. In [5] we learn that the courtesan Laïs cannot reasonably have been named in a play of 408 BC. Both Athenaeus and Harpocration observe that in Plut. 179 the name Nαΐς might be read instead of Λαΐς, thereby establishing a connection with a different hetaera, who was also mentioned in a speech of Lysias (fr. 299 Carey). Given the likelihood that this proposal goes back to Didymus’ Λέξις κωμική, Boudreaux argues that whereas the extant scholion presents the chronological ‘problem’, Athenaeus and Harpocration preserve Didymus’ ‘solution’ to it. However, [5] also underlines that a passage from a play of Plato Comicus of 392/1 BC, which apparently said that Laïs ‘no longer is’ (fr. 196), need not be taken at face value (presumably because it may simply mean that Laïs is past her prime). This needs to be stressed if one is dealing with a mention of Laïs in a play thought to postdate 392/1 BC. In other words, what the scholion, as we now have it, may be getting at is a structure of the type ‘X is not possible in 408 BC, but it is possible in 388 BC’, exactly as in [4] and [7]. In those two instances, the (or one of the) solution(s) envisaged is a textual transfer from the later version, and such a hypothesis has therefore been suggested to lurk behind [5] as well (cf. section III). That cannot be proved, of course, but neither can we affirm that a textual emendation (Λαΐς → Nαΐς) was advocated instead. In fact, for all the unpredictability of the epitomization processes that led to the scholiastic corpus in its current form, the absence of any hint at such an emendation in the extensive scholia on our passage (as well as in schol. Ar. Plut. 304a/304b Chantry), which persistently talk of the Corinthian Laïs, rather speaks against it. In other words, even if we grant that Didymus in the Λέξις κωμική wondered about changing the text of Plut. 179, for whatever reason,Footnote ¹⁹ and that he therefore might have done the same also in his ὑπόμνημα on Plutus, he cannot have done it in a way that settled the matter once and for all. Consequently, as long as the Λαΐς reading remained on the table, any post-Didymean commentator could also be the person who got worried about (perceived) chronological issues relating to it.

4. (iv) The background of Boudreaux’s reference to two uses of μήποτϵ (‘perhaps’) in the relevant dossier is the observation that, to judge by the scholia one may confidently attribute to Didymus, he frequently used this adverb in his explanations. As Boudreaux himself stresses elsewhere, this can never be a watertight criterion of attribution.Footnote ²⁰ For instance, in schol. Ar. Av. 303a we read μήποτϵ οὐχ ἕν ἐστιν, ἀλλὰ δύο, φησὶν ὁ Σύμμαχος (‘perhaps it is not one <bird> but two, says Symmachus’). On the face of it, μήποτϵ should here be attributed to Symmachus; and even if this scholion merely referenced Symmachus, although the latter had in reality copied Didymus (so that μήποτϵ would remain Didymus’ word), the case would still demonstrate that any Didymean μήποτϵ note was potentially subject to later editing. Hence, the presence of μήποτϵ could at best be used as an indication of Didymean authorship only of that part or sentence of a scholion, or set of scholia, in which the adverb occurs. More importantly, meanwhile, neither of the two μήποτϵ Boudreaux is talking about is to be found in any of the scholia with a distinct bearing on our discussion (nos [2]–[7]). Instead, they belong to the wider context of the one potentially relevant scholion whose text, in Boudreaux’s own judgement, is ‘so corrupt that one can hardly make use of it’ (cf. above, on schol. Ar. Plut. 550e, which follows schol. Ar. Plut. 550c/550d, in which the μήποτϵ occurrences are found). To now treat the same source as if it did provide positive evidence for Didymus regarding our Plutus as a play of 408 BC is paradoxical.

V. Blameless Didymus

All in all, it seems fair to say that the case for Didymus having been the scholar responsible for the dating error in the Plutus scholia is weak. Obviously, the fact that a case is weak need not mean that it is wrong. However, we should at least also ask whether there are any counter-arguments. In what follows, an attempt will be made to adduce some. Admittedly, none of them is unassailable on its own either, but when taken together they still seem sufficient to redeem Didymus’ name.

1. (i) We start with a closer look at the piece of evidence just mentioned, on whose elusiveness one must agree with Boudreaux:

   

   CHREMYLUS: So don’t we say that Poverty is a sister of Beggary?

   POVERTY: Yes, you say that, just as you say that Dionysius is similar to Thrasybulus!

   schol. Ar. Plut. 550a Chantry: ὑμϵῖς γ’ RΘ οἵπϵρ καὶ Θρασυβούλῳ REΘAld <Διονύσιον> [φατ᾿] ϵἶναι ὅμοιον M: τὰ μὴ ὅμοιά φησιν αὐτοὺς ὁμοιοῦν· ὡς ϵἴ τις λέγοι Διονύσιον τὸν ἐξώλη τύραννον ἐοικέναι Θρασυβούλῳ τῷ Λύκου, RVMEΘAld ἀνδρὶ φιλοπόλιδι VMEΘAld καὶ παντὸς κρϵίττονι λόγου, διά τϵ τὰς ἐπιφανϵῖς αὐτοῦ κατὰ τῶν πολϵμίων νίκας καὶ διότι κατέλυσϵ τὴν τῶν Λʼ τυραννίδα· VMEAld ϵὔπορος γὰρ ὤν, καὶ ἐξ αὐτῶν ἐκπϵσών, μϵτὰ ταῦτα Φυλὴν καταλαβὼν καθϵῖλϵν αὐτούς. EAld

   You say that, just as you say that Dionysius is similar to Thrasybulus: She is saying that they treat as similar things that are not similar; as if someone said that Dionysius the abominable tyrant resembles Thrasybulus son of Lycus, a patriotic man beyond description (?), both because of his illustrious victories over the enemy and because he brought the reign of the Thirty to an end: for being wealthy, and exiled by them, he subsequently took Phyle and deposed them.

   schol. Ar. Plut. 550c/550d/550e Chantry: [550c] (ἄλλως) VEAld μήποτϵ EAld ὁ μὲν ἀξιωματικὸς καὶ RVEAld αὐθάδης VEAld [μέγας R], ὡς Στράττις ἐν τῷ “Kινησίᾳ”. EAld Διονύσιος δὲ μαινόμϵνος, RVEAld καὶ ὡς ἔοικϵ [ἰχ]θυοπώλης. VEAld Πολύζηλος· “Xὠ μαινόμϵνος ἐκϵινοσὶ Διονύσιος | χρυσoῦν ἔχων †χραιδόνα† καὶ τρυφήματα | ἐν τῷ μύρῳ παρ᾿ Ἀθηναίων βαυκίζϵται”. V [550d] μήποτϵ δὲ καὶ πϵρὶ τὴν ὄψιν ἐπισκώπτϵι, ὡς ὁμοιούντων σφόδρα ἀνομοίους. VEAld

   [550e] μᾶλλον δὲ ἄν τις ὑπονοήσϵιϵν ἕτϵρον Διονύσιον, Θρασυβούλου τοῦ Kολ<λ>υτέως ‘ἀδϵλφόν’· ἔγγιστα συγγϵνϵίας ἐστί. †καὶ ἐν τῇ Ἰλιάδι† σαφέστϵρον οὐδὲν [ms. οἶδϵν] ἢ κατὰ Δίδυμον. EAld

   (Differently:) Perhaps the one [= Thrasybulus] is dignified and arrogant, as Strattis <says> in the Kinesias, and Dionysius crazy and, seemingly, a fish-seller [or: incense-seller?]; cf. Polyzelus: ‘And that crazy Dionysius, who with his golden bracelet (?) and his luxuries gives himself airs on the perfume market vis-à-vis (?) the Athenians’. But perhaps he is also jokingly referring to their looks, <implying> that they assimilate people who do not look similar at all.

   However, one might rather suspect <that> another Dionysius <is meant>, a ‘brother’ of Thrasybulus of Collytus; he is a very close relation. †also in the Iliad† nothing/not at all clearer [MS: knows more clearly] than according to Didymus.

   It hardly needs stressing how confused much of this is, especially the last part; any translation can only be tentative. But one or two observations are in order. Firstly, in this group of scholia consideration appears to be given to two competing approaches to the Aristophanic line: either (a) that the Dionysius and the Thrasybulus who are mentioned are the famous Sicilian tyrant and the equally famous Athenian general and democratic politician, respectively [550a/550c/550d]; or (b) that some lesser namesakes are meant [550e]. Moreover, option (a) is divided into two parts, one of which (a¹) concentrates on the political divergences between Dionysius and Thrasybulus [550a] whereas the other (a²) suspects some behavioural and/or visible differences in addition [550c/550d]. As long as one accepts the μήποτϵ argument (cf. section IV), (a²) should go back to Didymus. For our discussion this is of relevance because adhering to (a) is chronologically incompatible with believing in a composition date of 408 BC for the line in question: Dionysius only came to power in Syracuse two or three years later, a fact which might explain why some other voice, firmly believing in the 408 BC date, then thought the intrinsically unattractive (b) option had to be promoted instead.

   Furthermore, even if we cautiously disregard the μήποτϵ point, we should note that the last scholion [550e] ends by contrasting something, or some point of view, with what Didymus maintained (σαφέστϵρον ἢ κατὰ Δίδυμον), no matter whether the doubtful text contained a negative (cf. Holwerda’s conjecture σαφέστϵρον οὐδέν) or not (cf. the awkward σαφέστϵρον οἶδϵν of the manuscript). Unfortunately, the preceding phrase καὶ ἐν τῇ Ἰλιάδι is of little help here, since equally corrupt.Footnote ²¹ However, the scholion does not actually terminate with ἢ κατὰ Δίδυμον, as Chantry’s presentation suggests. In reality the text continues in manuscript E (the only witness, apart from its reproduction in the Aldine edition) with … ὡς ϵἴ τις λέγοι Διονύσιον τὸν ἐξώλη τύραννον ἐοικέναι Θρασυβούλῳ τῷ Λύκου, ἀνδρὶ φιλοπόλιδι καὶ παντὸς κρϵίττονι λόγου, διά τϵ τὰς ἐπιφανϵῖς αὐτοῦ κατὰ τῶν πολϵμίων νίκας καὶ διότι κατέλυσϵ τὴν τῶν Λʼ τυραννίδα, in other words a repeat of what we also read in schol. Ar. Plut. 550a. So, whatever one may think about the beginning of schol. Ar. Plut. 550e, the most natural translation of the final … ἢ κατὰ Δίδυμον ὡς ϵἴ τις λέγοι Διονύσιον τὸν ἐξώλη τύραννον κτλ. would be ‘… than, according to Didymus, “as if someone said that Dionysius the abominable tyrant etc.”’; and that would imply that the best part of schol. Ar. Plut. 550a is Didymean information as well, again despite its prima facie incompatibility with a 408 BC chronology. In the end, the evidence of [9] may thus indeed be inconclusive, but if one had to extract something from it, it could only be that Didymus placed our Plutus late in time.Footnote ²²

2. (ii) As already mentioned, there has been general agreement, ever since Moritz Schmidt discussed the matter in detail,Footnote ²³ that whenever the Aristophanic scholia record opinions of Hellenistic grammarians, including the early commentators Euphronius and Callistratus, this knowledge has been channelled through Didymus. Boudreaux concurs with this: ‘Le domaine de Didyme s’élargit encore si l’on admet que la plus grande partie des citations des grammairiens antérieurs lui sont dues. La démonstration en a été apportée pour Lycophron, Euphronius, Eratosthène, Ammonius. Il serait aisé de la poursuivre pour Aristophane de Byzance, Callistrate, Aristarque, etc.’.Footnote ²⁴ In other words, scholia that engage with one of those Hellenistic voices are to be taken as Didymean at their core. One such scholion is [1], the note on the Kϵραμϵικαὶ πληγαί which cites (and depends on) Euphronius. But [1] is also our only source that explicitly attributes a quoted passage to ‘the first Plutus’ (Ar. fr. 459), and it does so within the line of thought that leads to the Euphronius reference. This annotation therefore contains almost incontrovertible e contrario evidence for Didymus regarding ‘our’ Plutus as the second Plutus. Footnote ²⁵

3. (iii) In a similar vein, mention should be made of one further scholion, on Plut. 1194, whose Didymean pedigree is again virtually certain because it cites Eratosthenes:

   

   CHREMYLUS: But someone bring out here lighted torches, so you can take them and lead the procession for the god.

   schol. Ar. Plut. 1194 Chantry: ἀλλ’ ἐκδότω τις EΘNAld δϵῦρο Ald δᾷδας N: ὅτι ὁ Λυκόφρων, ὡς Ἐρατοσθένης φησίν, ᾠήθη πρῶτον δᾷδας τοῦτον ᾐτηκέναι. πϵποίηκϵ δὲ καὶ ἐν “Ἐκκλησιαζούσαις” αὐτό· ἀλλὰ γὰρ Στράττις, πρὸ ἀμφοτέρων τούτων τοὺς “Ποταμίους” διδάσκων, ϵἰς Φιλύλλιον ἀναφέρϵι τὸ πρᾶγμα· “ὑμϵῖς τϵ πάντϵς ἔξιτ’ ἐπὶ τὸ Πύθιον, | ὅσοι πάρϵστϵ, μὴ λαβόντϵς λαμπάδας, | μηδ’ ἄλλο μηδὲν ἐχόμϵνον Φιλυλλίου”. VEΘNBarbAld

   But someone bring out here torches: <To note> that Lycophron, as Eratosthenes says, thought that this one [= Chremylus] had been the first to ask for torches; yet he [= Aristophanes] also did it in Ecclesiazusae. However, Strattis, who was staging his Potamioi before both of these <plays>, refers the practice to Philyllius: ‘All of you who are here, go out to the temple of Pythian Apollo, without torches nor anything else that has to do with Philyllius’.

   Since Eratosthenes’ point in citing Lycophron must have been to correct or improve on him,Footnote ²⁶ the reference to Ecclesiazusae of 393/2 BC (that is, to Eccl. 1149–50) will have been his already. Accordingly, although the subject of πϵποίηκϵ in the second sentence must be Aristophanes, τοῦτον in the first sentence should designate Chremylus, not Aristophanes: after all, the fact that a similar stage motif also occurs in Ecclesiazusae would not otherwise contradict Lycophron’s claim. However, upon digging out this piece of information, presumably from Eratosthenes’ Πϵρὶ τῆς ἀρχαίας κωμῳδίας, Didymus in his turn corrected Eratosthenes by pointing to a comedy by Strattis, which in his view serves as an even earlier terminus ante quem. Now, if Didymus had believed that the Plutus was a play of 408 BC, the parallel in Ecclesiazusae would have been irrelevant to him for dating purposes. At best, one might argue that he mentioned it simply because it was highlighted in Eratosthenes (who himself, incidentally, must by implication have placed the extant Plutus after Ecclesiazusae). But in that case, it would still have been more natural to continue by saying, ‘However, Strattis, who was staging his Potamioi even before Plutus (which in fact is earlier than Ecclesiazusae) …’, rather than to group Plutus and Ecclesiazusae together, despite believing them to be separated by 15 years. And what is more, all we know of Strattis’ dramatic career strongly speaks against his having produced plays before the last decade of the fifth century;Footnote ²⁷ but if he did not, Didymus, who had access to and consulted the didaskaliai (cf. section IV and below), would not have maintained the opposite by emphatically placing the Potamioi before a play presumed to be of 408 BC.

4. (iv) As a complement to our consideration of [1] with its reference to ‘the first Plutus’, neither should we forget the two sources mentioned in section II which correctly attribute a word from the extant Plutus to the Πλοῦτος δϵύτϵρος: the Herodian-based schol. (A) Hom. Ψ 361a¹ on the optative μϵμνῇτο and a passage of Athenaeus (9.368d) discussing the word κωλῆ (‘ham’). Given the lasting influence of Didymus’ work on subsequent ancient Aristophanic scholarship, it is difficult to imagine that Herodian and Athenaeus, both working two centuries after Didymus, would have assigned such material (correctly) to the second Plutus if the leading authority in the field had maintained that the play they were excerpting, whether directly or indirectly (for instance via Didymus’ Λέξις κωμική), was the first Plutus. Footnote ²⁸

5. (v) While less specific, the last point is in some ways the most fundamental. We must ask what could have prompted Didymus to misdate the extant play to 408 BC. There are two options to consider. Either (a) he might have been misled by some text-internal evidence, or else (b) some external trigger might be involved.

   As for (a), it should first be stressed that the text does not contain anything that looks like a plausible reason to overturn the overwhelming evidence that speaks for a late date. Even so, some scholars thought that (a) was the right assumption. According to them, Plut. 972 and the scholion on that line (= [6]) are key. Given the learned quotation from Philochorus, it is indeed plausible that something is Didymean here. Hence, Jan van Leeuwen and Boudreaux argued that Didymus came across the passage in Philochorus in which reference is made to a novel institutional arrangement of 410/9 BC, whereby not only the composition of the law courts but also that of the Athenian Council was organized by sortition amongst individuals previously assigned to letter classes, brought this together with Plut. 972 and concluded that the line must be alluding to this innovation and therefore written at a date as close to it as is possible, hence in 408 BC.Footnote ²⁹ To impute such reasoning to anyone is bold, especially if that person so clearly recognizes what chronological difficulties his alleged conjecture entails and has to disregard that Plut. 277 already referred to the organization of the law courts by letter classes (cf. scholl. Ar. Plut. 277b–f, the core of which is itself of Didymean origin).Footnote ³⁰ But what is more, the presentation of [6] effectively disproves the van Leeuwen–Boudreaux theory. The Philochorus passage with its information about the Council is here introduced as an additional remark (οὐ μὴν ἀλλὰ καί) designed, not untypically for Didymus, to display great learning, yet not as something that would invalidate the more crucial acknowledgement with which the scholion starts, namely that ἔπινϵς takes the place, παρ’ ὑπόνοιαν (‘by para prosdokian humour’), of ἐδίκαζϵς (not: ἐβούλϵυϵς!). In other words, there is no doubt about [6] correctly decoding Plut. 972 in parallel with Plut. 277 (to which line it even adds something of a cross-reference); and if that is so, its source cannot have based any dating, however speculative or ill-judged, on what Philochorus had to offer.

   That said, we are left with option (b), that there might have been some text-external inducement for Didymus to date the play to 408 BC. Since Didymus was not in general overly impressed by what his predecessors thought, it is implausible that he would have simply made some earlier scholar’s mistake his own; and, of course, such an assumption would in any case only shift the issue slightly back in time. Alternatively, though, one might think of the didaskaliai,Footnote ³¹ whose book edition by Aristotle Didymus could no doubt still consult just as easily as Aristophanes of Byzantium did when he wrote the first dramatic hypotheses. If the didaskaliai recorded an Aristophanic Πλοῦτος in the year 408 BC, and if both ‘our’ Plutus and another version Didymus had access to both fit that date rather poorly, but ‘ours’ was (comparatively speaking) a less problematic match, that could have been an incentive to misplace it (while leaving the other version to 388 BC; cf. section III). However, not only are we operating with two unknowns here, but there are also, again, considerations of plausibility that speak against such a solution. A theory like this could only work reasonably well if the play of 408 BC had been lost not just before Didymus began to deal with Aristophanes, but much earlier, before Aristophanes of Byzantium could edit itFootnote ³² and Euphronius, Callistratus, Apollonius and all the other Hellenistic commentators of Aristophanes were able to annotate it; for otherwise we would have to postulate that, in addition to the play itself disappearing, any ὑπόμνημα on it must also have been lost before Didymus gathered together his source material. After all, if one has a commentary on Play X, one does not need to see Play X itself in order to decide whether or not ‘Play Y = Play X’ is true. Yet, apart from the fact that Ar. fr. 459 (in [1] above) could not then belong to the real first Plutus, because it was accessible to someone, we should also expect that such an early loss would have been known and discussed (cf. argum. II Ar. Pax on the other Peace, as well as the οὐ σῴζϵται/σῴζονται notes which were ‘probably taken over from [Callimachus’] Pinakes into Aristophanes’ hypotheses’).Footnote ³³ Furthermore, it would be an odd coincidence anyway if precisely this play had suffered a worse fate than the rest of Aristophanes’ output, including the two Thesmophoriazusae or the two versions of Clouds.

VI. What about Symmachus?

Although the above evidence shows that the suspicion that it was Didymus who misdated the extant Plutus is ill-founded, the fact remains that somebody did do it. As a complement to the preceding discussion, we may therefore ask who else it could have been. At first sight, it may look as if the last of the points made above similarly invalidated any alternative identification. If we mean to exonerate Didymus by arguing, among other things, that without pertinent text-internal or text-external evidence for a date of 408 BC no serious scholar could have erred like this, would the same argument not also apply to any other suspect? On closer inspection, the answer to this is a mere, ‘Yes, but …’.

Firstly, it is worth noting that the earlier four points (i–iv) are not applicable to any post-Didymean Mr X.Footnote ³⁴ In that sense, Didymus’ case would still be stronger on balance. Secondly, and more crucially, the parameters that are of relevance to point (v) may in fact have changed over time in such a way that the argument eventually loses its force. In order to demonstrate this without too much vagueness, from this point on I will tentatively replace ‘Mr X’ with ‘Symmachus’, not because proof of such an identification is possible, but merely because Symmachus appears to be the most likely alternative culprit if we want a specific name. Whoever made the mistake must have been an influential Aristophanic scholar as his opinion surfaces so prominently in the scholia; and after Didymus, Symmachus is the only known figure of whom this is true.Footnote ³⁵ Moreover, a scholion on Plut. 1011, according to which Symmachus (unhelpfully) contradicted Didymus’ explanation of the hypocoristic diminutives νηττάριον (‘little duck’) and φάττιον (‘little ringdove’) in that line, does suggest that Symmachus did not regard the Plutus as belonging to the very end of Aristophanes’ career; for he sought to bolster his own opinion that these are actually two personal names, Nιτάριος and Bάτος, with a reference to ‘subsequent plays’ in which a Nιτάριος would be lampooned:

Whereas Didymus ostensibly had at his disposal a singularly well-stocked library, in all likelihood that of Alexandria, the same need not be true of Symmachus (let alone of any other, more minor, Aristophanic scholar of the Imperial age). In particular, it is by no means clear that Symmachus, in the second century AD, could still access all the Aristophanic plays that had been collected in Alexandria: the selection and canonization process that eventually led to a restriction of the Aristophanic corpus to a mere 11 plays will have been under way by then. So, in comparison with Didymus’, Symmachus’ ability to cross-check any piece of information he came across would have been seriously curtailed. We must assume that what he was working from when writing his own commentaries on Aristophanes was (i) the texts of the comedies that were still commonly read in his day, which may have been significantly fewer than the 40+ plays held on record; (ii) Didymus’ comprehensive commentaries on the same plays (but presumably few if any of those written by Didymus’ Alexandrian predecessors); (iii) various minor post-Didymean ὑπομνήματα as well as a haphazard selection of primary and secondary literature (including reference works like Didymus’ Λέξϵις); and (iv) his own literary intuition and judgement. Against such a background, the question of whether there could have been a text-external trigger for Symmachus to misdate the Plutus to 408 BC looks rather different from Didymus’ situation.

If we assume, with Rogers and Sommerstein (cf. section III), that Symmachus came across two similar but not identical versions of the play of 388 BC, but that there was no didascalic information to explain this duplication, it would have been natural for him to infer that only one of them could really belong to 388 BC (as attested by the extant hypothesis) and that the other must therefore belong to some other year. The comic text itself would not have helped with this issue; and as long as we hold that the Didymean ὑπόμνημα did not misdate the Plutus, it too should not in principle have tripped up Symmachus. However, we have already seen that Didymus’ commentary did contain one entry which could, if misinterpreted, suggest an alternative production date: namely the entry on Plut. 972 (= [6]) with its information about the Council’s organization by letter classes. If Symmachus had disregarded the redundancy of the Didymean reference to Philochorus in the same way it has been disregarded by some modern scholars (cf. section V), he could have thought that Didymus’ citation must be of greater relevance than it actually was and extrapolated that the play he was looking at should therefore be dated to as soon after 410/9 BC as possible. The phrase τῷ πρὸ τούτου ἔτϵι ἀρξάμϵνοι would then be his own addition to the basic information taken over from Didymus: there is, after all, no reason to think that he must have copied Didymus’ note verbatim.Footnote ³⁶

The autoschediasm such a theory presupposes does, however, represent a major drawback. No matter how strong the wish to find a production date might have been, would anyone have built so much on so little? One way of getting around the assumption would be instead to suspect a misunderstanding of more limited scope.Footnote ³⁷ For example, if Symmachus had read in Didymus that the Council’s organization by letter classes had begun ‘the year before’, he might have erroneously taken this to mean ‘the year before the present play was staged’, when in reality what was meant was ‘the year before the dicastic courts were (also) organized by letter classes’. Since ancient commentators constantly rephrased what they found in predecessor works, he might then, in all innocence, have replaced an earlier formulation such as ἔτϵι πρότϵρον ἀρξάμϵνοι by his own τῷ πρὸ τούτου ἔτϵι ἀρξάμϵνοι.

On balance, though, a third, even simpler, scenario seems more attractive still, as it entails no misunderstanding of a Didymean note at all. The only reason for placing so much weight on [6] alone is that either of the preceding theories would allow us to deny altogether the erstwhile existence of a different Plutus I of 408 BC. Some scholars have indeed done so,Footnote ³⁸ but most do not and it may mean going one step too far. If Aristophanes did stage one Plutus in 408Footnote ³⁹ and another 20 years later, this could easily be gleaned, even in Symmachus’ day, from Aristotle’s compilation of the didaskaliai.Footnote ⁴⁰ The default assumption in that case should be that Didymus’ rich library contained both comedies (cf. point (ii) in section V, on [1] with Ar. fr. 459) and that he also wrote a commentary on both, just as he did for the other Aristophanic plays. However, as noted above, between the days of Didymus and those of Symmachus some plays (together with the corresponding commentaries) may have stopped circulating widely.Footnote ⁴¹ At the same time, it is perfectly plausible that the text of Plutus II (the play of 388 BC) that found its way into Symmachus’ hands was somewhat different from the one Didymus had read and worked on. Exactly how different we cannot tell: it could have been a matter of only the odd line here or there diverging,Footnote ⁴² but it is at least as likely that Didymus still dealt with a full version of the comedy, including all the choral songs with topical references. In either case, when consulting Didymus’ commentary next to his own script of Plutus II, Symmachus would have noticed these divergences even without having two copies of the play in front of him. The lemmata in Didymus’ commentary alone would have been sufficient to demonstrate that the identity was not complete. Now, knowing from the didaskaliai that there was a Plutus I of 408 BC and knowing from Didymus’ commentary that the play Didymus had looked at belonged to 388 BC, but not knowing that the real Plutus I of 408 had ‘disappeared’ in the meantime, Symmachus could reasonably infer from the fact that there were divergences between ‘his’ Plutus and the Plutus of Didymus’ commentary that ‘his’ Plutus must be the earlier play. Yet, this apparent play of 408 BC did of course show such substantial overlaps with the text Didymus’ commentary reflected that (a) certain ‘transfer’ hypotheses became unavoidable (cf. [4], [5], [7]), while (b) in other cases it was sufficient to diagnose that a line had been altered in the later version, as inferable from Didymus’ commentary lemmata (cf. [2], [3]). Finally, as for [6], while the substance of this note with its quotation of Philochorus would still be owed to Didymus, the phrase τῷ πρὸ τούτου ἔτϵι ἀρξάμϵνοι would again be Symmachus’; yet, far from being an autoschediastic addition in the strict sense, it would be a sensible attempt to highlight the significance of the institutional innovation mentioned by Philochorus for a play thought to be produced so soon afterwards.

All in all, then, it is possible to conclude that Symmachus, being less well equipped than Didymus, is both a priori a more likely source of the fundamental error that found its way into the scholia and a scholar whose mistake, however it may have arisen, is more easily forgiven than it would be in Didymus’ case. Of course we might still find that Symmachus (or whoever it was) should not be excused either because, for all the misgivings he had about the 408 BC date (cf. [4], [5] and [7]), he preferred to build spurious ‘transfer’ theories rather than thoroughly rethink the conclusion he had reached. However, we must not forget that, if something like our last scenario were correct, doing so would have meant calling into question the few certainties he had: that there was an Aristophanic Plutus of 408 BC and that the Plutus he was reading could not, as such, be the 388 BC play that Didymus had authoritatively annotated. In modern scholarship, too, there are few who would take such a step lightly.