III. Authenticity and Date

1. External Evidence

When J. J. Scaliger laconically pronounced Rhesus spurious in 1600,60 he could support his claim with one ancient piece of evidence. According to Hyp. (b) Rh. (64.25–6 = 430.23–4 Diggle), ‘some’ suspected that the play was not by Euripides because it rather showed a ‘Sophoclean stamp’ (τοῦτο τὸ δρᾶμα ἔνιοι νόθον ὑπενόησαν, Εὐριπίδου δὲ μὴ εἶναι· τὸν γὰρ Σοϕόκλειον μᾶλλον χαρακτῆρα). Unfortunately, there remains no clue to the date or identity of these doubters, and despite some Sophoclean echoes in language and dramatic structure (pp. 31–9, 41), it is not clear what Σοϕόκλειον … χαρακτῆρα was supposed to mean.⁶¹

Otherwise no indication survives that Rhesus was considered inauthentic in antiquity or Byzantine times. On the contrary, Hyp. (b) Rh. (64.26–7 = 430.24–5 Diggle) = Arist. fr. 626 Rose explicitly identifies the play with a genuine one of that name recorded in the didascaliae (ἐν μέντοι ταῖς διδασκαλίαις ὡς γνήσιον ἀναγέγραπται). This entry, which presumably caused the false attribution in Alexandria (below), must have been for a year between Euripides’ first performance in 455 (Peliades) and 438 BC (Alcestis),⁶² for Crates of Mallus is likely to have drawn on it⁶³ when excusing a supposed astronomical error in Rh. 528–31 as due to the poet’s youth (527–3 ln.). The correct interpretation of the passage goes back to the Alexandrian Parmeniscus, who also knew the play as Euripidean. His learning, seen also in the correct explanation of προταινί as Boeotian (523–5a n.), may have reached the ancestor of the V-scholia through Didymus.

The remaining old scholia in V share the view of the competing scholars,64 and in their often petty criticism of individual passages match those for the other ‘Selection’ plays.⁶⁵ There is certainly no evidence for a Hellenistic edition intended to prove Rhesus spurious, as Wilamowitz wished to reconstruct.⁶⁶

If we grant that neither the great Euripides nor a namesake wrote the Rhesus we have, we must assume the existence of two homonymous plays, one of which disappeared very early on and was supplanted in the tradition by the other one. Most probably this happened in Alexandria, where, despite all efforts and the relatively good documentation for Athenian drama, the librarians did not always achieve satisfactory results. We hear of several texts that did not survive (οὐκ ἐσῴζοντο).⁶⁷ For the same reason presumably Ar. Ran. 1269/70 = A. fr. 238 could not be firmly located in Aeschylean lyrics,⁶⁸ whereas Cratinus’ Satyroi and Eupolis’ Noumeniai are mentioned only once each.⁶⁹ Conversely, the Medicean catalogue of Aeschylus records Αἰτναῖαι and Αἰτναῖαι νόθοι (A. T 78 TrGF), that is, the Alexandrians had copies of both plays and found ways to distinguish the ‘genuine’ from the ‘spurious’. Questions unsolved (and unsolvable) since antiquity are the authorship of several comedies⁷⁰ and in particular whether the famous ‘atheistic’ Sisyphus speech, which is partially preserved in Sextus Empiricus (Adv. Math. 9.54) and elsewhere, belongs to Euripides’ satyric Sisyphus of 415 or an otherwise unattested play of that title by Critias (88 B 25 DK = TrGF 43 F 19; cf. TrGF V.2, 658–9).⁷¹ Likewise, the tragedies Tennes, Rhadamanthys and Pirithous are doubted in the Life of Euripides (E. T 1 IA § 9 ~ IB § 5 TrGF), and for the last one there is independent evidence for confusion with Critias (Ath. 11.496b = Critias 88 B 17 DK = TrGF 43 F 2).⁷² The ascription of Prometheus Bound to Aeschylus, on the other hand, may have been uncontested.⁷³

There was thus ample scope for error in the attribution of plays –more than in the fourth century, when little systematic research on drama was done,⁷⁴ memories of performances were still fresher and only the odd book-trader, actor or producer may have tried to pass off an obscure or anonymous text as the (re-discovered?) work of a fifth-century classic. In the case of Rhesus the Alexandrians will have acquired at least one such nameless, or wrongly inscribed, copy and identified it with the lost Euripidean tragedy in the didascaliae. They may have been particularly eager to fill a gap among the poet’s early plays, which by their very antiquity were most likely to have perished.⁷⁵

As has just been indicated, it is possible (though much less probable) that a spurious Rhesus had already before 300 BC become established under Euripides’ name. Proof of this, and indeed of authenticity, has sometimes been derived from the quotation in Hyp. (b) Rh. (64.29–32 = 430.26–431.29 Diggle) of a one-verse iambic prologue fragment after a hypothesis ascribed to Dicaearchus (fr. 81 Wehrli = 114 Mirhady):

Yet whereas Nauck’s Δικαίαρχος for δικαίαν (VL: om. Q) is certain,⁷⁷ we can hardly be sure that Dicaearchus really wrote those alphabetically arranged plot summaries which in their original form invariably added the first line for identification and are known since Zuntz as Tales from Euripides.⁷⁸ It is far more plausible that Sextus Empiricus’ attestation of Dicaearchan (Adv. Math. 3.3 = Dic. fr. 78 Wehrli = 112 Mirhady) represents a case of later mythography falsely credited to a famous scholar.⁷⁹ Diggle, most recently, endorsed a date between the second century BC and the first century AD, on the basis of the author’s extensive use of ‘Asianic’ rhythmical clausulae.⁸⁰

Nevertheless, the attestation of this and a second additional prologue in Hyp. (b) Rh. (64.33–65.47 = 431.30–44 Diggle), which the compiler (or his source) was able to define from autopsy as an actor’s interpolation ‘not worthy of Euripides’ , suggests that our Rhesus was revived at least twice in the fourth century BC. An audience ‘brought up’ on Euripidean expositions may have been irritated at the archaising opening parodos, which so perfectly sets a mood of disorder and suspense, but does not well prepare for the reversals of the plot (1–51n.), and the obvious solution was to supply an iambic speech or dialogue providing all the information that was missed. These revivals must have been a success – or the interpolations would not have got into circulation to survive in ‘wild’ copies until at least the second century BC. That Aristophanes of Byzantium did not know them can be deduced not only from our standard text, but also from the statement in his hypothesis that the sentry chorus ‘delivered the prologue’: Hyp. (c) Rh. 65.56–7 = 432.52–3 Diggle ὁ χορὸς ἐκ Τρωϊκῶν, οἳ προλογίζουσι. The Alexandrians did not as a rule omit passages they deemed spurious.⁸¹

The very possibility of multiple re-performances would lead us to place Rhesus relatively early in the fourth century, even if we are not bound by the notion that Dicaearchus knew it as Euripidean around 300 BC. A rough terminus ante quem (and evidence for an almost immediate non-theatrical interest in the play) may indeed be given by sources as different as vase-paintings and a scholarly text.

From the middle of the fourth century (ca. 360–340 BC) we have three splendid Apulian vessels with illustrations of Rhesus’ death.⁸² Their main pictures, showing Diomedes in flight or attack, while Odysseus is leading away the famous horses, follow Iliad 10. Yet one crater (Berlin, Antikensammlung, Inv. 1984.39) adds in the top right margin (next to Athena) a female figure in pathetic pose and below her a horned man, whom conch and reed identify as a river god. Given the influence that Attic tragedy seems to have exerted on the south-Italian vase painters, the couple is most easily interpreted as Rhesus’ divine parents, Strymon and the Muse, who were probably first introduced in our play (p. 13).

Secondly, a comment by Asclepiades of Tragilus (cf. n. 15), who may be the same as the tragic victor at the Lenaea of 352 BC (DID A 3b 54 TrGF), has with some plausibility been referred to Rhesus. The source is an almost identical note in Hesychius, Photius and the Suda on a fragmentary verse from Epicharmus (Hsch. ρ 272 Hansen ~ Phot. ρ 103 Theodoridis ~ Suda ρ 143 Adler):

As it is unlikely that Asclepiades in his wished to explain an obscure gloss from Sicilian comedy,83 several authorities have thought of the Muse’s prediction at Rh. 970–3.⁸⁴ Rhesus’ future status as a heroic Bacchus prophet on Mt. Pangaeus (970–1, 972–3nn.) could indeed be alluded to in ἄριστον αὐτὸν γεγονέναι ἀλήθειαν εἰπεῖν, especially since no other dramatisation of the myth can safely be , especially since no other dramatisation of the myth can safely be posited for Asclepiades’ days. But absolute certainty is impossible.

It should be stated that this interpretation of the external evidence, with its consequences for the dating of Rhesus and the way it entered the Euripidean tradition, represents the communis opinio of those who regard the play as spurious. Liapis’ theory that Rhesus was written for a Macedonian audience between 350 and 330 BC, which is untenable also on other grounds (pp. 18–21), requires him to dismiss Asclepiades and the Apulian vase and, more importantly, to assume that the two prologue fragments in Hyp. (b) Rh. belong to the genuine early work of Euripides and were transmitted through Dicaearchus.⁸⁵ Yet apart from the doubtful association of Dicaearchus with any kind of tragic hypotheses, it is hardly conceivable that this drama survived until the late fourth century BC (and with an alternative prologue in some copies), only to vanish soon after and become one of the very few Euripidean tragedies that did not reach Alexandria. Moreover, the second prologue seems specifically composed for our Rhesus, and not enough of the first is preserved to prove that it could not have been. It would require an extraordinary similarity of plot structure for the expositions to fit both plays and for the author of Hyp. (b) Rh. (or a predecessor) to attach them to the wrong one.⁸⁶

To sum up, if the extant Rhesus is not by Euripides, it belongs in all likelihood to the first quarter or third of the fourth century BC. Evidently popular on stage, it remained current long enough for at least one copy to reach Alexandria, where (most probably) it was falsely identified as Euripidean. The edition by Aristophanes of Byzantium and the scholarly work we can glimpse through the hypotheses and scholia secured its place in the corpus until in late antiquity it became one of Euripides’ ‘select’ plays (ch. III.4). Like the in Hyp. (b) Rh. (and as in the case of Prometheus Bound), modern scholars have to rely on internal criteria, from language, style and dramaturgy,⁸⁷ to show that Rhesus was not written by the man tradition claims.

2. Language and Style

Of the various internal grounds that can be advanced for or against the authenticity of a text language and style are often the most conclusive, but in the case of ancient literature hardest to assess on the basis of usually limited material for comparison. In particular, the pitfalls of judging by individual words and phrases, without taking account of the subject-matter, context or an element of chance, have been brilliantly (if very polemically) exposed by Douglas Young,88 and for Rhesus we need only remember the changes that subsequent discoveries have caused in the eighteenth- and nineteeth-century lists of λεγόμενα, and expressions shared with one other (group of) tragedian(s) only.⁸⁹ Also, regarding our play, it is obvious that almost any ‘intertextual’ argument from a work later than the mid-fifth century BC entails a suggestion of petitio principii, if the genuine Rhesus of Euripides indeed belonged in that period (p. 22). Yet whereas such linguistic idiosyncrasies as the author’s repetitiveness (a) and eclectic vocabulary (b) can only be described with an eye to general tragic practice, the sheer number, character and diversity of his poetic recollections (c) should suffice to settle the question of priority and establish a connection with post-classical drama (d).

(a) Repetition

Perhaps the most noticeable (and notorious) trait of the trimeter and, to a lesser extent, the tetrameter, anapaestic and lyric versification of Rhesus 90 is the high incidence of verbal repetition, which exceeds even what is found in Euripides⁹¹ and only occasionally seems unavoidable (41–2, 125–6a, 135b–6nn.)⁹² or to fulfil a specific function (149–50, 154–5, 543–5nn.). Entire phrases, moreover, tend to have equivalents in other drama and, whether then re-used or not, occupy the same metrical position before or after the main caesurae. Unfortunately, apart from certain set locutions,⁹³ we do not know how many of these, especially Euripidean, verse-openings and -ends (recorded throughout the commentary) formed part of a developed tragic koine,⁹⁴ as opposed to a single author’s recognisable predilections. One would be surprised, however, if Aristophanes’ ληκύθιον-joke (Ran. 1198–1247) were not also aimed at this ‘formulaic’ quality in Euripides’ style, which like his other mannerisms and relative simplicity of diction would have made exploitation by later contemporaries and successors all the easier.⁹⁵

(b) Vocabulary

In its lexicon Rhesus shows a curious blend of such Euripidean plainness and a liking for the archaic and precious, which manifests itself not only in the poetic adaptations discussed under (c), but also in the more or less incidental use of remarkable words or phrases from different categories.

Like the other tragedians and lyric poets, the author of Rhesus employed a variety of epic and epic-style expressions. Some are shared with one or more of his predecessors, so that it is hardly possible to specify the source,96 others were evidently suggested by the Homeric models for his plot (cf. 72–3, 233–5a, 509b, 557–8, 908–9nn.). An ‘intertextual’ relationship can also be detected at 29 (28–9n.) … Λυκίων ἀγὸν and, with reference to Achilles or Ajax, 492 (n.) θοῦρον, 494 (494–5n.) and 461–3 (n.) … ὑπομεῖναι. Similarly, both former and envisaged battles are marked as ‘epic’ by the appropriate language (408–10a, 477–8, 479, 480, 485–7nn.), while the Charioteer in his lament and messenger speech attempts in vain to meet heroic standards (756–803, 762–3a, 776–7nn.). In contrast, several verb-forms without parallel in tragedy or lyric seem merely intended to raise the overall stylistic level: 514 (513b–15n.) ἀμπείρας, 525 (523–5a n.) δέχθαι, 629 μεμβλωκότων, 811 (810b–12a n.) ἐξαπώσατε.

The same applies to the regular presence of words (especially compound adjectives and adverbs) which on our evidence were favourites of, or probably coined by, Aeschylus: e.g. 58 σύρδην, 77, 656, 737 τορῶς, 79, 158, 476 κάρτα, 222 ἀναιμάκτῳ, 390 προσήμενον, 646 πρευμενής, 724, 805 (nn.) (-ου), 932 ϕιλαιμάτους, 962 μελάγχιμον. The list is supplemented by such unique formulations as 8 (n.) γοργωπὸν ἕδραν (~ 554 ὄμματος ἕδραν), 260 (257b–60n.) … γόον, 288 (287–9n.) αὐτόρριζον ἑστίαν χθονός and 971 (970–1n.) ἀνθρωποδαίμων. Some of these may at least in part have Aeschylean precedents.97

Our poet’s taste for the unusual is further seen in various rarities, including, as Fraenkel (Rev. 234) pointed out, words that have pre-fourth-century equivalents only in Euripides’ latest plays. Thus two adjectives that once counted as (215 δίβαμος, 737 ἀμβλῶπες) are accompanied by e.g. 296 (~ Phoen. 92 προυξερευνήσω), 426 (~ S. fr. 384), 441 (440–2n.) (= Phoen. 45), Boeotian (523–5a n.) and what appear to be medical terms at 711 (710–11n.) ὕπαϕρον, 784–6 (n.) and 789 (n.) μυχθισμόν. In view of the last two classes, it is not surprising to find also Ionisms at 322 (322b–3n.) ἐξώστης, 633 (n.) (with a predicative participle) and perhaps 881 (879–81n.) κελεύθου .⁹⁸

Owing to the theme of Rhesus, the low-style sector is particularly represented by military expressions paralleled in the historians: 6 προκάθηνται, 125 (125–6a n.) κατάσκοπον, 136 (135b–6n.) ναυστάθμων, 146, 673 (-ούς), 311 πελταστῶν, 521 (521–2n.) ξύνθημα, 595 Τρωϊκῶν ἐκ τάξεων, 604 ποιούμενον, 664 τάξιν ϕυλάξων, 768–9 κἀϕεδρεύοντας νεῶν / πρύμναισι, 954 ἔϕεδρον. By the same token, a number of colloquialisms (partly, one suspects, of the soldierly kind) cluster in the comic-satyric chasing scene (675–91, 680, 686nn.) and the speech of the humble Charioteer (759, 770–2, 784–6nn.), where they strikingly contrast with his ‘epic’ pretensions. Otherwise few words or turns of phrase can reasonably be assigned to that register: 87–9 τί …; 149 … οἳ πάρεισιν ἐν λόγῳ, 195 μέγας ἀγών (in lyrics), 285 ϕαῦλος, 574, 729, 885 ἔα (ἔα), 625 (in a Euripidean reminiscence), 870 ἅλις, 874 αὖ. Most of these had entered tragic diction with Euripides (if not other mid- to late-fifth-century playwrights) and so do not contradict our poet’s general quest for a high-flown style.

(c) Poetic Borrowings

What really distinguishes Rhesus from the body of surviving tragedy is the manner and extent to which it relies on other drama, epic and lyric poetry, ranging from the adaptation of whole scenes to the collocation of memorable words or passages, and including several instances from later Euripides and Sophocles. Naturally the divisions merge so that, for example, Odysseus’ and Diomedes’ entrance dialogue in 565–94 (n.) features such linguistic overlaps as may, in any authors who share a literary language, be due to the description of similar situations (570–1, 574–9, 574nn.). But we also find at the beginning a fine example of more specific lexical association (565–6n.).

Likewise 675–91 (n.) illustrates what may be called ‘reverse composition’ in that two successive Rhesus passages seem inspired by one model. The agitated choral imperatives at 675b (n.) θένε θένε θένε and 685 (n.) can all be related to Ar. Ach. 280–3 (Χο.) οὗτος αὐτός ἐστιν, οὗτος· / βάλλε, βάλλε, βάλλε, βάλλε, / παῖε, μιαρόν. / ; ; and an analogous picture emerges if one compares 677/680 (680n.) · τοῦτον αὐδῶ. / δεῦρο and 730 (n.) σῖγα ὕϕιζ᾽· … with Ar. Ach. 238–40 (Χο.) σῖγα πᾶς. ἠκούσατ᾽, ἄνδρες, ; / αὐτός ζητοῦμεν. δεῦρο / ἐκποδών. Without connection to a particular scene add Rh. 48 + 261–3 (nn.) ~ IA 171–7 (cf. pp. 34, 36), Rh. 970 (970–1n.) + 972 ~ Cyc. 293–4 and perhaps Rh. 949 + 952–3 (nn.) ~ HF 911–12.99

In order to elucidate our poet’s ‘intertextual’ practice, it will nevertheless be well to take a systematic approach and, moving from large to smaller patterns, consider (i) ‘one-to-one’ scenic adaptations, (ii) scattered reminiscences of continuous tragic passages, (iii) ‘mosaic-like combinations of borrowed expressions’,¹⁰⁰ (iv) ‘noun with attribute’ phrases as a sub-category of the latter and (v) peculiarities of diction that appear to result from unconscious reception. Often we will discover contextual and purely verbal echoes side by side, and it should cause no surprise that war-plays and ‘slices from the great Homeric banquets’ (Ath. 8.347e ~ Eust. 1298.56–8 = A. T 112a, b TrGF) supply a significant proportion of its assumed sources. To judge by the number of parallels with A. Myrmidons, we might easily be able to expand the selection if more were preserved.¹⁰¹

(i) While different typological models lie behind the entry of Odysseus and Diomedes and their encounter with the chorus (above), other scenes and choral odes in Rhesus were evidently inspired by a single fifth-century precedent. The clearest example is Athena’s epiphany, which in structure and expression (rather than spirit and dramatic purpose) is heavily indebted to the Ajax prologue (595–674n.). A similar relationship exists between the ‘Dawn-Song’ in 527–64 (n.) and the first three stanzas of the Phaethon parodos (63–86 Diggle = E. fr. 773.19–42), although the lyrics were far more expertly converted into something new.102 On a smaller scale, the third stasimon of Ajax (1185–1222) is probably echoed in the chorus’ reverie of peace at 360–7 (n.), and we may also see the influence of Bacchae in Dolon’s disguise (201–23n.) and the Shepherd’s report of Rhesus’ approach (264–341n.).

Much less can be said with regard to Sophocles’ Poimenes (cf. 264–341, 342–79, 388–526nn.). Wilamowitz went too far in seeing there the fifth-century prototype of Rhesus,103 especially since only one slight verbal parallel survives among its certain remains (770–2n.).¹⁰⁴ But he was right to observe the analogies between the Idaean Shepherd and the Goatherd who announces the arrival of Cycnus’ army at Troy (S. frr. 502–4),¹⁰⁵ and also between the hero’s boasts (S. frr. 500?, 501, 507?) and those of Hector, Dolon and Rhesus.¹⁰⁶ That our play may generally have been inspired by a tragedy on (or recounting) Cycnus’ death is further suggested by Ar. Ran. 962–3, where ‘Euripides’ criticises ‘Aeschylus’ for having terrified his spectators ‘by creating characters like Cycnus and Memnon with bells on the cheek-plates of their horses’ (tr. Sommerstein). Aristophanes’ to describe the exotic Trojan allies arguably left its own trace in Rh. 306b–8 and 383–4 (nn.).107

(ii) Another category of scenic adaptation involves the openings of three, possibly four, tragedies which, like Ajax and Poimenes, resemble Rhesus in subject-matter and/or the basic situation.

Klyve (41–2) had already isolated from the many lexical and contextual echoes of Seven against Thebes a series that almost in order of appearance refers to the prologue (1–77) alone: Rh. 19 νυκτηγορίαν, 89 νυκτηγοροῦσι ~ Sept. 29 νυκτηγορεῖσθαι, Rh. 20–2 (n.) ~ Sept. 59–60, Rh. 514 ἐπ᾽ ἐξόδοισιν … ~ Sept. 33, 58, Rh. 632 … στρατοῦ ~ Sept. 36, 41 (+ 369), Rh. 932 ϕιλαιμάτους ~ Sept. 45 .108 An equally strong tendency of this kind seems to exist in connection with Persians, where the anapaestic-lyric parodos (1–154) provided not only numerous salient words and phrases,109 but also the notion of a vast army doomed to destruction with its overbold leader (264–341, 309–10nn.) and a model for bringing the chorus on stage at once (cf. p. 40 and 1–51n.).

The fact that this type of borrowing only works in one direction becomes relevant for the dating of Rhesus when a similar pattern emerges with regard to the disputed prologue anapaests of Iphigenia in Aulis (1–48 + 115–62):¹¹⁰ Rh. 12 θρόει ~ IA 143 (?), Rh. 16 ~ IA 2–3, Rh. 274 χειρῶν … βαστάζομεν ~ IA 36 πρὸ χερῶν … βαστάζεις and Rh. 529–30 ἑπτάποροι / Πλειάδες ~ IA 7–8 ἐπταπόρου / Πλειάδος. The impression that our poet used Euripides’ final play – and knew it more or less as we do – is corroborated by several further analogies, especially with its parodos (IA 164–230): Rh. 48 + 261–3 ~ IA 171–7, Rh. 356 πώλοις ~ IA 220–2 … βαλιούς, Rh. 467 … ἀπουσίας ~ IA 651 … ἀπουσία, 1172 … διὰ ἀπουσίας. By contrast, the occurrence of adjectival in Rh. 718 and IA 1306 may be coincidental, and the same applies to the conceptual parallel in Rh. 904–5 (n.) and IA 469–70.¹¹¹

Among lost plays, it is interesting to observe that three out of five verbal or syntactical overlaps with A. Myrmidons come from its likewise anapaestic-lyric parodos (cf. 404–5, 557–8, 763b–4a nn.).112 Given that we can only compare what exists, it would be imprudent to place too much weight on this; yet the use of the first two lines of the lyric part (A. fr. 132) in the parody of Aeschylean song at Ar. Ran. 1264–77 probably indicates that the passage was famous,¹¹³ and it seems safe to state that the beginnings of texts are generally better remembered than the rest. Our poet would have been no exception to that rule.

(iii) Fraenkel based his case against the authenticity of Rhesus above all on the author’s habit of combining notable expressions of different provenance, which in that form is not found in any of the ‘three great tragedians’. He did not usually spell out the ‘intertextual’ relationship between the passages, so that further systematisation is possible. At times his collection will silently be supplemented.

Some examples of this ‘mosaic-technique’ show a contextual overlap with both supposed models. The Charioteer’s grim depiction of Rhesus’ death at 790–1 (n.) δὲ σϕαγῆς / με δυσθνῄσκοντος αἵματος νέου recalls both the murder of Agamemnon at Ag. 1389–90 and, in the irregular compound participle δυσθνῄσκοντος, that of Aegisthus at E. El. 842–3. ‘Wily’ Odysseus, connects 498–9 ἔστι δ᾽ αἱμυλώτατον / κρότημ᾽ Ὀδυσσεύς with Ai. 388–9 and S. fr. 913, of which the first supplied the superlative epithet and the second the disparaging noun (498b–500n.). Similarly, Achilles in 122 (n.) acquired expressions used elsewhere for proud or overconfident warriors (Sept. 447–8, Ai. 221/2, 1088 + Or. 1568), while in describing Odysseus at 713b–14 (n.) our poet may have thought of Orestes and Pylades on their way to kill Helen (Or. 1125 + 1271–2).

More often only one of the two (or three) identifiable source passages resembles Rhesus in content. When in 404–5 (n.) Hector accuses Rhesus of having failed him as an ally (σὺ δ᾽ ὢν βαρβάρους / Ἕλλησιν ἡμᾶς προύπιες τὸ σὸν μέρος), the verb προπίνω, which in the same sense is previously found only in A. fr. 131.3 (Myrmidons), follows a variation of the contextually unrelated IT 31 … βαρβάροισι .114 Rhesus justifies his delay with a Scythian attack and the generally difficult journey to Troy (424–42). The account takes much from Persians (cf. 388–453n.), including two impressive ‘composite passages’: Rh. 430–1 (n.) ~ Pers. 816–17 + Alc. 850–1 (IT 300) and Rh. 440–2 (n.) ~ Pers. 500–1 + Phoen. 45–6 + E. El. 820, HF 959.115 An excellent specimen of literary ‘purple patches’ from outside tragedy being combined in such a way is Rh. 72–3 (n.) ~ Il. 8.512–15 + Pi. Pyth. 1.28 + Pi. Isthm. 8.49–50.

Even in the case of purely lexical adaptations a connection of thought can sometimes be detected. At 817–18a (n.) μάραγνά γ᾽ μόρος / σε τοίαδ᾽ two notable words from the Oresteia come together in one line: cf. Cho. 375 … and Eum. 186–7 … / δίκαι τε. The latter stems from Apollo’s list of cruel punishments meted out in barbarian lands (Eum. 186–90), which also supplied a precedent for Rhesus’ threat to impale Odysseus in 513–15 (512–17, 513b–15nn.). No such ‘intertext’ exists for 739–40 / ὑπασπίδιον ; Yet the metrical association with Ai. 408 may suggest that our poet also had in mind Ai. 1204 οὔτ᾽ ἰαύειν, where the epic verb is used in the same way (cf. 740n.).

(iv) In CQ n.s. 60 (2010), 345–51 I examined ‘noun + attribute’ collocations unique to Rhesus, in which each term can independently be traced to a common source. The intention was to defend 118 (n.) χνόας as an authorial conflation of S. El. 745–6 δ᾽ ἄξονος χνόας, / κἀξ , but a persistent and generally perhaps less conscious form of poetic borrowing also emerged. This section will review the earlier finds and add two further examples of a similar kind.

In one variety of the pattern a choice adjective qualifies a noun which separately stands nearby in its closest or only extant parallel. Thus (‘seafaring’) is restricted to Rh. 48 (n.) ναυσιπόρος στρατιά and IA 171–3 (ἔμολον …) ἐσιδοίμαν / ἡ- / , and although it may also have occurred elsewhere,116 other echoes of the IA parodos (cf. p. 34) suggest it was recalled from there and joined with the preceding στρατιά(ν). The same phenomenon (with ‘composition’) can be detected in Rh. 605–6 (n.) τὰς … / … ~ Tro. 562–6 (+ Andr. 399) and Rh. 618 (n.) κύκνου ~ Hel. 215 + E. El. 151–2.

A second class of attributive collocations tends to involve unusual nouns. Here Fraenkel (Rev. 231) had already analysed 569 (568b–9n.) as a fusion of Sept. 245 and 249 ἀραγμός, and Fantuzzi (MD 36 [1996], 183–5) did the same for 715 (715–16n.) in semantic relation to Od. 4.245 ἐοικώς and 4.247–8 ἄλλῳ δ᾽ αὐτὸν ϕωτὶ … ἤϊσκε / ΔΕΚΤΗΙ (taken to mean ‘beggar’ rather than as a proper name). Similarity of context also accounts for Rh. 276–7 (n.) ~ Or. 688–90, whereas Rh. 118 (n.) ~ S. El. 745–6 (above) is explicable by recollection of the famous ‘fake’ S. messenger speech in S. El. 680–763.

On the analogy of these, and especially the linguistically awkward 118, it may not be far-fetched to see in 111 (110b–11n.) νυκτὸς ἐν (with preceding in 109) a somewhat muddled adaptation of Ag. 22–4 ὦ χαῖρε , / καὶ / . More straightforwardly again 255b–7a (n.) / ἔχων ἐπὶ γαίας / seems to echo both Hec. 1058–9 and A. fr. 57.8–11 (Edoni).

(v) Finally, some light may be thrown on other difficult expressions by supposing they betray mechanical adaptation or even misunderstanding of tragic material.

The clearest examples of this are probably 389 (388–9n.) σ᾽ προσεννέπω, where reference to Ai. 624–5 … ἁμέρᾳ, / γήρᾳ explains rather than justifies the application of … to Rhesus’ last-minute appearance, and 770 (770–2n.) μελούσῃ καρδίᾳ λήξας ὕπνου, which Fraenkel (Rev. 238) had recognised as a hasty reworking of Sept. 287 μέλει, δ᾽ οὐχ κέαρ (+ 288–9 … / μέριμναι).

To these we may add Rh. 201 (201–2n.) ~ Tr. 262–3, Rh. 604 (603–4n.) πλατεῖαν ἐσδρομὴν ποιούμενον ~ E. fr. 495.29–30 (Captive Melanippe) [λόγ]χῃ δι᾽ / [παίσ]ας, Rh. 636–7a (n.) ~ Cho. 659–60, Rh. 780 (n.) … δόξα τις ~ OT 911 … δόξα μοι (in a different sense) and Rh. 870 (n.) · ἅλις γὰρ τῶν ὄχλος, which owes much of its peculiarity to the fact that μὴ (~ Alc. 690, IA 1419) lacks a sense-defining adverbial supplement. On a more general level, the choral entrance-announcement for Aeneas at 85–6 (n.) is too close to Hec. 216–17 to be fully appropriate to the situation, while the Muse’s ‘mortal’ pain at the loss of her son in 980–2 (980–2, 980nn.) may come from Hyps. fr. 60 ii.90–1 Bond = E. fr. 757.921–2.

(d) Conclusion and Outlook

The preceding enquiry should leave no doubt that, arguably from reading as well as theatre practice, our poet was familiar with the whole range of fifth-century drama and, alongside epic and lyric poetry, exploited even its most original language as a means to express his own thoughts and embellish the vernacular of his trade. Some of his poetic habits can also be related to post-classical tragedy, although comparison there is even more severely hampered than in the case of earlier works by the scarcity of what remains.

In a manner redolent of the Rhesus ‘Dawn-Song’ and the Phaethon parodos, Carcinus the Younger (floruit 380–76 BC)117 largely followed the third stasimon of Helen (1301–68, especially 1301–37) when he composed his iambic account of Demeter’s search for Kore in Sicily (Carc. II TrGF 70 F 5),¹¹⁸ and it is instructive to see how two other passages contributed a resounding epithet and a useful ‘verse-end formula’ each: Hel. 517–19 Μενέλαος οὔ- / μελαμϕαὲς / ἔρεβος ~ Carc. II TrGF 70 F 5.3 γαίας εἰς μελαμϕαεῖς ¹¹⁹ and Cyc. 95 πόθεν πάρεισι Σικελὸν Αἰτναῖον πάγον ~ Carc. II TrGF 70 F 5.6 καὶ Αἰτναίοισι Σικελίας πάγοις.

Parallels for our poet’s ‘mosaic’ technique are harder to find in genuine fourth-century fragments than in certain actors’ interpolations. The second iambic prologue to Rhesus, which is composed of various ‘fifth-century’ lines and half-lines,¹²⁰ provides an extreme example. More to the point, Fraenkel compared (supposed) insertions in Phoenissae¹²¹ and the undoubtedly spurious end of Seven against Thebes (1005–78), where in addition to borrowings from the play itself,¹²² the influence of Antigone and Phoen. 1625–82¹²³ extends to the smallest matters of diction.¹²⁴ The difference from Rhesus lies in the straightforward transposition of a tragic motif (Creon’s and Antigone’s dispute over the burial of Polynices) to an earlier dramatisation of the same story and, as mutatis mutandis in Phoenissae, the all but exclusive concentration on the relevant sources for theatrical and linguistic inspiration alike. One would like to know how the celebrated playwrights of the fourth century treated the classic material.

3. Dramatic Technique

Rhesus is well known for rapid action and theatrical excitement, which must have been a major cause of its initial survival. As in its language, we observe, among several ‘unconventional procedures’ discussed in the commentary,125 a highly idiosyncratic mixture of ‘old-style’ and ‘progressive’ elements, aimed apparently at indulging a later audience’s tastes with a maximum of spectacle and naturalistic turns.

Both these objectives underlie the sentry chorus, which has been created against the testimony of Il. 10.416–20 δ᾽ ἃς εἴρεαι, ἥρως, / ῥύεται στρατὸν οὐδὲ ϕυλάσσει. / ὅσσαι ἐσχάραι … / οἳ δ᾽ ἐγρηγόρθασι / ἀλλήλοις. Their role is the only suitable one in the circumstances. ¹²⁶ It allows them to take an active part in the play and, together with their excitable character, easily motivates the direction and manner of their stage-movements. The much-criticised lack of military common sense in that all the watchmen go respectively to rouse Hector (1–51n.) and their Lycian relief (527–64n.)¹²⁷ would either have passed unnoticed or been accepted as a piece of dramatic necessity.128

An archaising impression is given immediately by the anapaestic-lyric parodos, which does not resemble anything Euripidean and, for all its divergences in form and spirit, rather harks back to the choral openings of Aeschylus’ Persians, Supplices, Myrmidons, Nereids, Niobe and especially perhaps the spurious Prometheus Unbound (1–5ln.). Not for nothing, moreover, the first words of the turbulent, and possibly ‘scattered’ (σποράδην), epiparodos recall the Erinyes on the hunt (675–91, 675b nn.), and it may also be relevant that both Myrmidons and Nereids belong to the Iliadic(-Cyclic) Achilles trilogy 129 and show other connections with our play (cf. pp. 13–14, 34–5).

From the historical perspective, this handling of the chorus – almost as if our poet had followed Aristotle’s precept that it ‘should be treated as one of the actors, form an integral part of the whole and join in the action’ (Poet. 1456a25–7 καὶ τὸν χορὸν δὲ ἕνα δεῖ τῶν ὑποκριτῶν, καὶ μόριον εἶναι τοῦ καὶ συναγωνίζεσθαι) – may have been a reaction against its gradual loss of significance, and particularly Agathon’s alleged introduction of (‘interludes’).¹³⁰ We need not believe in Aristotle’s linear account of this decline (Poet. 1456a27–30), and there is good reason to think that, as in Middle Comedy, at least some of the post-classical tragic titles in the plural (e.g. Agatho Mysoi, Dicaeog. Kyprioi, Moschio Pheraioi, Lyc. Marathonioi) designate their, however detached, chorus.¹³¹

Another feature redolent of early tragedy is the neglect of the stage-building as part of the dramatic landscape (‘Scene and Setting’, 114 with n. 1).¹³² In accordance with Iliad 8 and 10, Hector’s sleeping-place in the temporary Trojan camp, which forms the centre of the action, is called εὐναί, κοῖται (1n.) and (9n.) rather than ‘hut’ or ‘tent’ as in other military plays (cf. Ai. 3–4, Hec. 53, Tro. 32–3, IA 1, 12, 189–90)¹³³ and indeed also in Rhesus when the Greeks are concerned (45, 61, 255). Likewise, the Trojans and their commander are said to rest in full armour by or under the cover of their shields (20–2, 123–4, 740nn.), while Rhesus has been assigned a bivouac (κατηύνασται) outside the ranks (611, 614). In contrast to Eum. 566–1047, however, where the presumably newly-invented skene fades into oblivion behind the Areopagus court,¹³⁴ the stage-building here retains its physical use as for Athena and the Muse (595–674, 882–9nn.).

Given this archaising mise-en-scène, it would have been fitting for Rhesus to enter on a chariot, drawn by his splendid horses, at 380–7 (n.). To judge by Ar. Ran. 962–3 (… οὐδ᾽ αὐτούς, / Κύκνους Μέμνονας κωδωνοϕαλαροπώλους), Aeschylus was famous for this device, which would have made particularly good sense on the pre-skene stage and, among his extant plays, is certainly found in Persians (155–8, 607–9) and Agamemnon (783–809, 906, 1039, 1054, 1070).¹³⁵ Yet, as is argued in the note above, we cannot be sure that Rhesus shared in the apparent fourth-century revival of chariot-borne arrivals. The text gives no conclusive evidence, and for strategic reasons our poet may not have wished to show the horses on stage. The vivid picture drawn by the Shepherd in 301–8 perhaps had to suffice.

Traces of new developments include the management of Athena, the Charioteer and the Muse. While Athena’s mid-play epiphany as such appears to have had Aeschylean and Sophoclean as well as later Euripidean precedents (cf. 595–674n.), her transformation into another deity, Aphrodite, has often been regarded as impossible for the fifth century 136 and so far is paralleled only once in the mythical burlesque of Middle Comedy (642–74n.).

The multiple functions of the Charioteer as victim, messenger and accuser (728–55, 756–803, 804–81, 833–81nn.) also transcend the conventions of classical tragedy, as does the Muse, who unlike any other deus ex machina sings a strophic dirge from ‘on high’ (882–9, 890–914nn.). In terms of staging, however, an even bolder move may already have been attempted in Prometheus Bound, if the chorus of Oceanids (minor deities like the Muse) performed their ‘flying’ parodos and initial conversation with Prometheus (PV 114–283) on the skene-roof or somehow suspended from the crane.137

The two epiphanies in Rhesus, as well as the brief, ‘purpose-oriented’ entries of Aeneas (85–148n.) and Paris (595–641, 642–74nn.), also contribute to the play’s very large number of speaking characters (11), which is matched only by the late-Euripidean, and much longer, Phoenissae (11) and Orestes (10).¹³⁸ By the same token, its structurally well-placed pair of messenger-scenes seems to follow Iphigenia in Tauris (238–343, 1284–1434), Helen (597–757, 1512–1618), Phoenissae (1067–1283, 1335–1479), Orestes (852–956, 1368–1536), Bacchae (660–774, 1024–1152) and, as we have it, Iphigenia in Aulis (414–542, 1532–1629) rather than isolated earlier examples of such doubling as Sophocles’ Trachiniae (180–99 + 335–496, 734–820) and Antigone (223–331, 384–445, 1155–1256 + 1278–1346). The fact that our poet adopted these techniques from far more expansive plays goes some way towards explaining our impression of hectic and at times truncated action, although given the underlying concept of divine responsibility (ch. I), it may also be ‘an index of the complexity of events and the separateness and weakness of each character’s impact on the overall [plot].’¹³⁹

For lack of evidence we cannot tell how far the eclectic stagecraft of Rhesus was typical of post-classical tragedy. But various ancient sources suggest that the late fifth and fourth centuries saw an increasing preference for spectacle, both in new productions and in revivals of old plays (especially, it seems, Aeschylus).¹⁴⁰ Rhesus is well placed in this development and appears to have been popular enough to be revived itself, albeit with a choice of two iambic prologues prefixed to the parodos. Whoever wrote them meant to add clarity without depriving the audience of the first exciting scene.

4. Rhesus and the ‘Euripidean Selection’

Although Rhesus had been edited in Alexandria under a famous poet’s name, it might still have perished before reaching the medieval tradition. Yet by the second or third century AD it had become part of the ‘Euripidean Selection’, the standard repertoire often plays (comprising also Alcestis, Medea, Hippolytus, Andromache, Hecuba, Troades, Phoenissae, Orestes and Bacchae) that were a mainstay of the ancient and Byzantine school curricula and survive in a separate line of transmission together with ‘old’ scholia.141 It is likely that this canon resulted from a prolonged and self-reinforcing reduction process rather than, as Wilamowitz supposed, the choice of a single influential school-master.¹⁴² Of the 75 or so Euripidean dramas that the Alexandrians collected some will already have been better known and so perhaps received priority treatment from exegetes; conversely, the existence of appropriate, and regularly modified, commentaries is liable to increase the popularity of a text. In time the range of plays performed and widely read diminished until it focussed on the ten in question. Barrett further suggests that seven to ten dramas (we have seven each of Aeschylus and Sophocles) were the content of a late-antique codex.

Spectacular stage action apart, the qualities which ensured the pre-Hellenistic survival of Rhesus will also have recommended it to a reading audience: a varied plot based on a well-known episode from ‘Homer’, relatively easy language (despite its many idiosyncrasies) and its brevity compared with other tragedies. Deviations from Iliad 10 invited literary adaptation, while for the last two reasons it may have been preferred to Aeschylus’ and Sophocles’ now lost Homeric tragedies,¹⁴³ especially when it came to finding suitable texts for education. By the same token Prometheus Bound stood at the head of the Byzantine triad of Aeschylus,¹⁴⁴ and the Batrachomyomachia ‘was adopted as a school text, to make a short and entertaining introduction to Homer’.¹⁴⁵

It is to some extent possible to trace the progress of Rhesus into the ‘Selection’ and beyond by way of its ancient literary reception and quotations in scholarly and other sources. For practical reasons the two categories will be treated separately, although they partly overlap in time and supplement each other.

(a) Literary Reception

A number of Greek and Latin authors, both poets and prose writers, show the influence of Rhesus, suggesting that the drama was familiar in literary circles. In some cases, notably Menander, Virgil and Longus, the allusions must also have been intended to be recognised by the audience or readers.

In pre-Alexandrian times already Menander may have referred to Rhesus, if Fantuzzi and Konstan are right to see an echo of the ‘guessing-game’ between Hector and Dolon (161–83) in Peric. 271–91.146 There Moschion’s slave Daos claims credit for the move of his master’s beloved Glycera into the house of his foster mother and asks for a reward. After going through several options for a ‘good life’ (including the command of the Greek land forces), they settle on a cheese stall, and Mos-chion asks Daos to investigate the situation in the house (Peric. 295–6 εἰσιὼν δέ μοι σύ, Δᾶε, κατάσκοπος / γενοῦ …). Despite obvious comic distortion, the subject matter of the debate and its connection with a ‘spying mission’ – occurs nine times in Rhesus (125–6a n.) – makes adaptation of the Dolon scene a distinct possibility.

Accius (170 – ca. 86 BC) perhaps also referred to Rhesus, although too little of his Nyctegresia, another tragedy based on Iliad 10, survives to tell whether any part of it came from our play.147 More promising is a fragment from his Antigone (Acc. Ant. fr. III Dangel = IV TRF³), which bears a striking resemblance to Rh. 532–3 (n.).

In the first century BC Parthenius of Nicaea wrote his Παθήματα, a collection of thirty-six mythological vignettes dedicated to C. Cornelius Gallus. His tale of Rhesus and Arganthone (no. 36) corresponds to our play in significant points (cf. p. 17), and there may even be a verbal echo in the (‘tribute’) the hero imposed on the peoples he conquered: Rh. 435 (434–5n.) ~ Erot. Path. 36.1. Both Parthenius and Asclepiades of Myrlea (who may have been the source) were scholars, of whom first-hand acquaintance with Greek tragedy could be expected.

By far the most extensive (and subtle) adaptation of the material stems from Virgil. The relationship of Iliad 10, Rhesus and the Nisus and Euryalus episode in Aen. 9.176–458 has been the subject of comprehensive studies¹⁴⁸ and can only be outlined here. While ‘Homer’ remains the main source of the nocturnal adventure, during which Nisus and Euryalus (on their way to Aeneas through enemy territory) devastate the Rutulian camp, but eventually fall victim to recklessness and are killed by an arriving cavalry unit, Rhesus appears to be responsible for numerous details. The Charioteer’s speech (756–803) especially provided Virgil with models for the careless (and drunken) disorder among the Rutulians (762–9n.), the deaths of Remus and Rhoetus (789, 793–5a nn.) and perhaps also Nisus’ and Euryalus’ reaction to the challenge of Volcens (774b–5n.), which itself resembles that of the chorus to Odysseus and Diomedes at 682 (n.). Before their capture Nisus had restrained his younger companion from further killing in much the same way as Athena does for the Greeks in 668–72 (668–9n.). At the outset of the expedition observe how Aletes’ praise for Nisus and Euryalus (Aen. 9.247–50) reverses the expression, but not the sentiment, of Rh. 245b–9a (n.).

While it seems improbable that Ov. Her. 11.111 reflects Rh. 896–7 (895–8n.), the verbal similarities between both D. Chr. 55.14 and Rh. 854b–5 (n.) and Aristid. Or. 1.106 Lenz–Behr and Rh. 335 (n.) may at least be due to unconscious reminiscence. The first two passages especially are related in content, and the relevant part of the Rhesus lines (οὐδ᾽ … ᾖσαν) pervades the later lexicographical tradition (below). But coincidence in the choice of (rather ordinary) words cannot be ruled out.

Around 200 AD, finally, Longus transferred to the pastoral world the motif of Dolon’s wolf-disguise. When the cowherd Dorkon plans to sneak up on Chloe dressed as a wolf, the actual donning of the hide (1.20.2) reads almost like a prose version of Rh. 208–11a (n.) – to the point that the military background of the scheme (cf. 201–23n.) is recalled in the comparison of the animal’s ‘gaping jaws’ (χάσμα) with a hoplite’s helmet (208–9n.). The near-homonymy Dolon – Dorkon will also have played a part.149

(b) Ancient and Medieval Quotations, Lexica

Owing to its small number of and other generally applicable utterances, Rhesus is under-represented in the anthological tradition,150 and for similar reasons presumably held no appeal to Plutarch or Athenaeus.¹⁵¹ Quotations and scholarly allusions begin only with Nicanor in the second century AD¹⁵² – that is when the ‘Selection’ was already in the ascendancy – and therefore, like our manuscripts and papyri (ch. IV), rather testify to the drama’s relatively subordinate status within the canon. In the fourth and fifth centuries one verse each of Rhesus occurs in Ap(h)-thonius and St. Basil of Caesarea,¹⁵³ while Orus in his Atticist lexicon quotes 854b–5 with an important textual variation.154 Of Byzantine authors, Eustathius and Tzetzes used Rhesus passages in their commentaries and poetic works.155

Although Rhesus could probably never compete with plays like Medea or Hippolytus, let alone the Byzantine triad Hecuba, Orestes and Phoenissae,¹⁵⁶ it continued to be read and annotated for private and educational use. Witness to this are not only the ‘old’ scholia preserved in V, but also the combined lexicon of Cyril and Hesychius, which contains a large number of entries related to our play.157 Most of these, especially in the Cyril part, are explanations of unusual words or phrases, excerpted from a basic Euripides glossary that was meant to render the ‘Selection’ more accessible at a time when few in or outside school would have been familiar with all the intricacies of the tragic language. The fact that proportionately as well as in the absolute number of glosses Rhesus is surpassed only by the much longer and already prevalent Orestes and Phoenissae shows that it had a firm place in the late-antique curriculum.

The ancient and medieval history of Rhesus, as sketched at the beginning and end of this chapter on ‘Authenticity and Date’, provides a salutary reminder that dramatic and literary tastes change over time. Modern scholars who deny Euripidean authorship because of the play’s poor quality usually fail to account for its survival not by chance, but as a canonised work of the most celebrated Greek tragedian. Authors like Virgil and Longus found in it aspects, even expressions, worth borrowing, while from the fourth century BC into the middle ages theatre-goers, schoolmasters and general readers would have appreciated it for some of the very features later condemned. Prometheus Bound represents a parallel in more than one way.