David Kovacs’ predecessor as Loeb editor of Euripides, A. S. Way, first published his verse translation in 1894-1898. Although his Loeb edition (1912) postdates Murray’s OCT, Way described its Greek text as “eclectic,” and he furnished it with barely any critical notes. To say nothing more, this work has long been due for replacement, and K. is to be congratulated on this second volume, in which he continues to meet the high standards he attained in volume 1 (Cambridge, Mass. 1994). The book fully deserves the wide circulation for which one hopes it is destined.
With regard to the translation, K. wrote in volume 1 that his goal was to translate “as literally as respect for English idiom allowed.” While the result indeed earns highest marks for accuracy, it is also pleasant and readable. As previously, K. provides occasional blank verse in passages of stichomythia; otherwise the medium is fairly unadorned prose. A certain amount of high-flown diction (e.g. “beetling,” “bedfast,” “mere” (= “marsh”), “twain”) seems more prominent in Hippolytus than in the other translations. In the not very numerous notes to the English pages, K. concisely identifies names, places, and customs. A few notes offer brief, non-technical justifications of textual choices or stage directions. Each play is also equipped with a short, workmanlike introduction.
With regard to the Greek text, K. set forth his editorial principles and explained his system of reporting variants in volume 1 (p. 36-9). There he cautions that “The notes to the Greek pages are not intended to be an apparatus criticus,” and he refers his readers to Diggle’s OCT for reports of manuscripts. His notes “primarily list conjectures adopted.” (They also, appropriately, include conjectures made by K. himself but not promoted to the text, and a small number of such conjectures made by others.) K. has written literary studies of Euripides, including two monographs covering three of the four plays included in the volume under review: The Andromache of Euripides (Chico, Calif. 1980) and The Heroic Muse: Studies in the Hippolytus and Hecuba of Euripides (Baltimore 1987). Nevertheless, he has devoted his greatest energies to textual studies, which have been appearing for nearly 20 years. In the rest of this review, I concentrate on the present volume’s contributions as a critical edition. It is impossible to do this fairly without making frequent reference to K.’s Euripidea Altera (Leiden 1996), which contains textual studies on the four plays of the Loeb volume 2, as well as the four plays to be included in volume 3. I thank Professor Kovacs for kindly allowing me to see Euripidea Altera (hereafter EA) in advance of publication. I emphasize that I am not reviewing that book (a review will appear soon in this journal) and that I have learned much from it that is not reflected in my references to it here, which are limited by the goal of characterizing the Loeb text.
As readers of K.’s textual studies are aware, he is enviably well versed in the centuries-long tradition of critical work on Greek tragedy. In constituting the Greek text of the present volume, he draws heavily on this tradition; the more recent scholars to whom he owes most are Diggle, Willink, and West. As these affinities might suggest, and as his earlier work certainly proves, K. is not a timid editor. Boldness is not necessarily demonstrated by the roughly 30 places in the present volume where K. prints a conjecture and Diggle an obelus, for these may be explained by what K. notes is a Loeb editor’s greater need to print a continuously readable text. On the other hand, the cases where Diggle retains the paradosis while K. prints a conjecture, about 90 in all, as compared to only about 30 in the other direction, are fair evidence that K. is generally more willing than Diggle to diagnose corruption. And of course Diggle himself is hardly a conservative editor. (As a modest illustration of this non-controversial point, I record that over a randomly selected 200-line stretch of Heraclidae (382-583), there are 16 places where both Diggle and K. print conjectures while the Teubner editor A. Garzya retains the transmitted reading.) The frequency with which which K. deletes verses (in the present volume most notably in Hecuba) and posits lacunae (in the present volume most notably in Heraclidae, though with an important exception to be discussed shortly) is also comparatively high. Overall, K.’s text is different from Diggle’s in about 50 places in each play (somewhat more in Andromache). In what follows, I comment on a small selection of these passages.
Beginning with Heraclidae, I first note that readers will encounter more of K.’s Greek here than they might have expected: 12 trimeters in all (one of them in the apparatus only). K. also prints three trimeters supplied by other modern scholars. It is only fair to add that he holds it to be a scholar’s duty to write a Greek supplement when claiming that a text is lacunose. Here are comments on a few specific passages. 107-8. In EA, K., arguing for Elmsley’s πέλει over transmitted πόλει, wonders “why any dative is wanted with an expression as generally valid as ‘It is a godless thing’.” But perhaps the threat to the city, typical of suppliant tragedy, is worth repeating; the chorus’ answer to the herald here indirectly responds to Iolaus’ challenge at 69-72. 147-8. In EA, K. considers the absence of a parallel “for an expression like ‘seeing some folly ἐσσὲ” a blow against the reading of L. But he knows that Diggle and Wilkins do not try to defend it that way, but rather construe ἐς σὲ with ἦλθον. K. has called this “a desperate attempt to avoid an obvious correction,” but that is not the same as refuting it. 169. While K. daggers this verse in his Loeb text, in the apparatus and in EA he makes a second attempt to emend it (cf. GRBS 29 (1988) 122), this time diagnosing a one-line lacuna. West’s tentatively proposed change of μόνον to μόνην is simpler and better. The herald imagines an objection “Hope alone will find the best result.” This yields good sense without our having to take “hope” as “acting optimistically,” as West apparently suggests. Hope occurs with a variety of verbs in combinations that provide little or no challenge to understanding. For example, a little later in this play Iolaus reproaches hope with delighting him while οὐ μέλλουσα διατελεῖν χάριν (433-4). “Hope will find the best” differs little from “hope will accomplish something to be grateful for.” A pithy, “proverbial” expression would suit the context well; with K.’s additional line, the sentence is flabby. 223. K. succeeds in undermining Jackson’s proposed emendation. 320-2. By diagnosing and then supplementing another one-line lacuna, K. achieves parallelism to Her. 1329-33. But the objections to the paradosis made in EA strike me as over-subtle. 619. K. gives good reasons for preferring Lesky’s στένει to Elmsley’s φέρε. 640. K. now prints Willink’s improvement of K.’s own earlier conjecture. Objections to the transmitted text are strong. 661-2. Objecting to the form of the double question, K. indicates a lacuna between these lines and supplies a trimeter to fill it. But I do not see why he regards τί as “part of” the participial phrase. The first question is “Why is he staying away?” just as K. says it should be. 1015. In K.’s solution to this old crux, Eurystheus will in effect say “Thank you” if he is spared by Alcmene. This ill suits his attitude of defiance (especially 983-5) and makes him appear to imagine mercy as a real possibility, as nothing else in the speech does.
The ending of Heraclidae presents a series of puzzles, which K. treats at length in EA. His solution posits no fewer than four lacunae of one trimeter each, deletion of one verse (1047; the deletion, proposed by Wecklein, is said by Wilkins to produce a “disastrous” result), and emendation of the intractable dogs at the end of 1050. After these operations, K. believes he can dispense with the lacuna indicated after 1052 by Hermann (according to Murray and Diggle, but K. argues that insufficient evidence exists for recovering Hermann’s intentions). Scholars will have to study the intricate argument in depth; here I offer comments on only part of it. First, the dogs. K.’s χύσιν is ingenious but implies that Hyllus’ servants (whom Alcmene commands as if they were her own) will first kill Eurystheus and then offer libations to him; but the cult he is destined to receive will come from Athenians. With Willink’s θανόντι in the next line, it becomes barely possible to understand Athenians as subject of δοῦναι, but it is hard to see why Euripides would have had Alcmene be vague on this point. (K.’s solution also depends on his interpretation of 1040-4, where he sees libations as after all belonging to the cult of Eurystheus.) Second, the lacunae in the stichomythia (after 962, before and after 970). K. is not the first to suspect gaps in these places, but he makes the best use of them to ease difficulties that have been felt later in the scene. With his supplement after 970 (Servant: “It would, I admit, have been right to kill him then”), he softens the Servant’s statement of the city’s will in a way that could help make it credible that the chorus eventually accept, as quickly as they do, Alcemene’s proposal to kill Eurystheus but hand over his corpse to the Athenians. Similarly, K.’s supplements after 962 and before 970 introduce the idea that Alcmene “sophistically” twists at 1022-5: Eurystheus’ body must be returned to the Argives (alive, the Servant means, whereas Alcmene will offer a corpse). But it remains troublesome that the chorus take her word for it, for as K. earlier pointed out forcefully, the one thing we know for sure about “the city’s will” at the start of the scene is that it forbids killing Eurystheus. The injunction against killing has multiple sources: the rulers of the land, according to the Servant (964); Hellenic law, according to Eurystheus (1010); the city, according to the chorus (1019; cf. 1020, 1024, 1026). Now one could say that the first of these carries the greatest authority, both because it is first and because the Servant (assuming, as K. does, that the line was rightly reassigned to him by Tyrwhitt) appropriately speaks for the “rulers.” In this case, the decision of the chorus at 1053-5 leaves them quite exposed. But would that be changed much by what happened in the lacuna as usually conceived? In favor of K., one could say that the identity of “the rulers” is vague to start with, and that Alcmene, Eurystheus, and the chorus all collaborate in shifting the emphasis to “the city,” for whom the chorus at 1053 may more easily presume to stand. A final note: at 973-4, K. possibly acquires a little misleading help for his interpretation from his Loeb translation. Servant: “There is no one to put him to death.” Alcmene: “I shall.” This makes it sound as though Alcmene has overcome a practical objection to her proposal, whereas the Servant’s line was of course an emphatic prohibition.
K.’s most noteworthy intervention in the text of Hippolytus is the assignment of 669-79 to the Nurse, in which he follows W. D. Smith and S. Østerud. In EA, he refers to their discussions and his own in The Heroic Muse 134 n. 80. Assignment of the short lyric passage to the Nurse makes it possible to have Phaedra exit at 600 and re-enter at 682. In a note on p. 182-3, K. details the not inconsiderable consequences for plot and character. The possibility is not mentioned by Barrett (Smith wrote in 1960), Diggle, or Stockert. I must admit that I am not convinced either.
Other passages of Hippolytus. 139. Of recent editors, only K. retains πένθει. He defended it at The Heroic Muse 129 n. 34. 241. A strong case for the reading preserved in an Oxyrhynchus papyrus. 271. Diggle may or may not have “thought it sufficient reply [to Barrett] to insert a comma after οὐκ οἶδ'” (K. in EA), but W. Stockert offers rather more at Prometheus 20 (1994) 217. 1012-15. These verses are suspected by all recent editors, but only K. deletes them (cf. 691 and 1049, the former deleted also by Barrett). K.’s reasons, which leave me not quite convinced, were given at GRBS 23 (1982) 45-7. 1105-6. In his commentary, Barrett proposed a solution to the notorious problem of the masculine participles in this stanza. K. and Stockert accept it, rightly in my opinion (cf. BMCR 95.9.25). 1459. K. overstates the case when he writes (in EA), “Neither [transmitted] reading is possible.” Barrett called ἀθηῶν suspect but printed it; W. Stockert, who emends, likewise only calls this reading weak ( art. cit. 233). K. believes the genitive plural lacks authority, since we can explain corruption to it but not from it. That argument is good but not obligatory. The question is then whether proposed conjectures improve sufficiently on the paradosis to warrant intervention. K. and Stockert, who both accept Sommerstein’s reasons for rejecting Fitton here and Fitton and Huxley at 1123, printed by Diggle, arrived independently at Ἐρεχθέως (though Stockert relegated it to his appendix of coniecturae minus probabiles). The problem, as both are aware, is that one might expect mention of both Trozen and Athens, as at 1094-7 and 1158-9. The former is cited by K. along with 973-5 as a point at which “Athens displaces Trozen as the place of chief importance for the action.” But this is unconvincing, since at 1094-7 Hippolytus goes on to mention Trozen, and at 973-5 Theseus says “neither Athens nor the borders of any land ruled by my spear.” So although Ἐρεχθέως may be an improvement on what Euripides wrote, it must be regarded as rewriting, since it does not meet the expectation that aroused suspicion in the first place.
Andromache 194-200. K.’s interventions, first proposed in 1977, are not mentioned by Diggle, and the transmitted text is defended by G. Goebel, CP 84 (1989) 32-5. In EA K. presents a lengthy rejoinder. 240-2. In EA, K. writes, “It can be proven with nearly mathematical certainty that 242 was never intended as the reply to 241,” and sure enough, his proof contains semi-formal logical propositions. In the end, he supplies two new trimeters. As always, K.’s arguments require careful consideration; I merely register here my suspicion that Euripides indeed wrote passages that do not withstand the level of scrutiny K. sometimes brings to bear when he is arguing for deletion. I also note that when he is defending, as at Hec. 441-3, K. is ready to admit with Weil that “such faults [in this case Hecuba interrupting her farewell to Polyxena to curse Helen] are not rare in Euripides” ( EA 62). In the present case, I would retain the paradosis and plead that the vagueness of πρῶτα and hence of 241 as a whole facilitates the discontinuity in strict logic detected by K. 833.“Diggle” in K.’s apparatus refers to Diggle’s first emendation of this passage, now in his Euripidea (Oxford 1994) 213; Diggle himself prints his second (cf. ibid. 268). 929. K. gives good reasons for following the ms. assignment to Orestes, though the one-line interruptions he cites are less perfect parallels than one would like, Tro. 889 coming after a new-fangled and elaborate invocation of Zeus, IT 673 after a line that invites a response. 1027-46. K. discussed these lines, in which his text differs from Diggle’s in 8 places, in The Andromache of Euripides 38-43. 1097. Diggle’s objection to “the magistrates filled themselves up into the council-chambers” is not overcome by the passages K. cites in EA, for both lack any suggestion that “filling” can be seen as a movement in a direction. 1279-82. K. accepts Sommerstein’s defence ( CQ 37 (1987) 51-64) of these lines, which Diggle follows Stevens in deleting.
K.’s text of Hecuba is most noteworthy for issues concerning suspected interpolation. To begin with, there are 33 verses deleted by both Diggle and K.; while all but three of these are retained in the conservative Teubner edition by S. Daitz (Leipzig 1973), C. Collard weighs in on the side of deletion in most cases (Warminster 1991). In places where K. differs from Diggle, he does not always favor deletion: he defends 92-7 (though continuing to voice suspicion), 211-15 (with some rewriting), 441-3 (despite earlier misgivings), and 756-7 (along with 758-9 in that order, as defended by D. J. Mastronarde, CP 83 (1988) 156-7). On the other hand, he deletes 14 lines that Diggle retains; several of these are his own suggestions, both old and new: 265-6, 279-81 (279 iam Hartung), 967, 1179 (in susp. voc. Wecklein). (K. no longer advocates deletion of 52-4.) I do not think many scholars will join K. in the deletions not shared with Diggle, but his diagnosis of faults may prompt useful discussion.
The volume was carefully proofread. I noted the following misprints. P. 48 ( Hcld. 402): read σωτηρίαν. P. 91 (apparatus to Hcld. 819-23): Wilamowitz deleted 819-22. P. 131 (apparatus to Hipp. 70-1): for Hec. 337 read Hec. 377. P. 138 ( Hipp. 147): read πελανῶν. P. 160 ( Hipp. 354): read ἀνέχομαι. P. 198 ( Hipp. 770): read δειρᾷ. P. 354 ( An. 909): read ἄνδρ’ ἔχειν. P. 362 ( An. 977): read ἔς. P. 380 ( An. 1188): for ἒ ἒ read ἒ ἔ. P. 406 ( Hec. 81): conjectural status of Doric forms (if intended) not noted.