Ancient Evidence for the Colometry of Aeschylus' "Septem" Fleming, Thomas J Greek, Roman and Byzantine Studies; Summer 1975; 16, 2; ProQuest pg. 141 Ancient Evidence for the Colometry of Aeschylus' Septem Thomas J. Fleming I N THIS ESSAY I hope to demonstrate three points: (1) that the colometry of Byzantine manuscripts of Aeschylus is supported and corrected by an ancient papyrus fragment of Septem; (2) that this colometry is defensible on metrical grounds; (3) that adoption of this colometry obliges us to reexamine two textual problems. The earliest witness to the text of Septem 563-67", 626-30 is to be found in a papyrus of the second century, P.Oxy. XXII 2333 (1956). One of the two scraps which comprise this papyrus contains the ends of lines 626, 627, 629 and 630. In a photograph (PLATE 1) the visible letters are displayed as follows: OYCI\ITAC I\ICEY HI EKT09EN NW When the papyrus was first published, it was restored to read:1 626 [KAvovTE:e Oeo£ S£Ka£ ]9Ve A£'TaC [1JJLE'TEpae 'TEAE£O we '7TO ]A£e EV[ 'TUxJn ~ [Sopmova KaK EK'Tpe'7TOV'T£C (ee) yac] [£7T£JLoAoVe f3aAwv '7TVpywv S] EK'TOOev 630 [Z£vc ec/>e K(tVot K£pav]yw The editor (c. H. Roberts) comments on his alteration of the tradi­ tional word order in 629: "The pOSition of the surviving letters in relation to the preceding and following lines makes it certain that 27-29 letters have been lost before eK'TOOeV, while the margin to the right precludes us from reading f3aAwv after it as do all other MSS" (italics mine). 1 The editor did not include the missing 628 in his reconstruction. For the sake of clarity I have printed 628 as it appears in Murray's first edition (OCT, Oxford 1937), against which the papyrus was collated originally. 141 142 EVIDENCE FOR THE COLOMETRY OF AESCHYLUS' SEPTEM It is not entirely clear by what process of calculation the figure H27-29 letters" was reached. So far as one can tell from a photograph, the first letter of EKT09EN would appear to have been, roughly, under the 24th letter of 627.2 Since, in the part of the papyrus con­ taining 644-55, the lines do not seem to vary by more than two (at most three) in number of letters when they have been aligned with the 24th letter of 646, we might, therefore, suppose that about 22-24 letters have been lost before E in 629. This accords far better with the 21 letters of the editor's restoration than his own figure of 27-29 letters. Is it necessary, however, to accept this altered word order? An altered colometry might provide a less drastic solution to the prob­ lem. If fJa'Awv began 630 (as the papyrus suggests), this would leave us only 16 letters in 629. Addition of yae from the end of 628 would produce a line of only 19 letters, but if we should take -TEe from the end of ~KTPE1TovTEe (the smallest possible addition), the sum would be raised to 22 letters, a total within the range of possibility which the disposition of the letters suggests. This colometry of 628-29: .. . l.KTP­ E1Tovl- ... €KTOOEV I would have an additional advantage. The last let­ ter, (0, of 630 appears directly under the T in 629. In the editor's reconstruction, T is the 24th letter of 629 and (0 the 19th of 630, a difference of five letters. In the colometry which I propose, T is the 25th letter of 629 and (0 the 24th letter of 630, a difference of only one letter. One more argument might be adduced in favor of this new colom­ etry. Although no trace of 628 has been preserved on the papyrus, there is just enough room beneath EY in 627 and enough at the right of 628 to have left some trace of the final yae. No trace, however, of 628 would have survived in a colometry where 628 was divided ••• ~KTpE1Tovl-· The most important piece of evidence supporting the editor's reconstruction is implied in his phrase, Has do all other MSS." Since there is a good deal of support for the view that the colometry of Byzantine MSS of the tragedians represents the tradition of ancient I It is difficult to determine the exact relation of the letters in the two lines. since that part of the scrap which contains 629-30 appears slightly askew in the photograph. with the result that the first letter of EKT09EN appears beneath the 25th (missing) letter of 627; that is. 629-30 are misplaced, by the space of one letter. to the right. This adjustment. however. is not vital to the argument. THOMAS J. FLEMING 143 editions,S it would be dangerous to reconstruct the papyrus in such a way as to do violence to the traditional division of the lines. Collation for colometry of a selection of older, oV El 8ol"1c Tl).,OC, which, however, can as easily be interpreted as the protasis of a conditional sentence which, after interruption, is completed by 262. 18 A. M. Dale, The Lyric Metres of Greek Drama (Cambridge 1968) 73. FLEMING PLATE 1 P.Oxy. 2333 fro A (SECOND CENTURY): AESCHYLUS, Septem 621-38 (ReproduCEd by permission of the Egypt Exploration Society) THOMAS J. FLEMING 147 problem lies in the ending v- - - -v-, for which there is only one Aeschylean parallel, at Agamemnon 247""" 258, where the intial anceps of the iamb is short (v- - v-v-) and the sequence of four longa avoided.19 If we accept the colometry of the paradosis as representing ancient editions, only KAVWV (which we may interpret with Verrall) or KAvovc' (which we might construe, albeit awkwardly, as a nominati­ vus pendens, "as I hear"), and KAvovc~ (whose metrical difficulty is almost justified by the improvement in sense) appear to deserve an editor's attention. The same metrical arguments apply to Hermann's supplement EC yac in the responding 628. Support has been drawn from the MSS: Ele is superscript in M lover bnfLoAove, and 7Tpoe is written after yae in nine of Dawe's MSS.20 In his edition Hermann recognized 7TpOC as an intrusion, and it is worthy of note that the phrase 7Tpoe Tove E7T£f'oAovC occurs among the scholia. Although Verrall and (it would seem) Wilamowitz21 believed Ele to be similarly intruded, Hermann's supplement has won approval because the loss of €C can be explained as haplography and because of the frequency of the combination EKTP€7TEtV Elc, which occurs in Aeschylus (Ag. 1464), Euripides, Thu­ cydides, Antiphon, Aristotle and Herodotus, who uses it twice to refer to the diversion of water. The same verb, however, is found with a variety of constructions: with 7TPOC in Sophocles (Aj. 53), with a direct object in Sophocles (El. 350) and (middle voice) in Demosthenes (19.225), and with various prepositions. Although we are able to understand the reason for the intrusion of EtC by scribes, we cannot assume that this unmetrical scribal conjecture has hit the mark. If we reject Hermann's supplement on the grounds of colome try and meter, is the Greek possible to interpret? The answer seems to lie in the construction found in Sophocles and Demosthenes, that is, to take E7T£f'oAovc as direct object. We might translate "You gods, hearing our just prayers, bring to fulfillment-that the city be vic­ torious-spear-Iabored woes, by turning aside the invaders of our land."22 19 Such a sequence seems almost unparalleled in the lyric iambs of Aeschylus. In the few (dubious) examples, there is metron-diaeresis. so Dawe, op.cit. (supra n.6) 275. 21 op.cit. (supra n.16) 112. 22 This interpretation was first suggested to me by the late Professor Douglas Young. 148 EVIDENCE FOR THE COLOMETRY OF AESCHYLUS' SEPTEM Two difficulties arise from this interpretation. In the first place, it may be objected that the intrusion of a purpose clause between verb and object is too violent a disruption of the sense, but see Septem 337 and Euripides' Cyclops 121, which is cited by Schwyzer,23 and an even closer Euripidean parallel at Supplices 1145. In Septem 628 the paren­ thetical purpose clause seems to take the form of an interjected hope, like similar interjected clauses in Aeschylus as, for example, Septem 5. Such a parenthetical purpose clause would have easily been mis­ understood by scribes who were able to make sense of this trouble­ some passage only by interpolating Etc or 71'poc. In the second place, KAVW is commonly described as requiring the genitive case when it means 'listen to', as it seems to do in 626.24 Aeschylus, however, provides at least three examples of this verb followed by an accusative case where the gods are asked to listen to a prayer (Sept. 171, Cho. 125-26) or a plea (Ag. 814). It might be added that the papyrus gives evidence against any attempt to reinterpret A'TeXC as a Doric genitive, since it appears to support S'Kalovc M2 I and most MSS with the exception of Q 0 LI Npc, which have S'Kalac. SlKawe is treated as an adjective of two terminations twice in Euripides (Heracl. 901, IT 1202). KUhner-Blass I 537 list other adjectives in -awe which suffer the same variation and cite aTrJc AaOpalov from Aeschylus (Ag. 1230), as well as nine examples of adjectives of various formations that are otherwise found with three terminations but which Aeschylus employs with only two. Page appears to be convinced by the evidence and prints the masculine form. In sum, there appears to be nothing against interpreting KAt5w with an accusative in the sense 'to listen to'. Although it is still quite possible that the solution to some of these problems eludes us, it is unlikely that any remedies which depend on Triclinian colometry can be correct. The starting point of any dis­ cussion of the text of this passage must be the text and colometry of our earliest witness.25 MIAMI UNIVERSITY january, 1975 23 Ed. Schwyzer, "Die Parenthese," AbhBerlin 1939 no.6, p.17. 24 See for example LS] s.". u and KUhner-Gerth I p.358. 25 I wish to express my gratitude to Professors P. A. Stadter and E. C. Kopff for their suggestions and criticisms, to Dr R. A. Coles and Professor E. G. Turner for providing a photograph of the papyrus, and to the Egypt Exploration Society for their permission to reproduce the photograph.