BMCR 2003.06.49

La Tradizione Metrica della Tragedia Greca

, , , La Tradizione Metrica della Tragedia Greca. Napoli: Università degli Studi di Salerno, Dipartimento di Scienze dell'Antichità, 2002. 74. EUR 7.50.

This book includes three essays, one by each of the authors, each analyzing a particular ode from a Greek tragedy. There is also a brief introduction by Paola Volpe Cacciatore. The odes discussed are the first and second stasima of Iphigenia in Aulis and the second of Trachiniae. The object of each essay is to determine whether the colometry of the manuscripts and, where available, papyri, is correct. These are practical analyses, not works of metrical theory, though they raise several significant theoretical questions. None of the essays has much to say about literary interpretation of these songs; this limitation is presumably a result of the authors’ desire to keep the essays and the book to manageable size. In what follows, I will discuss each of the three essays, then consider the larger questions that they raise.

The first essay, by Carmela Concilio, considers IA 751-800. Two papyri, P. Leid. inv. 510 and P. Köln II 67, contain portions of this ode, not overlapping; the Leiden papyrus has musical notation, which Concilio does not discuss. Only two manuscripts include this play, one (L, Laur. gr. 32.2) which was emended by Demetrius Triclinius and one (P, Palat. gr. 287) which is later than L and not independent of it. Concilio agrees with Zuntz that P was copied from L after Triclinius’s first round of corrections.1 She prints a text of the ode using the manuscript colometry, which as she says “può essere accolta così com’ è, senza bisogno di emendamenti metrici” (p. 10), except that in lines 797-798 she prefers the papyrus version. She then gives a metrical analysis, with a type of apparatus showing where other scholars disagree about the division or the identification of the cola. Next she gives text and analysis for the papyrus texts. Finally, there is a detailed metrical commentary on the ode.

The manuscript colometry seems straightforward enough, predominantly choriambic dimeters and short aeolic cola. Although there are some mildly unusual features, Concilio generally shows that Euripides uses them elsewhere; in other words, they are plausible in one of his late plays.2 The two papyri, both from the epode, have a different colometry from the manuscripts (and a different text as well). In 795-800, P. Köln II 67 divides differently, and Concilio finds its sequence of choriambic dimeters (the second acephalic) more coherent than the manuscripts’ iambo-choriambic passage. In 784-793, P. Leid inv. 510 begins with the same colometry as the manuscripts, but diverges at 790; the text is also rather difficult here. Concilio says (p. 18) that, leaving aside the textual questions, either colometry is acceptable. This much is certainly true, but, first of all, the textual questions cannot be ignored. Second, and perhaps more important, there is a question of method. Is it enough to say that both versions of the passage are metrically sensible, without making a definite decision? Is it possible that both are correct, or is there only one right analysis for any given lyric? I will revisit this question at the end of this review.

The second essay is the longest, on IA 543-589, by Massimiliano D’Aiuto. P. Köln II 67 includes the beginnings of 15 lines of this ode, the last four of the antistrophe and the first 11 of the epode.3 As D’Aiuto points out (p. 24), the manuscript colometry is generally coherent and quite acceptable; it is aeolo-choriambic, with a long series of choriambic dimeters making up the bulk of the strophe and antistrophe, and more varied cola, with much resolution, in the epode. The papyrus presents a rather different colometry, still fundamentally aeolo-choriambic, possibly a bit simpler in the epode but much more complex in the antistrophe. Here lines 569 and 572 are enoplians while 570 and 571 are not really recognizable, and D’Aiuto does not label them. The ends of these lines are missing from the papyrus, and textual problems make it impossible to determine what was originally there; the remaining fragment of line 571 does show that the papyrus had a different text from the manuscripts. The manuscript text of the antistrophe is certainly corrupt (it is ungrammatical), but the papyrus does not allow us to fix either the textual problem or the metrical one. Many scholars, however, have re-divided the lines of the strophe to produce three choriambic dimeters and a pherecratean, with the first choriambic dimeter acephalic. The papyrus gives the same first line as the manuscripts, but moves one syllable down from the second line to the third, making line 570 only six syllables long. As we do not have the strophe in the papyrus, we cannot tell whether this is a mistake. D’Aiuto observes that the papyrus colometry, to the extent we can determine what it is, is inferior to that of the manuscripts, and the manuscript colometry is acceptable without the modern re-arrangements.

Similarly, D’Aiuto defends the manuscript colometry in the first three lines of the strophic pair, against most modern scholars. In the manuscripts, these lines appear as glyconic, hipponactean, reizianum; word ends match in the two stanzas. Every scholar whose colometry D’Aiuto cites, however, re-divides to produce glyconic, glyconic, pherecratean, with dovetailing, a “strofetta” (p. 33) of a type familiar throughout Greek poetry. D’Aiuto deplores the tendency to regularize colometry, replacing unusual line-divisions by more common ones. Although he does not deny that dovetailing can occur, he points out that it is only unambiguously present when word-ends in strophe and antistrophe do not match. When they do, they usually determine the colon boundaries, or at least the colometry is ambiguous (p. 34). He observes correctly that glyconics and hipponacteans often go together, and that the reizianum is as suitable a clausula as the pherecratean; therefore, he would retain this sequence here, respecting the word-ends, rather than re-dividing to produce the more familiar pattern.

In lines 2-3 of the epode (574-575 of the play), D’Aiuto once again prefers the manuscript colometry to the papyrus. The manuscripts offer choriambic dimeter and pherecratean, while the papyrus divides ἐτρά ‐ φης to produce a choriambic dimeter and an enoplian, with the choriamb in the first half of the dimeter (i.e. type A) and the first four syllables of the enoplian all long. While this is not an impossible analysis, it would be the only choriambic dimeter A in the entire ode, which contains many choriambic dimeters of type B (with the choriamb in the second half). On the other hand, the papyrus divides a word across lines, an unusual mistake (whereas joining a word that should be divided is easy), which may indicate that the papyrus reflects a deliberate metrical decision.

In summary, the colometry shown in this papyrus is not as good as that of the later manuscripts, at least in this ode. D’Aiuto does not follow up the possible implications of this observation. Was the papyrus simply carelessly copied, or could it reflect a different ancient analysis from the one in our manuscripts? In any case, this case study shows that one should not blindly assume that papyrus colometry is always right as against the later manuscripts, just as one would not make the analogous assumption about the text itself.

In the third essay, Sara Polizio studies Trach. 633-646, a passage for which we do not have papyrus evidence. The manuscripts, however, present slightly different colometries, which Polizio has collated herself. This passage is only the first strophic pair of the ode. Its main metrical problem is the complicated responsion between strophe and antistrophe, not atypical of Sophocles but sometimes difficult to analyze or interpret. Modern editors have joined together the short cola of the manuscripts into longer lines (in 1-2 and 6-7) and in some cases have re-divided. Their analyses are generally aeolic. Polizio interprets the manuscript colometry as predominantly iambic, which accounts for some of the inexact responsion. One difficulty remains, at 636 ~ 643, apparently a dodrans with the second long element of the choriamb resolved in the strophe. Polizio rejects Triclinius’s emendation, πάρ for παρά, since Sophocles does not use this form elsewhere in lyrics. Nor does he generally resolve choriambs, but she notes Ajax 605 ~ 616 as one possible example and Ant. 787 ~ 798 as another, more difficult as the text is uncertain. Polizio’s principle is good: she accepts a metrical anomaly in preference to a linguistic anomaly, other things being more or less equal.

Like her two colleagues, Polizio gives us an apparatus of textual variants and one of modern editors’ colometries, but she adds a third showing variations in manuscript colometries. As we might expect, the natural groupings of manuscripts based on colometry (at least in this passage) are more or less the same as those based on text: A (Paris gr. 2712), U (Ven. Marc. gr. 467), and Y (Vindob. phil. gr. 48) together, the Triclinian group together, and the Z group together.4 Polizio presents the textual and metrical variants separately rather than trying to give these two different sorts of information in the same space. Her notation is slightly more elegant than that of the other two essays.

Three major questions arise from these essays, and from metrical work in general. First, how do we know what the correct colometry is, or whether there is only one? And what does it mean for one colometry to be “better” or “more coherent” than another? Second, what is the best method in metrical study: analysis and possible emendation of the transmitted colometry, or independent establishment of period and colon boundaries based on a metrical theory? Finally, when we talk about metrical analyses, how can we present a collation of colometries?

The coherence of a colometry or an analysis is at least in part a measure of how convincing it is to readers. If one analysis considers a given stanza as a series of glyconics, pherecrateans, and hipponacteans, while a second calls it a mix of dactyls, dochmiacs, and unidentifiable cola, most readers would consider the first analysis more coherent. In other words, coherence is one of those phenomena that we recognize when we see it. Can this be made more precise? When D’Aiuto points out that in his stanza of IA the manuscripts offer a more coherent colometry than the papyrus, why do we all agree that he is right?

Many readers, however, would consider the modern dovetailed analysis of IA 543-545 ~ 558-560 “more coherent” than the manuscripts’ version. And in IA 790-793, Concilio says that both the manuscript colometry and that of the papyrus are acceptable. Is there necessarily a single correct colometry? If there is, we can make three assumptions: that this colometry is what the playwrights wrote and what their choruses sang, that Aristophanes of Byzantium knew or reconstructed it, and that it is reflected in the tradition that has come down to us. If not, however, we cannot be certain that we can recover the original colometry. This may seem like an abstract philosophical question, but what is at stake is our ability to understand Greek meter.

This, then, opens up the methodological question. It is by now generally accepted that the colometry of the lyrics in earlier medieval manuscripts of Greek drama reproduces the colometry established by Aristophanes of Byzantium, more or less as accurately as the words of those manuscripts reproduce the words of the text available to the Alexandrian scholars.5 By this I mean that although there are mistakes in transcription — putting a word or part of a word on the wrong line, getting two words in reverse order — they are usually just mistakes, not deliberate emendations on the part of the copyists. It is not until the early 14th century AD that scholars, most famously Demetrius Triclinius, begin adjusting the transmitted colometry, in particular correcting errors that affect strophic responsion. We thus have a very good idea of the Alexandrian colometry. It is also clear that the Alexandrian scholars had evidence we do not: musical scores, many more lyrics to compare, and perhaps even theoretical writings. Their colometry is therefore correct. But is it the only correct colometry?

If we believe that it is, then the task of the metrical scholar is to determine what that colometry was, by rectifying any mistakes in transmission (as textual critics do for the words of the text), and to determine how it works and what it means. If on the other hand we believe that the Alexandrian colometry is simply the work of scholars like ourselves, with more evidence but without necessarily any certain knowledge of the original,6 then we are at liberty to re-divide and re-analyze as necessary to produce a colometry that we consider correct. Naturally, the process is circular: we know what is correct based on what we have seen in other texts, but those other texts have been corrected by our colleagues and predecessors. We risk homogenizing and over-simplifying, as D’Aiuto implies has happened at IA 574-575. Moreover, the process never ends: every new theoretical advance calls all prior analyses into question.7 Some scholars find this exhilarating (as I do), others, impossible.

My third major question concerns notation. Studies like the present essays are relatively rare, and I suspect one reason for this is the difficulty of presenting the evidence. These three authors want to compare several different colometries and analyses for a given passage. One way to do this is to set out all the different versions side by side. For the major versions — that of the manuscripts and that of the papyrus — they do this. It would take far too much space to add seven or eight modern scholars’ texts and analyses beside them. More important, however, this hypothetical presentation would be very difficult to read. Just as we use the compact notation of the apparatus criticus for textual variants, we need a similarly concise notation for metrical variants. We need also to be able to discuss the identification of the cola. A colon like -uu-u- might be identified as aeolic or as a type of dochmiac. The identification is significant because it depends on (and also partially determines) the reader’s interpretation of the metrical context of the stanza.

In this book the authors seem to have experimented with presentation of the colometric apparatus, as each one uses a slightly different method. Concilio includes the line divisions and the identifications of cola in one group. Her notation is difficult to read; she uses a hyphen-sized dash to indicate ellipsis (“the words from ἀκούω through πλοκάμους” is written ἀκούω‐πλοκάμους) and a slash apparently to indicate a line division. Where an edition does not include a metrical analysis, she prints the editor’s name in parentheses. D’Aiuto generally does not indicate the various editors’ line divisions explicitly, as they are almost always implicit in the colon identifications. Where the line division is not obvious from the identification, he prints the last word of a line followed by a vertical bar. He does not distinguish between editors who have published separate analyses and those who have not. Polizio uses a convention like D’Aiuto’s, but in her apparatus of manuscript variants she does not print colon identifications as she does in the apparatus of modern editorial colometries.

All of these apparatus are dense and difficult to read, more difficult than the familiar textual apparatus (perhaps precisely because the latter is familiar). It is not obvious, however, what system would be better. A textual apparatus fundamentally presents one-dimensional information: the word at a certain point in the linear stream of text has the following variations. Metrical analysis is more nearly two dimensional, as what is important is not only the words one after another but their arrangement into lines and also the labelling or interpretation of that arrangement. The standard handbooks of textual criticism (e.g. West) explain how to construct a textual apparatus, and insist that the textual critic working on poetry must be a competent metrist, but say nothing about recording or presenting metrical variations. Metrical works (e.g. Parker, The Songs of Aristophanes) regularly discuss colometric variations, but rarely attempt to present them in compact form. A good notation for metrical collation would make work of this kind easier to do and to present.

The book is nicely produced and has no obvious typographic errors. The bibliography at the end includes only works that are frequently cited by author and date; even some works that are cited by two or more of the essays do not appear here. This is unfortunate, as readers must comb the footnotes to appreciate the full range of scholarship cited.

These essays are useful contributions to the study of Sophocles and especially Euripides. The practical application of metrical scholarship helps us to confirm or refine our theories. It is good to have books like this one.


1. Gunther Zuntz, An Inquiry into the Transmission of the Plays of Euripides, Cambridge: 1965.

2. Concilio does not discuss whether Euripides himself actually wrote this ode. The question is not really important to her analysis; even if he did not write it, the actual author was presumably contemporary or not much later, so Euripidean metrical practice is a sensible basis for evaluation. Diggle in the OCT marks this ode and the speeches immediately preceding and following vix Euripidei.

3. This is the same papyrus that includes parts of Concilio’s ode; the two lyrics appear on the papyrus in the opposite order from the play.

4. Lloyd-Jones and Wilson characterize the Z family as “a group of manuscripts that has still not been very fully explored” (OCT preface, p. x), and Polizio’s collation of them here is welcome.

5. See for example Barrett’s commentary on Hippolytus, p. 84 ff, for a mainstream statement of the case, relatively unaffected by passion or partisanship.

6. Of course, Aristophanes was a native speaker, though of a somewhat later stage of the language than the playwrights, but this does not in general confer metrical competence in the same way as it confers grammatical or lexical competence. We can see this in the continuing debate among poets and scholars about the principles of English meter. One particularly vivid presentation of this debate is David Baker, ed., Meter in English (Arkansas: 1996).

7. New theories are indeed proposed from time to time. One of the most original and significant such works is Kiichiro Itsumi’s series of articles from 1982-1992 on the meters of the aeolic family, especially “Enoplian in Tragedy,” BICS 38(1991-1993), 243-261, a beautiful and creative paper, which the present authors make good use of.