In the past few years, when I have had the need to look up a passage in an obscure author or to search out particular words or names, I simply put the PHI Latin disk in my computer and start hunting. The fact that the texts are not of equal quality and that there is no apparatus is of course a drawback, and requires a certain amount of skepticism and circumspection (the use of TLL and dictionaries); but for no part of Latin literature—or Greek, for that matter—is modern technology so far less successful than in presenting the text of fragments. Look, for instance, at the fragments of Ennius’ Annales on the PHI disk. Every word of Ennius that survives verbatim is there, but there is no information at all about how or why we have it; no apparatus; and above all no testimonia or paraphrase. From the surviving words of Ennius, we are remarkably poorly informed about the dream, just as from the surviving words of the great orators Marcus Antonius and Lucius Licinius Crassus we learn virtually nothing about them—despite the fact that the testimonia and paraphrases fill pages and pages of Malcovati’s edition. The naked text is devoid of sense or utility, except for the most uninteresting sort of word-search.
In his comments on Marx’s edition of Lucilius, A.E. Housman made the point as well as it can be made: “What more than anything else enables the critic and commentator of an ancient author to correct mistakes and to elucidate obscurities is their context; and a fragment has no context.”1 Housman, of course, was intent on emending and not on interpreting; and hence he diminishes the value of the context that in fact fragments do have—the context in the source that supplies the quotation. Not as good, of course, as knowing the train of thought from which a corrupt phrase springs, but a great deal better than the bare words of the PHI disk.
Fragmentary texts are a challenge to the editor or interpreter, and unfortunately not all those who find themselves called to the challenge are up to the task. There have been some superb editions of fragmentary texts in the last generation: from my own interests, I think immediately of Lloyd-Jones and Parsons’ Supplementum Hellenisticum —and anyone who wants to see real progress need only compare the exiguous notes of Powell’s Collectanea Alexandrina (good though they are) with the detailed commentaries of the Supplementum —or of E. J. Courtney’s Fragmentary Latin Poets, as far above the awful Teubner texts of Morel or Buechner as it is possible to be. That is not to say, of course, that all modern editions of fragments represent an improvement or a higher standard, but I dwell on successes rather than failures.
The present volume is not an edition of fragments but a set of articles about them. In Frustula poetarum, Alessandro Perutelli has assembled nine papers (eight of them previously published) on poets whose main shared characteristic is that they do not survive whole. They range from Caecilius Statius and Lutatius Catulus in the second century BCE on one end to Lucan’s Orpheus of the mid-first century CE on the other; they cover drama, lyric, epigram, satire, epic, and didactic. Some papers deal with the entire surviving oeuvre of a minor poet (Lutatius Catulus); others deal with the fragments of a single play or poem. What all of them show, however, is the ability to tease information out of one context (the quoting source) and to use it to enhance another (the work, life, or literary context of the poet quoted).
The kind of unity that P.’s volume possesses is not the kind that will attract most Latinists: we look for discussions of texts that interest us rather than having an immediate interest in the problems of restoring or interpreting fragments. Hence, few will read P.’s book cover to cover, and that is perhaps as it should be. The virtues of P.’s work are consistent throughout, and anyone who has tended to take on trust the text of poetic fragments as presented by Morel, Ribbeck, Marx or Vahlen should read at least one of P.’s articles. Several of them are truly exemplary. His discussion of the Aethrio of Caecilius Statius—the only previously unpublished paper in the volume—destroys forever the idea that it has anything to do with Plautus’ Amphitruo, and he does so by careful and learned analysis of the context of quotation in Nonius Marcellus, of syntax, and of sense. His lengthy discussion of Lucan’s extremely fragmentary Orpheus (the latest text discussed) is in part an explication of the ways in which Statius used the Orpheus in Silvae 2.7 (the Genethliacon Lucani), in turn using Statius to recover the structure of the lost poem. Some of the other papers, while learned and sensible, offer less that is new: I did not learn a great deal about Varius from P.’s discussion of fr. 4, and the discussion of Varro’s Bimarcus seemed to dwell rather too much on the drawbacks of the various editions and too little on the text itself. Any extended examination of fragments will tend to become a collection of notes on the various details (of language, meter, source) that aid reconstruction, and P.’s work is no exception. That does not, however, mean that one can not admire the skill and learning that go into all the details, and that remains true even when, on occasion, I found his results less than completely convincing.
Rather than dwell on the details, however, I will single out two arguments that seem to me of particular importance. In the first place, his two adjacent papers on Lutatius Catulus and on Laevius together make an extremely interesting argument about the relationship of archaic Roman poetry to its Greek and Roman pasts, and about the relationship of first century Roman poetry to its own predecessors. P. rightly points out the important connections between Catulus’ epigrams (and those of Porcius Licinus and Valerius Aedituus) and Roman comedy, particularly Terence; he shows how comic language and scene somehow become the vehicle for the early epigrammatists to transmute Callimachean epigrams into a Roman context. Laevius too has dramatic tendencies, and he too stands between the truly archaic version of Hellenistic poetry and the later and less expressionistic (to use P.’s own term) world of the neoterics. P.’s argument concentrates on the details of particular lines, but he does not lose sight of a very interesting reading of Roman literary development between about 150 BCE and Catullus—a period about which we know far too little.
The other argument is textual and concerns the ways in which fragments are quoted by (in particular) Nonius Marcellus, our single most important source for lost works of the pre-Augustan period. P. observes, quite rightly, that the quotations in Nonius sometimes are attached to a lemma that must, on any reasonable reading, be corrupt: in other words, that the one word in the fragment that would seem to be guaranteed by the context is itself wrong. Other editors, of course, have encountered this; but it is frequently, and is frequently explained as, corruption that has occurred in the transmission of Nonius himself. But P. makes the argument that in some cases the corruption is earlier than Nonius, and that, as a result, the way a fragment should be printed as part of the text of Nonius is not the same as the way it should be printed as an independent text. It seems unlikely that Nonius will be re-edited soon—it is one of W. M. Lindsay’s finest editions—and hence this may be a theoretical more than a practical problem. It is nonetheless a problem for textual theorists: the same text may be right in one context, and wrong in another. Textual critics of modern literature, particularly those like Jerome McGann who emphasize the social context of the production of texts, have long known this; but most classical textual critics remain far too wedded to strong conceptions of textual truth and too unwilling to recognize the variability of texts in antiquity. We are too poorly supplied with texts to be able to see this very often, but it has clear relevance to a range of intertextual problems: was Virgil’s text of Catullus/Lucretius/Homer the same as ours, and how should we deal with that issue?
Perutelli does not explore the theoretical problems of Nonius’ text; he is too averse to speculations as unfounded as those in the previous paragraph to venture so far afield. At the same time, however, his ability to recognize and to define the textual and literary problems in the fragments he discusses is everywhere apparent. If you find the scraps of PHI useless and frustrating, Frustula Poetarum is a good place to start learning how to interpret them.
1. “Luciliana” in Collected Papers (Cambridge 1972) 2:662.