Bryn Mawr Classical Review 98.1.23

Glenn W. Most (ed.), Collecting Fragments-Fragmente sammeln. Aporemata: Kritische Studien zur Philologiegeschichte 1. Göttingen: Vandenhoeck and Ruprecht, 1997. Pp. x, 338. DM 60,-/ÖS 438,-/SFr 54,50. ISBN 3-525-25900-X.

Contributors: A. C. Dionisotti, Rudolf G. Wagner, Ewen Bowie, Richard Kannicht, Peter Bing, Giovanni Benedetto, A. S. Hollis, Anthony Grafton, Guido Schepens, G. W. Bowersock, Peter Schaefer, S. C. Humphreys, I. G. Kidd, A. Laks, Franco Montanari, Ann Ellis Hanson, Hans Ulrich Gumbrecht.

Reviewed by John Gibert, University of Colorado,

Word Count: 3,709.

The 17 essays in this book originated partly in a symposium held in Heidelberg in 1995. The publication is the first in a new series called Aporemata, which (to quote the back cover) "will investigate historical and methodological aspects of various philological activities." The second volume, Editing Texts-Texte edieren (ed. G. Most), was also to appear last year. The third, Fragmentsammlungen philosophischer Texte der Antike (ed. W. Burkert), and fourth, Commenting on Texts-Texte kommentieren, are announced for 1998.

The topic is an appealing one. Collections of fragments attract attention for the intrinsic interest of their contents, for the historical and methodological aspects mentioned in the blurb, and even for a theoretical discussion that has been picking up steam lately. (Most's preface notes this last context, which also informs Hans Ulrich Gumbrecht's final reflections; another, comparative dimension of fragment collecting is addressed only by the inclusion of R. Wagner's essay on exegeses of Chinese classics.) Many share A. S. Hollis's "addiction" to the peculiar intellectual challenges of identifying, restoring, ordering, and interpreting fragmentary texts, and many are drawn to the collections themselves, often among the most impressively learned and sumptuously produced works in our field.

The volume achieves good coverage -- on the Greek side. Latin authors are included in the survey chapter of A. C. Dionisotti, but only very occasionally thereafter. This reviewer would have welcomed more on dramatic fragments to accompany the contribution by R. Kannicht, but it may be that these have received more than their share of attention in the past. As noted, philosophical fragments will be covered in Aporemata 3, but the present volume has two good essays in this area. Taken together, Dionisotti and Grafton make a substantial contribution to the history of classical scholarship. The five essays on historical fragments form a strong and coherent section, in which special attention is given to Jacoby's Fragmente der griechischen Historiker. The contributors were up to their task. I enjoyed reading this volume and learned much from it. In what follows, I offer brief summaries of all the essays and a few critical remarks.

The first essay, "On fragments in classical scholarship", by A. C. Dionisotti, is wide-ranging, learned, and intelligent. After some definitions, the author uses the example of Gallus to show the limitations of "the bio-bibliographical tradition". This is an important first gesture, for that same tradition became the indispensable foundation of fragment collecting. Dionisotti writes that "the mainstream of reading, excerpting, and bio-bibliography. . . originated with the great expansion of learning in the twelfth century, but its development. . . belongs specially to the thirteenth" (9). In Petrarch's time, however, and for some time afterwards, fragment collecting as we know it was little favored, for "the thing to do about a missing work was look for it, not collect its fragments" (21). The survey continues with remarks on history, law, and epigraphy; early in the age of printed books, it is argued, the impetus for fragment collecting proper came from these fields. Hence the favor shown to Cicero and Sallust, for example. Still, work had begun on other authors and genres: Politian published work on Eupolis' Demoi, and left behind unpublished work on Sappho. Dionisotti concludes with the forgeries of Annius of Viterbo, which are interestingly discussed also by Anthony Grafton later in the volume.

Dionisotti has turned up memorable stories and statements of early fragment collectors. These words of Thomas of Ireland (from his Manipulus florum (1306), quoted on p. 10) exemplify a defensive attitude often encountered: "Propter has autem modicas spicas agrum fertilem originalium non despicias: improvidus enim est qui neglecto igne se per scintillas nititur calefacere, et qui contempto fonte per roris guttas sitim conatur extinguere." Dionisotti also traces the usage of Latin fragmentum (12-18), which was long burdened by its materiality and resulting low status. He wonders why, in view of this, Jacobus Cappellus and Guy de Fontenay gave the name Fragmenta to their collections (both Paris, early 16th century). The answer, he suggests, has to do with figurative readings of Jesus' words to the disciples after the miracle of loaves and fishes: colligite fragmenta, ne pereant. Such figurative readings had begun with Augustine and continued widely. Dionisotti notes, however, that striking extensions of fragmentum are found as early as Petrarch, whose lead Poggio followed with "powerful imagery [in which] the fragmentary text and its venerable author are one, maimed, injured, wounded, or mutilated" (20).

"Twice removed from the truth: fragment collection in 18th and 19th century China," by Rudolf Wagner, is the volume's only contribution to the comparative dimension of the project outlined in the editor's preface. It concerns a tradition of exegesis of sacred texts and touches on issues of orthodoxy, tolerance, and political freedom. In "The Theognidea: a step towards a collection of fragments?," Ewen Bowie studies how archaic poets and poets of Old Comedy cited earlier texts (53-7) and then extends the survey into later periods for elegy (58-61). In applying his results to the Theognidea, he ponders some definitions of "fragment" and develops a distinction between compilers and transcribers. The former could, at least in many cases, have chosen to quote more of the original and were therefore not working with fragments. The transcribers, in contrast, were dependent on selections made by their predecessors and are thus said, a little paradoxically, to be "assembling a collection of fragments" (64).

Richard Kannicht's essay, "TrGF V Euripides", focuses on anthologies and gnomologies. The figures he presents on the inclusion of tragic passages in such works show, not surprisingly, that Euripides predominates by a wide margin, just as he is already quoted by Aristotle much more often than any other tragedian. Figures are also given for quotation by Chrysippus, Cicero, Plutarch, and Stobaeus. Kannicht then proceeds to modern anthologists and offers illuminating remarks on Johannes Meursius, Hugo Grotius, and especially Dirk Canter, whose work is becoming better known through the researches of J. A. Gruys. Kannicht demonstrates that Grotius did not help, indeed may have hindered, publication of Canter's work, in the hope of ensuring that the humanistic-edifying mode of anthologizing would not lose ground to the antiquarian-learned mode. For discussion of these divergent approaches, Kannicht refers to the influential article of R. Kassel on "Fragmente und ihre Sammler", in Fragmenta Dramatica, ed. H. Hofmann (Göttingen 1991), 243-53 (reprinted in his Kleine Schriften). At the Heidelberg symposium, Kannicht concluded with an account of his work in progress on TrGF V, particularly in regard to the reconstruction of whole plays. Part of this work has appeared elsewhere.

Next come three essays on Callimachus. In the second, "Il trattamento dei frammenti nell' edizione callimachea del 1761 attraverso la corrispondenza inedita di J. A. Ernesti con D. Ruhnkenius", Giovanni Benedetto discusses the part taken by Dutch scholars in the genesis of Ernesti's Callimachus. Interest in this collaboration lies not in the quality of the product but in the light cast on the epochal shift, towards the end of the 18th century, from the methods of the Dutch school (Hemsterhuys, Valckenaer, Ruhnken) to the nascent German ideal of a comprehensive Altertumswissenschaft. A few decades after its appearance, Ernesti's edition figured in some polemics between Dutch and German scholars, the issue being whether Ernesti had limited Dutch influence on the edition in order to avoid cutting a poor figure himself. The transition from the Dutch to the German paradigm is traced in the standard histories of classical scholarship. The Latin scholarly correspondence cited in extenso by Benedetto was published in the first half of the 19th century, and much of it was apparently gone over in a 1953 Dutch dissertation. I am not sure Benedetto has added much by his review of this dossier. Ernesti's shortcomings as an editor and scholar of classical authors are recognized. As for the collaboration with the Dutch scholars, it is certainly acknowledged openly and warmly by Ernesti himself. If some mystery remains as to why he did not avail himself of it still more (especially in the case of the slighted Valckenaer), Benedetto has done little to clear it up. We are left, then, with the speculation already regarded as most likely on general grounds, that Ernesti was simply not up to the task of producing an edition on Dutch lines, or not interested in it, or, as seems most likely, both.

A. S. Hollis, in "A Fragmentary addiction", provides a miscellany of examples and experiences derived from his own experience as an editor of Callimachus' Hecale and, recently, fragmentary Latin poets. He suggests plausible contexts for unplaced fragments, reflects on the separation of testimonia and fragmenta, and accepts a challenge issued in an unpublished lecture many years ago to find 9 continuous lines of an extant author containing as many hints as to his identity as are contained in the famous Gallus fragments (116-17; his choice for this experiment is Catullus 92.1-95.1). When discussing the ancient grammarians' practice of citing errors and anomalies, he raises this entertaining question: "When Varro Atacinus scanned Phaethon as a spondaic disyllable instead of a trisyllabic anapaest, could he have been unaware that he was securing immortality for at least one of his lines?" (p. 121).

The stimulating essay by Peter Bing, "Reconstructing Berenike's Lock", differs from Hollis's in having a sustained argument and from Benedetto's in having to do with issues still relevant to the interpretation of an important Callimachean passage. After relating two examples of convincing intervention by Politian in the textual criticism of Catullus 66 (Chalybon restored at 48, Oarion defended at 94, both times on the assumption that Catullus transferred a Greek form directly into Latin), Bing describes his project: (1) to survey reconstructions of Callimachus' Lock of Berenice based on the assumption that Catullus was exceptionally faithful, from Politian to Pfeiffer, and (2) to show that "not one has proved correct in the light of subsequent discoveries" (81). The chief target turns out to be Pfeiffer himself. One of Bing's points concerns the layout of fr. 110, with its implication that Callimachus' plokamos and Catullus 66 corresponded line for line. Although Pfeiffer had made no secret of his belief that this was so in an earlier article, Bing judges such an arrangement inappropriate "in an edition which openly acknowledges none of its presuppositions" (87 n. 37; it should be noted that the article is often cited in the edition.) Bing may be right that this is unusual for an edition, but is it true that few readers are actively aware of it or its bearing on interpretation? Bing also tells of Pfeiffer's 1923 use of P.Oxy. 1793, recanted in the 1949 edition. According to Bing, Pfeiffer had gone "badly astray" and had a "chastening experience." This seems a bit dramatic. It appears that Pfeiffer merely corrected his calculation about the number of verses in each column and went on with his work more cautiously; he still believed that Catullus had translated Callimachus line by line. Bing closes (92-4) with the notorious puzzle of the 10 Catullan lines (79-88) on the wedding ritual with no surviving counterpart in Callimachus. He shares the widespread view that Pfeiffer's reconstruction is not only unprovable but introduces new difficulties. While he notes that Hollis has found yet another way to maintain that the lines are Callimachean, Bing prefers to leave the matter open. Overall, his study of this unusual situation, in which scholars' eagerness to solve a fragmentary puzzle is fired by the belief that there exists a sure guide to reconstruction, is both welcome and provocative.

Next come the five essays on historical fragments. Anthony Grafton, in "Fragmenta Historicorum Graecorum: fragments of some lost enterprises", deals with the extensive and usually unacknowledged efforts of Renaissance and late humanist collectors of historical fragments, with particular attention to the work of Joseph Scaliger on ancient chronology. The result is neatly summed up on p. 143: "The supposed pioneers of fragment-hunting around 1800 [especially Fr. Creuzer, singled out by C. Mueller for praise in the preface to FHG] accepted many of the same unconsidered assumptions that Scaliger and his contemporaries had made two hundred years before: that only the fragment, not its wider context, mattered; that the compilatory writers who preserved fragments were idiots whose own opinions should not be studied but ignored; that the chief questions of interest in the study of fragments were problems of textual and historical authenticity."1

The essay by Guido Schepens, "Jacoby's FGrHist: problems, methods, prospects", is an extended meditation on the choices faced by an editor of fragments, with an emphasis on Jacoby's broad definition of history and genos-oriented approach. Schepens describes Jacoby's design for Part IV and gives a sample of notes and lists from cards. They are full of unsettled questions about the boundaries of genres and their appropriateness to a collection of historical fragments. Schepens also discusses the roots of Jacoby's criterion of "Historikerqualitaet" in the departmentalization of the 19th century German university. He then offers a progress report on the continuation project and a preview of its features, of which I mention two: (1) Fragments that "can on more or less plausible grounds (to be explained in the commentary) be attributed to a given work or to a certain book [will be presented] at their putative place [rather than in a separate section after all fragments that are assigned title and/or book number by the citing source, as in Jacoby], but with an indentation at the left margin in order to draw attention to the hypothetical nature of their allocation (cf. Wehrli, Die Schule des Aristoteles)" (165 n. 61); and (2) English translation.

In "Jacoby's fragments and two Greek historians of pre-Islamic Arabia", G. W. Bowersock tests the usefulness of Jacoby's collection (and thus of his method) in the case of two writers of uncertain date, Glaucus and Uranius (FGrHist 674 and 675). Almost all of their "fragments" come from Stephanus of Byzantium, whose work must be studied carefully if our knowledge of the quoted authors is to be secure. Bowersock concludes (185), "The methodological problems of Jacoby's Fragmente illustrated here may arguably not warrant a wholesale condemnation of his enterprise. His collection can be viewed as a kind of ladder borrowed from Wittgenstein's philosophy: one uses it to climb up and then throws it away. Or again it may be seen to resemble navigational software for the Internet. We can perhaps locate useful material more quickly with it than without it. But one point is absolutely secure, and that is the necessity to leave Jacoby behind and to examine the original sources for historical fragments (however defined) before bringing any scholarly research on them to a conclusion."

Peter Schaefer's essay, "Die Manetho-Fragmente bei Josephus und die Anfänge des antiken 'Antisemitismus'", studies past applications of Quellenkritik to a passage of Josephus' Contra Apionem which preserves excerpts from the Aegyptiaca of Manetho of Heliopolis (3rd century BCE, FGrHist 609). At issue is whether Jacoby and others are right to dissociate certain phrases and paragraphs of antisemitic tendency from Manetho in what Schaefer takes to be an effort to establish a later date for the beginnings of ancient antisemitism. After demonstrating the difficulties in which this effort has involved the historians and finding parallels in Hecataeus of Abdera (FGrHist 262) for the particular features being denied to Manetho, Schaefer concludes that Manetho could indeed be responsible for the tendency of the passage in question. This kind of problem arises constantly in the study of prose fragments, and Schaefer's clear, subtle, and reasonable study exactly suits the aims of this volume. For full discussion of the wider issues involved in this example, including the thesis that allegations of misoxenia and misanthropia are specifically Greek contributions to ancient antisemitism, see Schaefer's new book, Judeophobia: Attitudes toward the Jews in the Ancient World (Harvard 1997).

In S. C. Humphries, "Fragments, fetishes, and philosophies: towards a history of Greek historiography after Thucydides", the emphasis is on some rather broad historiographical issues rather than on fragments in the ordinary sense. Humphries' thesis is that "historians disguise from themselves the problematic character of their relationship with the past by according to certain relics the privileged status of 'documents' or 'evidence'" (209). That is, they fetishize those certain relics, and so are behind the times, for "scholars in various branches of [Classics] are currently struggling to escape ... from [a master-narrative] derived from Enlightenment rationalism and nineteenth-century evolutionism" (207-8). Specifically, she is concerned with the blend of positivism and "romantic conceptions of authorship" that she feels shapes the practice of ancient historians and is represented, if in very different ways, in the works of both Jacoby and Momigliano. Little influence has been exerted by an alternative model exemplified by Mazzarino's Il pensiero storico classico (1965-6), the merits of which she describes as follows (213): "(1) the constant comparisons between historical works and texts in other genres, especially philosophy; (2) close attention to ancient historians' readings of their predecessors and contemporaries; and (3) a serious attempt to understand what Bakhtin would call the 'chronotopes' of ancient historiography, i.e. the way time and space are conceptualized". Humphries sketches the effects such a model would have on the interpretation of neglected portions (essentially narratives concerning relics) of various authors (e.g. Josephus, Livy).

In "What is a Posidonian fragment?", I. G. Kidd illustrates problems in drawing the boundaries of fragments with examples from his own editorial work. This preview of the results of his current work on that prodigious quoter Plutarch may be of special interest (233): "I believe now that the vast majority of his quotations are first hand (Plutarch kept notebooks of his reading and referred constantly to them when writing), and however exact or loose in wording (among those now checkable) they are (not entirely, but mostly) faithful." He concedes that what Plutarch then does with the passages he quotes "is, in our eyes, scandalous." Among other authors who would repay such study, Kidd names Galen and Diodorus Siculus.

A. Laks, "Du témoignage comme fragment", opens with reflections on the distinction between fragment and testimonium and offers some pertinent suggestions for the layout of editions. He then turns to his main purpose, a study of doxographical method. He reviews Diels' epoch-making reconstruction of Aëtius from pseudo-Plutarch and Stobaeus; this work, accomplished in Doxographi Graeci, became the foundation of Die Fragmente der Vorsokratiker. Laks, while asserting its essential soundness, is after something else as well, an understanding of how, in many doxographical lists, the linear succession of items conceals a series of dichotomies. The branches of this "arbre diérétique", forming a structure of possible answers to the question that interested the doxographer, are the main (if not quite only) principle of doxographical composition. Here Laks is following pioneering work by J. Mansfeld and D. Runia; however, he notes Mansfeld's observation that Aristotle himself recommends such a procedure (Top. 1.14, quoted on p. 259).

In "The fragments of Hellenistic scholarship", Franco Montanari concentrates on Aristarchus and Homeric philology. While many of the contributors wrestle with the problems of properly delimiting fragments, Montanari's material provides some extreme cases: "fragments" consisting of the adoption of a transmitted reading of Homer, "fragments" consisting of a lectional sign, even "fragments" consisting of blank space (sc. within a text whose layout signifies editorial choice). Montanari considers some problems of subscriptions, epitomes, and the identification of ipsissima verba within a lexicon or corpus of scholia. With regard to a famous controversy, whether the injunction "to explain Homer from Homer" truly belongs to Aristarchus, he adduces a neglected piece of evidence from the scholia to Pindar and refers to a forthcoming article for full discussion.

Ann Ellis Hanson, in "Fragmentation and the Greek medical writers", begins by surveying fragmentary Greek medical writing and the authors, texts, and editions from which it must be extracted. She notes that computer word searches have made possible secure attribution of many of the rapidly growing number of medical fragments found on papyrus, and that recent work has firmly established that medical recipes are highly conservative textually. She then develops her main thesis, that in the case of such works "stemmatic" compositional models, which end by attributing nearly everything to anonymous lost works, are less helpful than "accretive" models, wherein the inclusion of material in extant collections indicates that it continued to meet the expectations of doctors and patients. The "accretive" paradigm, she maintains, also fits the conditions of orality better. She then studies the scribal use of dicola, on-line spaces, paragraphoi, and the like in some medical papyri and argues that these contain reflections of orality.

Taking some remarks of Walter Benjamin as his starting point, Hans Ulrich Gumbrecht, in "Eat your fragment: about imagination and the restitution of texts", first draws a broad contrast between medieval and modern notions of presence (and hence representation) and then advances an historical thesis linking an alleged "crisis in representation" c. 1800 (which should have been better documented) with the simultaneous surge of scholarly interest in architectural and textual fragments. Gumbrecht argues that the features of the latter lead scholars to dwell on their materiality, and that this activity fulfills a cultural and in some cases religious need. It also bears profoundly on the functioning of the imagination, which, however, has always been regarded with unease by scholars engaged in the restitution of ancient monuments and texts. The relationship of this closing essay to the rest of the book seems to me problematic, not just because it deals with no Greek or Latin texts, or because it focuses on the interplay between what it calls "gestures of rationality" and imagination, but because its simple historical thesis does less than justice to the details of the scholarly enterprise carefully traced by Dionisotti and Grafton in particular, but to some extent by most of the other contributors. On the other hand, the conditions around 1800 are surely an important part of the whole story.

The production of the volume leaves something to be desired. There are misprints in all the languages used, many dozen in all, and some of the English, in particular, has been left in an unidiomatic state. Pages began falling out of the back of my copy as I neared the end, so that Gumbrecht ... indices ... fragmentary.  


1. Grafton's essay has 70 learned footnotes.