BMCR 2006.01.24

Marginal Scholarship and Textual Deviance

, Marginal scholarship and textual deviance: the Commentarium Cornuti and the early scholia on Persius. Bulletin of the Institute of Classical Studies. Supplement 84. London: Institute of Classical Studies, 2005. xiii, 241 pages. ISBN 9780900587962. £45.00.

1 Responses

James Zetzel has been struggling for many years with perhaps one of the most forbidding texts within a genre that seems no less forbidding to most of us. With respect to the scholia on Persius, he rightly felt that Otto Jahn’s 1843 edition finally had to be superseded by a modern, better-documented one, and this edition Z. and Clausen have recently completed for the Bibliotheca Teubneriana.1 Marginal Scholarship and Textual Deviance comes as a welcome and indeed necessary supplement to that volume, but it contributes much more: we now have the first, and a very fine and extremely dense, monograph dealing with many major problems of scholiastic research and editing.

Z.’s study focuses on two main points. On the one hand, we get detailed information — formerly given in a Latin praefatio — about the manuscripts and the making of Clausen’s and Zetzel’s Commentum text (a decision similar to that of J. B. Hall’s in publishing his Prolegomena to Claudian separately as a BICS supplement in 1985). On the other, there are considerations of a general character, featuring mainly the textual characteristics and nature of, respectively, “commentaries” and “scholia”, the different editorial approaches these two kinds of material demand, and the relationship between ancient and medieval commentaries. There also is a concluding chapter in which Zetzel touches upon modern (and post-modern) notions about texts, their (in)stability in transmission, and the implications of this for future editing. This last and highly sceptical part of the book has not prevented him, however, from adding three valuable appendixes of minor Persius scholia that have — unfortunately, in my opinion — not been included in the Teubner volume.

Reassessing the various chapters requires attention to minute detail. Yet this is due to the book’s line of argument, because Zetzel has chosen, first, to lead us into particular problems of the manuscript evidence, second, to widen the horizon again to cover further, but no less complex material, and only then to draw conclusions about what editing scholia is all about. Indeed, there seems to be no other possibility of proceeding in a field of research the pitfalls of which become clear only to those who are already lost in it. If my own experience in working with scholia does not mislead me, the order of chapters here presented is an exact rendering of how Zetzel must have felt like after having concluded his edition, while yet having to explain it to others. Anyone working in this area has known moments of true despair, and I therefore am the more grateful and full of respect for his concluding with such an overview.

Chapter I, “Scholia and Stemmatics” (pp. 1-9), sets the terminological framework. From the start, Zetzel warns us not to treat commentaries, which are continuous texts, and scholia, which are not, as one and the same scholiastic matter. He is careful enough, though, to add that this division at the same time oversimplifies the evidence, as there are many instances of combination of the two “ideal types which do, in fact, sometimes exist in reality” (p. 5). This shows right from the start that any decision taken about what type of commentary should be included in (or excluded from) an edition needs to remain open to revision, and this is the main problem the whole book has to tackle: even if one is dealing with commentaries only, one will not be able to exclude the other, greater (and even more heavily disordered) half of the evidence. Moreover, Z. makes it clear at once that in the case of Persius, the surviving scholiastic material will never allow reconstruction of an ancient commentary (p. 8 f.). This means that “Cornutus”, the putative author (?) of the “Commentum”, is to be looked for in the earlier Middle Ages (see below on ch. 6).

Chapter II deals with “The Major Manuscripts of the Commentum” (pp. 10-48). This is, in a sense, the additional praefatio which the Teubner volume lacks. In his discussion of the four major witnesses used in establishing his Teubner text,2 Zetzel makes three major points.

1. R has recently been overestimated, in that it preserves an undoubtedly easily readable version of the “Commentum”, but one that has obviously undergone revision. It would thus have been unwise to prefer all its readings to others, especially those of the Leiden ms. (L, ca. AD 1000) and its Munich relatives.

2. Z. argues convincingly that a passage (on Sat. 1.11-23) commented upon twice in mss. MLU has not only been inserted by their common ancestor from a second source, but was also part of R’s source before it was cut out again by the Londiniensis. This adds not only to the image of R as having been revised, but also throws more light on the creation of what is now our “Commentum” text from several series of very similar scholia.3

3. The most arduous and, I think, controversial point is Z.’s attempt to explain the presence of lemma symbols (.SS.) in many places in MLU, though almost nowhere in R. Z. proposes to interpret these symbols as remainders of a major change of the mise en page probably created by the compiler of the whole, integrating the entire text of Persius’ Satires word by word into his commentary in order to save his readers from the need to consider a second ms. with the text. This is fascinating because it implies a sort of acrostic Persius text in capital letters scattered across the pages of the commentary and thus quickly readable, but I am not quite sure if it works. First, it seems to be a singular phenomenon (apart from an example in Jerome, cited p. 65 n. 2), but what makes me really unsure about it are the inconsistencies between the extant mss.: why , for example, should it be that in R all lemmatization has ceased to exist a relatively short time after it was introduced in the first place? Why, on the other hand, does L preserve the vast majority of lemmata, that is, an entire “second text”, although in its left column there is already a complete copy of the Satires ? And must we not deduce, from the fact that in some places the same lemma appears twice (see p. 72 ill.), that this occurrence results from the combination of two different series of scholia, a venture hardly to be undertaken by one scribe at the same time as the lemmatization of a commentary text? And why, again, should marginal scholia have continuous lemmata, anyway? I do not mean to doubt the principle behind Z.’s proposal, nor do I have an alternative explanation for the recurrent .SS. signs. It is clear, though, that here we are touching upon the crux of the matter: how , and over how much time, did the creation of the “Commentum” as we have it take place? Even though there are only a handful of mss. dating from a period of less than a century (s. X ex. – s. XI m.), it seems that there must have been many successive steps: series of scholia, merging of these series, lemmatization, probably disruption again, and so, finally, our few “text” mss. Z. is well aware of these difficulties, but by constantly reevaluating the evidence while the argument proceeds he may cause some readers to feel disorientated, as the chapter concludes without a summary.

Chapter III, “The Later Tradition of the Commentum” (pp. 49-64), focuses exclusively on continuous (“text”) commentaries. There are actually many more mss. than are mentioned here, even ranging far into the 15th century,4 but few are as closely related to each other as are MLU (and perhaps N). Therefore, Z. is surely right in saying that only this extraordinary similarity of MLU makes an edition possible at all (p. 52), but he might have drawn some further conclusions at this point, giving at least some information about the provenance of these mss. Even if there is no itinerary of “Commentum” mss. to be drawn, this aspect of Überlieferungsgeschichte tends to be particularly illuminating. For example, there are strong topographical links between MLU, in that all three mss. are from the Rhine region (L) or from southern Germany (MU). This does not yet clear up their origin, but I am convinced that even in a kind of text that is subject to permanent change as much as these are, we can positively identify local workshops and thence draw new conclusions about the circolazione dei testi in Medieval Europe.5

Chapter III also deals with the editiones principes, about which Z. proposes some new details, and, in a particularly important passage, with the famous Glossae veteres published by Pithou in 1585. This is both a relief and a delusion at once, because now we know for sure that these are not excerpts from some very old ms. (similar to the famous Montpellier Juvenal, Fac. méd. 125), as has long been thought, but merely excerpts from Elie Vinet’s 1563 edition made by Scaliger and Pithou, the printing exemplar of which Z. has been lucky to identify in Paris, namely BN Dupuy 394 ter. Additional scholia not found in Vinet come indeed from Montpellier 125 (only on a few lines though) or are to be interpreted as conjectures by the French scholars.6 Z. has printed the Glossae veteres as Appendix I (pp. 162-179).

With Chapter IV, “The Shaping of the Commentum” (pp. 65-85), we are moving away from strictly prefatory matters towards the crucial question of how the “Commentum” has been and is now organized. We are thus back with the lemmata system already mentioned, and Z. proposes a lot of new and valuable ideas about the kind of composite text we have before us.

Z.’s argument is based on the assumption that the “Commentum” is “a continuous and ordered text” that came to life as “the system of lemmata was superimposed on pre-existing exegetical material” (p. 65), in other words, scholia. This seems likely enough, but his subsequent attempts to explain this process of “shaping” the text remain somewhat confusing to me. First we are told that the system of lemmata mentioned above “was never executed with… the precision that it requires” (p. 65), which means nothing less than shifting the responsibility for inconsistencies from scribes of the MLU generation (see above) to the earlier one of the Commentum compilers themselves. Next, Z. has to consider the intrusion of independently transmitted (and at times superfluous) paraphrases and glosses into the text as well. This implies a compiler who does no fewer than three things simultaneously: merge two marginal versions of an ancient commentary, combine this with a Persius text scattered as lemmata, and add glosses. I find it hard to believe that a single “‘Cornutus’ created this book” (p. 66), for two reasons. First, Z. himself rightly calls this work a “process”. This implies a longer period of time than is reasonably to be spent by a single person on such a project (I just cannot imagine a Carolingian scholar indulging for years in creating a Commentum of this kind). Second, why do we have to cling to the “Cornutus” fiction — which is almost done away with by Z. later on anyway — and attach this name to one single step in the development of the text? If a Cornutus (or “Cornutus”) was the creator of “this book”, then why do we have to explain the existence of further, partly even superfluous, glosses by mistakes made by this very same man (p. 83)? Is it not much easier to imagine that further glosses have been added by others, after the compiler’s creation had begun to circulate? Why attribute all we have to a single person, when assuming several revisions after the big one provides an equally plausible explanation? I would therefore agree perfectly with Z. that by content, the “Commentum” has been a variorum commentary ever since Late Antiquity (cf. p. 75), but insist that it is also the product of various revisors of the 9th to 11th century.

Again, it must be said that all the above-mentioned difficulties in Zetzel’s book never spring from mere speculation. On the contrary, his attempts to explain the structure of the text, and to find out more about its creator, always stem from highly precise observations concerning minute details of the “Commentum”. In short, the great variety of solutions Z. offers is the consequence of the fact that whatever theory one may apply, there are always one or two marginal glosses that will contradict it. This is why the more parts of the whole ms. material one takes into consideration, the more twists and turns become necessary for reaching a solution that can account for the whole. In a way, then, the middle part of Zetzel’s book retraces the winding path an inquiry into the “Commentum” has to take, but I fear the average reader may get lost somewhere in the middle.

Chapter V, “Marginal Scholarship” (pp. 86-126) takes us to the most complex part of the evidence and is therefore the longest and most intrinsically problematic of all. To give an idea of the yet more floating borders between formerly fixed ‘continents’ it will suffice to note that at the beginning of ch. 5, Z. insists on the division between commentaries and scholia (p. 86) whereas at the end he says “the Commentum can be shown to exist in multiple forms” (p. 125). In a sense, this is again true, because at least part of the scholia that must have constituted the sources for the reviser were never fixed into the Commentum text, but have continued to circulate in marginal form. On the other hand, it is of little help, if not arbitrary, to label sets of scholia as different ” Commentum versions” after going at great length to fence in a “protected” Commentum text. As there are too many different Cornutuses around in ch. 4, there is too wide a sense of what to subsume under the “Commentum Cornuti” label here.

Moreover, extending the inquiry to the scholia implies having to deal with what D. Robathan’s Persius article in the Catalogus Translationum et Commentariorum has called “A tradition” (our MLURN, among others) and “B tradition” mss. Whereas it has by now become undoubtedly clear that this classification is too strict and that there are indeed many textual overlappings between the two classes, I was very surprised at Zetzel’s statement that “the differences of content are relatively minor compared to differences of wording” (p. 86). What, if not “differences of wording”, makes for a clear difference at all between two manuscripts? Even if it is not advisable, or perhaps is even impossible, ever to publish the B scholia as a whole, it remains true that every element in a word that differs between one ms. and another is a difference of content, and has to be taken into account. Nor is it impossible to observe relations between single manuscripts of scholia or even subgroups of these, though this is of course far more difficult than with texts. Instead, due to Z.’s main (and very legitimate) goal of reevaluating the connections between A-texts and B-scholia, the transmission of the latter is mostly omitted.7 To be sure, no one will want to draw a stemma of scholia mss. nowadays, but I think it is possible, as with MLU, to identify temporary workshops of medieval Persius exegesis even by studying scholia mss. — which implies a minute analysis of all extant scholia mss. I am not saying Z. should have done this as well, because the task would probably have taken him many more years, with little hope of any great results, but I would not want him to dismiss this strain of research as “not worthwhile” altogether (as his judgment about the two most important B mss. on p. 88 seems to imply) merely because it does not help reconstruct a “Commentum” text that, after all, is not as fixed as he (or we) would like it to be.

What, then, are the Persius scholia ? Zetzel’s presentation of these matériaux flottants is instructive because he uses three very different pairs of mss.8 On these very specialized issues I would only like to give two comments.

1. To compare W (from Cologne) and O (from England) means to create a link between very different corners of the Persius galaxy, even though they may have many things in common. The more interesting aspect of W, to me, seems to be that its text visibly undergoes a process of rearrangement from scholia to (something like a) text,9 which is very similar to what happened in MLU. Thus, the main question in my view is not so much whether there have been different sets of scholia.10 Rather, we will have to explain how the creation of A, B, and related traditions could have happened simultaneously. This works only if we accept the notion of different contemporary workshops in different places as, say, Auxerre, Reims, and Cologne that reevaluate and develop the same text.

2. With the second pair of witnesses, KQ, Z. analyses the main mss. of the B tradition. The question here is: how much is there in the B tradition that is genuine? Z. shows convincingly that the percentage of genuine material gets smaller the further we proceed. This may indeed mean that the early creators of this set of scholia got increasingly tired of adding their own contributions to a basic material that is mostly identical to the A tradition (p. 106), but it could just as well have been the other way round. I do not think, though, that the existence of abbreviated majuscule lemmata in Q should be explained as the result of an influence from the Commentum text (p. 111). This clearly overestimates the “system of lemmata” mentioned above, and it complicates things further. Therefore, truly important finds such as the following risk getting drowned in the mass of detail: if it is true that the quotations from Solinus in KQ are the version used not by Heiric of Auxerre (who is believed to be “Cornutus”), but by Lupus of Ferrières, this may mean that the B tradition is, after all, something different from A. This is one of the very few windows into the otherwise dark ninth century (p. 120).

So what, then, is the “Commentum”? After ch. 5, one gets the impression that Z. now wants to include both text and scholia under that label. But he does not stop there: on p. 126 he says that all extant mss. are “overlapping fragments of a text that never existed” (my italics). Now this seems sheer desperation — even if, again, he is right in some respect. Yet after having insisted so far on the notion of “text”, this move is abrupt, to say the least, and one might feel like “having been beaten up for some time by a terribly intelligent person” (as Eduard Hanslick put it, after first listening to Brahms’ Fourth Symphony).

No wonder, then, that in ch. VI, “The Origins of the Commentum” (pp. 127-143), Z. is rather succinct. Nonetheless, this is a condensed and easily readable summary of what we can possibly say about the early medieval attempts to comment on the satirist. Surely, Heiric (b. 841) remains a good candidate for “authorship” (or what you may call it), and Remigius’ role is assessed correctly. To him, around AD 900, A and B texts were already available (p. 140 f.). And lastly: what about Cornutus? I can understand that at this point, neither Z. nor the reader care too much about his identity. But anyway, to call him just “a red herring” or indeed “a random name” (p. 136. 8) is, again, a little brusque, and also a bit disappointing. Moreover, it does not really explain why, after all, we should explicitly call the text version presented in the Teubner series a “Commentum Cornuti”. At the end, the only Cornutus we seem to get a hold on is the (ancient) scribe with his diptych, beautifully reproduced from Leiden B.P.L. 82 on the front cover.

Ch. 7, “Textual Deviance” (pp. 144-161), draws conclusions from the entire enquiry. They are radically sceptical, but perfectly understandable. While working on the Teubner edition,11 Z. must have realized that a text like the Commentum, after all, can hardly be published in this series’ traditional form. This is very honest, but it has led to some new problems. Thus, from the insight that the transmission of scholiastic material is “more like the Epicurean rain of atoms through the void” (p. 126), we have moved to a situation where several pieces of the whole have landed in various places: a “classical”, long-awaited and, I think, extremely reliable critical edition on the one hand, which has nevertheless appeared in a series that does not normally publish medieval texts, and on the other, a BICS supplement that not only contains supplements to that edition (the “preface” and the appendices), but is also a major achievement in critical thinking about what philologists can, or should (not) do. As Z. moves from almost-stemmatics featuring MLUR towards the question of whether editing is at all possible, he has us think twice about our profession, which is undoubtedly healthy and worthwhile. And yet, his invective against the Lachmann method dominating ch. 7 may seem a bit over the top: after all, the problems arising in the Commentum edition have long been known to most philological disciplines,12 and it is perhaps more due to the fact that scholiastic editions have become, ( secundum Remigium,) rarae aves nowadays that we feel their impact so strongly in this case.

This said, I sincerely hope that after his Cornutus experience, Zetzel’s choice of title, “Marginal Scholarship”, does not mean we should leave aside this kind of occupation from now on, but that we have here a classic example of the double entendre possible in English titles’. Zetzel has asked, and treated with great subtlety, many questions that are not marginal at all, and I hope, too, that his book will become known as a standard introduction to a whole field of textual criticism and editorial technique. It is not a handbook, but it is full of important insights, clues, and warnings to all who are going to deal with the fascinating and exasperating stuff we call scholia (or commentaries).

[[For a response to this review by James E. G. Zetzel, please see BMCR 2006.01.46 ]].


1. Commentum Cornuti in Persium recogn. et adnot. crit. instrux. W. V. Clausen et J. E. G. Zetzel, Monachii et Lipsiae: Saur, 2004.

2. M = Clm 23577, L = Leid. BPL 78, U = Clm 14482, R = Lond. Royal 15.B.XIX.

3. The doublet had already been explained thus by Z. in “Medievalia et Humanistica” 10, 1981, 25-6, but his new treatment is much clearer. — There is a fifth ms. regrettably overlooked by the editors, Nürnberg, Melanchthon-Gymnasium, cod. Ebnerianus Latinus 4 o 36 s. X-XI which adds to the explanation proposed by Z. This ms. (ν which is closer to MLU than to R, also transmits the double passage, but has omitted the scholion marked by Z. as 1,9,4 (cum … ludunt), thus trying to revise the doublet too, but in a different way than R does.

4. For a list, see Claudia Wiener, Persius-Kommentierung vom 10. bis zum 15. Jahrhundert, Würzburger Jahrbücher 26, 2002, 171-173. Others adduced by Z., as Bern 265 (pp. 55 f.), have up to now been classified as belonging to the B tradition of Persius scholia.

5. See my: Persiuskommentierung im ottonischen Köln, Würzburger Jahrbücher 28a, 2004, 147-161.

6. A shade of doubt remains because, though the Dupuy ms. bears page numbers, these are quite, but not exactly, the same as in Vinet’s edition. Maybe there were several drafts. Z.’s appendix is the more welcome as neither Pithou’s edition nor the 1991 microfiche reproduction are easily available.

7. Despite this, there is a very valuable Appendix III (pp. 193-227) reproducing for the first time those (“B”-) scholia not found in “A” mss., taken from six mss. belonging in turn to various groupings. It is a pity that this appendix has not been included in the Teubneriana instead, where Z. himself (193) says it actually belongs.

8. K = Krivoklát (not Prague) I d 31, Q = Leid. Voss. lat. Q. 18, W = Wien 131, O = Bodl. Auct. F.1.15, Paris BN 8049 and 8272. The marginal scholia on Sat. 3.99-5.12 in Vat. Reg. lat. 1560 fol. 72r-80v should have been mentioned, as this is the manuscript which also contains the Remigius fragment, a much discussed, but never before completely published text. Z. now gives an edition as Appendix II which is very welcome.

9. In the roughly contemporary Cologne Cathedral ms. 199. See note 5.

10. Or even “multiple commentaries” (p. 99), which I do not believe to have been the case: If our only remaining text — MLUR — is as poor as it is, we cannot possibly propose the existence of many different Persius commentaries around AD 1000. Where could they have gone?

11. And probably even more during the recent APA Panel discussing “Whither the APA/Harvard Servius?”, from which I quote Z.’s own statement: “Servius is a text, but DS is part of a process that extends over centuries…” APA 2004 Servius Panel So is the Commentum Cornuti.

12. A good summary of the main questions can be found in: Joachim Heinzle, Zur Logik mediävistischer Editionen (Einige Grundbegriffe), editio 17, 2003, 1-15.