Seneca Tragicus now joins the very small number of authors who have a book devoted to listing conjectures on their text: one thinks of R.D. Dawe’s Repertory of Conjectures on Aeschylus (1965) and W.R. Smyth’s Thesaurus Criticus ad Sexti Propertii Textum (1970: one of the earliest Mnemosyne supplements, at number 12). Billerbeck and Somazzi aim to collect systematically those conjectures recorded over the past four centuries, beginning with the second edition (1619-20) of Delrius’ Syntagma Tragoediae Latinae. However, they also include much material from the earlier editions: two appendices investigate the editions of Ascensius (1514) and Avantius (1517) and their relationship to other early editions and to the recentiores.
The Repertorium lists some 10,000 conjectures. This figure indicates the scale of the task which Billerbeck and Somazzi have undertaken. It is also an index of the sheer volume of scholarly industry on these texts over many centuries. Only a very few of the 10,000 conjectures, of course, have any chance of winning general acceptance or appearing in the text of a critical edition. If conjecture were an independent activity aimed solely at correcting the text, like the work of a police detective solving crimes, one would have to conclude that it is a low-yield pursuit. But in fact conjecture is only one outcome of the whole philological enterprise, of intensive engagement with all aspects of a text including word choice, style, metrics and so on.
One purpose of such a repertory is to attribute conjectures correctly to their first proponents. This salutary process can have uncomfortable consequences. Both Zwierlein and I are divested of several of our conjectures by earlier critics;1 in weaker moments one entertains suspicions of “anticipatory plagiarism”, in Mark Golden’s delightful phrase. Usually Billerbeck and Somazzi list the first proponent and ignore any repetition, but occasionally diverge from that policy;2 sometimes one can surmise the reason, but a word of explanation would have been helpful.
Another function of a repertory is to enable scholars puzzling over a passage to consider the whole range of existing conjectures. Negatively, this enables them to avoid repeating conjectures or making inferior conjectures of their own. Positively, the process of emendation is often a collaborative one, in which one critic’s proposal is refined by later critics, or serves as inspiration for them. Many instances can be found in Repertorium : at Pha. 596, for example, Axelson’s admovimus (for the transmitted amavimus) was improved by Watt to iam movimus, which fits the context well while providing an explanation of the corruption (omission of m through haplography or abbreviation).
One useful feature of Repertorium is to identify those conjectures which are “contra metrum” and therefore automatically disqualified: a warning to readers not to consider the conjectures in question, and a reminder to future emenders to ensure that their conjectures scan. Such a reminder should hardly be necessary, but the point is sometimes overlooked, notably by the prolific Giardina.
Billerbeck and Somazzi claim to record not only conjectures but also “punctuation decisive for the interpretation of the text” (4). This is of course an impossible promise, given the variety of editors’ usage. In fact Repertorium restricts itself largely to noting question-marks, but even then not consistently (e.g. my own at Oct. 448 is missed). Surely we need an amnesty in this regard, since there is after all no question of restoring an ‘original’ punctuation, and punctuation marks mean different things in different centuries. Editors should simply punctuate as best reflects their understanding of the text, and changes in punctuation need not be attributed unless they radically alter the sense.
Completeness and accuracy are major desiderata for such a repertory, and Billerbeck and Somazzi score well on these counts.3 They have sought out obscure and inaccessible sources, such as the handwritten materials left by Hoffa and Düring, both killed in the 1914-18 war. Of course they are not perfect. As my mortification for the day, I checked Bothe’s Annotationes to HO in his 1819 edition, and found three conjectures omitted in Repertorium, in contrast to some 110 recorded and many others set aside as pre-empted.4 My purpose here is not to take the compilers to task; scholars are no more immune from errors than the scribes whose mistakes they attempt to correct. But it would be nice to have some mechanism for correcting errors and omissions as they are noticed.
That leads me to the issue of electronic publishing. In the 21st century it seems almost quixotic to publish a database (which is what Repertorium is) in book form. An electronic format would allow the user to search the collection, for example by scholar, by type of conjecture (lacuna, transposition), by Latin word or phrase, and so on. It would also allow the compilers, or later collaborators, to correct and augment the database, and above all to update it periodically as new conjectures are made, whereas a printed book freezes the database at an arbitrary point in time.
An electronic format would also make it possible to include advocacy or refutation of an original conjecture by subsequent critics, or their unwitting repetition of it. Such discussions can be as valuable as the conjecture itself. For example, Bentley’s numerous conjectures were made simply as marginal notes in his copy of Gronovius, with no more than an occasional citation of a parallel passage; explanations and justifications by later scholars can be invaluable. The possibilities for including such material are shown by the design of Monika Asztalos’ electronic Repertory of Conjectures on Horace, currently in progress.
Of course these possibilities are not lost for Seneca’s tragedies by the publication of the Repertorium. In the first place, Repertorium itself must surely exist in electronic form, which would serve as a basis for the kind of database I envisage. Second, Billerbeck and Somazzi must have a card index, or indeed an electronic database, of countless discussions which they eliminated from the book on the principle of listing only the first proponent of a conjecture. It would be a great service to future critics if they could make this further material available electronically — or (since they are deservedly emeriti of this project) if they could find collaborators who would do so. They have given us a “possession forever”, but one which can never be complete, and which would be even more valuable if it could grow with the years.
1.Zwierlein’s cometen at Oct. 232 was anticipated by Viansino, and some of his other conjectures were made by Peiper: Ag. 819 (lacuna), HO 1598 et, Oct. 262 illos, Oct. 501 alto. I lose Tro. 965b-66 (attribution to Andromache) to Peiper; Ag. 810 educans to Viansino; HO 1221 specus to Bentley; HO 1402 vae to Richter.
2.We find Peiper 1867 credited with attribution of Herc. 1032-34 to Amphitryon, but “praeeunte H. Weil … 1865”; perhaps the point is that Weil much later discarded the idea. Often the earlier source cited is one of the Renaissance editions, eg. Med. 161 ” nusquam Viansino (1965) (praeeunte Ed. Caietani ):” perhaps such entries the possibility that the original source of the reading was one of the recentiores.
3.I noticed very few errors, though I have not searched diligently. At Med. 64 the proponent of a lacuna was Peiper rather than Richter. At Pha. 594 sata is not what Trevet read, but is Heinsius’ conjecture, mindlessly inserted into Trevet’s commentary by his editor Chiabò.
4. HO 576 nitentes … vestes“secundum vett. edd.”; 1012 exstitit; 1819 regno ut iacent. Other omissions that happened to catch my eye: Pha. 326 harena Zwierlein; Ag. 210 caede for morte uncertainly Fitch, Annaeana 158; Thy. 467 ducenda or tangenda Bentley (according to Hedicke); HO 998 and 1167 del. Axelson, Korruptelenkult 10. (When lines are omitted by E (as are these HO lines) or A, Repertorium sometimes does not record who first regarded them as spurious. This seems to me a mistake: such lines are often agreed to be genuine, so the decision to delete is not automatic, but a matter of judgement.)