This slim book – just over one hundred pages – is another byproduct (cf. BMCR 2017.07.18) of Conte’s Teubner editions of the Aeneid (20091, on which see BMCR 2010.10.03; 20192) and the Georgics (2013, on which see BMCR 2014.02.47). It is loosely organised, but five distinct thematic blocks can be singled out: a series of essays on some of Conte’s more prominent predecessors, Heinsius, Heyne, Ribbeck and Sabbadini (pp. 1–45); a paper on the recentiores of the Aeneid (46–57); a number of textual and interpretative notes on individual passages (58–66, 82–91); a paper on parataxis in the Aeneid (67–81); and a more theoretical ‘addendum’ on textual criticism in general (92–106).
The addendum seems a good place to start, as it offers important insights into Conte’s text-critical principles. Its main thrust is to challenge Tarrant’s conception of textual criticism as a form of rhetoric;[1] Conte feels ‘concerned that Tarrant’s authority might encourage […] a critical relativisation, as if the truth was just a matter of opinion and could be established by means of good rhetorical arguments. In marked opposition to this, I believe that the truth cannot be relativised, even though every step of the progress that must be made to reach it is relative. Trusting evidence and testimonies is the only path that leads to the truth’ (p. 105). To reach this conclusion, Conte discusses several cases of a textual problem for which a (seemingly) definitive solution has been obtained with the help of new (or previously neglected) evidence. While it is true that textual choices based on hard evidence are, ceteris paribus, more cogent than ones based on speculation, and while no-one will deny that new evidence can be useful, Conte’s positivism seems to present a rather one-sided view of the challenges and practices of textual criticism.
The editor’s job involves two fundamentally distinct tasks. The first is recensio, the reconstruction of the paradosis, or more strictly the elimination of readings that can be demonstrated to be descendants of transmitted alternative readings and hence to have no evidential value. This is largely a mechanical operation, and in the ideal case of a tidy closed tradition it leads to the reconstruction of the archetype. The second task is examinatio, which asks whether or not any given transmitted reading represents the original. (Maas presents divinatio as the third stage of the editor’s work, but in fact it is closely related to examinatio: it is as a rule impossible to know that a reading is wrong without having at least some idea of what the right reading should be like.) In the simplest case of a unanimously transmitted reading, the question is whether it can be accepted as original or should be replaced by a conjecture. Textual evidence is ex hypothesi in favour of the transmitted reading; against it and in favour of a conjecture, the editor can only produce various kinds of comparative evidence (factual, linguistic, metrical, stylistic, intra- and intertextual, and so on). While textual evidence is limited, objective and immediate, it is a matter of judgement to decide which pieces of comparative evidence are relevant and what their weight is in any given case. The proper place of rhetoric in textual criticism is precisely here: in assessing the various kinds of comparative evidence.
Conte’s treatment of Lucretius 3.1 O tenebris tantis (pp. 100–101) is an illuminating example of his ‘evidence-based’ approach. Lachmann was the first to realise that the archetype must have read O, but he favoured the widely accepted humanist conjecture E nevertheless, even though O had already been adopted by Wakefield. Conte presents the case as if Lachmann failed to grasp that O is the paradosis, and as if once scholars realised that, the problem was automatically solved: ‘The fact that the archetype can be reconstructed in the case of the Lucretian passage above is the ultimate proof: the textual critic is faced with hard evidence guaranteed by inductive rules. […] Pertinent [linguistic] parallels have been finely brought forward by Timpanaro and Kenney, but the key point is that the restoration stems from a cogent logical reconstruction, not from an opinion’ (p. 101). This is a rather distorted account of the case: what actually happened was that Lachmann made a textual decision based on his understanding of Lucretius’ style, but Timpanaro and Kenney proved him wrong. Conte fails to realise that the question had been not whether O is the paradosis, but whether the paradosis is correct. His treatment is also misleading in that it presents E as the starting point and O as an innovation: while this might make sense in historical terms (from a certain perspective), logically O has undisputed priority. Finally, he never stops to ask what the editor should do if the paradosis were E instead of O. The same confusion of historical and logical order is apparent in Conte’s discussion of conjectures confirmed by papyrus finds: a new papyrus with a previously unattested reading does not provide (or confirm) a solution to a textual problem, but fundamentally changes its input parameters. To reiterate, while no-one will deny that new evidence, of any kind, is always welcome in textual criticism, Conte seems to undermine the fact that our evidential basis for the text of most (if not all) ancient authors is fundamentally defective; the editor simply cannot afford to pick and choose only those problems that allow of clear-cut solutions.
Virgil’s text is probably the least corrupt among those of Latin authors and the best attested by ancient witnesses, which means that more often than not the editor’s job consists in choosing between ancient variants, not between a unanimously transmitted reading and a conjecture. In the paper on the recentiores Conte attempts to expand the evidential basis of the Aeneid’s text further by arguing that mediaeval manuscripts may preserve authentic readings not attested by the ancient witnesses (direct or indirect). While this is inherently plausible, most (if not all) of the examples of a true reading found in a mediaeval manuscript (Conte discusses about a dozen) could easily be conjectures or lucky errors. A complete list of valuable readings solely offered by mediaeval manuscripts would have been far more useful, as this would at least make clear the scope of their input. Instead, Conte boosts their importance by somewhat misleadingly presenting them as a case of recentiores non deteriores: the fact that a late manuscript may preserve a number of otherwise unattested genuine readings does not necessarily make it an equal of an older manuscript (this is certainly true in practice: in his editions Conte does not cite the recentiores as systematically as the late-antique manuscripts).
A fitting illustration of the danger of excessive trust in ‘hard evidence’ supplied by the Virgilian recentiores may be found in Conte’s discussion of Aeneid 10.882–884 dixit telumque intorsit in hostem; | inde aliud super atque aliud figitque uolatque | ingenti gyro (pp. 5–6). The problem (not explained by Conte) is that the double –que construction invites taking the two verbs in 883 as a closely coordinated pair, whereas only the former can be construed with inde aliud super atque aliud. Harrison’s solution is to take inde aliud super atque aliud with intorsit from the preceding line.[2] Conte enthusiastically embraces uolutatque (for uolatque), offered by a mediaeval manuscript and endorsed by Heinsius. In doing so, he ignores the facts that uolutare means ‘to roll’ rather than ‘to hurl’ and appears never to be used of a weapon, and that gyrus means ‘circle’ rather than ‘arc’ (Mezentius is not throwing boomerangs) and likewise appears never to be used of a weapon’s trajectory (but is commonly used of a horseman’s course).
While the aspiration to find decisive arguments is generally commendable, its drawback is that in the absence of such arguments the editor may nevertheless be compelled to view whatever solution s/he accepts as definitive. In principle Conte admits that not all problems can be solved to everyone’s satisfaction (‘an aporetic discussion’ is offered on pp. 82–85), yet in practice he routinely speaks of ‘right’ and ‘wrong’ textual choices. When he is right (and he often is), little harm is done; when, however, doubts exist over the accepted solution, such overconfidence can be quite detrimental. For instance, Conte uses Georgics 2.69–72 to illustrate Heyne’s superiority as an editor of Virgil: Heyne retained the transmitted hypermetric verse ending arbutus horrida at 69 and printed castaneae fagus [sc. flore incanuit] at 71. Conte speaks as if Heyne’s (and his own) text were as solid as a rock (pp. 23–24: ‘Wagner, a zealous as well as a short-sighted student […], changed the text and accepted the transposition. Ribbeck followed his mistake. […] Once again, Wagner’s revision of Heyne’s text is wrong, and this contributed to Ribbeck’s disorientation, leading him away from the correct solution which had been proposed by Heyne’), but both decisions are far from incontestable. In the former case, Conte completely ignores Goold’s detailed objections against the admissibility of this instance of hypermetric verse.[3] In the latter, while he is right that the vulgate castaneae fagos [sc. gessere] makes poor sense, it is hard to believe that, given the existence of the fourth-declension plural fagūs, fagus can readily be interpreted as the nominative singular; the need to supply flore incanuit from the following clause strains the wording, and the resulting parallelism (‘the beech has grown white with the chestnut’s snowy bloom, the ash with the pear’s’, tr. Goold) feels tastelessly overblown; Scaliger’s castaneas fagus [sc. gessere] appears superior. To follow Heyne’s textual choices in this passage might perhaps be defensible, but to use them as an illustration of his editorial sagacity is self-defeating.
If Heyne is the ideal of eclectic conservatism, Conte also has words of admiration for Heinsius, even though his ‘conjectural abilities were not cut out for the relatively well-preserved text of Vergil’ (p. 4).[4] The first truly critical (and sceptical) editor of Virgil, Ribbeck is predictably Conte’s bête noire, yet one would be more interested in a comparison of Ribbeck’s radical first (1859–1866) and more cautious second (1894–1895) editions, than in the protracted chastisement of his belief in a second redaction of the Georgics (p. 38–42). Sabbadini, a fundamentalist conservative, is the least interesting figure in the cast of Conte’s predecessors.
Though no doubt intended to justify his editorial approach in general and some of the textual decisions in particular, Conte’s latest Virgilian book serves rather to lay bare his weaknesses as a textual critic. And this is a good thing: to make further progress in the editing of Virgil, we should be more attentive to the failures than to the successes of the current editions. Conte has mostly succeeded in offering a comprehensive (but judicious) account of textual evidence (recensio); we should hope that the next editor will be no less meticulous in questioning this evidence (examinatio).
Notes
[1] Developed in R. Tarrant, Texts, Editors, and Readers: Methods and Problems in Latin Textual Criticism (Cambridge, 2016).
[2] S.J. Harrison, Vergil: Aeneid 10 (Oxford, 1991), 279.
[3] See G.P. Goold, ‘Hypermeter and elision in Virgil’, in C. Damon et al. (edd.), Vertis in usum: Studies in Honor of Edward Courtney (Munich, 2002), 76–89. Cf. G.B. Conte, Critical Notes on Virgil: Editing the Teubner Text of the Georgics and the Aeneid (Berlin, 2016), 4, expressing surprise that Pomponius Laetus’ normalised reading nucis arbutus horrida fetu ‘is still found in the excellent text of Goold’ (i.e., in his 2000 Loeb) and apparently unaware of its discussion in the 2002 article.
[4] In passing it may be noted that Heinsius’ corrections deripiunt for diripiunt at Aeneid 1.211 and deripientque for diripientque at 4.593 (praised by Conte on p. 8) are anticipated in sixteenth-century editions.