Georg Luck’s Teubner is the only truly “critical” edition of Tibullus now available, and for the remainder of the Corpus Tibullianum its only rival is Tränkle’s Appendix Tibulliana. When I reviewed Luck’s first edition at CR 39 (1989) 211-212, I expressed the hope that there could be a second edition correcting some minor deficiencies of the first. A second edition has now appeared, and it more than surpasses expectation: almost everything has been rethought and redone and, in general, improved (only 2.2, 4.11, and 4.14 survive without alteration in either text or apparatus). The new Roman and Greek fonts used are less appealing, to my eyes at least, but the introduction has been expanded significantly, and both text and apparatus criticus have been corrected and revised in the light of subsequent scholarship and of the reviews of the first edition.
The changes to the introduction appear to have been made in two stages, since we have not only a slightly altered version of the original (besides the correction of some misprints, note especially the addition of a footnote “6a”) but a Praefatio Editionis Alterius as well, with an update on the most recent scholarship. The most extensive changes, however, are to be found in the list of manuscripts; thanks largely to F.-R. Hausmann’s researches published in Kontinuität und Wandel (Berlin 1986), this has leapt from 136 items to 193. Unfortunately, the resetting of this section has replaced most of the commas with periods, so that (for example) shelf-marks of mss in the Laurenziana now appear in the form “pl. 38.39” instead of “pl. 38,39” (this particular one also has a period after it instead of the comma used elsewhere).
This second edition differs from the first in dozens of places and, in general, offers an even less conservative text. I have counted 16 passages where Luck abandons a conjecture to return to the paradosis (an awkward term in Tibullus, which I use here to mean the agreement of AV; these are at 1.1.5; 1.2.81; 1.8.30, 39; 2.3.61; 2.5.72, 73; 3.1.9; 4.1.72 [ bis ], 82, 88, 99, 133, 195; 4.6.16). I have counted 31 passages where he makes a different choice among conjectures and/or ms readings (which are difficult to distinguish, given our inadequate knowledge of the Tibullan tradition); these are 1.1.78; 1.3.14; 1.5.13, 33, 34, 61; 1.8.17, 18, 59; 2.1.21; 2.4.12, 43; 2.5.4, 72, 99; 2.6.10; 3.4.3, 33; 3.5.11; 3.6.13; 4.1.3, 26, 78, 98, 129, 164, 181, 202; 4.7.1; 4.8.8; 4.10.3). But I have counted 53 passages in which Luck did not emend before but does so now (1.1.17, 35; 1.2.76; 1.3.4, 54; 1.4.1, 28; 1.5.11, 47, 57, 67, 72; 1.6.1, 21, 59; 1.7.4, 15, 64; 1.8.53; 1.9.1; 1.10.9, 46, 50; 2.1.31, 32; 2.3.58; 2.4.43; 2.5.25, 71; 3.1.13; 3.2.15, 18; 3.3.35, 38; 3.4.11, 18, 26, 35; 3.6.20; 4.1.21, 91, 104, 136, 144, 155 [ bis ], 167, 175, 183, 196; 4.9.2; 4.12.3; 4.13.9). In two cases the paradosis has apparently been abandoned without a corresponding note in the apparatus (“referunt” has replaced “referent” in 1.4.65, and “petet” has replaced “petit” in 2.1.71). Conjectures not mentioned in the apparatus of the first edition but now adopted include Lee’s punctuation of 1.1.39f., Lee’s “nunc” at 1.5.47, and Camps’s “Heu” at 2.4.43; many more previously neglected (or unavailable) conjectures are now recorded in the apparatus, including “rapiat” (Delz) and “iudicat” (Baehrens) at 1.4.44, “num” (Huschke) at 1.5.60, Lee’s version of 1.6.42, “tetigisse” (Broekhuysen) at 1.6.52, “dente putat” (Statius) at 1.6.80, “lege” (Della Corte) at 1.9.25, “Delphice” (attributed to “nescioquis”) at 2.3.27, “vota” (Kraus) at 2.3.59, “taetra” (Watt) at 2.3.62, and many more. Though he now emends more than in the first edition, Luck has withdrawn a number of his own conjectures, whether from the text or the apparatus (at 1.4.34; 1.5.47; 1.8.64; 1.9.3; 2.4.41; 3.1.13; 3.4.3; 4.1.38; 4.1.94; 4.8.1; 4.8.8); some others have been reassigned to those now known to have conjectured them first (“urbem” at 1.5.55 goes to Castiglioni, “Dis et in” at 3.6.21 to Ayrmann). On rare occasions Luck now advances a new conjecture of his own (the interesting “serta” for “terra” at 2.3.62; “longa” offered diffidently in the apparatus at 3.1.3).
The differences from the first edition involve more than just a different selection of readings. Luck is now more inclined to capitalize personifications, and so we have “Mors” in 1.3.65 and 1.10.33, “Fata” in 1.4.36, “Amor” in 2.3.28, “Sol” in 2.4.17 (Luck had already printed “Luna” in 18), and two different groups of “Sorores” at 3.3.38 and 3.4.46; but should we not also have “ite procul, durum Curae genus, ite Labores” at 3.6.7? Also newly capitalized are 2.3.66 “Pater” (the first edition already had “Pater” at 3.6.3), 2.5.118 “Triumphe”, and 4.2.20 “Kalendis”; we also have “Rubro de Litore” at 4.2.19 (but “Rubro … mari” still at 2.4.30), and Alba has become suitably “Longa” at 2.5.50. In orthography, Luck has now been thoroughly converted from “tent-” to “tempt-” and from “heu heu” to “eheu”; he is less rigorous over “tunc” and “tum”, but his note referring to Housman’s discussion at Lucan 1.499 has been moved appropriately to 1.2.79, the first place where the question arises (hence it no longer needs to stand in its original location at 2.3.71). Luck now prints “postmodo” (2.5.102) and “necopinata” (4.9.4) as a single word; “umentes” (1.9.38) and “umorem” (3.2.21) have lost the initial “h” that they had in the first edition, just as “harundinis” has gained one at 2.5.31 (though “arundine” still stands at 2.6.23). The punctuation is frequently different, of course, since the text is as well, but in most cases this is simply a matter of stronger vs weaker articulation (comma for period or vice versa, for example): more substantial alterations include the placing of quotation marks around 1.5.21-34 (just before “Haec mihi fingebam”), the enclosing of “nec me … pudebit” at 1.6.31 in a parenthesis, and the repunctuation of 3.3.2-16 (question marks at 2, 12, and 16 have been replaced with commas and the period at 10 replaced with a question mark). However, a comma is still needed after “Phoebe” in 2.5.121, and surely the colon at the end of 2.6.25 should be at the end of 26; I would punctuate with a colon, not a period, at the end of 3.6.62. For some unexplained reason, the lacunae that Luck now marks after 1.10.50 and after 2.3.58 are signalled by a set of three asterisks rather than the series of dots used elsewhere.
Of course the apparatus criticus is now different as well. Besides the places where it has necessarily been changed to reflect a different text, it also contains new material, including (besides the new conjectures mentioned above) new parallels cited in support (Ovid’s Metamorphoses appears especially often in this connection, though one such reference has also been dropped, at 1.4.43). Sometimes the reports of ms readings have been changed as well; without access to the relevant documents I must assume that these are all accurate corrections (for example, V is now reported at 1.7.11 as reading “garumma” rather than “garumna”; X is no longer reported for “discordibus” at 2.3.37; the report of G has been changed at 3.2.9, 3.4.82, 3.6.46 and 59, 4.1.27, etc.). I note with special interest that Luck now reports yet another reading that is shared by W (=Vienna, Österreichische Nationalbibliothek 224, a Florentine ms of about 1460-70) with Scaliger’s Fragmentum Cuiacianum; this gives me still more hope that it might be possible one day to determine whether Scaliger’s great “discovery” was nothing more than an interpolated fifteenth-century manuscript after all.
The second edition has also removed some errors that appeared in the apparatus of the first (2.1.58, a colon added after “hauserat GV”; 2.1.12, the “+” deleted after “( ex codd.)”; 2.3.75, a comma dropped after “precor”; 3.1.10, ” codd.” for ” cod.“; 3.5.7, a superfluous “)” deleted; 4.2.20, “4, 3, 20” corrected to “4, 3, 10”; 4.7.8, “VX” corrected to “HX”). In view of the spacing problems endemic in the apparatus of the new edition (see below), it is ironic that some of these corrections involve errors in spacing (at the end of the entry for 1.10.30, for example). In the text of the Vita, an erroneous repetition of “ro cod. Vat. lat. 2794, a. 1434” has been removed from the apparatus and “iudicio” has been properly deitalicized in the entry for line 4; but I wish that Luck had not corrected “utiles”, which (along with “epistolae”, a synonym for “elegiae” in mediaeval Latin) reveals a mediaeval addition to the authentic ancient elements preserved here. I was especially pleased to see the apparatus acknowledge the possibility that we have lost the end of 2.6, but credit for this hypothesis really belongs to Michael Reeve rather than to George Goold. A very few errors still persist uncorrected from the first edition; at 2.3.43 the apparatus still says “ex duo Palatinis”; in the apparatus at 2.5.97 the “e” in “e Z+” still seems to be in bold; and the apparatus at 4.12.1 still repeats “Q (II)” in reporting the mss that read “iam”.
Unfortunately, the second edition has also introduced a few fresh errors. In the text, for “Ed” at 1.1.46 read “Et”; 1.3.47 is not properly indented; at 1.6.34 read “inest” for “inset”; at 2.1.43 there is too much space after “abiere”; at 2.5.22 read “Ilion” for “Illion”; at 3.4.74 there is too much space after “coniugiumque”; at the end of 3.6.47 there should be either a comma or no punctuation at all; at 4.1.167 there is too much space after “tenens”; and at 4.5.7 there should be no comma after “te” (which is the object of “rogo”, not of “per”). In the apparatus, there is now a superfluous colon at the beginning of the note on 1.1.5; at 1.1.55 there should be a colon after ” Mueller“; at 1.5.47 “huic” should not be italicized; at 1.7.53 “Or.” should be “Ov.”; at 1.8.17 “pall” (representing “pallentibus”) is followed by an unnecessary space and an anomalously large dash; at 1.8.45 the reference should be to Propertius 3.25.13, not 3.25.33; at 2.6.21 there is an unnecessary space after “f”. Occasionally there is a clash between text and apparatus; at 1.2.37 the arrangement of the apparatus implies that Luck reads “Ne” rather than the “Neu” printed in the text, just as it implies at 2.1.87 that he reads the conjecture “nam” rather than “iam.” The most serious new problem introduced in the apparatus is a persistent failure to provide adequate space between different readings or conjectures reported for the same line, but only rarely does this create a situation that is positively misleading. At 4.8.6 ” Postg. versum sic refinxit Heyne” is easily misunderstood until one realizes that “Postg.” pertains to the preceding “semper amice” while the rest belongs to the following entry; usually a little patience is sufficient to sort out the problem, which in itself would be no more than a minor nuisance were it not so prevalent.
Luck’s second edition, then, provides more raw data on the tradition and offers a still broader conspectus of conjectures than the first. Some might criticize this as (to paraphrase Housman) filling his apparatus with the dregs of the Italian Renaissance, but it is appropriate given our current inadequate understanding of the Tibullan tradition (the truth may indeed be out there: we’re just not entirely sure where to look for it) and given that most of the corruptions are not desperate cruces of the sort common in Propertius but rather involve the endings of nouns and verbs or clusters of unstable participles like “ne”, “neu”, “neue”, “num”, “non”, and “nunc” that perpetually mutate into each other in the copying of Latin mss. If there is to be a third edition on schedule in 2008, I have four desiderata for it (or indeed for any other future edition). First, perhaps editors will one day read J.D. Morgan’s demonstration ( CQ 42 [1992] 533-538) that “Molorchus” was really “Molorcus” and so print “Molorceis” at 4.1.12. Second, I hope that editors will cease to impose on Tibullus the two Priapea that are increasingly edited with the Corpus, or at least that, if they do not, they will put them after the Vita with which our manuscripts conclude, not before, and so not distort the form in which the Corpus has been passed down to us. Third, I hope that editors will either retain the division into 3 books given by the oldest mss (“Book 4” is only the invention of an Italian humanist) or else devise an entirely new presentation that clearly distinguishes the components as the autonomous elements they are: Tibullus’ Delia; Tibullus’ Nemesis; the elegies of “Lygdamus”, whom I have proposed ( LCM 18.4 [1993] 51-53) to identify with Messalla’s nephew Ser. Sulpicius Postumius; the Panegyric; the “Garland of Sulpicia”, which I have speculated is a joint effort of Postumius and his sister Sulpicia; and the epigrams of Sulpicia. (The hypothesis recently advanced by Niklas Holzberg in CJ 94 [1998-99] 169-191, that everything from 3.1 to the end is the work of a Tibulli sectator anxious to have his work taken as Tibullus’ youthful efforts, founders on two principal deficiencies: his failure to appreciate fully the difficulties of the transmission, especially with regard to the dubious status of “Book 3,” and the fact that, since nothing from 3.1 on — least of all the epigrams of Sulpicia — really resembles the long rambling elegies of Tibullus, none of it could really have been expected to pass as his work.) Finally, and most importantly, I hope that someone will use the resources provided in this edition to make the thorough investigation of the manuscript tradition that is so badly needed, and that she or he will coordinate the results with those already achieved for the texts of Catullus and Propertius that frequently circulated with it. When one speaks of the Tibullan “paradosis,” one generally means the readings of A and V; but no-one really knows how these mss are related to each other (not to mention the other 190 or so full mss), or whether there was an archetype, or how the Fragmentum Cuiacianum is related to the humanist tradition, or why there are so many agreements between the Florilegium Gallicum and Pontano’s ms. Even if it settled nothing else, such an investigation could rationalize the apparatus criticus; if Puccius’ and Petreius’ collations of Catullus and Propertius are anything to go by, the readings that Luck reports from them come from the same pool of recentiores that he also reports by siglum, and unnecessary duplication could well be avoided.