This welcome and important Oxford Classical Text volume, produced by a learned and industrious veteran of Latin philology, Robert A. Kaster, offers a new critical edition of three prose works by Seneca: De beneficiis, De clementia and Apocolocyntosis. As first argued by M. C. Gertz in his epoch-making edition of De beneficiis and De clementia,[1] and established beyond any reasonable doubt by later scholars,[2] the manuscripts of both treatises derive from an extant archetype, Vatican Pal. Lat. 1547 (N; c. 800). Kaster shares this view and accepts almost in toto the findings of Mazzoli and Malaspina about the stemmatic relationship of the descendants of N. [3] In this respect, his edition of De beneficiis and De clementia is all but revolutionary. However, Kaster has made the commendable effort to collate a number of recentiores and carefully examine the most important printed editions, which has allowed him to backdate several misattributed conjectures. Kaster’s OCT is undoubtedly destined to become the new standard edition of De beneficiis, because it supersedes the last two major critical editions of Seneca’s treatise, by Hosius and by Préchac.[4] On both these editions Kaster (p. xviii) gives an assessment that I fully share: Hosius’ edition lacks independence of judgement and provides an apparatus that is cluttered by derivative readings (Hosius wrongly refused to believe that N was the archetype); Préchac’s edition, based on the right assumption that N is the source of all later witnesses to the text of De beneficiis and De clementia, is marred by impossible conjectures (suffice it to mention <discerni haec duo> dixerim at 1.1.1 and the grotesque nulla <aulla> at 2.1.4) and by equally absurd attempts to defend the paradosis. Kaster’s edition, which wisely dispenses with virtually all of Préchac’s conjectures and incorporates some (more or less) recent fine emendations (for instance Watt’s <non> reddam at 4.10.1 and Lucarini’s dari at 4.35.1), may be regarded, as it were, as an updated and improved version of the aforementioned edition by Gertz, which is by far the most important critical edition of De beneficiis that has been published to date (Gertz not only was the first editor to base his text of De beneficiis and De clementia on the archetype N, but also excelled in ingenuity and critical judgement and was able to solve a great many philological problems suo Marte). Gertz laid the textual foundation on which Kaster’s edition has been built, but the latter has provided us with the most accurate and reliable critical apparatus of De beneficiis produced so far—an apparatus that, where necessary, is supplemented by a helpful Appendix critica (it should be remembered that Gertz did not personally inspect N but drew on a far from flawless collation by R. Kekulé and that he had a very sketchy knowledge of the later textual transmission). Furthermore, Kaster’s text hic illic improves on Gertz’s.
To give an idea of the work done by Kaster on the text of De beneficiis, I offer a comparison between his philological choices and those of Gertz, limited, breuitatis causa, to the first book. In the passages marked with an asterisk Kaster prints the same text as Préchac, in those marked with the symbol + the same text as Hosius (in one or both of his editions) and Préchac; the symbol # denotes passages where Gertz’s text relies on an inaccurate collation of N. Note that many controversial passages are discussed by Kaster in his useful Studies on the text of Seneca’s De Beneficiis (Oxford 2022). Kaster’s text is surely superior to Gertz’s at 1.1.5-6 properantes … compressus; 1.2.3+; 1.3.7 Phrixianis (a sound emendation made independently by Menghi and Lucarini);[5] 1.8.1 offerret (a conjecture proposed in the apparatus by Gertz that Kaster wrongly credits to Buck); 1.9.3-4* abominanda condicio … certissimum; 1.11.5+ quaeque; 1.11.6+# ut [Weidner : aut N] rustico libros, ut; 1.13.2+# non dilutum; 1.13.3+ iudicando quid uinceret; 1.14.1 {qui} … dat<um> (F. Skutsch); 1.15.1+ at sit (Gertz’s at adsit is cacophonous). On the contrary, Kaster’s text seems to me to be inferior to Gertz’s at 1.4.6 ingerere (Madvig’s incidere, accepted by Gertz and Préchac, gives a better clausula); 1.10.2 pudicitia (<im>pudicitia is necessary); 1.15.4+ cum delectant (Gertz’s quae delectant is brilliant). Some further remarks on the first book: at 1.8.2 Lucarini’s <im>parem (accepted by Kaster) deprives in my view the passage of its pointe (Aeschines, by giving himself to Socrates, succeeded in overcoming both the diuitiae and the μεγαλοψυχία of Alcibiades[6]); at 1.10.4 Kaster’s deletion of enim cannot be right (one who has lost a beneficium cannot keep the best part of it); at 1.12.3 I hesitantly suggest quibus momentis quaedam grata aut ingrata sint; at 1.15.2 my text would run ne <ille> [suppl. Gertz] … exaudiit [ego : -diat N : -diuit Préchac; but exaudiu- never occurs in Seneca’s paradosis].
Kaster’s edition of De clementia compares not unfavourably with Malaspina’s recent and valuable Teubner edition,[7] which, as Kaster acknowledges (p. xix), continues to be indispensable due to its extensive and accurate critical apparatus. Although a conservative critic himself, Kaster is in several passages less conservative than Malaspina, and not seldom rightly so. Also in this case, I limit my comparison to the first book. Kaster’s text seems to me to be superior to Malaspina’s at 1.7.1, where Kaster rightly adopts Madvig’s text and interpunction; at 1.8.4, where he prints circa (a conjecture found in some recentiores); at 1.9.1, where he reads with Madvig rei publice <clade> and puts a full stop after mouit; at 1.10.1, where he accepts Gertz’s Sallustios; at 1.13.3 (interpunction); at 1.14.2 (Haase’s deletion is certainly right); at 1.18.2 (in homine<m>), and at 1.25.2 (ille dirus). On the contrary, I regard Malaspina’s text as preferable at 1.11.1 (deduxerit eius). In the first book I disagree with both Kaster and Malaspina at 1.1.5, where Wesenberg’s innocentiae (not even recorded in Kaster’s apparatus) should in my opinion stand in the text (cf. Sen. epist. 114.7); at 1.6.1, where I would accept Baehrens’ creatur instead of the transmitted aratur, and at 1.20.1, where Bentley’s deletion of si seems necessary (cf. 1.21.1; 1.22.1). At 1.20.2 I would amend the paradosis as follows: et ut appareat non minorem agi rem periclitantis quam iudicis scire [ego : sciat N], “and so that it is clear that he realizes that the interests of the accused are no less at stake than those of the judge” (note that scire gives an excellent clausula; for the syntax, see Sen. epist. 114.7: corrupit … apparet enim mollem fuisse). At 1.25.3 the text should read, in my opinion, auersatio [Wesenberg probante Kaster : aeuersio N], o<dia, insi>diae [ipse suppleui].
As for the Apocolocyntosis, although Kaster has done valuable work by collating thirteen recentiores and tracing back to them some conjectures hitherto ascribed to modern scholars, he has missed the opportunity to thoroughly re-examine the textual transmission in its entirety. Kaster, in fact, too confidently accepts the findings and the stemma codicum of Eden and Roncali,[8] which, however, are not unquestionable. For example, as lucidly shown by Stagni in an important article apparently unknown to Kaster, it is far from certain that s (the lost ancestor of a group of recentiores) depended on S (St. Gall 569; s. IXex).[9] As regards the constitutio textus, to provide an appraisal of Kaster’s performance, I append a spicilegium of remarks on textual problems in the Apocolocyntosis, in which Kaster’s edition is systematically compared not only with the last major critical edition of the pamphlet (Roncali’s Teubner), but also with Schmeling’s Loeb and Vannini’s incisive revision of Roncali’s text (both Vannini’s and Schmeling’s work are completely overlooked by Kaster): [10]
Apoc. 2.1.1: De Nonno’s actum, accepted by Vannini and Schmeling, is far better than Froidmont’s orbem, printed by Kaster; pace Roncali, the transmitted ortum cannot be retained.
Apoc. 4.1.20: laudatum is suspect, but Maehly’s laudantum is impossible, because, as Kaster observes in the apparatus, “noster tantum forma -tium gen. plur. participii praesentis utitur”. Perhaps laudatum is a Perseverationsfehler (cf. 4.1.18 laudant) and Seneca wrote inceptum.
Apoc. 5.2: nuntiatur, retained by Kaster (as well as by Roncali, Vannini and Schmeling), is intolerable because it cannot govern quaesisse se and intellegere se (se clearly shows that a personal agent is involved). To the conjectures mentioned in Kaster’s apparatus I add one of my own: nuntiant (nuntiãt > nuntiat’).
Apoc. 5.3: Kaster is rightly baffled by ut qui etiam … timuerit (printed without alterations by Roncali, Vannini and Schmeling) and obelizes etiam. In his apparatus, however, he does not mention the best attempt at emendation proposed so far: Baehrens’ ut qui (“because”) uicta for the transmitted ut qui etiam. I suggest ut qui <d>omita (otiam > etiam); for domita … monstra cf. Sen. Herc. fur. 33, 444; for the accusative and a periphrastic form of infinitive (esse being omitted) after timeo cf. Liv. 10.36.3 and Oakley 2007, ad locum. Another possibility is ut qui euicta.
Apoc. 6.1: Munati, a conjecture by Beatus Rhenanus accepted by Kaster (and Schmeling), is in my opinion preferable to the transmitted Marci, retained by Vannini and Roncali, who interprets it not as a personal name, but as marci (marcus is a Gallic word indicating a kind of vine).
Apoc. 6.1: Licinus, a good conjecture by Bücheler and a felix error by the copyist of a recentior, is rightly accepted by Kaster, as well as by Roncali, Vannini and Schmeling; by mistake, however, Kaster’s apparatus does not provide the reading of the archetype (Licinius).
Apoc. 6.2: iusserat illi collum praecidi can be retained only if iusserat is taken with illo gestu eqs.; standing alone, as in Kaster’s text (as well as in Roncali’s and Vannini’s), it reads “like a botched interpolation” (Eden 1984, ad locum). However, a conjunction is needed to separate illo gestu from iubebat, which of course implies murmure (see the beginning of the paragraph). Eden suggests iubebat <atque> illo eqs.; I propose iubebat <et> illo gestu solutae manus, {et} ad hoc unum satis firmae, quo … solebat, iusserat illi collum praecidi.
Apoc. 7.4: Bücheler’s Tib<ur>i seems, pace Kaster and Schmeling, necessary; it is accepted by Roncali and Vannini (see also Mariotti 2000, 516).
Apoc. 7.5: the transmitted contulerim, retained by Roncali and Schmeling, is unacceptable, but the cacophonous tum tulerim by Richmond (put in the text by Kaster) cannot be what Seneca wrote. Vannini adopts Haase’s {con}tulerim. The best solution seems to me Mariotti’s <e>go tulerim (cf. Mariotti 2000, 516-517; Traina 2003, 329).
Apoc. 8.3: Kaster prints parum est quod templum in Britannia habet, quo nunc barbari colunt ut deum et orant μωροῦ εὐιλάτου τυχεῖν. From his apparatus one gathers e silentio that quo nunc is the transmitted reading, but the paradosis is actually quod nunc. This error aside, Kaster’s text is unconvincing. Kaster favours Ussing’s conjecture ut deum et for the transmitted (and solecistic) et ut deum (retained by Roncali, Vannini and Schmeling); I would read, with Mariotti 2000, 520-522, templum in Britannia habet, quod hunc [Lipsius] barbari colunt et {ut deum} [del. Mariotti] orant μωροῦ εὐιλάτου τυχεῖν.
Apoc. 9.1: di<cere non de>cere nec disputare, conjectured independently by Ussani and Vannini,[11] is worth mentioning in the apparatus.
Apoc. 9.3: Kaster deservedly revives Helm’s brilliant conjecture quas alit.
Apoc. 9.6: sententiam uincere is hardly Latin; Vannini’s {sentent}iam uincere,[12] accepted by Schmeling, is worthy of a place in the critical apparatus.
Apoc. 10.3: soror mea is ascribed to Bücheler by Kaster, but according to Roncali it was already in s.
Apoc. 12.1: Kaster is probably right to accept Lipsius’ concentus (the manuscript reading, conuentus, is retained by Roncali, Vannini and Schmeling).
Apoc. 12.2: cum maxime for tum maxime, Kaster’s only conjecture in the Apocolocyntosis, may be right but—for obvious reasons—is far from certain.
Apoc. 13.2: Kaster adopts Bücheler’s ‘celerius’ inquit Mercurius ‘i et eqs.’, but ‘celerius i’ inquit Mercurius ‘et eqs.’, the reading of one of the main witnesses to the text of the Apocolocyntosis (L = British Library, Add. 11983, s. XIIin), seems better to me.
Apoc. 14.4: alicuius cupiditatis spem sine effectu is in my view intolerable and cannot mean “a hope of gratifying some desire without result” (so Schmeling). I propose alicuius cupiditatis s<atiandae s>pem sine effectu.
Apoc. 15.2: Kaster may be right to read with Maehly ab ipso instead of the transmitted ab illo, put in the text by Roncali, Vannini and Schmeling.
All in all, Kaster’s edition of the Apocolocyntosis compares favourably with Roncali’s edition because the latter, although still useful due to its extensive apparatus, is too conservative and sometimes marred by veritable Korruptelenkult.[13] As we have seen, however, Kaster’s edition is neither particularly innovative nor always satisfactory in its choice of readings and emendations.
There are some misprints and minor errors in this OCT volume; they are mostly harmless (see, for example, p. vii, l. 12: “I should two add notes”; p. xl, l. 9: “delle tradizione”; ll. 28–29 “manoscrita … Academia”); note, however, that both in the preface and in the conspectus siglorum (pp. xxii; 263) the shelfmark of S (St. Gall 569) is misgiven. In the text of Apoc. 10.4 retulit should be corrected to rettulit.
Between the text and the critical apparatus of all three works edited by Kaster stands a helpful apparatus fontium et testium.
Works cited
Eden, P. T. 1979, The Manuscript Tradition of Seneca’s Apocolocyntosis, CQ 29, 149-161.
Eden, P. T. (ed.) 1984, Seneca, Apocolocyntosis, Cambridge.
Gertz, M. C. (ed.) 1876, L. Annaei Senecae Libri De beneficiis et De clementia, Berlin.
Hosius, C. (ed.) 1900, L. Annaei Senecae De beneficiis libri VII; De clementia libri II, Leipzig (21914).
Lucarini, C. M. 2021, Coniectanea (IV), Philologia antiqua 16, 370-387.
Madvig, J. N. 1873, Adversaria critica, II, Copenhagen.
Malaspina, E. (ed.) 2001a, L. Annaei Senecae De clementia libri duo, Alessandria.
Malaspina, E. 2001b, La ‘preistoria’ della tradizione recenziore del De clementia (a proposito di Paris, Bibl. nat., lat. 15085 e Leipzig, Stadtbibl., Rep. I, 4, 47), RHT 31, 147-165.
Malaspina, E. (ed.) 2016, L. Annaeus Seneca, De clementia libri duo, Berlin & Boston.
Mariotti, S. 2000, Scritti di filologia classica, Roma.
Mazzoli, G. 1982, Ricerche sulla tradizione medievale del De beneficiis e del De clementia di Seneca. III. Storia della tradizione manoscritta, BollClass 3,165-223.
Menghi, M. (ed.) 2008, Seneca, Sui benefici, Roma & Bari.
Oakley, S. P. 2007, A Commentary on Livy, Books VI-X. Vol. IV: Book X, Oxford.
Préchac, F. (ed.) 1926-1927, Sénèque, Des bienfaits, Paris, 2 vols.
Reynolds, L. D. 1983, The Younger Seneca, De beneficiis and De clementia, in Texts and Transmission: A Survey of the Latin Classics, ed. by L. D. Reynolds, Oxford, 363-365.
Roncali, R. (ed.) 1990, L. Annaei Senecae Ἀποκολοκύντωσις, Leipzig.
Schmeling, G. (ed.) 2020, Petronius, Satyricon. Seneca, Apocolocyntosis, Cambridge, MA.
Stagni, E. 1994, Ἀποκολοκύντωσις. Appunti sulla tradizione di Dione Cassio – Xifilino, RFIC 122, 298-339.
Traina, A. 2003, La lyra e la libra, Bologna.
Ussani, V. 1913, Sul Ludus de morte Claudii, RFIC 41, 74-80.
Vannini, G. (ed.) 2008a, Seneca, Apokolokyntosis, Milano.
Vannini, G. 2008b, Tre note per il testo dell’Apokolokyntosis, Philologus 152, 166-171.
Notes
[1] Gertz 1876, vi-vii. Gertz was set on the right path by his great teacher J. N. Madvig (Madvig 1873, 406–407).
[2] For the bibliography, see Reynolds 1983, 363 n. 1.
[3] Mazzoli 1982; Malaspina 2001a; 2001b; 2016. See Kaster’s Preface, v-xv.
[4] Hosius 1900 (2nd ed. 1914); Préchac 1926-1927.
[5] Menghi 2008 ad locum, unknown to Lucarini 2021, 371, to whom Kaster attributes the emendation.
[6] On Alcibiades’ magnitudo animi, see Nep. Alcib. 1.
[7] Malaspina 2016.
[8] Eden 1979; 1984; Roncali 1990, Praefatio.
[9] Stagni 1994, 303-304 n. 2.
[10] Roncali 1990; Vannini 2008a; Schmeling 2020.
[11] Ussani 1913, 74; Vannini 2008b, 166–168.
[12] Vannini 2008a, 168–170.
[13] The most conspicuous cases are 8.2 illum deum (the best conjecture proposed so far is illud, nedum by J. F. Gronovius); 8.3 corrigit (read, with Neubur, corriget), and 9.2 quantum uia sua fert (read, with Beatus Rhenanus, quantumuis uafer et). In all these passages Kaster and Vannini have made the right choice (not so, in my opinion, Schmeling).