¹ Kenney, E.J. (ed.), Lucretius: De Rerum Natura Book III (Cambridge, 1984), 107Google Scholar. I would like to thank Professor John T. Kirby and CQ's anonymous reader for comments on earlier versions of this essay. Professor Bruce Gibson also made several significant improvements to the piece, guiding it through the editorial process with skill and care. Charles Murgia, who died as this essay was in its final stages of revision, read an early draft and critiqued it with characteristically refreshing frankness. Wrestling with his challenges inspired significant revisions, although I feel certain that he would remain unconvinced by some of my suggestions. I dedicate the piece to his memory.

² Here Müller, K. (ed.), T. Lucreti Cari De Rerum Natura Libri Sex (Zurich, 1975), 102Google Scholar chooses the obelus; Kenney (n. 1), 44 and Bailey, C. (ed. and tr.), T. Lucreti Cari: De Rerum Natura, Libri Sex (Oxford, 1947)Google Scholar, 1.314 accept Frerichs' emendation, which was first published at Frerichs, H., Quaestiones Lucretianae (Oldenburg, 1892), 14Google Scholar; Rouse, W.H.D. (ed. and tr.), Lucretius: De Rerum Natura, rev. Smith, M.F. (Cambridge, MA, 1992), 206Google Scholar prints Saunders' conjecture et quaecumque ipsa volutat, which was first aired in Saunders, T.J., ‘A note on Lucretius III 240’, Mnemosyne 28 (1975), 296–8CrossRefGoogle Scholar.

³ Bailey (n. 2), 1.314, Kenney (n. 1), 107, and A. Ernout (ed. and tr.) Lucrèce, De la nature, rev. C. Rambaux (Paris, 1978), 1.94, among others, retain the transmitted text; the chief exception (see n. 5 below) is Heinze, R. (ed.), T. Lucretius Carus: De rerum natura Buch III (Leipzig, 1897), 7Google Scholar.

⁴ I suggest below that mente of line 240 enters the text under the influence of mens in 239. The most popular emendation for mens (239) is Bernays' res (published at J. Bernays [ed.], T. Lucreti Cari De Rerum Natura Libri Sex [Leipzig, 1852], 68 and accepted by H.A.J. Munro [ed. and tr.], T. Lucreti Cari De Rerum Natura Libri Sex [Cambridge, 1864], 1.114 as well as in Munro's subsequent editions). This emendation yields the phrase recipit res (rendered by Munro as ‘the fact of the case does [not] admit’), but as Heinze (n. 3), 81 points out, ‘recipere kann in dieser Bedeutung nur von dem gesagt werden, der an sich etwas geschehen lässt’.

⁵ Heinze (n. 3), 81 is the commentator who has seen this point most clearly. Although he puts forward no solution himself (calling the lines ‘korrupt und noch nicht geheilt’), he effectively dispenses with the major proposals of the nineteenth century, including those of Lachmann, K. (ed.), T. Lucreti Cari De Rerum Natura Libri Sex (Berlin, 1850), 1.87Google Scholar (recipit quem) and Bernays (n. 4), 68 (recipit res).

⁶ Bailey (n. 2), 1.315.

⁷ Rouse (n. 2), 207.

⁸ Ernout (n. 3), 1.94.

⁹ A partial exception is Giussani, C. (ed.), T. Lucreti Cari Libri Sex De rerum natura. Volume Terzo: Libro III e IV (Turin, 1897), 29–30Google Scholar, who does not transfer the negative to the main verb, writing instead: ‘Di nessuna delle quali sostanze la mente può capacitarsi …’ But his rendering of recipit as ‘può capacitarsi’ is not exactly literal.

¹⁰ I discuss below recipit in our passage, which is defined at OLD ² s.v. 7b. It is worth pointing out initially, though, that the larger entry under which (OLD ² s.v. 7) the present use of recipit appears is ‘To be willing to take, accept’, which does not prima facie seem at odds with ‘to accept (as one of several possibilities)’, vel sim.

¹¹ Several German critics have noticed this problem. See e.g. Goebel, P.E., Quaestiones Lucretianae Criticae (Salzberg, 1857), 24Google Scholar: nam recipere mentem ea tantum quae ratio admittat recte Lachm(annus) negavit. (cf. Lachmann [n. 5], 2.153) The mind may ‘allow’ more than reason alone permits.

¹² Here I pass over another question raised by several renderings. Translators often seem not only to negate recipit but also to give it an opposite sense, a phenomenon sometimes called litotes, sometimes οὐ-adherescent in Greek. (See e.g. R. Kühner, Ausführliche Grammatik der griechischen Sprache. Zweiter Teil: Satzlehre, rev. B. Gerth [Hanover, 1904], 2.180 [§510 n. 3] and H.W. Smyth, Greek Grammar [Cambridge, MA, 1956], 610–1 [§§2691–7].) Ernout (n. 3), 1.94, for example, renders recipit with ‘refuse d'admettre’ rather than ‘n'admet pas’. While this question is largely rhetorical (rather than purely grammatical or syntactical), it is unclear that recipit (or non recipit) here permits such a translation.

¹³ For the pleonachos tropos, see Epicurus, Ep. Pyth. 85–8 and Lucr. 5.509–33, with Long, A.A. and Sedley, D., The Hellenistic Philosophers (Cambridge, 1987), 1.95–6Google Scholar. For the tradition of multiple explanation in Lucretius and later Latin poetry, see Hardie, P., Lucretian Receptions: History, The Sublime, Knowledge (Cambridge, 2009), 231–63Google Scholar.

¹⁴ See also Lucr. 6.703–11, where the poet asserts that if you should see a dead man from a distance, you would not know whether he died by disease, poison, sword or cold. Presumably, these various causes of death are mutually exclusive. ‘Accepting’ one of them as ‘true’ (rather than merely possible or credible) would preclude the others.

¹⁵ It is perhaps worth pointing out that, according to Epicurean theory, it is precisely the mind (rather than the senses) that introduces error. See e.g. Lucr. 4.386.

¹⁶ Krebs, J.P., Antibarbarus der lateinischen Sprache. Zweiter Band, rev. J.H. Schmalz (Basel, 1907), 480Google Scholar (s.v. recipere 2) asserts that recipere in the construction recipere alicuius sententiam, with the sense ‘jemandes Meinung annehmen’ is only a late Latin usage, citing Sulp. Sev. Chron. 2.39.1 in support of this claim. If this assertion is true, it would be striking to see mens recipit mean something as strong as ‘the mind allows only that’ or ‘the mind adopts the position that’ already in Lucretius.

¹⁷ These words are almost entirely supplemented from the text of August. De civ. D. 22.6.10–16. The Vatican palimpsest (Vat. lat. 5757) only preserves the ‘r’ of recepit. See the apparatus of Powell, J.G.F. (ed.), M. Tulli Ciceronis De re publica, De legibus, Cato Maior De senectute, Laelius De amicitia (Oxford, 2006), 61Google Scholar.

¹⁸ At least one corrector also writes percipi. See Croiselle, J.-M. (ed.), Pline L'Ancien, Histoire Naturelle, Livre XXXV (Paris, 1985), 61Google Scholar. Given the frequent variation in manuscripts between e and ae, we may explain recipi as resulting merely from the omission of the initial p.

¹⁹ quoniam dixi de partibus, in quas omnis rerum naturae materia dividitur, quaedam in commune sunt dicenda, et hoc primum praesumendum, inter ea corpora, a quibus unitas est, aera esse. quid sit hoc et quare praecipiendumfuerit, scies si paulo altius repetiero, et dixero aliquid esse continuum, aliquid commissum. Hine, H.M. (tr.), Seneca. Natural Questions (Chicago, 2010), 164Google Scholar renders the gerundive praecipiendum with ‘needs to be established at the outset’. Cf. Lucr. 5.533.

²⁰ For the infallibility of the senses in Epicureanism, see Lucr. 4.469–521 and Diog. Laert. 10.31.

²¹ Heinze (n. 3), 81: ‘Das überlieferte mens recipit könnte verstanden werden “der Geist nimmt an, d. h. billigt, hält für richtig”: aber statt dessen müsste man, da der Geist ja doch haüfig genug auch Falsches für richtig hält, mindestens erwarten “kann nicht billigen”.’

²² It is worth pointing out that when recipit governs, as it does in 3.239–40, an accusative and infinitive construction, its meaning is more commonly ‘promises’ or ‘gives assurance that’ rather than ‘admits’ (and it typically takes a future infinitive rather than the present infinitive found in Lucretius); see e.g. Cic. Phil. 5.51: promitto, recipio, spondeo, patres conscripti, C. Caesarem talem semper fore civem qualis hodie sit qualemque eum maxime velle esse et optare debemus. See further OLD ² s.v. 10b and M. Leumann, J.B. Hofmann and A. Szantyr, Lateinische Grammatik. Band, Zweiter: Syntax und Stilistik (Munich, 1965), 356Google Scholar (§195.A.d), where verbs of permission (Erlauben) that govern an accusative and infinitive construction are listed. There is no mention of recipere.

²³ It seems this conclusion would hold even if e.g. OLD ² distinguishes too finely between the entries at s.v. 7b and those at s.v. 9, where the definition provided is ‘to admit of, give scope for, be subject to, permit’. The passages cited at 9 almost uniformly present a negative (or virtual negative such as vix) before recipere, with the exception of passages such as Caes. B Civ. 3.51.5, where fortasse suggests that the sense of recipere is not ‘allows only’ but ‘allows as a possibility’.

²⁴ See Lucr. 2.111, 2.993, 5.320, 5.323, 6.23, 6.146, 6.574.

²⁵ Confusion between recipere and percipere is also found at e.g. Sen. Q Nat. 3.11.5.

²⁶ Cicero similarly uses percipere of learning. See e.g. De or. 1.159, 1.252.

²⁷ See Lucr. 3.135, 4.111, 4.115, 4.270, 4.723, 4.880, 6.46, 6.536 for the other instances of this imperative.

²⁸ These lines virtually repeat 1.948–50. There, however, perspicis replaces percipis and qua constet compta figura stands in place of ac persentis utilitatem.

²⁹ In addition to learning, percipere also describes several other ‘grasping’ activities in Lucretius. At 3.79–80 the poet writes of how hatred of life ‘seizes’ humans: vitae | percipit humanos odium lucisque videndae. Similarly, at 3.28–9 he describes how ‘a quasi-divine pleasure and awe seize’ him at the discoveries of Epicurus: me … quaedam divina voluptas | percipit atque horror … Elsewhere percipio (4.729) names the action of simulacra striking our eyes. Other uses characterize how the sun's fire ‘seizes’ the air (5.605), fever ‘seizes’ limbs (6.804), or the various senses ‘grasp’ (6.985). In total, there are – if we do not include the current passage (where I argue for percepit) – eighteen instances of percipio in Lucretius.

³⁰ Cicero also writes of the animus ‘grasping’ its own nature (Tusc. 1.47): … cum autem nihil erit praeter animum, nulla res obiecta impediet, quo minus percipiat, quale quidque sit.

³¹ On ante oculos in Lucretius, see D. Fowler, Lucretius on Atomic Motion: A Commentary on De rerum natura 2.1–332 (Oxford, 2002), 197 on 2.113.

³² For other early Latin instances of percipio governing an accusative and infinitive, cf. e.g. Lucil. 30.1012 (Marx): et sua percipere<t> retro rel<l>icta iacere …; Plaut. Asin. 162, Curc. 159.

³³ Leumann, Hofmann and Szantyr (n. 21), 357 (§195.B): ‘Die klass. Sprache verschmäht die altlat. A[kkusativ]. c[um]. I[infinitiv]. nach percipio …’

³⁴ See Lindsay, W.M., Notae Latinae (Cambridge, 1915), 244Google Scholar (quia) and 263 (quoniam). The manuscript O in fact shows quoniam in the abbreviated form qm at (e.g.) 1.362 (fol. 10r.), suggesting that the proposed error would not be impossible from what we know of the transmission of Lucretius' text. Similarly, Q shows qm at 1.32 and qa with a superscript i above the q at 1.993.

³⁵ Confusion between P and R is found elsewhere in the manuscripts of Lucretius. See e.g. Lucr. 2.843: tepopis OQ, where the true reading is clearly teporis.

³⁶ Bailey (n. 2), 112. See also Kenney (n. 1), 99 (on 3.174), 109 (on 3.258).

³⁷ Bailey (n. 2), 2.1029–30 (ad loc.).

³⁸ For Munro's conjecture, see Munro (n. 4), 1.114; for Postgate's, see Postgate, J.P., ‘Lucretiana’, Journ. Phil. 24 (1896), 131–47Google Scholar, at 137 (with the remarks of Giussani, C., ‘Bibliografia’, Riv. Fil. 24 [1896], 97–115Google Scholar, at 112, Giussani [n. 9], 29–30, and Brieger, A., ‘Bericht über die Lucrez-Litteratur, die Jahre 1890–1895 umfassend’, Jahresb. 89.2 [1897], 120–205Google Scholar, at 142); for Goebel's, see Goebel (n. 11), 24; for Frerichs', see Frerichs (n. 2), 14.

³⁹ For Saunders' conjecture, see Saunders (n. 2); for Purmann's, see Richter, W., Textstudien zu Lucrez (Munich, 1974), 40–2Google Scholar (without acknowledgment) or Eichner, M., Annotationes ad Lucretii Epicuri interpretis de animae natura doctrinam (Sorau, 1884), 14Google Scholar.

⁴⁰ For Eichner's proposal, see Eichner (n. 39), 16; for Lambin's, see Lambinus, D., Titi Lucreti Cari De rerum natura libri sex (Frankfurt, 1583), 315Google Scholar; for Lachmann's, see Lachmann (n. 5), 1.87.

⁴¹ See also Lucr. 3.168–76, where physical wounds and blows impact on the mind.

⁴² See Aët. 4.8.10. Cf. J. Annas, Hellenistic Philosophy of Mind (Berkeley, 1992), 165. Epicurus, Nat. 28 12.VI.2–VII.9 (Sedley), while it mentions imagistic belief, may also describe mental activity that is not imagistic. See Atherton, C., ‘Epicurean philosophy of language’, in Warren, J. (ed.), The Cambridge Companion to Epiureanism (Cambridge, 2009), 197–215CrossRefGoogle Scholar, at 204.

⁴³ Heinze (n. 3), 81.

⁴⁴ For the soul as mixture, see e.g. Lucr. 3.258, where he speaks of ea (i.e. the four elements of the soul) quo pacto inter sese mixta and Aët. 4.3.11, discussed below.

⁴⁵ For Plutarch's statement, see Adv. Col. 1118E: τὸ γὰρ ᾧ κρίνει καὶ μνημονεύει καὶ φιλεῖ καὶ μισεῖ, καὶ ὅλως τὸ φρόνιμον καὶ λογιστικὸν ἔκ τινός φησιν (sc. Κωλώτης) ‘ἀκατονομάστου’ ποιότητος ἐπιγίνεσθαι. See Bailey (n. 2), 2.1027 for a summary of several positions on the relationship between the animus and the fourth nature. A clear statement of the view that the animus and the fourth nature are identical is Eichner (n. 39). Boyancé, P., ‘La théorie de l’âme chez Lucrèce’, in Classen, C.J. (ed.), Probleme der Lukrezforschung (Hildesheim, 1986), 131–50Google Scholar, at 138 (an essay first published in 1958) asserts by contrast that the anima contains the fourth element as well. Konstan, D., A Life Worthy of the Gods: The Materialist Psychology of Epicurus (Las Vegas, 2008)Google Scholar, 10 n. 11 suspends judgement on this point.

⁴⁶ Scholion on Ep. Hdt. 66: λέγει (sc. Ἐπίκουρος) ἐν ἄλλοις καὶ ἐξ ἀτόμων αὐτὴν συγκεῖσθαι λειοτάτων καὶ στογγυλωτάτων, πολλῷ τινι διαφερουσῶν τῶν τοῦ πυρός· καὶ τὸ μέν τι ἄλογον αὐτῆς, ὃ τῷ λοιπῷ παρεσπάρθαι σώματι· τὸ γὰρ λογικὸν ἐν τῷ θώρακι, ὡς δῆλον ἔκ τε τῶν φόβων καὶ τῆς χαρᾶς. See Long and Sedley (n. 13), 2.74 for discussion of the possibility that this passage describes the nameless fourth element.

⁴⁷ Heinze (n. 3), 81. Here πάθη represent not so much ‘emotions’ as ‘irrational affective sensations’ that offer a basis for judgement.

⁴⁸ Konstan (n. 45), 1–25 is an extensive discussion of Epicurean πάθη.

⁴⁹ See Diog. Laert. 10.31: ἐν τοίνυν τῷ Κανόνι λέγων ἐστιν ὁ Ἐπίκουρος κριτήρια τῆς ἀληθείας εἶναι τὰς αἰσθήσεις καὶ προλήψεις καὶ τὰ πάθη …

⁵⁰ See, too, Ep. Hdt. 37–8, 68, 82; RS 24. It also seems possible that Epicurus fr. 29.27.1 explicitly links the four elements of the soul to the πάθη.

⁵¹ This fact means, of course, that volutat is in this instance the lectio difficilior, but both volutat and voluptas are common enough in early Latin that sense rather than sheer statistical frequency should determine the reading. Indeed, that volutat occurs only here perhaps suggests that it is un-Lucretian in style. An apparent parallel to this line occurs at Stat. Achil. 1.198–200: At Thetis … | quae nato secreta velit, quibus abdere terris | destinet, huc illuc divisa mente volutat. While such phrasing, with its close resemblance to Lucretius' line (as transmitted), may suggest that Statius had Lucretius before him, divisa mente volutat is also strongly reminiscent of Homeric turns of phrase such as διάνδιχα μερμήριξεν (Il. 1.189). It has often been suggested that Statius spoke (and read) Greek, so a direct Homeric borrowing (rather than a Lucretian one) is not unlikely. On the Greek character of Naples, where Statius was born, and Statius' connection to the city, see Hardie, A., Statius and the Silvae: Poets, Patrons and Epideixis in the Graeco-Roman World (Liverpool, 1983), 2–14Google Scholar. On Statius' connection with the Greek language, see especially Holford-Strevens, L., ‘In search of Poplios Papinios Statios’, Hermathena 168 (2000), 39–54Google Scholar, who argues for understanding Statius as a Greek poet working in Latin. See also Marshall, A.R., ‘Allusion and meaning in Statius: five notes on Silvae 1’, Mnemosyne 61 (2008), 601–18CrossRefGoogle Scholar, at 611 n. 17, whose focus is upon bilingual word play (between Greek and Latin) in Statius' Silvae.

⁵² See Lucr. 1.1, 1.140, 2.3, 2.172, 2.258, 2.966, 3.28, 3.1081, 4.1075, 4.1081, 4.1085, 4.1201, 4.1207, 4.1263, 5.178, 5.1433, 6.94. (voluptas occurs twenty-four times in Lucr., not counting the present passage, where I believe it lurks in corrupted form.)

⁵³ For a similar construction, see Lucr. 5.948–50: denique nota vagis silvestria templa tenebant | nympharum, quibus escibant umore fluenta | lubrica proluvie larga lavere umida saxa … Cf. Lucr. 3.375, 839, 865.

⁵⁴ Following Munro (n. 4), Flores, E. (ed.), Titus Lucretius Carus: De rerum natura. Edizione critica con introduzione e versione. Volume primo (Libri I–III) (Naples, 2002), 250Google Scholar has suggested that quaedam(que) of line 240 originates in quaedam of line 241.

⁵⁵ It is also possible that e fit was misread as est, and the resulting word was abbreviated and then lost.

⁵⁶ On this textual problem, see Ingleheart, J., A Commentary on Ovid, Tristia, Book 2 (Oxford, 2010)Google Scholar, 291 and Luck, G. (ed.), P. Ovidius Naso. Tristia (Heidelberg, 1967–77)Google Scholar, 2.132 (who retains honesta voluntas but punctuates as follows: nec liber indicium est animi, sed honesta voluntas | plurima mulcendis auribus apta feret.). Cf. Liv. 21.4.6 and Plin. Ep. 2.17.24, two additional passages with manuscript variation between voluntas and voluptas where voluptas is generally held to be the true reading.

⁵⁷ See Hall, F.W., A Companion to Classical Texts (Oxford, 1913), 155–6Google Scholar, who comments that ‘it is the general similarity between two words rather than the similarity between the one or two letters in which they differ that has brought about the confusion between them’.

⁵⁸ Cf. Plin. Ep. 3.8.4: quare ego vero honestissimae voluntati tuaepareo.

⁵⁹ Lucr. 2.257–8, two lines from the poet's famous description of the clinamen, may also provide evidence of related corruptions. As printed in most editions (with the emendations of Denys Lambin), those lines show voluptas misread as voluntas, although this corruption seems due not simply to a confusion of two common words but also of two adjacent line endings.

⁶⁰ Konstan (n. 45), 8–9. It is worth pointing out that one of the chief texts for Konstan's argument comes from earlier in Lucretius' third book, Lucr. 3.136–51.