¹ The only explicit dissenter from the communis opinio that Cicero is in some way denigrating Sallustius’ poem is Hamblenne, P., ‘Au Salluste inconnu’, RBPh 59 (1981), 60–70CrossRefGoogle Scholar, at 67–70, who takes the effort implied as reflecting the Empedoclea’s sophistication rather than its turgidity. Comparing the antithesis of uir and homo in a literary context at Att. 2.2.2 = SB 22, he renders the sentence ‘C'est pour un surdoué que je te prendrai si tu as lu les Empedoclea de Salluste; je ne te prendrai pas pour un homme (ordinaire)’.

² Housman, A.E., ‘Notes on Martial’, CQ 13 (1919), 68–80CrossRefGoogle Scholar, at 72 = Diggle, J. and Goodyear, F.D.R. (edd.), The Classical Papers of A.E. Housman (Cambridge, 1972)Google Scholar, 3.982–95, at 986.

³ Neither Housman nor other critics note that he is partially anticipated by Munro, H.A.J. (ed.), T. Lucreti Cari De rerum natura libri sex (Cambridge, 1866 ²), 329–30Google Scholar (‘The sentence seems to me clearly to require something to be joined with virum te putabo, in order to contrast with si Salustii cet.’), but instead of Housman's lacuna, Munro emends to multae tamen artis esse cum inueneris, uirum te putabo.

⁴ e.g. Watt, W.S., Review of Nuuminen, P., Quo modo Cicero de Lucretio (et quodam Sallustio) iudicaverit (Turku, 1953)Google Scholar, CR 5 (1955), 209–10Google Scholar, at 210, who criticizes Nuuminen for not mentioning ‘the possibility that Cicero intended to make a new point, in which hominem is contrasted not with virum but with pecudem or bestiam’. Cf. Vesey, W.T., ‘Virum te putabo, hominem non putabo’, CR 42 (1928), 111–12.Google Scholar

⁵ I am grateful to CQ’s anonymous reader, whose scepticism and reference to the electronic resources unavailable to Housman prompted me to verify and prove this point.

⁶ It should be acknowledged that in all these examples the adjectives produce part of the antithesis, but it is hardly coincidental that the pejorative or milder adjective always qualifies homo, the positive or more effusive one uir.

⁷ For this use of omnino (OLD s.v. 4) in Cicero, cf. de hominum genere aut omnino de animalium loquor, Fin. 5.33.

⁸ D.R. Shackleton Bailey, Review of Diggle and Goodyear (n. 2 above), Cambridge Review 94 (1973), 189–90Google Scholar, at 190, quoted in id. (ed.), Cicero: Epistulae ad Quintum Fratrem et M. Brutum (Cambridge, 1980)Google Scholar, 191; id. (ed.), Cicero: Select Letters (Cambridge, 1980)Google Scholar, 150.

⁹ On Ciceronian humour, see esp. Haury, A., L'ironie et l'humour chez Cicéron (Leiden, 1955)Google Scholar and Corbeill, A., Controlling Laughter: Political Humor in the Late Republic (Princeton, 1996)Google Scholar. In the letters: Hutchinson, G.O., Ciceronian Correspondence: A Literary Study (Oxford, 1998), 172–99Google Scholar.

¹⁰ White, P., Cicero in Letters: Epistolary Relations of the Late Republic (Oxford, 2010), 99–104CrossRefGoogle Scholar, shows how surprisingly little Cicero discusses contemporary literature at any length or in any depth in his letters, but the following examples show that he still tends to be witty and pointed, even in short compass.

¹¹ On wordplay in Roman comedy, see now Fontaine, M., Funny Words in Plautine Comedy (Oxford, 2010)Google Scholar.

¹² Hollis, A.S., Fragments of Roman Poetry c. 60 b.c.–a.d. 20 (Oxford, 2007)Google Scholar, 80. eumque addictum iam tum puto esse Calui Licini Hipponacteo praeconio (‘I think he was long since sold at auction by Licinius Calvus’ Hipponactean auction notice’), Fam. 7.24.1 = SB 260. The first line of the lampoon (Calv. poet. fr. 3 FPL/FLP = fr. 36 FRP) is preserved by Porph. ad Hor. S. 1.3.1: Sardi Tigelli putidum caput uenit (‘The fetid person of Tigellius the Sardinian is for sale’).

¹³ sed quaero locupletem tabellarium, ne accidat quod Erigonae tuae, cui soli Caesare imperatore iter ex Gallia tutum non fuit. quid si canem tam bonam non haberet? (‘But I'm asking for a reliable courier, so that the same thing doesn't happen as did to your Erigona, for whom alone the journey from Gaul wasn't safe while Caesar was in command. What if she hadn't had such a good dog?’), Q. fr. 3.7(9).6–7 = SB 27. Since Accius’ Erigona certainly and Sophocles’ Erigone probably dealt with the daughter of Aegisthus and Clytemnestra rather than of Icarius, we might wonder whether Quintus’ tragedy was on the same theme. Marcus may be privileging humour over accuracy, or there may be a further joke in the deliberate misidentification of a play which he never had the opportunity to read.

¹⁴ Griffin, M.T., ‘Philosophical badinage in Cicero's Letters’, in Powell, J.G.F. (ed.), Cicero the Philosopher (Oxford, 1995), 325–46Google Scholar.

¹⁵ Fam. 7.14.1 = SB 38, with Griffin (n. 14 above), 332–3.

¹⁶ Hor. Ars P. 464–5; Lucr. 1.731–2. Brink and Rudd, Bailey and Brown ad locc. (respectively) each cite Emped. fr. DK B112.4–5.

¹⁷ Murley, C., ‘Cicero's attitude toward Lucretius’, CPh 23 (1928), 289–91Google Scholar, at 290.

¹⁸ An earlier attempt to make the antithesis one between divine and human is that of Hayman, H., ‘Mr Munro's Lucretius’, The Contemporary Review 5 (1867), 222–36Google Scholar, who suggests emending sed cum ueneris to si deum inueneris, so that Cicero's verdict on Lucretius is ‘“if you shall have found him to be divine, I shall deem you a man indeed”’ and involves ‘a further by-play of words between deum and virum besides that between virum and hominem’. However, the emendation is motivated by disquiet at Cicero's perceived ambivalence towards the DRN and, being focussed entirely on Lucretius, would presumably not relate to Empedocles’ ‘divinity’.

¹⁹ TLL V.1.909.32–8. The pairing of diuus with homo is almost entirely an inclusive one involving the genitive plural of each, as at Enn. Ann. 591 Sk. diuumque hominumque pater rex.

²⁰ TLL V.1.1649.3–1658.21; deified mortal: Serv. ad Verg. Aen. 5.45; between mortal and god: TLL V.1.1649.70–1650.5; principes: ibid. 1655.6–1658.21.

²¹ TLL V.1.890.42–71, citing eleven examples from Cicero. Its use of Amphiaraus at Nat. D. 1.32 relates so closely to hero-cult as to be scarcely metaphorical, and some examples are softened by a quasi (Nat. D. 2.32) or a word indicating similitude or approximation rather than identity (Lig. 38, Marcell. 8), but most would stand as parallels for its proposed usage here.

²² Another famous and much-imitated contemporary example, as Bruce Gibson reminds me, is of course Lucretius’ assertion about Epicurus at DRN 5.8: deus ille fuit, deus (‘he was a god, a god’). This would be a further parallel for the proposed usage of deus in Q. fr. 2.10(9). It is not inconceivable that Cicero might be alluding to Lucretius, to whom he has just referred, though this would assume a more detailed knowledge (on the part of both brothers) of the whole DRN than Marcus’ tone suggests, and than might be considered likely so soon after its initial circulation. It would also rather blunt the point of the allusion to Sallustius, though that argument would be perilously circular.

²³ It might be added that it would be redundant to express the antithesis hominem non putabo, whose point has already been made in deum te putabo, except in allusion to Empedocles’ (and perhaps Sallustius’) equally redundant assertion that, as well as being a god, he was also no longer a mortal. I am grateful to Steve Heyworth for pointing out the redundancy and hope this provides a satisfactory explanation for it.

²⁴ In similar vein, CQ’s anonymous reader objects that ‘diuus is archaic and poetic, and thus an implausible reading in Cicero's prose’. This is indeed another factor in favour of deus, though it might be argued that the use of such anomalous diction could be designed to draw attention to the allusion to Sallustius’ certainly poetic and perhaps (like Lucretius’) archaizing opus. It might also make it more likely to be emended away through banalization.

²⁵ On polar errors, see esp. Young, D., ‘Some types of scribal error in the manuscripts of Pindar’, GRBS 6 (1965)Google Scholar, 267, Kopff, E.C., ‘An emendation in Herodotus 7.9.3.2’, AJPh 96 (1975), 117–20Google Scholar, Briggs, W.W. Jr., ‘Housman and polar errors’, AJPh 103 (1983), 268–77Google Scholar.

²⁶ I draw my examples from the useful list provided by Oakley, S.P., A Commentary on Livy, Books VI–X, Volume IV: Book X (Oxford, 2005), 558–9Google Scholarad Liv. 7.26.9 (addenda and corrigenda).

²⁷ Pace Oakley (n. 26 above), the case at Mur. 75, where Vat. lat. 10660 writes defendere for reprehendere, strikes me as less obviously a polar error than a misguided attempt at correction, since Cato might be expected to be defending rather than criticizing the practices of his ancestors (instituta maiorum).

²⁸ For the impact of monks’ religious preoccupations on their activities as copyists, see Ogilvie, R.M., ‘Monastic corruption’, G&R 18 (1971), 32–4Google Scholar.

²⁹ Sedley, D., Lucretius and the Transformation of Greek Wisdom (Cambridge, 1998), 1–34CrossRefGoogle Scholar, esp. 22, id., ‘The proems of Empedocles and Lucretius’, GRBS 30 (1989), 269–96Google Scholar; cf. Furley, D.J., ‘Variations on themes from Empedocles in Lucretius' proem’, BICS 17 (1970), 55–64Google Scholar, Gyrani, M., Empedocles Redivivus: Poetry and Analogy in Lucretius (New York and London, 2007)Google Scholar.

³⁰ Cic. Amic. 24. The other references are at Fat. 39, Nat. D. 1.29, 1.93, De or. 1.217, Tusc. 1.19, 1.41, Acad. Pr. 1.44, Luc. 33 (bis), 62, 118, Rep. 3.19.

³¹ On ‘double allusion’, ‘the simultaneous allusion to two antecedents, one of which is based on the other’, see esp. McKeown, J.C., Ovid, Amores, vol. I: Text and Prolegomena (Liverpool, 1987), 37–45Google Scholar.

³² Sedley (n. 29 [1998]), 22 makes a similar case for the recognizability of allusions to Empedocles/Sallustius’ proem in Lucretius’: ‘many could be assumed … at least to have read the opening’. Of course, the incipit and proem are marked, privileged portions of any text and might be considered especially memorable even by those who have read the whole.

³³ See n. 20 above; Apul. Apol. 43 alluding to πᾶν τὸ δαιμόνιον at Pl. Symp. 202e. CQ’s anonymous reader rightly notes that this distinction between deus and diuus is mainly attested in later texts, but there is a suggestion of a connection between diuus and deification or heroization at Cic. Leg. 2.22: suos leto datos diuos habento (following Powell and Dyck in accepting Davies's suos for MS nos). Empedocles as δαίμων: DK B115, with Gain, F., ‘Le statut du daimon chez Empédocle’, Philosophie Antique 7 (2007), 121–50Google Scholar.

³⁴ A BTL search found 281 instances of the phrase in all cases.

³⁵ della Corte, F., ‘Gli Empedoclea e Ovidio’, Maia 37 (1985), 3–12Google Scholar. Cf. Hardie, P., ‘The speech of Pythagoras in Ovid Metamorphoses 15: Empedoclean epos’, CQ 45 (1995), 204–14CrossRefGoogle Scholar = Lucretian Receptions (Cambridge, 2009), 136–52Google Scholar.

³⁶ We might even wonder whether some phrase in Sallustius to the effect that the reader should ‘believe’ Empedocles to be a god is the common source of Horace's formulation haberi … cupit Empedocles and Cicero's odd repetition of putabo.

³⁷ I am indebted to Rick Benitez, Anthony Hooper and the Empedocles Reading Group of the Inspired Voices research cluster at the University of Sydney for inspiration, and to Steve Heyworth, CQ's anonymous reader and CQ’s editor, Bruce Gibson, for helpful comments and suggestions. Since none of them was ultimately convinced by the conjecture, they should certainly not be held responsible for the article's remaining shortcomings.