1. THE MEANING OF RECIPIT MENS

The first of these questions has not bothered all editors, many of whom retain the transmitted text.Footnote ³ Some have been concerned by the apparent repetition mens (239) … mente (240), a point addressed briefly below, but I believe there is nothing amiss with this first instance of mens.Footnote ⁴ The greater difficulty with the phrase is whether recipit can be given an acceptable sense.Footnote ⁵ Bailey (following Ernout–Robin) asserts that recipit can mean ‘admits’ or ‘allows’, apparently believing that such a rendering resolves all problems. In defence of this position, he cites Lucr. 1.377 scilicet id falsa totum ratione receptum est, where receptum est must mean something like ‘has been admitted as credible’. While this line clearly proves that recipere can describe the allowance of a given belief in Lucretius, it does not remove all questions. It is worth considering – in addition to whether recipit characterizes belief – whether its sense when it does so is appropriate to the structure and argument of the passage.

The lines run broadly as follows. Lucretius first acknowledges that the nature of the animus, or rational part of the soul, has been found to be triple (that is, consisting of aer, calor or vapor, and ventus or aura). Then he asserts that these elements ‘all together’ are nevertheless not sufficient for generating sensation. He subsequently explains, in lines that have been understood and emended variously, why the previous three psychic components are inadequate. Lucretius finally concludes, as a result of this explanation, that there must be (necessest, 241) a further, fourth element of the soul. From a logical perspective, then, this last deduction provides perhaps the greatest clue to the content of the preceding lines, 239–40. Whatever precisely they say, these lines offer strong grounds for setting aside a theory of the soul based in only three elements.

Accordingly, available renderings of recipit mens do much to make 239–40 a firm basis for the deduction of 241–2, giving a clear sense that an untenable possibility is being excluded. Bailey, for example, translates ‘… the mind does not admit …’.Footnote ⁶ The Rouse–Smith Loeb offers ‘… the mind cannot admit …’.Footnote ⁷ Ernout similarly provides ‘l'esprit … refuse d'admettre …’.Footnote ⁸ Yet, as careful readers already will have noticed, these translations produce their sense of logical exclusion not simply by rendering recipit with ‘admits’ or ‘allows’. They also universally introduce a negation, as if to translate mens non recipit or mens recipere non potest. Since non itself does not appear in line 239, this negation must instead be extracted from nil (which is here clearly an accusative subject in an accusative and infinitive construction rather than an adverb). Thus, with only minor variations, translators agree in rendering nil horum quoniam recipit mens posse creare not with the literal ‘since the mind admits that none of these can create’ but rather with ‘since the mind does not admit that any of these can create’.Footnote ⁹ This latter version certainly provides a clear, logical basis for the subsequent deduction, but is it in fact warranted? Under many conditions, it may be, but the two translations are not necessarily equivalent: if recipit only means ‘admits’ as in ‘admits as one credible possibility (among several)’, then it may be too weak to bear the second rendering.Footnote ¹⁰ Put differently, for these two translations to be logically equivalent, mens recipit needs to mean ‘the mind accepts as credible that (none of these can create sense-bearing movements)’, while also firmly ruling out any contradictory position (e.g. that ‘one of these elements can create etc.’).Footnote ¹¹ The mere fact of ‘acceptance’ must exclude all contradictory hypotheses – even if those hypotheses fit equally well with available information.Footnote ¹² Moreover, although it need not impact on the translation per se, the obvious implication of the passage is that whatever ‘the mind allows’ is in fact true, as a mistaken assumption would hardly provide adequate reason for introducing the nameless ‘fourth nature’.

To clarify the force of these points, it is perhaps not out of place to invoke the practice of pleonachos tropos, or multiple explanation, within Epicurean physics.Footnote ¹³ In laying bare the causes of celestial phenomena, Epicurean authors are often content to produce multiple possible explanations – explanations compatible with the visible evidence but perhaps not with each other – without then designating one explanation as correct. In the present context, then, it is not impossible to suppose that recipit means ‘allows’ or ‘accepts’ in the limited sense of allowing that a given possibility is consistent with the available information. Yet this sense, if correct, does not – as the various translations that negate or otherwise alter it implicitly suggest – provide the tight, argumentative logic desired. To pursue this idea a bit further, it is entirely consistent with the practice of pleonachos tropos that some explanations may be, at least in the case of any one world, outright erroneous. As Lucretius remarks (5.526–9), nam quid in hoc mundo sit eorum ponere certum, | difficile est; sed quid possit fiatque per omne | in variis mundis varia ratione creatis, | id doceo …Footnote ¹⁴ But in the present context an erroneous explanation – whether ‘allowed’ by the mind or not – would hardly provide a sound basis for deduction.Footnote ¹⁵

So, can recipit – as the various translations imply – mean ‘accepts only (that)’, where ‘only’ is shorthand for indicating the exclusion of other contradictory hypotheses? The direct philological evidence is limited, but it at least permits the suspicion that past interpreters have strengthened the term through over-translation. recipit appears at OLD ² s.v. 7b in a citation of the very lines under discussion, with the definition ‘to accept as credible (a belief, theory, etc.)’. Five passages illustrate this sense, including the two (1.377, 3.239) from Lucretius already mentioned. None of these passages seems to require ‘accepts only’ as a translation, and in some instances that rendering would make little sense.Footnote ¹⁶ As we have seen already at 1.377, Lucretius himself uses recipio to describe an account that has been ‘accepted’ on false reasoning (falsa ratione). In that passage, Lucretius resists the assertion that fish swimming in water represents an instance of bodies moving bodies without void. He appears to contest the mere credibility of this explanation, not whether it is the sole correct one. His target, in other words, is to show that the given explanation is not consistent with the available data – not to demonstrate that it has been presumed (wrongly) to exclude other contradictory assertions. At Rep. 2.19 Cicero similarly employs recipio to characterize the ancient ‘acceptance’ of fabricated myths, which ‘this age’ now rejects: antiquitas enim recepit fabulas, fictas etiam non numquam incondite, haec aetas autem … eludens omne quod fieri non potest respuit.Footnote ¹⁷ The broader context praises Romulus because belief in his divinity arose in a more educated age; the immediate context clearly disparages ‘antiquity’ as gullible. Recepit here seems unlikely to mean ‘accepted only’ as in ‘antiquity accepted myths alone as credible’ (thus excluding e.g. opposing myths or more ‘scientific’ explanations). The chief point is that earlier men uncritically ‘allowed’ even obviously false myths, not that they exclusively allowed them. Finally, at Ep. 87.23 Seneca the Younger spells out the unfortunate consequences that one must logically ‘accept’ (recipiat) if he believes that sacrilege and theft are not bad tout court but bad because they do more bad than good: quisquis hoc dicit, necesse est recipiat sacrilegium, sicut malum sit quia multa mala facit, ita bonum quoque ex aliqua parte esse, quia aliquid boni facit … The logical consequence that sacrilege is ‘part good’ is certainly not the only one that the hypothetical interlocutor ‘accepts’ (for he clearly also accepts the partially opposed hypothesis that it is ‘bad’); it is simply the one that Seneca views as most unappealing.

Apart from Lucr. 3.239, the passages just examined constitute the only secure instances of recipio listed at OLD ² s.v. 7b. Also listed under that heading is a passage of Pliny the Elder (HN 35.56), where recipi may not be the true reading: quod si recipi necesse est, simul apparet multo vetustiora principia … If it must be ‘accepted’ (recipi), Pliny asserts, that the ‘perfection’ (absolutione) of the art of painting dates from the death of Romulus, then the beginnings of the art are also much older. If recipi is retained here, this passage comes closest amongst the examples to the sense ‘accepts only or exclusively’, perhaps merely because of the nature of chronology. Contradictory chronologies cannot normally be ‘accepted’ as equally plausible. But the majority of the manuscripts here read praecipi, which gives perfectly good sense.Footnote ¹⁸ Much as at Sen. Q Nat. 2.2.1, it would mean ‘to take as a premiss or point of departure’.Footnote ¹⁹ It is perhaps unwise, then, to rely too heavily on this passage in determining the sense of recipit mens.

What does this evidence show? Taken together, these four texts provide only limited support for the traditional rendering of line 3.239, ‘the mind does not admit that any of these can create’. Instead, examples such as Cic. Rep. 2.19 – the only external parallel from Lucretius' own time – suggest that recipere can be used to describe the acceptance of fabulae, or myths, as credible. Given the many variations and competing versions in surviving myths, we may question whether recipere in this passage intrinsically conveys any sense of exclusive belief or acceptance. Similarly – a point that was raised hypothetically when discussing the pleonachos tropos – these passages (especially those from Lucretius and Cicero) collectively prove that recipio often describes mistaken belief. When recipio is employed to characterize belief, then, there are often other possible (i.e. correct) beliefs beyond the ‘accepted’ ones. The mere fact of multiple (possible) beliefs does not entail that recipere means ‘to allow as credible’ in a limited rather than in a strong, exclusive sense; yet – since Epicureanism is firmly committed to the infallibility of the senses – any error, any mistaken belief, must be introduced in the mental process named by recipere, which is at least mildly troubling for an argument grounded in the conclusions of that process.Footnote ²⁰

Perhaps now we can see the point of Heinze's remark that ‘the transmitted mens recipit could be understood as “the mind accepts, that is, endorses, holds as correct”: but instead of this – because the mind also frequently enough considers falsehoods true [emphasis added] – we ought to expect “cannot accept”’.Footnote ²¹ This comment not only presciently predicts the course pursued by most twentieth-century translators (who, as we have seen, turn ‘accept’ into ‘cannot accept’, vel sim.) but it also suggests the following conclusion: while recipit mens in itself is hardly indefensible as Latin, recipit sounds the wrong note for the argument of the passage, because the word so frequently characterizes erroneous belief, rather than irrefutable deduction. As we have seen, the sense of recipit mens appears to be much closer to ‘the mind considers plausible that’ than ‘the mind accepts (e.g. based on irrefutable evidence) only that’.Footnote ²² If this assessment is correct, we are left with a much more anaemic argument than is sometimes supposed: the ‘fourth nature’ must exist since the mind (possibly erroneously) ‘considers plausible the idea that none of the three elements can create sense-bearing movements’.Footnote ²³ Further occurrences of recipio in De rerum natura, which all describe physical taking or receiving rather than mental belief, do nothing to change this conclusion.Footnote ²⁴

2. NIL HORVM QVIA PERCEPIT MENS POSSE CREARE?

The evidence discussed in the previous section suggests that Heinze was not overzealous in obelizing recipit mens. While the foregoing analysis may make recipit only unlikely rather than strictly impossible, a pair of simple corrections, which alters no more than a few letters, can strengthen the argument and make 3.239 easier to construe.

First, I submit that recipit, which is reported as recepit in our chief witnesses, in fact conceals percepit.Footnote ²⁵ Unlike recipere, percipere is regularly used in Lucretius of ‘learning’ or ‘grasping’ mentally.Footnote ²⁶ Consider the following examples: nunc age, iam deinceps cunctarum exordia rerum | qualia sint et quam longe distantia formis, | percipe … (2.333–5); nunc age dicta meo dulci quaesita labore | percipe … (2.730–1); quod procul a vera quam sit ratione repulsum | percipe (6.768). In each of them, Lucretius uses a common locution (a second-person imperative that occurs in the same form twelve times in De rerum natura) to implore Memmius (or the reader more generally) to ‘learn’ or ‘grasp’ his various lessons.Footnote ²⁷ As these examples already suggest, percipere seems to be used regularly with regard to ‘learning’ the central truths of Epicureanism, a fact that is perhaps most clear at 4.23–5: si tibi forte animum tali ratione tenere | versibus in nostris possem, dum percipis omnem | naturam rerum ac persentis utilitatem.Footnote ²⁸recipere, as we have already seen, can characterize mistaken belief, but in these examples there is no hint that percipere typically does so. On the contrary, percipere often describes precisely the type of learning Lucretius envisions for his reader.Footnote ²⁹

If percepit replaces recipit, then the sense of the line will be: ‘since the mind has grasped that none of these can create (sense-bearing movements etc.)’. Line 4.24 (cited above) suggests that one can ‘grasp’ the whole nature of things; here, more restrictedly, the mind will have ‘grasped’ its own nature.Footnote ³⁰ Within the present argument, this change means that the three-element theory of the soul is rejected not because the mind ‘admits’ (or ‘does not admit’) but rather because the mind, which in Epicurean theory is itself a body, ‘has grasped’ Lucretius' immediately preceding statements about its nature, or perhaps more exactly ‘has perceived’ their truth through (infallible) sense perception. At 3.182–5 the poet remarks that nil adeo fieri celeri ratione videtur | quam sibi mens fieri proponit et inchoat ipsa; | ocius ergo animus quam res se perciet ulla, | ante oculos quorum in promptu natura videtur. This claim is grounded in visual perception: since nothing is seen to operate more quickly than the mind, then nothing visible (ante oculos) goads itself on more quickly.Footnote ³¹ Similarly, at 3.208–20 Lucretius argues that the ‘fine texture’ of the soul may be recognized by observing a dead man (who no longer possesses his): mors omnia praestat | vitalem praeter sensum calidumque vaporem (214–15). This argument is again based in what we perceive: vapor calidus and sensus vitalis, as touch and sight teach, are the only things absent from a corpse. The relevance of these claims to 3.239 is as follows: if we accept them, then the three psychic elements (aura, vapor, aer) described at 3.231–6 cannot be sufficient for characterizing the mind, precisely because all of them are perceptible and may be perceived not to possess sensation individually. Hence, a ‘fourth nature’ must be posited.

There are also syntactical reasons to prefer percepit. Unlike recipere, which here (if retained) governs an accusative and infinitive construction for the first time in the sense ‘allows’ (a construction with only one other instance in Classical Latin, Sen. Ep. 87.23), percipere governs an accusative and infinitive construction in several attested early Latin instances. Plaut. Mostell. 727 provides a close parallel (where this verb governs an accusative and infinitive construction): … vix tandem percepi super his rebus nostris te loqui.Footnote ³² As Leumann–Hofmann–Szantyr makes clear, the accusative and infinitive construction with percipere was normal in early Latin, if more limited in the Classical period.Footnote ³³

The alert reader will already have noticed, though, that substituting percepit alone for recipit would produce an un-metrical line. This simple change, if correct, thus necessitates another: for quoniam, I suggest reading quia, a minimal modification. If common abbreviations were used in the later transmission of Lucretius' text, then qa (quia) easily could have been misread as qn or qm (quoniam).Footnote ³⁴ In other words, by postulating a misreading of no more than two or three letters, it is possible to turn line 239 from an unusual extension of recipit into a regular, early Latin expression. While Lucretius' text is hardly without neologisms and syntactical innovations, it seems unduly conservative to cling to the difficult recipit with an easy alternative available.Footnote ³⁵

One final point must be mentioned in consideration of these suggestions. A reader trained in the Latin hexameter will notice that the newly emended line (with quia percepit) is a so-called versus vitiosus; that is, it apparently has no caesura (either masculine or feminine) in its third foot. While such a recognition at first argues against the emendation, if one studies other instances of such verses in Lucretius, this line may be seen to be broadly similar. For example, line 3.715 reads haud erit ut merito immortalis possit haberi, also apparently lacking any caesura. Considering this and similar verses, Bailey suggests instead that at times Lucretius has a ‘quasi-caesura’: ‘Lucretius’, he writes, ‘seems to have a quasi-caesura in the third foot produced by a break in a compound word between its prefix and its root’.Footnote ³⁶ Thus, at 3.715, a ‘quasi-caesura’ is felt in im-mortalis. Similar is 6.197 (complerunt, magno indignantur murmure clausi), where a ‘quasi-caesura’ may be felt in in-dignantur. In the proposed emendation of 3.239, then, I would suggest that a ‘quasi-caesura’ is felt in per-cepit, explaining any apparent metrical anomaly in the line.

III. CONCEALED PLEASURE

Whether one accepts quia percepit in place of quoniam recipit or not, the graver difficulty is clearly in line 240. Here, as we have already explored above, Lucretius lays the groundwork for presenting the nameless, ‘fourth nature’. Because the other three elements of the soul – ‘air’ (aer), ‘wind’ (ventus) and ‘heat’ (calor) – cannot account for sensation, Lucretius posits in lines 241–4 that a fourth element is necessary, and then in 245–51 he describes how this nameless ‘nature’ generates sensation throughout the body. Motion begins with the fourth nature itself and spreads to the ‘heat’ (calor), then to the ‘wind’ (ventus), then to the ‘air’ (aer); from there blood is stirred up, flesh gains sensation, and finally pleasure (voluptas) or an opposite passion (contrarius ardor, i.e. pain) is given to bones and their marrows. As this description makes clear, Lucretius' account of sensation explains ultimately how pleasure and pain, the two varieties of sensory experience most important to Epicurean hedonistic calculus, are generated.

Within this larger trajectory, line 240, as we have already seen, plays a key role, naming the psychic faculties unaccounted for by air, wind and heat alone. While it is clear that one of the faculties discussed in 240 is sensation, or ‘sense-bearing movements’, the second half of the line has been debated extensively. There is clearly something amiss with the transmitted text, which – whether it is taken as quaedam quae mente volutat (O) or an available variation thereon – does not provide good sense. Bailey offers a lucid exposition: ‘… the majority of modern editors (except Heinze who thinks the meaning was “however you turn it over in your mind”) seem agreed that the general sense of the line is “the motions of sensation and thought”, i.e. of the anima and the animus. It is true that in the rest of the paragraph it is only the sensation of the anima which is in question, but Lucr. may well at the outset have included the motions which caused thought in the mind … An emendation is required which means quite simply “and thought”.’Footnote ³⁷ After rejecting several possibilities, including Munro's et homo quae mente volutat and Postgate's nedum quae mente volutas, Bailey pronounces Goebel's et quae mens cumque volutat and Frerichs' et mens quaecumque volutat the best available options, printing (like Kenney) the conjecture of Frerichs.Footnote ³⁸

Recent proposals generally follow Bailey's lead. Saunders, for example, argues for et quaecumque ipsa volutat (adopted by the Rouse–Smith Loeb), offering a neat stylistic refinement of Frerichs' conjecture, while Richter (apparently unwittingly) revives an old conjecture of Purmann, aut quae quis mente volutat, which provides similar sense with a disjunction rather than a conjunction.Footnote ³⁹

Before we proceed any further, it is worth teasing out the chief assumptions in Bailey's analysis, both because they will aid our own investigation and because agreement about them is perhaps less univocal than Bailey makes it seem. On the one hand, although Bailey recognizes (as we, too, have already seen) that the broad sweep of the passage addresses only sensation (which he attributes to the anima, or irrational soul), he nonetheless feels that Lucretius may have mentioned ‘thoughts’ (which belong to the animus, or rational soul) in 240 before moving on to a fuller exposition of sensation. He also sees a firm distinction between the sensation of the anima and the thought of the animus. His preferred emendations are thus the palaeographically and metrically plausible proposals that, while keeping the two components of the line distinct (i.e. ‘sensation’ and ‘thought’, not ‘sensation’ from which ‘thought’, vel sim.), none the less maintain a balance between anima and animus. Frerichs' and Goebel's conjectures are both close to the paradosis and mean ‘thoughts’ or ‘whatever the mind turns over’.

Yet several older emendations raise questions left unexplored by Bailey. First, it is not obvious that the first and second portions of 240 name two fully distinct items. Even if the second half of the line concerns mental activity – a conclusion that is far from certain given the presence of mens just above in line 239 – suggestions such as Eichner's qui mentem cumque volutant, Lambin's quaedam qui mente volutent, and Lachmann's quaedam vis menti' volutat, which all posit no conjunction (or disjunction) between sensiferos motus and whatever lurks beneath the transmitted quaedamque mente, demonstrate that many have not felt the division between sensation and thought, anima and animus, to be as firm as Bailey suggests.Footnote ⁴⁰ At 3.152–3, for example, Lucretius remarks: verum ubi vementi magis est commota metu mens, | consentire animam totam per membra videmus, lines which show that the mens (animus) and anima can ‘feel together’ (consentire), even if no one would argue that the anima (in the limited sense ‘irrational soul’) ‘thinks’ independently.Footnote ⁴¹ Moreover, whatever the connection between the two parts of the line, there is some concern that Bailey's preferred emendations do not portray thinking entirely in line with Epicurean theory. ‘Thoughts’, as we learn at Lucr. 4.722–817, function much like other visual observations: haec fieri ut memoro, facile hinc cognoscere possis. | quatenus hoc simile est illi, quod mente videmus | atque oculis, simili fieri ratione necesse est (4.749–51). On this theory, the mind serves a selective, judging function. Potential thoughts, a thinner form of simulacra (‘thin image-films’), strike our minds, which then focus on the desired images.Footnote ⁴² But it is unclear, at least from the evidence of the fourth book, that any part of the mind actively ‘produces’ thoughts (as creare in 239 would imply). Presumably, this is Heinze's objection when he remarks that ‘die Dinge, quae mente volutamus, [sc. sind] bestimmte Gegenstände … die natürlich nicht durch die Seelensubstanz erzeugt werden [emphasis added]’.Footnote ⁴³

Arguably the key question for resolving these difficulties is: what explanatory role does the fourth nature, which is first introduced here as a remedy for the inability to create sensation and ‘thoughts’ (or whatever truly completes 240), play in Epicurean theory? The soul is regularly characterized as a ‘mixture’ whose various components none the less may have specific spheres and influences.Footnote ⁴⁴ Thus, at 3.288–93 Lucretius associates calor (‘heat’) with anger (ira), aura (‘wind’ or ‘breath’) with fear (formido), and aer (‘air’) with calm (vultus serenus). These attributions seem clear enough, but the characteristics of the fourth nature have occasioned more debate. Aët. 4.3.11 (Usener 315) echoes Lucretius' description of the soul, while also making a clear declaration about the fourth nature: Ἐπίκουρος (sc. τὴν ψυχὴν) κρᾶμα ἐκ τεττάρων, ἐκ ποιοῦ πυρώδους, ἐκ ποιοῦ ἀερώδους, ἐκ ποιοῦ πνευματικοῦ, ἐκ τετάρτου τινὸς ἀκατονομάστου· τοῦτο δ' ἦν αὐτῷ τὸ αἰσθητικόν. ὧν τὸ μὲν πνεῦμα κίνησιν, τὸν δὲ ἀέρα ἠρεμίαν, τὸ δὲ θερμὸν τὴν φαινομένην θερμότητα τοῦ σώματος, τὸ δ' ἀκατονόμαστον τὴν ἐν ἡμῖν ἐμποιεῖν αἴσθησιν· ἐν οὐδενὶ γὰρ τῶν ὀνομαζομένων στοιχείων εἶναι αἴσθησιν. Just as wind is associated with movement, air with calm, and heat with bodily warmth, the fourth, nameless element creates (ἐμποιεῖν) sensation. Lucretius states similarly (3.245) that it is the fourth nature ‘which first distributes sense-bearing motion (sensiferos motus) through our limbs’. It seems clear, then, that the fourth element is aimed to explain the creation of sensation. But some have also wanted to give the fourth element other, more purely intellectual functions.

When, for example, Lucretius remarks (3.279–81) that … nominis haec expers vis facta minutis | corporibus latet atque animae quasi totius ipsa | proporrost anima et dominatur corpore toto, he seems to grant the fourth nature a governing role in the soul-body assemblage. These words echo what he says elsewhere (3.136–9) about the animus, or rational soul: nunc animum atque animam dico coniuncta teneri | inter se atque unam naturam conficere ex se, | sed caput esse quasi et dominari in corpore toto| consilium quod nos animum mentemque vocamus. Such verbal similarities, coupled with an anti-Epicurean remark of Plutarch that attributes judgement, memory, love, hate and more generally intellectual and rational faculties to some ‘nameless’ quality, have occasionally inspired scholars to identify the fourth nature either with the animus or with an element belonging solely to the animus.Footnote ⁴⁵ On the other hand, if – as Long and Sedley have suggested – it is possible to read a scholion to Epicurus, Ep. Hdt. as a description of the fourth nature, then that nature has both rational and irrational components, not unlike the larger soul.Footnote ⁴⁶ Yet whatever the precise connections between the fourth nature, animus and anima, the most obvious feature of all this evidence is that neither Lucretius nor any other surviving Epicurean text directly strives to clarify these relationships.

Thus, on the one hand, we have a strong set of evidence connecting the fourth nature to sensation; on the other, we have conflicting reports about the link between the fourth nature and the rational soul. While none of this evidence precludes a link between thoughts and the fourth nature – indeed, in so far as thinking is a form of perception underwritten by sensation, such a link should, on some level, be expected – it nevertheless makes the direct link between the fourth nature and ‘thoughts’ in line 240 seem even more out of place. The issue is not merely that Lucretius spends the ensuing lines discussing the way in which sensation spreads through the anima, ultimately yielding pleasure or pain, but rather that the particularity of the line-ending ‘thoughts’ seems jarring. Other passages that link the ‘nameless’ nature to rational activity speak of governing, judging, recalling – even loving and hating – but they say nothing directly about ‘thoughts’, which again seem not so much to be created (creare, 239) by the animus as judged or selected by it. As Heinze puts it, ‘… bei Munros Deutung (sc. et homo quae mente volutat) ist die Ergänzung des ganz allgemein gesagten sensiferos motus, d. i. αἴσθησιν im weiteren Sinne der specifisch seelischen Thätigkeit, von der auch im Folgenden ausschliesslich die Rede ist, durch das specielle “Gedanken” sehr unglücklich: man würde mindestens die πάθη genannt zu sehen erwarten’.Footnote ⁴⁷

This series of observations perhaps explains why no existing emendation has inspired full confidence. Collectively, past proposals – while often palaeographically plausible – have done little to correct the misplaced focus of the latter half of 240. As we have seen, ‘thoughts’ are not clearly created by the fourth nature and in any case they seem too particular for the otherwise general explanation of the line. Further, as Heinze points out, there is a more common Epicurean complement to sensation – πάθη, the most famous of which are pleasure and pain (precisely what is mentioned at 3.251).Footnote ⁴⁸ In Epicurus himself, it is common to associate sensations (Greek: αἰσθήσεις) and πάθη, as they form, according to Diogenes Laertius, two of the three criteria of truth (by which we judge the external world), discussed in the now lost Canon.Footnote ⁴⁹ At Ep. Hdt. 55, for example, Epicurus remarks that βέλτιον γὰρ … τὰ κατὰ τὰ πάθη καὶ τὰς αἰσθήσεις γινόμενα ἀποδοθήσεται. Or again at Ep. Hdt. 63, he comments … δεῖ συνορᾶν ἀναφέροντα ἐπὶ τὰς αἰσθήσεις καὶ τὰ πάθη – οὕτω γὰρ ἡ βεβαιοτάτη πίστις ἔσται – ὅτι ἡ ψυχὴ σῶμά ἐστι λεπτομερές κτλ.Footnote ⁵⁰ This evidence suggests that it would be unsurprising if, instead of thoughts, πάθη lurk somewhere in the latter portion of line 240.

Is it possible, then, to interpret the corrupt line such that it provides both a better complement to sensiferos motus and a better introduction to the ensuing lines? I believe it is. The key is to recognize that line-final volutat may conceal the more obviously relevant voluptas, the most frequently discussed πάθος in Epicurean thought (which perhaps may stand pars pro toto for all πάθη). While volutat, as the many emendations that retain it suggest, certainly can be understood, it is in fact a surprisingly uncommon word in Lucretius' lexicon, occurring only here in his text.Footnote ⁵¹voluptas, by contrast, is common in De rerum natura, and it regularly occurs precisely in this line-final position.Footnote ⁵² If, then, voluptas is correct, the remaining task for the critic is to explain and understand the quaedamque mente of the manuscripts. Here I suggest, exempli gratia, that quaedamque mente conceals quibus e fit deinde.Footnote ⁵³ While at first brush such a suggestion may seem a somewhat extravagant rewriting, the changes may be explained economically – if also somewhat speculatively. The corruption begins when a scribe, perhaps anticipating quaedam in the next line, mistakes an abbreviated quibus (qb·) for quaedamque (qdq· or q·).Footnote ⁵⁴ Then deinde is misread as mente under influence from mens in the preceding line, and e fit falls out for metrical reasons.Footnote ⁵⁵ An attempt to restore sense to the now-corrupt line turns voluptas into volutat.

An illuminating parallel for the final portion of this hypothesis is Ov. Tr. 2.357: nec liber indicium est animi, sed honesta voluptas. Manuscripts here show both voluntas and voluptas at the conclusion of the line.Footnote ⁵⁶voluptas recently has been preferred as the true reading, which, if correct, perhaps demonstrates the manner in which voluptas was corrupted in our passage.Footnote ⁵⁷ Not unlike mente volutat, honesta voluntas is good Latin, paralleled elsewhere in extant literature:Footnote ⁵⁸ it enters Ovid's text not so much because of confusion between the letters p and n as because honesta voluntas is itself an attractive phrase; one familiar word or phrase is substituted mistakenly for another. Similarly, in Lucretius, once mente enters the text, volutat becomes an attractive modification, which gives some sense (even if not a convincing one) to an already badly corrupt text.Footnote ⁵⁹

Some may be unconvinced by the details of this proposal, yet even if the suggested quibus e fit deinde voluptas succeeds only as a diagnostic emendation, one that allows future editors to explain more fully the errors in the line, it is worth emphasizing the degree to which it yields an improved sense for Lucretius' text. Recent work on sensation in Epicureanism by David Konstan has confirmed that the πάθη of pleasure and pain belong to the anima, the irrational part of the soul, rather than to the animus or mens, its rational component.Footnote ⁶⁰ Thus, the newly emended line better introduces Lucretius' ensuing text, which also focusses on the anima. Moreover, as we have explored, the new emendation avoids positing a poorly attested causal connection between the fourth nature and ‘thoughts’ or providing a description of Epicurean thinking that seems at least mildly at odds with other extant accounts. Finally, while voluptas may seem to lend a greater specificity to the line than sensiferos motus, the fact that the same term is picked up at 3.251 and made to stand (together with pain) for a whole range of πάθη makes it an appropriate conclusion to the line. The advantages, then, of reading voluptas for volutat are numerous and far-reaching; and while it may always be difficult to obtain certainty in emending the corrupt central portion of line 240, the proposed quibus e fit deinde offers the requisite sense. In conclusion, then, the full proposal is that Lucr. 3.237–42 should be emended to read as follows: