SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

The Proems of 
Empedocles and Lucretius 

David Sedley 

Lucreti poemata ut scribis ita sunt, multis luminibus ingeni, multae 
tamen artis. sed cum veneris, virum te putabo si Sallusti Empedoclea 
legeris, hominem non putabo. 

Writing to his brother in 54 B.C., Cicero supplies two unique testi­
monies (Ad Q. Fr. 2.9.4). In the first sentence he echoes Quintus' 
admiration for Lucretius' poem, thus providing the sole allusion to 
the De rerum natura likely to be more or less contemporary with 
its publication. In the second, he attests the publication of an Em­
pedoclea by a certain Sallustius, presumably a Latin translation or 
imitation of Empedoc1es.1 

But even more striking than the two individual testimonies is 
their juxtaposition. Editors have traditionally printed a full stop after 
sed cum veneris, apparently understanding "But when you come 
... (sc. we will discuss it)." This suppresses any overt link between 
the two literary judgements. On the more natural and fluent read­
ing that can be obtained simply by substituting a comma for the full 
stop, as printed above,2 the letter becomes an explicit comparison 
between the DRN and the Empedoclea: 

Lucretius' poetry shows, as you say in your letter, many flashes of 
genius, but also much craftsmanship. On the other hand, when you 
come, if you have read Sallustius' Empedoclea, I shall consider you a 
man, though I won't consider you human. 

1 Cj. Cicero's title Aratea for his translation of Aratus. 
2 The punctuation transmitted by the MSS. is unlikely to have ancient authority. 

Of the Latinists on whom I have tried out this redivision, some have found it ac­
ceptable, others have expressed reservations either about the structure of the re­
sultant sentence, or about the loss of the aposiopesis sed cum veneris .... As to the 
latter worry, this telescoped expression does not in fact seem to occur elsewhere in 
Cicero's letters (although there are partial parallels at Att. 12.Sa and 14.20.3). As 
regards the former, I cannot help wondering how many scholars would have felt 
the need to argue for repunctuation if the MSS had chosen to divide the text as I 
have printed it. 

269 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

270 THE PROEMS OF EMPEDOCLES AND LUCRETIUS 

If this is right, the two works were being directly compared at the 
time of their publication, and Cicero, at least, judged the Lucretian 
poem vastly superior. 

Why did this particular comparison suggest itself? It is well recog­
nised that Empedocles is, along with Homer, Ennius, and others, 3 

an important literary influence on Lucretius, and it has even been 
claimed that he was a philosophical influence. 4 But I do not believe 
that the depth and significance of the poem's Empedoclean charac­
ter have yet been properly understood. If what I shall argue in this 
paper is right, Cicero's comparison of the DRN with the Empedo­
clea will turn out to be an entirely natural one, which Lucretius 
would have welcomed and indeed invited. My case will be centred 
on the relation of Lucretius' proem to the proem of Empedocles' 
On Nature. 

1. Empedocles' Two Poems 

There is plentiful evidence that it was principally if not exclusively 
in the hexameter poem known in antiquity as the On Nature (II£pi 
cpuCJ£ro<;) or the Physics (Ta cpuCJtKa) that Empedocles expounded 
his world system, centred on the cosmic cycle in which four 
enduring elements-earth, air, fire, and water-are periodically 
united into a homogeneous sphere by Love, then again separated 
out into a cosmos by Strife. s But there is a longstanding scholarly 
tradition, deriving primarily from Diels' editions published in 1901 

J The range of literary influences on Lucretius was considerably enlarged by the 
findings of E. J. Kenney, "Doctus Lucretius," Mnemosyne Su. IV 23 (1970) 366-92. 

4 D. J. FURLEY, "Variations on Themes from Empedocles in Lucretius' Proem," 
BICS 17 (1970) 55-64 (hereafter 'Furley' [reprinted in Cosmic Problems (Cam­
bridge 1989) 172-82]), discussed below; also J. Bollack, "Lukrez und Empedokles," 
Die neue Rundschau 70 (1959) 656-86; cf nn.25 and 40 infra. 

5 The traditional belief that zoogony took place in both halves of this cycle, for 
which see especially D. O'Brien, Empedocles' Cosmic Cycle (Cambridge 1969), 
has been powerfully challenged by J. Bollack, Empedocle (Paris 1965-69); U. 
Holscher, "Weltzeiten und Lebenszyklus," Hermes 93 (1965) 7-33; F. Solmsen, 
"Love and Strife in Empedocles' Cosmology," Phronesis 10 (1965) 109-48 (= 
Kleine Schr. I [Hildesheim 1968] 274-313); and A. A. Long, "'Empedocles' Cosmic 
Cycle in the 'Sixties," in A. P. D. Mourelatos, ed., The Presocratics (New York 
1974) 397-425; it is now ably defended by D. W. Graham, "Symmetry in the 
Empedoclean Cycle," CQ N.s. 38 (1988) 297-312. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

DAVID SEDLEY 271 

and 1903,6 of attributing all the fragments concerning Empedocles' 
theories of the pollution and transmigration of the individual spirit, 
or daimon, to a second hexameter poem, the Ka8ap/J.o{, or Puri­
fications. The original ground for this segregation was the belief that 
the physical doctrine of the cosmic cycle and the 'religious' doc­
trine of transmigration belonged to radically distinct and probably 
incompatible areas of Empedocles' thought. But Empedoclean 
studies have now reached a curious stage. On the one hand, the old 
dogma has been subjected to searching criticism, and is regarded by 
many as an anachronistic imposition on fifth-century thought.7 On 
the other hand, the conventional apportionment of fragments 
between the two poems, which was founded on that dogma, 
remains largely unchallenged,8 as if it had some independent 
authority. I believe that it has none. 

If we simply stick to the hard and the relatively hard evidence for 
what was in the Ka8ap/J.ot, a different picture will emerge. We do 
at least have its opening lines. 9 Empedocles addresses the citizens of 
his native Acragas and tells how he walks about as "a divine god, no 
longer a mortal," garlanded and revered. Wherever he goes, people 
follow him in their thousands, pressing him with enquiries, some 
requesting oracles, others asking to "hear a healing utterance for 
diseases of every kind." Why should we not suppose that the poem 
was simply a response to these requests, a set of purificatory 

6 H. Diels, Poetarum philosophorum fragmenta (Berlin 1901) and Die Frag­
mente der Vorsokratiker (Berlin 1903). 

7 E.g. C. H. Kahn, '"Religion and Natural Philosophy in Empedocles' Doctrine 
of the Soul," A GP 42 (1960) 3-35; J. Barnes, The Presocratic Philosophers II (Lon­
don 1979) 93; C. OSBORNE, '"Empedocles Recycled," CQ N.S. 37 (1987 [hereafter 
'Osborne']) 24-50; reservations in A. A. Long, '"Thinking and Sense-perception in 
Empedocles: Mysticism or Materialism?" CQ N.S. 16 (1966) 256-76. 

8 The two main challenges are those of N. VAN DER BEN, The Proem of Em­
pedocles' llepl 4>vO'ew., (Amsterdam 1975 [hereafter 'van der Ben']) and Osborne. 
I regard both as significant landmarks, even though my own conclusions will have 
relatively little in common with theirs. Osborne argues that the two titles name 
one and the same poem. Despite her arguments, I feel that D.L. 8.77 remains an 
insuperable obstacle to her view, as do the considerations advanced in my next 
section about the provenance of B115. In what follows I shall be trying to differ­
entiate sharply between the two poems in terms of their content. 

9 Empedocles B112 D.-K. I cite the fragments by their widely used Diels-Kranz 
numbers, although a better text is now available in M. R. WRIGHT, Empedocles: 
The Extant Remains (New Haven 1981 [hereafter 'Wright']). 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

272 THE PROEMS OF EMPEDOCLES AND LUCRETIUS 

oracles and "healing utterances"'?10 That would fit comfortably with 
the remaining evidence, which strongly implies that in the Ka9ap­
J.lOt Empedocles recommended celibacy and vegetarianism, along 
with advice on how to purify oneself of the pollution of previous 
meat eating. ll He also, in some context within the same poem, 
advised cleansing oneself with water drawn from five springs.12 In 
all these cases the impression given is not of any discursive exposi­
tion of the daimon's wanderings, but of instructions for ritual purifi­
cation. The one Empedoclean 'doctrine' that our sources tie to the 
Ka9apJ.lot is the thesis that the foetus achieves full human form in 
"seven times seven'" days (B 153a, cf AS3). Even here, significantly, 
the source that assigns this doctrine to the Ka9apJ.lot tells us only 
that Empedocles there "hints'" at it (aivtt't£'tat). This tends to con­
firm that the Ka9apJ.lot contained no expository account of gesta­
tion or birth,13 and that the mystical figure of "seven times seven '" 
came up incidentally in some related context-perhaps, for exam­
ple, in the course of purificatory advice concerning childbirth. 
Next, we have a two-line fragment assigned to Book 2 of the 
Ka9apJ.1O{14 in which Empedocles picks out a certain kind of tree, 
one with densely packed roots but well spaced branches. There is 
little to go on here, but it is perhaps a little easier to imagine a ritual 
context for these lines-e.g. the choice, or avoidance, of the right 
kinds of leaves-than one expounding the migrations of the 

10 For the scope and content of the relevant notions of pollution and purifica­
tion, see R. Parker, Miasma (Oxford 1983). I have no particular suggestion to make 
about the function of the ,. oracles." The evidence of a purificatory role for oracles 
is meagre (Parker 86), and I would guess that it is simply Empedocles' assumed 
divinity that makes this an appropriate designation for his pronouncements. 

11 Hippol. Haer. 7.30.3=preamble to BllO, and Porph.Abst. 2.31=B139, where lCa-
0ap~C; and 0\(1 'troY lCa9apIlrov respectively appear to allude to the poem, even if 
not to its title as such. In B 139 the words ,. Alas, would that the pitiless day had 
destroyed me before I schemed the wicked deeds of meat-eating for my lips" are 
usually taken to express Empedocles' own regrets, but Porphyry may seem rather 
to be reporting them as the formula that Empedocles recomends us to utter by 
way of self-purification. 

12 B143, where again Theon of Smyrna's lCa9apIlov seems to allude to the poem 
without actually naming it. I am unconvinced by van der Ben's argument (36£) for 
assigning the fragment to On Nature. 

13 See Wright 298. 
14 From a Herodian palimpsest=fr.152 Wright. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

DA VrD SEDLEY 273 

daimon.15 Finally, the biographer Satyrus quoted fragment BIll as 
confirming the suspicion that Empedocles dabbled in magic (D.L. 
8.59). Since Apuleius claims that it was Empedocles' Ka8apJiol that 
brought upon him just such a suspicion,16 there is a strong likeli­
hood that B 111 is from this poem.17 Significantly, this is once again a 
fragment offering not doctrinal exposition but ritual advice: how to 
influence the weather and to summon up the dead. 

My suggestion, then, is that the Ka8apJiol was just what its title 
and opening suggest, a set of oracles and purifications, consisting of 
ritual advice rather than doctrinal exposition. This is supported by 
the three other reports known to me of lCa8apJiol as a kind of com­
position. In all three cases these works are fathered on archaic fig­
ures of semi-legendary status. First, Epimenides the Cretan is said 
to have written lCa8apJiol, in verse and perhaps also in prose (3A2-
3 D.-K.). Although their content is not reported, it can hardly be a 
coincidence that Epimenides was celebrated above all for his ritual 
purifications, an exrertise that led the Athenians to send for him to 
purify their city 0 plague (A 1, 2, 4, 8). Second, the author of the 
pseudo-Pythagorean Carmen aureum says: "But abstain from the 
foods that I spoke of in my Ka8apJio{ and Absolution of the 
Soul. "18 This perfectly matches the purificatory theme we have al­
ready detected in Empedocles' work. Finally, the remark at Aris­
tophanes Frogs 1033 that Musaeus taught "healing and oracles" is 
glossed by a scholiast with the comment that Musaeus "composed 
absolutions [?], initiations, and lCa8apJioi. " (2A6 D.-K.). Healing and 

15 Cf B140 on avoiding laurel leaves. (According to Theophr. Hist.pl. 1.604, all 
plants have their roots more densely packed than their parts above ground, but 
some, e.g. the olive tree, have a particularly dense mass of slender roots.) Since, 
however, in EmpedocIes' view the daimon can be incarnated in plants (B117, 127), 
the possibility cannot be absolutely ruled out that the context was an expository 
one about transmigrations. 

16 Apul. ApoL 27=3A6a D.-K. 

17 The second person singular in B111 cannot be used, as traditionally ithas been, 
to settle the question of its provenance in favour of On Nature: on the impotence 
of this criterion see Osborne 31£. Now that we know that Empedocles wrote at 
least two books of KaBap~oi (fr.152 Wright), there is even less need to suppose that 
the second person plural address to the Acragantines, which opened Book 1, was 
retained throughout. 

18 Carmen aureum 67f: at D. Young, Theognis (Leipzig 1971) 103f. The author 
concludes the poem three lines later with an unmistakable echo of the opening of 
Empedocles' Kaeap~oi: EO"<H:m aeava'tO~ eEO~ a~pot~ oUlC£n eVTrt6~(cf Emped. 
Bl12A ). 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

274 THE PROEMS OF EMPEDOCLES AND LUCRETIUS 

oracles are precisely the two services mentioned by Empedocles at 
the opening of his Ka9ap)lot. Then why look further for the con­
tent of the poem? We are required to do so neither by the frag­
ments attributable to it, nor by the evidence for Ka9ap)lot as a 
genre. 

II. The Provenance of Empedocles B 115 

There is a decree of necessity, an ancient resolution of the gods, 
sworn by broad oaths, that when one of the daimons who have a 
share of long life defiles ... his own limbs, or does wrong and swears 
a false oath, for thirty thousand years he must wander, away from 
the blessed ones, being born during that time as every form of mor­
tal creature, exchanging for each other the arduous paths of life. The 
might of the aether drives him to the sea, the sea spits him out onto 
the threshold of land, the earth sends him into the rays of the gleam­
ing sun, and the sun hurls him into the whirling aether. One receives 
him from another, and all hate him. I too am now one of these, a 
fugitive from the gods and a wanderer, who trust in raving Strife 
(B115).19 

These lines, which are crucial for explaining the daimon's migra­
tions, have been assigned to the Ka9ap)lot by every editor of Em­
pedocles since Diels. Recently the attribution has been questioned 
by N. van der Ben,20 and subsequently defended by D. O'Brien.21 
But this renewed debate has so far focused excessively on the con­
texts in which the lines are quoted by our sources, as if one could 
settle the question of their provenance by counting the allusions in 
those contexts to lC&.9apat<; and cognate terms and likewise those to 
the cosmic cycle. Given the improbability that any ancient reader 
of Empedocles might have expected the physical poem and the 
Ka9apJlot to conflict doctrinally, the provenance of the lines will 
have mattered less to those who cited them than their value as evi­
dence for Empedocles' views on the lC&.9apat<; of the soul-a topic 
on which Platonism had conferred an absolutely pivotal philosophi­
cal importance. 

19 I have avoided engaging with the textual difficulties of this passage, which are 
well discussed by Wright. They do not affect any of the issues I am addressing 
here. 

20 Van der Ben 16ff; also Osborne, in so far as she rejects any distinction be­
tween the physical poem and the KaSapJlot. 

21 D. O'Brien, Pour interpreter Empedocle (Paris/Leiden 1981). 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

DAVID SEDLEY 275 

Plutarch reports that Empedocles "prefaced" these lines "at the 
beginning of his philosophy," EV apxft 't'i1~ CPtAo(Jocpia~ 1tpoa1tocprovn­
(Ja~ (De exil. 607c). Is this too vague to be helpful? "Philosophy" 
certainly might describe the content of the physical poem. It might 
also be appropriate to the KaeapJ.lOt, on the traditional view of that 
poem's content. But it is very much less appropriate if, as I have 
suggested, the Ka8apJ.1ot was not a doctrinal work but a set of 
purificatory pronouncements. Indeed, if that suggestion is correct, 
Plutarch's expression "at the beginning of his philosophy" would 
immediately gain a much clearer sense. If EmpedocIes wrote two 
doctrinal poems, the words "his philosophy" are a desperately 
vague way of referring to either one of them. But if he wrote just 
one, they become an entirely natural way of referring to that one. 22 

Plutarch's description in no way indicates that these were the 
very opening lines of the poem to which they belonged, just that 
they preceded the philosophy proper. Hence there is little value in 
the argument23 that since we have the opening of the KaeapJ.1ot 
and it differs from these lines, they must have opened the physical 
poem instead. Much more mileage can be got out of the content of 
the disputed lines. First, it is hardly insignificant that they name five 
of the six cosmic entities on which Empedocles' physical system is 
based: the daimon's wanderings are graphically described in terms 
of its being tossed into and out of each of the four elements in turn; 
and Strife is named as the cause of its downfall. This at least sup­
ports the coherence of the passage with the physical poem. But far 
more important, and strangely absent from the debate about its pro­
venance, is the following consideration. In these disputed lines, Em­
pcdocles is himself a fallen daimon: "I too am now one of these, a 
fugitive from the gods and a wanderer, who trust in raving Strife." 
Is it credible that these words came in the introductory passage of a 
poem in whose opening lines Empedocles had moments earlier de­
scribed himself as "a divine god, no longer a mortal"? 24 Without the 
straitjacket of the old prejudice that science and religion do not mix, 

22 Cf. Osborne 29ff. 
23 Van der Ben 16. 

24 Bl12A, reinforced by 113.2 ('"if I am superior to frequently-perishing mortal 
men"), if, as Sextus' juxtaposition of 113 with 112 suggests, it is also from the 
Kaeap~i. In Empedocles' world, even the generated gods perish eventually, i.e. at 
the end of each cosmic cycle: hence they are not immortal but "'long-lived" (B21.12, 
23.8; cf 115.5 on the daimones). By contrast, mortals are "frequently-perishing," 
1tOA:\)<p9EP£roV: see Wright 269. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

276 THE PROEMS OF EMPEDOCLES AND LUCRETIUS 

it is hard to believe that anyone would ever have thought of assign­
ing the former text to the Ka6ap)lot. The most natural interpreta­
tion is that B 115 comes from a poem in which Empedocles classed 
himself as a fallen daimon still working through his long cycle of 
transmigrations, whereas in the Ka6ap)lot, opening as it does with 
his confident self-rroclamation as a god, "no longer a mortal,» he 
presented himsel as having now completed the cycle and re­
covered his divinity. I therefore feel a reasonable degree of con­
fidence in placing Empedocles' major fragment on the wanderings 
of the daimon somewhere in the proem to the On Nature. This 
conclusion will prove important at a later stage in my argument. 
Earmarking it for future use, we can now at last turn to Lucretius. 

III. Lucretius and Empedocles 

Numerous echoes of Empedoclean passages have been recog­
nised in Lucretius' poem, with varying degrees of certainty.25 It is 
no part of my purpose to catalogue these. But two observations 
seem in order. First, the 450 or so extant lines of Empedoc1es rep­
resent less than one-tenth of his poetic output, if we are to trust 
Diogenes Laertius' figure of 5,000 lines in total,26 and even on the 
most conservative estimates of Empedocles' total output,27 not 
more than one-fifth. Or supposing (as I am inclined to suppose) 
that Lucretius' interest was exclusively in the On Nature, what is 
extant of that is still likely to be barely one-fifth-roughly 400 lines 
out of 2,000. 28 This raises the probability that if we had Empedocles' 
poems intact a great deal more Empedoclean influence would 
come to light, and our understanding of the DRN be immensely en­
riched. 

Second, I would suggest that Lucretius is likely to owe rather 
more to Empedocles in terms of poetic techniq ue than is generally 

25 E.g. Furley; also: W. Kranz, "Lukrez und Empedokles," Philologus 96 (1944) 
68-107; C. J. Castner, "De Rerum Natura 5.101-103: Lucretius' Application of Em­
pedoclean Language to Epicurean Doctrine," Phoenix 41 (1987) 40-49. I have not 
seen F. Jobst, Ober das Verhiiltnis zwischen Lukretius und Empedokles (diss. 
Munich 1907), but I understand from Don Fowler that it anticipated Kranz's most 
important findings. For other studies, see Tatum (n.32 infra) 178 n.5. 

26 D.L. 8.77; for discussion see Osborne 28f. 
27 Cf. Wright 21. 
28 2,000 lines seems to be the figure for the length of the physical poem given by 

the Suda, S.'ll. "Empedocles" (=3111.2 D.-K.), despite the slightly odd grammar. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

DAVID SEDLEY 277 

recognised. For example, at 1.271-97 Lucretius argues for the 
corporeality of air by means of an intricate analogy between the 
destructive power of wind and that of water. David West has 
observed that the number of distinct points of correspondence 
between the description of the wind and the description of the 
water greatly exceeds that normally found in the similes of Homer 
and Apollonius.29 Lucretius is thus, in West's terminology, a practi­
tioner of the "multiple-correspondence simile/' a legacy that he 
was to pass on in turn to Virgil. What I would myself add is that, 
although Homer and Apollonius may offer no adequate model for 
the technique, Empedoc1es does. In his description of the eye's 
structure and function as analogous to those of a lantern (B 84), 
Empedoc1es reinforces the idea with a set of carefully engineered 
correspondences between the two halves of the simile.30 As in 
Lucretius, so already in Empedoc1es, the multiplicity of correspon­
dences has an argumentative motive, and not merely a descriptive 
one: the more correspondences there are, the more persuasive the 
analogy will be. Here then is a technique, singularly germane to 
philosophical poetry, which has almost certainly passed from Em­
pedoc1es, through Lucretius, into the Latin tradition. 

Lucretius' reverence for Empedoc1es is evident in the paean of 
praise with which he prefaces his criticism of Empedoc1es' four­
element theory at 1.716-41: 

Of these [sc. the four-element theorists] the foremost is Empedocles 
of Acragas, born within the three-cornered land coasts of the island 
[Sicily] around which the Ionian Sea, flowing with its great windings, 
sprays the brine from its green waves, and from whose boundaries 
the rushing sea with its narrow strait divides the coasts of the land of 
Italy with its waves. Here is destructive Charybdis, and here the 
rumblings of Etna give warning that they are again gathering the 
wrath of their flames so that her violence may again spew out the fire 
flung from her jaws and hurl once more to the sky the lightning 
flashes of flame. Although this great region seems in many ways 
worthy of admiration by the races of men, and is said to deserve visit-

29 D. West, "Virgilian Multiple-Correspondence Similes and their Antecedents," 
Philologus 114 (1970) 262-75. 

30 These are contained principally in the close linguistic parallelism of lines 4f 
with the final two lines. For textual problems in the passage, Wright should be 
consulted. For comparable prose uses of complex analogy in Hippocratic authors, 
cf G. E. R. Lloyd, Polarity and Analogy (Cambridge 1966) 345-48. I discuss the 
passage more fully in a forthcoming paper. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

278 THE PROEMS OF EMPEDOCLES AND LUCRETIUS 

ing for its wealth of good things and the great stock of men that for­
tify it, yet it appears to have had in it nothing more illustrious than 
this man, nor more holy, admirable, and precious. What is more, the 
poems sprung from his godlike mind call out and expound his il­
lustrious discoveries, so that he scarcely seems to be born of mortal 
stock. 

But this man and the greatly inferior and far lesser men whom I 
mentioned above, although in making their many excellent and god­
like discoveries they gave responses in a holier and much more certain 
way than the Pythia who makes her pronouncements from Apollo's 
tripod and laurel, nevertheless came crashing down when they dealt 
with the elementary principles of things. Great as they were, their fall 
here was a great and heavy one. 

This is remarkable praise to lavish on a philosopher who did, after 
all, radically misconceive the underlying nature of the world. 
Where does the emphasis lie? Lucretius speaks highly both of Em­
pedocles' "illustrious discoveries'" (praeclara reperta), and of his 
poetry, which is so sublime as almost to prove his divinity-an 
honour that in the end Lucretius will reserve for Epicurus alone. 31 

With regard to Empedocles' "discoveries,'" I am inclined to agree 
with those who hold that Lucretius is commending above all the 
clarity of their exposition, especially by contrast with the obscuri­
ties of Heraclitus denounced in the preceding passage. 32 This, I 
would further suggest, is supported by the closing remarks in the 
passage translated above, where Lucretius expresses his approval 
both of Empedocles and of his "lesser'" colleagues in the pluralist 
tradition33 for revealing their findings "in a holier and much more 
certain way than the Pythia who makes her pronouncements from 
Apollo's tripod and laurel'" (738f). This has standardly been under­
stood as crediting those philosophers with an authority comparable 

31 First at 3.15. It is unwise to be too confident that Lucretius is alluding to 
Empedocles' own profession of divinity at the beginning of the Kaeap~o{, if, as I 
would maintain, his interest is otherwise focused entirely on Empedocles' 0 n 
Nature. But the legend of Empedocles' plunge into Etna in a bid to establish his 
own divinity was probably well enough known by this date to give the remark 
extra point (cf. Wright 15f and Hor. Ars P. 463-66; for its origin as a biographers' 
fiction, see A. Chitwood, -The Death of Empedocles," AlP 107 [1986] 175-91). 

32 E. D. Kollmann, -Lucretius' Criticism of the Early Greek Philosophers," 
StClass 13 (1971) 79-93, and especially W. J. Tatum, -The Presocratics in Book One 
of Lucretius' De Rerum Natura," TAPA 114 (1984) 177-89. 

33 The reference is vague, but perhaps picks up the proponents of two elements 
in 1.712f as well as the four-element theorists of 714f. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

DAVID SEDLEY 279 

to that of an oracle. 34 It seems to me more likely that it expresses a 
contrast-between, on the one hand, the clear and unambiguous as­
sertions of the pluralists, and, on the other, the Delphic ambiguities 
so characteristic of Heraclitus. 35 If so, we must be wary of reading 
into this eulogy of Empedoc1es any special admiration for his teach­
ing as such. It is much more as an eloquent and straight-talking 
expositor of his teaching that he is canonised. Empedoc1es' language 
may be densely metaphorical (as is Lucretius' own), but at least, as 
Lucretius sees it, it lacks the multi-layered evasiveness and trickery 
of Heraclitean prose. About Lucretius' evaluation of Empedocles' 
actual teachings I shalrsay more below. 

What purpose is served in this passage by the fulsome praise of 
Sicily? One object, no doubt, is to compare Empedoc1es favourably 
with that other wonder of Sicily, Etna. 36 But it also has the job of 
illustrating why Sicily was the birthplace of the four-element 
theory.3? The four elements are intricately worked into the trav­
elogue. Empedoc1es was born within Sicily's "land coasts," terra­
rum ... in oris (717)-and here terrarum is not an "otiose addition" 
(Bailey), but Lucretius' way of identifying the land of Sicily with the 
element earth. The elements water and fire are abundantly in evi-

34 It is prima facie implausible that an Epicurean should speak with implicit 
approval of oracular authority. The textual support for such a reading is usefully 
assembled by M. F. Smith in the revised Loeb edition of Lucretius (London/ 
Cambridge [Mass.] 1975), 60£ n., but it does not persuade me. In Epicurus S V 29, 
XP'l0'J.lC!l5£tv is associated with unintelligibility, thus favouring the alternative inter­
pretation proposed here. The Ciceronian texts (Fin. 2.20 [Smith's -1.7.20" is a mis­
print], 102; Nat.D. 1.66) do use oracula of philosophical pronouncements (some of 
them Epicurean), but only in the mouths of Epicurus' critics. That leaves the 
epigram of Athenaeus (ap. D.L. 10.12), which speaks of Epicurus not as himself 
oracular but as inspired either by the Muses or by the Delphic oracle. 

35 For certus;;:;::'unambiguous' see OLD s.'ll. 9. The same sense fits perfectly into 
5.111£, where these lines recur: Lucretius is saying that his prediction that the 
world will one day perish is a firm and unambiguous one, unlike those associated 
with the Delphic oracle. For Heraclitus' 'Delphic' ambiguity, cf. his 22B93 D.-K. 
As for sanctius, in a comparison with an oracle this must primarily imply 'holier', 
but the basic meaning of sanctus (from sancire) is 'ratified' or 'confirmed', and it 
also has connotations of 'above board' or 'honourable' (OLD S.'ll. 4). 

36 If the thesis developed below is right, it may not be too fanciful to see in the 
imminent new eruption of Etna (722ff) a hint at the scheduled rebirth of Empedo­
clean poetry. 

37 This was well spotted by L. MacKay, - De Rerum Natura 1.717 sqq., JJ Latini­
tas 3 (1955) 210, and J. M. Snyder, -Lucretius' Empedoclean Sicily," CW 65 (1972) 
217f. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

280 THE PROEMS OF EMPEDOCLES AND LUCRETIUS 

dence in the descriptions of the surrounding sea, of the whirlpool 
Charybdis, and of the flames of Etna (717-25). Finally (725), those 
flames are borne "to the sky'" (caelum). Now the sky, as the abode 
both of air and of the heavenly bodies, might in principle symbolise 
either of the elements air and fire. What surely clinches its identifi­
cation with air, and thus completes the catalogue of four elements, 
is the fact that Empedocles himself uses "sky'" (oupavoc;) as a name 
for his element air (B22.2). 

And the Empedoclean influence goes deeper still. The very idea 
of using individual phenomena like sea, rain, wind, and sun to 
symbolise the four elemental stuffs is thoroughly Empedoclean. So 
too is the poetic device of interweaving the four elements into the 
language of a descriptive passage: we have already seen Empedocles 
do the same at B 115, when he described the tossing of the fallen 
daimon from aether (:::air) to sea, to land, to the sun's rays, and then 
back once more into the eddies of the aether. 

At the very least, then, Lucretius' description of Sicily reveals his 
intimate knowledge and exploitation of Empedoclean poetry. And 
it would be unwise to rule out the further possibility that it is itself a 
direct imitation of a lost passage of Empedocles. 

IV. The Enigma of Lucretius' Proem 

We are now ready to turn to the most hotly and inconclusively 
debated passage in Lucretius, the proem to Book I.J8 It is struc­
tured as follows: 

1-20: praise of Venus as Aeneadum genetrix and the life force of all 
nature; 

21-28: prayer to Venus to inspire Lucretius' poem, because she alone 
is responsible for making things pleasing, and because Mem­
mius has always been her favourite; 

29-43: prayer to Venus to intercede with her lover Mars and bring 
peace to the Roman republic; 

44-49: it is not in the divine nature to concern itself with our affairs; 

38 The huge bibliography on this passage prominently includes F. Giancotti, II 
preludio di Lucrezio (Messina 1959); K. Kleve, "Lukrez und Venus (De rerum 
natura I, 1-49)," SymbOslo 41 (1966) 86-94; E. J. Kenney, Lucretius (:Greece and 
Rome: New Surveys in the Classics 11 [1977]) 13-17; D. CLAY, Lucretius and 
Epicurus (Ithaca 1983 [hereafter 'Clay']) 82-110; and all the major commentaries. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

DAVID SEDLEY 281 

50-61: programmatic address to Memmius about the content of the 
poem; 

62-79: praise of Epicurus' intellectual achievement; 
80-101: attack on the evils of religion, as illustrated by the sacrifice of 

I phigeneia; 
102-35: warning to Memmius not to be enticed by false religious tales 

about the survival and transmigration of the soul; 
136-45: the difficulty of Lucretius' poetic task. 

The most enigmatic feature of the proem lies in the first three 
subdivisions, 1-43. How can Lucretius, as an Epicurean, praise 
Venus as a controlling force in nature, and even beg her to inter­
vene in human affairs? In Epicureanism, the gods emphatically do 
not intervene in any way in human affairs-as Lucretius himself 
paradoxically goes on immediately to point out (44-49=2.646-51). 

To respond that the proem's treatment of Venus is allegorical is 
not in itself a solution to the puzzle. As Lucretius himself warns at 
2.655-60, allegorical use of divinities, e.g. 'Neptune' for the sea and 
'Ceres' for corn, is permissible only if one avoids any false religious 
implications. Although Venus might, on this principle, get away 
with symbolising nature, or even perhaps Epicurean pleasure, 39 the 
opening address to her as ancestress of the Romans can hardly be 
judged equally innocent, nor can the prayers to her to intervene in 
Roman affairs and to inspire Lucretius' poetry. It is not that these 
allegorical explanations do not carry any weight at all. I think there 
is much truth in them. But the most they can do, for readers who 
have read on and been surprised to learn that this is an Epicurean 
poem, is mitigate their bafflement. The question remains, what can 
have impelled Lucretius to start out so misleadingly, totally disown­
ing the attitude to the gods that the rest of the poem will so ener-

39 The suggestion of E. Bignone, Storia della letteratura latina II (Florence 1945) 
437-44 (which faces the difficulty that Lucretius' Venus controls all natural 
coming-to-be [esp. 21ff], not just animal reproduction). Cf the recent proposal of 
E. Asmis, "Lucretius' Venus and Stoic Zeus," Hermes 110 (1982) 459-70, that 
Venus is here an Epicurean deity invented to take over the role assigned to Zeus 
by the Stoics. I myself share the doubts of Furley, in "Lucretius and the Stoics," 
BICS 13 (1966) 13-33 (=Cosmic Problems 183-205), whether Lucretius was at all 
concerned to counter the Stoics. So far as concerns theology, the only two even 
superficially Stoic views criticised in the DRN are that the world is animate and 
divine, and that it was created for the sake of men (5.110-234); and I doubt if these 
are meant to be Stoic, since both are introduced by Lucretius as purported grounds 
for the thesis, foreign to mainstream Stoicism, that the world (although apparently 
created) is everlasting. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

282 THE PROEMS OF EMPEDOCLES AND LUCRETIUS 

getically promote? It is hardly an exaggeration to say that he spends 
the remainder of the poem undoing the damage done by the first 
forty-three lines. 

v. Furley's Thesis 

In short, the opening of the proem simply is not like Lucretius. 
But it is very like Empedocles. In an outstandingly important study 
of the proem, David Furley has observed the high level of 
Empedoclean content to be found in it. 40 My object here will be to 
augment his observations with further evidence of Empedoclean 
echoes, and, finally, to propose a very different explanation for 
their presence here. 

First, notice the by now familiar technique of working the four 
elements into a descriptive passage. The poem begins as follows 
(1-5): 

Ancestress of the race of Aeneas, delight of men and gods, nurturing 
Venus, who beneath the gliding beacons of the sky pervade the ship­
bearing sea and the crop-carrying lands, because it is due to you that 
every race of animals is conceived and born into the sunlight .... 

Planted in the text already are references to the sky (which we have 
seen to represent the element air in Empedoclean imagery),41 to the 
heavenly bodies and the sunlight (i.e. fire), to the sea, and to the 
land. We then launch into a second catalogue of the same four (6-9): 

From you, goddess, and your approach the winds and the clouds of 
the sky flee away. For you the creative earth pushes up sweet flowers. 
For you the sea's surface laughs, and the sky, made calm, shines with 
diffused light. 

Again, the four elements feature: the winds and clouds of the sky, 
the earth, the sea, the sunlight. And if all this is still not enough, we 
need only move on to 29-43, Lucretius' prayer to Venus to inter­
cede with her lover Mars. It has long been recognised that here we 
have a striking allusion to the joint-protagonists of Empedocles' 
physical poem, Love and Strife-whom Empedocles himself some­
times calls Aphrodite and Ares. 

40 Supra nA. The range and depth of Empedoclean nuances in the proem are 
further enriched by Clay 22f, 49ff, 82-110, 253-57. 

41 I offer this as a ground for going beyond Furley and detecting all four 
elements even in 1-5. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

DAVID SEDLEY 283 

Furley has noted two other Empedoclean echoes in the proem, 
to which we will come shortly. But first the question must be 
asked: why should an Epicurean poem start with an Empedoclean 
prologue? 

It is here that I part company with Furley. He argues that 
Lucretius' act of piety to Empedocles is the acknowledgement of a 
philosophical debt. Although Lucretius was himself a committed 
follower of Epicurus, Furley suggests, he recognised Empedocles 
as the inaugurator or champion of two traditions to which, as an 
Epicurean, he too adhered. The first of these is the insistence on 
absolutely unchanging physical elements. The second is the rejec­
~ion of a .teleological world-view, with all its implications of divine 
InterventIOn. 

But this could hardly explain Lucretius' decision to open with a 
tribute to Empedocles. No reader of the proems to Books 3, 5, and 
6 can doubt that Lucretius' other philosophical debts pale into 
insignificance when compared with his acknowledged dependence 
upon Epicurus. Why then would he give his putative philosophical 
obligation to Empedocles the undeserved and thoroughly mislead­
ing prominence that it gains from arosition at the poem's opening? 

Moreover, the unwritten rules 0 philosophical allegiance in the 
ancient world do not normally permit the imputation of authority 
to anyone other than the founder of your own school, or, at most, 
to his own acknowledged forerunners. 42 The Epicurean school was 
second to none in observing this principle. It seems certain that 
Empedocles was not regarded by Epicurus or his successors as any 
sort of philosophical forerunner; and even an acknowledged fore­
runner like Democritus was treated with scant respect in the 
school. 43 If Lucretius broke with the standard pattern of philosophi-

42 Argued in D. Sedley, "Philosophical Allegiance in the Greco-Roman World," 
in]. Barnes, M. T. Griffin, edd., Philosophia Togata (Oxford 1989) 97-119. 

·43 For Democritus as an acknowledged precursor of Epicurus, see Pluto Adv.CoL 
1108E-F. Epicurean attacks on Empedocles include those of Hermarchus (see F. 
Longo Auricchio, Ermarco, Frammenti [Naples 1988] 66-73, 92-99, 125-50, and P. 
A. Vander Waerdt, "Hermarchus and the Epicurean Genealogy of Morals," 
TAPA 118 [1988] 87-106, esp. 89f n.13) and Colotes (Plut. I111F ff); see also Cic. 
N at.D. 1.29; and Diogenes of Oenoanda 5 1I.2-4 Chilton, with the further passages 
assembled by Vander Waerdt. Here and in what follows I am talking about the 
treatment of other philosophers by Epicurus' followers. I do not believe that Epi­
curus himself was so ungenerous: see D. Sedley, &Epicurus and his Professional 
Rivals, JJ in ]. Bollack and A. Laks, edd., Etudes sur l'epicureisme antique (= 
CahPhiloll [1976]) 119-59. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

284 THE PROEMS OF EMPEDOCLES AND LUCRETIUS 

cal allegiance, that would be a most exceptional departure for which 
we would need strong evidence. His declarations of absolute 
loyalty to Epicurus, as the very first philosopher to liberate men 
from fear of the divine,44 hardly encourage any such speculation. 

Even on the two philosophical issues picked out by Furley, ele­
ment theory and anti-teleology, it is doubtful whether Lucretius or 
any other Epicurean would have been as generous in acknowl­
edging Empedocles' contribution as Furley proposes. Indeed, so 
far as concerns element theory, Lucretius is emphatic at 1.734-41 
(translated supra 278) that this is not a topic on which Empedodes 
acquitted himself with distinction. 

That there is something, singular or plural, that somehow persists 
through all cosmogonical and other changes is common ground for 
all physical philosophers from Anaximander on. No doubt Emped­
odes' elements were more emphatically unchanging than those of 
his predecessors. At least, he says that as the elements intermingle 
they both become different things at different times and remain 
always alike (B 17.34f). He probably means that they form different 
compound substances but nevertheless retain their own distinctive 
properties in the mixture. But other interpretations were possible­
for example, that in mixtures the elements do retain their original 
properties, but that these remain dormant until the compounds 
separate out again. And, at any rate, I see little sign that Lucretius 
was prepared to give him the benefit of the doubt on this point. In 
criticising the four-element theory, he makes no gesture of respect 
even for the well-advertised indestructibility of Empedodean ele­
ments (B 8, 9, 12): on the contrary, his principal ground for rejecting 
the theory is that stuffs like earth, air, fire, and water are inevitably 
perishable (1.753-62). As for their unchangeability, he mentions this 
as no more than a possible interpretation of the theory, and one that 
would rob it of what little explanatory power it has (770-81). 

Does Empedocles fare any better in Lucretius' eyes as a cham­
pion of anti-teleology? It cannot be denied that Aristotle casts him 
in that role: in defending the teleological structure of organisms, 
Aristotle contrasts his view with the zoogonical thesis of Emped­
odes that originally a set of randomly composed monsters sprang 
up-graphically described by Empedocles as «ox-children man­
faced "45-of which only the fittest survived. This anticipation of 

44 Especially 1.62-79 (see below), 3.1-22,5.9-13. 
45 B61.2. Cf. Arist.Phys. 198b32, 199b10-12. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

DAVID SEDLEY 285 

one of the principles of Darwinism has earned Empedocles wide­
spread respect, including, it is sometimes suggested, the respect of 
the Epicureans. For Lucretius testifies (5.837-77) that they adopted 
a similar-sounding theory of the survival of the fittest as their basis 
for the origin of species. 

I would not want to deny the probability of a historical link be­
tween the Empedoclean and Epicurean theories. But it is a large 
leap from that to the supposition that the Epicureans acknowledged 
a debt to Empedocles. Indeed, it can be precisely in those cases 
where a school is drawing on the ideas of another that it is most at 
pains to minimise the resemblance and to stress its own originality. 
This appears to have been the Epicurean attitude to the Empedocle­
an theory of evolution. Plutarch (Adv. Col. 1123 B) tells us explicitly 
that the Epicureans derided Empedodes' "ox-children man-faced." 
And well they might, for Empedodes' monsters were· themselves 
the bizarre product of random combinations of limbs and organs 
that in an even earlier stage had sprung up and wandered about on 
their own (A 72, B57)! There is nothing like this in the Epicurean 
theory, as we hear about it from Lucretius; and I can see no attempt 
in Lucretius Book 5 to restore to Empedodes the credit which the 
Epicurean school traditionally denied him. 46 

Indeed, since Lucretius certainly knew Empedodes' physical 
poem at first hand and did not have to rely exclusively on Aristo­
telian-influenced doxography,47 it certainly should not be assumed 
that he read Empedodes as a pioneering opponent of teleology. If 
Aristotle chooses Empedocles rather than the far more suitable 
Democritus for that role, it is surely because Empedodes, perhaps 
alone among the Presocratics, has actually supplied him with an 
illustration of what a non-teleological explanation of an organism 

46 Furley (61 with n.15) supports his thesis with the claim that Lucretius 5.837-41 
is a translation of Empedocles B 57. Although it may pointedly recall the Em­
pedoclean lines, it is hardly a translation. Where Empedocles describes isolated 
limbs, Lucretius describes whole organisms with congenital defects-and that 
represents a crucial difference between the two zoogonical theories. 

47 Cf Clay 22f, 289f nn.43f. W. RosIer, "'Lukrez und die Vorsokratiker: dox­
ographische Probleme im 1. Buch von 'De rerum natura'," Hermes 101 (1973) 
48-64, stresses Lucretius' use of doxography in his critique of Empedocles at 1.714-
829; but this is, I believe, a special case, in so far as the passage is almost certainly 
based on Epicurus' own criticism of earlier physical theories in On Nature 14 and 
15, which in turn probably relied heavily on Theophrastus' $u0"1.1crov 06~m: see D. 
Sedley, "'The Character of Epicurus' On Nature," Aui del XVII Congresso In­
ternazionale di Papirologia I (Naples 1984) 381-87. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

286 THE PROEMS OF EMPEDOCLES AND LUCRETIUS 

would look like. 48 It does not follow that Empedocles' own inten­
tion, taken in context, came over as anti-teleological. 49 As is well 
known, he is supposed to have postulated four stages of animal evo­
lution, of which the compounding of the ox-children man-faced 
was only the second. Either in the first stage, that of solitary animal 
parts, or perhaps in the third stage, that of the so-called "whole­
natured forms," he described the creation of individual animal parts 
in terms that could hardly have won him the friendship of an anti­
teleologist like Lucretius. In B 84, already mentioned above, Em­
pedocles describes how Aphrodite SO cunningly created the eye, just 
like a man fitting together a lantern for the preconceived purpose 
of lighting his way at night. Even if one strips from this the 
ngurative personincation of Love as a divine artisan, one is left with 
the impression of an intelligent and purposive creative force. The 
architectonic role of Love in Empedocles' cosmic cycle makes it a 
very hard task indeed to portray him as a pure mechanist. 

Why, then, does Lucretius nevertheless speak approvingly of Em­
pedocles' "discoveries" (1.732f)? To see this in perspective, it is 
important to note that only four lines later he speaks with equal ap­
proval of the "discoveries" of other, unnamed natural philosophers 
whom he brackets with Empedocles. Lucretius is not, in effect, 
singling out Empedocles as a uniquely important authority but is 
expressing an Epicurean's qualified respect for the work of the 

. Presocratic natural philosophers in general. Following Epicurus, he 
applauds the Presocratic tendency to seek physical, as opposed to 
theological, explanations for such cosmic phenomena as celestial 
motions, eclipses, and earthquakes. The Epicurean school's method 
of handling these phenomena was to catalogue with approval all the 
available physical explanations of each, adding that any or all might 
be correct, so that to choose between them would be arbitrary and 
unscientific. Both Epicurus, in his Letter to Pythocles, and Lucre­
tius, in Books 5 (509-770) and 6, thus come to list as possibilities a 
range of explanatory theses deriving from Presocratic philosophers, 
including Empedocles. For example, both Epicurus (Ep. ad Pyth. 
101) and Lucretius (6.204) accept as one of the possible explanations 
of lightning the thesis of Empedocles (A63) that it is fire from the 

48 So too, in a different way, at Arist. Part.An. 640a19ff. 
49 I pass over the further complication that teleology was not yet at this date a 

recognised issue on which sides had to be taken. 
50 B86 confirms that Aphrodite was the artisan in question. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

DAVID SEDLEY 287 

sun trapped in the clouds. It is, I am convinced, only at this level of 
detail that the Epicureans, Lucretius included, are prepared to ap­
plaud the "discoveries" of Empedocles. 

VI. Empedocles as Literary Forebear 

If, then, Lucretius is not thanking Empedocles for the content of 
the DRN, perhaps he is thanking him for its form. There are, after 
all, well-recognised formal correspondences between the two hex­
ameter poems. De rerum natura no doubt translates DEpt <pUO'ECOC;, a 
conventional title for many Greek cosmological texts, including 
Empedocles' physical poem. Indeed, one late source reports Em­
pedocles' title as DEpt <pUO'Eroc; 'trov OV'tCOV ,51 which would be closer 
still to De rerum natura. Lucretius' poem is addressed to a friend, 
Memmius, as Empedocles' On Nature is to his friend Pausanias. 
And both at certain points turn to address an invocation to the 
muse Calliope. 52 

My hypothesis will be as follows: the proem of the DRN is, and is 
meant to be recognised as, an imitation of the proem to Em­
pedocles' DEpt <pUO'Eroc;. 

The letter of Cicero with which we opened constitutes strong evi­
dence that contemporary readers could be expected to recognise 
this imitation, if such it was. For it attests a literary climate in which 
Empedocles was on the list of familiar Greek authors,53 either 
through direct acquaintance or through Latin translations. (Even if 
educated Romans shared Cicero's inability to struggle through to 
the end of Sallustius' Empedoclea, they could be assumed to have 
read the opening.) And it shows us Lucretius being thought about 
by his contemporaries in an Empedoclean context. 

51 The Suda S.71. "Empedocles"=31 A2 D.-K. There is no independent evidence to 
confirm this title, but it seems not implausible for a fifth-century work. Cf Melis­
sus' title nepl cpU(Jeco~ t\ 7tepl 'taU ov'to~, whose genuineness is guaranteed by its sat­
irisation in Gorgias' nept 'tou J.1'h ov'to~ t\ 7tept cpU(Jeco~. Viewed as a variant on this, 
nept cpU(Jeco~ 'troY ov'tcov is entirely apposite for a pluralist manifesto like that of Em­
pedocles. 

52 Emped. B3, 131; Lucretius 6.92-95. See Clay 253-57. 
53 For a judicious discussion of Ennius' possible use of Empedocles, see O. Skutsch, 

The Annals of Quintus Ennius (Oxford 1985) 160, 164 n.18, 260, 394ff, 758. Ovid's 
extensive use of Empedocles in the speech of Pythagoras in Met. 15 will be the 
subject of a forthcoming paper by Philip Hardie, which he has kindly let me see in 
draft. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

288 THE PROEMS OF EMPEDOCLES AND LUCRETIUS 

On my hypothesis, Lucretius' purpose is to establish from the 
outset the precise Greek literary mantle he is assuming (rather as 
Vergil's Aeneid announces with the opening words arma virumque 
cano that it will be a combined Iliad-Odyssey). Literary pedigree 
was a matter of immense importance to Roman poets, and Lu­
cretius, in his poetic manifesto at 1.921ff and his appreciation of 
Ennius' pedigree at 1.117ff, shows himself to be no exception. 

To amplify the hypothesis: Lucretius is imitating Empedocles' 
proem but adapting it, as he goes along, (a) to a Roman patriotic 
theme and (b) to Epicurean philosophy, at the same time steering 
us gently away from Empedodes' actual doctrines. His object? To 
announce himself as the Roman Empedocles-the great Roman 
poet of nature. In short, he is laying claim to a literary, not a philo­
sophical, heritage. For there can be little doubt that it was to Emped­
odes, rather than to the only other plausible candidate, Parmenides, 
that Lucretius looked as his great Greek forebear in the tradition of 
cosmological poetry. 34 

A glaring weakness of this hypothesis will already be obvious. We 
do not have the proem to Empedodes' On Nature.33 How then can 
we say anything at all about its resemblance or otherwise to 
Lucretius' proem? My answer is twofold. First, we are not alto­
gether without evidence about its content, as I hope to show. And 
second, if the proposed hypothesis proves capable of explaining 
features of Lucretius' proem that otherwise remain inexplicable, 
that in itself would provide some degree of confirmation. 

1.1-49. I shall begin my defence of the hypothesis with an exam­
ination of the opening lines (translated supra 282): 

Aeneadum genetrix, hominum divomque voluptas, 
alma Venus, caeli subter labentia signa 
quae mare navigerum, quae terras frugiferentis 
concelebras .... 

The linguistic case for a direct Empedoclean model seems to me a 
rather strong one. The first two words are, of course, a distinctively 

54 This was the comparison that regularly occurred to Roman readers: see e.g. 
Quint. Inst. 1.4.4; Lactant. Div.Inst. 2.12.4. 

55 Van der Ben offers his own wholesale reconstruction of Empedocles' proem. 
Most of it rests, in my view, on pure guesswork. My grounds for rejecting it will 
simply be the arguments I offer below for accepting a different reconstruction, 
based largely on Lucretius. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

DA VrD SEDLEY 289 

Roman invocation. But hominum divomque voluptas has the hall­
mark of Empedoc1ean ancestry. The identical phrase recurs, with a 
small change of syntax, at 6.92-95, in an address to Calliope that has 
long been recognised as an Empedoclean touch on Lucretius' 
partY" Hominum divomque could translate some variant on the 
regular hexameter ending avbprov t£ 9£rov t£, used in Homer's 
formulaic designation of Zeus as "father of men and gods." Such 
reworkings of Homeric locutions and concepts are an integral 
feature of Empedoc1es' poetry.57 And voluptas picks up r1l90cruvll, 
'Delight', used by Empedoc1es, like' Aphrodite', as a title for his 
goddess Love (B 17.24). Next, alma recalls ~d()ropo<;, 'life-giving', an 
attested Empedoclean epithet for Aphrodite (B 151). We then 
proceed, in 2-9, to the elaborate double interweaving of the four 
elements into the hymn. For Lucretius to expect any reader to 
identify these as the Empedoclean four in the very opening lines of 
the poem, without any prior clue, would be wildly optimistic. It is 
far more credible that he found them already present in his Em­
pedoc1ean original. We have already noted that interweaving the 
four elements into a descriptive passage is an authentic Emped­
oc1ean device. 

Line 3 is remarkable for its pair of compound adjectives. Lucre­
tius has a well-known penchant for these quasi-Greek formations,58 
and indeed they elsewhere sometimes combine with Greek loan 
words to build up an evocative context that transports his reader to 
the Greek world. 59 But it is an unusual feature of this pair that both 
accurately translate actual Greek compound adjectives-respec­
tively, navigerum=va:ucri1t0pov and !rugi!erentis=Ka.P1tOCPOpou<; (or 
a participial equivalent from Ka.P1tOCPOPEtV). The line practically 
tumbles into a Greek hexameter unaided: va.ucrt1t0POV 1tov'tov Ka.t 
ya.ia.<; Ka.P1tOCPOpoucra.<;, or the like. 

In line 21 Venus emerges as the controller of all natura, with the 
strongest possible indications that the term here signifies not so 
much 'nature' as 'birth' (through its derivation from nasci, 'to be 
born'). As Diskin Clay has shown, this unmistakably reflects Em-

56 6.94, Calliope, requies hominum divomque voluptas. 

57 See the seminal study of Bollack (supra n.S: I 277ff). Aristotle, in his lost IlEpl 
1t01.11trov, called Empedocles 'Qn1puro; (D.L. 8.57). 

58 Lucretius' compound adjectives are conveniently catalogued by C. Bailey, 
Lucretius, De Rerum Natura I (Oxford 1947) 132ff. 

59 E.g. 2.618-20,4.580--89. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

290 THE PROEMS OF EMPEDOCLES AND LUCRETIUS 

pedodes' characteristic use of q)'6(Jt~ with precisely the same shift 
from the more familiar 'nature' to 'birth'.60 

Beyond these linguistic and conceptual considerations, it is in any 
case eminently plausible that Empedodes' poem should have 
opened with a hymn to Aphrodite. Hesiod's Works and Days, with 
its opening hymn to Zeus, would constitute ample precedent with­
in the tradition of didactic poetry; and it goes without saying that 
Aphrodite would be Empedodes' preferred divinity. In B 128 he 
makes it a mark of the Golden Age, in which among other things 
there was no animal slaughter, that Aphrodite was the only divinity 
worshipped: 

Nor did they have Ares or Strife as a god, nor was Zeus or 
Cronos or Poseidon their king, but Cypris was queen .... Her 
they propitiated with pious images .... 

I am not suggesting that this fragment itself comes from Emped­
odes' proem. But it does reveal a feature of his religious thought 
that Lucretius could himself use to advantage-namely the idea that 
the identity of a person's divinities is a function of that person's 
own moral state.61 If you are a peaceful person, Ares is not your 
god, but Aphrodite is. Lucretius, as an Epicurean, must hold the 
somewhat similar view that the gods' true nature is peaceful, and 
that men's tendency to endow them with angry and warlike tem­
peraments is a projection of their own moral maladjustment. 62 

This may offer us a lead on the much debated lines 44-49, in 
which Lucretius presents the correct Epicurean view of the gods as 
tranquil and detached. These lines occur also at 2.646-51, where 
they appear superficially much more at home, and many editors be­
lieve that they are an intrusive gloss in the proem:63 it seems anoma­
lous for Lucretius to stress the total detachment of the gods from 
human affairs directly after his prayer to Venus to intervene and 
save the Roman republic from war. 

But now imagine an Empedodean proem in which Aphrodite, as 

60 B8. See Clay 83-95, with the parallels cited at 308 n.29. 
61 Cf. B17.23, where Love is ·she by whom mortals think friendly thoughts and 

perform peaceful deeds." 
62 See A. A. Long and D. N. Sedley, The Hellenistic Philosophers I (Cambridge 

1987) 139-49. The point stands whether or not, as argued there, Lucretius was 
wrong to understand Epicurus' gods as objectively real organisms. 

63 See the arguments marshalled by E. Courtney, cQuotation, Interpolation and 
Transposition," Hermathena 143 (1987) 7-18, at 11ff. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

DAVID SEDLEY 291 

Love, is asked to propitiate Ares, as Strife. What Empedocles 
would have intended by this is not so much an attempt to interfere 
with the inevitable yrogression of the cosmic cycle, as a plea to men 
to let their peacefu tendencies calm and suppress the bloodthirsty 
side of their nature. 64 If so, Lucretius would welcome this essen­
tially moral use of myth and prayer, and could readily apply it to the 
current war-torn state of his own country. But since Empedocles 
regards Ares/Strife as a real, if less palatable, god, Lucretius might 
very naturally want to add an Epicurean corrective: that Venus' 
hoped-for propitiation of Mars represents no more than men's 
return to the one true conception of the divine nature as tranquil 
and detached, instead of angry and warlike. Hence the connexion of 
thought found in the text: Venus, make Mars peaceful, because that 
alone is the true nature of divinity. Or, translated into Epicurean 
moral terms: Romans, let your belief in a peaceful god overcome 
your belief in a warlike god, because peacefulness alone is the true 
nature of godlike happiness. The connexion of thought could no 
doubt have been made clearer; but I would be very reluctant to rob 
Lucretius of this important Epicurean modification of Empedo­
clean theology. 

By this stage I am no longer suggesting direct translation or line­
by-line imitation of Empedocles' proem on Lucretius' part, but the 
deployment of the same sequence of themes as occurred in it, for 
increasingly Epicurean purposes. 

1.50-61. The next section of Lucretius' proem is a programmatic 
address to Memmius. He asks Memmius to give him his full 
attention-perhaps an echo of the passage that contained Empedo­
cles' surviving line "Listen to me, Pausanias, son of wise An­
chiteus" (B 1). He then proceeds to outline the content of the poem. 
He will explain to Memmius the character of the heaven and the 
gods, and the elements (56£) 

unde omnis natura creet res auctet alatq ue 
quove eadem rursum natura perempta reso/vat. 

64 Eustathius (Od. 310.33ff, ad Hom. Od. 8.367) may imply that Empedocles 
used the myth of the union of Aphrodite and Ares as an allegory for friendship; 
and since there is no stage within the cosmic cycle itself at which Love and Strife 
unite, the likeliest location for that piece of symbolism would indeed be his proem. 
However, Eustathius' words may mean no more than that some allegorists pro­
posed an Empedoclean interpretation of the myth; cf. Heraclit. Alleg.Hom. 69.8, 
and F. Buffiere, Les Mythes d'Homere et La pensee grecque (Paris 1956) 168-72. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

292 THE PROEMS OF EMPEDOCLES AND LUCRETIUS 

from which nature creates, increases, and nurtures all things, and into 
which that same nature once more resolves them when they are de­
stroyed. 

After this he spends four lines naming his cosmic first principles 
(genitalia corpora, semina rerum, etc. ). 

This dual process, whereby things combine and are once more 
dissolved into their constituents, closely echoes Empedocles' own 
programmatic description in Bt7, a passage that is explicitly attested 
as coming from the opening of his physical poem.6S It begins: 

o UtA. ' EP£ro' tOtE J.lEV 1a.p EV llu~i11h1 J.lOVOV dvat 
EK 1tA£ovrov, tOtE 0' a'\) ot€qm 1tAEOV' E~ EVO~ dvat. 

I will tell a double tale. For at one time there grew to be just one 
thing from many, and at another it grew apart once more to be many 
out of one. 

The symmetrical two-way nature of the process is emphasised re­
peatedly in similarly balanced antitheses for a further fifteen lines, 
after which Empedocles, like Lucretius, proceeds to name the cos­
mic principles underlying the process-the four elements, plus 
Love and Strife. The formal parallelism of the two programmatic 
passages is striking. 

It begins to look highly plausible that Empedocles' proem to On 
Nature, having opened with a hymn to Aphrodite, then continued 
with a programmatic address to Pausanias, of which Bt7 formed a 
part. 

1.62-79. Lucretius' next section is his praise of Epicurus'66 in­
tellectual achievement. At a time when mankind was wretchedly 
oppressed by religion, a certain Greek became the first (primum 
Graius homo, 66) to stand up against its tyranny. Such were his 
mental powers that he was able to break through the "flaming walls 
of the world" and traverse with his intellect the measureless uni­
verse. By reporting back to us the laws that bind and limit natural 
processes, he has broken the power of religion. 

Once more there is a clear Empedoclean model, B 129, almost cer­
tainly referring to Pythagoras: 

65 Simpl. In Phys. 161.14f; see preamble to »17 in D.-K. 
66 I am unpersuaded by the proposal of L. Edelstein, ·Primum Graius homo 

(Lucretius 1.1-149)," TAPA 71 (1940) 78-90, that the reference is a general one to 
the Pre socratic physical tradition. The proems to Books 3, 5, and 6 supply ample 
evidence of Lucretius' belief that Epicurus was the first to make the crucial break­
through, scientific as well as moral. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

DAVID SEDLEY 293 

There was among them a man of extraordinary knowledge, pos­
sessing a vast treasury of understanding, and master of every 
kind of wise deed. For when he reached out with all his under­
standing he easily saw everything that is, over ten and twenty 
generations of men. 

As Furley has pointed out,67 the Lucretian passage unmistakably 
recalls the Empedoclean. Both men are great historical figures, too 
august to be named. And both are praised for their intellectual 
achievement in breaking through the boundaries of ordinary 
human experience-Pythagoras for his recollection of his former 
incarnations,68 Epicurus for his grasp of the nature of the infinite 
universe beyond our own world. 

Doctrinally, it should be noticed, Lucretius and Empedocles are 
veering ever further apart. Epicurus' discoveries, which secured his 
victory over religion, are taking the place of an Empedoclean re­
li&ious. doctrine that is anathema to Lucretius, the doctrine of trans­
mIgratIOn. 

1.80-101. There follows Lucretius' direct attack on the evils of 
religion, illustrated with the example of Agamemnon's sacrifice of 
his own daughter Iphigeneia. Furley is right to point out the clear 
reminiscence of Empedocles B 137, in which Empedocles attacks 
the sin of animal slaughter with the example of a father unwittingly 
sacrificing his own son, who has transmigrated into the body of an 
ox. There is no detailed linguistic imitation, but the close functional 
parallelism of the two pathetic scenes of sacrifice should leave little 
doubt that the one passage is written with the other in mind. 
(Lucretius' description does not, incidentally, appear to be directly 
modelled on any of the accounts of Iphigeneia's sacrifice extant in 
Greek tragedy.) 

1.102-35. Lucretius continues with a warning to Memmius not to 
be confused by superstitious tales, such as those about the survival 
and transmigration of the soul. In the course of making the point, 
he names Ennius as the author of such a confusion-although, he 
adds, Ennius must be given his due as the great innovator who 
brought Greek poetry to the medium of the Latin language. 

67 See also W. Burkert, Lore and Science in Ancient Pythagorean ism (Cam­
bridge [Mass.] 1972) 137; P. R. Hardie, Virgil's Aeneid: Cosmos and Imperium 
(Oxford 1986) 39 n.17. 

68 For the tradition of Pythagoras' multiple incarnations, see Burkert (supra 
n.67) 137ff. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

294 THE PROEMS OF EMPEDOCLES AND LUCRETIUS 

Why did Lucretius choose to include the topic of transmigration 
in his proem? In view of all the Empedoclean echoes we have al­
ready witnessed, it can hardly be a coincidence that Empedocles 
likewise outlined his beliefs about transmigration in his proem. It is 
here that I can at last call upon the findings of the earlier part of this 
paper, in which I defended the attribution of B 115, Empedocles' ex­
planation of his doctrine of transmigration, to the proem of his On 
Nature. If I am right, and Lucretius' attack on transmigration is an 
intended counterpart to Empedocles' exposition of the doctrine at 
the corresponding point in his own proem, he has now moved yet 
further in distancing himself philosophically from his principal 
literary mode1. 69 Where previously we saw him adapting themes 
from Empedocles' proem to his Epicurean philosophy, he is now 
presenting his own matching passage not as an adaption of Emped­
ocles, but as a direct antidote to his teachings. 

1.136-45. Finally we come to the closing section of Lucretius' 
proem, in which he stresses the ma~nitude of his poetic task-a 
task made harder, he says, by the deficiencies of the Latin language 
and the novelty of the subject matter. It is overwhelmingly tempt­
ing to correlate this with the group of fragments (B 8-11, 15) in 
which Empedocles deplores the imprecision of ordinary language 
in speaking of things' being born and dying where there is in reality 
only combination and separation, but adds that he will follow the 
convention. The shared theme of how to cope with the deficien­
cies of one's own language constitutes a strong link between the 
two passages. We have no explicit attribution of these fragments to 
Empedocles' proem, but B 8 is at least cited by Simplicius as coming 
from the opening book of On Nature; and even without the Lucre­
tian parallel the proem has always seemed the likeliest location. 

69 Undoubtedly the dream passage in Ennius' own proem is also directly echoed 
in these lines: see 5kutsch (supra n.53) 12, 155; Clay 310 nA8; J. H. Waszink, "The 
Proem of the Annales of Ennius." Mnemosyne SER. IV 4 (1950) 215-40. esp. 224f. 
50 Lucretius is here distancing himself from Ennius' beliefs, while revering his 
poetry. in a way that pointedly parallels his treatment of Empedocles. (The point is 
redoubled if, as seems likely. Ennius' beliefs were themselves influenced by 
Empedocles: cf Hardie [supra n.67] 17-22. 79-83.) Indeed. anyone who may 
doubt the appropriateness of my distinction between a 'literary' and a 'philo­
sophical' debt to Empedocles should note that just such a distinction operates here 
with regard to Ennius. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

DAVID SEDLEY 295 

VII. Empedocles' Proem 

A little earlier we arrived at the informed guess that Empedocles' 
proem to On Nature opened with a hymn to Aphrodite, followed 
by a programmatic address to Pausanias. We can now, in the light of 
our subsequent findings, ask how it went on. 

Lucretius' proem offers the following sequence of topics in its 
latter part (62-145): (a) Epicurus' intellectual achievement and defeat 
of religion; (b) the evils of religion; (c) the folly of uncritically be­
lieving religious tales, such as those about transmigration; (d) the 
magnitude of Lucretius' poetic task. The thematic link between the 
first three is a perfectly satisfactory one, and the last is, if not di­
rectly connected, still an appropriate enough topic to address in a 
proem. And yet there is something disquietingly specific, not to 
say arbitrary, about the third topic. Why go to such lengths to 
criticise the transmigration thesis in particular, when there are 
countless other offending doctrines? Is it merely in order to in­
troduce a heavily qualified tribute to EnniusFo My preferred ex­
planation has been that the choice and sequence of topics was in 
some measure dictated by a further consideration, Lucretius' desire 
to reproduce the thematic structure of Empedocles' proem. One in­
cidental by-product has been the materials for a scissors-and-paste 
reconstruction of the latter part of Empedocles' own proem. Now 
stand back and look at the result. We have supplied Empedocles 
with the following fluent sequence of topics: (a) Pythagoras' 
achievement in recalling past incarnations: an appeal to authority for 
the doctrine of transmigration; (b) the evils of animal slaughter, 
illustrated by unwitting sacrifice of a deceased and transmigrated 
son: the moral importance of the doctrine of transmigration; (c) the 
origin and nature of transmigration itself; (d) the folly of being 
misled by ordinary usage into supposing that anything literally dies. 

70 I do not mean to deny that direct reaction to Ennius plays a significant part in 
this passage (see previous note). My question concerns the overall thematic struc­
ture of the passage. I would tentatively add that Lucretius is unlikely to have been 
impelled by contemporary philosophical developments to pick transmigration as a 
target. To judge from the evidence of Cicero's Tusculan Disputations, the current 
revival of interest in Plato's immortality doctrine played down reincarnation in 
favour of discarnate survival. Nor does transmigration appear to be an attested 
feature of first-century B.C. neo-Pythagoreanism (for which see J. M. Dillon, The 
Middle Platonists [London 1977] 117-21, and cf P. A. Vander Waerdt, ·Peripa­
tetic Soul-Division, Posidonius, and Middle Platonic Moral Psychology," GRBS 26 
[1985] 373-94, esp.388Q. 



SEDLEY, DAVID, The Proems of Empedocles and Lucretius , Greek, Roman and Byzantine 
Studies, 30:2 (1989) p.269 

296 THE PROEMS OF EMPEDOCLES AND LUCRETIUS 

The thematic coherence of this sequence is extraordinary. It is 
much more tight-knit than the corresponding passage in Lucretius, 
and tells a complete story of its own, one thematically parallel to the 
Lucretian passage, yet utterly unlike it in detailed content. What is 
more, the denial of literal birth and death with which it ends not 
only gives a philosophical basis to the transmigration doctrine that 
precedes it, but also rrepares the ground for the physical exposition 
to follow, which wi! likewise be founded on the Parmenidean tenet 
that nothing literally comes to be or perishes. 71 This emergence of a 
reconstructed Empedoclean proem with a coherence and vitality of 
its own is an additional windfall, which lends welcome support to 
my hypothesis about Lucretius' proem, quite apart from what it 
promises to teach us about Empedocles himself. 

The nature of my case has been essentially cumulative. Every 
main stage of Lucretius' proem has proved to correlate with an Em­
pedoclean original. The first part reads as if it were closely imitating 
an Empedoclean hymn, while the remainder sustains a virtually un­
broken series of thematic links with known or attested passages of 
Empedocles. Moreover, everyone of those Empedoclean originals 
can plausibly be located in the proem of his On Nature, either on in­
dependent evidence, or through its thematic coherence with pas­
sages that have already been located there. 

Lucretius is the servant of two masters. Epicurus is the founder 
of his philosophy; Empedocles is the father of his genre. It is the 
unique task of Epicureanism's first poet to combine these two 
loyalties. And that task is what gives his proem its very distinctive 
character. 72 

CHRIST'S COLLEGE, CAMBRIDGE 

August, 1989 

71 See especially B12. The Parmenidean tenet seems to be applied by Empedocles 
indiscriminately to the soul's survival and to the permanence of the elements: both 
equally are separated. not destroyed. How coherent this conflation is is another 
question. Cf especially Kahn (supra n.7). 

72 Ancestors of this paper have been presented to audiences at Berkeley, UCLA, 
Stanford, and the University of Wales, and a near-final draft was delivered at the 
conference on "Tradition and Innovation in Epicureanism" at Duke University in 
April 1989. I am grateful for invaluable comments received on all those occasions, 
and also in writing from David West, Don Fowler, Geoffrey Lloyd, Robert 
Wardy, Catherine Osborne. Samuel Scolnicov, Philip Hardie, Paul Vander 
Waerdt, Michael Reeve, Ted Kenney, and Bill Furley.