Bryn Mawr Classical Review 1999.10.29
David Sedley, Lucretius and the Transformation of Greek Wisdom. Cambridge: Cambridge University Press, 1998. Pp. 234. ISBN 0-521-57032-8. £35.00.
Reviewed by Gordon Campbell, Jesus College, Oxford (email@example.com)
Word count: 6901 words
David Sedley has done perhaps more than anyone else in recent years in showing the closeness of the relationship between Lucretius and Empedocles, and especially how Empedocles' Physics provides Lucretius with his main poetic model. In this new book he argues that Lucretius takes only his poetic, rather than any philosophical inspiration, from Empedocles, and that Epicurus' On Nature is the sole philosophical source for De Rerum Natura. Further, he argues for a traditional theory that Lucretius' work does not reflect any knowledge of contemporary philosophy and that he does not engage with any philosophical questions later than those of Epicurus' day, describing Lucretius as a 'fundamentalist'.
The book is thus an exercise in Quellenforschung, the investigation of sources, and as such Sedley, as he admits (p. xvi), is returning to a mode of scholarship very nearly abandoned in Lucretian studies in recent years. There was a time when debate was fierce over whether Lucretius' polemics could be proved to be anti-Stoic, in particular, which would show that he engages with philosophy post Epicurus, or whether Lucretius' targets are always the same as those of Epicurus himself. More recently the re-contextualization of Epicurus' arguments by Lucretius has seemed more important: a process in which an anti-Platonic argument of Epicurus', for instance, may well be seen by a first-century BC Roman reader as anti-Stoic, thus making the question of the original focus of the polemic less relevant. Accordingly, with the rise of this sort of thinking the investigation of Lucretius' sources has been somewhat neglected. Sedley justifies his approach with an admonition to Lucretian scholars, one that must be guiltily accepted, that despite the great advances in knowledge of Epicurus' On Nature in recent years, little of it has found its way into Lucretian studies. This book sets out to redress the balance. If the following review addresses mainly the parts of the book that fall most within my own area of interest, in Lucretius 5 and cosmology, I hope the author, and the readers of BMCR, will forgive me.
The book is broadly in three sections: the first two chapters examine the relationship between Lucretius and Empedocles: chapter one reconstructs Empedocles' proem to book one of the Physics and argues that Lucretius' proem to book one of D.R.N. closely follows the order of topics in Empedocles' proem, while the second chapter studies Lucretius' literary and stylistic debt to Empedocles. Chapter three provides a bridge between the Empedoclean and Epicurean sections of the book. Here Sedley describes Lucretius as a 'Fundamentalist' and shows that he fails to engage with contemporary critics of Epicureanism.
The fourth chapter reconstructs the contents of books I-XV of Epicurus' On Nature, which is then in chapter five compared to the structure of D.R.N. Here Sedley shows how closely the two match and argues that On Nature I-XV provide the sole Epicurean source for D.R.N.
Chapter six examines the Theophrastean influence on Epicurus and thus indirectly on Lucretius, and chapter seven shows how Lucretius re-worked book one of D.R.N. and re-ordered the topics he had originally found in On Nature I and II.
Sedley builds up his argument cumulatively to support the central thesis of the book that Lucretius is a philosophical fundamentalist who adheres closely to Epicurus' teaching in On Nature I-XV but who derives his poetic technique largely from Empedocles and does not hesitate to edit and re-order his Epicurean source material and so produces a masterpiece that transcends Epicurus. Thus each chapter is linked by this thread of argument and each builds upon the proofs of the preceding chapter, making the book as a whole a cumulative work. This technique will inevitably lead to attacks on each individual piece of evidence used to build up the entire structure and thus on the central thesis, and it does sometimes seem that Sedley is too ready to trust the solidity of the proofs that he has established in building up the structure of his argument. However, he adduces so much detailed material and argues so cogently at each stage that, even if we are not entirely convinced by the argument of the book as a whole, there is so much argument of excellent quality in it that often the central thesis may be set aside while still leaving the details of great value. In a similar way as we tend to do with Thomas Cole's excellent Democritus and the Sources of Greek Anthropology, we may disagree with the conclusions while relying upon the work as a crucial contribution to the field.
In chapter one Sedley's purpose is to show that Lucretius serves two masters, Epicurus and Empedocles, but, unlike David Furley ('Variations on Themes from Empedocles in Lucretius' Proem', BICS 17 (1970) 55-64) who sees Empedoclean philosophical influence on Lucretius, Sedley considers that Empedocles provides only poetic inspiration. In this argument he seems to be returning to the old idea of an opposition between philosophy and poetry in Lucretius that, as he says in his introduction, has somewhat died away in recent years. In separating Empedoclean and Epicurean influences in this way I feel he creates an opposition that is far too clear-cut to be really sustainable, and that we may indeed see Empedoclean influence in the philosophy as well as the poetry, especially since as I see it the poetic medium strongly informs Lucretius' mode of thought. We should also not rule out the possibility that Lucretius sees a stronger Empedoclean influence on Epicurus than we can, and in a circular progression injects further Empedoclean material into his poem on that authority. To take one of Sedley's examples: he argues that Lucretius' praise of Empedocles (1.716ff) in terms only slightly less lavish than the praise of Epicurus (1.62ff) can refer only to Empedocles' value as a poetic inspiration rather than as an anti-teleological forerunner of Atomism, and points to fragment DK 31 B84 in which Empedocles describes Aphrodite constructing the eye for the purpose of seeing in teleological terms. The charge of teleology against Empedocles may be answered by noting how difficult it is to describe the creation of the eye especially, even in evolutionary theory (to which Empedocles does not have recourse), without using teleological language (witness Richard Dawkins' use of his 'Blind Watchmaker' metaphor). Teleological language is still thought by many biologists to be a useful explanatory tool, and teleology itself a useful heuristic method, even in explicitly non-teleological contexts (see Ernst Mayr, 'Cause and Effect in Biology', Science 134 (1961) 1501-6). Gould and Lewontin (especially in 'The Spandrels of San Marco and the Panglossian Paradigm', Proceedings of the Royal Society of London. B205 (1978) 581-98) have long waged a fierce campaign against teleology since they consider that modern biology is riddled with it and that it does great damage to scientific thought.
Further, Sedley explains that Aristotle presents (Physics 198B) Empedocles' scheme of the origin of species as a non-teleological one, simply as an example of what such a scheme would look like if there were one, and that accordingly we need not see Empedocles' scheme as non-teleological. This can hardly account, however, for Simplicius' (in Phys. 371.33-372.11) more detailed explanation of Empedocles' scheme, which makes it quite clear that species and their characters arise by chance combinations of limbs, Aphrodite/Love being unable to ensure correct parts fit together. This is thus inevitably a non-teleological scheme. Simplicius' comment: 'all the ancient physicists seem to be of this opinion who say that material necessity is the cause of generation; and among the later ones this is also the case with the Epicureans', directly links Lucretius' undoubtedly anti-teleological origin of species closely with Empedocles and Empedocles with the other ancient Physicists, who, Aristotle complains, concentrate on the material cause and neglect teleology (De Part. An. 640b 5ff). Further, Plato in Timaeus (91d 6-e1) describes the origin of birds as a transformation of 'light-minded' men who think that visible evidence is all that is necessary for cosmology. I see this as an oblique reference to Presocratic materialists who deny teleology, and so it seems to me that Sedley's claim that in Empedocles' day teleology was not an important issue is unlikely to be true.
If we trust Simplicius (as we ought to since, as Professor Wright says on p. 51 of her edition, he alone of the commentators ever gives us extra quotations from Empedocles and so must be assumed to be working from a text of Empedocles) the Epicurean and Empedoclean theories of the origin of species are fundamentally the same in that they both rely on chance in the formation of animals and their characters. Lucretius' list of monsters thrown up at random by the new earth (5.837ff) has long been recognised as highly Empedoclean, and the fact that some of these monsters are born without eyes should indicate that the eye, in Empedocles as in Lucretius, is an organ that is formed randomly by chance as are all other organs and characters. It seems unlikely that Epicurus would present his zoogony in such Empedoclean language given his famous antipathy toward poetry, and so we should see the Empedoclean treatment of the origin of species in Lucretius as his own return to (one of) the sources of Epicurus own doctrine. This gives a much more complex and involved relationship between Lucretius, Empedocles and Epicurus than Sedley presents since both Lucretius and Epicurus are influenced separately by Empedocles.
Recently some new fragments of Empedocles, the first fragments from a direct transmission, have been edited and published by Alain Martin and Oliver Primavesi (L'Empedocle de Strasbourg. (P. Strasb. gr. Inv. 1665-1666), Berlin/Strasbourg, 1999) and Sedley had a preview of them in writing his book. He takes the evidence of the Strasbourg fragments and, together with an analysis of the proem to book one of D.R.N., produces a reconstruction of Empedocles' proem to book one of the Physics that follows the order of topics in Lucretius, and he shows how each of those topics is found in Empedocles. Empedocles' proem would thus begin with an invocation to Aphrodite, which Lucretius uses as the basis for his invocation to Venus, and the rest of the proem would match closely the topics of Lucretius' proem. This involves a re-allocation of certain fragments from their previous placement in the Katharmoi, to the proem of the Physics: the praise of Pythagoras and his achievements (DK 31 B129) which is paralleled by Lucretius' praise of Epicurus in his proem (1.62-79), the evils of meat-eating and animal sacrifice, exemplified by a father unwittingly sacrificing his son transmigrated into an ox (DK 31 B137) and paralleled by the sacrifice of Iphigenia in Lucretius (1.80-101), and the origins and nature of transmigration itself (DK 31 B115) which Lucretius argues against in his proem (1.112-135). Perhaps most importantly, the inclusion of the Pythagorean fragments in the proem shows that Empedocles' physics and his religion are by no means incompatible, as was once thought. This may have significant consequences for our understanding of the relationship between the physics and the psychology in Empedocles' cosmic system, and also for the Pythagorean status of Plato's Timaeus, in which, of course, we find a 'Pythagorean' account of cosmogony with significant appropriation and inversion of Empedocles' physics.
This rearrangement, if correct, would produce a highly satisfying proem to the Physics, and a close match between Lucretius' and Empedocles' proems would go a long way towards explaining Lucretius' opening hymn to Venus and his prayer for her to intercede and grant peace to the world by seducing Mars, which has long puzzled scholars given the Epicurean insistence on the gods' lack of interest in the world. Sedley shows how Venus and Mars represent Empedocles' two cosmic forces, Love (Aphrodite) and Strife (Ares), and that the prayer to Venus is very probably modelled on an Empedoclean original. The fact that this would produce a false polarisation of Love and Strife as respectively one creative and one destructive force, which does not match Empedocles' cosmology, may be turned to good account in explaining Lucretius' hieros gamos of Venus and Mars: Lucretius would have good authority for beginning his cosmological poem with an ethical metaphor that, as we read further, we learn is not literally reconcilable with the cosmological truth. Sedley quotes D.R.N. 1.56-7, pointing to their apparently Empedoclean nature in support of this argument:
unde omnis natura creet res auctet alatque
quove eadem rursum natura perempta resolvat
from which nature makes all things, increases and nourishes them, and into which the same nature again reduces them when dissolved
Perhaps a stronger argument could be made by saying that Lucretius appears to argue implicitly against Empedocles' two cosmic forces, Love and Strife, by saying that in fact it is nature that creates all things and the same nature (eadem ... natura) that dissolves all things into their component parts. This would act as a corrective to the Empedoclean picture we have been given just a few lines earlier of Venus as creatrix and Mars as the destructive force: natura in fact has both creative and destructive roles. And so the Empedoclean origin of the union of Venus and Mars would be strengthened.
However, having praised the reconstruction, I must point out that Sedley's technique of reconstruction may be open to the charge that it employs a certain circularity: we would like to see Lucretius following Empedocles' proem because this would have great value in explaining certain problems in Lucretius, and so we build a proem for Empedocles that matches that of Lucretius, but the fact that we find the topics of Lucretius' proem in Empedocles need not ensure that Lucretius found them in the proem. He may well have constructed his proem out of various Empedoclean materials he found scattered throughout Empedocles' Physics, using a technique very similar to that of Sedley, and have produced a condensed survey of Empedocles that would characterise his work for the reader as 'Empedoclean' and so grant it a glamour and authority that a plain Epicurean presentation would lack. However, I do not think that the technique of reconstruction is fundamentally flawed. As I see it, the chief problem is that any reconstruction of Empedocles' work is based necessarily on slender evidence, and the question must be whether we should, because it is slender, refrain from using it or, as Sedley has done, use it as the basis for a reconstruction since it is the best evidence we have. On this problem, rather than making an absolute judgement, I would be tempted to treat the problem from the point of view of Epicurean epistemology and say that we are investigating unknowable things and that accordingly we should provide multiple explanations, or at least as many explanations as the evidence will reasonably stand. This does not devalue Sedley's reconstruction of the proem, but does allow for other possibilities. In the context of cosmology, it may also be tempting to follow Timaeus when he says that his account is only a 'likely story', and that the investigation of sources, just as of cosmogony, should be treated as a paidia phronimos, a pastime, but a serious one.
Perhaps the most obvious difficulty with the reconstruction of Empedocles' proem arises from the new Strasbourg fragments. The first, very fragmentary, lines of new fr. 'a' appear to match closely the final lines of DK 31 B17, which Simplicius tells us is from the beginning of book one of the Physics, and line 30 of new fr. a(ii) contains the stichometric letter gamma, used by scribes to number each 100 lines copied, which places it as line 300. Attaching new fr. 'a' to the end of fr.17, this would make fr.17 begin at line 233 of book one of the Physics. Clearly fr.17 is part of the didactic address to Pausanias, and so proem and didactic address would take up 232 lines of the first book of the Physics, a book which is generally thought to have been only about 1,000 lines in length. Sedley claims that a proem and didactic address of 232 lines is not too long, appealing to Hesiod's Works and Days. However neither Works and Days nor any other didactic poem expends anything like this number of lines on proem and address (cf. Aratus' Phaenomena, Manilius' Astronomica, Vergil's Georgics, Nicander's Theriaca, Ovid's Ars Amatoria). Lucretius' proem and didactic address take up some 145 lines, a number that is unusually large, and so 232 lines would appear to be unparalleled. Given this, I feel that the burden of proof must lie with those who claim that such a long proem is possible. It may well be that other works may be adduced to support a 232 line proem, but Sedley does not do so, and even if we include an invocation to Aphrodite at the beginning of the proem as we probably should this does not avoid the problem, since Lucretius manages to fit his invocation to Venus and all the other elements of the proem into 145 lines. Further, Lucretius, in a work more than three times the length of the Physics, reaches the point in D.R.N., which the Strasbourg fragments tell us should match Empedocles' Physics 1.262 (fr.17.31-35 = new fr. a(i) 1-5), the appropriation of the Parmenidean tenet, at D.R.N. 1.150ff. Sedley appeals to Empedocles' wordy repetitious style to help explain the discrepancy, and Empedocles certainly is more repetitious generally than Lucretius, but we should also consider the apparent brevity of Empedocles' treatment of the impiety of sacrifice in DK 31 B137, where he seems to use only six lines to do what Lucretius does in 17 lines with the sacrifice of Iphigenia. Similarly, the praise of Pythagoras (fr. 129) seems to receive a briefer treatment than Lucretius' praise of Epicurus, and this is what we should expect in a work less than a third the size of Lucretius'. These discrepancies strongly suggest that Empedocles' proem may possibly be quite different in its structure than Lucretius'.
We can be fairly confident from Lucretius' evidence, and from the evidence of other didactic poems, that Empedocles' proem began with an invocation to Aphrodite, but the evidence adduced for the rest of Sedley's reconstruction, apart from that of Lucretius' proem, is really very slight, and although I feel it is a very likely story we should not place absolute reliance on it.
The examination of Lucretius' and Empedocles' language and poetic techniques in chapter two, 'Two Languages Two Worlds', is one of the most valuable parts of the book. Here Sedley explains more clearly than has ever been done before how Lucretius forges his philosophic vocabulary, and shows that his practice of using a variety of live metaphors derives directly from Empedocles, contrasting Cicero's technique of finding direct Latin equivalents to Greek technical terms. This chapter may be read separately from Sedley's main thesis and does not suffer from being thus excerpted. Of course, Sedley uses his findings to prove part of his thesis: especially that Lucretius' choice and revision of Latin terms for eidola in the two proems to book four indicate that Lucretius was in the process of revising book four when he died, and that, had he lived, he would have gone on to revise books five and six as well. This links in with chapter five, 'Lucretius Plan and its Execution'. The arguments in chapter four are cogent and may be accepted without agreeing with the argument of chapter five that Lucretius would have undertaken a wholesale re-ordering of topics in books five and six. Chapter two brilliantly illuminates Lucretius' linguistic and poetic technique, the relationship between poetry and prose, and Latin and Greek in the subsections 'Prose and Verse Contrasted', 'Evoking Greece', and 'The Familiar and the Exotic' and may be recommended unreservedly as a major contribution to the understanding of Lucretius' poetics, which must become required reading for all students of Lucretius.
In chapter three, 'Lucretius the Fundamentalist', Sedley argues that since Lucretius does not directly engage with contemporary attacks on Epicurus' doctrine, he is in his 'hardcore scientific and philosophical beliefs' a fundamentalist who ignores later philosophical developments.
On this point there would seem to be a divide between those, like myself, who consider that it would be highly unlikely that Lucretius does not engage in any way with contemporary philosophical opponents and so regard negative evidence as inconclusive, and those who feel that positive evidence that Lucretius does engage with contemporary opponents is necessary. The latter approach is exemplified by David Furley ('Lucretius and the Stoics', BICS 13 (1966): 13-33 = pages 75-95 in C.J. Classen, Probleme der Lukrezforschung. Hildesheim 1986) who has examined each case where Lucretius is claimed to be engaging with a source post-Epicurus, and has shown that in each case the target of the polemic is, or may well be, pre-Epicurean. However, Lucretius' technique of not naming the targets of his polemics (except for the Presocratics in book one) is characteristically Epicurean (see Knut Kleve in Fondation Hardt Entretiens 24 (1978) 59-60) and makes it particularly difficult to be sure whether an opponent later than Epicurus is being targeted. His failure to become involved in detailed arguments with the Sceptics and Stoics over their attacks on Epicurean theory that Sedley points out may indicate only that there is little place for them in a cosmological didactic poem. Certain passages certainly lend themselves much more naturally to an anti-Stoic analysis than to any other, even if we are quite unable to prove that the original targets were the Stoics. The warnings about the dangers of allegorising the myths of the gods at 2.644ff are very similar to what we now know was a contemporary Epicurean style of anti-Stoic polemic (cf. Velleius' attack on Chrysippus at Cicero N.D. 1.40, and Cicero's Epicurean source for this, Philodemus De Pietate pt. 2 columns 127-8 Obbink (forthcoming, = P. Herc. 1428 cols. 5-6 ed. A. Henrichs, 'Die Kritik der stoischen Theologie im P. Herc. 1428', Cronache Ercolanesi 4 (1974) 5-32)).
Further, it is difficult to avoid the anti-Stoic implications of the comparison between the achievements of Epicurus and the deeds of Hercules, popularly the patron hero of the Stoics, in 5.22ff. Of course, the Stoics did not really make as much use of Hercules as they were popularly thought to. But without going as far as insisting on a Derridan reading that meaning is realised at the point of reception, the popular association between Hercules and Stoicism, and the various representations in ancient literature of Heracles at the crossroads choosing between Virtue and Pleasure will inevitably be seen by the ancient reader as representing an opposition between Epicureanism and Stoicism. We must assume that Lucretius was fully aware of such literary nuances in his own day and so his presentation of Epicurus and Hercules as rival culture heroes, and his implicit invitation to the reader to choose between them, can hardly be thought not to be his own opposition between Epicureanism and Stoicism. If I am forced into discussing authorial intention, I feel that the burden of proof must lie with those who say that Lucretius did not have Stoicism in mind when he wrote this passage. The same may be said about the many passages where it cannot be proved that the Stoics rather than Plato are the targets, and we may reasonably say that the sort of catch-all polemics Lucretius uses, that may for example target either the Stoics or Plato, or both, are a deliberate part of his technique and show that he relies upon the reader to supply the target according to the reader's own knowledge and taste. Accordingly the arguments over whether we can prove one way or the other whether the Stoics are the targets of Lucretius' polemics are bound to be fruitless.
One of Sedley's main examples is 'The Location of the Mind', in which he claims that Lucretius is wildly out of date and out of touch with his own school and with contemporary science in insisting on Epicurus' out-of-date doctrine that the mind is located in the breast. Here Sedley suggests that Lucretius is in fact ignorant of contemporary arguments, while elsewhere he argues that Lucretius knows of them but ignores them. This is where the choice of the word 'fundamentalist' to describe Lucretius begins to seem unfortunate. The word has so many negative connotations that I am sure Sedley would not wish to load onto Lucretius, but it seems to give us only this choice that either he behaves as we see modern religious fundamentalists doing and deliberately rejects all developments of philosophy, even of Epicureanism, since Epicurus' day, or he is actually ignorant of them, as Sedley claims he is ignorant that the location of the mind had been proved to be the head. The latter possibility seems highly unlikely, and here again I feel the burden of proof must lie on those who would claim this, but the former paints Lucretius as somehow unusual in adhering to his master's doctrines. Although Sedley would hardly wish to suggest such a thing, the use of the word 'fundamentalist' may lead readers to infer that Lucretius is distorting his master's doctrine as religious fundamentalist's do. We may say reasonably that Lucretius is orthodox on this point of the location of the mind, but then orthodoxy is especially important for the Epicureans, and so Lucretius is not at all unusual in this. Sedley adduces the fact that Demetrius of Laconia argues against the theory of the location of the mind in the head, and he claims that, since this shows that the mind's location was a live issue for the Epicureans in the first century, Lucretius' failure to mention it must prove he is ignorant of the whole argument. This may be countered by saying that Lucretius is also orthodox in another way: that he will not argue for argument's sake, but only to aid his didactic technique. The foundation of the Epicurean approach to investigation may be seen in the Letter to Pythocles 85-6, where Epicurus reminds Pythocles that the only purpose of Epicurean science is firm conviction and peace of mind. Accordingly investigation, and thus philosophical argument, beyond that point is counterproductive, and given this attitude we may expect the Epicureans to gloss over weak points in the master's doctrine. Demetrius of Laconia is concerned with establishing Epicurean orthodoxy and so such a polemic is for him unavoidable, but Lucretius in book three is engaged in quite a different and more important polemic: that the soul is mortal. He could certainly introduce a polemical digression on the location of the mind, as he does against the 'Harmonia' theory of the soul (3.94ff), but establishing the location of the mind is not nearly as crucial as establishing that the soul is material, which is the purpose of the argument against the 'Harmonia' theory, and that thus it must perish as must all material things. Sedley does remind us that there are two places in book 3 where Lucretius has been thought to show a knowledge of the theory that the mind is located in the head (3.138 and 3.788 ff), but rejects the notion (p. 71 n.46). However, 3.136-40 still seem to me to give a sly ironic reference to the rival theory:
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.
idque situm media regione in pectoris haeret.
(Now I say that the mind and spirit are held in conjunction together and compound one nature in common, but that the understanding, which we call the 'animus' and 'mens', is as it were the head and rules in the whole body. And it is sited and stays in the mid-region of the breast.)
Sedley's argument that caput here should be read as 'source', as of a river, seems an unnecessarily forced reading. In the context of the location of the mind in parts of the body, caput will be much more naturally read as referring to a part of the body.
3.788ff seem to give a similarly ironic reference:
sic animi natura nequit sine corpore oriri
sola neque a nervis et sanguine longius esse.
quod si posset enim, multo prius ipsa animi vis
in capite aut umeris aut imis calcibus esse
posset et innasci quavis in parte soleret
(So the mind cannot arise alone without body, nor exist far from the sinews and blood. But if it were able to, the mind itself could much more easily exist in the head or shoulder or the heels and be born in any part.)
The choice of the head as one of the places impossible for the mind to exist must surely be pointed. To put it another way, Lucretius is saying obliquely: 'if you think that the mind can be located in the head you may as well say that it could be in the shoulders or the feet'.
Even if we were to reject the above two examples, Lucretius' failure to engage with the rival theory gives us only negative evidence about his state of knowledge and his orthodoxy, and this reliance on negative evidence seems perhaps the weakest part of Sedley's technique.
Chapter four concentrates on an examination of Epicurus' On Nature and in particular, a reconstruction of the contents of the first fifteen books. With information from the Herculaneum papyri, and citations in other authors, Sedley presents a detailed argument and displays the results, very usefully, in tabular form. Again, as with chapter two, Sedley provides us with material and information, which although adduced to aid his general argument, is extremely useful in itself, and often illuminating. The reconstruction is careful to avoid any Lucretian evidence in order to avoid a charge of circularity, and the reason for this becomes clear in chapter five, 'Lucretius' Plan and its Execution', where the reconstruction is compared topic by topic to Lucretius. The purpose of this, beyond summarising On Nature for scholars, is to examine Lucretius' technique of composition in D.R.N. and from the close matching of the topics of D.R.N. and On Nature I-XV, he concludes that Lucretius' only philosophical source is On Nature I-XV. This argument is very persuasive, but may be open to challenge. Sedley shows how closely the topics of On Nature 1-15 match those of D.R.N., but still the possibility of intermediary sources seems plausible. I would contend that even if Lucretius carefully follows Epicurus' On Nature I-XV this does not preclude his borrowing from and engaging with other writers. It is only if we insist on a strict divide between poetry and philosophy that we can say that Lucretius' only philosophical source is On Nature. I prefer to see the work as having both poetry and philosophy so closely interwoven into its structure that the one cannot easily be removed without the other, very much in the same way as Lucretius describes the interrelationship between body and soul in book 3. I have already suggested above that Lucretius returns to one of Epicurus' own forerunners and probable sources in importing extra Empedoclean elements into the zoogony in book 5, and this seems to me to go beyond simple poetic influence. If we allow Lucretius to engage philosophically with pre-Epicurean sources we may also allow him to engage also with post-Epicurean ones.
Recently P.H. Schrijvers has shown that Lucretius may well be engaging with various sources later than Epicurus, especially Palaephatus, and with the Peripatetic philosopher Dicearchus in his culture history in book five (see his 'Intertextualité et Polémique dans le De Rerum Natura (V 925-1010). Lucrèce vs. Dicéarque de Messène', in Philologus 138 (1994) 288-304, now pages 81-101 in his Lucrèce et Les Sciences de la Vie. (Leiden 1999)). Sedley disagrees with Schrijvers but refuses to answer him in detail on the question of Dicearchus, and restricts his arguments against Schrijvers to one (admittedly long) footnote. Schrijvers argues very persuasively that Lucretius draws upon Dicearchus, and that both Dicearchus and Lucretius draw upon Hesiodic and Homeric material in constructing their culture histories (cf. Dicearchus fr. 49 Wehrli with 5.937ff and Works and Days 116ff; 5.945ff and Odyssey 13.103ff, 5.958ff and Odyssey 9.108ff), and that Lucretius imitates Works and Days 116-9 in 5.937-8, 943-4 importing a strongly Golden Age feel into his pre-history. How likely is it that Epicurus turned to Hesiod for help in book 12 of On Nature? Lucretius, on the other hand sees no great divide between poetry and philosophy and will appropriate freely even from sources where the original context is quite the opposite of his argument, as he does in borrowing Odyssey 6.42-6 for D.R.N. 3.18-24, describing the blessed life of the gods, but inverting the underlying teleological message for his own purposes simply by re-contextualizing the lines. Overall I find Schrijvers' arguments that Lucretius was in touch with contemporary philosophy much more compelling than Sedley's that he was not.
The argument that books five and six are unfinished and would have been revised had Lucretius lived is less open to question, but Sedley's claim that Lucretius would have re-ordered the topics in them may be challenged. As it stands the order of topics in book five parallels that found in Plato's Timaeus: cosmogony, then astronomy, then zoogony. As Sedley argues, the astronomy in Lucretius' account interrupts the 'natural' flow of the narrative, separating cosmogony from zoogony, which we should expect to be described consecutively since they are seen as part of the same process. He argues that Lucretius followed the order of topics he found in On Nature 11-12, that this order was adopted by Epicurus from Timaeus in order to argue against it point by point, but that the table of contents to book 5 (5.64ff) indicates that Lucretius was planning to move the astronomy to after the pre-history and so avoid the awkward interruption between cosmogony and zoogony.
In Timaeus this order is natural since Plato presents the birth of the stars as one facet of zoogony, referring to them as one of the four types of living creature which the world needs in order to be complete, the other three being birds, land animals and sea-creatures. Thus, an explanation of the movements of stars and planets sits comfortably within the zoogony. However, although the order of topics in Lucretius five is problematic, of all changes he might have made, the one I would perhaps least expect him to make would be to move the astronomy away from the cosmogony.
I would claim that the order found in Lucretius is also natural, since the stars and planets are formed by the same process as the earth, air and sea (5.416ff), thus the birth of the stars is a part of the cosmogony: the atoms form compounds, which by the process of attraction of like to like separate out from one another and go to form the four parts of the world: earth, the heaviest, sinks down, water, next heaviest, above that, air rises up, and aether, lightest of all, rises highest and forms the stars and planets. The description takes us on an upward journey, which matches the chronological sequence of creation, and the rising up of the aether beyond the storms of the air is the reason given for the smooth flowing progression of the heavens. This is the crucial point that the astronomy needs to make: that the regular motions of the stars do not reflect any divine control of the world, since the origins of religion lie in a false judgement that the gods control the universe, which arises from observation of the order of the heavens and a failure to apply the correct ratio (5.76ff and 5.1183). Further, Lucretius makes it clear in introducing the cosmogony that the astronomy will follow, cf. 5.416-18:
sed quibus ille modis coniectus materiai
fundarit terram et caelum pontique profunda,
solis lunai cursus, ex ordine ponam.
(But next in order I will describe in what ways that assemblage of matter established earth and sky and the ocean deeps, and the courses of sun and moon.)
It is the orbits of the sun and moon that are to be explained by the following cosmogony, not simply their origins, and so it is imperative that Lucretius should follow on from the cosmogony with the astronomy. Thus, being part of the anti-teleological programme, this order, with astronomy following closely on cosmogony, may be originally Democritean (cf. the report of Leucippus DK 67 A1 (Diogenes Laertius 9.31)), or alternatively an even earlier Presocratic order as fr. DK 28 B10 (Clement Strom. 5.138) of Parmenides suggests:
'And you shall know the nature of aether and all the signs in it and the destructive works of the pure torch of the shining sun, and whence they came into being; and you shall hear of the wandering works of the round-eyed moon and of her nature; and you shall know too of the surrounding heaven, whence it grew and how Necessity guiding it fettered it to hold the limits of the stars'. (Trans. K.R.S.)
Parmenides promises to present an explanation of the movements of the stars as part of his exposition of their births and natures. Cf. also DK 60 A4 (Hippolytus Ref. 1.9) where Archelaus is also reported as if he presented the astronomy between the cosmogony and the zoogony.
Plato's practice in Timaeus of appropriating scientific language and theory and subverting it to his own purposes makes it more likely that he has borrowed a traditional order rather than inventing a new order of exposition. His presentation of the birth of the stars as part of the zoogony rather than, as traditionally, part of the cosmogony would seem to confirm this mischievous perversion of Presocratic science.
Of course, Lucretius' order of topics does interrupt the smooth flow of the narrative and necessitates a sudden drop back down to earth at 5.772ff to explain the zoogony, which in the historical chronology follows directly on the formation of earth and may even take place while the aether is still separating out from earth as 5.806 suggests (cf. Empedocles DK 31 B62), but the alternative is to hold over the astronomy until after the culture history, which is too far from the cosmogony for the clarity of the link between them to be sustained, and thus some of the anti-teleological force of the astronomy would be lost.
The order of topics in book five is inevitably problematic for Lucretius, since there is always a tension between the historical chronological sequence and the narrative sequence. This arises from the fact that Lucretius is not writing a history for its own sake but presenting a set of important Epicurean topics in the context of their development in history. Thus, the section on justice (5.1011-27) seeks to expound the true nature of justice, and the pre-historic setting explaining its origins is a part of the didactic technique. In a pure culture history, we should expect the order of presentation of topics to match their chronological development, but this could break up the more important didactic order. Thus, fire, family life, clothing, religion, and language should all be explained as arising during the development of the first societies and of justice, but since justice is the crucial topic, fire, language, and religion are held over and given separate sections of their own. But, since the crucial didactic aim of removing wonder at the order of the heavens is best served by interrupting the smooth transition from cosmogony to zoogony rather than interrupting the link between cosmogony and astronomy, narrative considerations must take second place. The birth of the cosmos is an important subject in its own right, since it is necessary to prove that it arose by chance and was not constructed by God, but it also serves as a historical introduction to the equally, or more, important anti-teleological explanation of the motions of the stars: there is no need to think that the gods control the motions of the heavens because they can be explained rationally, and the mechanics of the cosmogony are part of this rational explanation. Thus we cannot reasonably claim that the astronomy is in the wrong place.
I hope the above arguments show two things at least: firstly, that the reconstruction of Lucretius' sources is very much a live issue and, given the often slender nature of the evidence available to us, that different interpretations may be possible. Further that Sedley's book should be regarded as a spur to further investigation and discussion. The true value of such a book must lie in its ability to stimulate argument rather than in its ability to build an unassailable structure impervious to criticism. Criticisms along the lines that such reconstructions are not well enough founded on evidence seem to me based on false reasoning: if we had enough evidence to make an entirely solid reconstruction, we would not need to make one at all. This book has certainly inspired me to investigate the Herculaneum papyri of Epicurus and Philodemus much more closely, and to rethink my ideas of the relationships between Lucretius and Empedocles, and Lucretius and Epicurus. From the amount of time and thought I have found myself expending on problems thrown up by this book, and from the number of times I have recommended it to others, I must consider it one of the most stimulating works on Lucretius for many years.