^* DK: Diels-Kranz, Fragmente der Vorsokratiker, 6th edition.D.L.: Diogenes Laertius, Lives of the PhilosophersKRS: Kirk, G. S., J. Raven, M. Schofield, The Presocratic Philosophers.Cambridge,1983Google Scholar.Bollack (19659): J. Bollack, Empedocle, vols. 1–3. Paris, 1965, 1969 Diels(1898)Google Scholar:Diels, H.,‘Uber die Gedichte des Empedokles’, Silzungsberichte der Preussischen Akademie (SBB) 1898,396–415Google ScholarH. Diels, Kleine Schriften, ed. Burkert, 127–46.Guthrie(1965)Google Scholar: Guthrie, W. K. C.,History of Greek Philosophy, vol.2.Cambridge,1965.O'Brien(1969)Google Scholar: ‘Brien, D. O, Empedocles′ Cosmic Cycle.Cambridge,1969.O'Brien(1981)Google Scholar: ‘Brien, D. O, Pour interpreter Empedocle.Paris-Leiden, 1981.Osborne (1987)Google Scholar: Osborne, C. J., Rethinking Early Greek Philosophy.London,1987.Solmsen (1975)Google Scholar: Solmsen, F.,‘Eternal and Temporary Beings in Empedocles′ Physical Poem’, AGP 57(1975),123–45.Sturz(1805)Google Scholar: Sturz, F. W.,Empedocles Agrigentinus. Leipzig, 1805Google Scholar.Van der Ben(1975):Van der Ben, N.,The Proem of Empedocles′‘ Peri physios.Amsterdam,1975Google Scholar.Williams(1982): Williams, C. J. F., Aristotle's De generatione et corruptione. Oxford,1982Google Scholar.Wright(1981): Wright, M. R., Empedocles, the Extant Fragments.Yale,1981.Google Scholar

¹ Most obviously the poem ‘On Nature’ is regarded as ‘scientific’ and the ‘Katharmoi’ as ‘religious’ in content. Those who hold that Empedocles wrote other works would probably place the ‘Hymn to Apollo’ in the latter class and the medical treatise in the former. For the present purpose I shall ignore these minor works attributed to Empedocles and confine myself to the question of whether the major works ‘On Nature’ and ‘ Katharmoi’ can be divided into religious and non-religious in any meaningful way. For the minor works, see Diogenes Laertius 8.57–8, 8.77 and DK 31A23 and 31B134. Empedocles does not appear among Diogenes Laertius′ examples of philosophers who wrote only one work, nor among those who wrote more than one (D.L. 1.16) but since Diogenes believes that Empedocles wrote various minor works neither omission is of any significance for the question in hand.

² I include here Sturz (1805), whose suggestion (pp. 71 ff.) that the Katharmoi are certain books of the physics does not question the idea that these books are distinguishable in terms of content from the physics proper.

³ Compare the role played by ‘ Of the Standard of Taste’ in the sentence ‘ Hume produced in his essay " Of the Standard of Taste " the most mature aesthetic document…’.

⁴ Compare the role played by ‘on taste’ in the sentence ‘The problem he faced in the essay on taste was how to escape what appeared to be the conclusion of this view’.

⁵ D.L. 8.54

⁶ Athenaeus 14.620 .

⁷ Theo Smyrnaeus, p. 104.1 . Herodian Palimpsest (Empedocles fr. 152 Wright)

⁸ Hippolytus, Refutatio 7.30.3 .

⁹ This point is the fourth in a series of four questions addressed to Marcion. The second point uses a very similar ambiguity: Does Hipploytus mean that Marcion is secretly teaching the theories (logoi) of Empedocles or the very words (logoi) of Empedocles? He means and implies both: the first is the more plausible charge, the second the more damning charge and Hippolytus′ refutation is built upon the slide from one to the other.

¹⁰ Note also the occurrence of in connection with Empedocles in Porphyry, De abstinentia 2.31 and Theo Smyrnaeus, p. 15.7, neither of which is conclusive evidence concerning titles for his work.

¹¹ D.L. 8.60: Suda s.v. Empedocles (DK 31A2) . Galen, De elem. sec. Hipp. 1.9 (1.487K) TO yap TUIV . Compare also the Hippocratic On ancient medicine 20 which, if genuine, suggests that . was already used to describe the works of the Presocratics, Empedocles included, at a much earlier date.

¹² It is unclear what the neuter plural (TO.) refers to. It could mean his verses (en-q) or books (j8ij8Aia). Compare Simplicius, Phys. 144.26 on Parmenides: . Here it is similarly unclear whether is a title for Parmenides′ poem or a description of the content of the specific sections Simplicius is concerned with.

¹³ If we were looking for a fifth-century title this would be more plausible thanon its own. Apart from V.M. 20 (above, note 11) there is no good evidence forbeing used without a limiting genitive to mean ‘nature in general’ in the Presocratic period. See Schmalzriedt, E.,Peri Physeos(1970) and Wright (1981), pp.85–6. The credentials ofas a title for Empedocles are increased by the fact that Lucretius′ De rerum natura (which is often thought to be a deliberate echo of the titles of Epicurus and Empedocles) is a direct translation ofGoogle Scholar

¹⁴ Aetius 1.30.1; Simplicius, Phys. 157.27, 300.20, 381.29; Tzetzes, Chil. 7.522, ex il. 53.23.

¹⁵ Aristotle, Meteor. J4, 382al. Aristotle also uses the phrase Physics 196a22, but this may designate a particular section of the work rather than the whole thing.

¹⁶ In addition to those listed in nn. 14 and 15 see for example Simplicius, Phys. 32.2, 331.10 and schol. ad Dionys. Thrac. p. 166.13 (DK 31A25).

¹⁷ Common sense suggests thatas a title implies that the contents had some connection with ‘natural philosophy’. The implications of ‘Katharmoi’ as a title are more difficult to determine (see Guthrie [1965], pp. 244–5). The best evidence for the subject matter of Empedocles′katharmoi (as opposed to those of Orpheus and Musaeus, Plato, Republic364e) are the hints in Hippolytus and Porphyry. Hippoiytus, Ref.7.30.4 mentions prohibitions on marriage, reproduction and meat-eating as the katharmoi of Empedocles, and gives an explanation of the reasons in terms of the cycle of one and many under Love and Strife. Porphyry mentions katharmoi in connection with the sin of meat-eating and gives as an example Empedocles′ confession of his own guilt, B139 (De abstinentia2.31). These hints pose some difficulties (though not insuperable ones) for the suggestion (put to me by David Sedley) that the Katharmoiwas a medical work and was not the locus of any of the main doctrines we attribute to Empedocles. On this see further below nn. 51 and 98.

¹⁸ E.g. Melissus, (Simplicius, Phys. 70.16, De caelo 557.10), Protagoras, (Plato, Theaet. 161c, Sextus Empiricus, Adv. Math. 7.60), Gorgias, (Sextus, Adv. Math. 7.65).

¹⁹ E.g. Simplicius, Phys. 70.16: Cf. also Simplicius, De caelo 557.10 and Sextus Empiricus, Adv. Math. 7.65.

²⁰ With the examples in n. 19 compare Olympiodorus in Plat. Gorg. p. 112: Suda s.v. .

²¹ See Sextus Empiricus, Adv. Math. 7.60 and Plato, Theaetetus 152a with 161c.

²² B112, D.L. 8.62.

²³ Lobon, fr. 19 Cronert. The ascription is made on the basis of the characteristic counting in verses.

²⁴ ‘A man best known for his forgeries’ (Guthrie [1965], p. 135 n. 3). It need not follow, of course, that everything he says is a lie.

²⁵ If five books are in question in Diogenes, a thousand lines per book seems surprisingly high. See Wright (1981), p. 21. But longer books are attested in the Hellenistic period, e.g. Apollonius Argonautica (12851781 lines per book). See Skutsch, O.,The Annals of Q. Ennius(Oxford,1985), p.369. However there is no evidence that Diogenes had five books in mind: indeed more than five would be compatible with our present evidence.Google Scholar

²⁶ In particular Van der Ben (1975), pp. 12 and 14–16 writes as if Diogenes had written ‘five books’ (in response see O'Brien [1981], pp. 5–6, Zuntz, Mnemosvne 18 [1965], p. 365).

²⁷ See Wright (1981), p. 20 and n. 104.

²⁸ As punctuated here the Suda's two thousand verses are not necessarily confined to the physics but would cover all Empedocles′ verse works. This means that if the medical treatise in 600 verses mentioned by Diogenes is distinct from the prose works mentioned by the Suda it might be included in the Suda's total of two thousand verses. However,‘ 600 verses’ in Diogenes could be a loose way of referring to ‘lines’ of a prose work. The usual punctuation makes a parenthesis after , but such a parenthesis beginning repeating stylistically improbable. I owe this point to David Sedley.

²⁹ Good evidence in Simplicius, Phys. 157.27, 300.20, 381.29 for first and second books. Tzetzes, Chil. is the only extant reference to a third book of Physics. Diels argued that this really referred to the Katharmoi.

³⁰ E.g. if we read instead of . Compare Diels′ suggestion that it should read , Diels (1898).

³¹ Note the construction in Diogenes′ sentence.

³² Diogenes′ source might have given both titles even if they denoted one and the same work; similarly it might have mentioned a number of books for the physics with or without a number for the katharmoi books.

³³ Plutarch quotes frequently from Empedocles. 41 out of 133 fragments in Wright are quoted wholly or in part by Plutarch in extant works. In addition there is evidence that he wrote a work in ten books on Empedocles, now lost (Hippolytus, Ref. 5.10.6; Lamprias, catalogue 43).

³⁴ Van der Ben (1975), pp. 1620, enumerates arguments in favour of the conclusion that philosophia must mean the physics. Arguments that it could mean the Katharmoi are found in Diels (1898), O'Brien (1981), pp. 14–15, Wright (1981), pp. 81–2 and 270–1.

³⁵ Van der Ben (1975) discusses only those passages he believes to belong to the proem, but he suggests that other ‘ Katharmoi’ fragments also belonged to the main books of Peri Phvsios.

³⁶ E.g. P. Kingsley (unpublished thesis, Cambridge. 1979), O'Brien (1981), Wright (1981), p. 271. Van der Ben's position differs from mine in that he continues to operate with a strong distinction between what is properly‘ physics’ and what is properly‘ religion’. In moving material to the proem rather than the main thematic material of the physics, he is creating a further divide between proem (religious) and main books (physical) as well as that between Peri Physios and what remains for the Katharmoi.Google Scholar

³⁷ In addition to the fact that the testimonia fail to make a clear distinction between the physics and katharmoi some actually make it impossible to do so. Cf. e.g. Plutarch, De Is. el Osir. 48, 370d; De tranq. an. 15, 474b. We do not possess an extended interpretation in Plutarch's extant works, but Hippolytus, Ref. 7.29–30 is a good example of an exposition which interprets the entire system in terms of the banishment of the daimones. See Osborne (1987) and below section B.

³⁸ An exception might be made for one or two fragments (e.g. B134) which may be from the hymn to Apollo if such a thing existed.

³⁹ Long, H. S.,‘The Unity of Empedocles′ Thought’,AJP 70(1949),142–Google Scholar58 pointed out that the division of fragments was largely based on speculation concerning the content of the two works. Among others who argue for the unity of Empedocles′ thought may be mentioned Prier, R. A.,Archaic Logic(1976), pp.1–22, 120–5, 130Google Scholar; Kahn, C. H.,‘Religion and Natural Philosophy in Empedocles′ Doctrine of the Soul’, AGPh 42(1960),3–35, p. 24; Wright (1981), p. 57CrossRefGoogle Scholar; Darcus, S. M.,'Daimon Parallels the Holy Phren in Empedocles‘, Phronesis 22(1977), 175–90.CrossRefGoogle Scholar

⁴⁰ This ‘ fact’ is often used as a criterion for assigning verses to one or the other poem on the basis that they include singular or plural second person forms.

⁴¹ Both start and it is natural to assume that the friends are the same. However, Bl 14 continues with a concern for men in general, so that even here an audience of Acragantines cannot be reliably assumed.

⁴² We might compare Hesiod's Works and Days which names Perses in the third person at line 10, addresses him by name in the second person at line 27 and frequently thereafter; even this, however, begins by addressing the Muses first.

⁴³ For instance the Pindaric odes: the victor is usually named within the first 28 lines of a. victory ode but is often not addressed directly until 100 lines later (e.g. Ol. 1, 10). In Pythian‘ 1 Hieron is named at line 32 a n d addressed directly from 85–100, but the poem also apostrophises Zeus (29, 67), Apollo (39) and the Muse (58).

⁴⁴ KRS, p. 313.

⁴⁵ Plutarch, De exilio 607c. (See above, section iii.)

⁴⁶ Apart from B114 see particularly B136, 137, 145, 124, 4, 8, 9, 11 and 15.

⁴⁷ Line 3 of B112 does not, strictly speaking, belong to B112 since it is not quoted by Diogenes Laertius with the other lines, but is inserted by most editors of Empedocles. It is quoted by Diodorus Siculus without any context of its own, 13.83.2. However, since he uses it of the hospitality of the Acragantines we may consider its content relevant even if it did not occur in precisely this position.

⁴⁸ (unacquainted with evil) is ambiguous. It may well be that the primary meaning of (ca/coTTjy at this period i s ‘ misfortune’, bad things that happen (regardless of agency). The superficial meaning of the phrase is clearly suggesting that the Acragantines are prosperousignorant of evil because it never befalls them. But, I suggest, two further meanings may be in play: (1) they are ignorant of evil in the sense that they are unaware that their apparent prosperity and good fortune is in fact misfortune - that they are suffering evil and (2) they are ignorant of evil in the sense of wrong-doing - they are unaware that they are doingbad things when they think they are doing good things. The range of possible meanings and of irony is dependent on the context; we cannot be sure what that context was. See Dover, K. J.,‘The Portrayal of Moral Evaluation in Greek Poetry’,JHS 103(1983).35–48CrossRefGoogle Scholar

⁴⁹ I make no claim here concerning the original location of Bl 14, simply observing that the parallel address would invite interplay with Bl 12 regardless of their respective locations.

⁵⁰ Clement, Strom. 5.9 identifies & positive valuation of pistis in these lines, though on what grounds is unclear. On the other hand, it is undeniable that connotes bitter jealousy and that the struggle between truth and deep-rooted belief is characterised in harsh terms of Homeric warfare (seeWright [1981], pp. 267–8).Google Scholar

⁵¹ This interpretation of Bl 12 as the opening lines of the poem depends upon the claim that Katharmoi and Physics are the same work, so that this passage can be juxtaposed with those which undercut it. This is the claim that I am making and one of its recommendations is its productivity in the interpretation of Bl 12 and the resultant explanation of the excessive pomposity of these opening lines. For the alternative that the Katharmoi was a non-philosophical work (which would remove this fragment to a quite separate context), see below n. 98.

⁵² Plutarch, De exilio 607c. See above, section (iii).

⁵³ , Plutarch, De exilio 607c.

⁵⁴ Refulatio 7.29. For justification of this confidence in Hippolytus′ commentary see Osborne (1987).

⁵⁵ Simplicius, Phys. 1184.8–18.

⁵⁶ B26.1, B27.1, 3,4, B30 and B31 are all quoted by Simplicius in the same context as Bl 15.1–2; Phys. 1183.28–1184.18.

⁵⁷ Thirty thousand seasons may not mean anything specific in itself, but it presumably stands in for a determinate period of time.

⁵⁸ Simplicius′ suggestion that the alternation of rule of Love and Strife is itself governed by necessity is not supported by the texts he quotes, Phys. 1184.6–18.

⁵⁹ This is the passage on which Simplicius is commenting (see note 58). Aristotle's criticism has been much discussed (especially by O'Brien [1969], ch. 4).

⁶⁰ He allows that there are some grounds for this simple claim in the analogy with their presence in human affairs, 252a28–30.

⁶¹ Note that Aristotle's statement need not imply that the world is alternately in motion and at rest, but that Love and Strife are each alternately active and at rest, being at rest during the time that the other power is active: ‘that it applies to things of necessity that Love and Strife are alternately ruling and moving and are at rest in the intervening time.’

⁶² 252a26.

⁶³ Plutarch, in Defac. in orbe lun. 926d–927a, describes a total separation of elements as the work of Strife, but it is not clear that this is ‘total Strife’ as opposed to the first stage of Strife's work. The couplet quoted by Plutarch (fr. 26a Bignone, fr. 19 Wright) describes a state in which the elemental masses cannot be distinguished and does not appear to fit the context he provides for it. It has often been assimilated to fr. 27 which appears to describe the sphere under Love, but nothing in Plutarch can justify the assimilation.

⁶⁴ 64 Most editors read ( ‘ b u t a t will’).

⁶⁵ There is no clear evidence for a zoogony created by Strife in the verses we possess, but on this reconstruction it would be satisfactory to suggest that there was one, so that Love's zoogony does not create any new differences over and above those that were there under Strife. Nothing else hangs on it and it is possible that Strife might create nothing more coherent than detached limbs, which are thereby more diverse and count as more differences than the whole creatures produced under Love.

⁶⁶ These distinct bits and pieces would obviously be made out of combinations of the four elements, since Empedocles repeatedly emphasises that the further differences are a result of these things ‘running through each other’. Thus there is a sense in which what is there when Love returns is the four elements and nothing more, and the mortals formed when Love returns are made of combinations of the four elements; but there is also a sense in which what is there is not pure elements at all but a whole lot of other different things.

⁶⁷ Both this and the traditional interpretation of B35 are circular: if we assume that the work of Strife is to make four elements then we assume that at Strife's peak at the start of B35 there are four elements and it transpires that Love creates a greater plurality than Strife; if we assume that the work of Strife is to make an extreme plurality of distinctions then we assume that at Strife's peak at the start of B35 there are a great many different things and it transpires that Love reduces the plurality and creates more unified beings. Lines 14–15 of B35 are very unclear and it is impossible to determine what are supposed to be mixed or unmixed. Simplicius has (confused) not (unmixed) in line 15. Most editors read aKprjTa on the basis of a (?mis)reading by Theophrastus which clearly puzzled him and others: Plutarch 677d; Athenaeus 10.423f.

⁶⁸ B17.1–2, 7–8, 9–10, 16–17; B20.2–5; B26.5–9.

⁶⁹ On sacrifice and bloodshed, B136, 137, 139. On sex see Hippolytus, Ref. 7.29.22. Given Empedocles′ one world view all sexual relations are incestuous. On the consequent miseries, B145. Marriage and the shedding of blood in childbirth are conceived as parallel to sacrifice (see H. King, ‘Sacrificial Blood’, HeliosWomen in Antiquity, special issue 1986, discussing B70).

⁷⁰ B128 in particular describes mortals living in a state of purity in accordance with love, and it is natural to suggest that this belongs to the period of increasing unity just before the return to the one. B130 probably also describes the same stage. The source of enlightenment which will set deluded mortals on the path of purity may be the wise man described in B129.

⁷¹ 273d6.

⁷² 273d6.

⁷³ The phrase suggests‘ mixing and remixing’ of things already mixed rather than mixture and separation out of. and into, pure elements. An individual mortal thing is born when its peculiar mixture is formed; it dies when its elements are remixed to form other products. No reference is made here to separation into elements.

⁷⁴ It is noteworthy that in the simile of mixing paints the painters are mentioned - it is they who mix the paints and produce the representations (B23.1–5) - whereas by contrast in the real world no one does the mixing: the things grow up of their own accord , B21.10) and run through each other; they mix themselves up. Whether they do this ‘in Love’ or ‘in Strife’ is unclear: both are mentioned in B21.7–8; in Strife the things become different shapes which seems to be much like what happens when they mix, for then they ‘become various’ B21.7, 14.

⁷⁵ See for example Aristotle, Metaphysics A4. 985a25ff.; De caelo B13. 295a29ff.; Plutarch, Defac. in orbe lun. 926d–927a. See above n. 63.

⁷⁶ Ps. Plutarch (DK 31A30) does mention separation from a primal mixture, but the basis of his testimony is not known. The reference to mixture is absent from Aetius 2.6.3 (DK 31A49).

⁷⁷ This view is expounded in detail by M. C. Stokes, One and Many in Presocratic Philosophy(1971), ch. 6, esp. pp. 162–3. The result would be that Love is actually incapable of producing a unity since one p a r t must always differ from another however small those parts. It begins to look as though to achieve homogeneity by the smallness of the parts in the mixture Empedocles would have to make Love complete an infinite division; Love's ‘ o n e ’ becomes an extreme plurality. Solmsen (1975) suggests that Empedocles could not allow the elements to lose their identity completely, and that the most he can afford to say is that they disappear for the eye (p. 136) on the grounds that he has repudiated genesisand phlhora;but (a) as Solmsen himself notes (1975, p. 128) it is only genesis e nihilothat is repudiated and (b) the resolution of phusisand thanatosinto mixing and interchange of mixed things in B8 is specifically applied to mortal things and need not extend to elements.

⁷⁸ See particularly B17.17. Note that the entire passage B17.17–35 describes the world of plurality; it describes the way in which Love is operative among mortals in this world and likewise the way in which the elements operate in this world, each with its own (line 28). This gives a better sense to line 27: they are ‘ o f like age’ not in being immortal but in coming to birth at the same time and persisting throughout the period of Strife's dominance. B6 need not imply‘ that the elements are fundamental to more than Strife's dispensation: we have no context for these verses and need not mean that they occurred early in the poem; it could be the start of any new section (compare B38.1).

⁷⁹ There is no reason to suppose that Empedocles is included in this error, pace Williams (1982), p. 144, Joachim, Aristotle on Coming to Be and Passing Away (1922), p. 179. Philoponus takes the plural to mean which is surely right. 80 This must refer to the denial of alteration within the current phenomenal world.Google Scholar

⁸⁰ This must refer to the denial of alteration within the current phenomenal world.

⁸¹ Philoponus, Gen. et Corr. 19.3

⁸² Degen. et corr. 315a8–19 (tr. Williams, 1982.

⁸³ Degen. et corr. 315a4–14 tr. Williams, 1982).

⁸⁴ It might be thought that Aristotle is importing inappropriate distinctions here since it might appear plausible to claim that the elements do not lose their character any more in the one than in the ordinary mortal compounds: just as earth, air, fire and water are not apparent as such in trees and the like so they would not be apparent in the one; but nevertheless Empedocles would want to maintain that the elements were still unchanged in reality, the appearance being the result of the mixture. The sphere or ‘one’ would then be simply a huge ‘mortal’. This view, a version of which must be maintained by anyone who holds that the one is a mixture of the same type (only more so) as the mortals, does not do justice to the strong distinction made by Empedocles between the one and the many and the divine status accorded to the one. The distinction that Aristotle assumes between the way the elements behave in mortal compounds and the change they undergo in moving into and out of the‘ one’ seems to correspond with the emphatic contrast pervading Empedocles′ account.

⁸⁵ This must surely be right if Love's work is to abolish differences, but whether Aristotle's criticism is fair depends on whether he is right to suggest that the elements could not change into one another when the ‘one’ occurred as an intervening step, or whether Empedocles meant to exclude such change only within a single world of Strife. This relates to the question of whether the ‘one’ or the elements are Empedocles′ prior principles, which Aristotle proceeds to discuss, 315a9–25; if the elements are merely passing phases of a basic ‘one’ it is unlikely that the same part of the one need retain the same character in successive plural worlds; if the one is a passing phase of basic elements it is more plausible that they might be envisaged returning to their former character. This issue also affects the notion of personal immortality: if Empedocles takes the kinship of life seriously the true self is the one daimon in its true home under Love, and the differences under Strife are unfortunate aberrations; thus personal immortality is preserved by the repeated return to the one not by repeated return to the same elemental characters and mortal compounds.

⁸⁶ That is to say, the state of unity under total Love was not a primal mixture. There may be a primal mixture in the sense that the first creation of the differences between the four elements may convert the one under Love into a mixture of four elements, which Strife then structures to make a world. See below, section (x).

⁸⁷ Simplicius (Phys. 160.26–161.13), who quotes the passage, clearly sees that two groups of products are in question, one lot created by Love, the other by Strife. He suggests that the former are in the intelligible world, the latter in the sensible world. The text and precise meaning of line 9 are in doubt.

⁸⁸ E.g. Metaphysics 985a21–63.

⁸⁹ Metaphysics 1000a26–8.

⁹⁰ B136, 137, 139

⁹¹ B115.13; B31; cf. Simplicius, Phys. 1124.1.

⁹² B30. See above, section (i).

⁹³ Simplicius, Phys. 1184.2ff., quoting B31.

⁹⁴ Reading in Bl 15.3 with DK, KRS and most editors. as in Plutarch, is retained by Wright (1981). in line 4 is Diels‘ suggestion.

⁹⁵ For the idea that there are equal quantities of the four elements in the world, see B17.27.

⁹⁶ See above, n. 86.

⁹⁷ All this suggests that the type of scheme involving a single cosmogony between successive worlds of Love is more in the right spirit than a double cosmogony with a strictly symmetrical four-stage cycle. Whether we would need two proper zoogonies within the world of Strife depends on how far towards chaos Strife's work gets before the daimones come to their senses, and whether Strife can make anything resembling a coherent creature in the period in which difference is increasing. See above, n. 65. The primary opposition is between the world as ‘one’ and the world as ‘more than one’, the latter including all the various degrees of dominance of Strife or Love within ‘this world’. For this type of two-stage cycle see e.g. Bollack (1965–9) and KRS.

⁹⁸ I have argued for the unity of Empedocles′ philosophy on the basis that physics and Katharmoiare one and the same poem. David Sedley has suggested an alternative view which is that all our fragments come from the physics except the few that are explicitly assigned in the sources to the Katharmoi,proposing that the Katharmoiwas not a philosophical work but a set of medical purifications. This position would change little in the interpretation offered here, the main difference being the exclusion of Bl 12 (see above, n. 51). On the other hand, I prefer to maintain the idea that the Katharmoiwas the title of the philosophical poem because (1) it fits well with the theme of banishment of the daimones in purification of blood guilt which is central to the work, (2) Hippolytus implies that the alternation of one and many belonged with the Purifications,and (3) the material in Bl 12 is rich in connotations if it belongs in a context which questions the values of society.