¹ Panzerbieter, F., Beiträge zur Kritik und Erklärung des Empedokles (Meiningen, 1844)Google Scholar. See O'Brien, D., Empedocles' Cosmic Cycle (Cambridge, 1969), pp. 157f.Google Scholar, on the origin of the now traditional interpretation of the cosmic cycle; Zeller's account ‘remains essentially unchanged’ in Zeller, E., Die Philosophie der Griechen, 6th ed., Nestle, W., ed., Teil 1, Hälfte 2 (Leipzig, 1920), pp. 969–79Google Scholar.

² Tannery, P., ‘La cosmogonie d'Empèdocle’, Revue philosophique de la France et de I'èlranger 24 (1887), 285–300Google Scholar; Arnim, H. von, ‘Die Weltperioden bei Empedokles’, in Festschrift Theodor Gomperz (Vienna, 1902)Google Scholar.

³ Bollack, J., Empedocle, vol. 1 (Paris, 1965)Google Scholar; Holscher, U., ‘Weltzeiten und Lebenskyklus’, Hermes 93 (1965), 7–33Google Scholar; Solmsen, F., ‘Love and Strife in Empedocles' Cosmology’, Phronesis 10 (1965), 109–48CrossRefGoogle Scholar.

⁴ O'Brien, D., Empedocles' Cosmic Cycle (Cambridge, 1969)Google Scholar.

⁵ Cf. Lüth, J. C., Die Struktur des Wirklichen im empedokleischen System über die Natur (Meisenheim, 1970)Google Scholar, which, however, does not take account of O'Brien's study (presumably because it appeared when Lüth's book was in press). Lüth mainly follows Hölscher (see pp. 5, 32, and n. 42, 59f.), but does give one original argument, pp. 61–5. His interpretation contrasts with Long's in an interesting way because he takes the theme of B17.1–13 to be living things rather than the elements. However, his reading of the passage seems forced, in particular because he takes the subject of lines 9–10 and 12 to be the elements while the subject of 11 and 13 is living things. The only evidence for this shift of subjects is the predicate adjective dxiVijToi, which can be understood in another way (see n. 9 below).

Mansfeld, J., ‘Ambiguity in Empedocles B17, 3–5: an Interpretation’, Phronesis 17 (1972), 17–39CrossRefGoogle Scholar, proposes understanding δοίη of line 3 as meaning ‘ambiguous’. On this reading the confluence of elements satisfies the reference of уένεσɩς while doing away wth the sense of the term. This interpretation seems forced. It requires us to read into B17 a sophisticated semantic theory for which there is no evidence in the text, a theory which is patently anachronistic. And even given the semantic theory, Empedocles is made to say something wrong-headed. For to say that the confluence of all things creates and destroys genesis is to make a use/mention error: what is destroyed on Mansfeld's reading is not genesis but ‘genesis’.

See also Ben, N. van der, The Proem of Empedocles' Peri Physios (Amsterdam, 1975), 29–32, 78–84Google Scholar.

Note that while many distinguished scholars have attacked the traditional view, there are serious problems with all the revisionary reconstructions of Empedocles, and there is almost no consensus among anti-traditionalists. See criticisms of other anti-traditionalists by Mansfeld, 20–4, and A. A. Long (1974, 409–12; see following note).

⁶ Long, A. A., ‘Empedocles’ Cosmic Cycle in the ‘Sixties,’ in Mourelatos, A. P. D., ed., The Pre-Socratics (Garden City, N. Y., 1974)Google Scholar.

⁷ Barnes, J., The Presocratic Philosophers, rev. ed. (London, 1982)Google Scholar, Wright, M. R., Empedocles: The Extant Fragments (New Haven, 1981)Google Scholar. The traditional view is rejected by Malcolm Schofield in Kirk, S., Raven, J. E. and Schofield, M., The Presocratic Philosophers (Cambridge, 1983), p. 288 n. 1Google Scholar, and Osborne, Catherine, ‘Empedocles Recycled’, CQ 37 (1987), 24–50, at 38ffCrossRefGoogle Scholar.

⁸ Wright's reply (1981, pp. 48f. n. 147) to Long and other critics is so compressed as to be unintelligible.

⁹ According to Hölscher (1965, 1 IF., 15), Aristotle holds the traditional view because of a misreading of B26. Solmsen (1965, 130–2) holds that he is guilty of a 'misconstruction' based on B26, but that he is inconsistent in his interpretation. Bollack (1965, 102–6) and Long (1974, 420–4) hold that the tradition does not contradict the single creation view. On the other hand, proponents of the double creation view generally hold that the doxographic tradition supports them.

¹⁰ Cf. Wright (1981), p. 46. Here I mean 'cosmos' to denote 'just the physical frame of the world', as Long expresses it (1974, 400).

¹¹ The pronoun is not stated; it could be either singular or neuter plural. Empedocles may be imitating the practice of Parmenides B2, B8, who omits the subject of ἔστɩ - probably because he wishes the verb to be understood as part of a sentence frame: ‘x is’ or ‘x is F’ (see Mourelatos, A. P. D., The Route of Parmenides [New Haven, 1970], pp. 47ff.)Google Scholar. Empedocles does not have the same intent, but he probably does wish the vagueness of the phrase to convey a notion of identity of the one and the many: it is ultimately indifferent what pronoun one uses, for the one becomes the many and the many the one. The same remarks apply to ‘it’ in line 2. I have supplied pronouns corresponding to the number of beings at the beginning of the change.

¹² θρεϕθεῖσαδɩέπτη is an emendation of θρεϕθεîσαδɩέπτή(θρεϕθεîσα: Panzerbieter, δɩέπτη: Scaliger). The reading works remarkably well (cf. O'Brien 1969, 164–7). Alternative readings have been unconvincing, e.g. θρεϕθεîσα δɩέπτή (von Arnim, Hölscher) or downright tendentious, e.g. δρυϕθεῖσα' ἀποδρύπτεɩ (Bollack). For criticisms, see O'Brien, ibid., and Long (1974, 409f.).

¹³ This line, missing in quotations, is supplied from a parallel portion of B26 quoted ih Aristotle, Ph. 8.1.250b30 (see Ross text).

¹⁴ It is unclear why Empedocles shifts to the masculine-feminine form ἀκίνητοɩ here. Bignone, E., I poeti filosofi della Grecia: Empedocle (Turin, 1916)Google Scholar, ad he. suggests that θεοί be understood; cf. Diels in apparatus of Diels-Kranz, who refers to B6. H. Munding, ‘Zur Beweisfiihrung des Empedokles’, Hermes 82 (1954), 129–45, at 134, argues that the antecedent is θνητοί, as does Lüth (1970, 63f.)Google Scholar.

¹⁵ All translations are my own except as noted.

¹⁶ Simpliciusin Ph. 161.14, Clement, Strom. 5.15 cited in the prologue to B17 in Diels-Kranz.

¹⁷ Heraclitus Bl; Aristotle, Rh. 3.5.1407b16 = 22A4; Sextus, Adv. Math. 7.132 = A16. Cf.Kahn, C. H.. The Art and Thought of Heraclitus (Cambridge, 1979), pp. 7–9Google Scholar. Heraclitus' proem does not detail a formal structure, but it does prefigure key themes(logos, human ignorance, the limitations of perception), and it does promise analyses of reality - ὁκοίων ἐуὼ δɩηуεῦμαɩ κατὰ ϕύσɩν δɩαɩρέωνἕκαστον καì ϕάζων ἔϰεɩ - and anticipate diagnoses of human cognitive frailties.

¹⁸ Owen, G. E. L., ‘Eleatic Questions’, CQ 10 (1960), 84–102CrossRefGoogle Scholar, Mourelatos (1970 [see n. 11], 95 and n. 2). Cf. Wright (1981, 48), who, although she cites the programmatic remarks of Parmenides as a model, does not anchor Empedocles' programme to any one passage.

¹⁹ The Origins of European Thought (Cambridge, 1954), pp. 31 OffGoogle Scholar.

²⁰ P. 340 and nn. 3 and 4 cite fr. 179, Nem. 4.44, Pyth. 4.275, schol. on Nem. 7.79, Nem. 8.15, Pyth. 9.77.

²¹ The etymology is a false one (see Frisk, H., Griechisches etymologisches Worterbuch, vol. 2 [Heidelberg, 1973] s.v. ὔμνος) but revealing of the poet's intuitive connection of song composition with weavingGoogle Scholar.

²² Solmsen (1965, 140), Long (1974, 411).

²³ Of the use of δέ for οὖν Denniston, J. D., The Greek Particles, 2nd ed. (Oxford, 1954), p. 170Google Scholar, says: ‘In general, there are few examples, and none are striking’.

²⁴ Cf. Hymn. Horn. 4.1–3, 5.1–6, 6.1–5 with an etymological play on Aphrodite's name, line 5–and 9.1–6.

²⁵ For the motif of the races of mortal creatures, cf. B26.4, BU5.7.

²⁶ Contrast also Kirk, Raven and Schofield (1983), p. 288 n. 1.

²⁷ The one scholar who has seriously considered the literary structure of the work in relation to its content is Munding (1954) [see n. 14 above]. He recognizes some important structural patterns (esp. 136), but he fails to see their significance. In the end he does not make effective use of the structural evidence.

²⁸ See Kahn, C., ‘A New Look at Heraclitus’, American Philosophical Quarterly 1 (1964), 189–203Google Scholar.

²⁹ Frye, N., Anatomy of Criticism (Princeton, 1957), pp. 73fGoogle Scholar.

³⁰ Millerd, C. E., On the Interpretation of Empedocles (Chicago, 1908)Google Scholar.

³¹ A. H. Griffiths, cited in Long (1974, 401 n. 9), has noticed how the passage echoes Iliad 6.146ff.

³² Parmenides B1.7f. mentions δοɩοì κύκλοɩ, a motif that seems to resonate with any number of pairs in the poem: the mares, the doors, the two routes, the two bodies of the Way of Opinion.

³³ In the first occurrence a prepositional phrase is functioning as an adverb.

³⁴ As another instance of poetic mimesis note how Empedocles builds to a thematic climax in line 13 by inserting ἀκίνητοɩ after the caesura: the term jtfdvides an oxymoron while the three consecutive long syllables retard the metre to a halt, after which the image of the circle resolves the conundrum. I am indebted to Daniel R. Blickman for the analysis.

³⁵ Contrast most recently Osborne (1987, 40) [see it. 7 above]: ‘the process of increasing plurality is open-ended, and thus there need be no superficial symmetry between the work of Love and that of Strife.’

³⁶ The appearance here of the subjunctive уένηταɩ is difficult after the aorist indicative ἴκετο in the preceding line. However, we seem to have here a switch to a present general statement, which is not too extreme a turnabout if ἴκετο is understood as a gnomic aorist. Cf. Wright ad he. She is wrong, however, to claim that the occurrence of уένηταɩ implies that Love moved into the centre from elsewhere. Empedocles simply uses the verb to mean ‘comes to be confined’. Cf.n. 38 below.

³⁷ I shall assume a masculine gender for Strife.

³⁸ The text is corrupt, but the general sense is one of contrast as the context indicates. I am reading Wright's ζωρά τε πρ ìνκε κρήτο.

³⁹ This is the interpretation of Guthrie, W. K. C., A History of Greek Philosophy, vol. 2 (Cambridge, 1965), p. 179Google Scholar, and O'Brien (1965, pp. 116f.); cf. n. 36 above. It makes good military sense as a conflict between a besieger and a besieged, with alternate advances and retreats. Wright's view (1981,207) that Love' strikes at the centre from her position at the ἔσϰατα τέρματα κύκλου leads to the curious supposition that there was a power vacuum in the centre of the cosmos before Strife moved in, and envisions the curious strategy of engaging the centre of a position without breaching the perimeters. Or, if we suppose that Strife has not quite secured the centre in his advance, we find Love rushing in to engage the front of a phalanx when the rear is exposed. Perhaps we could save the conception by comparing the surreptitious capture of an acropolis by a faction; but nothing in the text supports this interpretation. In general, the Guthrie-O'Brien view seems to be less problematic and to accord with the sort of territorial invasions and repulses that Anaximander already had found to be expressive of the strife of elements.

⁴⁰ Met 1.4.985a21ff. = A37; GC 2.6.333b19ff. = A40.

⁴¹ Word and Object (Cambridge, Mass., 1960), p. 91Google Scholar.

⁴² See Mourelatos, A. P. D., ‘Heraclitus, Parmenides, and the Naive Metaphysics of Things’, in Lee, E. N. et al. , edd., Exegesis and Argument (Assen, 1973)Google Scholar.

⁴³ Even Solmsen grants this much, p. 140.

⁴⁴ I am grateful to Daniel R. Blickman, A. P. D. Mourelatos, and especially A. A. Long for helpful suggestions.