¹ Millerd, C. E., On the Interpretation of Empedocles (Chicago, 1908), 67.Google Scholar

² Aëtius 2.20.13 = Emped. A56a Diels-Kranz. Apart from being unnecessary and making the passage even more nonsensical, Bernardakis' supplement κυκλοτερος τς 〈αὐ〉γς is also contra-indicated by the final περ τν γν.

³ On the term ‘archetypal’, archetypos, cf. Schwyzer, H.-R. in Mansfeld, J. and de Rijk, L. M. (edd.), Kephalaion: Studies in Greek Philosophy and its Continuation Offered to Prof. de Vogel (Assen, 1975), 217–20Google Scholar, and for the idea in later Greek philosophy of two suns – one archetypal, the other its visible copy – cf. e.g. Lewy, H., Chaldaean Oracles and Theurgy (2nd rev. edn, Paris, 1978), 151–2Google Scholar with Boyancé's, P. comments, Études sur le Songede Scipion (Limoges, 1936), 73–4Google Scholar: the indebtedness to popular Platonism is obvious. For the idea of two celestial hemispheres rotating around the earth see below, with the refs in n. 14.

⁴ Diels-Kranz, , ad loc. (Die Fragmente der Vorsokratiker⁶ [Berlin, 1951]Google Scholar, i. 293.32, apparatus).

⁵ Even without the statements of later writers there would be no reason to doubt that Empedocles placed his earth at the centre of the cosmos; cf. Aristotle, De caelo 295a 13–34, Philo, De providentia 86.28–30 Awgerian, Guthrie, W. K. C., A History of Greek Philosophy ii (Cambridge, 1965), 198Google Scholar, Burkert, W., Lore and Science in Ancient Pythagoreanism (Cambridge, MA, 1972), 305Google Scholar, Tigner, S. S., Isis 65 (1974), 433–47CrossRefGoogle Scholar. The ‘crystalline sun’ mentioned by Aëtius is plainly a reference to the fiery hemisphere or ‘archetypal’ sun: compare the reports in Emped. A51, plus A1 § 77.

⁶ Millerd (as in n. 1), 68, refuted by Guthrie, , History ii. 194–5.Google Scholar

⁷ The quotation is from Guthrie, , History ii. 193Google Scholar. Cf. the detailed – and, from the point of view of the most elementary optics, totally erroneous – diagrams in Bollack, J., Empédocle iii (Paris, 1969), 259, 270Google Scholar; and (with more hesitation) Burkert, op. cit., 343 with n. 27. Burkert's apparent claim – ibid, with n. 24 – that the word anaklasis (Aëtius 2.20.13; cf. Plut. De Pyth. orac. 400b, and also ps.-Plutarch, Strom. 10 antanaklasis) can refer to the phenomenon of refraction as well as the phenomenon of reflection is also incorrect. The word is always used of the bending back of light towards its source; in the famous passage of Theophrastus, De igne 73 = Gorgias B5 Diels-Kranz, νάκλασις π τν λείων refers specifically to the reflection of light off smooth surfaces (as understood by Diels, H., Kleine Schriften [Darmstadt, 1969], 167–8Google Scholar; for τ λεῖα in the context of mirrors cf. e.g. Plato, Tim. 46a). Wright, M. R., Empedocles (New Haven, 1981), 201Google Scholar, like Guthrie, confuses the phenomena of ‘convex refraction’, which focuses light, and convex reflection, which dissipates it. Plutarch, loc. cit. = Emped. B44 is important as indicating that the reports about anaklasis in later writers derive in the first instance from Empedocles' own use of the word antaugein, and so providing independent confirmation that these reports ultimately refer to what was thought to be a phenomenon of reflection. See further below, with n. 34.

⁸ For the shape of Empedocles' earth cf. Millerd (as in n. 1), 63 n. 6; also Burkert, , Lore and Science 305.Google Scholar

⁹ Aëtius 4.14.1 = Emped. A88. In particular, knowledge of the use of concave mirrors for focusing sunlight (cf. e.g. Pliny, Hist. nat. 2.111.239) will have been widespread precisely because in practical life it was so useful. The quotation is from Dicks, D. R., Early Greek Astronomy (London, 1970), 55.Google Scholar

¹⁰ B41. For the language cf. Iliad 8.68, Plato, Phaedrus 246a, Timaeus 41a, Porphyry in Eusebius, Praep. evang. 3.12, and Lewy (as in n. 3), 49 with n. 156. For the masculine subject – i.e. Sun as opposed to Moon – compare B40, 42.1, 47 and Wright, , Empedocles 200.Google Scholar

¹¹ 2.23.3 = Emped. A58b; Guthrie, , History ii. 196.Google Scholar

¹² Strom. 10 = Emped. A30. The construction of the final clause is difficult: no doubt because pseudo-Plutarch had trouble understanding his source. But the general sense is clear, and Diels' emendation of κατ τν θροισμν in the final clause to κατά τι τν θροισμν serves (as noted by Bollack, , Empédocle iii. 218Google Scholar) no function whatever.

¹³ Night: Plutarch, Quaest. Plat. 1006e–f = Emped. B48 (for ϕαέεσι cf. Kranz, ad loc, and Hesiod, fr. 252.4 Merkelbach–West: as with B41 [above, n. 10], the reference is plainly to our familiar sun). Solar eclipses: B42.

¹⁴ Bouché-Leclercq, A., L'Astrologie grecque (Paris, 1899), 155–7, 276–96Google Scholar; Cumont, F., Lux perpetua (Paris, 1949), 193Google Scholar; Weisser, U., Das ‘Buch über das Geheimnis der Schöpfung’ von Pseudo-Apollonios von Tyana (Berlin, 1980), 193 n. 119.CrossRefGoogle Scholar

¹⁵ The discrepancy is well noted by Kafka, G., Philologus 78 (1923), 214Google Scholar. Cf. Emped. A59, B42.

¹⁶ For Empedocles' sun as made of fire cf. B21.3–4 (Wright, , Empedocles 177Google Scholar), B22.2, B27.1–2, B41 with Macrobius' comments, Sat. 1.17.46, B71.2, B115.9–11; Diogenes Laertius 8.77 = Emped. A1, Aëtius 2.6.3 = A49b, and Philo, De providentia 86.18–23 Awgerian; Wright, , Empedocles 23Google Scholar.

¹⁷ Aëtius 2.20.12 = Philolaus A19; cf. e.g. Burnet, J., Early Greek Philosophy⁴ (London, 1930), 298 n. 1Google Scholar, Guthrie, , History ii. 194.Google Scholar

¹⁸ Cf. e.g. Baltussen, H., Theophrastus on Theories of Perception (Utrecht, 1993), 145 and n. 42Google Scholar on Theophrastus' fondness for postulating a hypothetical τις, ‘someone’, in his refutations of Presocratic views (note in particular Theophr. De sensibus 12, 14, 23); ibid., 146 and n. 48, 155 and n. 85 on his use of future verb forms (as λέξει here) in criticizing them; and ibid., 151 on his forcing of the evidence in his ‘eagerness to refute wherever and whenever possible’.

¹⁹ Cf. esp. McDiarmid, J. B., HSCP 61 (1953), 85–156Google Scholar = Furley, D. J. and Allen, R. E. (edd.), Studies in Presocratic Philosophy i (London, 1970), 178–238Google Scholar; also Stratton's, G. M. comments, Theophrastus and the Greek Physiological Psychology before Aristotle (London, 1917), 52–4, 60–2Google Scholar, Guthrie, , History ii. 160Google Scholar (Aristotle), 232–4, 438–9, 444–5, Stevenson, J. G., JHS 94 (1974), 138–43CrossRefGoogle Scholar, and the note above. For Aristotle's predominant role in determining the methodology and interests of ‘doxographic’ tradition as known to us see J. Mansfeld in Fortenbaugh, W. W. and Gutas, D. (edd.), Theophrastus: His Physical, Doxographical, and Scientific Writings (New Brunswick, 1992), 63–111Google Scholar. The modern desire to whitewash both Aristotle and Theophrastus by playing down the element of wilful distortion in their treatment of the Presocratics (so e.g. Berti, E. in Cambiano, G. (ed.), Storiografia e dossografia nella filosofia antica [Turin, 1986], 101–25Google Scholar; Baltussen, op. cit., 132–94, 252–3) is misplaced, and sidesteps the essential issues. Certainly their misrepresentations can be said to be justified within a framework of Aristotelian dialectic; but that says nothing about their effect–both direct and indirect–on our understanding of the Presocratics themselves.

²⁰ ps.–Plutarch, Strom. 10 = Emped. A30; Philo, De providentia 86.18–23 Awgerian; Aëtius 2.6.3 = A49b. Cf. also the term περιαπλωθναι, Aëtius 5.26.4 = Emped. A70.

²¹ Emped. A30, A51 plus A1 §77. In describing this solidifying effect of fire Empedocles no doubt had in mind, among other things, the phenomenon of salt being crystallized by the heat of the sun: cf. Emped. B56, Philo, De prov. 86.38–41 Awgerian, Longrigg, J., CQ 15 (1965), 249–51CrossRefGoogle Scholar, esp. 251 with n. 3.

²² A53, A54.

²³ A30; Guthrie, , History ii. 186 n. 3Google Scholar, 187, Bollack, , Empédocle iii. 219Google Scholar. Cf. e.g. Iliad 5.91, 7.343, Odyssey 24.344, and for the idea of an ‘up’ and ‘down’ to Empedocles' universe, O'Brien, D., Empedocles' Cosmic Cycle (Cambridge, 1969), 298–9.Google Scholar

²⁴ Strom. 10 = Emped. A30.

²⁵ Aëtius 2.6.3 = Emped. A40b; Philo, De prov. 86.21–3 Awgerian. For the text of the Armenian Philo passage, with translation and commentary, see my forthcoming Ancient Philosophy, Mystery and Magic: Empedocles and Pythagorean Tradition (Oxford, 1995), ch. 2.Google Scholar

²⁶ D.L. 8.77 = Emped. A1 ad fin.

²⁷ Emped. B41 with Apollodorus, , FGrH 244Google Scholar F95 §46 = Macrobius, Sat. 1.17.46; above, n. 10, 13.

²⁸ A50b (cf. A 5 8 b ); A75.

²⁹ A30, A60, B43, 45, 47.

³⁰ Dreyer, J. L. E., History of the Planetary Systems from Thales to Kepler (Cambridge, 1906), 288–9Google Scholar. Aëtius' attribution of the idea of a crystalline heaven to Anaximenes (2.14.3 = DK 13 A14b: on the different sources see West, M. L., Early Greek Philosophy and the Orient [Oxford, 1971], 102–3)Google Scholar is almost certainly mistaken and due to confusion in the doxographic lemmata: cf. Longrigg, J., CQ 15 (1965), 249–50CrossRefGoogle Scholar, Kirk, G. S., Raven, J. E. and Schofield, M., The Presocratic Philosophers² (Cambridge, 1983), 152, 155–6.Google Scholar

³¹ Cf. esp. Boyce, M., A History of Zoroastrianism i ² (Leiden, 1989), 132–3Google Scholar; Livingstone, A., Mystical and Mythological Explanatory Works of Assyrian and Babylonian Scholars (Oxford, 1986), 86, 88Google Scholar; and for transmission to the Arabs, Heinen, A. M., Islamic Cosmology (Beirut, 1982), 78–83, 86Google Scholar and passim. See also Eisler, R., Weltenmantel und Himmelszelt (Munich, 1910), i. 94Google Scholar, Kingsley, P., Journal of the Royal Asiatic Society³ 2 (1992), 339–46CrossRefGoogle Scholar. It will be noted that the later concept of the crystalline heaven as transparent, entirely invisible and merely to be inferred by the mind (cf. e.g. Moore, E., Studies in Dante, Third series² [Oxford, 1968], 12)Google Scholar is a pale reflection indeed of this original, intensely visual idea.

The presence of the idea of a crystalline heaven in the Book of Revelation needs to be understood by referring in the first instance not to the same idea in Empedocles (Boll, F., Aus der Offenbarung Johannis [Leipzig, 1914], 17 and n. 1Google Scholar), but to the fact that this notion of a crystalline or bejewelled heaven was so prevalent in the ancient Near East that it had already exerted a powerful influence on Jewish apocalyptic tradition (Kingsley, loc. cit.; cf. also Clemen, C., NJb 35 [1915], 27Google Scholar). As far as the Zoroastrian evidence is concerned, the testimony of Curtius Rufus regarding Darius III's use of the symbol of a sun enclosed in crystal during his military campaigns (imago solis crystallo inclusa: History of A1exander 3.3.8) is particularly significant. It tends to confirm not only the familiarity of this sun-in-crystalline-heaven motif in Zoroastrianism by the 4th century b.c., but also the antiquity of the explanation of the Avestan word asmān – as meaning both ‘sky’ and ‘crystal’ – which is known to us from later sources (cf. Bailey, H. W., Zoroastrian Problems in the Ninth-Century Books² [Oxford, 1971], 120–48).Google Scholar

³² Emped. B44. For ‘Olympus’ as the vault of heaven cf. Parmenides B1 1.2–3, Philolaus A16, Kerschensteiner, J., Kosmos (Munich, 1962), 50, 57–8, 218Google Scholar, Burkert, , Lore and Science 244–5 with n. 31, 36.Google Scholar

³³ Any assumption that the word must have the more specific sense of ‘shining back’ and so ‘reflecting’is refuted by Chaeremon, , TrGF 71 F14.6Google Scholar(ντηύγει)and Xenophon, Cyneg. 5.18 (δι τν νταύγειαν…ντιλάμπει), where the idea is certainly not of reflection but of something shining out brightly in contrast to its surroundings. Similarly the word ᾽Ανταύγης, used in Orphic poetry as a name for the sun (fr. 237.4 Kern, Orphic Hymn 6.9; cf. also 70.7), hardly means ‘Reflector’ – which makes no sense – but ‘He who shines out brightly’. Again, although the description of the world of the stars in the pseudo-Hippocratic On Sevens as νταύγεια κα μάνωσις κα ραιωτάτη τς ϕύσιος λαμπηδών ‘brilliance and rarefied substance and the most refined illumination in nature’ (for the text cf. West, M. L., CQ 21 [1971], 368 §1.2Google Scholar; ραιωτάτη must, however, be retained), was interpreted in late antiquity as implying that stars shine with the reflected light of the sun (ps.-Galen, CMG XI.ii.i. 10.16–17) from the context it is clear that originally the reference was just to the brilliance – so the splendor of the Latin translation – and rarefied nature of the stellar sphere. It is also important to note that the term antaugeia can, even in technical writing, be used when referring not to reflection or reflected light but to the initial source of illumination prior to the stage of reflection: cf. e.g. Aëtius 2.20.12 = Philolaus A19, Mugler, C., Dictionnaire historique de la terminologie optique des Grecs (Paris, 1964), 36Google Scholar. Note as well the use of the word νταυγάζειν in Philo, De spec. leg. 1.321 (v. 77.19 Cohn-Wendland), meaning not ‘to reflect’ but ‘to illuminate’ (like the sun). Pindar, Olymp. 3.20, provides a particularly close parallel to our Empedocles passage with its use of the verb ντιϕέγειν to portray the moon ‘shining directly in front of someone’, ‘shining in someone's face’. Compare, similarly, the verb ντιλάμπειν, ‘to shine straight at’, ‘to shine out brightly’, in Xenophon, loc. cit., in Plutarch, De recta rat. aud. 41c, and used of the sun in Arrian, Tact. 21A (ἥλιος κατ προσώπου ντιλάμπων); and the word ντιϕωτισμός, ‘direct light’, ‘light shining straight at one’ (not ‘reflection of light’, LSJ) in Plutarch, Quaest. conviv. 625e. For the same spread of meanings – ‘radiate’, ‘shine brightly’, ‘illuminate’, but also ‘shine with reflected light’ – compare the Latin refulgere, and the modern Greek ανταύγεια/ανταυγάζειν.

³⁴ Cf. esp. Plutarch, De Pyth. orac. 400b (νταύγεια…νταγεῖ); also Aëtius 2.20.13 and 2.21.2 = Emped. A56 (νταύγεια…νάκλασις), and ps.-Plutarch, Strom. 10 = A30 (DK i.288.29–30) with Guthrie, , History ii. 192–3.Google Scholar

³⁵ See above, n. 3.

³⁶ Aëtius 2.8.2 = A58a. The idea which has sometimes been voiced (e.g. Karsten, S., Empedoclis Agrigentini carminum reliquiae [Amsterdam, 1838], 425Google Scholar, Diels, , Doxographi Graeci [Berlin, 1879], 338Google Scholar.8 app.) that Empedocles is here referring to the tilting of the earth's poles is totally unhistorical. From the Presocratics (Anaxagoras A1 §9, A67, Archelaus A4 §4) through to Psellus (De omnifaria doctrina 160 Westerink), there was no possibility of ‘the tilting of the poles’ referring to anything but the tilting of the celestial poles; the notion of a tilting of the terrestrial poles (e.g. Milton, Paradise Lost 10.668–71) only came to have any sense, or application, after the time of Copernicus.

³⁷ Aëtius 2.13.11 = A54; cf. Guthrie, , History ii. 420 n. 2Google Scholar, Burkert, , Lore and Science 311Google Scholar with n. 65, West, M. L., JHS 100 (1980), 208.Google Scholar

³⁸ Cf. e.g. Kerschensteiner (as in n. 32), 136; West, M. L., The Orphic Poems (Oxford, 1983), 214–15.Google Scholar