¹ Hemsen, , Anaxagoras Clazomenius … disquisitionem historico-philosophicam (Gottingae, 1821) 20Google Scholar; Schaubach, , Anaxagorae Clazomenii fragmenta, etc. (Lipsiae, 1827) 27–28Google Scholar; Brandis, , Handbuch der Geschichte der Griechisch-Römischen Philosophie (Berlin, 1835–1860) i 190.CrossRefGoogle Scholar

² Millerd makes a half-hearted move in this direction. She writes that Anaxagoras, and Empedocles, show ‘no large measure of consideration of each other's theories’, On the Interpretation of Empedocles (printed dissertation, Chicago, 1908) 13.Google ScholarGigon, writes, ‘Die Stellen, di man für Abhängigkeit des Anaxagoras von Empedokles angeführt hat, sind nicht zwingend’, ‘Zu Anaxagoras’ in Philologus xci (1936–1937) 2–3.Google Scholar Two other writers consider, but only in passing, that there was no connexion between the two philosophers: Bignone, , Empedocle, studio critico, etc. 148 n. 1Google Scholar, and Jöhrens, Otto, Die Fragmente des Anaxagoras (printed dissertation, Göttingen, 1939) 82.Google Scholar Professor Guthrie also doubts that there was any relation between the two, A History of Greek Philosophy ii 128. Lanza writes, following Diog. Laert. viii 56, ‘I rapporti Empedocle-Anassagora paiono limitarsi ad un tipo di condotta personale’, Anassagora, testimonianze e frammenti, etc. (Firenze, 1966) 20 n.

³ Zeller, , Die Philosophie der Griechen, Teil i Abteilung 2 (6th edn. by Nestle, ) 1259–64.Google Scholar Zeller prefaces his argument with an honest admission of the difficulties. Burnet, , Early Greek Philosophy ⁴261–2 and 272–4Google Scholar, cf. Greek Philosophy Part 1 Thales to Plato 76–7. Burnet's extension of Zeller's argument at one point is modified by Nestle, in his edition of Zeller, 1246 n 2, cf. 1264 n. 2. Burnet's argument that Anaxagoras' theory of sensation shows ‘polemic’ against Empedocles is a slight extension of Zeller, 1264, and is repeated by Raven, , The Presocratic Philosophers 394Google Scholar (cf. section V below, especially 113).

Zeller's conclusion is followed explicitly by Dümmler, , Akademika, etc. (Giessen, 1889) 217Google Scholar; and Baeumker, , Das Problem der Materie, etc. (Münster, 1890) 73Google Scholar; also Überweg-Prächter, , Die Philosophie des Altertums ¹²98.Google Scholar The same assumption, without direct reference to Zeller, underlies several accounts of Anaxagoras: by Wellmann, , in Pauly-Wissowa, , RE i (1894) coll. 2076–7Google Scholar; Giussani, edition of Lucretius (Torino, 1896) ii 147; Covotti, , ‘Il “fisicissimo” del V° Secolo avanti Cristo, Anassagora di Clazomene’ in Atti della reale Accademia di Scienze morali e politiche (di Napoli) xliii (1915) 226–9Google Scholar; Robin, , La pensée grecque, édition revue et corrigée (1932) 147–8Google Scholar, cf. 119; Peck, , ‘Anaxagoras and the Parts’ in CQ xx (1926) 69–71Google Scholar, cf. CQ xxv (1931) 33; Bailey, , The Greek Atomists and Epicurus 33–5Google Scholar; Cornford, , ‘Anaxagoras' Theory of Matter’ in CQ xxiv (1930) 15 and 94–5Google Scholar; apparently Vlastos, , ‘The Physical Theory of Anaxagoras’ in Philos. Rev. lix (1950) 37–9Google Scholar; also Romano, Anassagora, in the series Pubblicazioni dell' Istituto Universitario di Magistero di Catania, serie filosofica, saggi e monografie, no. 52 (Padova, 1965) 19–20.

There is further discussion, leading to the same conclusion as Zeller and Burnet, in Unger, , ‘Die Zeitverhältnisse des Anaxagoras und Empedokles’ in Philologus Supplementband iv (1884) especially 550Google Scholar and cf. 516; Bidez, La Biographie d' Empédocle in Recueil de Travaux, etc. de l'Université de Gand (1894) 170–1; Millerd, in a modified form, as cited in the preceding footnote; Bignone, , Empedocle 148 n. 1Google Scholar, cf. 312 n 4; Capelle, , ‘Anaxagoras’ in Neue Jahrb. xliii (1919) 197Google Scholar (Gigon's argument, as cited in the preceding footnote, against Capelle hardly does more than expand Bignone, 440); and Jöhrens, , Die Fragmente des Anaxagoras 82–4Google Scholar, and in an Anhang, ‘Das zeitliche Verhältnis Empedokles-Anaxagoras’ 93–4.

Ciurnelli says that this chronology is ‘per noi … un fatto certo, sia sull'esame intrinseco del sistema, sia per l'attestazione esplicita di Aristotele’, La Filosofia di Anassagora, in the series Problemi d'Oggi, collana di filosofia e storia della filosofia (Padova, 1947) 19 n. 1.

Guthrie apparently does not come down on either side, but writes that, ‘To consider Empedocles before Anaxagoras is certainly the logical order’, History ii 128 n. 4, cf. 275 n. 1.

⁴ Tannery does not argue the point: his opinion is left to appear in the details of his description, Pour l'histoire de la science hellène (2nd edn. by Diès, ) 316, 325, 327.Google Scholar Apparently Sturz, , Empedocles Agrigentinus, de vita et philosophia eius, etc. (Lipsiae, 1805) 18.Google ScholarKarsten, , Empedoclis Agrigentini carminum reliquiae, etc. (Amstelodami, 1838) 47–55.Google Scholar Also Fazelli, , De rebus Siculis decades duae (1st edn., Panormi, 1558) 134–5.Google Scholar For Cleve's view see n. 18 below.

Millerd, , Interpretation of Empedocles 13Google Scholar, refers this view to ‘some modern treatments, notably those of Tannery and Gomperz’. Gomperz does not hold this view; it is not implied in his note to 183 (Griechische Denker i 447), which Miss Millerd mentions, 13 n. 3: cf. below n. 20. I have been unable to trace the other ‘modern treatments’ to which Miss Millerd refers. For a fault of scholarly doxography tending in the opposite direction see n. 18 below.

⁵ Cf. 109–10 and 113 below.

⁶ Davison, , CQ n.s. iii (1953) 39–45.Google ScholarDiano, , ‘La data di pubblicazione della syngraphè di Anassagora’ in Anthemon, scritti di archeologia e di antichità classiche in onore di Carlo Anti (Firenze, 1955) 235–52.Google Scholar

⁷ References to modern discussions are given by Guthrie ii 128. When Guthrie writes, 128 n. 2, that there are ‘no dissentients’ on the dating of Empedocles, he apparently overlooks the article by Unger already cited in n. 3, ‘Die Zeitverhältnisse des Anaxagoras und Empedokles’. Unger seeks to establish an absolute chronology, whereby Empedocles was born in 520 and published his poem before 472, while Anaxagoras lived from 533 to 462 and wrote his work in 466. Diek describes Unger's thesis as ‘eine bedauerliche Verirrung des scharfsinnigen Gelehrten’, ‘Gorgias und Empedokles’ in SBB, 1884, 344 n. 2. Bidez, , Biographie 130Google Scholar, writes, ‘M. Unger aboutît à une chronologie impossible; il propose, sans scrupule, les transformations de texte les plus considérables … C'est de la haute fantaisie’. On the other hand Cieve finds Unger's article a ‘very instructive essay’ and ‘a brilliant, detailed research … in which he [Unger] accounts for all his statements with really convincing reasons’, The Philosophy of Anaxagoras, an attempt at reconstruction (New York, 1949) ix–x and 117 n. 18. (This work is a rewriting of Die Philosophie des Anaxagoras, Versuch einer Rekonstruktion, by Loewy-Cleve, , published at Wien in 1917.Google Scholar) Cieve repeats his praise of Unger, in The Giants of pre-Sophistic Greek Philosophy, an attempt to reconstruct their thoughts (The Hague, 1965) i 170–1.Google Scholar Although the evidence for Anaxagoras is difficult, I agree with Diels and Bidez. There is a severe, but it seems to me not unfair, assessment of Cleve's work on Anaxagoras by Gershenson, D. E. and Greenberg, D. A., Anaxagoras and the Birth of Physics (New York, London, Toronto, 1964) 433–5.Google Scholar

Works that have been referred to in the Introduction will for the most part be cited from now on by the author's name and page reference alone.

⁸ Pp. 1020–1, cf. 1022; 2nd. edn. (1856) i 560–1. Diels also questioned the accuracy of Diogenes' quotation of Alcidamas, , Doxographi Graeci, note on 477.18.Google Scholar He later withdrew this, in SBB, 1884, 357 n. 1, but without making entirely clear what his later view entailed.

⁹ Alcidamas is mentioned in passing by Gigon, 3 n. 1, at greater length by Cleve, , Anaxagoras 116–17Google Scholar, and somewhat ambiguously by Kahn, , Anaximander and the Origins of Greek Cosmology 164.Google Scholar But since the republication of Cleve's discussion, Alcidamas has again been ignored. He is conspicuous by his absence in Raven and Guthrie.

¹⁰ 4th edn. (1876) i 743. This argument comes from Brandis, i 190. It is omitted in Zeller's 5th and 6th editions.

¹¹ Zeller adopts this suggestion from Karsten, 49. Timaeus also connected Empedocles with Pythagoras, ap. Diog. Laert. viii 54, and Hermippus and Neanthes connected him with Pythagoreans, ibid. 55 and 56, as Theophrastus also may have done, ap. Simpl. Phys. 25.21 = fr. 3 Diels, Dox. 477.18 (DK 31A7); cf. Diels' note ad loc., which is elaborated in SBB, 1884, 357 n. 1.

¹² This point is made by Cleve, , Anaxagoras 117 n. 17.Google Scholar

¹³ References have been given, n. 10 above.

¹⁴ Pythagoras is estimated to have died not later than 490, which is about the time that Empedocles was born: references to modern discussion are cited by Guthrie, i 173 and ii 128. For Empedocles' date see n. 7 above.

Diels appears to allow that Alcidamas would have said that Empedocles heard Pythagoras himself; but he is not altogether clear, cf. n. 8 above.

¹⁵ viii 54, cf. n. 11 above. Diels seems to suggest that Timaeus followed Alcidamas; but this is not necessary.

¹⁶ Pauly-Wissowa, , RE i 2(1894) coll. 1533–9.Google Scholar

¹⁷ Anaxagoras 116–17.

¹⁸ The temporal interpretation was taken by Carus, , Anaxagoreae Cosmo-Theologiae indagantur fontes (Leipzig, 1797) 5 n. 12Google Scholar, reprinted in Nachgelassene Werke iv, ed. Hand, F. (Leipzig, 1809) 698 noteGoogle Scholar; Hemsen, 59–60, cf. 20; Schaubach, 28 and 48; Brandis, i 242, cf. 190; and Clemens, , De philosophia Anaxagorae Clazomenii, etc. (printed dissertation, Berolini, 1839) 23Google Scholar; who have been followed by Unger, 550 cf. 516; Dümmler, 217; Bidez, 170–1; Robin, 119, cf. 147–8; Bailey, 34; Capelle, 197 n. 4; Cherniss, , Aristotle's Criticism of Presocratic Philosophy 219 n. 5Google Scholar; Jöhrens, 93–4; Ciurnelli, 19 n. 1; McDiarmid, , ‘Theophrastus on the Presocratic Causes’ in HSCP lxi (1953) 145 n. 90Google Scholar; Zafiropulo, , Empédocle d'Agrigente (Paris, 1953) 26Google Scholar; with hesitation by Raven, 363; and also by Romano, 19.

Gigon, 2, says that Zeller and ‘alle Späteren’ choose the chronological interpretation. This is inaccurate. Zeller, 1261 n. 2, states both ambiguities, and discusses them, but leaves the question open, as is recognised by Bidez, 170 n. 1. Burnet at first adopted this interpretation, EGP ¹ (1892) 286 n. 47, cf. 215 n. 21,² (1908) 303 n. 1; but sixteen years before Gigon wrote he had abandoned it, see the following footnote.

Cleve, , Anaxagoras 116–19Google Scholar, attempts to combine the temporal interpretation of ὕατερος with acceptance of Alcidamas' evidence, by arguing that Anaxagoras taught Empedocles but ‘began publishing later than did Empedocles who apparently was somewhat impatient in this respect and too eager to become famous’. This thesis is filled out with a good deal of further fanciful biographical material, in part a highly speculative extension of what Diels, , SBB, 1884, 357 n. 1Google Scholar, calls ‘eine alberne Faselei’, the account of λογοκλοπία in Timaeus, ap. Diog. Laert. viii 54 (DK 31A1). A distinction in time between Anaxagoras' oral and written teaching is in itself perhaps not impossible; but there seem to me no means of demonstrating it, nor any need to assume it.

Another piece of embroidery on the chronological interpretation has been to the effect that Anaxagoras ‘n'a publié son livre, d'après Aristote, que vers la fin de sa vie’, Mugler, , ‘Le problème d'Anaxagore’ in REG lxix (1956) 317Google Scholar, following Zafiropulo, , Anaxagore de Clazomène (Paris, 1948) 273.Google Scholar There is in fact no such implication in Aristotle. Zafiropulo and Mugler seem not to appreciate that in the passage quoted above τούτου is to be understood with ὕστερος as well as with πρότερος.

¹⁹ This was the view in antiquity of Alexander, , Met. 27. 26–28.21Google Scholar Hayduck. Brandis, i 242 n. i, allowed the possibility of this view, and it seems to be the sense preferred by Burnet, after he had abandoned the simply temporal interpretation, EGP (3rd and 4th edn., 1920 and 1930) 262 n. 1. Burnet has been followed by Ross, edition of the Metaphysics, ad loc. (p. 132), who nonetheless allows the possibility of the simply temporal interpretation, and most recently by Kahn, , Anaximander 163–5Google Scholar, whose argument, as we shall see, has been rejected by Guthrie.

²⁰ Breier proposed this as the general sense of the passage, but apparently without committing himself specifically on the meaning of ὕστερος, Die Philosophie des Anaxagoras von Klazomenä nach Amtoteles (Berlin, 1840) 85–6. Breier's interpretation was adopted, with specific reference to ὕστερος, by Bonitz, , edition of the Metaphysics (Bonnae, 1848–1849) ad loc. (ii 67)Google Scholar, who was approved by Diels, , ‘Über die Gedichte des Empedokles’, SBB, 1898, 412 n. 1Google Scholar, and evidently by Millerd, 13, and followed by Gigon, 2–3. This sense of ὕστερος seems to be preferred by Raven, 363, to the sense ‘inferior’, cf. n. 26 below.

Gomperz attempts to take ὕστερος in a logical sense, without chronological or evaluative implication, Griechische Denker i 447. He is criticised by Millerd, 13. Cf. Guthrie's remark, quoted above n. 3 sub finem.

²¹ Anaximander 163–5.

²² History ii 128 n. 4, cf. 275 n. 1. Cf. Cherniss, ACP 223 n. 26 and 234. The passage is Met. 1075b3–4: αὕτη δ' (se. Love) It is for this reason that in another place Aristotle calls mind an ἀρχή and Love a στοιχεῖον, Met. 1091b11–12. The same criticism is perhaps latent at De gen. et corr. 314a16–17, where Empedocles is said to have had six elements in all: στοιχεῖα is understood from line 15.

²³ The inconsistency is that Love is cause of separation (of a single element) as well as of mixture, while Strife is cause of mixture (of a single element) as well as of separation. Empedocles would probably not have granted that the disintegration of a single element, in order to be mixed widi other elements, would be separation, nor that the isolation of a single element, in order to be divided from the other elements, would count as mixture. That is, Empedocles would have thought of mixture and separation not as processes affecting a single element, but only as processes affecting a number of elements. Aristotle repeats his criticism briefly at De gen. et corr. 333b20–21. Empedocles' attitude to the question can probably be seen in fr. 22, which I have tried to analyse in Empedocles' Cosmic Cycle (Cambridge, 1968) especially 311–12 (henceforward ECC).

²⁴ Phys. 203b12–13; De part. anim. 640b4–8; Met. 988a32–34, 988b6–11. Anaxagoras and Empedocles are probably intended as sponsors of respectively a single and double moving cause at Met. 987a5–13, where no names are given. Love and Strife and mind are given as instances of a cause other than material necessity at Phys. 198b10–16. Love and Strife are reckoned as moving causes without comparison with mind at De gen. et corr. 314a16–17, cf. Met. 989a25–26.

²⁵ Especially for the passages 300b25–31 and 301a11–20 certain details of the paraphrase are taken from an analysis in ECC 10–14.

²⁶ Thus Kafka speaks of Empedocles', aim as ‘eine Reihe genialer Intuitionen mit poetischer Freiheit zu einer dramatischen Darstellung des Weltgeschehens zusammenzufassen’, Die Vorsokratiker, Geschichte der Philosophie in Einzeldarstellungen, Abteilung 2, Die Philosophie des Abendlandes im Altertum Band 6 (München, 1921) 103Google Scholar, repeated in Philologus lxxviii n. F. xxxii (1923) 203. Robin, 134, writes, ‘synthèse hésitante et confuse … beaucoup trop passionnée pour être assez systématique’. There are similar remarks in Millerd, 21–2, Gilbert, , Griechische Religionsphilosophie (Leipzig, 1911) 224Google Scholar, and elsewhere.

It is probably some such attitude to Empedocles that makes Gigon, 2, write no more of the possibility of Anaxagoras' being said to be inferior to Empedocles than that ‘Die dritte Bedeutung (i.e. “unvollkommener”) fällt von vorneherein weg’. Perhaps for the same reason, Raven, 363, writes that ὕστερος may mean more up-to-date ‘or even … inferior’ (my italics). The same attitude seems to affect Zeller, 1261 n. 2 sub finem, when he writes that Aristotle may have meant ‘superior’ if he had in mind the whole of Anaxagoras' teaching, ‘Indessen ist es auch möglich, dass er bei dem Prädikat das Ganze der anaxagorischen Lehre im Auge hat, in der er allerdings einen wesentlichen Fortschritt gegen die Früheren erkennt’. This will not be true if, as Zeller evidently intends, Empedocles is to be included among ‘die Früheren’. Zeller is influenced in part by his highly questionable interpretation of Love and Strife as ‘ganz mythische Gestalten’, 1259, whereby Empedocles ‘der mythischen Kosmogonie noch annähert’, 1262, whereas ‘in der Idee des Geistes (i.e. Anaxagoras' Mind) tritt … ein neues und höheres Prinzip in die Philosophie ein’, 1262.

²⁷ This may be checked from Bonitz's Index.

²⁸ Met. 986b18–34, a favourite passage for Simplicius.

²⁹ E.g. on the void, Phys. 214b12ff; on the lack of a moving cause, Met. 985b4–22; cf. 1071b31–37.

³⁰ See De caelo 279b14–17 and 280a11–23.

³¹ E.g. De gen. et coir. 333a16–334b2, especially 333b22–26.

³² Met. 1000a18–26. The compliment, not untypically, is followed by various adverse criticisms. The contents of the above paragraph are repeated with more detail in ECC 71–4.

³³ Met. g88b29–30, 989a5–12. Similarly, Aristotle tells us that earth alone of the four elements was not identified with soul, except on the theory that the soul consisted of all four elements, De anima 405a2–b10, a theory to which Aristotle supposes that Empedocles was committed, 404b7–15, cf. 410a21–22, and De gen. et corr. 334a9–10.

³⁴ De gen. et con. 328b31–329a5 and 330b7–20. Philoponus, 207.19 and 227.14 Vitelli. In Aristotle's later passage Plato is mentioned in connexion with a more sophisticated version of three elements.

³⁵ Cherniss, , ACP 219 n. 5.Google ScholarMcDiarmid, , HSCP lxi (1953) 145 n. 90.Google Scholar There is essentially the same argument in Jöhrens, 93–4.

³⁶ Breier, as cited above n. 20. Also Bonitz, as cited above n. 20. The argument is answered in part by Zeller, 1261 n. 2.

³⁷ E.g. De caelo 308b30–32.

³⁸ Cf. Zeller, 1261 n. 2, speaking of (his understanding of) the phrase in Aristotle, ‘nur einen rhetorischen, nicht einen logischen Gegensatz'.

³⁹ This latter possibility is roughly the view taken by Bidez, who appears to be alone in giving the question any attention, 170–1. (There is a passing remark in Bonitz, as cited above n. 20). Bidez believes that Anaxagoras is later than, and influenced by, Empedocles. He supposes that Aristotle's knowledge is gained from ‘détails perdus pour nous’, and not simply because Aristotle ‘a comparé seulement la théorie des principes chez les deux philosophes’. This would seem to be an acknowledgment that the internal evidence which we have is not sufficient to determine which philosopher, if either, influenced the other. On the other hand, Bidez seems to introduce an internal argument when he associates with his belief in the priority of Empedocles the view that Empedocles' ‘théorie des principes’ is ‘moins compliquée’ than that of Anaxagoras. (For some instances where Empedocles' theories were in fact apparently the more complex, see n. 65 and p. 111 below.)

⁴⁰ This possibility is rejected by Bidez, 170–1.

⁴¹ If we accept Theophrastus' statement that Empedocles was Simpl. Phys. 25.19–20 = fr. 3 Diels, Dox. Gr. 477.17 (DK 31A7), then it would be probable that Theophrastus, and so Aristotle, knew the absolute date of birth. McDiarmid, , HSCP lxi (1953) 145 n. 90Google Scholar, suggests that Theophrastus may have ‘merely drawn his own conclusion from Aristotle's remark’. Bidez, on the other hand, 23, calls Theophrastus' remark ‘une des données les plus sûres’.

⁴² Diog. Laert. ix 41 (DK 59A5, 68B5).

⁴³ Ap. Diog. Laert. viii 51 and 74 = fr. 71 Rose (DK31A1).

⁴⁴ In the Phaedo Socrates expected to hear from Anaxagoras, 98A, add. Koehler, Ficinus Cf. Laws 889b. Aristotle writes in a similar vein, Met. 982b12–17: Eclipses were also of course of superstitious significance, cf. Pliny, , NH ii 46.Google Scholar

⁴⁵ Tannery, , Science hellène, 325 and 327Google Scholar, states that Empedocles is dependent on Anaxagoras on this question, but without fully discussing the evidence or bringing out its implications. Subsequently the point has been overlooked (for Guthrie, see n. 49 below).

⁴⁶ Elsewhere Plutarch refers to the theory that the moon owes her light to the sun as De facie 929B (DK 59B18).

The ability to give the correct explanation of an eclipse in difficult circumstances is a feature of several of Plutarch, 's Lives, Dion 24Google Scholar (cf. 19); Pericles 35; Aemilius 17; cf. Agesilaus 17. In the present instance the eclipse and Nicias' superstition are mentioned by Thucydides, vii 50, and Diodorus, xiii 12.6. The excuse for Nicias' not knowing the correct explanation is probably Plutarch's own addition, an indication of his favourable attitude to Nicias. This is in part the opinion of Busolt, , ‘Plutarchs Nikias und Philistos’ in Hermes xxxiv (1899) 296–7.Google Scholar

⁴⁷ There are several sources which attribute the innovation of a written discourse to various early thinkers. Anaxagoras' place in this tradition is usefully analysed by Diano in the article already cited, n. 6 above, ‘La data di pubblicazione della syngraphè di Anassagora’: for the passage from Plutarch, Life of Nicias, see especially 244–7.

⁴⁸ See especially 123 below.

⁴⁹ Professor Guthrie, whose attention I had drawn to the passage from the Cratylus in connexion with Aristotle's ambiguity, quotes it in his note, ii 128 n. 4, as a corollary to Kahn's thesis, but without adding the complementary evidence from Plutarch and Hippolytus, and without allowing it to affect his interpretation of Aristotle's words: cf. also i 286 n. 1 and ii 66.

The continuation of Plato's account is taken into consideration in my next article, p. 127.

⁵⁰ De facie 929C–E (cf. DK 31B42 and 43).

⁵¹ ii 24.7 (DK 31A59), ii 28.5 (DK 31A60). The former entry is reconstituted by Diels.

⁵² It should perhaps be noted that the correct explanations of a lunar and a solar eclipse need not necessarily have gone hand in hand (although they appear in fact to have done so), for in the passage already cited from the Life of Nicias Plutarch says that it was common knowledge that some eclipses of the sun were in some way connected with the moon, whereas the explanation was not known of eclipses of the moon.

⁵³ Aet. ii 21.2 (DK 31A56). According to Aetius Empedocles has a theory of ‘two suns’, but the one in question in this entry, is apparently the same as τὸν φαινόμενον, see Aet. ii 20.13 (DK ibid.). Plutarch's description of the same theory, De Pyth. orac. 400B (cf. DK 31B44), although obscure, need not, I think, in any way imply that the sun is smaller than the earth.

⁵⁴ Fr. 48. This was apparently added to an explanation in terms of dark aer, for Pseudo-Plutarch, Stromateis 10 (DK 31A30), records that night is the hemisphere which is ‘mixed of aer and a little fire’; and Empedocles' aer is said to be dark, Plut. Quaest. conu. 720 E. It is not altogether clear how to correlate this with the account of Empedocles' ‘two suns’, Aet. ii 20.13 (DK 31A56).

⁵⁵ Aet. ii 31.1 (DK 31A61). I have attempted to determine the meaning of this entry in an article, ‘Anaximander's Measurements’ in CQ n.s. xvii (1967) 428–9.

⁵⁶ I may perhaps anticipate this much from my later article: that if you are content simply to deny the validity of the evidence on derived light before Empedocles and Anaxagoras, which is more or less what Tannery, Heath and Guthrie do, then it should still follow that, in the opinion of Plutarch and Hippolytus, Anaxagoras gave his explanation of eclipses before Empedocles, and that on grounds simply of fact there is no reason to question their evidence. Boll alone takes the alternative view, that since there were theories of derived light before Anaxagoras, then Anaxagoras cannot have been the first to give the correct explanation of an eclipse. See pp. 118–121 below.

⁵⁷ Quaest. conu. 722A–B (DK 59A74).

⁵⁸ For Plutarch's work on Empedocles, see [Lamprias], Cat. 43, as corrected by Treu, , Der sogenannte Lampriascatalog der Plutarchschriften (Waidenberg in Schlesien, 1873)Google Scholarad loc., and Hippolytus, Ref. v 20.6. Throughout his extant writings Plutarch quotes some hundred verses or part verses from Empedocles. This compares with some hundred and fifty verses or part verses quoted by Simplicius; some hundred verses or part verses quoted by Aristotle; and some fifty verses or part verses from each of Sextus, Clement and Hippolytus. Diogenes and Porphyry are the next main sources of verbatim quotations: each quotes some thirty verses. The whole question of Plutarch's knowledge of Empedocles I have discussed in more detail elsewhere, ECC 31–6.

⁵⁹ Cf. Diels, , Doxographi Graeci 146Google Scholar, on Diog. Laert. ix 20 (DK 21A1), and Hippol. Ref. i 14.1 (DK 21A33); and 165, on Diog. Laert. ix 30 (DK 67A1) where Diels comments, ‘inventionis nota biographum potius vel Variarum historiarum collectorem redolet’.

⁶⁰ Cf. Diels, Dox. 145–6.

⁶¹ Cf. Diels, Dox. 137–9 and 146 ff.

⁶² Theophrastus recognised Anaxagoras as the innovator of the moving cause, Simpl. Phys. 27.2–5 = Theophr. fr. 4, Diels, Dox. Gr. 478.18–21 (DK 59A41); cf. Diog. Laert. ii 6 (DK 59A1).

⁶³ Anaxagoras, fr. 17. Empedocles, frr. 8–12 et alib. Parmenides, especially fr. 8.38–41.

Burnet's Supplement, EGP 206 n. 1, of the lacuna in fr. 9.3 by νέμονσι, which strangely is not mentioned in the apparatus of Diels-Kranz, seems to me decidedly preferable as against Reiske's λέγουσι, followed by Sturz and Diels, or the other supplements that have been proposed: δοκέουσι by Karsten; φάσκουσι by Mullach, Fragmenta philosophorum Graecorum i ad loc.; and φασί with different words preceding by Panzerbieter, Beiträge zur Kritik und Erklärung des Empedokles (Meiningen, 1844) ad loc., and Friedländer, ap. DK ad loc.

Νέμειν in the sense of νομίζειν is found, not too infrequently, see LSJ, s.v. Burnet's supplement makes best use of the v in the manuscripts immediately preceding the lacuna. Also, if we allow at least a superficial similarity between νέμειν and νόμος, then νέμουσι answers well to νόμωι in line 5. Further, there is a very similar play on words in Plutarch's paraphrase of the fragment, Adu. Col. 1112F: (the last four words as corrected by Reiske).

The reading affects my argument in that νέμονσι has, at least superficially, a closer similarity than has λέγονσι, or the other supplements, to νομίζονσιν, which is the word used by Anaxagoras in precisely this context, fr. 17, and by Parmenides (νενόμισται) in a similar context, fr. 6.8. (Alternatively, the similarity may be taken as an additional argument for νέμουσι.)

⁶⁴ These are mostly noted by Zeller, 1260–4, especially 1260.

⁶⁵ Mind in Anaxagoras and Love and Strife in Empedocles. I count Love and Strife as causes of ‘movement or rest’, and not simply, as is usual, a pair of moving causes, because it seems to me that Love, unlike Strife, is primarily cause of rest and not of movement, in that when things are fully mingled they are at rest, and the period of rest under Love in the sphere lasts for as long as the period of change, separation and movement, when Strife has control of the elements. The evidence for these points I have tried to present elsewhere, ECC 101–3. From the present point of view this means that Empedocles' reaction to Parmenides is more complex than that of Anaxagoras; cf. n. 39 above.

⁶⁶ Anaxagoras: Diog. Laert. ii 9 (DK 59A1); Aet. ii 8.1 (DK 59A67). Empedocles: Aet. ii 8.2. (DK 31A58).

⁶⁷ Anaxagoras: Aet. ii 8.1. (DK 59A67); Irenaeus, Adu. haeres. ii 18.2 Harvey (DK 59A113); cf. Hippol. Ref. i 8.12 (DK 59A42); and Diog. Laert. ii 9 (DK 59A1); cf. also Zeller, 1246 n. 6. Empedocles: fr. 62 et alib.; cf. ECC 200–9.

⁶⁸ Anaxagoras: Aet. ii 8.1 (DK 59A67). Empedocles: Aet. v 26.4, ad init. (DK 31A70). This refinement is not noted by Zeller.

⁶⁹ Anaxagoras: Plut. Quaest. nat. 911D (DK 59A116); Nicolas Damascenus = [Arist.] De plantis 5. 9–13, 6. 17–18, 9. 17–21 Meyer (DK 59A117); cf. fr. 11. Empedocles: Aet. v 26.4 (DK 31A70), ad init., the first two passages cited above from Nicolas Damascenus; Simpl. De anima 72.2–3; cf. perhaps fr. 110.10, which is specifically said to include plants by Sextus, Adu. math, viii 286.

⁷⁰ Anaxagoras: Cens. vi 8 (DK 59A111). Empedocles: Cens. vi 6–7 (DK 31A81); Aet. v 11.1 (DK ibid.). Censorinus' belief that Empedocles used right and left in his account of sexual differentiation is open to question.

⁷¹ There is a straight conflict—one might suspect a conscious contradiction—on two points. Anaxagoras thought that stones were formed by cold, fr. 16; Empedocles thought that they were formed by the hot, [Arist.] Problem. 937a11–16 (DK 31A69). Anaxagoras, at least according to Theophrastus, thought that we perceived by unlikes, De sens. 27 (DK 59A92); Empedocles thought that we perceived by likes, fr. 109.

Cherniss, , ACP 301 n. 40Google Scholar, cf. 91 n. 387, following Heidel, , ‘Qualitative change in Pre-Socratic philosophy’ in AGPh. xix n. F. xii (1906) 369–72Google Scholar, denies the truth of Theophrastus’ report, wrongly as it seems to me. Cherniss' attitude stems in part from an exaggerated interpretation of Aristotle's classification of the attraction of likes or unlikes in the Presocratics, , ACP 91–2Google Scholar, a question which I have tried to analyse elsewhere, ECC 301–13.

⁷² Theophrastus' account includes: Plato, 500.7–13 (Diels, Dox. Gr.); Empedocles, 500.23–501.11; Alcmaion, 506.28–507.3; Anaxagoras, 507.8–14; οἱ πολλοί, 509.17–24 and 32–33; Cleidemus, 510.4–5; Diogenes, 510.19–22 and 511.9–12; Democritus, 513.17–27.

⁷³ Alcmaion explains how some eyes see μᾱλλον than others; Democritus explains how some eyes are ‘better’ than others; and Diogenes explains how some eyes see ὀξύτατα. Diogenes' account of day and night vision seems to be simply a copy of Anaxagoras'.

⁷⁴ 507.8–14 Diels. The three classes in Theophrastus' account are (i) line 10, Diels; (ii) line 10; and (iii) line 11. Number (iii) I take to be the same eyes as (i), but with regard to night time and not to day time: so that ἁπλῶς δέ means in effect the same as τοῖς μὲν πολλοῶς. What happens to number (ii) in day time is not explicitly stated. The notion of dark and bright eyes is expressed in Theophrastus by ὁμόχρως and ἀλλόχρως.

⁷⁵ 500.23–501.11, cf. 503.4 ff. and 504.14 ff. Diels; Arist. De gen. anim. 779b15–20 (DK 31A91); and [Arist.] Problem. 910a12–15 (not in DK). What follows is an explanatory paraphrase of Theophrastus. This seems to be required since there are misapprehensions in the only two detailed exegeses known to me of this part of Theophrastus, those of Beare, , Greek Theories of Elementary Cognition, etc., 19–21Google Scholar, and Stratton, , Theophrastus and the Greek physiological psychology before Aristotle (London and New York, 1917) 71–73 and 163–6.Google Scholar

⁷⁶ It would be possible theoretically to achieve the same effect by saying that the small amount of water present in the air by day is able to fill only some of the large number of water pores. But it seems preferable to adopt the explanation given a few lines later in the text, 501.6–7, where it is applied of course with opposite effect:

Stratton takes ἔλαττον to mean that there is less fire in these eyes than in fiery eyes, and not, as I have done, that in the one kind of eye there is less fire than water. There is the same notion in Beare's translation of 500.29 Diels, which follows one of the less plausible of several attempted emendations. But this notion is directly opposed to Aristotle's description in the De gen. anim., 779b16–17,

Stratton argues that the best eyes are said to have equal amounts of fire and water, so that eyes which see better in the day time cannot have less fire than water. But on this same supposition eyes which see better at night should have more water in them than fire. This interpretation however would make in line 4 (a description of eyes which see well at night) mean in effect a preponderance within the eye of water, and in lines 5–6 mean in effect a deficiency of fire. Stratton, whose view appears to be endorsed by Taylor (quoted by Stratton), has evidently misunderstood the whole paradox of Empedocles' theory, in that although by fire we see fire, nonetheless it is eyes with a preponderance of water which see best in the day time.

⁷⁷ It is natural to write ἀήρ instead of ὕδωρ in line 9 for the dark element outside the eye: for a similar explanation of aer as opposed to aither in fr. 100 see ECC 291–2.

⁷⁸ The last sentence quoted above from Theophrastus is correctly translated by Beare; but Taylor, as cited by Stratton, somehow supposes that ὕδωρ at the end of the last clause is the water of which the eye is composed. This is clearly wrong, and leads to a cumbersome interpretation of ἀποκριθῆι.

⁷⁹ The reason for the qualification ‘probably’ is that the functioning of bright eyes in the day time is not explicitly stated for Anaxagoras, see n. 74 above.

⁸⁰ Beare, 38, explains this difference between Anaxagoras and Empedocles in terms of outward-flowing fire as the active principle in Empedocles' theory of vision, an interpretation of Empedocles' theory which I discuss, and seek to reject, in an article which will be published in next year's JHS.