¹ Wilpert, P., ‘Die aristotelische Schrift “Über die Philosophie”’, Autour d'Aristote: Receuil … A. Mansion (Louvain 1955) 99–116 (quotation, 102–3)Google Scholar; cf. also ‘The Fragments of Aristotle's Lost Writings’, in Düring, I. and Owen, G. E. L., edd., Aristotle and Plato in the Mid-Fourth Century, Stud. Gr. et Lat. Goth, xi (Göteborg 1960) 257–64Google Scholar. A similar concern in connection with specific problems has been voiced by others, e.g., Mansion, A., ‘L'immortalité de l'âme et de l'intellect d'après Aristote’, RPhLouvain li (1953) 450Google Scholar and Moraux, P., ‘Quinta Essentia’, RE xlvii (1963) 1219Google Scholar.

² To my knowledge, the only published rejections of this assumption are Ross, W. D., Aristotle's Physics (Oxford 1936) 96–7Google Scholar, and Furley, D. J., ‘Lucretius and the Stoics’, BICS xviii (1966) 22–3Google Scholar; but neither has affected the state of the question. See, for example, the recent reconstruction by Chroust, A. H., ‘A Tentative Outline for a Possible Reconstruction of Aristotle's Lost Dialog On Philosophy’, AntClass xliv (1975) 553–69Google Scholar, esp. 561–3. One authority, Effe, B., Studien zur Kosmologie und Theologie der aristotelischen Schrift ‘Uber die Philosophie’, Zetemata I (Munich 1970) 127–8Google Scholar, is so convinced of the truth of this assumption that he is prepared to doubt Cicero's reliability as a witness to Aristotle's De Philosophia on the grounds that Cicero, Nat. D. ii 42Google Scholar does not acknowledge the existence of the fifth element.

³ Jaeger, W., Aristotle: Fundamentals of the History of His Development (Oxford 1948) 139, 142–54Google Scholar; Bignone, E., L'Aristotele perduto e la formazione filosofica di Epicuro (Florence 1936)Google Scholar.

⁴ Jaeger (n. 3) 139, n. 1. Bignone (n. 3) ii 352 n. 1, accepts this interpretation of the passage but does not use it as a proof that the fifth element was discussed in De Philosophia. This approach is adopted as proof by a number of more recent writers, e.g., Berti, E., La filosofia del primo Aristotele (Padua 1962) 369Google Scholar; Chroust, A. H., Aristotle: New Light on his Life and on Some of his Lost Works (London 1973) ii 183–4Google Scholar (originally published as ‘The concept of God in Aristotle's On Philosophy [Cicero, , De Natura Deorum I.13.33Google Scholar]’, Emerita xxxiii [1965] 205–28)Google Scholar; P. Moraux (n. 1) 1196–1209; Pepin, J., Théologie cosmique el théologie chrétienne (Paris 1964 151–2Google Scholar; and Pötscher, W., Strukturprobleme der aristotelischen und theophrastischen Gottesvorstellung, Philosophia Antiqua xix (Leiden 1970) 34, 41CrossRefGoogle Scholar.

⁵ Bernays, J., Die Dialoge des Aristoteles in ihrem Verhältnis zu seinen übrigen Werken (Berlin 1863) 99–100Google Scholar.

⁶ For the polemical purpose of his work, see Bernays (n. 5) 30–42; cf. also Berti (n. 4) 19–21.

⁷ Schönmann, G. F., ed., De Natura Deorum (Leipzig 1850, 1857, 1865, 1876) notes on i 13.33Google Scholar. It should be noted that Jaeger cites Plasberg, a commentator on Cicero, to support his interpretation. The other major commentators on this work concur; cf. J. B. Mayor (Cambridge 1891) i 122; and A. S. Pease (Cambridge, Mass. 1955) i 242.

⁸ Cf. von Arnim, H., Sto. Vet. Fr. (Leipzig 1903–1905) ii frr. 413, 527, 555, 558, 580Google Scholar, cf. 434. The use of ‘fire’ and ‘ether’ as alternative names for the celestial element is stated explicitly by Diog. Laert. vii 137 ( = von Arnim ii a fr. 580). In general, Stoic texts use ‘fire’ and ‘ether’ interchangeably.

⁹ Cf. Paulson, J., Index Lucretianus (Göteborg 1911)Google Scholar s.v. ‘Aether’. The meaning of the term in Lucretius is not completely clear, perhaps because the traditional elements play only a small part in the Epicurean system (cf. C. Bailey's edition [Oxford 1947] iii 1393–4). On the development of the Latin philosophical vocabulary before Cicero and Cicero's attitude toward and use of this vocabulary see Gigon, O., ‘Cicero und die griechische Philosophie’ in Aufstieg u. Niedergang d. röm. Welt ed. Temporini, H., Sect. 1, iv 1 (Berlin 1973) 250–4Google Scholar.

¹⁰ The statement in the Physics presents some problems. It occurs in a chapter whose authenticity has been questioned (cf. Wicksteed, P. H. and Cornford, F. M., Aristotle: The Physics [Loeb 1963] 314–19)Google Scholar. Furthermore, it is unclear whether by οὐρανός Aristotle means the universe as a whole, as the previous lines suggest (Phys. iv 5.212b17–20), or the heavenly region consisting of the spheres of the heavenly bodies, as a parallel passage in Cael. ii 4.287a32–b4 suggests. If οὐρανός means the universe as a whole, the passage would seem to presuppose a four-element cosmology (cf. Solmsen, F., Aristotle's System of the Physical World [Ithaca 1960] 301)Google Scholar, and ‘ether’ then refers to the celestial fire. If, on the other hand, οὐρανός means the heavenly region, the region of the celestial element, it probably presupposes Aristotle's standard five-element cosmology (cf. H. J. Drossaart-Lulofs [n. 33] 127). In that case ‘ether’ refers to the sublunar element, fire. In either case, the word ‘ether’ refers to fire, not to the fifth element.

¹¹ In this Jaeger is followed by Chroust (n. 4) ii 403–4. n. 78.

¹² Bignone (n. 3); Heitz, E., Die verlorenen Schriften des Aristoteles (Leipzig 1865) 179–89Google Scholar

¹³ Cf., e.g. Heitz (n. 12) 185–6 and J. Pepin (n. 4) 222–3. I have presented the proof in its bluntest form. It is usually toned down somewhat and qualified with a word like ‘probably’. For example, Pepin, after admitting these expressions could be derived from De Caelo by an imprecise doxographer, continues, ‘But nothing prevents us from supposing that these expressions in reality belong to an earlier state of his terminology’. He then chooses the latter interpretation because Cicero's divergence from the extant treatises has convinced him that the doxographic tradition is inspired by De Philosophia as well as De Caelo. Jaeger (n. 3) 144, n. 2 uses the doxography only as evidence for Aristotle's terminology, since he has already accepted the presence of this element on the basis of the text of Cicero.

¹⁴ Cf., e.g. Heitz (n. 12) 186–8; and Pepin (n. 4) 223 and n. 2, for explicit discussion of attribution. Most writers, however, simply accept the attribution without discussion.

¹⁵ Philo, De Gig. 2.7–8; De Plant. 3.12; Aet. Mund. 14.45 (collected by Walzer, R., Aristotelis Dialogorum Fragmenta [Florence 1934] 87–8Google Scholar as De Phil. fr. 22). Cf. also Plato, , Epin. 984d–5bGoogle Scholar.

¹⁶ Jaeger (n. 3) 143–6.

¹⁷ Mariotti, S., ‘Nuove testimonialize ed echi dell' Aristotele giovanile’, Atene e Roma xlii (1940) 56Google Scholar n. 22; Pepin (n. 4) 223.

¹⁸ Cf. the list of the books and articles in Moraux (n. 1) 1215 and Berti (n. 4) 103–7. The book by Pepin (n. 4) is another work in this vein. The method has even been applied to Aristotle himself to attribute Cael. i 3.270b16–25 and Meteor. i 3.339b19–30 to De Philosophia because these passages do not fit into their contexts very well (Effe [n. 2] 39–41).

¹⁹ Cf. Bignone (n. 3) ii 406–503. Bignone, 425 and n. 3, takes the term ‘ether’ in Lucr. v 128 and 143 as further confirmation of the presence of the fifth element in De Philosophia. He apparently has not noticed that in Lucr. v 143 it occurs in a series consisting of earth, fire, water, and ether, and therefore more likely designates air than Aristotle's fifth element.

²⁰ Cf. Wilpert, , ‘Lost Writings’ (n. 1) 262–3Google Scholar.

²¹ Cf. Easterling, H.J., ‘Quinta Natura’, MusHelv xxi 1964) 79–80Google Scholar.

²² The doxographic tradition as a whole is drawn from too many Aristotelian sources to allow us to trace given placitum back to a single work. Diels, H., Doxographi Graeci (Berlin 1929) 215CrossRefGoogle Scholar lists some of the references to extant treatises.

²³ The term ‘nameless’ (ἀκατονόμαστον), which late doxographic accounts claim is Aristotle's own term for the substance of the heavens (cf., e.g., Clem. Rom. Recogn. 8.14 [ = De Phil. fr. 27 Ross] and Psellus, , De Omnifaria Doctrina 131Google Scholar [p. 69 Westerink]), is itself as inadequate as the term ‘fifth body’ for grounding the hypothesis that the fifth element was discussed in a lost work. Aristotle's non-commital account of the traditional name ‘ether’ and his explicit criticism of Anaxagoras' use of this term (Cael. i 3.270b16–25), combined with his own preference for descriptive paraphrases rather than a single name, is sufficient to account for the doxographical term. Furthermore, it occurs only in the later doxographies and may well be influenced by the doxographies regarding the substance of the soul. In that event, the term loses its value as independent evidence for Aristotle's cosmological doctrine and must be taken as part of the doxography of Aristotle's doctrine of the nature of the soul.

²⁴ Cf. Nuyens, F., L'évolution de la psychologie d'Aristote (Louvain/Paris 1948)Google Scholar. The evolution of Aristotle's psychology has been the subject of much debate since Nuyens. For a brief survey see Fortenbaugh, W., ‘Recent Scholarship on the Psychology of Aristotle’, CW lx (1967) 318–20Google Scholar.

²⁵ These suggestions have been made in attempts to save Cicero's credibility. Lefevre, C., ‘“Quinta Natura” et psychologie aristotelicienne’, RPhLouvain lxix (1971) 5–43Google Scholar, accepts the double change, whereas Pepin (n. 4) 245–7 suggests that Aristotle regarded the fifth element as incorporeal. Pepin's suggestion, which cannot be proven for De Philosophia and flies in the face of Cael. i 3Google Scholar, seems to be an act of desperation to save Cicero's reputation.

²⁶ Cf., e.g., Moraux (n. 1) 1213–26 and Easterling (n. 21) 73–85.

²⁷ One possible ground for misunderstanding is that Cicero or his source learned that Aristotle considered the soul to be of a nature totally distinct from that of the four corporeal elements of the body (whether the view of the Eudemus or the view of the extant treatises that the soul is the first actuality of the body) and then mistakenly identified this nature with that of the element of the celestial bodies, which in the treatises is regarded as a fifth corporeal element distinct from the earthly four. If this misunderstanding did not afflict Cicero, it certainly did later writers. (For full discussion see Easterling [n. 21] 73–85; cf. also Effe [n. 2] 148–55; and Moraux [n. 1] 1213–24.)

Another possibility is that Cicero's statement reflects a distortion of the unique account of how faculties of soul are passed on to offspring (Gen. An. ii 3.736b29–737a1Google Scholar; on this passage cf. Solmsen, F., ‘The Vital Heat, the Inborn Pneuma and the Aether’, JHS lxxvii [1957] 119–23)CrossRefGoogle Scholar. Here Aristotle claims that the faculty of soul is ‘associated with (κϵκοινωνηκέναι) body that is different from and more divine than the so-called [four] elements’, a body that he goes on to identify as ‘the pneuma … and the natural substance within the pneuma that is analogous to the element of the stars’. He does not actually claim that the soul consists of pneuma or of the unnamed substance in the pneuma, but he does bring the soul into close association with it by saying it shares in (κϵκοινωνηκέναι) this corporeal substance, and he clearly states that the unnamed component of the pneuma that is significant in the transmission of psychic faculties is analogous to the element of the stars. It would not be too difficult for a reader of this text to identify the soul with this unnamed substance in the pneuma and then go on to identify this substance with the fifth element that constitutes the heavenly bodies.

²⁸ For a summary of the controversy with relevant bibliography see Berti (n. 4) 395–9 and Chroust, , Aristotle (n. 4) ii 194–205Google Scholar. To the works cited there add Pepin (n. 4) 226–34; Moraux, P., Aristote: Du Ciel (Budé, 1965) li–lviGoogle Scholar; Berti, E., ‘Studi recenti sul Peri Philosophia di Aristotele’, Giorn. Metafisica xx (1965) 310–11Google Scholar; Bos, A. P., On the Elements: Aristotle's Early Cosmology, Bijdragen tot de Filosofie iii (Assen 1973) 138–40Google Scholar; Lefevre (n. 25); and Chroust, A. H., ‘The Akatonomaston in Aristotle's “On Philosophy”’, Emerita xl (1972) 461–8Google Scholar.

²⁹ The Eudemus has frequently been suggested, e.g., by Mariotti, S., ‘La “quinta essentia” nell'Aristotele perduto e nell'Accademia’, RFIC xviii (1940) 179–89Google Scholar; Gigon, O., ‘Cicero und Aristoteles’, Hermes lxxxvii (1959) 143–62Google Scholar, esp. 153, 156 ( = Studien zur antiken Philosophie [Berlin 1972] 305–25Google Scholar, esp. 315, 318–19); Prolegomena to an Edition of the Eudemus' in Düring, and Owen, (n. 1) 32Google Scholar; and Grilli, A., ‘Cicerone e l'Eudemo’, Par. Passato xvii (1962) 98–100Google Scholar.

³⁰ On the history of Aristotle's books see Düring, I., Aristotle in the Biographical Tradition (Göteborg 1957) 393–5Google Scholar; Chroust, A. H., ‘The Miraculous Disappearance and Recovery of the Corpus Aristotelicum’, Class, et Med. xxiii (1962) 50–67Google Scholar; and the thorough, fully documented discussion by Moraux, P., Der Aristotelismus bei den Griechen i (Berlin 1973) 3–44CrossRefGoogle Scholar. The history of Aristotelianism in the first century B.C. is now fully discussed by Moraux, , Aristotelismus iGoogle Scholar; for a brief summar y cf. Moreau, J., Aristote et son école (Paris 1962) 279–83Google Scholar.

³¹ Cf. Plezia, M., ‘De Andronici Rhodii Studiis Aristotelicis’, Polska Ak. Archivum Filologiczne xx (Krakow 1946)Google Scholar; Düring (n. 30) 420–5; and Morau x (n. 30) 45–94. Whether Andronicus worked in Rome or in Athens and whether before or after the death of Cicero are matters of current debate (cf. Moraux 45–58).

³² See Moraux, P., ‘Xenarchos (5)’, RE xviii (1967) 1423–6Google Scholar; Aristotelismus (n. 30) 198–206.

³³ This summary has been preserved in a Syriac translation, ed. with English trans, and comm. by Drossaart-Lulofs, H. J., Nicolaus Damascenus: On the Philosophy of Aristotle (Leiden 1965)Google Scholar. For the life and philosophical activity of Nicolaus see Drossaart-Lulofs 1–5, 20–3 and Moraux, , Aristotelismus (n. 30) 445–50Google Scholar. For his summary of De Caelo see Drossaart-Lulofs 82–7 and comm. 152–65, and Moraux 475–6. Unfortunately his summary of Cael. i I–iiGoogle Scholar I is missing in the Syriac MS.

³⁴ Quis Heres 283; De Plantat. 3, cf. 12; Quaest. Gen. 3.6; 4.8; Quaest. Ex. 2.73. Philo himself seems to consider the question of the nature of the heavenly bodies insoluble (De Somm. 1.21–4). He uses both the Stoic view that the heavens consist of a special kind of fire (Quis Heres 133–6; Mos. 2.148) and the Peripatetic view that they consist of a fifth element. For full discussion see Drummond, J., Philo Judaeus: The Jewish and Alexandrian Philosophy in its Development and Completion (London 1888) i 273Google Scholar–9, and Moraux (n. 1) 1235–6.

³⁵ Moraux, P., Les listes anciennes des ouvrages d'Aristote (Louvain 1951)Google Scholar has shown that the catalog of Diog. Laert. v 22–7Google Scholar goes back to the Hellenistic period. He conjectures that it represents the holdings of the Peripatetic library about 200 B.C. Düring, I., ‘Ariston or Hermippus?’, Class, et Med. xvii (1956) 11–21Google Scholar; Biog. Trad. (n. 30) 67–9, 90–2 has pointed out weaknesses in Moraux's conjecture and argues again for the traditional ascription to Hermippus (third century B.C.). The question remains unsettled (cf. Moraux, , Aristotelismus [n. 30] 4, n. 2)Google Scholar.

³⁶ Cf. Moraux, , Listes (n. 35) 313–20Google Scholar; Düring (n. 35) 20–1; and Notes on the History of the Transmission of Aristotle's Writings, Acta U. Got. lvi (1950) Pt 3, pp. 35–70Google Scholar, esp. 57–70.

³⁷ Cf. Moreau (n. 30) 272–8, esp. 272; or Brink, K. O., ‘Peripatos’ RE Suppl. vii (1940) 931–8Google Scholar for a survey of the Peripatos in this period. The problem of the decline of the Peripatos is discussed by Lynch, J. P., Aristotle's School: A Study of a Creek Educational Institution (Berkeley 1972) 135–62Google Scholar.

³⁸ The case for the availability of Aristotle's treatises apart from the manuscripts from Skepsis is convincingly presented by Moraux, , Aristotelismus (n. 30) 3–44Google Scholar. The evidence for Cicero's access to the treatises is also collected and discussed by Moraux 33–41.

³⁹ Luck, G., Der Akademiker Antiochos (Bern/Stuttgart 1953) 36–40Google Scholar finds Antiochus behind Cicero's statements about the fifth nature as the substance of the soul; and Düring, , Notes (n. 36) 60Google Scholar suggests that Cicero may have learned about Aristotle in general from Posidonius and could have seen Aristotle's works in a library at Rhodes.

⁴⁰ See O. Gigon, ‘Cic. u. Arist.’ (n. 29); ‘Cic. u. gr. Phil.’ (n. 9) 240–50; and Moraux, , Aristotelismus (n. 30) 41–3Google Scholar.

⁴¹ Cf. Easterling (n. 21) 73–85. Cicero's statements that the human soul or mind consists of some fifth substance embody an idea also attributed to Critolaus, a Peripatetic of the second century B.C. (Aet. i 7.21; Tert. De An. 5.2; Macrobius, In Somn. Scip. i 14.20)Google Scholar. This fact only compounds the problem. If this doxography is reliable, Critolaus could be the source, directly or indirectly, of the misinterpretation of Aristotle (Cicero knew and approved Critolaus' views on the virtue of the soul [Tusc. v 51Google Scholar]). Or the doxographies could have confused Aristotle's view with that of Critolaus. Finally, the doxography may have misinterpreted Critolaus in the same way it did Aristotle.

⁴² Jaeger (n. 3) 143–6 cf. Chroust (n. 4) ii 186–7. Though Jaeger himself does not use this as proof or even confirmation for the presence of the fifth element in De Philosophia, his theory has become part of the overall reconstruction of the position of the fifth element in De Philosophia and must be dealt with in this context. On the possible source of the Stoic remodeling see Hahm, D., The Origins of Stoic Cosmology (Columbus, Ohio 1977) 144, 176Google Scholar n. 18, 267–73.

⁴³ Proof for divinity of celestial bodies: Cic., Nat.D. ii 42Google Scholar; Sext. Emp. Adv. Phys. i 49Google Scholar; cf. Aët. v 20.1. Proof for spirits: Philo De Somn. 1.135; De Gig. 2.7–8; De Plantat. 3.12; Apuleius De Deo Socr. 8.137; Plato, Epin. 984d–5bGoogle Scholar. Proof for eternity of cosmos: Philo Aet. Mund. 14.45. Hybrid proof: Sext. Emp. Adv. Phys. i 86Google Scholar. Cf. the discussion of Reinhardt, K., Kosmos und Sympathie (Munich 1926) 62–4Google Scholar.

⁴⁴ Cf. Reinhardt (n. 43) 62–86, esp. 62–8.

⁴⁵ For a full discussion of the attribution of these proofs to Aristotle see Effe (n. 2) 7–17.

⁴⁶ Simpl. In Cael. 289.1–15 (= DePhil.fr. 16 Ross). Simplicius' evidence must be used with caution because he did not have first-hand knowledge of De Philosophia. Cf. Cherniss, H., Aristotle's Criticism of Plato and The Academy (Baltimore 1944; New York 1962) 119 n. 7, 587Google Scholar; Gnomon xxxix (1959) 38–9Google Scholar; and Tarán, L., AJP lxxxvii (1966) 467Google Scholar.

⁴⁷ Effe (n. 2) 19–20 tries to escape this conclusion by suggesting that Philo's version might here be assimilated to the Platonic and Stoic view or else that Philo might have drawn from a section of Aristotle that focused on the human body and simply failed to mention the celestial fifth element. Effe would treat this fragment like Cic., Nat.D. ii 42Google Scholar ( = De Phil. 21a) an explain away its four-element cosmology.

⁴⁸ See above, n. 34.

⁴⁹ Cf. above, nn. 17, 18.

⁵⁰ On the problem of Stoic contamination and Ciceronian distortion in these passages see K. Reinhardt (n. 43) 61–92; Cherniss (n. 46) 592, 595–602; Moraux (n. 1) 1213, 1223–4; Chroust, A. H., ‘Some Comments on Cicero, De Natura Deorum II. 15.42–16.44: A Fragment of Aristotle's On Philosophy’, Class. Folia xxix (1975) 103–13Google Scholar; and Hahm (n. 42) 176, n. 18.

⁵¹ For some of the attempts to find an alternative ground for postulating the existence of a fifth element see Jaeger (n. 3) 139 and n. 1, cf. 143, 153; and Seeck, G. A., ‘Über die Elemente in der Kosmologie des Aristoteles’, Zetemata xxxiv (Munich 1964) 122Google Scholar.

⁵² This is done by Berti (n. 4) 368–70; Gräser, A., ‘Zu Aristoteles πϵρὶ ϕιλοσοϕίας (Cicero, Nat. deor. II 16 44)’, MusHelv xxvii (1970) 16–27Google Scholar; ‘Aristoteles' Schrift “Über die Philosophie” und die zweifache Bedeutung der “causa finalis”’, MusHelv xxix (1972) 44–61, esp 61Google Scholar; Effe (n. 2) 132–6, and Bos (n. 27) 48–9, 62–3, 99 and n. 34 (cf. 61, n. 90). Though each has a slightly different interpretation of Aristotle's view, all minimize the difference between the celestial mechanics of De Caelo and of Cic., Nat.D. ii 44Google Scholar. Proponents of this approach fail to notice that the closer the celestial mechanics of Cic. Nat.D. ii 44 approaches that of De Caelo, the less reason there is to attribute the reference to De Philosophia at a and the more likely it is to be a Stoic adaptation of De Caelo.

⁵³ The theory of natural movements is also found in the context of a four-element cosmology in Philo, , Aet. Mund. 28–34Google Scholar, a passage which has been attributed to De Philosophia on other grounds (= fr. 19b, Ross; cf. above, n. 45). If this reference is, in fact, drawn from De Philosophia, we have further confirmation that in De Philosophia Aristotle held a theory similar to the one, presented in De Caelo iii–iv.

⁵⁴ Presumably this voluntary movement of the heavenly bodies is analogous to the movement of human beings discussed in Mot. An. 6–7.700b4–70132. As men need sensation and intellect to perceive a goal and move toward it, so the heavenly bodies are endowed with sensation and intelligence (Cic., Nat.D. ii 42Google Scholar [ = De Phil fr. 21a]; cf. fr. 24), which they doubtless use for the same purposes. The difficult question whether the Prime Mover is involved cannot be discussed here; cf. Jaeger (n. 3) 140–5; von Arnim, H., ‘Die Entstehung der Gotteslehre des Aristoteles”, SBWien, phil.-hist. Kl. ccxii 5 (1931) 7–9Google Scholar ( = Hager, F. P., ed., Metaphysik und Theologie des Aristoteles, Wege der Forschung ccvi [Darmstadt 1969] 1.15)Google Scholar; Guthrie, W. K. C., ‘The Development of Aristotle's Theology’, CQ xxvii (1933) 162CrossRefGoogle Scholar–71 ( = Hager [above] 75–95); Aristotle: On the Heavens (Loeb 1939) xxv–xxviiGoogle Scholar; Ross, W. D., Aristotle's Physics (Oxford 1936) 95–6Google Scholar; Cherniss (n. 46) 591–602; Gräser, , ‘Aristoteles Schrift … “causa finalis”’ (n. 52) 44–61Google Scholar; Chroust, , Aristotle (n. 4) ii 180–4Google Scholar; and Easterling, H. J., ‘The Unmoved Mover in early Aristotle’, Phronesis xxi (1976) 252–65CrossRefGoogle Scholar.

⁵⁵ Cf., e.g., Solmsen (n. 10) 287–303, 397–8. The main point still at issue is whether De Caelo iii–iv represents a view different from, and therefore earlier than, the view of De Caelo i. Solmsen (n. 10) 293–303, and Seeck (n. 51) 97–8, 123–6, have argued convincingly that in De Caelo iii–iv Aristotle shows no knowledge of the fifth element and that these books, or at least the theories reflected in them, were originally conceived before De Caelo i. There is still, however, some support for the traditional view that De Caelo iii–iv were written at the same time as De Caelo i, but with attention focused so rigidly on the sub-lunar world that these books make statements that are misleading and appear to preclude the existence of the fifth element (cf. e.g., Moraux, , Du Ciel [n. 28] xxxviiiGoogle Scholar and n. 4, cxxxv and n. 3; and Bos [n. 28] 70). If the latter view is correct, the first phase will be reflected only in De Philosophia, rather than in both De Philosophia and De Caelo iii–iv; the ultimate reconstruction of the evolution of Aristotie's cosmological thought will be unaffected.

It should be noted that on the basis of a subtle and painstaking analysis of the various discussions of elements in Gen. Corr. and Cael. Seeck has questioned whether the evolutionary model is a complete and sufficient explanation for the great variety of theories of elements he claims to have discovered in these two works. His chief grounds for doubt are some subtle problems he sees in the relation between the fifth element and the various theories of elements that he has identified in De Caeli iii–iv. Nevertheless, Seeck has no doubt whatsoever that the fifth element is later than the theories of De Caelo iv (cf. esp. 97–8, 123–6, 157). Hence his concerns about the theoretical sufficiency of the evolutionary model need not trouble us here.

⁵⁶ See the attempts cited in nn. 51 and 52.

⁵⁷ For a discussion of the absolute date of De Philosophia see Chroust, A. H., ‘The Probable Date of Aristotle's Lost Dialogue On Philosophy’, JHistPhilos iv (1966) 283–91Google Scholar ( = Aristotle [n. 4] ii 145–58Google Scholar). Chroust concludes that it was most likely published before 347 B.C. He also conjectures it is later than the Protrepticus therefore after 350 B.C. Düring, I., Aristoteles: Darstellung und Interpretation seines Denkens (Heidelberg 1966) 49–50Google Scholar, on the other hand, prefers an earlier date, i.e. c. 360–55 B.C. Since the relative order of De Philosophia and other early works is still an open question, it is preferable to set Plato's Timaeus (i.e. c. 360 B.C.) as the terminus post quem, and consider any reinterpretation that brings De Philosophia closer to the Timaeus as support for an earlier, rather than a later, date within the period of c. 360 to 347 B.C.

⁵⁸ Some of the significant contributions to this controversy are: Jaeger (n. 3) passim; Wilpert, P., ‘Die Stellung der Schrift “Über die Philosophie” in der Gedankenentwicklung des Aristoteles’, JHS lxxvii (1957) 155–62CrossRefGoogle Scholar; DeVogel, C. J., ‘The Legend of the Platonizing Aristotle’, in Düring, and Owen, (n. 1) 248–52Google Scholar; and Düring, Aristoteles (n. 55) passim.

⁵⁹ I would like to thank Friedrich Solmsen, George Kerferd, and David Furley for reading and commenting on an earlier version of this paper. I have benefited greatly from their criticisms and encouragement.