¹ By ‘Prime Mover’ I mean the first of the 47 or 55 unmoved movers of Metaphysics Λ.8, the mover of the sphere of fixed stars. Of course, the presence of a multiplicity of unmoved movers in Aristotle's system affects how we view his theology. It need not, however, affect how we interpret his argument for the existence of God, since none of the premises in this argument requires that God be unique. Whether Aristotle's overall metaphysics can consistently admit a multiplicity of immaterial movers that are the same in kind but different in number is a different question, which has still to be satisfactorily answered. The best treatment of the problem remains that of Merlan, P., ‘Aristotle's unmoved movers’, Traditio 4 (1946), 1–30CrossRefGoogle Scholar. So far as I know, none of the many recent discussions of individual forms in Aristotle's metaphysics tackles directly the problem of multiple unmoved movers, though one would think that this is a natural testing ground for any interpretation of Aristotelian individual forms.

² Versions of what I am calling the conventional interpretation are advanced in Lloyd, G. E. R., Aristotle: The Growth and Structure of his Thought (Cambridge, 1968), pp. 144–5Google Scholar; Gerson, L. P., God and Greek Philosophy (London, 1990), pp. 125–6Google Scholar; implied by Elders, L., Aristotle's Theology: a Commentary on Book Λ of the Metaphysics (Assen, 1972), p. 187Google Scholar.

Ross, W. D., Aristotle's Metaphysics (Oxford, 1924), vol. i, p. cxliGoogle Scholar, offers an interpretation similar to the one I advance below in that he sees Aristotle inferring the PM's identification as God from its identity as νος. His précis has in common with the conventional interpretation that it sees Aristotle offering no argument for attributing thinking to the PM: ‘All physical activity being excluded by the immaterial nature of the first mover, Aristotle can only ascribe to it mental activity….’

³ Indeed, this seems to be Gerson's point when he criticizes Aristotle for offering no argument ‘for the claim that the actuality of the unmoved mover is that of a life, specifically the life of thought’ (op. cit., p. 126). As I shall try to show, however, this claim is precisely what Λ.7 attempts to argue for.

⁴ Similar characterizations of first philosophy: Metaphysics A.2, 982b7–10; A3, 983a24–b1; Γ.2, 1003b16–19; E.I, 1025b3; Z.1, 1028b2–7; K.S, 1065a23–4.

⁵ On the essential connection between time and motion, see Physics IV. 11, 218b21–219a10; VIII.1, 251b10–28; De Caelo 1.9, 279a14–16.

⁶ De Caelo I.9 gives an explanation in terms of natural place: κα ἄπαυστον δ κνησιν κινεῖται εὐλγως· πντα γρ παεται κινομενα ταν ἔλθῃ εἰς τν οἰκεîον τπον, το δ κκλῳ σώματος αὐτς τπος θεν ἤρξατο κα εἰς ὃν τελευτᾷ (279b1–3). When compared with Physics VIII and Metaphysics Λ, the De Caelo's natural place explanation invites developmental speculation, for it appeals neither to a heavenly soul nor to an external unmoved mover as cause of the heavens' motion. Guthrie, W. K. C., ‘The development of Aristotle's theology – I’, CQ 27 (1933), 162–71CrossRefGoogle Scholar, sees the De Caelo as representing a ‘materialist’ stage in Aristotle's thought, when the natural motion of aether was sufficient to explain the continuous circular motion of the heavens. According to Guthrie, the introduction of the unmoved mover in Physics VIII and Metaphysics Λ is meant to complement the earlier De Caelo account by providing a final/efficient cause of the heavenly motion. Waterlow, S., Nature, Change and Agency in Aristotle's Physics (Oxford, 1982), pp. 233–57CrossRefGoogle Scholar, contends on the other hand that the De Caelo account highlights a genuine difficulty for Physics VIII: in order to make plausible the idea that the heavens are moved by something (viz. an unmoved mover), Aristotle must maintain without empirical justification that the heavenly sphere is a self-mover analogous to animal self-movers.

⁷ I assume here that Aristotle accepts the principle of plenitude for eternal objects. In De Caelo 1.12 he advances detailed arguments for the position that a perishable entity cannot exist forever (see esp. 283a25–9). For discussion see Sorabji, R., Necessity, Cause, and Blame (Ithaca, 1980), c.8Google Scholar, and Judson, L., ‘Eternity and necessity in De Caelo 1.12’, Oxford Studies in Ancient Philosophy 1 (1983), 217–55.Google Scholar

⁸ Aristotle's reference to a number of such substances shows clearly that he has in mind here a plurality of unmoved movers, though he is not necessarily thinking of the specific account found in chapter eight (pace Elders, op. cit., pp. 145–6). For two different accounts of how there can be plural unmoved movers without a material principle to individuate them, see Merlan, , Traditio 4 (1946), 1–30CrossRefGoogle Scholar, and Owens, J. ‘The reality of the Aristotelian separate movers’, Review of Metaphysics 3 (1950), 319–37, at 330–4.Google Scholar

⁹ Following Ross' (ad loc.) construal of the meaning of these lines I have ignored the second κα in my translation (treating μσον as a predicate of τ κινομενον κα κινοôν), and printed ἔστι τονυν rather than τονυν ἔστι. These are minimal changes to the text of the MSS. and express clearly what must be Aristotle's thought here, that (in Ross' words, op. cit., vol. ii, p. 374) ‘a κινομενον κα κινοôν is something intermediate, which presupposes τι οὐ κινομενον κινει.’

¹⁰ Indeed, the argument for the necessity of at least one eternal unmoved mover in Physics VIII.6 proceeds from assuming an unbroken succession of finite changes. See the discussion in Waterlow, op. cit. [n. 6], pp. 223–5.

¹¹ According to Sauve, S., ‘Unmoved movers, form, and matter’, Philosophical Topics 15 (1987), 171–96CrossRefGoogle Scholar, no explanation in terms of moved movers alone will provide an intrinsic, and hence non-accidental, cause of motion. We thus have one kind of straightforward answer why Aristotle demands an unmoved mover of the cosmos: the eternal motion of the heavens would otherwise be an accident. However, Sauve's main contention – that forms are unmoved movers whose exercise supervenes on the exercise of material moved movers – does not apply to the PM, which is the heavenly sphere's object of desire, not its form.

¹² Thus, while the faculty of desire (τ ρεκτικν) is what moves the animal, its object of desire (ρεκτν) is properly speaking the unmoved mover, and this will be true even if nothing external to the animal corresponds to what it desires. See De Anima III. 10, esp. 433b10–18.

¹³ The emphasis here is on ‘of explanation,’ since, as I have just noted, Aristotle must accept the fact of an infinite series of moved movers.

¹⁴ Solmsen, F., Aristotle's System of the Physical World (Ithaca, 1971), p. 178Google Scholar, claims that in Physics VIII this point is motivated by Aristotle's desire to rule out the possibility of Plato's self-moving soul being the first principle of the cosmos. Judson, L., ‘Heavenly motion and the unmoved mover’, in Gill, M. L. and Lennox, J. (edd.), Self-motion from Aristotle to Newton (Princeton, 1994)Google Scholar argues persuasively that in Physics VIII and Metaphysics Λ the heavenly soul's susceptibility to incidental motion rules out its being the Clark, PM. S., Aristotle's Man (Oxford, 1975), p. 180CrossRefGoogle Scholar, sees a difference between the PM and the subordinate unmoved movers on this count. He contrasts Λ.8, 1073a23–5 (the PM κνητον κα καθ᾿ αὑτ κα κατ συμβεβηκς) with 1073a33–4 (the subordinate movers are unmoved καθ᾿ αὑτ), and adduces Physics 259b21ff., which attributes accidental motion to ‘certain principles (ρχα) of the heavenly bodies, as many as experience more than one motion’ (b30–1). Clark's distinction will depend on the legitimacy of identifying these principles with the subordinate unmoved movers of Metaphysics Λ.8.

¹⁵ I am deliberately avoiding the difficulties presented by De Anima I.3–4, which denies that the soul is moved, but rather makes the soul the source of movement and change in the body. Furley, D., ‘Self-movers’, in Rorty, A. O. (ed.), Essays on Aristotle's Ethics (Berkeley and Los Angeles, 1980), pp. 55–67Google Scholar [originally published in Lloyd, G. E. R. and Owen, G. E. L. (edd.), Aristotle on Mind and the Senses (Cambridge, 1978), pp. 165–79]Google Scholar, argues that because its object of desire is intentional, the animal may be regarded as a self-mover, although, since there must actually be an external object, ‘the movement of an animal does not provide an example of a totally autonomous beginning of motion’ (65).

¹⁶ I here follow the convention of using ‘intentional existence’ to refer to Franz Brentano's conception of ‘intentional inexistence,’ according to which mental objects do not have to exist outside the mind, and which he took to be the distinguishing mark between mental and physical phenomena: ‘Every mental phenomenon is characterized by what the Scholastics of the Middle Ages called the intentional (or mental) inexistence of an object, and what we might call, though not wholly unambiguously, reference to a content, direction to an object (which is not to be understood here as meaning a thing), or immanent objectivity. ⃜This intentional inexistence is characteristic of mental phenomena’ (Psychology from an Empirical Standpoint, trans. Rancurello, A., Terrell, D., AND McAlister, I., [New York, 1973’, pp. 88–9Google Scholar). For a very interesting discussion of how ancient interpretations of Aristotle set the stage for the medieval and Brentanian conceptions of intentionality, see Sorabji, R., ‘From Aristotle to Brentano: the development of the concept of intentionality’, in Blumenthal, H. and Robinson, H. (edd.), Aristotle and the Later Tradition (Oxford, 1991), pp. 227–59.Google Scholar

¹⁷ Furley, op. cit. [n. 15], p. 67, makes a similar point: ‘What about delusions, hallucinations, etc.? Aristotle could reply that although animals may on occasion move in pursuit of a purely imaginary goal, these cases are parasitic on genuine cases. They would not pursue the imaginary goal unless there were similar goals in reality.’

¹⁸ Λ.7, by arguing for the actual existence of a transcendent PM, removes this difficulty from occurring at the level of the heavenly soul. The conception of the PM as νος, however, will invite its recurrence at the level of the PM itself. If the divine νος is a faculty like other νοο, then either its object of thought is the cause of its actualization, in which case it is not the first cause, or it produces its object of thought from itself, in which case it is thinking a fantasy. This dilemma is faced in Λ.9, and resolved with the characterization of the PM as νησις νοσεως νησις, which finally eliminates potentiality from the being of the PM, and thereby breaks down the distinction between its essence and that of its thought object. I discuss the argument of Λ.9 in greater detail in ‘Nησις νοσεως in Metaphysics Λ.9’ (unpublished).

¹⁹ As we shall see, his ultimate purpose requires it too. Identifying the PM as an object of love provides the requisite form of final causality for explaining it as an unmoved mover. But to show that this unmoved mover is God, Aristotle must be able to demonstrate that it is νος, which cannot be done simply from its status as ρεκτν. Aristotle's strategy is to identify the first objects of thought and desire so as to show that if the PM is the first object of love, it is necessarily the first object of thought; as I shall argue, from the PM as νοητν, he can get to the PM as νος.

²⁰ Cf. De Motu Animalium 6, 700b23–5: ὥστε κινει πρωτον τ ρεκτν κα διανοητν· οὐ πν δ τ διανοητν, λλ τ των πρακτν τλος.

²¹ Aristotle sometimes divides ρεξις into three species: πιθυμία, θυμς, and βολησις (De Anima 414b2, De Motu Animalium 700b22, Eudemian Ethics 1223a26–7). Elsewhere, as here, he is content to divide desire simply into rational (βολησις) and irrational (πιθυμα and θυμς) kinds (De Anima 432b5–6, 433a22–30). See Nussbaum, M., Aristotle's De Motu Animalium (Princeton, 1978), pp. 334–6.Google Scholar

²² The question naturally arises, why does the first heaven rotate if the PM is its goal? The PM not only does not rotate, it does not move. We should recognize first that the heaven's rotation is not the PM's only direct effect. It causes itself primarily to be thought and desired, and this effect is more immediate than (though not temporally prior to) the heavenly rotation. The first heaven's contemplation of the PM is thus part of the way in which it maintains a state of being which is like the PM's state of being. Indeed, from the perspective of the heavenly sphere itself, this is the most important way, since it is by contemplating that it emulates the internal nature of the PM. But since the first heaven is also a spherical body, its rotation causes its material capacity – the capacity for circular motion – to be actualized as well. If the sphere did not rotate, it would fall short of the PM as τλος, for at least one of its capacities would remain unrealized.

²³ For the συστοιχα see also Metaphysics A.5, 983a23–6; Γ.2, 1004b27–1005a5; Physics I.5, 189a1; III.2, 201b25–6 ( = Metaphysics K.9, 1066a15).

²⁴ Lindbeck, G., ‘A note on Aristotle's discussion of God and the world’, Review of Metaphysics 2 (1948), 99–106Google Scholar, denies substantiality to God on the grounds that his ‘actual existence cannot be inferred from the desire which the world has towards him’ (104). Aristotle, however, does not infer God's substantiality from his desirability, but from his intelligibility. Owens, , Review of Metaphysics 3 (1950)Google Scholar, observes brilliantly that the actuality of the first νοητν would ‘be immediately identified with the substance held to be producing it’, and hence ‘would have to be a thinking of itself, and so would be substance in its own right, with no dependence whatsoever on the celestial soul’ (329). Owens also denies that this thought can be thought by anything else, presumably because this would attribute to God the potential of being thought by something else. This worry is a red herring, however, for such a potentiality would be harmless to the PM's actuality as a substance and, therefore, to its role as a cause. There is no reason why the celestial soul, e.g., cannot think it, so long as the thought produced in the celestial soul is not itself the PM.

²⁵ Ryan, E., ‘Pure form in Aristotle’, Phronesis 18 (1973), 209–24CrossRefGoogle Scholar, argues that there are no pure forms – including the unmoved movers – in Aristotle. As it applies to the PM, his argument involves denying the equivalence, which I am assuming, between νοητν and εἶδος. Ryan claims, contra Ross (op. cit., vol. ii, p. 380), that Aristotle's statement at 1072b21, that νος is τ δεκτικν τοô νοητο κα τς οὐσας, is not ‘parallel’ to his statement at De Anima 429a15–16, that νος is δεκτικν τοô εἴδους κα δυνμει τοιοôτον λλ μ τοôτο. But his reason for denying a parallelism between the two passages – viz. that the De Anima passage is about human thinking – is insufficient, for there is nothing in Metaphysics 1072b18–21 that does not apply to human thinking. In fact, the description there of νος becoming intelligible by ‘touching and thinking’ does not strictly apply to the PM, which cannot become anything at all. As we shall see, Aristotle is using his theory of human νος in A.7 to argue that the νοητν that moves the heavens is νος in the first place. Λ.9's identification of the PM as νησις νοσεως νησις will then move beyond the theory of human νος by removing from the PM all implied potentiality. So, far from the De Anima description not fitting into Λ.7 or 9, the Λ.7 argument will not work without it. When we consider Aristotle's unconscious assumption of an equivalency between νοητν and εἶδος in De Anima III.4 (429a15, a28, a29), there is no way of understanding τοô νοητοô κα τς οὐσας at 1072b21 except as equivalent to τοô εἴδους.

²⁶ Partisans of a species-form interpretation of primary substance in Metaphysics Z will no doubt object that the species-form itself has such a way of existing. While my own sympathies lie much more with those who defend particular forms, my interpretation here depends on neither view. For even if one believes that Aristotle accepts such things as species-forms and identifies them as the primary substances, there is nothing to suggest that he would plausibly consider the PM to be a species-form.

²⁷ Thus the PM as intelligible object is numerically identical with the νος thinking it. It will have to be essentially identical as well to satisfy the argument of chapter 6, but this issue is not settled until chapter 9, with the definition of the PM as a pure actuality (cf. note 18 above). At this point in Λ.7 Aristotle is using the De Anima's conception of self-thinking νος to establish that the PM is a thinker as well as a thought. See Sorabji, R., Time, Creation and the Continuum (London, 1983), pp. 144–5Google Scholar, for a discussion of the numerical identity of the intellect and its object.

Thinking in itself is of what is best in itself, and the purest thinking is of the purest good. And νος thinks itself through participation in the intelligible. For it becomes intelligible by touching and thinking, so that νος and its object of thought are the same. (1072b18–21)

²⁸ …λλ᾿ αὐτῇ καθ᾿ αὑτν εἰλικρινεῖ τῇ διανοᾳ χρώμενος αὐτ καθ᾿ αὐτ εἰλικρινς ἔκαστον πιχειροι θηρεειν των ντων, παλλαγες τι μλιστα ϕθαλμων τε κα των κα ὡς ἔπος εἰπεῖν σμπαντος τοô σώματος, ὡς ταρττοντος κα οὐκ ντος τν ψυχν κτσασθαι λθειν τε κα ϕρνησιν ταν κοινωνῇ (Phaedo 66a1–6).

²⁹ Cp. De Anima II.2, 414a10, and the senses as receptacles: 11.12, 424a18; III.1, 425b23.

³⁰ Aristotle was certainly influenced on this point by Plato's arguments in the Timaeus for the featurelessness of the ὑποδοχ, which would otherwise obscure the sensibles coming to be in it (50d2–e4, cf. 51a6–b2). For Aristotle the intellect can only be forms. But insofar as the actuality of anything in the world (even down to the four elements) is expressed by the account of its form, the intellect can think the essence of anything. Thus a stone cannot be in the soul, but its form can (431b29). Cf. Alexander, , De Anima, 84.18–21.Google Scholar

³¹ See Wedin, M., Mind and Imagination in Aristotle (New Haven, 1988), c. 5, esp. p. 166CrossRefGoogle Scholar, for a very clear discussion of what Aristotle means by saying that νος is nothing actual before it thinks. Modrak, D., ‘The nous-body problem in Aristotle’, Review of Metaphysics 44 (1991), 755–74Google Scholar, discusses the general problem of νος' separateness from matter.

³² This is a consequence of his theory of the sense as a δναμις and a λγος of the sense-object. Sensory excesses destroy the proportion which is needed for the sense to function (II.12, 424a25–32; III.2, 426a27–b7). Cf. 422a21–3, 31–3; 424a14–15.

³³ On retaining the manuscript reading (δ αὑτν) of 429b9, and for a thorough history of the editing of this line, see Owens, J., ‘A note on Aristotle, De Anima III.4, 429b9’ in Aristotle: The Collected Papers of Joseph Owens, ed. Catan, J. R., (Albany, 1981), pp. 99–109Google Scholar (originally published in Phoenix 30 [1976], 107–118).Google Scholar

³⁴ The details of how this is supposed to work are controversial. From the De Memoria it is clear that human habitual knowledge is mediated by images (ϕαντσματα), which are the proper objects of memory (450a24; cf. 450a14). Thinking, in turn, requires images, within which the objects of thought are somehow located (De Anima 431a16–17, 431b2, 432a12–14; De Memoria 449b31). Since memory is associated with perception and imagination rather than with the intellect (449b30–450a25), the memory of scientific knowledge must be preserved through the images produced by the faculty of imagination. For discussion of how the De Memoria may shed light on Aristotle's theory of thinking, see Sorabji, R., Aristotle on Memory (London, 1972), pp. 2–8Google Scholar, and Wedin, op. cit. [n. 31], pp. 136–41.

³⁵ This interpretation of νος' self-thinking originates with Alexander. See De Anima 86.5–29, and also the section on νος in the work Bruns published as De Anima Libri Mantissa, 108.2–9, 109.4–23.

³⁶ ‘Aristotle's philosopher-god’, Phronesis 14 (1969), 63–74Google Scholar. This is not to say that the self-thinking νος is not aware of itself, which is entirely another issue. This means only that, on Aristotle's theory, νος’ self-thinking is not its self-consciousness. Indeed, nowhere in the technical discussions of νος does he ask the question, ‘how do we know that we are thinking?’ (He does, of course, ask the equivalent question for the senses: De Anima III.2, 425b12–15, with which compare De Somno 455a15–20.) Thus, I see no textual evidence for Oehler's, K. claim, ‘Aristotle on self-knowledge’, Proceedings of the American Philosophical Society, 118 (1974), 493–506Google Scholar, that, ‘Mind thinks itself not because it thinks the forms which it has become but insofar as it is conscious of its object, i.e. of the forms. Self-consciousness is consciousness being conscious of itself by means of the consciousness of its object’ (498). Indeed, Wedin, op. cit. [n. 31], pp. 265–8, offers an interesting argument to the effect that the PM is not capable of thinking first person propositions and hence is not capable of self-consciousness at all.

³⁷ At this point, one might object that if this is what Aristotle has in mind for God's self-thought, then why would the apparent problem of Λ.9 arise at all, i.e. why would he have to argue that God thinks of himself rather than something else? I address the details of Λ.9 in a paper currently in progress, but the short answer is that he does not argue there that God thinks of himself, if by this is meant anything other than the De Anima notion of self-thinking. Λ.9 seeks to examine aporiae about the νος-God of Λ.7 that arise from the conception of νος itself; specifically, the problematic notion is νος’ nature as a faculty, which may be exercised or not. The different candidates for νος' thought-object proposed at 1074b22–3 (itself, one other thing always, different things at different times) are possible answers to the question τ νοει; only as it is posed at 1074b21, independently of deciding whether the nature of God is νος ( = δναμις) or νησις ( = νργεια). 1074b28–33 then outlines the undesirable consequences that follow if we identify the nature of God as νος ( = δναμις), and the conclusion at 1074b33–4 redefines it as νησις νοσεως – the actuality of an actuality, i.e. a pure actuality – in order to avoid these consequences. God's self-thought is a corollary of this redefinition, but it is not to be understood in any terms other than those available from De Anima (pace Modrak, , Review of Metaphysics 44 [1991], 772).Google Scholar

³⁸ Owens, J., The Doctrine of Being in the Aristotelian Metaphysics, 3rd ed. (Toronto, 1978), pp. 457–60Google Scholar, discusses the wider implications of this feature of Aristotle's metaphysical thought.

³⁹ This statement of course applies only to an individual νος' relationship to itself, νος as a natural entity is an actual form and can be studied as part of the overall project of studying the soul, as Aristotle himself does in the De Anima. Clearly, though, the kind of understanding one possesses after having studied the intellect in the manner of De Anima III is no more to be described as self-knowledge or -awareness than the knowledge of other functions of the soul.

⁴⁰ Indeed, it is sometimes raised as a criticism of Aristotle's conception of God that His thought is contentless. Consider the eloquent statement of Fuller, B., ‘The theory of God in Book XII of the Metaphysics’, Philosophical Review 16(1907), 170–83, at 176–7CrossRefGoogle Scholar: ‘God's essence, we are told, is thought of thought. But thought of thought of what? By draining thought and form of its filling of sensible reference, we have apparently deprived it of all that gives it value and relevance. It is reduced to mere reflection upon itself, with no other self than the barren act of reflection to reflect upon…. It is like consciousness without anything but its mere name tobe conscious of, and therefore meaningless.’

⁴¹ Cf. Metaphysics Θ.8, 1049b18–29, and De Anima III.4, 429a13–15. See also the passages cited in note 29 above.

⁴² An earlier version of this paper was delivered in October 1990 at the New York meeting of the Society for Ancient Greek Philosophy; I am grateful for questions and comments I received on that occasion. I would also like to thank Michael Frede, David Furley, Lindsay Judson and Steven Strange for their generous and helpful criticisms.