Alexander of Aphrodisias on Celestial Motions

ISTVçN M. BODNçR

A number of features of the doctrine of Alexander of Aphrodisias on heav-

enly motions are beyond reasonable doubt. First and foremost of these is
that he identiŽ ed the nature of the heavenly spheres with their soul, 1
thereby he could entirely collapse natural motion with voluntary motion
into one in their case.2 Moreover the celestial element, which Alexander
tends to call yeÝon sÇma, divine body is removed from the components of
the everchanging sublunary world to the extent that it can be a legitimate
question whether the substrate of celestial bodies can be called matter,3

Accepted August 1996

1
For this Alexander had to criticise and reject proposals of his predecessors. See
Simplicius in De caelo 380.5ff. and 29ff.
The characterisation of the Alexandrean proposal on Merlan p. 181, that Òwe could
interpret Alexander as saying that the celestial motion is neither purely corporeal (ethe-
real) nor caused exclusively by the soul, but rather a combination of the two, [. . .]Ó
is unlikely to be correct, as this description would equally be true of SimpliciusÕ (or –
as Merlan notes – of JulianusÕ and of HerminusÕ) celestial theories as well, who did
not equate celestial nature with celestial soul. A more plausible claim would be that
Alexander as it were rephrased some motifs of HerminusÕ solution at a higher register:
whereas Herminus claims that celestial soul is responsible for the inŽ nity of celestial
motions, presumably allowing for some determination of the motion by the material
component of the spheres, Alexander will attribute the eternity of the motion to the
external Ž rst mover, and assign all the other determinations to the nature/soul of these
spheres.
2
Living beings move by nature on AristotleÕs account, too. Nevertheless, the nat-
ural motion of the animal is not natural to all (nor even to any of ) its bodily parts.
On account of their material composition the bodily parts possess separate natural
motions which can be different or can even run in the face of the natural motion of
the animal (Physics 8.4 254b14-20, cf. De philosophia fr. 19b Ross (part of 916
Gigon) = Philo De aeternitate mundi 6.28-7.34). On AlexanderÕs identiŽ cation of the
nature of the celestial spheres with their souls these natural motions are not subordi-
nated the one to the other in the case of the heavens, as in the case of other living
beings, but they are just identical.
3
Quaestio 1.10 argues for both horns of the dilemma: if matter is what is recep-
tive of opposites in turn, the divine bodies will lack a material principle. On the other
hand if matter is also deŽ ned as the ultimate inarticulate substrate, this way of speak-
ing about matter would include the substrate of the divine body as well. What
Alexander stresses here and also in quaestio 1.15 is that provided there is a special
celestial matter, there is no overarching concept of matter encompassing both domains

© Koninklijke Brill, Leiden, 1997 Phronesis XLII/2

 ALEXANDER OF APHRODISIAS ON CELESTIAL MOTIONS 191

and Alexander can refer to perishable entities as ¦nula, material in con-

trast to this sublime element.4 After identifying the contribution of the
nature of the celestial spheres with that of their soul, Alexander follows
Aristotle in setting out a celestial hierarchy, on top of which there is or
there are the separate unmoved mover(s), which move(s) by being ob-
ject(s) of striving and desire for the less perfect entities of the heavens.
This much seems to be Ž rmly settled. A number of further issues, how-
ever, call for detailed examination. In this paper Ž rst I set out to clarify
the contributions of the striving of the different celestial spheres, then I
turn to describing the interaction between the various motions of the celes-
tial system, and I discuss whether the theory Alexander propounded could
have been a fundamental revision, or rather an alternative exposition of
the original, Aristotelian celestial theory deploying homocentric spheres.

1. Planetary motions in quaestio 1.25

For a satisfactory interpretation of the interrelationship between celestial
motions the single most important task is to clear the apparently aberrant
evidence of quaestio 1.25. For this I will propose an alternative articula-
tion of the translation for a key passage of this text. As I hope to show
this alternative articulation draws with it an interpretation of the text
which is sufŽ ciently different to avoid the problems mentioned by Sharp-
les [1982] and [1992].
First I would like to set out the text as it is in the standard edition of
Bruns in the Supplementum Aristotelicum, with the two modiŽ cations pro-
posed in the Appendix of Sharples [1982], for which see nn. 8 and 12,
moreover I introduce a full stop in l.27, instead of the comma of previ-
ous editors:
40.23: pleiñ-
nvn d¢ sfairÇn oésÇn tÇn toè yeÛou sÅmatow ² m¢n prÅth te kaÜ ¤jv-
25 t‹tv pl°n te kaÜ mÛan kineÝtai kÛnhsin ¤keÛnhw ¤f¡sei t°w oésÛaw, aß d¢
metŒ taèta ¥ptŒ kinoèntai m¢n kaÜ toætvn ¥k‹sth ¤f¡sei te kaÜ ôr¡jei
tinòw oésÛaw, õpoÛaw kaÜ ² prò aétÇn. oé m¯n mñnhn t®nde kinoèntai, Žllƒ
¤j ¥autÇn kineÝtai ¥k‹sth, ¶n kineÝtai kaÜ perif¡retai Žn‹palin ¤keÛnú
tÒ t¯n y¡sin te kaÜ t‹jin toiaæthn ¦xein, kineÝtai d¢ kaÜ deut¡ran kÛnhsin
30 êpò t°w prÅthw periferom¡nh t¯n aét¯n ¤keÛnú.

of corporeal existents, and so celestial and sublunar matter differ without being com-
posed of a putative more basic entity plus some speciŽ c difference.
4
See De mixtione 229.6-9, where the context suggests that even if the divine body is
not among material entities (¦nula), it may possess some matter different from theirs.
192 ISTVçN M. BODNçR

24 d¢] te a 25 kÛnhsin ¤keÛnhw] n ¤k in lit. V 26 ¥k‹sth GFSBLa Sp.: ¥k‹sthi

V 27 tinòw libri, Zeller (p. 827, n. 5), Sharples [1982] et [1992]: t°w Bruns [1890]
(p. 226) et Bruns mñnhn t®nde] mñnú t»de Bruns [1890], nd in lit. V 28.29 ¤keÛnú
tÒ t¯n GFSLa Sp.: ¤keÛnh t» tÇn V1: ¤keÛnh t» tÇ V 2: ¤keÛnú tÒ B 30 ¤keÛnh V 1
(corr. V2)

As I think there is no need for further emendation and that a slightly modiŽ ed
version of SharplesÕ [1982] translation is perfectly suitable for an inter-
pretation of this text, here I quote this translation, 5 inserting a slightly dif-
ferent set of markers in the text from the ones Sharples [1982] employed: 6
(0) {There are several spheres of the divine body, and} the Ž rst and outermost
[sphere] is moved in a simple and single motion by desire for that being (i.e. the
Unmoved Mover). But the seven after it7 {also} are (1), on the one hand, each
of them moved by desire and appetite for some being8 of the sort by which that
before them is {also} moved. (2) However, they are not only moved with this
motion, (3) but [they are moved] of themselves, each of them, (4) with {the}
motion in which they are moved and carried around in the opposite direction to
that [outermost sphere], through having the sort of position and arrangement that
they do. (5) But [each of them] is also moved in a second motion by the Ž rst
[sphere], being carried round with the same motion as it. (Sharples [1982]
pp. 209f.)

The problem with this text can be summarised as follows: ÒIt seems clear
that (ii) [= (2) + (3)] and (iii) [= (4)] relate to the independent motions of
the inferior spheres from west to east; in which case both (iv) [= (5)] and
(i) [= (1)] must relate to the daily motion of the whole heaven from east
to west. It is, admittedly, most odd, that the Unmoved Movers are men-
tioned in (i) [= (1)], in connection with the daily motion shared by all the
spheres, rather than in (ii) [= (2) + (3)] in connection with the peculiar

5
I add some phrases in curled brackets from Sharples [1992], and change the inde-
Ž nite article in (4) to Òthe.Ó
6
Where my (1) = SharplesÕ (i), (2) + (3) = (ii), (4) = (iii) and (5) = (iv).
In fact I could have done with markers running up to (4) only, as (3) + (4) form an
inseparable grammatical unit. I chose to further subdivide SharplesÕ markers as instead
of shifting them, so that I can easily refer to SharplesÕ subdivisions of the text as well.
7
This is the only point where one might feel tempted to tamper with the text, as
both the preceding sphere and the Ž rst substance were feminine, and so we should
expect metŒ taæthn or metŒ taætaw. One could readily conjure up different lines of
excuses for this neuter pronoun, moreover such slight inconcinnities are by no means
alien to the Alexandrean writings, so the desire to emend should be restrained.
8
Here t°w seems to be an unwarranted conjecture of BrunsÕ. Although, as I intend
to argue elsewhere, all the spheres move on account of their desire for the same super-
cosmic entity, there is no need for Alexander to enunciate this doctrine abruptly here,
and the õpoÛaw of the next clause goes better with the tinñw of the manuscripts and
editions than with the t°w of Bruns.
 ALEXANDER OF APHRODISIAS ON CELESTIAL MOTIONS 193

motions of the lower spheres.Ó (Sharples [1982], p. 210) In fact if this

construal of the text were correct we would have three, as instead of two
moments in the motions of the planetary spheres. 9 Their diurnal motion,
according to (5) would be caused by the Ž rst sphere, then (1) mentions a
motion which the spheres perform Òby desire and appetite for some being
of the sort by which that before them is moved,Ó and then, contrasted to
(1) the motion described in the continuous stretch of text (2) - (4), i.e. a
motion which is contrary to the diurnal motion of the heavens, and which
is performed by the planetary spheres Òof themselves, each of them.Ó Even
if we were willing to take the odd, and un-Peripatetic notion that the diur-
nal motion of the planetary spheres is due to their desire for some emi-
nent entities, the contrast drawn in (2) and (3) seems to be ill-placed.
There, on this construal, Alexander would place the proper motion of these
spheres in a curiously strange category: they would be motions the spheres
perform on their own, and as the contrast seems to demand it, presum-
ably without a desire for a higher entity. 10 The fact that we encounter prob-
lems with both kinds of motions warrants that something has to be done
with the text.
One remedy is to resort to an emendation. Sharples [1992], p. 85, n. 262
proposes that some reshuf ing of the text will do the job. Take Òby desire
and appetite for some being of the sort by which that before them is
movedÓ from ll. 26f., supply something like <by the sphere before them>
or even <by the Ž rst sphere>11 instead, and attach the removed clause to

9
For reasons of convenience I will not speak about the Òspheres of the planets,
and the Sun and the MoonÓ but will use Òplanetary spheresÓ as a shorthand form
instead.
10
Which is somewhat more than SharplesÕ [1982] complaint: ÒOn any interpreta-
tion the passage from quaestio 1.25 seems to leave the proper motions of the lower
spheres oddly unexplained; [. . .]Ó (p. 210). A celestial system where diurnal motions
and proper motions alike are caused by desire for presumably different higher entities
would contain no contradiction. This construal, however, does not allow readily for
such a scheme.
11
Cf. l. 27 or l. 30 for these phrases. Note that this phrasing comes perilously close
to postulating action at a distance, which was on the whole avoided by Peripatetic phi-
losophers. On the other hand the phrase Sharples suggests, that each sphere should be
moved <by the sphere outside it> has its difŽ culties too, as it is far from trivial how,
in a system in which there are only seven spheres, and hence no counteracting ones,
a planetary sphere could transmit only the diurnal component of its motion to its neigh-
bour below.
Even though I argue that these emendations are super uous, the problem does not
vanish by merely rejecting them. AlexanderÕs text at l. 30 already indicates that the
diurnal motion of every planetary sphere is caused by the Ž rst sphere, and this phras-
ing as I have signalled almost inevitably implies the forbidden action at a distance. I
194 ISTVçN M. BODNçR

the end of (3).12 This emendation certainly will bring quaestio 1.25 in line
with Peripatetic practices. After rewriting (1) and (5) will relate about the
diurnal motion of the planetary spheres, somehow caused by the Ž rst
celestial sphere, whereas in the chunk of text (2) - (4) we will have the
proper motions of the planetary spheres, moving on their own accord (2),
by desire and appetite for some higher entity (the transposed clause of (1))
in a contrary direction to the motion of the Ž rst sphere (4).
My contention, however, is that much the same understanding can be
elicited from SharplesÕ [1982] translation, and on this account the need
for an emendation can be avoided. Besides the general principle of parsi-
mony with emendations, in the run of the argument I will mention two
further hints in favour of my interpretation.
Perhaps the best way to delineate my understanding of the text is to
give yet another paraphrase of it.
The Ž rst heavens we learn move with a single and simple motion, as
a result of the desire for a higher substance (0). Then in the next Ž ve
clauses motions of the planetary spheres are mentioned: (1) The seven
spheres also have a motion originating from desire for a substance like
the one strived for by the sphere before them. (2) Nevertheless this is not
the only movement they perform. (3) We are next told that each of them
moves on its own account. Then (4) and (5) again speak about two mo-
tions: (4) one is performed contrariwise to some earlier mentioned motion
and is determined by the position and order of these spheres, (5) while
the other (or literally: the second) is the same as that motion, and it is
performed by being carried around by the Ž rst (sphere or motion). As (5)
uses as a frame of reference the motion of the Ž rst celestial sphere men-
tioned in (0), the motions of (4) should run contrariwise to this motion.
This much is common ground. The crucial further consideration here is
how the relationship between (2) and (3) is construed. (2), by claiming

shall return to the discussion of the transmission of motions among the heavenly
spheres in section 2 below: there I argue that Alexander had to hold a celestial the-
ory which satisŽ es both requirements: that the diurnal motion of every sphere is caused
by the sphere of the Ž xed stars, and that only adjacent entities can be directly causally
related, see esp. pp. 200-201.
12
In fact this change necessitates some reworking on (4) as well, as by the inser-
tion of a considerable chunk of text the relative pronoun ´n at the beginning of (4)
gets separated from (3) by a further subordinate clause. Presumably this is why
Sharples [1992], p. 85 returns to BrunsÕ interpunction, and has a full stop after (3)
(rejected by Sharples [1982], p. 210, n. 94), and so the resulting translation (ÒHowever,
they do not only move with this [movement], but each is also moved of its own accord.
It is moved and carried round in this [movement] etc.Ó) keeps an open slot for the
transposition of the clause.
 ALEXANDER OF APHRODISIAS ON CELESTIAL MOTIONS 195

that the motion mentioned in (1) is not the only motion of the planetary
spheres speaks in fact about two motions. A clause coming after (2) can
either switch over to the second motion introduced by (2), or after men-
tioning the second motion of the planetary spheres can return for further
elucidation to the proper motion of the spheres mentioned already in (1).
As my articulation of the translation already suggests I would opt for the
second possibility. The further elaboration on (1), which comes in (3) + (4)
adds that the proper motion is on the one hand elicited by desire for a
higher entity (1), this is nevertheless the motion which the spheres per-
form of themselves (3) and in fact in an opposite direction to the diurnal
motion of the heavens (4). Then, after Ž nishing the further speciŽ cation
of the proper motion of the planetary spheres in (3) + (4), Alexander turns
in (5) to the characterisation of the common, diurnal motion of these
spheres, which was Ž rst mentioned in (2). The two features (5) mentions
characterise this motion in contradistinction to what has been said about
the proper motion of the planetary spheres in (3) + (4): this second motion
is imparted to the spheres by the Ž rst sphere, i.e. this motion is not per-
formed by the spheres on their own (cf. 3), and that it is carried around
with the same motion as the Ž rst sphere, contrariwise to the proper mo-
tions, which are performed in the opposite direction (4).
Apart from the internal coherence and the overall doctrinal soundness
of the text there are two further indications which favour an interpretation
on the above lines. One of them is that if there were a switch from one
motion to another between (1) and (3), the text should say something like
this: the planetary spheres perform not only the motion which is imparted
to them on account of their desire for higher being(s), but they also (ŽllŒ
kaÛ) perform this or that revolution. The text, however, says instead: oé
m¯n mñnhn t®nde kinoèntai, Žllƒ ¤j ¥autÇn kineÝtai ¥k‹sth ¶n kineÝtai kaÜ
perif¡retai ktl. the fact that this ÒalsoÓ is missing here makes the con-
strual I proposed above possible and plausible. 13 On the other hand when
in (5) we read kineÝtai d¢ kaÜ deut¡ran kÛnhsin ktl. the kaÛ here can intro-
duce the other motion, which (2) mentioned in contrast to the proper
motion of the planetary spheres.

13
The translation of the passage in Sharples [1992]: ÒHowever, they do not only
move with this [movement], but each is also moved of its own accordÓ (p. 85, my
italics) supplies this missing, and indeed unnecessary Òalso,Ó which is absent in the
Greek text. Similarly the translation of this passage by Bruns [1890], p. 226 contains
this super uous ÒalsoÓ: Òindessen werden diese sieben SphŠren nicht nur durch dies
eine Streben bewegt, sondern auch aus sich heraus wird noch eine jede SphŠre [. . .]
bewegtÓ (my italics again).
196 ISTVçN M. BODNçR

The other indication in the text which suggests that it is only at (5) that
the other motion is taken up is that Alexander, contrary to an almost ter-
minological usage, labels the diurnal motion of the planetary spheres as
a second motion. 14 This is only tolerable if this section from (1) to (5),
contrary to standard use actually introduces the diurnal motion of the plan-
etary spheres after their proper motion. This consideration rules out any
rearrangement of the text which wants (2) and then (3) to switch over to
the proper motion of the planetary spheres after their diurnal one.15 As on
the transmitted text (1), (3) and (4) all rehearse stock attributes of the
proper motion of the planetary spheres, it is very well understandable that
(5), after the lead of (2) will relate about the diurnal revolution of these
spheres that they also perform this as a second motion.

2. Planetary spheres and homocentric models

The account we get here, that the planetary spheres perform their proper
motions on their own, by desire for some higher entity or entities and that
the diurnal motion of the heavens is transmitted to them in some yet un-
speciŽ ed manner is still a long way from a satisfactory celestial theory in
the Peripatetic vein; it is conspicuously different from the complex system
of nested deferent plus counteracting spheres propounded in Metaphysics
12.8. Nevertheless Alexander might have claimed that his way of talking
about the heavens is completely legitimate, as it arguably has parallels in
AristotleÕs description of celestial phenomena elsewhere.16 In fact Alex-
anderÕs talking of 8 spheres is safer as it is less theory dependent. The

14
Cf. the same quaestio a dozen lines later, when it refers to the proper motions
of the planetary spheres with these words: ¤peid¯ t°w toætou [sc. toè êpò sel®nhn
sÅmatow ] x‹rin tetagm¡nhw metabol°w kaÜ ŽidÛou katƒ eädow diamon°w keÝntai
kineÝsya i toçw kukloforhtiko è sÅmatow aß metŒ t¯n prÅthn te kaÜ Žplan°
kaloum¡nhn ¥ptŒ sfaÝrai t¯n kÛnhsin t¯n deut¡ran, ktl . (41.11-4).
It is important to notice, that Alexander at 40.29 says that the diurnal motion is a
second motion of the planetary spheres, and not that it is the second motion belong-
ing to them.
15
On SharplesÕ rearrangement for instance the text would mention the diurnal and
proper motions of the planetary spheres in this order: diurnal – proper – proper – diur-
nal, or in the standard terminology: Ž rst – second – second – Ž rst. As in this sequence
Alexander would have mentioned the diurnal motion of these spheres Ž rst, he could
have no plausible excuse to call at the end of the section this motion the second.
16
See most particularly De caelo 2.10 291a34-b6, where planetary motion is ana-
lyzed into the simple and Ž rst revolution and the motion along the proper circle of the
planet. Accordingly Alexander when commenting upon this passage (Simplicius in De
caelo 472.8ff.) resorts to a model along the lines of quaestio 1.25, deploying one sphere
 ALEXANDER OF APHRODISIAS ON CELESTIAL MOTIONS 197

very fact that upon enumerating his spheres Aristotle comes up ¤nnoÛaw
xarÛn with two possible candidates for the total count of spheres17 illus-
trates graphically the dangers inherent in the project of establishing the
Ž nal and exhaustive representation of the heavens. Moreover Alexander
just had to heed AristotleÕs words at 1073b16f. where Aristotle stresses
that in matters celestial one should trust those who investigate the Ž eld
with more exactitude, to be on the alert against uncritically following the
original Aristotelian sphairopoiia too closely. By AlexanderÕs time the
theory of homocentric spheres had become obsolete and each contem-
porary astronomical theory – even if it was Aristotelian in inspiration –
violated basic Aristotelian principles to a certain extent by introducing
eccentric and/or epicyclic motions. Alexander then understandably could
use a non-committal formulation, where he could remain silent or vague
on the minutiae of celestial phoronomics. 18

per planet (here: the sphere of Saturn), and the cause of the diurnal motion of such a
sphere is the revolution of the sphere of the Ž xed stars.
Another telling example is De caelo 2.12 292b1ff., where Aristotle speaks about the
(up to a point) increasing number of activities of stars, and elucidates it in a frame-
work which deploys some version of a homocentric theory. This dual aspect of
planetary phenomena – that they are the manifestations of a unitary living being,
possessing a soul and performing several motions, and at the same time as analyzed
into the orbital motions of different spheres, which are assigned a single revolution
each – rests ultimately on the account of the construction of the World Soul in the
Timaeus (35B-36D, esp. 36B-D), where a unitary entity is created by the combination
of a circle of the Same and Ž ssions of the circle of the Other. From the description
of cognition at 37A-C, and of human souls endowed with revolutions of the Same and
the Other at 41D and 43A-44B, it is clear that any soul will have the same basic struc-
ture in a Platonic theory. This in turn means that the Eudoxan (and later the Callip-
pean) planetary theory could be viewed also as a psycho-analysis of these entities. The
Ž rst planetary spheres in each set, performing the diurnal motion of the heavens, were
the circles of the Same in each individual case (cf. Aristotle Metaphysics 12.8
1073b18f., b24f. and [Alexander] ad loc., in Meta 703.16-22, where we learn that not
only the sphere of the Ž x stars, but the Ž rst spheres of planetary systems, performing
the same motion, can also be called Žplan®w), and the other spheres, particular to each
planet, described the internal conŽ guration of the circle of the Other speciŽ c to that
celestial being.
17
In this respect it is immaterial whether Aristotle intended 47 or any other Ž gure
as the end result of his computations. Whatever his intended result was, it was admit-
tedly different from his Ž rst Ž gure of 55.
18
The more so as he could draw on SosigenesÕ detailed study for a comparative
assessment of homocentric celestial theory. For Sosigenes see Simplicius in De caelo
499-510, where even if SimpliciusÕ acquaintance with SosigenesÕ book is Ž rst hand,
AlexanderÕs commentary presumably contained a reference to, if not a recapitulation
of, SosigenesÕ book.
198 ISTVçN M. BODNçR

My contention then is that Alexander after talking about the sphere of

Saturn, Jupiter etc., could keep the possibility open to talk, on occasion,
about the very same entity as a bundle of different spheres, in whatever
pattern they are arranged by a viable astronomical theory. As precedents
or parallels19 to this usage Adrastus (as preserved by Theon [= T] and
Calcidius [= C]) and Alcinous can be quoted. Although Adrastus repeat-
edly talks about 7 planetary spheres (T 141.15, T 142.18 and T 143.13),
or on one occasion he says that each planet is carried in its proper sphere,
in the singular, 20 this will not stop him from expounding an epicyclic plan-
etary theory, where what counted as the sphere of a planet will be further
broken down into a deferent and an epicycle, or a hollow and a solid
sphere in AdrastusÕ terminology. 21
The problems of this dual aspect are well borne out by the almost self-
refuting formulation Alcinous 14.7 uses. There he speaks within the same
sentence about seven spheres being within the planetary sphere.22 Never-
theless this paradoxical way of expression does not originate with Alci-
nous. What he does here is nothing else than restate in terms of spheres
the hallowed example of Timaeus 36C-D, where the Demiurge creates the
World Soul by further dissecting the circle of the Other into the circles of
the 7 planets, thereby giving rise as Alcinous puts it to the different plan-
etary circles within Òthe wandering circle of otherness.Ó 23 Although Alci-
nous, writing a handbook of Platonic philosophy and not of astronomy,
stops at this level of analysis, just as the planetary sphere, or the circle
of the Other could be further subdivided into the spheres/circles of the
individual planets, there is no reason why these in turn could not be even

19
We have no indications for AlcinousÕ date. If he was identical with Albinus, he
was a precursor of AlexanderÕs. Otherwise we have to rest content with the claim that
his usage can provide a parallel to the one we should expect from Alexander.
20
T 148.7 = C 122, 15f., cf. T 128.9f. = C 111.14, T 129.1f. = C 111.17 speaking
about the sphere of the Sun. It may be claimed that some indication of the internal
structure of the unitary planetary spheres can already be detected from the point on
where Adrastus starts to talk about the motion according to the circle and in the sphere
of the planet (T 148.6f. = C 122.15f., T 150.13f. = C 125.6f.). Cf., moreover, T 152.3f. =
C 125.16f., according to which the planets are in different spheres and circles. Note
that on these occasions CalcidiusÕ translation drops the mentioning of either circles or
of spheres.
21
See T 181-9.
22
Cf. 15.4, speaking again about the sphere of the planets.
23
Cf., also PlatoÕs repeated claims in the Timaeus that the world is a single visi-
ble living being, containing within itself all the other living beings (30C-31A, 41B-C
and 69C).
 ALEXANDER OF APHRODISIAS ON CELESTIAL MOTIONS 199

further analyzed into components, which perform some distinguished –

e.g. basic and single – motion. 24
The cases of Adrastus and Alcinous (or of Simplicius) then suggest that
there is nothing intrinsically implausible in talking about the very same
entity as a single sphere and as a combination of different spheres on dif-
ferent occasions. Before I would turn to the scanty evidence which sug-
gests that Alexander also did so it is important to delimit the range of
possible alternatives available for Alexander. My basic presupposition in
charting these alternatives is that Alexander would have accepted only a
celestial theory which observed the basic principles of the Aristotelian phi-
losophy of nature. This in turn means that the expositions of these two
interpreters of Plato could not have suited his purposes: Alcinous does not
give any detailed discussion of planetary motion, whereas on AdrastusÕ
scheme the transmission of diurnal motion to the nested deferents of the
planets remains unexplained.
As we have seen quaestio 1.25 declared that Ò[each of the planetary
spheres] is also moved in a second motion by the Ž rst [sphere], being car-
ried round with the same motion as itÓ (40.29f.). 25 Alexander does not

24
To Adrastus and Alcinous SimpliciusÕ repeated claims also could be added, that
the whole heaven can be regarded as a single sphere and not only as eight spheres.
Moreover Simplicius also accepts the Aristotelian system of 55 spheres as a valid
description of the celestial domain.
Note, however, that these claims in SimpliciusÕ case are combined with the doc-
trine of the interpenetration of celestial spheres, which is unacceptable to Alexander.
Simplicius voices his conviction that the whole of the heavens forms a unity at in De
caelo 25.5-10 and 552.3f. Then at in De caelo 510.19ff. and 515.2ff. and at in Phys.
531.3ff., 616.23ff., 623.32ff., 643, 18ff. and 966.5ff. he expounds that each of the heav-
enly spheres is a complete (though as is also clear from in De Caelo 469.11ff. and in
Phys. 616.23, immaterial) sphere and not just a vault (cÛw at in Phys 616.23ff. and
966.5ff.), and so they are present in each other at the same place. For AlexanderÕs
rejection of any bodily interpenetration see e.g. Simplicius in Phys 868.29ff. On the
larger issue of bodily interpenetration in neo-Platonism see Sorabji ch. 7.
25
Cf., furthermore, Simplicius in Phys 1357.6ff., where Alexander claims that the
diurnal motion of the planetary spheres is caused by the mover of the Ž rst sphere
rather than by the Ž rst sphere itself. It is important to note that (pace Sharples [1982],
p. 210) this correction need not reject what Alexander says at quaestio 1.25 40.29f.,
rather it can provide a more proper formulation of the very same claim: the diurnal
motion of the planetary spheres may be caused by the Ž rst sphere, nevertheless it can-
not be an effect of the motive power within that sphere: as an inhabitant of a Ž nite
body it would not be able to induce eternal and so inŽ nite motion (cf. further n. 1
above). Rather it is the Ž rst mover of the sphere of Ž xed stars which is responsible
both for the motion of this sphere, and for the further motions in the universe effected
by this sphere. Cf. On the principles of the All, (a treatise extant only in Arabic translation)
200 ISTVçN M. BODNçR

specify how this interaction would be possible. As the interpenetration of

the celestial spheres is ruled out, and as (physical) action at a distance is
not part of the standard Peripatetic scheme, he had three basic options
open. First – contrary to the indications of quaestio 1.25, and to my argu-
ments in n. 25 above – he could have claimed that the overall uniformity
of the diurnal motion of the different celestial spheres goes back to the
same source, without these motions being causally subordinated among
themselves.26 Second, he could bring in direct contact the Ž rst sphere and
all the planetary Òspheres,Ó in a way analogous to one of PtolemyÕs pro-
posals in bk 2 of his Planetary hypotheses. There Ptolemy, after propound-
ing a model of hollow and solid spheres delineates an alternative account,
in which planetary revolutions are performed by spherical segments (prÛs-
mata, called manshâr¨t in the Arabic translation) and these planetary seg-
ments27 can be directly enveloped by the diurnal revolution of the common
aether (ch. 8 120.23ff. Nix).28 Ptolemy then can compare this motion to
the way several objects swim in the  ow of the same river (ch. 6 117.34f.
Nix), or to  ying birds drifting with the same wind (ch. 7 119f. Nix).29
Third, the diurnal motion can be transmitted to the planets through the

130.41f. and 132.19ff. Badawi, where the planetary spheres are also moved interme-
diately – presumably through the mediation of the motion of the Ž rst heavens – by
the Ž rst mover.
26
Averro‘sÕ celestial theory may have been along these lines, see his Commentary
on Metaphysics bk L¨m 1606f. comment 37, where he attributes the diurnal motion
of the planets to the desire they have for the motion of the Ž rst body (cf. GenequandÕs
summary on p. 41, according to which the object of desire in this case is the Ž rst mov-
ing body itself, and not its motion). He does not claim that he would be following
Alexander on this point, and as his explanation of the transmission of celestial motions
can be largely conditioned by his insufŽ cient grasp of the principles at work in a celes-
tial system containing counteracting spheres (see his discussion at 1673-7 in comment
47), it is not particularly likely that he would take over this line of thought from
Alexander, who presumably was clear about the function of counteracting spheres.
27
The details of this setup are elaborated for each planet in later chapters: Sat-
urn in ch. 12 (129.12ff. Nix), Jupiter, Mars and Venus in ch. 13 (131f. Nix), the Sun
in ch. 14 (133.7ff. Nix), Mercury in ch. 15 (137.7ff. Nix), and the Moon in ch. 16
(140.31ff. Nix).
28
Cf. ch. 17 142.4f. Nix, where the domain of this common aether is referred to
as the sphere of the rest of the aether, in which the spherical segments are embedded.
29
Cleomedes (Meteora I 2.11-9) and Proclus (in Rem publ. 234.8, and possibly also
l.3) mention similar models. PtolemyÕs case remains unique, however, in that he pro-
poses this physical model and at the same time allows for an analysis of planetary
motions into different constituent revolutions. In contrast Cleomedes advocates a the-
ory deploying eccentric circles, and not epicycles, and he invariably speaks about a
single circular motion, a single sphere of the Sun and the Moon, whereas Proclus uses
 ALEXANDER OF APHRODISIAS ON CELESTIAL MOTIONS 201

motions of the adjacent planetary spheres. In this case Alexander has to

make sure that only the diurnal component of the motion is transmitted,
and the rest is Ž ltered out. Again, all sorts of Ž ltering mechanisms can be
conceived for this purpose, but surely the most likely one to use is some
analogon of counteracting spheres, which prevent the undesired compo-
nent of motion from being carried over to further planets. 30
Not only is the third option intrinsically the most plausible of these, we
have some further indications suggesting a homocentric theory for Alexander.
First of all we can deploy an argumentum e silentio: when Simplicius
goes out of his way to summarise what Sosigenes had said about celes-
tial theory he does not say much about what AlexanderÕs commentary
contained on Aristotelian sphairopoiia. He only mentions that Alexander
(and also Porphyry) were baf ed by the elementary mistake of subtraction
at the Ž nal enumeration of celestial spheres in Metaphysics 12.8. The point
of this comment is apparently that in the only case where Sosigenes could
not provide an overwhelmingly persuasive answer – he suggested that
after all the silly equation 55 - 6 = 47 should just be a scribal error – the
two standard, major later commentaries were not satisŽ ed with this pro-
posal, but had nothing better to say. The fact that Simplicius invokes
Alexander at this juncture suggests that he did not Ž nd anything about
homocentric models in AlexanderÕs commentaries on De caelo and the
Metaphysics with which the account of Sosigenes could have been sup-
plemented. Furthermore SimpliciusÕ silence can be taken to suggest that
AlexanderÕs (and PorphyryÕs) comments were in the same vein as Sosi-
genesÕ.31 Moreover if AlexanderÕs commentary on De caelo 2.12 contained

his simile of the motion of an ant on the surface of a revolving cylindrical [sic!, as
also in Cleomedes loc. cit.] surface not in order to elucidate, but rather to repudiate
celestial models where planetary motions are broken up into several constituent revo-
lutions. Planets on ProclusÕ understanding occupy an intermediate place in the cos-
mos, their motions are Ònot simple, but variegatedÓ (233.1f.) and Òirregularly regularÓ
(234.27f.), they incorporate some linear component in their basic cyclical motion
(233.6-20). Planetary models then, he claims, gain their currency not from the fact that
Òthey [the celestial bodies] are actually moving in that manner, but from the fact that
we are unable to observe this motion unless we deploy hypotheses of this sort, through
which we are able to <generate/account for> the irregular motions through addition of
the postulated regular ones.Ó (233.18-234.2, cf. also his concluding remarks to the
Outline of astronomical hypotheses 236.10-238.27).
30
This option is not necessarily equivalent yet with AristotleÕs theory of homo-
centric carrying plus counteracting spheres. In fact an epicyclic model can also be simi-
larly outŽ tted, adding a counteracting sphere after each of the deferents, so that its
effect remains restricted only to the planet it belongs to.
31
This argument is taken from Sharples [1982], p. 209, n. 91.
202 ISTVçN M. BODNçR

some description of a theory of homocentric spheres, similar to, albeit pre-

sumably more concise than SimpliciusÕ extract from Sosigenes, then
AlexanderÕs talk of homocentric spheres could follow quite closely on his
comments on De caelo 2.10 (quoted by Simplicius at in De caelo 472.8ff.),
where he spoke about the (single) sphere of Saturn.32
Besides SimpliciusÕ silence some further indications for AlexanderÕs
propounding a homocentric theory may come from Averro‘sÕ commentary
on Metaphysics bk 12. Averro‘s in his introduction (1393) states that he
possessed a commentary by Alexander on about two thirds of book 12,
and Genequand (p. 7) argues that presumably the comments on ch. 8
already were missing from the copy Averro‘s had. Nevertheless he had
access to some information on homocentric theory, which he attributed to
Alexander (1673). Strangely enough he completely misunderstands these
apparently sound remarks,33 thinking that each counteracting sphere al-
ways followed immediately after the sphere whose motion it was devised
to cancel. Accordingly Averro‘s voiced his concerns at 1674 that this is
a setup which cannot work, and promptly proposed some modiŽ cations.
Some pages above, at the beginning of this discussion, after stressing
the complexity and inaccessibility of the subject, and remarking that this
model is not any longer in favour in astronomical circles, Averro‘s men-
tions that Alexander and Themistius34 also acknowledged something with
regard to this theory, but they did not understand the reason he mentioned
in this connexion (1667f.). I use this somewhat unspeciŽ c formulation, be-
cause Genequand proposes two interpretations of the clause about Alex-
ander and Themistius, and acknowledges that yet a third one, proposed
earlier by Freudenthal, is also possible. GenequandÕs two proposals are
that (1) Alexander and Themistius acknowledged the difference between
the homocentric and the epicyclic theory, or that (2) they acknowledged

32
Note, moreover, the wording of this testimony: the passage starts out to speak
about the motion imparted by the Ž rst sphere to the sphere of the Saturn, and then
states the general claim that the motion of the celestial spheres by each other (êpƒ
Žll®lvn) is not forced. Although strictly speaking not false, this general formulation
is super uous on a model where transmission of motion occurs only between the Ž rst
sphere and the planetary spheres, and may be already looking forward to a discussion
of the more detailed homocentric model.
33
Suggesting that Averro‘s could not have had much to go by, which is either
another indication that he did not have access to the commentary on Metaphysics 12.8,
or alternatively these comments were so condensed that Averro‘s was unable to deci-
pher them.
34
Whose paraphrase of the De caelo was available for Averro‘s in its entirety, see
his introduction to the commentary at 1393.
 ALEXANDER OF APHRODISIAS ON CELESTIAL MOTIONS 203

that the old theory was no longer understood, 35 whereas (3) FreudenthalÕs
proposal was that Alexander and Themistius acknowledged the old theory
itself.36 Of these (2) (without any further qualiŽ cations, see n. 39 below)
cannot be right: Alexander, a pupil of SosigenesÕ, could not have possibly
complained about the lack of comprehension of the homocentric model,
and Themistius does not in fact do so either. (1) is also somewhat inept:
nobody can possibly disregard the difference of the two models unless she
misunderstands at least either of them. Moreover, again, there is no word
in ThemistiusÕ paraphrase of Metaphysics L about epicyclic theory. There-
fore I submit that from the three possible interpretations mentioned by
Genequand the old one of Freudenthal seems to make the most sense.
According to that Alexander and Themistius embraced the old, homocen-
tric theory. This is certainly correct in ThemistiusÕ case, he in his in Meta
L 26.34-28.7 propounds the homocentric theory, without indicating any
concerns or reservations. Averro‘sÕ testimony then would afŽ rm that Alex-
ander had a similar attachment to this model. On this reading the last
clause of the sentence should indicate that although Averro‘s had access
to some texts where these two commentators propounded the homocentric
theory in their own voice, nevertheless he did not Ž nd any critical dis-
cussion and rejection by them of the rival theory of epicycles, and there-
fore he had to supply this himself.37 This comment would certainly be true
of ThemistiusÕ procedure, and it reinforces once more the conclusions we
drew from SimpliciusÕ silence about AlexanderÕs treatment of the techni-
calities of celestial theory: in his works, available to Simplicius (and a
fortiori in the passages which were available to Averro‘s) Alexander did
provide some discussion of homocentric spheres, but it was nowhere near
the elaboration and detail SosigenesÕ treatise devoted exclusively to this
subject. 38

Eštvšs University, Budapest*

35
Genequand, p. 179, n. 161.
36
In Freudenthal fr. 33, on p. 111: ÒZu ihr [sc. der frŸheren Methode der Berech-
nung] aber haben sich Alexander und Themistius bekannt, [. . .].Ó
37
So Freudenthal p. 111, n. 4.
38
It is important to mention that in all probability even according to the two alter-
native versions I rejected Alexander should have had a section on homocentric spheres.
Provided Alexander acknowledged the difference of the two models this he could have
most likely done in a discussion of the two theories. Then SimpliciusÕ and Averro‘sÕ
silence – both of them ardent followers of the homocentric model – will pledge that
Alexander remained a faithful student of AristotleÕs on this count as well. On the other
hand the most likely context for Alexander (a disciple of SosigenesÕ) to acknowledge
204 ISTVçN M. BODNçR

References**

Alcinous The Handbook of Platonism. Tr. with intr. and commentary by J. Dillon.
Oxford: Clarendon 1993.
Alexander On the principles of the All. French translation in Badawi, English transla-
tion of some selected passages in Rosenthal.
Averro‘s Commentary on AristotleÕs Metaphysics Book L¨m, quoted by the page num-
bers of BouygesÕ edition, given in the margins of Genequand.
Badawi: B., A. La transmission de la philosophie grecque au monde arabe. (ƒtudes
de philosophie mŽ dievale, vol. 56) Paris: Vrin 1968 [2nd, revised and enlarged edi-
tion: 1987].
Bruns [1890]: B., I. ÒStudien zu Alexander von Aphrodisias. III. Lehre von der
Vorsehung,Ó Rheinisches Museum 45 (1890), pp. 223-235.
Calcidius: Timaeus a Calcidio translatus commentarioque instructus. Ed. J.H. Waszink
(Plato Latinus, vol. 4) London: Warburg Institute-Leiden: Brill 21975.
Freudenthal: F., J. ÒDie durch Averroes erhaltenen Fragmente Alexanders zur Meta-
physik des Aristoteles,Ó Abhandlungen der Kšniglichen Preu ischen Akademie der
Wissenschaften zu Berlin, 1884.
Genequand: G., Ch. Ibn RushdÕs Metaphysics. A translation with introduction of Ibn
RushdÕs Commentary on AristotleÕs Metaphysics, book L¨m. (Islamic philosophy
and theology, vol. 1) Leiden: Brill 1984.
Goldstein: G., B.R. ÒThe Arabic version of PtolemyÕs Planetary hypotheses,Ó Transactions
of the American Philosophical Society 57 (1967) part 4.
Merlan: M., Ph. ÒPlotinus Enneads 2.2,Ó Transactions of the American Philological
Association 74 (1943), pp. 179-91 (Photomechanically reprinted in: Merlan, Ph.
Kleine Schriften. (ed. by F. Merlan) Hildesheim-New York: G. Olms., pp. 396-408).
Ptolemy Planetary hypotheses in Opera quae exstant omnia, vol. ii, Opera astronom-
ica minora. Leipzig: Teubner 1907. The Greek text of bk 1 is edited and translated
into German by J.L. Heiberg, the Arabic version of bk 2 is translated into German
by L. Nix. A part of bk 1, extant only in Arabic, which is missing in this edition
appeared (in an English translation) in Goldstein.
Rosenthal: R., F. The Classical Heritage in Islam. London-New York: Routledge 1975
(Translation of: R., F. Das Fortleben der Antike im Islam. Zurich-Stuttgart: Artemis
1965).

that the old astronomy is no longer understood can be polemical: we, Aristotelians,
still understand it properly, the fault lies with those uninitiated astronomers who con-
jure up new models for the explanation of celestial phenomena instead of carefully
and attentively studying the work of past masters.
* Section 1 of this paper was written in 1994/95 as part of a larger project on peri-
patetic celestial theory while I held a junior fellowship at the Center for Hellenic
Studies, Washington D.C., then in March/April 1996 I had the opportunity to revise
section 1 and to write section 2 at Fondation Hardt, Vandœuvres, Switzerland. I am
indebted for the help and support I received at both institutions, and am especially
grateful to Bob Sharples for written comments on earlier drafts.
** Bibliographical details of editions of ancient texts have been supplied only
in case they differ from the indications in Liddell – Scott – Jones, A Greek-English
Lexicon.
 ALEXANDER OF APHRODISIAS ON CELESTIAL MOTIONS 205

Sharples [1982]: S., R.W. ÒAlexander of Aphrodisias on divine providence: Two prob-
lems,Ó Classical Quarterly 32 (1982), pp. 198-211.
—— [1992]: Alexander of Aphrodisias Quaestiones 1.1-2.15. Translated by S., R.W.
Ithaca, NY: Cornell UP or London: Duckworth.
Sorabji: S., R. Matter, space and motion. Theories in antiquity and their sequel Ithaca, NY:
Cornell UP or London: Duckworth 1988.
Zeller: Z., E. Die Philosophie der Griechen in ihrer geschichtlichen Entwicklung. Part 3
section 1: Die nacharistotelische Philosophie. Leipzig: O.R. Reisland 1923
(Photomechanical reprint: Hildesheim: Georg Olms 1963).