ABU MA{SAR, AL-KINDI AND

THE PHILOSOPHICAL DEFENSE

OF ASTROLOGY

In his Fihrist, Ibn al-Nadim provides us with a tantalizingly brief

intellectual biography of the greatest astrologer of Islam, Abu Ma{sar
al-BalÌi. Al-Kindi, the first Islamic philosopher and a contemporary
of Abu Ma{sar (the latter died in 866 A.D., al-Kindi in about
870 A.D.), features prominently in the account. According to Ibn
al-Nadim, Abu Ma{sar began his career as «one of the followers of tra-
dition (aÒÌab al-hadi†)» who bore a grudge against al-Kindi and
encited the crowd against him. But al-Kindi enticed him towards the
study of astrology with the bait of «the sciences of arithmetic and
geometry». Abu Ma{sar fell in with al-Kindi, but did not perfect him-
self in the aforementioned sciences1. This anecdote is intriguing in
several respects. It leads one to wonder about the relationship between
al-Kindi and the Islamic mainstream. May we infer that Abu Ma{sar’s
initial hostility to al-Kindi was grounded in a traditionalist rejection
of the Greek philosophy about which al-Kindi was so enthusiastic2.
The account in the Fihrist also prompts speculation on the relation-
ship between al-Kindi and Abu Ma{sar after the latter’s philosophical
conversion. But above all, the anecdote reminds us that at this early
period astrology was a discipline integral to the study of philosophy3.
In what follows I will pursue some of the issues raised by this
passage in the Fihrist by examining an outstanding example of the

1. IBN AL-NADIM, al-Fihrist, ed. G. FLÜGEL, Leipzig 1871-2, p. 277.1-8.

2. Some trace of this antagonism may be discerned in al-Kindi’s vigorous defense of
Greek thought in the first chapter of On First Philosophy. On the translation movement
under the {Abbasid dynasty, see D. GUTAS, Greek Thought, Arabic Culture, London 1998.
For a survey of the philosophical translations, see G. ENDREß, «Die wissenschaftliche Lit-
eratur», in: Grundriß der arabischen Philologie, Bd. III Supplement, ed. W. FISCHER,
Wiesbaden 1992, pp. 24-61. For al-Kindi’s circle see further G. ENDREß, «The Circle of
al-Kindi», in: The Ancient Tradition in Christian and Islamic Hellenism, ed. G. ENDREß
and R. KRUK, Leiden 1997, pp. 43-76.
3. For an in-depth discussion of the cultural and political forces that made astrology
so important within the translation movement, see GUTAS (1998), pp. 108-110.

©RTPM 69,2 (2002) 245-270

246 RECHERCHES DE THÉOLOGIE ET PHILOSOPHIE MÉDIÉVALES

philosophical justification of astrology: the first treatise of Abu

Ma{sar’s most significant work, the Kitab al-MudÌal al-kabir {ala
{ilm aÌkam al-nugum (the Great Introduction to the Science of Astrol-
ogy, hereafter MudÌal)4. Along with al-SharaÌsi5, Abu Ma{sar was
al-Kindi’s most significant disciple, yet there has been no serious
research into the relationship between Abu Ma{sar’s thought and the
works of al-Kindi. I will attempt to begin the study of that relation-
ship here, focusing chiefly on similarities between the MudÌal and
al-Kindi’s closely related writings on the nature and effects of the celes-
tial bodies. Most important here are two epistles that have survived
in Arabic: Kitab fi }l-ibana {an al-{illa al-fa{ila al-qariba li-}l-kawn wa
}l-fasad (On the Explanation of the Proximate, Agent Cause of Genera-
tion and Corruption) and Risala fi }l-ibana {an sugud al-girm al-aqÒa
(On the Explanation of the Bowing of the Outermost Body)6. As will

4. ABU MA{SAR AL-BALÎI, Liber introductorii maioris ad scientiam judicorum astrorum,

ed. R. LEMAY, Naples 1995-6. The Arabic edition for all passages to be cited here is in
volume II.
Forthcoming are a new Arabic edition by K. YAMAMOTO and a new English transla-
tion by C. BURNETT (a shorter version of this article will appear as a preface to Burnett’s
translation). In what follows I will cite the MudÌal by Book (Qawl) and chapter (faÒl)
number, with section numbers from Burnett’s forthcoming translation in brackets, and
line numbers from Lemay’s edition prefaced by the abbreviation L. Translations of the
MudÌal have been taken from Burnett with slight modifications, but all other translations
are my own.
5. On whom see F. ROSENTHAL, AÌmad b. a†-™ayyib as-SaraÌsi, New Haven 1943.
6. Both treatises are in the first volume of AL-KINDI, Rasa}il al-Kindi al-Falsafiyya, ed.
M. ABU RIDA in two volumes, Cairo 1950-1953: Proximate Agent Cause pp. 214-237, On
the Bowing of the Outermost Body pp. 244-261. A new Arabic edition, with French transla-
tion, of On the Bowing of the Outermost Body, can be found in R. RASHED and J. JOLIVET,
Oeuvres Philosophiques & Scientifiques d}al-Kindi, volume II: Métaphysique et Cosmologie,
Leiden 1998, pp. 176-199. Citations to al-Kindi will be to page and line number of the
Abu Rida edition, marked AR, all citations being to volume I unless otherwise designated.
Where an edition has already appeared in Rashed and Jolivet (a series of which only two
out of five volumes has so far appeared) page and line numbers will also be given to this
edition, marked RJ. All of these citations are to volume II of Rashed and Jolivet.
Other works by al-Kindi important for the study of his attitude towards astrology and
the heavens include: R. fi }l-ibana {an anna †abi{a al-falak muÌalifa li-†aba}i{ al-{anaÒir al-arba{a
(That the Nature of the Celestial Sphere is Different from the Natures of the Four Ele-
ments): AR volume II pp. 40-46; De radiis (On Rays): preserved only in Latin and ed.
M.T. D’ALVERNY and F. HUDRY, «Al-Kindi, De Radiis», in: Archives d’histoire doctrinale et
littéraire du Moyen Âge 41 (1974), pp. 215-259; his actual works in applied astrology, which
include the Forty Chapters, on which see C. BURNETT, «Al-Kindi on Judicial Astrology:
The Forty Chapters» in: Arabic Sciences and Philosophy 3 (1993), pp. 77-117. See also
O. LOTH, «Al-Kindi als Astrolog», in: Morgenlandischen Forschungen (1875), pp. 261-309.
 ABU MA{SAR, AL-KINDI 247

become clear, the close doctrinal and textual parallels between the
MudÌal and al-Kindi support the idea that Abu Ma{sar was strongly
influenced by al-Kindi.
Our study of the MudÌal will also show that Abu Ma{sar has
rather sophisticated solutions to philosophical problems posed by the
practice of astrology. I will concentrate on three such problems. First,
how does astrology fit into philosophy as a whole? In answer to this
question Abu Ma{sar is determined to show not only that astrology is
an empirical science on the Aristotelian model, but also that it is one
of the highest, if not the highest, of such sciences. Second, what rea-
son is there to think that astrology can work? Again, Abu Ma{sar
depends on philosophical sources, explaining how the celestial bodies
influence the sublunar world with an account that is both physical
and metaphysical. Third, and perhaps most interesting, is the follow-
ing question: if we can foretell human actions using astrology, does
this mean that humans are not free? I will argue that, unlike the vast
majority of Islamic thinkers on the question of human freedom (but
like al-Kindi), Abu Ma{sar thinks that freedom is compatible with
determinism.

1. Astrology as an Aristotelian master science

In describing the exalted status of the art of astrology, Abu Ma{sar

shows himself to be familiar with the notion of a science ({ilm, which
translates episteme) in the Aristotelian sense. It is not clear whether he
closely studied the translations of Aristotle produced in the 9th cen-
tury, or was familiar with Aristotelian doctrine via more indirect
and incomplete sources7. His use of Aristotelian epistemology in this

7. R. LEMAY, Abu Ma{shar and Latin Aristotelianism in the Twelfth Century, Beirut
1962, contains (in chapter II) the most extensive discussion of Abu Ma{sar’s Aris-
totelianism, though it is unfortunately based only on the Latin translations by John of
Seville and Hermann of Carinthia. Lemay makes much of the Aristotelianism of
the MudÌal. He admits that «a large proportion of Aristotelian doctrines are quoted
from some post-classical compendia with the resulting promiscuous mixture of pure
Aristotelianism with Stoicism and later neo-Platonism» (44), but seems to presume
basic familiarity on Abu Ma{sar’s part with texts ranging from De interpretatione to the
Metaphysics.
248 RECHERCHES DE THÉOLOGIE ET PHILOSOPHIE MÉDIÉVALES

respect is often superficial enough to be explained by the latter

hypothesis. For example, in the midst of the rather pedantic opening
chapter of the MudÌal, he says that astrology belongs to the «theo-
retical part (guz} al-{ilm)» of philosophy, presumably as opposed to
the practical part (MudÌal I.1 [14] L 68): the distinction is Aris-
totelian but too commonplace to indicate use of any specific text8.
More sophisticated is Abu Ma{sar’s treatment of the method used
in astrology, and by implication in science generally. This method
can be divided into two stages: first, the gathering of empirical data,
and second, the use of analogy (qiyas). Abu Ma{sar’s empiricism
comes out especially clearly in the second chapter of the first treatise,
where he repeatedly emphasizes that the experiences (tagarib) of ordi-
nary folk such as farmers and sailors amply prove that the heavens
have an effect on the sublunar world. He later takes quite seriously
an objection against astrology that points out that one lifetime is not
enough to assemble the necessary experience of celestial influence in
order to predict future events (MudÌal I.5 [25] L 904-911). In his
reply to the objection he in no way retreats from the empiricist
position he has taken throughout: astrology is possible, he says,
because of the observations made by previous generations (MudÌal I.5
[26-31] L 912ff.).

An opposing stance is taken by David PINGREE in his entry on Abu Ma{sar in the Dic-
tionary of Scientific Biography, volume I, New York 1970, pp. 32-39. Pingree holds that,
although the MudÌal is «largely Aristotelian», it is almost exclusively based on informa-
tion taken from the astrological tradition in Îarran, i.e. the tradition of Hermetic doc-
trines whose importance Lemay minimizes (44, 130-131). Pingree is surely right that
Abu Ma{sar, like al-Kindi, was influenced by the Îarranian Hermetic tradition (cf. Ibn
al-Nadim, al-Fihrist, pp. 318.14-320.9, for a report on the Sabeans at Îarran that stems
from al-Kindi himself ). But we cannot exclude the possibility that he read the Arabic ver-
sions of such texts as the Metaphysics (translated for al-Kindi by one Us†ath), De meteoro-
logica and De caelo (translated by YaÌya b. al-Bi†riq), the Posterior Analytics and De inter-
pretatione (translated by Hunayn b. IsÌaq), all available around the time of al-Kindi’s
circle.
8. Al-Kindi also uses this distinction: see his treatise On the Five Substances, preserved
only in Latin: «The wise Aristotle, at the beginning of his dialectics, says that the knowl-
edge of anything falls under philosophy… philosophy is divided into theoretical and
practical» (AR volume II, p. 9.2-9). One might of course think that astrology is in fact a
practical science, since it assists us in practical deliberation. Abu Ma{sar’s insistence that
it is theoretical may have something to do with his desire to show that it is the highest
science (as we will see below); Aristotle denies that the highest science can be a practical
science (Metaphysics A.2, 982b11).
 ABU MA{SAR, AL-KINDI 249

The use of analogy allows the astrologer to extrapolate from

this empirical data, reaching general conclusions about the effects
produced by the heavens (see MudÌal I.2 [3] L 110-111). This is
how the ancients first established the laws of astrology. After making
their observations, «the philosophers looked at these things that the
planets were indicating time after time, and on this basis they
worked out by analogy what was hidden from them» (MudÌal I.5
[31] L 946-8). Abu Ma{sar says that the science of medicine is like
astrology in this respect (MudÌal I.2 [24-25] L 271-273, 288-299).
Analogy makes science possible by allowing us to go beyond the par-
ticular instances we have experienced, predicting future events on the
basis of general laws9. Abu Ma{sar is of course exactly right that
something like analogy is needed to supplement the fruits of experi-
ence, and with this insight he seems to have expanded upon Aristo-
tle’s own account of how scientific induction works in the Posterior
Analytics10.
One of the significant results of these methodological considera-
tions is that we can come to have an understanding of the nature of
the heavens merely by observing their effects in the sublunar realm.
Abu Ma{sar is explicit that this is only one example of a general
philosophical method: «for everything that is caused, its cause is
prior to it in level; the effect is posterior to the cause, and we know
the cause through the effect» (MudÌal I.4 [4] L 531-2)11. Here he is
endorsing a philosophical principle of the utmost importance in the
Islamic milieu: causes are known by their effects. Normally the prin-
ciple appears in attempts to know God on the basis of what He has
created. As Thomas Aquinas will later do in upholding the superior-
ity of knowledge propter quid (knowledge of effects through causes)
to knowledge quia (knowledge of causes from their effects), texts

9. See further Abu Ma{sar’s response to the second class of objectors to astrology in
the MudÌal, who claim (I.5 [4] L 88ff.) that astrology can grasp only the universal, not
the particular. Abu Ma{sar says that the objection is mistaken, because it is in fact pre-
cisely through the universal that the particular is known.
10. On the possible connections of his use of analogy and the concept of qiyas in
Islamic jurisprudence, see J.-C. VADET, «Une défense de l’astrologie dans le MudÌal
d}Abu Ma{sar al-BalÌi», in: Annales islamologiques 5 (1963), pp. 131-180, at 141-143.
11. Abu Ma{sar seems to present this as a direct quote from Aristotle; for possible
sources in Aristotle see LEMAY (1962), p. 73.
250 RECHERCHES DE THÉOLOGIE ET PHILOSOPHIE MÉDIÉVALES

from al-Kindi’s circle frequently say that God is known through His
effects, but only imperfectly. Thus the pseudo-Aristotelian transla-
tion of Proclus’ Elements of Theology, the Kalam fi }l-Ìayr al-maÌ∂
(On the Pure Good, later known as the Liber de causis), teaches that
God, the First Cause, «is described only by the second causes, which
are lit by the illumination of the First Cause… thus it happens that
the First is the one for which description fails»12.
Abu Ma{sar prefers instead to emphasize the positive benefits of
passing from effects to causes, until we reach God Himself: «see how
we have perceived the Creator as the Mover of things, from visible,
known things, perceptible to the senses!» (MudÌal I.3 [11] L 515-
516). Nevertheless Abu Ma{sar normally contents himself with
astrology — the study of the stars, rather than God — as the pinna-
cle of human knowledge. Indeed Abu Ma{sar adapts arguments he
must have found in a metaphysical and theological context, applying
them to the stars instead of God. A good example is his use of an
argument developed by al-Kindi, which presses grammar into the
service of theology:
Proximate Agent Cause, AR 215.10-14: The arrangement and order of this
world… is the greatest proof of the perfection of rule (tadbir, which can also
mean «Providence»: see below, section III). But together with all rule there
is a ruler. [And in this is proof ] of greatest wisdom, but together with all
wisdom there is someone wise. For this belongs to the genitive construction
(mu∂af).

Essentially this is a version of the cosmological argument: the world

is characterized by order, and there must be an agent or cause respon-
sible for that order. Al-Kindi however makes the unusual move of
providing a linguistic premise to get from the presence of order in
the world to the existence of a cause for that order. The premise is
that a word like «rule» or «wisdom» carries with it an implied
«genitive construction», which is to say that wisdom, for example, is
something that can only exist by belonging to something else. Simi-
larly rule belongs to or is exercised by the agent who is ruling. Abu
Ma{sar borrows this argument in the MudÌal to establish that there

12. Die pseudoaristotelische Schrift über das reine Gute, ed. O. BARDENHEWER, Frank-
furt a.M. 1882), Proposition 5, pp. 69.9-70.3.
 ABU MA{SAR, AL-KINDI 251

is an external cause of elemental combination in the sublunar realm,

the cause being in this case not God, but the stars:
MudÌal I.4 [5] L 538-9: Natured things do not subsist except through the
proportion of natures13, and the natures do not subsist in the natured things
except through the proportion of composition. And composition does not
occur except through a composer, and the natured thing [does not occur]
except through a cause.

Abu Ma{sar’s use of essentially theological styles of argument to prove

the efficacy of astrology is in keeping with the ambitious claims he
makes for this science: it is «the noblest of all the arts (Òina{at) in
importance and the most splendid in rank» (MudÌal I.2 [26] L 303-
304). He has several reasons for placing such a high value on astrol-
ogy. First, it is the study of noble substances, for the stars are, unlike
sublunar things, immune to generation and corruption14. Second,
astrology is the most useful of sciences because it allows one to pre-
dict future events and discern the principles of the stars’ causation,
which are ultimately responsible for all events in the sublunar realm
(see section 2 below).
But the third and most rigorous reason Abu Ma{sar gives for the
exalted status of astrology is that it is a master science in the Aris-
totelian sense. Such a science, as explained in the Posterior Analytics
and elsewhere15, has two features that distinguish it from the lower
sciences. First, other sciences serve as a propaedeutic for the master
science. The main example of a lower science that relates to astrology
in this way is astronomy, «the science of the quality and quantity
of the higher spheres and the spheres of the planets» (MudÌal I.2 [2]

13. Here «natures» refers to the four elements, and «natured things» to what is com-
posed from the elements. For the influence of this contrast in the Latin west, see T. SIL-
VERSTEIN, «Elementatum: its Appearance among the Twelfth Century Cosmogonists», in:
Mediaeval Studies 16 (1954), pp. 156-162.
14. That is, they are not subject to corruption «for as long as God wills» (MudÌal I.2
[26] L 301), a caveat that strongly implies that Abu Ma{sar would follow al-Kindi’s rejec-
tion of the eternity of the physical universe. See similarly MudÌal I.5 [3] L 686-7: the
planets cause generation (kawn) in the sublunar world, which is «perpetual until the time
that God wishes». Compare Proximate Agent Cause AR 231: the forms of generated
things endure «for a limited period of time willed for being (kawn) by the Creator of
being», and AR 220, where al-Kindi says that both the planets and the elements endure
only «for the period of time appointed for them by God».
15. ARISTOTLE, Posterior Analytics I.2, Metaphysics I.2, III.2 996b11ff., Ethics I.2.
252 RECHERCHES DE THÉOLOGIE ET PHILOSOPHIE MÉDIÉVALES

L 95-96)16. Thus he says that «it is necessary for the man who spec-
ulates in the first branch of the science of the stars [sc. astronomy] to
speculate after that in its second branch [sc. astrology], because they
are interconnected sciences» (MudÌal 1.5 [23] L 889-92). Second, a
master science proves principles only assumed in the lower sciences.
Hence Aristotle’s analogy of the race-track in Ethics I.4: we get to the
master science by going through the lower sciences, and only after we
have reached the master science do we turn back, using it to ground
the lower sciences so that the entire system is demonstratively justi-
fied. Astrology bears this relationship to all the arts (MudÌal I.2
[12ff ] L 177ff ), but most notably to medicine, about which Abu
Ma{sar asserts explicitly that «the science of the stars is the principle
(awwaliyya) of the science of medicine» (MudÌal I.5 [36] L 1012-13,
cf. I.5 [39] L 1041 where it is called the «cause» of the science of
medicine). For only by understanding astral causation will a doctor
be able to grasp the causes for the conditions of the bodies that he
treats.
All of this is very like the claims made for metaphysics, not astrol-
ogy, by al-Kindi and Aristotle himself. Aristotle admits that, were it
not for the existence of immaterial substances, then Physics (which
for Abu Ma{sar culminates in the science of astrology) would be the
highest science (Metaphysics E.1, 1026a28ff.). But because there are
immaterial substances, the master science that justifies the principles
of all others is «first philosophy», i.e. metaphysics or theology. For this
science alone studies being in itself, and thereby the causes and prin-
ciples of all things (see e.g. Metaphysics A.2, 982b5-8, and generally
Metaphysics G). In al-Kindi’s circle the primacy of metaphysics is
defended by assimilating it to theology, and arguing that as the study
of God this science is the study of the cause of everything17.

16. This definition resonates strongly with al-Kindi’s description of the mathematical
sciences (riya∂at), which include astronomy, in his R. fi kamiyya kutub Aris†u†alis (On the
Quantity of the Books of Aristotle), AR 363-384, as the «science of quantity and quality»
(AR 372.14).
17. See AL-KINDI, K. fi }l-falsafa al-ula (On First Philosophy), AR 97-162, [RJ 9-99]:
«the philosophy that is most noble and highest in rank is First Philosophy, i.e. the science
of the First Cause, Who is the Cause of all truth» (AR 98.1-2 [RJ 9.13-14]). The Pro-
logue to the Theology of Aristotle, which I believe to have been written by al-Kindi (see my
forthcoming book The Arabic Plotinus, London 2002, section 2.1.2), similarly portrays
the study of God as the highest philosophical enterprise.
 ABU MA{SAR, AL-KINDI 253

Abu Ma{sar, on the other hand, gives astrology pride of place and
only once in the opening treatise engages in anything remotely
resembling theology, describing God as the unmoved mover on the
model of Aristotle’s god in Physics VIII and Metaphysics L (MudÌal
I.3 [11] L 510-517). This lends some plausibility to Ibn al-Nadim’s
somewhat dismissive remark in the Fihrist that Abu Ma{sar «did not
perfect himself» in the philosophical disciplines but studied only
astrology. Perhaps, though, there is more to his praise of astrology
than a desire to advertise the value of his own profession, or a lack of
competence in theology and metaphysics. As just mentioned, Abu
Ma{sar believes we can know God’s nature as an unmoved mover by
studying the stars. But if he is a good enough Aristotelian to think
that this is all we can know about God through rational means,
astrology will already yield the best possible theological knowledge.
Furthermore, as we will see below (section 3), astrology studies
divine Providence as it is exercised through the stars, which are the
instruments of God’s will. From this perspective as well, astrology
can lay claim to the exalted position of a divine science.

2. How the stars influence the lower world

A fundamental tenet of Abu Ma{sar’s astrology is that the stars phys-

ically cause sublunar events. They are not mere signs or symbols of
what is to come, despite his use throughout of the astrological tech-
nical term «indication (dalala)», which has no particularly causal
connotation. The causal link between the stars and their effects here
on earth is one of necessitation or «compulsion (i∂†irar)» (MudÌal
I.3 [3] L 417-418). The stars are themselves determined in their
motions by God: «it is now clear that God the Creator gave to the
stars natural indications and movements, and that it is as a result
of the powers of their natural movements in the four elements that
the composition of ‘natured’ things takes place» (MudÌal I.3 [7]
L 542-4). Abu Ma{sar’s universe is then completely determined by
God’s will, with the stars serving as intermediate causes between God
and sublunar events.
This is in substantial agreement with al-Kindi, who glosses the
Qur}anic assertion that the stars «prostrate» themselves before God
254 RECHERCHES DE THÉOLOGIE ET PHILOSOPHIE MÉDIÉVALES

by explaining that the stars move with a regular motion out of obe-
dience to the divine will, and thus bring about the change of seasons
and all generation and corruption here below (On the Bowing of the
Outermost Body, AR 246-7 [RJ 180-1]). Even more explicitly, he says
that the mixture of elements in the lower world «exists through the
motion of these heavenly bodies, through the will of the Creator,
great be his praise. For it is quite clear that the effects of God’s will
are more proximate, and that God’s will is their proximate cause.
And what are the remaining things other than the necessary out-
comes (lawaÌiq) of this wondrous generation, i.e. the generation of
natural and psychic (nafsani) things?» (Proximate Agent Cause, AR
226.8-11).
Through what causal mechanism, then, do the stars influence the
sublunar world? The answer comes in the third chapter of the first
treatise of the MudÌal:
MudÌal I.3 [3] L 419-424: For the outermost sphere surrounds this world
and moves with its stars over this world, with an eternal circular motion,
so that by its constant moving of the stars, and by their [sc. the stars’]
movement over this world, heat is produced in the earthly world which
is connected to them, and it [sc. the lower world] becomes hot. And
when this world becomes hot, it is made more subtle and moves, and
as a result of its movement there occur changes in these bodies from one
into another, and generation and corruption occur, by the permission of
God.

In other words, the heavens produce their effects by causing friction,

since they are actually touching the outside of the lower world (i.e.
the spheres of fire and air that surround the spheres of water and
earth). By constantly turning around this lower realm they heat and
agitate the matter within it, and this gives rise to all sublunar events.
Notice the emphasis on the production of «generation and corrup-
tion (kawn wa fasad)», which encourages comparison with al-Kindi’s
On the Proximate Agent Cause of Generation and Corruption, an epis-
tle whose very title asserts the primacy of generation and corruption
among the effects of the stars18.

18. For Abu Ma{sar’s similar emphasis on generation and corruption, see also MudÌal
I.2 [25] L 287ff; I.3 [2] L 411; I.5 [3] L 680; I.5 [23] L 889.
 ABU MA{SAR, AL-KINDI 255

Such comparison yields striking similarities, not least in the

explanation of the stars’ causation in terms of heat produced by fric-
tion:
Proximate Agent Cause AR 224.5-18: The cause producing heat in the ele-
ments is from the first element, which moves them through motion
in time, place and quality… When more bodies move over a part [of the
elements] more quickly, with greater proximity, from lower down, and
more intensely, then the heat of this part is increased. But when this fails
to occur then it stays according to its own nature. If [the moving body] is
more distant from a part of the earth and the composed things upon it,
then earth and water stay according to their nature, that is, they stay
cold… Whenever stars are distant from the zenith (samt) over the earth,
cold predominates, but when they are near, heat [predominates]. Then
moisture and dryness, which follow upon heat and cold, come about,
and from these four qualities the remaining qualities come about. Every-
thing that comes about does so in accordance with the deficiency or
excess of these.

This passage is especially useful because it fills a gap in the account

provided by Abu Ma{sar, who does not explain how we get from the
production of heat to the production of all four contraries and the
elements that are characterized by them. Here al-Kindi explains that,
since the proximity of a heavenly body results in heat, the absence
thereof explains cold, and he sees dryness and moisture simply as fol-
lowing from heat and cold.
Now, it must be said that neither al-Kindi nor Abu Ma{sar are
quite consistent in holding to this elegantly simple explanation of
astral causation. For example, they both think that when the moon is
close to a place on the earth, the effect in that place is not heat, but
moisture (Proximate Agent Cause AR 232.17-19, MudÌal I.2 [24]
L 280)19. Worse still, al-Kindi is not consistent between his various
works in the explanations he gives of the stars’ causation. While he
ascribes their effects to friction in Proximate Agent Cause, in the more
explicitly astrological and magical De Radiis he explains the stars’

19. Al-Kindi’s statement here is particularly puzzling, since he has just said that
when the moon is near to the earth it heats the air and disperses cloud vapor, so that it
rains more when the moon wanes (Proximate Agent Cause AR 231.18-232.6). Ptolemy’s
Tetrabiblios says that the moon causes all four of the contraries, hot, cold, wet and
dry, depending on what quarter it is in. See PTOLEMY, Tetrabiblios, ed. and trans.
F.E. ROBBINS, Cambridge 1940, I.8.
256 RECHERCHES DE THÉOLOGIE ET PHILOSOPHIE MÉDIÉVALES

actions in terms of a more general theory of cosmic rays20. This gives

us good reason to think that, of al-Kindi’s works relevant to astrol-
ogy, it is Proximate Agent Cause which is closest to the thought of
Abu Ma{sar, and especially to the more theoretical considerations of
the first treatise of the MudÌal.
(This may be an appropriate moment to mention some of the
miscellaneous parallels we find between the two texts just mentioned.
This is by no means necessarily an exhaustive list of parallels, but
should be sufficient to demonstrate the close relation between the
two works. (a) At Proximate Agent Cause AR 224.14-15, al-Kindi
says that heat and cold are the «active» contraries, while dryness and
moisture are «passive». Al-Kindi repeats the point in On the Expla-
nation that the Nature of the Celestial Sphere is Different from the
Natures of the Four Elements (AR volume II, 41.10ff ). We find the
same in the MudÌal (II.2, L 196-197). (b) In arguing that the sun has
an influence on the lower world, al-Kindi mentions (AR 225.2-8)
that those who live in hot climes have blackened skin and curlier
hair, a phenomenon also noted by Abu Ma{sar (III.3, L 244)21.
(c) Both authors argue that if the paths of the celestial bodies did not
vary, there would be no difference in climate from day to day and
season to season: Proximate Agent Cause AR 233; I.2, L 168-169,
III.3, L 286ff. (d) The incorruptibility of the celestial sphere is
argued for in the MudÌal at I.3, L 411, and by al-Kindi at Proximate
Agent Cause AR 219-20. Both base this on the fact that the heavens
are not made out of the four elements and thus have no contraries
predicated of them.
Though Abu Ma{sar and al-Kindi drew on the same sources, these
parallels make it likely that al-Kindi’s treatise was itself a source for
the MudÌal. That this was the direction of any borrowing is indi-
cated by the dating of the two works. The MudÌal was written in

20. This inconsistency has been pointed out by H. WIESNER in her PhD dissertation,
The Cosmology of al-Kindi, Harvard University 1993. For a study of the theory of rays
in al-Kindi see P. TRAVAGLIA, Magic, causality and intentionality. The doctrine of rays in
al-Kindi, Turnhout 1999, which incidentally contains a most useful bibliography and list
of works for al-Kindi.
21. Compare remarks made by Ptolemy in his Tetrabiblios, III.11, IV.10. In general
the Tetrabiblios seems to stand behind both al-Kindi’s and Abu Ma{sar’s works on astrol-
ogy. For a comparison between Ptolemy and the MudÌal see C. BURNETT, «The Certi-
tude of Astrology», forthcoming in a Festschrift in honor of David King.
 ABU MA{SAR, AL-KINDI 257

849-850 or thereafter22. Proximate Agent Cause would have been

composed earlier, either during the reign of al-Ma}mun (reigned
813-833) or more likely al-Mu{taÒim (reigned 833-842) — more
likely because it seems to be addressed to Mu{taÒim’s son AÌmad, for
whom al-Kindi produced numerous treatises.)
The heating of matter and the elements, then, serves as the physical
explanation of how the stars affect the lower world. Yet Abu Ma{sar has
more to say about what one might call the metaphysical aspects of
astral causation. As we have seen, Abu Ma{sar thinks that all properties
and events in the sublunar realm are due to the stars. As an Aris-
totelian, he divides these effects into two groups: substantial forms and
accidental forms. The stars relate to these two types of form in rather
different ways. In the case of substantial forms, the stars have only an
indirect influence, because they can only exercise causality by produc-
ing heat, cold, moisture and dryness. Since the four elements are com-
posed from these four contraries (fire is hot and dry, water is cold and
moist, etc.) all possible elemental combinations can be brought about
by the stars. Substantial forms emerge from these combinations, a
point that Abu Ma{sar makes by distinguishing between «natures», i.e.
the elements (fire, air, etc.), and what is «natured», i.e. composite sub-
stances (a man, horse, etc.). The stars are the cause that «natures the
natured from the natures and composes the composed, and separates
the species of animals, plants and mineral from each other» (MudÌal
I.4 [5] L 541-2). In this way the stars can bring about the production
of individual members of a species, such as a given man23. This can
occur either by combining material causes of the same species (as a seed
for a plant) or through spontaneous generation (see MudÌal I.4 [18]
L 650ff.)24, but in both cases the stars are the cause of the mixture.

22. PINGREE (1970), p. 35.

23. Cf. Abu Ma{sar’s response to the anti-astrology objection that the stars affect only
species, and not individuals, at MudÌal I.5 [4-6] L 688ff. His response is that the indi-
viduals are composed of elements and the elements are directly affected by the stars, so
that the stars indirectly cause the generation and corruption of individuals in the manner
just described.
24. Interestingly, Abu Ma{sar criticizes Aristotle here, who he wrongly thinks denied
the possibility of spontaneous generation: «thus there is a refutation of what he [sc. pre-
sumably Aristotle] said: that nothing comes to be except from something of its own
genus» (MudÌal I.4 [14] L 661-2). Assuming that he does here mean Aristotle, we can
infer that Abu Ma{sar was ignorant of the work On the Generation of Animals, the main
source for Aristotle’s views on spontaneous generation.
258 RECHERCHES DE THÉOLOGIE ET PHILOSOPHIE MÉDIÉVALES

Once the substantial forms have emerged, however, it is they and

not the stars that explain substantial properties. But there is a role
for astral causation even at this stage. Abu Ma{sar argues that we
can discern the effect produced directly by the heavens using a
process of elimination: «we know that whatever is in something,
but not as a result of what is proper to it [i.e. because of its form],
results in natured things from the power of the movements of the
stars, by the permission of God» (MudÌal I.4 [10] L 571-2, cf. I.4
[13] L 590-1: «what is not due to the form and nature is undoubt-
edly due to… the celestial power»). What is left for the stars to do,
if they do not case “proper” or essential properties? They (directly)
cause the accidents that belong to a particular substance, such as
hot, cold, healthy and sick (MudÌal I.4 [13] L 592). These acci-
dents (a{ra∂) are, says Abu Ma{sar, distinguished by the fact that
they admit of opposites as well as increase and decrease, and by the
fact that the same substance can partake of the accident and its
contrary at different times25.
An interesting corollary of this theory is that the stars are respon-
sible for all that differentiates individuals within the same species
from one another. For, reasons Abu Ma{sar, any property that an
individual has from its substantial form will belong to everything
else in its species, so there must be some other cause of differentia-
tion between the individuals. The heavens are the cause for this dif-
ferentiation, indeed the sole cause, which underscores the fact that
all accidents are the result of astral influence (see MudÌal I.4 [14]
L 602-3, I.4 [17] 645-9). Notice that he assumes the principle of
the Identity of Indiscernibles, perhaps inherited from the Stoics,
which states that two distinct individuals must have at least one
different property: «we see qualities and properties in every individ-
ual that are not in another of the same species» (MudÌal I.4 [18]
L 650-1).
Abu Ma{sar believes, then, that both substantial forms and acci-
dents are brought about by the stars. While he does not explain his
theory of substantial forms and accidents as explicitly as he might,
Abu Ma{sar’s discussion develops themes that are barely touched

25. Here Abu Ma{sar’s source seems to be ARISTOTLE, Categories chapter 5, 3b25ff.
and 4a10ff.
 ABU MA{SAR, AL-KINDI 259

upon by al-Kindi. What we do find in the Kindian corpus seems

compatible with Abu Ma{sar’s theory. Certainly, as we have already
seen, al-Kindi agrees that the stars influence this world by mixing the
elements in varying proportions. He also thinks that substantial
forms emerge from this process: «[the stars] are the cause of our gen-
eration and the agent that makes us, that is, our species» (On the
Bowing of the Outermost Body AR 255.4, but following the text at
RJ 192). But he does not seem to distinguish between the stars’ pro-
duction of substantial forms and their production of accidents, con-
centrating instead, especially in Proximate Agent Cause, only on the
processes of generation and corruption.
A rare point on which al-Kindi and Abu Ma{sar actually disagree
is the question of whether the stars affect the lower world directly.
Though I have contrasted the stars’ direct causation of accidents with
their indirect production of substantial forms, for Abu Ma{sar astral
causation is never, strictly speaking, «direct». This is because he
defines direct action as involving physical contact, like wood being
burned by fire (MudÌal I.3 [7] L 449-450). In this case there is
no intermediary (mutawassi†) between the agent and the patient.
The stars do not have «direct» action on earthly things in this sense,
because they are at a distance and bring about their effects through
an intermediary, namely the elements. He compares the action of the
stars to that of a magnet (MudÌal I.3 [7] L 459-470). This notion of
action at a distance can be found in al-Kindi as well, in On the
Bowing of the Outermost Body (AR 249.10ff. [RJ 185]), where he also
uses the example of a magnet. But in Proximate Agent Cause he con-
trasts the action of the stars, which is «proximate», to that of God,
Who is the «remote» cause of sublunar affairs (AR 219), apparently
affirming the direct relationship between stars and the lower world
that Abu Ma{sar denies.
Still, the difference of opinion here may be less than it seems,
since al-Kindi is contrasting astral causation to divine causation,
whereas Abu Ma{sar is contrasting the action of the stars to the
action of sublunar things upon one another. It is also worth not-
ing that Abu Ma{sar’s terminology is thoroughly Kindian: the ter-
minology of «mediating» causes (tawassu†) is found prominently
in his brief epistle on the same issue of the difference between
mediated divine causation and immediate causation between
260 RECHERCHES DE THÉOLOGIE ET PHILOSOPHIE MÉDIÉVALES

effects26. Furthermore, it seems likely that both have in mind the

contemporary debate over mediated actions among authors of the
Kalam. Theologians had raised the question of whether what they
called «engendered» acts (tawallud), such as an arrow’s striking its
target, are in any way the act of the orginal engendering agent, in
this case the archer. This example of arrow and archer is in fact
used by al-Kindi at Proximate Agent Cause 219, and he affirms that
both archer and arrow are agent causes27. Abu Ma{sar gives the
more generic example of something’s being thrown, which is a
«voluntary act (fi{l bi-iradatihi)» on the part of the thrower:
«there results from his movement of the thing thrown a certain act
in the thing thrown» (MudÌal I.3 [7] L 454-455). Thus on the
essential philosophical issue Abu Ma{sar and al-Kindi do agree:
they both hold that mediated agency remains agency, so that a
cause is still responsible for its indirect effects28.

3. Providence, determinism and freedom

As we have seen there is, according to Abu Ma{sar and al-Kindi,

an unbroken chain of causation flowing from God, through the

26. Al-fa{il al-Ìaqq al-awwal al-tamm wa }l-fa{il al-naqiÒ alla∂i huwa bi-}l-magaz (On
the True, First Complete Agent and the Deficient Agent that is Metaphorically [an Agent]),
AR 182-184.
27. For a survey of Mu{tazilite and other positions on this question see H. DAIBER, Das
theologisch-philosophische System des Mu{ammar Ibn {Abbad as-Sulami (Beiruter Texte und Stu-
dien 19), Beirut 1975, pp. 367ff. For further discussion of the issue in relation to al-Kindi
see my «Al-Kindi and the Mu{tazila», in: Arabic Sciences and Philosophy 13 (2003), pp. 1-34.
28. In fact Abu Ma{sar’s position is a bit subtler than I have sketched it here: he dis-
tinguishes between two types of indirect acts, one through a proximate medium (striking
something with a thrown object), and another through a remote medium (the magnet).
Astral causality is of the latter sort. It seems however that the general point that engen-
dered acts are caused by the original agent would hold true for either sort of indirect act,
however. His examples and terminology suggest that he just means to contrast the case
where A gives an act to B (thrower to stone), to the case where A gives an act to B which
gives a further act to C (magnet to air to iron; or stars to elements to particular bodies).
If one holds that A and B are both genuine agent causes in the first case, one will pre-
sumably hold the same for A, B and C in the second. Al-Kindi also uses the magnet
example but highlights a different feature of magnetic action. For him it is an example of
a particular type of agent causality, in which the agent engenders an act in its effect with-
out itself being in motion; thus he compares it to love, for the beloved can cause motion
in the lover without moving him or herself.
 ABU MA{SAR, AL-KINDI 261

heavenly bodies, to the earthly realm. It follows that astrology is in

fact the study of divine Providence, since it is through astral causa-
tion that God’s will is brought to fruition29. In fact al-Kindi says that
inquiry into the stars’ influence on generation and corruption can
show «how universal Providence (tadbir) [is achieved] through prior,
divine wisdom» (Proximate Agent Cause AR 219.10-11)30. Although
this is an advantageous aspect of the theory, it carries with it an obvi-
ous difficulty. If the stars cause all sublunar events, including human
actions, does this negate human freedom?
Abu Ma{sar squarely confronts this objection in the fifth chapter
of the MudÌal’s first treatise, a chapter that is devoted to refuting var-
ious objections against the science of astrology. The third objection
reads as follows:
MudÌal I.5 [7] L 725ff.: The third category are [critics] among the adher-
ents of tradition and rational speculation. They reject the science of the
judgments of the stars and say that the planets do not have an indication
over things that come to be in this world. Their argument for this is their
claim that the stars do not indicate the possible.

It is not immediately clear who Abu Ma{sar has in mind when he

refers to ahl al-Ìadi† wa }l-naÂam, which I have translated «the
adherents of tradition and rational speculation»31. Their objection
against astrology is that if the stars «indicate» sublunar events, then
the stars must necessitate those events. But then all that happens in
the sublunar realm will be necessary, which is taken here to be an
absurd consequence.
29. R. Lemay’s discussion of the MudÌal is misleading on this point, insofar as he
presumes that «Providence» must mean irregular intervention by God in natural affairs:
«Abu Ma{sar’s cosmological system leaves as little room for the intervention of Providence
as that of Aristotle, perhaps even less, if we consider the intervention of the planets and
heavenly bodies in general as an added safeguard against the unfettered will of God»,
LEMAY (1962), p. 114. Far from presenting an obstacle to divine Providence, for Abu
Ma{sar the celestial bodies are in fact the instrument of that Providence. However
Lemay’s point is a valid one if we are worried, not about Providence generally, but about
the possibility of miracles. It does not seem that al-Kindi or Abu Ma{sar leave any room
in their universe, governed as it is by strict naturalistic causality, for the miraculous in the
sense of a departure from natural laws caused directly by God.
30. Similarly, at the end of the epistle he says the celestial bodies are the cause «that
gives order through the will (irada) of the Creator of this order (tartib), which is the rea-
son for generation and corruption» (Proximate Agent Cause 236.15-16).
31. Other manuscripts have ahl al-Ìadi† wa }l-gadal, «adherents of tradition and dis-
putation».
262 RECHERCHES DE THÉOLOGIE ET PHILOSOPHIE MÉDIÉVALES

Who would pose such an argument against astrology? The phrase

ahl al-Ìadi† wa }l-naÂam suggests that Abu Ma{sar has in mind con-
temporary mutakallimun, who used rational argument in the service
of Islamic theology. If this is right then the objectors must be the
Mu{tazila, who against other groups rejected determinism and
upheld the existence of human freedom. In support of this identifi-
cation is evidence that the Mu{tazila criticized astrology. A leading
Mu{tazilite, Abu }l-Hudhayl, is said to have given a demonstration of
the impossibility of astrology before the caliph al-Ma}mun32. Most
intriguingly, in one report he refutes an astrologer by appealing pre-
cisely to the astrologer’s inability to predict human actions33.
This sets the context for Abu Ma{sar’s response to the objection,
which is among the most philosophically interesting passages in the
MudÌal. He agrees with the objector that it would be absurd for all
sublunar events to be necessary. Indeed, he goes so far as to argue
against another group of philosophers: determinists, who reject the
existence of the possible or contingent (al-mumkin), leaving only the
necessary and the impossible. He maintains, appealing to Aristotle as
an authority, that there must be possibility in addition to necessity
and impossibility, because «our actions are possible» (MudÌal I.5
[10] L 754). In this Abu Ma{sar is signalling his basic agreement with
the Mu{tazila34. He affirms that human beings are unique among
creatures in having freedom of choice (iÌtiyar): «the power (quwwa)
by which [man] chooses one thing instead of another by the deliber-
ation (fikra) belongs only to man, and not to any other animal»
(MudÌal I.5 [18] L 739-40). Other animals act «by nature» rather
than by choice, because choice requires a rational soul (MudÌal I.5
[21] L 870-1; see below for full quotation). That humans are free is
proved by the fact that they deliberate and take advice when they are
trying to make a decision, which they would not do if the choice
were not up to them (MudÌal I.5 [12-13] L 778ff.).

32. See Josef VAN ESS, Theologie und Gesellschaft um 2. und 3. Jahrhundert Hidschra
Berlin 1993, Bd. V, texts XXI 74-75.
33. Ibid., text XXI 76, which comes from Ibn Murta∂a.
34. Given the context, the group of determinists Abu Ma{sar has in mind here are
probably those mentioned by Aristotle in Physics II.4. However it is natural to connect
the passage to the determinism of those who upheld absolute predestination in Islam, the
so-called Jabarites, who held that humans do all things under compulsion (gabr) from
God.
 ABU MA{SAR, AL-KINDI 263

Al-Kindi, too, affirms the existence of freedom among creatures,

though he disagrees with Abu Ma{sar by ascribing free choice to the
stars as well as to humans, an ascription that Abu Ma{sar explicitly
denies (MudÌal I.5 [21] L 872; see also I.3 [11] L 506: the sun does
not arrive over a place «from its own choice (iÌtiyar)»). Indeed one
of al-Kindi’s few explicit discussions of freedom comes in the context
of his explanation of the fact that the stars «bow down» before God,
which he interprets as meaning that they are «obedient» to Him.
Something can only be obedient, he points out, if it has freedom of
choice; like Abu Ma{sar he ties the possibility of choice to the pos-
session of a rational soul (On the Bowing of the Outermost Body AR
246.8-9 [RJ 180])35. The presence of a definition of free choice, this
time in the human case, in al-Kindi’s collection of definitions for
philosophical technical terms also gives us some reason to think that
he agreed with the Mu{tazila that creatures retain their freedom,
despite the universality of divine power and Providence36.

35. Again, the disagreement between the two is less than it seems, because Abu
Ma{sar does agree that the stars have rational souls (MudÌal I.5 [21] L 872). He thinks
they lack choice simply because the purpose of choice is to avoid harm, and the stars are
never confronted with harm. Here al-Kindi is the one whose position is more compati-
ble with Aristotle, since in the Aristotelian system the heavens rotate out of desire to
emulate God, which seems to imply a certain degree of choice or will. The significance
of this passage for al-Kindi’s views on human freedom have been pointed out by
T.-A. DRUART in «Al-Kindi’s Ethics», in: Review of Metaphysics 47 (1993), pp. 329-357,
at 345-346.
It may also be worth mentioning in this context the question of whether the stars
choose to bring about, and are therefore responsible for, evils that come about in the sub-
lunar realm. This is a question raised by Plotinus in Enneads IV.4, and his discussion of
it was known to al-Kindi via the sixth chapter of the Theology of Aristotle. On this ques-
tion Abu Ma{sar agrees with the author of the Theology, who writes: «no blameworthy
thing at all comes from the heavenly world to the earthly world, and the planetary lords
are not a cause for any of these evils that exist here, because they do not act through will
(bi-irada)» (A. BADAWI, Plotinus apud Arabes, Cairo 1955, 75). See further MudÌal I.3
[11] L 501ff, where Abu Ma{sar explicitly addresses the problem of whether the stars will
their effects.
36. Fi Ìudud al-asya} wa rusumiha (On the Definitions and Descriptions of Things), AR
165-179, at AR 167.1. The definition of iÌtiyar («a willing (irada) preceded by reflection
(rawiyya) together with discrimination») may come ultimately from Andronicus of
Rhodes: see F. KLEIN-FRANKE, «Al-Kindi’s On Definitions and Descriptions of Things»,
in: Le Muséon: Revue des Études Orientales 95 (1982), pp. 191-216, at 202. I have further
discussed al-Kindi’s relationship to the Mu{tazila on the question of human freedom in
my forthcoming paper, «Al-Kindi and the Mu{tazila», where I further defend the con-
clusion (see below) that al-Kindi was a compatibilist.
264 RECHERCHES DE THÉOLOGIE ET PHILOSOPHIE MÉDIÉVALES

How, then, does Abu Ma{sar reconcile his belief in human free-
dom with the certainty of astrological predictions? We might expect
that Abu Ma{sar would simply make an exception for human
actions, exempting them from the causal influence of the stars. This
would in effect be to give in to the Mu{tazilite criticism of astrology
by restricting its scope to events uncaused by human agency. But in
fact the position he takes is quite the opposite:
MudÌal I.5 [20] L 860-2 (my emphasis): Just as the stars indicate the possi-
bility and choice that belong to a man, so they indicate that a man will only
choose what the stars indicate, because his choice of a thing or its opposite will
be by the rational soul whose mixture with the animal soul in individuals is
determined by the indications of the stars.

We could hardly ask for a more explicit affirmation of determinism

— indeed the passage is so provocative that a medieval scribe wrote
a warning next to it in the Latin version37. The stars do, then, deter-
mine our actions, even though we exercise «choice». Much as we saw
in the case of the generation of substances, the actions and the dis-
positions of the soul emerge from material combinations brought
about by the heavens. Al-Kindi agrees with this theory of the soul’s
actions, for he says «it is found that the actions of the soul follow on
the mixtures of bodies, and the mixtures vary with the variation of
the stars» (Proximate Agent Cause 224.19-225.1).
In addition to this causal sort of determinism, there also looms
another kind of determinism, first described by Aristotle at De inter-
pretatione 19 in his celebrated example of the sea battle. Let us sup-
pose, modifying the example to fit this astrological context, that the
stars «indicate» that there will be a sea battle tomorrow, and that the
astrologer is able to determine with certainty that they so indicate.
If this is the case, then whether or not the stars actually cause the
sea battle to come about, then the sea battle must necessarily occur.
For the astrologer cannot know today that there will be a sea battle
tomorrow unless the battle will in fact occur. Abu Ma{sar, quite pos-
sibly with De interpretatione in mind, raises the problem as well and
puts it in the mouth of his imaginary objectors: «if someone were to
say today that something will happen tomorrow, and if that thing

37. «Cave, hic sermo durus est». See LEMAY (1962), 129. One of the two Latin trans-
lators, Hermann of Carinthia, simply omitted the passage.
 ABU MA{SAR, AL-KINDI 265

were to happen tomorrow, it will have happened because its occur-

rence is necessary» (MudÌal I.5 [9] L 742-3). Abu Ma{sar follows
Aristotle here and insists that there is a class of events that do occur
and are possible, not necessary38.
Abu Ma{sar’s position, then, seems to be as follows: human actions
are the result of a power of free choice (quwwa and iÌtiyar being the
key terms here), yet they are also determined by astral causation. This
would have struck his contemporaries among the mutakallimun as
outright self-contradiction, since for both the Mu{tazila and their
opponents freedom requires the lack of an external determinant
cause39. But nowadays we think of this as a perfectly respectable posi-
tion, known as compatibilism: the view that free will and determin-
ism are compatible. Of course Abu Ma{sar does not explicitly
endorse compatibilism in the abstract, but taking him to be a com-
patibilist is the only way to make sense of his dual support of the
universality of astral causality and human freedom40.
There is an obvious objection to this interpretation, namely that
as we have just seen Abu Ma{sar is at pains to show that human
actions fall into the class of the «possible» (al-mumkin), whereas

38. Unfortunately the recent publication of a relevant article on Abu Ma{sar came too
late for me to consider its findings in this paper. The article confirms the link between
Abu Ma{sar’s discussion of possibility and Aristotle’s De interpretatione: Carmela BAF-
FIONI, «Una citazione di De Interpretatione, 9 in Abu Ma{sar?», in: Aristotele e Alessandro
di Afrodisia nella tradizione araba, edited by Cristina D’ANCONA and Giuseppe SERRA,
Padua 2002, pp. 113-132.
39. For example, some Mu{tazila distinguish between «acts» (al-}af{al) and deter-
mined events (musabbabat, i.e. outcomes of }asbab, or causes). See R.M. FRANK, «The
Autonomy of the Human Agent in the Teaching of {Abd al-Jabbar», in: Le Muséon 95
(1982), pp. 323-355, at 324-5. Again I refer the reader to my article «Al-Kindi and the
Mu{tazila» for further discussion of this issue. There I defend the claim, which I merely
take for granted here, that the Mu{tazila were united by their incompatibilism even
though they disagreed with one another on many other points.
40. This approach should be contrasted to the highly critical, and less satisfactory,
interpretations of previous scholars, who take Abu Ma{sar to task for inconsistently
upholding both freedom and determinism. Lemay for example complains that «the omi-
nous inclination toward elimination of contingency so characteristic of Arabian philoso-
phy is already at play in Abu Ma{sar’s desire for harmonizing Aristotle’s cosmology and
the conclusions of astrological science», adding that given the «unlimited scope» of celes-
tial causation «man’s free will appears drastically curtailed», LEMAY (1962), pp. 125-6.
We need not agree with these conclusions if we take him to be a compatibilist; indeed as
we shall see immediately below Abu Ma{sar is quite keen to defend the existence of con-
tingency as well as human freedom.
266 RECHERCHES DE THÉOLOGIE ET PHILOSOPHIE MÉDIÉVALES

most compatibilists would be happy to say that human actions are

«necessary» within a given causal framework. For example, my
deciding to consult an astrologer is necessary in the sense that, given
the causal conditions that obtain (my upbringing, my having been
convinced of the efficacy of astrology by reading the MudÌal, etc.),
it cannot be otherwise than that I consult the astrologer. Is this not
precisely what Abu Ma{sar means to deny by saying that our actions
are possible?
The apparent inconsistency dissolves when we examine more
closely what Abu Ma{sar means by «possible». His notion of possi-
bility emerges from the following, typically thorough explanation of
modality:
MudÌal I.5 [10] L 753-66: The necessary and the impossible are known at
[all] three times [sc. past, present and future] through the necessity and
impossibility in their natures. But our actions are different from this,
because they are possible. For example, our knowledge that the sun was
shining in the past, is shining now, and will shine in the future… Like-
wise, if we say that fire, air, water and earth have been and will only be as
they are now, it is known that this is true for [all] three times. This, then,
is the first situation [sc. necessity]. As for the situation of impossibility,
this is like our saying that a man was flying, a man is flying now, and it is
possible that he will fly in the future… Our actions, on the other hand,
are not like this, because if a man says «I was doing good in the past and
I am doing good now», he is not able to say «I shall without doubt do
good in the future»… Since a man does not know what he will wish to do
with a knowledge in which there is no doubt, this is not necessary
(i∂†iraran), but possible.

This passage shows that Abu Ma{sar is working with what has been
called the «statistical» understanding of modality: the necessary is
simply what obtains at all times, the impossible is what never obtains,
and the possible is what obtains at some times but not at others41.
He shares this conception with al-Kindi, as is clear from the latter’s
definitions of necessity, possibility and impossibility: «Necessary (al-
wagib): that which is always in act in that to which it is ascribed. Pos-
sible (al-mumkin): that which is sometimes in potency and sometimes

41. For an illuminating discussion of the role this and other conceptions of
modality play in later Arabic philosophy, see T. KUKKONEN, «Possible Worlds in the
Tahafut al-Tahafut», in: Journal of the History of Philosophy 38 (2000), pp. 329-347 and
479-502.
 ABU MA{SAR, AL-KINDI 267

in act in that to which it is ascribed. Impossible (al-mumtani{): that

which is never in act or in potency in that to which it is ascribed»42.
Given that Abu Ma{sar is operating with this notion of possibility,
there is no difficulty in his holding that human actions are possible
(not necessary), and at the same time being a compatibilist. For he
can still say that human choices could not have been otherwise,
given that they were determined, without holding that they are nec-
essary: they are «possible» because they only happen some of the
time. He gives an example of this in the passage just cited: since a
given man will be good only some of the time, his being good at any
one time is merely possible. As he says a bit further on, «our actions
are not equal within the species [i.e., we do not act either well or
badly all the time by virtue of being human], but are changing from
good to evil and evil to good at different times… they are, therefore,
possible» (MudÌal I.5 [11] L 774-7).
This cannot be the whole story, however, since we have not yet
seen how Abu Ma{sar would distinguish between a possible event
that is a free choice and a possible event that is not, such as rainfall43.
Here, like many present day compatibilists, Abu Ma{sar appeals to
the notion that a free action is one that is preceded by deliberation
(fikra) and rational thought (see MudÌal I.5 [15] L 811-15, I.5 [18]
L 835). Indeed this is why he insists that only a creature with a ratio-
nal soul can exercise choice: «choice, which happens through delib-
eration on things, is found only in man, and not in any other animal,
because man has a rational soul with which he deliberates on a thing
or its opposite, and then chooses one of the two alternatives»
(MudÌal I.5 [21] L 868-70). His conception of freedom has to do
with the nature of a free action — that it should stem from rational
consideration of one’s options — not with whether or not the action
is ultimately caused by an external agent.

42. These definitions are from his epistle On the Definitions and Descriptions of
Things. See the text provided in KLEIN-FRANKE (1982), p. 211 (the relevant definitions
are absent from the text in AR but are attested by other manuscripts).
43. That he does want to draw such a distinction is made clear at MudÌal I.5 [15]
L 809-815, where he contrasts two kinds of possibility, one of a body to receive different
affections (such as water’s possibility to be heated), and the other stemming from free
choice. There is perhaps some equivocation here, and throughout this section of the
MudÌal, between «possibility» and «potentiality», encouraged by the equivocal term
quwwa (compare the Greek dunamis).
268 RECHERCHES DE THÉOLOGIE ET PHILOSOPHIE MÉDIÉVALES

Though al-Kindi has far less to say on this subject than Abu Ma{sar,
at least in his surviving works44, his scattered remarks on the nature of
freedom suggest that he, too, was a compatibilist. Here I will mention
only his characterization of the free choice of the heavens themselves,
in a passage (mentioned in passing several times above) where he
explains in what sense the celestial bodies are «obedient»:
On the Bowing of the Outermost Sphere, AR 246.7-247.8 [RJ 180-1]:
The meaning of «obedience» is «execution of the order of a commander».
Now, execution of the order of a commander is only by choice (iÌtiyar), and
choice belongs to complete souls, that is, rational [souls]. Therefore the stars
are endowed with obedience… For they follow a course of motion with no
alteration, and this has been existent for them in the preceding periods of
time up until the present… Therefore they are clearly obedient to what their
Creator wills, great be His praise, namely an act through which their great
essences last as long as is decreed for them, and through which the motions
decreed for them last.

While it is clear from this passage that al-Kindi wants to ascribe free-
dom to the stars, it is not quite clear whether his background assump-
tions are compatibilist or incompatibilist. The former seems much
more likely, however: presumably, al-Kindi does not believe that the
stars could have chosen to move in some other fashion. What he means
is that they voluntarily, though necessarily, submit to God’s will.
The reason it is a voluntary submission is that it is chosen ratio-
nally, or as he says a bit further on, the celestial body «is alive and dis-
cerning (mumayyiz), so it is clear that its obedience is due to choice
(iÌtiyariyya)» (AR 246.10 [RJ 181]). The stars’ freedom, like human
freedom in Abu Ma{sar, has nothing to do with the absence of exter-
nal causation or restraint, and everything to do with the nature of the
choosing. Indeed, if we bring into consideration other texts by al-
Kindi, it becomes clear that his conception of the stars’ freely chosen
motion is compatibilist. For example, his work On the True Agent45
says that only God is an agent in the proper sense, namely that His

44. We have in the Fihrist (255 ff.) a list of titles of works by al-Kindi, some of which
might have shed light on this subject, e.g. «On the Examination of the Statement Claim-
ing that Natural Things Perform Only One Act by the Necessity of [their] Innate Nature
[or: Creation]» (K. fi bă qawl al-mudda{i anna al-asya} †ab{iyya taf{alu fi{lan waÌidan
bi-}igab al-Ìilqa) and «On Ability» (R. fi }l-isti†a{a). See G.N. ATIYEH, Al-Kindi: the
Philosopher of the Arabs, Rawalpindi 1966, pp. 152-206.
45. See above, footnote 26.
 ABU MA{SAR, AL-KINDI 269

acts are not caused by another agent. If al-Kindi were an incompati-

bilist, the position he takes here would require him to say, with the
critics of the Mu{tazila, that God is the only free agent. The fact that
he ascribes free choice to the stars suggests that he is instead working
with a conception of freedom like Abu Ma{sar’s, on which something
can be determined towards an action and still choose it freely.
A further point worth making about the passage just cited is that al-
Kindi’s view is perhaps in tension with the requirement that chosen
actions be merely possible, where «possible» means happening only
some of the time. For, as he says explicitly here, the heavens choose to
obey God through their motions for the entire period of their exis-
tence. On the statistical model of modality affirmed by al-Kindi as well
as Abu Ma{sar this means that their obedience is necessary, not possi-
ble. The only way to avoid this conclusion would be to invoke the fact
that the stars only exist for a limited period of time, given al-Kindi’s
rejection of the world’s eternity. So nothing they do is «necessary» in
the sense that it «always» obtains. By the same token of course noth-
ing in the created world would be «necessary» in the strict sense of
being eternal. This would essentially nullify the contrast Abu Ma{sar
wants to make between «necessary» things that always occur, e.g. fire’s
being hot, and «possible» things that happen only sometimes. But of
course the issue presents no difficulty for Abu Ma{sar who, as we saw
above, does not think that the stars act out of free choice.
Admittedly Abu Ma{sar’s discussion of freedom plays a relatively
small role in the MudÌal. Yet it shows him taking up a position that
is quite distinctive in the context of mainstream Islamic discussions
of freedom, which as already mentioned took place within an incom-
patibilist framework. The departure of Abu Ma{sar and al-Kindi
from this framework is almost certainly due to their extensive
engagement with translations of ancient philosophical works. In this
case their compatibilism may well have its ultimate inspiration in
Stoicism rather than in Aristotle, given that the late ancient Peripatetic
tradition had interpreted Aristotle as an incompatibilist46. But this
question of sources should not distract us from the philosophical
originality of the MudÌal, which lies in its use of various ideas culled
46. See R.W. SHARPLES, Alexander of Aphrodisias on Fate, London 1983, pp. 21-23.
Alexander’s On Fate was available in the Arabic tradition: see H.-J. RULAND, Die arabis-
che Fassungen zwei Schriften des Alexander von Aphrodisias, Saarbrücken 1976.
270 RECHERCHES DE THÉOLOGIE ET PHILOSOPHIE MÉDIÉVALES

from what we would consider more traditionally «philosophical»

contexts, reworked into a defense of astrology47. Indeed, in this brief
examination of the beginning of the MudÌal, we have seen Abu
Ma{sar produce innovative arguments in the fields of epistemology
(astrology as a science, using Aristotelian metaphysics as a model),
physics (the stars as causes of elemental combinations), and meta-
physics (the production of substantial forms and accidents, as well as
the issue of free will). Mainstream philosophy in Islam, as repre-
sented by al-Farabi and Ibn Sina, would within a few generations go
on to reject astrology. But we can see that, just as the introduction of
Aristotelianism into the Latin world was partly thanks to the transla-
tion of astrological texts48, so at this early stage astrology had a sig-
nificant part to play in the Arabic reception and adaptation of Greek
thought49.

King’s College London P. ADAMSON

Strand
UK-London, WC2R 2LS
47. Of course this task was made much easier for al-Kindi and Abu Ma{sar by the
interest of Neoplatonists in astrology, magic, theurgy, and so on. Here one immediately
thinks of the hypothesis that the Îarranians from whom Abu Ma{sar and al-Kindi
apparently took at least some of their astrological doctrines represent a tradition originat-
ing with the residence of the Neoplatonist commentator Simplicius in Îarran. This
hypothesis has, however, been thrown into considerable doubt recently: see J. LAMEER,
«From Alexandria to Baghad: Reflections on the Genesis of a Problematical Tradition»,
in G. ENDRESS and R. KRUK (edd.), The Ancient Tradition in Christian and Islamic
Hellenism, Leiden 1997, pp. 181-191. There is a wealth of literature on the historical
accounts surrounding the transmission of Greek science from Alexandria to Baghdad; a
recent and important study is D. GUTAS, «The Alexandria to Baghgad Complex of
Narratives: a Contribution to the Study of Philosophical and Medical Historiography
among the Arabs», in: Documenti e Studi sulla Tradizione Filosofica Medievale 10 (1999),
pp. 155-193. Even if we dismiss the role of Îarran in this tradition, there can be no
doubt that plentiful Greek sources, whose influence we understand more clearly, encour-
aged the study of the stars (Plato, Aristotle, Ptolemy, etc.) as well as the belief that the
stars affect sublunar events, as would have been clear from Arabic versions of De caelo, the
Enneads, and so on.
48. In addition to many works by Charles Burnett, see LEMAY (1962).
49. I would like to thank Charles Burnett for first bringing the philosophical interest
of the MudÌal to my attention, and also all the members of our Arabic reading group
at the Warburg Institute in London (Catarina Belo, Charles Burnett, Moya Carey,
Muhammad Ilkhani, David Juste, Sayira Malik, Peter Pormann, Sajjad Rizvi, and Sophia
Vasalou) for stimulating discussion of the text. I have benefited greatly from comments
by Prof. Burnett on previous versions of this paper, as well as from comments by the
anonymous referees at this journal.