Avicenna's Metaphysics

in Context

Robert Wisnovsky

Cornell University Press

Ithaca, New York

 © 2003 by Robert Wisnovsky

All rights reserved. Except for brief

quotations in a review, this book, or parts
thereof, must not be reproduced in any form
without permission in writing from the publisher.
For information address Cornell University Press,
Sage House, 512 East State Street, Ithaca, New York 14850.

First published 2003 by Cornell University Press.

Library of Congress Cataloging-in-Publication Data

Wisnovsky, Robert, 1964-

Avicenna's metaphysics in context I Robert Wisnovsky.
p. cm.
Includes bibliographical references (p. ) and indexes.
ISBN 0-8014-4178-1 (cloth)
1. Avicenna, 980-1037-Contributions in metaphysics. 2. Metaphysics.
I. Title.

B751.Z7W57 2003
181'.5-dc21
2002041521

Printed in Great Britain

For Laila
 Contents

Acknowledgements vii

A Note on Transliteration and Citation ix

Introduction

PART I. AVICENNA AND THE AMMONIAN SYNTHESIS

1. Aristotle
Perfection in the Definitions of the Soul and of Change 21

2. Alexander and Themistius

Attempts at Reconciliation 43

3. Proclus, Ammonius and Asclepius

The Neoplatonic Turn to Causation 61

4. Proclus, Ammonius and Philoponus

Neoplatonic Perfection and Aristotelian Soul 79

5. Greek into Arabic

The Greco-Arabic Translations and the Early Arabic Philosophers 99

6. Avicenna on Perfection and the Soul

The Issue of Separability 113

PART II. THE BEGINNINGS OF THE AVICENNIAN SYNTHESIS

7. Essence and Existence (A)

Materials f rom the Kalam and al-Farabi 145

8. Essence and Existence (B)

Shay'iyya or Sababiyya? 161

9. Essence and Existence (C)

The Question of Evolution 173

10. Causal Self-Sufficiency vs. Causal Productivity 181

VI Contents
11 . Necessity and Possibility (A)
Materials from the Arabic Aristotle 197

12. Necessity and Possibility (B)

Materials from al-Farabi 219

13. Necessity and Possibility (C)

Materials from the Kalam 227

14. Necessity and Possibility (D)

The Question of Evolution 245

Conclusion 265

Appendix I
Tables of Greco-Arabic Translation 269

Appendix II
Transcriptions of Lemmata from MS Uppsala Or. 364 277

Bibliography 279

Index of Lemmata 293

General Index 297

 Acknowledgements

So many people have helped me reach the stage of publishing a book about
Avicenna's metaphysics that it is difficult to know how to thank them all. I
suppose I should just describe my intellectual trajectory from its beginning.
Elizabeth Fine taught me Greek in high school, and introduced me to the
beautiful strangeness of ancient texts. When I was a sophomore at Yale, Dimitri
Gutas convinced me to enroll in first-year Arabic, and then, perhaps feeling
responsible, was kind enough not to fail me. At Yale Abbas Amanat introduced
me to Islamic history, and Gerhard Bowering supervised my senior thesis on the
Ikhwan a~-~ara". A postgraduate year at Oxford studying with Fritz
Zimmermann showed me how serious medieval Arabic thought could be. Even
though this book preserves not a single sentence from my Princeton Ph.D.
dissertation, "Avicenna on final causality'', I owe a great debt to my doctoral
supervisor, Hossein Modarressi: he guided me gently through the most difficult
Avicennian texts, corrected my translations with care and suggested further
avenues of inquiry. At Princeton Sarah Waterlow Broadie was the second reader
of my dissertation, and Fadlou Shehadi and Parviz Morewedge were my external
examiners. Also at Princeton, Stephen Menn introduced me to medieval
philosophy, Michael Cook introduced me to kaliim, and Jim Morris introduced
me to Sufi thought; and courses on Aristotle's psychology and natural
philosophy with, respectively, Myles Burnyeat and Pierre Pellegrin showed me
the high level of philological and philosophical rigor which scholars of ancient
philosophy demand of themselves. The world of the late-antique commentators
on Aristotle opened up when Richard Sorabji hired me as his postdoctoral
research assistant at King's College London. Through my work there I was
introduced to Duckworth's editor, Deborah Blake, who has waited patiently for
this book to ripen.
Wolfhart Heinrichs, Bill Graham, Bill Granara, Roy Mottahedeh, Bashi Sabra,
Muhsin Mahdi, John Murdoch, Ayman El-Desouky and Wheeler Thackston have
all been supportive and stimulating colleagues here at Harvard. In particular, my
good friend and colleague Carl Pearson and I have had countless conversations
about the history of Aristotelian and Neoplatonic philosophy, and I learned
something important from every single one of them. Not only Carl but Sylvia
Berryman, Ian Crystal, Richard Sorabji, Victor Caston, Everett Rowson,
Therese-Anne Druart, Michael Marmura, Dimitri Gutas and Jules Janssens each
spent much of their precious time reading and criticizing earlier versions of this
work. Above all others, however, Stephen Menn has served as a paradigmatic
mentor: patient and generous, but also severe and exacting. He knows how much
I owe him.
viii Acknowledgements
I am also grateful to many other friends, colleagues and students who have
guided me on particular issues and pointed me towards useful and interesting
materials; these include Rahim Acar, Amos Bertolacci, Allan Gotthelf, Dag
Hasse, Angela Jaffray, Yaron Klein, Christian Lange, Craig Martin, Sait
Ozervarli, Amy Proferes, Ahmed al-Rahim, David Reisman, Will Robins, Jim
Robinson, Joe Saleh, Jeff Spurr, Abe Stone, Moshe Taube and Hikrnet Yaman.
Hussein Rashid helped me enormously with the logistics of producing camera-
ready copy.
My twelve-year-old son Simon Wisnovsky was born at the same time I started
to wrestle with Avicenna's Iliihiyyiit. My own experience, combined with that
of his beloved Red Sox, should teach him that it is good simply to persevere. I
am strengthened by the love I receive from Simon and from my two-year-old
daughter Jasmine Parsons, as well as from my parents Mary and Joe Wisnovsky,
from my brother Peter Wisnovsky, and from my in-laws Sheila, Emma and
Anne Parsons. What I cannot even begin to express is my gratitude to my wife
Laila Parsons. A historian of the modem Middle East, she started out skeptical
of my project's value. After many long rigmaroles together, and also some acute
editing on her part, she has now come to believe in its significance, and this has
meant more to me than anyone else's opinion. I dedicate this book to her.
 A Note on Transliteration and Citation

Throughout this volume individual Arabic and Persian letters have been
transliterated according to the system adopted by the International Journal of
Middle Eastern Studies. However, I have retained the sun letters when
transliterating the Arabic definite article, and elided the hamzat al-wa~l, and for
everything other than titles of books or individual names and terms, I have fully
vocalized the transliterated text; thus wa-s-sababu Ji dhalika, but also
Aris!ii!alis, ar-Tablca. The Greek long vowels have been transliterated as e and
o, and the iota subscript as -ei, -oi and -ai.
References to Plato's and Aristotle's works follow the standard page and line
numbering of Stephanus and Bekker, respectively. When abbreviating their
works I have adhered to the system adopted by the Ancient Commentators on
Aristotle Project; thus Aristotle's Posterior Analytics is abbreviated as An.
Post., and Plato's Parmenides is abbreviated as Parm. Citations of the Greek
commentaries on Aristotle's works all refer to the editions contained in the
Commentaria in Aristotelem Graeca (CAG) (Berlin, 1882-1909), and follow
their page and line numbering. My bibliography contains the particular editions I
used when citing the Greek commentaries on Plato's works. When I refer to a
commentary I follow the standard procedure of using the Latin preposition in
before the abbreviated title of the work commented on; thus the commentary by
Syrianus on Aristotle's Metaphysics appears as Syrianus, in Metaph., and the
commentary by Proclus on Plato's Timaeus appears as Proclus, in Tim.
 Introduction

My aim in this book is to present a history of the metaphysics of Abu c Ali: al-
I:Iusayn ibn c Abdallah ibn Sina (born before 980 AD - died I037 AD), known in
the West by his Latinized name Avicenna.
Since 1937, when Amelie-Marie Goichon published IA distinction de
{'essence et de {'existence d'apres lbn Stna (Avicenne), no serious book-
length study specifically devoted to Avicenna's metaphysics has appeared. This
is surprising enough given how influential Avicenna's metaphysical ideas were,
but what makes it astonishing is that tremendous advances have taken place
since Goichon' s time in fields relevant to this topic. Those fields include - from
the ultimate to the proximate, to use Avicenna's terms - the study of late-
antique Greek philosophy and the study of classical Islamic doctrinal theology,
or kalam; the study of the Greco-Arabic philosophical translations and the study
of the philosophy of al-Farabi; and the study of Avicenna's metaphysics and the
study of his intellectual biography. I shall briefly review those advances in the
order just described.

Two changes stand out in fields outside of, but contiguous to, Arabic
philosophy. The first is the increasing accessibility of Greek philosophical texts
from the late-antique period, the period that immediately preceded the Greco-
Arabic translation movement and the birth of Arabic philosophy. Although
many of these works have been available in their original Greek since the end of
the nineteenth century, when the Teubner editions of the texts of Proclus and
other Neoplatonists appeared, and when the Royal Prussian Academy published
the 24-volume Commentaria in Aristotelem Graeca, it is only fairly recently
that many of those texts have been translated, with notes and commentary, into
modem European languages. I am thinking in particular of the Bude series of
translations of the Neoplatonists into French, and the Ancient Commentators on
Aristotle series of translations of the CAG into English. These translations, as
well as the interpretive work that accompanies the translations and follows in
their wake, have enabled even those scholars of Arabic philosophy who do know
Greek to gain a vista of the whole late-antique philosophical landscape before
venturing into the dark hollow of a Neoplatonic treatise or commentary.
The second major advance has been the increasing availability of kalam texts
from the century or so immediately preceding Avicenna's birth. In the period
when Goichon wrote her book those interested in classical Islamic theology were
forced to rely on Islamic doxographies from later periods. These doxographies
often reflect post-A vicennian concerns and their value as sources for pre-
Avicennian kalam history is correspondingly limited. Since Goichon's time
many kalam texts from the tenth century have been edited, and the often
2 Introduction

fragmentary evidence of kaltim developments in the eighth and ninth centuries

has now been collated and translated into German. 1 Even though what we have
available to us is still only a small fraction of the literary output of the classical
mutakallimun, we are much better able than those in Goichon' s time to
construct an outline of the state of kaliim in the period before Avicenna.
In the field of Arabic philosophy itself two important advances have taken
place since Goichon wrote her book. The first is the large number of editions
now available of the Arabic translations from the Greek, be they of Aristotle's
texts or of texts wrongly attributed to Aristotle. Although some of these editions
are more reliable than others, they nevertheless provide a wealth of evidence of
how the translators went about rendering often obscure Greek terms into Arabic.
The word choices made by different translators in different places at different
times were philosophically significant because subsequent Arabic philosophers
scrutinized these texts with great care. A Greco-Arabic Lexicon is currently under
preparation, and several fascicles have already appeared. 2 Once completed, the
Lexicon will help scholars gain an even more precise understanding of the
history of the translations.
The second important advance in the field of Arabic philosophy has been the
increasing availability of the philosophical works of al-Farabi (d. 950). Avicenna
referred to al-Farabi as the Second Teacher, the first being Aristotle, and scholars
have known for a long time that Avicenna was influenced by al-Farabi's
thought, either directly, through reading al-Farabi's works themselves, or
indirectly, through reading texts written by philosophers from the generation
between al-Fara.bi's death and Avicenna's birth. With more of al-Farabi's texts
now edited, scholars can determine with greater precision than Goichon could
which philosophical items Avicenna borrowed from al-Farabi and which ones he
chose to leave behind.
Within Avicenna studies there have also been two important advances. Not
only are edited versions of many more of Avicenna's texts now available than in
Goichon's time, but individual scholars have been working on particular
problems of Avicenna's metaphysics, presenting the results of their research in
the form of articles in journals and collections, or as chapters in books which
cover other philosophers in addition to Avicenna. Goichon was breaking new
ground simply by basing her analysis on Avicenna's Arabic texts instead of their
Latin translations. Now, close readings of brief passages or even chapters that
bear on particular metaphysical topics have enabled scholars to paint detailed
pictures of important aspects of Avicenna's metaphysics. 3
The other advance that has been made in Avicenna studies is also a result of
the far greater number of Avicenna's texts now available in edited form. Scholars

1 van Ess 1991-1997.

2 Endress and Gutas 1992-.
) e.g., Marmura 1979, 1981, !984a, 1984b and 1992; and Jolivet 1984 and 1991.
 Introduction 3
of Avicenna's thought are now able to establish a fairly precise intellectual
biography by collating those newly available texts with what Avicenna wrote
about his life and career in his autobiography and in other works, and with what
his student al-Jlizjani wrote about him. In the past fifteen years or so, efforts to
date particular texts and to periodize Avicenna's career have intensified, aIXl
scholars can now appeal to evidence of intellectual change from an earlier text to
a later one, in order to explain what once seemed to be perplexing
inconsistencies in Avicenna's thought. What is more, the close scholarly
attention paid to Avicenna's intellectual biography means that we understand
much better than before how Avicenna viewed his place in the history of
philosophy. 4 It is now clear that Avicenna saw himself as the heir to a long
tradition of Aristotelianism, contrary to the claim by some modem interpreters
that mystical vision was the engine that drove his thought.

This is all, in a way, to help the reader locate me in my context, that is, the
scholarly context in which this book appears. But of course I mean more by
context than that, because my aim is to build upon all the scholarly advances I
just described, with a view towards locating Avicenna's metaphysics in its own
context more precisely than Goichon was able to do 65 years ago. I shall first
give a summary of the contents of the book and then say how I think I have
accomplished this goal.
My book revolves around the answers Avicenna and his predecessors gave to
two pairs of questions: what is the soul and how is it related to the body as its
cause, and what is God and how is He related to the world as its cause? The two
pairs of questions are similar because any claims I make about how the soul
causes the body will inevitably shape how I define the soul, just as any claims I
make about how God causes the world will inevitably shape how I define God.
In other words, my challenge is to come up with a definition of the soul which
satisfies both my intuitions about what the soul is in and of itself, as well as
my intuitions about what the soul is relative to its effect, the body. Similarly,
my challenge is to come up with a definition of God which satisfies both my
intuitions about what God is in and of Himself, as well as my intuitions about
what God is relative to His effect, the world.
In order to respond to these challenges, Avicenna and his predecessors invented
new concepts and distinctions and reinterpreted old ones. Part I, "Avicenna aIXl
the Ammonian Synthesis", is about the complicated history of one such
concept, entelekheia, a Greek term often translated as "actuality". Aristotle
invented the word and used it most prominently to define the soul (psukhe) as
well as to define change (kinesis). (Strictly speaking, Aristotle's definitions of
the soul and change are really descriptions, not definitions, but that need not
concern us here.) The soul, Aristotle claims, is the first entelekheia of a natural

4 Gutas 1988.
4 Introduction

instrumental body possessing life potentially. Change, he says, is the

entelekheia of the potential insofar as it is potential. It is not immediately clear
what Aristotle meant when he defined the soul and change in these ways. One of
the main reasons for this is that the concept of entelekheia is so
underdetermined. If I am unsure what exactly an entelekheia is, I will not be
able to get a good sense of what a soul is or what change is. Most modem
commentators on Aristotle agree that the best way to make sense of Aristotle's
definitions of the soul and of change is to treat his uses of entelekheia in each
case individually. They put to the side the question of what Aristotle thinks an
entelekheia might be, and concentrate instead on what Aristotle thinks the soul
is an entelekheia of, and on what Aristotle thinks change is an entelekheia of.
The rationale behind this approach is that it is more important to detennine
Aristotle's concept of the soul and his concept of change, than it is to determine
his concept of entelekheia.
The late-antique commentators were more reluctant than their modern
counterparts to give up on entelekheia. They felt that their main challenge was
to define what entelekheia meant in such a way that Aristotle's uses of it to
define the soul and to define change could be seen to be identical at some basic
metaphysical level. (Whether the underlying identity that binds together the two
uses of entelekheia is strong or weak was a subject of disagreement.) This was
a daunting task and occasionally the commentators fell back on the approach that
would eventually be taken up by modem scholars. Why should the
commentators have set themselves up for such frustration? They did this, I
believe, because their hermeneutical commitment to viewing Aristotle's
philosophy as systematic rather than aporetic was much stronger than that of
Aristotle's interpreters today. Aristotle had divided entelekheia into two types,
first and second, in his definition of the soul. According to most interpretations,
the first entelekheia refers to a capability to perform some function, while the
second entelekheia refers to the active exercise of that capability. The
commentators, committed to viewing Aristotle as a system-thinker rather than a
problem-thinker, felt an obligation to interpret Aristotle's definition of change in
light of the first entelekheia!second entelekheia distinction: to decide whether
change was a first or second entelekheia, and once that was done, to figure out
what the remaining entelekheia applied to.
In Chapter 2 I survey the attempts by Alexander of Aphrodisias (fl. 205 AD)
and Themistius (fl. 365 AD) to deal with the issue of how consistent Aristotle's
use of entelekheia to define the soul is with his use of entelekheia to define
change. Particularly significant, to my mind, are Alexander's promotion of the
term teleiotes - completeness - as a gloss of entelekheia, and his appeals to
Aristotle's discussion of what it is to be teleion, complete, in order to come up
with a list of meanings for teleiotes that might then be transferred to
entelekheia. I then argue that Themistius invents a new concept - I have called
it "endedness" - which he adds to the semantic range of teleiotes. In other
 Introduction 5

words, Themistius wants to interpret teleiotes in such a way that it refers not
only to being teleion, or complete, but also to the various ways in which a
thing can be directed towards or serve as a telos, or end.
In Chapters 3 and 4, I describe the attempts by Neoplatonic Aristotle-
commentators from Syrianus (d. ca. 437 AD) to Philoponus (d. ca. 570 AD) -
but whose central figure was Ammonius (d. ca. 521 AD), son of Hermeias - to
add yet another new understanding of teleiotes. In addition to completeness and
endedness, teleiotes was also held by the Neoplatonists to mean perfection. The
Neoplatonists used teleiotes, now understood as perfection, to refer in a strict
sense to superiority in the cosmic hierarchy. And since everything in the world
was seen to be striving for perfection, the Neoplatonists also used teleiotes to
refer to the goal or final cause of each thing's upward reversion (epistrophe)
towards the well-being (to eu einai) that was peculiar to its species.
Concurrently, the Neoplatonic commentators promoted a new understanding of
to teleion as the perfect. The perfect was seen by the Neoplatonists to refer not
so much to something which is complete in a quantitative or qualitative way, or
to something which is ended in the sense of being directed towards or serving as
a telos, but instead to something which is full of existence. Because a thing
which is full of existence can pass the existence it receives from what is above it
in the cosmic hierarchy directly down to what is below it in the cosmic
hierarchy, the Neoplatonists used the perfect to refer to the efficient cause of the
existence (to einai) each thing received from above in the downward procession
(proodos) of being.
By this stage it becomes clear that teleiotes, a term which had originally been
fashioned to solve a problem of Aristotelian exegesis - squaring Aristotle's use
of entelekheia to define the soul with his use of entelekheia to define change -
had accrued so many new meanings that it was now an important concept in its
own right; and that its usefulness in solving that particular exegetical problem
was beginning to take a back seat to its usefulness in other contexts. In
particular, teleiotes and to teleion were now being used by the Neoplatonists to
integrate their metaphysics of reversion and procession with the Aristotelian
concepts of final causation and efficient causation. Partly as a result of this
development, discussions of causation came to be seen by the Neoplatonists to
fall into the domain of metaphysics rather than into the domain of natural
philosophy, as Aristotle had thought.
I argue in Chapters 3 and 4 that this new understanding of perfection and the
perfect had profound consequences for the way the Neoplatonists responded to the
metaphysical questions and challenges I outlined above. In Chapter 3, I discuss
how the Neoplatonists' concepts of perfection and the perfect shaped their answer
to the question, what is God and how is He related to the world? The
Neoplatonists were in agreement that God should be described, as far as one was
able, as One: as utterly simple and unique. But if, relative to the world, God is
also seen to be both the original source of procession, and the ultimate goal of
6 Introduction
reversion, His combination of efficient causality and final causality would entail
duality where there ought to be only unity. I argue that despite their vigorous
efforts the Neoplatonists were unable to resolve this dilemma satisfactorily.
In a similar way I discuss in Chapter 4 how the Neoplatonists' new
understanding of perfection and the perfect played out in their discussions of the
soul and its relationship to the body. Since Alexander's time teleiotes had been
the term commentators turned to when they wished to gloss Aristotle's use of
entelekheia to define the soul. Yet now that teleiotes was seen by the
Neoplatonists to be related to its effect as a final cause, the soul, as a teleiotes,
should be seen to be related to the body as its final cause. Seeing the soul as the
final cause of the body was not at all difficult to reconcile with Aristotle's
conception of the soul, since he himself had said that the soul was related to the
body as its final cause as well as its formal and efficient cause.
But what the Neoplatonists meant by final cause was significantly different
from what Aristotle had meant by final cause. In particular, the Neoplatonists
had divided Aristotle's four causes - the efficient, the final, the formal and the
material - into causes which are separate from or transcend their effects, and
causes which are inseparable from or immanent in their effects. The
Neoplatonists reckoned that the efficient and final causes transcend their effects,
whereas the formal and material causes are immanent in their effects. The result
is that if the correct way to understand the soul's being an entelekheia is to see
the soul as a teleiotes; if a teleiotes is related to its effect as a final cause; if
therefore the soul is related to the body as a final cause; and if final causes are
separate from their effects; then I can claim that the soul is separate from the
body.
For a Neoplatonist this is a welcome conclusion, since a separate soul is far
more consistent than an inseparable one with the Platonic picture of the soul as
pre-existing the body's birth and surviving the body's death. But Neoplatonists
whose job it was to comment on texts where Aristotle discusses the soul faced
the exegetical challenge of integrating their understanding of the soul as a
teleiotes, along with all its happy consequences, into Aristotle's descriptions of
the soul, according to which - by Peripatetic accounts, at least - the soul is not
separate from the body. I argue that the solutions some Neoplatonic
commentators came up with to face this challenge are sophisticated and in some
cases quite justifiable readings of Aristotle's theory of the soul.
In Chapter 5 I examine how all these Greek terms - entelekheia, teleion,
telos and teleiotes - were translated into Arabic in the ninth and tenth centuries.
The tendency among scholars has been to view the translations as a disruption in
the history of philosophy: before the translations there was Greek philosophy,
and after the translations there was Arabic (or Islamic) philosophy. Of course
there is no denying that the translations did cause a certain amount of disruption.
In fact translation movements have often forced important conceptual changes to
take place in the history of philosophy. In the case of the Greco-Arabic
 Introduction 7

translation movement, the ways the translators chose to render a particular Greek
term or phrase into Arabic sometimes exposed new ways for an Arabic
philosopher to conceptualize old problems which had stumped his Greek
predecessors, sometimes flushed out problems which had been hidden to his
Greek predecessors, and sometimes created entirely new problems which his
Greek predecessors had never been forced to confront.
All this is true. But I argue that we can also detect a great deal of
philosophical continuity between the translators and their immediate
predecessors. One piece of evidence of this continuity is the fact that two distinct
concepts in Aristotle's philosophy - actuality (entelekheia) and end (telos) -
were conflated in the Arabic version of his Metaphysics and appear together as
perfection (tamiim). In other words, a Neoplatonic premise - that an
entelekheia, since it is a teleiotes, is related to its effect as a final cause - is
embedded in the very wording of the Arabic version of an Aristotelian text. To
my mind this shows that the translators, particularly those active in the early
stages of the translation movement, did not stand outside of the history of
philosophy but were an integral part of the Neoplatonic continuum. I end
Chapter 5 by discussing the philosophical impact the translators' word choices
had on the writings of early Arabic philosophers such as al-Farabi.
In Chapter 6 I examine the ways in which Avicenna stood as heir to this long
series of interpretations of perfection and the perfect. In particular, I consider
Avicenna's application of the concept of perfection to the soul, and - as in
Chapter 4 - focus on the role perfection plays in his discussions of the soul's
causal relationship with the body. Avicenna follows the later Greco-Arabic
translators in using kamiil rather than tamiim when defining the soul.
Nevertheless, he uses kamiil and tamiim interchangeably in other contexts, and in
most cases, I argue, he understands kamiil and tamiim in a Neoplatonic sense,
that is, in the way that a Neoplatonist understood teleiotes, and not in the way
that Aristotle understood entelekheia. This meant that Avicenna, like the
Neoplatonists, understood that the manner in which kamiil and tamiim related to
their effects was as final causes; and in addition to the term al- cillatu l-
ghii' iyya, which he takes from the later translators, Avicenna refers to the final
cause as al- cillatu t-tamiimiyya and al- cillatu l-kamiiliyya.
Avicenna also accepts and promotes the Neoplatonic distinction between the
efficient and final causes, which are separate from their effects, and the formal
and material causes, which are inseparable from their effects. I argue that
Avicenna's clear acceptance of the Neoplatonic premise that the final cause is
separate from its effect was crucial to his argument that the human rational soul
survives the death of the body it originated in, and that in this respect his theory
is similar to that of Neoplatonic commentators such as Philoponus, whose
theories I described in Chapter 4.
I also argue that one other striking piece of evidence of Avicenna's
Neop!atonic inheritance can be detected in the different ways he distinguishes
8 Introduction

between first and second perfections. In some contexts, the model he follows is
that of Aristotle, holding that the first perfection refers to a capability to perform
some function, while the second perfection refers to the active exercise of that
capability. At other times, Avicenna says that the first perfection should be
understood as that which is necessary for something to exist, while the second
perfection should be understood as that which is necessary for something to exist
well. To my mind this latter method of distinguishing between first and second
perfections is clearly an echo of the Neoplatonists' distinction between existence
(to einai) and well-being (to eu einai), the former passed down from above in
the course of procession (proodos), the latter striven for from below in the
course of reversion (epistrophe). I argue that Avicenna saw no fundamental
discrepancy between the two ways he distinguished between first and second
perfections, but that if pushed, he would regard the Neoplatonic method (i.e.,
necessary to exist/necessary to exist well) as basic, and the Aristotelian method
(i.e., capability to perform some function/exercise of that capability) as
derivative. This suggests that Avicenna's cosmology is best understood as being
structured along Neoplatonic lines of procession and reversion.

In Part II, 'The Beginnings of the Avicennian Synthesis", I begin by focusing

on the role played by perfection and the perfect in more general discussions of
causal relations. One of the many challenges facing an interpreter of Aristotle's
teleology is determining how the final cause operates in tandem with the other
three causes: the formal, the material and the efficient. Sometimes Aristotle
simply identifies the final and formal causes and claims that they operate in
tandem with the material cause. In organic things the form is the end and the end
is the form. It is for the sake of the form "cow'', for example, that the materials
that make up a cow - flesh, blood, bone, sinew - come together. Those material
elements, left on their own, would not make up a cow, Aristotle reckons; it is
only the form that explains their coming together. (Yet not every end is a form.
The Unmoved Mover, for example, is an end, but it is not a form, at least
according to most interpreters.)
In other contexts Aristotle says that the final cause operates in tandem with
the efficient and not the material cause. Health is the cause of hard work, he
says, and hard work of health. But they are not causes of each other in identical
ways. Hard work is the efficient cause of health, and health is the final cause of
hard work. Among Neoplatonic commentators there was little debate over the
fact that the efficient was the cause that most deserved to operate in tandem with
the final. This was because, as mentioned already, both the efficient and the final
causes were seen to be separate from their effects, while the formal and the
material causes were seen to be inseparable from their effects. Since the
Neoplatonists felt strongly that something that was separate from matter was
superior to something that was inseparable from matter, it would therefore be
 Introduction 9
demeaning for the final cause to be thought of as operating in tandem with the
material cause.
Avicenna follows the lead of the Neoplatonists in regarding the efficient and
the final as the causes that operate in tandem with each other. And like the
Neoplatonists he saw as one of his primary challenges determining the precise
nature of the complementarity between efficient and final causes. Avicenna goes
to great length to show how the final cause and the efficient cause do not
compete over the same explanandum but instead explain different explananda in
a complementary way. The final cause, Avicenna says, has explanatory priority
with respect to essence, while the efficient cause has explanatory priority with
respect to existence.
I shall use Aristotle's example to explain what Avicenna means. Yesterday I
worked hard in my garden with a view to attaining a state of health; with a view
to becoming healthier, in other words. The idea of health was in my mind
before I started working hard, because the idea of health was what got me off the
sofa and out into the garden in the first place. But health only came into
existence concretely- in my body, that is - after I had finished the hard work,
because the hard work was what caused my body to be in a state of health. Put in
Avicenna's terms, when health acts as a final cause, when it is conceived in the
mind as a universal idea or essence to be striven for, it comes before the
efficient cause, hard work. But the health that comes into concrete existence in
my body comes after the efficient cause, hard work. The final cause, health, is
prior to the efficient cause, hard work, with respect to essence, because the idea
or essence of health comes before its effect, hard work. But the efficient cause,
hard work, is prior to the final cause, health, with respect to existence because
the existence of hard work comes before its effect, health.
Because Avicenna often uses the term shay,iyya, "thingness", for essence in
many of his discussions of how the final and efficient causes operate in tandem
with each other, I try to explain what he meant by that strange term. In Chapter
7, I focus on ninth- and tenth-century debates about what it is to be a thing, and
about how things are related to existents. These debates took place among
mutakallimun - that is, those engaged in formulating Islamic doctrine, or kaliim
- and philosophers such as al-Farabi discussed the issue as well. I then examine
Avicenna' s assertions about things and existents in contexts outside the problem
of final-efficient complementarity. I compare what Avicenna says in those
contexts with what the mutakallimun and al-Farii.bi say, and find that in some
important respects al-Farabi's position is closer to the Mu<tazilite
mutakallimun's position than it is to Avicenna's position; and that Avicenna's
position is closer to the position of the Ash<arite and Mii.turidite mutakallimun,
than it is to al-Fii.rii.bi's position. This is surprising, given our expectation that a
philosopher is much more likely to agree with another philosopher than he is to
agree with a mutakallim, and given our expectation that a mutakallim is much
10 Introduction

more likely to agree with another mutakallim than he is to agree with a

philosopher.
In examining those passages where Avicenna talks about thingness and
essence in contexts outside the problem of final-efficient complementarity, I list
Avicenna's most important assertions about how thingness and essence are
related to existence, and find those assertions to be incompatible. Sometimes
Avicenna claims that essence and existence are always found together, that one
always entails the other, and that they should be seen to be on an equal footing,
with neither prior to the other. Other times he seems to favor essence, implying
that it is logically prior to existence. And sometimes he even appears to
entertain the notion that there is a certain category of essence which is in no way
connected to existence. This implies that essence and existence might not always
be found together.
Next I tum to the passages where Avicenna talks about thingness and essence
in the context of final-efficient complementarity. Chapter 8 focuses on a
pressing philological question: whether, in those particular passages, the
instances where Avicenna appears to use the term shay'iyya, "thingness",
should be emended to read sababiyya, "causality", instead. I argue that the
weight of evidence compels us to retain shay'iyya. Chapter 9 attempts to
explain how Avicenna's application of the essence/existence distinction to the
problem of final-efficient complementarity arose partly from his attempt to solve
a specific problem of Aristotelian exegesis, namely, how to avoid circularity
when holding that health and hard work are the causes of each other.
To Chapter Ts list of assertions about how thingness and essence are related
to existence in passages outside of the context of final-efficient complementarity,
I then add the assertions Avicenna makes about how thingness and essence are
related to existence in passages devoted to the problem of final-efficient
complementarity. I find that while Avicenna's assertions remain incompatible,
there seems to be evidence of a development in his thinking. In works from his
middle period Avicenna uses the terms shay' iyya (thingness) as well as essence
(miihiyya), whereas in works from his late period he relies solely on miihiyya
(essence). I discuss the various reasons why Avicenna' s views on this issue
might have evolved. I argue that the most important reason was that with
miihiyya rather than shay'iyya Avicenna reckoned that he would have an easier
time both upholding the universal applicability of the final cause, and
establishing the final cause's primacy among the causes. This is because there
were reasons for Avicenna to think that miihiyya was itself more universally
applicable than shay'iyya; and because miihiyya, if seen to be logically prior to
existence, would provide a basis on which Avicenna could argue for the final
cause's primacy.
In Chapter 6 I concluded that important pieces of evidence suggest that
Avicenna's cosmology is best understood along Neoplatonic lines of procession
and reversion. In Chapter 9 I described how Avicenna used his distinction
 Introduction 11

between essence and existence to attack the problem of how final causes and
efficient causes operate in tandem. The way Avicenna went about treating this
issue suggested to me that he was partly motivated by a desire to resolve an
issue of Aristotelian exegesis: avoiding the trap of circularity he might fall into
when he held that the final and the efficient are causes of each other. In Chapter
IO I argue that Avicenna's attempts to solve the problem of final-efficient
complementarity by using his essence/existence distinction should also be
viewed in the broader context of his Neoplatonic cosmology of procession and
return.
As I described in Chapters 3 and 6, the mechanism of procession is built
around the idea that when something becomes perfect, when it has become so
full of existence that it can pass the existence it receives from above directly
down to what is below, the relation of the perfect thing to the imperfect thing
below it is the relation of an efficient cause to its effect. I argue that Avicenna's
assertion that the efficient cause is prior to the final in terms of existence, is
compatible with this premise.
The mechanism of reversion, by contrast, is built around the idea that the
perfection possessed by something higher in the cosmic hierarchy serves as a
goal for something lower than it in the cosmic hierarchy. The perfection
possessed by the higher thing acts as a final cause on the lower thing, in the
sense that the lower thing strives to attain the higher thing's perfection by
becoming as similar to the higher thing as it can be. Not so obvious is how this
idea might be compatible with Avicenna's assertion that the final cause is prior
to the efficient cause with respect to essence. Given that perfection and essence
can each be understood as referring to a thing's substantial form, then perfection
and essence will be identifiable at some level. It is by being as human as I can
be that I revert upwards. This is because my being as human as I can be entails
my being as rational as I can be, since rationality is what sets the species
"human" apart from other species of animal such as "cow". My being as rational
as I can be consists simply in my thinking about universal intelligibles as much
as I can; and by thinking about universal intelligibles as much as I can, I revert
upwards to the higher level of the Active Intellect, which is thinking about
universal intelligibles all the time.
The perfection possessed by the Active Intellect serves as my goal while I am
thinking about universarintelligibles, because I am not thinking about universal
intelligibles for their own sake. Rather, I am thinking about universal
intelligibles because by thinking about universal intelligibles I imitate the
Active Intellect and thereby come to possess, or at least participate in, the
perfection which the Active Intellect possesses. In the sense that it causes me to
fulfill my humanity, in the sense that it causes my substantial form or essence
to inhere in me to the greatest degree possible, the Active Intellect, as a final
cause, is prior, in terms of essence, to the efficient cause of my thinking about
universal intelligibles. This is because I - meaning my human rational soul -
12 Introduction

am the efficient cause of the preparatory activities that enable me to think about
universal intelligibles, and I would never have begun to serve as the efficient
cause of those activities in the first place if the goal of reversion had not been
there first.
With some effort, therefore, Avicenna can integrate his use of the
essence/existence distinction (inherited from the mutakallimun) to explain the
problem of final-efficient complementarity (inherited from Aristotle), into his
cosmology of return and procession (inherited from the Neoplatonists). But
Avicenna still faces the same theological problem I described in Chapter 3. In
Chapter 3 I argued that the Neoplatonists were unable to resolve the following
dilemma. On the one hand, God is utterly simple and unique. On the other hand,
God is both the original source of procession and the ultimate goal of reversion,
and God's combination of efficient causality and final causality entails duality in
Him rather than simplicity.
Avicenna tries to deal with this problem by arguing that God is perfect - that
is, that He possesses perfection - yet is at the same time above perfection. In
the sense that God possesses perfection God is related to the world as a final
cause. God's perfection acts as a goal for all other beings. What is more, if Gcx:l
possesses perfection and is therefore perfect, and if the perfect is related to what
is below it as an efficient cause, then God will also act as an efficient cause of
all other beings. God's final causality is therefore identical, at least
extensionally, with His efficient causality.
There is a flaw in this picture, however. The perfect acts as an efficient cause
of what is below it by being so full of existence that it can pass the existence it
receives from above directly down to what is below. Since there is nothing
above God, He cannot receive any existence from above and pass it down to the
Intellect below. Therefore Avicenna is forced to say that God does not just
possess perfection, He is above perfection. Being above perfection, by
Avicenna's reckoning, means having too much existence - so much so that it
spills over and starts the cascade of procession downwards. But Avicenna's
proposal - holding that God possesses perfection yet is also above perfection -
not only fails prima facie to solve the duality problem, it raises the problem of
consistency in his theology, and it flushes out yet another problem embedded in
the premise that God is both an efficient and a final cause.
The consistency problem is this: if Avicenna thinks of God as being above
perfection, then at least in one sense he seems to be holding God's efficient
causality to be prior to his final causality. Perhaps it even forces him to hold
that God is not a final cause at all, with the topmost of the celestial Intellects
now serving in that role. How are we then to reconcile this picture with
Avicenna' s many unambiguous statements that the final is the primary cause and
the noblest of the causes, that metaphysics consists in understanding the final
cause, and that God Himself serves in the role of ultimate final cause?
 Introduction 13
The second problem is that as the possessor of perfection, as the ultimate final
cause of reversion, God's quality of causal self-sufficiency takes center stage. As
that which is above perfection, as the original efficient cause of procession,
however, God's quality of causal productivity takes center stage. In the former
case God is seen as isolated and transcendent. In the latter case God is seen as
involved and creative. How are we then to reconcile these two aspects of the
divine, the transcendent and the self-sufficient on the one hand, and the involved
and productive on the other?
In Chapters 11 to 14 I describe how Avicenna went about devising a new way
to solve the problem of retaining both of these divine characteristics -
transcendent self-sufficiency and involved productivity - without implying any
unwanted duality in God. To do this Avicenna constructed what I call a matrix of
distinctions. The distinctions he used to build this matrix include the following:
the distinction between necessary existence and possible existence; the
distinction between in itself and through another; and the distinction between
uncaused and caused. Ultimately Avicenna would reduce this matrix of
distinctions to a single, overarching distinction: the distinction between an
uncaused necessary of existence in itself (wajib al-wujUd bi-dhiitihi), and a
caused necessary of existence through another (wajib al-wujud bi-ghayrihi).
Although Avicenna's matrix was new, many of the distinctions he used to
construct it were old. Chapters, 11, 12 and 13 are an inventory of the sources that
he drew from to come up with his matrix. In Chapter 11, I review what Aristotle
has to say about necessity, in the Arabic versions of Metaphysics 5.5, the
chapter of his philosophical lexicon where he distinguishes between different
ways of being necessary; Physics 2.9, where he discusses hypothetical necessity,
the kind of necessity that operates down here in the world of coming-to-be and
passing-away; Metaphysics 12.7, where he might be misunderstood as
advocating the idea that God is a necessary being, when in fact he does not; and
in De Interpretatione 12-13, where he discusses the relations between the modal
operators "necessarily" (or "necessary that"), possibly (or "possible that") and
"impossibly" (or "impossible that").
I argue that Avicenna owes a conceptual debt to Metaphysics 5.5, since he
takes from that chapter the idea that what is necessary can be seen to be self-
sufficient (that is, as not being dependent upon a cause); to be productive (that
is, as necessitating its effect); and to be produced (that is, as necessitated by its
cause). I also argue that Avicenna owes a terminological debt to De
Interpretatione 12-13, since in those chapters, unlike Metaphysics 5.5 (where
to anankaion - the necessary - is translated into Arabic as al-mu¢farr), the
translator chose to render anankaion einai - necessary that it be - as wajib an
yujada. Crucially, the Arabic translator of the De Interpretatione used the
existential verb w-j-d to render what in Greek was a merely copulative einai, I
end Chapter 11 by explaining how this translator's decision allowed Arabic
14 Introduction

interpreters such as Avicenna to think of possibility as contingency, that is, as

causedness.
In Chapter 12 I discuss how al-Farabi contributed to Avicenna's matrix of
distinctions. In his Commentary on the De lnterpretatione al-Farabi came up
with a distinction between mumkin al-wujild bi-nafsihi and wiijib al-wujild bi-
nafsihi. But this distinction was introduced by al-Farabi in a discussion of De
lnterpretatione 9's problem of future contingents, and of the effect that God's
foreknowledge of those future events might have on the question of determinism.
I argue that this is not the context in which Avicenna devises and employs his
matrix of distinctions, and that the influence of al-Farabi's distinction on
Avicenna was probably only terminological, and not conceptual.
I examine what the mutakallimiln have to say in Chapter 13. In contrast to al-
Farabi, who seems to have been most influenced by Mu'tazilism, Avicenna's
most important theological source was the Sunni kaliim of the Ash'arites and
Maturidites. In order to explain the nature of the eternality possessed by God and
His attributes, Sunni mutakallimiln moved away from earlier, pre-A vicennian
attempts to argue that when we define an eternal thing as "that whose existence
has no beginning in time" (mii la awwala li-wujildihi), what we really mean is
that an eternal thing is uncaused; to later, post-Avicennian attempts to argue
that when we define an eternal thing as "that which has never ceased to be nor
will ever cease to be" (mii lam yazal wa-lii yaziilu), what we really mean is that
an eternal thing cannot possibly not exist, and that therefore an eternal thing is
necessary of existence. This transformation was motivated by the Sunni
mutakallimiln's desire to reconcile their use of eternality in proofs of the world's
createdness, with their use of eternality in discussions of how the divine
attributes (:;ifiit) relate to the divine Self (dhat).
I argue that evidence from a number of texts from the last quarter of the tenth
century, as well as from one of Avicenna's earliest works, the lfikma
<Aril¢iyya, which he composed in 1001, shows that the separate influences of
the Arabic Aristotle, al-Farabi, and the mutakallimiln were combining to suggest
the notion of the necessary of existence in itself to a number of thinkers.
However, none of those other thinkers focused as intently as Avicenna did on
creating a coherent matrix of distinctions that supported that notion.
Chapter 14 addresses the question of evolution in the ways Avicenna
constructed his matrix of distinctions and the uses he put it to. As I just
mentioned, Avicenna's new matrix of distinctions is first articulated in his
/fikma cAril(iiyya, one of his earliest works, and I examine his discussion there
and in the Mabda' wa-ma<ad, another early work. I conclude by arguing that
the tension which one scholar has detected between the different ways in which
Avicenna applied the matrix of distinctions in the Shifii', Najiit and Jshiiriit -
works from his middle and late periods - is prefigured by a tension between the
different ways he applied the matrix of distinctions in the Ifikma '"Aril(iiyya and
in the Mabda' wa-maciid.
 Introduction 15

My treatment of Avicenna's metaphysics in this book is not definitive. Many

more books will be required before justice is done to a subject as complex, wide-
ranging and historically important as this. But I am still confident that I have
moved the study of Avicenna's philosophy forward. Apart from the particular
hypotheses argued for in each chapter, I hope by the end of the book to have
convinced the reader of the general idea that Avicenna's metaphysics is best
understood as bridging two periods in the history of philosophy. In one sense
Avicenna's metaphysics is the culmination of what I call the Ammonian
synthesis. By "Ammonian synthesis" I refer to the tendency, epitomized by the
Aristotle-commentator Ammonius and his students, to incorporate the larger
project of reconciling Plato's and Aristotle's philosophies (I call this the greater
sumphonia), into the smaller project of forging a single, consistent doctrine out
of the sometimes incongruent theories found in Aristotle's many treatises (I call
this the lesser sumphonia).
All of this integrating spurred the thinkers of the Ammonian synthesis to
invent new metaphysical tools, a crucial example of which is the concept of
perfection. As a consequence, Avicenna's appeal to kamal and Aristotle's appeal
to entelekheia in explaining the soul's relationship to the body are superficially
similar but profoundly different. To ignore perfection's complex history is to
ignore the context in which Avicenna's metaphysics was born and developed.
The lesson to be drawn is that further scholarship on other issues in Avicenna's
philosophy will benefit from direct comparison with parallel treatments of those
same issues by earlier philosophers of the Ammonian synthesis.
This point may seem underwhelming, if not obvious; but the tendency of
some scholars has been to compare Avicenna directly with Aristotle, skipping
(or treating only perfunctorily) the profound developments - particularly among
the Neoplatonists - that took place during the 1,300 years that separated the two
thinkers. And even when scholars do take Neoplatonism into account, their
attention has been so fixed on pseudo-Aristotelian works such as the Theology
of Aristotle, that whatever Neoplatonism is found in the philosophy of a thinker
such as Avicenna is assumed to be directly attributable to those notorious
treatises. One thing which my book shows is that Neoplatonism came to
philosophers such as Avicenna not so much through a funnel (viz., the
Theology of Aristotle and a few other pieces of pseudepigraphia) but through a
sieve (viz., the many Aristotle-commentaries by philosophers of the Ammonian
synthesis).
But what do I mean when I claim that Avicenna's metaphysics is the
culmination of the Ammonian synthesis? By 1001, the Aristotle-commentators
of the Ammonian synthesis had been so successful in folding the larger project
of reconciling Plato and Aristotle into the smaller project of reconciling
Aristotle with Aristotle, that to the 21-year-old Avicenna, then composing his
first philosophical summa, an Ammonian interpretation (most importantly, of
the human rational soul's separability) seemed like an entirely natural reading of
16 Introduction

Aristotle's texts. At the same time the young Avicenna grew up and lived in the
provinces - in Bukhara, at the periphery of the Islamic world - and was almost
entirely self-taught in philosophy. The 21-year-old prodigy thus found himself
confronted by all the philosophical problems entailed by the Ammonian
synthesis (most importantly, the conceptual duality that results from God's
operating as both efficient and final cause), yet without having first been
inculcated with the hermenutical commitments (most importantly, the need to
reconcile Plato and Aristotle) which had originally created those problems, and
which had still served to motivate his predecessor al-Fiirabi, a student of the
Aristotelian school in metropolitan Baghdad and the author of a treatise on the
harmony of Plato and Aristotle.
This freedom from hermeneutical commitments enabled Avicenna to attack the
old Ammonian problems by laying the groundwork for a new synthesis, an
Avicennian synthesis between the Islamic theology of the classical
mutakallimun and the philosophy propounded by the Arabic heirs (most
importantly, al-Fiirabi) to that earlier, Ammonian synthesis. The kalam
influence on Avicenna's distinction between essence and existence will be
obvious enough, I hope. And though the language which Avicenna used to
formulate his distinction between necessary of existence in itself and
necessary of existence through another was undoubtedly shaped by the Arabic
Aristotle, particularly in the digested form contained in al-Fiirabi's writings, his
attempt to apply that new distinction to the relationship between God and the
world, was prefigured by earlier kalam debates about how best to describe the
relationship between God and God's eternal attributes. The shelf-life of the
Avicennian synthesis was even longer than that of the Ammonian synthesis.
After some of Avicenna's works had been translated into Latin in Toledo in the
second half of the twelfth century, his metaphysical ideas came to exert profound
influence on Latin scholastic philosophy. More significantly, working out the
implications of Avicenna's metaphysical distinctions was to become the
predominant challenge of post-classical Islamic thought, particularly in the areas
of epistemology, cosmology and theology; and evidence of the impact of
Avicenna's metaphysics can be found in Arabic philosophical works from as late
as the end of the nineteenth century. The history of the Avicennian synthesis is
the subject of another book, however.

I ought to say a few words about the general approach I took when writing this
book. One of the worries I had in devoting so much attention to the Aristotelian
and Neoplatonic background to Avicenna's thought was that I risked being
accused of secret membership in what Mahdi once described as the "charmed
circle of source-hunters".5 The source-hunter, by Mahdi's reckoning, sees

5 Mahdi 1990, 700.

 Introduction 17

nothing original in the Arabic philosophical text he is reading, only worn coins,
vestiges of earlier (and greater) Greek thinkers.
My aim in devoting such a great chunk of the book to the Aristotelian and
Neoplatonic background to Avicenna's thought is not to prove that he had no
original thoughts, or to support the traditional Western notion that the Arabic
philosophers simply received Greek philosophy and then, three centuries later,
handed it back - slightly scuffed - to Europe, where it presumably belonged. On
the contrary, I hope that by describing as accurately and completely as I can the
intellectual context in which Avicenna was situated, a more precise picture of his
originality will emerge.
Sabra has spoken of the twin ills of reductionism and precursorism in the
study of the history of Arabic science, and I believe his analysis applies just as
acutely to the study of the history of Arabic philosophy. According to a
reductionist, as with a source-hunter, Arabic philosophy will appear to be
nothing but a "reflection, sometimes faded, sometimes bright or more or less
altered, of earlier (mostly Greek) examples".6 Precursorism is the opposite
tendency. A precursorist sees nothing but glorious originality in an Arabic
philosophical text, and traces later, supposedly original, European ideas back to
their true Arabic roots.
The middle road between reductionism and precursorism is contextualism, and
this is the approach I have tried to follow. More specifically, a contextualist
approach requires that Arabic philosophy not be viewed through the lens of
Greek philosophy or through the lens of medieval and early-modem European
philosophy. Instead, contextualism demands that I try as hard as I can to see
Arabic philosophy on its own terms.
In the case of a book on Avicenna's metaphysics, this means that my
translations of and citations to Greek sources will, wherever possible, take into
account the Greek sources' Arabic translations, since Avicenna read the Arabic
translations and not the Greek originals. It also means that in addition to
Aristotle I take into account the Greek commentaries on Aristotle, many of
which were also translated into Arabic, since Avicenna read commentaries as
well as the Aristotelian texts themselves. 7 What is more, I should take into
account the works of al-Farabi:, since Avicenna thought of him as his Second
Teacher. Finally, I should take into account works of kaliim, and particularly
tenth-century kaliim, since Avicenna studied some theology as a youngster. 8
Only after I have amassed the extant raw materials which Avicenna had before
him, can I begin the task of the historian of philosophy: reconstructing the
spectrum of philosophical options available to Avicenna on a given issue, and

6 Sabra 1987, 223-224.

7 wa-naiara Ji n-nu~iisi wa-sh-shuru~i: Ibn Khallikan, Wafaytit al-a'yiin wa-anbii' abnii' az-
zamtin 1(#182),420,10-1 I.
8 [qad] ~ji;.a ashyii'a min u~iili d-dini: Ibn Khallikan, Wafayiit al-a'yiin wa-anbii' abnii' az-

zaman I (#182), 420,4-5.

18 Introduction

explaining why he chose one option and not others. What I have found in
Avicenna's metaphysics is - to use Sabra's terms once again - an active
appropriation of his sources rather than a passive reception of them.
 Part I

Avicenna and the Ammonian Synthesis

 I. Aristotle
Peifection in the Definitions of the Soul
and of Change

I should begin Part I with a word of warning. I intend to offer not so much my
own interpretation of Aristotle's understanding of entelekheia, but rather a
survey of entelekheia-interpretations of philosophers from Alexander of
Aphrodisias to Avicenna. Although my hope is that Aristotle specialists will be
interested in reading about how issues relating to entelekheia played out in late-
antique and early-Arabic philosophy, I cannot say whether this will contribute to
their debates about what Aristotle really thought about those issues. My goal is
simply to begin contextualizing Avicenna's theory: to start proving, in other
words, that his understanding of perfection and the perfect, and his application of
those concepts to "hot" issues in late-antique and early-Arabic metaphysics, can
be seen as justifiable, and even compelling, readings of Aristotelian philosophy,
given the peculiar evolution of the interpretation-history which he was heir to,
and given the subtle conceptual shifts which resulted from the word choices
made by the Greco-Arabic translators. This is why I have gone into such detail
describing the theories of Avicenna's predecessors, and why I have based my own
translations, wherever possible, on the Arabic versions of his predecessors'
Greek works.

Aristotle invented the term entelekheia, sometimes translated into English as

"actuality", but also as "fulfillment", "perfection" and "completion". Aristotle
uses the term most often when he applies it adverbially in constructions such as
the dative entelekheiai and the prepositional phrase kat' entelekheian. When
used in this adverbial way entelekheia is roughly synonymous with energeia,
"activity", in its analogous constructions energeiai and kat' energeian. The
terms energeia and entelekheia are most interchangeable when they are placed
in opposition to the term dunamis, "potentiality", similarly applied in the dative
dunamei and in the prepositional phrase kata dunamin. In their adverbial forms
entelekheiai and energeiai mean something like "in actuality", while dunamei
means something like "in potentiality".
Aristotle uses the term entelekheia less often - but more interestingly - in a
nominal, as opposed to an adverbial sense. In other words, Aristotle not only
says that something is in a state of actuality (X is entelekheiai), but that
something is itself an actuality (X is entelekheia), in which case entelekheia
is used as a predicate, or - most rarely of all - that actuality is itself something
or other (entelekheia is X), in which case entelekheia is the subject. When
used in this nominal rather than adverbial way, entelekheia can be - and has
been - understood as referring not only to the state of being actual, but also to
22 Part I: The Ammonian Synthesis

the process of becoming actual. In other words, when used in the nominal
rather than adverbial way, entelekheia has been interpreted by some scholars as
referring not only to the state of actuality, but also to the process of
actualization.
Of these nominal uses of entelekheia, the two most prominent occur in De
Anima 2.1, where Aristotle defines what psukhe ("soul") is, and in Physics 3 .1,
where Aristotle defines what kinesis ("motion" or, more generally, "change") is.
In the De Anima translation incorrectly attributed to Isl;iaq ibn I:Iunayn and edited
by Badawi, Aristotle says:

Ll
Aris~ii~alis,Fi n-nafs, 29,9-30,4
Aristotle, DA 2.1, 412a6-28
It is held that substance is one of several classes of things, and
that one [type] is like matter, which in itself does not subsist as a
particular thing; another is the shape and form by which a thing is
made particular, such that it may be said to be this [thing, i.e., as
opposed to that thing]; and the third is the compound of these two.
Matter is a potentiality, and form is an "entelekheia'', meaning
perfection ffa-1-hayulii quwwatun mina 1-quwii wa-~-~uratun hiya
an{alakhiyii ya<nl t-tamiim = esti d'he men hule dunamis to d 'eidos
entelekheia], this [i.e., entelekheia] being of two types: the first
is like knowledge, and the other is like contemplating.
Now bodies are said to most warrant being substances, and
especially natural ones, for these are prior to the rest. Among
natural bodies are those which possess life, and those which do
not possess life, life referring simply to that which nourishes
itself, grows and decays. So every natural, living body will
doubtless be a composite substance. And since it is the body
which we describe as being such [i.e., that it has life in it], it
follows that it [body] will not be a soul, because body is not
predicated of a substrate, but instead is like substrate or matter.
The soul, then, must be substance as form of a natural body
which possesses life potentially. Substance is "entelekheia", and
"entelekheia" is of two types, the first being like knowledge,
because sleeping and waking come about only through the soul's
existence, and waking is equivalent to contemplating, whereas
sleeping is equivalent to possessing [the capability to think]
without activity [i.e., without actively using that capability]. As
far as coming-to-be is concerned, knowledge is [also] likely to be
prior. On account of this the soul is an "entelekheia'', that is, the
first perfection of a natural body possessing life potentially [min
ajli dhiilika ~iirati n-nafsu an{aliishiyii wa-huwa awwalu tamiimi
jirmin {abf<iyyin dhl flayiitin bi-l-quwwati = dio he psukhe estin
entefekheia he prate SOmatOS phusikOU dunamei zoen ekhontOS]. I

1 I address the problem of the Arabic De Anima translations in Chapter 5.

 I. Aristotle 23
I should remind readers that my English translations are, wherever possible,
drawn from the Arabic versions of the Greek texts, which is why I have rendered
tamiim here in Ll (and kamiil in L2) as "perfection" rather than as "actuality",
since "perfection" is the English word that most precisely evokes Avicenna's
understanding of those two Arabic terms. When discussing Aristotle's own uses
of entelekheia, however, I have more often than not simply transliterated that
Greek term; but I have also felt free to translate it variously - as actuality,
actualization, perfection or completion - in order to illustrate the ambiguities
and apparent inconsistencies which worried the Aristotle-commentators.
In any case, two issues can be distinguished already. The first is that Aristotle
seems to think that entelekheia is indicative of the type of substance a thing is.
Substance, according to Aristotle, may be viewed as matter, as form, or as the
composite of matter and form. In Ll entelekheia is clearly meant to be tied to
substance as form. More particularly, entelekheia seems intended to indicate the
respect in which substance as form may be contrasted with substance as matter:
just as matter is linked to a substance in its state of potentiality, so form is
linked to a substance in its state of actuality.
So far so good. The second issue, however, clouds this picture. Aristotle
thinks that entelekheia is divisible into two types, or modes - first entelekheia
and second entelekheia - and that the soul is a first entelekheia. One of the
examples Aristotle gives of first and second entelekheiai - knowledge
(episteme) and contemplating (theorein) - helps us, though not much, and the
whole distinction comes across as underdetermined. Aristotle has already said that
the soul is substance as form, and we are meant to see that as representing
entelekheia - actuality - in contrast to the potentiality represented by substance
as matter. Yet in the distinction between knowledge and thinking, it is
contemplating, not knowledge, which seems closer to representing actuality.
How then can Aristotle claim that the soul is an entelekheia because it
represents actuality rather than potentiality, and then claim that the soul is a first
entelekheia, which seems much closer to potentiality than it does to actuality?
Although scholars throughout the ages have offered various interpretations of
Aristotle's distinction, there is something close to a consensus now that when
Aristotle distinguishes between first and second entelekheiai, he means to refer
not so much to the difference between potentiality and actuality, but to the
difference between possessing (hexis) the capability to perform some function,
and using (khresis) or exercising that capability.2 The example Aristotle gives is
thus meant to be understood this way: when we are merely knowledgeable, that
is, when we possess the capability to contemplate but are not actively
contemplating, our souls are in a state of first actuality. When we are
contemplating, that is, when we are actively using our capability to

2 How entelekheia applies to the soul is treated by Ross 1961, 166-167 and 213; Hamlyn 1968,

82-85; and Ackrill 1972-73. Ackrill is criticized by Charlton 1980.

24 Part /: The Ammonian Synthesis

contemplate, our souls are in a state of second actuality. 3 As we shall see, one
reason why Aristotle is keen not to reduce the distinction between first and
second entelekheiai to a simple distinction between potentiality and actuality is
that he wants to avoid implying that the transitions from ignorance to
knowledge, and from knowledge to contemplating, represent changes, and
particularly alterations or changes of quality (alloioseis). Instead, the contrast
between ignorance and knowledge and knowledge and contemplating should be
seen as indicating the presence or absence of a relation, namely, the relation to
the object of thought. 4
Given that the transition from first entelekheia to second entelekheia ought
not to be seen as a change, it is a bit surprising that Aristotle should have
chosen to use entelekheia to define change. lsJ:iaq ibn I:Iunayn's translation of
Physics 3.1 has Aristotle saying that:

L2
Aris~iiFilis,
af-Tablca, 170,15-171,2
Aristotle, Phys. 3.1, 20la!0-12
So if each one of these types [of change, i.e., change of substance,
change of quality, change of quantity, and change of place] is
divisible into what is in [a state of] perfection and what is in [a
state of] potentiality, change will consist in the perfection of
what is in [a state of] potentiality in so far as it is such [wa-idha
qussima kullu wal.zidin min hlidhihi 1-ajnasi illi mli bi-l-kamlili wa-
mli bi-1-quwwati klinati 1-1.zarakatu kamlila mli [or, <wa-hiya>
kamlilu mli] bi-1-quwwati bi-mli huwa ka-dhiilika = dieiremenou de
kath' hekaston genos tou men entelekheiai tou te dunamei he tou
dunamei ontos entelekheia hei toiouton kinesis estin].

What is different about Aristotle's nominal use of entelekheia in L2 to define

change and his nominal use of entelekheia in L1 to define the soul? First of all,
Aristotle seems unconcerned here in L2 with showing how entelekheia might
refer to substance as form, as opposed to substance as matter. This is in contrast
to Ll, where an explicit connection is made between entelekheia and substance
as form. In fact, in Physics 3.1, 20la4, Aristotle does make a contrast using
form, but the contrast he makes is between form and privation (steresis), not
between form and matter (hule) as in Ll .5 What is more, this contrast (Physics
3.1, 20la4) is meant to illustrate the contraries between which one type of

3 Other places where Aristotle opposes episteme (knowledge) and theorein (contemplation) are

Metaph. 9.6, 1048a34 and 9.8, 1050al2-14; Phys. 8.4, 255a34; DA 2.5, 417a29; GA 2.1, 735all ;
andEN7.3, 1146b31-35.
4 Phys. 7.3, 247bl-248a7; at DA 2.5, 417b5-16 Aristotle allows for the possibility that the

transition from ignorant to knowledgeable might be a special sort of alloiosis.

5 It is true that the tenn Aristotle uses for form here at Physics 3.1, 20 I a4, is just morphe, most

often translated as "shape", and not morphe kai eidos as in LI; but in Physics 1.7-9 Aristotle often
contrasts eidos with steresis, so it is reasonable to interpret the contrast at 201a4 as likewise
indicating form as substance, and not some subsidiary notion of fonn as shape.
 I. Aristotle 25
change - change of substance - takes place, and is not meant to illustrate the
difference between actuality and potentiality as in Ll.
Second, it is hard to avoid drawing the conclusion that when Aristotle uses
entelekheia nominally in L2, saying that "change is an entelekheia", he might
well have meant for entelekheia to refer not so much to a state of being as to a
process of becoming, a meaning which would be counterintuitive in Ll, that is,
in the context of defining the soul. In other words, it is unclear in L2 whether
change simply refers to the state of being changed which the changing thing
itself possesses; or whether change refers to something outside of the changing
thing - to something which the changing thing participates in but does not
possess as a state - namely, the process by which the potentially changed thing
comes to be actually changed. The main problem with seeing change as a
process in which the changing thing participates is that it violates Aristotle's
clearly stated rule that 'There is no change above and beyond [i.e., outside of] the
changing thing (ouk esti de tis kinesis para ta pragmata = wa-laysat
~arakatun khiirijatan cani 1-umuri anfusihii)" (Phys. 3.1, 200b32-33 = af-
Tabzca, 168,10-11).
Whether Aristotle intended that his definition of change in L2 be read as
supporting the view that change is a state of being or a process of becoming was
a major source of disagreement among medieval Latin thinkers. 6 In fact, it is
still debated by Aristotle scholars: the process view held sway until the middle of
the twentieth century; the state view then replaced it as the standard
interpretation; and the pendulum now seems to have begun swinging back to the
process view. 7 For my own purposes, what is important is not so much
uncovering the correct interpretation of Aristotle's use of entelekheia to define
change. Rather, my concern is that if a commentator decides to interpret
Aristotle's use of entelekheia to define change as supporting the process view
of change, he will have committed himself to saying that entelekheia in general
includes process as well as state in its range of referents.
The third major difference between Aristotle's uses of entelekheia in Ll and
in L2 is that in L2 Aristotle does not introduce the distinction between first and
second perfections, unlike in Ll, where the distinction is crucial to his definition
of the soul. On the other hand, Aristotle does introduce a distinction in Physics
3.2 that - at first glance, at least - could be seen to be analogous to the
first/second entelekheia distinction introduced in LL In Physics 3.2 (and echoed

6 Maier 1958 (Avicenna's theory is treated at 68-73). Following in Maier' s footsteps are

McCullough 1980, Weisheipl 1981, and Des Chene 1996, 17-52.

7 The "process" or "actualization" interpretation of Aristotle's definition of kinesis is

expounded by Ross 1936, 537; and Peck 1972. The most ingenious - and certainly the most
influential - recent defense of the "state" or "actuality" view has been the criticism of Ross'
interpretation by Kosman 1969. Kosrnan's view is also held, or modified, by Owens 1978; Gill
1980, 130-133; Waterlow 1982, 112-119; and Hussey 1983, 55-65. Kosman's view is criticized in
turn by Kostman 1987, and Graham 1988.
26 Part/: The Ammonian Synthesis

at De Anima 2.5, 417al6-17) Aristotle says that change is a state of being active
(energeia) though in an incomplete or imperfect way (ateles):

L3
Aris~ii~iilis,
af-Tabf'a, 183,8-11
Aristotle, Phys. 3.2, 201 b32-33
Now change is thought to be a kind of activity, although an
incomplete activity. The reason for this is that what is potentially
changed - that whose activity is change - is incomplete [wa-l-
J:zarakatu mavzunun annahii fi'lun mii ilia annahu fi'lun ghayru
tammin wa-s-sababu ft dhiilika anna mii huwa bi-l-quwwati
mutalJarrikun wa-huwa lladhi l-J:zarakatu fi'luhu fa-huwa ghayru
tammin = he te kinesis energeia men tis einai dokei ateles de
aition d'hoti ateles to dunaton hou estin energeia].

I shall leave aside for the moment the thorny question of the equivalence of
entelekheia and energeia, a question which arises because - as mentioned at the
beginning of the chapter - the times when the two terms seem most
interchangeable are those when they are used adverbially, rather than nominally
as in L3. 8 I shall also continue translating teleion and ateles as "complete" and
"incomplete" until I discuss their use by the Neoplatonists, when the concepts of
entelekheia and teleiotes become assimilated, and "perfect" and "imperfect" will
become more apt.
Any attempt to square the incomplete/complete distinction implied in Physics
3.2 with the first/second distinction articulated in De Anima 2.1 will still have
to confront a number of problems. This is mainly because Aristotle first
introduces the incomplete/complete distinction not in L3 - where, in fact, it is
merely implied - but in Physics 3.1, 20la4-9, where he lists four examples of
paired contraries between which the four types of change take place. In the case
of change of substance, the contraries are - as mentioned above - form and
privation; in the case of change of quality, the contraries are black and white; in
the case of change of location, the contraries are up and down; and in the case of
change of quantity, the contraries are complete and incomplete. The
incomplete/complete distinction seems primarily applicable, therefore, to one
type of change - change of quantity - and only derivatively applicable (if
applicable at all) to other types of change, or to change in general.
To sum up: if - as a Greek or Arabic commentator - my aim is to harmonize
Aristotle's use of entelekheia in Ll and his use of entelekheia in L2, I will
have to confront three major challenges.

8 On the etymology of entelekheia and its differences with energeia, see Ross 1924, 245-246;

Chen 1958; Blair 1967; Franzen and Georgulis 1972; and Graham 1989. Most recently, Stephen
Menn (Menn 1994) has argued that a chronological progression can be detected in Aristotle's uses
of the terms dunamis, energeia and entelekheia.
 1. Aristotle 27

(1) Whereas in Ll entelekheia is identified with form and linked to substance as

form, in L2 it is not; and even when Aristotle does say, in Physics 3.1,
201a4, that form is one of the contraries involved in a particular type of
change - change of substance - the other contrary he mentions is privation
(steresis), not matter, with which form is contrasted in LL

(2) Whereas in LI entelekheia's identification with form is meant to contrast

the state of actuality or perfection with the matter's state of potentiality, in L2
entelekheia can plausibly refer either to the state of actuality or perfection
possessed by the changing thing or to the actualization or process of
becoming actual or perfect, in which the changing thing participates.

(3) Whereas in Ll the distinction between first entelekheia and second

entelekheia refers to the difference between possessing a capability to perfonn
a function and exercising that capability, no distinction between two types of
entelekheia is articulated in L2; and even the distinction between the
incomplete and the complete hinted at in L3 and articulated in Physics 3.1,
20la4-9, seems applicable only, or at least primarily, to the contraries of
quantitative change, and not to the states of possessing and using.

In the rest of this chapter I shall focus on the third challenge, reconciling the
first/second distinction of De Anima 2.1 with the incomplete/complete
distinction implied in Physics 3.1-2. This is for several reasons. First, Avicenna
regards the discussion of the first challenge as a dead end, since he insists that
determining whether or not something is a perfection will not tell us whether or
not that thing is a substance. In other words, Avicenna holds that perfection is
not necessarily linked to substantiality, a view I shall discuss in Chapter 6.
Second, the early Peripatetic commentators' discussions of the second challenge
came to revolve around their discussions of the third challenge; this will become
clear as we proceed through Chapter 2. What is more, the later Neoplatonic
discussions of the second challenge also came to revolve around a new, fourth
challenge: determining whether or not (and if so, how) perfection was linked to
causality; this will become clear in Chapters 3 and 4. The main point of listing
the challenges above, of course, is simply to show that Aristotle's uses of
entelekheia in Ll and L2 are not transparently identifiable, but warrant
aggressive interpretation.
In order to come to grips with the third challenge - in order, that is, to
reconcile the first/second distinction articulated in De Anima 2.1, with the
incomplete/complete distinction hinted at in Physics 3.2 and articulated in
Physics 3.1 - I, as a commentator, shall probably end up having to advocate
either of two positions. On the one hand, I could ( 1) throw my hands up and say
that the two distinctions are simply not identifiable. Try as I might to reconcile
them, I must conclude in the end that the two distinctions apply to two different
28 Part I: The Ammonian Synthesis

species of the overarching genus of entelekheia. The incomplete/complete

distinction applies to one species of entelekheia, namely the kind whose
specific difference is "predicable-of-change". And the first/second distinction
applies to another species of entelekheia, namely the kind whose specific
difference is "predicable-of-soul".
If, on the other hand, I wish to argue that the two distinctions are in fact
reconcilable, I shall have to choose between two further alternatives. The first
alternative (2a) has me saying that the two distinctions are identifiable in a
strong sense. In other words, I could argue that either of the two distinctions
applies equally to both species of the genus entelekheia. The
incomplete/complete distinction applies equally to the species of entelekheia
whose specific difference is "predicable-of-change" and to the species of
entelekheia whose specific difference is "predicable-of-soul". Similarly, the
first/second distinction applies equally to the species of entelekheia whose
specific difference is "predicable-of-soul" and to the species of entelekheia
whose specific difference is "predicable-of-change".
The second alternative (2b) has me saying that the two distinctions are
identifiable in a weak sense. In other words, I could argue that only one of the
distinctions applies to both species of the genus entelekheia, while the other
applies merely to one species of entelekheia, either to the species whose
specific difference is "predicable-of-change", or to the species whose specific
difference is "predicable-of-soul". I could, for example, argue (2b[i]) that the
first/second distinction can be construed broadly enough as to apply not only to
the species of entelekheia whose specific difference is "predicable-of-soul", but
also to the species of entelekheia whose specific difference is "predicable-of-
change". The incomplete/complete distinction, by contrast, can only be
construed so narrowly that it applies merely to change and not to the soul. In
short, everything which is subject to the incomplete/complete distinction will
also be subject to the first/second distinction; but not everything which is
subject to the first/second distinction will be subject to the incomplete/complete
distinction. For example, changes are subject to both the incomplete/complete
and the first/second distinctions; but souls are subject only to the first/second
distinction and not to the incomplete/complete distinction.
Conversely, I could argue (2b[ii]) that the incomplete/complete distinction can
be construed broadly enough as to apply not only to the species of entelekheia
whose specific difference is "predicable-of-change'', but also to the species of
entelekheia whose specific difference is "predicable-of-soul". The first/second
distinction, by contrast, can only be construed so narrowly that it applies merely
to the soul and not to change. In short, everything which is subject to the
first/second distinction will also be subject to the incomplete/complete
distinction; but not everything which is subject to the incompleteicomplete
distinction will be subject to the first/second distinction. For example, souls are
subject to both the first/second and the incomplete/complete distinctions; but
 1. Aristotle 29
changes are subject only to the incomplete/complete distinction and not to the
first/second distinction.
Let me give another example of (2b) in order to illustrate the kind of weak
identity which I am referring to. Take the two distinctions cause/effect and
agent/act. The cause/effect distinction can be construed broadly enough as to
apply both to the species of the genus existent whose specific difference is
"productive/produced with will" and to the species of the genus existent whose
specific difference is "productive/produced without will". The agent/act
distinction, by contrast, can only be construed so narrowly that it applies merely
to the species of the genus existent whose specific difference is
"productive/produced with will". Everything which is subject to the agent/act
distinction will also be subject to the cause/effect distinction; but not everything
which is subject to the cause/effect distinction will be subject to the agent/act
distinction. For example, my taking my son bowling is subject to both the
cause/effect distinction and the agent/act distinction, since my taking my son
bowling is voluntary; but the brick's falling to the ground is subject only to the
cause/effect distinction, since the brick's falling to the ground is involuntary.
What goes through my mind, as a commentator, when I am faced with
choosing between these various positions? Position (1) involves my giving up
on reconciling the incomplete/complete distinction implied in Physics 3.2 with
the first/second distinction articulated in De Anima 2.1, and logically speaking
this has to be my safest route. Any serious attempt to identify the two
distinctions will, it seems, require me to pare down so many rough edges -
including, as mentioned above, whether the basic distinction is one between the
two contraries of quantitative change or one between possessing a capability and
using that capability - that what was left of each distinction would be trivial.
Instead, my claim would be more limited: entelekheia is an overarching concept
which, when predicated of change, is distinguishable into incomplete and
complete, and when predicated of the soul, is distinguishable into first and
second.
In spite of the obvious advantages of this first option, the late-antique and
Arabic commentators tended to prefer position (2), a position which left them
the further choice between holding (2a) that the two distinctions were identifiable
in a strong sense, and holding (2b) that the two distinctions were identical in a
weak sense, with one distinction subsumed under the other. Needless to say, the
commentators' hermeneutical commitment to viewing Aristotle as a system-
thinker rather than as a problem-thinker (let alone as a thinker whose ideas may
have evolved during his lifetime, or even - God forbid - as a thinker whose
views on certain topics were hopelessly inconsistent) gave them a big push in
the general direction of position (2), in that it inspired them to persevere in
reconciling Aristotle's often incongruous applications of terms and distinctions
between terms.
30 Part I: The Ammonian Synthesis

Hermeneutical commitments aside, adopting position (2) - either strongly

identifying the two distinctions (2a), or subsuming one under the other (2b) -
forced the commentators to confront not only the obvious philosophical
problems pushing them to adopt position (1), but also the problematic
implications that arose when they chose position (2) and attempted to solve
those problems. What I mean is that the commentators' solutions to this
challenge were often Pyrrhic, in that they led to further, unsavory commitments
on the issues of the soul and change. The direst of these implications was first,
that the soul - or at least the intellectual part - could be described as undergoing
some kind of change when it makes the transition from knowledge to
contemplation; and second, that change could be described as a process in which
the changing thing participates, rather than as a state which the changing thing
possesses. What I am arguing, in other words, is that these two unsavory
commitments were the horns of a dilemma that emerged as a direct result of
trying to square the first/second distinction with the incomplete/complete
distinction.
Let me be more precise. Because the incomplete/complete distinction seems to
apply only, or at least primarily, to quantitative change, since that is the only
context in Physics 3 in which Aristotle articulates the distinction explicitly, I
could draw the conclusion that when Aristotle describes in Ll the distinction
between a soul in a state of first perfection and a soul in a state of second
perfection, he is imagining that when I make the transition from the state of
possessing the capability to think to the active exercise of that capability, my
soul is undergoing a change from incomplete quantity or magnitude to complete
quantity or magnitude. This is absurd, of course; commentators on Aristotle are
agreed that the soul undergoes no type of change, let alone a quantitative change,
when it passes from first to second perfection. The reason, mentioned above, is
that according to Aristotle change occurs between two contraries, and knowledge
and contemplation are not contraries, but rather indications of the presence or
absence of a relation. In this case, the relation which is present or absent is the
relation to the object of thought. When I am contemplating - when, that is, I
am in a state of second perfection - the relation between my soul and the object
of thought is present. When I am not contemplating - when, that is, I am in a
state of first perfection - the relation between me and the object of thought is
absent.
For example, when I am contemplating the Pythagorean theorem, my soul is
in a state of second perfection, since the relation between my soul and the object
of thought - the Pythagorean theorem - is present. But when I am watching
television, my soul is in a state of first perfection, since the relation between my
soul and the object of thought - the Pythagorean theorem - is absent. No change
takes place: as just mentioned, the transition from knowing the Pythagorean
theorem (i.e., the state of possessing the capability to think about the
Pythagorean theorem), to contemplating the Pythagorean theorem (i.e., actively
 1. Aristotle 31
exercising that capability to think about the Pythagorean theorem) is not a
change from one contrary to another, but rather the appearance of a relation
where previously there had been none. In short, if I say that the first entelekheia
of De Anima 2.1 is nothing other than the incomplete entelekheia of Physics
3.1, I shall have to conclude that when I am watching television and not
contemplating the Pythagorean theorem - when, in other words, I am in a state
of first perfection - my soul is itself a change or is somehow undergoing a
change. This is because my soul's being in a state of first entelekheia is
nothing other than my soul's being in a state of incomplete entelekheia, arxl
incomplete entelekheia is - according to the implication of L3 - what change
consists in. This is clearly unacceptable.
On the other hand, if I were to say that the complete entelekheia of Physics
3.1 is nothing other than the second entelekheia of De Anima 2.1, I would have
to conclude that once something has finished changing quantitatively - say, a
fruit which has expanded in magnitude to its full, complete and mature size - it
is somehow actively using a capability which it merely possessed when it was
incomplete and immature. This is because the fruit's being in a state of complete
entelekheia is nothing other than the fruit's being in a state of second
entelekheia, and second entelekheia is - according to the standard interpretation
of LI - using a capability which earlier one merely possessed. This too is
unacceptable, for it is difficult to understand what it would mean to say that a
ripe piece of fruit is actively using a capability which it merely possessed when
it was unripe.
It seems hopeless, then, to argue in favor of the strong identity of position
(2a), that is, to argue that the first/second distinction means nothing more or less
than the incomplete/complete distinction, and vice versa; and that everything
which is subject to the first/second distinction will also be subject to the
incomplete/complete distinction, and vice versa.
This leaves me with position (2b ), arguing that one distinction can be
construed broadly enough as to apply both to the soul and to change, while the
other can only be construed so narrowly that it applies merely to the soul or to
change. To argue in favor of this position I shall have to choose between (2b[i))
expanding the notions of possessing a capability and using that capability so
that they were no longer restricted to intellection, nor even to being ensouled in
general; and (2b[ii)) expanding the notions of incompleteness and completeness
so that they were no longer restricted to quantitative change, nor even to change
in general.
According to (2b[i]), possessing and using can, if construed broadly enough,
be understood as applying not only to knowledge and contemplation, and to
being asleep and awake (Aristotle's other example of the distinction in LI), but
also to incomplete and complete quantitative changes, and to incomplete arxl
complete changes in general. According to (2b[ii]), incompleteness arxl
completeness can, if construed broadly enough, be understood as applying not
32 Part I: The Ammonian Synthesis

only to quantitative changes, and to changes in general, but also to non-kinetic

transitions from possessing to using, such as intellection. Once I have
undertaken either expansion, (2b[i]) or (2b[ii]), entelekheia could then be held to
be basically divisible into the expanded distinction, an expanded distinction
which would now subsume the non-expanded distinction.
Which distinction ought I, as a commentator, to choose for expansion,
possessing/using or incomplete/complete? Neither is without difficulties. On the
one hand, expanding the possessing/using distinction (2b[i]) seems like the
obvious choice, since it is, after all, the only distinction which Aristotle
explicitly applies to entelekheia; the incomplete/complete distinction is only
expressly articulated in the context of the contraries of quantitative 'change, and
not in the context of change in general; and even when Aristotle does imply that
the incomplete/complete distinction applies to change in general, he uses the
term energeia and not entelekheia, and says merely that change is an
incomplete energeia, leaving us to guess what a complete energeia might be.
On the other hand, natural philosophy is a broader science than psychology,
since its subject is all natural things, and not just natural things with souls.
Because the incomplete/complete distinction seems to apply to at least one type
of change, and perhaps to change in general, expanding the incomplete/complete
distinction (2b[ii]) would appear to be the better choice, if entelekheia is to
cover inanimate as well as animate things. One problem with this line of
argument is that Aristotle's assertion at Parts of Animals 1.1, 641a22-b10, that
the intellect is not a subject to be investigated by the natural philosopher,
appears to eliminate the very transition - from knowledge to contemplation -
which (2b[ii]) must account for.
Nevertheless, what appears to have been the decisive factor in prompting the
commentators to choose (2b[ii]), expanding the incomplete/complete distinction,
is that Aristotle simply provides more textual raw material with which to
construct an interpretation along the lines of (2b[ii]) than he provides to
construct an interpretation along the lines of (2b[i]). This was true for the
ancient commentators, as we shall see, and it remains the case today: the most
influential recent treatment of Aristotle on kinesis, that of Kosman, is based on
an expanded notion of completion or perfection.
This raw material came in two forms. First, Aristotle devotes an entire chapter
of Metaphysics 5 - his philosophical lexicon - to discussing "the complete" (to
teleion), a discussion which itself allows for an expanded notion of
completeness (and, by implication, of incompleteness). Second, Aristotle uses
two terms for completion or completeness - teleiosis (which appears about two
dozen times in the Aristotelian corpus) and teleiotes (which appears only four
times) - in contexts similar to those in which he uses entelekheia.
Here is what Aristotle says about to teleion:
 I. Aristotle 33
L4
Aris~ii~iilis,Ma ba<da f-!abica !I, 621,9-622,13
Aristotle, Metaph. 5.16, 1021bl2-1022a2
The complete [at-tamm =to teleion] is spoken of in one way as
that outside of which no part whatsoever can be found; for example
a time-span: each thing's complete time-span consists of that
outside of which no time at all can be found which is part of that
time-span. Similar to this is that to which some kind of excellence
applies, and that which is spoken of in the sense of goodness,
whenever it has a superiority in its genus; for example a doctor is
complete and a piper is complete when they have no deficiency in
their special kind of excellence. In this way ("complete"] is said -
by means of transference [bi-ntiqali l-qawl = metapherontes] - of
bad things. For we say "a complete liar" and "a complete thief'
because we also apply the term "good" to them, as when we say a
good thief and a good liar. Every single one of these complete
things and all their substances are then said to be complete when,
with respect to their particular kind of excellence, no part of their
natural magnitude is deficient. Those things in which their
worthwhile end exists are also said to be complete [wa-ayrj.an
tuqalu tiimmatan allatl lahii tamamun farj.ilun = eti hois huparkhei
to telos, spoudaion on, tauta legetai teleia], for they are complete
by virtue of the end which they possess; and since an end is a kind
of limit ffa-idhi t-tamamu ghiiyatun ma mina l-ghiiyiiti = h6st' epei
to telos ton eskhat6n ti esti] we transfer it to bad things, saying
that a thing is completely ruined and completely destroyed when
there is no deficiency in the destruction, but it is instead at the
utmost degree of badness. For this reason passing away is also
said, by way of transference, to be a completeness, because both
are limits. For the end, and what something is on account of, are
limits. ffa-t-tamiimu wa-lladhi bi-sababihi yakunu sh-shay' u
ghiiyatun = telos de kai to hou heneka eskhaton]. Those things
which are themselves said to be complete are thus spoken of in
these ways: some in the sense that no goodness is deficient in
them, nor any superiority, nor is any part of them found outside
them, and generally in the sense that nothing at all is superior in
every one of the genera, nor is it outside them. As for the other
ways, they are in respect of these, either by doing, or possessing,
or being equivalent to something like this, or by being in another
kind of relation to things which are said to be complete in the
primary way.

I shall discuss in Chapter 5 the implications of the Arabic translations of teleion

as tlimm and of telos as tamlim in L4. For the moment what is of interest is
Aristotle's series of claims about to teleion. In the first sentence he says that to
teleion is that which has quantitative completeness or wholeness. That is, a
thing is teleion when it has parts and when none of its parts is missing (though
the example Aristotle gives in the second sentence is of temporal rather than
spatial magnitude). Aristotle next claims that teleion applies to degrees of
excellence in a particular field of endeavor: a doctor is complete when there is no
34 Part I: The Ammonian Synthesis

degree of excellence in doctoring beyond what he has attained. In general, he

says, things which have achieved their end - their telos - may be said to be
complete. Even bad things such as dying or thieving can be said to be complete
once they achieve their telos; although to be precise, they are "complete" only in
a metaphorical or derived sense of the word. Here telos is assumed to be
synonymous with peras ( = nihiiya, "terminus"), a position which Aristotle
openly adopts in the next chapter, Metaphysics 5.17. There he says that a
thing's terminus is that outside which nothing can be found, and he explicitly
identifies the terminus with the telos of each thing. 9
In Metaphysics 5 in general, it is sometimes unclear whether the first sense
of a philosophical term which Aristotle lists is the philosophically basic sense,
or simply the most commonly applied or primitive sense. In the case of
Aristotle's discussion in Metaphysics 5.4 of the term phusis ("nature"), for
example, the first sense offered ("the coming to be of growing things") is clearly
the primitive sense, and the last sense offered ("the substance of things which
have, in and of themselves, the principle of change") is clearly the most
philosophically basic sense. That also seems to be the pattern with his
discussion of to teleion. As expressed in the first sentence, the primitive sense
of teleion is quantitative wholeness: the quality possessed by a whole none of
whose parts is missing. Advocates of (2b[ii]) can appeal to that primitive sense
of completeness to cover the contraries of quantitative change, given that the
notion of quantitative wholeness seems easily applied to spatial magnitude.
The example Aristotle gives in the second sentence - a span of time - will
cover change in general, if change is held to be a process in which the changing
thing participates. Advocates of (2b[ii]) who are inclined to see change as a
process (2b[ii]') will argue that since change is an incomplete activity, change
will consist in the dynamic quality of incompleteness that characterizes an
unfinished time-span. The quality of incompleteness here is dynamic because it
is temporal: with the passing of every moment, the process becomes measurably
less incomplete, and more complete.
Even if I do not want to interpret change as a process, the use of teleion to
describe a thing possessing a state - as opposed to a thing participating in a
process - finds support in L4. In fact, as mentioned above, Aristotle indicates at
the end of L4 that the sense in which teleion refers to a thing that possesses a
state - the state of having attained its telos - is more philosophically basic than
the sense in which teleion refers to a thing that has undergone spatial or
temporal completion. Advocates of (2b[ii]) who are inclined to see change as a
state (2b[ii]") can then argue that since change is an incomplete activity, change
will consist in the particular type of incompleteness that a changing thing
possesses. The quality of incompleteness here is static because it is logical, not
physical: the changing thing possesses the state of incompleteness in the sense

9 Metaph. 5.17, 1022a4-l3 =Ma ba<da f-fabi<a II, 628,3-12.

 1. Aristotle 35

that incompleteness is taken here to be the essential differentiating characteristic

of a logical category - the category of "changing thing" - rather than taken to be
a measurable characteristic of spatial and temporal quantities, such as the growth
of the fetus in the womb or the time-span of any process of development.
In short, seeing change as a process in which the changing thing participates
will require me to interpret the incompleteness of change as admitting of
measurable degrees: at the beginning of the change, the incompleteness is at its
greatest, and towards the end of the change, the incompleteness is at its least.
Seeing change as a state possessed by the changing thing, on the other hand,
will require me to interpret the incompleteness which the changing thing
possesses as not admitting of measurable degrees: the changing thing possesses
the differentiating characteristic of one logical category, while the changed thing
possesses the differentiating characteristic of another logical category. Neither
category admits of measurable degrees of difference, since the categories are
logical, not physical: a thing is either in a state of incompleteness or it is in a
state of completeness.
Regardless of how decisively the evidence provided by L4 supports the state
interpretation of change over the process interpretation of change, an advocate of
either (2b[ii]') or (2b[ii]") could still mine that text for raw material with which
to fashion an overarching incomplete/complete distinction which would subsume
entelekheia as knowledge and contemplation as well as entelekheia as change.
This is because telos, "end", can refer not only to "limit" or "terminus" (peras)
but also to "purpose" or "that for the sake of which" (to hou heneka). What I
mean is that a commentator could argue that when Aristotle asserts in L4 that
"Those things in which their worthwhile end exists are also said to be complete,
for they are complete by virtue of the end which they possess (eti hois
huparkhei to telos, spoudaion on, tauta legetai teleia · kata gar to ekhein to
telos teleia = wa-aytj.an tuqiilu tiimmatan allatl lahii tamiimun fiitj.ilun fa-
innaha tiimmatun bi-tamiimiha lladhi huwa laha)" (Metaph. 5.16, 102lb23-25
=Ma bacda f-{abica II, 622,2-3), he is referring to telos in the sense of purpose
or that-for-the-sake-of-which (to hou heneka) and not to telos in the sense of
limit or terminus (peras). This line can further be read as implying that to
teleion refers to something which is "ended", either in the sense of being
directed towards an end but not having attained it, or in the sense of serving
as an end for something else.
The result of having so many meanings of teleion to choose from is that a
commentator had more tools at his disposal with which to subsume the
first/second distinction under the incomplete/complete distinction, regardless of
whether he chose to advocate the process view of change (2b[ii)') or the state
view of change (2b[ii] " ). Change could be held to be an entelekheia ate/es
because it has not reached its telos qua "limit" or "terminus". The changing
thing either remained in an unfinished process which had not reached its terminus
(2b[ii] ') or itself possessed the state of not having reached its terminus (2b[ii] " ).
36 Part I: The Ammonian Synthesis

On the other hand, knowledge and contemplation could each be held to be

teleia because both are "ended", that is, both have a telos qua "end" or
"purpose". Knowledge, the first entelekheia of the De Anima, is "ended" in the
sense of "having an end": it refers to a capability which is directed towards an
end. Contemplation, the second entelekheia of the De Anima, is "ended" in the
sense of "being an end": it refers to a function which serves as an end. Thus:

2b[ii]'
Phys. l't: entelekheia ateles =incomplete process (e.g., "construction")
DA l't: entelekheia teleia =ended capability (e.g., "knowledge")

Phys. 2"d: entelekheia teleia = completed process (e.g., "constructed")

DA 2"d: entelekheia teleia =ended exercise of capability (e.g., "contemplating")

or

2b[ii]"
Phys. 1'1: entelekheia ate/es= state of incompleteness (e.g., "being
constructed")
DA l't: entelekheia teleia =ended capability (e.g., "knowledge")

Phys. 2"d: entelekheia teleia =state of completeness (e.g., "having been

constructed")
DA 2"d: entelekheia teleia =ended exercise of capability (e.g., "contemplating")

Adding teleion to the mix certainly helps a cormnentator identify the complete
entelekheia of the Physics with the second entelekheia of the De Anima. Both
are teleia, after all; and even though they are teleia in different ways - the
former is teleia in the sense of being completed, the latter is teleia in the sense
of serving as an end - the shared characteristic of being teleia is enough to
contrast them with the incomplete entelekheia of the Physics. Of course, that
still leaves the commentator the task of accormnodating the first entelekheia of
the De Anima, which ought, according to this approach, to be viewed as ateles
rather than teleia.
In spite of that very loose end, the commentators persevered in viewing
teleion, both in the sense of quantifiably whole and in the sense of ended, as a
useful tool in flushing out the evasive entelekheia. Fueling their enthusiasm
for teleion was the possibility of appealing to Aristotle's use of the abstract
nouns which derive from it - namely teleiotes and teleiosis, translatable as
"completeness" and "completion" - in contexts similar to those in which he uses
entelekheia. As I mentioned earlier, Aristotle employs the terms so rarely that
it would be foolhardy to claim that he applied them in any strict or consistent
way in his work. But as with to teleion, so teleiotes and teleiosis were seen by
 1. Aristotle 37
commentators as rough diamonds, so to speak, ready to be cut into sparkling
new concepts with which to gloss entelekheia, and thereby to sort out the
confusion surrounding Aristotle's seemingly inconsistent uses of that term.
Of the two terms, teleiotes is used less often. Aristotle seems to have
understood teleiotes and ateleia as the abstract nouns corresponding to the
substantive adjectives to teleion and to ateles which, as we saw, he applied to
the contraries of change of quantity in Physics 3.1, 20la4-9. For example, in
Physics 8.7 - foreshadowed at Physics 3.6, 207a21 - teleiotes and ateleia are
linked to change of quantity. The Physics 8.7 passage is meant to distinguish
between change of location and the other types of change:

LS
Aristiitalis, ar-Tabi'a, 887,6-12
Aristotle, Phys. 8.7, 261a32-37
We say that it appears that none of the other changes [i.e., other
than circular locomotion] can be continuous. This is because all of
the various types of change are only ever from one opposite to
another opposite. For example, the two limits of coming-to-be
and passing-away [i.e., change of substance] are the existent and
the non-existent [mawjudun wa-lii-mawjudun =to on kai to me on];
the two limits of alteration [i.e., change of quality] are the
contrary affections; and the two limits of increase and decrease
[i.e., change of quantity] are greatness or smallness, or
completeness or incompleteness of magnitude [aw kamiilu l- 'i'{,am
wa-naq~iinuhu = € teleiotes megethous kai ateleia].

Given Aristotle's apparently restricted use of teleiotes in passages focusing on

change of quantity, teleiosis might have appeared to a commentator as a better
candidate for the abstract noun corresponding to a new, expanded notion of to
teleion. In a passage in Physics 7.3, for example, Aristotle uses teleiosis in a
way that might help show how an expanded incomplete/complete distinction
could cover the first/second distinction. In Physics 7.3 Aristotle is at pains to
show that while the conditions preparatory to the acquisition of knowledge may
involve changes - and specifically changes of quality, or alterations (alloioseis
= istif:ziUiit) - the acquisition of knowledge is not itself a change. This is
because, as mentioned above, states of the soul are indicative merely of the
presence or absence of a relation, and therefore the transition between them
cannot be categorized as a change of quality:

L6
Aristiitalis, ar-Tabl'a, 760,8-761,6
Aristotle, Phys. 7.3, 246al0-16
Again, states [T 760,8: al-hay'iit = hexeis], whether of the body or
of the soul, are not changes of quality. For some are excellences [T
760,9: far;J.ii'il = aretai] and others are defects, and neither
excellence nor defect is change of quality. Excellence is a kind of
38 Part /: The Ammonian Synthesis
completion [T 761,4: kamiilun ma = teleiOsis tis], for anything
which has acquired its own special excellence and is then in its
natural state is - at that moment - complete [T 761,5: kiimilan =
teleion].

What made teleiosis even better qualified to serve as the abstract noun for an
expanded notion of to teleion was that Aristotle used teleiosis not only to
describe states which do not involve or presuppose change, as in L6 (as well as
in L4, where excellence - arete - is also held to be a kind of teleiosis), but also
to describe changes themselves. In other words, it seemed that teleiosis could
serve in precisely the contexts - the states of the soul as well as change - in
which entelekheia figured so prominently. It might, then, seem surprising that
most commentators chose to use teleiotes instead of teleiosis to gloss
entelekheia.
The reason they did so, I believe, is that when Aristotle did use teleiosis in
the context of change, he meant for it to describe types of change that are more
easily understood as processes rather than as states, as completions rather than
completenesses. Aristotle used teleiosis most often to describe natural
phenomena such as fetal development, a process which involves not only growth
- that is, change of quantity - but also concoction or ripening - that is, change
of substance or of quality (the commentators disagree). It would be a reasonable
extrapolation to expect that Aristotle also thought of an infant's growth to a
mature, adult size, as a teleiosis which takes place after birth. And given
Aristotle's use of teleiosis to describe concoction generally, one has to assume
that he would also regard the functioning of the soul's nutritive faculty in
transforming food into blood, as well as pubescence (i.e., the sexual maturation
of children into adults) as post-natal teleioseis.
Admittedly, it is sometimes unclear whether in a few of those cases Aristotle
also meant for teleiosis to describe the end-state of the process of fetal
development rather than the process itself. But it is certain that in most instances
teleiosis was meant to refer to a process. 10 What is more, if a fetus'
development can be assigned to any of the four types of change, change of
quantity will be the leading candidate; and if concoction can be assigned to any of
the four types of change, it is probably substantial change, and possibly change
of quality. It would be perfectly reasonable for a commentator to suppose, then,
that Aristotle meant for teleiosis to refer to the process of change generally - the
passage between the beginning-state of ateleia and its contrary, the end-state of
teleiotes - and not to only one of the four types of change.

' 0 Passages where teleiosis refers to a process are: GA 3.7, 757a32 (teleiosin = kamal at Ff
kawn al-l:zayawan, 117,21); 4.4, 770b26 (teleioseis = tamam at 152,25); and 5.10, 777b27 (teleioseis
=at-tamam wa-1-kamal at 171,6); HA 6.3, 56Ja6; and 7.4, 583b24, 584a34 and 584b26; Meteor.
4.2, 379bl8 and 21 and 4.3, 380a19; Prob/. 10.41, 895a27; 20.7, 923b4; and 20.24, 925b29.
=
Passages where teleiosis refers to the end-state of a process are: GA 3.2, 753a10 (teleiosin at-
=
tamam 106,16); and 4.8, 776bl(bis) (releiosis tamam at 167,2l[bis]).
 1. Aristotle 39
The problem is that if I, as a commentator, wish to use teleiosis as a gloss of
entelekheia, I will also commit myself to understanding change as a process.
And in fact, some of the commentators who allowed that change could be seen as
a process used teleiosis to gloss entelekheia when applied to change, and
teleiotes to gloss entelekheia when applied to the soul. By contrast, those
commentators wishing to see change as a state used teleiotes to gloss
entelekheia when applied to both change and the soul, reserving teleiosis to
describe only the specific changes of quantity, quality and substance that obtain
during fetal development and concoction - primarily, it seems, out of deference
to Aristotle's use of teleiosis in those contexts.
The conclusions one has to draw, therefore, are that Aristotle was inconsistent
in his use of the terms teleiosis and teleiotes, since both seemed applicable to
the end-state of a change of quantity; and that Aristotle used teleiosis to describe
processes (fetal development, ripening or maturation, and concoction generally),
the end-states of such processes, and states of excellence or perfection which
involve no change but only the presence or absence of a relation. If, because of
Aristotle's broader use of teleiosis, I choose to use teleiosis instead of teleiotes
as a gloss of entelekheia, I shall also commit myself to viewing entelekheia
as a process as well as a state.
How did the commentators deal with this problem? As I just mentioned, a
commentator committed to seeing change as a process (2b[ii]') could tum to
teleiosis as a gloss of entelekheia when he reckoned entelekheia referred to a
process, namely the process of change; and could tum to teleiotes as a gloss of
entelekheia when he reckoned entelekheia referred to a state, be it the state of
having changed, the capability to function or the exercise of that capability.
Thus:

2b[ii]'
Phys. 1": entelekheia ateles = teleiosis qua "process of completion" =
incomplete process (e.g., "construction")
DA I": entelekheia teleia = teleiotes qua "state of being directed towards an
end"= ended capability (e.g., "knowledge")

Phys. 2"d: entelekheia teleia = teleiotes qua "state of completeness" =

completed process (e.g., "constructed")
DA 2"ct: entelekheia teleia = teleiotes qua "state of serving as an end" = ended
exercise of capability (e.g., "contemplating")

On the other hand, a commentator committed to seeing change as only a state

(2b[ii]") could tum to teleiotes as a gloss of entelekheia regardless of whether
entelekheia referred to the state of changing, the state of having changed, the
capability to function or the exercise of that capability. Thus:
40 Part I: The Ammonian Synthesis
2b[ii],,
Phys. 1•t: entelekheia ateles = teleiotes qua "state of being completed" = state
of incompleteness (e.g., "being constructed")
DA 1st: entelekheia teleia = teleiotes qua "state of being directed towards an
end"= ended capability (e.g., "knowledge")

Phys. 2°d: entelekheia teleia = teleiotes qua "state of having been completed"
=state of completeness (e.g., "having been constructed")
DA 2°d: entelekheia teleia = teleiotes qua "state of serving as an end" = ended
exercise of capability (e.g., "contemplating")

Although the commentators' varied uses of Aristotle's concepts of teleiosis and

teleiotes succeed in providing common ground on which to reconcile the
incomplete/complete distinction of the Physics with the De Anima's first/second
distinction, the horns of the process/state dilemma remain sharp. What I mean is
this: I could, as a commentator, use teleiotes to gloss all instances of
entelekheia (2b[ii)"). That is, I could insist on understanding entelekheia as
only ever referring to a state. This is consistent with Aristotle's insistence in
Physics 3.1 that change is only ever in the changing thing.
The problem then arises, how do I distinguish the teleiotes that is possessed
by a changing thing from the teleiotes that is possessed by a changed thing? I
could, I suppose, claim that the teleiotes of the changing thing is active, while
the teleiotes of the changed thing is inactive. That would be fine, but what am
I to do with the teleiotes of one who possesses knowledge but is not actively
contemplating, and the teleiotes of one who is actively contemplating? It seems
that the criteria of activity and inactivity will get me into trouble here, since the
teleiotes of one who possesses knowledge but is not actively contemplating is
inactive, which implies that it is the end-state possessed by a changed thing; and
since the teleiotes of one who possesses knowledge and is actively
contemplating is active, which implies that it is the state of change possessed by
a changing thing.
The result of using the criteria of activity and inactivity is that I have
succeeded in implying precisely what I wanted to avoid: that knowledge is the
end-state possessed by a changed thing, that acquiring knowledge is therefore a
change, and that contemplating is itself a change. I could perhaps reconcile the
inactive end-state of the changed thing with the inactive state of someone who is
knowledgeable but is not actively contemplating, by saying that even though the
acquisition of knowledge is not, properly speaking, a change, the conditions that
are preparatory to the acquisition of knowledge - such as walking to class or
taking a book out of the library - are changes. Nevertheless, I am left with
seeing activity as binding together change and contemplation.
My only recourse, it seems, is to use Aristotle's own criterion in L3 and say
that teleiotes, when understood as the state possessed by the changing thing, is
 1. Aristotle 41

incomplete, in contrast to the complete teleiotetes of the changed thing, of

someone who is knowledgeable but is not actively contemplating, and of
someone who is actively contemplating. That still leaves me with the problem
of explaining what an incomplete completeness is: either it is a self-
contradiction, and therefore as meaningless as saying "a non-white whiteness" or
"an unsmall smallness"; or it simply means an incompleteness, in which case
the idea of teleiotes no longer serves to bind together the entelekheia that is
descriptive of change and the entelekheiai that are descriptive of the changed
thing's end-state and of the soul's states of possessing a capability and using that
capability. It seems that I am left, after all that work, with option (1), namely,
giving up on reconciling the first/second distinction articulated in LI and the
incomplete/complete distinction implied in L2.
The one way I could, as a commentator, still use the notion of completeness
to bind together the entelekheia of LI and the entelekheia of L2, is to appeal
to the apparent distinction between teleiotes and teleiOsis. I can use teleiotes to
refer to the end-state possessed by a changed thing, to the state of someone who
is knowledgeable but is not actively contemplating, and to the state of someone
who is actively contemplating. But because, as shown above, using activity and
inactivity as criteria to distinguish between change itself and the end-state
possessed by a changed thing leads to so many other problems (either implying
that contemplating is a change, or propounding the absurd notion of an
incomplete completeness), I can choose instead to call change a teleiosis rather
than a teleiotes (2b[ii] '). I will have retained the notion of completeness as an
overarching gloss of entelekheia, but I will have paid a heavy price: because
teleiosis was so clearly used by Aristotle to refer to processes such as fetal
development and growth, my use of teleiosis will commit me to seeing change
as a process in which the changing thing participates, rather than seeing change
as a state possessed by the changing thing. Nevertheless, I will have avoided
seeing contemplation as a change, since contemplation is a teleiotes and not a
teleiosis. I will also have avoided propounding the seemingly contradictory
notion of an incomplete completeness and retained the idea of completeness as a
gloss to explain entelekheia, though with a rider: I must now distinguish
between teleiosis as process of completion and teleiotes as state of
completeness. In any case, historically, the attempt to resolve this dilemma is
one of the main sources of the view that change is best understood as a process,
an interpretation that, as mentioned above, reverberated throughout the Middle
Ages and still does among Aristotle scholars.
My main purpose in Chapter 1 has been to show that Aristotle's uses of his
new term entelekheia to define both motion and the soul are opaque and appear
at first (or even second) glance to be inconsistent, and that they therefore invited
energetic exegesis along the lines I have described. As I shall show in Chapters
2, 3 and 4, Alexander subscribed to the "completeness" rather than the
"endedness" understanding of teleiotes; Themistius also subscribed to the
42 Part /: The Ammonian Synthesis

"completeness" understanding of teleiotes but hinted that the distinction between

the two types of endedness - I shall call them "autotelic" and "allotelic"
endedness - might provide an escape route from a further complication; the
Neoplatonists subscribed generally to the understanding of teleiotes as endedness
but collapsed the two different types of endedness which Themistius appears to
have advocated; and the Arabic translations collapsed endedness and end. This is
the legacy left to Avicenna.
 2. Alexander and Themistius
Attempts at Reconciliation

Reconstructing the views of Alexander of Aphrodisias (fl. 205 AD) is hampered

by the loss of his conunentary on the Physics. Nevertheless, we can get some
idea of what he thought by sorting through the bits and pieces of physical
doctrine contained in his Quaestiones as well as the citations contained in
Simplicius' conunentary on the Physics, and then comparing that material with
what he says about entelekheia, teleiosis, teleiotes and to teleion, in the
context of the soul, of which there is much evidence, and with what he says
about those concepts in other conunentaries. In Alexander there seems to be an
acute awareness of the problems that arise from Aristotle's use of entelekheia to
define the soul and to define change, and a willingness to use the Aristotelian
terms teleiosis and teleiotes in ways different from Aristotle's in order to help
resolve those problems.
Alexander is quite consistent in his uses of teleiosis and teleiotes. Apart from
one example, Alexander always uses teleiOsis, and never teleiotes, when he
wishes to describe processes such as fetal development, growth to maturity, and
concoction; in other words, when he wishes to refer to quantitative, substantial
or qualitative changes that occur over periods of time. 1 Like Aristotle, Alexander
also uses teleiosis to refer to states. But unlike Aristotle, who almost always
uses teleiosis to refer to states such as "excellences" (aretai) that present
themselves following transitions (such as not-contemplating to contemplating)
which are not changes of quality (alloiOseis), Alexander uses teleiosis to
describe the state of completeness following changes of quantity, quality and
substance. 2 In this respect Alexander seems to be following Aristotle's rare use
of teleiosis in the Generation of Animals to refer to the state of completeness
following a developmental change.
But in contrast to Aristotle, who appealed to teleiosis far more often than to
teleiotes, Alexander assigns teleiotes the more prominent role. Not only does
Alexander use teleiotes to describe states such as aretai, he appeals to teleiotes
whenever he wishes to gloss entelekheia. Alexander even goes so far as to
claim (incorrectly) that Aristotle himself identified entelekheia and teleiotes,
saying in one passage, "The soul, therefore, is substance as form (kata to

1 Alexander, in Metaph. 1.5 (ad 985b26), 38,18; in Meteor. 4.2 (ad 379bl0), 186,14. 16.31.35

and 187,7(bis); and 4.3 (ad 380al2), 188,24.25 and 189,5.34; DA, 36,6; Quaest. 3.14, 111,17.23.24;
in Sens. 6 (ad 446b2), 125,15. At in Meteor. 4.2 (ad 379bl0), 188,13 Alexander uses teleiotes to
describe concoction, though given the uniqueness of its usage in that sense, my guess is that its
ap~arance there is the result of a scribal error.
Alexander, in Metaph. 5.16 (ad 1021b12), 410,19.
44 Part /: The Ammonian Synthesis

eidos). Aristotle calls the form 'perfection' (teleiotes) and 'actuality'

( entelekheia)". 3
Alexander's new uses of teleiosis and particularly teleiotes can be seen more
clearly in his commentary on L4. In that commentary Alexander distinguishes
between three senses in which something may be said to be to teleion. The first
sense (in Metaph. 5.16, 410,18-28) refers to something which has attained its
teleiosis, here understood as referring to temporal completeness, and thus to a
kind of quantitative completeness. As I mentioned above, Alexander's use of
teleiosis in this sense finds only very limited support in Aristotle.
The second sense (in Metaph. 5.16, 410,28-411,18) in which something is
said to be teleion is when it possesses an excellence (arete) compared to other
similar things. That is to say, a thing is teleion when it possesses a superiority
to all other things alike it in species. Alexander's use of "species" (eidos) here,
in contrast to the "genus" (genos) which appears in the text of L4, is possibly
significant, since Alexander seems to be making two claims. First, excellence or
superiority is a teleiotes in the sense that it signifies the furthest limit of
formation, the maximum extent to which a form can inhere in a thing. Second,
excellence or superiority is a teleiotes in the sense that it signifies the full
exercise of the capability represented by the inherence of that form.
Although Alexander does not make this explicit, the implication of his use of
"alike in species" rather than Aristotle's "alike in genus" is that, in logical
terms, Alexander saw the teleiotes here as playing something like the role of the
specific difference: the essential characteristic that differentiates one species of a
genus from another species of that genus, the essential characteristic that makes
a species the species it is. In the case of the species "human", the specific
difference "rational" is what sets my species apart from other species subsumed
under the genus "animal". Alexander's point appears to be that the more rational
I am, the closer to my teleiotes as a human I shall be. However, I hesitate to
make too much of Alexander's choice of terms here, because a strict
understanding of genus, species and specific difference comes into ancient
philosophy only after the appearance of Porphyry's /sagoge. My aim here is
simply to show how Alexander's choice of species rather than genus provided
Peripatetic cover for a post-Porphyrean move by Ammonius and his successors,
which I shall describe in Chapters 3 and 4.
The third sense of to teleion, according to Alexander, refers to something
which is in possession of its own end (telos). Whereas the first and second

3 Alexander, DA (Mantissa), 103,3-4. At Alexander, DA 16,1 the claim that Aristotle held the

soul to be teleiotes as well as entelekheia is repeated. The editor suggests that this line might be a
glossator's interpolation, though given its appearance in the Mantissa, where the kai is unlikely to
be epexegetical (since the kai introduces entelekheia, not teleiotes), it seems reasonable to retain
it. According to Todd 1974, 214n.27, Aristotle' s " identification" of entelekheia and teleiotes is also
attested to in a scholium (Schol. in Arist. 358a 19). Alexander pairs teleiotes and entelekheia
(without, however, claiming that Aristotle did it first) at DA 24,1-2 and 52,2-3.
 2. Alexander and Themistius 45
senses of teleion refer to quantitative completeness (teleiosis) and qualitative or
substantial completeness (teleiotes), respectively, the third sense refers to a kind
of endedness. Alexander calls this third sense a teleiotes. But here Alexander
holds to a strict understanding of telos: he says (in Metaph. 5.16, 411,18-38, as
well as in Metaph. 5.17, 413,31-34) that it refers to the terminus or limit of a
process.4 Alexander's strict understanding of telos here prevents him from using
teleiotes to refer to endedness in the sense of being directed at an end or
serving as an end, senses which Themistius would later appeal to in an effort to
reconcile the first and second entelekheiai of De Anima 2.1 with the first and
second entelekheiai implied in Physics 3.1 and 3.2.
The general impression one gets from Alexander's works is that teleiotes has
replaced teleiosis in importance as a philosophical term. In this respect
Alexander's use of the two terms is the reverse of Aristotle's. What is
Alexander's rationale for elevating teleiotes over teleiosis? First of all, as I
discussed above, Aristotle used teleiosis not only to describe states such as
excellence, he used it even more often to refer to changes - fetal development and
concoction - which cannot possibly be applied to the soul. Alexander wants
another term for completeness which can be used to gloss entelekheia but
which will not imply that the soul is either a concoction or the result of a
concoction. To do this he needs not only to understand teleiotes in a far broader
sense than Aristotle understood it, but also to choose a term which covers all
transitions or transformations leading up to the teleiotes. This new term must
cover the four changes articulated in Aristotle's discussion of kinesis, as well as
transitions such as not-seeing to seeing and not-contemplating to contemplating,
which Aristotle insists are not changes, and specifically, not changes of quality.
The term Alexander chose is metabole, used by Aristotle in a separate discussion
of change in Physics 5.1 and 5.2.
In those chapters Aristotle settles on a new categorization of change to replace
the one he had originally proposed in Book 3. In Book 5 Aristotle now says that
genesis, or substantial change, is in a different category from the other types of
kinesis introduced in Physics 3. This is because genesis involves a transition
from one contradictory to another (ta kat' antiphasin), that is, between two
things so related that one and only one can be true at the same time (e.g., the
transition from not-being to being). Quantitative change, qualitative change and
locomotion, however, involve transitions between one contrary and another (ta
enantia), that is, between two things within the same genus which differ from
each other to the greatest possible degree (e.g., the transition from small to

4 Here Alexander may also have had in mind Metaph. 10.4, 1055al0-18, where Aristotle says

that everything which possesses a telos is called teleion, the sense of telos in this passage being
identical to peras (i.e., end as limit rather than end as purpose or function). Aristotle also identifies
to teleion in that passage as that which is greatest in a genus, not a species, reinforcing for
Alexander the connection to L4.
46 Part /: The Ammonian Synthesis

large, black to white, and Boston to Los Angeles). Aristotle introduces the term
metabole to subsume both genesis and the three remaining types of kinesis.
Alexander's innovation is to understand genesis so broadly - as applying not
only to the coming-to-be of substantial forms in bodies, but also to the coming-
to-be of intelligible forms in intellects - that he can now use metabole to refer
to quantitative change, qualitative change and locomotion; to the coming-to-be
of substantial forms; and to the coming-to-be of intelligible forms. In other
words, Alexander uses metabole to cover all types of kinetic change, and all
types of coming-to-be, including non-kinetic transitions such as that from not-
contemplating to contemplating.
Let me be more specific. In his comments on Metaphysics 2.2 (ad 994al9-
b6), Alexander says (in Metaph. 2.2, 153,14-28) that the teleiotes is the end-
state following various metabolai: the metabole from boy to man; the
metabole from learner to one with scientific knowledge; and the metabole from
not-contemplating to contemplating. That is to say, the teleiotes is the end-state
following each of these types of metabole, regardless of whether the metabole is
with respect to growth (auxesis), as in the first case; with respect to quality
(alloiosis), as in the second case; or with respect to form (kat' eidos), as in the
third case.5
The first problem with Alexander's account is his calling the transition from
learner to possessor of knowledge an alloiosis. As mentioned above, Aristotle is
pretty insistent that while the conditions preparatory to this transition may be
alloioseis, the transition itself is not an alloiosis. But this really is an
unsolvable problem for Alexander. He cannot claim that the transition from
learner to someone with scientific knowledge is a genesis, because the transition
consists neither in the coming-to-be of a substantial form (the possessor of
knowledge is as human as he was when he was a learner), nor in the coming-lo-
be of an intelligible form (this is reserved for the transition from not-
contemplating to contemplating). The result is that Alexander is left with
alloiosis. To be fair, there is a line in De Anima 2.5 (417b5-7) where Aristotle
seems to allow for the possibility that the transition from learner to possessor of
knowledge can be seen as a special kind of alloiosis, so Alexander does have
some Aristotelian cover.
The other problem arises from Alexander's nifty move of understanding
metabole kat' eidos as referring to the coming-to-be in the intellect of an
intelligible (noiiumenon). When Aristotle discusses genesis in Physics 5 and

5 Alexander recapitulates this list of examples at in Metaph., 154,25-155,5 and 155,15-30. One

slight lexicographical problem can be detected: unlike in his comments on Metaphysics 2.2,
Alexander's comments on Metaphysics 5.16 use teleiosis, not teleiotes, to describe the completed
state following a quantitative change. Perhaps this is because Aristotle himself uses teleiosis in
Metaphysics 5.16, though Aristotle uses it in the context of the second type of completeness, the
completeness that is reducible to superiority or excellence (arete) in a genus or species. Other
places where Alexander uses teleiotes to refer to an end state following a metabole are Quaest.
l.7, 16,2-3 and 3.3, 86,21-34; and DA 82,2.
 2. Alexander and Themistius 47
distinguishes it from the three kinetic changes, he had in mind the coming-to-be
of a substantial form, not the coming-to-be of an intelligible form. Alexander
could argue in his defense that Aristotle himself maintains that when someone
thinks of an intelligible form, his intellect "becomes" that intelligible form
(e.g., DA 3.2, 426al6 and 3.4, 430a3). Alexander could then claim that he is
simply making the logical inference that the intellect undergoes a kind of
substantial genesis; and that therefore the transition from not-contemplating to
contemplating is a metabole kat' eidos.
Where this line of defense seems most vulnerable is in Alexander's claim that
because the intellect undergoes a metabole kat' eidos, the intelligible form is
the teleiotes of the intellect's metabole from not-contemplating to
contemplating. This is an idea found elsewhere in Alexander's work. 6 But it
seems to me that Alexander also has another type of teleiotes in mind here, and
is conflating the two types of teleiotes into one. The other type of teleiotes
Alexander has in mind here is the good of each thing, a quasi-moral objective.7
The presence of the intelligible form in the intellect following the transition
from not-contemplating to contemplating signifies a teleiotes both in the sense
that intellectual activity is the human good, and in the sense that intellectual
activity is the end-state following the intellect's metabote kat' eidos. The two
types of teleiotes are not identical: the former teleiotes signifies the full and
complete inherence of a substantial form (in this case, "humanness"), while the
latter teleiotes signifies the full and complete inherence of an intelligible form.
This problem becomes acute when I try to distinguish what happens during
self-intellection from what happens during my intellection of other concepts. At
the most abstract level, I can see that (1) if I am thinking about "intellect"
(nous), the coming-to-be in my intellect of the intelligible form of "intellect"
(the nooumenon of nous) will be numerically identical to the coming-to-be in
my intellect of the substantial form of "intellect" (the eidos of nous). By
thinking about "intellect" my intellect becomes an intellect. Therefore the
teleiotes qua complete inherence of the intelligible form "intellect" in my
intellect will be one and the same as the teleiotes qua complete inherence of the
substantial form "intellect" in my intellect.
What about (2) when I am thinking about "human"? Is the coming-to-be in
my intellect of the intelligible form "human" (the nooumenon of anthropos)
numerically identical to the coming-to-be in my intellect of the substantial form
"human" (the eidos of anthropos)? On the one hand, I am committed to the
view that my intellect "becomes" whatever it is thinking about. By thinking
about the intelligible form "human", my intellect will itself become human. In
that sense the coming-to-be in my intellect of the intelligible form "human" is

6 e.g., Alexander, DA 84, 19-24.

7 Alexander, in Metaph. 1.1(ad980a21) 1,4-9 and 411,5; DA (Mantissa) 151,31-152,4; and De
jato 17, 198,1-2 and 36, 211,15.
48 Part I: The Ammonian Synthesis
identical to the coming-to-be in my intellect of the substantial form "human".
Therefore the teleiotes qua complete inherence of the intelligible form "human"
in my intellect will be identical to the teleiotes qua complete inherence of the
substantial form "human" in my intellect.
On the other hand, common sense dictates that it is not my intellect but my
soul-body composite, which most warrants possessing the substantial form
"human". In this sense the coming-to-be in my soul-body composite of the
substantial form "human" is identical neither to the coming-to-be in my intellect
of the intelligible form "human" nor to the coming-to-be in my intellect of the
substantial form "human". Therefore the teleiotes qua complete inherence of the
substantial form "human" in my soul-body composite will be identical neither to
the teleiotes qua complete inherence of the intelligible form "human" in my
intellect nor to the teleiotes qua complete inherence of the substantial form
"human" in my intellect.
Of course, there is a sense in which I, as a human, have achieved a kind of
perfection whenever I contemplate anything. But as mentioned above, this is a
different type of teleiotes. It refers to a quasi-moral good, and not to an end-state
following a transition. My soul-body composite possesses its teleiotes qua
complete inherence of the substantial form "human" as long as it is alive,
regardless of whether I am contemplating or not. But my soul-body composite
possesses its teleiotes qua moral good only when I am actively contemplating.
The reason I have gone into such detail about this issue is that Alexander's
expansion of metabole to cover the transition from not-contemplating to
contemplating as well as coming-to-be in a strict sense, and his concurrent
expansion of teleiotes to cover the possession of intelligible form as well as
substantial form, provided much of the conceptual raw material for later
Neoplatonic ideas about perfection. As I shall show in Chapters 3 and 4, the
Neoplatonists held that during the act of self-intellection, the identity of two of
Alexander's types of teleiotes (teleiotes qua intelligible form of "intellect", and
teleiotes qua substantial form of "intellect") with the third type of teleiotes
(teleiotes qua good of "human") was so strong, that this last type of teleiotes
came to be seen more as the existential good of "human" and less as the moral
good of a human. In other words, self-intellection came to be seen as the primary
mechanism of reversion, of ascending to the next higher level of being, partly
because during self-intellection so many types of perfection are attained. I shall
also explain in Chapters 3 and 4 how the Neoplatonists used the concept of the
specific difference to make the move from seeing teleiotes as a moral good to
seeing it as an existential good.
For my present purpose what is important is once again pointing out
Alexander's use of eidos ("species") instead of Aristotle's genos ("genus") in
commenting on L4. Aristotle's definition of to teleion as that which has
achieved excellence or superiority in a genus is now understood by Alexander as
referring instead to that which has achieved excellence or superiority in a
 2. Alexander and Themistius 49
species. Achieving excellence or superiority in a genus can be interpreted as
referring to the full and complete inherence of a substantial form in the thing in
which it inheres. When related to the genus, the substantial form "human" fully
and completely inheres as long as the human is alive.
Achieving excellence or superiority in a species, on the other hand, can be
interpreted as referring not to the full and complete inherence of the substantial
form "human", but rather to the full and complete exercise of the specific
difference, the essential characteristic which makes the species "human" what it
is: rationality, or the capability to think. When I am actively contemplating, I
am not achieving excellence or superiority in my genus, "animal". I achieve
excellence or superiority in my genus, "animal", when I am actively moving or
sensing, because I am using the two capabilities which are the essential
differentiating characteristics of the genus "animal". But when I actively
contemplate I am achieving excellence or superiority in my species, because I
am using my capability to think, which is the essential differentiating
characteristic of the species "human".
Another respect in which Alexander's terminology prefigures that of the
Neoplatonists can be detected in his assertion that the transition from intellectual
inactivity to intellectual activity is not so much a transition from dunamis to
energeia as it is a transition from disposition (epitedeiotes) to perfection
(teleiotes). This is because the transition is from one thing to itself, and not
from one thing to another. 8
Most importantly, of course, the result of Alexander's broadened understanding
of metabote is that teleiotes can now serve in all but one of the roles Aristotle
appears to demand of entelekheia. As the end-state following a metabole,
teleiotes can serve as the complete entelekheia implied in Physics 3.1 and 3.2,
because metabole covers the genesis of a substantial form as well as the three
kineseis: kinesis with respect to quantity, kinesis with respect to quality and
kinesis with respect to place. Moreover, as the end-state following a metabole,
teleiotes can now serve as the second entelekheia of De Anima 2.1, because
metabole now covers the genesis of an intelligible form as well as the genesis
of a substantial form. Finally, as we saw above, Alexander uses teleiotes to
gloss the first entelekheia of De Anima 2.1, indicating that he saw teleiotes as
referring to the substantial form in and of itself, that is, to form as substance
(ousia) or essence (to ti en einai), and not merely to form as that which comes-
to-be as the result of a metabole.9
But what about change itself? In Physics 3.1 Aristotle defined kinesis, not the
end-state following a kinesis, as an entelekheia. Replacing kinesis with a more
encompassing metabole won't solve this problem. This is because Alexander

8 See Quaest. 3.2, 81,7-82,17 and 3.3, 84,2-27.

9 Other places where teleiotes seems to refer to substantial form qua essence are Alexander,
DA 6,30-7,8; 9,11; and 10,8; Quaest. 2.3, 50,19; and in Metaph. 3.2 (ad 996a20-21), 181,22; 5.1 (ad
I013a20), 347,15-18; and 5.4 (ad 1014bl6), 359,13-25.
50 Part /: The Ammonian Synthesis
has expanded the notion of metabole to allow teleiotes to serve as its end-result,
and not to allow teleiotes to serve as the metabole itself. Yet Alexander seems
to have meant for teleiotes to include within its range of referents the first or
initial entelekheia implied in Physics 3.1 and 3.2, given that he sometimes
uses teleiotes interchangeably with energeia. 10
Inasmuch as we are able to reconstruct his views from Simplicius' and
Philoponus' citations, Alexander was aware of this problem, and says that in
relation to the actual X which is a potential Y, the energeia of the actual X qua
potential Y is a teleiotes. In other words, when I place a log on the fire, the
energeia that is the wood's kinesis from potential fire to actual fire is, in
relation to the actual wood's being a potential fire, a teleiotes. This conflation
of entelekheia as teleiotes and energeia as teleiotes is something Simplicius
takes Alexander to task for, but Philoponus approves of. I I
Simplicius' criticism seems to be based on Aristotle's assertion in L3 that
change is an incomplete energeia. Since Alexander says that change is the
energeia which takes place when an actual X changes to an actual Y, and that
change as energeia is the teleiotes of the actual X qua potential Y, we will be
forced to conclude that, given Aristotle's description of change in L3 as an
incomplete energeia, change is an incomplete completeness. And this rather
meaningless conclusion, as discussed in Chapter 1, is one of the positions we
wanted to avoid being forced into.
According to Simplicius, Alexander's canonical position is that kinesis is the
first or initial transformation (metabole prote) or the first or initial actuality
(entelekheia prote) of what is in potentiality, the last or final (hustate) being
the metabole at the point of teleiotes, completion. He calls the initial metabote
or entelekheia the "route" or "path" (hodos) to the final. 12 In other words,
Alexander avoids using teleiotes to describe change because doing so would
make him vulnerable to the criticism that he held change to be an incomplete
completeness; and as a result he is forced to see change as a process, as the route
or path from potentiality to actuality.
So despite his desire to use teleiotes in the sense of both energeia as well as
entelekheia - to refer to the change itself as well as the end-point that is the
culmination of a change - Alexander seems resigned to giving up on that goal.
In short, Alexander is tom between this:

10 See, for example, Quaest. 3.2, 81,7-82,17 and 3.3, 84,2-27; ap. Simplicium, in Phys. 3.1 (ad

201a9), 414,15-416,6.
11 ap. Simplicium, in Phys. 3.1, 415,29-416,2; and ap. Philoponum, in DA 2.1 (ad 4!2a16),

216,9-22; and 2.4 (ad 415al6), 264,15-19 and (ad 416a34), 284,3.
12 ap. Simplicium, in Phys. 3.1(ad201all),416,27-31.
 2. Alexander and Themistius 51
2b[ii]"
Phys. 1st: entelekheia ateles = teleiotes qua "state of being completed" = state
of incompleteness (e.g., "being constructed")
DA 1st: entelekheia teleia = teleiotes qua "state of having been completed" =
state of completeness (e.g., "having come to know")

Phys. 2"ct: entelekheia teleia = teleiotes qua "state of having been completed"
=state of completeness (e.g., "having been constructed")
DA 2"ct: entelekheia teleia =teleiotes qua "state of having been completed" =
state of completeness (e.g., "having come to contemplate")

which is unsatisfactory because change will now be held to be an incomplete

completeness; and this:

2b[ii)'
Phys. 1st: entelekheia ate/es =teleiosis qua "process of completion" =
incomplete process (e.g., "construction")
DA 1st: entelekheia teleia = teleiotes qua "state of having been completed" =
state of completeness (e.g., "having come to know")

Phys. 2"ct: entelekheia teleia =teleiotes qua "state of having been completed"
=state of completeness (e.g., "having been constructed")
DA 2°ct: entelekheia teleia = teleiotes qua "state of having been completed" =
state of completeness (e.g., "having come to contemplate")

which is unsatisfactory because change will now be seen as a process above and
beyond the changing thing rather than as a state possessed by the changing
thing.
Regardless of Alexander's inability to solve this dilemma, his use of prote
("first" or "initial") and hustate ("last" or "final") to distinguish between
entelekheia as energeia and entelekheia as teleiotes is historically significant.
Assuming that Simplicius is quoting Alexander accurately, the source of a long
tradition among the commentators - including Avicenna and continuing into the
Latin Middle Ages - of asserting that change is a first or initial perfection, has
been detected in the fragments of Alexander's lost commentary on the Physics. It
would also appear that in his frustration at being unable to decide between
position 2b[ii] ,, and position 2b[ii] ', Alexander falls back on position 2b[i],
namely, construing the first/second distinction broadly enough as to make it
apply to change as well as to transitions such as intellection, that is, the
transition from knowing to contemplating.
On the other hand Simplicius may not be reporting Alexander's views
correctly, but imposing a later articulation of the distinction backwards to
Alexander. My own guess is that when Simplicius quotes Alexander as saying
52 Part I: The Ammonian Synthesis

that kinesis is a proten metabolen e proten entelekheian (ap. Simplicium in

Phys. 3.1, 416,28-29), the e proten entelekheian is probably a scribe's or
Simplicius' own gloss. This is because Alexander nowhere else calls motion a
first entelekheia, although he clearly implies that motion is a first metabole. If
this is in fact the case, Themistius will then deserve the honor of being the first
commentator to describe kinesis as a "first" entelekheia.

Themistius (fl. 365 AD) says in his Paraphrase of Physics 3.1 that in each
instance of change the entelekheia is twofold. One type of entelekheia, which
Themistius calls entelekheia prate, "first" or "initial" actuality, refers to the
actuality possessed by the potential thing while it is still changing; while, in
other words, the changing thing retains some part of its potentiality. The other
type of entelekheia, which Themistius calls entelekheia hustate, "last" or
"final" actuality, refers to the actuality possessed by the fully changed thing; to
the actuality, in other words, of the formerly potential but now actual thing,
which retains none of the potentiality it possessed while it was still changing. 13
By Themistius' reckoning, kinesis is a first, or initial, actuality, and consists
in the potential thing's passage (poreia) to the form which lies at the end of the
change. Last or final entelekheia, by contrast, refers to the change (here
Themistius uses metabole) when it is at the point of full inherence of the form
(eis to eidos). Themistius corrects Aristotle by saying that since entelekheia
properly speaking, and in an absolute sense, is form (all' epeide kai to eidos
entelekheian legomen ten kurios te kai haplos), kinesis is really a passage to
an entelekheia, rather than an entelekheia itself. The conclusion one has to
draw, Themistius says, is that kinesis is an entelekheia only in an incomplete
or imperfect sense (houtos oun he entelekheia oukh hos kurios oude haplos
all'hbs ateles). And in an apparent slap at Aristotle in L3, Themistius asserts
that change can be correctly called an incomplete actuality, and it can be correctly
called a complete activity, but it is not correct to call change an incomplete
activity. Themistius' careful distinction between entelekheia and energeia
seems (again, if Simplicius is reporting accurately) to have been taken from
Porphyry. 14
But because he also follows Alexander in using teleiotes to gloss
entelekheia, Themistius is faced with the possible problem that, having
described change as an energeia teleia, teleiotes now appears as a gloss of both
entelekheia and energeia. 15 Another problem is that Themistius seems to have
felt obliged somehow to accommodate Aristotle's bald assertion in L3 that

13 Themistius, in Phys. 3.1(ad20 1a7-16),69,6-20 and (ad201al6-27), 70,5-13.

14 ap. Simplicium, in Phys. 3.1(ad201a9), 415,6-8.
15 Themistius pairs teleiotes with entelekheia at in Phys. 3.6 (ad 206a20-31), 92,9-12 (in a

discussion of the infinite), and pairs teleiotes with energeia at in Phys. 8.4 (ad 254b32-255a32),
218,27-29 (in a discussion of locomotion). The only time Themistius uses teleiosis in the in Phys. is
when he is quoting Aristotle's assertion in Phys. 7.3, that aretai are teleioseis.
 2. Alexander and Themistius 53
kinesis is an incomplete activity (energeia ate/es). Themistius says in his
comments on Physics 3.2 that one type of energeia is complete, the other
incomplete, and that it is possible to speak of energeia properly and in an
absolute sense, yet for it still to be incomplete. 16 What can Themistius mean by
upholding, in his comments on Physics 3.2, precisely what he attacked in his
conunents on Physics 3.1?
It is in response to these problems that Themistius appeals to two
distinctions. The first distinction is that between the two different ways in which
teleiotes can be understood to refer to endedness: being directed towards an
end as opposed to serving as an end. I have already described this first
endedness distinction in Chapter 1.
But Themistius adds a second distinction to the mix, differentiating between
two other modes of endedness, which I shall call "allotelic" endedness and
"autotelic" endedness. Themistius says that in one sense an energeia is teleia
when it is static, or at rest, and when the energeia consists in, for example, the
functioning of an organ. When the eye makes the transition from not-seeing to
seeing, there is no kinesis, and the eye remains static. The energeia in this case
is teleia not in the sense that it is complete or terminated, but rather in the
sense that the activity of seeing is autotelic, that is, in the sense that this
energeia is an end in itself. When the eye rotates, on the other hand, the
energeia of the eye is a kinesis, and is ateles in the sense that for the eye, the
activity of rotation is allotelic, not autotelic. That is to say, the rotation of the
eye is not an end in itself but only a means to another end, namely, the end-state
following the kinesis kata topon (i.e., locomotion) of rotation. 17
What evidence might Themistius have adduced from the Aristotelian corpus in
order to support his new distinction between autotelic and allotelic endedness?
Answering this question will require me to leave Themistius aside for a while
and return to Aristotle. To my mind the main sources of the distinction between
autotelic and allotelic endedness can be found in the Nicomachean Ethics. At
the very beginning of that work, in EN 1.1, 1094a3-5, Aristotle says that some
ends are nothing more than activities (energeiai), while other ends are products
(erga) which are above and beyond (para) activities. A commentator could infer
from this that some activities serve as their own ends, that is, that some
activities are ends in themselves. Other activities, by contrast, are directed at ends
which are produced as a result of those activities. In short, activities are either
autotelic or allotelic.
Reinforcing this inference are a number of assertions and arguments Aristotle
makes later on in the Nicomachean Ethics. At EN 1.7, 1097a25-b6, Aristotle
expands on the distinction he made in EN 1.1, saying now that some ends are
means to other ends, while other ends are ends in themselves. Wealth is an end

16 Themistius, in Phys. 3.2 (ad 202a3-7), 75,13-14.

17 Themistius, in Phys. 3.2 (ad 201b29-202a3), 74,11-23.
54 Part /: The Ammonian Synthesis

in the former, allotelic sense, while happiness is an end in the latter, autotelic
sense. Aristotle then goes on to claim that the more a thing is an end in itself
and not sought or chosen for the sake of something else, the more teleion that
thing will be.
The result is that instead of a twofold distinction we now seem to be presented
with a threefold distinction, between what one scholar has recently tenned
"subservient", "subordinate" and "ultimate" senses in which something may be
said to be teleion. 18 To apply this threefold distinction once again to energeiai,
a subservient energeia will be chosen or sought only for the sake of some
product that comes into being as a result of the energeia. In other words, a
subservient energeia is entirely allotelic. For example, a surgeon's act of
cutting is never an end in itself, but only ever a means to a further end, namely,
the patient's health. A subordinate energeia will serve as an end in itself in one
respect, but in another respect be chosen or sought for the sake of an end which
is above and beyond it. In other words, a subordinate energeia is both autotelic
and allotelic. In one respect health is an end in itself, since a surgeon's act of
cutting is directed at nothing above and beyond health; and in another respect
health is sought not for its own sake but rather for the sake of something above
and beyond health, namely, happiness. An ultimate energeia will only ever be
sought or chosen for its own sake, and never as a means to a further end. In other
words, an ultimate energeia is entirely autotelic. Happiness is only ever sought
for its own sake, and never for the sake of something above and beyond it.
Given Aristotle's claim that the more a thing is chosen or sought for its own
sake and not for the sake of something which is above and beyond it, the more
teleion it will be; and given his claim later in the chapter (EN 1. 7, 1097b6-2 l)
that happiness is not simply the thing which is most teleion but also the thing
which is most autarkes ("independent" or "self-sufficient"); a commentator could
fairly infer that what being teleion and being autarkes had in common was the
fact that it is in virtue of satisfying these two criteria that happiness enjoys the
characteristic of being to the greatest degree uncaused. This is because
happiness, being entirely autotelic, will never be the effect of a final cause which
is above and beyond it; that is to say, happiness is uncaused because it never
depends upon a further final cause to bring it into being. 19
It is true that Aristotle describes to autarkes in this chapter (EN 1. 7,
1097b14-15) merely as "that which makes a life uniquely worthy of choice and
lacking in nothing", a description that would seem to support an understanding
of teleion as quantitatively complete rather than autotelically ended. But when

18 Keyt 1978.
19 Here our commentator would disagree with Heinaman 1988, who takes teleion in the
Nicomachean Ethics to refer to supremacy in a genus, i.e., the second sense of teleion which
Aristotle offers in L4. This position is modified by Santas 1989, who argues that this sense of
teleion fits best when Aristotle is articulating his own conception of the good from EN 1.7 onwards
(as opposed up to EN 1.7, when Aristotle is merely talking about the concept of the good).
 2. Alexander and Themistius 55

Aristotle rev1Slts the question of happiness in EN 10.6-8, he articulates a

position which supports an understanding of to autarkes as uncaused. 20
I hesitate to put my oar into these choppy waters, since it bears upon a
fundamental disagreement between Aristotle scholars who think that Aristotle
held an "inclusivist" or "comprehensive" view of happiness, and those who think
that Aristotle held an "exclusivist" or "dominant" view of happiness. According
to the inclusivist view, happiness consists in the sum of all activities which are
in accordance with the virtues; this view seems to govern Aristotle's discussions
of happiness in EN 1. According to the exclusivist view, happiness consists in
the one activity - contemplation - which is in accordance with the highest
virtue; this view seems to govern Aristotle's discussions of happiness in EN 10.
The suggestion that to teleion should in many instances be understood as that
which is ended - and that to teleiotaton, by extension, should be understood as
that which is ended to the highest degree, namely, that which is most autotelic -
is more in accordance with the exclusivist view of happiness. This is because
understanding to teleiotaton as that which is most autotelic - and hence as that
which is most uncaused by any final cause which is above and beyond it -
derives from the sense of teleion which is most reconcilable with autarkes,
understood as that which is causally self-sufficient and thus uncaused. In other
words, an exclusivist could argue that happiness consists solely in
contemplation, by appealing to the fact that contemplation is the most autotelic
human activity. That is, given the fact that contemplation is never the effect of a
final cause which is above and beyond it, contemplation is the human activity
which is to the greatest degree causally self-sufficient.21 By contrast, an
inclusivist will tend to understand teleion as "quantitatively complete", since an
inclusivist argues that happiness consists in the sum of all activities which are
in accordance with virtue. In other words, happiness is teleion in the sense that
happiness can be viewed as a whole none of whose parts is missing.
Regardless of whether the inclusivist or exclusivist view more accurately
reflects what Aristotle really thought happiness was, the fact remains that the
Neoplatonists gravitated towards the exclusivist view, for reasons I shall discuss
in Chapters 3, 4 and 6. For the moment, I should just mention Aristotle's claim
(EN 10. 7, l l 77b26-78a2) that contemplation, the epitome of the completely
happy life, is the most divine of all human activities and is thus the one activity
that makes us humans in some sense immortal. This assertion provided the
Neoplatonists with a clear piece of Aristotelian evidence in support of their

20 I agree with Curzer 1990 that Aristotle' s description of being autarkes and being teleion - the

two main criteria for happiness - in EN 10.6-8 is a reconsideration, rather than a continuation, of
his description of to autarkes and to teleion in EN 1.7.
21 Aristotle says as much at EN 10.7, l 177bl6-20. Surprisingly, exlusivists such as Heinaman

1988 fail to appeal to the notion of autotelic endedness to support their own attack against the
inclusivists.
56 Part /: The Ammonian Synthesis

notion of reversion (epistrophe), according to which all things have a irreducible

tendency upwards along the chain of being.
Once a commentator felt justified in construing teleion as "autotelically
ended" - being an end in itself - he could be confident in also construing ateles
as "allotelically ended". He could support this move by appealing to Aristotle's
assertion (EN 10.4, 1174a14-23) that pleasure, like seeing, is an autotelic
(teleion) activity. This is in contrast to kineseis, which always appear to be
allotelic (ateles), given the fact that kineseis - e.g., housebuilding - are directed
towards an end - shelter - which is above and beyond the activity of the kines is
itself. I admit that this is not the sense of teleion and ateles which Aristotle
himself had in mind in this passage. In fact Aristotle is quite clear that he is
thinking in terms of completeness and incompleteness rather than in terms of
autotelic endedness and allotelic endedness. Nevertheless a commentator who,
like Themistius, had committed himself to construing teleion and ateles - at
least in some cases - in terms of autotelic and allotelic endedness, could easily
apply that new understanding to the difference between energeiai such as seeing,
which is an autotelic activity (though not the most autotelic activity), and
kineseis such as the eye's rotation, which is entirely allotelic.
How did adding autotelic and allotelic endedness to his quiver of distinctions
help Themistius integrate the entelekheiai of Physics 3.1 and 3.2 with those of
De Anima 2.1? In his synopsis of De Anima 2.1, Themistius starts by saying
that in contrast to matter, which is pure disposition, form (eidos) is a teleiotes,
in the sense that it is form which represents the fulfillment (anaplerosis) of a
thing's innate disposition, and it is form which represents the "leading-on"
(proagoge) of a thing towards its end. Themistius goes on to say that qua
fulfillment, the eidos is the thing's substantial form, and qua "leading-on", the
eidos is the thing's structure or shape. A thing composed of form and matter,
Themistius says, is said to be complete (teleion einai) in view of its substantial
form (para tou eidous), and to have "end-in-itself-ness" (entelos ekhein) in
view of its structure or shape (para tes morphes).
I take this to mean that a thing may be said to be teleion in the sense of
"completed" when its substantial form has fully inhered in its subject. This is
what Themistius had in mind in his Physics paraphrase when he said that the
end-state following a metabole is an entelekheia in the proper sense. A thing
may also be said to be teleion in the sense of "ended". It is in view of its
structure or shape that a thing's function becomes apparent. The shape or
structure of the eye is ended in the sense that the matter of the eye is arranged in
the way it is in view of the function - seeing - which the eye performs. To sum
up: when the form of a thing is seen as its substance or essence, it is a teleiotes
because it refers to the state of completeness following a change; when the form
of a thing is seen as its shape, on the other hand, it is a teleiotes because it
refers to the endedness of the potentiality or capability represented by possessing
that shape.
 2. Alexander and Themistius 57
In both of these senses, the soul is an entelekheia, which Themistius takes to
mean "the state of completeness-and-endedness" (ten hexin tes teleiotetos = [fa-
laysa yadullu qawluhu anfaliikhiyii ft lisiini l-Yuniiniyyfn caza shay'in
ghayri] malakati l-kamiil). 22 The soul is the substantial form of the body and in
that sense is the end-state of the metabole by which the matter of the body came
to be informed by the soul. The soul is also a structure of faculties or
capabilities which is ended in the sense that these faculties or capabilities are
either allotelic, in that their activities are directed towards the coming-to-be of
new actualities (as with the faculty of reproduction, whose activity is directed
towards the coming-to-be of offspring); or autotelic, in that their activities are
ends in themselves (as with the faculty of intellection, whose activity is directed
at no end other than itself).
In other words, Themistius has combined the idea of completeness with the
idea of endedness (both in the sense of being directed at an end and in the sense of
serving as an end), and then added the allotelic/autotelic distinction, to create a
new matrix of all the different ways of being a teleiotes. With the
autotelic/allotelic distinction in hand, Themistius thinks he is better able than
Alexander to construct a scheme flexible enough to accommodate all the senses
in which an entelekheia can be understood as a teleiotes.
In particular, applying the allotelic/autotelic distinction to the directed-at-an-
end half of the directed-at-an-end/serving-as-an-end distinction allows Themistius
to hold that change is in one sense teleion - meaning "directed at an end" - and
in another sense ateles - meaning allotelic, i.e., "directed at an end other than
itself'. This move allowed Themistius to skirt the major roadblock - reconciling
Aristotle's assertion in L2 that change was an entelekheia with his assertion in
L3 that change was ateles - that had stood in the way of Alexander's idea of
appealing to teleiotes to effect the larger reconciliation of Aristotle's use of
entelekheia in L1 and his use of entelekheia in L2. Now, Themistius can hold
that change is a teleiotes ateles without making a fool of himself. This is
because that phrase need no longer be understood - and ridiculed - as some
paradoxical "incomplete completeness'', but can now be construed as referring
instead to allotelic endedness, to the quality of being directed at an end other than
itself.
What is more, by applying the autotelic/allotelic distinction to the serving-as-
an-end half of the directed-at-an-end/serving-as-an-end distinction, Themistius is
able to accommodate entelekheiai such as post-change end-states, which serve
as an end for the changes that produced them; and he is also able to accommodate
entelekheiai such as intellection which serve as their own end. Thus:

22 Themistius, in DA 2.1 (ad 412a2-19), 39,3-22 at 39,20 (= Thamas\iyiis, Sharl:i Kitab an-nafs li-

Aris{u{alis, 42,9-44,4 at 44, I).

58 Part /: The Ammonian Synthesis

2b[ii]"
Phys. l":
l. entelekheia ateles = teleiotes qua "state of being completed" = state of
incompleteness (e.g., "being constructed")
2. energeia ateles = teleiotes qua "state of being directed towards an end other
than itself'= state of allotelic endedness (e.g., "being constructed into a
house")
DA 1":
la. entelekheia teleia = teleiotes qua "state of having been completed"= state
of completeness (e.g., "having been born") [identical to Phys. 2"d (l)]
1b. entelekheia ate/es =teleiotes qua "state of being directed towards an end
other than itself'= state of allotelic endedess (e.g., "capable of reproducing
another")
2a. entelekheia teleia = teleiotes qua "state of having been completed" = state
of completeness (e.g., "having come to know")
2b. entelekheia teleia = teleiotes qua "state of being directed towards an end
which is nothing other than its own exercise" =state of autotelic endedess
(e.g., "capable of contemplating")

Phys. 2"d:
1. entelekheia teleia = teleiotes qua "state of having been completed" = state
of completeness (e.g., "having been constructed")
2. entelekheia ate/es = teleiotes qua "state of serving as an end for something
other than itself'= state of allotelic endedness (e.g., "having been constructed
into a house")
DA 2"d:
l. entelekheia ate/es = teleiotes qua "state of being directed towards an end
other than itself'= state of allotelic endedness (e.g., "reproducing")
2. entelekheia teleia = teleiotes qua "state of serving as its own end" = state
of autotelic endedness (e.g., "contemplating")

The fly in the ointment is Themistius' uncertainty over how interchangeable

entelekheia and energeia are in the definition of kinesis. Nevertheless, the
upshot is that when teleiotes is interpreted - as Themistius seems to have done
- as referring to completeness as well as to the various types of endedness
(endedness in the sense of being directed towards an end as opposed to
endedness in the sense of serving as an end, and autotelic endedness as opposed
to allotelic endedness), that very elastic term can serve in all the roles Aristotle
demands of an entelekheia.
One objection which could be raised to Themistius' interpretation is that the
fundamental idea of endedness is based on an incorrect reading of Aristotle. This
is because when Aristotle used the term telos in L4 it is pretty clear that he
meant for it to be understood as referring to limit or terminus rather than purpose
 2. Alexander and Themistius 59

or that-for-the-sake-of-which. I would argue in Themistius' defense that the

passage in question (Metaph. 5.16, 1021b23-30) can be read as two distinct
sentences. The first (Metaph. 5.16, 1021 b23-25) is neutral as to the sense of
telos (either qua to hou heneka or qua peras is appropriate), while the second
(Metaph. 5.16, 1021b25-30), it is true, does foreground the "limit" sense of
telos. To repeat what I said in Chapter 1, a commentator could understand
"Those things in which their worthwhile end exists are also said to be complete,
for they are complete by virtue of the end which they possess (eti hois
huparkhei to telos, spoudaion on, tauta legetai teleia · kata gar to ekhein to
telos teleia = wa-ayrj,an tuqiilu tiimmatan allat'i Lahti tamiimun farj,ilun fa-
innahii tiimmatun bi-tamiimihii lladhi huwa lahii)" (Metaph. 5.16, 1021b23-25
= Mii ba'da !-fabi'a JI, 622,2-3), as referring to telos in the sense of purpose or
that-for-the-sake-of-which (to hou heneka), with just as much reason as he could
understand it as referring to telos in the sense of limit or terminus (peras).
 3. Proclus, Ammonius and Asclepius
The Neoplatonic Turn to Causation

In trying to speak clearly about a very complex and opaque topic I may have
given the impression that Themistius' expanded understanding of endedness
represents a full-fledged theory. It does not. But however patchy it may in fact
have been, the ways in which Themistius understood endedness nevertheless
provide evidence of a subtle but important shift in ancient discussions of
entelekheia. This is because Themistius' new use of teleiotes to glue together
Aristotle's problematic applications of entelekheia to both psukhe and kinesis,
is one of the first indications that teleiotes had begun to have causal
significance. Specifically, Themistius' appeal to the various senses in which
teleiotes refers to endedness - both in the sense of serving as an end and being
directed towards an end, and in the sense of autotelically and allotelically -
is a sign that Aristotle' s concepts of actuality and activity were on their way to
being viewed less as states of being and more as causes of being, and
particularly as final causes of being. Themistius' move towards viewing
teleiotes (and hence entelekheia) in causal terms, shows the influence of
Plotinus and especially Porphyry, whose distinction between a complete
energeia and an incomplete entelekheia I have already mentioned. 1 The precise
nature of this turn from actuality and activity to causation is the topic to be
examined in this chapter.
To be sure, I am not proposing that the Neoplatonists made the move to
causation simply for its own sake. The move to causation seems instead to have
been a surface tremor arising from a deeper, tectonic shift; to have been little
more, in fact, than a new use of a couple of Aristotelian and Alexandrian terms -
teleion and teleiotes - as part of a larger effort to reconcile Platonic and
Aristotelian cosmologies and theories of the soul. In particular, Neoplatonists
applied teleion and teleiotes in their new, causal sense in order to blur the sharp
distinctions between the Platonic (as understood by Plotinus) core doctrines of
the separability of the soul and God's efficient and final causation of the world's
existence, and the Aristotelian (as understood by Alexander) core doctrines of the
inseparability of the soul and God's final causation of the world's motion.
As far as the issue of the separability of the soul is concerned Plotinus ·is well
known for having devoted a section of his Enneads to attacking Aristotle's
application of entelekheia to the soul. The reason was that the definition of

1 Another hint in Themistius of Neoplatonic influence (and one which comes into full bloom in

Proclus) is the idea that the contraries between which change takes place can be distinguished into
better (ameinon) and worse (phauloteron): the better contraries are to eidos, in the case of
substantial change; to teleion, in the case of quantitative change; to melan, in the case of qualitative
change; and to ano, in the case of locomotion. Their base counterparts are, respectively, he
steresis, to ate/es, to leukon, and to kato.
62 Part /: The Ammonian Synthesis

soul as entelekheia did seem to commit Aristotle - at least in the eyes of

interpreters such as Alexander - to an openly anti-Platonic position, namely,
that the soul is inseparable from the body just as form is inseparable from
matter. This of course is at odds with Platonic psychology, which held that the
soul pre-exists the body and survives its death.
Here is Plotinus' attack on Aristotle's entelekheia, as expressed in its Arabic
paraphrase, the Theology of Aristotle:

L7
(Ps.-)Aris~utalis, Uthulujiya 3, 54, 7-55, l 9
corr. Plotinus, Enneads IV.7.8(5),1-18 2
If they say: The principal philosophers agree that the soul is the
perfection of the body [tamamu l-badani], and perfection is not a
substance, and therefore the soul is not a substance, since the
perfection of a thing is from the substance of the thing, we reply:
We must investigate their argument that the soul is a perfection,
and in what sense they call her "entelekheia" [innahu yanbaghi an
naft.zasa 'an qawlihim inna n-nafsa tamiimun ma wa-bi-ayyi l-
ma'ani sammawha anralashiya]. We say that the principal
philosophers have mentioned that soul is to body as a form
whereby the body becomes ensouled, just as matter by means of
form becomes body: but though the soul is form to the body, yet
she is not a form to every body qua body but is a form only to the
body that is alive in potentiality. If the soul is a perfection after
this description she is not of the class of bodies, for if she were a
form to the body like the form that is in a brazen idol, then if the
body were divided and split up she too would be divided and split
up, and if one of the members of the body were cut off part of her
would be cut off too, and this is not so. Therefore the soul is not a
perfecting form [laysati n-nafsu idhan bi-$uratin tamiimiyyatin]
like the natural and the artificial form; no, she is a perfection
because it is she that perfects the body so that it comes to possess
perception and mind. We say: If the soul is a form, adhering and
not separating, Like the natural form [in kiinati n-nafsu $Uratan
liizimatan ghayra mufaraqatin3 ka-$-$iirati !-fabi'iyyati], then how
does she withdraw in sleep and separate from the body without
quitting it? Similar too is her action in the waking state when she
returns to herself - for she often returns to herself - and casts off
bodily affairs, though that is apparent from her activity only by
night because of the quiescence of the senses and the cessation of
their activities.
If the soul were a perfection of the body qua body, she would not
separate from it [law kiinati n-nafsu tamaman li-1-badani bi-annahu

2 The translation, with minor changes, is that of Lewis 1959, 209-211. The italicized lines are

those bits of the Uthu/Ujiyii which Lewis saw as corresponding directly to words, phrases and
sentences in the Enneads.
3 I have transliterated Lewis' "separating" (it should really be "separable" or "separated") as

the passive participle mufaraqa rather than the active participle mufariqa because in general the
terrn translates the Greek exeiremenon, which is the passive participle of exairein.
 3. Proclus, Ammonius and Asclepius 63
badanun la-mii foraqathu] and would not know the remote, but
would know only things present, like the cognition of the senses;
then she and the senses would be one thing, and that is not so
because the soul recognizes a thing although it be remote from
her, and recognizes the impressions that the senses receive, and
distinguishes them, as we have often said. The function of the
senses is merely to receive the impressions of things, while
recognition and discrimination belong to the soul.
We say that if the soul were a natural perfecting form [law kiinati
n-nafsu ~uratan tamiimiyyatan rablciyyatan], she would not oppose
the body in its desires and many of its activities, indeed she would
not oppose it in anything at all, and, if any affection were
impressed on the body, that affection would be in the soul too, and
man would have sense-perception only, because the province of
the body is sense-perception; knowledge, thought and reflection
are not within its province. The Materialists recognized that, and
consequently they are compelled to acknowledge a second soul and
a second mind that does not die. Now we say that there is no
second soul other than this rational soul which is in the body at
the moment, and she it is of whom the philosophers say that she
is the "entelekheia" of the body [anraliishiyii 1-badani]. though
they speak of her as "entelekheia" and a perfecting form in a way
different from that in which the Materialists speak [ghayra
annahum inna-mii dhakarii annahii anfaliishiyii wa-~uratun
tamiimiyyatun bi-naw<in iikhara ghayri n-naw<i lladhi dhakarii 1-
jirmiyyuna]. I mean that she is not a perfection like the natural
perfection, which is passive, but is an active perfection, that is,
she makes perfection. In this sense do they say that she is the
perfection of the natural organic body, potentially ensouled.

Others have analyzed Plotinus' attack.4 For my present purpose, what is

interesting is that in the Greek original from which this passage of the
Uthulujiyii is derived, Plotinus himself does not use the term teleiotes to
describe the soul as an alternative to the despised entelekheia. This is perhaps
because, as we saw, Alexander had tied the two terms so closely in his own
works on the soul, works in which Alexander upheld the soul's inseparability
from the body even more adamantly than Aristotle himself had. Yet in the
Arabic version, the term tamiim - the standard early translation of teleiotes - is
used extensively. As I shall show, this is because in the six centuries between
Plotinus and the Arabic translation movement, teleiotes had become such a
standard way for Neoplatonists to describe the soul while avoiding the use of the
troublesome entelekheia, that the Arabic term tamiim, which was used to
translate both teleiotes and entelekheia during the early period of translation,
when the UthU!Ujiyii was being put together, appears everywhere in that text. If I
am right, it is one of the ironies of the history of philosophy that the term
teleiotes, first introduced and promoted by the arch-Peripatetic Alexander - a

4 See Bruni 1960; Hager 1964; Verbeke 1971; and Emilsson 1991.
64 Part /: The Ammonian Synthesis

philosopher fully committed to the idea that the soul was inseparable from the
body - was seized on by later Neoplatonists as a way to introduce and promote
Platonic-Plotinian ideas about the separability of the soul.
As I mentioned, the Neoplatonic tum consisted in a move from Aristotle's
more natural-philosophical uses - in the context of motion and the soul - of
entelekheia and energeia, terms indicative primarily of states of being (viz.,
actuality and activity); to the Neoplatonists' more metaphysical uses of both
teleiotes, a term which came to be used to express the idea of the final cause of
being, and of teleion, a term which came to be used to express the idea of the
efficient cause of being. This tum to causation took place over several
generations of Neoplatonist philosophers. These include Syrianus (d. ca. 437);
his student Proclus (d. ca. 485); Proclus' student Ammonius (d. ca. 514); and
Ammonius' students Asclepius (fl. 525) and Philoponus (d. ca. 570), through
whose apo ph6nes ("lecture-note") commentaries on the Metaphysics and De
Anima, respectively, we can reconstruct Ammonius' own philosophy.
The transformation that occurs during this period is a complex and profound
story, and I do not want to give the impression that what I shall present here is
an exhaustive description of how and why it happened; that is a task for
specialists in Greek Neoplatonism. For now, what is important about this move
- a manifestation of what I referred to in the Introduction as the "Ammonian
synthesis", by which I mean the project of the Aristotle-commentator
Ammonius and his students to integrate the greater sumphonia of reconciling
Aristotle and Plato into the lesser sumph6nia of reconciling Aristotle and
Aristotle - is that it is crucial but unappreciated background both to the word
choices made by the various Greco-Arabic translators, and, partly for that reason,
to the theories of the soul and of causation of the Arabic philosophers al-Farabi
and Avicenna.
The gradual subsuming of entelekheia under teleiotes occurred because a
Neoplatonist's use of entelekheia could be attacked as committing him to two
unsavory theories (unsavory to a Neoplatonist, that is): to an ontology of the
soul according to which the soul is understood as being inseparable from the
body, and to a cosmology according to which God's causation of the world is
understood as an instance of the final causation of motion or change, rather than
as an instance of the efficient and final causation of existence. By contrast, using
teleiotes rather than entelekheia allowed a Neoplatonist - as I shall show - to
advocate the separability of the soul from the body, and the causation of
existence rather than of motion; but it also left a number of loose ends, which
the later Neoplatonists, including Avicenna, tried to tie up.
The tum to causation begins with the identification of being in a state of
activity (kat ' energeian or energeiai) or being in a state of actuality (kat'
entelekheian or entelekheiai) and being perfect (teleion). Now "that which is
in a state of activity" or "that which is in a state of actuality" are not included
among the meanings of to teleion offered by Aristotle in L4. Nevertheless it is
 3. Proclus, Ammonius and Asclepius 65
not too much of a stretch to connect to teleion with those two senses, for the
following reasons.
First of all, if - as Alexander tells us - Aristotle thought that an entelekheia
such as the soul was nothing other than a teleiotes, and if to teleion is the
substantive adjective whose abstract noun is teleiotes, then that which is kat'
entelekheian or kat' energeian will similarly be to teleion. Second, and more
importantly, a characteristic of something which is kat' entelekheian or kat'
energeian is that it stands as a cause to the thing which it is in contact with and
which is changing from being kata dunamin to being itself kat' entelekheian
or kat' energeian. That is perfectly Aristotelian: an actual or active fire is the
cause of the potential fire's changing into actual or active fire.
But does a perfect thing likewise stand as a cause to something which is
imperfect? Again, even though standing as a cause to something which is ateles
is not one of the meanings of teleion offered by Aristotle in L4, connecting the
two concepts is quite straightforward. As was discussed above, one of the
meanings of teleiosis is the process of concoction by which an immature living
thing (an embryo or a child) becomes fully developed or mature (a fetus or an
adolescent). The teleiotes, or state of complete development, possessed by a
fetus which is at the point of being born refers mainly to its possession of all its
component parts, a numerical wholeness accommodated in Aristotle's list of
meanings in L4. The teleiotes of an adolescent, however, refers not to its
possession of all its parts - it has precisely as many parts as the fully developed
fetus - but rather to its sexual maturity. That is, teleiotes in this instance refers
to the sexually mature thing's ability to reproduce, to bring something like it
into existence; and hence teleiotes refers to its causality.
The Neoplatonists reckoned that if teleiotes can be seen to refer to causality
in this way, it follows that something possessing teleiotes - namely, to teleion
- can be seen to refer to cause. 5 In fact, it is a Neoplatonic cliche to speak of

5 The perfect's necessary causation of other things is alluded to by Plotinus, Enneads, V.1.6,37-
38 (kai panta de hosa ede teleia genniii); V.4.1,23-39 (esp. 26-28: ho ti d 'an ton allon [viz., other
than the First] eis teleiosin iei, horomen gennon kai auk anekhomenon eph' heautou menein, all'
hereron poioun) (the entire passage is compressed into al-ffl'ilu 1-awwalu yabqa 'ala l:ziilihi sakinan
qfl'iman tflmmanfa-yaJ:uiuthu min tamflmihifi'lun: R. ft 1-'ilm al-ilahl, 179,22); and V.9.4,3-12 (au
gar de, h6s oiontai, psukhe noun teleotheisa genniii; pothen gar to dunamei energeiiii estai, me tau
eis energeian agontos aitiou ontos? .... Dia dei ta prota energeiiii tithesthai kai aprosdea kai teleia;
ta de atele hustera ap' ekeinon, releioumena de par' auton ton gegennekoton diken pateron
teleiounton, ha kat' arkhas acete egennesan .... ) (the beginning of the passage corresponds to wa-
laysa ka-ma "{.anna nasun anna n-nafsa idha tammat wa-kamilat waladati I- 'aqla: R. Ji l-' ilm al-ilflhi,
168,17). Syrianus introduces the causality of the perfect into an Aristotelian context by explicitly
equating the perfect's causality of the imperfect with the causality enjoyed by what is entelekheiiii
towards what is dunamei: in Metaph. 14.5 (ad 1092al 1), 185,29-186,14. The two concepts are
further integrated by Proclus: Theol. Plat., Vol. 1.6, 27,15-17; 5.20, 73,14-15; Inst. Theol., Prop. 24,
28,8-13; 25, 28,21-30,4; 27, 30,25-32,2; 77, 72,20-74,7; in Ale. I, 235,10-12; in Parm., 797,35-38;
and in Tim., Vol. I, 297,2-18; 2, 56,12-68,5; 3, 22,20-28. Asclepius lists productivity as one of the
criteria of the perfect in his commentary on L4: in Metaph. 5.16 (ad l02lbl2), 338,14-339,7
66 Part /: The Ammonian Synthesis

how the perfect is "fecund" or "fertile" (gonimon). 6 By contrast, something

imperfect, which does not possess teleiotes but which has a disposition or
suitability (epitedeiotes) to attain that teleiotes, can be seen to refer to effect. 7
While connecting actuality or activity with perfection via their common
causality may have been a nifty move, it raised many questions, the answers to
which had profound ramifications. Analyzing how late-antique and early-Arabic
philosophers up to Avicenna tidied up the messy after-effects of this move, and
dealt with the further implications of their answers to those questions, is the
subject of the rest of Part I.

Perhaps the first and most obvious question raised by connecting the perfect and
perfection to causality is, what kind of cause is the perfect, and what kind of
cause is perfection? In other words, how exactly can what is perfect be described
as producing its effect; and how exactly can perfection be described as producing
its effect? Before trying to answer that question, it would be a good idea to say a
few words about the term "cause". Recently, the trend among scholars of ancient
philosophy has been to emphasize the fact that the Greek terms aitia and aition
ought not to be seen solely as referring to a cause which produces its effect, but
should sometimes be seen as referring to an explanans or explanatory factor
which explains its explanandum or thing to be explained.8 In other words,
instead of asking how the perfect and perfection produce their effects, we should
perhaps be asking how the perfect and perfection explain their explananda. But
it is unclear whether our understanding of late-antique discussions of causation
should automatically be revised in light of the recent work done on classical
theories of causation and explanation.
On the one hand, the more general Neoplatonic discussions of the fecundity
and productivity of the perfect seem to indicate that what the Neoplatonists are
primarily focused on is how the cause produces its effect in the real world. On
the other hand, Neoplatonic talk about perfection's acting as a final cause, as
something which stands as a goal or object of desire in relation to its effect (in
contrast with the perfect, which stands as a generator or producer in relation to
its effect), will sound less counterintuitive to us if we imagine the Neoplatonists
to be thinking of perfection as an explanans rather than as a cause. In this
respect the debates about whether Aristotle, in particular, thought in terms of
causes or explanations can also be applied to the Neoplatonic discussions. This
is because those Aristotle scholars who today find his theory of final causality

(theoria) and 340,32-341,25 (Lexis); at 339,3-7, Asclepius says that we call "perfect" that which is
Ca!Jable of perfecting another (to dunamenon allo teleioun).
6 On the fecundity of the perfect see Proclus, Theo/. Plat., Vol. 2.7, 50,3-51,6; 4.3, 16,3; 5.5,
24,17; and in Tim., Vol. 1, 25,14-18.
7 See Proclus, in Remp., Vol. l, 218,3-4; Inst. Theo/., Prop. 79, 74,18-26; in Parm., 668,4-19.
8 Hocutt, 1974; Mure 1975; Moravcsik 1975; Frede 1980; Sorabji 1980, 3-69; Schofield 1991;

Moravcsik 1992; and Freeland 1992.

 3. Proclus, Ammonius and Asclepius 67
philosophically incoherent, metaphorical, or superfluous, tend to assume the
correctness of the "cause" view.
By contrast, those Aristotle scholars interested in defending the coherence of
teleology, its essential role in Aristotle's thought, and its complementarity with
efficient causality, tend to advocate the "explanation" view. There is nothing
surprising in this, of course, since in a competition between the efficient cause
and the final cause over which is more cause-like, the efficient will always be
judged the winner by those with modem intuitions that what a cause causes is an
event. Because the cause of an event must exist before the event, and because the
efficient cause is temporally prior to its effect, while the final cause is either
simultaneous with or temporally posterior to its effect, the efficient cause will
appear the more cause-like to modem eyes. We are forced in the end to conclude,
rather spinelessly, that the Neoplatonists are inconsistent on whether "cause" or
"explanation" is at issue: sometimes they seem to be referring to causation,
particularly when speaking of the fecund efficient causation of the perfect; and
other times they seem to be referring to explanation, particularly when speaking
of the complementarity between the way that the perfect operates and the way
perfection operates. What I shall do, then, is begin by referring to the perfect and
perfection as causes, and then later tum to describing the complementary ways
they explain their respective explananda.
What does seem clear is that Neoplatonists such as Syrianus, Proclus,
Ammonius and his students, interested not only in Platonizing Aristotle but also
in Aristotelianizing Plato, worked very hard to apply Aristotle's four-cause
theory to the cosmology they inherited from Plotinus and which derived many of
its basic principles from Platonic works such as the Phaedo, Phaedrus,
Philebus, Timaeus and Parmenides, the last two of which served as the
culmination of the Neoplatonic curriculum. I have mentioned above that the
Neoplatonists thought that the perfect operated as an efficient cause and that
perfection operated as a final cause. What prompted them to choose the efficient
cause for the perfect and the final cause for perfection? Why were the other two
Aristotelian causes - the material and the formal - not in the running?
Answering that question is complicated by three corollary factors. The first
factor is that later Neoplatonists actually had two extra causes to choose from
apart from the four Aristotelian causes. One was the paradigmatic or exemplary
cause, invented to accommodate the separate Ideas, which - unlike the
inseparable formal cause - transcended their effects. The other extra cause was the
instrumental cause, inherited from the Middle Platonists and designed to serve in
the role of efficient cause in the sublunary world. 9
The second factor is that according to the scheme which appears to have been
invented by Syrianus and which was expanded upon by Proclus, the final,

9 On the history of the paradigmatic cause, see Hankinson 1998, 326-327 and 351; on the

instrumental cause, see 342 and 380-381.

68 Part /: The Ammonian Synthesis

paradigmatic and efficient causes were causes in a proper or strict sense (kurios)
and inhabited - in the order just listed, from highest to lowest - the superlunary
world of eternal existence. The formal, instrumental and material causes, by
contrast, were not really causes at all but rather conjoint causes or causal factors
(sunaitia) and inhabited the sublunary world of coming-to-be and passing-
away.10
The third factor is that among the Neoplatonic commentators on Aristotle
such as Syrianus, Ammonius and his students, the rigid assignment of cosmic
roles to each of the causes came to talce the form of a general law whenever it
was applied in interpreting Aristotelian discussions of the four causes. The law
held that final and efficient causes are transcendent of or extrinsic to their
effects, while formal and material causes were immanent in or intrinsic to their
effects. Applying this law to Aristotle's texts required less effort than might be
expected. For though Aristotle nowhere makes such a canonical distinction
between transcendent and immanent causes, in Physics 2.3 he contrasts the
material cause and the efficient cause, saying that the material cause is a
"constituent" (enuparkhon) of its effect.1' This is repeated in Metaphysics 5. l,
where Aristotle takes the next step and actually makes a distinction between
principles that constitute (enuparkhon) their effects and those that are not
constituent of (me enuparkhon) and outside (ektos) their effects. 12 In
Metaphysics 12.4, Aristotle contrasts the matter and form, which he calls the
constituent causes, with the extrinsic efficient cause. 13
In his comments on Metaphysics 5.l, Alexander extended the category of me
enuparkhon to cover ends as well as agents. 14 Syrianus, Proclus and Asclepius
then took the next step in systematizing the distinction, saying that the formal
and material causes are constituents of their effects, while the final and efficient
causes are outside their effects. 15

10 Syrianus, in Metaph. 13.I (ad 1076al0), 82,2-13; Proclus, in Parm., 983,1-3 (following

Westerink's suggestion of relika for Cousin's releia kai); in Tim., Vol. 1, 2,1-4,5; 4,26-28; 17,15-30;
263,19-264,3; and 3, 126,11-13.
11 enuparkhOn at Phys. 2.3, 194b23 (= wa-huwafihi, af-Tabi< a, 101,1). Philoponus (in Phys. [ad
Joe+ 243,30-244,3) concludes that the efficient cause is therefore extrinsic to its effect.
1 enuparkhontos at Metaph. 5.1, !0!3a4 (= wa-huwa ft sh-shay'i at Ma ba< da {-!abica II,
473,8), and me enuparkhontos at !013a7 (= wa-laysa huwa fihi at Ma ba<da f-tabica //, 474,3);
touton de hai men enuparkhousai eisin hai de ektos at Metaph. 5.1, 1013al9-20 (= wa-ba<{lu /.
ibtidii'atifi 1-ashya'i wa-ba<{luha hiya kharija at Ma ba<da f-fabi°a II, 474,11-12).
13 cf. ta enuparkhonta aitia at Metaph. 12.4, 1070b22 ("' al-asbabu hiya 1-mawjudatu ... ft llati

takunu at Ma ba'da f·fabi'a Ill, 1522,10).

14 Alexander, in Metaph. 5.1 (ad 1012b34), 345,37-346,16.
15 See Syrianus, in Metaph. 3.1 (ad995b27), 7,8-10 (where he cites Alexander as the source of

the distinction); and 3.2 (ad 996a21), 13,30-14,38; Proclus, Inst. Theol., Prop. 75, 70,28-72,4; in
Tim., Vol. I, 237,9-11 and 239,24-240,l; 3, 4,17-18; 196,14-16; and 202,13-14; Asclepius, in
Metaph. 1.9 (ad 99!a8), 84,7-33 and (ad 99lal9), 87,29-30; 5.1 (ad 1013al7), 305,2-17 (where
the distinction is stated most canonically: some principles - the formal and material - are
constituents of the effect [hai men enuparkhousi], while others - the final and efficient - are
extrinsic to the effect [hai de ektos )). As I shall show in Chapters 6 and 11 , Avicenna inherited this
 3. Proclus, Ammonius and Asclepius 69
With these three extra factors at play - the addition of the paradigmatic and
instrumental causes to Aristotle's final, efficient, formal and material; the
assigning of each of the six causes to a place in the cosmic hierarchy; and the
distinction between final and efficient causes, which transcend their effects, and
formal and material causes, which are immanent in their effects - determining
which type of cause the perfect is and which type perfection is had ramifications
beyond that simple act of identification.
The less complicated identification seems to be what kind of cause the perfect
is. As mentioned above, since the perfect is called "fecund", and therefore seems
to be productive of its effect's corning-to-be (genesis) in the way a father is
productive of his son, it fits most comfortably in the role of efficient cause. But
if the perfect produces its effects as an efficient cause, and if, strictly speaking,
an efficient cause transcends its effect and belongs to the superlunary world of
eternal existence, it will follow that the perfect similarly transcends its effects
and belongs to the superlunary world of eternal existence. Does that mean that
nothing down here in the sublunary world can be correctly described as being
perfect? If it does, then we are left with two alternatives: either the perfect will
fail to serve in the sublunary role of the actual or the active, namely, causing the
potential to change into the actual or the active; or the actual or the active will
be held to exist in the sublunary world only in a metaphorical sense, with true
actuality and activity reserved for superlunary beings.
Which alternative I chose depended partly on the extent to which I, as a
Neoplatonist, was committed to integrating Aristotelian four-cause theory into
my cosmology. If my commitment is mild, then I shall tend to assign the four
Aristotelian and two Neoplatonic causes to various stations in the hierarchy -
final cause at the top of the superlunary world, then paradigmatic, then efficient,
then formal at the top of the sublunary world, then instrumental, and then
material - and not worry too much about how that might complicate my
attempts to account for sublunary change. This is the impression Proclus
usually gives. Proclus sees the perfect as belonging, strictly speaking, to the
superlunary world of real efficient causation, and as applying specifically to the
Demiurge, the lowest of the three major inhabitants of the intelligible world and
the one responsible for causing the coming-to-be of sublunary entities.
(Sometimes the One is spoken of as if it were an efficient cause, other times as
if it were above any type of causation.)
Proclus derives the distinction between causes which are transcendent of their
effects and causes which are immanent in their effects, from the fact that the
final, paradigmatic and efficient causes inhabit the superlunary world and are thus
transcendent of their effects in the sense of being separate from matter, whereas
the formal, instrumental and material causes inhabited the sublunary world and

distinction and used it in his own work. I discuss the history of this distinction at greater length in
Wisnovsky 2003a.
70 Part /: The Ammonian Synthesis

are thus immanent in their effects in the sense of being inseparable from
matter. This seems to indicate that it was usually more important to Proclus to
assign to each element of the cosmic hierarchy a station and a type of causation
appropriate to that station - and then maintain, if forced, that the
transcendent/immanent distinction applied in an equivocal way to sublunary
causes and effects - than it was to apply the transcendent-immanent distinction
univocally, that is, to hold that the distinction applies to causes down here in the
sublunary world just as it applies to causes up there in the superlunary world.
If, on the other hand, integrating Aristotle's four-cause theory into the
Neoplatonic cosmos is my main goal, then I shall tend to apply the
transcendent-immanent distinction univocally, to sublunary entities as well as to
superlunary ones, at the cost of losing the rigid identification of each element of
the cosmic hierarchy with a particular type of cause. The upshot is that if the
efficient cause operates univocally in both the superlunary and sublunary realms,
and if the perfect is nothing more than the state of being which is enjoyed by an
efficient cause relative to its effect, then - assuming that there are no efficient
causes which are not perfect - the perfect will operate univocally in both the
superlunary and sublunary worlds.
Regardless of what a Neoplatonist would say if forced to choose between an
equivocal or a univocal understanding of the perfect, it is clear that the
procession (proodos) of being from the One, at the very top of the superlunary
hierarchy, through the Demiurge, at the bottom of the superlunary hierarchy,
through the world of coming-to-be and passing-away, and all the way down to
matter - the procession, that is, from things which are more perfect in their
existence to things which are less perfect in their existence - was generally seen
by the Neoplatonists as a procession from efficient causes to their effects. 16
Now that the perfect has been shown to produce its effect as an efficient cause,
albeit with some problematic implications, what kind of cause will perfection
be? On the one hand, perfection can be understood simply as the abstract noun
corresponding to the adjective teleion. Since something teleion operates, as we
have seen, as an efficient cause, then teleiotes will simply mean efficient
causality. On the other hand, teleiotes is spoken of by the Neoplatonists as
something to be attained by means of imitation (homoiotes) and participation
(methexis), and therefore seems to operate on its effects the way a final cause
does. In other words, when a perfect X encounters an imperfect X and the
imperfect X becomes a perfect X, the perfect X stands as an efficient cause to the
imperfect X, while the perfection of X - the state of perfect X-ness in which the
perfect X participates - stands as a final cause relative to the imperfect X. 17

16 Proclus, Theo/. Plat., Vol. 2.6, 40,9-27; 3.2, 7,21-27; Asclepius, in Metaph. 4.1 (ad 1003a21),

223,34-36. Ammonius' theory of procession and reversion, as reported in Asclepius' apo phOnes
corwnentary, is helpfully discussed by Verrycken 1990.
e.g., Proclus, Inst. Theo/., Prop. 32, 36,3-4; and 153, 134,23-35.
 3. Proclus, Ammonius and Asclepius 71

Of course it is possible that efficient causation - the state or quality of being

an efficient cause - is itself something to be striven for, and that it therefore
causes its effects as a final cause. But strictly speaking this is incorrect,
according to the Neoplatonists. On the contrary, teleiotes can be divided into
two types. The first type of teleiotes refers to a thing's perfection taken in and
of itself. A perfect X strives to maintain its own teleiotes in the sense that it
strives constantly to fulfill the characteristic of X-ness, and thereby ascend to the
cusp of the next level up in the cosmic hierarchy. In this first sense teleiotes
seems to refer most directly to one of the original Aristotelian senses of teleion
articulated in L4: the sense in which a thing which is teleion possesses a
superiority or "unsurpassedness" in its genus or species. A first teleiotes will
thus act on its effect as a final cause by serving as a goal: in this case, the goal
of self-perfection. The second type of teleiotes refers to a thing's perfection in
relation to things other than itself, and specifically in relation to things which
are inferior to it in the hierarchy of beings. A perfect X which is also an
imperfect Y strives to attain the state of perfect Y-ness in which a perfect Y
participates. Again, this type of teleiotes - perfect Y-ness - acts on its effects as
a final cause, by serving as the goal of change.
Proclus helps us out by illustrating what he means. In his commentary on the
Republic, Proclus says that examples of a thing taken in itself are "human" or
"soul", and examples of a thing taken in relation to others are "human master of
slaves (despotes)" and "soul master of body". 18 Elsewhere, making an allusion
to Phaedrus 247E4-6, Proclus claims that the distinction between the two types
of perfection is analogous to the distinction between the solidity of ambrosia
(teleiotes in relation to self) and the liquidity of nectar (teleiotes in relation to
others). 19
I shall discuss in Chapter 4 how Proclus' distinction between first teleiotes
and second teleiotes caused Philoponus to rework Alexander's and Thernistius'
entelekheia schemes. For the moment, I shall focus on the role teleiotes plays
in Neoplatonic cosmology. For just as the procession from the One, through the
Derniurge, through the world of coming-to-be and passing-away, all the way
down to matter - the procession, that is, from higher things which are perfect in
their existence to lower things which are imperfect in their existence - was seen
by the Neoplatonists as a procession from efficient causes to their effects, so the
reversion or turning upwards (epistrophe) from matter, through the world of
coming-to-be and passing-away, through soul, through Intellect, through the
Ideas contained in the paradigmatic cause, all the way to the Good - the
reversion, that is, from things which are imperfect in their existence towards

18 Proclus, in Remp.. Vol. !, 207,14-208,2; see also 2, 74,7-9; and Inst. Theo!., Prop. 153,

134,23-35.
19 Proclus, Theo/. Plat., Vol. 4.15, 47,7-48,9; 4.16, 49,1.
72 Part /: The Ammonian Synthesis

their own perfection - was seen as a reversion from effects towards their final
causes. 20
A number of interrelated clusters of questions immediately arises. First, how
does procession relate to reversion? Are they symmetrical, in the sense that what
happens during reversion is simply the reverse of procession? Or is reversion
something other than the mirror image of procession, and if so, what does each
refer to? Second, in what sense do the One and the Good refer to the same entity?
If they are different, is the Good superior to the One, or vice versa? If the One
and the Good are the same, how can They - or It - operate both as an efficient
cause and as a final cause, yet remain utterly simple and unitary? And third, how
does all this relate to the sense in which the perfect and perfection are causes?
Specifically, has the move to causation forced so many new meanings on perfect
and perfection that they no longer refer to anything coherent?
The short answer to the first cluster of questions is that the Neoplatonists
regarded procession and reversion as different but complementary, rather than as
symmetrical. Sometimes, when referring to the two aspects of a single entity -
when referring, that is, to its status as an element of procession and to its status
as an element of reversion - its ousia ("being") was contrasted with its teleiotes
("perfection").21 Other times, when describing what was passed down during
procession, the term to einai ("existence") was used, and contrasted with to eu
einai ("well-being"), which is what is attained during reversion. The distinction
between to einai and to eu einai appears to have originated with Alexander, and
the Neoplatonists seized upon it with much the same enthusiasm as they seiz.ed
upon Alexander's promotion of teleiotes. 22

20 Plotinus, Enneads 1.8.2,2-4 (esti de touto [viz., the Good] eis ho panta aneretai kai 'ou panta
ta onta ephietai' [quoting Aristotle, EN 1.1, 1094a3] arkhen ekhonta auto ka'keinou deomena);
Syrianus, in Metaph. 13.5 (ad 1079b24), 117,28-32; Proclus, in Remp., Vol. l, 286,11-287, 10
(citing Laws X, 903C8); Theol. Plat., Vol. 1.22, 104,3-12; 2.2, 20,17-25 (like Plotinus, citing EN I. I,
1094a3 and 10.2, 1172bl4-15); 2.8, 55,10-14 (citing Letters II, 312E); Inst. Theol., Prop. 8, 8,29-
10,13 (citingPhilebus20D); 12, 14,1-23; in Ale./, 61,3-5; in Parm., 621 ,3-7; 810,2-3.8-10.24-31;
845,19-25; 1124,16-20 (citing Laws IV, 716C); in Tim., Vol. l, 281,27-29; 285,21-286,3 (ad 2888
and citing Resp. VI, 50989); 355,28-357,23; 360,5-362,16 (ad 2906-E4); 368,15-369,9 (ad 29E4-
30Al); 370,19-21; Ammonius, in /sag., 24,2-5; Asclepius, in Metaph., 1.1 (ad 982al), 15,7-10; 1.6
(ad 987b9), 47,9-10; 1.6 (ad 988a8), 51,31-52,7; 1.9 (ad 992a24), 103,9-11; 1.9 (ad 992b9),
106,33-107,4; 1.9 (ad992bl8), 108,23-26; 2.2 (ad994al9), 123,7-17; 3.2 (ad 996a18), 151,22-32;
7.17 (ad 104la9), 450,25-7.
21 Plotinus, Enneads, lll.4.1,8-17; Syrianus, in Metaph. 3.4 (ad 100la29), 46,27-34; 13.5 (ad

1079b24), 117,28-32; Proclus, in Remp., Vol. I, 37,28-29; 98,23-25; 206,6-26; 236,19-237,3;

252,29-31; 270,13-272,7; 275,29-276,22; Theol. Plat., Vol. 1.15, 70,5-6; l.25, 109,8-9; 3.6, 21,16-
17; 3.27, 94,15-16; Inst. Theol., Prop. 31, 34,28-36,2; 35, 38,9-18; 36, 38,30-32; 37, 40,7-9; 38,
40,25-26; 43, 44,29-31; in Ale. /, 1,7-4,5; in Parm., 831,25-28 and 1210,3-11; in Tim. , Vol. l,
404,16-18 and 41 2,30-413,3; 3, 4,17-18 and 213,31-214,6; Asclepius, in Metaph. 3.2 (ad 996al8),
151,29-32; Philoponus, in DA (Proem.), 17,29-30; 1.1 (ad 403a29), 56,33-34; 2.2 (ad 413al l),
228,22-26; 2.8 (ad 420bl6), 380,27-381,14; (Ps.-)Philoponus, in DA 3.12 (ad 434b8), 602,21-32;
3.13 (ad 435bl9-22), 606,25-607,14; (Ps.-?)Simplicius, in DA 2.8 (ad 420b22); 3.1 (ad 425al0),
181,14-16; 3.12 (ad 434a27), 317,19-22; 3.13 (ad 435bl9), 329,24-26.
22 At in Metaph. 1.9 (ad 992a24), 121,19-20, Alexander argues that the Ideas cannot be final

causes because the Ideas are held by Platonists to be the causes not of well-being (tou eu einai) but
 3. Proclus, Ammonius and Asclepius 73
The second cluster of questions posed more of a problem. As mentioned
above, Proclus sometimes speaks as if the superlunary world were rigidly
stratified according to causality: the final causality of the Good is at the top, then
the paradigmatic causality of the Ideas, then the efficient causality of the
Demiurge. If the efficient causation of the universe springs from the Demiurge,
and if the final causation of the universe culminates in the Good, then there will
be no identity between the ultimate final cause and the initial efficient cause in
the superlunary world, and the final cause's status will be superior to that of the
efficient cause. 23 But the complementarity I referred to in the last paragraph was
not between the coming-to-be which the Demiurge causes here in the sublunary
world, and the well-being which the Good causes in everything; but rather
between existence and well-being. If efficient causation is to apply to eternal,
superlunary existence as well as to sublunary coming-to-be, and if procession is
to begin with the One and not with the Demiurge, then the One will have to be
regarded as an efficient cause in some sense. And if a duality in the supreme

of being in a general or absolute sense (tou holos einai): since the cause of something's being in a
general sense must precede its effect, and since a final cause cannot precede its effect, the Ideas
cannot act as final causes. (Asclepius also appeals to the distinction between to einai and to e u
einai in his attempt to soften Alexander's criticism: in Metaph. 1.9 [ad 992a24], 103,6-104,9.) In
his DA (3.3, 81,15-20), Alexander contrasts a man's faculties of nutrition and touch, which are
inseparable from that man, and a man's faculty of intellection, which comes to the person only
when he is mature: the former are responsible for the man's existence (pros to einai), the latter for
the man's well-being (pros to eu einai). Themistius follows Alexander by saying that touch is the
only one of the five senses necessary for an animal's being, the other four being necessary for the
animal's well-being (in DA 3.12, 124,17-21 and 3.13, 126,13-15). These two Alexander passages
are discussed briefly by Dooley 1993, 12n.5 and !63n.358. At Quaest. 2.19, devoted to the problem
of whether Providence extends to the whole universe or just to the edge of sublunary world,
Alexander appears to identify the idea of a thing' s well-being (to eu einai) with its proper being
(he oikeia ousia) (63,13-14.16); and Alexander even seems to anticipate one of the ways
Avicenna would later describe the difference between I" and z•d perfections when he
distinguishes between teleiotes with respect to being and teleiotes with respect to well-being
(63,22). As I shall discuss in Chapter 11, Alexander also applies the to einai/to eu einai distinction
to Aristotle's division (Metaph. 5.5, 1015a20-26 (=Ma ba<da f-fabr<a ll, 515,7-14); see also Phys.
2.9, 199b33-15 and 200a31-b8 [= ar-Tabi<a, 158,4-160,7 and 163,9-164,7]) between a thing's
absolute necessity, in virtue of which the thing exists baldly, and a thing's hypothetical necessity, in
virtue of which the thing exists well (Alexander, in Metaph. 5.5, [ad 1015a20], 360,19-28 and
361,11-13.24-29). Aristotle does use to eu to describe organs (e.g., the larynx) which are not
necessary in the sense that they contribute not to being but only to well-being. But a hard distinction
between to einai and to eu einai is nowhere found in Aristotle (apart from the definitely spurious
Econ. [1343bl8-19] and the probably spurious MM [2.11, 44,5-6 and 47,6]). Plato, Crit. 48b5-6,
contrasts to zen and to eu zen; and mentions to eu in what appears to be a causal context at TimJJeus
68E5-6. Plotinus seems to follow Alexander's DA comments when at Enneads, 11.1.5,20-21 he
makes a distinction between a lower type of soul, associated with the fonnation of our bodies, and
which is responsible for our existence; and a higher, celestial soul, associated with our reason, and
which is the cause of our well-being (he gar al/e psukhe, kath ' hen heme is, tou eu einai, ou tau
einai aitia).
23 Proclus, in Tim., Vol. I, 294,9-296,12 (alluding to Meteor. 1.2, 339a26); Theo!. Plat., Vol. 2.9,

57,25-58,10; 59,13-21; 61,2-9; in Parm., 887,36-888,35 and 1168,16-1169,11; in Tim., Vol. I,

260,19-262,2 and 305,6-16 (alluding to Resp. 50986-9 and Philebus 27Al-B2); 355,28-357,23;
360,5-362,16; 368,15-369,9; 370,19-21.
74 Part I: The Ammonian Synthesis

principle is to be avoided, the final causality of the Good cannot be seen as being
superior to the efficient causality of the One, but as identical to it. 24 At the level
of the other inhabitants of the intelligible realm - Gods, Intellects, Demiurges -
the final causality that leads upwards to the Good and the efficient causality that
proceeds downwards from the One are not identical but complementary, in the
way alluded to above: final causality refers to the respect in which an explanans
explains the "well-being" (to eu einai) of an explanandum, and efficient
causality refers to the respect in which an explanans explains the "existence" (to
einai) of an explanandum. 25
The third question then arises, how do all these different elements fit together
in a way which allows us to construct a coherent theory of perfection and the
perfect? The One and the Demiurge in the superlunary realm, and active things
here in the sublunary realm, are all efficient causes of different types; do those
different types of efficient causation correspond to different senses in which
something may be said to be perfect? Or is "perfect" a univocal term which
applies in exactly the same way to superlunary and sublunary things? Similarly,
the Good in the superlunary realm, and actuality in the sublunary realm, are final
causes of different types; do these different types of final causation correspond to
different senses in which something may be said to be a perfection? Or is
"perfection" a univocal term which applies in exactly the same way to
superlunary and sublunary things?
In fact, all the different things to which the perfect and perfection have been
applied can, with a little effort, be accommodated under Proclus' distinction
between first and second perfections. Remember that a thing's perfection can be
considered in relation to the thing itself, in which case it is a first perfection; and
considered in relation to something else, in which case it is a second perfection. I
shall focus on second perfection first. The "something else" in relation to which
a thing is trying to attain its perfection in the second sense, is clearly the thing' s
effect. I am trying to attain my second perfection, to be perfect in the second
sense, when I am being a good father; that is, when I am trying to attain
perfection in my relation to others who stand below me in the cosmic hierarchy
as my effects, namely my children, by means of my passing existence down to
them, as well as by means of my serving as a goal for them to imitate. In short,
I attain my second perfection by perfecting my efficient causation of others

24 Proclus, Inst. Theo/., Prop. 13, 14,24-16,8; 33, 36,11-16; I 13, 100,l 1-12; in Pann., 788,12-28

and 1109,4-14 (citing Letters II, 312E); in Ale./, 181,11-182,11; Asclepius, in Metaph. 1.4 (ad
985b4), 33,34-5; 3.2 (ad 996al8), 151,18-20.26-28; Simplicius, in Cae/. 1.8 (ad 277b9), 271,4-21;
and in Phys. 2.7 (ad 198a22), 365,20-21 and 367,24-29; and 8.10 (ad 267bl7), 1360,24-1363,24.
The extent to which Asclepius, Simplicius and Philoponus paint a coherent picture of Amrnonius'
theology is discussed in Verrycken 1990.
25 Syrianus, in Metaph. 13.4 (ad 1078b32), 107,38-108,7; Proclus, Theo/. Plat., Vol. 6.15, 75,1-2;

in Parm., 826,30-35 and 842,26-28; in Tim., Vol. 3, 226,10-18; Asclepius, in Metaph. 1.3 (ad
984b20), 28,18-29,8; 1.6 (ad 988al 1), 52,21-25; and 3.2 (ad 996al8), 151,9-12; Damascius,
Treatise on First Principles 3, I 17,4-118,18.
 3. Proclus, Ammonius and Asclepius 75

(teleiotes here understood as the quality of being teleion qua actual and active)
and by serving as a final cause for others to imitate (teleiotes here understood as
the quality of being teleion qua superior in a genus or species).
The first perfection, by contrast, is related solely to the thing itself. I am
trying to attain my first perfection, to be perfect in the first sense, when I am
being a good human; that is, when I am trying to attain perfection as an
individual member of my species - rational animal - by being as rational as I
can be, by thinking about philosophy or, even better, by thinking about myself
thinking about philosophy. As an element of the procession of existence, I am
trying to attain my second perfection in the sense that I am trying to attain
perfection as an efficient cause in relation to things below me in the hierarchy,
just as the Demiurge does in relation to the world as a whole. As an element of
the reversion towards well-being, I am both trying to attain my second perfection
by serving as a final cause for things below me in the hierarchy to imitate; as
well as trying to attain my first perfection by thinking about myself thinking
about myself and thereby moving upwards in the hierarchy, just as the Demiurge
does by contemplating the Ideas contained in the paradigmatic cause, according to
whose model He continually creates the world.
These two senses of perfection - a first perfection which is relative to what is
above and upwards, and a second perfection which is relative to what is below -
are, I believe, part of Proclus' attempt to explain the dynamic tension between
procession and reversion, and they derive from the new, Neoplatonic tying
together of causality and being perfect. Does the distinction between first and
second perfections also apply univocally to the One and the Good (or perhaps
better, the One-Good, since they are held by Neoplatonists to be distinguishable
only conceptually)? As the ultimate cause of the rest of the superlunary world
and of the entire sublunary world, the One-Good could, it seems, be described as
being in a state of second perfection, given that the One-Good can be related - as
cause - to something other than itself. The One-Good could also be described as
being in a state of first perfection, if we assume its primary activity to be the
self-intellection, the turning inwards (though in this case not upwards, since
there is nothing above it), that first perfection consists in, at least in the case of
rational beings. It seems appropriate then to apply the scheme of first and second
perfection to the One-Good, with the proviso that its first perfection is only a
turning inwards and not upwards. 26
Yet for many Neoplatonists the causal sense of first and second perfection -
actualizing oneself and actualizing things below one - implied too much
involvement in the messiness and multiplicity of both the superlunary and
sublunary worlds to be properly applicable to the One-Good. Rather, the sense in
which the One-Good is perfect and possesses perfection appears, on the surface,

26 For an example of the first perfection/second perfection distinction applied to God, see

Asclepius, in Metaph. 2.4 (ad 1000a5), 195,37-196,10.

76 Part I: The Ammonian Synthesis

to derive more from one of the original, primitive Aristotelian notions of to

teleion, namely, that which is quantitatively complete (i.e., whole, not missing
anything, not lacking anything) than it does from the Neoplatonists' new causal
sense of teleiotes. Under the surface, however, the tum to causation can be
detected, the only difference being that in the case of the One-Good's perfection
the tum to causation consists in a tum from quantitative completeness to causal
self-sufficiency, that is, to immunity from causedness; whereas in the case of
perfection in procession and reversion, the turn to causation consists in a tum
from actuality and activity to causal productivity and self-actualization.
What I mean is that the sense in which a thing is quantitatively complete in
the old Aristotelian sense could not really be applicable to the One-Good, since
it seemed too similar to the idea of wholeness, and wholeness implied having
parts - having all of them, to be sure, but still having parts - a quality which
clearly infringed on the One-Good's utter simplicity. Instead, the One-Good's
teleiotes signified an absence of lack or need understood as causal independence
rather than as quantitative completeness, as referring to the quality of not
requiring a cause, and not to the quality of being in possession of all one's
parts.
As a result the terms "perfect" (teleion) and "self-sufficient" (autarkes), and
"perfection" (teleiotes) and "self-sufficiency" (autarkeia), came to be seen as
mutually implicative, if not actually synonymous. 27 The connection between the
two ideas was held by the Neoplatonists to spring from a passage in the
Philebus, where the three elements of the Good are said to be the desirable, the
sufficient and the perfect (to epheton, to hikanon, to teleion). These terms were
understood by the Neoplatonists as referring, respectively, to the Good's final
causality, self-sufficiency and perfection, and the sense one gets from Proclus at
least is that each of the three terms was thought to imply the other two. 28
Finally, Neoplatonists could always cite EN 1.7, 1097b7-8 (to gar teleion
agathon autarkes einai dokei), which I discussed in Chapter 2 - and, more
ambiguously, Metaph. 14.4, 109lbl6-19 - as evidence that Aristotle followed
Plato in tying perfection (or at least goodness) to self-sufficiency.
Yet holding that the One-Good was perfect and self-sufficient raised its own
problems. While being perfect and self-sufficient might be the qualities which
best insured the One-Good's transcendence, they were less useful in explaining
its productivity. Something self-sufficient is certainly eternal, and it may even
act as a final cause by serving as a model for lower things which seek to attain

27 Plotinus, Enneads, Vl.7.2,48-49 (ei oun teleion, ouk estin eipein hotoi elleipei, oude dia ti

touto ou paresti) (although this was rendered into Arabic as the more innocuous fa-in kana l-'aqlu
tamman kamilan fa-innahu lam yaqdir qa' ilun an yaqula inna-hu naqi~un Ji shay' in min /;liilatihi:
UthUlujiya 5, 73,7-8); Proclus, Theo/. Plat., Vol. 4.25, 74,6-75; 5.7, 28,3; Inst. Theo/., Prop. 9, 10, 14-
16. cf. (Ps.-)Plato, Definitiones, 412b3: "Self-sufficiency: completeness in the possession of goods
(autarkeia teleiotes kteseos agathOn)".
2 g Proclus, Theo/. Plat., Vol. 1.22, 101,14-19; 3.22, 79,9-22; in Ale./, 153,10-15; Olympiodorus,

in Phaedonem, 30,14-20; and in Ale. I, 42, 18-43,3.

 3. Proclus, Ammonius and Asclepius 77

its quality of perfection and self-sufficiency; but it will only be productive of an

effect below it if it passes some of the existence it does not need down to
something which is not perfect and self-sufficient, and which does need some
existence.
As a result the One-Good was. held to be "the most self-sufficient"
(autarkestaton) or even "above self-sufficiency" (huperautarkes), with the other
inhabitants of the superlunary world now being described as being self-
sufficient. 29 The tying of perfection to self-sufficiency also applied to the
universe as a whole, which was described as being teleion and autarkes. 30 When
referring to the self-sufficiency of superlunary beings other than the One-Good,
however, the term authupostaton was often used instead of autarkes.
Authupostaton seems to have meant "self-substantiating" or "self-constituted" as
opposed to "self-sufficient", and the distinction, though obscure, was held to be
meaningful. The basic difference seems to be that autarkes is a negation while
authupostaton is an affirmation. In something autarkes, self-causation is
denied, since it is held to be above all types of causation, even self-causation; in
something authupostaton, however, self-causation is affirmed. Another way of
distinguishing them is that authupostaton refers to the self-sufficiency of
something like the universe as a whole, which has parts and which therefore is
complete or perfect in the sense of not lacking any of its parts. By contrast,
autarkes, strictly speaking, referred to the self-sufficiency of the One-Good,
which is utterly simple and which is complete or perfect in the sense of not
needing anything else to cause it. 31
As will become clear in Chapter 10, Avicenna wrestled with a number of
issues left unresolved by the Neoplatonists' application of the concepts of the
perfect and perfection to problems of cosmology. The Neoplatonists' tying
together of the Good's final causality, self-sufficiency and perfection, came into
conflict with their desire to see the One as somehow productive - and hence as
the efficient cause - of the world's existence. This is because the Good's final
causality, self-sufficiency and perfection were attributes that best guaranteed its
transcendence of the world, while the One's productivity and efficient causality
were attributes that best guaranteed its involvement in the world. In addition, the
Neoplatonists bequeathed to Avicenna an ambivalence about the univocity or
equivocity of perfection and the perfect: were these terms applicable to
superlunary and sublunary entities in different or identical ways? For now,
however, I shall stick with the Neoplatonists and examine how this new way of
looking at perfection and the perfect influenced the ways in which the soul - the

29 Syrianus, in Metaph. 14.4 (ad 1091bl6), 183,10; Proclus, Inst. Theo!. , Prop. IO, 10,31-32; 127,

112,25-34.
Jo Proclus, in Tim., Vol. I, 289,17; 2, 89,31-90,2.
JI Proclus, Inst. Theol., Prop. 42, 44,11-24; 43, 44,25-32; 45, 46,12-19; in Parm., 1145,31-
1146,35. Also see Whittaker 1975.
78 Part /: The Ammonian Synthesis

most interesting and problematic example of a perfection - was integrated into a

new, Neoplatonic entelekheia scheme.
 4. Proclus, Ammonius and Philoponus
Neoplatonic Perfection and Aristotelian Soul

I have just shown how the Neoplatonists transfonned the concepts of perfection
and the perfect from referring, respectively, to a state of being (being in actuality
or being in activity) and to a thing which possesses that state of being (the
actual or the active), to referring, respectively, to a state which acts as a final
cause and to a thing which acts as an efficient cause. The Aristotelian rationale
for this transfonnation was that something which is in a state of actuality or
activity is in some sense productive of the thing which is in a state of
potentiality and which is changing, as a result of its contact with the actual or
active thing, into something actual or active itself. Thus the actual or active
stands as a cause - and specifically as an efficient cause - relative to the potential
thing. The state of actuality, on the other hand, stands as a final cause - as a
state to be attained - relative to the potential thing. Keeping a safe distance
between themselves and the term entelekheia (and its adverbial forms
entelekheiai and kat' entelekheian) as a result of Plotinus' attacks on that
term, the Neoplatonists used teleiotes - a term Alexander of Aphrodisias had
promoted as a gloss of entelekheia - and teleion. One of the advantages of
understanding teleion and teleiotes as causes was that the two tenns could be
applied quite neatly to the complementary cosmic forces of, respectively,
procession and reversion, the fonner expressing the downward direction of the
efficient causation of existence, the latter expressing the upward direction of the
final causation of well-being.
The Neoplatonists went further and divided teleiotes into first and second
perfections. The first refers to the type of perfection that obtains when a thing is
related only to itself, and which serves as the goal of that thing's constant
striving to attain or maintain its self-perfection and thereby ascend to the next
level up in the cosmic hierarchy. The second refers to the type of perfection that
obtains when a thing is related to something other than itself, that is, to the
imperfect things that are below it in the cosmic hierarchy. A thing strives to
attain its second perfection both in the sense that it strives to perfect its efficient
causation of the things below it, and in the sense that it strives to serve as a goal
for lower things to imitate in their own attempts at reversion. In the case of first
and second perfection, the teleiotes operates on its effect as a final cause.
Given the contortions which commentators such as Alexander and Themistius
subjected themselves to in their efforts to square Aristotle's use of entelekheia
in his definition of soul (where the distinction between first and second
entelekheiai is most clearly articulated) and his use of entelekheia in his
definition of motion (where the distinction between initial and ultimate
entelekheiai is only implied), the following question arises: how does the
80 Part I: The Ammonian Synthesis

Neoplatonists' distinction between first and second teleiotetes square with the
entelekheia schemes proposed by Alexander and Themistius? For the sake of
clarity, this question ought to be split into three subsidiary questions: (1) Is the
Neoplatonic distinction nothing other than a new formulation of the De Anima's
distinction between first and second entelekheiai? (2) Is it instead another
attempt, in the tradition of Alexander and Themistius, to create a basis for
reconciling the De Anima's explicit distinction between 1" and znd entelekheiai
and the Physics' implicit distinction between complete and incomplete
entelekheiai? (3) Or is it a new distinction, formulated by Neoplatonist thinkers
whose overriding interest may not have been soothing Peripatetic anxieties about
Aristotle's seemingly inconsistent uses of entelekheia?
I shall begin answering the first question ("Is the Neoplatonic distinction
nothing other than a new formulation of the De Anima's distinction between
first and second entelekheiai?") by recalling that the De Anima 's distinction
between first and second entelekheiai is most plausibly interpreted as referring
to a distinction between an inactive possession (hexis) of a capability to perform
a certain function, and the active exercise (energeia) of that capability. Proclus
occasionally refers to both knowledge and contemplation as teleioteres, which
might lead us to believe that he is following Aristotle's De Anima examples in
some way; but Proclus does not identify the former as a first and the latter as a
second perfection, and he seems to have Plato in mind as much as Aristotle. 1
Other times Proclus contrasts the static quality of first perfection with the
active quality of second perfection, saying at one point that just as the perfection
of the gods is of two sorts - the first intellective, the second providential, the
first at rest and the second in motion (dittes gar ouses tes theias teleioteros,
res men noeras, res de pronoerikes, kai tes men en stasei, res d'en kinesei) -
so human souls have two perfections: one which consists in the contemplation
of the eternal, divine things which the gods possess, the other which consists in
a providential concern for things in the world of coming-to-be and passing-
away. 2
Commenting on Plato's definition of philosophy - becoming as much like a
god as is humanly possible - Ammonius seems to be referring to Proclus'
distinction between first and second perfections when he claims that just as gods
have both cognitive (gnostikas) activities and providential (pronoetikas)
activities, so philosophers must similarly contemplate all things and look after
inferior things. Although Ammonius asserts that the difference between an
activity which is self-related and one which is related to others consists in the
difference between striving to fulfill one's differentia in respect of substance and
striving to fulfill one's differentia in respect of perfection, Ammonius'

1 Proclus, Theo!. Plat., Vol. 4. 13, 43, 19-20 (alluding to Phaedrus 247C8); in Ale. /, 187,8-18; in

Tim., Vol. I, 79,9-13.

2 Proclus, in Tim., Vol. 3, 324,6-12.
 4. Proclus, Ammonius and Philoponus 81
distinction can in fact be reduced to the difference between Proclus' first and
second perfections: the perfection of human is to live rationally (first perfection)
and with phronesis (second perfection); the perfection of horse is to run quickly
(first perfection) and to be useful in war (second perfection). 3
On the surface, then, it seems that the Neoplatonists' second teleiotes and the
De Anima 's second entelekheia might have something in common, in that they
both express the quality of being active rather than being at rest: my mind is in a
state of second entelekheia when it is actively contemplating, and it is in a state
of first entelekheia when it is inactive. Certainly, if a second teleiotes referred
to nothing other than the quality of productivity and efficient causality possessed
by something which is teleion, then the De Anima's second entelekheia and
the Neoplatonists' second teleiotes would indeed have something in common. A
perfect thing is related to an imperfect thing as an efficient cause - that is, a
perfect thing is productive of an imperfect thing's emergence into perfection -
only when the perfect thing is active. This is an Aristotelian axiom which the
Neoplatonists take for granted, and as we saw above it serves as one of the bases
for their linking of to teleion to the efficient cause.
But as was also discussed above, when the Neoplatonists made their
distinction between first and second teleiotetes, they meant for the second
teleiotes to refer to something more than just the quality of efficient causality
possessed by what is teleion. What is also meant by the second teleiotes is the
quality or state of perfect X-ness - that is, the quality of unsurpassedness in a
species or genus - which a perfect X possesses and which an imperfect X strives
to attain. And in fact, the sense in which the second perfection is understood to
refer to the goal of serving as a final cause for others to imitate, as opposed to
the sense in which it is understood to refer to the goal of acting on others as an
efficient cause, was held to be the more basic sense of second perfection. In other
words, I am striving to attain my second perfection not so much to possess the
quality of efficient causality, as to possess, or to come as close as possible to
possessing, the more important qualities associated with perfection: causal self-
sufficiency and eternality.
The reason for this is that if a second teleiotes were nothing other than the
quality of efficient causality which a perfect thing possesses, procession and
reversion would be symmetrical rather than complementary. The efficient cause
responsible for procession and the final cause responsible for reversion would
compete over the same effect or explanandum, with reversion losing out to
procession since reversion will refer to nothing more than the attaining of
efficient causality. In other words, reversion would consist simply in gaining the
ability to proceed. On the contrary, the efficient causation of procession and the
final causation of reversion complement each other in the sense that the efficient
cause and the final cause each explain a different type of effect or explanandum.

3 Amrnonius, in !sag., 3,24-4,4.

82 Part /: The Ammonian Synthesis
The efficient cause involved in procession explains pure existence or existence in
an absolute sense, while the final cause involved in reversion explains well-
being, that is, special existence, X-ness as opposed to Y-ness. If the efficient
causation of procession and the final causation of reversion are to be
complementary in a meaningful way, therefore, a Neoplatonist's second
teleiotes will have to refer to something more than simply the efficient
causality or activity which to teleion possesses.
What about Proclus' comment that the first perfection is static and the second
active? Since activity is what allowed the Neoplatonists to make their move
from Aristotelian energeia and entelekheia to their new concept of to teleion
as efficient cause, Proclus' assertion that the second perfection is active while the
first is static would seem to support the idea that second perfection is nothing
more than efficient causality, or at least that efficient causality was the essential
characteristic of second perfection.
The fact of the matter is that Proclus is inconsistent in his application of the
idea of activity to perfection and the perfect. More often than not, in fact, it is
the first teleiotes - the turning inwards and upwards that occurs during self-
intellection - which seems to involve more activity than the second teleiotes. 4
Even though my body may be still, my soul is very active when I think about
myself thinking about myself, in my attempt to pull myself up to the next level
of the cosmic hierarchy, however temporarily I may reside there. By contrast, I
may be asleep or watching television when my graduate student is hard at work
using my intellectual teleiotes as a goal to model himself after.
The classic example of this type of relationship is, unsurprisingly, the
relationship between Socrates and his protege Alcibiades. Socrates' intellectual
perfection (teleiotes) serves as the goal of Alcibiades' teleiosis, his journey
towards his own intellectual perfection; but it is never clear whether Socrates
needs to be actively using his capability to think in order to serve in the role of
model. As long as Alcibiades keeps Socrates' intellectual teleiotes in mind as a
goal to attain through imitation, it seems possible that Socrates can lie back and
relax, or be actively engaged in some other activity.5 Certainly, it is a condition
of serving as a second teleiotes relative to something else that at some point in
the past one actively sought one's own teleiotes. At some point in the past
Socrates needed to have been active in attaining his own teleiotes before his
teleiotes can then serve as a model for Alcibiades to imitate. But Socrates'
active pursuit of his own perfection was a first teleiotes, not a second teleiotes.
When Socrates is engaged in a dialogue with Alcibiades, teasing out the truth
and impelling his protege to think hard and behave more philosophically,
Socrates may well be actively using his own capability to think. But Socrates'

4 Proclus, in Remp., Vol. I, 21,18-21and212,2-8; in Tim., Vol. I, 243,26-246,9; 2, 117,28-29;

5 Proclus, Theol. Plat., Vol. 4.16, 49,2-3; 4.37, 109,1-13; in Ale. /, 121 , 1-123,20 (teleiosis is also
seen as something bestowed at 129,16; 132,4; 140,8; and 141 ,11).
 4. Proclus, Ammonius and Philoponus 83
active engagement in contemplation during the dialogue is more easily
interpretable as a first teleiotes than it is as a second teleiotes, since Socrates'
goal is always first and foremost attaining or maintaining his own self-
perfection, not bestowing perfection on others; and this is why, I believe,
Proclus calls self-perfection the first, or primary, teleiotes.
By contrast, serving as a goal for imperfect beings like Alcibiades will only
ever be a second perfection for Socrates, that is, it will only ever be a perfection
in a secondary sense, a mere by-product of the more important goal of self-
perfection. The upshot is that the criterion of inactivity or activity, which could
be useful in distinguishing the De Anima's first and second entelekheiai, is
inapplicable to the Neoplatonists' distinction between first and second
teleiotetes, since the Neoplatonists' distinction rests more on the difference
between the primary goal of attaining one's self-perfection, and the secondary
goal of serving as a goal for lower things to imitate.
The two pairs of perfections - the De Anima's first and second entelekheiai,
and the Neoplatonists' first and second teleiotetes - cannot therefore be held to
be strictly identical, so the first question mentioned above ("Is the Neoplatonic
distinction nothing other than a new formulation of the De Anima's distinction
between first and second entelekheiai?") can be answered with a no. In fact the
two pairs of distinctions now appear sufficiently incompatible that the second
question mentioned above ("Is the Neoplatonic distinction another attempt, in
the tradition of Alexander and Themistius, to create a basis for reconciling the
De Anima's explicit distinction between 1st and 2nd entelekheiai and the
Physics' implicit distinction between complete and incomplete entelekheiai?")
is hardly worth pursuing, since even if it could be shown that the Neoplatonic
distinction squares more easily with the Physics' implicit distinction than it
does with the De Anima's explicit distinction, the Neoplatonic distinction will
still be so different from the De Anima distinction that it becomes implausible
to suppose that it was ever meant to serve as a new means of reconciling the
Physics and De Anima distinctions.
My answer to the third question ("Is the Neoplatonic distinction a new
distinction, formulated by Neoplatonist thinkers whose overriding interest may
not have been soothing Peripatetic anxieties about Aristotle's seemingly
inconsistent uses of entelekheia") is therefore a yes. But despite the
incompatibility of the Neoplatonic distinction between first and second
teleiotetes, and the De Anima's distinction between first and second
entelekheiai, the term teleiotes came to be used by the Aristotle-commentator
Philoponus, the student of Proclus' student Ammonius, in his discussions of the
soul. Only now, as we shall see, teleiotes was promoted in order to underscore
the sense in which the soul acted as a final cause, a move that was to have
important ramifications on the issue of the soul's separability.
84 Part I: The Ammonian Synthesis
In spite of his hostility, inherited from Plotinus, towards Aristotle's definition
of the soul as an entelekheia, and in spite of the fact that his distinction
between first and second teleiotetes differed from Aristotle's De Anima
distinction between first and second entelekheiai, Proclus shows a little
flexibility - perhaps ambivalence is a better term - towards Aristotelian
psychology at certain points in his work.6 Philoponus, involved in setting down
his teacher Arnrnonius' lecture-notes on the very text of the De Anima, was
forced, in a way that Proclus was not, to show even greater flexibility in order to
reconcile Proclus' new understanding of teleiotes with the De Anima's use of
entelekheia.
The way this happened took several steps. The first step was to follow
Alexander's lead in simply glossing entelekheia with teleiotes. Philoponus
claims that when the soul is defined as an entelekheia, it means that the soul is
a teleiotes and a form (eidos ). 7 But Philoponus makes a distinction between
entelekheia and teleiotes, saying that teleiotes signifies something more
elevated than what is signified by entelekheia. This is because entelekheia can
refer to form (eidos) in both of form's guises - that is, both qua essence or
formula (kata ton logon) and qua shape or structure (kata ten morphen) -
whereas teleiotes only ever refers to form as essence, and never to form as
shape.
As Philoponus points out, Aristotle himself sometimes speaks of the shape of
something - of a statue, for example - as being its entelekheia. But while
being snub-nosed and paunchy (einai simon progastera) may be the form qua
shape of Socrates, and hence an entelekheia in that sense, being snub-nosed and
paunchy is not Socrates' teleiotes, but rather something merely accidental.8
Philoponus reckons that the reason why teleiotes is a particularly appropriate
term with which to describe the soul, is that a teleiotes, understood as referring
only to the substantial form or essence, is always the cause of the coming-to-
perfection (teleiosis) of the matter (in this case, the body) whose teleiotes it is. 9
Entelekheia, by contrast, can refer to both shape and essence; and because shape

6 Sometimes (e.g., in Tim., Vol. 2, 285,23-31) Proclus even allows that there is a sense in which

the soul, as an entelekheia, is inseparable from the body which it animates, while in itself
remaining separate. Other times he appears unsure of why precisely Aristotle's definition is
unsatisfactory. At Theol. Plat., Vol. 4.24, 71,10-72,11 and in Tim, Vol. 3, 254,10-31, Proclus
attacks Aristotle for assuming as self-evident the distinction between "living" and "ensouled" (lo
zoion kai to empsukhon) in his definition of soul in the De Anima, when in Proclus' opinion the two
amount to the same thing. Yet at in Tim., Vol. 1, 412,11, Proclus himself clearly advocates making
such a distinction.
7 Soul as entelekheia and soul as teleiotes and eidos are identified at Philoponus, in DA

(Preface), 9,26-27; 1.5 (ad 411a26), 193,34-35; 2.1 (ad 412a3), 203,14; 2.1 (ad 412b6), 218,25-26.
8 Philoponus, in DA 2.1(ad412a9),211,11-26.
9 Philoponus, in DA 2.1 (ad 412a3), 208,36-209,4. Asclepius, however, asserts that morph€ is an

end (telos), thereby implying that not just form qua essence but also form qua shape can serve as a
cause: in Metaph. 5.24 (ad 1023a34), 349,10.
 4. Proclus, Ammonius and Philoponus 85
is not the cause of the coming-to-perfection of matter, entelekheia will not
always refer to the soul's causation of the body's coming-to-perfection.
Philoponus' use of teleiotes in this way is important. Not only does it reflect
the Neoplatonic move to causation in the understanding of perfection, it also
frees him to analyze the soul-body relationship as first and foremost a
relationship between cause and effect. This is in contrast to Aristotle's view that
the relationship between soul and body is fundamentally a relationship between
two types of substance (substance-as-form and substance-as-matter), and hence a
relationship between two states of being (being-in-actuality and being-in-
potentiality). In other words, Philoponus has, like Alexander, used teleiotes to
gloss entelekheia in describing the soul; but what Philoponus meant by
teleiotes is different from what Alexander meant. This is because Philoponus
uses teleiotes in the Neoplatonic way, understanding teleiotes as referring not
merely to a type of being - actuality as opposed to potentiality - but to a way of
causing. As discussed above, the way a teleiotes causes its effect, or explains its
explanandum, is as a final cause.
If the soul is a teleiotes, and if a teleiotes acts on its effect (or explains its
explanandum) as a final cause, it follows that the soul acts on the body either
exclusively, or at least primarily, as a final cause. But what does that really
mean? I believe that what Philoponus has in mind is the idea that when we call
the soul a teleiotes, we are asserting that the soul is the explanans and that the
body is the explanandum insofar as the body is viewed in terms of its upward
reversion. As mentioned above, Philoponus has said that when we speak of the
soul as a teleiotes, what the teleiotes causes or explains is the teleiosis of its
effect or explanandum. I have already discussed Proclus' use of teleiosis in his
discussions of the causal relationship between Socrates and Alcibiades: Socrates'
teleiotes serves as the goal, or final cause, of Alcibiades' teleiosis, or upward
journey towards intellectual perfection. Unsurprisingly the distinction between
teleiotes and teleiosis is most explicitly articulated in Proclus' comments on
the First Alcibiades, even though Syrianus may have been the first to promote
the idea that the distinction between teleiosis and teleiotes was somehow basic;
the distinction appears in Asclepius as well. 10

10 The evidence for the distinction between teleiotes and teleiosis in Syrianus is ambiguous: at in

Metaph. 38,21, teleiosis corresponds to energeia, yet at 40,7, teleiotes corresponds to energeia. In
Proclus the distinction is clearer though not held to be basic: teleiosis is clearly meant to be
understood as a temporal process at Theo/. Plat., Vol. 1.6, 30,10 and 4.34, 102,1-3 (alluding to
Timaeus 37C6-D7); yet at in Parm., 1237,3-17, Proclus qualifies that claim by saying that while
teleiosis may be a temporal process, it must be understood as one that is completed, not as one that
is open-ended. As I mentioned in the text, the distinction between teleiosis and teleiotes is used
most often in the Commentary on the First Alcibiades, where Alcibiades is said to be at the starting
point of the process of self-perfection (teleiosis), the culmination of which is the attainment of the
state of perfection (teleiotes): in Ale. /, 4,19-5,12; 6,17-7,3; 10,14-11,3. In more general terms,
teleiosis is simply a process which is undergone (in Ale. /, 172,19 and 224,13-228,20; in Parm.,
667,26-32; 988,6-10; 997,16-33; in Tim., Vol. 2, 16,18-20 and 3, 24,1); whereas teleiotes is simply a
state which is attained (in Ale./, 174,6 and in Ale. /, 177,14; in Parm., 988,6-10; and in Tim., Vol. 3,
86 Part /: The Ammonian Synthesis
Philoponus also uses teleiosis more generally, as a synonym of genesis, with
teleiotes understood as the substantial form that inheres following the genesis
of the composite. 11 It should be said, however, that when commenting on the
Physics Philoponus employs both teleiotes and teleiosis to define kinesis. In
his explanation of Aristotle's Physics 3.1 definition of kinesis as an
entelekheia Philoponus glosses entelekheia with teleiotes, not teleiosis,
saying that change is a first entelekheia in the sense that it is the teleiotes of
the potentiality itself, in contrast to the second entelekheia, which refers to the
teleiotes of the thing in which the potentiality used to exist before and during
the change. 12 In this respect Philoponus anticipates the instinct of modem
commentators to put to one side the question of what exactly Aristotle thinks an
entelekheia might be, and concentrate instead on what Aristotle thinks change
might be an entelekheia of.
Later on in his Physics commentary, however, Philoponus uses teleiosis
when he describes change as a "leading-on" (proagoge). 13 The upshot is that just
as with his understanding of teleiotes, Philoponus' understanding of teleiosis
seems to have been shaped more by Proclus' use of that term than by Aristotle's
or Alexander's: as referring to a thing's upward reversion towards its teleiotes,
be the teleiotes an intellectual form or a natural substantial form. The soul is
the teleiotes of the body, and hence acts on the body as a final cause acts on its
effects, by impelling the body's teleiosis or upward reversion. 14

24, l 2). When Asclepius uses teleiotes and teleiosis in commenting on the first line of the
Meraphysics ("All men by nature desire to know"), and explains that the state of having attained
gnosis is the soul's teleiotes and its good, while the soul's teleiosis is the process of getting to that
point, he very probably had in mind both Proclus' use of teleiotes and teleiosis in discussing the
causal relationship between Socrates and Alcibiades, as well as Alexander's use of teleiotes to
explicate that very same line: Asclepius, in Metaph. I.I (ad 980a21), 5,35-6,2; and Alexander, in
Metaph. I.I (ad 980a21), 1,4-9.
11 Philoponus, in DA I.I (ad 402a9), 27,12-13; 1.3 (ad 406al2), 100,24-30; 2.4 (ad 415b21),

275,2-9; and 2.5 (ad 417b2), 303,13-15; but see also in DA 2.5 (ad 417b2), 301,9-12, where
genesis seems to be understood as a teleiotes rather than a teleiosis.
12 Philoponus, in Phys. 3.1(ad200bl2), 342,10-343,12 (ar-Tabi'a, 171,8-13). Philoponus had

read Thernistius' Paraphrase of the Physics and even quotes Thernistius' distinction between
entelekheia prote and entelekheia hustate at Philoponus, in Phys. 3.1 (ad 20lal6), 351,8-15 (= a{-
Tabi'a, 176,6-9).
13 Philoponus, in Phys. 3.3 (ad 202b2 I), 385,2-4.
14 One difference between Philoponus and Proclus is that Proclus was very reluctant to see the

soul as a hexis of the body, presumably because a hexis is inseparable from that whose hexis it is.
At in Tim., Vol. 1, 406, 14-22 (ad 3084-6), Proclus argues that because the soul is the final cause of
the body just as the intellect is the final cause of the soul, and because a hexis cannot be a final
cause, the soul cannot therefore be a hexis. Of course what Proclus means is that in no sublunary
thing is the teleiotes ever really a possession (hexis); on the contrary, the most that can be said of
the sublunary thing's relation to the teleiotes it constantly strives to attain is that it is a
"participation" (methexis) in the teleiotes. Here Philoponus would disagree with Proclus, given
Philoponus' identification of teleiotetes such as gnosis as hexeis. Nevertheless Philoponus would
still hold, with Proclus, that perfection acts on its effect as a final cause; at in DA 2.4 (ad 4l5a16),
264,15-19 (repeated at in DA 2.4 [ad 416a34], 284,3), Philoponus cites Alexander's authority in
holding that perfection is the final cause of the potential thing. Simplicius also seems to understand
 4. Proclus, Ammonius and Philoponus 87
As mentioned above, seeing the soul-body relationship as first and foremost a
relationship between a cause and an effect, rather than as a relationship between
one type of being and another type of being, freed commentators such as
Philoponus to exploit the whole range of different ways that Aristotle and the
Neoplatonists discussed the final cause, in support of the larger Neoplatonic
project of reconciling Aristotelian and Platonic psychology. What advantages
would a Neoplatonic commentator gain by seeing the soul primarily as a final
cause? First, given the apparent irreconcilability between Proclus' distinction
between first and second teleiotetes and Aristotle's De Anima distinction
between first and second entelekheiai, there might be some way to appeal to an
Aristotelian distinction between the different ways a final cause is spoken of in
order to create a new teleiotes-distinction which could then be reconciled with
Aristotle's De Anima distinction. What I mean is this: if the De Anima
distinction between first and second entelekheiai cannot be reconciled with
Proclus' distinction between first and second teleiotetes, there might be ·an
advantage, given that a teleiotes acts as a final cause, to seeing whether a
distinction between different types of final cause could be reconciled with the De
Anima distinction.
The second advantage a Neoplatonic commentator would gain by seeing the
soul primarily as a final cause has to do with the sticky problem of the soul's
separability from the body. A Neoplatonic commentator on the De Anima faced
a big challenge when reconciling the Platonic doctrine of the soul's separability
and immortality, with Alexander's interpretation of Aristotle's psychology,
which holds that the soul is as inseparable from the body as form is from matter.
But given the Neoplatonic distinction between the efficient and final causes,
which transcend their effects, and the formal and material causes, which are
immanent in their effects, there might be some way to use the soul's being a
teleiotes, and hence its acting as a final cause, to advocate the position that the
soul is separate from the body.
Before reaping those benefits, the Neoplatonic commentators will first have to
deal with the fact that even though Proclus' teleiotes was clearly understood as
acting on its effect like a final cause, it is not so clear that Aristotle meant for
the soul to be seen as a final cause. If anything, given his insistence that
defining the soul as an entelekheia pointed to its being substance as form,
Aristotle probably meant for the soul to be viewed first and foremost as the
formal rather than the final cause of the body, and this was how Alexander
understood things. But a Neoplatonic commentator could point out that when
Aristotle starts talking, in De Anima 2.4, about the ways in which the soul is
the cause of the body, he maintains that the soul is the cause of the body not
only as form, but also as agent and end. There Aristotle says:

teleiotes as acting first and foremost as a final cause at (Ps.-?)Simplicius, in DA 1.1 (ad 402a4),
8,19-21 and in DA 3.4 (ad 429b10), 231.21-24.
88 Part /: The Ammonian Synthesis
L8
Aris!ii!al!s, Fi n-nafs, 38,3-5
Aristotle, DA 2.4, 4 l 5b9- l 3
[This is] because soul is a cause in the three ways we have
mentioned above; that is, it is a cause as the origin of motion, and
as that for the sake of which the body exists [wa-min ajli dhalika
kiina 1-jirmu = kai hou heneka], and it is [a cause as] the being
Uawhar = ousia] of bodies possessing souls.

In this passage Aristotle is not simply covering all his bases, for he is
consistent in maintaining that form, agent and end often coincide in natural
things. For example, in Physics 2.7 he says:

L9
Aris!ii!alis, af-Tabl<a, 137,15-138,3
Aristotle, Phys. 2.7, 198a22-28
It is clear that the causes are these, and that this is the extent of
their number. Since the causes are four, the natural philosopher
should know about each of them, and the question "On account of
what?" should be reduced to them such that the response to it is in
accordance with the doctrine of natural philosophy; I mean the
matter, the form, the mover, and that for the sake of which. Often
three of them devolve into one: for what the thing is, is
sometimes also that for the sake of which the thing is [fa-inna mi'i
sh-shay'u qad yakunu ay<f.an mi'i min ajlihi sh-shay'u = to men gar
ti esti kai to hou heneka hen esti], and the thing from which the
motion first comes, and that in a single sense.

This passage was interpreted by the commentators as implying that form and end
were identical in more profound ways than form and agent were.15 I shall try to
explain, by way of an example, what the commentators seem to have had in
mind here. The formal cause and final cause of my son's coming-to-be are
identical in number because the very same form of humanity instantiated in my
son's body at conception serves both as a formal cause - as his essence, in other
words - and as a final cause - as that in view of which the matter that made up
his body was set in motion. I, however, am the efficient cause of my son's
coming-to-be, with the result that my son and I are identical only in the weak
sense that we are both subsumed under the species "human" (that is, we both
possess the form of humanity), and not in the strong (numerical) sense of
identity, according to which the instantiation of the form of humanity in my son
is the very same instantiation of the form of humanity in me.
The notion that the formal and final causes are identical to each other in a
stronger sense than they are to the efficient cause, is supported in other passages
where Aristotle simply collapses the end into the form. For example, in a

15 See, for example, Philoponus, in Phys. 2.7 (ad 198al4), 297,30-298,6 and (ad 198a24),

301,7-302,3; and Simplicius, in Phys. 2.7 (ad 198a22), 363,32-364,15 and 265,14-18.
 4. Proclus, Ammonius and Philoponus 89

passage intended to encourage philosophers to search for the nearest or most

immediate cause of a thing, Aristotle appeals to his four-cause scheme to explain
the coming-to-be of a human being:

LlO
Aris~a~alis, Ma ba<da !-fabi'a II. 1073.15-1074,2
Aristotle, Metaph. 8.4, 1044a32-bl
When it is asked, "What is the cause?", it is necessary, since
"cause" is spoken of in many ways, that all of the ways in which a
cause can be, be spoken of. For example, "What is the cause as
matter of man?" It is the menstrual fluid. The cause as mover? The
semen. The cause as form? It is the essence [to ti en einai :::: ma
huwa bi-l-anniyyati]. The cause as that for the sake of which? It is
the end [wa-l- <illatu llati ka-lladhi min ajlihi fa-hiya t-tamiimu :::: ti
d'h6s hou heneka; to telos ]. (It is likely that these two are one
thing.) ·

But while this passage clearly supports a strong identification of the formal arxt
final causes, it seems to undermine the claim that the efficient cause is identical
to the formal and final causes in only a weak sense. This is because the example
of the efficient cause in this passage is not the father, but the father's semen. In
fact, Aristotle implies that strictly speaking, it is not even the father's semen -
viewed as a composite of matter and form - which is the efficient cause of the
son's coming-to-be, but only the form of humanity contained in the father's
semen. This is because the matter of the father's semen is not transferred at
conception to the son, but dissolves and evaporates at conception; the form of
humanity contained in the father's semen, by contrast, is transferred to the son
(Generation of Animals 2.3, 737a8-16 =F7 kawn al-J:iayawiin, 64,13-19).
The result is that the form of humanity which was contained in my semen arxt
transferred at conception to my son is the very same form of humanity which
was instantiated in my son's body at conception and which served as that in view
of which the matter that made up his body was set in motion. It turns out,
therefore, that the efficient cause is numerically identical - that is, identical in a
strong sense - to the formal and final causes. I suppose it could be argued that
although the very same form of humanity contained in my semen came to be
instantiated in my son's body at conception, that form of humanity acted as an
efficient cause only while it was contained in my semen. Once the form of
humanity was transferred from my semen and came to be instantiated in my
son's body at conception, it no longer served·as his efficient cause, but started to
serve as his formal and final cause.
This argument is not really cogent, however, for the most one could hope to
conclude is that the efficient cause and the formal and final causes are not
numerically identical at the same time: before the moment of my son's
conception, the form was contained in my semen; after the moment of
conception, the form was contained in him. And once employed, the criterion of
90 Part I: The Ammonian Synthesis

temporality could just as easily divide the formal cause from the final cause. The
final cause - the form of humanity in view of which the matter that made up my
son's body was set in motion - operated only during the process of formation.
Once that process of formation was completed, the form of humanity, now fully
instantiated in my son's body, operated only as a formal cause.
Partly in response to this confusion, Aristotle holds that some efficient causes
are intrinsic to their effects, others extrinsic (ton de toiouton enion men en
autois he arkhe tes kineseos estin ... enion d'exo = wa-min hiidhihi l-anwiici
mii btidii'u l-J:iarakati fihi ... [ wa- ]min khiirijin ). 16 And in other passages
Aristotle implies that the semen is not so much an efficient cause as something
which the efficient cause (i.e. the father) uses as an instrument (hOs organoi =
mithla iilatin). 17 To sum up, then, while Aristotle holds that the formal, final
and efficient causes may sometimes be reduced to the same thing, it is unclear
whether the formal cause can be identified in a stronger way with the final cause
than either can with the efficient cause. It seems that this is true if the efficient
cause is something extrinsic to the effect, as a father is to his son; but false if
the efficient cause is something intrinsic to the effect, as the form of humanity
contained in the father's semen and transferred at conception is to the son.
Even the strong identification of the formal and final causes can be
undermined. For in one very important case, the final cause cannot be held to be
identical to the formal cause: the Unmoved Mover is the final cause of celestial
activity, but It is not the formal cause or essence of the spheres. Because the
Unmoved Mover is not confined to a material body, and because form, by
Aristotle's reckoning (Physics 2.1, 193b4-5), is never separable from matter _in
the real world, but only when speaking in purely logical terms (kata ton logon),
the Unmoved Mover cannot be a formal cause. In other words, because Aristotle
holds that the Unmoved Mover acts on Its effects as a final cause, and because
the Unmoved Mover is separate from matter, there must therefore be some final
causes which are separate from matter and hence not collapsible into formal
causes.
To the commentators, however, it still seemed that of form, agent and end,
form and end were most closely identifiable. It also seemed that when push came
to shove - as in LIO, the Metaphysics passage - Aristotle was willing to
collapse end into form, but not vice versa, at least in the case of sublunary
causation. In other words, assuming form and essence are the same, Aristotle
would be prepared to say that end was subsumable under essence, but not that
essence was subsumable under end. There were some indications, therefore, that
of form and end, form was more basic. So despite Aristotle' s claim in De Anima
2.4 that the soul is the cause of the body as form, end and agent together, his
discussions elsewhere of how form, agent and end coincide implied that form was

16
17
GA 1.18, 724a3 l-35 =Fi kawn al-hayawiin,
•
29,13-16.
GA 1.22, 730b20 =Fi kawn al-fzayawiin, 47,7.
 4. Proclus, Ammonius and Philoponus 91
still somehow the most basic of these three non-material causes. The perception
that the form was primus inter pares was reinforced, of course, by Aristotle's
assertion in De Anima 2.1 that the soul is substance as form.
Nevertheless, Aristotle's willingness to see the soul as a final cause - however
subsumable final causality may be under formal causality - gave the
conunentators at least some flexibility to interpret the soul's being a teleiotes
as referring to its being a final cause. In order to explain more precisely how the
soul served as a final cause, and thereby reinforce the connection between the
soul and the Neoplatonic understanding of teleiotes, the commentators turned to
another Aristotelian distinction, that between two types of "that for the sake of
which": to hou ("that in view of which") and to hOi ("that for the benefit of
which").
Aristotle articulates the distinction in De Anima 2.4 and Metaphysics 12.7,
and alludes to it in Physics 2.2 and Eudemian Ethics 8.3. He asserts that to hou
heneka, that for the sake of which, may be spoken of in two ways: in the
genitive, as to hou (or tinos), "that in view of which"; and in the dative•. as to
hOi (or tini), "that for the benefit of which". 18 Aristotle' s distinction is very
compressed and far from transparent. 19 It was also translated variously into
Arabic: in the De Anima translation edited by Badawi, it appears as wa-ma ' na
min ajli caza wijhatayni iJ:idahuma lahu wa-l-ukhra fihi; in the Arabic version
of Themistius' Paraphrase of the De Anima, it appears as wa-lladhi bi-
sababihi yuqalu 'ala {i.arbayni aJ:iaduhuma lladhi min qibalihi wa-l-akharu
lladhi lahu; and in the Metaphysics translation it appears as wa-dhalika anna
ma min ajlihi yujadu li-shay'in wa-li-dhi shay'in (reading li-dhl for Bouyges'
li-dha). 20
The contexts in which Aristotle introduces this distinction are quite different.
In the Physics, the distinction is meant to show that nature is similar to art in
that the form - that in view of which an artifact is created, such as the shape of
an axe - must be taken into account by the natural philosopher, and not just the
matter. Nature is also like art because the matter the artisan uses in creating the
artifact is determined by the use which the "customer" or "end user" - that for
the benefit of which - makes of the artifact, given its form. Thus an axe is
made of iron and not clay, given the function - cutting - which the form "axe"
represents to the person who benefits from using it. 21

18 to men hou, to de hOi: DA 2.4, 415b2 and esti gar tini to hou heneka [kai] tinos: Metaph. 12.7,

1072b2-3; cf. Phys. 2.2, 194a35-36 and EE 8.3, 1249b9-I 9.

19 The fullest treatment of Aristotle's distinction between to hou and to h6i is that of Gaiser

1969.
20 Aris\ii\alis, Ff n-nafs, 37,15-16; Themistius, in DA 2.4 (ad 415b2-3), 50,11 (= Thamas\iyiis,

Sh~r(l Kitiib an-nafs li-Aristiitiilis, 68,12-13); Aris\ii!alis, Mii ba<da f-fabl 0 a Ill, 1599,3.
The example Philoponus uses is that of a door: "Now 'end' is two-sided: 'that in view of
which' and 'that for the benefit of which'. For example, the form of a door is a 'that in view of
which' end, since the artisan strives for this, being what people call an 'aim' [skopon]. As for 'that
for the benefit of which', well, the door is not on account of itself, in order that its form should
92 Part I: The Ammonian Synthesis

So far so good. But how exactly is the distinction useful in explaining the
final causation of natural things, such as souls, as opposed to artifacts, such as
doors or boats? When, in the De Anima passage, Aristotle applies the distinction
between the final cause qua to hou and the final cause qua to hoi to living
things, the distinction seems to be based not so much on the difference between,
respectively, form and user, but rather on the difference between what is eternal
and divine and what is mortal. 22 There seem to have been two ways to reconcile
the Physics' form/user criterion with the De Anima's eternal/mortal criterion,
depending on whether one understands the "that for the benefit of which" in the
De Anima to be referring to the soul-body composite, or to the soul alone.
The first interpretation, advocated by Themistius, runs along the following
lines: that in view of which nature creates living bodies is the form of divine
immortality which nature directs them to attain as a species, through the faculty
of reproduction in their souls. That for the benefit of which nature creates
living bodies, on the other hand, are the ensouled bodies themselves - that is,
the soul/body composites - which attain the form of immortality through
reproduction. 23 The second interpretation, which strikes me as more sensible, is
advocated by Philoponus, and goes like this: as with the first interpretation, that
in view of which nature creates living bodies is the form of divine immortality
which nature directs them to attain as a species, through the faculty of
reproduction in their souls. But in contrast to the first interpretation, that for the
benefit of which nature creates living bodies are the souls that use the bodies as
instruments in order to attain immortality as a species. Philoponus also
expresses the stipulation that that for the benefit of which does not apply to
things - such as metals and rocks - which possess no souls to use their bodies
as instruments; instead, only that in view of which applies to inanimate
things. 24 What is more, commenting on the issue of formal-final-efficient
identity discussed in L9, Philoponus says that substantial form is identical with
end only in the "in view of which" sense (i.e., qua to hou), and not in the ''for
the benefit of which" sense (i.e., qua to hOi). 25

come into being, but rather in order that it serve a function for man; thus the door comes into being
for the benefit of man, being in some sense an end, so that it [i.e., the door] may protect the
house"; Philoponus, in Phys. 2.2 (ad 194al2), 230,5-10. Olympiodorus follows Philoponus at
Ol~mpiodorus, in Meteor. 4.2 (ad 379b8-10), 284,7-19.
2 In the comments to his edition of the De Anima, Ross says "In II. 1-2 A. says that eternity and

divinity are the hou heneka in the sense of being that at which all things aim (to hou]; in II. 20-21 he
says that soul is the hou heneka in the sense of being that in whose interests [to hOi] the bodies of
animals and plants exist" (Ross 1961, 228).
23 Themistius, in DA 2.4 (ad 415b3-7), 50,15-26 (= Thiimast_iyus, Sharl:i Kitab an-nafs li-

Aristufatis, 69,2-12).
24 Philoponus, in DA 2.4 (ad 415b2 and 15), 269,25-270,32 and 274,1-23. See also (Ps.-?)
Simplicius, in DA 2.4 (ad 415b2), 110,31-36 and (ad 415bl 7 and 20), 111,22-112,2.
2 Philoponus, in Phys. 2.7 (ad 198a14), 297,30-298,6 and 301,18-302,2. Proclus claims that

divine souls can only be final causes qua to hou because divine souls are separate from bodies and
hence are not tied to material instruments; Proclus, in Tim., Vol. 3, 254,10-31 (ad 4107).
 4. Proclus, Ammonius and Philoponus 93
In any case, the distinction between to hou and to hOi was useful in helping
Philoponus understand more precisely the ways in which the soul could be seen
to be a final cause, and thereby reinforced the sense in which the soul could be
seen as a Neoplatonic teleiotes. Thus:

2b[ii]'

Phys. l't:
l. entelekheia = teleiotes qua "state of being completed" = (the potential X's)
state of completeness (e.g., "[the bricks' and mortar's] being constructed")
2. entelekheia = teleiosis qua "process of being completed" = process of
completion (e.g., "construction")
DA l":
la. entelekheia = teleiotes qua "state of having been completed" = state of
completeness (e.g., "having been born")
1b. entelekheia = teleiotes qua "resulting substantial form" = essence (e.g.,
"humanity")
le. entelekheia = teleiotes qua "final cause qua to hou" = state in view of
which teleiosis occurred (e.g., "being a perfect human")
2a. entelekheia = teleiotes qua "state of having been completed" = state of
completeness (e.g., "having come to know")
2b. entelekheia = teleiotes qua "resulting substantial form" = essence (e.g.,
"knowledge")
2c. entelekheia = teleiotes qua "final cause qua to hou" = state in view of
which teleiosis occurred (e.g., "being perfectly knowledgeable")
3. entelekheia = teleiotes qua "final cause qua to h8i" = user (e.g., "the soul
[uses the bodily organs, in view of being perfectly knowledgeable, to cause the
teleiosis of coming-to-know]")

Phys. 2nd:
1. entelekheia = teleiotes qua "state of having been completed" = state of
completeness (e.g., "having been constructed")
2. entelekheia = teleiotes qua "resulting substantial form" = essence (e.g.,
"houseness")
3. entelekheia = teleiotes qua "final cause qua to hou" = state in view of
which teleiosis occurred (e.g., "being a perfect house")
DA 2"d:
1. entelekheia = teleiotes qua "state of being completed" = (the potential X's)
state of completeness (e.g., "[the intellect's] contemplating")
2. entelekheia = teleiosis qua "process of being completed" = process of
completion (e.g., "contemplation")
94 Part /: The Ammonian Synthesis

The major difference between Themistius' and Philoponus' schemes is that as a

result of the new, Prodan understanding of teleiotes, Philoponus can now see
teleiotes as referring simply to "final cause" and not, as Themistius had, as
referring to all the different senses of endedness. In other words, Philoponus now
replaces "endedness" with "end" in the semantic field of teleiotes. By this
choice, however, Philoponus is forced to forfeit the very handy notion of
endedness, which provided Themistius with four new senses of teleiotes with
which to reconcile Aristotle's seemingly incompatible uses of entelekheia in
De Anima 2 and Physics 3: allotelic endedness, which can mean both the "state
of being directed towards an end other than itself' and the "state of serving as an
end for something other than itself'; and autotelic endedness, which can mean
both the "state of being directed towards an end which is nothing other than its
own exercise" and the "state of serving as its own end".
On the other hand, by understanding teleiotes in the Prodan rather than
Thernistian way, as end rather than endedness, Philoponus gains two new senses
of teleiotes, the final cause qua to hou and the final cause qua to hoi. Still, the
deal seems hardly fair. Philoponus has lost four senses of teleiotes and gained
only two, and therefore he must surely be less able to reconcile Aristotle's
seemingly incompatible uses of entelekheia than Themistius was. I believe this
is true. What Philoponus' move does show, I think, is that he was more worried
about reconciling Platonic and Aristotelian doctrines about the separability of the
soul than he was with reconciling Aristotle's seemingly incompatible uses of
entelekheia. In other words, Philoponus had integrated the greater sumphOnia of
harmonizing Plato's and Aristotle's psychologies, into the lesser sumphOnia of
sorting out what Aristotle means when, for example, he distinguishes between
final causes which are to hou and those which are to hoi. But how exactly does
understanding teleiotes as end help Philoponus create a compromise between
Plato and Aristotle on the issue of the soul's separability?
The first thing to be stressed is that Aristotle's position on the soul's
separability - or at least the intellect's separability - is sufficiently
underdetermined that Neoplatonic commentators such as Philoponus (and, as we
shall see, Avicenna as well) could" point to a number of passages in Aristotle's
works where Aristotle allows for the possibility of, while not expressly
advocating, some version of separability. The following are the most prominent
of those Aristotelian passages:

Lil
Aris!ii!iilis, Fi n-nafs, 20, I 0
Aristotle, DA 1.4, 408b 18-19
As for the intellect, it seems to come into being [ya<ta>kawwinu
enginesthai] as a cause of the soul, permanent and never passing
away [ghayra fiisidin = kai ou phtheiresthai].
 4. Proclus, Ammonius and Philoponus 95
L12
Aris~ii~a!Is,Ff n-nafs, 31,13-15
Aristotle, DA 2.1, 413a3-9
It has become clear that neither the soul nor any of its parts is
separable from the body, because the entelechy of some living
things belongs only to their parts. But it is not yet clear if the
soul is the entelechy of the body as the pilot [is] of the ship
[riikibi s-safinati =plater ploiou].

L13
Aris~ii~alis,Ff n-nafs, 72, 19-21
Aristotle, DA 3.4, 429a22-25
It is doubtless the case that the intellectual [part] of the soul,
called "intellect" (namely, that by which it reasons and holds
opinions) does not actually exist in anything before it apprehends
the thing by means of its comprehension. For this reason it does
not follow that it is intermingled with the body [li-dhiilika iii
yajibu an yakuna mukhiilafan li-l-jirmi = dio oude memikhthai
eulogon auton toi somati].

Ll4
Aris~ii~alis, Ff a 'lfii ) aHzayawiin, 10,24-11,8
Aristotle, PA 1.1, 64la32-64lbl0
When someone considers what has just been said, a difficulty
arises, because it is difficult [to decide] whether the natural
philosopher should wish to speak about all soul or about one
[particular type or part of] soul. If it is about all soul, then there
will be no difference at all between that person and the natural
philosopher. For the philosopher's intellect has as its object
intelligible things. Natural philosophy would have as its object
knowledge of all things; because all correlatives are the object of
the same study (as is the case with sensation and sensible things),
it will belong to the same science to study intellect and
intelligible things, given the fact that they are correlative. Yet
neither all soul nor all its parts is an origin of movement: the
principle of growth is what is found in trees; the principle of
alteration is what is found in the [faculty of] sensation; and the
principle of locomotion is [found in yet] another thing, which is
not what the intellect is found in. This is because locomotion is
also found in other types of animal, but intellect is found in none
of then:i [i.e., in none of the other types of animal]. It is therefore
clear to us that the discussion is <not> about all soul, because not
all soul is nature, but instead about only some part of it, since the
soul contains many parts.

L15
Aris~ii~a!Is,Ff kawn al-l}ayawiin, 63,22-23
Aristotle, GA 2.3, 736b28-30
It follows from our discussion that we should say the intellect
alone enters from outside [i.e., the body], and that only it is
96 Part /: The Ammonian Synthesis
divine, on account of the fact that its activity is not at all
associated with any bodily activity.

L16
Aris!li!lilis, Ma ba<da r-tabfca II, 706,6-7
Aristotle, Metaph. 6.1, 1026a5-6
[It is clear] that it belongs to the natural philosopher, in [his
branch of] knowledge, to investigate a certain type of soul,
namely, the one which is not without matter ffa-anna li-f-fabfciyyi
fl z-cilmi an yan?,urafi nafsin ma wa-hiya llatf la takunu min ghayri
<un$urin = kai dioti kai peri psukhes enias theoresai tou phusikou,
hose me aneu tes hules estin]

L17
Aris1a1alis, Ma bacda t-tabf<a ll/, 1486,13-1487,2
Aristotle, Metaph. 12.3, 1070a24-26
Now whether anything [viz., any form] remains afterwards [viz.,
after its effect] requires investigation, because there is nothing
preventing certain things [from remaining after their effects]. An
example is whether the soul is this way; not all of it, though, but
the intellect, since not all of it [viz., not all of the soul] can
possibly [remain after its effect].

Most modern commentators would argue that the balance of evidence still tips
towards inseparability, given the heavy weight of Aristotle's tying of
entelekheia to form in DA 2.1. Nevertheless, Aristotle's position will, I hope,
seem more underdetennined than has generally been supposed. 26 It is easy to
imagine that for a Neoplatonic commentator, the scale could be tipped back
towards separability given that since Alexander's time the soul was held to be a
teleiotes; given that since Proclus' time a teleiotes was held to cause its effect
as a final cause; and given that since Syrianus' time final causes were held to
transcend their effects.
As seen above, moderate Neoplatonic commentators such as Philoponus -
commentators who sought concessions not only from Aristotle and Alexander
but from Plato and Plotinus as well - argued that the soul uses the body as an
instrument and that in this sense the soul is a final cause qua to h6i (as user and
beneficiary, that is) of the body. As a final cause qua to hOi, the soul can be
seen to be immanent in its effect in order to accommodate Aristotle and
Alexander. In order to accommodate Plato and Plotinus, however, the intellect or
human rational soul - the type of soul which is eternal and divine and which acts
as a final cause qua to hou - will be held to be separate. The fact that the final
cause qua to hOi seems to be immanent in its effect while the final cause qua to

26 An example of the tendency among some scholars of ancient philosophy to see Aristotle's

theory of the soul-body relationship as more consistent than I am presenting it as being, and to
blame the Neoplatonists for willfully corrupting it, is Blumenthal 1990. For a recent defense of the
idea that Aristotle allowed for the possibility of the soul's - or at least the intellect's - separability,
see Bolton 1978.
 4. Proclus, Ammonius and Philoponus 97
hou seems to transcend its effect is reflected in Philoponus' reluctance to commit
himself to holding that a teleiotes in and of itself will imply either separability
or inseparability. 27
How does a separate teleiotes act as a final cause, drawing its effect up to its
higher ontological level? First we must be precise about what a teleiotes is a
final cause of. I have already discussed the fact that to Neoplatonists such as
Philoponus what a teleiotes causes or explains is a thing's corning-to-
perfection, its teleiosis. So strictly speaking, it would be incorrect to say that
the soul is the teleiotes of the body; rather, the soul is the teleiotes of the
body's corning-to-perfection. Even more precisely, the soul is the final cause of
the body insofar as the body is seen as an element of reversion, as something
striving upwards towards self-perfection. The point is that not only the body but
also the soul and even the human intellect can be regarded as each undergoing
comings-to-perfection (teleiOseis) whose ultimate end is a separate teleiotes.
This separate teleiotes is apprehended during self-intellection, the activity which
the Neoplatonists regarded as the final step in a hurnan's reversion to a higher
level.
I have already spoken in Chapter 2 about why the specific difference, rather
than the species, should be seen as corresponding to the teleiotes in this
context. Neoplatonists such as Arnrnonius agreed, arguing that only the specific
difference can be seen to be truly perfective (teleiOtike), that is, cause-like.28 In
short, what seems to happen during self-intellection is that my species changes
from "rational animal" to "active intellect". In other words, once the specific
difference ("noetikos") of my original species ("anthropos") has been attained as
a teleiotes, my former specific difference ("noetikos") becomes my new genus
("nous"). The new specific differences which set one species of the genus "nous"
apart from another species of the genus "nous", are "active" (energeiai) and
"potential" (dunamei).
The reversion carries on upwards in the sense that the specific difference
"active" then becomes its own genus - (intellectual) activity - with the specific
difference now being "eternal" or "intermittent". The highest a human can attain
in life is intermittent (i.e., non-eternal) intellectual activity. In death (according
to Avicenna, as we shall see) or following the completion of a certain number of
transmigrations (according to Proclus and some other Neoplatonists), eternal
intellectual activity can be attained.

27 At in DA 2.2 (ad 412b9), 219,9-10, Philoponus says that taken in and of itself, the teleiores of

a thing will be inseparable from that thing; but at in DA 2.1 (ad 4!2a3), 206,18-23, he says that
taken as the thing's perfecting activity (ten heautou energeian teleiotiken) - taken as a cause, that
is - the teleiotes will be separable from the thing.
28 Ammonius, in Int., 71,5-11; Proclus was slightly less certain, sometimes (e.g., Inst. Theo/.,

Prop. 108, 96,9-22) maintaining that one can become perfect through imitating either a genus or a
specific difference, other times (e.g., in Ale. I, 18,4-7; 247,10-11; 248,2-4; 278,3-13) implying that
it is only during self-intellection - that is, during intellection of the specific difference "rational" -
that a thing reverts in a proper sense.
98 Part I: The Ammonian Synthesis

In any case the notion that self-intellection is the primary mechanism of

reversion, and that this reversion consists more generally of an ascension to
teleiotetes - to final causes, that is - by means of imitation, is present in
virtually all Neoplatonic thinkers. 29 Self-intellection is the only truly divine
activity since it is the only truly self-sufficient, and hence perfect, activity, a
precept which appears to have come from Aristotle (EN 10.7, 1177b15-26), as I
mentioned in Chapter 2. 30 The view that because self-intellection requires
nothing else to bring it about, it is complete and perfect in the sense of self-
sufficient, will, as I shall show in Chapters 6 and 11, prove influential on
Avicenna's theories of perfection and the perfect. For now, I shall discuss how
all these Greek terms were rendered into Arabic by the translators of the eighth,
ninth and tenth centuries.

29 Plotinus, Enneads, 1.8.1,7-10; Proclus, Theo/. Plat., Vol. l.14, 66,3-5 (where he quotes

Parmenides' famous - and variously interpreted - dictum in Fr. 3 that being is a necessary
precondition of, or perhaps even identical to, thinking [to gar auto noein estin te kai einai], and also
alludes to Laws X 897B); Inst. Theo/., Prop. 32, 36,3-4 and 186, 162,13-23; in Ale. I, 291,3-5 and
308,4-5; in Parm., 627,12-14; 771,26-30; 789,31-790,4; 1148,1-17; Philoponus, in DA 1.3 (ad
407b9), 138,25-139,2 (citing the beginning of Plotinus, Enneads II.2); and (Ps.-?)Simplicius, in DA
2.1 (ad 412a10), 84,13-22 and 2.1 (ad 412a22), 88,13-89,12.
30 See also Proclus, Theo/. Plat., Vol. 3.6, 22,3-6; Inst. Theo/., Prop. 114, 100,16-27.
 5. Greek into Arabic
The Greco-Arabic Translations and the Early Arabic
Philosophers

In this chapter I shall show that the philosophical tendencies of the Ammonian
synthesis, which I have just described at length, had an impact on the choices
made by the Greco-Arabic translators when rendering into Arabic Greek words
relating to the idea of perfection and the perfect, as well as on the ways in which
the early Arabic philosophers such as al-Kindi and al-Farabi formulated their own
ideas about perfection and the perfect. I hope it will become clear that the
translation movement was less a transparent, lexicographical process, than it was
part and parcel of the Ammonian synthesis described in the previous chapter.
What I mean is that the Greco-Arabic translators chose words based on certain
current philosophical presuppositions, namely, those of Ammonius and his
successors in the sixth century.
Scholars such as Walzer, Frank, Endress, Gatje, Daiber and Arnzen have each
briefly discussed the translation into Arabic of the terms entelekheia, teleion
and teleiotes, some in the form of glosses to their editions of the Arabic
translations of Greek works, others in the form of footnotes to their studies of
the translation history of individual texts; none pretends to offer the reader a
diachronic analysis. 1 My objective for this chapter is to propose - very
tentatively, because the evidence is still sketchy - a narrative of the Arabic
fortuna of an interrelated cluster of Greek terms, those just mentioned as well as
the terms telos, telikos and teleiotikos.
According to Gutas' useful summary, the Greco-Arabic translation movement
can be divided into five stages: (1) the first and earliest period, in which Euclid's
Elements and Aristotle's Rhetoric were translated; (2) the period of the circle of
translators surrounding the philosopher al-Kindi (d. 875); (3) the period of
I:Iunayn b. Isl)aq (d. 873) and his school; (4) the period of Qus~a b. Luqa (d. ca.
912); and (5) the period which Gutas calls the "stage of scholarly emendation,
i.e., on the basis of contents rather than original texts". 2 Based on the evidence I
will present here, the only modification to Gutas' scheme I would suggest is
incorporating the retranslating efforts of Isl)aq b. I:Iunayn (d. 910), which seem
to bridge the fourth and fifth periods: though chronologically they fall in fourth
period, lsl)aq's efforts are distinct from QusWs; and though their aim of
emendation was that of the fifth and final period, lsl)aq seems to have had access
to new texts, at least in the case of the De Anima.

1 Walzer 1953, 95-96; Frank 1958-59, 242n.2 and 249; Endress 1966, 121; Glitje 1971, 34, 39

and 43; Daiber 1980, 5, 157, 369 and 456-7 (who also mentions the Syriac intermediary renderings
of entelekheia); and Arnzen 1998, 375-7.
2 Gutas 1998.
100 Part /: The Ammonian Synthesis

Inevitably, the borders between any such periodization are fuzzy, and one can
detect transitions in word choice even within the work of a single translator. One
overall conclusion I am inclined to draw is that much of the effort of the fourth
and fifth stages - and particularly that of lsl,laq b. I:Junayn - was devoted to
cleaning the mess left by the disparate and sometimes conflicting efforts of the
translators of the al-Kindi circle. The mopping up was undertaken with a view to
creating a more systematic Greco-Arabic lexicon, one which superseded the old,
al-Kindi-circle terminology because the new one could be used consistently in
more than one translated text.
This effort at standardization did succeed, in the sense that philosophers such
as Avicenna came to use the revised Aristotelian terminology in much of his
work, particularly in canonical passages such as his definitions of important
concepts such as motion (al-flaraka) and the soul (an-nafs). Nevertheless,
enough of the al-Kindl-circle vocabulary not only remained in some of the
unrevised translations, but had infiltrated the work of earlier philosophers (I am
thinking of al-Kindi himself but most of all al-Farabi), that many of the earlier
terms and the ideas expressed by those terms came to be present in Avicenna's
work, though often in less prominent passages than the ones containing the
standardized vocabulary of the later periods. Some of the most important of these
vestigial, early-period terms are those that express the concepts of perfection and
the perfect.
The first translation which required a standard rendering of any of the Greek
terms I have been discussing was that of Euclid's Elements, which used the
Arabic tiimm to translate the Greek teleios in Euclid's discussions of perfect
number (arithmos teleios). 3 This very early use of tiimm to translate Euclid's
teleios may have prompted Ustat, a member of the al-Kindi circle and - to the
best of our knowledge - the translator of most of the version of the
Metaphysics contained in the lemmata of Averroes' Tafsir Mii ba<da !-{abr<a
(Great Commentary on the Metaphysics) and edited by Bouyges, to render to

3 According to De Young 1981 (Vol. 1, Pt. 1, 1-3), there were three translations of the
arithmetical sections of Euclid's Elements, the first two by al-J:Iajjiij b. Yusuf b. Ma\ar, which are
not extant, and the last by lsl:iiiq b. J:Iunayn and revised by Thlibit b. Qurra, which is the translation
edited and translated into English by De Young. There are two versions of this last translation, of
which one (majmu' "a") lists the definition as Bk 7, Def. 24, and reads al-'adadu t-tiimmu (De
Young 1981, Vol. 1. Pt 1, 5,9); and the other (majmu' "b") lists the definition as Bk 7, Def. 23, and
reads al-'adadu lladhl yuqiilu t-tiimmu lahu (De Young 1981, Vol. I. Pt 2, 321,6). We can surmise,
however, that tiimm was the translation of teleios in the earlier versions by al-J:Iaijiij, because
tiimm was also used in the Leningrad MS Akademia Nauk C2145, which is a more literal
translation than the Isl:iiiqffhlibit version and which therefore fits Ibn al-Nadim's description of al-
l:lajjiij's translation; see De Young 1981, Vol. 1, Pt I, 28. Also see the translation by Thlibit b.
Qurra (d. 901) of Nicomachus of Gerasa's Introduction to Arithmetic, where teleios is again
rendered into Arabic by tiimm: Niqumlikhis al-Jliriisini, K. al-madkhal ilii 'ilm al- 'adad, 36,6-40,23;
teleios is rendered into Arabic as at-tiimm at 36,10; 37,5.15; 39,6.19; 40,15.16.21; "superabundant"
is translated as ziiyid (= zii'id) 'a/ii t-tamiimi, and "deficient" is translated as niiqi~ 'ani t-tamiimi.
Avicenna himself discusses perfect (tiimm), superabundant (zii'id) and deficient (niiqi~) numbers
in K. ash-shifii'/Riy&jiyyiit 2: al-l:fisiib, 30,21-23 and 32,7-33,22.
 5. Greek into Arabic 101
teleion of Metaphysics 5.16 as at-tiimm, as is apparent in L4. The same
translator may also have reasoned that just as teleion was the adjective deriving
from the noun telos, so telos should be rendered into Arabic by the noun tamiim
from which the adjective tiimm is derived; this is made clear in Table 1 in
Appendix I.
More specifically, Table 1 shows that telos was rendered as ghiiya in Book 2
(which is probably just an artifact of Averroes' use of IsJ:iaq b. I:Iunayn's later
translation of Book 2 in addition to Ustat's); then, in Book 3, as ghiiya and as
tamiim, and also as ghiiya wa-tamiim, a hendiadys that also turns up in the
Arabic version of the Parts of Animals;4 and then Ustat used only tamiim in the
rest of the Metaphysics. As Table 2 shows, Ustat also rendered entelekheia
(mainly in its adverbial use in the dative case) as ficl in Books 1-8; and as
tamiim in Book 9 onwards, except for once in Book 12, where an instance of
entelekheia in the dative is rendered as ft l-kamiil (which is probably just an
artifact of Averroes' use of Abu Bishr Matta's later translation of Book 12 up to
1072bl6). My guess is that Ustat began by rendering entelekheia as ficz,
believing it to be virtually interchangeable with energeia. But when confronted
with Aristotle's attempt in Metaphysics 9 - his longest sustained discussion of
potentiality and actuality - to distinguish between energeia and entelekheia,
Ustat changed tack and began rendering entelekheia as tamiim.
The result of Ustat's switch was that an attentive but Greekless reader of the
Arabic Metaphysics such as Avicenna (who, after all, claimed in his
autobiography to have read the Metaphysics 40 times), turned to canonical
passages on the four causes in Metaphysics 5.2, and found telos, or final cause,
rendered as tamiim, and contrasted with the form, the matter and the agent; and
then turned to canonical passages on substance and actuality in Metaphysics 9
and found entelekheia, or actuality, also rendered as tamiim, and contrasted with
potentiality.
My point is that in the translation of Aristotle's Metaphysics itself there
existed a lexical momentum pushing close readers such as Avicenna to assume
that the ideas of actuality and final cause fell together naturally. This conflation
of telos and entelekheia could be nothing more than an accidental result of a
translator's word choice. But to my mind, the fact that Ustat seems not to have
felt it necessary to revise his translation in order to separate the two concepts is a
sign that he was influenced by the profound philosophical shift that had occurred
among the Neoplatonists who glossed entelekheia with teleiotes, and then
spoke of the teleiotes as operating like a telos; and that in this sense Ustat, like
other early Greco-Arabic translators, cannot be extracted from the philosophical
continuum of the Ammonian synthesis.

4 PA 2.3, 650a27 = FT a'f!ii' al-J:iayawiin, 32,23; 3.6, 669a13 = 80,3; 3.9, 672a5 = 87,12; and

=
3.14, 675a16 96,6; cf. also I.I, 641b25 = 11,21.
102 Part /: The Ammonian Synthesis

The philosophical shift from entelekheia to teleiotes is also evident in the

word choices of other translators in the al-Kindi circle, another of whose
members was the compiler of the Uthilliljiyii, al-I:Iim~i. 5 It seems from L7 that
given al-I:Iim~i's use of the transliteration anfaliishiyii to render the entelekheia
of the Enneads, his use of tamiim probably rendered a term other than
entelekheia - teleiotes seems the obvious candidate - in whichever bits and
pieces of Greek Neoplatonic source-text the non-Enneads material in the
Uthilliljiyii is a translation of. Of course al-l:lim~i's use of tamiim in this context
may not have been a direct translation from some ghostly Greek original at all,
but rather his own importation into the Uthilliljiyii of the Arabic version of a
Greek term - again, presumably teleiotes - that he had seen as a gloss of
entelekheia in another text circulating among al-Kindi's team of translators.
It does seem that the point of the L7 passage is that the soul may be called the
tamiim of the body, as long as it is clear that a tamiim can be separable or
inseparable, unlike a nasty anfaliishiyii, which is usually understood as being
inseparable. To my mind this shows that al-l:lim~i was aware of Proclus' and the
Ammonian commentators' use of teleiotes in the context of the soul, as
opposed to Alexander's use of teleiotes; for even though Alexander used
teleiotes to gloss entelekheia, he was clear that the soul was inseparable from
the body just as form was inseparable from matter.
Another interesting feature of L7 is the use of the term tamiimiyya, as in ~ilra
tamiimiyya, meaning "perfecting form". This almost certainly rendered, or
echoed, eidos teleiotikon (or perhaps eidos telesiourgon), rather than eidos
telikon, "final form", which makes no sense there, or anywhere for that matter.
What the compiler seems to have wanted to do first in L7, is to distinguish
between form qua shape (eidos qua morphe) and form qua essence or substance
(eidos qua ousia), a distinction which appears, as I mentioned in Chapter 4, in
Philoponus' commentary on the De Anima. Though an entelekheia can refer to
either type of eidos, a teleiotes will refer only to eidos as substance, and not to
eidos as shape. The soul is an entelekheia only in the sense that it refers to
eidos as substance, not in the sense that it refers to eidos as shape; therefore, it
is fine to call the soul a tamiim-teleiotes because it refers only to the substantial
sense of eidos.
In the third paragraph of L7 the compiler then argues that even though the
Materialist philosophers - perhaps Alexander is in the frame here - sometimes
spoke of this latter type of form - i.e., eidos qua ousia - as "perfecting", what
they really had in mind was in fact a passive, natural perfecting form, one which
is inseparable from the body in which it inheres. Their "perfecting form" is
therefore not truly perfecting in a Neoplatonic sense, since the Neoplatonists
consistently used eidos teleiotikon to refer to something separable, with

5 On the very murky history of this work see Zimmermann 1986, in tandem with the review by

Rowson 1992.
 5. Greek into Arabic 103
genuine agency: to separate Forms, not to inseparable forms. 6 If the soul is to
be spoken of as a perfecting form, it must be so in that truly active, productive
sense, the sense in which the Forms are active and productive, and not in the
natural, passive sense which the Materialists have in mind. Again, the influence
of the Neoplatonic turn to causation is clear: the soul is a perfection in the sense
that it is a cause of the body just as a separate Form is the cause of its effect, and
not in the sense that it is an inseparable form, be it a morph€ or even a natural
ousia.
The renderings of the relevant Greek terms into Arabic in the Arabic versions
of Proclus' writings, mainly from Proclus' Elements of Theology and often
disguised as essays by Alexander of Aphrodisias, are listed by Endress in his
Proclus Arabus.7 They generally follow the pattern described above, using tamm
to translate teleion, and either transliterating entelekheia as anfalashiya or
rendering it as tamam. But the most innovative and important feature of these
translations is the introduction of the hendiadys tamm kamil (sometimes tiimm
wa-kiimil) as a translation of teleion, and, less often, tamam wa-kamiil as a
translation of entelekheia and teleiotes. To the passages listed by Endress I can
add the following other instances where k-m-l has been introduced into the
picture. In one "Alexander" translation, which Hasnawi has argued is in fact an
excerpt from Philoponus' De aeternitate mundi contra Proclum, the
Aristotelian definition of motion is cited, and kamal is used once, and tamam
used four times, to render entelekheia. 8 In the Arabic version of what is
genuinely Alexander's Quaestio 3.3 (De sensu = Fl l-J:iiss), we find the
hendiadys mutammima mukmila and ila tamamihi wa-kamalihi. 9 The hendiadys
also appears as a translation of teleiosis, and of verbs stemming from teleiosis,
in the Arabic versions of Aristotle's Generation of Animals and History of
Animals. 10 The use of this hendiadys added the Arabic root k-m-1 to the quiver of
terms used by the translators, who in the fourth and fifth stages of the translation
movement tended, as I shall soon show, to discard t-m-m almost entirely and use
k-m-l instead.
The final important text which appears to be from this early period, but about
which there has been much discussion among scholars, is the text edited by
Badawr and thought by him to have been the final, revised translation of Is~aq b.

6 Syrianus, in Metaph. 13.4 (ad 1078b32), 107,38-108,8; Proclus, Inst. Theo/., Prop. 45, 46,12-

19; and in Remp., Vol. 1, 270,13-272,7 and 286,11-287,10; and Philoponus, in DA 2.1 (ad 412a3),
206,18-23.
7 Endress 1973, 67 and 115-117.
8 Maqiilat Aliskandar ft anna /-ft'/ a'ammu mina l-fiaraka 'a/ii ra' y Aris[u, 293,8 (kamiil) and

294,9.11.13.14 (tamiim). For the arguments that its attribution to Alexander is wrong and that it is in
fact a translation of excerpts from Philoponus' De aeternitate mundi contra Proclum, see Hasnawi
1994, 68-75, and Zimmermann 1994, 22-27.
9 Maqiilat Aliskandar ft l-fiiss, 176,68-69; 178,90; 190,185; and 192,194 (Quaest. 3.3, 83,37;

84,28; 86,22.24).
10 GA 1.8, 718b7 =Fi kawn al-fiayawiin, 11,16; 4.1, 763b22 = 134,5; 4.4, 773al4 = 159,6; 4.7,

= = =
776a4 166,19; and 5.10, 777b27 171,6; HA 6.3, 561a6 Tibii' al-fiayawiin, 249,2.
104 Part /: The Ammonian Synthesis

I:Iunayn, which the Fihrist mentions in its compressed and cryptic description of
the Arabic reception of the De Anima. As Frank has convincingly argued, the
version edited by Badawi cannot be, as Badawi thought, IsJ:iaq's later, revised
edition. 11
Above and beyond the arguments offered by Frank and others, it is clear that
IsJ:iaq used an entirely different vocabulary to render terms such as entelekheia
and teleiotes - particularly the terms kamiil and istikmiil - than the translator of
the BadawI text had. For example, as Table 3 makes clear, BadawI's Fi n-nafs
either renders entelekheia (mainly in the dative case, that is, in its adverbial
sense) as fief or simply transliterates it as anraliishiyii (and once as
anraliikhiyii), occasionally glossing it with tamiim. The rendering of
entelekheia as fief is just as Ustat had done in the first eight books of his
Metaphysics translation, and the transliteration of entelekheia as anraliishiyii
is just as al-l:limsi had done in his compilation of the Uthu!Ujiyii. The BadawI-
text' s translator's use of tamiim to gloss anraliishiyii in Ll not only echoes the
Greek commentators' gloss of entelekheia with teleiotes, it is also a sign of
the second period of translation, as manifest in the use of tamiim in L7, the
Uthiilujiya's discussion of the soul's definition. A similar use of tamiim to
render entelekheia is found in the translation of a short, late-antique paraphrase
of the De Anima and in a commentary on that paraphrase, both of which have
recently been re-edited by Arnzen, who ascribes them to Ibn al-Bitrlq (fl ca. 835)
and dates them to around 820-845, that is, to the second stage of the translation
movement. 12
In contrast to the Badawi text, it is obvious from Tables 4 and 5 that in his
Physics translation Isl:iaq viewed the root k-m-l, introduced during the second
stage alongside t-m-m in the aforementioned hendiadys, as the best Arabic term
for entelekheia, since that way it could be more easily distinguished from telos,
which IsJ:iaq now rendered with ghiiya. In other words, IsJ:iaq's purpose in using
k-m-l and ghiiya, respectively, for entelekheia and telos, while reserving t-m-m
for teleiotes, was to prevent the Greekless reader of the Arabic Physics from
conflating two distinct Aristotelian concepts, entelekheia and telos, each of
which - as I just showed - had commonly been rendered as tamiim by his
predecessors. 13

11 Frank 1958-59.
12 Anonymous, Paraphrase of the De Anima, 183,18-22 (= [Ps.-]lsl:iliq b. I:Iunayn, K. an-nafs,
129, 13-16); Anonymous, Commentary on the De Anima, 213,15-217,16 (= [Ps.-]Isl:iii.q b. J:lunayn,
K. an-nafs, 139,20-1140,23) and 219,7-16 (=[Ps.-]lsl:iii.q b. J:lunayn, K. an-nafs, 141,3-10).
13 -More specifically, as Tables 4 and 5 make clear, Ist:iaq's Physics renders ente/ekheia as

istikma/ in Books 1, 2 and 8, and as kama/ elsewhere in the Physics, most notably in the discussion
of motion in Phys. 3.1 and 3.2; by contrast telos is rendered as ghaya in Books I and 2; as tamam
and akhir in Books 3 and 4; and as akhir, intiha', ghaya and inqii;Jii, in Books 5-8. The Arabic
paraphrases of Philoponus' commentary on the Physics seem to show that lsl:iliq wanted to reserve
kama/ for entelekheia, tamam for teleiotes, and ghaya for telos. Philoponus' views are
paraphrased in Arabic and attributed to "l:i" (which the editor, Badawi, takes to be shorthand for
Yal:iyii. al-Nal:iwi, i.e., John Philoponus) at at-Tabi<a, 171,8-13: "He (i.e., Aristotle) means by kamiil
 5. Greek into Arabic 105

The upshot is that given IsJ.iaq's preference for istikmiil, kamiil, and ghiiya in
his Physics translation; given the use of istikmiil for entelekheia and kamiil for
teleiotes in the Arabic version of Themistius' Paraphrase of the De Anima, as
can be seen in Table 3; and finally, given the appearance of istikmiil and kamiil
in the lemmata contained in Avicenna's Marginal Notes on Aristotle's De
Anima, which can also be seen in Table 3; the evidence supports either a strong
hypothesis or a weak hypothesis regarding Isl_laq's revised version of the De
Anima. The strong hypothesis is that IsJ.iaq's final, revised version of the De
Anima is nothing other than the Arabic version of Themistius' Paraphrase. The
weak hypothesis is that IsJ.iaq's final, revised version is not the Arabic version of
Themistius' Paraphrase, but the one which exists in the majority of Avicenna's
lemmata, the similarity in terminology between it and the Arabic Themistius
being nothing other than a manifestation of the new way entelekheia was being
translated during the fourth and fifth stages.
Other pieces of evidence suggest that the progression from tamiim to istikmiil
in the translation of entelekheia may have had a number of intermediate stages,
or, perhaps better, may have had an intermediate period of uncertainty, lasting
from the end of the second stage to the beginning of the fourth stage, from about
850 to 900, during which a number of possible alternative translations of
entelekheia contended with each other, with istikmiil eventually emerging as the
victor. The great translator ijunayn ibn lsJ.iaq, responsible above all for the many
translations of Galen's works into Arabic, provides some circumstantial evidence
of this shift: in his own work, the Masii'il, ijunayn used the hendiadys tiimm
kiimil as well as the new istikmiil. 14
The move from t-m-m to k-m-l can also be detected in the Essay on the
Definitions and Descriptions of Things, attributed to al-Kindl. where two
versions of the Aristotelian definition of the soul are offered, with the term
tamiimiyya used to render entelekheia in the first and istikmiil in the second. 15
Assuming that al-Kindl was in fact the author of that work, one translator in al-
Kindi's circle may have suggested, with teleiotes in mind, that al-Kindi use

in this sense the passage from what is potential to what is actual, not the state of completeness (la t-
tamam) .... Kamal is of two types: initial and final. The final consists of the termination of what is
potential [once it has arrived] at the [state of being] actual. The first consists in the procession
towards (at-ta!arruq ila) the final perfection, potentiality being preserved with it; this [first kamal]
is motion"; and at 187,2-3: "Because perfection (al-kamal) applies to the end (al-ghaya) as well as
to the passage to it (i.e., to the end, as-sulilk ilayhii)". Alternatively, the Fihrist claims that it was
Qus\ii, not IsJ:iiiq, who translated the first four books of Philoponus' Physics commentary, so the
quotations above could be Qus\ii's own work, or perhaps even YaJ:iya b. 'Adi's revision of Qus!ii' s
translation of Philoponus. In any case, the quotations above do seem to show an early-tenth-
century desire to avoid conflating the ideas of entelekheia, telos and teleiotes.
14 mithla l-afa 'ili t-tammati l-kamilati ... al-fi'lu l-mustakmilu: I:Iunayn ibn IsJ:iaq, al-Masii'il ft t-

!ibb li-1-muta'allimln, 65,1-2. cf. also MuJ:iammad b. Zakariya ar-Razi's (d. ca. 930) pairing of
tamm and kamil in his description of how the fetus leaves the womb when it is tiimm af·filra and
kiimil al-ma'na: Maqiilafi-mii ba'da t-tabl'a, 127,5-6.
15 (Ps.-?)al-Kindi, R. ft }iudild al-~shyii' wa-rusilmihii, 113,7-8. On al-Kindi's debt to the

Uthilliljiyii, see now Adamson 2000.

106 Part I: The Ammonian Synthesis

tamamiyya to define the soul, the idea being that the adjective tflmm could be
derived from the abstract noun tamamiyya just as easily as it could be derived
from tamam, in the same way that teleion could be derived from teleiotes as
well as from telos. Another translator in al-Kindi's circle may then have
suggested that he use the new term istikmal as an alternative.
The shift from t-m-m to k-m-l is also apparent in Qusta b. LUqa's translation
of Ps.-Plutarch's Opinions About Natural Philosophy Which Philosophers
Agree On, where change is defined in the Aristotelian way with entelekheia
again appearing as tamflmiyya; yet in his rendering of the Aristotelian definition
of the soul, Qusta used kamal. 16 Qusta's use of kamal to render entelekheia in
the Aristotelian definition of the soul is mirrored in a contemporaneous source
from the Jabirean corpus (ca. 900), the Kitab al-l}udud. 11 Qusta's translation
also seems to have been the source for Sacdiya (d. 942) and al-Masclidi (d. ca.
953), who list the Aristotelian definition of the soul and use kamal. 18 A final
piece of evidence supports the hypothesis that Isl_laq's istikmal supplanted
Qusta's kamal in the first quarter of the tenth century as the standard translation
of entelekheia, as a result of its presence in what appears to be the final, revised
translation of the De Anima by Isl_laq b. ijunayn: in his Essay in Response to
Questions Asked, al-Farabi defines the soul in the Aristotelian manner but uses
the term istikmal, not kamal. 19 The related uncertainty over whether tamam or
ghaya was a better translation of telos (and, correspondingly, whether al- cillatu
t-tamamiyyatulas-sababu t-tamamiyyu or al- cillatu l-gha'iyyatulas-sababu l-
gha'iyyu was a better translation of to aition telikon, "final cause"), can be seen
as early as lbn al-Bitrlq's second-stage translation of the late-antique De Anima
paraphrase mentioned earlier. 20
The result of all this is that in his Physics translation, and in what appears to
be his later De Anima translation, Isl_laq used kamal or istikmal to render
entelekheia, and ghiiya to render telos, in an attempt to avoid his predecessors'
overuse of tamiim. Despite Isl_laq's best efforts, however, a number of the Greek
philosophical texts which were to have the greatest impact on Avicenna's

16 motion is a tamiimiyya: (Ps.-)Plutarch, Ff l-iirii' af-fabi'iyya allatf tart;fii bihii l-faliisifa, 120,6-7
(= Daiber, Aetius Arabus, 21,4); the soul is a kamiil: (Ps.-)Plutarch, Ff l-iirii', 157,2-4 (= Daiber,
Aetius Arabus, 50,12-13).
17 Jabir, K. al-1,iudud, 113,3-7.
18 Sa'diya, K. al-iimiiniit wa-l-i'tiqiidiit, 189,11 ; al-Mas'iidi, Muruj adh-dhahab wa-ma'iidin al-

jawhar, 52, 362, 10.

19 l,iadda Arisfu n-nafsafa-qiila innahii istikmiilun awwalu li-jismin fabi'iyyin aliyyin dhr hayiitin

bi-l-quwwati: al-Fiirabi, R.fijawiib maso'il su'ila 'anhii, 99,1-2.

20 At Anonymous, Paraphrase of the De Anima, 185,7-8 (= [Ps.-]IsJ:iaq b. Hunayn, K. an-nafs.

129,23-24), Aristotle investigates what al-'illatu l-ghii'iyya of nourishment is. Yet at Anonymous,
Paraphrase of the De Anima, 193,3-4 (= [Ps.-]lsJ:iaq b. l:lunayn, K. an-nafs, 132,12-13), the final
cause of locomotion is listed as [al-'illatu t-]tamomiyya. In the Commentary (ad 415b8), Ibn al-
Bi\riq uses 'illatun tamiimiyyatun and 'illatun ghii'iyyatun within three lines of each other:
Anonymous, Commentary on the De Anima 225,1-10 (= [Ps.-]IsJ:iaq b. l:lunayn, K. an-nafs, 142,12-
19).
 5. Greek into Arabic 107

philosophy - the Arabic version of Aristotle's Metaphysics above all - fit

squarely in the second-stage tendency to conflate the concepts of actuality and
final causation by using t-m-m to render Aristotle's entelekheia and telos as
well as the commentators' telikon. What is more, as a result of the third- and
fourth-stage attempts to restrict the use of t-m-m to the instances of teleiotes in
the commentaries on the Physics and the De Anima, the Neoplatonic concept of
teleiotes as final cause came into circulation and appeared to be consistent with
the earlier translators' use of t-m-m to render both entelekheia and telos. And
given that Avicenna's main discussions of causality in general, and final
causality in particular, are contained in his metaphysical rather than natural-
philosophical works - unlike Aristotle's canonical discussions of causation,
which are more prominent in the Physics than in the Metaphysics - the Arabic
Metaphysics' use of t-m-m to render both entelekheia and telos provided even
greater momentum pushing Avicenna in the direction of seeing actuality or
perfection as being related to its effect as a final cause.
Just as important is the fact that the early translators' uses of t-m-m had
infiltrated many of the discussions of causation and cosmology of Avicenna's
predecessors al-K.indi, the Ikhwan a~-~ala 0 and al-Farabi, the last of whom played
an important role in shaping Avicenna's thought on this subject. In discussions
of causation, for example, al-K.indi lists the final cause as al- cillatu t-
tamiimiyyatu in his On First Philosophy, a practice echoed in the Epistles of
the Ikhwan ~-~ala and in the Jabirean text The Book of Research. 21 In
discussions of cosmology, al-Kindi describes God in the title of one work as the
perfect agent (al1iicilu t-tiimmu), though in that same work he uses ghiiya rather
than tamiim to assert that God is the end of every other cause (huwa ghayatu
kulli cillatin); and in his On First Philosophy al-Kindi argues that the eternal -
which by his reckoning referred only to God - is necessarily perfect lfa-l-
azaliyyu tammun i<j.firaran), an echo of which can again be found in the
Jabirean text The Book of Research. 22 Al-K.indi also uses naq$ and tamiim to
label the contraries between which an istif:iiila (alloiosis) or taghayyur
(metabole) occurs, in contrast to the more Peripatetic quwwa and Ji cl, a sign
that thinkers such as al-K.indi were beginning to investigate the nature of the
parallelism between the Aristotelian potentiality/actuality distinction and the
deficiency/perfection distinction, an Arabic variant of the disposition/perfection
distinction introduced by Alexander and promoted by the Neoplatonists. 23

21 al-Kindi, F7 ljalsafat al-ulii, 11,8.11 (the final cause is also listed at 11,4 as mutammima,

which seems to be a sign of the second-stage "Alexander" translations); lkhwan a~-~afa', Rasii' il,
Vol. 2, 79, 15-17; 89,9-10; 115,23-24; 155,6-7; 3, 358,18 and 359,4-6 (the Ikhwan also seem to use
tiimm and kiimil interchangeably at Vol. 3, 371,2 and 379,6.11-12); Jabir, K. al-ba/:ith, 525,10.
22 al-Kindl, Risiilat al-Kindl ft 1-facil al-/:iaqq al-awwal at-tiimm wa-L-facil an-niiqie alladhl huwa

bi-L-majiiz, 169,6-7; and F7 l-falsafat al-ulii, 29,4. Jabir, K. al-ba/:ith, 524,16-525,7 (admittedly, this
passage is a bit murky).
23 al-Kindl, K.ft ljalsafat al-ulii, 27,24-29,4; and Risiilat al-Kindl ilii A/:imad b. al-Mu<taeim ft 1-

ibiina <an sujud al-jinn al-aqeii wa-ta<arihi Ii-I/ah <azza wa-jalla, 179,11-23.
108 Part /: The Ammonian Synthesis

Although al-Farabi uses Qus~a's kamiil and IsJ:iaq's istikmiil and ghiiya, and is
clearly the product of the more purely Aristotelian fourth and fifth stages of the
translation movement, the influence of the more Neoplatonic second-stage
translations of the al-Kindi circle is also obvious in his writings, particularly in
his use of tiimm and niiqi~ to describe two contrary types of existence. As far as
his influence on Avicenna is concerned, the most important instance of al-
Farabl' s use of tiimm in this metaphysical way occurs in al-Fiirabl's On the
Philosopher's (i.e., Aristotle's) Aims [lit., Scopes] in Each Book of the
Treatise Designated by Letters (i.e., the Metaphysics), the little book which
- according to Avicenna's autobiography - helped Avicenna move to a higher
stage of understanding what Aristotle's Metaphysics was all about. In his book
al-Fiirabi says that metaphysics, or universal science, examines things which are
common to all beings, such as existence (al-wujud) and oneness (al-waJ:zda), as
well as the different types (anwac) and concomitant properties (lawaJ:ziq) of
existence and oneness, such as priority and posteriority (at-taqaddum wa-t-
ta'akhkhur), potentiality and actuality (al-quwwa wa-l-ficl), and the perfect and
the deficient (wa-t-tamm wa-n-naqi~). 24
It is slightly confusing that al-Farabi should use the substantive adjectives
tamm and naqi~, and not tamam and naq~ (or naq~an), the nouns most parallel
to taqaddum and ta'akhkhur, quwwa and ficl; perhaps those last four
substantives were understood as lawii}:ziq, "concomitant properties" of existence,
while the substantive adjectives at-tamm and an-naqi~ were meant to be
understood as anwiic, "types" or "species" of existent. An awareness of the
problem of delineating the precise nature of the parallelism between the actuality
(al-ficl)/potentiality (al-quwwa) distinction on the one hand, and the perfect (at-
tamm)/deficient (an-naqi~) distinction on the other, is more apparent in al-Fiirabi
than it was in al-Kindl. In a sense, al-Fiirabl's efforts to reconcile the two pairs
of concepts, the former Aristotelian and the latter Neoplatonic, are very much in
the tradition of the Ammonian synthesis. 25
In view of the fact that actuality (al-ficl) is tied to form (a~-~ura) (since a
thing is in actuality only when it possesses its form), while potentiality (al-
quwwa) is tied to matter (al-miidda) (since a thing is in potentiality when it
only possesses its matter), al-Farabi asserts that being in actuality is more
perfect in respect of existence (akmalu wujudan) than being in potentiality. 26
Al-Farabi appeals to the criterion of causation to defend this assertion: the matter

24 al-Farabi, MaqiUa ... fi aghrcu;i al-~akim ft kulli maqalatin min al-kitab al-mawsum bi-1-~uruf,
35,8-11.
25 That al-Farabi saw himself as part of the tradition of the Ammonian synthesis is evident

throughout his treatise On the Harmony between the Opinions of the Two Philosophers, Plato the
Divine and Aristotle, but particularly in his claim that Ammonius' arguments about God's being
both an efficient and final cause were so well known that they did not require citation: K. al-jam'
bayna ra'yay al-~akimayni Af/ii!iin al-iliihi wa-Aris!ii!i'llis, 24,24-25, 1.
26 al-Farabi, K. arii' ah/ al-madinat al-]afiila, 12, 64,9-13; K. as-siyiisat al-madaniyyat al-

mulaqqab bi-Mabadi' al-mawjudat, 38,10 and 39,1-3; and R.fi al-'aql, 23,1-8.
 5. Greek into Arabic 109

exists only in order that the form which inheres in it may exist, but in no sense
is the form for the sake of the matter. 27 In other words, the actual is a more
perfect type of existent than the potential because a thing which is in actuality is
a cause, while a thing which is in potentiality is not; and because causes are
better than effects, a thing which is in actuality is more perfect and more
excellent than a thing which is in potentiality. Al-Fara.bi's use of causality as
the distinguishing feature of perfection and the perfect is of course a familiar
Neoplatonic move. But al-Farabi not only follows the Neoplatonists' move to
causation here. He pushes the Arnmonian synthesis forward by attempting to
incorporate more fully the Neoplatonists' distinction between first and second
perfections into Aristotle's distinction between first and second actualities.
Al-Fara.bi's distinction between first and second perfections - actually, the
terms he uses are Themistius' , "initial" (awwal) and "ultimate" (akhir) - rests
on the criterion of causation, just as Proclus' distinction had done. And in one
passage in his Siyasa madaniyya al-Farabi clearly echoes Proclus' distinction,
saying that terms signifying perfection (kamiil) and excellence (jatj,ila) are
divided into those that point to some characteristic which the thing in question
has in and of itself, such as "existent" and "one"; while others point to some
characteristic which the thing has in relation to something else other than it,
such as "just" and "generous". 28
However, al-Fii.rabi seems to have Aristotle's distinction between first and
second entelekheiai in mind when in other passages he departs subtly from
Proclus' distinction. Proclus had said that a first perfection referred to a thing's
perfection when the thing was viewed in and of itself, and that a second
perfection referred to a thing' s perfection when the thing was viewed as a cause
of something else. Al-Farabi asserts, however, that a thing's initial perfection
refers to the state something is in when what normally would issue as an effect
from it does not do so. The ultimate perfection, on the other hand, refers to the
state the thing is in when what normally would issue from it as an effect in fact
does so.
The criterion of causation is present in both distinctions, but it is applied
differently in each. For Proclus, something to which the term "first perfection"
is correctly applied may well be a cause - in fact, it is certainly a cause - but it
is not viewed qua cause. Only something to which the term "second perfection"
is correctly applied is viewed qua cause. Al-Farabi, on the other hand, asserts
that something which is in a state of first perfection is not, in fact, a cause at
all: no effect issues from it. By contrast, something which is in a state of second
perfection is a cause: the effects that normally issue from it cannot help but
issue from it. In short, both distinctions rest on the criterion of causation, but
whereas for Proclus the criterion of causation is understcxxl as referring to

27 al-Farabi, K. as-siylisat al-madaniyya, 36,11-12; 38,12; and 39,5.

28 al-Farabi, K. as-siylisat al-madaniyya, 49,12-15.
110 Part /: The Ammonian Synthesis

whether or not a thing is viewed as the cause of some effect, al-Fiirlibi held
that the criterion of causation referred to whether or not a thing is in fact the
cause of some effect. That al-Flirlibi had Aristotle's distinction between first and
second entelekheiai in mind seems clear from the fact that he uses one of
Aristotle's De Anima examples - being asleep and being awake - to illustrate
what he means. That he also had the Physics distinction in mind seems clear
from the fact that in his distinction he uses the terms that Themistius invented -
prote!awwal and hustate!akhfr - in order to distinguish between the state of
change and the state of having changed. 29
What I am getting at is that al-Flirlibi's distinction between first and second
perfections, based on the criterion of not-being-a-cause/being-a-cause, seems
intended to reconcile not only the Aristotelian distinction with the Neoplatonic
distinction, but also to reconcile the Physics distinction with the De Anima
distinction. Al-Fiirlibi's reasoning seems to proceed along the following lines.
The state of being changed into something actual (the first entelekheia of the
Physics commentators) retains some potentiality, and is therefore not the cause
of anything; possessing the capability to perform a function (the De Anima's
first entelekheia) is an inactive state, and is therefore not the cause of anything;
a thing taken in and of itself (Proclus' first teleiotes) is not (viewed as) the
cause of anything. Second perfections are slightly trickier. The state of having
changed into something actual (the second entelekheia of the Physics
commentators) is the final cause of the change; contemplating (the De Anima's
second entelekheia) is the final cause of the acquisition of knowledge; and a
thing's perfection taken in relation to a lower thing (Proclus' second teleiotes)
is the final cause of that lower thing, at least when both cause and effect are
viewed as elements of reversion.
Al-Flirlibi's attempt to merge the three distinctions into one is interesting, and
it certainly earns him a place of honor in the Ammonian synthesis, but what
interests me more is the cosmological implication of his attempted
reconciliation, an implication which throws into stark relief the underlying
difference between his distinction and that of Proclus. According to al-Fiirlibi's
way of distinguishing first and second perfections, the eternal things above the
sphere of the moon will only ever be in a state of second perfection: they are
always actual, and as a result they cannot help but cause their effects. 30 By
contrast, sublunary things down here in the world of generation and destruction
are sometimes in a state of second perfection, when they are actual and thus are
causes; while at other times they are in a state of first perfection, when they are
potential and thus are not causes.

29 al-Fariibi, K. iirii' ahl al-madinat al-fii<Jila, 5, 49,5-8; 23, 105,9-11; 24, 108,7-8; K. as-siyiisat

al-madaniyya, 64,15-66,8.
30 al-Fariibi, R.fi al-'aql, 32,8-33,3.
 5. Greek into Arabic 111

For Proclus, however, it was a thing's first perfection - its active turning
inwards and (thereby) upwards towards self-perfection - which made it a cause of
other things; which created the causality, in other words, of its second perfection.
Proclus' superlunary beings were primarily engaged in seeking their first
perfection, in pursuing their inward and upward epistrophe, which is why, of
course, Proclus referred to it as a first, or primary, perfection. Their second
perfection, consisting in their providence towards things below them - their
causation of them, in fact - is almost a by-product of their more primary act of
reversion. Al-Farabi seems to have this Neoplatonic axiom in mind when he
insists that superlunary things - and particularly God - do not derive any
perfection from their causation of things below them. 31 But if this is the case, in
what sense will superlunary entities be in a state of perfection? Either they will
be in a state of first perfection, which according to al-Farabi means that they will
not in fact be causing their effects, which is impossible; or they will be in a
state of second perfection, in which case they will in fact be causing their effects.
But their being in a state of second perfection derives from their being causes of
effects, which in tum contradicts al-Farabi's assertion that the higher derives no
perfection from its causation of the lower. Al-Farabi is forced to conclude that
God's act of self-substantiation - Proclus' first perfection - is what earns Him
the attribution of perfection; and that His causation of things below Him is
nothing more than a consequence and concomitant (tiibi<atan wa-liifiiqatan) of
His own, self-substantiating perfection. 32
Despite its problems, al-Farabi's distinction had some influence on Avicenna's
cosmology, as I shall show in Chapters 6 and 11. For now, what is important is
al-Fariibi's inference that because contingent, sublunary things are sometimes in
a state of first perfection and other times in a state of second perfection, since
they are sometimes not causing and other times causing, their existence is less
perfect than that of superlunary things, which are always in a state of second
perfection and thus are always causing. Al-Farabi's clear implication is that all
eternal things - spheres, souls, intellects and God - can be characterized as being
in a state of perfection, and particularly second perfection, an implication which,
as discussed immediately above, involved him in a serious contradiction. 33
Leaving that problem aside, however, how is one supposed to solve the further
problem of distinguishing between the various eternal things? For al-Kindi this
was not a problem, since there were no eternal things other than God. Al-Farabi,
on the other hand, faced a dilemma; most of the time he said simply that God is
the most perfect with respect of existence, that the intellects were existentially
less perfect than God, that the souls were existentially less perfect than the

31 al-Farabi, K. ara' ah/ al-madfnat al-fiit;iila, 7, 55, 1-56,8 (and implied at I, 38,12-13).
32 al-Farabi, K. as-siyiisat al-madaniyya, 47,14-50,7. .
33 al-Farabi, Shar}_i li-kitab Arisfufiilfs ft a/-cibara, 193,13-15 (ad 23a21-26); K. ara' ah/ al-

madfnat al-ftit;iila, 2, 40,2-1 O; K. as-siyiisat al-madaniyya, 64, 15-66,8.

112 Part I: The Ammonian Synthesis

intellects, and so on. 34 But if all superlunary things satisfy the criterion of
necessary causation; if all superlunary things cannot help but cause their effects;
and if causation is the distinguishing mark of existential perfection; how then
will God be more perfect than an intellect or a soul or a sphere? God is a cause,
an intellect is a cause, a soul is a cause and a sphere is a cause; things which are
causes are perfect; therefore God, the intellects, the souls and the spheres are
perfect. Again, how is God more perfect than an intellect?
Sometimes al-Farabi says that the answer lies in God's lack of any
potentiality and causedness whatsoever. 35 But if this is the case, and the
intellects, souls and spheres possess some potentiality or causedness, they will
not satisfy al-Farabi's own criterion of being in a state of perfection. Other
times, again appealing to Proclus' concept of first teleiotes, al-Farabi explains
that God's self-causation or self-substantiation (tajawhur) is more perfect than
his causation of others, and hence primary to Him. Al-Farlibi's implication is
that since only in God is self-causation prior to the causation of others, God's
existence is more perfect - in the sense of more self-sufficient - than other
existences.36
Ultimately al-Farabi was stumped by the problem of degrees of perfection.
Should perfection be understood univocally, as referring to a single attribute
which sublunary things, superlunary things and God all possess? Or should
perfection be understood equivocally, as referring to different attributes, one of
which sublunary things possess, another of which superlunary things possess,
and a final one which God only possesses? As I shall show in Chapter 10,
Avicenna grew so frustrated by the problem, and was so dissatisfied with his
attempted solution, that he ended up deciding to side-step perfection entirely.

34 K. iirii' ah/ al-madfnat al-fii{iila, 5, 48,4-5; and 9, 59,3-11; K. as-siyiisat al-madaniyya, 34,1-2;

40,3-8; 42,15; 49,15-50,7; 50,8-11; 51,7-52,4; 53, 1-10.

35 al-Farli.bi, K. iirii' ahl al-madfnat al-fiidila, l, 37,6-38,13.
36 al-Farabi, K. as-siyiisat al-madaniyya: 47,11 -49,11.
 6. Avicenna on Perfection and the Soul
The Issue of Separability

In this and subsequent chapters I shall show how several of the most important
elements of Avicenna's psychology and cosmology can be better understood
when they are seen to be shaped by the doctrinal presuppositions of the
Ammonian synthesis as well as by the word choices of the early Greco-Arabic
translators. In other words, the Ammonian synthesis provided the philosophical
momentum, and the early Greco-Arabic translations provided the lexical
momentum, which together propelled Avicenna towards holding that the ideas of
actuality and perfection on the one hand, and of the final cause on the other, were
at least mutually implicative, if not identical.
Like earlier philosophers of the Ammonian synthesis, Avicenna struggled to
find the best way to reconcile Aristotle's use of perfection in his definition of the
soul with the Neoplatonists' broader understanding of perfection as signifying
superiority in the hierarchy of being, and hence causation. And like earlier
philosophers of the Ammonian synthesis, Avicenna aggressively interpreted a
number of other Aristotelian concepts and distinctions in order to use them as
tools to fashion a new synthesis. But unlike late-antique Greek thinkers such as
Ammonius and Philoponus, Avicenna was helped along by the early Greco-
Arabic translations, and particularly by the translations of the Metaphysics and
the Uthulujiyii, which, by using the same Arabic root - t-m-m - to render the
Greek entelekheia, teleiotes and telos, did part of the job for him.
One of the lessons which I hope will be drawn from the preceding chapters is
that the passage of Neoplatonic ideas from Greek into Arabic seems to have been
diffuse rather than concentrated, sieve-like rather than funnel-like. What I mean is
that important Neoplatonic ideas found their way to thinkers such as Avicenna
on board a large number of different vehicles, particularly the Aristotle-
commentaries by philosophers such as Syrianus, Ammonius, Philoponus,
Simplicius and Olympiodorus (the Arabic translations of which - though only a
few now appear to be extant - are attested to in the catalogues of lbn an-Nadim,
lbn al-Qiffi and lbn Abi U~aybi c a); and that these Neoplatonic ideas were not all
stuffed into the two or three mislabeled vehicles which we do have at hand: the
Plotinian Theology of Aristotle, the Prodan Liber de causis and the works of
pseudo-Alexander. The fact that Avicenna received important elements of
Aristotelianism already Neoplatonized in the commentaries, is particularly
important given how uncomfortable he seems to have been with many of the
ideas expressed in the UthUlujiyii. 1 This will become most apparent later on in

1 In fact, there is some evidence, based on a sentence in a letter to one of his disciples, Kiyli, in

which Avicenna describes the scope of his K. al-i~iif (wa-awi;faJ:itu sharJ:ia l-mawiii;fici l-mushkilati
ft lju$ft$i ilii iikhiri Uthulftjiyii 'alii miifi Uthiiliijiyii mina 1-ma(ani: R. ilii Kiyii, 121,19-20), that
114 Part I: The Ammonian Synthesis

this chapter, where I shall show that Avicenna's position on the soul and its
relationship with the body owes much more to the Aristotle-commentator
Philoponus than it does to Plotinus or Proclus. In Chapter lO, Avicenna will
also be shown to owe a great deal to the Aristotle-commentators Syrianus,
Ammonius and Asclepius, particularly with regard to the issue of God's being
both an efficient and final cause. Where Avicenna does appear to borrow directly
from the Arabic Plotinus and Proclus is in his use of the phrase "above
perfection" (jawqa t-tamiim) to describe God.

How does Avicenna use perfection to define the soul? Somewhat surprisingly,
Avicenna does not use istikmiil, the term cited in the lemmata of his Marginal
Notes on Aristotle's De Anima - and which, as I discussed above, is probably
the term lsJ:iaq b. l:funayn employed to render entelekheia in his revised
translation of the De Anima - opting instead for Qusta's kamiil. Avicenna uses
kamiil to define the soul in two ways, the first as part of the standard
Aristotelian definition, the second as part of his own modified definition, in
which the various types of soul or faculties of soul are defined in a series. The
standard Aristotelian definition appears in the On the Soul (Fr n-nafs) section of
the Natural Philosophy (!abi'iyyiit) part of his great summa, the Book of
Healing (Kitiib ash-shifii'):

LIS
Ibn Sinii, Kitiib ash-shifii'!Tabf'iyyiit (6): Ff n-nafs l.l, 12,6-8
Avicenna, Liber de anima seu sextus de naturalibus I-III, 29,61-63
So the soul which we are defining is a first perfection of a natural
instrumental body [which the soul uses] to perform the activities
of living [fa-n-nafsu llatf naf.zudduhii hiya kamiilun awwalu li-
jismin rabi'iyyin iiliyyin lahu an yaf"ala af"iila l-hayiiti = ldeo
anima quam invenimus in animali et in vegetabili est perfectio
prima corporis naturalis instrumentalis habentis opera vitae]. 2

This definition is repeated almost exactly in Avicenna's Investigation of the

Soul's Faculties (Mabf:iath 'ani l-quwii n-nafsiiniyya), in his Essay on the
Soul, its Survival and its Reversion (Risiila fi n-nafs wa-baqii' ihii wa-

Avicenna may have realized that the Uthiiliijiya was wrongly attributed to Aristotle. Gutas 1988,
63-64 (following Kraus 1942, 272-73, n.3), translates 'ala mii ft Uthiiliijiyii mina 1-ma!'ani as
"despite the fact that the Theologia is somewhat suspect'', thus giving Avicenna credit for sensing
its false provenance. Zimmermann 1986, 184, translates the sentence "for all one may find to
object to in the Uthiiliijiyii", maintaining that the sentence only refers to what he calls the "chaotic
scattiness" of the work, not its hazy provenance. In any case it seems absolutely clear that
Avicenna was uncomfortable philosophically, if not philologically, with parts of the Uthiiliijiyii, so
the question of its influence on him must be answered on a case-by-case basis.
2 On the reception of Avicenna's definition of the soul by medieval Latin thinkers see now

Hasse 2000, 236. Apart from my detailed discussion in Chapter 8 of the Latin translation of L36,
Latinists can consult van Riet's notes for explanations of the Latin translation's divergences from
the Arabic in this and other passages.
 6. Avicenna on Perfection and the Soul 115

ma'adihli), in his Book of Definitions (Kitab al-budud), and in his Discussions

(al-Mubiibathlit); and in an abbreviated and less formal way, in the
Demonstration (al-Burhlin) section of the Logic (Manfiq) part of his Shifa'. 3
Avicenna's slightly modified version appears in a later chapter of the Fi n-nafs
of the Shi/a', in the Risa/a ft n-nafs wa-baqa'ihli wa-ma'adiha, and in the
Salvation (Kitab an-najat):

Ll9
lbn Sina, Ki tab an-najat, 258,2-10
The soul is, as it were, a single genus divisible into three parts.
The first part is the vegetative soul; it is the first perfection of a
natural body possessing organs in so far as it reproduces, grows,
and nourishes itself .... The second part is the animal soul; it is the
first perfection of a natural body possessing organs in so far as it
perceives particular things and moves intentionally. The third part
is the human soul; it is the first perfection of a natural body
possessing organs in so far as it performs activities that come
about as a result of deliberate choice and rational inference, and in
so far as it perceives universal things. 4

Two questions immediately arise. First, why did Avicenna use kamal instead of
istikmiil, which was the term contained in the Arabic text of Aristotle which he
commented on? Second, how does the first way of defining the soul, as
articulated in LI 8, relate to the second, as articulated in LI 9?
To answer the first question, two possibilities present themselves. The first is
that Avicenna wanted to use the same term - kamal - which appeared in lsl,laq's
translation of the definition of change in Physics 3.1. But given that it is
unlikely that Avicenna knew that the same Greek term - entelekheia - was used
in the Physics 3.1 definition of change and in the De Anima 2.1 definition of
the soul, there is no reason why Avicenna should want to give himself the same
headache - using the same term to define change and the soul - which had so
plagued the Greek commentators Alexander and Thernistius. It seems instead that
in his uses of kamiil and istikmal Avicenna was trying to make a distinction
between the type of perfection - istikmal - which resulted from non-kinetic
transitions such as sensation and intellection (that is, transitions that involve the
presence or absence of a relation); and the type of perfection - kamal - which
either resulted from kinetic changes and alterations (the soul, for example) or

3 lbn Sina, MabJ:iath 'ani l-quwli n-nafsliniyya, 2, 153,13; R. ft n-nafs wa-baqa'iha wa-ma'iidiha,

I, 56,4; K. al-J:iudiid, 14,2-3; al-MubaJ:iathat, 129,11 (= MubaJ:iatha 5, #343); K. ash-shifli'!Man{iq

(5): al-Burhan 2.9, 181,11-12.
4 lbn Sina, K. ash-shifa'fTabiciyylit (6): Fi n-nafs 1.5, 39,13-40,7; R. Ji n-nafs wa-baqli'ihli wa-

ma 'iidiha 2, 57,3-9.
116 Part /: The Ammonian Synthesis

which described the kinetic change or alteration itself (change, for example). 5
And indeed, Avicenna also uses kamal, not istikmal, to define motion:

L20
lbn Sina, Kitiib ash-shifii'/Tablciyyiit ( 1 ): as-Samiic a!-!ablcl
2.1, 83,5
So motion is a first perfection of what is potential in so far as it is
what is potential ffa-l-J:iarakatu hiya kamiilun awwalu li-mii huwa
bi-l-quwwati min jihati mii huwa bi-l-quwwati]. 6

In trying to reserve istikmal for non-kinetic transitions such as sensation and

intellection Avicenna follows the lead of the Arabic version of Alexander's De
lntellectu, which employs istikmiil in this way. 7
The second question ("How does the first way of defining the soul, as
articulated in L18, relate to the second, as articulated in Ll9?") will take a little
longer to answer. I believe that Avicenna's use of perfection in both his standard,
Aristotelian definition of the soul as well as in one that appealed to the different
capacities or faculties represented by each type of soul - vegetative, animal or
human - is a sign of the tectonic shift that had taken place as a result of the
new, Neoplatonic conception of perfection. What I mean is that Avicenna is
pulled in two directions: by Aristotle's use of perfection to define the soul, and
by the Neoplatonists' use of perfection to describe how beings which are higher
in the cosmic hierarchy relate to beings which are lower in the manner that
causes relate to effects.
Avicenna wants to use the more general Neoplatonic concept of perfection to
describe how the higher faculties of the soul relate as causes to the lower ones,
while at the same time retaining the more specific Aristotelian use of perfection
to describe the relation of the soul as a whole to the body. The tension in
Avicenna's work between these two competing tendencies becomes more
apparent in two further clusters of discussions: one on the problem of how form
and perfection are related, the other on how first and second perfections are

5 Ibn Sina, K. ash-shifa'fTabfciyyat (6): FT n·nafs 2.2, 66,15-17. In the Metaphysics (/lahiyyat)
of the Shifa', Avicenna allows that istikmal can cover changes as well as non-kinetic transitions
when he takes Aristotle to task for not having made clear enough that an istikmiil can either be
instantaneous (dujatan), as is the case with non-kinetic transitions such as "not-knowing to
knowing", or can proceed over time (salikan), as is the case with the transition from "boy to man":
K. ash-shifa'!llahiyyiit(2) 8.1, 332,15-18 and 333,1-11.
6 This formulation is repeated, with minor variations, at K. an-najat 2: Tabfciyyat, 170,7-14, and

K. a/-hidaya 2.1, 138,1. On Avicenna's definition of motion see Haschmi 1956; Hasnawi 1998; and
now Hasnawi 2001. On the medieval Latin translations of Avicenna's (and Aristotle' s) definitions
of motion see the classic treatment of Maier 1958a (Avicenna's theory is treated at 12-20). The
"Mari nus" Latin translation of this passage reads unde motus est perfectio prima eius quod est in
potentia secundum quod est in potentia, according to Jules Janssens, who is currently preparing an
edition of that text. Maier appears to follow a later translation.
7 Alexander, R. ft aJ-<aql < a/a ra'y AristiiJiilis, 33,18-20; 38,2-3 (where kamila is used for the
exercise of the capability to walk, and istakmala is used for the exercise of the capability to think);
and 39,15-17.
 6. Avicenna on Perfection and the Soul 117

distinguished. Once those problems have been dealt with, the question of how
the soul relates to the body as cause to effect can be treated properly; and once we
answer that question, we can better understand Avicenna's position on the thorny
issue of the separability or inseparability of the soul.
In some respects Avicenna follows earlier philosophers of the Ammonian
synthesis in his attempts to delineate the precise relationship between form and
perfection. Philoponus, as I showed above, held that while entelekheia and
eidos may be identical in the sense that both refer to form as shape as well as
form as substance, teleiotes refers only to the type of form - form as substance,
not form as shape - which could be correctly applied to the soul. Avicenna, by
contrast, understands kamiil ("perfection") to be a far more broadly applicable
term than ~ura, "form". 8 For example, in his Marginal Notes on Aristotle's De
Anima, Avicenna makes a clear distinction between form and perfection:

L21
Ibn Sina, at-Taclfqat <ala J.zawashl Kitab an-nafs li-Arisfu
(ad DA 2.4, 4l5b2), 91,3-16
The Easterners say that it [i.e., the soul's being the form of the
body] is not distinct from [its being] a "perfection'', but is rather
identical to [its being] a perfection [al-istikmal]; and then they
leave it at that ....
It seems that he [Aristotle] means by "perfection" [bi-l-
istikmali] something other that what he means by "form'', even
though it [perfection] is a logical consequence of it [what he
means by form], for a thing is called a form only relative to its
substrate, while it is called a perfection relative to all that is
perfected by it [ila kulli ma yukmalu bihi]. The substrate and the
matter are sometimes perfected by it [the perfection] though not in
the sense that it [the perfection] is a part of it [the substrate]; and
the composite which is given the nature of a species is sometimes
perfected by it [the perfection]. It is known that it [the perfection]
is the part from which its [the composite's] actions issue, such
that the perfection has an influence on the form, because the
perfection is the perfection of a species while the form is the form
of a matter. The matter is what is formed - not perfected - by form

Because "form" implies "being impressed upon" [al-in!iba c] he

[Aristotle] casts it aside for another term which applies to what is
separable as well as to what is inseparable. Even though his
present aim with "form" is what covers both of them [i.e., that
which covers what is separable and what is inseparable], he drops
it in favor of a new term which does not have any discernible
implication [i.e., of having to be impressed on matter].

8 Avicenna also maintains that when we predicate the term "perfection" of something, we are

not making any claim about whether that thing is a substance or not; on this see now Sebti 1999.
118 Part /: The Ammonian Synthesis

I should explain that Avicenna's method in his Kitii.b al-in~af, a kind of

problem-by-problem commentary on the entire Arabic Aristotelian corpus, was
for him first to articulate the positions of "Easterners" (i.e., Khurasanis and
Transoxanians) and "Westerners" (i.e., Baghdadis) on issues of interpretation that
arose as a result of ambiguities in Aristotle's works, and then to "judge fairly"
(in~af) between them. 9 His Marginal Notes on Aristotle's De Anima probably
represents a working draft of the section of the Kitab al-in~af devoted to
problems arising from Aristotle's De Anima, and position (2) in the passage
above is either the Westerners' position - a position which, as we shall see,
Avicenna sides with - or a draft of Avicenna's "fair judgment", the Westerners'
position having been lost or unreported in what is available to us now.
The philosophical reasons Avicenna gives in his Marginal Notes on
Aristotle's D,e Anima for rejecting the Easterners' simple identification of form
and perfection are similar to those offered in L7, the Uthulujiya passage, in
which the term used for perfection is tamam rather than kamal or istikmal.
These reasons are reiterated in Avicenna' s Fi n-nafs:

L22
Ibn Sina, Kitiib ash-shifii 0fl'abl°iyyiit (6 ): Fl n-nafs
1.1, 6,13-7,2 and 7,8-10
Avicenna, Uber de anima seu sextus de naturalibus I-Ill,
19,27-20,37 and 20,44-21,3
Now while every form is a perfeclion, not every perfection is a
form [thumma kullu ~uratin kamiilun wa-laysa kullu kamiilin
~uratan = Deinde dicimus quod omnis forma est perfectio, sed non
omnis perfectio forma] . For the king is lhe perfection of the state
and the captain is the perfection of the ship [fa-inna l-malika
kamiilu l-madlnati wa-r-rubbiina kamiilu s-safinati = magister enim
fabricandi perfectio est civitatis et carpentarius lignorum perfectio
est navis], yet neither is the form of the state or the ship. The kind
of perfection which is itself separable [from that whose perfection
it is] will not really be a form possessed by matter and contained
in malter [fa-mii kiina mina l-kamiili mufiiraqata dh-dhiiti lam yakun
bi-l-/:laqlqati ~uratan li-1-miiddati wa-fi l-miiddati = Et cuiuscumque
perfectionis est essentia per se separata, ipsa certe non est forma
materiae nee in materia]. The form that is contained in matter is a
form which is impressed upon il and which subsists through il,
except on the rare occasion when it is used technically and the
perfection of a species is spoken of as the form of a species -
though in facl it has been settled that technical usage dictates that
something be a form relative to matter; an end and perfection
relative to the whole [wa-bi-l-qiyiisi ilii l-jumlati ghiiyatan wa-
kamiilan = comparatione vero totius collectionis sit finis et
perfectio ]; and an efficient principle and moving power relative to
setting-in-motion ....

9 See Gutas 1988, 130-140.

 6. Avicenna on Perfection and the Soul 119
Therefore it is clear that when, in the course of defining the
soul, we say that it is a perfection [kamal = perfectio ], it will be
the most indicative of its meaning. What is more, it [i.e.,
"perfection"] encompasses all the types of souls in all their
aspects, with the soul that is separable from matter not being an
exception to it [i.e., to being a perfection].

What precisely Avicenna means when he links together the ideas of perfection
and species is not immediately clear. Elsewhere in the Fi n-nafs of his Shi/a'
and in a long passage from his Risiila ft n-nafs Avicenna helps us a bit by
explaining that the soul is a potentiality (quwwa) relative to its activity; a form
(~ilra) relative to the matter it inheres in, if it happens to be an enmattered soul;
and a perfection (kamiil) relative to the species "animal" or "human". But,
Avicenna argues, perfection is the most complete (atamm) and the most
generally applicable (a<amm) of the three terms.
That perfection is more complete and more broadly applicable than
potentiality needs no argument, Avicenna says. Presumably this is simply
because the soul can be described not only as being in a state of first
entelekheia (i.e., as a potentiality or capability), but also as being in a state of
second entelekheia (i.e., as an activity or an exercise of that capability).
Showing that perfection is more complete and more broadly applicable than
form is trickier, however. Perfection is more complete than form because,
Avicenna argues, perfection is more closely related to the nature (fablca) of the
thing in question than form is. The nature of "human", for example, refers more
to a human in actuality - a human that is both matter and form, that is - than it
does to a human in potentiality, which is just the matter. Since perfection is
related to the matter/form composite, while form is related to matter alone,
perfection will be related more closely to the nature of the thing than form will.
As far as the breadth of its applicability is concerned, Avicenna's motivation
is clearer: to find a way for the term "perfection" to be understood in a way
which is broad enough to cover separable entities - captains of ships, generals of
armies, kings of countries, and rational souls - as well as inseparable entities,
such as vegetative and animal souls, and more generally, the forms of natural
things. Since form, strictly speaking, is immanent in the thing in which it
inheres, while perfection can either transcend or be immanent in the thing whose
perfection it is, perfection will be more broadly applicable than form, and
therefore perfection is a better term than form with which to define the soul. As
Avicenna says, the rational soul is not imprinted on matter as form is. 10 In fact,
in his Book of Definitions (Kitiib al-J:iudild), just as in L22, Avicenna
recognizes that "form" is spoken of so loosely that it might appear to be
extensionally identical with perfection, and thus to include in its range of

10 Ibn Sina, R. fl n-nafs wa-baqii'ihii wa-ma (iidihii I, 50,11-53,2 and 54,12-55,2.

120 Part /: The Ammonian Synthesis

referents not only inseparable perfections but separate perfections as well. 11 What
the passages above and the Ristila ft n-nafs passages are meant to show is that
technically speaking, this is incorrect.
Now I shall examine the second problem mentioned above, namely, how
Avicenna distinguishes between first and second perfections. Generally speaking,
Avicenna has three methods of distinguishing first from second perfections. The
first (1) is cause of activity vs. activity itself; the second (2) is form vs.
function; and the third is split into two further subdivisions, (3a) necessary for
existence vs. necessary for well-being, and (3b) necessary for existence vs. not
necessary for existence.
The first way that Avicenna splits first and second perfections (1) can be found
in his Mabf.zath <ani l-quwa n-nafsaniyya and in his Risa/a ft n-nafs. In those
texts Avicenna bases the distinction between first and second perfections on the
difference between the cause or principle of an activity, and the activity which
follows upon that principle. In the Mabf.zath Avicenna says:

L23
lbn Sina, Mabf:iath <ani l-quwii n-nafsiiniyya 2, 153, l l-13
Perfections are divisible into two types: [they are] either the
principles of activities and effects [mabiidi' u l-afiiclli wa-l-iithiiri],
or [they are] the activities and effects themselves, the one being
"first" and the other being "second". The first is the principle [al-
mabda,u], and the second is the activity and effect. The soul is a
first perfection [kamiilun awwalu] because it is a principle, and not
originating from a principle.

This is very close to what Avicenna says in the Ristila ft n-nafs:

L24
Ibn Sina, Risiilafi n"nafs wa-baqii'ihii wa-ma<adihii I , 55,12-56,1
Now a body's perfection may be a principle, or it may follow upon
a principle [qad yakunu mabda'an wa-qad yakunu bacda l-mabda,i],
for sensation and locomotion are also a perfection of the animal
body. Now the soul is a principle of these [activities], and for this
reason we say that the soul is a body's first perfection.

Although defining the first perfection in terms of its causation of the second
perfection may smack of the Neoplatonists' teleiotes, this would, I think, be a
superficial reading. Avicenna's distinction seems rather to be a slightly modified
version of Aristotle's. Avicenna's second kamal and Aristotle's second
entelekheia are identical in that they both refer to activity. But whereas
Aristotle' s activity follows after a first entelekheia understood as capability,
Avicenna's activity follows after a first kamal understood as cause.

11 Ibn Sina, K. al-lzudud, 16,6-13 and 17,7-8.

 6. Avicenna on Perfection and the Soul 121
An example of the second method of distinguishing first and second
perfections (2) occurs in the Fi n-nafs section of the Shifii':

L25
Ibn Sina, Kitab ash-shifa'/Tablciyyiit (6): Ff n-nafs 1.1, 11,7-11
Avicenna, liber de anima seu sextus de naturalibus I-Ill,
27,40-28,45
"Perfection" [al-kamiil = perfectio] has two aspects: a first
perfection and a second perfection. The first perfection is that by
which a species becomes a species in actuality [bi-1-ficli = in
effectu], such as the shape [ka-sh-shakli = sicut figura] that
"sword" has. The second perfection is a kind of acting or being-
acted-upon that attaches as a consequence to the thing's species,
such as the cutting that "sword" has, and such as the distinction-
making, deliberating, sensing and moving that "human" has. 12

Although Avicenna's distinction between first and second kamiils in this passage
seems quite close to Aristotle's in Ll, there is a subtle difference. Avicenna's
distinction appears to derive from a slightly later De Anima passage (DA 2.1,
412bl l-413a3), where Aristotle says that cutting is to the power of an axe just
as seeing is to the power of an eye. But whereas Aristotle clearly meant for the
example in this later passage to be understood as illustrating the distinction
between the capability to perform a function and the function itself - and hence
to be understood as a refinement of the capability/exercise distinction introduced
earlier, in LI - it seems to have been interpreted by Avicenna instead as referring
to a distinction between form and function.
Avicenna might have reasoned along the following lines: given that Aristotle
says at the very beginning of Ll that the soul is entelekheia in the sense of
substantial form; given that Aristotle goes on to say later in Ll that the soul is
a first entelekheia; and given that in Ll and in the later passage (412bll-413a3)
Aristotle illustrates what first entelekheia means by using an example that is
most easily interpreted as referring to a capability to perform a function; form
and capability to perform a function are therefore equivalent. The problem with
this line of reasoning is that when Avicenna gives an example of first kamiil, he
says that it should be understood as "shape" (shakl = morphe). The implication
is that soul is form not in the sense of substantial form, but in the sense of
shape, and this is clearly not what Aristotle meant, given that in Ll he linked
the soul to form as substance.
Avicenna's third method of distinguishing between first and second perfections
is itself divided into two versions. The first version (3a) bases the distinction on
the difference between what is necessary for a thing to exist, and what is

12 As with its definition of the soul, the FT n-nafs distinction between form and function in L25 is

repeated in somewhat compressed form in the Burhlin: Ibn Sina, K. ash-shifii'/Man!iq (5): al-
Burhlin 2.9, 182,16-23. On the Latinfortuna of L25 see Hasse 2000, 235.
122 Part /: The Ammonian Synthesis

necessary for its well-being, or for it to exist well. This 1s articulated in a

passage in the Najar:

L26
Ibn Sina, Kitab an-najat, 161,6-14
As a result of the separate principle, natural bodies seek their own
perpetuity as well as that of their perfections [istibqa'un li-
dhawatiha wa-stibqa'un li-kamaliitiha]. Now their perfections are
either first perfections, upon whose disappearance the things
whose perfections they are cease to exist; or second perfections,
whose disappearance does not lead to the ceasing-to-exist of the
things whose perfections they are, but instead leads to the thing's
ceasing to be in a state of well-being (ilii rtijaci ~alaJ:ii f.zalatihi).
The separate principle does not itself go about seeking the
perpetuity of these second perfections, but [does so] instead by
means of placing powers in the [natural] bodies; for it is the first
perfections and principles from which these second perfections
issue. The activities of natural bodies derive from their second
perfections, given that their activities come about through these
powers.

The distinction here in the Najat is clearly infonned by the Neoplatonic

distinction between to einai, the existence that is passed along in the downward
course of procession, and to eu einai, the well-being that is striven for in the
upward course of reversion. This distinction also seems to be in line with a
distinction Avicenna makes in several of his biological works, where he splits
the final cause (al- cillatu t-tamamf, in this case) into a necessity and a benefit
(<f.arii.ra wa-manfa<a), a distinction which itself derives from the Aristotelian
distinction between ananke and opheleia. 13
The second version of Avicenna's third method (3b) of distinguishing between
first and second perfections is found in his Notes (Ta<ifqiit) and in the
Metaphysics of his Book of Science for cAla' ad-Dawla (llahiyyat-i
Danishnama-yi cAla'l). In these two works Avicenna bases the distinction not
on the difference between what is necessary for a thing's being and what is
necessary for its well-being, as with (3a), but rather on what is necessary for a
thing's being and what is extra, or a bonus. Avicenna puts it this way in the
Tactiqiit:

13 e.g., K. ash-shifii'l['abfciyyiit (8): al-l:fayawiin 12.5, 207,5 =al-Qiinunfi f-fibb ( 1) 1.4.2, 35,23.

Avicenna's formulation may derive from Aristotle's statement at GA 1.4, 717a16-17 (= Ff kawn
al-f:iayawiin, 7,3-4), that the final cause of a thing is either "by necessity" (dia to anankaion = li-
lladhf huwa liizimun bi-t;ffiriirin) or "for the better" (dia to beltion = li-lladhl huwa amthalu wa-
ajwadu).
 6. Avicenna on Perfection and the Soul 123
L27
Ibn Sina, at-Taclfqat
21,16-19: Just as He [i.e., God] bestows upon everything what it
needs as far as its existence and perpetuating [its existence] are
concerned, so He also bestows upon it something above and
beyond its need for these. An example is the fact that He bestows
upon man philosophy and astronomy, since no man needs
astronomy as far as his existence and perpetuating [his existence]
are concerned. What is required as far as his existence is concerned
is the first perfection [fa-ma la budda minhu ft wu)Udihi huwa l-
kamalu l-awwalu], while that other is the second perfection.
21,22-22,2: The Necessary of Existence is also the cause of every
existent, and has bestowed upon every existent its existence's
perfection, which is what it needs as far as its existence and
perpetuating [its existence] are concerned; and He has also added to
these two what it [the existent] does not need. The Qur'an hints at
this concept when it says [20:50]: "He [i.e., Moses] said: 'Our
Lord is He Who gives to everything its creation, then guides [it]'".
The guidance is what he [man] does not need as far as his existence
or the perpetuation [of his existence] are concerned, while the
creation is what he needs as far as his existence or the perpetuation
[of his existence] are concerned. Also when it says [quoting
Abraham] [26:78] : "Who created me [i .e., Abraham], then guided
me". Philosophers call what something needs as far as its
existence or the perpetuation of its existence are concerned the
first perfection, and what it does not need as far as its existence or
the perpetuation [of its existence] are concerned the second
perfection. 14

It seems that just as we were faced with the challenge of reconciling the various
ways Aristotle explicitly (as in De Anima 2) and implicitly (as in Physics 3)
distinguished first from second entelekheiai, so we must try to make some
sense of all the different ways - (1 ), (2) and (3a) and (3b) - Avicenna has divided
first and second perfections. One question I should try to answer first is whether
these different ways of distinguishing first and second perfections might reflect a
progression or development in Avicenna's thought.
The question of development arises because each of the different ways of
dividing first and second perfections comes in a cluster of texts - (1) in the
Mabf:iath and the Risiila ft n-nafs; (2) in the Ff n-nafs and Burhiin of the Shifii';
(3a) and (3b) in the Najiit, Tacllqiit and Diinishniima-yi cAlii'f - and because
each text coincides chronologically with the other texts in its cluster. The first
cluster of texts (1) appears to date from Avicenna's earliest period of

14 The terms used in the Diinishniima-yi 'Alii,f are iifirinish-i farft;fa and iifirinish-i ziyiidat,

referring to God's creation of what is required for an existent to exist (i.e., existence), and God's
creation of something extra which is not required for an existent to exist (e.g., philosophy or
astronomy): Diinishniima-yi 'Alii,f: Iliihiyyat, 100,9-11 (cf. also 109,1). The Qur'anic quotations
cited in the Ta'llqiit passage are also cited in the Diinishniima passage at 100,7-9.
124 Part I: The Ammonian Synthesis

philosophical activity, from 1000-1014.' 5 The second cluster of texts (2), taken
from the Kitiib ash-shifii", dates from what Gutas has referred to as Avicenna's
"Middle Period", between 1020 and 1027. 16 The third cluster of texts, (3a) and
(3b), date from the very end of the middle period, around 1027 (3a); and from the
period immediately following, which Gutas has called the "Period of Eastern
Philosophy", between 1027 and 1030 (3b). 17
Despite the close correspondence of texts within each cluster, it would be a
mistake, in my opinion, to argue that a development took place in Avicenna's
thought. Apart from anything else, someone arguing for a developmentalist
interpretation of Avicenna's theory of perfection would have to confront the fact
that the Najiit passage (L26) turns out to be identical to, and hence was copied
and pasted from, !Jikma cAru¢iyya, 51 v5-10, and therefore dates from around
100 l, during Avicenna's early period. 18
I think instead that it would be more accurate to argue that the difference
between the ways the distinction between first and second kamiils is expressed in
cluster (1) and cluster (2) can be attributed instead to the type of work Avicenna
was writing. The texts in cluster ( l) are independent, stand-alone essays, works
in which Avicenna felt able to modify Aristotle's distinction in a more dramatic
way than he would have done in cluster (2). This is because the texts in cluster
(2) are taken from the Shi/ii", Avicenna's most Aristotelian work, in which
Avicenna wanted to veil his modifications by using examples, such as the axe
and cutting, which were drawn from the Aristotelian texts themselves.
By contrast, the difference between the way Avicenna expresses the first
kamiil/second kamiil distinction in clusters (1) and (2), and the way he expresses
the distinction in clusters (3a) and (3b), is attributable not so much to a
difference between types of work as it is to a difference between subject-matter.
In clusters (3a) and (3b ), Avicenna is not writing about the soul in particular but
about natural philosophy (3a) and cosmology (3b) more generally. And whereas
the ways he expressed the distinction in clusters (1) and (2) are obviously derived
- though with modifications - from Aristotelian psychology, the ways he
expressed the distinction in clusters (3a) and (3b) are clearly inspired - again,

15 Gutas 1988, 82-87, argues for a very early dating (ca. 1000) of the Mab~ath 'ani 1-quwii n-

nafsiiniyya (which he refers to, following Landauer, its first editor, as Maqiila ft n-nafs 'alii sunnat
al-ikhti~iir, The Compendium on the Soul). He dates the R. ft n-nafs wa-baqii'ihii wa-ma'iidihii
(which he refers to as l:fiil an-nafs al-insfmiyya) at around 1014 (Gutas 1988, 99-100). For an
argument against Gutas and in favor of a later dating of this work see Michot I 997a, 240n.6.
16 Gutas 1988, 101-112.
1 7 On the dating of the Najiic to around 1027, see Gutas 1988, 112-114. Gutas (Gutas 1988, 141-

144) actually dates the Ta'llqiillo the last period of Avicenna's life, that is, to around 1035-1037.
But I believe that Janssens 1986 and 1997 has shown enough of a textual correspondence between
the Ta' liqiit and the Diinishniima-yi 'Alii'i - verified by the correspondence I just demonstrated
between L27 and the corresponding Diinishniima passage - to throw the burden of proof on those
who would deny that they came from the same period, namely, 1027-1032.
18 Gutas 1988, 87-93 and 112-114, discusses the Najiit's borrowings from Avicenna's earliest

works.
 6. Avicenna on Perfection and the Soul 125
with slight modifications - by Neoplatonic cosmology, and particularly by the
Neoplatonists' application of the distinction between to einai and to eu einai to
their cosmology of procession and return.
In other words, just because each cluster might - with some rough edges - be
distinguished from the others chronologically, a doctrinal progression from one
cluster to the next need not be postulated. I cannot see that the way Avicenna
distinguished between first and second kamiils in clusters (3a) and (3b) is
somehow a response to, or more coherent than, the ways he distinguished
between first and second kamiils in clusters (1) and (2). On the other hand, I cb
see that there is a sense in which the ways the distinction is articulated in
clusters (1) and (2) can be reduced to the ways the first kamiil/second kamiil
distinction is articulated in (3a) and (3b ), and that in this sense the (3a) and (3b)
cluster is the most philosophically basic articulation of the distinction. To be
more precise, I believe (3a) to be the most basic articulation of the distinction.
This is because the way the distinction is articulated in (3b) seems to be nothing
more than (3a) with some religious terminology (khalq and iifirinish) and a
couple of Qur"anic quotations (20:50 and 26:78) thrown in for good measure.
How can the distinction expressed in cluster (1), namely, the principle of an
activity as opposed to the activity that follows from that principle; and the
distinction expressed in cluster (2), namely, a form as opposed to the function
associated with possessing that form; be reduced to the distinction expressed in
cluster (3a), namely, what is necessary for existence as opposed to what is
necessary for well-being? Well, an activity that follows from a principle and a
function such as "cutting" are quite obviously reconcilable. Just as "activity" in
general will either be allotelic, in which case it leads to an end outside itself, or
autotelic, in which case it leads to no end other than itself, so "function" -
which means nothing more than an activity whose endedness is being stressed -
will either be allotelic (e.g., reproduction) or autotelic (e.g., intellection).
What about "principle" in cluster (1) and "form" in cluster (2)? I think that
what they share becomes more obvious if the term mabda, in (1 ), which I have
translated as "principle", is understood as referring more to the idea of "starting-
point" than to the idea of "cause". Let us also assume that when Avicenna said
in cluster (2) that the first kamiil is form as shape, what he was doing was
simply emphasizing the endedness of the soul. What I mean is that the soul is
not simply the substantial fonn that results from a previous activity -
reproduction - it is also the springboard for future activities. In the sense that it
is the springboard for future activities, the soul is more like a form as shape -
the arrangement of matter structured with a view to performing some function -
than it is like a fonn as substance. With this in mind, the form can be seen as
the springboard from which there arise activities associated with possessing that
fonn. My soul is the springboard from which there arise the activities or
functions associated with my soul, such as moving, perceiving, deliberating and
so on. Similarly, the form of an axe is the starting-point from which the
126 Part /: The Ammonian Synthesis

activities associated with that form, such as chopping, derive. That is to say,
just as without the axe's form, the axe's activity of chopping could never take
place, so without the human soul's form (i.e., its complex structure of
capabilities), the human soul's activities of nourishing, growing, reproducing,
moving, sensing, acting and thinking could never take place.
Both (1) and (2) can be reduced to (3a), then, because the "starting-point" in (1)
and the "form" in (2) are both required for a thing - a fonn/matter composite,
that is - to exist. By contrast, the thing does not have to perform the activity or
function that springs from possession of that form in order to exist. An axe that
sits in the garage and is never used for chopping is still an axe, because it
possesses its form. A human who watches television all day and never exercises
his intellectual capability is still a human, because he possesses his soul. But an
axe that is often used for chopping, and a human who thinks constantly about
philosophy, not only exist but exist well, since well-being results from the full
and repeated exercise of a thing's capabilities.
Now that I have reduced the ways in which the distinction between first and
second kamiils is expressed in the other clusters to the way it is expressed in
cluster (3a), the question arises, how does motion or change, which Avicenna
defines in L20 as a first perfection, fit into the picture? I have to admit that I am
stumped by this question. I cannot see that motion or change is a starting point
as opposed to the activity that springs from that starting-point; if anything,
change is itself an activity, the second kamiil as articulated in cluster (1). I also
cannot see that change is a form as opposed to a function, since again, change is
surely closer to function than it is to form, and hence change ought to be seen as
a second kamiil, at least as it is articulated in cluster (2). And as I just discussed,
a thing need not be active - in a state of change, that is - in order to exist, so
change fails to fulfill the requirement of being a first kamiil according to clusters
(3a) and (3b). My guess is that Avicenna's calling change a first kamiil is
nothing other than a vestigial remain of the Physics 3 commentary tradition he
inherited from Philoponus, Themistius and perhaps Alexander, if Simplicius is
quoting him accurately. In other words, in the context of his Physics, Avicenna
is happy to apply the Arabic version of the commentators' distinction between
entelekheia prate and entelekheia hustate to, respectively, the state of change
which the changing thing possesses and the resulting state of actuality which the
changed thing possesses, without feeling the need to reconcile that distinction
with the numerous articulations of the first kamiil/second kamiil distinction
found elsewhere in his work.
I believe that the reason for this is that like Philoponus but unlike Themistius
and Alexander, Avicenna was a part of the Ammonian synthesis. That is to say,
Avicenna was heir to the effort, begun by Ammonius, to fuse together the
greater sumphOnia (reconciling Plato and Aristotle) and the lesser sumphOnia
(reconciling Aristotle with Aristotle). What that meant in practical terms is that
Philoponus' appeals to Aristotle's distinction between final causes which are to
 6. Avicenna on Perfection and the Soul 127

hou and final causes which are to hOi, were intended to make the soul, or at least
the intellectual part of it, separable, and hence reconcilable with Plato's
conception of the soul. In fulfilling that primary objective Philoponus assigned
a lower priority to the Alexandrian and Themistian objective of constructing a
scheme which could reconcile Aristotle's uses of entelekheia in De Anima 2
with his uses of entelekheia in Physics 3. Nevertheless, that lesser objective
still motivated Philoponus to some extent. In Avicenna, by contrast, the lesser
objective had almost totally disappeared: ensuring that the kamiil which he
appealed to in his Physics was the same concept which he appealed to in his De
Anima does not appear to have motivated him.
In other words, it is part of the legacy of the Ammonian synthesis that
Avicenna's larger aim of creating a coherent picture of the soul's relationship
with the body, by applying the language of the Arabic Aristotle and appealing to
the sense in which kamiil could be understood as a final cause, subordinated the
lesser aim of constructing yet another scheme which could accommodate his own
use of kamiil to describe motion and his own use of kamiil to describe the soul.
Of course it is difficult to tell what Avicenna in commentator mode would have
said about how first perfection applies to change, and how its application to
change was consistent with its application to the soul, since the bits and pieces
of the Kitiib al-in:;af which are extant do not, sadly, include his comments on
Physics 3.

Now that I have discussed how Avicenna distinguishes between form and
perfection, as well as how he distinguishes between first and second perfections,
I must turn to his treatment of the soul's causal relationship to the body. What
then are Avicenna's assertions about the way the soul is the cause of the body?
Like Aristotle, Avicenna is willing to allow that the soul is cause of the body as
form, agent and end. 19 Yet in L22 above, Avicenna clearly ties perfection to the
final cause, saying that technical usage dictates that a thing may be considered to
be an end and perfection when it is spoken of relative to the whole or composite
(wa-bi-l-qiyiisi ilii l-jumlati ghiiyatan wa-kamiilan). So on the one hand, given
that the soul is the perfection of the body; and given that the soul is cause of the
body as form, agent and end; a perfection can be related to its effects as form,
agent or end, or as all three together. On the other hand, even apart from the clear
tying together of end and perfection in L22, there is a great deal of evidence in
Avicenna's writings that when we speak of a perfection in relation to its effect,
we are speaking of it only as a final cause, or at least primarily as a final cause.
How did this ambivalence arise? The reason can be found in the vacillations of
the Greco-Arabic translators. For even though Avicenna follows Isl:iaq in using
ghiiya - which translated telos in the Arabic Physics - in several of the more

19 Soul is form: K. ash-shifii'/Manfiq (5): al-Burhi'in 2.9, 182,11-12; soul is end: at-Ta'liqiit, 63,6;

soul is agent and end: Ta'liqi'lt, 63,28 and 64,2.

128 Part /: The Ammonian Synthesis

canonical passages on the four causes, such as that introducing lliihiyyiit 6.1 of
the Shifii', Avicenna also follows Ustat in using tamiim - which translated
entelekheia as well as telos in the Arabic Metaphysics - in many other
passages. For example, in his introduction to the four causes in the Burhiin (On
Demonstration) of the Manriq (Logic) of his Kitiib ash-shifa', Avicenna says:

L28
Ibn Sinii., Kitiib ash-shifii'/Manfiq ( 5 ): al-Burhiin 4.4, 294, 11-14
The causes are four. One is a thing's form, [consisting] in the
inner reality of [the thing's] existence in itself. Next is the thing
which must first exist to receive the form of its existence [so that]
when it actually bears [the form], it comes into being, namely, the
matter. The third is the principle of motion, namely, the agent.
Fourth is the thing for the sake of which the entity's form and
matter come together, namely, the end [wa-huwa t-tamiimu]. 20

And when he speaks of the final cause Avicenna uses the expressions al- cillat
al-ghli'iyya, al-cillat at-tamiimiyya, and al-cillat al-kamiiliyya, in such an
interchangeable way that it is hard to avoid the conclusion that he thought they
enjoyed a relationship of mutual implication, if not actual synonymy. Avicenna
often calls the final cause al- cillat at-tamiimiyya or as-sabab at-tamiiml; for
example, in the section of lliihiyyiit 6.5 of his Kitiib ash-shifa' where he argues
against the idea that there can be an infinite series of ends, Avicenna says:

L29
Ibn Sinii., Kitiib ash-shifa'/lliihiyyiit (2) 6.5, 289, 16-290,5
Avicenna, Uber de philosophia prima sive scientia divina V-X,
334,30-40
It seems that we have digressed from our purpose, so let us return
to it and respond to the objection raised [viz., "How is an infinite
regress of ends to be avoided?"]. We say that individual entities
which are infinitely many cannot be essential ends in nature.
Rather, essential ends are, for example, that the substance that is
man - or horse or palm-tree - should exist, and that this existence
should be a continuous and well-established existence. This
[perpetuity] is impossible in [the case of] the aforementioned
single individual, because a necessary concomitant of every entity
(I mean entities [derived] from corporeal matter) is passing-away.
And as this [perpetuity] is impossible in the individual, it is
maintained through the species. The primary purpose [here] is the
maintenance, for example, of human nature, or of something other
than [human nature], or of an indeterminate extended individual.
This [purpose] is the final cause [wa-huwa 1- cillatu t-tamiimiyyatu
causa perfectiva] of the act of universal nature.21

20 Other instances where Avicenna calls the final cause at-tamam are K: ash-shifa,/Man{iq ( 5 ):

al-Burhiin 4.4, 294,13-14 and 4.5, 297,15-16.

21 Other instances where Avicenna calls the final cause al-cillatu t·tamamiyya, as-sababu r-

tamiimi or al-mabda>u t-tamami are K. ash-shifa,11/ahiyyar (2) 8.3, 340,14-341,13; K. ash-

 6. Avicenna on Perfection and the Soul 129

Elsewhere, though less often, Avicenna uses al- cillatu l-kamaliyya and as-
sababu l-kama!l. 22
As was the case with Neoplatonic commentators such as Philoponus,
Avicenna's views about the final cause in general had important ramifications on
the particular issue of the soul's separability and inseparability. In the
introduction to the section on causes in the /lahiyyat of his Kitab ash-shifa',
Avicenna claims that the four causes can be divided into two subsets: one subset
is the matter and the form, which are intrinsic to or immanent in their effect; the
other subset is the agent and the end, which are extrinsic to or transcend their
effect:

L30
Ibn Sina, Kitab ash-shifa'!llahiyyat (2) 6.1, 258,1-8
Avicenna, liber de philosophia prima sive scientia divina V-X,
292,27-36
We say that a thing's cause must either be intrinsic to its
constitution and a part of its existence, or not [imma an yakuna
dakhilanfi qiwamihi wa-juz'an min wujudihi aw la yakuna = vel est
intra essentiam rei et pars esse eius ·vel non]. If it is intrinsic to
[the thing's) constitution and a part of its existence, then either it
is the part on account of whose existence alone it is not necessary
that [the thing] be in actuality, but rather only in potentiality, and
is called "hyle"; or it is the part whose existence is identical to the
[thing's] coming to be in actuality, and is "form". If it is not a part
of [the thing's] existence, then either it is that for whose sake [the
thing) is, or it is not. If it is that for whose sake [the thing] is, it
is the "end". If it is not that for whose sake [the thing] is, then
either [the thing's] existence comes from it without its being
contained in [the thing] except in an accidental way - and is the
"agent" - or [the thing's) existence comes from it with it being
contained in [the thing), and is also the "element" or "substrate".

I already discussed - in Chapters 3 and 4 - the Neoplatonic distinction between

immanent and transcendent causes, as well as the history of its incorporation
into Aristotelian philosophy, a history which appears to have begun with

shifii'fTabl'iyyi'lt (8): al-/fayawiin 12.5, 207,2.5 and 12.13, 256,17; al-Qiinun fi f-fibb (I), I.4.2,
35,23; R. fi n-nafs wa-baqii'ihii wa-ma'iidihii I, 53,13-54,2; K. ash-shifii'/Manfiq (5): al-Burhiin 2.9,
182,5; 4.4, 295,9 and 297,4.14; Diinishniima-yi 'Alii'f: /lahiyyiit, 54,7; and ar-Risalat al-'arshiyya fi
~a~ii'iq at-taw~ld wa-ithbiit an-nubuwwa, 18,2-3.
2 Instances where Avicenna calls the final cause al-'illatu 1-kamiiliyya and as-sababu 1-kamiilf

are K. ash-shifa'fTabl'iyyiit (6): Ff n-nafs 5.4, 228,7; R. fi n-nafs wa-baqii'ihii wa-ma'iidihii 9,

100,2.10; K. ash-shifii'!Man{iq (5): al-Burhiin 2.9, 182,12; and ar-Risiilat al-'arshiyya fi ~aqii'iq at-
taw~fd wa-ithbiit an-nubuwwa, 22, 18-20 and 29,3-4. In using al-'illatu 1-kamiiliyya Avicenna could
be following lbn Miskawayh, who also cites Aristotle as the source of the view that the final is the
noblest of the causes (wa-hiidhihi 1- 'illatu 1-akhlratu llati tusammii 1-kamiiliyyata wa-hiya ashrafu 1-
'ilali); ap. at-TawJ:iidi, al-ljawiimil wa-sh-shawiimil, #159, 342,17-343,2.
130 Part I: The Ammonian Synthesis

Syrianus. 23 We are given a hint that Avicenna probably viewed his distinction
between immanent and transcendent causes as being in line with a tradition of
interpreting Metaphysics 5.l, Aristotle's discussion of arkhe, "origin" or
"principle". In his earliest summa, the ljikma cAru¢iyya, Avicenna introduces
the same distinction between immanent and transcendent causes that he would
expand upon later in the Ilii.hiyyat of the Kitii.b ash-shifii.'; but in the /jikma
cAru¢iyya - much of whose metaphysics section seems to be an abridgement of
the "philosophical lexicon" Aristotle offers in Metaphysics 5 - Avicenna uses
mabda', the Arabic term that corresponds to the Greek arkhe, instead of 'ilia,
the Arabic term that corresponds to the Greek aition, which Aristotle discusses
in Metaphysics 5.2. 24 Avicenna says:

L31
lbn Sina, al-/jikmat aVArufiiyya, 4vl6-5r5
"Origin" [al-mabda'] is that upon which a thing's existence
depends, either its matter, if it is material; or its form, if it is
composed of matter and form; or its end, if it exists for some end;
whereas that which facilitates and particularizes, and the
instrument, attach as concomitants to the agent; and the paradigm
[exists] because of that very agent's imperfection with respect to
the act's issuing and existing from it. A composite substance has
two conjoint causes [sababiini muqiiraniini] - its matter and its form
- and two transcendent causes [sababiini mufiiraqiini] - its agent and
its end. An accident has three causes, of which form is not one.
Form and the separate [substances] have two causes: agent and
end.25

23 See Wisnovsky 2003a.

24 There is in fact a striking correspondence between the metaphysics section of Avicenna' s
/jikma 'Aru¢iyya and Metaphysics 5. In addition to mabda'larkhe, Avicenna discusses Aristotelian
terms such as "potentiality" (quwwaldunamis = /jikma 'Aru¢iyya 3r6-171Metaph. 5.12),
"necessary" (wi'ljiblanankaion = f:likma 'Aru¢iyya 4vl6-5r5/Metaph. 5.5), "perfect" (tlimmlteleion
= /jikma 'Aru¢iyya 4vl4-16/Metaph. 5.16), "universal" (kul/flholon = 4v6-12/Metaph. 5.26) and
"prior and posterior" (qabla + mutaqaddimlproteron and ba'da + muta'akhkhirlhusteron = /jikma
' Aru¢iyya 5r7-5vl5/Metaph. 5.11).
25 Unlike many other sections of the /jikma 'Aru¢iyya, this passage is not copied in an-Najiit.

Other passages where Avicenna makes the distinction between immanent (formal and material)
and transcendent (efficient and final) causes are K. al-hidliya, 243,5-244,3 (and implied at 237,4-
5); 'Uyun al-J:rikma, 52,3-11; K. an-najiit, 344,2- 14; Diinishniima-yi 'Alii' i: Iliihiyyiit, 53,9-54,9
(where the distinction is between causes which are inside the effect itself [andar dhat-i ma'IUI]
and those which are outside the effect itself [birun az dhiit-i ma'lUI]); and ar-Risiilat al-'arshiyya Ji
J:raqii' iq at-tawJ:rid wa-ithbiit an-nubuwwa, 18,5-11. Jolivet 1991 was the first to notice the distinction
in Avicenna, though he seems unaware of its Neoplatonic pedigree, nor of the apparent evolution
in Avicenna' s views and uses of the distinction, both of which I describe at greater length in
Wisnovsky 2003a. Although the distinction between intrinsic and extrinsic causes originates in the
Neoplatonists, Avicenna' s immediate source may have been al-Farabi, in whose epitome of the
Physics (preserved only in Latin) the distinction appears; cf. materia et forma (que due sunt intra
rem) et agens et finis (que due sunt extra rem): Epitome of Aristotle 's Physics, 475,7-8.
 6. Avicenna on Perfection and the Soul 131

But if final causes transcend their effects; if perfection acts on its effect - the
body - as a final cause; and if the soul is a perfection; then the soul will always
transcend the body. But this is not quite what Avicenna wants to prove, since he
thinks that vegetative and animal souls, at least, are immanent in their bodies.
How does Avicenna avoid the dilemma of being forced on the one hand into
holding that all souls transcend their effects just as ends do, and being forced on
the other hand into holding that all souls are immanent in their effects just as
forms are?
Part of the job was done for him by the Arabic translators' conflation of two
distinct ideas in Greek philosophy. The result of the use of the phrases al- c illatu
t-tamlimz and as-sababu t-tamlimz to render to aition telikon ("final cause") as
well as to aition teleiotikon ("perfecting cause") in the commentaries, was that
the difference between the two types of cause - the former an end, the latter a
Form which operated as an agent and an end - was lost in the Arabic, which
collapsed the latter into the former. The result was that Arabic philosophers such
as Avicenna allowed the final cause an even greater spectrum of causality than
the Neoplatonic philosophers had. For example, in the section of Iliihiyylit 6.5
of his Kitlib ash-shifli' where he discusses whether "end" and "good" are
identical, Avicenna says:

L32
Ibn Sina, Kitab ash-shifa'!Jlahiyyat (2) 6.5,
294,6-9; 294,12-295,12; and 296,3-5
Avicenna, Liber de philosophia prima sive scientia divina V-X,
339,33-37; 339,41-341,71; and 341,84-86
As for the question subsequent to this [viz. are the end and the
good identical?], it will be resolved through our saying that an end
which occurs in an agent's act is divisible into two categories: [I]
an end which is a form or accident in a patient receiving the act,
and [2] an end which is not in any way a form or accident in a
receiving patient, and is thus inevitably contained in the agent ....
An example of the first [category] is the form of humanity in
human matter: [the form of humanity) is an end of the active
potential for formation in the matter of man. It is toward [this end]
that its act and its setting-in-motion are directed. An example of
the second [category] is seeking shelter, for it is an end of the
person seeking to build the house and is thus the principle of
motion of his being [a builder], and is not in any way a form
contained in the house. It appears that the end of the proximate
agent (the one who is in closest contact with the setting-in-
motion of the matter) is a form contained in the matter, and that
that whose end is not a form in the matter is [conversely] not a
proximate principle of the setting-in-motion as such. So if it
happens that that whose end is a form contained in the utilized
matter, and that whose end is something other than a form in this
matter, were one thing, then its oneness would be accidental. [Say]
for example, that a man is building a house therein to take shelter:
132 Part / : The Ammonian Synthesis
in so far as he is a seeker of shelter, he is a motivator and a first
cause of the act of building. [But] in so far as he is a builder, he is
an effect of the seeker of shelter. Hence the end of the seeker of
shelter is different from the end of the builder. This being so, in
the case of the one man who is both a seeker of shelter and a
builder, his end in so far as he is seeker of shelter will be different
from his end in so far as he is builder.
Now that this has been settled, we say: as for the first category,
the end has a relationship to many things which are prior to it in
terms of actual occurrence [in time] as well as in terms of
existence. [This is so] because [the end] has: [I] a relationship to
the agent; [2] a relationship to the receiving patient when [the
receiving patient) is potentially [a receiving patient); [3] a
relationship to the receiving patient when [the receiving patient]
is actually a receiving patient; and (4) a relationship to the
motion. With regard to [l) the agent, [the end] is an end, and with
regard to [4] the motion it is a terminus and not an end. This is
because the end [is that] on account of which the thing exists, and
toward which the thing is directed; the thing will not cease to exist
once [the end] exists, but will instead be perfected by [the end].
The motion, by contrast, ceases to exist once it reaches its
terminus. With regard to [2] the potentially perfected receiving
patient, [the end] is a good which sets [the receiving patient] in its
proper state. [This is] because evil consists in the non-existence
of [the receiving patient's] perfection, whereas good, by contrast,
consists in [the receiving patient's] actual existence and
occurrence [in time]. And with regard to the actualized receiving
patient, [the end] is a form ....
Therefore, every end is on the one hand an end, and on the other
hand a good, either supposed or real. And this is the situation
concerning the good and the final cause [billu l-khayri wa-l-cillati
t-tamamiyyati = causae perfectivae].

Let me summarize Avicenna's new taxonomy of final causes as outlined in the

passage above. First of all, Avicenna makes a distinction between ends that are
contained in the patient - that is, ends that are immanent in the effect that is
acted upon by the cause - and ends that are contained in the agent - that is, ends
that are immanent in the cause that acts upon the effect. He claims that ends that
are immanent in the patient are forms contained in matter: that is to say, they
are natures, essences, and quiddities, such as "humanity" and "house-ness". Those
ends that are immanent in the agent, on the other hand, are the purposes or ends
the agent has in mind - seeking shelter, for example - and not the form of the
structure he is building.
Avicenna expands his taxonomy to cover not only actual agents and patients
but also potential patients and motion. Strictly speaking, Avicenna says, the
final cause may be related to several things: (1) when it is related to the actual
patient, the final cause is a form; (2) when it is related to the agent, the final
cause is an end; (3) when it is related to motion, the final cause is a terminus;
 6. Avicenna on Perfection and the Soul 133

and (4) when it is related to the potential patient, the final cause is a good. Final
causes of types (1), (2) and (3) are immanent in their effects; whereas final causes
of type (4) transcend their effects.
The upshot is that because Avicenna holds that the final cause is identifiable
with perfection, perfection will subsume form (i.e., number 1), not be identical
to it, as the standard interpretation of Aristotle holds. When viewed as a final
cause, therefore, perfection can perform the immanent/transcendent double-duty
demanded of sublunary souls more plausibly than form can. That Avicenna
should have seen his expansion of final causality to subsume formal causality,
as being entirely in line with Aristotelian thought should not surprise us, given
Aristotle's general implication that the end is superior to the other causes, a
position explicitly upheld by Avicenna.26
Avicenna's new understanding of perfection and the final cause allowed him to
subsume formal causality under final causality. But what about the efficient
cause, the third type of cause which Avicenna, following De Anima 2.4, says
the soul can be seen to be? This question is particularly acute because, as I
mentioned above, Avicenna's cilla tamiimiyya seems at least partly to have been
the product of the translators' conflation of the aition telikon, which (obviously)
operates as a final cause, and the aition teleiotikon, which operates as a final
and as an efficient cause. I believe that Avicenna's attempt to create a concept of
final causality broad enough to subsume not only the sense in which the soul is
the form of the body but also the sense in which the soul is the efficient cause of
the body forced him to appeal, as Philoponus had done, to the distinction
between the two types of final cause: that in view of which (to hou) and that for
the benefit of which (to h8i).
What does Avicenna himself have to say about the Aristotelian distinction
between the type of final cause which is "that in view of which" (to hou) and the
type of final cause which is "that for the benefit of which" (to h8i)? Avicenna
treats the distinction most directly in his Marginal Notes on Aristotle's De
Anima:

L33
lbn Sina, at-Ta cllqiit calii f}awiishl Kitiib an-nafs li-Aris!u (ad DA
2.4, 4 l 5b2), 94, 18-23
... i.e., "that for the sake of which" [alladhl min ajlihi] is of two
types: one is "that in view of which" [alladhl min qibalihi], such
as pleasure, the good, health and so on; and the second is "that for
the benefit of which" [alladhl lahu], it being that which strives for
the end, such as that which strives for pleasure or health. Among
things which are subject to passing-away the end is that they
imitate - as much as is possible for them - the eternal, divine
things, namely the divine thing which is perpetuity. The end "for

26 See for example Metaph. 1.2, 982b5-l l. For Avicenna's view that the end is superior to the

other causes, see K. ash-shifii,llliihiyyiU (2), 300,7-9.

134 Part /: The Ammonian Synthesis
the benefit of which" consists in the things which are subject to
change. Since the thing which is subject to passing-away cannot
persist forever and attain this imitation (this being sought by it as
an object of desire and [through] an act of will) as an individual, it
strives for it in its species. 27

I ought to remind readers that Aristotle's distinction between to hou and to h6i
appeared most prominently in two contexts. In Physics 2.2, the distinction is
meant to differentiate between a thing's form - that in view of which - and a
thing's user - that for the benefit of which. The form of a door is a that in view
of which; the user of a door is a that for the benefit of which. In De Anima
2.4, the distinction is meant to differentiate between etemality (or divinity) - the
form striven for by individual, mortal members of a species - and the individual,
mortal thing which uses its body to attain immortality through its species.
Although as far as I can tell Avicenna nowhere else articulates the to hou!to
h6i distinction as explicitly as he does in L33, the language he uses in
discussions of how the soul is related to the body, and of how the higher
faculties of the soul, such as the intellect, are related to the lower ones, make it
quite clear that he had Aristotle's distinction between to hou and to h6i in mind.
What I mean is that Avicenna felt that a single thing - the soul - could be
viewed as an efficient cause and as a final cause, though from different angles, or
more precisely, as two different ways of explaining two different explananda.
In the sense that the soul is a rriover, in the sense that it causes the body to
move, the soul is an efficient cause. The soul as efficient cause is the explanans
of the existence of the body's motion. In the sense that the soul uses or benefits
from the body's motion, the soul is a final cause qua to hoi: the soul as final
cause qua to hoi is the explanans of the fact that the body's motion is intended
to attain some end, namely, the final cause qua to hou. But since the soul as
efficient cause of the body is numerically identical to the soul as final cause qua
to hoi of the body, Avicenna can subsume efficient causation under final
causation, in order to reconcile his implication that the soul's being the
perfection of the body is equivalent to its being the final cause of the body, with
his explicit assertion that the soul is the formal and efficient cause of the body as
well as its final cause. What Avicenna seems to mean, therefore, is that when we
use the term "perfection" to describe the soul, we are saying that the soul is first
and foremost a final cause, and that the other ways in which it is a cause - as
formal cause and as efficient cause - can in fact be subsumed under its being a
final cause. Viewed as the formal cause of the body, the soul is inseparable,
since formal causes are immanent in their effects. Viewed as the efficient cause

27 The fact that both the Arabic Themistius and L33 use alladhi min qibalihi and a/ladhi lahu for

to hou and to h6i, respectively, is yet more evidence linking the Arabic Themistius to Is~iiq's
revised translation of the De Anima: Themistius, in DA 2.4 (ad 415b3-7), 50,15-26 (= Thiimas\iyiis,
Sharfi K. an-nafs li-Aris{u{alis, 69,2-12).
 6. Avicenna on Perfection and the Soul 135

and final cause qua to h6i of the body, however, the soul transcends the body,
since efficient and final causes transcend their effects.
This is all well and good. But is Avicenna not forced into a contradiction,
given that he is well known to have thought that vegetative and animal souls are
immanent in their bodies, while human rational souls transcend their bodies;
given that vegetative and animal souls are efficient causes and final causes qua to
h6i, since vegetative and animal souls both move and use their bodies in view of
attaining a final cause qua to hou, namely, species-etemality; and given that
efficient and final causes transcend their effects?
The difference between vegetative and animal souls on the one hand, and
human rational souls on the other, is that human rational souls are the only
souls in which a faculty - the theoretical intellect - is both a final cause qua to
h6i and a final cause qua to hou. What do I mean by this? Avicenna says in the
Tac liq at that the soul both is an end and has an end; and says in the
MubiiJ:iathiit that the soul both is a perfection and has a perfection. 28 In other
words, it is important to think of perfection not only in the way that Aristotle
did in the De Anima, where the soul as a whole was viewed as a perfection; but
also to think of perfection as the Neoplatonists did, where perfection was viewed
as a universal principle which described the way all things - not just souls -
revert upwards towards their well-being, towards being what they are as best they
can. 29
As far as the soul is concerned, the perfection it has will of course be
something higher than it. When discussing how the faculties of the human soul
relate to each other, Avicenna employs terms - "using", "serving" and
"instrument" as well as "attaining" and "end" - that make it clear that he saw
each faculty 's relationship with the one below it as the relationship between a
final cause qua to h6i (higher faculty) and the effect of a final cause qua to hOi
(lower faculty); and each faculty's relationship with the one above it as the
relationship between the effect of a final cause qua to hou and a final cause qua
to hou. In other words, faculty (Y) uses lower faculty (X), which in tum uses the
body, in view of attaining a final cause qua to hou. My theoretical intellect uses
my practical intellect, which in tum uses my animal faculty of locomotion, to
move my body to the library, in view of my theoretical intellect' s attaining
peace and quiet, which is itself in view of its thinking about philosophy, which
is itself in view of its attaining conjunction with the Active Intellect, which is
itself in view of its attaining immortality.

28 lbn Sina, at-Tacliqat, 63,3-64,4; and al-Mubal:zathat, 232,15-233,4 (= Mubal:zatha 6, #693) -

the guestion - and 233,12-19 (= Mubal:zatha 6, #694) - Avicenna's response.

29 The expansion of the to hou sense of the final cause to cover causal relations between the

intellect and the soul - in other words, the doctrine that just as the body is min qibali or min ajli the
soul, so the soul is min qibali or min ajli the intellect - can be found in compressed form in the
treatise on the soul by (Ps.-?)Porphyry; see Kutsch 1954, 268,5-6 (Arabic text).
136 Part I: The Ammonian Synthesis

In other words, just as individual animal souls use their vegetative faculty of
reproduction, which in turn uses their bodies, in view of attaining the form of
immortality through the species, so human rational souls use their animal
faculties of locomotion and sensation in view of attaining individual
immortality .30 The only difference is that each human rational soul can attain
the immortality which the Active Intellect possesses and which serves as the
final cause qua to hou of the activities of the theoretical intellect and its lower
faculties. Once the theoretical intellect has attained that immortality, it becomes
identical in species, rather than numerically identical, with the Active Intellect,
thus allowing the theoretical intellect to retain its individuality. At the point
when the theoretical intellect becomes identical in species with the Active
Intellect, the theoretical intellect ceases to be a final cause qua to h6i and
becomes a final cause qua to hou.31
The upshot is that even though vegetative and animal souls may act as
efficient causes and as final causes qua to h6i, moving and using their bodies,
they never act as final causes qua to hou, as goals in view of which their bodies
or lower faculties are employed. On the contrary, the final cause qua to hou of
vegetative and animal souls transcends them: it is the immortality which they
cannot attain as individuals but which they can attain through reproduction.
Human rational souls, by contrast, do contain within themselves a final cause
qua to hou: the theoretical intellect, and particularly the "acquired intellect" (al-
<aqlu l-mustafrul) in view of which all lower faculties of the intellect and of the
soul as a whole are employed. For it is when the theoretical intellect is fully
activated, when it is in the state of acquired intellect and hence in conjunction
with the superlunary Active Intellect, that the goal of individual immortality is
attained. Avicenna's position seems to be that it is when perfection is understood
not as an efficient and formal cause but first and foremost as a final cause - and
further, when perfection is understood not as any final cause but primarily as a
final cause qua to hou - that the transcendence of the human rational soul is best
guaranteed.
How does Avicenna's tying of perfection to the two types of final causality, as
well as to efficient and formal causality, square with his distinction, articulated
in cluster (3a), between first perfections and second perfections? I said previously
that when Avicenna made his distinction between first and second perfections, as
articulated in cluster (3a), he had in mind the Neoplatonic distinction between
procession and reversion. The first perfection, "what is necessary for a thing to
exist", clearly corresponds to the substantial form possessed by a thing

30 lbn Sina, K. an-najiit, Cairo, 1913, 274,3-8; K. ash-shifii' fTabl'iyyiit (6): Fi n-nafs 1.5, 50,13-

15; R. ft 1-kaliim 'ala n-nafs al-nii{iqa, 198,8-12 (reading li-l-muta~arrif for al-Ahwani's Ii+
ta~arruf); K. al-ishiiriit wa-t-tanblhiit, 160,8-9; and at-Ta'lfqiit, 176,8-11.
31 In addition to the passages cited in note 139, see lbn Sina, K. al-ishiiriit wa-t-tanblhiit 3,

125,16-18 and 126,12-17; K. al-hidaya, 302,9-305,7; and R. ft n-nafs wa-baqii'iha wa-ma'iidihii 2,

66,1-13
 6. Avicenna on Perfection and the Soul 137

composed of form and matter. Without the substantial form, the composite
would not exist as an actual, concrete thing. Here the substantial form will be
seen to act on its effect, or explain its explanandum, as a formal cause. The
reason why the substantial form - and hence the formal cause - is tied to
procession is because the inherence of substantial forms in the sublunary world
is the effect of an efficient cause: the Active Intellect, in its role as "Giver of
Forms" (wiihibu ,J-.Juwar). Thus the first perfection, as articulated in cluster (3a),
corresponds to the formal cause, and is itself the effect of an efficient cause.
The second perfection, as articulated in cluster (3a), "what is necessary for a
thing to exist well", clearly refers to the exercise of the capability associated
with that form. Which cause does that correspond to? I would like to say that it
corresponds directly to the final cause qua to hOi. But because, strictly speaking,
a second perfection ought to refer to an activity and not to an active thing, and
because a cause is more likely to be an active thing than an activity, it is
difficult to see how there will be a strict correspondence between the second
perfection and the final cause qua to hOi. 32
In a way this is the old question we saw in Philoponus: does teleiotes refer to
"endedness'', in which case it would refer to an activity (in this case, serving as
an encl); or to "end"; in which case it would refer to an active thing? Because of
the lexical momentum in Arabic tying kamiil and tamiim more closely to "end"
than to "endedness", the problem is perhaps slightly less acute, and the second
perfection, "what is necessary for a thing to exist well", can be understood both
as an active thing and as an activity, and thus be seen to correspond to the final
cause qua to hOi. Without actively striving to attain its goal, a thing would not
be able to exist well. Here the active thing will be seen to act on its effect, or
explain its explanandum, as a final cause qua to hoi, since by being an active
user of what is below it, it succeeds in existing well.
In other words, the reason why the final cause qua to hoi is tied to reversion
now appears to be because the final cause qua to hoi refers to the composite
when it is actively using, and thus benefiting from, something below it in the
cosmic hierarchy, in view of - that is, while striving to attain - something
higher than it in the cosmic hierarchy. Thus the second perfection, as articulated
in cluster (3a), corresponds to the final cause qua to h6i.
What about the final cause qua to hou and the efficient cause? How do they fit
into the scheme of procession and reversion? Answering this question will
require me to qualify my position that there is a direct correspondence between
the first perfection and procession, and the second perfection and reversion.
Instead, I should make the obvious point that procession and reversion are
dynamic rather than static. A single thing cannot be extracted from procession or
reversion and then held to be an element only in procession or only in reversion.

32 pace Lettinck 1994, 167-171. who views the to houlto hOi distinction as the straightforward

equivalent (and antecedent) of Avicenna's distinction between first and second perfections.
138 Part I: The Ammonian Synthesis
A single thing is always playing two roles: one role is as an element of
procession, and the other role is as an element of reversion.
As discussed above, when the substantial form is seen as an element of
procession, it will act on its effect, or explain its explanandum, as a formal
cause. When, on the other hand, the substantial form is seen as an element of
reversion, it will act on its effect, or explain its explanandum, as a final cause
qua to hou. This is because the substantial form acts on its effects not only as a
formal cause, giving actual, concrete existence to the composite in which it
inheres; but also as a final cause qua to hou, serving as the goal in view of
which a lower being exercises the capability associated with its own (lower)
substantial form.
I shall try to explain what I mean. Let Y be higher and X lower. Y's
substantial form will serve in two roles, both as an element of procession and as
an element of reversion. Seen as an element of procession, Y's substantial form
operates as Y's own formal cause: it is by virtue of Y's possession of the
substantial form "Y-ness" that it can be said to exist as a Y. Seen as an element
of reversion, on the other hand, Y's substantial form operates as a final cause
qua to hou, namely, as the final cause qua to hou of X. This is because X,
being in possession of a lower substantial form (viz., "X-ness"), constantly
strives to revert upwards by imitating Y and thereby coming to possess Y's
substantial form (viz., "Y-ness").
Similarly, the efficient cause, like the final cause qua to h6i, corresponds to
the composite when it is in a state of activity. Like the final cause qua to hoi,
an efficient cause can only be a cause when it is active. This is because the
active thing acts on its effect not only as a final cause qua to hoi, using what is
beneath it in the cosmic hierarchy as a means of attaining some higher goal; but
also as an efficient cause, moving what is beneath it in the cosmic hierarchy just
as the soul moves the body, or giving existence to what is beneath it within its
own species just as the father gives existence to the son.
Again, let Y be higher and X lower. Y will once more serve in two roles, both
as an element of procession and as an element of reversion. Seen as an element
of procession, Y operates as an efficient cause, by coming into contact with X,
passing existence down to X and making X so actual and perfect and full of
existence that X comes to possess Y's own substantial form (viz., "Y-ness").
Seen as an element of reversion, on the other hand, Y operates as a final cause
qua to h6i. This is because Y uses X in view of an even higher element in the
reversion - Z - whose substantial form (viz., "Z-ness") in tum operates as the
final cause qua to hou of Y.
With this scheme in mind, we can begin to answer our question, namely, how
does Avicenna's tying of perfection to the two types of final causality, as well as
to efficient and formal causality, square with his distinction, articulated in cluster
(3a), between first perfections and second perfections? The answer is that the
formal cause and the efficient cause are, respectively, the effect of a first
 6. Avicenna on Perfection and the Soul 139

perfection and the first perfection itself. Y's first perfection is its efficient
causality, an efficient causality which causes X's substantial form - X's formal
cause, in other words - to come into existence.
Moreover, the final cause qua to hoi and the final cause qua to hou are,
respectively, the effect of a second perfection and the second perfection itself.
The second perfection for which Y strives is Y's final cause qua to hou, namely,
the substantial form which Z possesses (viz., "Z-ness") and which Y imitates
and strives to attain. Y's final causality qua to hoi - its use of X in order to
imitate and attain "Z-ness'', is therefore the effect of Y's second perfection.
What does all this X, Y and Z-ness tells us about Avicenna's theory of the
separability of the human rational soul? It tells us that Avicenna's conception of
the human rational soul was one which requires us to view the soul as dynamic
rather than static, relative rather than absolute, and complex rather than simple.
In other words, the human rational soul is constantly switching between
activities of one sort or another - thinking, moving, using - each of which
stands in a cause-and-effect relationship with the faculties of the soul below it
and above it.
Simply put, the criterion of activity vs. inactivity will determine whether at
any given moment the soul is separate or not. Avicenna's view, of course, is
that in itself the human rational soul is mufaraqa, which means both separable
as well as separated. In itself, then, the human rational soul is separable; but
the human rational soul is in fact separated only when it is using its lower
faculties, and its lower faculties are using the bodily organs they are associated
with, in order to think. The human rational soul is in fact separated, in other
words, only when it is operating as the final cause qua to hoi of its lower
faculties and of the bodily organs those lower faculties are associated with, in
view of its own, true, higher final cause qua to hou, namely, the immortality of
the Active Intellect.
When it is asleep, however, or when it is using its lower faculties merely to
reproduce or to nourish, the human rational soul is not in fact separated, only
separable; when, in other words, the human rational soul is operating as the final
cause qua to h8i of its lower faculties and of the bodily organs those lower
faculties are associated with, in view of a false, lower final cause qua to hou,
such as sensory gratification. And since sensory gratification is not the true and
real final cause qua to hou of the human rational soul, nor, likewise, will the
human rational soul be truly operating as a final cause qua to hoi of its lower
faculties and of the bodily organs those lower faculties are associated with. The
general rule to be drawn is that whenever the human rational soul is not
operating as a true and real final cause qua to hoi, the human rational soul will
not in fact be separated. This is because it is only when the human rational soul
is operating as a real and true final cause qua to h8i - that is, whenever the
human rational soul is using its lower faculties and the bodily organs those
140 Part /: The Ammonian Synthesis

lower faculties are associated with, in view of its own, true final cause qua to
hou - that the human rational soul will in fact be separated.
Why have I gone into such detail in analyzing how Avicenna's theory of
perfection is related to his ideas about causality, and particularly final causality,
and in explaining what that relationship tells us about his views concerning the
soul's immanence or transcendence? There are two reasons. First of all, some
scholars have presented Avicenna's ideas about the human rational soul's
separability or separatedness as nothing but thinly veiled Platonism. For
example, in an article first published in 1981, Mul;lammad al-Mi~bal;li suggested
that Avicenna's use of the idea of "separable" or "separated" perfection (al-
kamiilu l-mufaraq) represented a dramatic tum away from Aristotle's psychology
and towards Plato's. 33 Other scholars, such as Fazlur Rahman, call Avicenna's
doctrine of the separability or separatedness of the human rational soul "entirely
non-Aristotelian", claiming that a passage from the Enneads should be seen as
its source. 34
But given Avicenna's rejection of the Platonic/Plotinian doctrine of the soul's
pre-existence and descent into the body; given the fact that Aristotle's position
on the soul's, or at least the intellect's, separability or separatedness is more
underdetennined than most modem scholars have allowed; given the radical
conceptual transformation which the concept of perfection underwent as a result
of the activities of Greek commentators and Greco-Arabic translators; and finally,
given Avicenna's inheritance of an increasingly hardened distinction between the
formal and material causes, which are intrinsic to or immanent in their effect,
and the final and efficient causes, which are extrinsic to or transcend their effect;
Avicenna's position on the soul's separability or separatedness should, I believe,
be seen as a sophisticated and justifiable reading of Aristotle by a philosopher
who stands as the culmination of the Ammonian synthesis, rather than as a
symptom of his being in thrall to some caricature of Platonism or
Neoplatonism.

33 al-Misbiihi 1980-81.
14 Rahdtan.1952, 108, citing Enneads IV.7.12. In fact the Arabic version of the Plotinian
chapter immediately following the passage which Rahman cites in support of his claim ( Uthiiliijiyii
I, 18,13-16) expounds the doctrine of the soul's pre-existence and descent from the world of mind
into the body, a position Avicenna is crystal-clear in rejecting. By Avicenna's reckoning the soul
comes into existence (budiith) at the same time as the body. Avicenna does maintain that it is not
the body which is the essential cause of the soul's coming into existence, but rather the superlunary
separate causes; but this in no way commits him to the Platonic/Plotinian position that the soul pre-
exists the body and then descends into the body. For Avicenna' s position that the soul does not pre-
exist the body see, for example, the chapter entitled Fa#un ft ithbiiti budiithi n-nafsi in K. an-najiit
2: Tablciyyiit, 300,2-302,8, esp. 300,6 (ja-mubiilun an takiina qad wujidat [i.e., an-nufos] qabla [.
badani) and 301,12-13 (ja-qad fabba idhan anna n-nafsa tabduth kullamii yabduthi 1-badanu f·
fiilibu li-sticmiilihii iyyiihu); see also K. ash-shifa,/[ablciyyiit (6): Fl n-nafs 5.4, 228,19-229,2. For a
criticism of Avicenna's account of how the body is responsible for the soul's individuation. see
now Druart 2000.
 6. Avicenna on Perfection and the Soul 141

The second reason is that by describing as precisely as I can how Avicenna

thought the soul was related to the body I hope to have shown how crucial an
appreciation of the complex and dynamic cosmology of procession and reversion
is to understanding what Gutas has named Avicenna's "metaphysics of the
human rational soul". The next four chapters explain how Avicenna tries to
integrate that cosmology of procession and return, a cosmology which he
inherited from earlier thinkers of the Ammonian synthesis, with the ontology of
the Muslim doctrinal theologians, or mutakallimun, in order to lay the
foundation for a new, Avicennian synthesis.
 Part II

The Beginnings of the A vicennian Synthesis

 7. Essence and Existence (A)
Materials from the Kalam and al-Farabz

As shown above, when Avicenna distinguishes between first and second

perfections on the basis of the difference between what is necessary for a thing's
existence, and what is necessary for a thing's well-being, he has in mind a
broader distinction, that between the downward procession of existence and the
upward reversion towards well-being. Avicenna, like Proclus and Ammonius,
maintained that the final cause was the Aristotelian cause best suited to
explaining the upward reversion of each thing towards its well-being, and that
the efficient cause was the Aristotelian cause best suited to explaining the
downward procession of existence from one thing to another. I have just shown
how the tension between procession and reversion plays out in the case of
Avicenna's position on the issue of the ontology of the soul. But how are
procession and reversion reflected in Avicenna's metaphysics and cosmology in
general? Specifically, what does linking the final cause to reversion tell us about
Avicenna's metaphysics of essence and existence, and about the ways in which
God is the cause of the world? Chapters 7 through 11 will be devoted to
answering those questions. 1
In an article published in 1984, Jean Jolivet suggested that the origins of
Avicenna's distinction between essence and existence lay not in ancient Greek
philosophy, as has generally been supposed, but in early Islamic doctrinal
theology (kalam), and specifically in the ninth- and tenth-century debates
between Muslim dogmatists (mutakallimun) over how the terms "thing" (shay')
and "existent" (mawjud) relate to each other. 2 In the next four chapters I shall
argue that Jolivet is correct in highlighting the kali'im background to Avicenna's
essence/existence distinction. I shall also argue, however, that an important and
previously overlooked testing-ground for the essence/existence distinction -
Avicenna's analysis of the relation between efficient and final causes - arose
from his efforts to solve a problem of Aristotelian exegesis by appealing to
Neoplatonic notions of procession and reversion.
Chapter 7 is an introduction to the problem. I shall survey early kali'im
discussions of things and existents, as well as al-Farabi's discussions, and
present evidence supporting Jolivet's hypothesis that these kalam discussions
were the backdrop against which Avicenna made his distinction between essence
(miihiyya, literally "whatness", and often Latinized as "quiddity") and existence

1 Chapters 7 through 9 derive from Wisnovsky 2000, though with alterations, some minor, others

significant.
2 Jolivet 1984.
146 Part II: The A vicennian Synthesis

(wujud). 3 While it is easy to imagine a progression from the mutakallimun's

use of mawjud ("existent") to Avicenna's use of wujud ("existence"), the route
from shay' ("thing") to miihiyya ("essence") seems less direct. I explore the
possibility that the concept of "thingness", shay'iyya, served to link the
mutakallimiln's use of shay' and Avicenna's use of miihiyya. In Chapters 8 and
9, I discuss some previously unexamined passages where Avicenna uses the term
shay'iyya to explain how the final cause can be seen to be prior to the efficient
cause. More particularly, Chapter 8 addresses a philological problem: whether or
not these passages should be emended to read sababiyya, "causality", instead of
shay'iyya, "thingness". I argue that shay'iyya should be retained. In Chapter 9,
I discuss what shay'iyya means in the context of Avicenna's attempt to unpack
a rather compressed assertion which Aristotle makes about how efficient and
final causes relate to each other. In so doing I hope to cast some light on how
exactly the concept of thingness bridges the mutakallimiln' s discussions of
things and existents, and Avicenna's distinction between essence and existence.
Chapter IO is devoted to showing how Avicenna's tying together of essence and
final causality on the one hand, and existence and efficient causality on the other
hand, can be better understood in the context of his inheritance of the Ammonian
cosmology of procession and reversion.

Given the opaque and scattered nature of most of the early sources available to
us, we cannot know for sure when and where thinkers writing in Arabic first
used the terms shay' and mawjild. It is fairly certain, however, that the
mutakallimiln' s first discussions of things and existents arose not out of some
general and spontaneous interest in ontology, but out of a desire to marshal
Qur'anic evidence in favor of their own positions and against their opponents'
on two basic theological topics: God's attributes (in this case, whether or not
God could be correctly spoken of as a thing, and if so, how) and God's creative
power (in this case, whether or not it is things which God causes to come into
existence and pass out of existence, and if so, how). In order to buttress their
arguments the mutakallimiln had to address the question of what shay', "thing",
means when it is used in the Qur'ii.n.
The term shay' appears in many places in the Qur'an, and mainly refers, in a
general and undifferentiated sense, to the objects of God's attributes. God is all-
powerful (qadir); over what is God all-powerful? Things, or, more precisely,
every thing. Similarly, things are the objects of God's knowing ('-l-m), creating

3 It is true that mtihiyya, "quiddity", is only one of several terms - including 1,zaqlqa, dhtit, ~ilra,

rabl'a and, as we shall see, shay'iyya - which, taken as a whole, constitute Avicenna's idea of
"essence" (cf. Goichon 1937, 48). As will become clear, mtihiyya is first among equals in this
cluster of terms. Goichon sees dhtit as the essential term for essence, following the lead of its Latin
translation as essentia; miihiyya, by contrast, was rendered as quidditas.
 7. Essence and Existence (A) 147

(kh-l-q), witnessing (sh-h-d), being in charge of (w-k-l), preserving (/:z-f-?-).

encompassing (J:z-y-f), reckoning (J:z-s-b) and supervising (r-q-b). 4
Despite the Qur'an's straightforward, quasi-pronominhl use of shay" (°amr is
used as an all-purpose direct object in much the same way), early commentators
and mutakallimun nevertheless tried to determine what precisely it meant to be a
thing. For example, it seems clear enough that because things are the objects of
God's attributes, they are other than God. The distinction between thing and
God is reinforced by the famous verse 42:11, laysa ka-mithlihi shay", "No
thing is [even] like a likeness of Him". Believers are also instructed not to verge
towards polytheism by associating things with God (3:64; 4:36; 6:151; 12:38;
22:26; 60:12). Finally, the divine act of creation is described in two famous
verses, 16:40 (inna-ma qawluna li-shay"in idha aradnahu an naqula lahu kun
fa-yakunu) and 36:82 (inna-mii amruhu idha arada shay' an an yaqula kun fa-
yakunu), as consisting in God's saying "Be!" to a thing, at which point the
thing then is.
But the strong suggestion that things are other than God weakens slightly
when one takes into account a number of verses that seem to indicate that God
Himself is a thing. For example, 6: 19, "Say: Which thing is the greatest in
terms of witnessing? Say: God is witness between you and me" (qul ayyu
shay'in akbaru shahiidatan qul Allahu shahidun baynf wa-baynaka), implies
that God is a thing, and 28:88, "Every thing will perish save His face" (kullu
shay'in halikun ilia wajhahu), implies that at least the divine countenance is a
thing.
This ambiguity was reinforced by the early Arabic grammarians, who held that
shay' was the most generally applicable of terms (acammu z-camm), applying
to all that may be placed in relation to a predicate (yaqacu caza kulli mil
ukhbira canhu). 5 That is, shay' refers to every mubtada' (subject) in relation
to which one could place a khabar (predicate). If we want to predicate any
attributes of God, therefore, we will be forced to maintain that He is a thing.
This grammatical consideration seems to have overridden the weight of Qur'anic
evidence, and apart from the very early arch-unitarian Jahm ibn ~afwan (d. 746),
who - as we shall see - held to a strictly understood interpretation of thing
(namely, that it was synonymous with makhluq, "created", and that therefore

4 q-d-r: 2:20, 2:26, 2:109, 2:148, 2:259, 2:284, 3:26, 3:29, 3:165, 3:189, 4:85, 5:17, 5:19, 5:40,

5: 120, 6:17, 8:41, 9:39, 11:4, 16:77, 18:45, 22:6, 24:45, 29:20, 30:50, 33:27, 35:1, 41 :39, 42:9, 46:33,
48:21, 57:2, 59:6, 64:1, 65:12; '"-1-m: 2:29, 2:23 1, 2:282, 4:32, 4:176, 5:97, 6:80, 6:101, 7:89, 8:75,
9: 115, 20:98, 21 :8 1, 24:35, 24:64, 29:42, 29:62, 33:40, 33:54, 40:7, 42:12, 48:26, 49:16, 57:3, 58:7,
64: 11; kh-1-q: 6:101, 6:102, 13:16, 16:48, 20:50, 25:2, 27:88, 32:7, 39:62, 40:62, 51:49, 54:49; sh-h-
d: 4:33, 5:117, 22:17, 33:55, 34:47, 41:53, 58:6 (and implied in 3:5, 14:38, 40:16); w-k-1: 6:102,
11: 12, 39:62; f;-f-~: 11 :57, 34:21; fJ-y-r: 4: 126, 41 :54; fJ-s-b: 4:86; and r-q-b: 33:52.
5 Sibawayhi (d. ca. 7%), al-Kitiib !, 47,7; and ap. az-Zabidi, Tiij al- 'ariis min jawiihir al-qiimiis

!, 18Sa20-21.27-28. The formulation ash-shay'u huwa mii yajiizu an yukhbara 'anhu is also found
in al-Khwarazmi (d. 997), Ma.fiitifJ al-'ulii.m, 22,14-15; in that same work al-Khwarazmi reports
(199,14-200,1) that shay' is also used by the "Algebrists" as a kind of universal variable.
148 Part /I: The A vicennian Synthesis

God was not a thing), consensus arose around the fence-straddling assertion that
God was a shay'un Iii k0;-l-ashyii'i, "a thing not like [other] things". 6
But whether or not it is right to call God a thing was not the only problem
which the term shay' forced the early mutakallimun to confront. More
seriously, the mutakallimun had to contend with the implication of verses 16:40
and 36:82 (mentioned above) that things were somehow there before God said
"Be!" to them; for otherwise, what would God be saying "Be!" to?7 It seemed
perfectly sensible to draw the conclusion, as most Mu<tazilites did, that thing
applies not only to what exists (al-mawjud) but also to what does not exist (al-
ma cdum); and that what does not exist in turn applies not only to what did not
exist and now exists, such as the world, but also to what does not now exist but
will exist, such as the Day of Resurrection.
Other mutakallimun, by contrast, held that shay' applies solely to what
exists. This was partly because the Mu<tazilite view could be seen as leading to
two unsavory alternatives. The first arises from the premise that even before it
exists one particular thing - my great-great-grandson - is and always has been
distinct from another particular thing - my great-great-granddaughter. In 68 years
God will say "Be!" to one non-existent thing, my great-great-grandson, and in 72
years God will say "Be!" to another non-existent thing, my great-great-
granddaughter. The same premise can be applied to things that have existed but
now do not. Thus my great-great grandfather, even though he is non-existent
now, remains one thing, and my great-great-grandmother remains another thing,
and both will remain distinct things up to, including and (presumably) beyond
the Day of Judgment. The problem is that all these pre- and post-existent
individual things must be or subsist somewhere before and after they exist. Let
us say they are all located somewhere outside God's mind; in that case they will
share the attribute of eternality with God, and the precept of tawflld, understood
as divine uniqueness, will be violated. Let us then say that they are all located
within God's mind; in that case they will introduce multiplicity into God, and
the precept of tawflld, understood as divine simplicity, will be violated.
Alternatively, and just as problematically, the Mu<tazilite conception could be
understood as implying that before and after they exist, things were and will be
undifferentiated, just one great eternal blob of Thing. In other words, when
things are non-existent, they are undifferentiated; when things are existent, they
are distinct one from the other. But this conception would expose the
Mu<tazilites to the charge that an eternal and undifferentiated non-existent Thing
was, for all intents and purposes, equivalent to the eternal and (in itself)

6 (Ps.-)Abu l::lanifa (ca. 950), al-Fiqh al-akbar (= Mulla ' Ali b. Sulian Mul).ammad al-Qari,

Sharf; al-fiqh al-akbar), 16,18.

7 The term amr is used synonymously with shay' to refer to the things which God is saying

"Be!" to, at 2:117, 3:47, 19:35, 36:82 and 40:68. In his Ta/sir (ad 2:117) ai-Tabari (d. 923)
describes the positions of various "interpreters" (mu'awwilun) on this issue (Ta/sir at-Tabari II,
542,1-550,7).
 7. Essence and Existence (A) 149
undifferentiated prime matter of the Eternalists (ad-dahriyya). God's sole
possession of etemality will be infringed upon, and the precept of tawJ:iid,
understood as divine uniqueness, will once again be violated. 8
To some extent the arguments just articulated are conjecture, for as I
mentioned above, contemporaneous evidence of early kaliim discussions of
things and existents is hard to come by. Al-Ashcan's (d. 935) Maqiiliit al-
isliimiyyln contains a few typically terse descriptions of the views held by early
mutakallimun:

(1) a thing is something created which has a likeness (Jahm ibn ~afwan and
various Zaydi Shicites);
(2) a thing is a body (Anthropomorphists - al-mushabbiha);
(3) a thing is what can be pointed to (Hisham al-Fuwali [d. ca. 842]);
(4) a thing is a thing before it comes into being (°Abbad b. Sulayman [d. 864);
certain Baghdadi Muctazilites); saying "Xis a thing" is the same as affirming
X; affirming X covers the time before X came into being as well as the time
after it came into being (al-Khayyat [d. ca. 913]); a thing is all that can be
known, and all that can be called to mind and of which predication is possible;
things are knowable as things before they come into being. Things can be
called things before they come into being (al-Jubba 0 i [d. 915]);
(5) a thing is an existent (al-Ashcan); a thing is a thing only when it exists
(Abu 1-I:Iusayn ~-~aliJ.ti; Ibn ar-Rawandi [d. ca. 950]). 9

From the end of the ninth century to the end of the tenth - the period, that is,
just before Avicenna started writing philosophy - the mutakallimun holding
positions l, 2 and 3 became increasingly isolated, and the debate about things
polarized into positions 4 and 5, with most Muctazilites holding position 4 and
virtually everyone else -Ashcarites and Maturidites in particular, but also some
Shicites - holding position 5. 10 With God's uniqueness and simplicity at stake,

8 On this see Wolfson 1946, reprised in Wolfson 1976, 359-372. The world-view of the
eternalists was shaped partly by Aristotelian notions of potentiality (quwwa) and actuality (ji</) then
coming into circulation. See, for example, al-Kindi (d. ca. 870), R. ilii Aflmad b. al-Mu<tafim ft 1-
ibiina 'an sujitd al-jirm al-aqfii wa-fii'atihi li-lliih 'au.a wa-jalla, 187,3-19; as well as the evidence
preserved in the Jiibirean corpus (ca. 900), where one can detect the infiltration of notions of
potentiality and actuality into the shay'lmawjudlma'dum paradigm: K. ikhriij miifi 1-quwwa ilii 1-fi'I,
2,14-3,1; K. al-khawau al-kabir, 357,1-2; K. al-miziin Qf·faghir, 427,3-5; 431,8-11; 445,13-14;
451,18.
9 See al-Ash'ari, Maqlillit al-isliimiyyin I-II, 158,1-163,8; 181,1-182,4; and 518,4-520,8.
Perhaps our earliest source regarding the debate over things and existents is the Zaydi Shi'ite al-
Qiisim ibn Ibrahim (d. 860), K. ad-dalil al-kabir, 74,10-76,7.
' 0 Mu'tazilites: al-Khayyii\, K. al-instiflir, 60,4-14; 107,6-108,8; 126,1-2; Sa'diyii al-Fayyiimi
(i.e., Saadia Gaon, d. 942), K. al-amiinlit wa-1-i'tiqiidiit, 11,10-12 and 213,13-15; ' Abd al-Jabbiir
(d. 1025), al-Mughni IV 247,12-13; V 202,8-203,9; 249,4-8; 251,3-4; 252,4-6; VIII 79,17-80,15;
V/12 75,5-77,7 and 135,3-136,9; VJ// 74,1-82,12; XI/ 20,1-3 and 48,5-6. Ash'arites: Ibn Fiirak (d.
1015), Mujarrad maqiillit al-Ash'ari, 42,7-18 and 252,4-256,22; and al-Biiqilliini (d. 1013), K. at-
tamhid, 193,17-194,6; 195,6-196,12; 266,4-9. (In fact, al-Ash'ari admitted in his K. al-'amad that
150 Part II: The A vicennian Synthesis

not to mention the nature of His causation of the world, the original dogmatic
formulae devised to encapsulate these two positions grew into fully articulated
school doctrines. The kalam discussion of things and existents came to
encompass general questions of ontology, and the metaphysical notions used in
the debate became more sophisticated.
This increase in complexity went hand in hand with the intensifying
philosophical activity of the period, activity that involved translating Greek
philosophical texts into Arabic as well as composing original philosophical
works in Arabic. In his Kitiib ft l-falsafat al-Ula, al-Kindi uses shay" in a brief
discussion of essence, existents and non-existents, but for the latter three terms
he uses dhiit, ays and lays respectively.' 1 Al-Farabi was the first faylasuf to
incorporate the term miihiyya ("whatness", "quiddity" or simply "essence") - a
term deriving primarily from the rendering of the Greek to ti en einai as mii
huwa in the Arabic translations of Aristotle's logical works - into the kaliim
problematic of things and existents. 12
AI-Farabi appears at first glance to come down fully on the side of the
Mu<tazilites on the issue of whether or not thing was a more overarching
category than existent, but closer examination reveals that his view is more

he himself had originally held the Mu'tazilite position: "We wrote a treatise on the topic of 'thing',
to the effect that things are things even if they be non-existent. We have retreated from it [i.e., the
position articulated in this book] and [now] contradict it, so whoever comes across it [the book], let
him not place any store by it"; reported by Ibn 'Asakir, Tabyin kadhib al-muftari fl-ma nusiba ila 1-
/mam Abi l-ljasan al-Ash'arl, 133,5-6.) Matutidites: al-Miitutidi (d. ca. 944), K. at-tawl}ld, 39,20-
43,21; 86,2-92,20; 104,8-106,!8; 242,3-16; and AbU 1-Layth as-Samarqandi (d. ca. 990), Sharl; al-
fiqh al-absa! li-Abl lfanifa, 119,1-2 (=lines 426-7). Shi'ites: ash-Shaykh al-Mufid (d. 1022), Awa' il
al-maqalat ft 1-madhahib wa-1-mukhtarat, 42,14-19. See also Frank 1980 and now Frank 1999;
Dhanani 1994, 27 n.4 and 29-30; McDermott 1978, 196-9; Klein-Franke 1994; Gimaret 1988, 142-
150; and van Ess 199!-!997, / 357f.; // 48lf. and 499-501; IV 45-49 and 432-435.
11 al-Kindl, K.fi 1-falsafat al-Ula, 41,3-43,18.
12 It is true that to ti en einai (and its variants) was translated as mahiyya only rarely, and

primarily in the context of distinguishing definition (al-J;add) from property (al-khiiffa): the former
is a statement that "indicates the thing's essence" (ad-diillu 'ala mahiyyati sh-shay'i), the latter one
that does not. See, for example, Top. !O!b38-102al (= Tublka, 474,17); 102a21 (= 475,14);
!03bl0-l 1 (= 481,11); and 154a30 (= 685,12-13). In the Organon to ti en einai was mostly
translated as ma huwa. See, for example, An. Post. 83a21 (= Tal;lilat thiiniyya, 375,6); 83b5 ( =
376,14); 84a26 (= 380,8); 97a25 (= 448,5) and 97b2 (= 449,4); !02a32-33 (= 476,12);
103b21.30.34 (= 482,4.12.15); 120b21(=552,15); 122a5-6.12-19.2!.32.36 (= 557,18, 558,6-11.13
and 559,4.8.9.10.ll[bis]); 128al4.19 (= 581,15 and 582,1) and 148al (= 656,8). It is also
translated as ma huwa sh-shay': An. Post. 79a25.27 .29 (= Tal;liliit thiiniyya, 354,2.4.5) and Top.
10!b22 (= Tubika, 474,l); and as ma sh-shay': An. Post. 82b37 (= Tal;lfliit thiiniyya, 373,15); and
Top. !O!b21 (= Tubika, 473,17); Top. 103b24.26 (= Tubika, 482,8.9) and 148al (= 656,7). In the
Metaphysics, to ti en einai was translated as ma huwa at 994bl 7 (= Ma ba'da f-fabi'a /, 34,8) and
993a!8 (= /, 161,13); as ma hiya and ma sh-shay' at 1016a33-34 (= II, 536,5-6); as ma huwa sh-
shay' at 994al 1 (= /, 17,3); as ma anniyyatu sh-shay' at !024b29 (= II, 685,1); and as ma
kaynunatu sh-shay' at 1013b22 (= //, 487,14). Mostly, however, to ti en einai in the Metaphysics
was translated as ma huwa bi-l-anniyya: at 1025b29-29 (= II 705,8); 1030a6 (= II 795,10);
!030al7 (= II 798,6); 1030a29 (= II 802,6); 103la9-10 (= II 818,10); 103lal8 (= II 821,16);
103lb30 (=II 831,7); !033b7 (= II 860,15-16); 1035bl6 (=II 903,8); 1037a21 (= II 936,8-9);
!037a33-bl (= II 939,15); 1044a36 (= II 1074,l); !045b3 (= II 1096,13); and !075a2 (= II
1693,1).
 7. Essence and Existence (A) 151

nuanced. Al-Farabi first establishes that his definition of mawjud as "that which,
outside the soul, is set apart by some essence, be it conceived of or not" (mii
huwa munf:ziizun bi-miihiyyatin ma khiirija n-nafsi tu~uwwirat aw lam
tuta~awwar), is broader and hence more basic than the other definitions of
mawjud he suggestsY Al-Farabi defines shay', on the other hand, as "all that
possesses some essence, in whichever way, be it outside the soul or conceived of
in any sense whatsoever, divided up or undivided" (kullu mii lahu miihiyyatun
mii kayfa kiina khiirija n-nafsi aw kiina muta~awwaran calii ayyi jihatin kiina
munqasimatan aw ghayra munqasimatin). 14 Thus defined, mawjud is not as
broadly applicable as shay', because shay' covers essences both outside and
inside the mind, whereas mawjud covers merely those outside the mind. In this
sense al-Farabi appears to be echoing the Mu 0 tazilite position.
On the other hand, al-Farabi again follows the Arabic translations of the
Organon and argues that mawjud has one technical use that shay' does not: as
the copula (riibif) that connects the subject (here the logical term al-maw{iuc as
opposed to the grammatical term al-mubtada') to the predicate (al-maf:zmul as
opposed to al-khabar) in categorical affirmative propositions (Ji l-aqiiwfli l-
jiizimati l-mujibati). 15 For example, al-Farabi says, when we wish to state that
Zayd is a just person, it might make sense to say "Zayd is existent as a just
person" (Zaydun mawjudun cadilan), while it makes no sense to say "Zayd is
thing as a just person" (Zaydun shay'un cadilan). In this technical, copulative
sense, mawjud is more broadly applicable than shay' , although al-Parabi
stresses that the first, non-copulative sense of mawjud is more basic. As he did
with miihiyya, al-Farabi has tried to incorporate an Aristotelian notion - here the
copulative or predicative sense of to einai, "to be" - into the kaliim problematic
of shay' vs. mawjud. 16 To sum up al-Fara.bi's position, then, shay' is more
broadly applicable than mawjud in the first, more basic sense, yet less broadly
applicable than mawjud in the second, more technical sense. 17
The kaliim discussions of things and existents and al-Farabi' s efforts to
Aristotelianize them provide the backdrop for an important discussion in
Iliihiyyiit 1.5 of Avicenna's Kitiib ash-shifii', where Avicenna takes some
important steps towards constructing a distinction between essence and

13 al-Farlibi, K. al-f:1uruf. 116,23-117,1.

14 al-Farabi, K. al-1.iurof. 128,6-8.
15 al-Farlibi, K. al-huruf. 125,12-13.
16 al-Farabi, K. al-~uruf. 128,12-13. In Metaph. 4.2, 1003b!O (= Ma ba'da !-{abi'a /, 301,9-10)

Aristotle claims that we can say that even something that is non-existent (to me on = alladhi laysa
bi-huwiyyatin) in some sense is non-existent (einai me on= fa-innahu laysa huwiyyatan). Generally
speaking, to on in the Metaphysics was translated into Arabic not as al-mawjud but as al-huwiyya
("being"). Al-Flirlibi's analysis is discussed by Jolivet 1984, 17; Abed 1991, 111-117; Zimmermann
1981, xliv-xlv, lx-lxiii and cxxx-cxxxiv; and Shehadi 1982, 45-69.
17 In his Introduction (to Logic, that is), al-Farlibi is neutral about which term is more broadly

applicable, saying simply that "Thing and existent are the most generally applicable [terms] by
which an individual or species may be characterized (ja-innii sh-shay'a wa-1-mawjuda humii
a<ammu mii yumkinu an yu~afa bihi shakh~un aw naw'un)": K. isiigha;7 ay al-madkhal, 122,7-8.
152 Part II: The Avicennian Synthesis

existence. 18 In /lahiyyii.t 1.5 Avicenna attempts to make the following points.

First (29,5-31,2), he argues that thing and existent are primitive, basic and
immediately apprehensible concepts. Because there are no terms more broadly
applicable than thing and existent, they are indefinable, in the Aristotelian sense
of definition at least; that is, there is no genus under which they can be
subsumed as species. Second (31,2-32,3), Avicenna shows that thing and
existent have different meanings. "Thing" is associated with terms such as "inner
reality" (J:iaqfqa) and "whatness" (mahiyya), and refers to an entity viewed in
light of essence; "existent" is associated with terms such as "affirmed" (muthbat)
and "realized" (muJ:ia~~al), and refers to an entity viewed in light of existence.
Third (32,3-34, 14), Avicenna explains that although they have different
meanings, thing and existent are co-implied (mutalazimiini). In this context,
Avicenna's purpose in describing shay' and mawjud as co-implied is to convey
the idea that although they have different meanings, neither shay' nor mawjud is
more broadly applicable than the other: the domains of objects to which each
term refers fully overlap. The reflexivity inherent in the sixth-form term talazum
has important connotations as far as Avicenna's distinction between essence and
existence is concerned. For at least in this context, Avicenna is saying that
neither shay' nor mawjud is logically prior to the other. There is no hint here
that being an existent is somehow subordinate to, or an accident of, being a
thing.
To sum up, then, the debate in the tenth century came to tum on how things
and existents relate to each other both extensionally (that is, whether or not the
domain of things overlaps with the domain of existents) and intensionally (that
is, whether or not thing and existent have the same meaning). Logically
speaking, these were the options:

(1) thing and existent are mutually exclusive both extensionally and
intensionally: "things are never existents and existents are never things; to be
a thing and to be an existent have different meanings" (no one from this period
seems to have held this view);
(2) thing is subsumed extensionally but not intensionally under existent:
"things are always existents but existents are not always things; to be a thing
and to be an existent have different meanings" (no one from this period seems
to have held this view);
(3a) existent is subsumed extensionally but not intensionally under thing:
"existents are always things but things are not always existents; to be a thing
and to be an existent have different meanings" (most Mu<tazilites)
(3b) existent is subsumed extensionally but not intensionally under thing*:
"existents are always things but things are not always existents (*although

18 Ibn Sinii, K. ash-shifii'llliihiyyiit (1), 29,5-34,14.

 7. Essence and Existence (A) 153

mawjud can be used as a copula whereas shay' cannot); to be a thing and to

be an existent have different meanings" (al-Farabi);
(4a) thing and existent are identical both extensionally and intensionally:
"things are always existents, and existents are always things; to be a thing and
to be an existent have the same meaning" (Ashcarites and Maturidites);
(4b) thing and existent are identical extensionally but different intensionally:
"things are always existents, and existents are always things; to be a thing and
to be an existent have different meanings" (Avicenna).

The debt Avicenna owes to the kaliim discussions of things and existents seems
self-evident, given the similarity of his position to that of the Ashcarites and
Maturidites. In this regard Jolivet is surely correct. But more precisely,
Avicenna's position in lliihiyyiit l.5 is a compromise between the Ashcarites'
and Maturidites' position on the one hand, and the Muctazilites' and al-Farabi's
position on the other. For although Avicenna advocates the extensional identity
of thing and existent just as the Ashcarites and Maturidites do, he also advocates
the intensional difference between thing and existent just as the Muctazilites and
al-Farabi do.
In any case, Joli vet reckons that the influence of the mutakallimun was greater
than that of the Aristotelian philosophers because no Greek antecedent can be
found for shay' (pragma is the possibility Jolivet explores and rejects), in
contrast to mawjud, whose roots lie in the Greek term to on (and similarly ta
onta for mawjudiit, and to einai for wujud). 19 Although Jolivet is correct in
asserting that to on is the Greek equivalent of mawjud, and that pragma is not
the Greek equivalent of shay', the fact remains that a Greek antecedent for shay'
does exist: ti ("something"). In fact, the question of the ontological status of the
non-existent - to me on - and its relation to the something - to ti - has an
ancient pedigree. The problem originates with some cryptic lines of Parmenides,
is discussed at length by Plato in the Sophist, is analyzed by Aristotle in the
Metaphysics and the Physics, and emerges as a coherent ontology among the
Stoics, who - like the Muctazilites and al-Farabi - put "something" (ti) at the
top of their ontological pyramid, above even "existent" (to on). 20

19 Jolivet 1984, 15-16. On this see also Hadot 1980.

20 Parrnenides, fr. VI, lines 1-2 (esti gar einai, meden d'ouk estin - "It is [possible] to be, but
there is no nothing") and VII,! (ou gar mepote touto damei einai me eonta - "That there are non-
existent things will never be proven") (In fact, Parrnenides comes across as a proto-Ash'arite
when al-Fiirlibi - perhaps distilling Aristotle's refutation of Parmenides at Phys. 1.3, !86a22-bl2
[= af-Tabi'a, 21,1-14; 23,15-25,2 and 26,1-4], or, as appears more likely, simply lifting from Abii
' Ali ibn as-Saml).'s commentary on that passage [= af-Tabi'a, 22, l-23,14; 25,3-22 and 26,5-8; see
particularly wa-mii laysa bi-mawjudin laysa bi-shay,in at 22,2) - quotes Parmenides as asserting
that "All that is other than existent will not be a thing" - wa-kullu mii huwa ghayru mawjiidin fa-
laysa bi-shay,in: al-Farabi, K. al-J:zurilf, 128,19); Plato, Sophist 236E onwards; Aristotle, Metaph.
4.2, 1003bl0 (=Ma ba'da f-fabl'a /, 301,9-10) (At Metaph. 14.2, 1089a5-28, Aristotle argues that
Plato's conception of to me on as "the false" - to pseudos - is useless in explaining coming-to-be
and passing-away, and proposes that a better conception of to me on is "the potential" - to kata
154 Part II: The Avicennian Synthesis

It would be tempting to assert a Stoic link here, but tracing Stoic influence on
Arabic kalam and falsafa has always been a tricky business. Many scholars have
tried, with varying degrees of success. 21 Less uncertain is the influence Stoicism
had on the Greek Aristotle-commentators, whom the Arabic philosophers in tum
read carefully. For example, in his commentary on Book 4, Chapter 1 of
Aristotle's Topics, Alexander of Aphrodisias criticizes the Stoics for elevating
"something" above "existent":

L34
Alexander, in Top. 4.1(ad121a!O), 301,19-25
Here is a way you can demonstrate how wrong the Stoics are in
holding that "something" is the genus under which "existent" is
subsumed: if it is a something, it will clearly be an existent as
well; and if it is an existent it will be definable as existent. Now
they wriggle out of this dilemma by holding that "existent" is said
of bodies alone, and on this basis they speak about "something"
being a higher genus than it, given that it ["something"] is
predicable of incorporeal as well as corporeal entities.

Now any attempt to use this particular passage as evidence of a direct filiation
between Stoic and Mu<tazilite ontology will be frustrated by the fact that only
Alexander's comments on part of Book 1 and Books 5-8 of the Topics survive,
at least according to lbn an-Nadim. 22 But elsewhere in Alexander's commentary
on the Topics, as well as in Simplicius' commentary on the Categories, the
Stoic ontology is articulated, although less starkly than in the in Top. 4.1
passage just translated. 23 To sum up: Jolivet is correct in asserting that the
kalam discussions of things and existents were likely to have been the most
important (and certainly the most immediate) influences on Avicenna, but wrong
in reasoning that this was because ancient philosophers had no such discussions.

Having briefly dealt with the question of Greek antecedents, I can focus on the
question of how exactly the kaliim discussions of things and existents influenced
Avicenna's distinction between essence and existence. For while it is easy
enough to imagine his making the move from mawjud to wujud, less obvious
is Avicenna's progression from shay' to mahiyya, the term he uses most
consistently to describe essence when contrasting it with existence. To some

dunamin- a position followed explicitly by al-Farabi, K. al-/:iurilf, 120,8-121,6 and 123,1-124,4; see
also Aristotle's distinction between absolute and incidental non-existence in Phys. 1.8 and 1.9,
19la23-192b7 [= af-Tabi<a, 66,1-76,4]). On the Stoic ontology, and on the way in which
incorporeals are said to "subsist" rather than "exist", see Hadot 1969; Graeser 1971 ; Goldschmidt
1972; Pasquino 1978; Brunschwig 1988 and 1990; and now Caston 1999.
21 e.g., Pines 1936, 116-117; van Ess 1966, 191-200; Jadaane 1968, 43-98; and Wolfson 1976,

355-372; these are reviewed briefly in Gutas 1994, 4959-4962.

22 Ibn an-Nadim, K. al-jihrist, 249,17-19.
23 Alexander in Top. 1.5 (ad 101b38) 42,13-25, and 4.6 (ad 127a26), 359,12-16; Simplicius in

Cat. 5 (ad 3b!0-23), 105,7-20, and 8 (ad 8b26), 222,30-33.

 7. Essence and Existence (A) 155
extent Avicenna's progression from shay' to miihiyya can be reconstructed by
analyzing the /liihiyyiit 1.5 passage devoted to showing how thing and existent
have different meanings.
In that passage Avicenna starts by asserting that every thing (shay') or entity
(amr) has an inner reality (baqiqa) by which it is what it is (31,5-6). This inner
reality (e.g., triangularity) is sometimes called "existence that is specific" (al-
wujud al-khii~~); specific, that is, to one class of things (triangles) as opposed to
another class of things (cats). Existence that is specific is distinct from the more
general type of existence which Avicenna calls "affirmative existence" (al-wujud
al-ithbiit'i) (31,7-8). To predicate affirmative existence of an entity is to assert
that the entity is, not what the entity is. To predicate existence that is specific,
on the other hand, is to assert what the entity is, not that the entity is. Since
existence that is specific is identical to inner reality, and since inner reality is
identical to whatness (miihiyya), it follows that existence that is specific is
identical to whatness (31, 10). Since existence that is specific is identical to
whatness, and since existence that is specific is distinct from affirmative
existence, it follows that whatness is distinct from affirmative existence. In other
words, essence is distinct from existence.24
A brief remark later on in Iliihiyyiit 1.5 raises the intriguing possibility that
shay' iyya, thingness, served to link Avicenna's distinction between things and
existents on the one hand and his distinction between essence and existence on
the other:

L35
Ibn Sina, Kitab ash-shifa'!lliihiyyiit (1) 1.5, 33, 16-18
Avicenna, Uber de philosophia prima sive scientia divina I-IV,
38,20-23
According to people who hold this view [that what is predicated
can be a non-existent, i.e., the Mu<tazilites], in and among all that
is predicated and known are some entities ['umuran = res] that,
when non-existent, possess no thingness [la shay'iyyata lahli =
non habent similitudinem]; Jet the person who is inclined to agree
with that go back to whatever dogmatic formulae they babbled out
unintelligibly, [formulae] which do not deserve any attention. 25

In this passage Avicenna is attacking those who maintain that there is a class of
non-existent entities that possess no thingness, and hence are not things. A later
commentator on the /liihiyyiit, Muna Mahdi Nariiql (d. 1764), advises us that
Avicenna's target here was a group of Mu<tazilites who, in maintaining that
some non-existents (viz., those that are impossible, al-mustablliit or al-

2 4 Avicenna's discussion of things and existents in Iliihiyyiit 1.5 is related to similar treatments in

A"iuinas' work in Craemer-Ruegenberg 1991.

s I shall discuss in Chapter 8 the various mistranslations of shay'iyya into Latin.
156 Part II: The Avicennian Synthesis

mumtani<tit) had no thingness and hence were not things, deviated from the
classical doctrine of their school. 26
Despite Avicenna's use of the term shay'iyya here, any attempt to prove that
the term served to bridge the ninth- and tenth-century kaliim discussions of
things and existents on the one hand, and Avicenna's discussions of essence and
existence on the other, will face two difficult challenges. The first challenge is
that shay'iyya would play a bridging role more convincingly if that term had
appeared often and in a wide variety of ninth- and tenth-century kaliim texts. It
does not. In fact, I have not seen a single instance of shay'iyya where I most
expected to find it: in a text written by a Muctazilite, a member of the school
most infamous for affirming the thingness of the non-existent. Nor, to my
knowledge, does shay>iyya appear in an Ashcarite text until al-Juwayni (d
1085), that is, not until a generation after Avicenna's death. 27
But shay'iyya does make a few appearances in the tenth century. It turns up
once (in a passage attributed to Jacfar a~-$iidiq (d. 765) commenting on the credal
formula shay'un lii ka-l-ashyii,i, "[God is] a thing not like [other] things") in
the Kiifi of the Shi<ite traditionist Abu Jacfar al-Kulayni (d. 940). 28 It is only in
the Kitiib at-tawbfd of al-Matwidi (d. 944) that shay'iyya is also used
consistently to describe and attack the doctrine of the Muctazilites. 29 Later in the
tenth century, al-Maturidi's use of shay'iyya in the context of anti-Muctazilite
polemics was echoed in the Shar/:t al-fiqh al-absa{ li-Abf f:lanifa of the l:lanafite
scholar Abu 1-Layth as-Samarqandi (d. 990). 30
Are these appearances of shay'iyya the first uses of the term, or are they
evidence of earlier Muctazilite use? At this stage of research it is difficult to tell.

26 Mulla Mahdi Naraqi, Shar~ al-iliihiyyiit min kitiib ash-shifii>, 265,2. Marmura l 984a (revised

slightly in Marmura 1991-92) argues that what troubled Avicenna specifically was the Mu<tazilite
Abii Hashim's and his followers' application of their shay'lmawjudlma'dum paradigm to the
problem of bodily resurrection. Also see Black 1997 for a discussion of the dilemmas that faced
Avicenna in distinguishing between fictional entities such as phoenixes, which have some kind of
mental existence even though they never exist as concrete objects in the outside world, and
im~ssibilities such as a square circle, which have neither mental nor concrete existence.
7 Strangely enough, Juwayni devoted a long section of his Shiimil to reviewing the debate about

things and existents, yet does not once use the term shay'iyya there: ash-Shiimil ft U$iil ad-dfn,
23,18-33,24; but see his use of the term in other contexts, at 113,2.4; 161,25.26.27; and 169,6.17.
Ash-Shahrastani (d. 1153) uses shay'iyya in several places: K. nihiiyat al-iqdiim ft 'ilm al-kaliim,
33,20; 150,7.8.10; and 151,2.7. For an argument that ash-Shahrastani is more properly described as
an lsma:cm, rather than an Ash'arite, see Steigerwald 1997. Al-Ghazali (d. 1111) himself mentions
the Mu'tazilite view without using shay'iyya: Tahiifut al-faliisifa, 358,8-11.
28 Abii Ja<far Mul,Jammad b. Ya'qiib al-Kulayni, al-U$iil min al-kafi, 82,4-85,9 at 83,6-8. In Ibn
Babawayh's slightly later collection of Imami Shi'ite sayings, shay' iyya is also used in the context
of affirming that God is a thing unlike other things, except this time in a passage attributed to Ja<far
a~-Sadiq's son, Miisa al-Kiii:im: Abii Ja'far Mul,Jarnrnad ibn Babawayh a~-Sadiiq, K. at-taw~id,
94,9-95,1.
29 al-Maturidi, K. at-tawl}id, 41,IO(bis).15; 86,3.16; 104,10.11.12.15; 132,19; 238,9.10;
242,8.9.l I.12.13.14(bis).18. See also the brief discussion of al-Maturidi's use of shay'iyya in
Gimaret 1980, 182-184 and 203.
lo Abii 1-Layth as-Samarqandi, Shar~ al-fiqh al-absa! li-Abi l:fanifa, 119, l.2 (=lines 426-7).
 7. Essence and Existence (A) 157

With such a tiny proportion of what was written by ninth- and tenth-century
mutakallimun now available to us, assertions about the term's history will
inevitably suffer from tentativeness. But the fact that the term seems to be
absent from the few Mu 0 tazilite and Ash 0 arite sources we do have from this pre-
A vicennian period indicates that al-Mlituridi might well deserve the honor of
being called Abu Shay'iyya. 31 What is interesting here is that because Avicenna
grew up outside Bukhara - that is, in Transoxania, an area where the Samarqandi
l:lanafism of al-Miituridi was strong - and because he was taught jurisprudence
by a l:lanafite scholar called Ismii 0 Il az-Ziihid, Avicenna may well have come
across the term shay'iyya during his youthful fiqh studies, if not in al-
Miituridi' s Kitiib at-tawl:zid, then perhaps in Ahli 1-Layth as-Samarqandi's
commentary on Abu l:lanifa's al-Fiqh al-absaf. Even if Ismacn az-Ziihid himself
was not a Maturidite in kaliim (Bukhiiran l:lanafites tended to be less rationalistic
and speculative in kaliim than those in Samarqand, and in any case there were
l:lanafites such as Ibn Furak who were Ash 0 arite in kaliim), it is perfectly
possible that Ismii 0 Il could have used a simple commentary such as Abu 1-
Layth's when teaching his young student the l:lanafite creed. 32
Of course it is also perfectly possible that Avicenna came up with the term
shay'iyya all by himself. Others have remarked on al-Maturidi's predilection for
abstract nouns ending in -iyya, an extreme example being his grafting of the
Arabic -iyya (equivalent to the English suffixes "-ness" or the more Latin-
sounding "-ity") onto the Persian hast ("is") to form hastiyya ("is-ness"). 33 Al-
Maturidi' s predilection for abstract nouns was a trait Avicenna certainly shared.
But given the presence of the term shay'iyya in a small number of tenth-century
kaliim texts as well as in Avicenna's lliihiyyiit 1.5, there is a possibility -
nothing more - that the term played a bridging role between the earlier kaliim

31 My own guess is that the Mu'tazilite al-Ka'bi al-Balkhi (d. 931), a student of al-Khayya\ and

of al-Jubba'i, and an active participant in debates in Khurli.san and Transoxania during this period,
may well be the missing link, since he is cited (and attacked) so extensively by his contemporary
al-Maturidi. But until al-Ka'bi al-Balkhi's Maqiiliit is fully edited - at the moment it remains in
manuscript in Yemen, apparently - this must remain only a tentative suggestion.
32 On Avicenna's I:Ianafite education, see Gutas 1987-1988. On Isma'il az-Zli.hid, see Ibn Sina,

Autobiography, 20,4-7, and al-Kha!ib al-Baghdadi, Tiirikh Baghdad 6, 310,17-311,18 (= #3355).

On Maturidism in Transoxania, see Madelung 1971 and 1991, and now Rudolph 1996. Other hints
of Avicenna's I:Ianafite background can be detected in the fragments from Avicenna's Lisiin al-
'arab, where Abii l:lanifa is mentioned by name and quoted at length in a discussion of the
particulars of prayer (Risiila ft 1-lugha, 15,2. I0.13); none of the other founders of the Sunni
madhiihib is mentioned, let alone a Shi'ite Imam. In fact Avicenna defines the Shi'ites as those
who incline towards 'Ali b. Abi Talib to an excessive degree (wa-lladh1na yamiluna ilii jiinibi
'Aliyyi bni Abi Tiilibin bi-akthara mina l-qadri 1-mu'tadili yusammuna anfusahum ash-Shi'ata):
Risa/aft 1-lugha, 16,7-9. What is more, at ljikma 'Arui;liyya 5r12-13, Avicenna offers, as an
example of priority with respect to perfection, the sense in which Abii Bakr (r. 632-634), the first
of the rightly guided caliphs, is prior in nobility to the second, 'Umar (r. 634-644) (wa-yuqiilu
"qabla"fi 1-kamiili ka-qawlinii innii Abii Bakrin mutaqaddimun 'ala 'Umarafi sh-sharafi). It is hard
to imagine a Shi'ite 21-year-old using Abii Bakr, and not 'AH, as an example of priority in respect
of Pfrfection or nobility.
3 Rudolph 1996, 212-13; compare al-Fli.rabi, K. al-~uruf, 11 J,17-21.
158 Part II: The Avicennian Synthesis

discussions of things and existents and Avicenna's distinction between essence

and existence.
The second challenge to my hypothesis still remains to be met. For even if
future research is able to establish a firmer link between kaliim uses of the term
shay'iyya and Avicenna's own use of it in Iliihiyyiit 1.5, the fact remains that
the concept of what it is to be a thing (of thingness, in other words) which
Avicenna articulates in Iliihiyyiit 1.5, appears to be inconsistent with his
discussions of things elsewhere. More importantly, Avicenna's concept of
thingness in Iliihiyyiit 1.5 does not seem to mesh with his concept of miihiyya,
or essence. Despite Avicenna's clear assertion in Iliihiyyiit 1.5 that thing and
existent are co-implied (mutaliizimiini) and extensionally identical, there are
several discussions of essence and existence elsewhere in the Iliihiyyat and in
other works from Avicenna's middle period, which suggest that shay' - and
miihiyya, as we shall see - is at the very least logically prior to mawjud, and
perhaps even a broader category than mawjud. If shay' is logically prior to
mawjud, then Avicenna's assertions about the reciprocal nature of the co-
implication (taliizum) between thing and existent - so clearly articulated in
llii.hiyyli.t 1.5 - will be undermined. Even more seriously, if shay' is a broader
category than mawjud, then their extensional identity will be repudiated.
The following are summaries of the relevant passages, including Iliihiyyiit
1.5, in chronological order:

(1) Ifikma cAru¢iyya, 2v8-10: "One" is a necessary accident of things (mina L-

a crii¢i l-Liizimati Li-l-ashyii.'i). Essence is a thing (bal takunu l-miihiyyatu
shay'an), be it a man or a horse, an intellect or a soul; that thing is only
subsequently characterized as being one or existent. 34
(2) Madkhal 1.2: The essences of things (miihiyyiitu l-ashyii') are sometimes
found in concrete objects in the outside world, and other times are conceived of
in the mind. But essence has three aspects: as a concrete, external existent; as
a mental, internal existent; and a third aspect, in which it is unrelated to either
concrete or mental existence. 35
(3) Madkhal 1.12: Genera and species may be divided into those which are
before a state of multiplicity (that is, those contained in the Active Intellect),
those which are in a state of multiplicity (that is, those contained individually
in sublunary concrete existents), and those which are after a state of
multiplicity (that is, those contained as abstracted universals in human
intellects). Taken in itself a genus or a species is a thing. "Animal", taken in
itself, is an intentional object (ma <nan), regardless of whether it is a concrete
or a mental existent, or whether it is general or specific. 36

34 This passage is identical to K. an-na)iU, 340,12-15.

35 Ibn Sina, K. ash-shifii'!Manfiq (1): al-Madkhal 1.2, 15,1-7.
36 lbn Sina, K. ash-shifii'!Mantiq (1): al-Madkhal 1.12, 65,4-12. On this passage see Marmura

1979.
 7. Essence and Existence (A) 159
(4) /liihiyyiit 1.5 (discussed above): "Thing" and "existent" are primary,
indefinable categories. Whatever is predicable of thing will also be predicable
of existent, and whatever is predicable of existent will also be predicable of
thing. Although they are co-implied, thing and existent have different
meanings.
(5) lliihiyyiit 5. l: A universal (kullf) such as "horseness", taken in and of itself
lfi nafsihi) - that is, without considering whether it is one or many (lii
wiiJ:iida wa-lii kathfra), a concrete existent in the outside world or a mental
existent inside the soul (lii mawjuda ft l-a<yiini wa-lii mawjuda ft n-nafsi), in
potentiality or in actuality (lii bi-l-quwwati wa-lii bi-l-ficli) - is a thing
(shay'). 31
(6) lliihiyyiit 7.1: "One" (al-waJ:iid) and "existent" (al-mawjud) are equally
predicable of things (qad yatasawaytini ft l-J:iamli caza l-ashyii'); all that
may be characterized by "one" may also be characterized by "existent", but the
two terms do not have the same meaning as each other. 38

From these passages a discrepancy seems to emerge over the issue of what
exactly is it to be a thing; over the question, that is, of thingness. The message
from the /lahiyyat 1.5 passage discussed above at length is that thing and
existent are extensionally identical yet intensionally different. In other words,
while thingness and existence have different meanings, neither term applies to
any entity that the other does not apply to. Yet in the ljikma <Aru(iiyya and in
lliihiyyiit 7 .1 Avicenna implies that thing is a universally applicable subject of
which one and existent may be predicated. In other words it is one and existent,
not thing and existent, which are extensionally identical yet intensionally
different. 39 Thing therefore seems to be a category which is more basic than, and
hence logically prior to, the categories one and existent.
This logical priority is reinforced by Avicenna's implication in Iliihiyyiit 5. I
that properly speaking, thing should be elevated above considerations of
existence. There Avicenna claims that a universal is a thing when taken in itself,

37 Ibn Sina, K. ash-shifii'/Iliihiyyiit (1) 5.1, 196,6-13.

Js Ibn Sina, K. ash-shifii'/lliihiyyiit(2) 7.1, 303,6-12.
J 9 Avicenna's assertion that mawjud and wal;id are extensionally identical but intensionally
different almost certainly derives from Aristotle's discussions in Metaph. 4.2, 1003b22-1004a2 (=
Ma bacda t-tabfca /, 310,1-311,4 and 316,9) where Aristotle discusses how "that which is" (to on=
al-huwiyya) and "that which is one" (to hen= al-wiil;id) can be seen to be extensionally identical
but intensionally different, in the sense that both are equally predicable of substance (ousia). See
also '"That which is one' is said of [the same things] 'that which is' is said of' (to hen legetai
hOsper kai to on = yuqiilu 1-wiil;idu ka-mithii ma tuqiilu 1-huwiyyatu ayt,lan) at Metaph. 7.16,
1040bl6-24 (= MBT II, 999,14-1000,5); and '"That which is' and 'that which is one' are the
predicates spoken most universally of everything" (to gar on kai to hen katholou kategoreitai
malista panton = fa-inna l-wiil;ida wa-1-huwiyyata maquliitun kulliyyatun aktharu dhiilika tuqiilu
<afiijamfci i-ashya'i) at Metaph. 10.2, 1053bl6-21 (= MBT Ill, 1268,11-1269,3). The assertion is
repeated by Themistius (fa-inna l-mawjuda shay'un wa-l-wiil;ida shay'un iikharu) in his paraphrase
of Metaphysics 12.7: Thamas\iyiis, Sharl:z l:zarf al-Lam min Kitiib Mii ba<da !-fabfca li-Aris{il, 15, 13.
160 Part II: The Avicennian Synthesis

that is, when it is separated from any consideration of the mode of its existence:
one or many, concrete or mental, potential or actual. Not only is a universal
taken in itself a thing, but - we are told in Madkhal 1.12 - a genus or a species
taken in itself is a thing. Even more importantly, Avicenna says in the /fikma
cAru¢iyya passage that an essence taken in itself is a thing, and that "one" -
and, by implication, "existent" - is merely a necessary accident of essence taken
in itself. Most starkly of all, Avicenna asserts in Madkhal 1.2 that essence has a
third aspect, unrelated to any type of existence whatsoever.
The clear message given in /liihiyyiit 1.5 that thing and existent are co-
implied, with neither being prior to the other, is therefore muffled by the many
hints from the other passages that thing is logically prior to existent. More
serious is the question of extension. For while all the passages are consistent in
maintaining that thing and existent have different meanings, two in particular -
/fikma cAru¢iyya 2v8-10 and Madkhal 1.2 - imply that there is a separate
category of things - essence taken in itself - to which existence need not apply.
In these passages Avicenna seems to have forced himself into advocating the
Muctazilite position that existent is subsumed extensionally but not
intensionally under thing; that is, that existents are always things but things are
not always existents.40
Is there any other evidence in Avicenna's writings to help us understand his
concept of thingness? Does that evidence allow us to reconcile Avicenna's
various positions on things and essences? These questions will be addressed in
Chapters 8 and 9.

40 Much ink has been spilled trying to detennine how Avicenna meant this extra aspect of

essence to be understood. Some maintain that it should be understood simply in the context of a
thought experiment, in which miihiyya can be isolated in a purely logical sense from its existence
as a concrete individual in the outside world and from its existence as a universal concept in the
mind. Others, citing remarks Avicenna makes in lliihiyyiit 5.2, believe that the passage provides
evidence that he thought existence was attached to essence merely as an accident. Western
interpreters in the Catholic tradition have generally understood Avicenna as implying the latter, but
consensus is now emerging that the fonner is also a feasible interpretation. For a sampling of
different views see Goichon 1937 passim; Gilson, 1948, 121-131; Rahman 1958; Morewedge 1972;
Rahman 1981; Burrell 1986; Nasr 1989; and Mannura 1992.
 8. Essence and Existence (B)
ShayJiyya or Sababiyya?

To get a fuller picture of the subtle tension in Avicenna's thought between

things and essences we must look for other instances of shay'iyya in
Avicenna's writings. After all, the isolated instance of thingness in Jliihiyyiit l. 5
- the only instance of shay'iyya that Jolivet cites - hardly constitutes a fully
articulated concept in Avicenna's thought. Where else, if anywhere, does
shay'iyya appear in Avicenna's works? It turns out that the Iliihiyyiit 1.5
passage, the one that so clearly echoes the earlier Mu 0 tazilite discussions of
things and existents, is not the only time when Avicenna uses the term
shay'iyya.
Far more often than its lonely appearance there, where the mutakallimun are
hovering in the background, Avicenna uses shay'iyya in a cluster of discussions
which have nothing to do with early kaliim debates and which are driven by a
problem specific to Aristotelian philosophy: the relation between efficient and
final causes. In Iliihiyyiit 6.5 Avicenna writes:

L36
Ibn Sina, Kitab ash-shifa'lllahiyyat (2) 6.5, 292,1-294,5
Avicenna, Liber de philosophia prima sive scientia divina V-X,
335,84-339,32
The subsequent objection [to the reality of final causation, viz.,
"How can the end be anything but posterior to the other causes?"]
will be solved by knowing that the end may be taken to be a thing
as well as taken to be an existent. Although a thing cannot be
other than an existent, the difference between thing and existent is
just like the difference between some entity and its concomitant.
You have already come to know and to verify this. Consider, once
again, the case of man: man has an inner reality, consisting of his
definition and his essence, which is not conditioned upon [his]
existence's being particular or general, concrete or in the soul, or
potential or actual.
[292,6] Each cause, insofar as it is that [particular] cause, has an
inner reality and a THINGNESS. In its THINGNESS the final cause is the
reason why the other causes actually exist as causes. In its
existence the final cause is the effect of the other causes' actually
[existing] as causes. It is as if the THINGNESS of the final cause were
the cause of the cause of its [own] existence; conversely it is as if
its existence were the effect of the effect of its [own] THINGNESS.
However its THINGNESS does not become a cause unless it occurs as
an image formed in the soul, or as something analogous to that.
The only cause of the final cause in its THINGNESS is another cause
which is different from the cause toward which one thing sets
[another] in motion, or toward which something is set in motion.
162 Part II: The Avicennian Synthesis
Know that something can be caused in its THINGNESS, and caused
in its existence. An example of what is caused in its THINGNESS is
twoness: contained in the definition of its being twoness is its
being caused by oneness. What is caused in its existence is so
plainly knwon [as to require no explanation]. Similarly, one thing
might possess another thing that occurs as an existent in its
THINGNESS, as twoness possesses numbemess; or else one thing
might be additional to another thing that is added to its THINGNESS,
as wood and stone contain quadrangularity. Natural bodies are
causes of the THINGNESS of many forms and accidents (I mean of
those [forms and accidents] which can only recur in them), as well
as being causes of the existence of some of them [the forms and
accidents] without [being a cause of] their [the forms' and
accidents'] THINGNESS, as is sometimes thought to be the case in
mathematics.
[293,4] It has thus become easy for you to understand [!] that the
final cause, with regard to its THINGNESS, is prior to the efficient
and receptive causes, and similarly, prior to form insofar as form
is a formal cause leading toward it [the final cause]. In addition, the
final cause is prior to the other causes in its existence in the soul.
As for [the final cause's being prior to the other causes] in the
agent's soul, this is because it [the final cause] comes to exist first
and then agency, seeking out a receptive patient, and the quality of
the form come to be represented as images. As for [the final
cause's being prior to the other causes] in the souls of those other
than the agent, one [cause] need not follow another in any kind of
necessary order. Therefore, in terms of THINGNESS and in terms of
existence in the intellect, there is no cause prior to the final
[cause]; instead, it is a cause of the rest of the causes' becoming
causes. However, the actual existence of the other causes as causes
is [itself] a cause of its [the final cause's] existence. The final
cause is a cause not insofar as it is an existent, but insofar as it is a
thing. In the sense that it is a cause, it is the cause of the causes,
while in the other sense it is the effect of the causes.
(293,12] This is [so] when the final cause is involved in
coming-to-be; when, however, the final cause is not involved in
coming-to-be but its existence is, instead, above coming-to-be (as
will be explained in its place), then none of the other causes is a
cause of it [the final cause], not even in the case of the One which
is Itself occurrence and existence. [In the former sense] therefore
the final cause will be caused by the rest of the causes not because
it is a final cause but rather because it is subject to coming-to-be;
and if it [the final cause] is not subject to coming-to-be it will not
be caused at all. When you consider its [the final cause's] being a
final cause, you will find it to be a cause of the rest of the causes'
being causes (viz., being an efficient cause and a receptive cause
and a formal cause) but not of their being entities and existents in
themselves. So what is essential to the final cause insofar as it is a
final cause is being a cause of the rest of the causes, while what is
accidental to it (insofar as it [the final cause] is understood to be
[actually] occurring in the world of generation) is being an effect
 8. Essence and Existence (B) 163
in the world of generation. It has thus been made clear to you [!]
how something may be cause as well as effect, given that it is
[both] an agent and an end, this being one of the Natural
Philosophers' principles.

Avicenna also mentions thingness in the same context in Najiit/lliihiyyiit 1. 11:

L37
Ibn Sina, K. an-Najii.t, 345,1-11
In terms of coming into existence (fi f:zu~iUi l-wujUdi] the end is
posterior to the effect, while it is prior to the rest of the causes in
terms of THINGNESS. It is clear that THINGNESS is something other
than existence in concrete reality [ghayru 1-wujudi ft 1-a'yii.ni]. For
an intentional object [al-ma'na] has an existence in concrete
reality and an existence in the soul as well as something common
[to both]: what is common [to both] is THINGNESS. The final
[cause] insofar as it is a thing is prior to the other causes and is the
cause of the causes in terms of their being causes. Insofar as it is
an existent in concrete reality it [the final cause] is posterior.
When the efficient cause is not itself the final cause, the agent is
posterior to the end in terms of THINGNESS. This is because the
other causes become actual causes only on account of the final,
whereas it [the final cause] is not on account of anything else, but
exists first in a kind of existence and then makes the [other] causes
become actual causes. Once this distinction is made, the result, it
seems, is that the primary agent and the primary mover in
anything is the end. 1

Before I try in to explain what Avicenna means when he uses shay'iyya in these
very difficult passages, and to determine whether or not this helps us reconcile
the apparent tension in his thought between thingness and essence, I must devote
this chapter to resolving a textual problem. The problem is that there are some
indications that sababiyya, "causality", and not shay'iyya, "thingness", was the
word Avicenna used in the passages above. This is in spite of the overwhelming
evidence provided by the manuscripts used to come up with the Cairo edition of
the Ilahiyyiit, which are unanimous in their reading of shay'iyya. 2 I shall argue
that the balance of evidence compels us to retain shay' iyya.
The textual problem in the Iliihiyyat might never have come to light had it
not been for the fact that the Latin translator Gundisallinus consistently rendered
what appears as shay'iyya in L36 - the long Iliihiyyat 6.5 passage - not as

1 The term shay'iyya occurs on 345,2(bis).4.7. Unlike the sections of the Najat immediately

surrounding L37, L37 itself is not identical to the analogous section in the I:Jikma 'Aril(iiyya (4v16-
5r5). Nor is there anything analogous to L37 in the Mabda' wa-ma'ad, the other early work from
which Avicenna lifted passages into the Najar. The conclusion must be that L37 was composed
ori~inally for the Najar.
The manuscripts examined for the Cairo edition are described in the introduction to Anawati
1978, 17-21. The use of shay'iyya in the Najar is confirmed in the 1593 edition: K. an-najar
mukhta~ar ash-shifa' li-bn Sina, 58,22-26.
164 Part II: The Avicennian Synthesis
realitas or entitas, as one would have expected, but as causalitas. 3 For
example, Avicenna's assertion that "Each cause, insofar as it is that [particular]
cause, has an inner reality and a THINGNESS", and which reads in Arabic:

wa-kullu cillatin fa-innaha min J:iaythu hiya tilka l-cillatu laha

/:laqlqatun wa-SHAY'IYYATUN,

was rendered into Latin as:

omnis autem causa, inquantum est ipsa causa, habet certitudine et

CAUSALITATEM.4

One explanation of this anomaly takes into account the fact that translating
Avicenna's philosophical works into Latin appears to have been a two-step
process. A first translator - the Jewish emigre lbn Dawud, or Avendauth - would
render the Arabic orally into the vernacular Spanish of twelfth-century Toledo; as
he went along, a second translator - the Catholic cleric Gundissalinus - would
render the vernacular Spanish into proper philosophical Latin.5 Given this
scenario, it is easy to imagine how the first translator might have tried faithfully
to render shay'iyya into the vernacular Spanish as cosita. The second translator
might then have confused cosita, "thingness", with causita, "causality'', terms
which were probably pronounced much the same.6
It appears more likely, however, that ta~J:i,if, a scribal misplacement of
diacritical marks, was the cause of the divergence between the Arabic and Latin
texts. The ta~J:i,if in this case would consist in a scribe's having mispointed
shay'iyya as sababiyya ("causality"), a far less mysterious term. (I take it as
obvious that shay'iyya is the lectio difficilior here.)
Confusing the pairs of terms is perfectly understandable. In a medieval Arabic
manuscript, after all, many words would have been written down without
diacritical marks. As a result, the skeleton (rasm):

) Avicenna, Liber de philosophia prima sive scientia divina V-X, 337,88.90.91.92.94.96.97.00.1.

3.5 and 338, 11 (= /lahiyyat, 292,6[bis].8.9[bis].10.12.14.16 and 293,2.3.4.8). On the translators,
see van Riet's comments in her introduction to Avicenna, Uber de philosophia prima sive scientia
divina I-IV, 123*.
4 Avicenna, Uber de philosophia prima sive scientia divina V-X, 336,87-337,88 (= /lahiyytit,

292,6). On the next page, Avicenna's assertion "that the final cause, with regard to its thingness, is
prior to the efficient cause", and which reads in Arabic anna 1-'illata l-ghti'iyyata ft SH-
SHAY'IYYATI qabla 1-'illati l-fii'ilati, was rendered into Latin as quod causa finalis in
CAUSALJTATE praecedit causas agentes: Avicenna, Uber de philosophia prima sive scientia
divina V-X, 337,5-6 (= Iltihiyytit, 293,4).
5 On this see van Riet's comments in her introduction to Avicenna, Uber de anima seu se.xtus de

naturalibus I-Ill, 95*-98*, and Burnett 1997, 12-14.

6 This is the theory advanced by d' Alverny with regard to two instances where, as she

understood it, cosa had been misunderstood as causa. These are mentioned (without references to
the text) in d'Alverny 1951, 133, and repeated in d'Alvemy 1952, 60. The instances of
mistranslation she probably had in mind are Avicenna, Liber de anima seu se.xtus de naturalibus f-
ill, 192,20 (= K. ash-shifii' fTabi'iyyat (6): an-Nafs, 103,I I) and 273,11(=154,19).
 8. Essence and Existence (B) 165

can just as easily be pointed as sababiyya ("causality"):

....
as it can be pointed as shay'iyya ("thingness"):

.. ,, .
~.
.. ..
But given the lack of extant Arabic /lahiyyat manuscripts from before 1284 (let
alone from before the Arabic-Latin translation movement in the middle of the
twelfth century), it is impossible to say whether a scribe mistranscribed shay'iyya
as sababiyya before the manuscript reached Avendauth's hands and Avendauth
then read the mispointed manuscript accurately; whether the manuscript reached
Avendauth with these words pointed as shay'iyya and he then "corrected" the
pointing to sababiyya before translating it; or whether the manuscript reached
A vendauth with these words unpointed and he then pointed them incorrectly as
sababiyya before translating them.
In any case, what inclines me towards attributing the divergence between the
Arabic and Latin texts to ta~J:i,if and not (as d' Alvemy thought) to mishearing, is
that what appears as shay'iyya in the Ilahiyyat 1.5 passage discussed earlier
(i.e., L35) was also incorrectly rendered into Latin but this time as similitudinem,
clearly the result of Avendauth's (or an earlier scribe's) mispointing - the
skeleton:

having been pointed as shabzh ("similitude"):

...
instead of as shay'iyya ("thingness"):

..
~
,, .
....
166 Part II: The Avicennian Synthesis

- rather than of Gundissalinus' mishearing. 7

On the other hand, the specter of ta~J:zif raises the chilling possibility that the
original Arabic texts - the ones written or dictated by Avicenna himself - should
in fact be read with sababiyya, not shay'iyya. This concern is heightened by
the fact that the oldest manuscript of the Latin translation of the lliihiyylit dates
from about 1240, while the oldest Arabic manuscript of the lliihiyylit dates from
1284.8
However, I believe that the Latin translation does not, in itself, provide
enough evidence to justify concluding that Avicenna wrote or dictated sababiyya
rather than shay'iyya. First of all, if causalitas were an indication that the
original Arabic should read sababiyya, not shay'iyya, then - by way of
consistency - we would also expect to see the instances of shay' in the
lliihiyylit 6.5 passage to have been rendered into Latin as causa. This is not the
case. The Arabic of the Iliihiyylit 6.5 passage immediately preceding the last
quotation:

wa-ammii sh-shakku lladhi yal!hi fa-yanJ:iallu bi-an yu<[ama anna l-

ghiiyata tufr<l{i.u SHAY,AN wa-tufrat;iu mawjudan,

was rendered accurately into Latin as:

Sed dubitatio quae sequitur hie solvitur hoc modo: scilicet, iam scis
quod finis ponitur RES et ponitur ens. 9

Second, sabab would' clearly have been out of place in almost every instance
where shay' was mistranslated into Latin as causa rather than res, so we can be
quite certain that in Jliihiyyiit 6.5, at least, the original Arabic text did not read
sabab. 10

1 Avicenna, Liber de philosophia prima sive scientia divina /-IV, 38,21 (= l/ahiyyat, 33,17).

Jolivet 1984 seems not to have noticed this mistranslation; Rahman 1958, 5n.2, mentions it, as does
Khodeiri 1959-61, 316. Elsewhere in the Latin translation of the /lahiyyat, bi-sababihi was
mistranslated as comparationis - van Riet sees this as a result of misunderstanding it as nisbatahu:
188,80 (= /lahiyyat, 166,4) - and sababun as comparationes: 493,96 (= /lahiyyat, 413,12). This
confusion also occurs in a passage of the Ff n-nafs, where sabab was translated into Latin as
comparatio (Avicenna Liber de anima seu sextus de naturalibus /V-V, 20,79 = K. ash-
shifi,'/Tabiciyyat (6): an-Nafs, 174,14).
Introduction to Avicenna, Liber de philosophia prima sive scientia divina /-IV, 125*-138*.
9 Avicenna, Liber de philosophia prima V-X, 336,84-85 (= l/ahiyyat, 292,1).
10 In addition to the two instances alluded to by d' Alverny, shay' is mistranslated as causa

elsewhere in the Latin version of the /lahiyyat: shay'ayni becomes duarum causarum at 99,62 ( =
l/ahiyyat, 86,3), and sa'ir al-ashyii' becomes ceterae causae at 394,84 (= l/ahiyyiit, 341 ,3). In the
Latin translation of al-Af'iil wa-l-in.fi'aliit, the fourth book of the Tabi' iyyat, shay' was also
occasionally mistranslated as causa: Avicenna, Liber quartus naturalium de actionibus et
passionibus qualitatum primarum, 23,72 (= K. ash-shifa'/Tablciyyat (4): al-Aj'al wa-l-infi'alat,
213,15); 28,54 (= al-Aj'al wa-l-infi<aJat, 217,3); and 79,73 (= al-Aj'al wa-l-infi<aJat, 256,6).
Conversely, asbab was mistranslated in the Ff n-nafs as rebus at 21,82 (=K. ash-shifa'/['abi'iyyiit
(6): an-Nafs, 174,15), and sabab as res at 21,83 (= K. ash-shifa'/['abl'iyyat (6): an-Nafs, 174,16).
 8. Essence and Existence ( B) 167
All in all, the evidence provided by the Latin translation of the Jlahiyyat (and
the parallels in the translation of the Fl n-nafs) allows us merely to hypothesize
that ta~J:i,if occurred in one or more of the manuscript lines that reached
Andalusia. It is not conclusive with regard to the question of whether the
original text should read sababiyya rather than shay'iyya.
The Latin mistranslation of shay'iyya in the /lahiyyat - if it was in fact a
mistranslation, as I believe - does, however, explain the absence of analogous
concepts in discussions of final and efficient causes by later Latin philosophers,
and the presence in those same discussions of causalitas. 11 Medieval European
thinkers would probably not have been struck to find Avicenna's concept of
thingness missing from their translations of the /Lahiyyat, since shay'iyya does
not appear in al-Ghaziili's (d. 1111) summary of Arabic philosophy, the
Maqa~id al-falasifa, through which - as Intentiones philosophorum - many
Latins were introduced to Avicenna's thought. The Maqa~id was popular due to
al-Ghaziili's fluency of style and liberality with examples, qualities which
Avicenna painfully lacked. It is an irony of the history of medieval philosophy
that during the period before Averroes' (d. 1198) Tahafut at-tahafut was
translated into Latin - that is, before al-Ghaziili was exposed as a philosophy-
basher - "Algazel" was seen by some Europeans as Avicenna's greatest
disciple. 12
The non-appearance of shay'iyya in the Maqa~id is, on the surface, evidence
inclining us towards thinking that the /lahiyyat 6.5 and Najat passages should
be emended to read sababiyya. But shay'iyya's absence in the Maqa~id has a
more straightforward explanation. Al-GhazaU based the Maqa~id primarily
(though not entirely) on Avicenna's Persian summa, the Danishnama-yi <Ata' I,

11 The Latin Avicenna's "theory" that the final cause is prior to the efficient in terms of

causality is cited explicitly by Henry of Ghent (d. 1293), Quadlibet XIII, 106,71-78; and by Duns
Scotus (d. 1308), Quaestianes super Libras Metaphysicarum Aristatelis I-IX, Bk. 5, Quest. I, Para.
51. Avicenna's "theory" (or the gloss on it contained in the Latin translation of al-Ghazali's
Maqii~id - see note 14 in this chapter) also seems to inform the opinions of Alexander of Hales (d.
1245), in Metaph. I, 12rC (Avicenna' s lliihiyyiit 6 is mentioned at 12vE and 104rD), in Metaph. 3,
54rD-55rD, and in Metaph. 5, 105rA-D; of Thomas Aquinas (d. 1274), Scriptum super Libras
Sententiarum 1.8, 52,5-6); in Metaph. 5 #775 and #782; in Phys. 2.3, #186; and Summa contra
gentiles 3.17.9; of Siger of Brabant (d. 1283), Quaestianes in metaphysicam 5.9, 203-4; of John
Buridan (d. 1358), in Metaph. 5, quaest. 1, Fols 26a-27a; of Albert of Saxony (d. 1390), Expasitio
et quaestianes in Aristotelis Physicam, 111,6-9; and of Francisco Suarez (d. 1617), Disputationes
metaphysicae Disp. XXVII, Sect. 2, paras 7-14, (172-177). Maier 1955, 212-213 and 302-303,
discusses Avicenna's "theory" and relates it to Buridan's; see also Biard 1997, 494-495. This is not
to say that Avicenna's concept of thingness was lost forever to European thinkers. In 1907 Horten
translated the instances of shay' iyya in /liihiyyiit 6.5 correctly into German as Dingheit (Horten
1907, 428-9), and the instances of shay'iyya in the Najar passage were translated by Carame into
Latin as entitas: Carame 1926, 34,21-35, 19 (shay'iyya is rendered as entitas at 34,23[bis) and
35,2.7).
12 e.g., Giles of Rome, introducing the section on al-Ghazlili from his Errores Philasophorum,

38,4-5: Algazel autem, ut plurimum Avicennam sequens et eius abbreviator existens .... "
168 Part II: The Avicennian Synthesis

in which a Persianized equivalent of shay,iyya does not make an appearance. 13

Avicenna's Persian statements in the Diinishniima that:

L38
lbn Sina, Iliihiyyiit-i Diinishniima-yi 'Alii,I, 54,9-10 and 55,2
The end is the cause of all the causes .... So when there is an end, it
is the cause of all the causes.

were rendered into Arabic by al-Ghazali in the Maqii#d as:

L39
al-GhazaH, Maqii$id al-faliisifa, 190, l 0.13
Part of what is special about the final cause is the fact that the
other causes become causes through it .... So when the final exists
among all the causes, it is the cause of the causes. 14

Since al-Ghazali: did not allow the straw-man philosopher of the Maqii:fid to
apply the concept of shay,iyya to the question of how efficient and final causes
relate to each other, he probably felt no need to criticize thingness in the
polemical Tahiifut al-faliisifa. And perhaps as a consequence of shay'iyya's
absence in the Tahiifut al-faliisifa, Averroes may have felt no need to raise the
topic in the Tahiifut at-tahiifut. Nor is Avicenna's distinction between thingness
and existence cited by Averroes in the latter's long commentaries on any of the
most canonical Aristotelian discussions of the four causes, Physics 2.3 and 2.7
and Metaphysics 5.1 and 5.2. 15 Therefore, shay'iyya's absence from al-
Ghazfili's (and hence Averroes') discussions of efficient and final causes does not
by itself constitute evidence in favor of either reading.
Some evidence that appears to favor sababiyya may be found in the writings
of two other thinkers trained in the Ash<arite theological tradition. The Ash<arite
mutakallimun ash-Shahrastani (d. 1153) and Fakhr ad-din ar-Riizi (d. 1210) wrote
extensively about Avicenna's metaphysics, among other topics. If shay,iyya
were to appear in one of their summaries or critiques of Avicenna's discussions
of causality, it might help confirm its inclusion in the Iliihiyyiit 6.5 and Najiit
passages.

13 On the Diinishniima's relationship to a!-Ghazali's Maq[4id, see Janssens 1986.

14 Interestingly, the Intentiones philosophorum - the Latin translation of the Maqii~id aljaliisifa
- contains several extra lines which are in neither the Danishniima nor the Arabic edition of the
Maqii~id; these include the assertion "Indeed, the final cause is last in terms of existence, yet first
and foremost in terms of intention" (Causa vero finalis est ultima in esse, et est prima et precedens
in intencione = Intentiones philosophorum, 38,5-6). My guess is that the extra line is a gloss by
Avendauth or Gundissalinus, given their intimate involvement in translating both al-Ghazali's
Maqii~id and Avicenna's Iliihiyyat into Latin. Stem 1962 describes how this precept was applied to
the arrangement of topics in philosophy books.
is Averroes in Phys. 2.3 and 2.7 (= Aristotelis opera cum Averois commentariis IV, 59D-63K)
and lbn Rushd, Tafslr Ma ba'da {-!abl'a II, 475,1-481,8; 483,7-487,8; 490,1-497,6).
 8. Essence and Existence (B) 169
The opposite seems at first glance to be the case. One major piece of evidence
tempting us to reject shay'iyya in favor of sababiyya in the Najat is the fact
that in ash-Shahrastani's doxography, the Kitab al-milal wa-n-niftal, ash-
Shahrastani paraphrases the Najat passage on final and efficient causes (i.e.,
L37) but appears to read sababiyya instead of shay'iyya. According to the
editor's critical apparatus, however, most of the Mital manuscripts read
shay'iyya. The Milal's importance as a piece of evidence one way or the other
is therefore limited. 16
More alarming is the evidence pointed to by ar-RazL Ar-Razi wrote a very
important commentary on as well as a shorter summary of Avicenna's Kitab al-
isharat wa-t-tanbihiit, a late work of Avicenna that greatly influenced
subsequent Islamic philosophy. 17 In the Ishiirat discussions of final and efficient
causes shay' iyya does not make an appearance. Instead, Avicenna uses mahiyya
- not shay'iyya - in the passage from the /shiirat's metaphysics section
(Nama{ 4: Fi l-wujud wa-'ilalihi) which parallels the /lahiyyat 6.5 and Najiit
passages, as well as the term 'illiyya (causality):

L40
lbn Sina, Kitab al-ishlirat wa-t-tanbfhlit, 139,14-20
Something may be caused with reference to its essence and its
inner reality, and it may be caused in terms of its existence .... The
final ... is the efficient cause of the causality of the efficient cause
[li-«illiyyati> l-'illati l-fa'iliyyati]. 18

In another passage from /shiiriit 4, Avicenna says:

L41
Ibn Sina, Kitab al-ishlirlit wa-t-tanbfhlit, 140,6-9
The final cause - that on account of which the thing is - is a cause,
through its essence [bi-mahiyyatihi] and its [being an] intentional
object [wa-ma'nlihu], of the causality of the efficient cause [li-
'illiyyati l-'illati l-fa'iliyyati], while it is an effect of it in its
existence. The efficient cause is a cause of its [the end's] existence
if it is one of the ends that actually come into being, but it [the
efficient cause] is not a cause of its (the end's] causality nor of its
[being] an intentional object.

In his summary of the /shiirat, entitled Lubiib al-isharat, ar-Razi has this to say
about the /shiiriit passages:

16 K. al-mi/al wa-n-nihal ll, 1092,6-1093,4.

17 In Chapter 9, n. 8 I.briefly discuss Michot's objections to giving the lshiiriit a late date.
18 I supply 'illiyya with MS F, to concur with the phrasing in L41.
170 Part II: The Avicennian Synthesis
L42
ar-Razr, Kitiib lubab al-ishlirat, 80,4-5
How clever the Shaykh was to say that the final cause is an
efficient cause of the causality of the efficient cause [li-'illiyyati 1-
'illati l-fii'ilati]. 19

To my mind Avicenna's use of 'illiyya in the Ishariit passages devoted to

efficient and final causality is simply evidence of what appears to be an almost
universal preference for 'illiyya over sababiyya in his works. 20 But a
counterargument might run like this. Given shay'iyya's absence in the /shariit
passages, and given instances elsewhere in Avicenna's works where he uses
'illiyya to describe priority in causality, we would be perfectly justified in
emending the Iliihiyyiit 6.5 passage (i.e., L36) and the Najat passage (i.e., L37)
to read sababiyya instead of shay'iyya. 21 In other words, Avicenna sometimes
uses 'illiyya to describe priority in causality, and other times (e.g., in the
Iliihiyyiit 6.5 and Najat passages) uses sababiyya to describe priority in
causality; shay"iyya need not come into the picture at all.
As attractive as this counterargument might appear, I believe it suffers from
flaws graver than those of my own argument. First of all, when Avicenna talks
about priority in 'illiyya it is clear from the contexts in which the term appears
that he is referring to the priority enjoyed by any cause - be it formal, material,
final or efficient- to its effect. As far as I know, Avicenna nowhere says that the
final cause in particular enjoys priority in 'illiyya to the efficient cause.
Second, Avicenna's assertion in the /shiiriit that the final cause is a cause
through its essence (bi-mahiyyatiha) of the causality of the efficient cause (li-
'illiyyati z-cillati l-fa'iliyyati), is perfectly reconcilable with his assertions in
Iliihiyyiit 6.5 that the final cause is both a cause through its own thingness and
of the efficient cause's thingness. As for the first prepositional phrase (i.e., bi-
miihiyyatiha), the lshiirat passages are meant to show that the final cause's
priority derives from its essence; and essence, as I have shown in Part 1, is
conceptually very similar - though not, it seems, identical - to thingness.
Because mahiyya is much closer in meaning to shay"iyya than it is to
sababiyya, the lshiiriit passages can just as easily be understood as providing

19 Ar-Rii.zi echoes this in his fully fledged commentary, Shar/.I al-ishiiriit, 193,36-194,20 (outside

boxJ and in his al-Mabii/.lith al-mashriqiyya /, 661,21-662,8.

2 e.g., /liihiyyiit 16,3; 166,12; 169,IO(bis); and 266,11. Goichon 1938, 150, says that sababiyya

is "tres rarement employe". In addition to the instance she cites (she refers to "Qa(iii', 57'', but I
have been unable to locate this), the only places where I have comes across sababiyya in
Avicenna's works are ar-Risiilat al-'arshiyyafi /.laqii'iq at-taw/.lfd wa-ithbiit an-nubuwwa, 24,6; and
'Uyun al-/.likma, 52,8. According to the editor's apparatus for the latter work, however, the
manuscripts contain many variant readings of sababiyya, and in any case, the term is not applied
either there or in the Risiila 'arshiyya to the final cause specifically but rather to the substrate
(maw(iu'), end and agent collectively.
21 See aqdam bi-1-'illiyya in K. ash-shifii'/Man{iq (5): al-Burhiin, 297,10-11; taqaddum fi 1-

'illiyya in Avicenna's De Anima - Arabic Text, 230,3, and in K. al-hidiiya, 241,l; and qabla ft l-
'illiyya in lfikma 'Aru¢iyya, 5rl3.
 8. Essence and Existence (B) 171
evidence in favor of, rather than against, retaining shay"iyya in the Najat and
Ilahiyyat 6.5 passages. As for the second prepositional phrase (i.e., li- cilliyyati
z-cillati l-faciliyyati), the apparent discrepancy between the final cause's being
the cause of the efficient cause's thingness and its being the cause of the efficient
cause's causality, is resolved by recognizing that the thingness (shay"iyya) of
any cause (cilia) is precisely its causality (cilliyya) .
My confidence in shay"iyya is strengthened by turning from the evidence in
later Ashcarism to that in later Shi'ite Avicennism. For example, in a~-Tiisi's
(d. 1274) commentary on the Isharat passages on final and efficient causes, he
follows Avicenna's lead and uses miihiyya as well as cilliyya:

L43
a~-Tiisi,
Sharf:i al-ishiiriit, 193,31-194,6
The essence of the end and its [being an] intentional object - I
mean its being some particular thing or other [kawnahii shay"an
mii] - is different from its existence .... [The final cause's]
causality [cilliyyatuhii] consists in the fact that it makes the agent
an actual agent and is thus a cause of the agency of the agent. The
agent is a cause of the fact that that essence [of the final cause]
becomes an existent. Thus the essence of the end is a cause of the
cause of its existence not in an absolute sense, but in a certain
respect, so no circularity need be implied by this. 22

It seems clear from his gloss on mahiyya and macna - that they refer to the
end's being some thing or other - that a~-Tiisi had read the Ilahiyyat 6.5 and
Najat passages carefully and was borrowing from them to flesh out the much
terser /sharat passage. For it is the distinction between essence and existence,
not one between causality and existence, that Tiisi (rightly) sees as underlying
Avicenna's distinction between the final cause's causality and the efficient
cause's causality.
Like a~-Tiisi, Mulla ,Sadra (d. 1640)- in his commentary on the /lahiyyat 6.5
passage - seems aware of the Isharat passage when he pairs shay"iyya with
miihiyya, and uses cilliyya:

L44
Mullii ~adrii, at-Tacllqiit 'a/ii Iliihiyyiit ash-shifii', 258,3
Therefore the final cause, through its essence and its thingness, is
a cause of the causality of the rest of the causes [Ii-<illiyyati sii"iri
I- 'ilali] .23

22 Compare similar uses of miihiyya by Avicenna's pupil Bahmanyar ibn al-Marzubiin (d. 1067)

(at-Taflfli, 546,1-9), as well as Sabzawan (d. 1878) (Sharf! ghurar a/-farii'id [also known as Sharfl-
i manplmah], 161,11-12).
23 In his K. al-mashii'ir, 7,13 and 20,19.20, Mullii Sadrii - in existentialist rather than

commentator mode - chucks shay'iyya into the dustbin of tenns denoting essence.
172 Part II: The Avicennian Synthesis

Even in his own philosophical treatises Mulla Sadra occasionally retains

shay'iyya. In the Kitiib al-asfiir, for example, he says:

L45
Mulla $adra, al-Asfiir al-arba<a II, 347,4-5
Although with respect to thingness [bi-~asabi sh-shay'iyyati] the
end is prior to the act, it follows nonetheless that with respect to
existence it is posterior to the act and subsequent to it.

To summarize Chapter 8, then, pieces of evidence from the later Latin and
Ash<arite traditions tempt us to reject shay'iyya and replace it with sababiyya.
Nevertheless, I feel the burden of proof still lies with anyone who would
advocate such an emendation, in view of the following factors: the apparent
unanimity of the extant Najat and Ilahiyyii.t manuscripts in reading shay'iyya;
the inconsistent and scattered mistranslations of sabab, shay' and shablh in the
Latin translations of the Ilahiyyat and the Fi n-nafs, which indicate ta~l:zif in the
Latin translators' Arabic manuscripts but not necessarily in the originals those
manuscripts were derived from; the principle of lectio difficilior, which clearly
favors thingness over causality; Avicenna's use of mii.hiyya - a term that is far
closer in meaning to shay' iyya than sababiyya - in the parallel /shariit
passages; Avicenna's clear preference for cilliyya over sababiyya when
discussing causality; Mulla Sadra's equation of shay,iyya and miihiyya; and.
finally, shay,iyya's appearance in Avicenna's Tacliqiit and Mubiiflathiit, as we
shall see in Chapter 9.
 9. Essence and Existence (C)
The Question of Evolution

Now that I have argued in favor of retaining thingness in the lliihiyyat 6.5
passage (i.e., L36) and in the Najiit passage (i.e., L37), albeit with a few
qualms, I must try to explain what exactly the term means in the context of
Avicenna's discussion of the relation between efficient and final causes. Once
this is done it will be easier to determine how these discussions contribute to our
understanding of Avicenna's concepts of thingness and essence, and of his
progression from the kaliim problematic of shay' vs. mawjud to his own
problematic of miihiyya vs. wujud.
I think that when Avicenna asserts in lliihiyyat 6.5 that the final cause is prior
in its shay'iyya to the efficient cause, he is using the term in one of the two
early kalam senses of thing: the notion that a thing (for example, the Day of
Resurrection, or my great-great-granddaughter) can subsist mentally in God's -
or anyone's - mind, before it exists in the real world. Here is an example of how
thingness thus conceived works in final causation. I am thirsty and a bottle of
soda is in the refrigerator. I want to quench my thirst by drinking that bottle of
soda. Quenching my thirst exists in my mind as a final cause. My motion to the
refrigerator - the efficient cause of my quenching my thirst - then comes into
concrete existence in the outside world. But the quenching existed first in my
mind - as a thing - before it existed concretely in the outside world. In this
sense the final cause is prior in its thingness (prior as a thing, that is) to the
efficient cause. This is why Avicenna asserts in /lahiyyat 6.5 that "its [i.e., the
final cause's) thingness does not become a cause unless it occurs as an image
formed in the soul" (292,9) and "In terms of thingness and in terms of
existence in the intellect, there is no cause prior to the final cause" (293,8-9).
Avicenna's assertions here are echoed in the Taclfqat:

L46
Ibn Sina, at-Ta<lfqat, 128,17-19
The end is prior in its thingness to all the causes and posterior in
the existence it derives from them. The end which is absolutely
non-existent [al-ghiiyatu l-ma<dumatu <aJa l-i{liiqi] is not a cause.
Instead, it must exist in the agent's mind [alternatively, "in the
agent himself' - ft nafsi l-ja<ili] in order to perform its action.'

When shay'iyya is used in this way, as the basis for the priority enjoyed by
things which are in the mind but which have not yet come into being in the
outside world, it helps Avicenna flesh out an Aristotelian assertion: that the

1 Also see lbn Sina, at-Ta'lfqiit, 158,14.

174 Part II: The A vicennian Synthesis

efficient and final causes can be seen to be causes of each other. In Physics 2.3
Aristotle briefly discusses the reciprocity between efficient and final causes:

L47
Aristiita!Is, af-Tabfca, 103,8-13
Aristotle, Phys. 2.3, l 95a8-l l
Sometimes things are causes one of the other [wa-qad takiinu
ashyii'u bac{iuhii sababun li-bac{iihii = esti de tina kai allelon aitia].
For example, hard work is the cause of the body's health and the
body's health is the cause of hard work, though not in an identical
way [ghayra anna dhiilika laysa min wajhin wiiJ:iidin =all' ou ton
auton tropon]; the body's well-being is a cause in that it is an
intended end, while hard work is a cause in that it is the origin of
motion .2

The interpretive challenge facing Avicenna here was to uphold Aristotle's fairly
straightforward idea - that the efficient can be seen as the cause of the final and
the final can be seen as the cause of the efficient - without falling into the trap
of circularity. For if the final cause is simply the cause of the efficient cause, and
the efficient is simply the cause of the final, each will be the cause of the cause
of itself, and circularity will result. Avicenna had to find some way to unpack
(and defend) Aristotle's compressed assertion by providing a metaphysical basis
for the distinction between the ways in which the final cause and the efficient
cause operate.
This is where the kaliim distinction between shay' and mawjud came in
handy. Understood as a thing, the final cause can have a shay'iyya in the mind
before it comes into existence concretely in the outside world. In terms of its
being a shay' in the mind - in terms of its shay' iyya in the mind, that is - the
final cause can be seen to be the cause of the efficient cause. On the other hand,
the efficient cause comes into concrete existence in the outside world before the
final cause comes into concrete existence in the outside world. My motion to the
refrigerator exists concretely before my quenching exists concretely. By
distinguishing between the final cause's priority as thing and the efficient cause' s
priority as concrete existent, Avicenna has managed to wriggle out of the hole of
circularity.
But while Avicenna's analysis in Ilii.hiyyii.t 6.5 helps him flesh out Aristotle' s
assertion about the reciprocity between efficient and final causes, it raises two
serious problems. The first is this: let us assume, as Avicenna says, that all
final causes are prior in their thingness to the efficient causes with which they
are paired. It follows that nothing without priority in thingness will be a final
cause; in other words, being prior in thingness will be a necessary condition of
being a final cause. But this description of how the final cause works will

2 This is echoed in Metaph. 5.2, !013b9-12 (=Ma bacda !-fabl'a II, 486,9-10). At Metaph. 1.3,

983a32, Aristotle asserts that the final cause is the "opposite" (antikeimene) of the efficient.
 9. Essence and Existence (C) 175

exclude all natural phenomena from the domain of entities that occur for a final
cause. This is because natural things - a tree, or a rock - possess no mind or
consciousness in which the shay'iyya of an intended end can exist before it
exists in the outside world as a concrete existent. Thus natural processes - a
tree's growth, or a rock's fall to the ground - will not operate according to final
causation, and final causation will be restricted to intentional acts. 3
Avicenna appears to be aware of the problematic implications of using
shay'iyya as the basis on which the priority of the final cause rests when he
appeals - again in his Tactiqat - to another Aristotelian assertion: that form and
end are often identical, particularly in natural phenomena. 4 In other words, the
completion of the natural process by which a form inheres in a properly disposed
matter can itself be regarded as a final cause, with the result that there is no need
to appeal to intentionality or consciousness:

L48
Ibn Sina, at-Taclfqiit, 128,17-25
The form is sometimes the same as the end, as in the case of
health: it is a form and it is the same as the end. The ends of natural
entities are the same as the existence of the form in the matter,
because an individual nature will move only in order for a form to
inhere in a matter.

But why has Avicenna not felt free to restrict the final cause to intentional action
and be done with it? Apart from violating a fundamental Aristotelian belief in
natural teleology, restricting the final cause's causality to intentional action
would also undermine almost universal Peripatetic assertions of the priority of
the science of final causality. As I mentioned in Chapter 6, Aristotle is quite
clear in holding that knowledge of the final cause is superior to that of the other
causes.5 Following him, Avicenna not only says that knowledge of the final
cause is the most excellent part of metaphysics (al-f:iikma), but implies that
teleology can be seen as its sum and substance.6
What is more, if final causation were restricted to intentional acts, Avicenna
would face a second problem: that of unrealized ends. In intentional acts a final
cause need not come into concrete existence for its effect to come into concrete
existence. Let us say, for example, that unbeknown to me my son has drunk the

3 For an example of this line of argument see ar-Riizi, Sharl:z al-ishiiriit, 194,1-6 (outside box).

Alexander also confronts this objection: ap. Simplicium, in Phys. 2.3 (ad 194b26), 310,25-311,37,
and 2.8 (ad 198bl6), 372,11 - 15
' I have discussed the nature of the identity between form and end at length in Chapters 3 and 6;
see Phys. 2.7, 198a25 (= Tabfca 137,20-138,I) (i.e., L9) and 198b3 (= 140,12-13); Metaph. 5.24,
1023a34 (= Mii ba'da f·fabfca II 655,9-10) and 8.4, 1044a36-bl (= /I 1074,1-2); and GC 1.7,
324bl3-18.
5 Metaph. 1.2, 982b5- l I. At the very least this is how Alexander of Aphrodisias understood the
above passage: in Metaph. l.2, 14,3-4; cp. in Metaph. 1.3 (ad 983a31-33), 22,7-14; 3.2 (ad 996b8),
184,21-4; 5.1(ad1013a7-8), 346,11-18; and 5.2 (ad 1013b3) 350,28-31.
6 Iliihiyyiit, 300,7-9.
176 Part II: The Avicennian Synthesis

bottle of soda in the refrigerator. My motion to the refrigerator will come into
concrete existence even if the bottle of soda is not there, that is, despite the fact
that my quenching will fail to come into concrete existence. In Mubiif:iatha 5
Avicenna seems to be grappling with this problem:

L49
lbn Sina, al-MubaJ:iathiit, 116,15-117,2
[MubiiJ:iatha 5, #277]
If it [the existence of the effect] were to come about as a result of
the thingness of something else whose existence is conceived, it
[what was conceived] would be a cause whether or not it existed [in
the outside world]; yet a thing's existence will never be causally
dependent upon that whose non-existence and whose existence is
all the same to it. As long as the cause of existence does not exist,
its effect will not exist. Were something to exist regardless of
whether some other thing existed or not, [the latter] would have no
effect in its existence in addition to the "effect" of straightforward
simultaneity [akthara min athari l-maciyyati s-siidhijati]; but
causality is more than simultaneity, even though it goes hand in
hand with simultaneity.7

In other words, if it makes no difference whether or not the final cause exists
concretely for its effect to exist concretely, the final cause will not fulfill the
basic criterion of causality: being that whose existence necessitates the effect's
existence. At best the final cause will be a cause only metaphorically.
Now Avicenna is clear that mental existence fully warrants being called
existence. So he could defend himself by saying that even in the case of
unrealized ends the final cause did exist and its existence did necessitate the
effect's existence; the final cause simply existed in the mind, not in concrete
reality. Nevertheless, the fact remains that in Mu<tazilite ontology the thing
(shay') which is in the mind alone qualifies as a non-existent (ma<dum), and not
as an existent (mawjud). Avicenna's use of shay'iyya in Iliihiyyiit 6.5 to
describe how the final cause operates could therefore be interpreted as implying
that an unrealized end - that is, a final cause which only ever "exists" in the
mind - is a non-existent thing (shay' ma<dum). And given that only an
existent, not a non-existent thing, can be properly spoken of as a cause, the final
cause will fail to satisfy the basic criterion of causality.
Avicenna seems aware of this problem in the Najiit passage on final and
efficient causes (i.e., L37). There he asserts that "an intentional object (al-
ma<na) has an existence in concrete reality and an existence in the soul as well
as something common [to both]: what is common [to both] is thingness."
Avicenna is implying that for thingness to serve as the basis on which the final

7 cf. 93,5-8 (= Mubii}J.atha 4, #177). Concern about how the final cause's possible non-

existence affects its causality is also evident in Awl)ad az-zaman Abu I-Barakat al-Baghdad!, K
al-mu'tabar fi 1-}J.ikma Ill, 52,12-53,5.
 9. Essence and Existence (C) 177

cause's priority rests, we will need to add the condition that the end must be
realized concretely. Only then will thingness link the intentional object's
existence in the mind with its concrete existence in the outside world. Things
which have a mental existence but no corresponding concrete existence -
unrealized ends, that is - will not be much use in pointing to the final cause's
priority.
It is my view that Avicenna moved from shay'iyya in Iliihiyyiit 6.5 and in
the Najiit to miihiyya in the /shiiriit precisely to skirt these various problems.
Avicenna starts using shay'iyya in order to pre-empt one Aristotelian problem -
the possibility of circularity in the relation between efficient and final causes -
and then later discards shay'iyya in favor of miihiyya because using shay'iyya
created the two further problems just discussed, each of which undennined the
primacy of the final cause.
Avicenna's chronological progression from shay'iyya to miihiyya can be
detected in his discussions of efficient and final causality, for his assertions about
the basis on which the final cause's priority rests change from shay'iyya and
f:zaqfqa in the Iliihiyyiit and the Najiit (middle period), to f:zaqfqa and miihiyya in
the Ishii.rat (late period). 8 This becomes even clearer when we look carefully at
the progression of Avicenna's thought within the middle period. According to
Gutas' careful reconstruction, Avicenna started writing the Kitiib ash-shifii' with
Tabfciyyiit 1 (i.e., as-Samiic af-fabfcf) but stopped after finishing only twenty
folios; then wrote Manfiq 1 (i.e., al-Madkhal); then completed Tabfciyyiit 1;
and then wrote the entire Iliihiyyiit. 9 After completing the rest of the Kitiib ash-
shifii' Avicenna wrote the Najiit.
The fact that Avicenna had already written the Madkhal but had not yet started
the Iliihiyyiit is clear from his discussion of the relationship between efficient
and final causality in Tabfciyyiit 1:

LSO
Ibn Sina, Kittib ash-shifti'fTablciyytit (I): as-Samac ar-!abrcr
1.11, 53,4-12
Avicenna, Liber primus naturalium: Tractatus primus de causis et
principiis naturalium, 95,2-19
The agent is, in a way, a cause of the end; how could it not be so,
when the agent is what makes the end occur as an existent? The end

a Michot l 997b has recently argued for an earlier dating of the /shiiriit, based on evidence
contained in Mubii/µltha 3. Given our current uncertainty about the Mubiil,iathiit - in particular, if
they constitute a single work from a single period or are, as it seems to me, a grab-bag of many
separate discussions recorded over Avicenna's lifetime - I feel Michot's conclusion remains
highly tentative, and the traditional late dating of the /shiiriit should be retained. Avicenna does use
the term shay'iyya once in the Man{iq of the /shiiriit (K. al-ishiiriit wa-t-tanbihiit, 15,6-8; shay'iyya
appears on line 8) but in the context of describing the function of the specific difference (al-f~l),
not in the context of final and efficient causality.
9 Gutas 1988, 101-112. Compare al-Jiizjiini's Prologue to the Shifii' in Man{iq (I), 2,13-3,17 and

his description of the composition of the Shifii' contained in his biography of Avicenna (lbn Sina,
Autobiography, 54,1-60,7).
178 Part II: The Avicennian Synthesis
is, in a way, the cause of the agent; how could it not be so, when
the agent acts only on account of it [the end]; otherwise, why
would it be acting? For the end sets the agent in motion towards
being an agent. For this reason, when it is said "Why do you
exercise?", and it is said "In order to be healthy", that will be an
answer; just as, when it is said "Why are you healthy?", and it is
said "Because I exercised", it will be an answer. Exercise is the
efficient cause of health, and health is the final cause of exercise
.... The agent is not a cause of the end's becoming an end, nor of
the end's essence in itself [wa-la li-miihiyyati 1-ghiiyati ft nafsihii
= nee ut finis habeat esse in se]; rather it is a cause of the existence
of the end's essence in concrete reality. The difference between
essence and existence is as you already know. The end is a cause of
the agent's being an agent, for it [the end, reading fa-hiya] is the
cause of [the agent's] being a cause, whereas the agent is not a
cause of the end in terms of [the end's] being a cause. This will be
made clear in First Philosophy.

Having taken into account all the new evidence from Chapters 8 and 9, I can
make the following additions (in bold) to my earlier chart (i.e., from Chapter 7)
of the progression in Avicenna's thought about thingness and essence:

(1) ljikma cAru<;iiyya, 2v8-10: "One" is a necessary accident of things (mina l-

acra<;ii l-lazimati li-l-ashya'i). Essence is a thing (bal takunu l-mahiyyatu
shay' an), be it a man or a horse, an intellect or a soul; that thing is only
subsequently characterized as being one or existent.
(2) Madkhal 1.2: The essences of things (mahiyyatu l-ashya'i) are sometimes
found in concrete objects in the outside world, and other times are conceived of
in the mind. But essence has three aspects: as a concrete, external existent; as a
mental, internal existent; and a third aspect, in which it is unrelated to either
concrete or mental existence.
(3) Madkhal 1.12: Genera and species may be divided into those which are
before a state of multiplicity (that is, those contained in the Active Intellect
and the other celestial intellects), those which are in a state of multiplicity
(that is, those contained individually in sublunary concrete existents), and
those which are after a state of multiplicity (that is, those contained as
abstracted universals in human intellects). Taken in itself a genus or a species
is a thing. "Animal", taken in itself, is an intentional object (macnan),
regardless of whether it is a concrete or a mental existent, or whether it is
general or specific (wa-laysafi nafsihi bi-cammin wa-la khii~~in).
(4) Tabi<iyyiit 1.11: The efficient cause causes the final cause to exist in
concrete reality, that is, to exist in an absolute or affirmative sense; the final
cause causes the efficient cause to exist as an efficient cause, that is, to exist
in a special sense.
(5) Ilahiyyat 1.5: "Thing" and "existent" are primary, indefinable categories.
Whatever is predicable of thing will also be predicable of existent, and
 9. Essence and Existence (C) 179
whatever is predicable of existent will also be predicable of thing. Although
they are co-implied, thing and existent have different meanings.
(6) Ilahiyyat 5.1: A universal (kulll) such as "horseness'', taken in and of itself
(ft nafsihi) - that is, without considering whether it is one or many (la
watzida wa-la kathira), a concrete existent in the outside world or a mental
existent inside the soul (la mawjuda ft l-a <yani wa-la mawjuda ft n-nafsi), in
potentiality or in actuality (la bi-l-quwwati wa-la bi-l-ficli) - is a thing
(shay').
(7) lliihiyyiit 6.5: The other causes are prior to the final cause in terms of
existence; the final cause is prior to the other causes in terms of thingness.
The other causes are the causes of the final cause's existence; the final cause is
the cause of the other causes' thingness. Thing is logically prior to existent
("the difference between thing and existent is just like the difference between
some entity and its concomitant" - 292,2-3); thingness is operative only when
existent in the soul, i.e., in intentional action.
(8) Jlahiyyat 7.1: "One" (al-watzid) and "existent" (al-mawjud) are equally
predicable of things (qad yatasawayani ft l-tzamli <a/a l-ashya'i); all that
may be characterized by "one" may also be characterized by "existent", but the
two terms do not have the same meaning as each other.
(9) Najiit: The other causes are prior to the final cause in terms of existence; the
final cause is prior to the other causes in terms of thingness; thingness is what
is common to both mental existence and concrete existence.
(10) Ishiiriit/Fi l-wujud wa-<ilalihi: The final cause is a cause, through its
essence, of the causality ("illiyya) of the efficient cause; the efficient cause is
a cause of the existence of the final cause.

Avicenna's use of mahiyya when discussing the final cause's priority in the
Isharat suffered from none of the obstacles strewn across the path of his earlier
uses of shay'iyya. Unlike shay'iyya, mahiyya was clearly identified with form,
so natural phenomena with forms but no intentionality could more easily be
accommodated in a universal teleology. Unlike a shay'iyya in the mind, which
smacked of the Mu<tazilites' non-existent thing, the mental existence of a
mahiyya was explicitly allowed for, so unrealized ends with mental but no
concrete existence could more easily satisfy the basic criterion of causality. And
unlike shay'iyya, mahiyya was clearly held to be logically prior to existence,
so the primacy of final causation could be more easily upheld. For all these
reasons Avicenna opted for mahiyya, and in so doing went some way to
resolving the tension between his inconsistent uses of the term shay°, and by
extension, between his concepts of thingness and essence.
Let me summarize what I have proven in Chapters 7, 8 and 9. I first argued
that the discussions of things and existents by the mutakallimun were the most
important source of Avicenna's distinction between essence and existence. This
is because Avicenna's position on how things and existents are related to each
180 Part II: The Avicennian Synthesis

other - at least as articulated in Iliihiyyiit I .5 - is a slightly modified version of

that propounded by the Sunni mutakallimun; the position of al-Farabi, by
contrast, is a slightly modified version of that propounded by the Mu<tazilites. I
then argued that the concept of "thingness" (shay'iyya) served to link the early
kaliim concept of "thing" (shay' ) with Avicenna's concept of "essence"
(miihiyya ). I showed that despite some textual indications otherwise Avicenna
uses the term shay'iyya to explain how the final cause is prior to the efficient
cause. I then argued that Avicenna's statements in Iliihiyyiit 6.5 and elsewhere
that the final cause is prior to the efficient cause in terms of thingness or
essence, while the efficient cause is prior to the final cause in terms of existence,
should be understood as a way to avoid the circularity that loomed behind
Aristotle's statement in Physics 2.3, 195a8-l l , that sometimes the efficient
cause is the cause of the final cause and the final cause is the cause of the
efficient cause, as health is the cause of hard work and hard work is the cause of
health. Finally, I argued that Avicenna's incompatible ideas about thingness and
essence, in the course of their application to the problem of distinguishing final
and efficient causality, approach resolution.
 10. Causal Self-Sufficiency vs. Causal Productivity

To the arguments just presented I shall now suggest that Avicenna's distinction
between essence and existence, and his assertions that the final cause is prior to
the efficient cause in tenns of essence or thingness while the efficient cause is
prior to the final cause in tenns of existence, should also be viewed as springing
from yet another problematic. Just as I argued that kaliim debates about the
relationship between "thing" (shayJ) and "existent" (mawjud') provided the
immediate tenninological and philosophical basis for Avicenna's distinction
between essence and existence, so I shall now argue that the Neoplatonists'
distinction between "well-being" (to eu einai) and "existence" (to einai) provided
the ultimate basis for Avicenna's distinction between essence and existence. And
just as I earlier argued that "thingness" served to link the early kaliim concept of
"thing" with Avicenna's concept of "essence", so I shall now also argue that the
concept of "perfection" (teleiotes = tamiim/kamiil) served to link the
Neoplatonists' concept of "well-being" (to eu einai) with Avicenna's concept of
"essence". And just as I earlier argued that Avicenna applied the distinction
between essence and existence to the relationship between final and efficient
causes in order to avoid the circularity that loomed behind Aristotle's statement
that sometimes the efficient cause was the cause of the final cause and the final
the cause of the efficient, so I shall now also argue that Avicenna applied the
distinction between essence and existence to the relationship between final and
efficient causes in order to avoid the circularity that loomed behind the
Neoplatonists' scheme of procession and reversion.
Although it is easy enough to see the transition from to einai to al-wujud,
Avicenna's main tenn for existence, less obvious is the route from the
Neoplatonists' to eu einai to miihiyya, Avicenna's main term for essence. As I
mentioned, I would like to suggest that the missing link between the
Neoplatonists' to eu einai and Avicenna's miihiyya was the concept of
perfection. How is this so? In Chapter 6 I showed that a substantial form can be
viewed as an element of procession, in which case it is both the formal cause of
the composite in which it inheres, and the effect of a higher efficient cause which
bestowed it on the composite. Alternatively, the substantial fonn can be viewed
as an element of reversion, in which case it is the final cause qua to hou of the
thing below it which is actively striving to attain it. The substantial form
viewed as an element of reversion is clearly what Avicenna had in mind when, in
his early treatise on (among other topics) procession and reversion, entitled The
Origin and the Destination (al-Mabda' wa-l-ma ciid), he identifies fonns
(,ruwar) with the perfections of bodies (kamiiliit al-ajsiim); when, in the Burhiin
of his ShifiiJ, he equates the essence of a thing with a thing's perfect fonn; and
when, towards the end of the lliihiyyiit of his Shifii', he asserts that reversion
182 Part II: The Avicennian Synthesis

(maciid) consists in the coming about of a thing's perlection. 1 These pieces of

evidence place Avicenna squarely in the Neoplatonic tradition described above.
What Avicenna does that is original is apply his own, new distinction between
essence and existence to the old Aristotelian problem of detennining the nature
of the complementarity between final and efficient causes. But Avicenna was
almost certainly led to use his essence/existence distinction in this way by the
Neoplatonists' own attempts to solve the complementarity problem by applying
Alexander's to einailto eu einai and ousia/teleiotes distinctions to procession
and reversion.
Let me explain. As I showed earlier, Avicenna held that the efficient cause is
the explanans of an explanandum insofar as the explanandum is viewed in light
of its existence in an absolute sense (al-wujud al-muf[aq) or in light of its
affirmative existence (al-wujud al-ithbiit'i). That is, the efficient cause is the
primary explanans when the explanandum is viewed as an existent (mawjud).
The final cause, by contrast, is the explanans of the very same explanandum
insofar as that explanandum is now viewed in light of its essence (miihiyya) or
thingness (shay'iyya); that is, the final cause is the primary explanans when
that very same explanandum is now viewed as having one type of existence as
opposed to another (al-wujud al-khii~~), as being one thing (shay') in contrast
to another thing.
When put in terms of procession and reversion, the efficient cause is now held
by Avicenna to be prior with respect to the downward procession of pure
existence from one thing to another below it. In other words, the efficient cause
is better suited to explaining a higher thing's passing absolute, undifferentiated
existence down to a lower thing. By contrast, the final cause is held to be prior
with respect to a thing's upward reversion towards its own well-being or
perlection. In other words, the final cause is better suited to explaining a thing's
striving to attain its substantial form or nature, or, more broadly, its striving to
attain the complete inherence of its essence (miihiyya) or "X-ness".
For example, the Active Intellect is related to me as an efficient cause is
related to its effect: by bestowing my substantial form - my humanity - to me,
the Active Intellect has given me existence in an absolute sense (or more
precisely, existence in an absolute sense has proceeded or issued from the Active
Intellect to me). In this respect the Active Intellect is the efficient cause of my
existence.
In another sense, the Active Intellect is related to me as a final cause (qua to
hou) is related to its effect: the perlect rationality that is possessed by the Active
Intellect and that signifies its etemality provides me with a model to imitate, and
thereby attain individual immortality, which is my ultimate goal. Again,
Avicenna's idea seems to be that the Active Intellect qua final cause best

1 lbn Sina, al-Mabda' wa-1-ma'ad, 78,14-15; K. ash-shifii'/Manfiq (5): al-Burhiin 4.5, 299,17-

18; and K. ash-shifii'!/lahiyylit (2) 9.7, 424,3.

 10. Self-Sufficiency vs. Productivity 183

explains my own reversionary struggle to perfect my humanity. That is, the

Active Intellect qua final cause best explains my striving to exist, as fully and
as well as I can, as a human: by being perfectly rational, by focusing entirely
on my rationality, which is the specific difference or essential characteristic that
sets my species apart from other species of animal such as "cat".
With this in mind, the Neoplatonists' move to distinguish between to einai
and to eu einai, and Avicenna's move to distinguish wujiid and mahiyya, were
similar in that both moves were driven partly by the desire to avoid a situation
in which the efficient cause and the final cause competed over the same
explanandum (or, more precisely, over the same aspect of a single
explanandum). What I mean is that if the efficient cause and the final cause are
simply different ways of explaining the same explanandum - the existence of
effect X - the final cause will inevitably lose to the efficient cause. This is
because the efficient cause enjoys a built-in advantage over the final cause:
whereas the efficient cause exists concretely in the extramental world before or at
the same time as its effect, the final cause comes into existence concretely in the
extramental world at the same time as or after its effect. Sometimes, when an
end is unfulfilled, the final cause never comes into existence concretely in the
extramental world at all: my motion to the refrigerator exists regardless of
whether or not I eventually fulfill the aim which first caused that motion to
exist, namely, quenching my thirst by drinking the can of soda that may or may
not be in the refrigerator. With the cards stacked in the efficient cause's favor, the
final cause will appear superfluous, and this will in turn undermine Aristotle's
and Proclus' many assertions about the supremacy of the final cause, a
supremacy that is most starkly evident in the fact both Aristotle and Proclus held
God, the ultimate cause, to be a final cause; in Aristotle's terms, the Unmoved
Mover, and in Proclus' terms, the Good.
If, on the other hand, the efficient cause and the final cause were not seen to
compete over the same explanandum - the existence of the effect - but were
instead seen to cause different explananda - or more precisely, different aspects
of the same explanandum - there would be complementarity rather than
competition between them. I could hold that the efficient cause was the primary
explanans of explanandum X insofar as explanandum X was viewed in light of
the downward procession of existence; while at the same time holding that the
final cause was the primary explanans of explanandum X insofar as
explanandum X was viewed in light of the upward reversion towards its well-
being, that is, towards explanandum X's perfection, towards perfect X-ness.
Holding that there is a complementarity between the way an efficient cause
explains its explanandum and the way a final cause explains its explanandum,
rather than holding that the efficient and final causes compete over the same
explanandum; and arguing that this complementarity is meaningful rather than
trivial because existence and essence are conceptually distinct; are, in my
opinion, Avicenna's attempt to come to grips with one of the major problems
184 Part II: The Avicennian Synthesis
facing late-antique and medieval philosophers who tried to defend the coherence
of Aristotelian teleology.
This is important partly because the question of whether or not Aristotle
really thought teleological explanation was essential to scientific explanation has
been the subject of fierce debate in recent Aristotle scholarship. 2 Nowadays the
charges that historians of ancient philosophy either level at Aristotelian
teleology or defend it against tend to revolve around the question of how
philosophically coherent Aristotle was in maintaining that the final cause is
compatible either with the material cause (generally understood by them as the
problem of compatibility between hypothetical necessitation and absolute
necessitation) or with the efficient cause (generally understood by them as the
problem of compatibility between teleological explanation and mechanistic
explanation). 3
To my mind these debates reflect post-scientific-revolution concerns that
would have appeared inappropriate to most Aristotelian philosophers from before
the scientific revolution. In other words, the modem assumption that
Aristotelian teleology was accepted blindly by most premodem philosophers
(until, that is, the scientific revolution taught us that it was more important to
answer the question "How?" than the question "Why?") confuses two issues.
Philosophers before the scientific revolution, though committed to teleology
generally, were worried about Aristotle's final cause, and particularly about how
it could be seen to be compatible with other forms of causation or explanation.
But the compatibility or complementarity problem that preoccupied most late-
antique and medieval philosophers was not one that was located in the first
instance in the sublunary world of coming-to-be and passing-away, such as the
problem of compatibility between hypothetical and material necessitation, or the
problem of compatibility between teleological and mechanistic explanation. The
compatibility problem that most worried Neoplatonizing Aristotelians such as
Avicenna a thousand years ago was one born of the larger project to harmonize
Plato's and Aristotle's cosmologies: the problem of compatibility between the

2 Following the taxonomy proposed by Hankinson 1995, 130-132, opinions on this issue can be

divided along the following lines, from weakest to strongest: (I) metaphorical teleology: those who
understand Aristotle as holding that the final cause is a cause only metaphorically, that it is a useful
heuristic device but one which does not reflect the way things are in reality, that natural things are
structured in such a way that it only appears as if they are directed towards some purpose; (2)
epistemological teleology: those who understand Aristotle as holding that the final cause is one part
of a full explanation of a natural thing, and thus accommodates our intuitions about nature's being
end-directed; but still, in reality, final causation will not resist being reduced to material causation;
(3) concrete teleology: those who understand Aristotle as holding that final causes are irreducible
potentialities for form, and that final causation is found to be operating in this way in the real
world. Proponents of (I) include Wieland 1975, and Nussbaum 1978, 59-99; proponents of (2)
include Sorabji 1980, 165-166; and proponents of (3) include Gotthelf 1976.
1 For various persepctives on this issue see Charles 1981, Boylan 1981, Lennox 1982, Bradie

and Miller 1984, Friedman 1986 and 1987, Cooper 1987, Balme 1987, Gotthelf 1989 and Sauve
Meyer 1992.
 10. Self-Sufficiency vs. Productivity 185

procession of being, in which the efficient cause plays the primary role, and the
reversion towards well-being, perfection and essence, in which the final cause
plays the primary role.

Given Avicenna's inheritance and adaptation of the idea of complementarity

between efficient and final causes, it should come as little surprise that like his
Ammonian predecessors, Avicenna thought that God was at one and the same
time efficient cause, the origin of all existence, and final cause, the good or end
of all essential perfection. Avicenna says as much in the lliihiyyiit of the Shifo.~:

LSI
lbn Sina, Kitab ash-shifli,llliihiyylit (2) 8.6, 355,6-356,5
Avicenna, Liber de philosophia prima sive scientia divina V-X,
412,55-413,78
The Necessary of Existence is perfect of existence [tammu l-wujudi
= perfectum esse], because He is not deficient in any part of His
existence or in any of the perfections of His existence [kamlilliti
wujudihi = perfectionibus sui esse]; nor does any part of the genus
of His existence fall outside His existence or exist on account of
something else, as is the case with other things such as "human"
(who is deficient in many of the perfections of his existence, not
to mention the fact that his humanity exists on account of
something other than him). In fact the Necessary of Existence is
above perfection [fawqa t-tamlimi =plus quam perfectum], because
not only does He possess the existence which He alone possesses,
but all [other] existence is an overflow that comes from His
existence, and is on account of it, and emanates from it. The
Necessary of Existence in itself is pure good, and good, on the
whole, is what everything desires. Now what everything desires is
existence, or rather, the perfection of existence of the type of
existence [which the thing has] [aw kamlilu l-wujUdi min babi l-
wujUdi =et perfectio esse inquantum est esse]; non-existence, on
the other hand, is not desired in so far as it is non-existence, but
rather in so far as existence or the perfection of existence follows
after it. So what is really desired is existence, and existence is a
pure good and a pure perfection [khayrun ma/:u.fun wa-kamalun
ma/:u.fun = bonitas pura et perfectio pura]. On the whole, the good
is that which each thing desires in respect of its definition and is
that by which its existence is perfected; evil, on the other hand,
has no essence [i.e., to be pointed to with a definition] but rather
is either the non-existence of a substance or the non-existence of
the substance's proper state. Existence is a [type of] goodness,
and the perfection of existence is the goodness of existence.
Existence which is accompanied by no non-existence (be it the
non-existence of a substance, or the non-existence of something
possessed by the substance) but which is instead perpetually in
act, is a pure good. The possible of existence in itself is not a pure
good, because in itself it does not necessarily entail its own
existence, and in itself it allows for non-existence; and whatever
186 Part II: The A vicennian Synthesis
allows for non-existence in any way will not be in every respect
free from evil and deficiency. Therefore there is no pure good other
than the Necessary of Existence in itself.

Unsurprisingly, Avicenna's works contain many assertions that the good is

equivalent to perfection, that God is "perfect of existence" (tiimm al-wujud), that
He possesses perfection or "pure perfection" (al-kamiil al-malµf), and that He is
only for the sake of His own perfection (li-ajli kamiili dhiitihi). 4 For my present
purpose, what is interesting about those assertions is the inference that if the
Necessary of Existence is to be the initial source of existence on the one hand,
and the ultimate good and perfection on the other; if He is to be both the
beginning of procession and the end of reversion; and if He is to have no end
above and beyond Himself; then He will need to be a final cause as well as an
efficient cause. In the Taclfqiit Avicenna is much more explicit:

L52
Ibn Sina, at-Taclfqiit
18,4: If the Necessary of Existence in himself is the agent, then he
is also the end and objective.
51,28: [In the First) there is no difference between end and agent.
62,3-5: Every end is a good, and the Necessary of Existence -
given that the end of whatever issues from Him is the absolutely
perfect good - is [Himself) the end of creation, since everything
terminates in Him, as He says [Qur0 lin 53:42) ["Does the apostate
not have Moses' and Abraham's knowledge) that the end point is
at your Lord?".
62,12-13: If the First is an end in terms of perfection and there is
no perfection above and beyond Him which may be related to Him,
then no perfection above His perfection is imaginable.
80,25-26: [The Creator) is the first and the last because He is agent
and end, His end being His self, and because [He) is the source of
everything and that to which everything returns.
160,5-6: One must know that He is Necessary of Existence, that
He is a principle, that He is a good, that he is an agent, that He is
an end, and that he is power.
178,25-26: In addition [to the agent), the end is a cause which is
distinct from the effect, although in the case of the Necessary of
Existence the end and agent are identical, for He is agent and end.

Although Avicenna's position that God is both efficient and final cause seems
like a natural consequence of his cosmology of procession and return, and is
perfectly in line with the tradition of the Ammonian synthesis, as described in

4 A thing' s good (nfki) is its perfection (kamiil): Diinishniima-yi <A/ii'f: /liihiyyiit, 117,7-9. The
Necessary of Existence is pure or absolute perfection: K. al-hidiiya, 262,2-3; ar-Risiilat al-
carshiyyafi J:iaqii'iq at-tawl,ifd wa-ithbiit an-nubuwwa, 21,2-4; and Diinishniima-yi cA/ii'f: lliihiyyiit,
148,9-10. God has no perfection or final cause other than Himself: ar-Risiilat al-'arshiyya ft
~qii'iq at-tawl,ifd wa-ithbiit an-nubuwwa, 22,18-20 and 29,3-4; aL-Mubiil,iathiit, 300,4-7 (=
Muba~tha 6, #840).
 10. Self-Sufficiency vs. Productivity 187

Chapter 3, many interpreters have accepted without question the view of the
Shicite philosopher and theologian a~-Tiisi (d. 1274), in his commentary on a
rather cryptic line in Avicenna's /shiiriit, that the Necessary of Existence is only
an efficient cause. 5
Partly as a result of a~-Tiisi's interpretation - or in my view, a~-Tiisi's
misinterpretation - the little scholarly attention that has been paid to Avicenna's
discussions of causality has focused on his theory of efficient causation, even
despite Avicenna's bald assertion, mentioned above, about the primacy of final
causality.6 This focus is symptomatic of a broader tendency among historians of
philosophy to reduce Neoplatonic cosmology to God's efficient causation of
existence, in order to contrast it with an Aristotelian cosmology in which God's
final causation of motion is paramount. As a result many historians of
philosophy have ignored Neoplatonic teleologies such as Proclus' and
Avicenna's, in which the final causation of existence - or rather, of "well-being"
or "special existence" (to eu einai = al-wujud al-khii~~. i.e., al-miihiyya) - is
meant to complement the efficient causation of existence (to einai =al-wujud).1

5 Na~ir al-din a\-Tiisi (d. 1274), SharJ:i al-lshariit, 194,13-17 (inside box); in his own corrunents

on !his passage (194,23-33, outside box) Fakhr al-din ar-Razi (d. 1210) had refused to corrunit
himself to naming one cause or the other. Avicenna's rather cryptic "Pointer" (lshiira) is "If there
is a First Cause, it will be a cause of every existence, and of the cause of the reality of every
existence in existence (in kiinat <illatun illii fa-hi ya cillatun li-kulli wujildin wa-li- ci/lati J:iaqiqati kulli
wujildin fl 1-wujildi)" (lshiiriit, 140, 10-11). Now in the three chapters immediately preceding this
one Avicenna distinguishes between causes of existence (the efficient and final causes) and
causes of essence (the formal and material causes), and also argues for the complementarity of
efficient and final causes (the efficient is the cause, in terms of its wujild, of the existence of the
final, and the final is the cause, in terms of its J:iaqiqa, mahiyya and ma<na, of the causality and
efficiency of the efficient) (lshariit, 139,14-140,9). In this light the line should be understood as
claiming that if there is a first cause, it will be a cause of existence (that is, it will be an efficient
and final cause), and it will also be the cause of the cause of essence (that is, it will be the cause of
the formal and material causes). The point is that given the proximity of these arguments to the line
in question (not to mention the explicit statements in the Tacliqiit, the clear implication of K. ash-
shifii,/lliihiyyiit 8.6 that as the Good, God is the culmination of all the universe's perfections; and
Avicenna's statement at the end of K. ash-shifti!/llahiyyiit 6.5 that the final is the highest of the
causes) 1 find it hard to understand how a close reader such as a\-Tiisi could argue so blithely that
God could not be a final cause because of the fact that the other causes precede it in existence
(reading bi-l-wujild for the text's bi-1-wujilb); with eternal things such as God, after all, there is no
temporal beforeness. The first half of Avicenna's apodosis must refer to God's being an efficient
cause, and the second half must refer to God's being a final cause; otherwise, there would be a
respect in which God's efficient causality was the effect of another cause, which is impossible. A
later commentator on the /shiiriit, al-Aghii J:Iusayn al-Khwiinsiiri (d. 1686), tries to correct what he
regarded as his predecessors' error on this issue, claiming that Avicenna simply meant to deny that
God was a panicular type of final cause, since the other causes are prior to the final not in an
absolute sense but only with respect to a certain category of final causes, namely, the non-essential
final causes found in the world of coming-to-be and passing-away; the earlier commentators
should be understood as denying merely that God was a final cause of that category, not as
denying that God was a final cause in an absolute sense: al-ljiishiya <a/ii shurilJ:i al-ishariit (2),
39,3-5.
6 See, for example, Gilson 1960 and 1962, and Marmura 1981 (esp. 65-72) and 1984b.
7 See, for example, Sorabji 1990. Even authors who take into account Neoplatonic teleology

tend either to ignore (e.g., Lovejoy 1936, 24-98) or to offer only cursory treatments of (e.g.,
I88 Part II: The A vicennian Synthesis

Apart from the question of whether or not God is a final cause as well as an
efficient cause, an important piece of evidence illustrating Avicenna's absorption
of Neoplatonic cosmology can be gleaned from LS I. Given the fact, discussed in
Chapter 3, that in Neoplatonic cosmology the perfect corresponds to or to
implies efficient causation, while perfection corresponds to or implies final
causation, it should be no surprise that in LS I, where Avicenna holds God to be
both the source of existence and the ultimate goal of all perfection, the extent to
which "perfect" and "perfection" may correctly be predicated of God is also
discussed.
In LSI Avicenna starts by saying that God is perfect (tiimm), in the sense that
He does not lack anything and is by implication causally self-sufficient. He then
adds that God is in fact "above perfection" (jawqa t-tamiim), by which Avicenna
seems to mean that God is not simply full of existence and hence causally self-
sufficient, but is also overflowing with existence, and hence a cause of others.
The rest of this chapter is devoted to examining the tension in Avicenna's
thought between these two conceptions of God, one as "perfect", complete in
Himself, self-sufficient and unrelated to any effect, the other as "above
perfection", overflowing with existence, and related to His effects; Avicenna's
sources for the two ideas; and Avicenna's attempt to side-step the problems
associated with them by using a new formula to describe God.
What precisely does Avicenna intend when he distinguishes between "perfect"
and "above perfection" in L51 ?8 Avicenna follows the Metaphysics translations
of teleion as tiimm and of telos as tamiim in his own discussion of the perfect,
lliihiyyiit 4.3, and begins his treatment of at-tiimm in a way that is quite similar
to Aristotle's treatment of to teleion in L4. However, Avicenna seems to follow
Euclid - and Euclid's commentator, Nicomachus - in the sense that Avicenna
thinks that numerical perfection (the quality by which a number is equal to the
sum of its divisors, as 6 is equal to 1 + 2 + 3) is the most primitive sense of the
term. This is in contrast to Aristotle, who introduces L4 with the idea of
quantitative completeness (the quality by which a whole is in possession of all
its parts).
Avicenna then proceeds, somewhat opaquely, to discuss the implications of
viewing numerical perfection as being the most basic sense of perfection.
Avicenna introduces the notion of fawqa t-tamiim first in this numerical context,
as equivalent to zii'id, "superabundant" (that is, a number the sum of whose
divisors is greater than itself), which suggests that he had Nicomachus' use of
huperteles in mind, even though Thabit b. Qurra, the translator of Nicomachus'

Blanchette 1992, 164 and 276-277) the contributions of late-antique philosophers such as Proclus
and Arnrnonius, let alone the contributions of Arabic philosophers such as al-Fiiriibi and Avicenna.
8 Avicenna also uses the phrase tiimm bal fawqa t-tamiim at at-Ta<liqiit, 16,26-27. At

Diinishniima-yi cAlii'i: /liihiyylit, 108,10-11, Avicenna says that the Necessary of Existence is
"most perfect" (tamiimtarin); at 116,9-117,5, Avicenna gives a synopsis of the distinction between
deficient, sufficient, perfect and above perfection.
 10. Self-Sufficiency vs. Productivity 189

Introduction to Arithmetic, rendered huperteles into Arabic as zii'id calii t-

tamiim rather than fawqa t-tamiim.9 Avicenna then abruptly dismisses the
numerical discussion he has just had as nothing but the silly notions of common
people:

L53
Ibn Sina, Kitiib ash-shifii'l/liihiyyiit (1) 4.3, 188,3-189, 11
Avicenna, Uber de philosophia prima sive scientia divina I-IV,
215,13-217,47
Having now brought [you] to this point, let us leave it behind,
since we are not accustomed to talking about things such as this,
which are based upon rhetorical conjectures and do not derive from
the methods of scientific syllogisms. So we say instead:
Philosophers have also extended "perfect" to cover the reality of
existence [ilii flaqiqati l-wujudi = ad certitudinem essendi]. In one
sense, they say, the perfect is that whose characteristic is not to
lack anything by which its existence is perfected [inna t-tiimma
huwa lladhi laysa shay'un min sha'nihi an yukmala bihi wujUduhu
bi-mii laysa lahu = (lacuna in Latin)]; on the contrary, everything
such as this has come to be contained in it. In another sense they
say that the perfect is that which has this characteristic but with
the condition that its existence, in itself, is in the most perfect
state it can possess [calii akmali mii yakunu lahu = quantum
perfectius esse potest]; that it alone has it [this existence]; that
there is no part of it which it does not have; that in the genus of
existence nothing superior to it is related to it in any primary way;
and that it is not on account of anything below it. "Above
perfection" [fawqa t-tamiimi =
Plus quam perfectum] is whatever
has the existence which it needs to have, and from which there is
existence left over for other things. It is as if it has the existence
which it needs to have as well as some extra existence which it
does not need and which is left over for [other] things, this being
something that comes from itself.
They then posited that this rank - namely, being that which is
above perfection [fawqa t-tamiimi =ultra perfectionem] - applies
to the First Principle, given that in itself and not on account of
anything else, a part of its existence is to emanate existence from
its own existence to all things. They held that the rank of
"perfect" belongs to a separate intellect [Ii- caqlin mina l- <uquli l-
mufiiraqati = intelligentiae ei quae ex intelligentiis separatis]
which is untouched by anything potential at the origin of its
existence in act, and which is not mindful of any other existence
which exists as a result of it; for in fact everything other [than it]
comes from the existence that issues from the First. They held that
below the perfect were the sufficient and the deficient [wa-ja <alu
dUna t-tamiimi shay' ayni l-muktafiya wa-n-niiqi~a = sufficiens et
insufficiens]. The sufficient is what is given [the ability] to
produce its own perfection. In a strict sense the deficient is what

9 Niqiimakhis al-Jarasini, K. al-madkhal i/ii <i[m al-'adad, 36,11; 37,5.15; 40,14.

190 Part II: The A vicennian Synthesis
needs something else to extend one perfection after another to it.
An example of the sufficient is the rational soul which belongs to
the universe, I mean the heavens. For in itself it [the rational soul]
performs the activities which pertain to it and causes the existence
of the perfections it needs to possess in a piecemeal fashion,
without all of them being brought together at once, and without
their remaining in perpetuity; otherwise those perfections which
are in its substance and form, would not be distinct from what is
potential, even though they contain within themselves a principle
which makes them emerge from potency into act, as you will learn
later on. As for the deficient, it is like the things which are in the
world of coming-to-be and passing-away.

Avicenna's frustration at the beginning of the passage could be directed at the

Neoplatonists' inconsistency over the question of what constituted a perfect
number, which differed depending on the context in which perfect number was
introduced. 10
In any case, Avicenna claims that what is above perfection is whatever has all
the existence it needs, and then some extra existence left over to spill onto
things below it. This obviously refers to God. That which is perfect in existence
is whatever does not exist on account of what is below it, and refers to the
celestial intellects. What the sufficient refers to is slightly unclear: we are told
that it refers to any of the celestial souls, since it has it in itself to attain its
perfection - by causing its sphere's eternal circular motion, one assumes -
without requiring the help of anything outside. In other words, the sufficient has
a potentiality which it is able to exercise without relying on external things,
whereas the perfect has no potentiality whatsoever. The deficient clearly refers to
sublunary things, things like horses and trees which are subject to coming-to-be
and passing-away.
I think that Avicenna's debt in this passage to the Neoplatonists is undeniable.
Avicenna has applied Neoplatonic, and particularly Prodan, ideas - discussed in
Chapter 3 - of to autarkes to the perfect, and of to authupostaton to the
sufficient, in an effort to create a clearer distinction between the ways in which
things are spoken of as being perfect. Just as Proclus had done, Avicenna applied
the concepts of to autarkes and to authupostaton to the various inhabitants of

10 As I mentioned at the end of Chapter 4, figuring out the perfect number of transmigrations of

the soul, as well as the perfect number of revolutions of the universe, was a Neoplatonic
preoccupation. The Neoplatonists seemed unable to decide on whether to understand perfect
number in the Euclidean sense (that is, as the sum of its divisors, as 6 is the sum of I, 2 and 3), or in
a more mystical, Pythagorean sense (with the number JO usually taken to possess true perfection).
The Euclidean sense is in the foreground at Iamblichus, Theo!. Arithm., 42,19 and 17,13; and
Proclus, in Parm., 767,30-32; and in Tim., Vol. 3, 168,17-20. The Pythagorean sense is in the
foreground at Syrianus, in Metaph. 13.8 (ad 1084a29), 149,27; Proclus, in Remp. , Vol. 2, 81,9 and
121,4 (where Porphyry is said to have claimed that 12 represents perfection); Theo!. Plat., Vol.
4.29, 87,7-8 (citing Resp. VIII 54683-4); in Tim. , Vol. 3, 94,32-95,6; and Asclepius, in Metaph. 1.5
(ad 985b23), 37,14-15 and 1.8 (ad 989b29), 65,18-19.
 10. Self-Sufficiency vs. Productivity 191
the superlunary realm partly in order to find some way round the problem of
differentiating between things which are all eternal and actual and, in that sense,
perfect. 11 Avicenna also seems to have followed Iamblichus' and Proclus'
attempts to Pythagoreanize (by which I mean their attempts to assign
ontological and cosmological value to categories of number) Euclid's and
Nicomachus' distinction between superabundant, perfect, and deficient number. 12
In short, Avicenna has used the concept of "that which is above perfection" in
order to distinguish the causal self-sufficiency of the Intellects, which is not in
itself productive of the existence of anything else, from the causal self-
sufficiency of God, which is in itself productive of the existence of other
things.13 The idea that God is above perfection is not found in Metaphysics 12,
where God and the celestial substances or movers (Aristotle never uses intellects
in the plural in this context) are often described as being completely actual, but
never as being above actuality. 14 Yet in the Arabic Liber de causis, a
paraphrase of sections of Proclus' Elements of Theology, but attributed to
Aristotle and called Kitab al-i<;lal:z ft l-khayr al-mal:z<;l li-Aristutalfs (Aristotle's
Book on The Exposition of the Pure Good), God is described as being above
perfection:

L54
(Ps.-)Aris~ii~iiJis, Kitab al-lf}af:z fl l-khayr al-mal:zfi li-Aris!fl!alls
21, 22,12-23,5 15
corr. Proclus, Inst. Theo/. 115, 100,28-102,12
The first cause is above every name predicated of it. This is in view
of the fact that deficiency is not worthy of it, nor even perfection
alone. [This is] because what is deficient is not perfect, and since
it is deficient it is unable to carry out a perfect act; and what is
perfect <in our opinion> - although it is self-sufficient [wa-in
kana muktafiya bi-nafsihi] - is unable to originate anything else,
nor to emanate from itself anything whatsoever. If this is so, we
shall return and say that the first cause is not deficient, nor perfect

11 In addition to the Prodan ideas of to autarkes and to authupostaton, the idea of "the self-

perfect" (to autoteles) - referring to something which has it in itself to come to perfection - also
suggests itself as an antecedent to Avicenna's muktafin and tiimm; cf. Proclus, Inst. Theo/., Prop.
64, 60,25-31.
12 See, for example, Iamblichus' distinction between huperteles, teleion and ate/es, at Theo/.

Arithm. 18.17; and Proclus' distinction between "the full" (to pleres) and "that which is above
fullness" (to huperpleres) at Inst. Theo/., Prop. 131, 116,15-27.
13 Al-Farabi uses mu/4afin at K. as-siyiisat al-madaniyya, 53, 11-54, l I, but differently from

Avicenna: al-Farabi's muktafin refers to the fact that celestial substances in general are sufficient
to produce - i.e., capable of producing - other things; compare al-Farabi, R. fl al«aql, 33,5-11,
where he argues that the Active Intellect cannot be God because the Active Intellect does not
possess the self-sufficiency (kifiiya) to be its own cause, as a result of the unreliability of receptive
sublunary matter and intellect always to receive its eternal emanation.
14 Though Asclepius distinguishes between the celestial Intellect, which is completely actual

and has no potentiality, and the First Cause, which is above actuality (huper energeian): in Metaph.
3.1~ad996a9), 148,30-34.
1 fawqa t-tamiim appears at 23,l; <ms b adds what is between brackets>.
192 Part II: The Avicennian Synthesis
alone; rather, it is above perfection [fawqa t-tamami] <because it
originates things and emanates goods to them in a perfect
emanation> because it is an infinite and inexhaustible good.

Although the terms huperousios ("above being"), huperzoos ("above life"),

hupernous ("above intellect") and autoteles ("self-perfecting"), all appear in the
corresponding Greek chapter of Proclus' Elements of Theology, the term
huperteleion does not; nor does it appear anywhere else in the Elements, as far
as I can tell. The term huperteleion can be found, however, in Proclus'
commentary on Plato's Timaeus, though applied in a different way. 16 I do not
know whether the abridger of the Liber de causis applied the phrase fawqa t-
tamam with Proclus' in Timaeum usage in mind, or whether he invented it in
light of the other "hyper"-qualities described in the Elements. In any case the
phrase also appears - far more often than in the Liber de causis - in Mlmar IO
of the Uthulujiya, sometimes on its own as fawqa t-tamam, other times as a
hendiadys with kamiil:

LSS
(Ps.-)Aris!ii!iilis, Uthulujiya JO, 134,16-135,1217
corr. Plotinus, Enneads V .2.1,4-8
I say too that the absolute One is above perfection [fawqa t-tamami
wa-l-kamali]. The sensible world is defective because it is
originated from the perfect thing, which is mind; mind is perfect
because it is originated from the true absolute One, which is above
perfection. It is not possible that the thing which is above
perfection should originate the defective thing directly, nor is it

1 ~ Proclus, in Tim., Vol. 3, 10,22-11,4. The reason why Proclus uses huperteleion in this context

is different from Avicenna's reason for usingfawqa t-tamiim in L53. In his Timaeus commentary,
Proclus is discussing the nature of time, and is trying to distinguish Eternity (ho aion) from Living-
in-Itself (to autozoion): Eternity is superior to Living-in-Itself, Proclus says, for various reasons;
what then to do with Tim. 3002, where Plato says that to autozoion is "the most beautiful of the
objects of intellection and perfect in every respect (kai kata panta teleion)"? Well, Proclus argues,
most beautiful does not mean best, and Eternity is best. As for the latter qualification, that which is
"perfect in every respect" is not necessarily first and foremost (protiston); for "the perfect"
possesses all things inasmuch as it possesses first, middle and final parts. Now, whatever is above
this divisibility into beginning, middle and end, will therefore be "above perfection" (epeita de ouk
anangke protiston einai to kata panta teleion; to men gar teleion panta ekhei, h6ste kai prota kai
mesa kai teleutaia; ho de estin huper tauten ten tomen, huperteleion an eie). Proclus' argument
here corresponds to his distinction between (I) the type of perfection that is prior to (meaning
above) being divided into parts; (2) the type of perfection that wholes possess, that is, the
perfection which is made up of parts; and (3) the type of perfection that each of the parts making
up a whole possesses: Proclus, Theo/. Plat. 4.25, 74,6-75,2 and 75,21 -76,13. This distinction is
articulated elsewhere as one between different types of "wholeness" (holotes): Inst. Theo/., Prop.
67, 64,1-14; and Theo/. Plat. 3.25, 87,26-89,2. It would be tempting to postulate some kind of
filiation between Proclus' tripartite distinction between types of perfection or wholeness, and
Avicenna's tripartite distinction - in Madkhal 1.12, mentioned in Chapter 7 - between types of
universal, but upon closer examination, the distinctions appear to have little in common apart from
beinr tripartite.
1 fawqa t-tamiim appears at 134,16 and !35,2(bis).5.7.8.9.l !. As with L7, the translation, with

some variations, is that of Lewis 1959, 291-2.

 JO. Self-Sufficiency vs. Productivity 193
possible for the perfect thing to originate anything complete like
itself, because origination is deficiency, by which I mean that the
originated is not of the rank of the originator but is beneath it.
The proof the absolute One is perfect and above perfection is that
it has no need of anything and does not seek to acquire anything,
and because of the intensity of its perfection and superabundance
another thing is produced from it. For the thing which is above
perfection cannot produce unless the thing be perfect: otherwise it
is not above perfection. For if the perfect thing produces
anything, then a fortiori the thing which is above perfection
produces perfection, because it produces the perfect thing than
which none of the things produced can be more powerful, more
splendid or more sublime. For when the true One which is above
perfection originates the perfect thing, that perfect thing turns to
its originator and casts its gaze on it and is filled with light and
splendor from it and becomes mind.

As was the case with L54, the Arabic phrase fawqa t-tamam in L55 does not
correspond to an instance of huperteleion in Plotinus' Greek. But since by
Plotinus' reckoning intellect is perfect, and because intellect looks up at what
originates it, it would not have been too much of a stretch for the compiler of
the UthUlUjiya to paraphrase him as holding that what originates mind must be
above perfection.
It is hard to avoid the conclusion that all the conceptual fine-tuning undertaken
by philosophers from Proclus to Avicenna in order to create a coherent scheme
of differentiating between the various types of perfection and self-sufficiency
enjoyed by superlunary beings, cried out for a clear, simple and watertight way
to distinguish God from other perfect, self-sufficient and eternally actual things.
The problem is that making such a clear, simple and watertight distinction forces
a Neoplatonic philosopher to decide which is more basic: God's transcendence of
the world or His involvement in the world. If I - as a Neoplatonist - were to opt
for transcendence, I should naturally be drawn to emphasizing God's perfection:
His self-sufficiency, His lack of nothing, and His not requiring any cause to
sustain His existence. If, on the other hand, I were to opt for God's involvement,
I should naturally be drawn to emphasizing God's being above perfection: His
productivity, His being related to the world as its cause. Is there any way for me
to strike a balance between God's transcendence and God's involvement and
thereby avoid having to chuck out one in favor of the other; apart from making
the fence-straddling assertion - as Avicenna did - that God is on the one hand
perfect and on the other hand above perfection?
This dilemma came to a head because Avicenna, like other philosophers before
him in the Ammonian synthesis, was committed to seeing God as both efficient
and final cause. But since God's perfection, the ultimate goal of reversion, acted
on its effect as a final cause acts on its effect, while God's being above
perfection, the initial source of procession, acted on its effect as an efficient
194 Part II: The Avicennian Synthesis

cause acts on its effect, Avicenna was forced to decide whether reversion
culminates in the celestial Intellects or in God. On the one hand, the Intellects
are perfect, Avicenna says, and hence possess perfection; therefore reversion
should culminate in them, or in the highest of them. On the other hand, God is
both perfect and hence possesses perfection, so reversion should culminate in
Him; yet God is also above perfection, so reversion should not quite reach Him.
Avicenna's ambivalence about whether God as well as the Intellects, or the
Intellects alone, should be regarded as being perfect and possessing perfection is a
sign that like Proclus before him, he was tom between viewing perfection and
the perfect as univocal terms or as equivocal terms. Avicenna seems to be
struggling with this uncertainty when he says in the TacllqiU:

L56
Ibn Sina, at-Ta'lfqiit, 62,11-24
Imagining one perfection coming above another [kamiilun fawqa
kamiilin] in those things that possess perfection, and the
dissimilarity each has compared with others, is possible only in
respect of the relation the perfections have to complete perfection
[ilii l-kamiili t-tiimmi], since the dissimilarity is [only] with
reference to that [complete perfection]. If the First is an end [or
"limit"] in terms of perfection [in kiina 1-awwalu ghiiyatan ft l-
kamiili] and there is no perfection above Him which may be related
to Him, then no perfection above His perfection will be
imaginable.

In other words, Avicenna, like Proclus, could not decide if a single basic
meaning of "perfection" and "perfect" ought to be applied consistently to all
subjects of which those terms are predicated - to sublunary souls and superlunary
souls, to sublunary intellects and superlunary intellects, to sublunary bodies and
superlunary spheres - or if, on the other hand, each subject of which "perfection"
and "perfect" is predicated should be understood as being a perfection or as being
perfect in different ways: as referring, for example, to activity or to actuality, to
efficient causality or to final causality, to numerical wholeness or to self-
sufficiency. Proclus and Avicenna devoted a great deal of effort to showing that
all those apparently different meanings of perfection and the perfect in fact
implied each other, or were reducible to one or the other of them. The effort to
determine what Aristotle scholars would call a "focal meaning" for perfection and
the perfect began, of course, with Alexander's comments on L4, as I showed in
Chapter 2. Nevertheless, it is hard to avoid concluding that as a result of their
very mutability - and hence their usefulness both with regard to the lesser
sumph6nia of reconciling Aristotle with himself and with regard to the greater
sumph6nia of reconciling Aristotle with Plato - the terms "perfection" and
"perfect" came to be applied to so many different things in so many different
ways that the terms were eventually rendered trivial.
 JO. Self-Sufficiency vs. Productivity 195

I believe that it is partly as a result of Avicenna's frustration with the fact that
perfection and the perfect had become such blunt philosophical tools, so
incapable of chiseling the fine distinctions required to differentiate between
various eternal things such as God and the Intellects, that Avicenna's new idea of
wajib al-wujud bi-dhatihi, "that which, in itself, necessarily exists'', grew to
enjoy such prominence in the middle and late parts of his career. Fleshing out
the philosophical background to Avicenna's concept of necessary existence -
actually a matrix of distinctions, between necessary existence and possible
existence, between in itself and through another, between uncaused and caused,
and between eternal and originated - and tracing its evolution in Avicenna's
works is my task in the next four chapters.
 11. Necessity and Possibility (A)
Materials from the Arabic Aristotle

In Chapters 7 to 10 I argued that while Avicenna's distinction between essence

and existence appears to have originated in Maturidite and Ashcarite kaliim, the
fact that he applied his distinction to the Aristotelian problem of the
compatibility between efficient and final causes is best understood in light of
intervening Neoplatonic attempts to construct a meaningful complementarity
between the procession of existence from an efficient cause to its effect, and the
reversion towards perfection of an effect to its final cause.
I also showed that several Neoplatonists, including Proclus, perhaps drawing
from Aristotle's discussions of happiness in the Nicomachean Ethics, claimed
that the idea of self-sufficiency- autarkeia in Greek, istighnii' in Arabic - was
implied in the idea of perfection; and furthermore that self-sufficiency should be
primarily construed to refer to uncausedness, and hence to a kind of existential
necessity. I then showed that Avicenna was unhappy with the idea of God's
possessing only perfection, because while that could explain His causal self-
sufficiency (that is, His freedom from being caused to exist by other things) it
could not explain his causal productivity (that is, His causing other things to
exist). For this reason Avicenna borrows from the Arabic Neoplatonica and
asserts that God is both perfect and above perfection (jawqa t-tamiim).
Avicenna's discomfort with the dualism - conceptual, if not real - that results
from holding that God is both final and efficient cause, and from holding that He
can be seen as both possessing perfection and being above perfection, led him to
search for an alternative to the old paradigm which had dominated metaphysical
speculation throughout the period of the Ammonian synthesis. The question to
be answered in the next four chapters is, how and why did Avicenna come up
with his alternative paradigm? More specifically, I shall examine Avicenna's
construction of a matrix of distinctions between "necessary of existence" (wiijib
al-wujild) and "possible of existence" (mumkin al-wujild), and between "in
itself' (bi-dhiitihi) and "through another" (bi-ghayrihi); his integration of older
distinctions - between uncaused and caused, and between eternal and originated -
into that new matrix of distinctions; and finally his use of that new matrix of
distinctions to confront the old challenge of reconciling one's assertions about
what God is with one's assertions about how God causes the world.
A number of scholars have treated this topic before, which is unsurprising
given its centrality in Avicenna's thought and its influence on subsequent
thinkers. Goichon views Avicenna's matrix of distinctions as basic, yet treats it
almost as a corollary to Avicenna's distinction between essence and existence.'

1 Goichon 1937, 156-180.

198 Part II: The Avicennian Synthesis

Davidson subjects it to a penetrating analysis, detecting two ways Avicenna's

matrix of distinctions is put to use in proofs of God's existence, one articulated
in the Najat and the /sharat, the other in the Shifa' (explicitly) and the
Diinishniima (implicitly). 2 But because Davidson's main concern is determining
how the question of the world's eternity played out in proofs of God's existence,
he is able to group the Avicennian texts together in this way, and thereby does
not address the question of whether there might be any evolution in Avicenna's
thinking about this issue. Nor does Davidson examine Avicenna's treatments of
the matrix of distinctions, and his application of it to the God-world
relationship, in early works such as the /fikma cAru{iiyya and the Mabda' wa-
maciid, both of which precede the Shifii', Najat, Danishnama and /sharat, and
from both of which Avicenna copied whole passages to paste into the pages of
his Najar. In focusing exclusively on works from Avicenna's middle and late
periods, Davidson follows Goichon, as well as other scholars who have weighed
in on the topic. 3 As far as I know Gutas is the only scholar who has raised the
issue of a possible evolution in Avicenna's thinking about necessary and
possible existence, and called for further research on it. 4
To a great extent the next four chapters revolve around that developmental
question. I shall show that there is richer evidence of philosophical evolution
from the works of Avicenna's early period (namely, the /fikma cArii.{iiyya and
the Mabda' wa-ma<ad) to the works of Avicenna's middle period, than there is
of a philosophical development from the middle-period works to the late-period
works. But the fact that the early-period works have been ignored (apart from
Gutas' brief discussion) insofar as they bear on the issue of the development of
Avicenna's matrix of distinctions and his application of that matrix to the God-
world relationship, has prevented us not only from determining whether or not
an evolution in Avicenna's thought on this topic can be detected. Our relative
ignorance of the early-period works hinders us from determining precisely the
sources Avicenna drew on when he first articulated the matrix of distinctions, and
first applied that matrix to the God-world relationship. Davidson has made some
useful sl.lggestions as to possible sources for Avicenna's discussions. But
without knowing how Avicenna first conceived of this issue, they remain
tentative. Before showing how Avicenna's thought evolves from his early period
to his middle and late periods, and showing how that evolution is linked to a
contemporaneous progression in his thought about essence and existence, I shall
present the lfikma cAru¢iyya passage in which he first formulates his matrix of
distinctions and applies them to the God-world relationship, and take stock of the
sources that appear to have shaped that very early discussion.

2 Davidson 1987, 281-310 and 350-361.

3 Hourani 1972, 74-75, claims that no progression is detectable; see also Hyman 1987.
4 Gutas 1988, 261-265.
 11. Necessity and Possibility (A) 199

In his autobiography Avicenna famously claimed that at the age of 18,

following an 18-month period of intensive philosophical study:

LS7
Ibn Sina, Autobiography
30,5-6: All that I knew at that time is just like what I now know:
until today I have not become greater in respect of it [i.e., in
knowledge] at all.
36,8-38,2: When I reached my eighteenth year I was done with all
these sciences. And while at that time I had a better memory for
[such] knowledge, I am more mature today; otherwise the
knowledge [itself] is one and the same thing, nothing new having
come to me afterwards [i.e., after the age of 18].

Avicenna's assertion is contradicted by the evidence of a progression in his

thought from shay"iyya to mahiyya, which I described in the previous chapter,
as well as by recent research on other aspects of his philosophy which lend
themselves to developmentalist interpretations. 5 So what could Avicenna have
meant? Gutas has already attacked the tendency of earlier interpreters to ascribe
Avicenna's assertion to an intellectual arrogance and immodesty supposedly
common among "Orientals".6 Gutas reasons that Avicenna must instead have
been referring to his theory of intuition, which by Gutas' reckoning is the "core
theory" of Avicenna's philosophical system. 7
My guess, however, is that Avicenna was referring to a metaphysical rather
than an epistemological insight: his distinction between necessary and possible
existence, and specifically his equation of that which is necessary through
another and that which is possible in itself. What I mean is that when, at around
the age of 50 (the age he composed the Autobiography) Avicenna looked back at
his philosophical career, and thinking of himself (to use Isaiah Berlin's
distinction) as a hedgehog rather than a fox, he had to admit that his big idea -
the idea which he knew was his most original and which he reckoned would be
his most influential - came to him at the very young age of 18. That is not to
say that this idea did not evolve in his subsequent thought; in fact I hope to
show in the next four chapters that his distinction between necessary and
possible existence did mature in subtle but important ways. Rather, what
Avicenna meant is that after the age of 18 he had no other ideas which were both
as big and as original as that one.
There is unequivocal evidence that Avicenna came to his distinction very early
in his career. Immediately after the second assertion above in L57, Aviceima
goes on to describe his first commissioned summa, The Compilation (al-
Majmu <a), which he composed for a neighbor of his in Bukhara, Abi.i 1-l:lasan

5 See, for example, Gutas 2001 and Hasse 2001.

6 Gutas 1988, 159-160, esp. 159n.28.
1 Gutas 1988, 160-176.
200 Part II: The Avicennian Synthesis

al-c Aru<;li, and which is known as al-/fikmat al- cAru{iiyya, or The cAru{iian
Philosophy. Avicenna asserts that he was 21 when he finished it, that is, around
the year 1001. In the section on metaphysics - which, as I noted in Chapter 6, is
modeled after Metaphysics 5, Aristotle's philosophical lexicon - Avicenna
devotes a paragraph to the necessary:

LSS
lbn Sina, al-ljikmat al-<Aru{iiyya, 3vl6-4rl2
"The Necessary" [al-wajib] is that which is necessary of existence
in respect of the way it is [a{i-{iaruriyyu 1-wujadi <afii ma huwa
<alayhi], this being either in itself [bi-dhatihi] (such as the
principle of existents [ka-mabda'i 1-mawjudati]) or through
another (such as the fact that two and two are four). The necessary
is either eternal (such as the principle of existents) or is at one
time and not another lfi J_ialin dUna J_ialin] (such as the eclipse of
the moon at that moment). All that is necessary of existence in
itself has no cause. Whatever has a cause will be neither necessary
of existence in itself nor impossible of existence in itself (for
otherwise, it would never come to exist); with respect to itself [fi
haythu dhiitihi], therefore, it [i.e., that which has a cause] will be
possible of existence [mumkinu l-wujudi], while [at the same time]
being necessary of existence through its cause. Now the existence
of whatever has no cause is not itself divisible into two states [wa-
ma la cillata lahu fa-inna wujada dhiitihi la yanqasimu min
J_ialatayni], in virtue of which it [i.e., the existence of whatever
has no cause] would come to be caused in both states; for there
would be no way out of being caused, nor any escape from
causedness. All that is subject to change is in these two states,
neither one of which it possesses in itself; rather, it possesses
both of them through a cause (there being no alternative to them
[i.e., there being no alternative to being in the two states]). Thus
all that is subject to change will itself be caused and possible,
whereas all that is necessary of existence in itself will be
necessary of existence in every respect, and no type of change
whatsoever will be attributable to it. "The possible" is the
existent which is not necessary [laysa bi-{iaruriyyin]. "The
possible" is said to be whatever is not impossible; and "the
possible" is said to be whatever is not impossible and which
exists and [then] is non-existent, and [in general] whatever does
not exist at some time or other.

I shall defer until Chapter 14 explaining what Avicenna could have meant in this
passage, and how the passage ought to be seen as a first, slightly incoherent
attempt to formulate his new matrix of modal distinctions, one which he
improved on in subsequent works such as the Mabda' wa-ma cad, the /lahiyyi:lt
of the Shifa", and the lsharat wa-tanbihat. My task now is to provide an
inventory of the sources which Avicenna drew from to express the new ideas he
articulates in the passage above. The sources include discussions of necessity and
 11. Necessity and Possibility (A) 201
necessitation contained in the Arabic Aristotle, which I shall discuss in this
chapter; the digested versions of those discussions contained in al-Farabi's
works, which I shall discuss in Chapter 12; and discussions of the etemality of
God's attributes contained in tenth-century kaliim works, w_hich I shall discuss in
Chapter 13.

Aristotle speaks at length about necessity in various contexts, but nowhere more
canonically than in Metaphysics 5.5, the chapter of his philosophical lexicon
on the necessary. Since the different ways Aristotle understands and applies the
idea of necessity in other works can, with a little effort, be subsumed under the
categories he sets out in Metaphysics 5.5, I shall begin my discussion with that
chapter:

L59
Aristatalis, Ma bacda !-fabFa II, 515,7-517,7
Aristotle, Metaph. 5.5, 1015a20-bl5
[1] The necessary is that without which a thing cannot exist, such
as the cause [al-mu¢farru huwa lladhl laysa yumkinu an yakiina sh-
=
shay' u ilia bihi wa-huwa ka-l- cillati anankaion legetai hou aneu
ouk endekhetai zen hOs sunaitiou], just as breathing and
nourishment are with respect to the animal: they are necessary
since the animal cannot exist without them. (2] Next, the
necessary is that without which a good, or the negation or non-
existence of an evil, cannot exist or come to be; such as when it is
said that ingesting a drug is necessary for an illness not to appear,
or as embarking on a boat to Athens [ila Athlniya = eis Aiginan]
is necessary for someone to obtain some money which is owed to
him. [3] Next, what is necessary for something is that which
compels and compulsion [al-qahiru wa-l-qahru =to biaion kai he
bia], this being what stands in the way of an act of will or impulse.
For something which compels is said to be necessary, and for this
reason it is also [said to be something] distressing [mutzzinun =
luperon]; as Adonis [Adlnls = Euenos] says, "Everything which is
necessary is painful and distressing". Compulsion is a kind of
necessity; as Qurqalis [Qurqalfs = Sophoktes] also says,
"Compulsion necessitated that I do this". Necessity is also
thought to obtain when someone cannot be dissuaded from
holding a firm belief in the correctness of his opinion, because it
is the opposite of motion which comes about through an act of
will of deliberation. [4] The necessary is also said of something
which cannot exist in any other way [ash-shay'u lladhl laysa
=
yumkinu an yakiina bi-nawcin akhara to me endekhomenon allos
ekhein]. It is by analogy to this "necessary" [bi-mithli hiidhii l-
mu¢farri = kata touto to anankaion] that all other things which are
necessary in some way [are held to be necessary]. For necessary
action or passion is said to obtain when an impulse cannot be
fulfilled because of something which compels it [not to be
fulfilled], since necessity is that on account of which a thing
cannot exist in any other way. And similarly with regard to the
202 Part II: The A vicennian Synthesis
causes of life and the good, for when the good on the one hand, or
life and coming to be on the other, cannot exist without certain
things, those things are necessary, and the cause is a kind of
necessity. [5] There is [also] the logical demonstration of
necessary things [fa-hiidhii i{ili/:iu l-mu{itarrliti = eti he apodeixis
ton anankaion], in the sense that they cannot exist in any way
other than what has been demonstrated in a simple [i.e., absolute]
fashion, the cause [of this] being the first [premises], if the things
which are combined [to form the syllogism, i.e., the major and
minor premises] cannot exist in any other way. [Corollary]
With some necessary things there is another cause [of their being
necess!J.ry]; with others there is no [other cause], but instead it is
on account of them that other things are of necessity. [Summary]
It is clear, then, that what is necessary in a primary and real sense
is the simple [al-mabsu{ = to haplounJ; for that [i.e., what is
simple] cannot exist in many ways, nor even in one other way (for
otherwise it will exist in many ways). So if there are certain things
which are simple, eternal and not subject to change [in klinat
ashyli'u mabsu!atan wa-azaliyyatan wa-ghayra mutaJ:iarrikatin = ei
aro estin atta aiaia kai akineta], none of them will be subject to
compulsion nor [will they be] outside nature [khlirija <ani !-!iblici =
para phusin].

Let me sum up this complex passage. Aristotle distinguishes five senses in

which something may be spoken of as being necessary. 8 The first two can be
combined to form one category, which I shall call the indispensable. The
example given in [ 1] is breathing and eating, which are indispensable if an
animal is to exist, or, to be more precise, to continue to exist. The example
given in [2] is ingesting a drug, which is indispensable if an ill person is to get
better. Together [l] and [2] form a single category, which Aristotle refers to
elsewhere - most prominently in Physics 2.9 and Parts of Animals 1.1 - as
hypothetical or conditional necessity.
Hypothetical necessity in Aristotle is distinguished from absolute or material
necessity. An explanation involving hypothetical necessity begins with the end-
state - continued existence, in the case of [l], and health or well-being in the
case of (2] - and then determines what causal conditions are required in order for
that end-state to be realized. An explanation of this form will generally run along
the following lines: "In order for X to exist, it is necessary that A, B and C
exist". In other words A, Band C - be they materials, processes or events - are
indispensable if X is to exist. In this light the main difference between [ 1] and
[2] is that category [l] is the kind of hypothetical necessity that takes into
account the end-directedness found in the organs and processes of a living thing,
while [2] is the kind of hypothetical necessity that allows for the end-directedness
found in the intentional actions of a human being. In another sense categories [l]

8 Here I follow Alexander (in Metaph. 5.5 [ad 1015a20], 360,19) rather than Ross 1924, 299,

who not only collapses [l] and [2] but also [4] and [5].
 11. Necessity and Possibility (A) 203

and [2] appear to prefigure Alexander's and the Neoplatonists' distinction

between existence (to einai) and well-being (to eu einai): [1] is what is
necessary for existence, and [2] is what is necessary for well-being.
Category [3] is the compulsory, and seems to need little explanation. I was
inclined to watch television at home all day, but something compulsory (in this
case, not losing my job) stood in the way of that impulse, and necessitated that I
come to the classroom to teach my students. Like category [2], category [3]
appears to be restricted to the intentional actions of human beings; but whereas
category [2] refers to the necessity possessed by something which ensures that an
end be realized, category [3] refers to the necessity possessed by something
which frustrates an end from being realized. It is possible, however, to imagine
cases where category [3] might apply to natural phenomena. A clump of lava
spewed from a volcano into the air is compelled to move in the direction - up -
which is contrary to the direction in which the lava's nature would naturally
impel it. No intentionality obtains in this case, yet the conditions required to
satisfy category [3] appear to be met.
Category [4] covers things which cannot be other than they are, which I shall
call the invariable. Necessary things which fall into this category will be simple
rather than complex; eternal rather than mortal; and immutable rather than
changeable. Nor can a thing which is necessary according to category [4] be
subject to the compulsion which is necessary according to category [3].
Category [5] is somewhat problematic. Attentive readers may have noticed that
the extremely compressed Arabic of L59 slightly distorts the Greek of categories
[3] and [4). However, the only instance where the Arabic translator has made a
decision which actually limits the philosophical options that had been open to a
Greek reader of L59 occurs in category [5]. This is where the ton anankaion in
the first phrase (eti he apodeixis ton anankaion) can be understood as a partitive
genitive, meaning "And demonstration is among the things which are
necessary", a sense which the Arabic translator could have rendered only by a
preposition such as min (e.g., fa-hadhii l-1\iiilJ,u [huwa] mina l-mu¢farriiti). As
the Arabic translation stands, however, the phrase fa-hadhii f\falJ,u l-mu¢tarriiti
can only be construed with the genitive of belonging, to mean 'There is [also]
the logical demonstration of necessary things".
The upshot is that the Arabic of L59 may have led its Greekless readers to
think that Aristotle meant for category [5] to refer to the necessity that is
possessed by the conclusion of a valid syllogism (what the Latins would name
necessitas consequentis), a necessity which follows from the necessity that is
possessed by that valid syllogism's premises; rather than to the necessity that
obtains in the act of inferring a conclusion from premises (what the Latins
would name necessitas consequentiae). What I mean is that in the following
valid syllogism:
204 Part II: The Avicennian Synthesis
"animal" holds of every "dog" [major premise]
"dog" holds of every "dachshund" [minor premise]

"animal" holds of every "dachshund" [conclusion]

the Arabic version of L59 probably pushed a close reader of the Metaphysics
such as Avicenna in the direction of thinking that the "necessary things" (al-
mu¢tarriit) of category [5] referred to a valid syllogism's conclusion and
premises, all of which possess necessity, with the conclusion's necessity now
seen as the product of the premises' necessity.
As I mentioned in the footnote, Ross collapses [4] and [5], while Alexander
keeps them separate. I think Alexander is correct in distinguishing them because
category [4] refers to what appear to be the essential characteristics of being
necessary, namely simplicity, etemality and immutability. Category [5], by
contrast, refers - at least in the Arabic version - to a derivative characteristic of
being necessary, namely, being the necessary effect of a cause, with the
conclusion's necessity now seen as the effect and the premises' necessity now
seen as the cause. In other words, category [4] refers to being necessary, whereas
category [5] refers to being necessitated, which I shall also refer to at times as
the inevitable.
To sum up, then, Aristotle speaks of five types of necessary things, the first
two of which may be reduced to the indispensable, the third of which is the
compulsory, the fourth of which is the invariable, and the fifth of which is the
inevitable. Aristotle then makes two very important general claims about the
necessary. First, he asserts that category [4], the invariable, is the most basic
type of necessary thing, that is, the category to which the other categories of the
necessary ultimately refer. As I shall show, this prompted later commentators
such as Alexander and Asclepius to make explicit Aristotle's implication that
category [4] refers to the kind of necessity possessed by divine things. Aristotle' s
claim that category [4] is the necessary in a primary sense, and the
commentators' interpretation of category [4) as referring to divine things, will
play some role in helping shape Avicenna's claim that God is necessary of
existence (wiijib al-wujud) in some primary or absolute way.
The second general claim Aristotle makes is that the necessary may refer either
to something which is necessary because a cause other than itself has made it
necessary, or to something which is itself necessary and which makes other
things necessary. This claim, I believe, is one of the main sources of Avicenna's
distinction between that which is necessary of existence in itself (wiijib al-wujud
bi-dhiitihi) and that which is necessary of existence through another (wiijib al-
wujud bi-ghayrihi).9
But how exactly do we get from Metaphysics 5.5 to Avicenna's idea that God
is the only being that is necessary of existence in itself (wiijib al-wujud bi-

9 Davidson 1987, 291, mentions this in passing.

 11. Necessity and Possibility (A) 205
dhiitihi)? Answering that question requires that I first tum to other passages in
Aristotle where the types of necessity distinguished in Metaphysics 5.5 (apart
from the compulsory, which is not relevant) are decompressed, as well as to
some of the commentaries on those passages. I shall then need to review how
the language of the Arabic translations of these other Aristotelian works is or is
not reflected in Avicenna's discussions of necessity.

Physics 2.9 is where Anstotle discusses most explicitly his distinction between
necessity in a hypothetical sense and necessity in an absolute sense. Here is the
passage in question:

L60
Aris!ii!iilis, ar-Tabica, 158,4-160,7 and 163,9-164,7
Aristotle, Phys. 2.9, 199b33-15 and 200a31-b8
We must investigate whether necessity [Q{i-~rora = to d'ex
anankes] is hypothetical [<an wa(,J<in = ex hupotheseos] or
absolute [<ala 1-i!liiq = haplos]. For they [viz., natural
philosophers such as Empedocles and Anaxagoras] believe that
necessity is found in coming-to-be in such a way that one might
think that a wall is a necessity because the heavy naturally
descends downwards and the light naturally rises upwards; and for
this reason the stones and foundation are at the bottom of it [i.e.,
the wall], while the mortar and cement are above [that] on account
of their lightness, and the timber is at the top because it is lightest
of all. Needless to say the wall will not exist without these
[materials]; yet its existence is not on account of these [materials]
(except of course in the sense that by "cause" you might be
referring to the matter), but rather on account of, and for the sake
of [min ajli = heneka], protection and safety. And it is in this
manner that all other things which are for the sake of something
proceed. It [the end] will not [come to] exist without things which
have the necessary and required nature, although its [the end's]
[coming into] existence is not for the sake of these [materials]
(except by way of [their being] the matter), but [the coming-to-be]
is for the sake of something else. An example of this [is the
question] "Why does the saw come to have this characteristic?" So
that it performs this [function], or for the sake of this [function].
To perform this [function, i.e., of cutting] - which is what it is for
the sake of - it can be fashioned only out of iron. Therefore it will
follow of necessity that it be made of iron ffa-yajibu idhan
=
<Jarilratan an yakilna min l}adldin ananke ara sideroun einai] if it
is saw and if it is to perform its function. Thus the necessity is
hypothetical [i.e., it derives from having first postulated a
condition], it [the necessity] not being like an intended end; this
is because the necessity is found in the matter, while the that for
the sake of which is found in the definition ....
It is clear, therefore, that the necessity found in natural things is
what is so called in the sense of the matter and its motions. The
natural philosopher must concern himself with both of these two
206 Part II: The Avicennian Synthesis
causes [i.e., the matter and the end], although he must be more
concerned with the cause that is "that for the sake of which",
because this is the cause of the matter, while the matter is not the
cause of the end, by which I mean the "that for the sake of which".
[the Arabic omits 200a35] For example, because a house has such-
and-such a characteristic, it is necessary that X and Y [first] come
to be, or that X and Y [first] exist; because health is such-and-such
[a thing], it is necessary that X and Y must [first] come to be, or
that X and Y must [first] exist [wa-li-anna ~-~iflflata kadha, fa-
wajibun fiariiratan an yakuna kildha wa-kildhil wa-an yujada kildha
wa-kadhil = kai epei he hugieia todi tade dei gignesthai ex anankes
kai huparkhein]. Similarly, if this is what "human" is, then it is
necessary that X must [first come to be or exist], and [in general
terms] if A is B, then it is necessary that X must [first come to be
or exist]. It is likely, in fact, that the necessary enters into the
definition as well. This is because if we define the activity of
sawing as dividing in such-and-such a manner, this [activity] can
come about only through the saw's having teeth with such-and-
such a characteristic, and these [teeth] can only have such-and-
such a characteristic by being made out of iron. For a definition
also contains parts which are analogous to matter. 10

Hypothetical and absolute necessity are distinguished along similar lines in

Parts of Animals 1.1} 1 What is made more explicit in the Parts of Animals
passages, however, is the linkage between absolute necessity and divine, eternal
things. Hypothetical necessity, by contrast, applies to natural things down here
in the world of generation and destruction. The distinction between absolute and
hypothetical necessity is also alluded to in Posterior Analytics 2.11 .12
In nature it is the postulating of an end which makes the materials necessary
since it is possible (at least according to Aristotle's reasoning) to imagine a
situation where all the materials that make up a human body are present - flesh,
bone, nerves, blood, etc. - but for the end - the concrete existence of the form
"human" - not to exist. This is because the materials must first be set in motion
towards attaining that end; and given the fact that it is in view of the end that the
materials have been made to come together in the first place, the end is what
makes the materials (hypothetically) necessary.
That is Aristotle's theory, at least. The problem is that in almost all cases
where Aristotle actually applies this theory to examples in the natural world, it
is possible to appeal only to absolute, material necessitation, and dispense with
the end. One well-known example is the shedding of deer antlers: Aristotle
claims that the shedding takes place both because of absolute, material necessity

10 cf. Metaph. 7.12, !038a6-8, where he says that in a definition, the genus is analogous to the

matter and the differentia is analogous 10 the form.

11 PA 1.1, 639b21-640a9 (=Fl a'(la' al-!zayawan, 6,24-7,16); cf. also PA I.I, 642a2-13 (=Fl

a'da', 12,8-19).
12 An. Post. 2.1 J, 94b27-95a9 (= Ta!z/ilat thaniyya, 433,8-434,11).
 11. Necessity and Possibility (A) 207
(the earthy matter that constitutes the antlers having grown too heavy) and for an
end (in order that the deer's head be lighter). 13 As I mentioned in Chapter 10,
scholars of ancient philosophy nowadays disagree over whether or not - and if
so, to what degree - the latter type of explanation (i.e., "In order that the deer's
head be lighter") is an essential component of Aristotle's theory of how we
know or understand something scientifically.
In any case, one final question arises from the Arabic of L60: should we view
certain phrases in Isl_taq b. I:Iunayn's translation, such as wiijibun ¢aruratan an
yakuna, as somehow prefiguring Avicenna's formula of wiijib al-wujud? This
might be tempting given the fact that the Metaphysics translation, in which to
anankaion is rendered as al-mu¢tarr, appears to be an unfruitful source, at least
lexically speaking. But because Aristotle is speaking in Physics 2.9 of
hypothetical and not absolute necessity, L60 will be valuable as a source only in
the limited sense that Avicenna's concept of wiijib al-wujUd bi-ghayrihi, the
necessary of existence through another, could be conceived as being necessary in
the way that organic matter in nature is necessary: as hypothetically necessary,
that is. A natural thing in the world of coming-to-be and passing-away is wiijib
al-wujud bi-ghayrihi because the materials that constitute it are caused to be
necessary by the end they serve. As I shall show in Chapter 13, however, the
main benefit Avicenna saw in wiijib al-wujud bi-ghayrihi was its use in solving
a kaliim problem: distinguishing between two or more eternal things.

I mentioned earlier in this chapter that Aristotle asserts in Metaphysics 5.5 that
the necessity possessed by simple, eternal and immutable things is the most
basic type of necessity, to which all other types may be referred. Aristotle's
implication that this most basic type of necessity is one of the essential
characteristics, if not the essential characteristic, of divine things, is made
explicit by later commentators on Metaphysics 5.5, including Alexander and
Asclepius. They were justified in making this connection because Aristotle
himself alludes to the Unmoved Mover's necessity in Metaphysics 12.7. Much
has been made of Aristotle's assertion at 1072b2 that God ex anankes ara estin
on, particularly by those who see in it the most obvious antecedent of
Avicenna's wiijib al-wujud.14 But it seems unlikely to me that Aristotle's little
assertion had anything more than a coincidental relation to Avicenna's formula.
The first reason why this is so is because what Aristotle means by divine
necessity is so different from what Avicenna means by divine necessity. This
will be brought out better by quoting the passage in full.

13 PA 3.2, 663b!0-14 (=Fi a'rj.ii' al-?iayawiin, 64,15-18).

14 Davidson 1987, 282, citing Ibn Sina, Shar~ ?tori al-Lam, 23,21-22.
208 Part II: The Avicennian Synthesis
L61
Aris!ii.~ii.lis,
Ma bacda !-fabf<a Ill, 1607,11-1608,5
Aristotle, Metaph. 12.7, 1072b4-l l
If something is moved it can be in various states. Therefore if
locomotion is the primary activity by which it [i.e., heaven] is
moved, it is possible that with respect to place [at least] it will be
in various states, even if with respect to substance there is no
variability. Because that which moves it is [itself] unmoved (since
it exists actually), it cannot possibly be in any state other [than
the one it is in]. The primary type of change is locomotion, and
[more specifically] circular motion; and this [Unmoved Mover] is
the mover of that [circular motion]. It is therefore of necessity that
it [the Unmoved Mover] exist as a good, and in this way [i.e., as a
good] it is of necessity that it be a principle [fa-mina l-ir;Jtirari
idhan yUjadu jayyidan <replacing Bouyges' mabda'an> <wa->
ayr;Jan huwa mina l-ir;Jtirari hakadha mabda'an =ex anankes am
estin on kai hei anankei kalos kai houtos arkhe].

Once again, the Arabic of the Metaphysics translation reflects an interpretation

on the part of the translator, an interpretation that may not accurately express
what Aristotle himself meant in this rather compressed passage. Attentive readers
will have noted that in L61 I emended Bouyges' text to readfa-mina l-i<jfirari
idhan yujadu JAYYIDAN instead of Bouyges' MAEDA' AN. This is an
emendation which the apparatus suggests could be justified on manuscript
grounds; it is certainly justified on the basis of content, since the Greek kalos
must have been rendered by a term like jayyidan, the skeleton of which is
almost identical to that of mabda'an. In any case what Aristotle probably meant
by ex anankes ara estin on kai hei anankei kalos is simply that the Unmoved
Mover must be necessary in the sense of simple and invariable: necessary in the
sense of L59's category [4), that is. Aristotle adds the epexegetical kai hei
anankei kalos to ensure that his reader understands that he means "necessary"
here in the nice sense of category [4), and not in the nasty sense of category [3),
which would entail that God be constrained or compelled in some way.
However, the Arabic of L61 - assuming that I am correct in emending
mabda'an to readjayyidan - lends itself to a rather more specific interpretation
of those crucial lines. According to this interpretation, Aristotle begins by
describing the necessity of the Unmoved Mover in line with the conception of
necessity articulated in L59 as category [4). The Unmoved Mover is invariable,
since it is only mover and not moved. Given that the Unmoved Mover is not
subject to change of any sort - not even the highest form of change, circular
motion - it is necessary in the sense that it is invariable.
Then Aristotle appeals to the conception of necessity articulated in L59 as
category [5). Given the fact that the Unmoved Mover is necessary in the sense
that it is invariable, the Unmoved Mover will, by necessity, act as a principle -
that is, it will cause its effects - as a good. In other words, it is only as a good,
as an object of desire or thought, that a thing can move another without itself
 11. Necessity and Possibility (A) 209

moving, and hence without being moved by another. According to this

interpretation, the sense of necessity appealed to at the end of the passage falls
under category [5] because it is a logically necessary consequence of the premises
(1) the Unmoved Mover is invariable; (2) the Unmoved Mover is a cause; and (3)
the only cause that can move things without itself moving is the final cause. Let
me recapitulate: given that the Unmoved Mover is not subject to change, and
given the fact that the Unmoved Mover causes motion, it follows of necessity
that the Unmoved Mover causes motion as a final cause, since only final causes
can move other things without themselves moving. To be more precise, it is
only a final cause in the sense of that in view of which (i.e., qua to hou), and
not in the sense of that for the benefit of which (i.e., qua to hOi), that can
cause other things without itself being caused. Aristotle makes this explicit at
12.7, 1072b2-3 (=Ma ba<da !-{abica Ill, 1599,3).
In other words, the Arabic Aristotle's aim in this passage does not come
across as a general one. It is not to introduce the notion that the Unmoved
Mover is a necessary existent, that it must be in some bald sense. Rather, it is
to draw, from what he has just discussed, the necessary conclusion that the
Unmoved Mover must be a final cause qua to hou.13
How then to reconcile the Unmoved Mover's being a final cause, and its
necessitation of other things? After all, there seems to be a contradiction. On the
one hand, we are given a description in Physics 2.9 of how final causes
necessitate their effects hypothetically, not absolutely. Hypothetical necessity is
the kind in which materials are made necessary only by postulating an end
towards which they are set in motion. In Metaphysics 5.5 hypothetical
necessity is covered by categories [l] and [2]. Aristotle also makes clear in
Physics 2.9 that hypothetical necessitation is the kind that is in operation
among natural things like animals and plants, as well as among artifacts. By
contrast, the way simple, eternal and immutable things necessitate their effects
seems much more likely to be the kind of absolute necessitation covered by
category [5] in Metaphysics 5.5. That is, simple, eternal and immutable things
should necessitate their effects the way that premises necessitate their
conclusions. The problem, then, is that if the Unmoved Mover is related to its
effects as a final cause, and if final causes necessitate their effects hypothetically,
the Unmoved Mover will necessitate its effects hypothetically and not
absolutely; not, that is, in the way premises necessitate conclusions.

15 The sense in which the necessity which Aristotle appeals to in the second part of the passage

is that of category [5] - that of a logically necessary inference - is brought out in Thiimas~iyiis,
Sharl;i J;iarf al-Liim min Kitiib Ma ba cda {-!abica li-Arisfu, 14,18-19. However Themistius
understands Aristotle to be making the general claim that "it follows necessarily that there exists an
unmoved mover" (fa-in kana yujadu shay'un yatal;iarraku wa-yul;iarriku ma<an wa-yujadu shay'un
yatal;iarraku faqar min ghayri an yul;iarrika, fa-yajibu ijaruratan an yujada mul;iarrikun ghayru
mutal;iarrikin), and not the specific claim that "it follows necessarily that the unmoved mover is a
final cause qua to hou".
210 Part II: The Avicennian Synthesis

One could argue, of course, that unlike natural things subject to coming-to-be
and passing-away, the Unmoved Mover is eternal. Because it already exists, its
future coming-to-be need not be postulated for its effect to exist, and therefore its
necessity is not hypothetical. But however convincing that might be, it is still
difficult to conceive of how any final cause, even an eternal one such as the
Unmoved Mover, could necessitate its effect in an absolute way.
It is in trying to cope with this problem that Alexander takes one step further
than Aristotle when he links together the ideas of necessity and necessitation by
arguing that the other categories of necessity derive their necessity from - that
is, they are necessitated by - the necessity that category [4) possesses. In other
words, the other categories of necessity are not only referable in a logical sense
to category [4) - that is, not only is category [4] logically prior to the other
categories of necessity - their necessity is caused by category [4]'s necessity.
This is an important move. Alexander can now apply Aristotle's distinction
between things whose necessity is the cause of other things' necessity and things
whose necessity derives from something outside themselves, to the distinction
between category [4] and the other categories. Aristotle - or at least the Arabic
Aristotle, as I showed earlier - seems to have meant for the distinction between
necessitating and necessitated to be applied respectively to premises, whose
necessity is the cause of the conclusion's necessity; and to conclusions, whose
necessity is derived from the premises' necessity. What I am getting at is that
Alexander is the first to draw an explicit connection between divine necessity a00
divine necessitation. Because of their significance for Avicenna, Alexander's
comments are worth quoting:

L62
Alexander, in Metaph. 5.5 (ad 1015b9-l l), 361,24-36
But eternal things are also necessary in this way [viz., in the way
necessary premises are necessary]. for they have no cause
extrinsic to them and they are of necessity causes of being for
other things too; for all things that are naturally constituted owe
to the gods the fact that they are immediately necessary because
they have a cause of their being that is eternal. Moreover, that
which either cannot be or be in a good state without something
else has from that other of necessity the cause both of its being
and of its well-being [ex anankes par' ekeinou ten aitian to men
tou einai to de tou eu einai ekhei] .... Only the eternal beings,
those that are divine, do not have another thing as cause of their
eternal being, and it is because of them that the things naturally
constituted are and come to be of necessity. 16

In other words, Alexander has appealed to his explicitly articulated distinction

between being (to einai) and well-being (to eu einai) in order to incorporate the

16 The translation is Dooley 1993, 32-33.

 11. Necessity and Possibility (A) 211
conception of necessity articulated in L59 as category [1] - the causation or
necessitation of being (to einai) - and the conception of necessity articulated in
L59 as category [2] - the causation or necessitation of well-being (to eu einai) -
into an overarching scheme whereby the necessity of [ 1] and [2] is caused by the
category-[4] necessity possessed by divine things. Alexander then adds to the mix
the intrinsic necessity/extrinsic necessity distinction of the corollary that appears
to modify category [5): the necessity of the premises is intrinsic to them, while
the necessity of the conclusion comes from outside, viz., from the premises'
necessity.
Let me recapitulate Alexander's understanding of this passage. The divine
things' necessity (i.e., category [4]) is the cause of all other things' necessity -
both the necessity of things which are necessary for being (i.e., category [1]) as
well as the necessity of things which are necessary for well-being (i.e., category
[2]) - just as the necessity of premises is the cause of the necessity of the
conclusions that follow from them (i.e., category [5]). Since the necessity that
premises possess is intrinsic to them, while the necessity that conclusions
possess comes from outside, the necessity possessed by divine things will be
intrinsic to them, while the necessity possessed by all other necessary things
will come from outside of them. Alexander's move to theologize category [4]
gains momentum from Asclepius' comments on L59, where Asclepius explains
that the necessity of simple things does not come from outside, but is in their
very being; thus we say that God is good of necessity, since God's substance is
goodness. 17 All of these interpretive trends seem to me to be reflected in the
Arabic translations of L59 and L6 l.

When I labeled category [5) in Aristotle's list of the different ways we speak of a
thing's being necessary as the "inevitable", I meant to distinguish between the
absolute necessity possessed by something simple, eternal and immutable, arxl
the necessity possessed by a conclusion which cannot help but follow from its
premises. By applying the necessitating/necessitated distinction, we can infer
that category [5) can apply to necessary things (such as premises) which
necessitate as well as to necessary things (such as conclusions) which are
necessitated. After all, the conclusion of a syllogism fulfills category [4]'s
criteria of invariability just as much as its premises do. Therefore it would be
correct to say that the necessity articulated in category [4] does not belong
exclusively to causes. Taken in itself, category [4) refers only to simplicity,
eternality and immutability. When those criteria are satisfied by a cause, it
signifies that they will be productive of the same criteria in an effect; when those
criteria are satisfied by an effect, it signifies that they will have been produced by
a cause which also satisfies those criteria.

17 Asclepius, in Metaph. 5.5 (ad 1015a33-34), 309,10- 15 (theoria) and 313,2-3 (lexis).
212 Part II: The Avicennian Synthesis

In itself, therefore, category [4] seems closer to the idea of self-sufficiency, of

immunity from causation, than it does to the idea of causal productivity, because
neither being a cause nor being an effect appears to be essential to category [4].
We are still faced, then, with the same disjunction between self-sufficiency and
productivity which forced Avicenna to appeal to the fence-straddling idea that
somehow God both possesses perfection yet is also above perfection. Where else
in Aristotle's writings could Avicenna find a way to fuse the notion of
necessitating more tightly into the notion of being necessary?
In Posterior Analytics l.30, Aristotle says:

L63
Aristiitalis, Ta/:lliliit thiiniyya, 397,4-9
Aristotle, An Post. 1.30, 87b 19-25
There is no demonstrative knowledge of something which arises
by chance, since something which is by chance is neither
necessary or for the most part [laysa huwa {iarilriyyan wa-lii cala
akthari l-amri = oute gar h6s anankaion outh' h6s epi to polu], but
is instead what falls outside of these two [categories].
Demonstration, however, applies to either of these two
[categories], because every [demonstrative] syllogism has either
necessary premises or premises which are for the most part. If the
premises are necessary, the conclusion will be necessary as well
[fa-in kanati l-muqaddimiitu {iarilriyyatan fa-n-natijatu hiya ay{ian
{iarilriyyatun = kai ei men hai protaseis anankaiai, kai to
sumperasma anankaion]; and if they are for the most part, then the
syllogism will be this way as well.

It seems true that the necessity which the premises possess can be reduced to
Aristotle's category [4]. At the beginning of An. Post. l.4, and again in l.6,
Aristotle argues that since demonstration treats only simple or absolute things, -
things which cannot be otherwise, that is - what comes to be known through
demonstration will be necessary knowledge. But what is being discussed in An.
Post. l .30 and in category [5] is necessitation, not necessity: a pair of necessary
premises necessitate the conclusion that follows from them, that is to say, their
necessity produces the necessity of the conclusion. Yet even though necessary
premises satisfy the main criterion of category [4], in that they cannot be
otherwise, what this means in a premise is different from the criteria of
simplicity, eternality and immutability which Aristotle listed in Metaphysics
5.5. For a premise to be necessary its predicate must hold of its subject
necessarily. To find out precisely what that means we must tum to De
Interpretatione 12-13.
For readers unfamiliar with the treatise, the De Interpretatione is meant to
determine what makes up a proposition. That is not an aim in itself, however:
Aristotle is interested ultimately in determining what kind of propositions can
serve as premises in a syllogism. A large part of the De Interpretatione is
 11. Necessiry and Possibiliry (A) 213
devoted to analyzing the elements of a proposition, such as verbs and nouns; to
distinguishing between different types of propositions, such as those that
exclaim ("Stop biting your fingernails!") and those that assert ("Biting one's
fingernails is bad"); and to distinguishing between assertoric propositions that
are affirmative ("Biting one's fingernails is bad") and those that are negative
("Biting one's fingernails is not good").
Finally, Aristotle hopes to distinguish between propositions which contradict
each other and those which are contrary to each other. Contradictory propositions
are so related that if one of the two is true the other will be false, and if one is
false the other will be true, as for example the propositions "Rob is white" and
"Rob is non-white". Contrary propositions, on the other hand, are so related that
though both may be false they cannot both be true, as for example the
propositions "Rob is white" and "Rob is black"; after all, Rob could be green, or
slightly pink.
This is complicated stuff, and problems arise. First, Aristotle has to decide in
fact whether the contradictory of "Rob is white" is "Rob is non-white" or "Rob
is not white". Then Aristotle has to decide what happens if he tosses in adverbs
or adjectives that in some way qualify the proposition, such as the temporal
operators "always", "sometimes" or "never"; and their parallel modal operators
"necessary'', "possible", and "impossible". How do you go about making the
contradictories of propositions that have these qualifiers in them?
The reason that the modal operators "necessary", "possible" and "impossible"
are particularly important to Aristotle is that he wants to winnow down the vast
number of propositions available to him to a smaller number which he can use
as premises in demonstrative syllogisms, the only type of syllogism which he
reckons is guaranteed to produce a scientific understanding of a thing. As I
mentioned above in discussing An. Post. 1.6, Aristotle feels that the
propositions that can be used as premises in a demonstrative syllogism will have
to be necessary in the sense of being primary and true in some unshakeable way.
After all, we learned already in Metaphysics 5.5 that it is only when the
premises of a syllogism are necessary that the conclusion produced from them
will be necessary. In De lnterpretatione 12-13 Aristotle therefore needs to come
to grips with what it means when we assert that X is P necessarily, that is,
when we assert that the predicate "P" holds of the subject "X" in a necessary
way. In short, Aristotle wants to determine as precisely as he can the properties
of necessary propositions.
Part of his effort in this direction, as I said above, is focused on determining
what the contrary and contradictory of "It is necessary that X is P" are. (I shall
leave aside for a moment the question of whether the contradictory of "X is P" is
"Xis not P" or "X is non-P".) Two options immediately present themselves as
possibly serving in the role of contrary and contradictory of "It is necessary that
X is P": "It is not necessary that X is P'', "It is necessary that X is not P".
Aristotle eventually determines that the contrary of "It is necessary that X is P"
214 Part II: The Avicennian Synthesis

is "It is necessary that X is not P"; and that the contradictory of "It is necessary
that X is P" is "It is not necessary that X is P". This is because, as I mentioned
above, a contrary pair of propositions is so related that though both may be false
they cannot both be true; while a contradictory pair of propositions is so related
that if either is true the other will be false.
In the case of the proposition "It is necessary that X is not P", the idea of
impossibility is being expressed. In the case of the proposition "It is not
necessary that Xis P", however, the idea of possibility is being expressed. With
that in mind, it is easier to see why the idea of possibility expressed in the
proposition "It is not necessary that X is P" is the contradictory of the necessity
expressed in the original proposition, "It is necessary that X is P". There is no
middle ground between the two propositions - either something is necessary or
it is possible - and therefore they are contradictories, since contradictory
propositions are so related that if either is true the other will be false.
By contrast, there is a middle ground between the necessity expressed in the
original proposition, "It is necessary that X is P", and the impossibility
expressed in the proposition "It is necessary that X is not P". This is because it
is possible to imagine X's being Q or R. Therefore, since it is possible to
imagine a situation where the propositions "It is necessary that X is P" and "It is
necessary that Xis not P" are both false (even though they cannot both be true
at the same time), the two propositions are contraries. This is because, once
again, a pair of contrary propositions are so related that though both may be
false they cannot both be true.
In short, the contradiction derives from a negation of the mode, whereas the
contrariety derives from a negation of the copula. One crucial ambiguity arises,
however. We have just decided that the possibility expressed in the proposition
"It is not necessary that X is P" is the contradictory of the necessity expressed in
the proposition "It is necessary that X is P". But the term Aristotle uses for
impossible, adunaton, seems to be a straightforward negation of the term he
uses for possible, dunaton. If a thing is either dunaton or adunaton, then the
contradictory of the possibility expressed in the proposition "It is not necessary
that X is P" will be the impossibility expressed in the proposition "It is
necessary that X is not P", instead of the necessity expressed in the proposition
"It is necessary that Xis P". The upshot is that "the possible" (to dunaton) and
"the impossible" (to adunaton) seem to be just as completely opposed to each
other as "the possible" (to dunaton) and "the necessary" (to anankaion) are.
Deciding whether or not possibility is "one-sided" (that is, opposed only to
impossibility) or two-sided (that is, opposed to both impossibility and necessity)
is a problem that Aristotle takes a while to work out in De lnterpretatione 12-
 11. Necessity and Possibility (A) 215

13. 18 As Ackrill points out, the first and third tables of propositions which
Aristotle sets out in Int. 13, 22a24-3 l will be truly compatible only if
possibility is understood as being two-sided, while the second and fourth tables
will be truly compatible only if possibility is understood as being one-sided.
Eventually Aristotle reconfigures the tables so that they are truly compatible,
but in so doing he is forced to chuck out two-sided possibility in favor of one-
sided possibility. 19
Now none of what Aristotle actually thought about these complex issues
concerns me in and of itself. What I am interested in uncovering instead is how
Aristotle's presentation of these issues, and his ambivalence over whether to
think of possibility as one-sided or two-sided, may have influenced the way Isl)aq
b. l:funayn, the Arabic translator of the De Interpretatione, chose to render the
relevant terms. First of all, I maintain that Aristotle's ambivalence regarding
whether or not to see possibility as one-sided or two-sided is reflected in a
gradual evolution in the Arabic translations of De Interpretatione 12-13, from a
rendering of einai as the more copulative (or at least amphibolous) k-w-n, to a
rendering of einai as the more existential w-j-d. This is because, as I shall show,
an existential rendering of einai makes it easier to see possibility as opposed to
necessity, in that it flushes out the sense in which dunaton means "contingent"
(and hence as being naturally opposed to "necessary") as well as "possible" (and
hence as being naturally opposed to "impossible").
The other factor pushing ls~aq b. I:Junayn to use w-j-d instead of k-w-n is that
in De lnterpretatione 13 Aristotle begins to use a kind of shorthand, saying, for
example, anankaion einai instead of the more explicitly copulative ex anankes
X esti Y. This may have led ls~aq to think it better to render the einai with the
more existential mawjud instead of the strictly copulative huwa or the
amphibolous yakunu. Let me show how the terms are transformed in Arabic.
In the first list of modal expressions we come across in De lnterpretatione
12, the one at 21a36-8 (= Fl l-'ibiira, 86,4-6), einai is rendered by the
copulative, or at least neutral, verb k-w-n. Thus:

to dunaton einai = mii yumkinu an yakuna

to me dunaton = mii iii yumkinu
to endekhomenon = mii yul:ztamalu an yakuna
to me endekhomenon = mii la yul:ztamalu
to adunaton = mii kiinafi l-mumtani '
to anankaion = mii klina ft t;l-t;larurl

Only one line later, however, at 2la39 (=Fl l- cibiira, 86,8), einai is rendered
by the more existential verb w-j-d:

18 The distinction between one-sided and two-sided possibility can be found at An. Pr. 1.3,

25a37-40 and 1.13, 32a18-21, though in both those passages Aristotle uses to endekhomenon
instead of to dunaton.
19 Ackrill 1963, 151-153.
216 Part II: The A vicennian Synthesis
to einai = mawjud
to me einai = lii mawjud

At 2lb23-27 (=Fi l-'ibara, 87,15-18), Is~aq begins to apply the new verb w-j-
d to modal propositions (yumkinu an yujada instead of yumkinu an yakuna and
yuf:ztamalu an yiljada instead of yuf:ztamalu an yakuna); and at 2lb27 (= Fil-
'ibiira, 87,18) Is~aq also begins using the term al-wajib to render anankaion -
instead of the previous a<f-<j.arurl - when drawing a contrast with al-mumtani'
(adunaton). And even though at 22a4-7 (= Fi l-'ibiira, 88,14-17), in his
discussion of the contradictories that arise from using the modal operator
"necessary", Is~aq uses both w-j-b and <j.-r-r (wiijibun <j.aruratan an yujada), in
the list of expressions at 22al2-13 (= fl l-'ibiira, 89,7-11), we have the
following:

dunatonlou dunaton = mumkinllii mumkin

endekhomenonlouk endekhomenon = mul;tamal/lii mul;tamal
adunatonlouk adunaton = mumtani'/lii mumtanic
anankaionlouk anankaion = wiijibllii wiijib
alethes/ouk alethes = l;aqqllii l;aqq

with the result that at least in these new passages wiijib has eclipsed <j.arurf.
More evidence of Is~aq's move from k-w-n to w-j-d and from <j.-r-r to w-j-b can
be seen in the fourth table of compatible propositions in De lnterpretatione 13,
at 22a28-31 (=fl l-cibiira, 90,11-15):

(I) ou dunaton me einai = laysa mumkinan allii yujada

(2) ouk endekhomenon me einai = laysa mul;tamalan allii yujada
(3) adunaton me einai = mumtani'un allii yujada
(4) anankaion einai = wlijibun an yujada

Why does using w-j-d to render einai allow Is~aq to interpret Aristotle as
advocating two-sided possibility in these passages more easily than using k-w-n
would? My sense is that when dunaton (= mumkin) is used in opposition to
anankaion (= wiijib), what is being implied is contingency rather than
possibility. That is to say, when dunaton (= mumkin) is used in opposition to
anankaion ( = wiijib), the contrast which is implied is one between necessity in
the sense of uncausedness and contingency in the sense of causedness. It is in
this respect that w-j-d has an advantage over k-w-n, because w-j-d carries greater
existential weight than k-w-n does. In order to acconunodate the idea of
contingency, dunaton (= mumkin) had to be understood with the more
existential w-j-d, because with an existential verb the idea of causedness - that
is, of being a thing whose existence is caused - is more justifiable than it would
have been with the more copulative k-w-n.
What about w-j-b? What advantage did wiijib have over <farurl in rendering
anankaion? I would guess that unlike <farii.rl (or the Arabic Metaphysics'
 11. Necessity and Possibility (A) 217

mu{i{arr), wajib accommodated a broad enough sense of logical necessity that it

could include in its semantic range the necessity which obtained in propositions
such as the ones being discussed in the De Interpretatione . This is because the
standard way to render the idea of syllogistic necessitation was by the use of w-
j-b. The idea that the conclusion follows necessarily from the premises was
generally expressed by the phrase yajibu X min A wa-B, where X is the
conclusion and A and B are the premises. Although the root l-z-m is sometimes
used instead of w-j-b to express logical entailment, I am not aware of any
instances where ¢-r-r is used in this way.
Since what is being discussed in De Interpretatione 12-13 is the necessity
that obtains when, in a proposition, a predicate holds necessarily of a subject;
since Aristotle is interested in determining the nature of that necessity because he
wants to employ necessary propositions as premises in syllogisms; and since
Aristotle is interested in using necessary propositions as premises in syllogisms
because they are the only ones, according to An. Pr. 1.6 and Metaph. 5.5
(remember category [5]), which produce conclusions which are necessary; Isl_iaq
turned to w-j-b rather than ¢-r-r, because w-j-b was the standard verb used to
render that very sense of syllogistic necessitation, that sense of the inferential
necessity (necessitas consequentiae) by which necessary premises cause or
produce the necessity possessed by the conclusion (necessitas consequentis).
What makes the De Interpretatione translator's use of the active participle
wiijib to render anankaion striking is that, as I shall show in Chapter 13, wiijib
had been used by mutakallimun primarily to render the idea of moral obligation
rather than logical necessity. It is true that knowledge CCilm) was sometimes held
by the mutakallimun to be wiijib, but this was generally construed to refer to the
quasi-religious obligation not to be ignorant. Lexically speaking, the Arabic
version of De Interpretatione 12-13 thus seems to have been the most
significant Aristotelian source from which Avicenna drew in his attempt to
construct a matrix of distinctions around the new concept of wiijib al-wujud.
Conceptually speaking, however, Metaphysics 5.5 is at least as significant a
source as the De Interpretatione passages. Chapters 12 and 13 will detennine
how two more immediate sources - al-Farabi and the tenth-century Sunni
mutakallimun - influenced the formation of Avicenna's new concept.
 12. Necessity and Possibility (B)
Materials from al-Farabr

I have just argued that one reason why Ist:iaq b. l:funayn chose the Arabic root w-
j-d to render einai in much of De lnterpretatione 12-13 was because the
existential momentum carried by w-j-d propelled an interpreter towards viewing
possibility as two-sided. This is because mumkin will more plausibly contain
the sense of contingency (that is, the sense of being caused) in its semantic
range when it is used in conjunction with an existential "is" verb, than it would
with a more purely copulative "is" verb. With mumkin understood as contingent,
it becomes easier to view it as the contradictory of wajib, now understood as
necessary (that is, as being uncaused). In other words, I will have an easier time
of arguing that "necessary" and "possible" are contradictory terms if I can reduce
their meanings, respectively, to "being uncaused" and "being caused". When
placed in opposition with mumtanic, on the other hand, mumkin can retain what
seems to have been its previously more basic meaning of "possible", and
therefore serve plausibly as the contradictory of impossible.
In his Commentary on the De lntepretatione, al-Farabi takes a couple of
steps further in this direction than the Arabic translator of the De
lnterpretatione had. The first step which al-Farabi takes - and Avicenna, as I
shall show, follows him in this respect - is to make the opposition between the
possible and the necessary the basic type of contradiction, with possible and
impossible now relegated to second place. In fact, al-Farabi refers to the possible
and the necessary (al-mumkin wa-t;f.-t;f.arurl) as the primary or basic modes (al-
jihat al-uwal), the other modes being reducible to them. 1 Aristotle, by contrast,
sees the possible first and foremost as the contradictory of the impossible in
contexts where possibility is clearly taken to be one-sided. In other words, when
the possible is viewed as one-sided rather than two-sided, Aristotle chooses the
impossible, rather than the necessary, to be the more basic contradictory of the
possible. It is only when Aristotle allows that possibility may be two-sided that
its opposition with necessity is explicitly allowed for. One effect of al-Farabi's
having taken the step to view possibility and necessity as the more basic pair of
contradictories, is that the idea of an opposition between necessary existence and
possible existence becomes more expressly articulated than it had ever been
before.
The second step which al-FarabI takes in his Commentary on the De
lnterpetatione, in addition to the now more expressly articulated opposition
between necessary existence and contingent existence, is to introduce another
distinction which will play a major role in Avicenna's discussions. This time,

1 al-Farabi, Shar~ li-kitiib Aris{fl{iilis ft 1-'ibiira, 163, 16-19 (ad 2 la34-37).

220 Part II: The Avicennian Synthesis

the distinction is between being necessary of existence in itself ({iariirl al-wujiid

ft nafsihi) and being possible of existence in itself (mumkin al-wujiid ft nafsihi).
To be sure, al-Farabi does not articulate the linkage between being possible of
existence in itself and being necessary of existence through another as expressly
as Avicenna later will, but the linkage is implied in al-Farabl's discussions. The
following is an inventory of such passages.
In the first passage, al-Farabi criticizes previous commentators for not
distinguishing clearly enough the subject matter of metaphysics from logic:

L64
al-Fariibi, Shar~ Kitiib Arisfi[!iilis Ji l- %iira (ad l 8a33),
84,12-19
Their approach, apart from the fault already mentioned, entails
their making logic an investigation into the nature and mode of
being of existing things. In this they go wrong. Their own
principles would oblige them to reject this approach. For in the
Categories, they refuse to give an account of entities in terms of
their [modes of] existence. Their refusal to make any part of logic
an investigation into the intrinsic nature of entities is more
appropriate [than the approach adopted by them here]. To examine
something as to whether its existence is a matter of intrinsic
possibility or necessity [fa-inna l-fa~~a cani sh-shay"i hal huwa
mumkinu l-wujudi Ji nafsihi aw ¢aruriyyu l-wujudi Ji nafsihi] is to
examine the mode of existence of the entity in question; which is
outside the proper domain of logic. We must therefore ignore the
commentators' line of interpreting this text and take the approach
outlined above. 2

In the secoHd passage, al-Farabi paraphrases Aristotle's claim at De

lnterpretatione 19a25-26/F7 z-cibara 74,16-75,l that there is a difference
between being necessary in an absolute sense (to haplos einai ex anankes =
anna wujudahu i;iariiratun caza l-i!laq) and being necessary as long as it exists
(ijariiratun idha wujida):

L65
al-Fiiriibi, Shar~ Kitiib Arisfufiills Ji l-%iira (ad l 9a25-26),
95,18-19
Aristotle divides necessary existence into two kinds, of which one
is necessary as long as it lasts, the other necessary without
qualification [qad qassama l-wujUda a<,l-¢aruriyya ilii ¢arbayni
¢arbun huwa tfarnriyyun mii diima mawjudan wa-l-iikharu ¢aruriyyun
<a/ii l-ifliiqi]. 3

2 The translation is Zimmermann 1981, 78. Zimmermann 1981, xxxviii-xli, contextualizes al-

FariibI' s criticism of previous commentators.

3 The translation is Zimmennann 1981, 90.
 12. Necessity and Possibility (B) 221

As I mentioned before, al-Fiirabi only implies the idea of something which is

necessary through another; and it is true that in this context, what is at issue is
not so much the opposition between causedness and uncausedness but the
relation between the temporal qualifiers "always" and "sometimes" on the one
hand, and the modal qualifiers "necessary" and "possible" on the other.
In contrast to Avicenna, al-Fiirabi seems mainly interested in articulating a
distinction between necessity and possibility which will be consistent with the
ideas of necessity and possibility that appear in the "tomorrow's sea-battle"
discussion of future contingent events in De Interpretatione 9. What is at issue
there is the precise manner in which the truth of my knowing or asserting
yesterday that there will be a sea-battle tomorrow makes the fact of its occurring
necessary. For if I do claim that the truth of my saying yesterday that there will
be a sea battle tomorrow makes the fact of its occurring necessary, I shall be
forced to deny that there is any contingency in the world; the basis on which I
claim to act voluntarily will be pulled out from under me, and everything will be
predetermined. As al-Fiirabi says:

L66
al-FlirlibI, Shar/:i Kitiib Arisfii!iilis fi l- 'iblira (ad l 9a32-b4),
98,12-15
Thus the existence of what exists in the future will be necessary
[{iariiriyya l-wujud] if an anterior statement is true. Everything will
be intrinsically necessary [{iaruriyyatan fi anfusihii] again, being
possible only in terms of our knowledge. Free will, deliberation
and everything else that Aristotle has mentioned will again be
eliminated. 4

If, on the other hand, I claim that the truth of my saying yesterday that there will
be a sea battle tomorrow does not make its occurring necessary, I shall be forced
to maintain that there is no necessary relation between a statement's truth and
the reality it truthfully describes; the basis on which I claim to understand a
thing scientifically is pulled out from under me, and I will be left with mere
opinion.
With the stakes so high it is not surprising that commentators energetically
strove to create new distinctions that could resolve the dilemma, or at least blunt
its horns. It is in this context that al-Fiirabi introduces a more explicit
distinction between being necessary in itself and being necessary through
another:

4 The translation is Zirnmennann 1981, 93.

222 Part II: The A vicennian Synthesis
L67
al-Fara bi, Sharft Kitab Arisruralis ft l-%ara (ad l 9a32-b4),
99,1-8 and 99,18
No, the right answer lies in arguing that if something follows of
necessity from something else, this does not mean that it is
necessary in itself [¢aruriyyan ft nafsihi]. For if it is true to affirm
something, it follows of necessity that it is the case. Yet it does
not follow that its being the case is intrinsically necessary [wa-
laysa yalzamu an yakuna l-amru ¢aruriyya l-wujudi ft nafsihi].
What follows is that its following from the truth of the affirmation
is necessary. It is not the case that, if something follows from
something else necessarily, the thing that follows is itself
necessary. Take the case of intrinsically possible conclusions [an-
nata,iju llatf hiya mumkinatun ft dhawatiha]. They follow
necessarily in the syllogisms that lead to them, yet a conclusion
following of necessity is not intrinsically necessary. For its
possibility is not eliminated by the necessity with which it
follows from the premises ....
If something follows from something else of necessity, this
does not mean that it is in itself a necessity [fi dhatihi
i¢firariyyan]. 5

Al-Farabi's point is that the conclusion of a syllogism is not intrinsically

necessary, but is made necessary by the necessity of the premises from which it
follows necessarily. This obviously brings to mind Metaphysics 5.5, and
particularly the concept of necessity articulated as category [5], that is, the
necessity that obtains when a conclusion is necessitated by its premises, or,
more strictly, when the necessity of a conclusion is the product of the necessity
of its premises. In the present context, al-Farabi is trying to find some way to
resolve the sea-battle dilemma by maintaining that future contingent events have
both possibility and necessity. The necessity of future contingent events derives
from their being necessitated by - that is, their following necessarily from - the
truth of our prior knowledge of and assertions about them. But that is not to
claim that in themselves these future events are necessary, for they are made
necessary by something else. The obvious next step for al-Farabi would then be
to claim that future events are in themselves possible. Once he had done that, he
could - like Avicenna - judge that something whose existence is possible in
itself will also be necessary of existence through another.
But al-Farabi resists being forced to articulate a concept of the possible of
existence in itself. One reason for this is that al-Farabi is not in fact talking here
about necessitas consequentis, the necessity possessed by the conclusion of a
valid syllogism and which is the type of necessity suggested by the genitive of
possessing in the Arabic translation of L59. Instead, al-Farabi merely suggests
that these future events can be said to be necessary only in the sense that their

5 The translation is Zimmermann 1981, 94.

 12. Necessity and Possibility (B) 223

following from the truth of our prior knowledge of or assertions about them is
necessary. In other words, al-Farabi is speaking about necessitas
consequentiae, the necessity that obtains in the act of inferring a conclusion
from premises, and not necessitas consequentis, the necessity possessed by the
conclusion itself. It is that latter type of necessity - necessitas consequentis -
which Avicenna has in mind when he equates the possible of existence in itself
with the necessary of existence through another.
Another reason why al-Farabi resists taking the step that Avicenna would later
take, is that al-Fiirabi, like contemporary Mu<tazilite thinkers of the early-tenth
century, was tom between maintaining God's omnipotence and God's justice. If
God is all-powerful, and therefore causes all our actions, including our evil
actions, how then can God be just when He punishes us for those evil actions?
One solution to this old dilemma would be to appeal to the distinction between
God's foreknowledge of future events and His causal determination of future
events. In the context of this discussion, it would seem perfectly natural to make
the claim that God's foreknowledge of future events makes them necessary of
existence, but not in themselves. Rather, they are necessary of existence in the
sense that they follow necessarily from the truth of God's foreknowledge of their
existence; in themselves, they are possible of existence. The question of God's
causation of the events, as opposed to His foreknowledge of them, would be left
open, in order that He not be directly implicated in our evil acts.
Again, al-Fiirabi is reluctant to take that extra step. Partly this is because he
seems to have thought that the concept of being possible of existence in itself
somehow undermined the idea that future events are possible of both existence
and non-existence. 6 Al-Fiirabi says that possibility lies:

L68
al-Flirabi, Sharf:t Kitiib Aris!u!iiils ft 1-<ibiira (ad l 9a32-b4 ),
99,22-25 and 100, 1-4
... not in the event itself, but in the fact that it follows from a true
statement [about it]. If so, a state of affairs can become one of
necessary existence ffa-qad yakunu sh-shay'u {iaruriyya l-wujudi],
having had, up to the time of its existence, the possibility of not
existing and not having existed. There is little objection to this if
you bear in mind that the element of necessity which comes into it
does not affect its possibility .... 7

In this sense the prior knowledge of or assertion about a future event causes that
event to be both necessary of existence and possible of existence, and therefore

6 For a discussion of the history of how the problem of future contingents in Int. 9 was viewed in
the context of discussions of God's foreknowledge of events, see Zimmermann 1981, Ixvii-Ixviii
and lxxxvi, n.l; andcxvi-cxvii.
1 The translation is Zimmermann 1981, 93-94.
224 Part II: The Avicennian Synthesis

the event will be both necessary of existence through another and possible of
existence through another.
Why should al-Farabi so resist the idea of a thing's being possible of existence
in itself? I believe the reason for al-Farabi's reluctance to use the terms
"existence" and "in itself' together in a single phrase is that he thinks it might
commit him to the idea that whatever had "existence in itself' - even things
which are possible of existence in themselves - would be uncaused. In the Kitiib
al-J:iuruf, he claims that one of the two basic senses of "that which exists in
itself' (al-mawjUd bi-dhiitihi) is that whose essence is independent (mii
miihiyyatuhu mustaghniyyatun) of a relation to anything else, namely, to a
cause. The opposite of "that which exists in itself' is "the existent which has
some cause or other".8 Given this assertion, al-Farabi may have worried that
calling anything a mawjud bi-dhiitihi or a mawjud bi-nafsihi - regardless of
whether it was mumkin al-wujud bi-dhiitihi or i;f.arur! al-wujud bi-dhiitihi -
would force him to see the "possible of existence in itself' as causally
independent. In other words, anything which can be spoken of as existing in
itself will be causally independent. The implication is that only God fits al-
Farabi' s description of mawjud bi-dhiitihi, strictly understood.
Al-Farabi resists describing God as necessary of existence in itself because for
him necessity and possibility still carry too many strictly logical connotations,
despite the increasingly existential understanding of those modes that resulted
from the infiltration of the root w-j-d into discussions. As discussed above, God
comes into the Farabian picture only in the sense that God's foreknowledge of
future events could be described as consisting in truthful, and hence necessary,
assertions, assertions which in a way cause the event to be necessary.
In his Siyiisa madaniyya, al-Farabi does at one point (47,11-12) make the
claim that the existence of the world follows necessarily (lazima i;J.aruratan)
from the existence of the First, and does use the terms al-mawjudiit al-mumkina
and al-wujud al-mumkin in referring to material things and to various types of
materiality (56,13 and 57,1-58,14). In the former case, however, this expression
is an isolated occurrence, at least in comparison to his extensive discussion in
the Commentary on the De lnterpretatione of how the existence of future
events follows necessarily from God's foreknowledge of them. And in the latter
case, it seems clear that for al-Farabi, necessary and possible are subordinate to
the primary existential distinctions which he lists in his Aghriii;f.. There, as I
mentioned in Chapter 5, actual and potential, perfect and deficient, and prior and
posterior are presented as the modes of existence the study of which is the
business of the metaphysician; necessary and contingent are not included.9 Nor,

8al-Farabi, K. al-/_luruf. 124,8-11.

9Maqiila ... ft aghriifl al-/_!akfm ft kulli maqiilatin min al-kitiib al-mawsum bi-1-/_luruf. 35,8-11. It is
true that at Int. 13, 23a22-23 (= Fi I- cibiira, 94, 16-95, 1), for example, Aristotle states that "It is
clear from what has preceded that what exists of necessity, will be actual" (phaneron de ek ton
eiremenon hoti to ex anankes on kar' energeian estin = qad ~ahara mimmii qulnii anna mii
 12. Necessity and Possibility (B) 225
as far as I know, does al-Farabi ever describe God as the necessary of existence in
itself, even though he does claim in the Ara' (37,14-38,1) that God is perpetual
of existence in His substance and self (da'imu l-wujudi bi-jawharihi wa-
dhatihi).
The upshot is that for al-Farabi, to be necessary in itself is still a modal
category that applies primarily to propositions, and only secondarily to beings;
and that the distinction between "necessary" and "possible", insofar as it can be
applied to beings, is subordinate to more basic distinctions, be they Aristotelian,
such as that between actual and potential, or Neoplatonic, such as that between
perfect and deficient. Yet al-Farabi's claim that God is the eternal of existence is
an echo of what were, in al-Farabi's time, intense discussions among the
mutakallimun about what it is for something to be eternal in itself. More
specifically, al-Farabi's use of the phrase "eternal of existence in itself' is a
reflection of the tendency of mutakallimun before, during and after his lifetime,
to appeal to that idea in order to solve the problem of the relationship between
God's self (dhat) and God's attributes (~ifiit). The history of that problem will be
the subject of the next chapter.

wujUduhu wiijibun {iaroratanfa-huwa bi-1-.ficl). Although Ackrill 1963, 153, claims that this passage
must be a later addition (he says the whole paragraph "reeks of notions central to the
Metaphysics"), the commentators appear to.have accepted it as a genuinely Aristotelian doctrine.
See, for example, Ammonius in Int. 247,3-5. By contrast Philoponus (in Phys. 4.12 (ad 22lb7-8),
755,26-29), as the mutakallimun and Avicenna will, ties the necessary of existence to the eternal,
rather than to the actual. In this respect Philoponus may have had in mind Aristotle's distinction, at
GC 2.9, 33Sa33-3S, between things which exist of necessity, such as eternal things (ta men gar ex
anankes estin hoion ta aidia), and things which do not exist of necessity, presumably refering to
natural things in the sublunary world, that is, things which come-to-be and pass-away (cf.
Philoponus, in GC 2.9 [ad 335a32], 284,14-21). The equation of what is "always" or "eternal" and
what is "necessary" is asserted again at GC 2.11, 337b33-338a5.
 13. Necessity and Possibility (C)
Materials from the Kalam

In Chapter 7 I argued that Avicenna's distinction between essence and existence

owed as much to preceding kaliim discussions about things and existents as it did
to preceding Greek treatments of this issue. In some texts, in fact, Avicenna's
position on how things and existents relate to each other is closer to that of
tenth-century Ash<arite and Miituridite mutakallimun than it is to that of al-
Fariibi, his fellow faylasuf In much the same vein I shall now argue that
Avicenna's matrix of distinctions - between necessary of existence (wiijib al-
wujud) and possible of existence (mumkin al-wujud), between in itself (bi-
dhiitihi) and through another (bi-ghayrihi), between uncaused and caused, and
between eternal and originated - can be understood fully only when viewed in the
context of the evolution of classical Sunni theology.
In order to explain the nature of the eternality possessed by God and His
attributes, Sunni mutakallimun moved away from earlier, pre-Avicennian
attempts to argue that when we define an eternal thing as "that whose existence
has no beginning in time" (ma la awwala li-wujudihi), what we really mean is
that an eternal thing is uncaused; to later, post-Avicennian attempts to argue
that when we define an eternal thing as "that which has never ceased to be nor
will ever cease to be" (mii lam yazal wa-lii yaziilu), what we really mean is that
an eternal thing cannot possibly not exist, and that therefore an eternal thing is
necessary of existence. Many of the different ways Avicenna constructed his
matrix of distinctions should, it seems to me, be seen as responses to earlier
kaliim ideas and as influences on later kaliim ideas. It is those earlier kaliim ideas
which I shall focus on here.

During the ninth and tenth centuries God's eternality (qidam) was held to be the
most important of the divine attributes. This was because when a theologian said
that God is eternal (qadfm) he used it for two distinct purposes. The Muctazilite
and then the Sunni mutakallimun based one of their proofs for the existence of
God on the contradictory nature of the opposition between the eternal and the
originated. Given the fact that nothing exists apart from what is eternal (qadfm)
and what is originated (muf:ulath); and given the fact that everything originated
(muJ:idath) requires an originator (muJ:idith); then, in order to avoid an infinite
regress, the chain of originated things and originators must terminate in an
originator that is not originated. Since there is nothing - apart from the eternal -
that is not originated, this ultimate originator must be eternal. This eternal,
ultimate originator is God.
The Sunni mutakallimun also used God's eternality for another purpose: to
stress their basic difference with the Mu<tazilites concerning the issue of the
228 Part II: The Avicennian Synthesis

reality of divine attributes such as God's knowledge ('ilm), power (qudra), life
(l:iayiit) and so on. The Sunnis, calling themselves "Upholders of the attributes"
(A~~iib a~-~ifiit) and following the early anti-Muctazilite mutakallim Ibn Kullab
(d. ca. 855), held that God's attributes (~ifiit) were entities whose existence was
real enough that they could not simply be identified with or subsumed under
God's self (dhiit), as early Muctazilites such as Abii 1-Hudhayl (d. 841) and an-
N~'.?3.m (d. ca. 840) had maintained. The Sunnis reasoned that since the
attributes were both real and divine, they were eternal just as God Himself is
eternal.
In short, God's eternality did double service to Sunni theologians, helping
them prove the world's need for an eternal Creator, and helping them argue for
the eternal, and in some sense separate, reality of God's attributes. I shall now
argue that a fundamental incompatibility existed between these two uses of
eternality - the first directed against atheist Materialists (dahriyya), who
believed in the eternity of the world, the second directed against the Muctazilites,
who denied the separate, eternal reality of the divine attributes - and that
resolving, or at least skirting, this problem of incompatibility served as the
main catalyst in spurring the Avicennian turn in Sunni theology. 1
One of the consequences of the use to which both Muctazilite and Sunni
mutakallimun put God's eternality in their proofs of His existence, was that the
term qidam (eternality) underwent a semantic shift, from referring to beforeness
to referring to uncausedness. 2 According to cAbd al-Jabbar (d. 1025), the
Muctazilite al-Jubbifi (d. 915), who taught al-Ashcan (d. 935) before al-
Ashcan•s defection to Kullabism, defined qadfm as "that which comes before in
existence" (mutaqiidimfi 1-wujud). 3 In other books, according to cAbd al-Jabbar,
al-Jubba 0 i defined qadfm as "that whose existence has no beginning" (ma la
awwala li-wujudihi). 4 Despite al-Ash<an's defection from the Muctazilites, both
of al-Jubba'i's definitions turn up in ~e works of later Ash<arites. Ibn Fiirak (d.
1015), for example, uses the "existential beforeness" definition, implying that it
is the one al-Ashcan himself used. 5 Another Ashcarite, al-I:Ialimi (d. 1012),
applied the "beginningless" definition to qadim; but al-I:Ialimi uses the term
ibtidii' ("origin") in place of awwal ("beginning"), reflecting, I believe, the
beginnings of the semantic shift I mentioned.6 This is because "origin" is
ambiguous: like the Greek term arkhe, the Arabic ibtidii' can mean both a
starting-point in time (awwal), as well as a principle (mabda').

1 For Avicenna's influence on subsequent kaliim discussions of God, eternality and the divine

attributes, see Wisnovsky 2003b.

2 On qidam in general see Gimaret 1988, 164-169.
3 'Abd al-Jabbar, al-Mughni V, 233,17-18; 234,15 and 235,l; see also at-taqaddum ft 1-wujud at

234,6.11.
4 'Abd al-Jabbar, al-Mughni V, 233,1-2 and 234,7-9.
3 lbn Fiirak, Mujarrad maqiiliit al-Ash'ari, 26,19-20; 27,19-20; and 42,19-20.
6 al-J:lalimi, K. al-minhiijfi shu'ab al-imiin (ap. al-Bayhaqi, K. al-asma>wa-Nifli.t, 29,12-30,5).
 13. Necessity and Possibility (C) 229
The rationale for this shift from beginninglessness to uncausedness is quite
plain. If, as a mutakallim, my main interest in the concept of eternality is using
it in proofs of God's existence, I shall want the contradictory nature of the
opposition between eternal (qadlm) and originated (muJ:idath) to be basic to the
meaning of the two terms. In other words, I shall have an easier time proving
God's existence if I define qadlm in such a way that it means not only
"beginningless" but also "uncaused". This is because "uncaused" will satisfy my
intuitions about what an opposite of the passive participle mu/:ldath
("originated") should look like, far more completely than "beginningless" will.
According to cAbd al-Jabbar, al-Jubbli 0 i had treated qadlm (eternal) and
mul:zdath (originated) as contradictories: holding a single thing to be qadlm and
mul:zdath results in mutual contradiction (yataniiqa<ju).1 And for a Muctazilite,
this seems obvious enough, given that God is the only thing which is eternal,
everything other than God being originated. As with all contradictories, there is
no middle ground between qadlm and mul:zdath. As a Sunni mutakallim,
however, I shall have made another commitment concerning eternality which
dulls this otherwise sharp opposition. For unlike my Muctazilite counterpart, I
believe that God Himself is not the only subject to which the attribute "eternal"
(qadim) may be predicated: God's attributes (~ifiit), after all, are said to be eternal
as well. I shall briefly review the history of the problem of the divine attributes'
eternality, and then explain why this produced a dilemma for Sunni theologians.

In order to buttress their strict understanding of God's unity, the Muctazilites ha:t
divided God's attributes into "attributes of the self' (~ifiit adh-dhiit) and
"attributes of the act" (~ifiit al-ficl). Attributes of the self, such as God's
"knowledge" ( cilm), could be predicated of God without referring to His creation.
Attributes of the act, such as God's "providing" (rizq), could be predicated of
God only with reference to His creation.
According to the Muctazilites, the attributes of the self, including God's
knowledge, power and life, were in no sense to be understood as separate entities.
Instead, God is "a knower in Himself' {"iilimun bi-nafsihi). 8 Abu 1-Hudhayl
went so far as to claim that an attribute of the self was identical to God,
asserting that God is "a knower through a knowledge which is identical to Him"
("iilimun bi-cilmin huwa huwa). 9 An-Nll.'?µm phrased it differently, saying that
God "never stops being" His attributes of the self. For example, God "never
stops being a knower in Himself' (lam yazal ciiliman bi-nafsihi). 10 Included in
Abii 1-Hudhayl's and an-Nll.'?+lim's lists of attributes of the self was God's being
eternal (qadzm). According to Abu 1-Hudhayl's formula, therefore, God is

7 'Abd al-Jabbar, al-Mughnf V, 233,5.

8 al-Ash'ari, Maqalat al-islamiyyln, 164,13-14. On Mu'tazilite theories of attributes, see
generally Maqalat al-islamiyyfn, 164, 10-165, 13 and 484,5-487, I 4.
9 al-Ash'ari, Maqalat al-islamiyyfn, 165,5.

' 0 al-Ash'ari, Maqalat al-isldmiyyfn, 486,11-12.

230 Part II: The Avicennian Synthesis

"eternal through an eternality which is identical to Him" (qadimun bi-qidamin

huwa huwa); according to an-Na?:?:iim's formula, God "never stops being eternal
in Himself' (lam yazal qadiman bi-nafsihi).
At first glance lbn Kullab's view seems like a combination of Abu 1-
Hudhayl's and an-N~?:iim's formulae. According to lbn Kullab, God "never stops
being a knower through a knowledge" (lam yazal <a/iman bi- cumin). However,
lbn Kullab resists Abu 1-Hudhayl's identification of the attribute with God,
saying instead that God is a knower through a knowledge which "He possesses
(lahu)", and which "is subsistent in [or 'with'] Him (qii"im bihi)". lbn Kullab's
new formula applies in an identical way to 29 further attributes which he lists in
addition to knowledge. In general, he claims, the attributes "belong to His self'
(li-dhiitihi) and are "neither identical to God nor other than He" (iii hiya lliihu
wa-lii hiya ghayruhu). 11 What is striking about the description of lbn Kullab's
doctrine is that "eternal" is not included in his list of 30 attributes. Instead, lbn
Kullab appends a rather cryptic parenthesis to the end of his list, saying about
God that "He is an eternal [thing] who never stops being [so] in [or 'with'] His
names and attributes [innahu qadlmun lam yazal bi-asmii"ihi wa-:jifiitihi)". 12
Why did lbn Kullab single out "eternal" for special treatment when Abu 1-
Hudhayl and an-N~ had seen the eternal as just another attribute of the
self? 13 My guess is that Ibn Kullab is hinting that qadlm is special because it is
a meta-attribute rather than a regular attribute. The most important characteristic
of meta-attributes such as qadlm is that they are attributable not only to God
Himself but also to some or all of God's regular attributes. For example, once a
Sunni mutakallim had determined that God was a mawjud (existent), and that
God therefore possessed the attribute of wujud (existence), he had to decide
whether or not each of God's attributes likewise possessed the attribute of
wujud.

11 al-Ash'ari, Maqalat al-is/amiyyln, 169,2-170,3.

12 al-Ash'ari, Maqalat al-islamiyyfn, 169,9-10.
13 I shall leave aside the problem of whether all or some of the attributes are eternal. As far as

the classical Ash' arites were concerned, only the attributes of the self could be called eternal, the
attributes of the act being originated. Their reasoning was that if an attribute of the act, such as
"providing" (rizq), were eternal, then the object of that act - namely, the creatures for whom God
provides - would also have to be eternal. According to the Maturidites, attributes of the act as well
as attributes of the self are eternal. The Maturidites explained the eternality of the attributes of the
act by appealing to a distinction which for all intents and purposes is the same as Aristotle's
distinction between first entelekheia and second entelekheia as articulated in L l. According to the
Maturidites, the transition from possessing the capability to provide (first entelekheia) to exercising
that capability (second entelekheia), like the transition from knowing how to write but not writing,
to writing, does not involve a change from one thing to another (as would be the case with one of
Aristotle's four types of change), but rather a transition from one thing to itself. On this issue see,
for example, the Maturidite scholar Abu 1-Layth al-Samarqandi (d. 983): "So the correct doctrine
is that God is characterized by all His attributes for eternity, be they of the self or of the act"
(thumma l-madhhabu f-faJ:iiJ:iu anna llaha ta<aJa mawfiifun bi-jaml'i fiflitihi ft l-azali dhatiyyatan
kanat aw ji</iyyatan): Sharh al-jiqh al-absat /i-Abl l:fanifa, lines 527-626 (= 138,4-160,3) at 568-569
(= 147,4-148,1).
 13. Necessity and Possibility (C) 231

The reason why qadlm came to be seen by lbn Kullab as a meta-attribute,

when the Mu 0 tazilites by contrast had seen it merely as one of several attributes
of the self, is that lbn Kullab was so firmly committed to affirming the
attributes' eternality. This was a stance he took in opposition to the 0 Abbasid
Inquisition (mitzna) of 833-848, during which time jurists were forced to confess
the Mu 0 tazilite dictum that the Qur 0 an was created. 14 By Ibn Kullab's reckoning,
the Qur 0 an, conceived of as God's attribute of speech (kalam), was not only
separate in some sense from Him, but also co-eternal with Him. For this reason,
it seems, lbn Kullab wished to distinguish God's attribute of eternality from His
other attributes.
But the special status which Ibn Kulliib assigned to qadlm confused his
followers, who had to decide between two alternatives. The first alternative was
to hold that Ibn Kullab's general principle of attributes, that "God is X through
an X-ness which he possesses", applied equally to "eternal"; in this case God
will be eternal (qadim) through an eternality (bi-qidamin) which He possesses.
The second alternative was to hold that God's eternality (and by extension, His
other meta-attributes, such as existence) was exempt from lbn Kullab's general
principle of attributes; in this case God will be eternal in Himself (bi-nafsihi),
and not through an eternality which He possesses. 15
Neither alternative was free from difficulties. Although the first alternative
enjoys the benefit of consistency, it forces its adherents into a sticky situation:
let us allow, for the purpose of argument, that God is eternal through an
eternality which He possesses. What about the other attributes, such as God's
knowledge, power and speech - are they eternal or not? If they are not eternal,
then God's attribute of speech will not be eternal either, and nor, therefore, will
the Qur 0 an be eternal. But this is precisely the Mu 0 tazilite position a Kullabite
is so keen to steer clear of.
If, on the other hand, the attributes are eternal; and if it is correct to infer that
since God is eternal through an eternality which He possesses, each attribute will
similarly be eternal through an eternality which it possesses; then all the
attributes will possess their own individual meta-attributes of eternality. Now
what about the meta-eternalities which each of the regular attributes possesses -
will each of these meta-etemalities also be eternal through a further meta-meta-
eternality which each meta-eternality possesses? It seems that if a Kulliibite goes
down this route a proliferation of eternalities will be hard to avoid. 16
The upshot was that in order to steer clear of the Mu 0 tazilite position that the
Qur 0 an is created, and in order to pre-empt the infinite-regress problem just

14 On this see Madelung 1974a; van Ess 1965-66 and 1982; and Wolfson 1976, 235-303. On the

attributes in general, see Wolfson 1959 and 1976, 112-234; and Allard 1965.
15 al-Ash'ari, Maqalat al-islamiyyln, 170,4-6; 171,16-172,3; 517,14-16.
16 For evidence that an infinite regress of meta-eternalities was a real worry to Sunni thinkers,

see the Ash'arite mutakallim and mystic al-Qushayri (d. 1072), Shar/:i asma' Allah al-flusnii, 55,8
and 392,5-7.
232 Part II: The A vicennian Synthesis

described, a follower of lbn Kullab will be forced to maintain that lbn Kullab's
general principle of attributes, that "God is X through an X-ness which he
possesses", does not apply to meta-attributes, and that God is eternal not through
an eternality but in Himself. True, the formula that God is eternal in Himself
smacks of Muctazilism, since for all intents and purposes it is identical to an-
N~?iim's formula mentioned above. Nevertheless, the specific concern with
upholding the Qur"an's uncreatedness at all costs and with avoiding a
proliferation of eternalities trumped the worries about sounding too much like
an-N~?iim.
Having decided in favor of God's being "eternal in Himself' (qadim bi-
nafsihi), a Kulllibite is still left with the problem of how to describe the
eternality which His attributes enjoy. Two options present themselves. On the
one hand a Kulliibite could claim that like God, each attribute (:fifa) is eternal in
itself (qad'ima bi-nafsiha). This raises a serious problem, however. First of all,
attributes are not, strictly speaking, selves (anfus or dhawat), but only things
which are predicated of selves. Second, the more explicitly a Kullabite affirms
that God's attributes are eternal, the more causally independent the attributes
might appear to be, given the trend, described earlier, towards seeing
uncausedness and not simply beginninglessness as basic to the idea of eternality.
In other words, if a Kullabite explicitly asserts that the attributes are all eternal
in themselves, he will run the risk of painting a picture in which all God's
attributes can be viewed as separate, uncaused divinities; and that in tum will
expose him to accusations of shirk, or polytheism. 17
On the other hand a Kulliibite could stick like glue to lbn Kulliib's rather
ambiguous formula - that God "is an eternal [thing] who never stops being [so]
in [or 'with' ] His names and attributes (innahu qadimun lam yazal bi-asma'ihi
wa-:jifiitihi)" - and thereby dodge any accusations of shirk which might arise as
a result of explicitly affirming that each of God's attributes is etemal. 18 The only
danger of adhering to lbn Kullab's formula - apart from its opaqueness - is that
it might tempt anti-Kullabites to misrepresent the Kullabite position by
claiming that Kullabites hold God to be caused to be eternal by His attributes.
This is because one of the derived meanings of the preposition bi- (at least in
theological and philosophical texts) is causal: "by", "through" or "by means of'.
In short, if a Kullabite holds that God is eternal bi-:fifatihi, he runs a risk -
admittedly slight - of being accused of implying that God is eternal through or
by His attributes, when what he means is simply that God is eternal in or with
His attributes.
When al-Ashcari abandoned Muctazilism in favor of Kullabism, these were the
dilemmas he found himself facing. If al-Ashcari held both that God was eternal

17 This is precisely the trap which the Mu'tazilite-influenced Shi'i mutakallim ash-Shaykh al-

Mufid (d. 1022) accuses al-Ash'ari of having fallen into: ash-Shaykh al-Mufid, Awii'il al-maqiiliit
ft 1-madhiihib wa-l-mukhtiiriit, 11,20-12,8.
18 al-Ash'ari, Maqiiliit al-isliimiyyln, 172,1-3.
 13. Necessity and Possibility (CJ 233

through an eternality (qadlm bi-qidamin) and that the divine attributes were
eternal, he might be seen to be committing himself to the position that each of
God's attributes was eternal through a further attribute of eternality (qadima bi-
qidamin); and an infinite regress of meta-eternalities would result. If, on the
other hand, al-Ashcan held that God was eternal in Himself and that the divine
attributes were eternal, he might be seen to be committing himself to the
position that each of God's attributes was eternal in itself (qadlma bi-nafsihii);
and a pleroma of causally independent attribute-divinities would result. Al-
Ashcan could, of course, avoid this particular dilemma by denying that the
attributes were eternal at all; but that was unacceptable because it would force
him to admit that the Qur 0 an, qua God's attribute of speech, was created.
Alternatively, he could adhere to lbn Kulllib's cryptic formula - that God "is an
eternal [thing] who never stops being [so] in [or 'with'] His names and attributes
(innahu qadfmun lam yazal bi-asmii"ihi wa-~ifiitihi)" - and leave it at that; but
philosophically speaking that would represent a retreat into ambiguity rather
than an advance towards greater precision, an ambiguity which opponents could
exploit.
Al-Ashcan clearly saw himself as a Kulla.bite as far as the eternality of the
divine attributes was concerned, claiming at one point that the "evidence points
to the eternity of the Creator and His knowledge". t 9 But al-Ashcan never decided
whether God should be seen as "eternal through an eternality" or "eternal in
Himself', probably because of the serious consequences that arose from choosing
one option or the other. In fact the Ashcarite mutakallim lbn Fiirak admitted that
opponents were correct in criticizing al-Ashcan for flip-flopping on this issue.
According to lbn Flirak, al-Ashcan adopts a strict-constructionist interpretation
of lbn Kullab's view (that God is "eternal through an eternality" - qadlm bi-
qidamin) in some texts, and a loose-constructionist interpretation (that God is
"eternal in Himself' - qadlm bi-nafsihi) in other texts. 20 (In commenting on al-
Ash<an's ambivalence lbn Fiirak confirms that the strict-constructionist
interpretation is the one that reflects lbn Kulllib's genuine opinion.) lbn Fiirak's
contemporary, al-Baqilllini (d. 1013), however, follows the loose-constructionist
line, holding that God is eternal in Himself (idh li-nafsihi kiina qadlman). 2 t
A Sunni mutakallim such as al-Bliqillani could feel that, having departed from
lbn Kullab's formula by maintaining that God is eternal in Himself and not
through an eternality, he had at least skirted the old Kullabite dilemma, namely,
affirming God's eternality without allowing an infinite proliferation of meta-

19 al-Ashcan, K. a/-luma<, 12,21-13,2.

20 Ibn Fiirak, Mujarrad maqiiliit al-Ash carf, 326,7-12; see also 28,12-17. The texts Ibn Fiirak
refers to explicitly are al-it,lii/:I ("' K. fi;/ii/:I al-burhiin fi r-radd <a/ii ah/ az-zaygh wa-{-{ughyiin: Ibn
cAsilir, Tabyln kadhib a/-muftarffi-mii nusiba ilii 1-/miim Abl 1-Ijasan al-Ash<ari, 130,3-4), where
al-Ashcan adopts the qadim bi-qidamin view; and al-Mukhtazan (Ibn cAsakir, 133,2-5), where he
follows the qadim bi-nafsihi line.
21 al-Baqilllini, K. at-tamhfd, 29, 18.
234 Part II: The A vicennian Synthesis

eternalities. This choice, however, forced Sunni mutakallimun not only to come
to grips with the obvious question of whether or not God's attributes were
similarly "eternal in themselves", but also to rethink their use of qadlm in
proofs of God's existence.
As I mentioned earlier, the Muctazilite al-Jubba'i had treated qadlm (eternal)
and mul:ufath (originated) as contradictories: holding a single thing to be qadlm
and mul:zdath results in mutual contradiction (yatanaqa¢u). But a Sunni
mutakallim, having avoided the old Kullabite dilemma by holding that God is
eternal in Himself (qadlm bi-nafsihi) rather than eternal through an eternality
(qadlm bi-qidam), will now be inclined to articulate the two opposites as eternal
in itself (qadim bi-nafsihi) and originated in itself (mul:zdath bi-nafsihi), and this
is precisely what al-Baqillani does (though using the slightly less causative li-
nafsihi).22
Strictly speaking, however, al-Baqillani's opposition between qadlm li-nafsihi
and mul:zdath li-nafsihi is one between contraries, not contradictories, because
there is a middle position between them. For although it is impossible to be
both qadlm li-nafsihi and mul:zdath li-nafsihi at the same time, it is possible to
be neither. One of God's attributes (#fat), for example, is eternal (qadlma) - it
most certainly is not originated (mul:zdatha) - but it is difficult to see how an
attribute can be eternal in itself (qadlma li-nafsiha). As I mentioned above, this
is because attributes in general, and the divine attributes in particular, are not
selves, but only things which are predicated of selves; and also because holding
that an attribute is eternal in itself would assign it too much causal
independence, given that uncausedness was so basic to the kalam conception of
eternality.
This raised a problem precisely because, as mentioned above, it was the
contradictory, not contrary, nature of the opposition between the terms qadlm
and mul:zdath, which served as the basis for the mutakallimun's proof of God's
existence: given that everything which is mul:zdath (originated) requires a·
mul:zdith (originator), we must terminate eventually in something which is not
mul:ufath in order to avoid an infinite regress. And since the only thing which is
not mul:zdath is qadlm, the ultimate mul:zdith will be qadlm. But this proof will
not work if the terms in use are contraries rather than contradictories. If the terms
are contraries, as qadlm li-nafsihi and mul:zdath li-nafsihi appear to be, there will
be things - the divine attributes, at least according to the Sunnis - which are
eternal but not eternal in themselves. Put another way, there will be things
which are eternal, but not causally independent. The proof will only work if a
new category is created - "eternal through another" (qadlm li-ghayrihi) - which
could then somehow be identified with "originated in itself' (mul:zdath li-
nafsihi). Only then will qadlm li-nafsihi and mul:zdath li-nafsihi cover all

22 al-Baqillani, K. at·tamhid, 29, 17-30,2.

 13. Necessity and Possibility (C) 235

possible ent1t1es, such that the contradictory nature of their opposition be

preserved.
The first of these two steps - creating a new category of "eternal through
another" (qadim li-ghayrihi) - was certainly acceptable to some Sunni
mutakallimun. In his Qur'iin commentary, entitled Interpretations of the
Sunnis, al-Maturidi (d. 944) claims that the verse "And therein [viz., in Heaven]
shall they dwell forever" (wa-humfiha khaliduna = Surat al-baqara 25) can be
understood as a refutation of the Jahmites who, in their zealousness to protect
God's being the First, the Last, and the Eternal a parte post (al-awwalu wa-l-
akhiru wa-1-baqi), felt constrained to maintain that Heaven would pass away.
Otherwise, the Jahmites reasoned, both Heaven and God would be Eternal a
parte post, and that would be a sin of tashbih, or "likening" - likening anything
created to God, that is.
Where the Jahmites got it wrong, al-Miiturldi reckons, is that they did not
make the requisite distinction between bi-dhatihi ("in itself') and bi-ghayrihi
("through another"). If they had understood and applied this distinction they
would have realized that God is eternal a parte post in Himself (al-baqi bi-
dhiitihi), just as He is First in Himself (al-awwalu bi-dhiitihi); whereas Heaven
and all that it contains are eternal a parte post through something other than
they (baqiyatun bi-ghayrihii).23
It would be too bold to call the distinction which al-MaturldI offers in this
passage a full-fledged and coherent theory. First of all, as the later Ashcarite
scholar al-BayhaqI (d. 1066) would point out, God's attribute of baqa' can be
seen to encompass His etemality a parte ante as well as His etemality a parte
post; whereas the baqa' of heaven and hell, which are created things, can refer
only to their eternality a parte post. 24 Nor is al-Miiturldi's bi-nafsihilbi-
ghayrihi distinction here meant to help explain how an eternal God and His
eternal attributes are eternal in different ways, the former eternal in Himself, the
latter eternal through another. Finally, even if we allow the creation of a new
category of thing, "eternal through another", identifying that new category with
"originated in itself' (muf.idath li-nafsihi) will still run counter to a
mutakallim's basic intuition that qadlm and muf.idath ought to be distinguished
on temporal as well as causal grounds.
In the metaphysics section of his earliest philosophical summa, the lfikma
cAru{iiyya, composed in 1001 when he was 21, Avicenna, like al-Maturldi
before him, embraces the distinction between "eternal in itself' and "eternal
through another" which is implicit in al-Baqilliini's attachment of the rider li-
nafsihi ("in itself') to qadim ("eternal"). However, Avicenna resists applying the
distinction between "in itself' and "through another" to muf.idath ("originated"),

23 al-Maturidi, Ta'wfliit ah/ as-sunna, vol. I, 76,16-77,4. Al-Maturidi also appeals to the bi-
dhiitihilbi-ghayrihi distinction at K. at-tawl;iid, 43,8-9.
24 al-Bayhaqi, K. al-asmii' wa-~-~if?it, 33,6-12.
236 Part II: The Avicennian Synthesis

let alone equating qadlm bi-ghayrihi ("eternal through another") with muJ:idath
bi-nafsihi ("originated in itself'):

L69
lbn Sina, al-/fikmat al-cArii.<jiyya, 4rl4-17
"Eternal" is said of all that has never been non-existent.
Something may be eternal in itself [qadiman bi-dhiitihi] and it may
be eternal through another [qadiman bi-ghayrihi]. Now
"originated" [al-mu~ath] and "brought-into-being" [al-
mutakawwin] consist in that which did not exist at some time or
other [alladhi kiina laysa ft waqtin ma], and which will not exist,
except through another. It cannot escape coming from matter,
because all that is brought into being has been preceded by the
possibility of some existence ffa-qad taqadaamahu imklinu
wujiidin]; otherwise it would not exist.

The problem remains, then, that even if by identifying "originated in itself' with
"eternal through another". we were seen to be referring to an intelligible category
of thing, an Ashcarite such as al-Baqillani will resist describing the attributes as
either mu/:idatha li-nafsihii ("originated in themselves") or qadlma li-ghayriha
("eternal through another"), because doing so raises further kaliim-specific
problems, some familiar, others new. On the one hand, if an attribute is said to
be muJ:idatha li-nafsihii or li-dhiitihii, its possessing a nafs or dhiit ("self') will
assign it too much ontological independence, when - as mentioned above - an
attribute is not, strictly speaking, a self but only something predicated of a self.
What is more, given that muJ:idath must mean originated in time if it is to be
useful in proofs of God's existence which rely on an equation of prior non-
existence and causedness, the attributes, if described as muJ:idatha, will be seen to
be temporally bounded rather than eternal, and the Qur"an, understood as God's
attribute of speech, will be seen to be created instead of uncreated.
On the other hand, if the attributes are qadlma li-ghayrihii, an alarming degree
of otherness (ghayriyya) will infect the relationship between God's attributes
and God's self, with the result that the attributes will fail to satisfy Ibo Kullab's
criterion of being neither identical to nor other than God. What is worse,
positing a significant degree of otherness between God and His attributes will
run the risk of allowing a host of eternal divinities to proliferate - entities which
are caused by something other than they. it is true, but which are still eternal and
separate. In sum, the ambiguous status of the attributes - eternal, yet not
causally independent - flushes out the problems latent in the pre-Avicennian
Sunni trend towards seeing uncausedness as basic to eternality.

In attempting to grapple with this problem the Ashcarites al-Baqillani and al-
l:falimi offer another meaning for qadlm, one that should be familiar from my
discussion in Chapter l l of modality and De Interpretatione 12-13: "that whose
 13. Necessity and Possibility (C) 237
non-existence is impossible". 25 They could even cite the founder of their school
for support of their new definition:

L70
al-Ash<ari:, Kitab al-Luma<, 11,14-15
If the opposite of knowledge were eternal [qadim], it would be
impossible for it not to exist [la-sta~ala an yab.tula].

At first glance, it is hard to see how this new definition does al-Baqillani and al-
l:falimi much good. For even if they apply the new understanding of the eternal
as that whose non-existence is impossible, to their argument for God's
existence, they will still run aground on the rocky shores of the divine attributes.
This is because their proof remains reducible to an appeal to the impossibility of
an infinite regress of causes and effects, so whichever phrase is chosen to
describe the eternal, the term's basic meaning must remain "uncaused" if it is to
be opposed to mul:zdath and thereby useful in their proof of God's existence. As
before, the problem arises when that newly chosen phrase - "impossible not to
exist" - is similarly applied to the attributes.
Nevertheless, there is a sense in which, if another little step is taken, and
being "necessary of existence" (wiijib al-wujud) - not just being "impossible of
non-existence" (mustal:zll al- cadam) - is held to be the basic criterion of
eternality, the horns of the old Kullabite dilemma will at least be blunted.
Eternality, let us recall, was held to be predicable both of God's self and of His
attributes, though perhaps in different ways. Necessity, by contrast, is not
merely a meta-attribute but also a mode of predication, since necessity can be
seen to govern the very act of predicating attributes of a subject.
What I mean is that in propositions such as "God is an existent", "God is a
knower" and "God is a provider", the modal qualifier "necessary that" (or
"necessarily") could be added to describe how each of the predicates or attributes
holds of its subject: for example, it is necessary that God be (or "exist as") a
knower (wiijibun an yakuna [or yujada] lliihu caliman). The result is that
"necessity of existence" can be seen to obtain in the copula which binds the
predicates (i.e., God's attributes) to the subject (i.e., God's self), and not simply
to be predicable of the subject and of each of the attributes in turn, as was the
case with meta-attributes. In short, necessity of existence blunts the horns of the
old Kullabite dilemma better than eternality does, because unlike eternality,
necessity of existence can be held to describe the divine attributes only insofar as
they are predicated of God's self.

25 al-Baqillani, K. at-tamhid, 29,5: "because the non-existence of the eternal is not possible (Li-

anna 1-qadima Iii yajuzu 'adamuhu)"; al-f::!alimi. K. al-minhiij ft shu'ab al-imiin (ap. al-Bayhaqi, K.
al-asmii' wa-~-~ifiit, 33, 1-5): "If He is existent neither from a beginning nor through a cause, then
neither extinction nor non-existence will be possible for Him (/i-annahu idhii kiina mawjudan Iii 'an
awwala wa-lii bi-sababin lam yajuz 'alayhi 1-inqif{ii' u wa-1-'adamu)".
238 Part II: The A vicennian Synthesis

Given the clear advantages of necessity over eternality, why were pre-
Avicennian Sunni mutakallimun so reluctant to take that little step and assert
openly that since eternality is now to be defined as impossibility of non-
existence, and since impossibility of non-existence is identical to necessity of
existence, God's etemality will refer at the most basic level to His necessity of
existence? The short answer is that they were uncomfortable using the term
wiijib in this way. When the early mutakallimiin wished to refer to a
proposition's being axiomatic or intuitively necessary, in the sense that a
necessary proposition expresses an a priori truth, they turned to the term ¢arurl.
For example, the proposition "the whole is greater than any of its parts" is
¢.ariiri.26 (The precise date when ¢ariira first came to be used this way is a
matter of mild debate among historians of early kaliim. 27)
By contrast, early mutakallimun had understood the active participle wi'ljib as
connoting religious or moral obligation (jar¢), and used wiijib with the
preposition 'ala to mean "morally incumbent upon". 28 It is true that many of
those same mutakallimun turned to the root w-j-b - most commonly in the form
of yajibu X min A wa-B ("[Conclusion] X follows necessarily from [premises]
A and B") - to render the idea of syllogistic necessitation. For example, in his
Vindication of the Science of Kaliim, al-Ash<ari uses wajaba in this way:

L71
al-Ashcari, Risalat istiJ.tsan al-khaw<j, ft cilm al-kalam, 92, 12-16
We say: If God were to resemble something, He would resemble it
either in every respect or in one respect. If He resembled it in
every respect, it would follow necessarily [wajaba] that He be
originated in every respect. And if He resembled it in one respect,
it would follow necessarily [wajaba] that like it He be originated to
the extent that He resembled it, given that each of two similar
things will be judged to be similar in the respect in which it

26 Examples of bi-¢-¢arilra, ¢arilratan, ¢arilrat al-'aql, i;Jarilrl (and, Jess often, bi-1-ii;ifirar, etc.,

which is usually paired with bi-f-fab', etc., "by nature", and contrasted with bi-1-ikhtiyar, etc., "by
choice") in an epistemological context in pre-Avicennian kalam include al-Ash' ari, Maqiiliit al-
islamiyyln, 136,10-15; 393,5-14; 480,6-10; an-Nashi' al-Akbar (d. 906), al-Kitab al-awsar ft I·
maqalat, 109,13-110,11 (= #148); Sa'diya al-Fayyumi (i.e., Saadia Gaon), Kitab al-amanat wa 1-
i'tiqiidat, 12,17-13,10; 16,19-20,18; and al-Maturidi, K. at-taw}Jid, 5,13; 7,11; 8,14.17; 42,20; al-
Ash'ari, K. al-Zuma', 41,10-42,15; Baqillani, K. at-tamhid, 7,4-10; 8,6-13; 9,2-15; 52,4-7; Ibn Furak,
Mujarrad maqaliit al-Ash'ari, 12,1-20; 13,25-14,20; 18,21-19,6; 20,9-21,13; 222,16-19; 247,17-
249,22; 284,15-18; 324,4; 328,14-19.
27 Rudolph 1992, 83n.37, and 1996, 252, argues that the distinction between i;Jarilri (necessary)

and muktasab (acquired) knowledge is a product of late-ninth/early-tenth-century Muctazilite

epistemology; see also Frank 1974, 142n.26. However, according to Tritton 1940, 253, the Jate-
Umayyad mutakallim Ja' d b. Dirham (d. 743) advocated a version of 'ilm i;larii.ri; and van Ess
1966, 114-118, argues that the i;Jarii.rlfmuktasab distinction had been in circulation since the very
be~innings of kalam. On this issue see also Abrahamov 1993.
8 On the distinction between wajib and i;Jarilri, see van Ess 1966, 118-119. Instances where

wajib andfari;l appear to be interchangeable include Ibn Furak, Mujarrad maqalat al-Ash'arl, 16,3-
6; 32,7-17; 180,17; and 199,8-9; and al-Baqillani, K. at-Tamhid, 187,J.
 13. Necessity and Possibility (C) 239
resembles the other. Yet it is impossible for what is originated to
be eternal, and for what is eternal to be originated.

The later mutakallimun could, of course, justify adding syllogistic necessity to

wiijib's semantic field by stressing the rather Mu 0 tazilite idea that knowledge of
the concepts of right and wrong was both morally incumbent and intellectually
necessary, since a human endowed with responsibility for his actions (a
mukallaf, that is) both possessed an intuitive knowledge of right and wrong, and
was under a moral obligation to act according to that intuitive knowledge. 29
Still, using the active participle wiijib - as opposed to the verb wajaba/yajibu -
to denote syllogistic necessity remained rare in pre-Avicennian Sunni kaliim.

Unlike most mutakallimun, a number of polymaths and litterateurs of the late-

tenth and early-eleventh centuries had no qualms about asserting that since God
is impossible of non-existence, He is necessary of existence. For example, the
lkhwan a~-$afii° (fl. ca. 985?) at one point refer to God as al-wajib al-wujud in
their Rasii"il. 30 lbn Miskawayh (d. 1030), who despite his death date was a
generation or two older than Avicenna, also argues that "If, as we asserted, the
existence in Him is essential, He could not possibly be imagined to be non-
existent; thus He is necessary of existence, and whatever is necessary of
existence will be perpetual of existence, and whatever is perpetual of existence
will be eternal (wa-idhii kiina l-wujudu fihi ka-mii qulnii dhiitiyyan fa-laysa
yajuzu an yutawahhama macduman fa-huwa wiijibu l-wujudi wa-mii kana
wiijiba l-wujudi fa-huwa dirimu l-wujudi wa-mii kiina dii"ima l-wujudi fa-
huwa azaliyyun)". 31 And the phrase wajib al-wujud appears with the rider bi-
dhiitihi, again in reference to God, in the K. al-amad calii l-abad of al- 0 Amiri. 32
Al-cAmiri in fact appears to be the most likely direct source of Avicenna's
distinction, for a number of reasons. First of all, al-c Amiri was the first to
predicate the entire expression wajib al-wujud bi-dhiitihi ("necessary of existence
in itself') of God. Second, though trained in Baghdad, al- 0 Amiri moved to
Bukhara, and was active in the same Samanid court, and studied in the same
Samanid library where, only a decade or so later, the 21-year-old Avicenna would
write his first philosophical summa, the Ifikma cArul)iyya, which is also the
first Avicennian text in which wiijib al-wujud bi-dhiitihi appears. Finally, in his
Kitiib at-taqrlr li-awjuhi t-taqdlr, al- 0 Amiri again makes the distinction
between necessary, possible and impossible of existence, and illustrates what he

29 Ibn fijrak, Mujarrad maqiiliit al-Ash'arf, 285,7-20; al-Baqillani, K. at-Tamhid, 8,4-5; 379,11-

380,15.
30 Ikhwan as-Sara', Rasii' il II, 471 ,1.
31 Rowson (al~Amad, 233) refers to the Cairo, 1907, edition of Ibn Miskawayh' s K. al-fawz al-

asghar: 15f.; in the edition I have access to (K. al-fawz al-asghar, Beirut, 1901), this sentence
ap~iars o_n 20~10-12. -· ·- . . . _ ..
al-' Amm, K. al-amad 'alii 1-abad, 78,12; wapb al-wu1ud appears without the nder b1-dhiit1h1
at 170,12.
240 Part II: The A vicennian Synthesis

means by necessary of existence with the example "2+2=4'', the same example
which, as will be shown in the next chapter, Avicenna later used in both his
l:fikma cArilt;liyya and in his Mabda' wa-macad. 33 (I admit that "2+2=4" is
such a prosaic example that its use in two or more texts does not constitute
definitive proof of some kind of filiation between them.)
Al-c Amiri also makes the distinction in the Kitab at-taqrfr li-awjuhi t-taqdfr
between that which is necessary of existence in itself(bi-dh-dhat) and that whose
existence is necessary as a consequence of a relation (bi-l-it;lafa), the same
distinction which Avicenna was to make in his Mabda' wa-macad. 34 (I shall
discuss the Mabda' wa-macad, and this passage in particular - L75 - in
Chapter 14.) Of course it is not certain whether Avicenna derived his matrix of
distinctions directly from the courtesy copy of al-Amad which al-cArniri
doubtless left in the Sii.mii.nid library, or whether Avicenna and al-< Amiri read the
same texts there and were independently influenced by them.
In fairness to the mutakallimiln I should mention that the appearance of
"necessary of existence" (wajib al-wujild) - sometimes with the rider "in itself'
(bi-dhatihi), other times without - in descriptions of God, is also a characteristic
of a small number of theological texts produced during the thirty years between
985 and 1015, including those by the Mu<tazilite <Abel al-Jabbii.r and the
Ash<arite belletrist ar-Rii.ghib al-I~fahii.ni. 35 In fact, some scholars have wondered
whether there might be a causal link between, on the one hand, cAbd al-Jabbii.r's
presence in Rayy between 1013 and 1015 and the presence of terms such as
wujilb al-wujud in cAbel al-Jabbii.r's works, and, on the other hand, Avicenna's
presence in Rayy around 1014-1015 and Avicenna's own evolving ideas about
necessary and possible existence. The assumption has been that if there were any
influence, it must have come from the older c Abd al-Jabbii.r and towards the
younger Avicenna. This perception is reinforced by Averroes' famous comment
in the. Tahafut at-tahafut that Avicenna pinched the idea of distinguishing
necessary and possible existence from the Mu<tazilites (wa-huwa fariqun
akhadhahu bnu Sina mina l-mutakallimi:na ... hadha huwa ctiqadu l-
Muctazilati qabla l-Ashcariyyati). 36
What truth might there be to Averroes' claim? In the Kitab al-majmil c ft l-
muJ:il! bi-t-taklif c Abel al-Jabbii.r does seem to toy with the idea that God's
etemality (qidam) can somehow be explained by referring to the necessity of His
existence (wujub al-wujild). 37 <Abel al-Jabbii.r also comes close to making a
distinction in that work between intrinsic and derivative necessity (al-wujilbu li-

33 al-' Amiri, Kitab at-taqrir li-awjuh at-taqdir, 28-30 [MS Princeton 2163 (3938), 26-76]; here I

am following the synopsis by Rowson in his commentary on the Amad, 232-233.

34 Ibn Sina, al-Mabda' wa-1-ma'ad, 3,2-15.
35 I have been told that the phrase wajib al-wujud also appears in the Longer Version of the

Uthulujiya, but I have not been able to check the manuscript.

36 Ibn Rushd, Tahafut at-tahafut, 276,4-9.
37 ap. Ibn Mattawayh, K. al-majmii' fl 1-mu~i{ bi+taklif /, 50,24; 99,21-22; 141,10-12; and

142,1.
 13. Necessity and Possibility ( C) 241
dhiitihi lii li-shay'in siwiihu). 38 In his Mughnf the two trends are joined
together a little more closely. 39 But as far as I can tell, in neither work does
cAbd al-Jabbiir articulate a matrix of distinctions (necessary of existence vs.
possible of existence, in itself vs. through another, uncaused vs. caused), nor
apply that matrix to the problem of God's relationship to the world, in a way
that even approached the clarity and completeness found in Avicenna's
discussions. It is difficult to be definitive about this question because we do not
possess the first three volumes of the Mughnl, the volumes which cover the
topic of taw/:ifd (divine oneness) and hence those which would be the obvious
context for c Abd al-Jabbiir to expound most fully on the issue of wiijib al-
wujud.
Of course, if we accept (as I think we should) that the Uppsala Ijikma
cArucjiyya manuscript we now have access to is really a copy of the very text
which Avicenna composed when he was 21, in the year 1001, then the matrix of
distinctions appears in Avicenna's work- in L58, that is - a dozen years before
Avicenna ever could have laid eyes on c Abd al-Jabbiir in Rayy, and we can
dismiss any claim of cAbd al-Jabbar's personal influence on purely historical
grounds. This is not to say, of course, that cAbd al-Jabbiir's works, or other
Muctazilite works which are now lost, may have been contained in the Siirnfurid
library and read by Avicenna before he composed the Ijikma cArucjiyya. But
given the lack of textual evidence that other Muctazilites thought of this matrix
of distinctions first; and given the hints at that matrix of distinctions by al-
e Amiri, who worked in the Samanid library only a dozen years before Avicenna;
my tentative conclusion has to be either that lbn Sina's work influenced c Abd al-
Jabbiir's, or, as seems more likely, that lbn Sina and c Abd al-Jabbiir came up
with the idea independently, the former in a sustained and precise way and
building on earlier work by al- c Amin, and the latter almost in passing. My
guess is that Averroes' comment might well be an expression of his discomfort
at the astonishingly rapid and widespread incorporation of Avicenna's matrix of
distinctions into Sunni-kaliim proofs of God's existence and into their
discussions of epistemology and God's attributes, in the century or so following
Avicenna's death. 40
The one Ash<arite contemporary of Avicenna who uses the phrase wiijib al-
wujud to describe God is ar-Raghib al-I~fahani. I shall not go into too much
detail concerning the uncertainty over Raghib's death date. Rowson, in the most
recent synopsis of this question, reckons that he flourished around 1010, and I
am convinced by his arguments, which build upon earlier work by Madelung. 41
In his J<tiqadat, Raghib argues that existents and originated things cannot help
but terminate eventually in an existentiator and originator (ilii mujidin wa-

38 ap. lbn Mattawayh, K. al-majmuc ft l-mul/i{ bi-t-taklif /, 51,20-21.

39 'Abd al-Jabbar, al-Mughnf IV, 250,4-15; and XI, 432,11-15.
•o See Wisnovsky 2003b.
41 Rowson 1995 and Madelung 1974b.
242 Part ll: The A vicennian Synthesis

muf.idithin), and that this existentiator and originator must be One, Eternal, and
Necessary of Existence in Himself (wa-anna dhiilika l-mujida wa-l-muf.iditha
yajibu an yakuna wiif.iidan azaliyyan wiijiba l-wujudi li-dhiitihi). 42 Later on,
Raghib explains what he means when he says that God is necessary of existence:

L72
ar-Riighib al-I~fahiini, al-l'tiqiidiit, 56,9-57,l l
Proof that He, may He be exalted, is an existent which is necessary
of existence [mawjiidun wiijibu 1-wujudi] consists in the fact that
whenever we assume or imagine Him to be an existent, it must be
in one of three ways: necessary of existence, impossible of
existence, or possible of existence [immii wiijibu 1-wujiidi aw
mumtani'u 1-wujudi aw mumkinu l-wujudi]. The necessary of
existence is that which, when postulated as non-existent, an
absurdity is logically entailed, e.g., [when] the occurrence of four
from the existence of two and two [is postulated as non-existent]
[fa-l-wiijibu l-wujudi huwa lladhi idhii fariij,a ghayra mawjudin
lazima minhu mul,ziilun ka-l,iu$uli arba'atin min wujudi ithnayni wa-
ithnayni]. The impossible of existence is that which, when
postulated as existent, an absurdity is logically entailed, e.g.,
[when] the occurrence of four from the existence of two and three
[is postulated as existent]. The possible of existence is that
which, when postulated as existent or non-existent, no absurdity
is logically entailed, e.g., [when] the coming of rain in the winter
[is postulated as existent or non-existent].
The necessary of existence is of two types: the necessary of
existence not in itself but through something else [wiijibu 1-wujudi
Iii li-dhiitihi bal /i-amrin iikhara], such as the existence of four
which follows necessarily from the occurrence of two and two; and
the necessary of existence in itself, not through anything else,
namely the Creator, may He be exalted [wa-wiijibu l-wujiidi li-
dhiitihi Iii li-shay'in iikhara wa-huwa l-biirl ta'iilii]. The necessary
of existence is that which, when postulated as non-existent, an
absurdity occurs [idhii fariij,a ghayra mawjudin l,ia$ala minhu
mul,ziilun]; nor does it need, in its existence, anything to make it
exist; and it is eternal, this being God, may He be exalted [wa-lii
mul,itiijunft wujudihi ilii shay'in yujiduhu wa-yakunu azaliyyan wa-
dhiilika huwa lliihu ta'iilii] .... The necessary of existence is that
which has no need, in terms of its existence, for anything other
than itself [wa-l-wiijibu l-wujudi huwa lladhi lii yaftaqiru ft wujudihi
ilii shay' in ghayri dhiitihi]. It is established therefore that it is
correct to say that there is only one necessary of existence in
itself, this being God, may He be exalted.

The wording of Raghib's explanation - which we can date, following Rowson's

conjecture, to around 1010 - is strikingly similar to that found in yet another
passage from Avicenna's al-Mabda' wa-l-ma'iid (L74), composed around 1013

42 ar-R~ghib al-lsfahani, al-l'tiqiidar, 48,15-20.

 13. Necessity and Possibility (CJ 243

(that is, twelve years after his first stab at the idea of necessary existence in the
Ijikma cAru{iiyya), and which I shall discuss in the next chapter.
As with cAbd al-Jabbar, because of the absence of whose Mughnl I-III we
cannot definitively solve the problem of who influenced whom, so the
uncertainty over Raghib's dates forces us to be agnostic about who came up with
the new formulation first, Avicenna or Raghib. Nevertheless, I believe the
burden of proof lies with those who deny that al-c Amiri is the most likely direct
source of Avicenna's matrix of distinctions, since he is the one author whose
works were almost certainly in the right place (Bukhara) at the right time (ca.
1000) to spark the young Avicenna's metaphysical imagination. My task in
Chapter 14, then, is to survey Avicenna's major discussions of necessary and
possible existence, mindful of the strong possibility that his views on this topic
may have evolved over his lifetime.
 14. Necessity and Possibility (D)
The Question of Evolution

As a way of starting this final chapter, I shall dissect the main reason which
prompted Avicenna - and al-c Amiri before him - to seize upon wii.jib with such
enthusiasm. In my opinion wii.jib came to enjoy prominence as a technical term
at the end of the tenth and beginning of the eleventh century, because it is more
syntactically amphibolous than its main rival, rjaruri, and therefore could be
used to signify a wider range of referents. What I mean is that unlike rjaruri,
wii.jib could be construed not only intransitively but transitively as well; and that
in its transitive sense wii.jib could be construed in the active voice as well as in
the passive voice.
Let me list the different types of necessity (wujub) which wii.jib could
plausibly refer to:

1. Intransitive

(i) The modal wujub of a proposition taken in itself: WA.JIB UN an yujada S P

("it is NECESSARY that S is [or 'exists as'] P"), or fa-yujadu S P WA.JIBAN
(''Sis [or 'exists as'] P NECESSARILY"), where Sis a subject and Pa predicate.

(ii) The absolute wujub of simple, eternal and immutable things: X huwa
WAJIBU l-wujUd ("X is NECESSARY of existence"), where X is a simple,
eternal and immutable thing.

(iii) The perfect wujub of a causally self-sufficient thing: X huwa WAlIBU l-

wujud ("X is NECESSARY of existence"), where X possesses perfection
understood as uncausedness.

2a. Transitive (active)

(i) The inferential wujub of a valid syllogism: YAJIBU C min A wa-B ("C
FOLLOWS NECESSARILY from A and B"), where C is a conclusion and A and B
are premises (i.e., necessitas consequentiae).

(ii) The productive wujub of a proposition taken as a premise: yajibu wujubu C

min WUJUBI A wa-min WUJUBJ B ("C's necessity follows from A's
NECESSITY and B's NECESSITY"), where C is a conclusion and A and B are
premises.
246 Part II: The Avicennian Synthesis

2b. Transitive (passive)

(i) The derived wujilb of a proposition taken as a conclusion (e.g., yajibu

WUJUBU C min wujilbi A wa-min wujilbi B ("C's NECESSITY follows from
A's necessity and B's necessity"), where C is a conclusion and A and B are
premises (i.e., necessitas consequentis).

(ii) The hypothetical wujub of generable and perishable things (e.g., li-an
yujada E WAJIBUN an yiljada M 1 wa-M2 wa-M3 ("In order that E exist, it is
NECESSARY that M 1 and M2 and M3 exist"), where E is an intended end, and M 1
and M 2 and M3 are the materials necessary for that end to exist, or to exist
well.

In other words, wiijib could serve in almost all the major roles a faylasilf or
mutakallim might require of it. Understood intransitively, wiijib could be
construed as referring to the necessary qua modal qualifier "necessary that" or
"necessarily" (i.e., qua wiijibun or wiijiban = Int. 12-13); the necessary qua
"invariable" (i.e., qua wiijib al-wujild = Metaph. 5.5 category [4]); and the
necessary qua "causally self-sufficient" (i.e., qua wiijib al-wujud = Neoplatonic
perfection). Understood transitively, wiijib could be construed as referring to the
necessary qua "necessary inference" (i.e., qua fjiib = Metaph. 5.5 category [5]);
the necessary qua "necessitating cause" (i.e., qua miljib = Metaph. 5.5 category
[5]); the necessary qua "necessitated effect" (i.e., qua mujab = Metaph. 5.5
category [5]); and the necessary qua "indispensable" (i.e., qua miljab = Metaph.
5.5 categories [l] and [2] and Phys. 2.9).
The only sense of the necessary which the older term ¢arilrl seemed to express
better than wiijib did was the compulsory (= Metaph. 5.5 category [3]). And in
fact Avicenna regularly refers to accidental material factors which prevent a cause
from producing its effect as (iarilriyyiit. But that was a relatively small price to
pay for all the benefits he accrued by using wiijib. Most important of these is
that wiijib was amphibolous enough to be used in a context where causal
relationships are not at issue as well as in a context where causal relationships
are at issue. The chasm between the self-sufficiency of that which possesses
perfection and the productivity of that which is above perfection could be bridged
by employing wiijib to describe both types of necessity. This allowed Avicenna
to use wiijib in the phrase wiijib al-wujud bi-dhiitihi to refer to God's intrinsic
self-sufficiency at the same time as he used it to refer to God's causal
productivity.
What is more, wiijib could express the derived necessity of an effect as well as
the productive necessity of a cause. This allowed Avicenna to use wiijib in the
phrase wiijib al-wujild bi-ghayrihi ("necessary of existence through another") to
refer to the necessity an effect derives from its cause. Avicenna thereby gained a
 14. Necessity and Possibility (D) 247

decisive advantage over the mutakallimun, because while Avicenna could make a
case for identifying wajib al-wujud bi-ghayrihi ("necessary of existence through
another") and mumkin al-wujud bi-dhiitihi ("possible of existence in itself'),
given that the latter referred to the effect's intrinsic contingency, there was little
chance - as I showed in the previous chapter - that a mutakallim could plausibly
identify qadim bi-ghayrihi ("eternal through another") and mul:ufath bi-dhiitihi
("originated in itself').

Having taken stock of the numerous advantages which impelled Avicenna to

construct his matrix of distinctions around the term wajib, I shall now tum to
the question of whether or not - and if so, how - Avicenna's uses of that matrix
of distinctions in different works might represent an evolution in his views. For
the sake of convenience I shall repeat what Avicenna says in the l:fikma
cAru(iiyya:

L73 (= L58)
Ibn Sina, al-lfikmat al-cArutjiyya, 3v16-4r12
"The Necessary" [al-wajib] is that which is necessary of existence
in respect of the way it is [a¢-¢arariyyu l-wujudi <aJa ma huwa
<alayhi], this being either in itself [bi-dhatihi] (such as the
principle of existents [ka-mabda'i l-mawjadati]) or through
another (such as the fact that two and two are four). The necessary
is either eternal (such as the principle of existents) or is at one
time and not another [fi 1:talin dUna 1:talin] (such as the eclipse of
the moon at that moment). All that is necessary of existence in
itself has no cause. Whatever has a cause will be neither necessary
of existence in itself nor impossible of existence in itself (for
otherwise, it would never come to exist); with respect to itself [fi
/:taythu dhatihi], therefore, it [i.e., that which has a cause] will be
possible of existence [mumkinu l-wujadi], while [at the same time]
being necessary of existence through its cause. Now the existence
of whatever has no cause is not itself divisible into two states [wa-
ma La cillata lahu fa-inna wujiida dhatihi La yanqasimu min
/:talatayni], in virtue of which it [i.e., the existence of whatever
has no cause] would come to be caused in both states; for there
would be no way out of being caused, nor any escape from
causedness. All that is subject to change is in these two states,
neither one of which it possesses in itself; rather, it possesses
both of them through a cause (there being no alternative to them
[i.e., there being no alternative to being in the two states]). Thus
all that is subject to change will itself be caused and possible,
whereas all that is necessary of existence in itself will be
necessary of existence in every respect, and no type of change
whatsoever will be attributable to it. "The possible" is the
existent which is not necessary [laysa bi-¢aruriyyin]. "The
possible" is said to be whatever is not impossible; and "the
possible" is said to be whatever is not impossible and which
248 Part l/: The A vicennian Synthesis
exists and [then] is non-existent, and [in general] whatever does
not exist at some time or other.

In this passage Avicenna presents both in itself/through another and

always/sometimes as alternative ways of distinguishing between types of
necessary existence. He opts for the in itself/through another distinction, and
holds it to be equivalent to uncaused/caused. This appears to be because of his
reluctance to advocate a bald parallelism between the modal operators
"necessary", "possible" and "impossible" and the temporal operators "always",
"sometimes" and "never". As Back has shown, in later works such as the Shi/ii'
Avicenna will distinguish between six ways necessity operates in predications
when it is tied to varying time periods. 1 Even at this early stage Avicenna seems
aware that if linking the modal operators to the temporal operators produces so
many alternative types of necessity, the in itself/through another distinction will
be easier to mesh with the desired uncaused/caused distinction than the
always/sometimes distinction will.
What I mean is that "necessary eternally" and "necessary sometimes" will
entail uncausedness and causedness, respectively, far less strictly than "necessary
in itself' and "necessary through another" will. As we saw with the
mutakallimun's struggle to determine the precise nature of the divine attributes'
etemality, there are things which are eternal but which are not uncaused. For
Avicenna, of course, it is the heavens rather than the divine attributes which fit
this category. But the fundamental challenge remains the same for the
mutakallimun and for Avicenna: to find a coherent way to distinguish between
eternal things which are uncaused and eternal things which are not uncaused.
So how exactly does one go about establishing a strict equivalence between
"being necessary of existence in itself' and "being uncaused", and between "being
necessary of existence through another" and "being caused"? It is in response to
this question that Avicenna begins to articulate - in the /fikma 'Aru{iiyya and
the Mabda' wa-ma'ii.d - the two approaches which Davidson would differentiate
in Avicenna's later works.
Avicenna claims in the /fikma 'Aru{iiyya passage above (i.e., L58/L73) that
"the existence of whatever has no cause is not itself divisible into two states [or
'modes'] (wa-mii Iii 'illata lahu fa-inna wujUda dhiitihi la yanqasimu min
J:iiilatayni), in virtue of which [two states] it [the existence of whatever has no
cause] would come to be caused in both states; for there would be no way out of
being caused, nor any escape from causedness". The question arises, how ought
we to interpret Avicenna's phrase "divisible into two states [or 'modes']"
(yanqasimu min J:iiilatayni)? Two alternatives present themselves. The first
alternative (1) takes the "two states" to refer to (i) the state of non-existence
which precedes X's coming-to-be or which follows X's passing away, and (ii)

1 Back 1992.
 14. Necessity and Possibility (D) 249

the state of existence currently possessed by X. The advantage of this

interpretation is that it coheres with the final assertion of L58/L73, that the
"possible" refers to that which "exists and [then] is non-existent (yujadu wa-
ya cdamu), and [in general to] whatever does not exist at some time or other".
Nevertheless, alternative (1) suffers from some grave flaws. First, AviceIUla
does not claim that Xis divisible into two states, but that the existence of X is
divisible into two states. It is difficult to imagine Avicenna's making the
assertion that the existence of something is divisible into non-existence and
existence. An advocate of interpretation (1) might argue that within X itself
existence is mixed with non-existence, since X is infected with some degree of
non-existence by virtue of the fact that at one point in the past X was non-
existent and at some point in the future X will once again be non-existent. But
even if we were to concede this point, it is X, and not X's existence, which
would be divisible into non-existence and existence.
The second flaw is that interpretation ( l) fails prima facie to solve one of the
main philosophical challenges facing the Sunnr mutakallimun and AviceIUla,
namely, figuring out a way to distinguish eternal things which are uncaused
(God's self; God) from eternal things which are not uncaused (God's attributes;
the heavens). Because interpretation (1) takes the two states to refer to the non-
existence and existence possessed by things that come-to-be and pass-away at
different points in time, the divisibility cannot apply to eternal things which
never come to be or pass away. The upshot is that if (1) is the correct
interpretation of what Avicenna meant when he said "divisible into two states [or
'modes']'', being necessary of existence in itself will entail uncausedness no
more strictly than being eternal will. This is because eternal things which are
not uncaused - God's attributes, or the heavens - will not be accounted for in
Avicenna's scheme, since they cannot be described as being divisible into the
two states of non-existence and existence.
Let us tum then to the second alternative, interpretation (2). Interpretation (2)
takes AviceIUla's assertion that "the existence of whatever has no cause is not
itself divisible into two states [or 'modes'], in virtue of which [two states] it [the
existence of whatever has no cause] would come to be caused in both states; for
there would be no way out of being caused, nor any escape from causedness'', to
be a somewhat tentative attempt to appeal to the old "every composite requires a
composer" axiom, common to both Neoplatonists and mutakallimun, the gist of
which is that everything simple is uncaused, while everything composite or
complex is caused. 2 Two further lines of interpretation - (2a) and (2b) - can be
distinguished.
The first line of interpretation - (2a) - holds that Avicenna is referring to the
fact that the existence of all beings in the world other than God is conceptually
divisible into the two modes (i) "necessary through another" (wiijib al-wujud bi-

2 On the history of this argument, see Davidson 1987, 146-153.

250 Part II: The Avicennian Synthesis

ghayrihi) and (ii) "possible in itself' (mumkin al-wujud bi-dhatihi). God's

existence, by contrast, is not conceptually divisible into two modes, since God's
existence is only ever "necessary in itself' (wiijib al-wujud bi-dhiitihi). The
conceptual divisibility of the existence possessed by all beings other than God
into the two modes "necessary of existence through another" and "possible of
existence in itself', entails that their existence be thought of as a composite of
those two modes, just as the conceptual divisibility of each sublunary substance
into form and matter entails that it be thought of as a composite of matter and
form.
According to interpretation (2a), Avicenna's train of thought is that since
being a composite entails being caused; and since everything composite requires
a composer (the same Arabic term - murakkab - means both "composite" and
"composed"); all existence which is conceptually divisible into, and thus a
composite of, the two modes "necessary through another" and "possible in itself'
will therefore be caused. By contrast, all existence which is not conceptually
divisible into and thus not a composite of the two modes, will be uncaused. The
only being whose existence fits that latter category is God.
As attractive as (2a) might be, it, like interpretation (1), suffers from a serious
difficulty. In the very next line after Avicenna states that the existence of
whatever is divisible into two states or modes will be caused, he asserts that "All
that is subject to change (kullu mutaghayyarin) is in these two states, neither
one of which it possesses in itself' (laysa wa-lii wiil:zidata minhumii lahu bi-
dhatihi). An opponent of (2a) could argue, then, that if the two states or modes
into which the existence of all beings other than God is divisible, were meant to
refer to "necessary through another" and "possible in itself', Avicenna would not
have claimed so baldly that such a thing possessed neither state in itself. The
reason, of course, is that a being whose existence is both "necessary through
another" and "possible in itself', does possess one of the two states in itself:
namely, being possible.
If the "two states" should not be taken to refer to the two modes "necessary of
existence through another" and "possible of existence in itself', I am left with
interpretation (2b). This alternative holds that the "two states" refer to (i) the
state of potentiality or imperfection that characterizes the existence of a
changeable thing - X - when X is understood as a potential or imperfect Y; and
(ii) the state of actuality or perfection that characterizes the existence of a
changeable thing - X - when X is understood as an actual or perfect X.
What makes interpretation (2b) plausible is that it coheres with Avicenna's
immediately preceding assertion that everything which is changeable (kullu
mutaghayyarin) is in these two states. As with interpretation (2a), interpretation
(2b) takes the two states - potentiality/actuality or imperfection/perfection -
which characterize the existence of a changeable thing at a given time, to entail
compositeness, and thus causedness.
 14. Necessity and Possibility (D) 251

Nevertheless, like interpretation (1), interpretation (2b) fails prima facie to

address one of the main challenges facing Avicenna and the mutakallimun, which
is, once again, figuring out a way to distinguish eternal things which are
uncaused (God's self; God) from eternal things which are not uncaused (God's
attributes; the heavens). This is because Avicenna's assertion that everything
which is changeable falls into these two states could be taken to restrict
"divisibility into two states" to sublunary things which come to be and pass
away. A defender of (2b) could argue, I suppose, that the existence of some
eternal things - celestial spheres and celestial souls - can still be characterized by
potentiality and imperfection, and that therefore their existence can be taken to be
divisible into, and hence a composite of, actuality and potentiality, or perfection
and imperfection.
But the defender of interpretation (2b) will still be unable to account for the
celestial intellects, which Avicenna clearly takes to be entirely actual and perfect,
at least according to L53. That is why, after all, he felt the need to appeal to the
idea of God's being not only perfect but also above perfection. If the existence
of the celestial intellects is not, therefore, a mixture of actuality and potentiality,
or a mixture of perfection and imperfection, they will not be divisible into, and
hence composites of, those states. And if their existence is not composed of
those states, it will be uncaused just as God's existence is uncaused. So while
interpretation (2b), with a little tweaking, is better than interpretation (1) at
accommodating at least some eternal things (i.e., the celestial spheres and the
celestial souls), interpretation (2b) still fails to entail a distinction between
caused and uncaused among eternal things which are all fully actual and perfect;
between the celestial intellects and God, that is.
The upshot is that none of these interpretations - (1), (2a) or (2b) - is entirely
satisfactory. Let us assume that this ambiguity, or perhaps confusion, is real,
and not the result of a scribal error which has corrupted the l:fikma cAru{i.iyya
manuscript we have and left it without some key phrases or sentences. We then
seem to have little choice but to view L58/L73 as nothing more than the young
Avicenna's first stab at his new matrix of distinctions. L58/L73 should be taken
as evidence, in other words, that at this very early stage of his career Avicenna
still inhabited the thought-world of tenth-century kalam, a world in which the
notions of eternality and uncausedness were bound so closely together that they
were held to be mutually implicative; yet at the same time he struggled to flesh
out and usefully apply al-c Amiri's new and exciting suggestion that God is
necessary of existence in Himself. 3

3 In fact Avicenna's assertion (/fikma <Arii(liyya 82r5-6) that "It has already been said that the
First is necessary of existence in itself (wa-qad qila inna 1-awwala wajibu al-wujiidi bi-dh-dhati)",
could be taken to refer to al-cAmiri's earlier use of that phrase to describe God. Alternatively the
qad qi/a may refer merely to Avicenna's own identification of the First with the necessary of
existence in itself in L58n3. If that were the case, however, I would have expected to see the
first-person qad qulna ("we have already said") rather than the impersonal qad qila.
252 Part II: The Avicennian Synthesis

The rest of this chapter focuses on Avicenna's attempts to sort out this
confusion by applying his matrix of distinctions in a more coherent way. In the
Mabda' wa-ma<ti.d, Avicenna will use a more purely ontological method of
constructing the matrix of distinctions; and then, in the Iliihiyyiit of his Shi/ii',
he will view his newly sharpened distinction between essence and existence as a
new and better way to provide the conceptual divisibility, compositeness arxl
causedness that he appealed to in the lfikma cAriujiyya. Finally, in the Ishiiriit
wa-tanbihiit, Avicenna will allude, in a very compressed manner, to many of the
different approaches he articulated in preceding works.

Although from later descriptions of the work Avicenna almost certainly

discussed metaphysical issues in The Available and the Valid (al-lfii#l wa-l-
malJ,~ul), a book he composed immediately following the lfikma cAru<jiyya, it
is sadly lost, and we must jump ahead 12 years, to 1013, when Avicenna
composed his next summa, The Origin and the Destination (al-Mabda' wa-1-
ma ciid). Given the fact that the lfii~il wa-malJ,~ul is not extant, and given the
fact that Avicenna's correspondence with al-Biriini (which was also composed
soon after the lfikma cAru<jiyya) contains no information directly relevant to the
distinctions between necessary existence and possible existence and between
being in itself and being through another, it is difficult to say whether or not
there were any intermediate steps which Avicenna took between his discussion in
the lfikma cAru<jiyya and his discussion in the Mabda' wa-maciid.
In any case, Avicenna takes a very different approach in the Mabda' wa-
maciid to describing and applying his matrix of distinctions. To be precise, he
employs some terminology (e.g., matiVidhiifuri<ja ghayra mawjudin) which is
identical to that employed by ar-Raghib al-I~fahiini in L72; refers to a concept
(viz., the inconceivability of the necessary's non-existence) which is identical to
that proposed by Ibn Miskawayh; and appeals to an example (viz., 2+2=4) which
is identical to that offered by al-c Amiri:

L74
lbn Sina, al-Mabda' wa-l-ma<ad, 2,4-17
1. On defining the necessary of existence and the possible of
existence
The Necessary of existence is the existent which, when postulated
as non-existent [matli furirj,a ghayra mawjudin], an absurdity
occurs. The possible of existence is that which, when postulated
as either non-existent or existent, no absurdity occurs. The
necessary of existence is the necessary [arj,-rj,aruri], while the
possible of existence is that in which there is no necessity at all
[alladhi la rj,arurata fihi], i.e., neither in its existence or in its non-
existence. This is what we mean by "possible of existence" in this
context. If what is in potentiality were what is meant by "possible
of existence", "possible" would be said of all that could correctly
 14. Necessity and Possibility (D) 253
[be said] to exist [<aJa kulli ~al:zl/Ji l-wujudi], this having been
explained in detail in the Logic.4
Next, the necessary of existence may be in itself [bi-dhatihi]
and it may be not in itself [la bi-dhatihi]. That which is necessary
of existence in itself [wiijibu l-wujudi bi-dhiitihi] is that on whose
account [li-dhiitihi] (and not on account of anything else,
whichever thing that might be) postulating its non-existence
becomes absurd. The necessary of existence not in itself is that
which becomes necessary of existence on account of postulating
[the existence of] something which is not [identical to] it, such as
the fact that four is necessary of existence not in itself, but [only]
when two and two are postulated; and the fact that being burned and
burning [i.e., burning something else] are necessary of existence
not in themselves, but [only] when contact between the natural
active potentiality and the natural passive potentiality (I mean
[the potentiality] to bum and be burned) is postulated.

In L74 the matrix of distinctions articulated so tentatively in L58/L73 is

restructured. The most striking difference between Avicenna's first discussion of
necessary and possible existence in the l:fikma cArurjiyya (i.e., L58/L73),
composed around 1001, and his discussion here in the Mabda, wa-ma<ad (i.e.,
L 74), composed around 1013, lies in the fact that in the l:fikma <Arurjiyya
Avicenna begins by distinguishing between the necessary of existence in itself
and the necessary of existence through another, whereas in the Mabda, wa-
ma<ad Avicenna begins with a distinction between the necessary of existence
and the possible of existence.
What is the significance of this change? My sense is that in the Mabda, wa-
ma ciid Avicenna is formulating an alternative approach to the idea of the
necessary of existence in itself, an approach which he would modify and partiaily
abandon in the Jliihiyyiit of the ShifiiJ, and then return to in the Najiit, into
which work this and the following MabdaJ wa-maciid passage were copied and
pasted. Let me explain. As I discussed above, Avicenna began in the l:fikma
cAru¢iyya with a distinction between the necessary of existence in itself, and the
necessary of existence through another. He then went on immediately to imply
that the former is uncaused and the latter is caused because the former is simple
and the latter composite; composed, that is, of two modes of existence, necessary
of existence through another and possible of existence in itself, if interpretation

4 This may refer to the Shorter Summary of Logic (al-Mukht~ar al-asghar ft 1-man{iq),

composed at roughly the same time as the same time as the Mabda' wa-ma<ad; see Gutas 1988,
112. Given the fact that Avicenna copied the Shorter Summary and pasted it into the first section of
the Najiit, which is devoted to logic; given the fact that this passage (i.e., L74) is itself copied and
pasted into the Najiit (K. an-Najiit, 366,3-367,1); and given the fact that Avicenna did not excise the
internal reference in the copied-and-pasted version in the Najiit, the Shorter Summary passage to
which Avicenna seems to refer in L74 above is almost certainly that reproduced in K. an-Najiit,
38,7-39,9.
254 Part II: The Avicennian Synthesis

(2a) is correct; or composed of the states of potentiality and actuality, or of

perfection and imperfection, if interpretation (2b) is correct.
In the Mabda' wa-maciid (2,5-11), on the other hand, Avicenna begins with
the claim that the basic distinction is between the necessary of existence, which
by definition is that whose non-existence is inconceivable; and the possible of
existence, which by definition is that whose non-existence and whose existence
are conceivable. In other words, Avicenna appeals to what he seems to think is
an a priori difference between necessary and possible existence. It is only in the
second half of the chapter (2, 12-17) that we encounter the distinction between the
necessary of existence in itself and the necessary of existence through another,
the distinction which fronted the discussion in the lfikma cAru{iiyya. In other
words, Avicenna now seems to feel that the ontological distinction between
necessary of existence and possible of existence is sufficiently intuitive that he
need not appeal to the latter's compositeness to explain its causedness, a move
which would have pushed the in itself/through another distinction to the fore, as
was the case with L58/L73, assuming interpretation (2a) or (2b) is correct.
One effect of this new move is that Avicenna can identify the necessary of
existence through another and the possible of existence in itself more strictly in
the Mabda' wa-maciid than he had been able to in the lfikma cAru<jiyya. In the
Mabda' wa-maciid Avicenna not only claims that the two modes "necessary of
existence through another" and "possible of existence in itself' appear to be
extensionally identical (that is, any being of which one mode is predicated will
also be subject to predication by the other mode), he hints that the two modes
are intensionally identical as well (that is, being necessary of existence through
another has the same meaning as being possible of existence in itself).
Remember that in the lfikma cAru{iiyya Avicenna highlighted the intensional
difference between being necessary of existence through another and being
possible of existence in itself, in order to flush out the compositeness, and hence
causedness, of any being of which both modes are predicated (assuming, that is,
that interpretation [2a] is correct). In the Mabda' wa-maciid, by contrast,
compositeness is no longer needed to explain the causedness of all beings other
than God, so the two modes can now be identified intensionally.
In the Mabda' wa-maciid a new argument is introduced to identify the
necessary of existence through another with the possible of existence in itself:

L75
Ibn Sina, al-Mabda' wa-l-maclid 1, 3,2-15
3. On the fact that the necessary of existence through another is
possible of existence in itself
All that is necessary of existence through another is possible of
existence in itself, because the necessity of the existence of
whatever is necessary of existence through another, is a
consequence of some relation and some connection [tiibicun li-
nisbatin mli wa-i(iiifatin]. Taking into consideration relation and
 14. Necessity and Possibility (D) 255
connection is something other than taking into consideration the
very thing itself which possesses a relation and a connection. So
necessity of existence is established [in this case] only by
considering this relation. Taking into consideration the self alone
will logically entail either necessity of existence, possibility of
existence or impossibility of existence ffa-<tibiiru dh-dhliti
waJ:ufaM. lii yakhlu immii an yakuna muqta{iiyan li-wujubi l-wujudi
aw muqta{iiyan li-imkiini l-wujiidi aw muqta{iiyan li-mtinii'i l-
wujudi]. Now it cannot logically entail the impossibility of
existence, because everything whose existence is impossible
either in itself or through another will not exist; nor can it entail
necessary existence, since we have already asserted that [with
regard to] that whose existence is necessary in itself, the necessity
of its existing through another is impossible. It remains therefore
that taking into consideration itself [alone], it will be possible of
existence; taking into consideration the appearance of the relation
to that other, it will be necessary of existence, and taking into
consideration the disappearance of the relation which is to that
other, it will be impossible of existence; [all this] given that its
self, in itself, with no conditions [attached], is possible of
existence. It is now clear that all that is necessary of existence
through another is possible of existence in itself. 5

Following upon the a priori distinction between being necessary of existence

and being possible of existence articulated the first Mabda' wa-maciid passage
(i.e., L74), Avicenna argues in the second Mabda' wa-maciid passage (i.e.,
L75) that we can conceive of the existence of something which is possible of
existence only in the context of a causal relation, a causal relation which makes
the possible of existence in itself necessary of existence through another. The
necessary of existence, by contrast, can be imagined to exist entirely on its own,
without any causal relation having been supposed.
The causedness of that which is necessary of existence through another and
possible of existence in itself therefore derives from the fact that its necessary
existence is only ever a consequence of a relation, and not from the fact that
possessing two mutually implicative modes (being necessary of existence
through another and being possible of existence in itself, according to
interpretation [2a]) or states (being potential and being actual, or being perfect
and being imperfect, according to interpretation [2b]) entails a compositeness
which requires a composer, and hence a cause.
The difference between the lfikma cAru(iiyya's appeal to compositeness to
explain the causedness of the necessary of existence through another and the
possible of existence in itself, and the Mabda' wa-maciid's appeal to our
intuition that whatever is necessary of existence through another will by
definition be contingent upon that relation with the other and hence caused, is
one that will play out in the middle part of Avicenna's career. As I mentioned

5 This passage is reproduced in K. an-najat, 367,7-368,9.

256 Part II: The A vicennian Synthesis

above, this is mainly because the Mabda' wa-maciid's treatment of this issue
was lifted and plopped into the Najiit.

How does Avicenna's construction of the matrix of distinctions change from the
l:fikma cAru~iyya and the Mabda' wa-maciid to the Iliihiyyiit of the Shifii'? In
Iliihiyyiit 1.6, Avicenna begins as he did in the Mabda' wa-maciid, by making
an ontological distinction; but he phrases it slightly differently:

L76
lbn Sina, Kitab ash-shija'/lliihiyylit (1) 1.6, 37,7-11
Avicenna, Liber de philosophia prima sive scientia divina I-IV,
43,8-15
Things which fall into [the category of] existence bear dividing in
the mind into two subdivisions [inna l-umura llatl tadkhulu ft l-
wujudi taf.itamilu ft l- <aqli l-inqisiima ilii qismayni = Dicemus igitur
quod ea quae cadunt sub esse possunt in intellectu dividi in duo].
One is that which, when considered in itself, its existence does not
follow necessarily [ma idha <tubira bi-dhatihi lam yajib wujuduhu =
quod, cum consideratum fuerit per se, eius esse erit necesse] (and
obviously its existence is not impossible either, for otherwise it
would not fall into [the category of] existence), and this thing is
found to be in the domain of possibility [fi f.iayyizi l-imkiini =in
termino possibilitas]. The other is that which, when considered in
itself, its existence follows necessarily. We say that the necessary
of existence in itself has no cause, and that the possible of
existence in itself has a cause.

After listing the various implications of the fact that the necessary of existence
in itself is uncaused - such as being necessary of existence in all respects, and
not being equivalent to anything else, or composed of several other things, or
essentially similar to anything else - Avicenna then goes on to explain why the
necessary of existence in itself is uncaused:

L77
Ibn Sina, Kitiib ash-shifii'/Iliihiyyiit (1) 1.6, 38,1-5
Avicenna, Uber de philosophia prima sive scientia divina I-IV,
44,24-31
Now the fact that the necessary of existence in itself has no cause
is obvious, because if the necessary of existence had a cause for its
existence, its existence would be through it [i.e., through that
cause]; and everything whose existence is through something,
when it is considered in itself apart from anything else, existence
will not follow necessarily for it. Everything which, when
considered in itself apart from anything else, and for which
existence does not follow necessarily, will not be necessary of
existence in itself. Thus it is clear that the necessary of existence
has no cause.
 14. Necessity and Possibility (D) 257

Even though Avicenna's structuring of his matrix of distinctions in the

Jliihiyyiit is very close to that found in the Mabda' wa-maciid, there seems
even so to be a subtle development. Avicenna does not appeal in the Iliihiyyiit to
the idea that a logical absurdity will occur when the necessary of existence is
held to be non-existent, an appeal which is prominent in the first Mabda' wa-
ma<iid passage (i.e., L74). Here in the lliihiyyiit, in fact, Avicenna seems to be
combining elements from the two Mabda' wa-ma<iid passages discussed above.
From the first Mabda' wa-maciid passage (i.e., L74) Avicenna draws the
ontological basis for the distinction; it appears in the lliihiyyiit passage as an
appeal to the permissibility in the mind (taf:itamilu ft [-<aql) of making the
distinction. From the second Mabda' wa-ma<iid passage (i.e., L75) Avicenna
draws the considered in itself/considered in relation distinction.
Later on in the chapter, however, Avicenna makes an important move: he
introduces the term miihiyya (essence) into the discussion. Avicenna makes a
distinction between a miihiyya which is sufficient for a thing to exist and a
miihiyya which is not sufficient for it to exist (38, 17-18). The existence of that
which has a miihiyya sufficient for it to exist will be uncaused (38,18-39,1);
whereas the existence of a thing which has a miihiyya insufficient for it to exist
will be caused (39,2-4).
Avicenna's distinction is less important for what it actually expresses than
what it may have brought to his mind. What I mean is that Avicenna tosses the
term miihiyya into this discussion almost as an afterthought. Nevertheless, his
appeal to miihiyya, however incidental it may be to his present argument, is
highly significant, because it is at this point, I believe, that it occurs to him that
he can appeal to the distinction between essence and existence - so clearly held
in the previous chapter (Iliihiyyiit 1.5) to be intensionally distinct yet
extensionally identical - in much the same way that he appealed to the notion of
divisibility into two modes or states in the lfikma cAru<Jiyya: to explain
causedness through compositeness.
It is true that in lliihiyyiit 1.6 Avicenna's equation of being necessary of
existence in itself and being uncaused rested on much the same basis as it had in
the Mabda' wa-ma<iid. But in lliihiyyiit 8.4 Avicenna will make the claim,
when discussing why the First is necessary of existence in itself, and why the
First is the only thing in the universe which is necessary of existence in itself,
that the First is not distinguishable into miihiyya (essence) and wujud
(existence) (343,10-15); that the necessary of existence cannot be a composite
(345,6-346,8); that therefore miihiyya and wujud are identical in the necessary of
existence in itself (346,8-12); and that everything else is caused, because in them
miihiyya must be combined with wujud in order to exist (346,13-347,16).
What Avicenna has done in the lliihiyyiit is to add the essence/existence
distinction to the matrix of distinctions he structured in the Mabda' wa-maciid.
As I just mentioned, the essence/existence distinction is not crucial to Iliihiyyiit
1.6' s articulation of the matrix of distinctions, since Avicenna could have left it
258 Part II: The Avicennian Synthesis

out of that passage without creating any fatal weaknesses in his arguments. So
why should it appear in the first place? To my mind Avicenna brought up the
essence/existence distinction in order to side-step the objections of critics who
could have argued either that the link between being possible of existence and
being caused was not as intuitive as Avicenna seemed to be taking for granted; or
that the link between being necessary of existence through another and being
caused was so intuitive as to smack of circularity. Were he ever to be attacked in
this way, Avicenna could now appeal to the causedness of everything composed
of essence and existence. And with space not an issue in the Shifa>, Avicenna
had the luxury of including arguments which he may have felt compelled to
excise in the more compressed MabdaJ wa-macad and Najat.

Most compressed of all was Avicenna's late work, the Isharat wa-tanbihat. Here
is what he says there about being necessary of existence:

L78
lbn Sina, Kitiib al-ishiiriit wa-t-tanblhiit, 140, 12-151,2
Warning: When you examine every existent in and of itself and
apart from any other, it is such that the existence it possesses
follows necessarily, or does not [follow necessarily). If it [i.e.,
the existence it possesses] does follow necessarily, then it [the
existent] is what is real in itself [al-~qqu bi-dhiitihi] and that
whose existence is necessary in itself, namely, the Self-subsisting
[al-qayyum]. If it [existence) does not follow necessarily, it
cannot be said to be impossible in itself after having been
assumed to be an existent. On the contrary, if a condition is added
[in qurina ... shar!un] to taking it in itself, such as the condition
that its cause be non-existent, then it would become impossible;
or [if] the condition that its cause exist [is added], then it would
become necessary. Now if no condition is attached - neither the
existence or non-existence of its cause - it will, in itself, be left
with the third alternative, namely, possibility, because taken in
and of itself, it will be the thing which is neither necessary nor
impossible. Every existent is thus either necessary of existence in
itself, or possible of existence in itself.

In the Isharat Avicenna's approach has changed. He introduces the distinction

between necessary and possible existence following discussions of two other
distinctions. The first distinction he discusses is that between essence and
existence (138,2-139,13). In this sense the Isharat follows the pattern of the
Sh ifa J, in which the discussion of necessary and possible existence (llahiyyat
1.6) follows immediately upon a discussion of essence and existence (llahiyyat
1.5).
The /sharat departs from the pattern of the /lahiyyat, however, in that a
discussion of the different types of causes (139,14-140,9) is interposed between
the Isharat' s discussion of essence and existence and its discussion of necessary
 14. Necessity and Possibility (D) 259

and possible existence. In that discussion of causation Avicenna distinguishes

between causes of essence (the formal and material) and causes of existence (the
efficient and final). Avicenna also makes the claim, examined in Chapters 8, 9
and 10, that the two causes of existence are complementary, in the sense that the
final is prior in terms of its essence and the efficient is prior in terms of its
existence. In other words Avicenna inserts, in summary fashion, the results of
his discussion of the four causes in Iliihiyyiit 6.l and 6.5. The only difference is
that in Iliihiyylit 6.1, Avicenna had distinguished between causes that are
intrinsic to the effect (formal and material) and causes that are extrinsic to the
effect (efficient and final), whereas in the lshiiriit Avicenna distinguishes between
causes of essence (formal and material) and causes of existence (efficient and
final).6
The reason why Avicenna structures his discussions in the Jshiiriit this way is
that he wishes to press his distinction between necessary and possible existence
into the immediate service of proving God's existence, a task he was willing to
separate into two different chapters - 1.6 and 8.4 - in the lliihiyyiit. To do this
he must introduce the distinction between essence and existence in the context of
causation. The underlying message of the Jshiirlit passage is that the distinction
between essence and existence is crucial to his argument for the existence of
God, since everything other than God requires something to combine its essence
with its existence.
The argument in lliihiyyiit 8.4 for the uncausedness of the necessary of
existence in itself, which rests on the fact that the necessary of existence in itself
is not a composite of essence and existence and thus requires no composer, is
merely an adjunct to the ontological argument for the uncausedness of the
necessary of existence in itself which appears in lliihiyyiit 1.6 and which derives
from the Mabda' wa-maciid. In the lshiirlit, by contrast, the distinction
between essence and existence and the discussion of the necessary of existence in
itself are joined together by a discussion of causation.
The essence/existence distinction thus appears to be more integral to the
lshiiriit' s discussion of the necessary of existence in itself than it had been to the
parallel discussion of the necessary of existence in itself in Iliihiyyiit 1.6.
Whether or not this is because Avicenna felt that pressing the essence/existence
distinction into service in this way really made his discussion of the necessary of
existence in itself more coherent, is very difficult to say, because the lshiiriit
does not contain philosophical arguments per se but rather reminders - almost
mnemonic devices, like bulleted lecture notes - of those arguments.
The most that can be said about this series of discussions in the lshiiriit,
perhaps, is that we find there, in extremely condensed form, threads from all the
strands of thought expressed in previous works: the /fikma cAril{iiyya's linking
of causedness to being composed of two modes or states of existence; the

6 On this development see again Wisnovsky 2003a.

260 Part /I: The A vicennian Synthesis

Mabda' wa-maciid's appeal to our a priori intuitions about the causedness of

the possible of existence; lliihiyyiit l.5's claims that essence and existence are
intensionally distinct but extensionally identical; and lliihiyyiit 6.5's and the
Najiit's distinction between the complementary ways final and efficient causes
cause their effects, the former in terms of essence, the latter in terms of
existence.

I mentioned at the beginning of Chapter 11 that Davidson detected two ways

Avicenna's matrix of distinctions is put to use in proofs of God's existence, one
articulated in the Najiit and the Ishiiriit, the other in the Shi/ii' (explicitly) and
the Diinishniima (implicitly). 7 In the Shifii' and Diinishniima, Davidson claims,
we find the distinction between necessary and possible existence juxtaposed with
Avicenna's argument for God's existence from the impossibility of an infinite
regress of efficient causes. In the Najiit and the Ishiiriit, by contrast, the
distinction between necessary and possible existence is juxtaposed with
Avicenna's argument for God's existence from the impossibility of a self-
sufficient totality of possible things. In the Shifii' and Diinishniima, the
distinction between necessary and possible existence is a rather redundant
supplement to what is a more purely cosmological argument for God's
existence. In the Najiit and Ishiiriit, however, the distinction between necessary
and possible existence is crucial to what amounts to a combined ontological-
cosmological argument for God's existence.
I also pointed out at the beginning of Chapter 11 that Davidson's analysis,
though philosophically acute, was based only on works from Avicenna's middle
and late periods. My claim was that ignoring the works from the early period,
such as the lfikma cAru{iiyya and the Mabda' wa-ma ciid, weakened Davidson's
analysis in two ways. The first is that it is in Avicenna's earliest works that we
can expect to find the rawest evidence of the sources of his thinking on this
complex of problems. Any conclusions about sources and influences drawn only
from Avicenna's middle and late works, will therefore end up being tentative.
The second weakness is that the earliest works are the only evidence we have of
Avicenna's starting-point, that is, th~origin from which his theories evolved. ·
What do these early works tell us about the sources of Avicenna's matrix of
distinctions and of the different ways he applied that matrix in his works? And
how might that information allow us to revise Davidson's hypothesis? In
Chapters 12-14 it has emerged that the background to Avicenna's matrix of
distinctions was very complex. The Arabic version of De Interpretatione 12-13,
mediated through al-Fiirabi's commentary, allowed Avicenna to think that the
primary modal opposition was between the contradictory terms mumkin
(possible) and wiijib (necessary), rather than between mumkin (possible) and
mumtanic (impossible). This move, combined with Isl,laq b. I:Iunayn's and al-

'Davidson 1987, 281-310 and 350-361.

 14. Necessity and Possibility (D) 261

Fiirabi' s use of the existential root w-j-d in contexts where Greek authors had
understood einai only in its copulative sense, encouraged Avicenna to hold that
not only "possible" but "contingent" was included in mumkin's semantic field;
and thus to think that "caused" might be an intuitively apprehended characteristic
of something which is mumkin al-wujud.
In his Commentary on the De Interpretatione al-Farabi also came up with a
distinction between mumkin al-wujud bi-nafsihi and wajib al-wu}Ud bi-nafsihi.
But al-Farabi introduced his distinction in the context of the problem of future
contingents, and of the effect that God's foreknowledge of those future events
might have on the question of determinism. This is not the context in which
Avicenna devises and employs his matrix of distinctions. Although al-Farabi and
the Arabic De Interpretatione may have supplied the terminology, the context
was supplied by Metaphysics 5.5 and by the mutakallimun.
The Arabic version of Metaphysics 5.5, mediated by Alexander and
Ammonius, provided Avicenna with the idea that the necessity which God
possesses consists in invariability: to be more specific, the qualities of
simplicity, eternality and immutability. From the Neoplatonists Avicenna
inherited a tradition which elevated simplicity above the other characteristics of
the divine. In Neoplatonic terms God's simplicity was often expressed as His
perfection, now understood as referring to His being whole and non-composite,
and hence causally self-sufficient.
Divine simplicity was not only a tenet of Neoplatonists. The mutakallimun
had vigorous debates about the extent to which divine simplicity was infringed
upon by holding that God's attributes of knowledge and power, for example,
enjoyed some kind of separate reality instead of being distinguishable from the
divine self only conceptually. From the Ash<arites and Maturidites Avicenna
inherited a tradition of discussions about how to distinguish between two or
more eternal things. God, they claimed, was eternal in Himself (qadim bi-
nafsihi). And though they could not quite bring themselves to say that the
attributes were eternal through another (qadima bi-ghayrihii.), al-Maturidi did
hold that Heaven was eternal a parte post through another (biiqiya bi-ghayrihii).
I believe that what this background tells us is that the distinction which
Avicenna was searching for above all others was one which could differentiate
between an eternal thing which is causally self-sufficient and an eternal thing
which is not causally self-sufficient. This is precisely the distinction which the
Neoplatonists and the mutakallimun were searching for. The only difference in
their searches was that Avicenna had in mind the heavens when he thought of
things which were eternal but not causally self-sufficient; the Neoplatonists had
in mind the full complement of gods, minds, souls and demiurges; and the
mutakallimun had in mind the divine attributes. The Neoplatonists, committed
as they were to integrating Plato's and Aristotle's cosmologies, were ultimately
incapable of overcoming the duality that resulted from seeing God as both final
and efficient cause, as causally self-sufficient but also causally productive. The
262 Part II: The A vicennian Synthesis

Sunni mutakallimun held back from making a clean jump to holding that the
attributes were eternal through another (qadima bi-ghayrihii) because they
worried that it would imply too much otherness in the relationship between
God's self and His attributes. It would also complicate their attempts to use
qadlm in the sense of uncaused in their proofs of God's existence.
The conclusion we can draw from Avicenna's early works about the sources of
his matrix of distinctions is that the Neoplatonists and the mutakallimun defined
the problem for Avicenna, while the Arabic Aristotle, mediated through al-
Farabi, provided the terminology and some of the conceptual tools he used to
solve it. Finally, by applying the phrase "necessary of existence in itself' to
God, al-c Amiri provided the young Avicenna with a crucial element of his
matrix. This conclusion about sources does not, in itself, invalidate Davidson's
hypothesis of a two-track approach in Avicenna's middle and late works towards
the use of the necessary of existence/possible of existence distinction in proofs
of God's existence.
What it does show is that the problem of the eternity of the world is less
overarching than Davidson makes it out to be, and consequently that the
opposing positions of the mutakallimun and the philosophers on this particular
issue should not be seen as signifying some basic or "natural'' opposition in
their philosophies. Avicenna's position on the distinction between essence and
existence was shown in Chapter 7 to be closer in important ways to that of the
Ash<arites and Maturidites than it was to al-Farabi's position, which in turn was
closer in important ways to that of the Mu<tazilites than it was to Avicenna's.
In the same vein I have shown that while Avicenna employs the terminology
and a number of important conceptual tools from the Arabic Aristotle, al-Fiiriibi
and the Neoplatonists to construct his matrix of distinctions, the question he
originally wanted to use that matrix to answer was posed to him by the
mutakallimun. The enthusiasm with which post-classical Sunni mutakallimun
and Shicite faliisifa - not to mention medieval Latin theologians and
philosophers - later employed Avicenna's matrix of distinctions in their proofs
of God's existence and in their discussions of epistemology and the divine
attributes, is a clear sign that Avicenna's matrix transcended the particular debate
over the createdness or eternality of the world.

Now to the question of development. Can the series of different ways Avicenna
articulates and employs his matrix of distinctions be described as an evolution in
his views? And how might our conclusions about that allow us to revise
Davidson's hypothesis? I believe that the different ways Avicenna structured his
matrix of distinctions do constitute evidence of evolution. But I must say
precisely what I mean by "evolution". I use the term carefully, to refer to the fact
that the different ways Avicenna structured his matrix of distinctions were both
responses to environmental stimuli and the results of a self-propelled algorithm.
 14. Necessity and Possibility (D) 263
External factors which helped to shape the ways Avicenna structured his
matrix of distinctions and employed it in different texts include, for example, the
length of the text and the intellectual level of the intended audience. In a
relatively brief compendium, such as the Mabda' wa-ma'iid or the Najiit,
Avicenna allowed himself less space to explore interesting implications of his
arguments than he did in the much lengthier Shifii'. In the Jshiirii.t, a text
intended for very advanced students of philosophy, Avicenna's argumentation is
so compressed that it can be interpreted in very different ways; this is much less
the case in the Mabda' wa-ma'iid, written for a civilian patron.
By "self-propelled algorithm" I simply refer to an internal continuum in all
Avicenna's texts: the degree of philosophical coherence Avicenna expected of
himself. Even this continuum changed in subtle ways. For example, in some
texts, such as the Mabda' wa-ma'iid, Avicenna felt that the link between being
possible of existence and being caused was so intuitive that he did not need to
appeal to an argument for compositeness to forge that link. In other texts, such
as the l:fikma 'Aru{i,iyya and the Jliihiyyii.t of the Shifii', he felt the need to
buttress that intuition with an argument for compositeness. In the l:fikma
'Arut;iiyya, that argument for compositeness was based on the premise that what
is necessary of existence through another and possible of existence in itself is a
composite of those two existential modes, or of the states actuality/potentiality
or perfection/imperfection. In the Jliihiyyiit of the Shifii.', the argument for
compositeness was based on the premise that what is necessary of existence
through another and possible of existence in itself is a composite of essence and
existence.
In short, as long as we resist thinking about Avicenna's intellectual
development in teleological terms, as long as we are clear that evolution does
not necessarily mean resolution, I think it can safely be said that the way
Avicenna structured his matrix of distinctions and employed it in different texts
evolved over his lifetime. Once again, this conclusion about evolution does not,
in itself, invalidate Davidson's hypothesis of a two-track approach in Avicenna's
middle and late works towards the use of the necessary of existence/possible of
existence distinction in proofs of God's existence.
What it does show is that the seeds of what Davidson sees as a two-track
approach can be found in two of Avicenna's earliest works, the l:fikma
'Arut;iiyya and the Mabda' wa-ma'iid. In addition, taking these two works into
account, and showing how Avicenna's thinking on this issue evolves over his
lifetime, weakens Davidson's hypothesis in the sense that the external and
internal factors which combined to produce a certain structure of the matrix of
distinctions in a particular text at a particular time now appear to be very
complex. Given this complexity, the discussions in which Avicenna's matrix of
distinctions appears will resist being divided according to a single, all-governing
criterion: usefulness or superfluousness in proofs of God's existence.
 Conclusion

What have I accomplished in this book? In an effort to analyze the sources and
evolution of Avicenna's metaphysics, I focused on the answers Avicenna and his
predecessors gave to two fundamental pairs of questions: what is the soul and
how does it cause the body; and what is God and how does He cause the world?
To respond to these challenges Avicenna invented new concepts and distinctions
and reinterpreted old ones.
In Part I I traced the complicated history of one such concept, "perfection", and
of the distinctions associated with perfection. I began by discussing how
Aristotle's varied uses of the Greek term entelekheia ("actuality") and of a
cluster of other terms related to entelekheia - such as energeia ("activity"), to
teleion ("the complete"), teleiosis ("completion") and teleiotes
("completeness") - were interpreted by late-antique commentators, both
Peripatetic and Neoplatonist. I showed that the commentators juggled all these
terms around in an effort to reconcile what appeared on the surface to be
incompatible views. Peripatetic commentators focused their attention on
reconciling Aristotle's seemingly inconsistent uses of entelekheia to define the
soul and to define change. Neoplatonic commentators focused their attention on
reconciling two larger inconsistencies, which appeared to divide Plato (at least as
understood by Neoplatonists such as Plotinus) and Aristotle (at least as
understood by Peripatetics such as Alexander). The first was the inconsistency
between the Aristotelian doctrine that God operated as a final cause, and the
Platonic doctrine that God operated as an efficient cause. The second was the
inconsistency between the Aristotelian doctrine that the soul is inseparable from
the body, and the Platonic doctrine that the soul is separable from the body.
Finally I described how the commentators' reconciliation work shaped the way
these terms were translated into Arabic in the ninth and tenth centuries, and I
explained the ways in which Avicenna, in his own discussions of the soul,
appropriated the conceptual and terminological legacy he inherited from earlier
thinkers and translators.
In Part II I argued that Avicenna's famous distinction between essence
(mahiyya) and existence (wujud) owes a great deal to earlier debates within
Islamic doctrinal theology, or kalam, concerning the relationship between
"thing" (shay') and "existent" (mawjud). I also described how Avicenna
integrated his distinction between essence and existence into the cosmology of
procession and reversion bequeathed to him by the Neoplatonic commentators;
and how this integration led to an unacceptable dualism in God, now seen to be a
composite of efficient and final causality, of involvement with and transcendence
of the world. Finally, I analyzed the background to and development of the
matrix of distinctions Avicenna devised to skirt this dualism, a matrix
266 Conclusion

constructed around the idea that God is the only being which, by virtue of itself,
exists necessarily (wiijib al-wujud bi-dhiitihi). Once again, Avicenna is shown
to have drawn much inspiration from preceding kaliim debates, this time about
how divine attributes (~ifiit) such as knowledge {'ilm) and power (qudra) relate
to the divine Self (dhiit), and about the status of what I have called "meta-
attributes" such as etemality (qidam), which are attributable not only to the
divine Self but to the divine attributes as well.
My general conclusion is that Avicenna's innovations are a turning point in
the history of metaphysics. Avicenna is the culmination of one period of
synthesis - I refer to it as the "Ammonian" synthesis, after the Neoplatonic
commentator Ammonius (d. ca. 521 AD), son of Hermeias - during which
philosophers succeeded in incorporating the larger, Neoplatonic project of
harmonizing Plato with Aristotle, into the smaller, Peripatetic project of
harmonizing Aristotle with himself. Yet Avicenna also stands at the beginning
of another period of synthesis, during which philosophers sought to fuse
together the Arabic version of the Ammonian synthesis with the ontology and
theology of the Muslim mutakallimun. Avicenna's metaphysical theories,
translated into Latin, greatly influenced European scholastic thought, but they
had an even more profound impact on Islamic intellectual history: the
philosophical problems and opportunities engendered by the A vicennian
synthesis continued to be debated in the Islamic world up to the end of the
nineteenth century.

It is up to the reader to judge whether I have proven or merely suggested the

particular hypotheses set out in the Introduction and argued for in the different
chapters, as well as the general hypothesis that Avicenna serves as the
culmination of the Ammonian synthesis and as the initiator of the Avicennian
synthesis. Assuming that I have proven them, what are the lessons that different
readers can take from my book?
Those scholars of late antiquity and of medieval Europe who ponder about
when the late-antique era ended and the medieval began, can infer from my book
that at least as far as the history of metaphysics is concerned, the decisive
moment occurred around 1001, in the Samanid library in the city of Bukhara in
the Central Asian province of Transoxania, far outside their traditional area of
focus. For historians of philosophy, the lesson is that Avicenna's thought
cannot be properly understood when it is ripped out of the context in which it
emerged, and thrust, lifeless, into comparison with earlier thinkers such as
Aristotle and later thinkers such as Aquinas. Avicenna read Aristotle very
carefully, to be sure; but he read an Arabic version of Aristotle, and viewed
Aristotle through the prism of the late-antique commentators. What is more,
some of Avicenna's most important ideas were inspired by preceding kaliim
debates. Similarly, Aquinas read Avicenna very carefully; but he read a Latinized
version of Avicenna, and viewed Avicenna through the prism of Averroes,
 Conclusion 267
among others. And again, some of Aquinas' most important ideas were inspired
by preceding debates in Christian theology. My point is that comparative work
which decontextualizes the philosophers under examination runs the risk of
comparing what in the end are mere metonyms for a cluster of philosophical
concepts or arguments selected precisely for their usefulness in serving the
comparativist's aim. Such metonymy rarely evokes the rhetoric, complexity,
evolution and ultimately underdetermined nature of the philosophy of thinkers as
creative and prolific as Aristotle, Avicenna and Aquinas.
Some Arabists may feel that the paucity of reliably edited texts of Arabic
philosophy and theology makes a project such as mine so tentative and
premature that we should postpone our analyses until some future date when all
the relevant works have been subject to critical edition. I can only say in
response that I suffer from no delusion about the definitiveness of my treatment
of Avicenna's metaphysics; far from it. What I do hope is that readers will find
my discussion and analysis thought-provoking, and mostly convincing, and that
proving my hypotheses wrong will itself spur the creation of newer and better
theories. As Sabra says, regarding the study of Arabic science:

As for the argument that "we do not yet know enough to ask the
big questions'', my answer is this: it is only by attempting to
formulate appropriate questions that can be fruitfully examined in
light of what we now know that we make it possible for others to
come up with deeper and more probing questions in the future. We
do not know much (that is for certain), but the day when we know
"enough" will never come. On the other hand, by altogether
abandoning all programs of full-fledged historical research, we
only tempt others to fill the vacuum with easy and useless
essentialist generalizations. 1

Ultimately, I shall judge my book a success if readers come away from it with
the sense of a moment in the thought-world of a philosopher from a thousand
years ago, a moment when the tools which had been crafted to solve old
metaphysical problems and to reconcile old metaphysical inconsistencies,
themselves became so interesting and powerful that problems and inconsistencies
associated with those tools now became the stuff of a new metaphysics: the
metaphysics of Latin scholasticism in Europe, and the metaphysics of post-
classical kaliim in the Islamic world.

1 Sabra 1996, 664.

 Appendix I
Tables of Greco-Arabic Translation

Tl telos in the Arabic Metaphysics

Metaph. = Aristatalis (ap. Averroem}, {Ta/sir} Ma ba<da !-faWa /-Ill [MBT /-Ill]
Metaph. 2.1, 993b21 ghaya: MBT / 11,3(bis)
Metaph. 2.2, 994b5 nihaya: MBT I 24,8
Metaph. 2.2, 994b9 eti de to hou heneka telos = fa-inna sh-shay'a lladhl bi-sababihi
takunu l-ashya 0 u huwa ghayatun: MBT I 30, 10
Metaph. 2.2, 994bl6 to gar telos peras estin = wa-dhalika anna n-nihayata hiya l-
ghiiyatu l-maqfudatu ilayhii: MBT / 33,12-13
Metaph. 3.1, 995bl ghaya: MBT / 166,7
Metaph. 3.2, 996a24 ghaya wa-tamiim: MBT / 183,14
Metaph. 3.2, 996a26 al-ghaya: MBT / 183,15
Metaph. 3.2, 996a26 tamam: MBT / 183,15
Metaph. 3.2, 996b12 at-tamam: MBT / 185,4
Metaph. 3.4, 999b10 ghaya wa-tamam: MBT / 236,4
Metaph. 5.2, 1013a33 hOs to telos = ka-t-tamami: MBT II 482,3
Metaph. 5.2, 1013a36 at-tamam: MBT II 482,6
Metaph. 5.2, 1013b2 at-tamam: MBT II 482,8
Metaph. 5.2, 1013bll at-tamiim: MBT l/ 483,1
Metaph. 5.2, 1013b25 hOs to telos = ka-t-tamami: MBT l/ 488,1
Metaph. 5.2, 1013b26 tamiiman: MBT II 488,3
Metaph. 5.4, 1015al 1 to telos tes geneseos = nihiiyatu t-takawwuni: MBT 11501,9
Metaph. 5.6, 1016a20 at-tamam: MBT II 532,6
Metaph. 5.16, 1021b23 at-tamiim: MBT 11622,3
Metaph. 5.16, 1021b25 bi-tamamiha: MBT 11622,3
Metaph. 5.16, 102lb25 at-tamiim: MBT 11622,3
Metaph. 5.16, 1021b29 at-tamiim: MBT l/ 622,7
Metaph. 5.16, 102lb29 tamam: MBT 11622,6
Metaph. 5.17, 1022a6 tamiim: MBT 11628,5
Metaph. 5.24, 1023a34(bis) telos men gar estin he morphe, teleion de to ekhon telos
= fa-inna f-furata tamiimun wa-t-tiimmu huwa lladhi lahu tamiimun: MBT 11655,9-
10
Metaph. 8.1, 1042a4 bi-t-tamami: MBT l/ 1023,3
Metaph. 8.4, 1044bl at-tamam: MBT 111074,2
Metaph. 9.8, 1050a8 tamam: MBT II 1186,6
Metaph. 9.8, 1050a8 at-tamiim: MBT 111186,7
Metaph. 9.8, 1050al 7 tamiimuhii: MBT II 1190,9
Metaph. 9.8, 1050al8 bi-t-tamiimi: MBT 111190,10
Metaph. 9.8, 1050a21 tamiim:MBTll1191,1
Metaph. 9.8, 1050a27 tamam: MBT II 1191,6
Metaph. 9.8, 1050a28 bi-tamiim: MBT 111191,6
Metaph. 9.8, 105la16 at-tamam: MBT 111210,10
Metaph. 10.4, 1055a12 at-tamiim: MBT //l 1301,5
Metaph. 10.4, 1055al3 at-tamiim: MBT //l 1301,6
Metaph. 10.4, 1055a!4 tamiiman: MBT //l 1301,6
270 Appendix I
Metaph. 10.4, 1055a15 at-tamam: MBT /ll 1301,7
Metaph. 12.8, 1074a23 tamam: MBT /ll 1678,4
Metaph. 12.8, 1074a30 tamam: MBT /ll 1678,10
Metaph. 12.10, 1075bl Ms telos = ka-t-tamami: MBT /ll 1720, 6

T2 entelekheia in the Arabic Metaphysics

Metaph. =Aristiitalis {ap. Averroem}, [Ta/sir} Ma bacda !-!abrca I-Ill [MBT I-Ill] 1

Metaph. 4.4, 1007b28 entelekheiiii = bi-l-/Fli: MBT I 381,12

Metaph. 4.5, 1009a36 entelekheiiii = bi-l-ficli: MBT I 409,17
Metaph. 5.4, 1015al9 entelekheiiii = bi-l-ficli: MBT II 507,15
Metaph. 5.7, 1017bl entelekheiili = bi-l-ficli: MBT II 555,18
Metaph. 5.11, 1019a7 kat' entelekheian = bi-l-ficli: MBT II 570, 1
Metaph. 5.11, 1019a8 kata entelekheian = bi-l-ficli: MBT II 570,3
Metaph. 5.11, 1019a10(bis) kat' entelekheian = bi-1-ficli: MBT II 570,3 and 4
Metaph. 7.9, 1034bl7 entelekheiiii = bi-l-ficli: MBT II 887,7
Metaph. 7.13, 1039a4 entelekheiiii = bi-l-ficli: MBT II 970,8
=
Metaph. 7.13, 1039a5(bis) entelekheiai bi-l-ficli: MBT II 970,8 and 9
Metaph. 7.13, 1039a7 entelekheia = al-ficl: MBT II 970,11
Metaph. 7.13, 1039al4 entelekheiiii = bi-l-ficli: MBT II 970,12
Metaph. 7.13, 1039a17 entelekheiiii = bi-l-ficli: MBT II 971,2
Metaph. 7.16, 1040b12 entelekheiiii = bi-1-ficli: MBT II 997,l
Metaph. 8.3, 1044a9 entelekheia =ji<[: MBT II 1064,10
Metaph. 9.1, 1045b34 kata ... entelekheian = bi-t-tamami wa-/-ji<li: MBT II 1103,10
=
Metaph. 9.1, 1045b35 peri ... entelekheias fi t-tamami: MBT II 1103,10
Metaph. 9.3, 1047a30 he pros ten entelekheian suntithemene =
alladhi tarakkaba
ma<a t-tamami: MBT II 1133,6
Metaph. 9.3, 1047b2 entelekheiiii = bi-t-tamami: MBT II 1133,11
Metaph. 9.7, 1049a5 entelekheiiii = bi-t-tamami: MBT II 1165,8
Metaph. 9.8, 1050a2 l-23 to gar ergon telos, he de energeia to ergon, dio kai
tounuma energeia legetai kata to ergon kai sunteinei pros ten entelekheian =fa-
inna [-<amala tamamun wa-l-jFla l-camalu wa-li-dhalika ismu l-ficli ay<,f.an yuqalu
<ata l- <amali bi-l-yunaniyyati wa-yasluku mas/aka t-tamami: MBT II 1191, 1-2
Metaph. 12.5, 1071a36 eti to proton entelekheiiii =
wa-ay<,f.an awwalu ft l-kamali:
MBT Ill 1549,7
Metaph. 12.8, 1074a35-36 to de ti en einai ouk ekhei hulen to proton; gar
entelekheia =wa-amma ma huwa bi-l-anniyyati l-Ula fa-laysa lahu <un$urun li-
annahu tamamun: MBT III 1683,8-9

1 Metaph. 11 (K), which recapitulates much of Book 9, does not appear to have been translated

into Arabic, so the renderings of entelekheia at 1065bl5.21.22.25.27.33.35, 1066a27.29, and

1066bl8 cannot be checked.
 Appendix I 271

T3 entelekheia, teleiotes and teleiosis in the Arabic De Anima

DA = Aristii~alis, Fi n-nafs [FN]
+ Themistius, in De Anima [TG] = Thiimastiyiis, Shar}J. Kitiib an-nafs li-
Aris!il!iilis [TA]
+ lbn Sina, at-Ta<lfqiit <atii }J.awiishf Kitiib an-nafs li-Aris{ii [MN]

DA 1.1, 402a26 entelekheia = bi-t-ji<li: FN 4,10

+MN: bi-l-istikmiili at 75,22 and bi-l-ficli at 76,2
DA 2.1, 412al0 entelekheia = anfaliikhiyii ya<nr at-tamiim: FN 29,11-12
+ TG (ad 412a6-l0), 39,9-10: to de eidos he teleiotes kai hOsper anaplerosis
tes euphuias kai tou dunamei pros to telos proagoge = wa-amma f-fiirara
fa-hiya l-kamiilu wa-ka-annahii tatimmatu t-tahayyu'i wa-[nqiyiidu} ma bi-
l-quwwati ilii ghiiyatihi: TA 43,5-6
+ TG (ad 412al0-l l), 39,17: tauten oun ten morphen kai to eidos ei tis
entelekheian onomawi = wa-hiidhihi l-khilqatu <wa-> ~-~uratu in
sammiihii musamm<ii> anfaliikhiyii ay istikmiilan: TA 43, l 2- l 3
=
+ TG (ad 412al0-ll), 39,18: teleiosis kamiil: TA 43,15
+ TG (ad 412al0-ll), 39,19-20: semainoi an ouden allo he entelekheia e ten
hexin tes teleiotetos = fa-laysa yadullu qawluhu anfaliikhiyii ft lisiini l-
yuniiniyyin <alii shay'in ghayri malakati l-kamiili mina l-khilqati: TA
43, 15-44, l
=
+ TG 39,20 (ad 412al0-l l): teleiotes kamiil: TG 44,l
+TG(ad412al0-ll), 39,21: entelekheia =istikmiil:TA44,l
+ TG (ad 412al3-l6), 40,3: anankaion ara ten psukhen eidos einai kai
entelekheian kai houtos ousian hOs eidos = fa-qad yajibu r,fariiratan an
takuna n-nafsu furatan wa-stikmiilan wa-takUna jawharan calii fariqi f-
fiirati: TA 45,7-8
+ TG (ad 412al6-20), 40,12-13: ananke pasa <to> eidos einai ten psukhen
kai entelekheian = fa-qad yajibu <}ariiratan min kulli wajhin an takuna n-
nafsu hiya f-fiiratu wa-l-istikmiilu: TA 45, l 8-46, l
+MN : bi-l-istikmiili at 91,3.8
DA 2.1, 412a2l entelekheia = anraliishiyii: FN 29,19
DA 2.1, 412a22 entelekheia = al-anfaliishiyii: FN 29,l 9
DA 2.1, 412a27 entelekheia = anraliishiyii wa-huwa awwalu tamiimin: FN 30,3
=
+ TG (ad 412a22-28), 41,8: entelekheia istikmiil: TA 47,16
=
+ TG (ad 412a22-28), 41.12: entelekheia istikmiilan: TA 48,l
+ TG (ad 412a22-28), 41,15: tei entelekheilii = istikmiil: TA 48,4
+ TG (ad 412a22-28), 41 ,20: entelekheian = istikmiil: TA 48,10
+ TG (ad 412a22-28), 41,24: entelekheia he prote = al-istikmiil al-awwal:
TA 48,14
+ TG (ad 412a22-28), 42,l: entelekheia = istikmiil: TA 49,16
+MN al-istikmiil at 91,3.l0.12(tris)
DA 2.1, 412b5 entelekheia = anfaliishiyii: FN 30,8
+ TG (ad 412b4-9), 42,18-19: all' homos eph' hapanton to eidos amphoin
aition kai hen legomen entelekheian = lakinnahu wa-in kiina ka-dhiilika
fa-Ji jamfcihii ~-fiiratu wa-sh-shay'u lladhi nusammfhi stikmiilan huwa
sababuhumii TA 51,l-2
DA 2.1, 4 l 2b9 entelekheia = al-anfaliishiyii: FN 30, 12
272 Appendix I
DA 2.1, 412b28 entelekheia = anfaliishiyii: FN 31,10
+ TG (ad 412bl 1-17), 42,33: entelekheia prate = istikmiiluhu l-awwalu: TA
51,17
+ TG (ad 4 I 2b25-27), 43,6: tou somatos nekrou entelekheia = istikmiilu l-
jismi l-mayyiti: TA 52, 12
=
+ TG (ad 412b27-413a3), 43,15: entelekheia istikmiil: TA 53,4:
+ TG (ad 412b27-413a3), 43,17: entelekheia = al-istikmiil: TA 53,5
+ TG (ad 412b27-413a3), 43,17-18: psukhe men oun amphO hai
entelekheiai kai mallon he prate = fa-n-nafsu hiya l-istikmiiliini jamr<an
wa-ka-annahii l-awwalu wa-th-thiini [sic]: TA 53,5-6
DA 2.1, 413a7 entelekheias = anfaliishiyii: FN 31,14
DA 2.1, 413a8 entelekheia = anfaliishiyii: FN 31, I 5
+ TG (ad 413a3-8), 43,24: entelekheia kai teleiotes = istikmiilan wa-
kamiilan: 53, 11-12
=
+ TG (ad 413a8-9), 43,27: entelekheia istikmiilan: TA 53,14
+ TG (ad 413a8-9), 43,27-30: houto de kai ho nous ekhein dokei. oupo gar
delon, ei kai houtos somatos tinos entelekheia ara toiaute haste
akhOristos einai, e hOste khOrizesthai, hOsper ho kubernetes tou ploiou.
houtos gar entelekheia men, alla khOriste = fa-inna t-<aq[a qad yuiannu
annahu laysa yajri majrii sh-shakli wa-l-mithiili li-annahu lam yubayyin
bacdu annahu istikmiilun li-jismin mii fa-in kiina istikmiilan yajri majrii
mii yufo.raqu bi-manzilati r-rubbiini cinda s-safinati fa-inna hiidhii huwa
stikmiilun illii annahu mufiiraqun: TA 53,15-54,2
+MN kamiil bi-l-ficli at 93,12 and al-kamiiliit at 93,17
DA 2.2, 413bl8 entelekheiai = al-anfaliishiyii llati hiya tamiimun: FN 33,5
+ TG (ad 413bll-27), 45,20: entelekheias = istikmiilayni: TA 58,10
+ TG (ad 413bl 1-27), 45,24: entelekheias = istikmiiliit: TA 58,13
+ TG (ad 413bl 1-27), 45,33: entelekheiai = bi-l-istikmiili: TA 59,4
DA 2.2, 4 l 4al 7 to de eidos entelekheia = wa-anna $-$Urata an.taliishiyii: FN 34, l 0
DA 2.2, 414al8 ou to soma estin entelekheia psukhes, all' haute somatos tinos wa- =
anna l-jirma laysa tamiima n-nafsi wa-li-dhiilika Lan yaqa<a <a[ayhi ma<nii l-
anfaliishiyii bali n-nafsu anfaliishiyii jirmin bi-#fati kadhii wa-kadhii: FN 34, 11-12
+ TG (ad 414al4-19), 46,29: to de eidos entelekheia = wa-$-$Uratu
stikmiilun: TA 61,3
DA 2.2, 414a26 entelekheia = anfaliishiyii: FN 34, l 8
DA 2.2, 414a27 entelekheia = al-anraliishiyii: FN 34,19
=
+ TG (ad 414a27-28), 47,5: entelekheia istikmiil: TA 61,16
=
DA 2.4, 415bl5 entelekheia anfaliishiyii: FN 38,9
+ MN kamiil at 95, 1
DA 2.5, 417a9 entelekheiai = bi-l-ficli: 42,2
DA 2.5, 417a21 peri dunameos kai entelekheias = al-quwwa wa-l-ficl: FN 42, l l
+ TG (ad 417a21-b2), 55,17: entelekheias = al-istikmiil: TA 80,6
+ TG (ad 417a21-b2), 55,23: entelekheiai = bi-l-istikmiili: TA 80,18
=
DA 2.5, 417b4 hupo tou entelekheiai ontos min dhl /-ji<Li: FN 43,2
DA 2.5, 417b7 eis entelekheian = al-anraliishiyii wa-huwa l-fi clu t-tiimmu: FN 43,4
DA 2.5, 417bl0 eis entelekheian = ilii l-anfaliishiyii: FN 43,7
DA 2.5, 417bl3 hupo tou entelekheiai ontos =min ... dhi l-anfaliishiyii: FN 43,9
+ TG (ad 417b8-16), 56,12: teleiotes = kamiil: TA 82,6
DA 2.5, 418a4 entelekheiai = bi-l-ficli: FN 44,4
+ TG (ad 418a3 l-b4), 59,9: energeia tis kai teleiotes =ji<tun wa-kamiilun:
TA 89,14
 Appendix I 273
DA 2.7, 418bl2 entelekheiai = bi-l-fi'li: FN 45,19
+ TG 59,33 (ad 418b9-12): entelekheia tis kai teleiotes = istikmiilun ma wa-
kamiilun: TA 91,2
+ TG 59,34 (ad 418b9-12): teleiotes = kamiil: TA 91,4-5
+ TG 59,35 (ad 418b9-12): teleiotes = kamiil: TA 91,6
DA 2.7, 418b30 entelekheiai = bi-l-fi'li: FN 46,11
+ TG (ad 418bl2-13), 60,2: entelekheiai = bi-l-istikmiili: TA 91,9
DA 2.7, 419all entelekheia =fi'l: FN 46,20
DA 2.10, 422bl entelekheiai :: bi-l-fi'li: FN 55,18
DA 2.10, 422bl6 entelekheiai = ilii l-fi'li: FN 56,9
+ TG (ad 422a34-bl0), 71,33: entelekheiai:: bi-l-istikmiili: TA 118,2
DA 3.4, 429a28 entelekheiai = bi-l-ficli: FN 73,2
+MN bi-l-ficli: 100,23
DA 3.4, 429b31 entelekheiai = bi-I-ficli: FN 74,11
+ TG (ad 429b31-430a2), 97,22: entelekheiai =?[The Arabic text diverges
from the Greek of Themistius from TG 91,7-97,27]
DA 3.4, 430al entelekheiiii = bi-l-fi'li: FN 74,13
+ TG (ad 430al0-14), 98,13: teleiotes = at-tamiim: TA 169,5
+ TG (ad 430al4-19), 106,10: teleiotes = kamiil: TA 193,1
DA 3.7, 43la3 entelekheiiii = bi-l-an{aliishiyii: FN 76,20
+MN bi-l-fi'li: 105,22-23
DA 3.8, 431b25-26 M d'entelekheiai eis ta entelekheiai = wa-mii kiina fl /:laddi l-fi ' li
kiinat qismatuhu li-dhawl l-af'iili: FN 78,21-79, I

T4 entelekheia in the Arabic Physics

Phys. = Aris1u1alis, a{-Tabl'a [71

Phys. 1.2, l 86a3 entelekheiai = bi-l-istikmiili: T 16, 13

Phys. 2.1, 193b7 entelekheiai = bi-stikmiilihi: T 86, 1 I
Phys. 3.1, 200b26 entelekheiai = calii l-kamiili: T 168,2
Phys. 3.1, 200b27 entelekheiai = bi-l-kamiili: T 168,3
Phys. 3.1, 201al0 entelekheiiii = bi-l-kamiili: T 171,1
Phys. 3.1, 20lall entelekheia =kamiil: Tl71,2
Phys. =
3.1, 201al7 entelekheiai bi-l-ficli: T 172,14
Phys. 3.1, 201a20 entelekheiai = bi-l-kamiili: T 177,3
Phys. 3.1, 201a21 entelekheiai = bi-l-kamiili: T 177,5
Phys. 3.1, 201a28 entelekheiai = bi-l-kamiili: T 178,1
Phys. 3.1, 201 a30 entelekheia = kamiil: T 178,4
Phys. 3.1, 201a34 entelekheia = kamiilahu: T 178,8
Phys. 3.1, 201b5 entelekheia = kamiil: T 178,15
Phys. 3.1, 201b6 entelekheia =fl l-kamiili: T 179,2
Phys. =
3.2, 202a7 entelekheia kamiil: T 184,7
Phys. 3.2, 202al l entelekheiai = bi-l-kamiili: T 184, 12
Phys. 3.3, 202a14 entelekheia = kamiil: T 191,5
Phys. 3.3, 202al6 entelekheian = kamiilan, 191,7
Phys. 3.3, 202b26 entelekheia :: kamiil: T 195,4
Phys. 3.5, 204a28 entelekheiai = bi-I-kamiili: T 224, l
Phys. 3.6, 206al 5 entelekheiai = bi-l-kamiili: T 250, 12
274 Appendix I
Phys. 3.6, 206bl3 entelekheifii = fi 1-kamiili: T 253,4-5
Phys. 3.6, 206b22 entelekheifii = bi-l-kamiili: T 254, 1
Phys. 3.6, 206b25 entelekheifii = bi-l-kamiili: T 254,4
Phys. 3.6, 207a22 entelekheifii = bi-l-kamiili: T 260,5
Phys. 4.5, 213a7 entelekheia = kamiil: T 333,4
Phys. 4.5, 213a8 entelekheifii = bi-l-kamlili: T 333,5
Phys. =
8.5, 257b7 entelekheiai bi-1-istikmlili: T 856,4
Phys. =
8.5, 257b8 eis entelekheian ilii l-istikmali: T 856,5
Phys. 8.5, 257b8 entelekheia = istikmlil: T 856,6
Phys. 8.5, 258b2 entelekheifii = bi-l-istikmiili: T 862, 1
Phys. 8.8, 263a29(bis) entelekheifii = bi-l-istikmlili: T 901,5.6
Phys. 8.8, 263b5 entelekheifii = bi-l-istikmiili: T 901, 15

TS telos in the Arabic Physics

Phys. = Aris~ii~lis, af-Tablca [Tl

Phys. 2.2, 194a27 al-ghiiya: T 95, 12

Phys. 2.2, 194a28 ghaya: T 95, 13
Phys. 2.2, 194a29 ghliya: T95,14
Phys. 2.2, 194a32 ghiiya: T 96,3
Phys. 2.2, 194a35 ghliya: T 96,7
Phys. 2.3, 194b32 al-ghiiyat al-maq~uda: T 102,2
Phys. 2.3, 194b36 al-ghiiyat al-maq~uda: T 102,6-7
Phys. 2.3, 195a2 al-ghiiya: T 102,9
Phys. 2.3, 195a10 ghliya maq~uda: T 103, 12
Phys. 2.3, 195a24 al-ghiiyat al-maq~uda: T 104, 12
Phys. 2.3, l 95a25 ghiiya: T 104, 13
Phys. 2.5, 197al al-ghiiya: T 121,8
Phys. 2.7, 198b3ghliya:T140,13
Phys. 2.8, 199a8 ghliya: T 147 ,5
Phys. 2.8, 199a25 al-ghiiyat al-maq~uda: T 150,12
Phys. 2.8, 199a31 al-ghiiya: T 151,20
Phys. 2.8, I 99a32 al-ghiiya: T 151,21
Phys. 2.8, 199b6 al-ghi'iyat al-maq$uda: T 152, 11
Phys. 2.8, 199b17 ghiiya: T 155,2
Phys. 2.9, 200a14 al-ghiiyat al-maq$uda: T 160,6
Phys. 2.9, 200a20 al-ghiiyat al-maq$uda: T 161,3
Phys. 2.9, 200a22 al-ghiiya: T 161,6
Phys. 2.9, 200a27 al-ghi'iyat al-maq$uda: T 162,12
Phys. 2.9, 200a34 al-ghiiya: T 163, 14
Phys. 3.3, 202a24 ergon de kai telos = wa-yakunu /-Camalu nafsuhu wa-t-tamiimu: r
192,2
Phys. 3.4, 203b9 iikhir: T 212,9
Phys. 3.6, 207a14 teleion d'ouden me ekhon telos; to de telos peras = wa-t-tiimmu lii
yakunu illii ma lahu tamiimun wa-t-tamiimu nihiiyatun: T259,10-11
Phys. 4.1, 209a22 tamiim: T279,15
Phys. 6.5, 236a10 kata to telos tes metaboles =min qibali ntihii,i t-taghayyuri: T
674,7
 Appendix I 275
Phys. 6.5, 236al2 metaboles telos = iikhir at-taghayyur: T 674,8
Phys. 6.9, 239bl3 iikhir: '[713,10
Phys. 6.9, 239b34 apo telous tau stadiou =min iikhiri l-mldiini: '[715,3
Phys. 8.8, 262b3 l ananke gar epi telos elthein = fa-wiijibun tjarflratan an yantahiya
ila ghiiyatihi [reading ilii ghiiyatihi instead of Badawi's ilii ghayri cillatihi, because
of the Greek]: T 899,4
Phys. 8.9, 265a30 al-inqi(ia 0 : T 917, 11
Phys. 8.9, 265a34 iikhir: T 918,3
Phys. 8.9, 265b4 iikhir: '[918,8
Phys. 8.9, 265bl3 inqi¢a 0 : '[919,7
Phys. 8.9, 265bl5 inqidii": '[919,9
Phys. 8.10, 267a9 telos de pauetai = wa-yakuffu bi-akharatin: T 930, 1
 Appendix II
Transcriptions of Lemmata from MS Uppsala Or. 364

4v16-5r5 (L31)
Loi_, -~~Lo 0 L.S 0 1 <G~Lo Loi~" 'l:i "~' .1~_, I ~ Lo _y. i.i....:J-1
~J Lo ~LiJ ..:,,lS ..:,,I ~Ii. ~I_, b~J b~lo u.o 11 t~.;-o ..:,,lS ..:,,I <G~
' ,,.,,
I.)
. j&.l.£11 c.l.l I.). ~ JGllJ j&.l.£JL.
. I i..)&- -=--:I-:
~
(1':11J I uu.J.I
- J~J
~

. ·II

0 ~ .u .);.).1 >"~1 <~> JUJI <__y.> Lo~~~<~> I~ J~

ill .u ..,.:.~1_, ~Li._, o.ku ..:,,u)..~.4 ..:,,~_, I <G~_, cG~~ ..:,,l.i)..L
~WI_, JLWI ..:,,~ 4.1 ..:.i~.;W.1_, b.;_,.dl_, b.;_,.dl ~ I~ '-:'~I

3v16-4r12 (L58/L70)
I i~ cGI~ L...1 &.l_, ~ ~ Lo u-1.c. .1~_,Jl ~J.J~I ~ ~l_,JI
II ~l.1 Loi ~l_,JI_, ~)~I_, ~I~~~ Loi_, d.1~_,.ll
Lo ~J I <Cii_, ~~I ._j~ JL:.. ..:,,_,.1 JI...:..~ C..1_, ..:..l-1>7"_,.ll j~
<G ll:i I .i~_,Jl ~l.J-:1 ~ 41.u w .u 41 )Li <Gll:i .i~_,.il ~I_,~
~ I <Glj ~ ~ L1 ~ ..i..::--~ ..:,,LS u ':II_, <Gll:i .i>7->11 ~ ':I_,
~ ':I <Glj -l~J ..:,,u .u Uc. I ':I ~., ~ .i~_,JI ~'" ~J .i~_,Jl
. I . .u
1.J'-""' :w• .< ~ 0Wl~I I L..:JS: ~~~~-
. ':I ~- . _Q ';I .I... I . . . ~L::.. .
'-""'
~ ~L:.. u-1:- ~~LI~ I ~J 4µ1 ~ ~ ':I_, ':l_,h.o ~
~ ~~':I_,~ ~ b~I_, I~ <GI~ .u ~ b~I_, ':/_,
~I_,~ cGl.l.:i -l>7-_,11 '-;-'71J ~ ~ ~J I ~J ..:..l~I J,h...o ~
~1 .. ~1 u.o~ ~1 J...o~ ~ <G4-7- I~ u.o .i~)I
(•i o o: ~ U ~ J~J ~J.l~ ~ (f.:JI .i~_,.11 I~ J'. all_,
. JL:..J I ~ 1.i~J-0 ~J ~..a:J ~~ ~ ..r.:C- ..:,,lS u I ~ J~J
278 Appendix II

4r14-17 (L69)
~ ..u_, ct.:il~ ~..u ~I ~ .u_, L..~ hi ~ r1 ~ J$.! ~..u J~
.:.i.Q_, ~ ~ wl.S 'f.ll 1~ ~I_, .::..~I Lo I_, :.,~ :i2 ~ b~ I ·1.o:.l.i

I ~:U ..i....i..Q ~ J$ :.,'i b.JL..:.,... ct.1 ~ -i_, o~ -ii~ :.,.1_, I Lo

11 t~~ ~ ~ t ..o->' ~ :_,1.S...'i1 <!11.:i_, ~~-ii_,/~" :.,1.S...1
I ,;A~ _JA ~ o.r;:i. JI UL...;'i~ 4-Al_,:A ~ ,;Al~I :_,...~I_, 'i ~'i
~ .J_,.L .J~_, ~I (.] rA~ 'i_, µ 1 JI l.l"4-:AJ4 .:.,l.S..ol :_,1.S...'il_,
b~UI ~"~_,I
- - v0.;-&- J~ Js u1c- ~ j->-"~ ~I~ 11.l"l.L.o
~

dy2:hll i~ I~_, b~lo .w I ~JS .Jb JS~

 Bibliography

PRIMARY SOURCES CITED IN THE NOTES

cAbd al-Jabbiir, al-Mughnf fi abwiib at-tawJ:ild wa-l- cad/ ed. M.M. l:filmi et al., Cairo,
1958f.
(Ps.-)Abii l:fanifa, al-Fiqh al-akbar ed. M.M. ash-Shacciir (= Mullii c Ali b. Sultan
MuJ:iammad al-Qiiri, Sharl:z al-fiqh al-akbar), Beirut, 1997
Albert of Saxony, Expositio et quaestiones in Aristotelis Physicam ad Albertum de
Saxonia attributae /, ed. B. Patar, Louvain, 1999
Alexander, DA (CAG Suppl. I.I), ed. I. Bruns, Berlin, 1887
Alexander, in Metaph. (CAG 1), ed. M. Hayduck, Berlin, 1891
Alexander, in Meteor. (CAG Ill.I), ed. M. Hayduck, Berlin, 1899
Alexander, in Sens. (CAG Ill.I), ed. P. Wendland, Berlin, 1901
Alexander, in Top. (CAG Il.2), ed. M. Wallies, Berlin, 1891
Alexander, Quaest. (CAG Suppl. I.2), ed. I. Bruns, Berlin, 1892
Alexander, Maqiilat Aliskandarfi anna 1-ficl a<ammu mina 1-J:iaraka <a/ii ra'y Arisru, ed.
cA. Badawi (Arisru Cinda [-<Arab), Cairo, 1947
Alexander, Maqiilat Aliskandar fi 1-J:iiss, ed. H.-J. Ruland (Die arabische Ubersetzung
der Schrift des Alexander van Aphrodisias iiber die Sinneswahrnehmung),
Gottingen, 1978
Alexander, Risiila fi /-<aql calii ra'y Aris!il!iilis, ed. cA. Badawi (Commentaires sur
Aristote perdus en grec et autres epftres), Beirut, 1971
Alexander of Hales, In XII Aristotelis Metaphysicae Libras, Venice, 1572
Algazel, Intentiones philosophorum, ed. J. Muckle (Algazel's Metaphysics),
Toronto, 1933
al-c Amiri, Kitiib al-amad <a/ii 1-abad, ed. E. Rowson, New Haven, Conn., 1988
Ammonius, in /sag. (CAG IV.3), ed. A. Busse, Berlin, 1895
Ammonius, in Int. (CAG IV.5), ed. A. Busse, Berlin, 1897
Anonymous, Late-Antique Paraphrase of and Commentary on the De Anima, ed. R.
Arnzen (Aristoteles De anima: eine verlorene spiitantike Paraphrase in arabischer
und persischer Oberlieferung: arabischer Text nebst Kommentar,
quellengeschichtlichen Studien und Glossaren ), Leiden, 1998
Aristii!iilis {ap. lbn Rushd}, [Tafsfr} Mii bacda !-fabz<a I-Ill, ed. M. Bouyges, Beirut,
1938-48
Aristiitiilis, Fi a<fiii> al-J:iayawiin, ed. R. Kruk (Aristoteles Semitico-Latinus: The
Arabic Version of Aristotle's Parts of Animals), Amsterdam, 1979
Aristiitiilis, Fi kawn al-J:iayawiin, ed. J. Brugmann and H. Drossaart Lulofs, Leiden,
1971
Aris!iitiilis, Fi l- 'ibiira, ed. cA. Badawi (Man#q Arisru /), Cairo, 1948-1952
Aristiitiilis, Fi n-nafs, ed. cA. Badawi, Cairo, 1954
Aristii!iilis, af-Tabf<a, ed. cA. Badawi, Cairo, 1964
Aris\ii\alis, TaJ:ilfliit thiiniyya, ed. cA. Badawi (Man[iq Aris[u II), Cairo, 1948-1952
Aris\ii!iilis, Tibiic al-J:iayawiin, ed. cA. Badawi, Kuwait, 1977
Aristiitiilis, Tubfka, ed. cA. Badawi (Man!iq Aris!ii II), Cairo, 1948-1952
(Ps.-)Aristii!iilis, Uthulujiyii, ed. cA. Badawi (Plotinus apud Arabes), Cairo, 1955
(Ps.-)Aris!ii!iilis, Kitiib al-fl;iii/:t fi l-khayr al-ma/:tr;i li-Aris!il!iilis, ed. A. Badawi
(Neoplatonici apud Arabes), Cairo, 1955
Asclepius, in Metaph. (CAG VI.2), ed. M. Hayduck, Berlin, 1888
al-Ashcari, Maqiiliit al-isliimiyyin 1-11, ed. H. Ritter, Istanbul, 1929-30
280 Bibliography
al-Ashcari, Kitab al-lumac and Risalat istif:isan al-khaw(i ft cum al-kalam, ed. R.J.
McCarthy (in his The Theology of al-Ash<arl), Beirut, 1953
Averroes, Aristotelis opera cum Averois commentariis, Venice, 1562-1574
Avicenna, Liber de anima seu sextus de naturalibus I-III, ed. S. van Riet, Leiden, 1972
Avicenna, Liber de philosophia prima sive scientia divina I-IV, ed. S. van Riet,
Leiden, 1977
Avicenna, liber de philosophia prima sive scientia divina V-X, ed. S. van Riet,
Leiden, 1980
Avicenna, Liber primus naturalium: Tractatus primus de causis et principiis
naturalium, ed. S. van Riet, Leiden, 1992
Avicenna, liber quartus naturalium de actionibus et passionibus qualitatum primarum,
ed. S. van Riet, Leiden, 1989
al-Baghdiidi, Aw~ad az-zaman Abii I-Barakat Kitab al-mu<tabar ft l-f:iikma Ill, no ed.,
Hyderabad, 1938-39
al-Baghdadi, al-Kha~ib Tarlkh Baghdad 6, no ed., Beirut, 1968
Bahmanyar ibn al-Marzuban, at-Taf:i~il, ed. M. Muµhhari, Tehran, 1997
al-Baqillani, Kitab at-tamhid, ed. R.J. McCarthy, Beirut, 1957
al-Bayhaqi, Kitab al-asma' wa-~-~ifat, ed. M.Z. al-Kawthari, Beirut, 1970
Buridan, John, in Metaphysicen Aristotelis questiones, Paris 1588 (i.e., 1518), repr.
Frankfurt a.M., 1964
Damascius, Treatise on First Principles, ed. L. Westerink, Paris, 1986-1991
Duns Scotus, Quaestiones super libros Metaphysicorum Aristotelis I-IX, New York,
1997-1998
al-Farabi, Epitome of Aristotle's Physics, ed. A. Birkenmajer, "Eine wiedergefundene
Ubersetzung Gerhards von Cremona", Beitriige zur Geschichte der Philosophie und
Theologie des Mittelalters (Suppl. 3.1), 1935, 472-481
al-Farabi, Kitab al-f:iurilf, ed. M. Mahdi, Beirut, 1969
al-Farabi, Kitab al-jam< bayna ra'yay al-f:iakimayni Aflarun al-ilahi wa-Arisraralis, ed.
F. Dieterici (Alfarabl's philosophische Abhandlungen), Leiden, 1890
al-Farabi, Kitab lira' ahl al-madlnat al-fa¢ila, ed. A. Nadir, Beirut, 1968
al-Farabi, Kitab lsaghilji ay al-madkhal, ed. and trans. D. Dunlop, "Al-Farabi's
Eisagoge", Islamic Quarterly 3 (1956), 117-138.
al-Farabi, Kitab as-siyasat al-madaniyyat al-mulaqqab bi-Mabtidi' al-mawjildat, ed. F.
Najjar, Beirut, 1964
al-Farabi, Maqala ... ft aghra(i al-f:iaklm ft kulli maqalatin mina l-kitab al-mawsilm bi-
l-f:iurilf, ed. F. Dieterici (Alfarabl's philosophische Abhandlungen), Leiden, 1890
al-Farabi, Risala ft l- <aql, ed. M. Bouyges, Beirut, 1938
al-Farabi, Risala ft jawab masa'il su'ila <anhti, ed. F. Dieterici (Alfarabl' s
philosophische Abhandlungen ), Leiden, 1890
al-Farabi, Sharf:i li-kitab Aris!il!tilis ft l-%ara, ed. W. Kutsch and S. Marrow, Beirut,
1971
(Ps.-)al-Fiirabi, Risala ft [.<ifm al-ilahi, ed. cA. Badawi, Neoplatonici apud arabes,
Cairo, 1955
Giles of Rome, Errores Philosophorum, ed. J. Koch, Milwaukee, 1944
al-Ghazali, Tahtifut al-falasifa, ed. M. Bouyges, Beirut, 1927, 358,8-11
al-Ghazali, Maqa~id al-falasifa, ed. S. Dunya, Cairo, 1961
Henry of Ghent, Quodlibet XIII, ed. J. Decorte, Leuven, 1985
}:l.unayn ibn Is~aq, al-Masa'il ft !-!ibb li-l-muta <allimln, ed. M. Abii Rayyan et al.,
Cairo, 1978
Iamblichus, Theologumena Arithmeticae, ed. F. Ast, Leipzig, 1817
 Bibliography 281
lbn cAsakir, Tabyfn kadhib al-muftarf fi-mii nusiba ilii l-/miim Abt l-/jasan al-Ash <arf,
no ed., Damascus, l 928
lbn Babawayh a~-~adiiq, Abii Jacfar Mul)ammad Kitiib at-tawf.ild, Bombay (?), l 903
lbn Fiirak, Mujarrad maqiiliit al-Ash<arf, ed. D. Gimaret, Beirut, 1987
lbn Khallikan, Wafayat ai-a<yan wa-anba, abna, az-zamiin, 6 vols, ed. M. cAbd al-
}:lamid, Cairo, l 949
lbn Mattawayh, Kitiib al-majmu c fi l-muf.ii! bi-t-taklif 1, ed. J. Houben, Beirut, 1965
lbn Miskawayh, K. al-fawz al-asghar, no ed., Beirut, 1901
lbn an-Nadim, Kitab al-fihrist, ed. G. Fliigel, Leipzig, 1871
lbn Rushd, Tafsfr Ma ba<da f-fabFa I-III, ed. M. Bouyges, Beirut, 1938-48
Ibn Rushd, Tahafut at-tahafut, ed. M. Bouyges, Beirut, 1930
lbn Sina, Autobiography, ed. and trans. W. Gohlman (The Life of lbn Sina), Albany,
N.Y., 1974
Ibn Sina, Danishniima-yi cAlii'f: Ilahiyyat, ed. M. Mucin, Teheran, 1331 (Shams)
lbn Sina, al-/jikmat al-cAru<;iiyya, MS Uppsala Or. 364
lbn Sina, /liihiyyiit-i Diinishnama-yi cAll2 3 f, ed. M. Mucin, Tehran, 1952
Ibn Sina, Kitiib al-f.iudud, ed. A.-M. Goichon, Cairo, 1963
Ibn Sina, Kitiib al-hidaya, ed. M. c Abduh, Cairo, 1974
lbn Sina, Kitab al-ishariit wa-t-tanblhat, ed. J. Forget (lbn Sfnii: Le Livre des
thioremes et des avertissements), Leiden, 1892
Ibn Sina, Kitab an-najat mukhta$ar ash-shifii, li-bn Slnii, Rome, 1593
lbn Sina, Kitab an-najiit, Cairo, 1913
lbn Sina, Kitiib ash-shija,//liihiyyat (1), edd. G. Qanawati and S. Zliyid, Cairo, 1960.
Ibn Sina, Kitiib ash-shija,//lahiyyat (2), edd. M.Y. Miisa, S. Dunya and S. Zli'id,
Cairo, 1960
lbn Sina, Kitiib ash-shifa'/Manfiq (1): al-Madkhal, edd. G. Qanawati, M. al-KhuQ.ayri
and F. al-Ahwani, Cairo, 1952
lbn Sina, Kitab ash-shifii,/Manfiq (5): al-Burhan, ed. A. cAfifi, Cairo, 1956
lbn Sina, Kitiib ash-shifli'/Riyli<;iiyylit (2): al-ljislib, ed. A. Ma+har, Cairo, 1985
Ibn Sina, Kitab ash-shifli'l{abf<iyyiit (1): as-Sama< af-fablci, ed. S. Zayid, Cairo,
1983
lbn Sina, Kitab ash-shifa'l[abiciyylit (4): al-Aj<iil wa-l-inji<a[at, ed. M. Qasim, Cairo,
1965
lbn Sina, Kitiib ash-shifii'l[ablciyyiit (6): an-Nafs, ed. F. Rahman (Avicenna's De
Anima - Arabic Text), Oxford, 1959
lbn Sina, Kitab ash-shifa'l[abfciyylit (8): al-ljayawiin, ed. c A. Mustan~ir, S. Zli'id and
c A Ismacn, Cairo, 1970
lbn Sina, al-Mabda' wa-l-ma<ad, ed. c A Nurani, Tehran, 1984
lbn Sina, Mabf.iath <ani l-quwa n-nafsliniyya, ed. A. al-Ahwani (Af.iwal an-nafs),
Cairo, 1952
lbn Sina, al-Mubaf.iathat, ed. M. Bidarfar, Qom, 1992
lbn Sina, al-Qanun fi !-fibb (1), ed. I. al-Qashsh, Beirut, 1987
lbn Sina, ar-Rislilat a/-<arshiyya fi J.iaqa'iq at-tawf.iid wa-ithblit an-nubuwwa, ed. I
Hila!, Cairo, no date
lbn Sina, Risiilafi n-nafs wa-baqii'ihli wa-ma<adiha, ed. A. al-Ahwiini (Af.iwal an-nafs),
Cairo, 1952
lbn Sina, Risiila fi l-kalam calii n-nafs an-nii!iqa, ed. A. al-Ahwani (Af.iwiil an-nafs),
Cairo, 1952
Ibn Sina, Risala ilii Kiya, ed. cA. Badawi (Aris!it <inda [-<Arab), Cairo, 1947
lbn Sina, Sharf.i J.iarf al-Liim, ed. cA. Badawi (Arisfit cinda l-cArab), Cairo, 1947
Ibn Sina, Risala fi l-lugha, ed. I. Yarsha~ir (Panj Rasii'il), Tehran, 1332 (Shams)
282 Bibliography

Ibo Sina, at-Tacliqiit, ed. <A. Badawi, Cairo, 1973

Ibo Sina, at-Tacliqiit calii J:iawiishl Kitiib an-nafs li-Aris{u, ed. cA. Badawi, (Aris{u cinda
1-cArab), Cairo, 1947
Ibo Sina, cuyan al-J:iikma, ed. CA. Badawi (Avicennae Fontes sapientiae), Cairo, 1954
lkhwan a~-Safli. 0 , RasiVil, no ed., Beirut, 1957
(Ps.-)lsl.1aq b. J:lunayn, Kitiib an-nafs, ed. A. al-Ahwani (Talkhl~ Kitiib an-nafs li-Abi
l-Walid ibn Rushd wa-arba' rasii' il), Cairo, 1950
Jabir b. J:layyan, Kitiib al-J:iudad, Kitiib al-ba/:ith, Kitiib ikhriij mii ft l-quwwa ilii l-ficl,
Kitiib al-khawii~~ al-kabTr and Kitiib al-m"iziin a~-~aghir, ed. P. Kraus (Jiibir ibn
/fayyiin: Essai sur l'histoire des idees scientifique dans l'islam. Vol. 1: Textes
choisis), Paris, 1935
al-Juwayni, ash-Shiimil ft u~al ad-din, ed. cA.M. cumar, Beirut, 1999
al-Khayyat, Kitiib al-insti~iir, ed. H.S. Nyberg, Cairo, 1925
al-Kh"'ansari, al-Agha J:lusayn al-lfiishiya 'alii shurul:i al-ishiiriit (2), ed. A. al-c Abidi,
Qum, 1999
al-Khwarazmi, MafiitiJ:i al-culum, ed. G. van Vloten, Leiden, 1895
al-Kindl, Kitiib ft l-falsafat al-ulii, edd. R. Rashed and J. Jolivet (Oeuvres
philosophiques et scientifiques d'al-Kindl II: Metaphysique et cosmologie),
Leiden, 1998
al-Kindl, Risiilaft ljiicil al-J:iaqq al-awwal at-tiimm wa-l-facil an-niiqi~ alladh"i huwa bi-
l-majiiz, edd. R. Rashed and J. Jolivet (Oeuvres philosophiques et scientifiques d'al-
Kindi 11: Metaphysique et cosmologie), Leiden, 1998
al-Kindl, Risiila ilii AJ:imad b. al-Mu'ta~im ft l-ibiina 'an sujiid al-jirm al-aq~ii wa-
!ii'atihi li-lliih 'azza wa-jalla, edd. R. Rashed and J. Jolivet (Oeuvres
philosophiques et scientifiques d'al-Kindl II: Metaphysique et cosmologie),
Leiden, 1998
(Ps.-?)al-Kindi, Risiilaft J:iudad al-ashyii' wa-rusumihii, ed. M. Abii ar-Rida (Rasii'il al-
Kindi al-falsafiyya [2nd ed.]), Cairo, 1978
al-Kulayni, Abii Ja<far MuI:iammad b. Ya<qub al-U~al min al-kiift, ed. cAli Akbar al-
Ghaffli.ri, Tehran, 1961
al-Mas<iidi, Muruj adh-dhahab wa-ma'iidin al-jawhar, ed. C. Barbier de Meynard et A.
Pavet de Courteille, Paris, 1864
al-MaturidI, Kitiib at-taw/:iid, ed. F. Kholeif, Beirut, 1970
al-Maturidi, Ta'wiliit ahl as-sunna (= Tafsir al-Miiturldi al-musammii Ta'wiliit ahl as-
sunna), Cairo, 1971
Mulla Mahdi Naraqi, SharJ:i al-iliihiyyiit min kitiib ash-shifii', ed. M. Mul)aqqiq,
Tehran, 1986
Mulla Sadra, al-Asfiir al-arba'a II, ed. H.H. al-Amuli, Tehran, 1995
Mulla Sadra, Kitiib al-mashiicir, ed. H. Corbin (Le Livre des Penetrations
metaphysiques), Parisffehran, 1964
Mulla Sadra, at-Ta'liqiit 'alii iliihiyyiit ash-shifii', in Ibo Sina, ash-Shifii' II, Tehran,
1886
an-Nashi' al-Akbar, al-Kitiib al-awsarft l-maqiiliit, ed. J. van Ess, Beirut, 1971
Niqiimakhis al-Jarasini, Kitiib al-madkhal ilii 'ilm al-'adad, ed. W. Kutsch (Thiibit b.
Qurra's arabische iibersetzung der Arithmetike Eisag6ge des Nikomachos von
Gerasa), Beirut, 1959
Olympiodorus, in Ale. 1, ed. L. Westerink, Amsterdam, 1956
Olympiodorus, in Meteor. (CAG Xll.2), ed. W. Stuve, Berlin, 1900
Olympiodorus, in Phaedonem, ed. W. Norvin, Leipzig, 1913
Philoponus, in DA (CAG XV), ed. M. Hayduck, Berlin, 1897
Philoponus, in GC (CAG XIV.2), ed. H. Vitelli, Berlin, 1897
 Bibliography 283
Philoponus, in Phys. (CAG XVI-XVII), ed. H. Vitelli, Berlin, 1887-1888
(Ps.-)Plutarch, FT l-iirii0 ar-rabZCiyya allatl tarrfii bihii l-faliisifa, ed. CA Badawi
(Aris!fl!iilis FT n-nafs), Cairo, 1954
Proclus, in Ale./, ed. L. Westerink, Amsterdam, 1954
Proclus, in Parm., ed. V. Cousin, Paris, 1864
Proclus, in Remp., ed. W. Kroll, Leipzig, 1899-1901
Proclus, Inst. Theol., (2"d) ed. E. Dodds, Oxford, 1963
Proclus, in Tim., ed. E. Diehl, Leipzig, 1903-1906
Proclus, Theol. Plat., edd. H. Saffrey and L. Westerink, Paris, 1968-1997
al-Qasim ibn Ibrahim, Kitiib ad-dalil al-kabir, ed. B. Abrahamov, Leiden, 1990
al-Qushayri, Sharf.r asmii 0 Allah al-f.rusnii, ed. r.cA. sacd and S.l:f.M. CAii, Cairo, 2001 ,
ar-Raghib al-I~fahani, al-J<tiqiidat, ed. Sh. al-c Ajali, Beirut, 1988
ar-Riizi, Fakhr ad-din Sharf.r al-ishiiriit, no ed. (Sharf.ray al-ishiiriit li-l-Khwiija Na$ir ad-
din ar-Tusi wa-li-t-<Imam Fakhr ad-din ar-Razi), Qum, 1983 or 1984
ar-Razi, Fakhr ad-din Kitab lubtib al-ishiirtit, ed. cA.S. cA~iyah, Cairo, 1936
ar-Razi, Fakhr ad-din, al-Mabtif.rith al-mashriqiyya /,ed. M. al-Baghdadi, Beirut, 1990
ar-Razi, Mul:iammad ibn Zakariya> Maqtila ft-ma bacda !-!abica, ed. P. Kraus (Rasa 0 il
falsafiyya li-Abi Bakr Muf.rammad ibn Zakariyii 0 ar-Razi, Part I), Cairo, 1939
Sabzawari, Sharf.r ghurar al-fara 0 id, ed. M. Mul:iaqqiq and T. lzutsu (Sharf.r-i man?,umah),
Tehran, 1969
Sacdiya al-Fayyiimi (i.e., Saadia Gaon), Kitiib al-amaniit wa 1-ictiqiidiit, ed. S.
Landauer, Leiden, 1880
as-Samarqandi, Abu 1-Layth Sharf.r al-fiqh al-absa! li-Abi l;lanifa, ed. H. Daiber (as The
Islamic Concept of Belief in the 4th/10th Century), Tokyo, 1995
ash-Shahrastani, Kitiib al-milal wa-n-nif.ral, ed. M. Badran, Cairo, no date
ash-Shahrastani, Kitiib nihayat al-iqdiim ft cilm al-kaliim, ed. A. Guillaume, no date
ash-Shaykh al-Mufid, Awii 0 il al-maqiiliit ft L-madhiihib wa-L-mukhtiiriit, ed. M.
Mul:iaqqiq, Tehran, 1993
Sibawayhi, al-Kitiib /,ed. l.B. Yacqiib, Beirut, 1999
Siger of Brabant, Quaestiones in metaphysicam (Cambridge MS), ed. A. Maurer,
Louvain, 1983
Simplicius, in Cael. (CAG VII.I), ed. J.L. Heiberg, Berlin, 1894
Simplicius, in Cat. (CAG VIII), ed. C. Kalbfleisch, Berlin, 1907
Simplicius, in Phys. (CAG IX-X), ed. H. Diels, Berlin, 1882-1895
(Ps.-?)Simplicius, in DA (CAG XI), ed. M. Hayduck, Berlin, 1882
Suarez, Francisco, Disputationes metaphysicae, edd. and trans, S. Rabade Romeo, S.
Caballero Sanchez and A. Puigcerver Zanon (Disputaciones metafisicas IV),
Madrid, 1962
Syrianus, in Metaph. (CAG VI.I), ed. W. Kroll, Berlin, 1902
a~-Tabari, Tafsir ar-Tabari, ed. M.M. Shakir, Cairo, 1954
at-Tawl:iidi, al-l;lawiimil wa-sh-shawiimil, no ed., Cairo, 1951
Thamas~iyiis, Sharf.r f.rarf al-Liim min Kitiib Mii bacda af-Tabica li-Arisfu, ed. cA.
Badawi (Arisfu cinda /-<Arab), Cairo, 1947
Thamas~iyiis, Sharf.r Kitiib an-nafs li-Arisfii!iilis, ed. M. Lyons (An Arabic Translation
of Themistius ' Commentary on Aristoteles De Anima), Norfolk, 1973
Themistius, in DA (CAG V.3), ed. R. Heinze, Berlin, 1899
Themistius, in Phys. (CAG V.2), ed. H. Schenk!, Berlin, 1900
Thomas Aquinas, in XII Libros Metaphysicorum Aristotelis Expositio, Turin, 1950
Thomas Aquinas, in VIII Libras Physicorum Aristotelis, Turin, 1954
284 Bibliography
Thomas Aquinas, Scriptum super libros Sententiarum, ed. and trans. E.M.
Macierowski (Thomas Aquinas's Earliest Treatment of the Divine Essence),
Binghamton, NY, 1998
Thomas Aquinas, Summa contra gentiles, Rome, 1934
a~-Tiisi, Na~ir ad-din, SharJ:i al-ishlirat, no ed. (SharJ:iay al-ishlirat li-1-Kh"iija Na~lr ad-
dln af-Tusl wa-li-l-</mam Fakhr ad-din ar-Razl), Qum, 1983 or 1984
az-Zabidi, Taj al- <ariis min jawahir al-qamiis /, ed. 'A. Shirl, Beirut, 1994

SECONDARY SOURCES CITED IN THE NOTES

Abed, S. 1991. Aristotelian Logic and the Arabic Language in Alfarabi, Albany, N.Y.
Abrahamov, B. 1993. "Necessary knowledge in Islamic theology", BJMES 20, 20-32
Ackrill, J. 1963. (trans. and comm.) Aristotle's Categories and De lnterpretatione,
Oxford
Ackrill, J. 1972-73. "Aristotle's definitions of psukhe", Proceedings of the
Aristotelian Society (New series) 73, 119-133
Adamson, P. 2000. "Two early Arabic doxographies on the soul: al-Kindl and the
'Theology of Aristotle'", The Modern Schoolman 77, 105-125
Allard, M. 1965. Le probleme des attributs divins dans la doctrine d'al-Ash<ari et de
ses premiers grands disciples, Beirut
d'Alverny, M.T. 1951. "L'introduction d'Avicenne en Occident", Revue du Caire
14/141, 130-13 9
d'Alverny, M.T. 1952. "Les traductions latines d'Ibn Sina et leur diffusion au Moyen
Age", Millenaire d'Avicenne, Cairo, 59-69
Anawati, G. 1978. (trans. and comm.) La Metaphysique du Shifti', Paris
Back, A. 1992. "Avicenna's conception of the modalities'', Vivarium 3012, 217-255
Balme, D. 1987. "Teleology and necessity" in A. Gotthelf and J. Lennox, eds.,
Philosophical Issues in Aristotle's Biology, Cambridge, 275-285
Barnes, J. 1995. (ed.) The Cambridge Companion to Aristotle, Cambridge
Biard, J. 1997. "Le systeme des causes dans la philosophie naturelle de Jean Buridan",
in A. Hasnawi et al., eds, Perspectives arabes et medievales sur la tradition
scientifique et philosophique grecque, Paris, 491-504
Black, D. 1997. "Avicenna on the ontological and epistemic status of fictional
beings", Documenti e studi sulla tradizione filosofica medievale, 8, 425-453
Blair, G. 1967. "The meaning of 'energeia' and 'entelecheia' in Aristotle",
International Philosophical Quarterly 111, 101-117
Blanchette, 0. 1992. The Perfection of the Universe according to Aquinas: A
Teleological Cosmology, University Park, PA
Blumenthal, H. 1990. "Neoplatonic elements in the de Anima commentaries", in
Sorabji, ed., Aristotle Transformed, 305-324
Bolton, R. 1978. "Aristotle's definitions of the soul: De Anima 11,1-3'', Phronesis
23/3, 258-278
Boylan, M. 1981. "Mechanism and teleology in Aristotle's biology", Apeiron 1512,
96-102
Bradie, M. and F. Miller, Jr. 1984. "Teleology and natural necessity in Aristotle",
History of Philosophy Quarterly 112, 133-146
Bruni, G. 1960. "Note di polemica neoplatonica contro l'uso e ii significato del
termine entelekheia'', Giornale critico della filosofia italiana 39, 205-236
 Bibliography 285
Brunschwig, J. 1988 "La theorie stolcienne du genre supreme et I'ontologie
platonicienne", in J. Barnes and M. Mignucci, eds., Matter and Metaphysics:
Fourth Symposium Hellenisticum, Naples, 19-127 (= J. Lloyd [trans.], "The Stoic
theory of the supreme genus and Platonic ontology", in Brunschwig 1994, 92-157)
Brunschwig, J. 1990. "Sur une fa~on stolcienne de ne pas etre'', Revue de theologie et
de philosophie 122, 389-403 (= J. Lloyd [trans.], "On a Stoic way of not being'',
in Brunschwig 1994, 158-169)
Brunschwig, J. 1994. Papers in Hellenistic Philosophy, Cambridge
Burnett, C. 1997. The Coherence of the Arabic-Latin Translation Programme in
Toledo in the Twelfth Century, Berlin
Burrell, D. 1986. "Essence and existence: Avicenna and Greek philosophy", MIDEO
17, 53-66
Carame, N. 1926. (trans.) Avicennae metaphysices compendium, Rome
Caston, V. 1999. "Something and nothing: The Stoics on concepts and universals",
Oxford Studies in Ancient Philosophy 17, 145-213
Charles, D. 1981. "Teleological causation in the Physics", in L. Judson, ed.
Aristotle's Physics: A Collection of Essays, Oxford, 101-128
Charlton, W. 1980. "Aristotle's definition of soul'', Phronesis 25/2, 170-186
Chen, C. 1958. "The relation between the terms energeia and entelecheia in the
philosophy of Aristotle", Classical Quarterly (New Series) 7/1, 12-17
Cooper, J. 1987. "Hypothetical necessity and natural teleology", in A. Gotthelf and
J. Lennox, eds., Philosophical Issues in Aristotle's Biology, Cambridge, 243-274
Craemer-Ruegenberg, I. 1991. " 'Ens est quod primum cadit in intellectu' - Avicenna
und Thomas von Aquin", in U. Tworuschka, ed., Gottes ist der Orient - Gottes ist
der Okzident, Koln, 133-142
Daiber, H. 1980. Aetius Arabus: die Vorsokratiker in arabischer Oberlieferung,
Wiesbaden
Davidson, H. 1987, Proofs for Eternity, Creation and the Existence of God in
Medieval Islamic and Jewish Philosophy, Oxford
Des Chene, D. 1996. Physiologia: Natural Philosophy in Late Aristotelian and
Cartesian Thought, Ithaca, NY
De Young, G.R. 1981. (ed., trans. and comm.) The Arithmetic Book of Euclid's
Elements in the Arabic Tradition, Harvard University Ph.D. dissertation.
Dhanani, A. 1994. The Physical Theory of Kaliim, Leiden
Dooley, W. 1989. (trans.) Alexander of Aphrodisias: On Aristotle Metaphysics 1,
London
Dooley, W. 1993. (trans.) Alexander of Aphrodisias: On Aristotle Metaphysics 5 ,
London
Druart, T.-A. 2000. "The human soul's individuation and its survival after the body's
death: Avicenna on the causal relation between body and soul", Arabic Sciences and
Philosophy, 10/2, 259-273
Emilsson, E. 1991. "Plotinus and soul-body dualism", in S. Everson, ed.,
Psychology (Companions to Ancient Thought, Vol. 2), Cambridge, 148-165
Endress, G. 1966. "Die arabischen Obersetzungen von Aristoteles' Schrift De Caelo"
(Dr.Phil. Diss.) Frankfurt am Main
Endress, G. 1973. Proclus Arabus: Zwanzig Abschnitte aus der lnstitutio Theologica
in arabischer Obersetz.ung, Beirut
Endress, G. and Gutas, D. (eds) 1992-. A Greek-Arabic Lexicon, Leiden
Frank, R. 1958-59. "Some fragments of lsQaq's translation of the De Anima",
Cahiers de Byrsa 8, 231-251
286 Bibliography
Frank, R. 1974. "Notes and remarks on the raba"i' in the teaching of al-Miitiiridi", in
P. Salmon, ed., Melanges d'islamologie (Annand Abel Festschrift), Leiden, 137-
149
Frank, R. 1980. "Al-ma'diim wal-mawjiid", MIDEO 14 (1980), 185-210
Frank, R. 1999. ''The Ash<arite ontology: I. Primary entities", Arabic Sciences and
Philosophy 912, 163-231
Franzen, W. and Georgulis, K. 1972. "Entelechie", in J. Ritter, ed., Historisches
Worterbuch der Philosophie (Bd. 2), Basel, 506-507
Frede, M. 1980. "The original notion of cause", in M. Schofield et al., eds, Doubt and
Dogmatism, Oxford, 217-249
Freeland, C. 1992. "Accidental causes and real explanations'', in L. Judson, ed.,
Aristotle's Physics, Oxford, 49-72
Friedman, R. 1986. "Necessitarianism and teleology in Aristotle's biology",
Biology and Philosophy 113, 355-365
Friedman, R. 1987. "Simple necessity in Aristotle's biology", in International
Studies in Philosophy 19/1, 1-9
Gaiser, K. 1969. "Das zweifache Telos bei Aristotles" in I. During, ed.,
Naturphilosophie bei Aristoteles und Theophrast, Heidelberg, 97-113
Gatje, H. 1971. "Die vorhandenen Ubersetzungen der Schrift iiber die Seele,
Avicennas Kitlib al-in~lif und die Autoren der Ubersetzungen", in his Studien zur
Oberlieferung der aristotelische Psychologie im Islam, Heidelberg, 28-44
Gill, M.L. 1980. "Aristotle's theory of causal action in Phys. III.3", Phronesis 25,
129-147
Gilson, E. 1948. L'Etre et !'essence, Paris
Gilson, E. 1960. "Avicenne et Ia notion de cause efficiente," Atti de/ XII Congresso
lnternazionale di Filosofia, Florence, 121-130
Gilson, E. 1962. "Notes pour l'histoire de la cause efficiente," AHDLMA 29, 7-31
Gimaret, D. 1980. Theories de l'acte humain en theologie musulmane, Paris
Gimaret, D. 1988. Les noms divins en Islam, Paris
Goichon, A.-M. 1937. IA distinction de /'essence et de /'existence d'apres /bn Sina
(Avicenne), Paris
Goldschmidt, V. 1972. "Huparkhein et huphistanai dans la philosophie sto"icienne",
Revue des Etudes Grecques 85, 331-344
Gotthelf, A. 1976. "Aristotle's conception of final causality", Review of
Metaphysics 30, 226-254
Gotthelf, A. 1989. "Teleology and spontaneous generation in Aristotle: a
discussion", in Apeiron 2214, 181-193
Graeser, A. 1971. "A propos huparkhein bei den Stoikern'', Archiv fur
Begriffsgeschichte 15, 299-305
Graham, D. 1988. "Aristotle's definition of motion", Ancient Philosophy 8/2, 209-
215
Graham, D. 1989. "The etymology of Entelecheia'', American Journal of Philology
110, 73-80
Gutas, D. 1987-1988. "Avicenna's madhhab with an appendix on the question of the
date of his birth", Quaderni di studi arabi 5-6, 323-336
Gutas, D. 1988. Avicenna and the Aristotelian Tradition, Leiden
Gutas, D. 1994. "Pre-Plotinian philosophy in Arabic (other than Platonism and
Aristotelianism): a review of the sources", Aufstieg und Niedergang der romischen
Welt 11.36.7, Berlin, 4939-4973
Gutas, D. 1998. "Tardjama", Encyclopaedia of Islam (New Edition), Vol. 10, 225-231
 Bibliography 287
Gutas, D. 2001. "Intuition and thinking: The evolving structure of Avicenna's
epistemology'', in R. Wisnovsky, ed., Aspects of Avicenna, Princeton, NJ
Hadot, P. 1969. ''Zur Vorgeschichte des begriffes Existenz: huparkhein bei den
Stoikern", Archiv fur Begriffsgeschichte 13, 115-127
Hadot, P. 1980., "Sur divers sens du mot pragma dans la tradition philosophique
grecque", in P. Aubenque, ed., Concepts et categories dans la pensee antique, Paris,
309-319
Hager, F. 1964. "Die Aristotelesinterpretation des Alexander von Aphrodisias und die
Aristoteleleskritik Plotins beziiglich der Lehre von Geist", Archiv fur Geschichte
der Philosophie 46, l 74-187
Hamlyn, D. 1968. (trans. and comm.) Aristotle's De Anima: Books II.Ill, Oxford
Hankinson, R.J. 1995. "Philosophy of Science", in J. Barnes, ed., The Cambridge
Companion to Aristotle, Cambridge, 109-139
Hankinson, R.J. l 998. Cause and Explanation in Ancient Greek Thought, Oxford
Haschmi, M.Y. 1956. "Die Bewegungstheorie Avicennas", Sudhoffs Archiv fur
Geschichte der Mediz.in 40, 115-118
Hasnawi, A. 1994. "Alexandre d'Aphrodise vs Jean Philopon: notes sur quelques
traites d' Alexandre 'perdus' en grec, conserves en arabe", Arabic Sciences and
Philosophy 4, 53-109
Hasnawi, A. 1998. "Le mouvement et !es categories selon Avicenne et Averroes", in
Oriens-Occidens: Sciences, Mathematiques et Philosophie d'Anriquite a l 'Age
classique, Paris, 2, 119-224
Hasnawi, A. 2001 . "La definition du mouvement dans la Physique du Shifii'
d'Avicenne", Arabic Sciences and Philosophy 1112, 219-255
Hasse, D. 2000. Avicenna's De anima in the Latin West, London
Hasse, D. 2001. "Avicenna on abstraction", in R. Wisnovsky, ed., Aspects of
Avicenna, Princeton, NJ
Heinaman, R. 1988. "Eudaimonia and self-sufficiency in the Nicomachean Ethics",
Phronesis 33, 31-53
Hocutt, M. 1974. "Aristotle's four becauses", Philosophy 49, 385-399
Horten, M. 1907. (trans.) Die Metaphysik Avicennas, Halle
Hourani, G. 1972. "Ibn Sina on necessary and possible existence", Philosophical
Forum 411, 74-86
Hussey, E. 1983. (trans. and comm.) Aristotle's Physics: Books III and IV, Oxford
Hyman, A. 1987. "Aristotle, Algazali and Avicenna on necessity, potentiality and
possibility", in K.-L. Selig and R. Somerville, eds, Florilegium Columbianum:
Essays in Honor of Paul Oskar Kristeller, New York, 73-88
Jadaane, F. 1968. L 'influence du stoi'cisme sur la pensee musulmane, Beirut
Jalbert, G. 1960. "Le necessaire et le possible dans la philosophie d' Avicenne",
Revue de l'Universite d'Ottawa 30, 89-101
Janssens, J. 1986. "Le Diinesh-niimeh d'Ibn Sina: Un texte a revoir?", Bulletin de
philosophie mediivale 28, 163-177
Janssens, J. 1997. "Les Tacliqiit d'Ibn Sina: Essai de structuration et de datation", in
A. de Libera et al., eds, Langages et philosophies: Hommage a Jean Jolivet, Paris,
109-122
Jolivet, J. 1984. "Aux origines de l'ontologie d'Ibn Sina", in. J. Jolivet and R.
Rashed, eds, Etudes sur Avicenne, Paris, 19-28
Jolivet, J. 1991. "La repartition des causes chez Aristote et Avicenne: le sens d'un
deplacement'', in J. Jolivet, Z. Kaluza, and A. de Libera, eds, Lectionum varietates:
Homage a Paul Vignaux (1904-1987), Paris, 49-65
Keyt, D. 1978. "Intellectualism in Aristotle", Paideia 1, 138-157
288 Bibliography

Khodeiri, M. 1959-61. "Lexique arabo-Iatin de la Metaphysique du Shifii"', MIDEO 6,

309-324
Klein-Franke, F. 1994. "The non-existent is a thing'', Le Musion 104, 375-390
Kosman, L.A. 1969. "Aristotle's definition of motion", Phronesis 14, 40-62
Kostman, J. 1987. "Aristotle's definition of change", History of Philosophy
Quarterly 4/l, 3-16
Kraus, P. 1942. "Plotin chez Jes Arabes", Bulletin de l'Institut d 'Egypte 23, 263-295
Kutsch, W. 1954. "Ein arabisches Bruchstiick aus Porphyries (?) PERI PSUKHES und
die Frage des Verfassers der 'Theologie des Aristoteles"', Melanges de l 'Universiti
Saint Joseph 31/4, 265-286
Lennox, J. 1982. "Teleology, chance, and Aristotle's theory of spontaneous
generation", Journal of the History of Philosophy 20/3, 219-238
Lettinck, P. 1994. Aristotle's Physics and its Reception in the Arabic World, Leiden
Lovejoy, A. 1936. The Great Chain of Being, Cambridge, Mass.
Lewis, G. 1959. (Trans.) (Ps.-)Aris!ii!iilis, Uthillujiyii, in Plotini Opera, Tomus II:
Enneades IV-V, edd. P. Henry and H.-R. Schwyzer, Paris, 1959
Madelung, W. 197 l. "The spread of Miituridism and the Turks", Actas do IV
Congresso de Estudos Arabes e Islfimicos Coimbra-Lisboa 1968, Leiden, 109-168
Madelung, W. 1974a. "The origins of the controversy concerning the creation of the
Koran," in J. Barra!, ed., Orientalia Hispanica: sive studia F.M. Pareja octogenario
dicata, Leiden, 504-525
Madelung, W. l974b, "Ar-Raghib al-I~fahiini und die Ethik al-Ghazalis'', in R.
Gramlich, ed., Islamwissenschaftliche Abhandlungen Fritz Meier zum 60sten
Geburtstag, Wiesbaden, 152-163
Madelung, W. 1991. "Maturidiyya", Encyclopaedia of Islam (New Edition), Vol. 6,
847-848
Mahdi, M. 1990. "Al-Fiiriibi's imperfect state," JAOS 11014, 691-726
Maier, A. 1955. "Finalkausalitiit und Naturgesetz", in her Metaphysische
Hintergrunde der spiitscholastischen Naturphilosophie (Studien zur
Naturphilosophie der Spiitscholastik, Bd. 4), Rome, 273-335
Maier, A. l 958a, "Matus est actus entis in potentia .. .. ", in her Zwischen
Philosophie und Mechanik (Studien zur Naturphilosophie der Spiitscholastik, Bd.
5). 3-57
Maier, A. l958b. "Fonnafluens oder fluxus formae?", in her Zwischen Philosophie
und Mechanik (Studien zur Naturphilosophie der Spiitscholastik, Bd. 5), 61-143
Marmura, M. 1979. "Avicenna's chapter on universals in the Isagoge of his Shifii"',
in A. Welch and P. Cachia, eds, Islam: Past Influence and Present Challenge,
Edinburgh, 34-56
Marmura, M. 1981. "Avicenna on causal priority", in P. Morewedge, ed., Islamic
Philosophy and Mysticism, Delmar, N.Y., 65-83
Marmura, M. l984a. "Avicenna on primary concepts in the Metaphysics of his al-
Shifii"', in R. Savory and D. Agius, eds, Logos Islamikos: Studia islamica in
honorem Georgii Michaelis Wickens, Toronto, 219-239
Marmura, M. 1984b. 'The metaphysics of efficient causality in Avicenna", in M.
Marmura, ed., Islamic Theology and Philosophy, Albany, 172-187
Marmura, M. 1991-92. "Avicenna and the Kaliim", ZGAIW 7, 172-206
Marmura, M. 1992. "Quiddity and universality in Avicenna", in P. Morewedge, ed.,
Neoplatonism and Islamic Thought, Albany, NY, 77-87
McCullough, E.J. 1980. "St Albert on motion as Forma fluens and Fluxus formae", in
J. Weisheipl, ed., Albertus Magnus and the Sciences, Toronto, 129-153
McDermott, M.J. 1978. The Theology of al-Shaykh al-Mufid, Beirut
 Bibliography 289
Menn, S. 1994. "The ongms of Aristotle's concept of energeia: energeia and
dunamis", Ancient Philosophy 14, 73-114
Michot, J. I 997a. "Avicenne, La definition de l 'time, section I de I' Epftre des etats de
l'iime: Traduction critique et lexique", in A. de Libera et al., eds, Langages et
philosophie: hommage a Jean Jolivet, Paris, 239-256
Michot, J. 1997b. "La reponse d' Avicenne a Bahmanyar et al-Kirmani", Le Museon
11011-2, 143-221
al-Mi~biiJ:ii, M. 1980-81. "Dawr majhum al-kamiilfi ta<rif an-nafs bayna Arisru wa-bn
Sfnii" ("The role of the concept of perfection in defining the soul: Aristotle and
Avicenna"), Majallat Kulliyyat al-iidiib wa-1- <utam al-insiiniyya bi-Fiis 4-5, 121-
147. (The article also appears, with minor variations, in A <miil nadwat al-fikr al-
<arabl wa-th-thaqiifat al-yuniiniyya bi-muniisabat murar alf sana <ala mlliid Ibn Sfnii
wa-thaliitha wa- <ishrlln qarnan <alii wafat Arisru [:::: Actes du colloque pensee arabe et
culture grecque organise pour ce/ebrer le millenaire de la naissance d'Ibn Sina et le
23eme siecle apres la mon d'Aristote: Les 7-10 Mai 1980], Casablanca, 1985,
427-450, and in the author's Daliiliit wa-ishkiiliit: Diriisiit ft l-falsafat al- <arabiyyat
al-isliimiyya, Casablanca, 1988, 113-136.)
Moravcsik, J. 1975. "Aitia as generative factor in Aristotle's philosophy", Dialogue
14, 622-638
Moravcsik, J. 1992. "What makes reality intelligible? Reflections on Aristotle's
theory of aitia", in L. Judson, ed., Aristotle's Physics, Oxford, 31-47
Morewedge, P. 1972. "Philosophical analysis of Ibn Sina's essence-existence
distinction", JAOS 92/3, 425-35
Mure, G. 1975. "Cause and because in Aristotle", Philosophy 50, 356-357
Nasr, S.H. 1989. "Existence (wujud) and quiddity (miihiyyah) in Islamic philosophy",
International Philosophical Quarterly, 29/4, 409-428
Nussbaum, M. (trans. and comm.) 1978. Aristotle's de Motu Animalium, Princeton
Owens, J. 1978 "Aristotle - motion as actuality of the imperfect", Paideia: Special
Aristotle Issue, 120-132
Pasquino, P. 1978. "Le statut ontologique des incorporels", in J. Brunschwig, ed.,
Les Stoi"ciens et leur logique, Paris, 375-386
Peck, A. 1972. "Aristotle on kinesis'', in J. Anton and G. Kustas, eds, Essays in
Ancient Greek Philosophy, Albany, NY, 478-490
Pines, S. 1936. Beitriige zur islamischen Atomenlehre, Berlin
Rahman, F. 1952. (trans. and comm.) Avicenna's Psychology, Oxford
Rahman, F. 1958. "Essence and existence in Avicenna", Mediaeval and Renaissance
Studies 4, 1-16
Rahman, F. 1981. "Essence and existence in Ibn Sina: the myth and the reality",
Hamdard Islamicus 4, 3-14
Ross, D. 1924. (ed. and comm.) Aristotle: Metaphysics I-II, Oxford
Ross, D. 1936. (ed. and comm.) Aristotle's Physics, Oxford
Ross, D. 1961. (ed. and comm.) Aristotle: De Anima, Oxford
Rowson, E. 1988. (ed., trans. and comm.) A Muslim Philosopher on the Soul and its
Fate: Af-<A.miri's Kitiib al-Amad <a/ii l-abad, New Haven
Rowson, E. 1992. "The Theology of Aristotle and some other pseudo-Aristotelian
texts reconsidered", JAOS 112/3, 478-484
Rowson, E. 1995. "al-Raghib al-l~fahani" , Encyclopaedia of Islam (New Edition)
VII/, 389-390
Rudolph, U. 1992. "Ratio und Uberlieferung in der Erkenntnislehre al-Ashcans und al-
Maturidis", ZDMG 142, 72-79
Rudolph, U. 1996. Al-Miiturfdf und die sunnitische Theologie in Samarkand, Leiden
290 Bibliography

Sabra, A.I. 1987. "The appropriation and subsequent naturalization of Greek science
in medieval Islam: A preliminary statement", History of Science 25, 223-243
Sabra, A.I. 1996. "Situating Arabic science: Locality versus essence", Isis 87, 654-
670
Santas, G. 1989. "Desire and perfection in Aristotle's theory of the good", Apeiron
22, 75-99
Sauve Meyer, S. 1992. "Aristotle, teleology and reduction'', Philosophical Review
101, 791-825
Schofield, M. 1991.' "Explanatory projects in Physics 2.3 and 7", Oxford Studies in
Ancient Philosophy (Supplement), 29-40
Sebti, M. 1999. "La signification de la definition avicenniene de I' ame comme
'perfection premiere d'un corps natural organique' dans le Livre I du Traite de l'iime
du Shifii 0 " , Bulletin d'Etudes Orientates 51, 299-312
Shehadi, F. 1982. Metaphysics in Islamic Philosophy, Delmar, NY
Sorabji, R. 1980. Necessity, Cause and Blame: Perspectives on Aristotle's Theory,
London
Sorabji, R. 1990. "Infinite power impressed: the transformation of Aristotle's
physics and theology", in his Aristotle Transformed: The Ancient Commentators
and their Influence, London, 181-198
Steigerwald, D. 1997. La pensee philosophique et theologique de Shahrastiini, Quebec
Stern, S. 1962. '"The first in thought is the last in action': The history of a saying
attributed to Aristotle", Journal of Semitic Studies 712, 234-252
Todd, R. 1974. "Lexicographical notes on Alexander of Aphrodisias' philosophical
terminology", Glotta 52, 207-215
Tritton, A.S. 1940. "Theory of knowledge in early Muslim theology", in M. Shafi,
ed., Woolner Commemoration Volume, Lahore
van Ess, J. 1965-66. "Ibn Kullab und die MiJ:ma", Oriens 1965-66
van Ess, J. 1966. Die Erkenntnislehre des cAdudaddin al-lei, Wiesbaden
van Ess, J. 1982. "Ibn Kullab", Encyclopaedia of Islam (New Edition), Suppl., 391-
392
van Ess, J. 1991-1997. Theologie und Gesellschaft im 2. und 3. Jahrhundert
Hidschra, Berlin
Verbeke, G. 1971. "Les critiques de Plotin contre I' entelechisme d' Aris tote", in
Philomathes: Studies and Essays in the Humanities in Memory of Philip Merlan,
The Hague, 194-222
Verrycken, K 1990. ''The metaphysics of Ammonius son of Hermeias", in R.
Sorabji, ed., Aristotle Transformed: The Ancient Commentators and their
Influence, London, 199-231
Walzer, R. 1953. "New light on the Arabic translations of Aristotle", Oriens 6, 91-
113
Waterlow, S. 1982. Nature, Change and Agency, Oxford, 1982
Weisheipl, J. 1981. ''The specter of Motor coniunctus in medieval physics", in Studi
sul XIV secolo in memoria di Anneliese Maier, Rome, 81-104
Whittaker, J. 1975. "The historical background of Proclus' doctrine of
Authupostata", in B. Dalsgaard Larsen et al., eds, De Jamblique aProclus, Geneva,
193-237
Wieland, W. 1975. "The problem of teleology'', in J. Barnes et al., eds, Articles on
Aristotle I: Science, London
Wisnovsky, R. 2000. "Notes on Avicenna's concept of thingness (shay 0 iyya)",
Arabic Sciences and Philosophy, 10/2, 181-221
 Bibliography 291
Wisnovsky, R. 2001 (ed.) Aspects of Avicenna (Princeton Papers: Interdisciplinary
Journal of Middle Eastern Studies, 9), Princeton, NJ
Wisnovsky, R. 2003a. "Towards a history of Avicenna's distinction between
immanent and transcendent causes", in D. Reisman and A.H. al-Rahim, eds, Before
and After Avicenna, Leiden
Wisnovsky, R. 2003b. "One aspect of the Avicennian tum in Sunni kalam", Arabic
Sciences and Philosophy 13/2
Wolfson, H. 1938. "The amphibolous terms in Aristotle, Arabic philosophy and
Maimonides", Harvard Theological Review 31, 151-173
Wolfson, H. 1946. "The Kalam problem of nonexistence and Saadia's second theory
of creation", Jewish Quarterly Review 36, 371-391
Wolfson, H. 1959. "Philosophical implications of the problem of the divine
attributes in the Kalam'', Journal of the American Oriental Society 19, 73-80
Wolfson, H. 1976. The Philosophy of the Kalam, Cambridge, Mass.
Zimmermann, F. 1981. (trans. and comm.) Al-Farabf's Commentary and Short
Treatise on Aristotle's De lnterpretatione
Zimmermann, F.W. 1986. "The origins of the so-called Theology of Aristotle", in J.
Kraye et al., eds, Pseudo-Aristotle in the Middle Ages, London, 110-240
Zimmermann, F. 1994. "Proclus Arabus rides again", Arabic Sciences and Philosophy
4, 9-47
 Index of Lemmata

Ll Aris!il!alis, Fi n-nafs, 29,9-30,4 =Aristotle, DA 2.1, 412a6-28: 22

L2 Aris!ilJalis, af-Tabica, 170, 15-171 ,2 =Aristotle, Phys. 3. I, 201al0-
12: 24
L3 Aris!ii!alis, ar-Tabica, 183,8-11 =Aristotle, Phys. 3.2, 201b32-33: 26
L4 Aris!ii!iilis, Mii bacda t-fabi<a II, 621,9-622,13 =Aristotle, Metaph.
5.16, 102lbl2-1022a2: 33
LS Aris!il!alis, af-Tab'ica, 887,6-12 =Aristotle, Phys. 8.7, 261a32-37: 37
L6 Aris!ii!alis, af-Tabfca, 760,8-761,6 [the Arabic translation is out of
sequence] = Aristotle, Phys. 7.3, 246al0-16: 37
L7 (Ps.-)Aris!ii!alis, Uthalujiyii 3, in Plotinus apud Arabes, ed. A. Badawi,
Cairo, 1955, 54,7-55,19 corr. Plotinus, Enneads IV.7.8(5),1-18: 62
L8 Aris!ii!alis, Fi n-nafs, 38,3-5 = Aristotle, DA 2.4, 415b8-l 3: 88
L9 Aris!ii!alis, af-Tabi<a, 137,15-138,3 =Aristotle, Phys. 2.7, 198a22-
28: 88
LIO ArisJa!alis, Mii bacda [-!abica II, 1073,15-1074,2 =Aristotle, Metaph.
8.4, 1044a32-bl: 89
L 11 Aris!ii!alis, Fi n-nafs, 20, 10 = Aristotle, DA l.4, 408b 18-19: 94
L 12 Aris!ii!alis, Fi n-nafs, 31, 13-15 = Aristotle, DA 2.1, 4 l 3a3-9: 95
Ll3 Aris!ii!alis, Fi n-nafs, 72,19-21 =Aristotle, DA 3.4, 429a22-25: 95
L 14 Aris!ii!alis, Fi a<(ia' al-J:iayawiin, 10,24-11,8 = Aristotle, PA l.1,
641a32-641b10: 95
L 15 Aris!ii!alis, Fi kawn al-J:iayawiin, 63,22-23 = Aristotle, GA 2.3,
736b28-30: 9 5
L16 Aris!a!alis, Ma ba<da f-fabica II, 706,6-7 =Aristotle, Metaph. 6.1,
1026a5-6: 96
L 17 Aris!a!alis, Ma ba<da !-!abica III, 1486,13-1487,2 = Aristotle, Metaph.
12.3, 1070a24-26: 96
L18 Ibn Sina, Kitiib ash-shifii'/[abi<iyyiit (6): Fi n-nafs I.I, 12,6-8 =
Avicenna, Uber de anima seu sextus de naturalibus I-Ill, 29,61-63:
114
L 19 lbn Sina, Kitiib an-najiit, 258,2-10: 115
L20 lbn Sina, Kitiib ash-shifii'l[abi'iyyiit (1): as-Samii' af-fablci 2.1, ed. S.
Zayid, Cairo, 1983, 83,5: 116
L21 lbn Sina, at-Ta'lfqiit calii J:iawiishi Kitiib an-nafs li-Arisfu (ad DA 2.4,
415b2), 91,3-16: 117
L 22 Ibn Sina, Kitiib ash-shifii'l[abi'iyyiit (6): Fi n-nafs I.I, 6, 13-7,2 and
7,8-10 = Avicenna, Uber de anima seu sextus de naturalibus I-Ill,
19,27-20,37 and 20,44-21,3: 118
L23 Ibn Sina, MabJ:iath cani l-quwii n-nafsiiniyya 2, 153,11-13: 120
L24 Ibn Sina, Risiilafi n-nafs wa-baqii'ihii wa-maciidihii 1, 55,12-56,1: 120
L25 lbn Sina, Kitiib ash-shifii'l[abl'iyyiit (6): Fi n-nafs I.I, 11,7-11 =
Avicenna, liber de anima seu sextus de naturalibus I-Ill, 27,40-
28,45: 121
L26 Ibn Sina, Kitiib an-najat, 161,6-14: 122
L27 Ibn Sina, at-Ta cliqiit, 21,16-19 and 21,22-22,2: 122
L28 lbn Sina, Kitiib ash-shifii'/Manfiq (5): al-Burhiin 4.4, 294, 11 -14: 128
294 Index of Lemmata
L29 Ibn Sina, Kitab ash-shifa'/llahiyyat (2) 6.5, 289,16-290,5 =
Avicenna, liber de philosophia prima sive scientia divina V-X,
334,30-40: 128
L30 lbn Sina, Kitab ash-shifii'lllahiyyat (2) 6.l, 258,1-8 =Avicenna, Liber
de philosophia prima sive scientia divina V-X, 292,27-36: 129
L31 lbn Sina, al-l:fikmat al-'Ariir;iiyya, 4vl6-5r5: 130
L32 Ibn Sina, Kitiib ash-shifii'llliihiyyiit (2) 6.5, 294,6-9; 294,12-295,12;
and 296,3-5 = Avicenna, Liber de philosophia prima sive scientia
divina V-X, 339,33-37; 339,41-341,71; and 341,84-86: 131
l..33 lbn Sina, at-Ta'lfqiit 'alii f.iawiishi Kitiib an-nafs li-Arisru (ad DA 2.4,
415b2), 94,18-23: 133
L34 Alexander, in Top. 4.1(ad121al0), 301,19-25: 154
L 35 lbn Sina, Kitiib ash-shifii'/lliihiyyiit (1) 1.5, 33, 16-18 = Avicenna,
liber de philosophia prima sive scientia divina I-JV, 38,20-23: 155
L36 Ibn Sina, Kitiib ash-shifii'llliihiyyiit (2) 6.5, 292,1-294,5 = Avicenna,
Liber de philosophia prima sive scientia divina V-X, 335,84-
339 ,32: 161
L37 lbn Sina, K. an-Najat, 345, l-l l: 163
L 38 lbn Sina, Jliihiyyiit-i Diinishniima-yi 'Alii'i, 54,9-10 and 55,2: 168
L39 al-Ghazali, Maqa~id al-faliisifa, 190,10.13: 168
L40 Ibn Sina, Kitiib al-ishariit wa-t-tanbihiit, 139,14-20: 169
L41 Ibn Sina, Kitiib al-ishiiriit wa-t-tanbihiit, 140,6-9: 169
L42 ar-Rlizi, Kitab lubiib al-ishiirat, 80,4-5: 170
L43 a'·Tiisi, Sharf.i al-ishiirat, 193,31-194,6: 171
L44 Mulla Sadra, at-Ta'lfqat 'ala llahiyyiit ash-Shifii', 258,3: 171
L45 Mulla Sadra, al-Asfar al-arba'a JI, 347,4-5: 172
L46 lbn Sina, at-Ta'liqat, 128,17-19: 173
L 47 Aris~ii~alis, ar-Tabi'a, 103,8-13 = Aristotle, Phys. 2.3, 195a8-l 1: 174
L48 Ibn Sina, at-Ta'liqiit, 128,17-25: 175
L49 lbn Sina, al-Mubaf.iathiit, 116,15-117,2 (= Mubiif.iatha 5, #277): 176
L 50 lbn Sina, Kitiib ash-shifii'/Tabi'iyyat (I): as-Samii' af-!abi' i I.I I,
53,4-12 = Avicenna, Liber primus naturalium/Ttractatus primus de
causis et principiis naturalium, 95,2-19: 177
LSI Ibn Sina, Kitiib ash-shifii'/lliihiyyiit (2) 8.6, 355,6-356,5 =Avicenna,
Liber de philosophia prima sive scientia divina V-X, 412,55-
413, 78: 185
L52 Ibn Sina, at-Ta'liqiit, 18,4; 51,28; 62,3-5; 62,12-13; 80,25-26;
160,5-6; 178,25-26: 186
L53 lbn Sina, Kitiib ash-shifii'l/lahiyyiit (1) 4.3, 188,3-189,11 =
Avicenna, Liber de philosophia prima sive scientia divina I-IV,
215,13-217,47: 189
L 54 (Ps.-)Aris~ii~alis, Kitiib al-ir;iiif.i fl 1-khayr al-maf.ir;J li-Arisfii!iilis 21,
22,12-23,5 corr. Proclus, Inst. Theo!. 115, 100,28-102,12: 191
L 55 (Ps.-)Aris~ii,li.lis, Uthiiliijiyii JO, 134, 16-135, 12 corr. Plotinus,
Enneads V.2.1,4-8: 192
L56 Ibn Sina, at-Ta' liqiit, 62,11-24: 194
L 57 Ibn Sina, Autobiography, 30,5-6; 36,8-38,2: 199
L58 lbn Sina, al-l:fikmat al-'Ariir;Jiyya, 3vl6-4rl2: 200
L59 Aris~a~alis, Mii ba'da !-!abi'all, 515,7-517,7 =Aristotle, Metaph. 5.5,
1015a20-bl5: 201
 Index of lemmata 295
L60 Aristiitalis, ar-Tabf<a, 158,4-160,7 and 163,9-164,7 =Aristotle, Phys.
2.9, 199b33-15 and 200a31-b8: 205
L 61 Arista~alis, Ma ba<da !-!abf<a Ill, 1607,11-1608,5 = Aristotle, Metaph.
12.7, 1072b4-l l: 208
L62 Alexander, in Metaph. 5.5 (ad l015b9-ll), 361,24-36: 210
L63 Aristiitalis, Tablflflt thaniyya, 397,4-9 =Aristotle, An Post. l.30,
87b19-25: 212
L64 al-Farabi, Sharb Kitab Aris{uffllfs fi l-cibara (ad 18a33), 84,12-19: 220
L65 al-Fiiriibi, Sharb Kitflb Aris!u!fllis fi l- cibara (ad I 9a25-26), 95, 18-19:
220
L66 al-Farabi, Sharb Kitab Aris{u{fllfs fi l-cibara (ad 19a32-b4), 98,12-15:
221
L 67 al-Farabi, Sharb Kitab Aris{utfllfs fi Vibara (ad 19a32-b4), 99,1-8 and
99,18: 222
L68 al-Farabi, Sharb Kitab Aris{u{{ilisfi l-cibara (ad 19a32-b4), 99,22-25
and 100,1-4: 223
L69 Ibn Sina, al-lfikmat al-cAru!fiyya, 4rl4-17: 236
L 70 al-Ash<ari, Kitflb al-Luma<, 11,14-15: 237
L 11 al-Ash<ari, Risalat istibsfln al-khaw!f fi cum al-kalam, 92,12-16: 238
L 72 ar-Raghib al-I~fahani, a/-/<tiqadat, 56,9-57,l l: 242
L73 (= L58) Ibn Sina, al-lfikmat al-cAru!fiyya, 3v16-4rl2: 247
L74 Ibn Sina, al-Mabda' wa-1-macad, 2,4-17: 252
L 75 Ibn Sina, al-Mabda' wa-1-ma<ad 1, 3,2-15: 254
L76 Ibn Sina, Kitab ash-shifa'//lahiyyflt (!) 1.6, 37,7-11 =Avicenna, Liber
de philosophia prima sive scientia divina I-IV, 43,8-15: 256
L 77 Ibn Sina, Kitab ash-shifii'//lahiyyat (I) 1.6, 38, 1-5 = Avicenna, Liber
de philosophia prima sive scientia divina I-IV, 44,24-31: 256
L78 Ibn Sina, Kitab al-isharat wa-t-tanbfhat, 140,12-151,2: 258
 General Index

cAbbad b. Sulayman, 149 Ammonius, 15, 44, 61, 64, 67-68,

cAbd al-Jabbar, 149, 228-229, 240- 70, 72, 74, 79-81, 83-84, 97, 99,
241, 243, 279 108, 113, 126, 145, 188, 225,
Abed, S., 151, 284 261, 266, 279, 290
above perfection, 12-13, I 14, 185, anankaion (Arabic: mu<ffarr/<fariiri/
188-193, 197, 212, 246, 251 wlijib), see Necessary
Abrahamov, B., 238, 283-284 ananke (Arabic: i<ffirlir!<fariira!wujiib),
Abii }:ianifa, 148, 157, 279 see Necessity
Abii 1-Hudhayl, 228-229, 230 Anawati, G., 163, 284
Ackrill, J., 23, 215, 225, 284 arete (Arabic: fa<fila ), see Excellence
Active Intellect, I I, 97, 135-137, Aristotle, passim; see esp. 21-42 and
139, 158, 178, 182-183 197-217
activity, 26, 34, 40-41, 47, 49, 52- arithmos teleios (Arabic: <adad
57, 61, 64-66, 75-76, 79, 82, 97- tlimm); see Perfect number
98, 120, 125-126, 137-139, 194 Asclepius, 61, 64-65, 68, 70, 72-75,
actuality, 3, 21-23, 25, 27, 50-52, 84, 85-86, 114, I 90-191, 204,
61-66, 76, 79, 82, 85, 87, 96, 207, 211, 279
102, 106-109, 113, 126, 194, al-Ash 0 ari, 149, 228-233, 237-239,
250-251, 263, 265 279-284, 295
Adamson, P., 105, 284 Ashcarites, 9, 156-157, 168, 171-
<ad.ad tlimm, see Perfect number 172, 197, 227-228, 231, 233,
adunaton (Arabic: mumtanicl 235-236, 240-241, 286
mustal.zil), see Impossible ate/es (Arabic: nliqi~), see
Albert of Saxony, 167, 279 Incomplete; Imperfect
Alcibiades, 82-83, 85 autarkeia (Arabic: istighnli'}, see Self-
Alexander of Aphrodisias, 4, 6, 21, sufficiency
41, 43-52, 57, 61-63, 65, 68, 71- autarkes (Arabic: mustaghnin), see
72, 79, 83-87, 96, 102-103, 107, Self-sufficient
115-116, 126, 154, 175, 182, authupostaton (Arabic: muktafin), see
194, 202, 204, 207, 210-211, Self-constituting
261, 265, 279, 285, 287, 290, autotelic, see Endedness
294-295 auxesis, see Growth
Alexander of Hales, 167, 279 Averroes (Arabic: Ibn Rushd), 100-
Algazel (Arabic: al-Ghazali), 167, 101, 167-168, 240-241, 266, 280
279 Avicenna (Arabic: Ibn Sina), passim
Allard, M., 23 I, 284 Avicennian synthesis, 16, 141, 266
alloiosis (Arabic: istil.zlila); see
Alteration Back, A., 248, 284
allotelic, see Endedness Badawi, 0 A., 22, 91, 103-104, 275,
alteration, 24, 37, 46, 95, 107, I 16 279-283, 293
d' Alvemy, M.T., 164, 166, 284 al-Baghdadi, Abii I-Barakat, 176, 280
al- 0 Amiri, 239-241, 243, 245, 251- al-Baghdlidi, al-Kha!ib, 157, 280
252, 262, 279 Balme, D., 184, 284
Ammonian synthesis, 6-7, 15-16, al-Baqillani, 149, 233-239, 280
64, 99, 108-110, 113-114, 117, Barnes, J., 284-285, 287, 290
126-127, 140-141, 186, 193, 197, al-Bayhaqi, 228, 235, 237, 280
266 Biard, J., 167, 284
298 General Index
bi-dhatihi, see In itself completeness, 4, 6, 26-59 passim,
bi-ghayrihi, see Through another 61, 65, 76-77' 79-80, 83-85, 93,
Black, D., 156, 284 96, 98-99, 100-105, 107, 110,
Blair, G., 26, 284 112-113, 117, 119, 182, 188,
Blanchette, 0., 188, 284 193-194, 241, 265
Blumenthal, H., 96, 284 completion, 21-23, 32, 34, 36-41,
body, 3-4, 6-7, 9, 15, 22, 37, 57, 62- 43-46, 50-52, 65, 82, 82-86, 93,
64, 71, 82, 84-90, 92, 95-97' 102, 97, 103, 175, 265, 271
114-117, 120, 127, 130, 133-135, compositeness, 22-23, 48, 86, 89,
138, 140, 149, 174, 206, 265, 285 92, 117, 119, 126-127, 130, 136-
Bolton, R., 96, 284 138, 177, 181, 249-259, 263, 265
Boylan, M., 184, 284 compulsory, 201, 203
Bradie, M., 184, 284 conception, 6, 54, 88-90, 116, 126,
Bruni, G., 63, 284 139, 148, 153, 208, 211, 234, 284
Brunschwig, J., 154, 285, 289 concoction, 38-39, 43, 45, 65
Bukhara, 16, 157, 199, 239, 243, conjoint causes, 23-24, 30-32, 35-
266 36, 40-41, 48-49, 51, 55, 68, 130
Buridan, John, 167, 280, 284 contemplation, 80, 83, 93
Burnett, C., 164, 285 contingency, 14, 216, 219, 221, 247
Burrell, D., 160, 285 contingent, 111, 214-217, 219, 222,
224, 255, 260
capability, 4, 8, 22-23, 27, 29, 30- contradictories, 41, 45, 213-214,
31, 36, 39-41, 44, 49, 56, 110, 21~ 219, 227, 229, 234, 260
119, 120-121, 126, 137-138, 230 contraries, 3, 17, 24, 26-27, 29-30,
Carame, N., 167, 285 32, 34, 37-38, 61, 71, 81, 86,
Caston, V., vii, 154, 285 107-108, 203, 213-214, 234, 258
causality, 6, 10, 12, 27, 65-67, 70, Cooper, J., 184, 285
73, 75-77, 81-82, 91, 107, 109, Craemer-Ruegenberg, I., 155, 285
111, 130-131, 133, 136, 138-140,
146, 163-172, 175-180, 187, 194, Daiber, H., 99, 106, 283, 285
265, 286, 288 Damascius, 74, 280
change, 3-5, 22, 24-46, 49-53, 56- <;iarura, see Necessity
57, 61, 64, 69, 71, 86, 106, 109- <;iarurf, see Necessary
110, 115, 126-127, 133, 177, Davidson, H., 198, 204, 207, 248-
200, 202, 208-209, 230, 247' 249, 260, 262-263, 285
250, 253, 255, 265, 288 De Young, G.R., 100, 285
Charles, D., 184, 285 Demiurge, 69-75
Charlton, W., 23, 285 Des Chene, D., 25, 285
Chen, C., 26, 285 Dhanani, A., 150, 285
co-implication, 152, 158-160, 179 dhat, see Self
coming-to-be, 13, 22, 37-38, 45-49, differentia, see Specific difference
57' 68-73, 80, 86, 88-90, 95, 103, divine attributes, 14, 148, 225, 228-
122, 153, 162, 184, 187, 190, 236, 266
205, 207, 210, 248, 279, 293 Dooley, W., 73, 210, 285
compatibility, see Complementarity Druart, T.-A., vii, 140, 285
complementarity, 9-12, 67, 72-74, duality, 6, 12-13, 16, 73, 197, 261,
79, 81 , 182-187, 197, 258-259 265, 285
complete, 4-6, 26, 32-40, 43-45, 48, dunamis (Arabic: quwwa), see
51-59, 61, 64-65, 68, 70-71, 75- Potentiality
77, 79, 81-82, 99, 101, 103, 106, dunaton (Arabic: mumkin), see
130, 188, 191-192, 265, 269, 274 Contingent
 General Index 299
Duns Scotus, 167, 280 existent, 26, 28-29, 37-38, 46, 71-
72, 86, 108-109, 113, 123, 145-
efficient cause, 5-9, 11-13, 64-65, 163, 171, 174-179, 181-182, 194,
67-82, 88-90, 133-40, 145-146, 200, 209, 224, 230, 237, 242,
162-164, 167, 169, 170-171, 173- 247, 249, 252, 258, 265
174, 176, 178-188, 194, 197, extensional, see Identity
259-261, 265 extrinsic (see Transcendent), 68, 90,
eidos (Arabic: ~ura), see Form 129-130, 140, 210-211, 259
<to> einai (Arabic: wujad), see
Existence fai;Jila, see Excellence
Emilsson, E., 63, 285 al-Fiiriibi; 1-2, 7, 9, 14, 16-17, 64,
end, see Final cause; Limit 99-100, 106-112, 130, 145, 151 ,
endedness, 4-5, 35-36, 39-42, 44-45, 153, 157, 180, 188, 191, 201,
53-59, 61, 94, 112, 125, 137 217, 219, 220-225, 227, 260-262,
Endress, G., 2, 99, 103, 285 280, 295
energeia (Arabic: ji</), see Activity; fawqa t-tamam, see Above perfection
Actuality ji<t, see Activity; Actuality
entelekheia (Arabic: an{alakhiyiil final cause, 5-13, 16, 54-55, 61, 64,
an{aliishiyii/fi cl/tamiimlkamiil), see 66-67, 69-76, 79, 81, 83, 85-94,
Actuality; Perfection 96-98, 101, 106-108, 110, 113-
epistrophe, see Reversion 114, 122, 126-139, 145-146, 161-
essence, 1, 9-12, 16, 49, 56, 84, 88- 164, 167-171, 173-188, 193, 197,
90, 93, 102, 145-146, 150-151 , 209-210, 265
154-156, 158, 160-161, 163, 169- form, 22, 24, 44-48, 52, 56, 61, 84-
171, 173, 178-179, 181-183, 185, 85, 102, 117, 119, 146, 271-272
187, 197-198, 224, 227, 252, formal cause, 8, 67, 88, 90, 134,
257-259, 262-263, 265, 286, 290 136, 138, 162, 181
eternality, 14, 68-69, 73, 76, 80-81, Frank, R., 99, 104, 150, 238, 285,
92, 96-97, 107, 110-111, 133- 286
134, 148-149, 182, 187, 190-191, Franzen, W., 26, 286
195, 197, 200-204, 206-207, 209- Frede, M., 66, 286
212, 227-240, 242, 245, 247-248, Freeland, C., 66, 286
250-251, 261-262, 266 Friedman, R., 184, 286
<to> eu einai, see Well-being function, 4, 8, 23, 27, 36, 39, 45,
excellence, 33, 37-39, 43-46, 48-49, 56, 63, 80, 91-92, 110, 120-121,
52, 109 125- 126, 177, 205
exeiremenon (Arabic: mufaraq), see future contingents, 14, 221-223, 261
Separated; Separable
exercise, 4, 8, 30, 36, 39-40, 44, 49, Gaiser, K., 91, 286
58, 80, 94, 116, 119, 121, 126, Galen, 105
137, 178, 190 Gatje, H., 99, 286
existence, 5, 8-14, 16, 22, 61, 64- genesis (Arabic: kawn/takawwun), see
65, 68-75, 77, 79, 82, 108-109, Coming-to-be
111-112, 120, 122-123, 125, 128- genos (Arabic: jins), see Genus
130, 132, 134, 138- 140, 145-147, genus, 24, 28-29, 33, 44-46, 48-49,
152-1 56, 158- 163, 168-169, 171- 54, 71 , 75, 81, 97, 115, 152, 154,
176, 178-179, 181 -183, 185-191, 158, 160, 178, 185, 189, 206, 285
193, 195, 197-200, 202-203, 205- Georgulis, K., 26, 286
206, 210-211, 216, 219-31, 234, ghayriyya, see Otherness
236-243, 247-260, 262-263, 265, al-Ghaziili, 167-168, 279-280, 288,
285-287, 289 294
300 General Index
Giles of Rome, 167, 280 hypothetical necessity, 13, 73, 202,
Gill, M.L., 25, 286 206, 209
Gilson, E., 160, 187, 286
Gimaret, D., 150, 156, 228, 281, 286 Iamblichus, 190-191, 280
God, 3, 5, 8, 12-14, 16, 29, 61, 64, Ibn c Asiikir, 150, 233, 281
75-77, 90, 107-108, 111-112, lbn Biibawayh a~-~adi.iq, 281
114, 122-123, 145-149, 156, 173, lbn Fi.irak, 149, 157, 228, 233, 238-
183, 185-188, 190-191, 193-195, 239, 281
197-198, 201, 204, 207-212, 223- Ibn Khallikiin, 17, 281
225, 227-242, 246, 249-251, 254, Ibn Kulliib, 228, 230-233, 236, 290
257-263, 265, 285 lbn Mattawayh, 240-241, 281
Goichon, A.-M., 1-3, 146, 160, 170, lbn Miskawayh, 129, 239, 252, 281
197, 281, 286 lbn an-Nadim, 113, 154, 281
Goldschmidt, V., 154, 286 lbn Rushd, see Averroes
Good, see God lbn Sina, see Avicenna
Gotthelf, A., viii, 184, 284-286 identity, 4, 29, 31, 48, 73, 88, 92,
Graeser, A., 154, 286 153, 158, 175, 254
Graham, D., 25-26, 286 i~?irlir, see Necessity
greater sumphonia (i.e., the harmony ijlib, see Necessitation
of Plato and Aristotle), 15, 64, 94, Ikhwiin a~-~afii.>, vii, 107, 239, 282
126, 194 'illiyya, see Causality
growth, 35, 38, 41, 43, 46, 95, 175 immanent, 6-9, 68, 87, 90, 96, 119,
Gutas, D., vii, 2-3, 99, 114, 118, 129-130, 132-135, 140, 211, 220,
123-124, 141, 154, 157, 177, 240, 246-247, 259, 291
198-199, 252, 285-287 immortality, 55, 87, 92, 134-136,
139, 182
Hadot, P., 153-154, 287 imperfect, 11, 26, 37, 52, 56-57, 61,
Hager, F., 63, 287 65-66, 70-71, 79, 81, 83, 100,
Hamlyn, D., 23, 287 108, 130, 191, 250-251, 253,
Hankinson, R.J., 67, 184, 287 255, 263, 288-289
happiness, 54-55, 197 impossible, 13, 111, 128, 155-156,
Haschmi, M.Y., 116, 287 165, 187, 200, 213-216, 219,
Hasnawi, A., 103, 116, 284, 287 234, 237, 239, 242, 247-248,
Hasse, D., viii, 114, 121, 199, 287 255-256, 258, 260
Heinaman, R., 54-55, 287 in itself, 13-14, 16, 22, 53-54, 56,
hendiadys of t-m-m and k-m-l, 101, 71, 84, 128, 139, 148, 158-160,
103-105, 192 166, 178, 185, 189-191, 195,
Henry of Ghent, 167, 280 197, 199-200, 204, 2ll-212, 220-
hexis, see State of being 225, 227, 232-236, 239-242, 245,
Hocutt, M., 66, 287 247-259, 262-263
<to> hOi, see That for the benefit of inactive, 40, 80-81, 110
which; Final cause incomplete, 26-32, 34-37, 39-41,
Horten, M., 167, 287 50-53, 56-57, 61 65, 76, 80, 83,
<to> hou, see That in view of which; 100, 108, 191
Final cause instrumental cause, 67, 69
<to> hou heneka, see That for the intellect, 32, 46-48, 80, 86, 93-97,
sake of which; Final cause 112, 134-136, 140, 158, 162,
Hourani, G., 198, 287 173, 178, 189, 191-193
l:funayn b. Is~iiq, 99, 105, 280 intelligible, 46-49, 69, 74, 95, 236,
Hussey, E., 25, 287 289
Hyman, A., 198, 287 intensional, see Identity
 General Index 301
intentional object, 158, 163, 169, lesser sumphonia (i.e., the harmony
171, 176, 178 of Aristotle and Aristotle), 15, 51,
intrinsic, see Immanent 58, 64, 94, 126, 194
IsJ.iaq b. I:Iunayn, 22, 24, 99-101, Lettinck, P., 137, 288
103-106, 108, 114-115, 127, 134, Lewis, G., 62, 192, 288
207, 215-217, 219, 260, 282, 285 limit, 33-35, 44-45, 58-59, 132,
isti~ala, see Alteration 156, 194, 269, 274, 283
istighna', see Self-sufficiency location, 26, 37
istikmal, see Actuality locomotion, 37, 45-46, 52-53, 61,
95, 106, 120, 135-136, 208
Jabir b. I:Iayyan, 282 Lovejoy, A., 187, 288
Ja<d b. Dirham, 238
Jadaane, F., 154, 287 Madelung, W., 157, 231, 241, 288
Ja<far a~-Sadiq, 156 ma<dum, see Non-existent
Jalbert, G., 287 Mahdi, M., vii, 16, 288
Janssens, J., vii, 116, 124, 168, 287 mlihiyya, see Essence, Thingness
Jolivet, J., 2, 130, 145, 151, 153- Maier, A., 25, 116, 167, 288, 290
154, 161, 166, 282, 287, 289 ma<nan, see Intentional object
al-Juwayni, 156, 282 Marmura, M., vii, 2, 156, 158, 160,
187, 288
kallim (i.e., Islamic doctrinal al-Mas<iidI, 106, 282
theology), 1-2, 9, 12, 14, 16-17, material cause, 6-8, 68-69, 87, 140,
141, 145-151, 153-154, 156-158, 184, 187
161, 168, 173-174, 180-181, 197, matrix of distinctions (i.e.,
201, 207, 217, 225, 227-236, Avicenna's), 13-14, 57, 195, 197-
238-240, 246-251 , 261-262, 265- 198, 200, 217, 227, 240-241,
267, 291 243, 247, 251-253, 255-257, 260-
kamlil, see Actuality; Perfection 263, 265
klimil, see Perfect al-Maturidi, 150, 156-167, 235, 238,
kawn, see Coming-to-be 261, 282, 289
Keyt, D., 54, 287 Maturidites, 9, 14, 149-150, 153,
al-Khayyat, 149, 157, 282 157, 197, 227, 230, 261-262, 288
Khodeiri, M., 166, 288 mawjud, see Existent
Khurasan, 118, 157 McCullough, E.J., 25, 288
al-Khwansari, 187, 282 McDermott, M.J., 150, 288
al-KhwarazmI, 147, 282 Menn, S., vii, 26, 289
kinesis (Arabic: ~araka), see Change; metabole (Arabic: taghayyur), see
Motion Change
Klein-Franke, F., 150, 288 Michot, J., 124, 169, 177, 289
knowledge, 22-24, 30-32, 35-37, 39- Miller, F., 184, 284
40, 46, 63, 80, 93, 95-96, 100, mistranslation of shay'iyya into
110, 156, 175, 186, 199, 212, Latin as causalitas, 164, 166-167
217, 221-223, 228-231, 233, 237- modal qualifier, 221, 237, 246
239, 261, 266, 284, 290 Moravcsik, J., 66, 289
Kosman, L.A., 25, 32, 288 Morewedge, P., vii, 160, 288-289
Kostman, J., 25, 288 morphe (Arabic: shakl), see Shape;
Kraus, P., 114, 282-283, 288 Form
al-KulaynI, 156, 282 motion, 22, 26, 41, 52, 61, 64, 79-
Kutsch, W., 135, 280, 282, 288 80, 88-90, 100, 103-104, 106,
116, 126-128, 131-132, 134, 161,
Lennox, J., 184, 284, 285, 288 173-174, 176, 178, 183, 187,
302 General Index
190, 201, 206, 208-209, 286, non-existent, 148, 150, 153, 155-
288-289 156, 173, 176, 179, 200, 236-
mu{ifarr, see Necessary 239, 242, 247-249, 252-254, 256
mufiiraq, see Separable nooumenon (Arabic: ma<qul), see
mul_zdath, see Originated Intelligible
miljab, see Necessitated Nussbaum, M., 184, 289
miljib, see Necessitating
muktafin, see Self-constituting Olympiodorus, 76, 92, 113, 282
Mulla MahdI NariiqI, 155-156, 282 <to> on (Arabic: mawjild), see
Mulla ~adra, 171-172, 282, 294 Existent
mumkin, see Contingent One, see God
mumkin al-wujild, see Possible of One-Good, see God
existence originated, 7, 72, 192, 195, 197,
mumtani<, see Impossible 227, 229-230, 234-238, 241, 247
Mure, G., 66, 289 otherness, 236, 261
Miisa al-Ka~im, 156 ousia (see Existence), 49, 72-73, 88,
mustaJ.zil, see Impossible 102, 159, 182
mutakallimiln (i.e., Islamic doctrinal Owens, J., 25, 289
theologians), see Kaliim;
Ash<arites; Maturidites; paradigmatic cause, 67, 71, 75
Mu<tazilites; Shi<ites Parmenides, 9, 67, 98, 153
mutaliizim, see Co-implication Pasquino, P., 154, 289
Mu<tazilites, 9, 14, 148, 150-151 , Peck, A., 25, 289
154, 156-157, 160-161, 176, 223, peras (Arabic: nihiiya), see Limit
227-229, 231-232, 234, 238-241 perfect, 5-8, 11-12, 21, 26-27, 64-
67, 69-72, 74-77, 79, 81-82, 93,
nafs, see Soul 97-100, 107-108, 111-112, 130,
niiqi~. see Incomplete; Imperfect 138, 181-183, 185-186, 188-195,
an-Nashi, al-Akbar, 238, 282 197, 224-225, 245, 250-251, 255
Nasr, S.H., 160, 289 perfect number, 100, 190
necessary, 8, 13-14, 16, 65, 73, 89, perfection, 5-8, 11-13, 15, 21-24,
98, 101, 112, 120-122, 125, 128- 27, 30, 32, 37-39, 44, 48-49, 51,
130, 136-137, 145, 158, 160, 62-63, 66-72, 74-77, 79-83, 85-
162, 174, 178, 195, 197-263 86, 98-100, 103-106, 108-127,
passim, 287 130, 132-40, 157, 181-183, 185-
necessary of existence, 13-14, 16, 186, 188-197, 212, 245-246, 250-
195, 197, 200, 204-205, 207, 251, 253, 261, 263, 265, 271-
217, 219-225, 227, 237, 239-243, 274, 289-290
245-259, 262-263 Peripatetic, 6, 27, 44, 80, 83, 107,
necessitas consequentiae, 203, 217, 175, 265, 266
223, 245 Philoponus, 7, 50, 64, 68, 71-72,
necessitas consequentis, 203, 217, 74, 79, 83-88, 91-98, 102-104,
222, 246 113, 117, 126, 129, 133, 137,
necessity, 13, 73, 107, 122, 184, 225, 282-283
197, 200-240 passim, 245-248, Pines, S., 154, 289
252, 254, 261, 284-287 Plato, 9, 15, 64, 67, 73, 76, 80, 94,
Neoplatonists, 6-7, 15, 64, 69-70, 96, 108, 126, 140, 153, 184, 192,
80, 82-83, 140, 193, 265, 288 194, 261, 265-266
nihiiya, see Limit possibility, 14, 24, 36, 46, 94, 96,
Niqiimakhis al-Jarasini, 100, 189, 146, 153, 155, 157, 166, 177,
282
 General Index 303
214-216, 219-224, 236, 243, 254, Ross, D., 23, 25-26, 92, 202, 204,
256-258, 287 289
possible of existence, 26, 185, 197, Rowson, E., vii, 102, 239, 240-242,
200, 214-216; 220, 222-224, 227, 279, 289
241-242, 247, 250, 252-259, 262- Rudolph, U., 157, 238, 289
263
potentiality, 21-22, 26, 49, 107, sababiyya, see Causality;
108, 119, 130, 149, 289 Mistranslation of shay'iyya into
process, 22, 25, 27, 30, 34-36, 38- Latin as causalitas
41, 45, 50-51, 65, 85, 90, 93, 99, Sabra, A.I., vii, 17-18, 267, 290
164, 175 Sabzawari, 171, 283
procession, 5, 8, 10-13, 70-73, 75- as-Samarqandi, 150, 156-157, 283
76, 79, 81, 105, 122, 124, 136- Santas, G., 54, 290
138, 140, 145-146, 181-183, 185- Sauve Meyer, S., 184, 290
186, 193, 197, 265 Schofield, M., 66, 286, 290
Proclus, l, 9, 61, 64-87 passim, 92, scribal misplacement of diacritical
96-98, 102-103, 109-112, 114, marks, 164-167, 172
145, 183, 187-188, 190-194, 197, sea-battle, 221-222
283, 285, 290-291, 294 Sebti, M., 117, 290
productivity, 13, 29, 65, 66, 69, 76- self, 14, 71, 130, 146, 150, 186,
77, 79, 81, 103, 191, 193, 197, 192, 200, 205, 222, 225, 228-
211-212, 246, 261 231, 236-237, 239, 241-242, 247-
proodos, see Procession 249, 251, 254, 261-262, 266
psukhe (Arabic: nafs), see Soul self-constituting, 77, 190-191
self-sufficiency, 13, 76-77, 81, 191,
qadfm, see Eternal 193-194, 197, 212, 246, 287
al-Qasim ibn Ibrahim, 149, 283 self-sufficient, 13, 54-55, 76-77, 98,
qidam, see Etemality 112, 188, 190-191, 193, 245-246,
quality, 13, 24, 26, 34, 37-39, 43, 260-261
45-46, 49, 57, 71, 75-77, 80-81 , semen, 89-90
162, 188 separability, 15, 61, 64, 83, 87, 94,
quantity, 24, 26, 30, 37-39, 43, 49 96-97, 117, 129, 139-140
Qur'an, 123, 125, 146-147, 186, separable, 62, 90, 95, 97, 102, 117-
231, 233, 235-236 119, 126, 130, 139-140, 160,
al-Qushayri, 231, 283 265, 272
Qusia b. Liiqa, 99, 105-106, 108, 114 ash-Shahrastani, 156, 168-169, 283,
quwwa, see Potentiality 290
shape, 24, 84, 102-103, 121, 269
ar-Raghib al-I~fahani, 240-242, 252, shay' , see Thing
283, 295 shay' iyya, see Thingness;
Rahman, F., 140, 160, 166, 281, 289 Mistranslation of shay'iyya into
ar-Razi, Fakhr ad-din, 168-170, 175, Latin as causalitas
187, 283-284 ash-Shaykh al-Mufid, 150, 232, 283
ar-Razi, Mutiammad b. Zakariya, 105, Shehadi, F., vii, 151, 290
283 Shi 0 ites, 149, 156, 157, 171, 187,
reversion, 5-6, 8, 10-13, 48, 56, 70- 262, 267
72, 75-76, 79, 81, 85-86, 97-98, Sibawayhi, 147, 283
110-111, 122, 136-138, 140, 145- :;iflit, see Divine attributes
146, 181-183, 185-186, 193, 197, Siger of Brabant, 167, 283
265 simplicity, 5, 12, 69, 72, 76-77,
118, 139, 148-149, 157, 193,
304 General Index
202-204, 211-212, 245, 249, 253, teleiotes (Arabic: tamiim/kamiil), see
261 Completeness; Endedness;
Simplicius, 43, 50-52, 72, 74, 86, Perfection
88, 92, 98, 113, 126, 154, 283 teleology, 8, 67, 175, 179, 184,
Socrates, 82-86 187, 263, 284-286, 290
Sorabji, R., vii, 66, 184, 187, 284, telos (Arabic: tamiimlghiiya/nihiiya),
290 see Final cause; Limit
soul, 3-7, 11, 15, 22-25, 28-31, 37- that for the benefit of which, 91-94,
39, 41, 43-45, 57, 61-65, 71, 73, 96, 126, 133-139, 209
77, 79, 82-88, 90-97, 100, 102- that for the sake of which, 33, 35, 59,
106, 112-117, 119-121, 124-127, 88-89, 91, 133, 205-206, 269
129-130, 133-140, 145, 151, 158- that in view of which, 33, 35, 59, 88-
159, 161-163, 173, 176, 178-179, 89, 91-94, 96, 126, 133-139, 181-
190, 265, 272, 284-285, 289 182, 209, 269
species, 5, 11, 28-29, 44-46, 48-49, Themistius, 4, 41, 43, 45, 52-53, 56-
71, 75, 81, 88, 92, 97, 108, 117- 58, 61, 71, 73, 79, 83, 86, 91-92,
119, 121, 128, 134-135, 138, 94, 105, 109-110, 115, 126, 134,
151-152, 158, 160, 178, 183 159, 209, 27 l, 273, 283
specific difference, 28-29, 44, 48-49, thing, 5, 9, 10-ll, 14-15, 22-23, 25,
80, 97, 177, 183, 206; 27, 30, 33-35, 40-41, 44, 47, 49,
state of being, 21, 23, 25-26, 37, 39- 51-52, 54, 56-57' 62-63, 65-66,
40, 51, 58, 70, 79-80, 86, 93-94, 71 , 73-76, 79, 81, 84, 86, 88-91,
105, 110 94-95, 97, 108-111, 117, 119,
Steigerwald, D., 156, 290 121-122, 126-127, 128-38, 145-
Stem, S., 168, 290 163, 169, 171, 173-174, 176,
Suarez, Francisco 167, 283 178-182, 184-186, 189, 192, 199,
substance, 22-24, 26-27, 34, 37-39, 201-204, 206-211, 213-214, 216,
43, 49, 56, 62, 80, 85, 87, 91, 221-222, 224, 227, 229-230, 232-
101-102, 117, 121, 125, 128, 236, 242, 245, 250, 253-254,
130, 159, 175, 185, 190, 208, 256-258, 261, 265, 288
211, 225, 250 thingness, 9-10, 146, 155-82, 199
sunaitia, see Conjoint causes Thomas Aquinas, 167, 283-284
~ura, see Form through another, 13, 195, 197, 199-
Syrianus, 9, 5, 64-65, 67-68, 72, 74, 200, 221, 224, 227, 234-236,
77, 85, 96, 103, 113, 129, 190, 241, 247-250, 252-255, 261, 263
283 <to> ti (Arabic: shay'), see Thing
<to> ti en einai (Arabic: miihiyya),
at-TabarI, 148, 283 see Essence
takawwun, see Coming-to-be Todd, R., 44, 290
taliizum, see Co-implication transcendent, 13, 68-69, 76-77, 129-
tamiim, see Completeness; Final 130, 133, 136, 140, 193, 265, 291
cause; Perfection translations and translators, 13, 100-
tiimm, see Complete; Perfect 101, 104-105, 163-164, 188, 203,
ta~l)if, see Scribal misplacement of 208, 215, 217, 219
diacritical marks Transoxania, 157, 266
at-Tawl}idi, 129, 283 Tritton, A.S., 238, 290
teleion (Arabic: tiimmlkiimil), see at-Tiisi, 171, 187, 283-284, 294
Complete; Perfect
teleiosis (Arabic: tamiim/kamiil), see Unmoved Mover, see God
Completion Ustat, 100-101, 104, 127
 General Index 305
UthulUjiyii, 62-63, 76, 102, 104-105, well-being, 5, 8, 72-75, 79, 82, 120-
113, 118, 140, 192-193, 279, 122, 125-126, 135, 145, 181-183,
288, 293-294 185, 187, 202-203, 210-211
Whittaker, J., 77, 290
van Ess, J., 2, 150, 154, 231, 238, Wieland, W., 184, 290
282, 290 Wisnovsky, R., 69, 129-130, 145,
Verbeke, G., 63, 290 228, 241, 259, 287, 290-291
Verrycken, K., 70, 74, 290 Wolfson, H., 149, 154, 231, 291
wujub, see Necessity
wiijib, see Necessary wujud, see Existence
wiijib al-wujud, see Necessary of
existence az-ZabidI, 147, 284
Walzer, R., 99, 290 Zimmermann, F.W., vii, 102-103,
Waterlow, S., vii, 25, 290 114, 151, 220-223, 291
Weisheipl, J., 25, 288, 290